Butler P 36

The Dynamics of Language Use
Pragmatics & Beyond New Series
Editor
Andreas H. Jucker
University of Zurich, English Department
Plattenstrasse 47, CH-8032 Zurich, Switzerland
e-mail: ahjucker@es.unizh.ch
Associate Editors
Jacob L. Mey
University of Southern Denmark
Herman Parret
Belgian National Science Foundation, Universities of Louvain and Antwerp
Jef Verschueren
Belgian National Science Foundation, University of Antwerp
Editorial Board
Shoshana Blum-Kulka Catherine Kerbrat-Orecchioni
Hebrew University of Jerusalem University of Lyon 2
Jean Caron Claudia de Lemos
Université de Poitiers University of Campinas, Brazil
Robyn Carston Marina Sbisà
University College London University of Trieste
Bruce Fraser Emanuel Schegloff
Boston University University of California at Los Angeles
Thorstein Fretheim Deborah Schiffrin
University of Trondheim Georgetown University
John Heritage Paul O. Takahara
University of California at Los Angeles Sandra Thompson
Susan Herring University of California at Santa Barbara
University of Texas at Arlington Teun A. Van Dijk
Masako K. Hiraga Pompeu Fabra, Barcelona
St.Paul’s (Rikkyo) University Richard J. Watts
David Holdcroft University of Berne
University of Leeds
Sachiko Ide
Japan Women’s University
Volume 140
The Dynamics of Language Use: Functional and contrastive perspectives
Edited by Christopher S. Butler, María de los Ángeles Gómez-González
and Susana M. Doval-Suárez
The Dynamics of Language Use
Functional and contrastive perspectives
Edited by
Christopher S. Butler
University of Wales, Swansea
María de los Ángeles Gómez-González

Susana M. Doval-Suárez
University of Santiago de Compostela
John Benjamins Publishing Company

Amsterdam/Philadelphia
TM The paper used in this publication meets the minimum requirements
8
of American National Standard for Information Sciences – Permanence

of Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data
International Contrastive Linguistics Conference (3rd : 2003 : Santiago de

Compostela, Spain)
The Dynamics of Language Use : Functional and contrastive perspectives /
edited by Christopher S. Butler, María de los Ángeles Gómez-González and
Susana M. Doval-Suárez.
p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; v. 140)
Selection of papers presented at the 3rd International Contrastive
Linguistics Conference held in Santiago de Compostela, Spain, 23rd-26th Sept.
2003.
Includes bibliographical references and indexes.
1. Contrastive linguistics--Congresses. 2. Functionalism (Linguistics)
-Congresses. 3. Sociolinguistics--Congresses. 4. Psycholinguistics-
Congresses. I. Butler, Christopher, 1945- II. Gómez-González, María de los
A. III. Doval Suárez, Susana Ma. (Susana María) IV. Title. V. Series.
P134.I514 2003
410--dc22 2005050112
isbn 90 272 5383 8 (Hb; alk. paper)
© 2005 – John Benjamins B.V.
No part of this book may be reproduced in any form, by print, photoprint, microfilm, or
any other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
Table of contents
Foreword vii
Contributors ix
I. Introduction
Functional approaches to language 3

On contrastive linguistics: Trends, challenges and problems 19
María de los Ángeles Gómez-González and Susana M. Doval-Suárez
The present book 47
Christopher S. Butler, María de los Ángeles Gómez-González
II. Form and function in a cognitive perspective
The relation of grammar to thought 57

Wallace Chafe
Communicative constructions in English and Spanish 79
Montserrat Martínez Vázquez
III. Information structure
Incremental Functional Grammar and the language of football commentary 113

J. Lachlan Mackenzie
The role of Theme and Rheme in contrasting methods of organization in texts 129
Michael Cummings
On clefting in English and Spanish 155
María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Anaphoric terms and focus of attention in English and Spanish 197
Maite Taboada
 Table of contents
IV. Collocations and formulaic language
Formulaic language: An overview with particular reference

to the cross-linguistic perspective 221
A contrastive analysis of entrenchment and collocational force
in variable-sized lexical units 243
László I. Komlósi and Elisabeth Knipf
V. Language learning
Designing vocabulary tests for English, Spanish and other languages 271
Paul Meara
Timing in English and Spanish: An empirical study of the learning
of Spanish timing by Anglophone learners 287
Francisco Gutiérrez Díez
Spanish and English intonation patterns: A perceptual approach
to attitudinal meaning 307
Rafael Monroy
VI. Discourse and culture
Emotivity in narrative discourse: Cross-cultural and cross-gender perspectives 327

Kyoko Takashi and Douglas Wilkerson
Cardinal transitivity in foregrounded discourse: A contrastive study
in English and Spanish 349
Pilar Guerrero Medina
English consciousness in 19th century Spain 371
Paloma Tejada Caller
Language index 395

Scholars index 397
Subject index 403
Foreword
We would like to take this opportunity to thank all the people who have helped in
bringing this book about. First, and foremost, we would like to express our gratitude
to the contributors: without them there would have been no book. But we wish to
thank them especially for being so patient, while we struggled to complete this book.
The papers collected together in this volume represent a selection from among
those heard at the 3rd International Contrastive Linguistics Conference (ICLC-3), held
in Santiago de Compostela, Spain (23rd–26th September 2003). We are particularly in
debted to the ICLC-3 Organising Committee and the Department of English Philology
of the University of Santiago de Compostela for their assistance on that occasion: that
was the first contribution towards this volume. We would also like to express our sin
cere gratitude to the conference coordinator, Professor Dr. Luis Iglesias-Rábade, who
first conceived the idea of organising a conference that could serve as an open forum
for the presentation and discussion of investigations in the field of Contrastive Lin
guistics. Special thanks are also due to the members of the SCIMITAR research group
(http://www.usc.es/scimitar) for their invaluable collaboration in this endeavour.
Finally, for years of financial support, we would also like to thank the Spanish
Ministry of Education, the Galician Government and the European Funds for Regional
Development.
María de los Ángeles Gómez-González
Contributors
Christopher Butler took a first degree in biochemistry at the University of Oxford and
taught biochemistry and microbial physiology for some years at what is now Notting
ham Trent University, UK, before making a career move into linguistics. He worked at
the University of Nottingham from 1971 to 1992, first in the Department of English
Studies and later in the Department of Linguistics, where he was Head of Department
from 1986 to 1992. He also directed the university’s Language Centre from 1986 to
1990. In 1992 he took up the post of Head of English Language and Linguistics at what
was then the University College of Ripon and York St. John, a constituent college of the
University of Leeds. From 1994 to 1998 he acted as Director of Research at that institu
tion. He was awarded a Professorship of Linguistics in 1994. In 1998 he retired from his
position in York in order to devote more time to research and writing, and now lives
permanently in Spain. He is a member of the research group SCIMITAR (Santiago
Centred International Milieu for Interactional, Typological and Acquisitional Research).
In 2000 he was made Honorary Professor in the Centre for Applied Language Studies
at the University of Wales, Swansea.
Many of Professor Butler’s publications are concerned with theoretical and de
scriptive issues in functional grammars, with particular reference to English and Span
ish. The most recent is a 2-volume book, Structure and Function: A Guide to Three Ma
jor Structural-Functional Theories, comparing Functional Grammar, Systemic Func
tional Grammar and Role and Reference Grammar. A second area in which Professor
Butler has published widely is the use of computational and statistical techniques for
the study of English and Spanish, especially through the use of corpora. A number of
publications, particularly in recent years, are concerned with both of the above aspects,
attempting to show how insights from corpus linguistics can help us to formulate more
adequate functional grammars. E-mail: cbutler@telefonica.net
Wallace Chafe received his doctorate in linguistics at Yale University and was then em
ployed in the Smithsonian Institution in Washington as a specialist in Native American
languages, before moving in 1962 to the Berkeley campus of the University of Califor
nia, where he was department chair in the late 1960s and early 1970s. In 1986 he moved
to the Santa Barbara campus, where he is now professor emeritus. He has worked with
Native American languages of the Iroquoian and Caddoan families, and has been in
volved in attempts to understand language from functional and cognitive perspectives.
He has studied differences between speaking and writing, applications of linguistics to
literature, and most recently the functions of prosody. Among his many writings have
 Contributors
been the books Meaning and the Structure of Language (1970) and Discourse, Con
sciousness, and Time (1994). He is currently preparing a book on laughter. E-mail:
chafe@linguistics.ucsb.edu
Michael Cummings is Professor of English Language and Literature, Department of

English, Faculty of Arts, York University, Toronto, Canada, where he teaches traditional
and functional grammar, linguistic stylistics, and at both undergraduate and graduate
level the history of the English language. His research interests include Systemic-
Functional Grammar, discourse, and Old English. He is co-author of The Language
of Literature: A Stylistic Introduction to the Study of Literature (Pergamon, 1983), and
co-editor of Linguistics in a Systemic Perspective (Benjamins, 1988) and Relations and
Functions within and around Language (Continuum, 2002). Other publications of par
ticular interest to the issue of Theme and Rheme include “A Systemic-Functional
approach to the thematic structure of the Old English clause”, in On Subject and Theme:
A Discourse Functional Perspective, Ruqaiya Hasan and Peter Fries (Eds.), Benjamins,
1995; and “Structural semantics as the basis for Theme/Rheme”, in The 21st LACUS
Forum 1994, Mava Jo Powell (Ed.), Linguistic Association of Canada and the United
States, 1995. E-mail: mcummings@gl.yorku.ca
Susana M. Doval-Suárez is Lecturer in English Language and Linguistics in the De

partment of English, University of Santiago de Compostela (Spain), where she received
her PhD (awarded the 2003 PhD Extraordinary Prize of the Faculty of Philology).
She is a member of the research group SCIMITAR and has participated in a num
ber of research projects related with English historical linguistics, Error Analysis and
the grammar-discourse interface, and financed with European, Spanish and Galician
funds (XUGA-20401B95, PB96-O955, PFF2000-0492, pgidt00pxi20402PR, BFF2002
02441, PGIDIT03PXIC20403PN). Her research interests include English orthography
(both from a diachronic and from a synchronic perspective), Psycholinguistics and
Second Language Acquisition, with particular reference to the acquisition of a second
orthographic system. In this connection, she has described the orthographic interlan
guage of L2-English learners from the encoding perspective following the theoretical
framework set by Luelsdorff (1986). She has recently published The Acquisition of L2
English Spelling (LINCOM Studies in Language Acquisition 10. Lincom Europa, 2004).
She is co-author of Análisis de los errores del examen de inglés en las pruebas de ac
ceso a la universidad en el distrito universitario de Galicia (Universidad de Santiago de
Compostela: Instituto de Ciencias de la Educación, 1999) and co-editor of Studies in
Contrastive Linguistics (Universidad de Santiago de Compostela, 2002). She has actively
participated as Secretary and Coordinator in the organisation of the last two editions
of the International Contrastive Linguistics Conference (ICLC-2 and ICLC-3) held in
Santiago de Compostela in October 2001 and September 2003 respectively. E-mail:
iasdoval@usc.es
Contributors 
María de los Ángeles Gómez-González received her European doctorate in Linguis

tics (awarded the 1996 PhD Extraordinary prize) at the University of Santiago de
Compostela (USC), where she is Tenured Lecturer in English Grammar and PhD
Programme Coordinator of the Department of English. She is also the principal in
vestigator of the research group SCIMITAR. Dr Gómez-González has directed dif
ferent research projects exploring the grammar-discourse interface financed with
European, Spanish and Galician funds (BFF2002-02441, PGIDIT03PXIC20403PN,
PGIDT00PXI2040PR, PB97-0507, PB94-619, PB90-0370).
Her research interests include the interfaces between grammar-discourse and lan
guage production-linguistic structure. She has analysed spoken and written language
with the intention of spelling out the functional and cognitive principles that lie behind
the production and perception of different types of natural discourse, mainly in En
glish, although comparisons with other languages are also drawn (mostly Spanish and
Galician). She has worked in the fields of discourse analysis and functional linguistics,
paying particular attention to the areas of grammar, information structure, coherence
and cohesion. Her latest books are The Theme-Topic Interface: Evidence from English
(John Benjamins, 2001) and A New Architecture for Functional Grammar (Mouton de
Gruyter, 2004 (co-edited with Professor Lachlan Mackenzie)). She is a member of the
editorial boards of the journals Annual Review of Cognitive Linguistics and Iconicity in
Language. E-mail: iadimly@usc.es
Francisco Gonzálvez-García is Lecturer in English Language and Linguistics in the

Department of English and German Philology, University of Almería, Spain. His re
search interests center on the areas of syntax, semantics, pragmatics and discourse
analysis, as well as on functionalist and constructivist approaches. His early work, de
riving from his PhD received at the University of Bologna (Royal Spanish College)
“The syntax-semantics interface in complex-transitive complementation in contem
porary English”, was geared towards an eclectic approach to the syntactic analysis and
semantico-pragmatic import of verbless complement clauses in English. This synthe
sising analysis is outlined in “A modality view of predicate selection in small clauses”
(1997, Texas Linguistic Forum, 38, 101–119). A much more ‘delicate’ version of this
account, drawing on the Goldbergian version of Construction Grammar and applied
to both English and Spanish subcategorized verbless complement clauses, has been
recently published as “Re-constructing object complements in English and Spanish”
(in M. Martínez Vázquez (Ed.) (2003), Gramática de Construcciones: Contrastes en
tre el inglés y el español. Huelva: Grupo de Gramática Contrastiva, pp. 17–58). He has
also explored the implementations of a constructionist analysis of argument structure
for the morphology and grammar of the utopian literary genre in “Finding, seeing,
thinking, and observing in English Utopian literature: Towards an understanding of
NP + XP constructions in the morphology and grammar of J. Swift’s Gulliver’s Trav
els” (2000, Atlantis, 22(2), 69–91) as well as for translation in “Literatura, gramática e
iconicidad: Algunas notas a propósito de las traducciones de los Sonetos de Shakespeare
en español e italiano” (in P. Raccah and B. Saiz Noeda (Eds.) (2001), Lenguas, Lite
 Contributors
ratura y Traducción. Aproximaciones Teóricas. Madrid: Arrecife, pp. 205–225). E-mail:

fgonza@ual.es
Pilar Guerrero Medina is tenured lecturer in English Grammar at the University of

Córdoba. She has conducted her research within a functionalist framework, mainly
focusing on the areas of semantic function assignment and linguistic representation of
states of affairs in S. C. Dik’s theory of Functional Grammar (FG).
In her doctoral dissertation (La transitividad como categoría prototípica en la
Gramática Funcional: estudio contrastivo inglés-español-alemán, Micropublicaciones
Universidad de Córdoba, 1997) she attempts to integrate the notion of prototypical
transitivity within the FG framework. Other FG publications are: “A prototype ap
proach to transitivity: its implications for the FG typology of SoAs” (in Olbertz, H.
et al. (Eds.) The Structure of the Lexicon in Functional Grammar. Studies in Language
Companion Series. John Benjamins, 1998), “Reconsidering aspectuality: interrelations
between grammatical and lexical aspect” (Working Papers in Functional Grammar 75,
2001) and “Object assignment in S. C. Dik’s Functional Grammar: marginal accessibil
ity in English” (Odisea. Revista de Estudios Ingleses 4. Universidad de Almería, 2003).
Her most recent publications explore the issue of “lexical ergativity”, adopting a
constructional approach to the study of argument structure: “Lexically ergative con
structions in English and Spanish” (in Iglesias Rábade, L. and S. M. Doval (Eds.) Stud
ies in Contrastive Linguistics. Proceedings of the 2nd International Contrastive Linguistics
Conference. Universidad de Santiago de Compostela, 2002) and “A constructional ac
count of ‘lexical ergativity’ in English” (in Carretero González, M. et al. (Eds.) A Life in
Words. A Miscellany Celebrating Twenty-five Years of Association between the English
Department of Granada University and Mervyn Smale (1977–2002). Universidad de
Granada, 2002). E-mail: ff1gumep@uco.es
Francisco Gutiérrez Díez is Professor of English Philology at the Faculty of Arts, Uni
versity of Murcia. He obtained his PhD in English Philology from the University
of Barcelona. Before moving to Murcia, he taught English and the teaching of En
glish as a foreign language at a teacher training college (Autonomous University of
Barcelona). He currently teaches translation at the University of Murcia. His research
interests are in the areas of phonetics and language teaching/learning, in particular
in the application of comparative phonetics to language teaching/learning. His early
work concentrated on intonation analysis (La función demarcativa de la entonación en
inglés, castellano y catalán, Servicio de Publicaciones de la Universidad de Murcia, 1995;
Modelos entonativos del inglés, Granada: Editorial Universitaria, 2000). He is currently
engaged in prosodic research and has published extensively on contrastive linguistic
rhythm (English-Spanish) and the learning of linguistic rhythm. E-mail: fgut@um.es
Elisabeth Knipf is Professor of German linguistics and chair of the German Linguis
tics Department, Institute of German Studies, Eötvös Loránd University Budapest,
Hungary. She is a habilitated doctor in German linguistics (ELTE Budapest, 2001)
Contributors 
and holds a PhD degree (University of Szeged, 1983) and a Candidate of Science
degree (Hungarian Academy of Sciences, Budapest, 1993). Her research interests lie
in the areas of lexicology, morphology, language variation, text-linguistics and socio
linguistics. Her early work on text-linguistics is outlined in Textgrammatik. Eine Ein
führung (Budapest: Nemzeti Tankönyvkiadó, 1996, co-author P. Canisius), while her
more recent interest in morphology and language change is represented in Lexikolo
gie I–II. Ein Textbuch zur Lexikologie (Budapest: Holnap Kiadó, 1998, co-edited with
R. Hessky), Grundlagen der deutschen Wortbildung (Budapest: ELTE Institut für Ger
manistik, Budapester Chrestomathie 11, 2000), Regionale Standards. Sprachvariationen
in den deutschsprachigen Ländern (Budapest: Dialog-Campus Kiadó, 2001, co-edited
with N. Berend), Die Substantivbildung in der Mundart (Budapest: ELTE Institut für
Germanistik, Budapester Beiträge zur Germanistik 42, 2003) and Lexikalische Seman
tik, Phraseologie und Lexikographie. Abrgünde und Brücken (Frankfurt am Main: Peter
Lang, 2004, co-edited with R. Brdar-Szabó). E-mail: knipfe@freemail.hu
László Imre Komlósi is Professor of theoretical and English linguistics and chair of
the English Linguistics Department at the University of Pécs, Hungary. He is a habil
itated doctor in English linguistics (University of Debrecen, 1997) and holds a PhD
degree in general linguistics (University of Szeged, 1981) and a Candidate of Science
degree in theoretical linguistics (Hungarian Academy of Sciences, Budapest, 1989). In
his research he has contributed to the fields of formal semantics, lexicology and the
mental lexicon, Functional Grammar, inferential pragmatics and Cognitive Seman
tics. The subtle changes in his academic and research interests can best be traced in
his dissertations: his early work in Montague Grammar was an attempt to incorporate
indexicality and pragmatic features in formal interpretations of natural language as
witnessed in his unpublished doctoral dissertation A Modification of a Montague-Type
Formal Semiotic System: Towards a Non-Realist Semantics and a Constructive Model-
Theoretic Linguistics, followed by research in the functional approach to semantics
and pragmatics reflected in his CSc. dissertation Semantic and Pragmatic Determi
nacy of Predication: The Functional Perspective on Linking, then research in inferential
pragmatics and cognitive linguistics as in his habilitationsschrift Inferential Pragmat
ics and Cognitive Structures: Situated Language Use and Cognitive Linguistics (Budapest:
Nemzeti Tankönyvkiadó, 1997) and the recent publication of Communication and Cul
ture: Argumentative, Cognitive and Linguistic Perspectives (Amsterdam: University of
Amsterdam Press – SIC SAT, 2003, co-edited with P. Houtlosser and M. Leezenberg),
in which a major article on cognitive semantics by the author appeared under the title
In Quest of Cultural and Conceptual Universals for Situated Discursive Practice. E-mail:
komlosi@btk.pte.hu
J. Lachlan Mackenzie is an Academic Consultant in Languages and Linguistics, hav

ing been for many years Professor of English Language and Director of Research at
the Faculty of Letters, Vrije Universiteit Amsterdam (Netherlands); he is a member
of SCIMITAR. His interests are in the area of the relationship between discourse and
 Contributors
grammar, in particular the incorporation of aspects of discourse analysis into Func

tional Grammar. He is also interested in exploring the implications of recent findings
in studies of language production for models of grammar. He is currently working,
together with Kees Hengeveld, on a book-length presentation of recent developments
in Functional Grammar, Functional Discourse Grammar. With Matthew Anstey, he is
editor of Crucial Readings in Functional Grammar, an anthology of articles in Func
tional Grammar, and with María Gómez-González, he is editor of A New Architecture
for Functional Grammar. He is one of the editors of the journal Functions of Language.
E-mail: lachlan_mackenzie@hotmail.com
Montserrat Martínez-Vázquez (PhD, University of Seville, 1990) is Professor of En

glish Language and Linguistics at the “Pablo de Olavide” University, Seville. Her main
areas of research are English Grammar and Contrastive Linguistics. She is coordinator
of a research group in Contrastive Linguistics at the University of Huelva (“Gramática
contrastiva Inglés-Español”) and of the projects “Sintaxis Contrastiva Inglés-Español”
(BFF2000-1271) and “Metáfora y Metonimia en el Metalenguaje” (BFF2003-04064)
funded by the Spanish Ministry of Science and Technology. She is editor of a series
of Working Papers in Linguistics published at the University of Huelva (Gramática
Contrastiva Inglés-Español, Gramática y Pragmática, Transitivity Revisited, Recent Ap
proaches to English Grammar and Gramática de Construcciones) and author of Sintaxis
inglesa: la atribución (Cáceres, 1991) and Diátesis: alternancias oracionales en la lengua
inglesa (Huelva, 1998) (winner of the AEDEAN 1998 award for Studies in Linguistics).
She has collaborated in the Diccionario Sintáctico del Verbo Inglés (Barcelona, 1996)
and published articles on English Grammar and English-Spanish Contrastive Linguis
tics. Her most recent publications concentrate on marked constructions as vehicles for
changes in the grammatical system (“Effected Objects in English and Spanish” Lan
guages in Contrast, 1, 2; “Delimited Events in English and Spanish” Estudios Ingleses de
la Universidad Complutense 9). E-mail: montse@uhu.es
Paul Meara is Professor in the Centre for Applied Language Studies at the University of
Wales Swansea, where he has worked since 1990. He is currently chair of the Institute
of Linguists Examinations Review Board. He runs a research group which specialises
in experimental work on vocabulary acquisition in a second language, and has pub
lished widely in this area. His most influential work, apart from the vocabulary tests
described in this paper, is his massive bibliographical archive, which can be accessed at
http://www.swan.ac.uk/cals/calsres/varga/
His most recent work has been a ground-breaking study of the structural prop
erties of lexicons, which makes extensive use of simulations based on random au
tonomous Boolean network models. Publication of this work is expected shortly.
E-mail: p.m.meara@swansea.ac.uk
Rafael Monroy is Professor of English and Linguistics at the Department of English

Philology of the University of Murcia where he is currently teaching English Phonet
Contributors 
ics and Stylistics. His areas of specialization include English phonetics and phonology,
adult foreign language learning and spoken discourse analysis. He is the author of As
pectos fonéticos de las vocales españolas, La pronunciación del inglés RP para hablantes de
español, Sistemas de transcripción fonética del inglés and several textbooks on English
pronunciation for Spanish speakers. He has also co-authored three Spanish dictionar
ies. Current research projects include Spanish intonation, interlanguage phonology
(English-Spanish) and English academic writing from a contrastive rhetoric perspec
tive. E-mail: monroy@um.es
Maite Taboada is Assistant Professor in the Department of Linguistics at Simon Fraser

University, in Canada. Maite works in the areas of discourse analysis, systemic func
tional linguistics and computational linguistics, concentrating on coherence, cohesion,
and information structure. Her book, Building Coherence and Cohesion: Task-Oriented
Dialogue in English and Spanish, was published by John Benjamins (Amsterdam) in
2004. The study is a cross-linguistic characterization of coherence and cohesion in
spoken language, where she examines the strategies that speakers use to build a con
versation. She has also published articles on rhetorical relations, Theme, and cohesion.
Maite has participated in research projects in natural language generation, soft
ware agents and machine translation. She was a post-doctoral research fellow at the
University of Alberta, and has also worked in industry, leading the development of a
natural language processing system. Some of her current research projects are: cohe
sion and referring expressions in Spanish and English, combining cohesion and Cen
tering Theory; discourse markers in relation to turn-taking in conversation; Rhetorical
Structure Theory; evaluation and appraisal in text (and how to implement a system to
perform automatic appraisal). E-mail: mtaboada@sfu.ca
Paloma Tejada Caller is Tenured Lecturer of English Historical Linguistics and Head
of the Department of English I, Universidad Complutense, Madrid. Her research in
terests concentrate on the wide area of historical linguistics with special reference to
English, and on that of typology and contrastive linguistics. Her theoretical interests
lie mainly around dynamic (re-)interpretations of linguistic change; intermediate tem
porary stages of language development and paths of change, historical ideology and
cross-linguistic intertwining of grammatical and lexical categories. She has published
several books: El cambio lingüístico. Claves para interpretar la historia de la lengua
inglesa (1999), Madrid: Alianza ISBN: 84-206-8657-3; Historia de la lengua inglesa
para españoles. Materiales de trabajo (2001), Madrid: Alianza. ISBN: 84-206-5775-1;
Guzmán Guerra, A. y Tejada Caller, P. ¿Cómo estudiar filología? (2000). Madrid: Alianza
ISBN: 84-206-5752-2. She is now finishing her 3-year project on the Spanish image of
English and Englishness. E-mail: ptejadac@filol.ucm.es
Kyoko Takashi (Wilkerson) received her PhD in Linguistics at Georgetown Univer

sity in 1990. She is currently Associate Professor at Nagoya University of Foreign
Studies, teaching courses in sociolinguistics and comparative linguistics. Her primary
 Contributors
research interests include studying how “power” and “solidarity” are expressed in vari
ous types of discourse, such as narratives written by men and women, child caregivers’
reports, and advertising. In doing so, she has examined reference terms for husbands
in Japanese, code-switching in advertising, and style shifts and schema shifts in narra
tives, and published the results in “A Sociolinguistic Analysis of English Borrowings in
Japanese Advertising Texts” (1990); “Language and Desired Identity: The case of con
temporary Japan” (1992); “A Study of Speaker’s Subjectivity in Causal Expressions”
(1997); “Reference Terms for Husbands in Japanese” (1997); “Rethinking Power and
Solidarity in Japanese Discourse” (1999); and “When Power and Solidarity Collide:
New Politeness Strategies in Japanese” (2000). E-mail: dktwil@nufs.ac.jp
Douglas Wilkerson is Associate Professor in the Department of Japanese Studies at

Nagoya University of Foreign Studies. He received his doctorate in East Asian Lan
guages and Literatures from Yale University. His research interests include cultural
and sociolinguistic aspects of Japanese discourse and language use. He has written
about the spiritual dimensions of social activities in contemporary Japan, and how this
dimension influences and is incorporated into both formal (ritual) and informal dis
course; how cultural identity, especially domestic identity, is created and manipulated
through and in Japanese travel advertising; and the cultural limitations of metaphor
ical conception, ways in which metaphors must be altered for adaptation in different
cultural contexts, and the constraining parameters for such adaptation. E-mail: wilk
erdk@nufs.ac.jp
P I
Introduction
Functional approaches to language*
. Introduction: Functionalism within the spectrum of linguistic theories
The term functional, as used to label linguistic theories, is often seen as opposed to
formal.1 Like many dichotomies in linguistics, this one is broadly valid, but somewhat
misleading in its over-simplification of what are actually very complex relationships.
A more accurate and realistic way of thinking about linguistic theories is in terms of a
spectrum with quintessentially formal theories such as Chomskyan linguistics at one
end and radically functional ones at the other. Between the two extremes there exists a
rich variety of approaches. Even this clinal model is somewhat naïve in its adoption of
a purely unidimensional classification. In reality, what we have is a range of theories,
some of which may share only the very broadest of aims and assumptions, while others
lie much closer together in the multidimensional space defined by a fairly complex set
of features. In what follows I shall discuss a number of such features, some of which
would probably be agreed upon by most if not all who regard themselves as function
alists, and even by some who would not, while others are prominent in some theories
but may even be absent from others.2
A further complicating factor is the geographical distinction between European
and American functionalism. Among those linguistic theories which were initially
developed in Europe, the most influential are probably Dik’s Functional Grammar
(FG: Dik 1997a, 1997b) and Halliday’s Systemic Functional Grammar (SFG: for re
cent accounts see Halliday & Matthiessen 1999; Halliday 2004). On the other side
of the Atlantic we find two rather different types of approach. Firstly, we have the
Role and Reference Grammar of Van Valin (Van Valin & LaPolla 1997; Van Valin
2005), which tends towards the more formal end of the formal-functional spectrum.3
Secondly, at the more radically functional end of the spectrum we find a cluster of ap
proaches which frequently used to be lumped together under the label of West Coast
Functionalism,4 some of which have recently evolved into what are becoming known
as usage-based approaches to language (see e.g. Bybee & Scheibman 1999; Barlow &
Kemmer 2000; Bybee 2001; Hopper 2001; Thompson 2002), characterised by the claim
that the linguistic representations of the grammar are closely linked to, and ultimately
extracted from, concrete usage events (Kemmer & Barlow 2000: viii ff.). Often, in the
 Christopher S. Butler
writings of American functionalists, the terms functionalist and functionalism are used
primarily with reference to this last group of approaches, very little reference being
made to European models.5
. The main tenets of functionalism
The central, and most general, tenet of functionalism is that language is first and fore
most a means of human communication in sociocultural and psychological contexts,
and that this fact must determine our view of how language should be modelled. In
other words, there is a strong consensus among functionalists that the linguistic system
is not self-contained, and so autonomous from external factors, but is shaped by them.
Admittedly there are some, described by Croft (1995: 492) as external functionalists,
who maintain that the linguistic system (the grammar, in its widest sense) is indeed
autonomous, while challenging some of the tenets of Chomsky’s Universal Grammar.
The majority of functionalists, however, subscribe to what Croft dubs integrative func
tionalism, believing that although the system shows some degree of arbitrariness, it is
so closely bound up with external motivating factors that it is useless to attempt to
account for it without reference to these factors. In principle, it follows that a func
tional theory must take fully into account the essential connection between language
and (a) cognition and (b) the social and cultural context of language use. We might
also expect a functional theory to have something to say about language change, and
its interaction with these basic functional motivations (see e.g. Gvozdanovic 1997).
A further point of major agreement among functionalists has to do with rejection
of the formalist claim that syntax is self-contained, and thus autonomous, with respect
to semantics and discourse pragmatics. Croft (1995: 491) distinguishes four positions
on this issue which he regards as falling within his (rather wide) view of functionalism.
Those whom he describes as autonomous functionalists basically accept the formal
ist view that syntax is both arbitrary and self-contained, though they recognise that
some constraints on sentence structure may derive ultimately from discourse. In my
view, such a position lies outside the mainstream of functionalism as I understand
it. Croft’s second category, mixed formal/functionalism, regards syntax as arbitrary
but not self-contained, and mixes purely formal and functional categories and fea
tures, as for example in Head Driven Phrase Structure Grammar. Croft recognises this
tendency even in the more recent grammars of Chomsky: once more, I would not
regard such grammars as being fundamentally functional in nature. Croft’s typologi
cal functionalists also reject the self-containedness of syntax, and distinguish between
functionally-based universal properties of grammars, usually expressed in terms of im
plicational hierarchies, and arbitrary language-specific features. Finally, Croft’s extreme
functionalism rejects not only the self-containedness, but also the arbitrariness, of syn
tax. Pace Croft’s view that no linguist subscribes to this extreme position, we shall see
that there are in fact functionalists who are predisposed to the view that all formal
distinctions are motivated by underlying semantic considerations.
Functional approaches to language 
The core of the functionalist position, then, is that both the language system as
a whole and the formal structures it uses are in large part motivated by external fac
tors such as cognitive structures, processing constraints and social factors. There is, of
course, no guarantee that all these pressures will act in the same direction: the state
of a language at any one time is the result of competing motivations, which may lead
to a situation which appears to be purely arbitrary (for a summary of formalist and
functionalist positions on this issue see Butler 2003a: 14–25).
The characteristics reviewed above are, in my view, the ones which define a cen
tral core of functionalism. However, even these are emphasised to different degrees in
different functional theories: indeed, Van Valin (2000: 335–336) goes so far as to say
that many concepts and methodologies found in the functionalist literature “are more
distant from each other than they are from many formalist ideas”. It is therefore worth
looking in rather more detail at how each of the fundamental claims relates to each of
the functional theories or approaches mentioned earlier.
There is no doubt that FG, RRG, SFG, ‘West Coast’ Functionalism and the emerg
ing usage-based approaches all subscribe to the claim that the communicative func
tions of language are crucial in moulding the shapes which languages take. The differ
ences between theories lie rather in whether they prioritise sociocultural or cognitive
motivating factors.
SFG has from its very beginning placed great emphasis on relating language to its
social contexts of use, and has developed sub-theories of register and genre in order
to model these relationships. There is also evidence that sociocultural factors are being
taken seriously in some work within usage-based frameworks (see e.g. Verhagen 2000).
On the other hand, other theories such as FG and RRG have so far paid relatively
little detailed attention to the sociocultural aspects of language use: indeed, Van Valin
& LaPolla (1997: 3, 15) say that the study of the use of language in different social
situations is not at present a priority for RRG.
The position on language and cognition is more or less a mirror image of that
summarised above for sociocultural factors. FG and RRG make a strong commitment,
in principle, to the study of language in relation to cognition, although in RRG this
commitment is honoured more in the breach than in the observance, and it is only
recently that we have witnessed the development of FG models which take psychologi
cal adequacy seriously (see Nuyts’ 1992, 2001 Functional Procedural Grammar (FPG),
Hengeveld’s Functional Discourse Grammar (FDG) (Hengeveld 2004a, 2004b, 2005;
Hengeveld & Mackenzie forthcoming a, b), Mackenzie’s (2000, 2004) variant of FDG
which he calls Incremental Functional Grammar (IFG), Bakker and Siewierska’s (2004)
model of the speaker).
Cognitive matters are given a very high priority in the ‘West Coast’ functionalism
of Givón and others, and in the usage-based models which have emerged from work
of this kind.6 Furthermore, such work is tending to blur the distinction between func
tional linguistics and cognitive linguistics, so much so that scholars such as Tomasello
(2003: 2) now talk about their approach as “Cognitive-Functional (Usage-Based) Lin
guistics”, and Croft (1999: 88, 2001: 9) labels Langacker’s Cognitive Grammar as a
 Christopher S. Butler
functionalist theory. This collapsing of functionalism and cognitivism is perhaps natu

ral within the context of American functionalism of the ‘West Coast’/usage-based kind,
but it is less obviously appropriate if we widen our view to take into account other
functional models such as FG, RRG and SFG. The difference in viewpoint is high
lighted by the fact that Croft (2001: 3, 9) appears to include both FG and RRG within
what he considers to be formal grammars, despite the fact that both of these theories
have signed up very firmly to the basic principles of functionalism I have proposed.
Furthermore, in SFG psychological/cognitive matters have historically been given a
rather low priority, although recent work by Halliday & Matthiessen (1999) does deal
with cognition, but expresses the view that it is to be explained in terms of linguis
tic processes, rather than vice versa. It seems preferable to take the view, put forward
by Van Valin (1991a: 6), Horie & Comrie (2000) and Butler (2003a: 59), that func
tionalism and cognitivism are indeed very closely related in many of their aims and
assumptions, but that cognitive theories such as Cognitive Grammar and the various
flavours of Construction Grammar differ from functional theories enough, in their
backgrounds, foundational assumptions and emphases, for us still to be able to recog
nise a distinction between predominantly functionalist and predominantly cognitivist
parts of what we might call functional-cognitive space.7
All the theories I have mentioned also share the view that the syntax of languages
is not autonomous from their semantics and discourse pragmatics. Again, however,
there are differences of degree and emphasis. The RRG view is that “[s]yntax is not
radically arbitrary [. . . ] but rather is relatively motivated by semantic, pragmatic and
cognitive concerns” (Van Valin 1991b: 9, emphasis in original). On the other hand,
Halliday (1978: 44), while admitting in principle the possibility of free variation in the
grammatical system, feels that we should be suspicious of any instances of it we find,
since he claims that they are very often reflections of a semantic distinction which
has not yet been discovered. Those who wish to include cognitive approaches within
the scope of functionalism will also find diverging views: the Cognitive Grammar of
Langacker (1987, 1991), for instance, subscribes to a strong view of the motivation of
syntax, whereas the variety of Construction Grammar proposed by Fillmore and his
colleagues (see Fillmore et al. forthcoming) is more agnostic on the meaningfulness of
all constructions.
With regard to language change, Dik (1997a, 1997b) makes some brief remarks
on the importance of hierarchies in the prediction of possible diachronic changes,
and discusses some concrete examples of historical grammatical changes. Detailed
studies of grammaticalisation in the English modals have been made by Goossens,
initially within FG (see e.g. Goossens 1985/1987a, 1987b, 1996), but later within a
more explicitly cognitive framework. The SFG view of language as a constantly chang
ing, dynamic, open system (see e.g. Matthiessen 1995: 49) leads naturally to discussion
not only of phylogenesis (the evolution of language in the human species, and within
this, the development of individual languages), but also of ontogenesis (the develop
ment of language in the individual) and logogenesis (the unfolding of language in a
text). Linguists working with a usage-based model also recognise the intimate rela
Functional approaches to language 
tionship between the usage of language, synchronic variation and diachronic change
(see Kemmer & Barlow 2000: xviii–xx). RRG, on the other hand, has so far given rather
little attention to matters of diachrony.
. Some further characteristics of functional approaches
I now turn to a number of other facets of functionalism, which derive from the ba
sic principles enunciated above, and are again visible to varying degrees in different
functional theories. A useful starting point here is the set of nine claims which Givón
(1995: 9) believes to be among the “cherished premises” of functionalism, and which
he himself considers to be “all valid – up to a point and within well defined contexts”.
Four of these deal with matters I have already mentioned here, the remaining five being
as follows:
– meaning is context-dependent and non-atomic
– categories are less-than-discrete
– structure is malleable, not rigid
– grammars are emergent
– rules of grammar allow some leakage (Givón 1995: 9).
These points are all concerned with the flexibility of meaning and structure in relation
to the demands made by the context of communication. Their most vociferous advo
cates are those usage-based linguists, such as Hopper, Thompson, Bybee, Barlow and
Kemmer, who view grammar as emerging from the requirements of discourse, rather
than as a stable system:
. . . the concept of grammar as emergent suspends provision for fixed structure,
and sees all structure as in a continual process of becoming, as epiphenomenal,
and secondary to the central fact of discourse. (Hopper 1992: 366)
We can, however, hear clear echoes of this position in SFG, especially in the view that
meaning is immanent, that is constructed through language itself, rather than tran
scendent, seen as a reflection of what is outside language (see Halliday & Matthiessen
1999: 17, 416).
A further aspect of flexibility is reflected in Givón’s statement that “categories are
less-than-discrete”. This is, of course, a major claim of cognitive linguistics, but is no
less important for the theories I have labelled as functional. Givón’s own work, and
that of many other American functionalists, has certainly paid considerable attention
to this area. Surprisingly, there is little evidence of a concern with non-discreteness in
mainstream FG, though the concept of prototype is exploited in work of Keizer (1992)
on reference, and also in work by Goossens (1994) and Guerrero Medina (1998) on
transitivity, all within the framework of FG. In RRG, the concept of prototype emerges
in the use of hierarchies of argument roles to predict more and less typical choices
for the macroroles of Actor and Undergoer, and also in the area of syntactic relations,
 Christopher S. Butler
modelled in terms of the privileged syntactic argument of a construction. Systemic lin

guists also fully recognise the inherent indeterminacy of language and the categories
needed to describe it (for discussion see Halliday 1996: 16–18). One way of dealing with
this is through the recognition that the choices we make when we speak or write are
essentially probabilistic in nature. It has also been accepted (Halliday & Matthiessen
1999: 46) that the features in terms of which languages can be described are not dis
crete, Aristotelian categories, but rather are clinal in nature, and might be characterised
in terms of fuzzy set theory. Furthermore, systemic linguistics conceptualises lan
guage itself as a flexible, multi-dimensional semantic space (Halliday & Matthiessen
1999: 68ff.; Martin & Matthiessen 1991).
We can also see something of the same concern for flexibility and adaptability in
the work of Mackenzie, whose IFG model arose out of the realisation that holophrastic
utterances, far from being oddball phenomena at the very margins of the grammar, are
actually extremely important, both in adult language and in child language acquisition.
The idea of building up an utterance from a focal holophrase, which is central to IFG,
allows considerable flexibility of structuring. Hengeveld also accepts the importance
of non-sentential utterances and builds them into his accounts of FDG.
The non-atomic nature of much of the language we use is amply demonstrated by
a large amount of research, both corpus-based and psycholinguistic, on what has come
to be known as formulaic language (see Wray 2002), which is claimed to be stored and
processed holistically rather than generated or analysed by the grammar. Formulaic
multi-word sequences, which form a considerable proportion of normal speech, pose
challenges for most functional theories, though they are accommodated more easily
within an emergent grammar approach (see e.g. Thompson 2002; Bybee 2001; Bybee
& Scheibman 1999). However, Butler (2003c) shows that formulaic sequences can be
accommodated within FDG, IFG, Nuyts’ FPG and Bakker & Siewierska’s dynamic ex
pression model, and Tucker (1996, in press) demonstrates how they can be handled
within SFG.
All this also lends credence to the idea that the rules of grammar, or at least those
presented in most grammars, whether functional or not, are indeed inadequate to
deal with the complexity of language use as evidenced in attested textual material –
as Givón’s final point reminds us, the rules leak, in ways which we may or may not be
able to patch up. In recent years there has emerged, in the corpus linguistics commu
nity, a group of linguists who claim that we should put aside existing theories and try
to approach with as few preconceptions as possible what corpora of text tell us (the so-
called ‘corpus-driven’ approach to linguistics – see e.g. Sinclair 1992; Tognini-Bonelli
2001). This approach, however, throws the baby out with the bath water: what we have
discovered through functional linguistics is far too valuable for us just to discard it and
start all over again. We do, however, very much need to take on board what studies
of authentic language tell us. For those functionalists who adopt the emergent gram
mar approach, this does not present a problem, since they themselves challenge the
view that the grammar can be seen as a stable if very flexible system (see e.g. Hopper
2001; Thompson 2002). For those functionalists who adopt a more moderate approach
Functional approaches to language 
which still preserves the idea of a grammar in its own right, the challenge is a daunt
ing one. What moderate functionalists need to do is to approach the corpus data in as
open and honest a manner as possible, and to use these data to test not only individual
claims about particular parts of the grammar, but also the fundamental principles on
which the grammar is based. This important enterprise is still at a fairly early stage (for
full discussion see Butler 2004).
The mention of authentic textual materials above brings us to a further impor
tant point. If functionalists are genuinely committed to accounting for language as a
communicational tool, they must recognise that their grammars cannot limit them
selves to the sentence level, but must give an account of the structure and function
of whole discourses in their contexts of production and reception. There is, how
ever, wide variation in the extent to which functional theories have actually taken on
this important challenge. SFG has, from its inception, been a strongly text-oriented
theory, with its roots in Malinowski’s (1923, 1935) work on texts in their cultural envi
ronment and Firth’s insistence (in Palmer 1968: 199) that theoretical constructs must
maintain renewal of connection with textual data. Indeed, Halliday’s Introduction to
Functional Grammar (Halliday 1985a, 1994a, 2004), undoubtedly the most widely read
of Halliday’s works, was written specifically for the needs of text analysts, and Halliday
characterises the orientation of the underlying theory as follows:
In general, [. . . ] the approach leans towards the applied rather than the pure, the
rhetorical rather than the logical, the actual rather than the ideal, the functional
rather than the formal, the text rather than the sentence. The emphasis is on text
analysis as a mode of action, a theory of language as a means of getting things
done. (Halliday 1994a: xxvii)
The last sentence of this quotation is particularly worth noting: many systemicists are
committed to a view of linguistics, and in particular text analysis, as a way of inter
vening in social processes such as education. Elsewhere, Halliday writes that SFG has
developed “both in reflection and in action – as a resource both for understanding and
intervening in linguistic processes” (Halliday 1994b: 4505), and that for him “linguis
tics cannot be other than an ideologically committed form of social action” (Halliday
1985b: 5).8
There is also a strong tradition of textual work in American functionalism, seen for
example in the work of Givón and others on topic continuity and referent tracking in
discourse (e.g. Givón 1983). Givón (1995: Chapter 7) offers a useful discussion of the
advantages and disadvantages of studying the communicative use of language through
analysis of the distribution of grammatical features in texts.
Until quite recently, FG has not given priority to the study of discourse, although
two groups of practitioners have always had an interest in the textual matters. A group
of classical scholars working at the University of Amsterdam, amongst whom the
names of Bolkestein, Kroon and Risselada are particularly prominent, have made valu
able contributions to textual studies in Latin and Greek. A second group of scholars,
associated with the Free University of Amsterdam, and including Mackenzie, Hannay
 Christopher S. Butler
and Keizer, have used textual analysis in order to further the study of information
distribution from an FG perspective. More recently, a plethora of discourse-oriented
models have emerged, due to the realisation that this was a weak area in FG, addressed
only very programmatically in Dik (1997b). Particularly noteworthy are Hengeveld’s
FDG and its variant, Mackenzie’s IFG, already mentioned in relation to the study of
language in relation to cognition. These models are explicitly presented as discourse
grammars rather than (just) sentence grammars, and this is reflected in their top-down
nature, and in the postulation of an interpersonal level in the grammar at which the
speaker formulates a series of discourse acts, grouped into moves, and furnished with
operators for illocution and other interpersonal meanings.9
RRG, on the other hand, has not articulated any theory of discourse structure as
such, although it does have important and interesting things to say about the interac
tion of discourse pragmatic and syntactic phenomena. Particularly well developed is
that part of the theory which deals with information focus and its realisation across a
wide range of languages (see Van Valin & LaPolla 1997: 199–236, 417–430, 484–492).
The theory also deals with reference tracking devices in languages of different types,
under the headings of switch-function and switch-reference.
Much of what I have said about various functional theories demonstrates a con
cern for typological adequacy, the criterion according to which a theory must be
capable of accounting for phenomena across the whole range of types found in the
languages of the world. In this sense, much of functional linguistics is inherently con
trastive, linking up with the types of study reviewed in the article by Gómez-González
& Doval-Suárez in the present volume. FG, RRG and ‘West Coast’ functionalism could
all be classified as typological functionalist theories, and test out their claims on a wide
range of language types. Indeed, Van Valin (1995: 461) comments that RRG arose out
of attempts to answer the question of what linguistic theory might look like if it were
based on the analysis of, say, Lakhota, Tagalog and Dyirbal rather than, as in so many
cases, largely on English and other familiar Indo-European languages.
SFG, on the other hand, has not historically given priority to typological matters.
The crucial stages in the development of the theory were based very largely on the
analysis of English, though it should not be forgotten that Halliday himself has worked
on Chinese at various stages of his career. Four of the major texts on SFG (Halliday
1985a/1994a/2004; Martin 1992; Matthiessen 1995; Fawcett 2000) are overwhelmingly
centred on English, though with some comments on other languages, and Halliday &
Matthiessen (1999) devotes just one chapter to Chinese, again with occasional mention
of other languages. There is ample evidence that systemicists are getting much more
interested in typological matters (see especially Caffarel, Martin, & Matthiessen 2004);
nevertheless, it is likely that the main lines of SFG are by now too firmly set for this
work to have a major impact on the fundamentals of the theory. Furthermore, the
methodological position taken on typology is very different from that in theories such
as FG and RRG: whereas the latter study particular areas of the grammar across a
wide range of languages, systemicists prefer to describe particular languages in some
Functional approaches to language 
detail, and only then to draw conclusions regarding possible typological similarities
and differences.
Finally, I turn to the area of language acquisition. In general, functionalists have
taken a constructionist approach in relation to acquisition, as opposed to the adapta
tionist view of most formalists. In other words, functionalists generally believe that the
child constructs his or her own grammar, largely on the basis of the language s/he is
exposed to, rather than that the core of the grammar is innately given and then merely
adapted in certain ways to match the linguistic input. This view characterises the ap
proach to acquisition taken in all the functional theories I have dealt with here. Two
aspects of this work are worthy of particular note. Firstly, a good functional theory
should enable us to make predictions about, for instance, the order and ease/difficulty
of acquisition of particular areas of the grammar, and this has certainly proved to be
the case for RRG (see Butler 2003b: 402–413 and the references given there). Secondly,
the FG linguists Hengeveld & Pérez Quintero (2001), basing their work on suggestions
by Boland (1999), propose that we should recognise a criterion of acquisitional ade
quacy which, like that of typological adequacy, acts as a testbed for the postulates of
a linguistic theory. Hengeveld & Pérez Quintero regard pragmatic and psychological
adequacy as extralinguistic and explanatory, restricting possible theories of grammar,
while acquisitional and typological adequacy are intralinguistic and descriptive, con
cerned with evaluating a theory in terms of its ability to make correct descriptions of
synchronic and diachronic facts in single languages and across languages.
It is important to note that functionalists do not reject the concept of innate
ness itself, since there are clearly many aspects of cognition and learning which are
built into the architecture of the human brain and are thus genetically determined.
Where functionalists differ from Chomskyan formalists is in the great importance at
tached to the part played by learning and experience in acquisition. In this respect, it is
instructive to compare functionalist constructivist approaches with the position artic
ulated recently by Jackendoff (2002), who, while fully taking on board the motivation
of language by cognitive and sociocultural factors, and also the partial predictability of
syntactic structure from meaning, still argues for a nativist view of Universal Grammar
in terms of a set of attractor structures which, through inheritance hierarchies, guide
the generalisations made by the child from the linguistic evidence available.
. Conclusion
Functionalist approaches, which occupy an important place within the spectrum of

linguistic theorising today, are characterised first and foremost by the claim that lan
guage should be seen primarily as a means of human communication in sociocultural
and psychological contexts, and that this fact must determine our view of how lan
guage should be modelled. Most who would call themselves functionalists would agree
that the language system is not autonomous from, or self-contained with respect to,
external factors, that syntax is not autonomous or self-contained with respect to se
 Christopher S. Butler
mantics and pragmatics, and that these claims must find full reflection in language
models. There is nevertheless considerable variation in the degree to which social and
cognitive explanatory factors are prioritised in different functional approaches, such
as FG, RRG, SFG, ‘West Coast’ Functionalism and recent usage-based models.
We may also recognise a number of further facets of functionalism, ultimately
derivable from the main tenets, and again variable across the spectrum of functionalist
approaches. These features are concerned with the flexibility of meaning and structure
in relation to the demands made on language in its contextually-appropriate use.
Notes
* The author wishes to acknowledge financial support from the Research Project Discourse
Analysis in English: Aspects of cognition, typology and L2 acquisition, awarded to the SCIMITAR
research group (http://www.usc.es/scimitar) and sponsored by the Spanish Ministry of Edu
cation, the FEDER funds and the Xunta de Galicia (XUGA) (grant numbers BFF2002-02441,
PGIDIT03PXIC20403PN). The author would also like to thank Francisco Gonzálvez-García and
an anonymous reviewer for their valuable comments on an earlier draft of this introduction.
. For a collection of articles on functionalism in relation to formalism in linguistics, see
Darnell et al. (1999a, 1999b). Further discussion can be found in Newmeyer (1998) and Butler
(2003a: Chapter 1).
. The following discussion is based on the characteristics of functional theories proposed by
Givón (1995: 9) and by Butler (2003a: 29). For a more detailed treatment of the area than can be
given here, see Butler (2003a, 2003b).
. For web-based resources relating to SFG, FG and RRG see the resource list at the end of this
introduction.
. This label, which clearly arose because of the initial geographical location of some of its ma
jor practitioners, such as Givón, Thompson, Haiman and DuBois, is perhaps less appropriate
nowadays, in view of the fact that some scholars associated with this group of approaches do not
even work in the US, let alone on its West coast.
. Exceptions to this include the discussion of similarities and differences between RRG and FG
in Van Valin (1990), Van Valin & LaPolla (1997), also the inclusion of neo-Firthian linguistics in
the account of usage-based models by Kemmer & Barlow (2000).
. We might also note here the important work of Chafe (e.g. 1994), which again combines
a thoroughly functional approach with detailed attention to cognition, as is clear from his
contribution to the current volume.
. For a wide-ranging discussion of relationships among functional and cognitive theories
and an attempt to map functional-cognitive space in terms of a complex set of features, see
Gonzálvez-García & Butler (submitted).
. For examples of the engagement of SFG with the processes of social change, see the papers in
Young & Harrison (2004).
. For further detail of these and other models of discourse within the overall framework of FG,
see Butler (2003b: 306–331).
Functional approaches to language 
References
Bakker, Dik & Anna Siewierska (2004). “Towards a speaker model of Functional Grammar”.
In J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional
Grammar (pp. 325–364). Berlin and New York: Mouton de Gruyter.
Barlow, Michael & Suzanne Kemmer (Eds.). (2000). Usage-Based Models of Language. Stanford,
CA: CSLI.
Boland, Annerieke (1999). “Functional Grammar and First Language Acquisition”. MA thesis,
Department of General Linguistics, University of Amsterdam.
Butler, Christopher S. (2003a). Structure and Function: A Guide to Three Major Structural-
Functional Theories, Part 1: Approaches to the Simplex Clause. Amsterdam and Philadelphia:
John Benjamins.
Butler, Christopher S. (2003b). Structure and Function: A Guide to Three Major Structural-
Functional Theories, Part 2: From Clause to Discourse and Beyond. Amsterdam and
Philadelphia: John Benjamins.
Butler, Christopher S. (2003c). “Multi-word sequences and their relevance for recent models of
Functional Grammar”. Functions of Language, 10(2), 179–208.
Butler, Christopher S. (2004). “Corpus studies and functional linguistic theories”. Functions of
Language, 11(2), 147–186.
Bybee, Joan L. (2001). Phonology and Language Use. Cambridge: Cambridge University Press.
Bybee, Joan L. & Joanne Scheibman (1999). “The effect of usage on degrees of constituency: the
reduction of don’t in English”. Linguistics, 37(4), 575–596.
Caffarel, Alice, James R. Martin, & Christian M. I. M. Matthiessen (Eds.). (2004). Language
Typology: A Functional Perspective. Amsterdam and Philadelphia: John Benjamins.
Chafe, Wallace (1994). Discourse, Consciousness, and Time: The Flow and Displacement of
Conscious Experience in Speaking and Writing. Chicago: The University of Chicago Press.
Croft, William (1995). “Autonomy and functionalist linguistics”. Language, 71, 490–532.
Croft, William (1999). “What (some) functionalists can learn from (some) formalists”. In M.
Darnell et al. (Eds.), Functionalism and Formalism in Linguistics. Vol. I: General Papers (pp.
87–110). Amsterdam and Philadelphia: John Benjamins.
Croft, William (2001). Radical Construction Grammar: Syntactic Theory in Typological
Perspective. Oxford: Oxford University Press.
Darnell, Michael, Edith Moravcsik, Frederick Newmeyer, Michael Noonan, & Kathleen
Wheatley (Eds.). (1999a). Functionalism and Formalism in Linguistics, Vol. I: General Papers.
Amsterdam and Philadelphia: John Benjamins.
Darnell, Michael, Edith Moravcsik, Frederick Newmeyer, Michael Noonan, & Kathleen
Wheatley (Eds.). (1999b). Functionalism and Formalism in Linguistics, Vol. II: Case Studies.
Dik, Simon C. (1997a). The Theory of Functional Grammar, Part 1: The Structure of the Clause
(2nd ed.) [Functional Grammar Series 20]. Ed. K. Hengeveld. Berlin and New York:
Mouton de Gruyter.
Dik, Simon C. (1997b). The Theory of Functional Grammar, Part 2: Complex and Derived
Constructions [Functional Grammar Series 21]. Ed. K. Hengeveld. Berlin and New York:
Mouton de Gruyter.
Fawcett, Robin P. (2000). A Theory of Syntax for Systemic Functional Linguistics [Current Issues
in Linguistic Theory 206]. Amsterdam and Philadelphia: John Benjamins.
 Christopher S. Butler
Fillmore, Charles, Paul Kay, Laura Michaelis, & Ivan Sag (forthcoming). Construction Grammar.
Stanford, CS: CSLI.
Givón, Talmy (Ed.). (1983). Topic Continuity in Discourse: A Quantitative Cross-Language Study
[Typological Studies in Language, Vol. 3]. Amsterdam and Philadelphia: John Benjamins.
Givón, Talmy (1995). Functionalism and Grammar. Amsterdam and Philadelphia: John
Benjamins.
Gonzálvez-García, Francisco & Christopher S. Butler (submitted). “Mapping functional-
cognitive space”. Submitted to Annual Review of Cognitive Linguistics.
Goossens, Louis (1985/1987a). “The auxiliarisation of the English modals”. Working Papers in
Functional Grammar 7. Amsterdam: University of Amsterdam. Slightly revised version in
“The auxiliarisation of the English modals: a Functional Grammar view”. In M. Harris &
P. Ramat (Eds.), Historical Development of Auxiliaries (pp. 111–143). Berlin and New York:
Mouton de Gruyter.
Goossens, Louis (1987b). “Modal shifts and predication types”. In J. van der Auwera & L.
Goossens (Eds.), Ins and Outs of the Predication (pp. 21–37) [Functional Grammar Series
6]. Dordrecht: Foris.
Goossens, Louis (1994). “Transitivity and the treatment of (non)prototypicality in Functional
Grammar”. In E. Engberg-Pedersen, L. Falster Jakobsen, & L. Schack Rasmussen (Eds.),
Function and Expression in Functional Grammar (pp. 65–80) [Functional Grammar Series
16]. Berlin and New York: Mouton de Gruyter.
Goossens, Louis (1996). “English modals and functional models: a confrontation”. Antwerp
Papers in Linguistics 86. Antwerp: University of Antwerp.
Guerrero Medina, Pilar (1998). “A prototype approach to transitivity: its implications for the FG
typology of SoAs”. In H. Olbertz, K. Hengeveld, & J. Sánchez García (Eds.), The Structure
of the Lexicon in Functional Grammar (pp. 215–232). Amsterdam and Philadelphia: John
Benjamins.
Gvozdanovic, Jadranka (Ed.). (1997). Language Change and Functional Explanations [Trends in
Linguistics, Studies and Monographs 98]. Berlin and New York: Mouton de Gruyter.
Halliday, Michael A. K. (1978). Language as Social Semiotic: The Social Interpretation of Language
and Meaning. London: Edward Arnold.

Halliday, Michael A. K. (1985a). An Introduction to Functional Grammar. London: Edward
Arnold.
Halliday, Michael A. K. (1985b). “Systemic background”. In J. D. Benson & W. S. Greaves (Eds.),
Systemic Perspectives on Discourse, Vol. 1: Selected Papers from the 9th International Systemic
Workshop (pp. 1–15) [Advances in Discourse Processes XV]. Norwood, NJ: Ablex.
Halliday, Michael A. K. (1994a). An Introduction to Functional Grammar (2nd ed.). London:
Edward Arnold.
Halliday, Michael A. K. (1994b). “Systemic Theory”. In R. E. Asher & J. M. Y. Simpson (Eds.),
The Encyclopedia of Language and Linguistics (pp. 4505–4508). Oxford: Pergamon Press.
Halliday, Michael A. K. (1996). “On grammar and grammatics”. In R. Hasan, C. Cloran, &
D. Butt (Eds.), Functional Descriptions: Theory in Practice (pp. 1–38). Amsterdam and
Halliday, Michael A. K. (2004). An Introduction to Functional Grammar (3rd ed., revised by
Christian M. I. M. Matthiessen). London: Arnold.
Halliday, Michael A. K. & Christian M. I. M. Matthiessen (1999). Construing Experience through
Meaning: A Language-Based Approach to Cognition. London and New York: Cassell.
Functional approaches to language 
Hengeveld, Kees (2004a). “The architecture of a Functional Discourse Grammar”. In J. L.

Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional Grammar
(pp. 1–21). Berlin and New York: Mouton de Gruyter.
Hengeveld, Kees (2004b). “Epilogue”. In J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A
New Architecture for Functional Grammar (pp. 365–378). Berlin and New York: Mouton de
Gruyter.
Hengeveld, Kees (2005). “Dynamic expression in Functional Discourse Grammar”. In C. de
Groot & K. Hengeveld (Eds.), Morphosyntactic Expression in Functional Grammar (pp.
53–86) [Functional Grammar Series 27]. Berlin and New York: Mouton de Gruyter.
Hengeveld, Kees & J. Lachlan Mackenzie (forthcoming a). Functional Discourse Grammar.
Oxford: Oxford University Press.
Hengeveld, Kees & J. Lachlan Mackenzie (forthcoming b). “Functional Discourse Grammar”. In
K. Brown (Ed.), The Encyclopedia of Language and Linguistics (2nd ed.). Oxford: Elsevier.
Hengeveld, Kees & María J. Pérez Quintero (2001). “Descriptive adequacy in Functional
Grammar”. In M. J. Pérez Quintero (Ed.), Challenges and Developments in Functional
Grammar (pp. 103–117) [Monographic section of Revista Canaria de Estudios Ingleses 42].
La Laguna, Tenerife: Universidad de La Laguna.
Hopper, Paul J. (1992). “Emergence of grammar”. In W. Bright (Ed.), International Encyclopedia
of Linguistics (pp. 364–367). New York and Oxford: Oxford University Press.
Hopper, Paul J. (2001). “Grammatical constructions and their discourse origins: prototype
or family resemblance?” In M. Putz, S. Niemeier, & R. Dirven (Eds.), Applied Cognitive
Linguistics I: Theory and Language Acquisition (pp. 109–129). Berlin and New York: Mouton
de Gruyter.
Horie, Kaoru & Bernard Comrie (2000). “Introduction”. In K. Horie (Ed.), Complementation:
Cognitive and Functional Perspectives (pp. 1–10). Amsterdam and Philadelphia: John
Benjamins.
Jackendoff, Ray (2002). Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford:
Oxford University Press.
Keizer, M. Evelien (1992). “Reference, Predication and (In)definiteness in Functional Grammar:
A Functional Approach to English Copular Sentences”. Ph.D. dissertation, Free University
of Amsterdam.
Kemmer, Suzanne & Michael Barlow (2000). “Introduction: a usage-based conception of
language”. In M. Barlow & S. Kemmer (Eds.), Usage-Based Models of Language (pp. vii–
xxviii). Stanford, CA: CSLI
Langacker, Ronald W. (1987). Foundations of Cognitive Grammar, Vol. I: Theoretical Prerequisites.
Stanford, CA: Stanford University Press.
Langacker, Ronald W. (1991). Foundations of Cognitive Grammar, Vol. II: Descriptive Application.
Stanford, CA: Stanford University Press.
Mackenzie, J. Lachlan (2000). “First things first: towards an Incremental Functional Grammar”.
Acta Linguistica Hafniensia, 32, 23–44.
Mackenzie, J. Lachlan (2004). “Functional Discourse Grammar and language production”. In
J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional
Malinowski, Bruno (1923). “The problem of meaning in primitive languages”. Supplement to C.
K. Ogden & I. A. Richards The Meaning of Meaning. London: Routledge and Kegan Paul.
Malinowski, Bruno (1935). Coral Gardens and their Magic, Vol. 2. London: George Allen and
Unwin.
 Christopher S. Butler
Martin, James R. (1992). English Text: System and Structure. Amsterdam and Philadelphia: John
Benjamins.
Martin, James R. & Christian M. I. M. Matthiessen (1991). “Systemic typology and topology”. In
F. Christie (Ed.), Literacy in Social Processes: Papers from the Inaugural Systemic Functional
Linguistics Conference, Deakin University, January 1990 (pp. 345–383). Darwin: Centre for
Studies of Language in Education, Northern Territory University.
Matthiessen, Christian M. I. M. (1995). Lexicogrammatical Cartography. Tokyo: International
Language Sciences Publishers.
Newmeyer, Frederick J. (1998). Language Form and Language Function. Cambridge, MA and
London: MIT Press.
Nuyts, Jan (1992). Aspects of a Cognitive-Pragmatic Theory of Language. Amsterdam and
Nuyts, Jan (2001). Epistemic Modality, Language, and Conceptualization: A Cognitive-Pragmatic
Perspective. Amsterdam and Philadelphia: John Benjamins.
Palmer, Frank R. (1968). Selected Papers of J. R. Firth 1952–1959. London: Longman.
Sinclair, John M. (1992). “Trust the text”. In M. Davies & L. Ravelli (Eds.), Advances in Systemic
Linguistics: Recent Theory and Practice (pp. 5–19). London: Pinter.
Thompson, Sandra A. (2002). “‘Object complements’ and conversation towards a realistic
account”. Studies in Language, 26(1), 125–163.
Tognini-Bonelli, Elena (2001). Corpus Linguistics at Work [Studies in Corpus Linguistics 6].
Tomasello, Michael (2003). “Introduction: some surprises for psychologists”. In M. Tomasello
(Ed.), The New Psychology of Language: Cognitive and Functional Approaches to Language
Structure, Vol. 2 (pp. 1–14). Mahwah, NJ: Lawrence Erlbaum.
Tucker, Gordon H. (1996). “So grammarians haven’t the faintest idea: reconciling lexis-oriented
and grammar-oriented approaches to language”. In R. Hasan, C. Cloran, & D. Butt (Eds.),
Functional Descriptions: Theory in Practice (pp. 145–178). Amsterdam and Philadelphia:
John Benjamins.
Tucker, Gordon H. (in press). “Systemic incorporation: on the relationship between corpus and
Systemic Functional Grammar”. To appear in G. Thompson & S. Hunston (Eds.), System
and Corpus: Exploring Connections. London: Equinox.
Van Valin, Robert D., Jr. (1990). “Layered Syntax in Role and Reference Grammar”. In J. Nuyts,
A. M. Bolkestein, & C. Vet (Eds.), Layers and Levels of Representation in Language Theory
(pp. 193–231). Amsterdam and Philadelphia: John Benjamins.
Van Valin, Robert D., Jr. (1991a). “Variations on a functionalist theme”. In Souvenir of
the International Summer Institute of Functional Linguistics, 1991, Hyderabad (pp. 1–8).
Hyderabad: Central Institute of English and Foreign Languages.
Van Valin, Robert D., Jr. (1991b). “Functionalist linguistic theory and language acquisition”. First
Language, 11, 7–40.
Van Valin, Robert D., Jr. (1995). “Role and Reference Grammar”. In J. Verschueren, J.-A.
Östman, & J. Blommaert (Eds.), Handbook of Pragmatics (pp. 461–469). Amsterdam and
Van Valin, Robert D., Jr. (2000). “Functional linguistics”. In M. Aronoff & J. Rees-Miller (Eds.),
The Handbook of Linguistics (pp. 319–336). Oxford: Blackwell.
Van Valin, Robert D., Jr. (2005). Exploring the Syntax-Semantics Interface. Cambridge:
Cambridge University Press.
Van Valin, Robert D., Jr. & Randy J. LaPolla (1997). Syntax: Structure, Meaning and Function.
Cambridge: Cambridge University Press.
Functional approaches to language 
Verhagen, Arie (2000). “Interpreting usage: construing the history of Dutch causal verbs”. In
Barlow & Kemmer (Eds.), Usage-Based Models of Language (pp. 261–286). Stanford, CA:
CSLI.
Wray, Alison (2002). Formulaic Language and the Lexicon. Cambridge: Cambridge University
Press.
Young, Lynne & Claire Harrison (Eds.). (2004). Systemic Functional Linguistics and Critical
Discourse Analysis: Studies in Social Change. London and New York: Continuum.
Resource list
Functional Grammar web page: http://www.functionalgrammar.com

Role and Reference Grammar web page: http://linguistics.buffalo.edu/research/rrg.html
Systemic Functional Grammar web page: http://www.wagsoft.com/Systemics
On contrastive linguistics
Trends, challenges and problems*
María de los Ángeles Gómez-González and
. Some terminological issues
The label contrastive has been used in linguistic inquiry mainly to refer to inter-
linguistic and inter-cultural comparisons; it has, however, also been used for compar
isons within languages/cultures (cf. Fries 1945; Hellinger & Ammon 1996; Lado 1957;
Trager 1949). The raison d’être of contrastive investigations is to compare (or contrast)
linguistic and socio-cultural data across different languages (cross-linguistic/cultural
perspective) or within individual languages (intra-linguistic/cultural perspective) in
order to establish language-specific, typological and/or universal patterns, categories
and features (cf. Altenberg & Granger 2002b; Garrudo-Carabias 1996; Hawkins 1988;
Hellinger & Ammon 1996; Johansson & Hofland 1994; Oleksy 1989).
An astonishingly varied assortment of collocations and corresponding areas of
study emerge when considering the various head nouns such adjectives as contrastive
or comparative most readily co-occur with in the literature. Thus, depending on
what particular authors feel to be the most appropriate description for the issue
under discussion, we find such labels as (Applied) Contrastive (Language) Studies,
Contrastive Linguistics, Comparative (Historical or Typological) Linguistics, Contrastive
(Interlanguage) Analysis, Contrastive (Generative) Grammar, Comparative Syntax, Con
trastive Lexicology/Lexicography, Contrastive Pragmatics, Contrastive Discourse Analysis,
or Contrastive Sociolinguistics, to mention but a few. A selective bibliography on con
trastive linguistics is offered in the COLLATE group website (see the reference list),
while a synthetic treatment of relevant terminology and the main trends in the field
can be found in e.g. Di Pietro (1971), Fisiak (1980, 1981b, 1983, 1984), Granger
(1996), Hartmann (1977), Hoey & Houghton (1998), James (1980), Jaszczolt (1995a,
b), Krzeszowski (1974, 1989, 1990), Mair & Markus (1992), Nickel (1971, 1972), Rein
(1983), Rusiecki (1976), and Sridhar (1981). By way of illustration, Fisiak suggests a
major distinction between theoretical and applied contrastive investigations, that is,
contrastive research performed for its own sake, as opposed to that performed for the
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
purpose of some application; whereas Krzeszowski proposes the opposition text-bound

vs. system-bound, reflecting the Saussurean distinction between parole and langue, or
language use and language system.
Leaving aside the pertinence of such divisions, which, although widespread in
research, have not truly established themselves and are therefore not unanimously
accepted, here we shall argue that behind this terminological jungle lie three main
factors that are closely interrelated and difficult to separate. One is, as Krzeszowski
(1989: 56–58) suggests, the apparent lack of a well established theoretical contrastivist
framework, independent of specific linguistic models employed in the descriptions of
the languages under analysis (functional, generative, constructional, cognitive, etc.).
Partly responsible for this situation appears to be the widespread conviction among
contrastivists that relevant theoretical problems arise and can be solved within the cor
responding domain of pure linguistics – a view that raises the problem of the adequacy
of linguistic models, as these may show different degrees of suitability for contrastive
investigations; while the contrastivist perspective often appears to be reduced to the
use of comparative procedures that simply tend to apply the principles and findings
derived from theoretical and descriptive linguistics. Johansson (1975: 15), Ringbom
(1994: 738–740) and Sajavaara (1996: 17–20), for example, make succinct presenta
tions of what are in their views the main problems of applied contrastive investigations.
Briefly, these authors claim that contrastive research cannot be equated with purely lin
guistic studies, as there are aspects or (sub)fields that are peculiar to it, but have often
been discarded or rejected by many for not being sanctioned as theoretical. Some cases
in point are such fields as: (1) error analysis, or the systematization of types of mother
tongue induced errors among second language learners in order to gain insight into
processes of transfer and the representation of rules; (2) performance analysis, that
is, the description of the total performance of learners; and (3) interlanguage studies,
which focus on the learner’s development and communicative strategies (cf. Ringbom
1994: 740–741).
Closely connected with this, the second factor is that virtually every aspect of lin
guistic analysis can be approached from a contrastive perspective, and accordingly
research in the field flows from numerous academic disciplines that are very different
from one another. Thus, the adjectives contrastive or comparative have been assigned
to disciplines traditionally regarded to pertain to theoretical linguistics (e.g. phonet
ics/phonology, morphology, lexicology, syntax, semantics, pragmatics, discourse anal
ysis or sociolinguistics), as well as to those fields that have been traditionally dubbed
as applied (mostly language learning/teaching, translation and interpreting), as will
be further explained in Section 2 below. Be that as it may, the point at issue here
is – as Johansson (2003: 32) comments – that we are not speaking about a unified
field of study.
And thirdly, behind this terminological profusion there seems to exist a differ
ence of scope with regard to the three main collocations the aforementioned terms
tend to cluster around, namely: (i) contrastive studies (CS), (ii) contrastive analysis
(CA), and (iii) contrastive linguistics (CL). It would seem that CS names the most gen
On contrastive linguistics 
eral field, embodying both the linguistic and the extralinguistic (e.g. cultural, ethno
graphic, semiotic, etc.) dimensions of contrastive research. Some investigations that
adopt this comprehensive perspective are Aijmer, Altenberg, & Johansson (1996a),
Chesterman (1998), Filipović (1974), Fisiak (1980, 1984), Krzeszowski (1985, 1989),
Marton (1979), Mukattash (2001), or Twaddell (1968), to mention only some.
By contrast, CA, though frequently used interchangeably with the other two collo
cations, seems to more accurately name the third of the three steps involved in classical
contrastive procedure: description, juxtaposition and comparison (cf. Jaszczolt 1995b;
Krzeszowski 1990: 35). Description includes the selection and preliminary character
isation of the items under comparison in the framework of a language-independent
theoretical model. Juxtaposition involves a search for, and identification of, cross
/intra-linguistic/cultural equivalents, while the comparison proper evaluates the de
gree and type of correspondence between items under comparison. A detailed account
of the history, strengths and weaknesses of CA can be found in e.g. Aarts (1982),
Chesterman (1998), Filipović (1984), Fries (1945), Krzeszowski (1990), Ringbom
(1994), Sajavaara (1977, 1996), and Van Buren (1974).
Lastly, CL could be said to restrict its domain to just contrastive linguistic research,
whether theoretical, focusing on a contrastive description of the languages/cultures in
volved, or practical/applied, intended to serve the needs of a particular application, as
will be discussed in turn.
. The revival of Contrastive Linguistics (CL)
The origins of CL as a regular linguistic procedure can be traced back to the middle of
the 15th century, and the appearance of the first contrastive theories to the beginning
of the 17th century (cf. Krzeszowski 1985, 1990). In the 19th century comparative in
vestigations used an empirical, historical methodology to discover genetic links and
language families; while in modern linguistics, J. Baudouin de Courtenay’s compara
tive studies of Slavic and other Indoeuropean languages were continued by the Prague
Circle, whose members also spoke about analytical comparison, or linguistic charac
terology, as a way of determining the characteristics of each language and gaining a
deeper insight into their specific features (Andreeva 1990; Firbas 1992: 3 ff.; Math
esius 1975). But it was not until after World War II that the discipline reached its
heyday. From its beginnings till the 1970s, CL basically served practical pedagogical
purposes in foreign and second language teaching/learning (Fries 1945; Lado 1957;
Nickel & Wagner 1979). It was mainly synchronic (rather than diachronic) – in fact,
some would exclusively use the term comparative linguistics to refer to the diachronic
study of genetically related languages – interlingual or cross-linguistic (rather than in
tralingual), involved two different languages (rather than more than two (varieties of)
languages/cultures), adopted a unidirectional perspective (taking one of the two lan
guages as frame of reference, usually English), focused on differences (rather than on
parallelisms), and was directed to foreign language teaching/learning (cf. Altenberg &
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Granger 2002b; Di Pietro 1971; James 1980; Hellinger & Ammon 1996; Krzeszowski
1990; Morciniec 2001; Rein 1983).
Now, in a time when we speak about the world as a global village, when there exists
a greater recognition of intra-/cross-linguistic/cultural variation, a growing awareness
has emerged of the need for multilingual/multicultural and intra-linguistic/cultural
competence and research. In addition, and as a side effect of this, there has been a
change of focus in linguistic research, which has shifted away from speculative au
tonomous theorizing in the direction of a more dynamic and practical view of language
processing and interaction. Undoubtedly, a further factor which has contributed to
this shift is the advent of large (bilingual or multilingual) text corpora and computer
search tools, which have opened up new fronts of research in the fields of linguistic
description (at all levels), computational linguistics, machine translation or informa
tion retrieval, to mention but a few. All these circumstances have given rise to lively
theoretical, descriptive, applied and methodological discussion which has turned CL
into a reactivated and at present expanding field.
This trend towards expansion was foreseen by Trager (1949), who suggested that
CL should move beyond structurally-oriented views – predominant in the United
States throughout the 50s and 60s (cf. Haugen 1953; Weinreich 1953) – and extend
its scope so as to describe the differences, as well as the similarities between two
or more linguistic systems, both cross-linguistically and intralinguistically, and both
synchronically and diachronically. Thus, on the diachronic level, issues regarding the
phylogenetic development of languages are high on the agenda of CL, as well as the
ontogenetic development of individual language acquisition (Kastovsky & Szwedek
1986; Lamiroy 1993). In addition, Sajavaara (1981, 1996) claims that in order to ac
count for an individual’s communicative competence, the goal of inquiry in CL must
also include discourse analysis, psycholinguistics, and sociolinguistics, a position also
endorsed by Kühlwein (1990), among many others, who argues for the integration of
structural and processual CL, the latter entailing the analysis of systems of knowledge
and knowledge about structural systems. Likewise, Liebe-Harkort (1985), following
Lado’s (1957) position, adds that languages cannot be compared without comparing
the cultures in which they are spoken (cf. Olhstain 1983; Odlin 1989). The same idea
is insisted upon by Kühlwein (1990), who is particularly interested in culturally dif
ferentiated semiotic systems that serve as the starting point for social and language
interaction. But in addition, he emphasizes the relevance of CL for foreign language
teaching, given its growing recognition of performance errors, interlanguage, trans
fer (i.e. the interference of L1 in L2), and the interaction of cognition and discourse
processes. An extreme form of this trend is represented by a recent view of contrastive
literature that reduces the key task of CL to predicting and thereby obviating learn
ers’ errors (Krzeszowski 1990: 243; Riegelhaupt 1989; Selinker 1992: 65–67), while this
procedure is openly criticized by other authors such as Garrudo-Carabias (1996).
For two reasons, the present introduction will make no attempt to exhaustively
systematize the ever growing mass of studies resulting from the revival of CL, nor will
it give a definition of each of the disciplines that could be included within the um
On contrastive linguistics 
brella term of CL. The first reason is that this task would require a book on its own;
the second is that CL encompasses many and rapidly growing (sub)fields, a situation
that is complicated even further by the often formidable problems posed by simul
taneously describing and relating several languages/cultures, some of which will be
detailed in Section 3 below. Instead, the reader is invited to find out for him/herself
about the most prolific trends in CL (yielding relevant results for teaching and other
practical domains) on the basis of the synopsis that follows, and obviously also the
papers included in this volume.
Firstly, we should distinguish between CL and typological research. Broadly, the
latter adopts a multilingual perspective to find linguistic universals, that is, fundamen
tally structural similarities and differences among languages. It has a great breadth of
coverage and can rectify any bias in monolingual work towards the major Indoeuro
pean languages (cf. Birnbaum 1986; Comrie 1986; Hawkins 1985, 1988; see Di Pietro
1971; James 1980; and Krzeszowski 1990 for various early approaches). Lying some
where between monolingual and typological research, CL combines the crosslinguistic
perspective – comparing a smaller number of languages – with analytical depth and
puts in a sharper focus mainly the differences between languages, their underlying re
sources, and the use of such resources in discourse. Some researchers have explored
what CL can do for typology. Thus, Brdar (1996) argues that the implementation of
contrastive methodology – working with a network of syntactic, morphological and
semantic parameters – helps in the verification/falsification of putative language uni
versals and parameter-setting, and that corpus-based bidirectional approaches would
make the typology-derived picture less neat and orderly but more realistic; whereas
Tarasova (1993) proposes a “content typology approach” to CL, isolating fundamental
categories of analysis (i.e. unit, class and system) and focusing on extralinguistic con
ceptual properties and language-specific realizations in order to allow comparisons of
semantic forms.
Now considering the different levels of linguistic description, most contrastive
phonetic studies focus on articulatory and acoustic comparisons between two lan
guages, as is the case with Agard & Di Pietro (1966a), Bald (1976), Moulton (1962),
Stockwell & Bowen (1965), or Strangert (1981), among others; while other investi
gations run the full gamut of contrastive phonological issues (cf. e.g. Carlisle 1988;
Eliasson 1984; Monroy 2001). A thorough list of bibliography and resources for
contrastive phonetics/phonology can be found on Joaquim Llisterri’s website (see
resource list).
Moving on to lexical CL (LCL), this research concentrates on cross-/intra
linguistic comparisons of “lexical items”, i.e. stable (multi)word pairings of form and
meaning, considering grammatical, semantic and pragmatic information involved in
the interdependence between lexical choice and contextual factors (cf. Sinclair 1998).
General overviews of LCL are provided in e.g. Altenberg & Granger (2002a, b), Atkins,
Levin, & Zampolli (1994), Viberg (1993), and Weigand (1998). Turning to case studies,
some trends could be mentioned regarding the languages under analysis, in Europe,
for example: (i) English-Swedish, the work by Aijmer (1996b, 1998, 1999) (modal
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
ity and modal particles), Altenberg (1998, 1999, 2001, 2002) (adverbial connectors
and causative constructions) and Viberg (1996a, b, 1998, 1999, 2002) (cognate verb
pairs); (ii) English-Norwegian, Johansson & Løken’s (1997) analysis of modal particles,
or Johansson’s (1997, 1998) account of the generic pronoun man and other particles
in English-German-Norwegian; (iii) English-French-Dutch, the Contrastive Verb Va
lency Dictionary that is being compiled by the Contragram group at the University
of Ghent (see website on the resource list), Paillard’s (2002) work on lexical units
and figures of speech, Salkie’s (1997) study of adversative conjunctions, and Simon-
Vandenbergen’s (1998, 1999) description of verbal equivalents; and (iv) with regard to
English-Spanish, notably the investigations and activities carried out by the Santiago
de Compostela centred SCIMITAR team, Martínez-Vázquez’s research group and the
Complutense study units (see websites on the resource list).
On the other hand, exponents of bilingual/multilingual grammars or bilin
gual/multilingual morphosyntactic aspects are presented in e.g. Aarts & Wekker (1987,
1988), Agard & Di Pietro (1966b), Aijmer (1996a), Contreras (1987), Devos (1995), Di
Pietro (1968), Ebeling (1999), Fente (1971), Fisiak et al. (1974), Haegeman (1997), Hill
& Bradford (1991), Jager & Wekker (1997), Kefer & van der Auwera (1992), Müller-
Gotama (1994), Russ (1981), Simon-Vandenbergen, Taeldeman, & Willems (1996),
Simon-Vandenbergen & Noël (1997), Simon-Vandenbergen, Defrancq, & Davidse
(1998), Stockwell, Bowen, & Martin (1965), Sveen (2000/2001) and Tops, Devriendt,
& Geukens (1999).
Work has also mushroomed regarding the nature of semantic diversity among
the planet’s languages and the implications of semantic diversity for general linguistic
theory. Here the big issue seems to be the testing of the Semantic Universals Hypothe
sis (SUH), that is, the question whether the semantic systems of the world’s natural
languages share (at least) some common properties. Representatives of this trend
on contrastive (lexical) semantics are e.g. Gil (1991), Jaszczolt & Turner (1996), van
Benthem & ter Meulen (1996), and mostly the research by Wierzbicka’s group (e.g.
Wierzbicka 1980, 1991, 1992; Goddard & Wierzbicka 2002; see the Natural Seman
tic Metalanguage website and the references therein). Broadly, while the first authors
focus on contrastive sentential semantics, Wierzbicka’s group moves a step beyond
and argues for the existence of a “universal semantic common measure” founded on
empirically established universal human concepts and their universal combinatory
properties which – they say – can provide an effective basis for CL.
Contrastive Discourse Analysis (CDA) and Contrastive Pragmatics (CP) are two par
tially overlapping labels referring to contrastive research that goes beyond clause/sen
tence level to explore the (textual features of) language in use under the assump
tion that the relations between texts and contexts are mutually reflexive – texts
not only reflect but also shape their contexts (cf. Fillmore 1984; Jaszczolt & Turner
1996; Littlewood 1983; Markkanen 1985; Oleksy 1984, 1989; Smith 1987). Wider in
scope, CDA covers such issues as: (1) discourse particles (e.g. Aijmer 1997; Aijmer
& Simon-Vandenbergen 2003); (2) rhetorical relations and rhetorical transfer across
languages/cultures (e.g. hedging and metadiscourse, generic conventions, author’s and
On contrastive linguistics 
addressee’s intentions, responsibility for textual clarity, etc.) (cf. e.g. Alharbi 1997;
Connor 1996; Clyne, Hoeks, & Kreutz 1988; Moder & Martinovic-Zic 2004; Rudolf
1996); in addition to (3) genre studies and information packaging across languages
and/or text-types, as well as their side effects in terms of coherence and cohesion (cf.
e.g. Clyne 1987a, b; Duszak 1997; Fareh 1995; Givón 1983; Gómez-González 2001;
Hasselgård et al. 2002; Johansson 1996, 2001; Svensson 2000; Swales 1990). CP, in
turn, has been committed since its beginnings to studying certain phenomena (of
ten with a philosophical slant) such as: (1) conversation from a speech act/implicature
point of view (cf. e.g. Blum-Kulka, House, & Kasper 1989; Kalisz 1986; Liebe-Harkort
1985); (2) deixis (Celle 2000; Rauh 1983); (3) politeness (cf. e.g. Brown & Levinson
1987; Dufon et al. 1994; Hickey 2003; Sifianou 1992; Turner 1996); and other prag
matically oriented aspects of speech behaviour (cf. e.g. Kasper & Blum-Kulka 1993;
Tannen 1984; van Benthem & ter Meulen 1996). Nevertheless, it would appear that
these studies have not yet provided a systematic account of the contrastive implications
of face-to-face interactions.
Also close to or overlapping with CP and CDA, the field “Contrastive Sociolin
guistics” (CSL) is similarly in the ascendant (cf. Ammon et al. 1987/1988; Hellinger &
Ammon 1996; Janicki 1980, 1986). The latter claims that contrastive sociolinguistics
should aim at the systematic comparison of sociolinguistic patterns and the develop
ment of a theory of language use, defining the field as “a systematic juxtaposition of
linguistic items as they are distributed in the multi-dimensional (multi-parameter)
social space” (Janicki 1984: 28; quoted by Hellinger & Ammon 1996: 8). In this vein,
such authors as Ammon (1989) or Hellinger (1992) propose corresponding typolo
gies of social patterns. However, it would seem that this definition leaves out all the
phenomena associated with the sociology of language which in principle should also
concern CSL. For this reason current definitions and developments in the field argue
for more comprehensive views, in which CSL is regarded as a branch of sociolinguistics
and aims at providing comparison of cross-/intra-/multi-cultural sociopragmatic data
along such research lines as multilingualism, language planning and language politics
(cf. Ammon et al. 1987/1988; Goddard & Wierzbicka 2002; Hellinger & Ammon 1996).
Now turning to the area of computational linguistics, efforts have been devoted
to, for example, the creation of different types of electronic dictionaries (Atkins 1994;
Cardey & Greenfield 2002; Colleman 2002; Dawes 2003) or the design of computer
tools for cross-linguistic research, especially in translation enquiries (Corness 2002;
King 2003) and machine translation, where the results have been disappointing, partly
due to the limitations of computational resources, but mainly owing to the complexity
entailed in translation processes (Steffens 1995).
This leads us to the fields of interpreting and translation, of which some expo
nents could be Astington (1983), Baker (1993, 1999), Ballard (1995), Fabricius-Hansen
(1996), Gellerstam (1996), Granger, Lerot, & Petch-Tyson (2003), Guillemin-Flescher
(1992, 2000), or Teich (1999) and in Spain, the activities developed by the already al
luded groups at the universities of Santiago de Compostela and the Complutense, as
well as Martínez-Vázquez’s research team.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Finally, it should be mentioned that most of the aforementioned contrastive re

search is based on the exploitation of corpora, a methodology which helps in achieving
authenticity and empirical adequacy in the investigations. It should be noted, however,
that corpus evidence itself has limitations, for it is necessarily based on a selection of
texts put together in a principled way and prepared for computer processing, which
by definition can never contain all the possible events/instances of a given language
(or language type/variant, etc.). Fabricius-Hansen & Solfjeld (1994), Ilson (1991) and
Sinclair (forthcoming), for example, offer interesting discussions about the problems
researchers are faced with when compiling the texts of a corpus.
But despite its limitations and problems, the power and utility of corpus anal
ysis have been widely attested in a huge amount of studies, of which the aforemen
tioned and the following constitute only a small fraction: Ahrenberg & Merkel (1996),
Baker (1993, 1999), Barlow (2000), Borin (2002), Botley, McEnery, & Wilson (2000),
Butler (2002), Fries, Müller, & Schneider (1997), Granger (2003), Granger, Lerot,
& Petch-Tyson (2003), Johansson (1997, 1998, 2002), Johansson & Oksefjell (1998),
Lewandowska-Tomaszczyk & Melia (1997), Mair & Hundt (2000), Maniez (2002),
McEnery & Wilson (1996), Sinclair, Payne, & Pérez Hernández (1996), and Svartvik
(1992). More extensive lists of corpus linguistics investigations may be found in e.g.
the online bibliographies created by Michael Barlow, Manuel Barbera, the Survey of
English Usage, or The Gateway to Corpus Linguistics (see resource list).
In closing this section, a relevant distinction should be made between “corpus
based” and “corpus-driven” approaches (see e.g. Francis 1993; Sinclair 1998; and
Tognini-Bonelli 2001). Although the former is usually taken as a cover term for corpus-
based studies, the main difference lies in the degree of importance conferred on the
data. Thus, “corpus-based” studies are broadly those that depart from a cross-/intra
linguistic difference or similarity on theoretical grounds using the data to confirm,
refute or enrich the theory. By contrast, in corpus-driven approaches the corpus evi
dence is used to discover cross-/intralinguistic correspondences and to make theoreti
cal statements derived from these.
In addition, several types of corpora should be distinguished. Thus, Johansson
(1998: 4–7), for example, differentiates two kinds of multilingual (including bilingual)
corpora: (1) comparable corpora, consisting of original texts in the same language,
of equivalent type, subject matter and communicative function, which can be either
domain-specific or general collecting different text types (Teich 2003); and (2) trans
lation corpora, containing texts in the original languages and their translations into
other languages. Translation corpora are unidirectional if the translations go only from
one language to (an)other language(s), and they are bi-/multi-directional if transla
tions from and to all the languages under comparison are considered (Steiner & Yallop
2001). Lastly, parallel corpora are, strictly speaking, aligned corpora in which transla
tions are linked in corresponding units (Johansson, Ebeling, & Hofland 1996; Kraig
2002; Veronis 2000), although the label “parallel corpora” has also been used as a
cover term for the other two types – especially in the natural language processing
community – (cf. Baker 1995, 1999; Hartmann 1996). There are many types of par
On contrastive linguistics 
allel corpora: comparable corpora of different historical periods, or for different social
and regional varieties, or learner language corpora. As Granger (1996, 1998) points
out, contrastive and learner corpora complement each other. The former compare lan
guages and formulate hypothesis about learning problems, while the latter identify the
characteristics of learner language and interferences of the mother tongue, which in
turn provide feed-back for contrastive descriptions. LINGUA (Johns 1998), EAGLES
(Calzolari, McNaught, & Zampolli 1996), and ICLE (Granger 1996, 1998) are three
examples of projects devoted to multilingual corpora (the first two) and learner lan
guage corpora (the latter). The advantages and disadvantages of each type of corpus
and their suitability for different types of study are addressed in e.g. Aijmer, Altenberg,
& Johansson (1996b), Granger (1996, 2003), Granger, Lerot, & Petch-Tyson (2003),
Johansson (1998, 2003), Laviosa (1998) and Teubert (1996, 2002).
. Looking ahead: Challenges and problems
Especially in Europe, a large number of research projects, conferences and journals

contribute to revitalizing CL, encouraging cross-fertilization of such fields as lin
guistics, computational linguists, translation/interpreting, lexicography and corpus
linguistics in order to implement the applications of the field. These range from the
production of (foreign language) teaching materials and computer tools, to bilingual
or multilingual dictionaries. Some cases in point are the periodical journals Lan
guages in Contrast (John Benjamins), Language Sciences, Vol. 18 (1996), Linguistics,
Vol. 34.3 (1996), International Journal of Lexicography, Vol. 9.3 (1996), and Contra-
gram, the online version of the Quarterly Newsletter of the Contrastive Grammar
Research Group at the University of Ghent (see resource list). Alongside the many
projects described in Danielsson and Ridings (1996), Filipović (1984), Gellerstam
(1996) or the EuroWordNet project (Vossen 1998), mainly in the field of lexicography,
machine-assisted translation and information retrieval, the activities and production
of the following research groups (and projects) also merit especial mention (see web-
sites on the resource list): (i) the Polish-English Contrastive Project (Fisiak 1971); (ii)
COLLATE, COntrastive Linguistics and LAnguage Typology in Europe, and CONTRA-
GRAM, its core research unit (University of Ghent), working on Dutch, French and
English (Colleman 2002); (iii) the SPRIK project and the English-Norwegian Parallel
Corpus project (ENPC) (English-German-Norwegian); (iv) the English-Swedish Par
allel Corpus (Altenberg & Aijmer 2000; Altenberg, Aijmer, & Svensson 1999, 2001);
(v) the Finnish-English Contrastive Project (Sajavaara & Lehtonen 1980: 9–10); (vi)
the PLUG project (Sågvall Hein 2002); (vii) the DELIC research group (taking French
as the reference language); and (viii) in Spain, the production and events associated
with the SCIMITAR team, Martínez-Vázquez’s group and the research units at the
Complutense in Madrid.
Many problems remain to be solved and many challenges lie ahead (cf. Altenberg
& Granger 2002b; Di Pietro 1971; Fisiak 1984; Garrudo-Carabias 1996; Hellinger &
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Ammon 1996; James 1980; Krzeszowski 1990; Morciniec 2001; Rein 1983). One such
problem is, as noted, the difficulty of comparing some aspects of language such as
the temporal, aspectual and modal systems of verbs, or such areas as gender or the
functional/semantic values of prepositions. Another issue, which can be seen as a side
effect of the first, is that it is virtually impossible to contrast two (or more) languages in
their entirety. At most, only subsytems can be compared at a microlinguistic level. But
these subsystems are again very numerous, which makes it very difficult to produce an
exhaustive contrastive analysis of two (or more) languages.
In addition, it is often claimed that contrastive research requires a holistic view
that approaches the items under study against the background of productive and
receptive/comprehensive message-processing systems. And, in order to produce de
scriptively adequate accounts, these systems are in turn expected to be functioning
in the context of human interaction and along the varying cognition processes un
dergone by speakers/learners (Kühlwein 1990; Towell & Hawkins 1994), taking into
account different types of linguistic knowledge (explicit, implicit, formal, informal)
and their relationships with other types of knowledge (James 1992). As a result and
related to the aforementioned, there seem to be many issues intervening and thereby
opening up corresponding lines of research within the general umbrella of CL. This is
the case of, for example, learning-psychological research on the relationship between
the working of different types (and intensities) of transfer, interference and avoid
ance strategies in second language performance depending on the speaker/learner
and his/her age or environment (Benson 2002) such as negative experience (Gass
1989: 394), under-representation of target forms (Levenston 1972), the influence of
perceived competence, the image of communicator and self-concept, or the manage
ment of control, to mention but a few factors (Sajavaara 1996). Other relevant aspects
seem to be the existence of structural differences between L1 and L2 or the impact
of consciousness/awareness on language processing (Laufer & Eliasson 1993). Along
side this, worthy of particular note is also the fact that different objectives seem to
require different methodologies; and, just as there are no two learners/speakers that
are exactly alike, some approaches will be more adequate for some learners/speakers,
whereas other perspectives will be more appropriate for others. It looks as if it is not
always an easy or even a feasible task to take all these aspects into account when un
dertaking contrastive investigations, which for this reason could be liable to criticism
on the grounds of deficient accountability or surfacey analyses.
A further and central point of controversy among contrastivists is the question of
equivalence or tertium comparationis. In order for two or more categories to be com
pared, that is, in order to determine whether they are (dis)similar, it is necessary that
they have some common ground, or tertium comparationis (TC). The extent to which
a TC can be found for a set of elements across/within languages/cultures determines
the extent to which these elements are equivalent. For only equivalent elements are
at all comparable. Thus, equivalence and TC are two sides of the same coin and for
that reason they tend to be used interchangeably. Depending on the TC adopted and
the kind of equivalence involved, which in turn are affected by the level or aspect un
On contrastive linguistics 
der comparison – as already noted – various kinds of contrastive investigations can be

distinguished.
But again it is not always an easy task to determine the TC of a comparison:
(dis)similar) meaning, (dis)similar) function, (dis)similar) grammatical environment,
(dis)similar) context? The same applies to the degree of equivalence that is at issue or
that is attained in a comparison: can it be absolute? Is it measurable, and if so, how can
it be measured? These and other related questions have been addressed by e.g. Chester-
man (1998), Hoey & Houghton (1998), James (1980), Janicki (1986), Jaszczolt (1995a,
b), Krzeszowski (1984, 1990), or Oleksy (1986). By way of illustration, Krzeszowski
(1990: 23ff.) differentiates seven types of equivalence: translation equivalence, statis
tical equivalence, system equivalence, semantico-syntactic equivalence, rule equiva
lence, substantive equivalence and pragmatic equivalence. Space constraints preclude
a detailed account of these equivalence types, but it would seem that, their positive
aspects notwithstanding, they have come in for criticism owing to an excessive formal
ism (for a critique of Krzeszowski’s equivalence types see e.g. James 1980; Oleksy 1986;
Van Buren 1976).
Alternative perspectives on this concept have produced the terms functional or
communicative equivalence, consisting of a cluster of functional and formal proper
ties (cf. Hansen 1985; Tognini-Bonelli 2002), or that of pragmatic equivalence, which
deals with both general pragmatic phenomena (e.g. politeness systems) and particu
lar phenomena such as the use of the past simple across languages, etc., in an attempt
to determine in which way (dis)similar phenomena evoke (dis)similar cognitive reac
tions in language users (cf. Becka 1978: 131–132; Fillmore 1984; Kalisz 1986; Oleksy
1984, 1986; Riley 1980). Unfortunately, however, it appears that so far no generally
accepted procedure has been developed to attest pragmatic equivalence cross- and
intra-linguistically.
Perhaps more widespread and wider in scope is the concept of translation equiv
alence explored by e.g. Astington (1983), James (1980), Chesterman (1998), Salkie
(1997) and Sinclair, Payne, & Pérez Hernández (1996). In essence, it is suggested that
the TC should not be founded on equivalence in a strict sense (identity of meaning or
a taxonomy of it) because cross-linguistic equivalence is not absolute, but rather a rel
ative concept with often fuzzy boundaries that are impossible to delineate satisfactorily
(e.g. the area of epistemic modality or pragmatic particles). Among other reasons, this
is mainly due to such complex relationships as (1) overlapping polysemy (items which
have the same meaning extensions crosslinguistically) (cf. Alsina & DeCesaris 2002),
(2) diverging polysemy (items which have different meaning extensions crosslinguis
tically) (cf. Viberg 1996a, b, 1998, 1999, 2002), and (3) lack of correspondence (items
that have no obvious equivalents in other languages) (cf. Johansson & Løken 1997;
Johansson 1998; Kittay & Lehrer 1992). In addition to these signs of the source/target
language/culture that make it difficult to find literal renderings of original texts – we
should not forget that if communication is culturally relative, so are texts (Tannen
1984: 194) – and often as a corollary, there are also traces of universal translation strate
gies, used in the process of transfer from one language/culture to another, which also
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
result in deviations from the source text. Such deviations may involve additions, omis
sions, and various kinds of “free” renderings that are either unmotivated or respond to
cultural and communicative considerations (cf. Salkie’s 2002 notion of “modulation”
to refer to shifts in perspective resulting from “unexpected” translations).
In view of these problems, there seems to be a trend towards viewing translation
equivalence as a matter of judgement, or, more precisely, as a question of translation
competence, interpreted as the ability to relate two or more things across/within lan
guages/cultures in a given context. In this relativistic vein, three procedures could be
mentioned to attest translation equivalence more or less objectively proposed by re
spectively: (1) Ivir (1983, 1987), (2) Krzeszowski (1990: 27), and (3) Altenberg (1999)
or Ebeling (1999). The first suggests the idea of back-translation, that is, to restrict the
comparison to forms in L2 that can be translated back into the original forms in L1,
with which part of the deviations just described are avoided. The second argues for
a quantitative notion of equivalence: the more highly used the translation, the more
relevant it is. And finally, the third approach, a combination of the other two, cal
culates the mutual correspondence (or translatability) of two items in a bidirectional
translation corpus: the higher the mutual correspondence value between an item x in
language A and an item y in language B is, the greater the equivalence between the
compared items is likely to be.
However, none of these methods seems to be infallible either. The first two miss
out valuable crosslinguistic/cultural evidence that could be provided by unexpected
or impossible translations; while the third one does not help us determine what is or
is not equivalent. Be that as it may, the issue underlying and explaining these short
comings seems to be that the notion of (translation) equivalence and its correlate,
the TC, remain a matter of judgement and a question of degree, ultimately reflecting
the researcher’s competence and intuitions, which for some could be taken as a sign
of subjectivism allegedly in conflict with the objectivity that should rule a scientific
approach.
In rounding off this outline, an outsider’s view of CL would possibly demand a
conclusive demonstration that knowledge of difference/similarity contributes to better
translations, greater accuracy and fluency in L1 and L2, and other improvements with
respect to what is offered by non-contrastive investigations. Once such is the case or
if that is already the case – as would probably be agreed upon by most if not all who
regard themselves as contrastivists – then we would urgently need a comprehensive
redescription of languages in the shape of grammars, teaching materials, dictionaries,
etc., based on evidence provided by contrastive (cross-/intra-linguistic) and corpus-
based/driven research (Sinclair 1998: 14; Teubert 1996: 238).
It is manifest that not all researchers are yet adequately equipped with comput
erized resources and tools to undertake this kind of study, neither are the available
corpus resources in many (variations of) languages adequately developed to satisfy
the requirements of modern contrastive corpus research. This situation therefore calls
for the elaboration of more comprehensive and more accessible multilingual corpora
(including variation within languages), which should be accompanied by more power
On contrastive linguistics 
ful and user-friendly software tools, which, in turn, would require more coordination
and cooperation across related disciplines. There are countless issues here, but we are
convinced that challenges and problems will only spur further activity in the field. Ac
cordingly, it is to be hoped that over the coming years, with increased technological
support, we will witness dramatic advances in CL.
Note
* We are grateful to Stig Johansson, AnneMarie Simon-Vandenbergen, Christopher Butler,

Lachlan Mackenzie and Montserrat Martínez-Vázquez for invaluable assistance in preparing
this overview.
References
Aarts, Flor (1982). “The Contrastive Analysis debate: problems and solutions”. Studia Anglica
Posnaniensia, 14, 47–68.
Aarts, Flor & Herman Wekker (1987). A Contrastive Grammar of English and Dutch / Contrastieve
Grammatica Engels. Engels-Nederlands. Leiden: Martinus Nijhoff.
Aarts, Flor & Herman Wekker (1988). “Contrastive Grammar: theory and practice”. In Papers
and Studies in Contrastive Linguistics, 23, 5–17, and in J. Fisiak (Ed.) (1990), Further Insights
into Contrastive Analysis (pp. 163–176) [Linguistic and Literary Studies in Eastern Europe
10]. Amsterdam and Philadelphia: John Benjamins.
Agard, Frederick B. & Robert J. Di Pietro (1966a). The Sounds of English and Italian. Chicago:
Chicago University Press.
Agard, Frederick B. & Robert J. Di Pietro (1966b). The Grammatical Structures of English and
Italian. Chicago: Chicago University Press.
Ahrenberg, Lars & Magnus Merkel (1996). “On translation corpora and support tools: a project
report”. In K. Aijmer, B. Altenberg, & M. Johansson (Eds.), Languages in Contrast. Papers
from a Symposium on Text-based Cross-linguistic Studies (pp. 183–200) [Lund Studies in
English 88]. Lund: Lund University Press.
Aijmer, Karin (Ed.). (1996a). A Wealth of English. Studies in Honour of Göran Kjellmer. Göteborg:
Acta Universitatis Gothoburgensis.
Aijmer, Karin (1996b). “Swedish modal particles in a contrastive perspective”. Language Sciences,
18, 393–427.
Aijmer, Karin (1997). “Epistemic modality as a discourse phenomenon: a Swedish-English
cross-language perspective”. In U. Fries, V. Müller, & P. Schneider (Eds.), From Ælfric to
the New York Times. Studies in English Corpus Linguistics (pp. 215–226). Amsterdam and
Atlanta: Rodopi.
Aijmer, Karin (1998). “Epistemic predicates in contrast”. In S. Johansson & S. Oksefjell (Eds.),
Corpora and Cross-linguistic Research (pp. 277–295). Amsterdam and Atlanta: Rodopi.
Aijmer, Karin (1999). “Epistemic possibility in an English-Swedish contrastive perspective”. In
H. Hasselgård & S. Oksefjell (Eds.), Out of Corpora. Studies in Honour of Stig Johansson (pp.
301–323). Amsterdam and Atlanta: Rodopi.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Aijmer, Karin, Bengt Altenberg, & Mats Johansson (Eds.). (1996a). Languages in Contrast. Papers
from a Symposium on Text-based Cross-linguistic Studies [Lund Studies in English 88]. Lund:
Lund University Press.
Aijmer, Karin, Bengt Altenberg, & Mats Johansson (1996b). “Text-based contrastive studies in
English. Presentation of a project”. In K. Aijmer, B. Altenberg, & M. Johansson (Eds.),
Languages in Contrast. Papers from a Symposium on Text-based Cross-linguistic Studies (pp.
73–85) [Lund Studies in English 88]. Lund: Lund University Press.
Aijmer, Karin & Annemarie Simon-Vandenbergen (2003). “The discourse particle well and its
equivalents in Swedish and Dutch”. Linguistics, 41(6), 1123–1161.
Alharbi, Lafi M. (1997). “Rhetorical transfer across cultures: English into Arabic and Arabic into
English”. Interface, 11(2), 69–94.
Alsina, Victòria & Janet DeCesaris (2002). “Bilingual lexicography, overlapping polysemy,
and corpus use”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 215–230).
Altenberg, Bengt (1998). “Connectors and sentence openings in English and Swedish”. In
S. Johansson & S. Oksefjell (Eds.), Corpora and Cross-linguistic Research (pp. 115–143).
Amsterdam and Atlanta: Rodopi.
Altenberg, Bengt (1999). “Adverbial connectors in English and Swedish: semantic and lexical
correspondences”. In H. Hasselgård & S. Oksefjell (Eds.), Out of Corpora. Studies in Honour
of Stig Johansson (pp. 249–268). Amsterdam and Atlanta: Rodopi.
Altenberg, Bengt (2001). “Contrasting delexical English make and Swedish göra”. In K. Aijmer
(Ed.), A Wealth of English: Studies in Honour of Göran Kjellmer (pp. 195–219) [Gothenburg
Studies in English 81]. Gothenburg: The University of Gothenburg.
Altenberg, Bengt (2002). “Causative constructions in English and Swedish. A corpus-based
contrastive study”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 97–116).
Altenberg, Bengt & Karin Aijmer (2000). “The English-Swedish Parallel Corpus: a resource
for contrastive research and translation studies”. In C. Mair & M. Hundt (Eds.), Corpus
Linguistics and Linguistic Theory: Papers from the Twentieth International Conference on
English Language Research on Computerized Corpora (ICAME 20), Freiburg im Breisgau,
1999 (pp. 15–33). Amsterdam and Atlanta: Rodopi.
Altenberg, Bengt, Karin Aijmer, & Mikael Svensson (1999). The English-Swedish Parallel Corpus
(ESPC): Manual. Department of English, Lund University.
Altenberg, Bengt, Karin Aijmer, & Mikael Svensson (2001). The English-Swedish Parallel Corpus
(ESPC): Manual of Enlarged Version. Department of English, Lund University.
Altenberg, Bengt & Sylviane Granger (Eds.). (2002a). Lexis in Contrast. Amsterdam and
Altenberg, Bengt & Sylviane Granger (2002b). “Recent trends in cross-linguistic lexical studies”.
In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 3–48). Amsterdam and
Ammon, Ulrich, Norbert Dittmar, Klaus Mattheier, & Peter Trudgill (Eds.). (1987/1988).
Sociolinguistics: An International Handbook of the Science of Language and Society. Berlin
and New York: Mouton de Gruyter.
Ammon, Ulrich (1989). “Towards a descriptive framework for the status/function (social
position) of a language within a country”. In U. Ammon (Ed.), Status and Function of
Languages and Language Varieties (pp. 21–106). Berlin and New York: de Gruyter.
Andreeva, E. G. (1990). “Contrastive linguistics abroad: present state of the art”. Vestnik-
Leningradskogo-Universiteta-Istoriya-yazykoznanie-litaraturovedenie, 45(4/2), 58–62.
On contrastive linguistics 
Astington, Eric (1983). Equivalences: Translation Difficulties and Devices French-English and
English-French. Cambridge: Cambridge University Press.
Atkins, B. T. Susan (1994). “A corpus-based dictionary”. In The Oxford-Hachette French
Dictionary (pp. xix–xxvi). Oxford and Paris: Oxford University Press/Hachette Livre.
Atkins, B. T. Susan, Beth Levin, & Antonio Zampolli (1994). “Computational approaches to the
lexicon: an overview”. In B. T. S. Atkins & A. Zampolli (Eds.), Computational Approaches to
the Lexicon (pp. 17–45). Oxford and New York: Oxford University Press.
Baker, Mona (1993). “Corpus linguistics and translation studies: implications and applications”.
In M. Baker et al. (Eds.), Text and Technology. In Honour of John Sinclair (pp. 233–250).
Baker, Mona (1995). “Corpora in translation studies – An overview and some suggestions for
future research”. Target, 7, 223–243.
Baker, Mona (1999). “The role of corpora in investigating the linguistic behaviour of
professional translators”. International Journal of Corpus Linguistics, 4, 281–298.
Bald, Wolf-Dietrich (1976). “Contrastive studies in English and German intonation: a survey”.
Papers and Studies in Contrastive Linguistics, 4, 37–47.
Ballard, Michel (Ed.). (1995). Relations discursives et traduction. Lille: Presses Universitaires de
Lille.
Barlow, Michael (2000). “Parallel texts in language teaching”. In S. P. Botley, A. M. McEnery, &
A. Wilson (Eds.), Multilingual Corpora in Teaching and Research (pp. 106–115). Amsterdam
and Atlanta: Rodopi.
Becka, Jan V. (1978). “Application of quantitative methods in contrastive stylistics.” Prague
Studies in Mathematical Linguistics, 6, 129–147.
Benson, Cathy (2002). “Transfer/Cross-linguistic influence”. ELT Journal, 56(1), 68–70.
Birnbaum, Henrik (1986). “Contrastive linguistics and language typology: the three-way
approach”. In D. Kastovsky & A. Szwedek (Eds.), Linguistics across Historical and
Geographical Boundaries: In Honour of Jacek Fisiak on the Occasion of his Fiftieth Birthday,
Vol 2. Descriptive, Contrastive and Applied Linguistics (pp. 1133–1146). Berlin and New
York: Mouton de Gruyter.
Blum-Kulka, Shoshana, Juliane House, & Gabriele Kasper (Eds.). (1989). Cross-cultural
Pragmatics: Requests and Apologies. Norwood, NJ: Ablex.
Borin, Lars (2002). “Alignment and tagging”. In L. Borin (Ed.), Parallel Corpora, Parallel Worlds:
Selected Papers from a Symposium on Parallel and Comparable Corpora at Uppsala University,
Sweden, 22–23 April, 1999 (pp. 207–218). Amsterdam and New York: Rodopi.
Botley, Simon P., Anthony M. McEnery, & Andrew Wilson (Eds.). (2000). Multilingual Corpora
in Teaching and Research. Amsterdam and New York: Rodopi.
Brdar, Mario (1996). “Raising cross-linguistic data: what can contrastive linguistics do for
typology?” Suvremena-Lingvistika, 22(1–2), 63–80.
Brown, Penelope & Stephen C. Levinson (1987). Politeness: Some Universals in Language Usage.
Butler, Christopher S. (2002). “Lexis and grammar across languages: functional and corpus-
oriented approaches”. In L. Iglesias Rábade, S. M. Doval Suárez, M. L. A. Gómez González et
al. (Eds.), Studies in Contrastive Linguistics. Proceedings of the 2nd International Contrastive
Linguistics Conference, Santiago, October 2001 (pp. 15–35). Santiago de Compostela:
Servicio de Publicacións, Universidade de Santiago de Compostela.
Calzolari, Nicoletta, John McNaught, & Antonio Zampolli (1996). EAGLES results now
available. See EAGLES website on resource list.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Cardey, Sylviane & Peter Greenfield (2002). “Computerised set expression dictionaries: design
and analysis”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 231–248).
Carlisle, Robert (1988). “The effect of markedness on epenthesis in Spanish/English
interlanguage phonology”. Issues and Developments in English and Applied Linguistics, 3,
15–23.
Celle, Agnès (2000). “De la deixis temporelle à la deixis discursive”. In J. Guillemin-Flescher
(Ed.), Linguistique contrastive et traduction, Tome 5 (pp. 87–111). Paris: Ophrys.
Chesterman, Andrew (1998). “Review of Krzeszowski: ‘Early contrastive studies in England”’.
Neuphilologische Mitteilungen, 99(1), 99–100.
Clyne, Michael G. (1987a). “Cultural differences in the organization of academic texts”. Journal
of Pragmatics, 11, 211–247.
Clyne, Michael G. (1987b). “Discourse structure and discourse expectations: implications for
Anglo-German academic communication”. In L. E. Smith (Ed.), Discourse across Cultures:
Strategies in World Englishes (pp. 3–83). New York: Prentice Hall.
Clyne, Michael, Jimmy Hoeks, & Heinz J. Kreutz (1988). “Cross-cultural responses to academic
discourse patterns”. Folia Linguistica, 22, 457–475.
Colleman, Timothy (2002). “The Contragram Verb Valency Dictionary of Dutch, French and
English”. In H. Gottlieb, J. E. Mogensen, & A. Zettersten (Eds.), Symposium on Lexicography
X. Proceedings of the Tenth International Symposium on Lexicography May 4–6, 2000 (pp.
63–76). Copenhagen: University of Copenhagen.
Comrie, Bernard (1986). “Contrastive linguistics and language typology”. In D. Kastovsky & A.
Szwedek (Eds.), Linguistics across Historical and Geographical Boundaries: In Honour of Jacek
Fisiak on the Occasion of his Fiftieth Birthday [2 Vols.] (pp. 1155–1164) [Vol. 2: Descriptive,
Contrastive and Applied Linguistics]. Berlin and New York: Mouton de Gruyter.
Connor, Ulla (1996). Contrastive Rhetoric: Cross-cultural Aspects of Second-Language Writing.
Cambridge and New York: Cambridge University Press.
Contreras, Heles (1987). “Small clauses in Spanish and English”. Natural Language and Linguistic
Theory, 5, 225–244.
Corness, Patrick (2002). “Multiconcord: a computer tool for cross-linguistic research”. In
B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 327–332). Amsterdam and
Danielsson, Pernilla & Daniel Ridings (1996). “PEDANT: parallel Texts in Göteborg”.
Unpublished paper. Department of Swedish, University of Gothenburg.
Dawes, Elizabeth (2003). “Towards a new dictionary of idioms”. In S. Granger, J. Lerot, &
S. Petch-Tyson (Eds.), Corpus-based Approaches to Contrastive Linguistics and Translation
Studies (pp. 207–220). Amsterdam and New York: Rodopi.
Devos, Filip (1995). “Contrastive grammar: tenets and criteria”. Papers and Studies in Contrastive
Linguistics, 30, 17–29.
Di Pietro, Robert J. (1968). “Contrastive Analysis and the notions of deep and surface grammar”.
In J. Alatis (Ed.), Contrastive Linguistics and its Pedagogical Implications. Report of the
Nineteenth Annual Round Table Meeting on Linguistics and Language Studies (pp. 65–80)
[Monograph Series on Languages and Linguistics 21]. Washington: Georgetown University
Press.
Di Pietro, Robert J. (1971). Language Structures in Contrast. Rowley, MA: Newbury House.
DuFon, Margaret A., Gabriele Kasper, Satomi Takatashi, & Naoko Yoshinaga (1994).
“Bibliography on linguistic politeness”. Journal of Pragmatics, 21, 527–578.
On contrastive linguistics 
Duszak, Anna (Ed.). (1997). Culture and Styles of Academic Discourse [Trends in Linguistics
Studies and Monographs 104]. Berlin and New York: Mouton de Gruyter.
Ebeling, Jarle (1999). Presentative Constructions in English and Norwegian. A Corpus-based
Contrastive Study. Oslo: Unipub Forlag.
Eliasson, Stig (1984). Theoretical Issues in Contrastive Phonology. Heidelberg: Julius Gross.
Fabricius-Hansen, Catherine (1996). “Informational density – a problem for translation and
translation theory”. Linguistics, 34, 521–565.
Fabricius-Hansen, Catherine & Kåre Solfjeld (1994). Deutsche und norwegische Sachprosa im
Vergleich. Ein Arbitsbericht. Arbeitsberichte des germanistischen Instituts der Universität
Oslo, No. 6. Oslo.
Fareh, Shehdeh (1995). “Foregrounding devices and their functions in Arabic and English
written discourses”. Papers and Studies in Contrastive Linguistics, 30, 119–130.
Fente, Rafael (1971). Estilística del Verbo en Inglés y Español. Madrid: S.G.E.L.
Filipović, Rudolf (1974). “The use of a corpus in contrastive studies”. In H. Raabe (Ed.), Trends
in kontrastiver Linguistik, Vol. 1 (pp. 51–66). Tübingen: Gunther Narr.
Filipović, Rudolf (1984). “What are the primary data for Contrastive Analysis”. In J. Fisiak (Ed.),
Contrastive Linguistics: Prospects and Problems (pp. 107–117). Berlin and New York: Mouton
de Gruyter.
Fillmore, Charles J. (1984). “Remarks on contrastive pragmatics”. In J. Fisiak (Ed.), Contrastive
Linguistics: Prospects and Problems (pp. 119–143). Berlin and New York: Mouton de Gruyter.
Firbas, Jan (1992). Functional Sentence Perspective in Written and Spoken Communication.
Fisiak, Jacek (1971). “The Poznan Polish-English Contrastive Project”. In R. Filipović (Ed.),
Zagreb Conference on English Contrastive Projects (pp. 87–96). Zagreb: Institut of
Linguistics.
Fisiak, Jacek et al. (1974). An Introductory English-Polish Grammar. Warsaw: PWN.
Fisiak, Jacek (Ed.). (1980). Theoretical Issues in Contrastive Linguistics. Amsterdam and
Fisiak, Jacek (Ed.). (1981a). Contrastive Linguistics and the Language Teacher. Oxford: Pergamon
Press.
Fisiak, Jacek (1981b). “Some introductory notes concerning contrastive linguistics”. In J. Fisiak
(Ed.), Contrastive Linguistics and the Language Teacher (pp. 1–11). Oxford: Pergamon Press.
Fisiak, Jacek (1983). “Present trends in contrastive linguistics”. In K. Sajavaara (Ed.), Cross-
language Analysis and Second Language Acquisition 2 (pp. 9–38). Jyväskylä: Department of
English, University of Jyväskylä.
Fisiak, Jacek (Ed.). (1984). Contrastive linguistics: Prospects and Problems. Berlin and New York:
Mouton de Gruyter.
Francis, Gill (1993). “A corpus-driven approach to grammar”. In M. Baker, G. Francis, & E.
Tognini-Bonelli (Eds.), Text and Technology. In Honour of John Sinclair (pp. 137–153).
Fries, Charles C. (1945). Teaching and Learning English as a Foreign Language. Ann Arbor:
University of Michigan Press.
Fries, Udo, Viviane Müller, & Peter Schneider (Eds.). (1997). From Ælfric to the New York Times.
Studies in English Corpus Linguistics. Amsterdam and Atlanta: Rodopi.
Garrudo-Carabias, Francisco (1996). “Los nuevos caminos del análisis contrastivo”. In M.
Martínez-Vázquez (Ed.), Gramática Contrastiva Inglés-Español (pp. 11–24). Huelva:
Servicio de Publicaciones de la Universidad de Huelva.
Gass, Susan (1989). “The development of L2 intuitions”. TESOL Quarterly, 17, 273–291.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Gellerstam, Martin (1996). “Translations as a source for cross-linguistic studies”. In K. Aijmer,

B. Altenberg, & M. Johansson (Eds.), Languages in Contrast. Papers from a Symposium on
Text-based Cross-linguistic Studies (pp. 53–62) [Lund Studies in English 88]. Lund: Lund
University Press.
Gil, David (1991). “Aristotle goes to Arizona and finds a language without ‘and”’. In D. Zaefferer
(Ed.), Semantic Universals and Universal Semantics (pp. 96–130). Berlin and New York:
Foris.
Givón, Talmy (Ed.). (1983). Topic Continuity in Discourse. A Cross-Language Study [Typological
Studies in Language 3]. Amsterdam and Philadelphia: John Benjamins.
Goddard, Cliff & Anna Wierzbicka (2002). Meaning and Universal Grammar: Theory and
Empirical Findings. Amsterdam and Philadelphia: John Benjamins.
Gómez-González, María de los Ángeles (2001). The Theme-Topic Interface: Evidence from
English. Amsterdam and Philadelphia: John Benjamins.
Granger, Sylviane (1996). “From CA to CIA and back: an integrated approach to computerized
bilingual learner corpora”. In K. Aijmer, B. Altenberg, & M. Johansson (Eds.), Languages in
Contrast. Papers from a Symposium on Text-based Cross-linguistic Studies (pp. 37–51) [Lund
Studies in English 88]. Lund: Lund University Press.
Granger, Sylviane (Ed.). (1998). Learner English on Computer. London: Longman.
Granger, Sylviane (2003). “The corpus-approach: a common way forward for Contrastive
Linguistics and Translation Studies?”. In S. Granger, J. Lerot, & S. Petch-Tyson (Eds.),
Corpus-based Approaches to Contrastive Linguistics and Translation Studies (pp. 17–31).
Amsterdam and New York: Rodopi.
Granger, Sylviane, Jacques Lerot, & Stephanie Petch-Tyson (Eds.). (2003). Corpus-based
Approaches to Contrastive Linguistics and Translation Studies. Amsterdam and New York:
Rodopi.
Guillemin-Flescher, Jacqueline (Ed.). (1992). Linguistique contrastive et traduction, Tome 3. Gap:
Ophrys.
Guillemin-Flescher, Jacqueline (Ed.). (2000). Linguistique contrastive et traduction, Tome 5. Paris:
Ophrys.
Haegeman, Liliane (Ed.). (1997). The New Comparative Syntax. Harlow: Addison Wesley
Longman.
Hansen, Klaus (1985). “Trends and problems in contrastive linguistics”. ZAA, 3, 117–132.
Hartmann, Reinhard R. K. (Ed.). (1977). Contrastive Analysis: Papers in German-English
Contrastive Applied Linguistics. Coleraine: The New University of Ulster.
Hartmann, Reinhard R. K. (1996). “Contrastive textology and corpus linguistics: on the value
of parallel texts”. Language Sciences, 18, 947–957.
Hasselgård, Hilde, Stig Johansson, Begljot Behrens, & Catherine Fabricius-Hansen (2002).
Information Structure in a Cross-Linguistic Perspective. Amsterdam and New York: Rodopi.
Haugen, Einar (1953). The Norwegian Language in America. A Study of Bilingual Behavior.
Philadelphia: University of Pennsylvania Press.
Hawkins, John A. (1985). A Comparative Typology of English and German: Unifying the Contrasts.
London: Croom Helm.
Hawkins, John A. (Ed.). (1988). Explaining Language Universals. Oxford: Basil Blackwell.
Hellinger, Marlis (1992). “Function and status change of Pidgin and Creole languages”. In U.
Ammon & M. Hellinger (Eds.), Status Change of Languages (pp. 264–281). Berlin and New
Hellinger, Marlis & Ulrich Ammon (Eds.). (1996). Contrastive Sociolinguistics. Berlin and New
On contrastive linguistics 
Hickey, Leo (Ed.). (2003). Politeness in Europe. Clevedon: Multilingual Matters.

Hill, Sam & William Bradford (Eds.). (1991). Bilingual Grammar of English-Spanish Syntax. New
York: University Press of America.
Hoey, Michael & Diane Houghton (1998). “Contrastive analysis and translation”. In M. Baker
(Ed.), Encyclopedia of Translation Studies (pp. 45–49). London: Routledge.
Ilson, Robert F. (1991). Assembling, Analysing and Using a Corpus of Authentic Language.
Budapest: Institutum Linguisticum Academiae Scientiarum Hungaricae.
Ivir, Vladimir (1983). “A translation-based model of contrastive analysis”. Jyväskylä Cross-
Language Studies, 9, 171–178.
Ivir, Vladimir (1987). “Functionalism in contrastive analysis and translation studies”. In R.
Dirven & V. Fried (Eds.), Functionalism in Linguistics (pp. 471–481). Amsterdam and
Jager, Sake & Herman Wekker (1997). “Aarts and Wekker hologrammed: contrastive grammar
in the computer age”. In J. Aarts, I. de Mönnink, & H. Wekker (Eds.), Studies in English
Language and Teaching (pp. 257–273). Amsterdam and Atlanta: Rodopi.
James, Carl (1980). Contrastive Analysis. London: Longman.
James, Carl (1992). “Awareness, consciousness and language contrast”. In C. Mair & M. Marcus
(Eds.), New Departures in Contrastive Linguistics. Proceedings of the Conference Held at
the Leopold-Franzens-University of Innsbruck, 10–12 May 1991 [2 Vols.] (pp. 183–197).
Innsbrucker Beiträge zur Kulturwissenschaft, Anglistische Reihe, Band 5. Institut für
Anglistik, Universität Innsbruck.
Janicki, Karol (1980). “Contrastive sociolinguistics”. In B. V. J. Fisiak (Ed.), Theoretical Issues in
Contrastive Linguistics (pp. 11–18). Amsterdam and Philadelphia: John Benjamins.
Janicki, Karol (1984). “Contrative sociolinguistics reconsidered”. Papers and Studies in Contras
tive Linguistics, 18, 15–29.
Janicki, Karol (1986). “Tertium comparationis in contrastive sociolinguistics”. In D. Kastovsky
& A. Szwedek (Eds.), Linguistics across Historical and Geographical Boundaries: In Honour
of Jacek Fisiak on the Occasion of his Fiftieth Birthday (2 Vols.) (pp. 1233–1246) [Vol. 2:
Descriptive, Contrastive and Applied Linguistics 2]. Berlin and New York: Mouton de
Gruyter.
Jaszczolt, Katarzyna M. (1995a). “Typology of contrastive studies: specialisation, progress and
applications”. Language Teaching, 28(1), 1–15.
Jaszczolt, Katarzyna M. (1995b). “Contrastive analysis”. In J. Verschueren, J.-O. Östman, &
J. Blommaert (Eds.), Handbook of Pragmatics: Manual (pp. 561–565). Amsterdam and
Jaszczolt, Katarzyna M. & Ken Turner (Eds.). (1996). Contrastive Semantics and Pragmatics (2
Vols.). Amsterdam and New York: Pergamon.
Johansson, Mats (1996). “Fronting in English and Swedish: a text-based contrastive analysis”.
In C. E. Percy, C. F. Meyer, & I. Lancashire (Eds.), Synchronic Corpus Linguistics: Papers
from the Sixteenth International Conference on English Language Research on Computerized
Corpora (ICAME 16) (pp. 29–39) [Language and Computers 16]. Amsterdam and Atlanta:
Rodopi.
Johansson, Mats (2001). “Clefts in contrast: a contrastive study of it clefts and wh clefts in
English and Swedish texts and translations”. Linguistics, 39, 547–582.
Johansson, Stig (1975). “The Uses of Error Analysis and Contrastive Analysis”. English Language
Teaching Journal, XXIX(3), 246–253.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Johansson, Stig (1997). “Using the English-Norwegian Parallel Corpus – a corpus for contrastive
analysis and translation studies”. In B. Lewandowska-Tomaszczyk & P. J. Melia (Eds.),
Practical Applications in Language Corpora (pp. 282–296). Łód´z: Łód´z University Press.
Johansson, Stig (1998). “On the role of corpora in cross-linguistic research”. In S. Johansson & S.
Oksefjell (Eds.), Corpora and Cross-linguistic Research (pp. 1–24). Amsterdam and Atlanta:
Rodopi.
Johansson, Stig (2002). “Towards a multilingual corpus for contrastive analysis and translation
studies”. In L. Borin (Ed.), Parallel Corpora, on Parallel and Comparable Parallel Worlds:
Selected papers from a Symposium Corpora at Uppsala University, Sweden, 22–23 April, 1999
(pp. 47–59). Amsterdam and New York: Rodopi.
Johansson, Stig (2003). “Contrastive linguistics and corpora”. In S. Granger et al. (Eds.), Corpus-
based Approaches to Contrastive Linguistics and Translation Studies (pp. 31–45). Amsterdam
and New York: Rodopi.
Johansson, Stig, Jarle Ebeling, & Knut Hofland (1996). “Coding and aligning the English-
Norwegian Parallel Corpus”. In K. Aijmer et al. (Eds.), Languages in Contrast. Papers from
a Symposium on Text-based Cross-linguistic Studies (pp. 87–112) [Lund Studies in English
88]. Lund: Lund University Press.
Johansson, Stig & Knut Hofland (1994). “Towards an English-Norwegian parallel corpus”. In U.
Fries, G. Tottie, & P. Schneider (Eds.), Creating and Using English Language Corpora (pp.
Johansson, Stig & Berit Løken (1997). “Some Norwegian discourse particles and their English
correspondences”. In C. Bache & A. Klinge (Eds.), Sounds, Structures and Senses. Essays
Presented to Niels Davidsen Nielsen on the Occasion of his Sixtieth Birthday (pp. 149–170).
Odense: Odense University Press.
Johansson, Stig & Signe Oksefjell (Eds.). (1998). Corpora and Cross-linguistic Research.
Amsterdam and Atlanta: Rodopi.
Johns, Timothy (1998). Multiconcord: the Lingua Multilingual Parallel Concordancer for
Windows.
Kalisz, Roman (1986). “More on pragmatic equivalence”. In D. Kastovsky & A. Szwedek (Eds.),
Linguistics across Historical and Geographical Boundaries: In Honour of Jacek Fisiak on the
Occasion of his Fiftieth Birthday (2 Vols.) (pp. 1247–1256) [Descriptive, Contrastive and
Applied Linguistics 2]. Berlin and New York: Mouton de Gruyter.
Kasper, Gabriele & Shoshana Blum-Kulka (Eds.). (1993) Interlanguage Pragmatics. New York:
Kastovsky, Dieter & Alexander Szwedek (Eds.). (1986). Linguistics across Historical and
Geographical Boundaries: In Honour of Jacek Fisiak on the Occasion of his Fiftieth Birthday (2
Vols.) [Descriptive, Contrastive and Applied Linguistics 2]. Berlin and New York: Mouton
de Gruyter.
Kefer, Michel & Johan van der Auwera (Eds.). (1992). Meaning and Grammar. Cross-linguistic
Perspectives. Berlin and New York: Mouton de Gruyter.
King, Philip (2003). “Parallel concordancing and its applications”. In S. Granger et al. (Eds.),
Corpus-based Approaches to Contrastive Linguistics and Translation Studies (pp. 157–168).
Amsterdam and New York: Rodopi.
Kittay, Eva F. & Adrienne Lehrer (1992). “Introduction”. In A. Lehrer & E. E. Kittay (Eds.),
Frames, Fields, and Contrasts. New Essays in Semantic and Lexical Organization (pp. 1–18).
Hillsdale, NJ: Lawrence Erlbaum.
On contrastive linguistics 
Kraig, Olivier (2002). “Translation alignment and lexical correspondences: a methodological

reflection”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 271–290). Amsterdam
and Philadelphia: John Benjamins.
Krzeszowski, Tomasz P. (1974). Contrastive Generative Grammar: Theoretical Foundations. Łódz: ´
Łódź University Press.
Krzeszowski, Tomasz P. (1984). “Tertium comparationis”. In J. Fisiak (Ed.), Contrastive
Linguistics: Prospects and Problems (pp. 301–312). Berlin and New York: Mouton de Gruyter.
Krzeszowski, Tomasz P. (1985). “The so-called ‘sign theory’ as the first method in contrastive
linguistics”. In U. Pieper & G. Stickel (Eds.), Studia Linguistica et Synchronica (pp. 485–501).
Berlin and New York: Mouton de Gruyter.
Krzeszowski, Tomasz P. (1989). “Towards a typology of contrastive studies”. In W. Oleksy (Ed.),
Contrastive Pragmatics (pp. 73–100). Amsterdam and Philadelphia: John Benjamins.
Krzeszowski, Tomasz P. (1990). Contrasting Languages: The Scope of Contrastive Linguistics.
Berlin and New York: Mouton de Gruyter.
Kühlwein, Wolfgang (1990). “Language processes in contrast”. In A. Michaux, A. Sinner, & J. J.
Webber (Eds.), English Studies 2: Essays in Honour of Marie-Thérèse Schroeder-Hartmann
(pp. 63–108). Publications du Centre Universitaire du Luxembourg.
Lado, Robert (1957). Linguistics across Cultures: Applied Linguistics for Language Teachers. Ann
Arbor: University of Michigan Press.
Lamiroy, Béatrice (1993). “La typologie des langues romanes et la dichotomie synchronie
diachronie”. In W. Raible & W. Oesterreicher (Eds.), Actes du XXe Congrès International
de Linguistique et Philologie Romanes [Tome III, Section 4] (pp. 211–221). Munich: Saur.
Laufer, Batia & Stig Eliasson (1993). “What causes avoidance in L2 learning: L1-L2 difference,
L1-L2 similarity, or L2 complexity?” Studies in Second Language Acquisition, 15, 35–48.
Laviosa, Sara (Ed.). (1998). The Corpus-based Approach. Special issue of META (Translators’
Journal). Montreal.
Levenston, Edward A. (1972). “Over-indulgence and under-representation: aspects of mother-
tongue interference”. In G. Nickel (Ed.), Papers in Contrastive Linguistics (pp. 115–121).
Lewandowska-Tomaszczyk, Barbara & Patrick J. Melia (Eds.). (1997). Practical Applications in
Language Corpora. Łód´z: Łód´z University Press.
Liebe-Harkort, Marie Louise (1985). “Contrastive pragmatics”. Papers and Studies in Contrastive
Littlewood, William (1983). “Contrastive pragmatics and the foreign learner’s personality”.
Applied Linguistics, 4(3), 200–206.
Mair, Christian & Marianne Hundt (Eds.). (2000). Corpus Linguistics and Linguistic Theory:
Papers from the Twentieth International Conference on English Language Research on
Computerized Corpora (ICAME 20), Freiburg im Breisgau, 1999. Amsterdam and Atlanta:
Rodopi.
Mair, Christian & Manfred Markus (Eds.). (1992). New Departures in Contrastive Linguistics.
Proceedings of the Conference held at the Leopold-Franzens-University of Innsbruck, 10–12
May 1991 (2 Vols.). Innsbrucker Beiträge zur Kulturwissenschaft, Anglistische Reihe, Band
5. Institut für Anglistik, Universität Innsbruck.
McEnery, Tony & Andrew Wilson (1996). Corpus Linguistics. Edinburgh: Edinburgh University
Press.
Maniez, François (2002). “The use of electronic corpora and lexical frequency data in solving
translation problems”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 291–206).
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Markkanen, Raija (1985). Cross-language Studies in Pragmatics. Jyväskylä Cross-Language

Studies. Department of English, University of Jyväskylä.
Marton, Waldemar (1979). “Some more remarks on the pedagogical use of contrastive studies”.
Papers and Studies in Contrastive Linguistics, 9, 35–45.
Mathesisus, Vilém (1975). A Functional Analysis of Present-day English on a General Linguistics
Basis. Transl. L. Dusková, Ed. J. Vachek. Prague: Academia.
Moder, Carol M. & Aida Martinovic-Zic (Eds.). (2004). Discourse Across Languages and Cultures.
Monroy, Rafael (2001). “Profiling the phonological processes shaping the fossilised IL of
adult Spanish learners of English as a foreign language: some theoretical implications”.
International Journal of English Studies, 1(1), 157–217.
Morciniec, Norbert (2001). “Problems of contrastive linguistics”. Studia Linguistica, 20, 125–136.
Moulton, William G. (1962). The Sounds of English and German. Chicago: Chicago University
Press.
Mukattash, Lewis (2001). “Some remarks on Arabic-English contrastive studies”. Poznan Studies
in Contemporary Linguistics, 37, 115–126.
Müller-Gotama, Franz (1994). Grammatical Relations. A Cross-Linguistic Perspective on their
Syntax and Semantics. Berlin and New York: Mouton de Gruyter.
Nickel, Gerhard (Ed.). (1971). Papers in Contrastive Linguistics. Cambridge: Cambridge
University Press.
Nickel, Gerhard (1972). Reader zur kontrastiven Linguistik. Frankfurt am Mein: Athenäum
Fisher Taschenbuch Verlag.
Nickel, Gerhard & Karl-Heinz Wagner (1979). “Contrastive linguistics and language teaching.
Studies in Descriptive Linguistics, 2, 31–53.
Odlin, Terence (1989). Language Transfer. Cross-linguistic Influence in Language Learning.
Oleksy, Wieslaw (1984). “Towards pragmatic contrastive analysis”. In J. Fisiak (Ed.), Contrastive
linguistics: Prospects and Problems (pp. 349–365). Berlin and New York: Mouton de Gruyter.
Oleksy, Wieslaw (1986). “Some recent approaches in Contrastive Studies”. In D. Kastovsky & A.
Szwedek (Eds.), Linguistics across Historical and Geographical Boundaries: In Honour of Jacek
Fisiak on the Occasion of his Fiftieth Birthday (2 Vols.) (pp. 1405–1420) [Vol. 2: Descriptive,
contrastive and applied linguistics]. Berlin and New York: Mouton de Gruyter.
Oleksy, Wieslaw (Ed.). (1989). Contrastive Pragmatics. Amsterdam and Philadelphia: John
Benjamins.
Olhstain, Elite (1983). “Sociocultural competence and language transfer. The case of apology”.
In S. Gass & L. Selinker (Eds.), Language Transfer in Language Learning (pp. 232–249).
Rowley, MA: Newbury House.
Paillard, Michel (2002). “From figures of speech to lexical units: an English-French contrastive
approach to hypallage and metonymy”. In B. Altenberg & S. Granger (Eds.), Lexis in
Contrast (pp. 175–188). Amsterdam and Philadelphia: John Benjamins.
Rauh, Gisa (Ed.). (1983). Essays on Deixis (pp. 229–275) [Tübinger Beiträge zur Linguistik 188].
Tübingen: Narr.
Rein, Kurt (1983). Einführung in die kontrastive Linguistik. Darmstadt: Wissenschaftliche
Buchgesellschaft.
Riegelhaupt, Florence (1989). “Contrastive linguistics revisited: the positive effects of L1 and L2
in second language acquisition”. Contrastes, 18/19 Dec., 11–23.
On contrastive linguistics 
Riley, Philip (1980). “Towards a contrastive pragmalinguistics”. In J. Fisiak (Ed.), Theoretical

Issues in Contrastive Linguistics (pp. 121–146). Amsterdam and Philadelphia: John
Benjamins.
Ringbom, Håkan (1994). “Contrastive analysis”. In R. E. Asher & J. M. Y. Simpson (Eds.), The
Encyclopedia of Language and Linguistics (pp. 737–742). Oxford: Pergamon Press.
Rudolf, Elisabeth (1996). Contrast. Adversative and Concessive Relations and their Expressions
in English, German, Spanish, Portuguese on Sentence and Text level. Berlin and New York:
Mouton de Gruyter.
Rusiecki, Jan (1976). “The development of contrastive linguistics”. Interlanguage Studies Bulletin,
1, 12–44.
Russ, Charles V. J. (Ed.). (1981). Contrastive Aspects of English and German. Heidelberg: Julius
Groos.
Sågvall Hein, Anna (2002). “The PLUG project: parallel corpora in Linköping, Uppsala,
Göteborg: Aims and achievements”. In L. Borin (Ed.), Parallel Corpora, on Parallel
and Comparable Parallel Worlds: Selected Papers from a Symposium Corpora at Uppsala
University, Sweden, 22–23 April, 1999 (pp. 61–78). Amsterdam and New York: Rodopi.
Sajavaara, Kari (1977). “Contrastive linguistics past and present and a communicative
approach”. In K. Sajavaara & J. Lehtonen (Eds.), Contrastive Papers. Jyväskylä Contrastive
Papers 4 (pp. 9–30). Department of English: University of Jyväskylä.
Sajavaara, Kari (1981). “Contrastive linguistics past and present and a communicative
approach”. In B. V. J. Fisiak (Ed.), Theoretical Issues in Contrastive Linguistics (pp. 33–56).
Sajavaara, Kari (1996). “New challenges for contrastive linguistics”. In K. Aijmer et al. (Eds.),
Languages in Contrast. Papers from a Symposium on Text-based Cross-linguistic Studies (pp.
17–36) [Lund Studies in English 88]. Lund: Lund University Press.
Sajavaara, Kari & Jaako Lehtonen (Eds.). (1980). Papers in Discourse and Contrastive Discourse
Analysis [Jyväskylä Contrastive Studies 5]. Department of English, University of Jyväskylä.
Salkie, Raphael (1997). “Naturalness and contrastive linguistics”. In B. Lewandowska-
Tomaszczyk & P. J. Melia (Eds.), Practical Applications in Language Corpora (pp. 297–312).
Łód´z: Łód´z University Press.
Salkie, Raphael (2002). “Two types of translation equivalence”. In B. Altenberg & S. Granger
(Eds.), Lexis in Contrast (pp. 51–72). Amsterdam and Philadelphia: John Benjamins.
Selinker, Larry (1992). Rediscovering Interlanguage. London: Longman.
Sifianou, María (1992). Politeness phenomena in England and Greece. Oxford: Clarendon Press.
Simon-Vandenbergen, Anne-Marie (1998). “I think and its Dutch equivalents in parliamentary
debates”. In S. Johansson & S. Oksefjell (Eds.), Corpora and Cross-linguistic Research (pp.
Simon-Vandenbergen, Anne-Marie (1999). “The Semantics of English find in contrast with
Dutch vinden”. In G. Tops, B. Devriendt, & S. Geukens (Eds.), Thinking English Grammar:
To Honour Xavier Dekeyser, Professor Emeritus (pp. 409–424) [Orbis Supplementa series,
Vol. 12]. Leuven and Paris: Peeters.
Simon-Vandenbergen, Anne-Marie, Bart Defrancq, & Kristin Davidse (Eds.). (1998). Objects:
Contrastive and Typological Issues [Languages in Contrast 1, 2]. Amsterdam and
Simon-Vandenbergen, Anne-Marie & Dirk Noël (1997). “English as, French comme and Dutch
als: conjunctions, prepositions, or what?” In J. Aarts, I. de Mönnink, & H. Wekker (Eds.),
Studies in English Language Research and Teaching (pp. 123–140). Berlin and New York:
Mouton de Gruyter.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Simon-Vandenbergen, Anne-Marie, Johan Taeldeman, & Dominique Willems (Eds.). (1996).

Aspects of Contrastive Verb Valency [Studia Germanica Gandensia 40]. Gent: University of
Gent.
Sinclair, John M. (1998). “The lexical item”. In E. Weigand (Ed.), Contrastive Lexical Semantics
Sinclair, John M. (forthcoming). “Corpus and text – basic principles”. In M. Wynne (Ed.),
Developing Linguistic Corpora: A Guide to Good Practice. Oxford: Oxbow Books.
Sinclair, John, Jonathan Payne, & Chantal Pérez Hernández (Eds.). (1996). Corpus to Corpus:
A Study of Translation Equivalence. Special issue of International Journal of Lexicography, 9,
179–276.
Smith, Larry E. (Ed.). (1987). Discourse across Cultures: Strategies in World Englishes. New York:
Prentice Hall.
Sridhar, Shikaripur N. (1981). “Contrastive analysis, error analysis and interlanguage: three
phases of one goal”. In J. Fisiak (Ed.), Contrastive Linguistics and the Language Teacher (pp.
9–38). Oxford: Pergamon Press.
Steffens, Petra (Ed.). (1995). Machine Translation and the Lexicon. Third International EAMT
Workshop, Heidelberg, April 26–28 1993. Berlin and New York: Springer.
Steiner, Erich & Colin Yallop (Eds.). (2001). Exploring Translation and Multilingual Text
Production. Berlin and New York: Mouton de Gruyter.
Stockwell, Robert P. & J. Donald Bowen (1965). The Sounds of English and Spanish. Chicago:
University of Chicago Press.
Stockwell, Robert P., J. Donald Bowen, & John W. Martin (1965). The Grammatical Structures of
English and Spanish. Chicago: Chicago University Press.
Strangert, Eva (1981). Rhythmic Patterns of Swedish in a Cross-linguistic Perspective. Department
of Phonetics: University of Umeå.
Svartvik, Jan (Ed.). (1992). Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82,
Stockholm, 4–8 August 1991. Berlin and New York: Mouton de Gruyter.
Sveen, Andreas (2000/2001). “Comparative Scandinavian syntax”. Languages in Contrast, 3(1),
125–148.
Svensson, Mikael (2000). “Sentence openings and textual progression in English and Swedish”.
In C. Mair & M. Hundt (Eds.), Corpus Linguistics and Linguistic Theory: Papers from the
Twentieth International Conference on English Language Research on Computerized Corpora
(ICAME 20), Freiburg im Breisgau, 1999 (pp. 355–370). Amsterdam and Atlanta: Rodopi.
Swales, John M. (1990). Genre Analysis: English in Academic and Research Settings. Cambridge:
Tannen, Deborah (1984). “The pragmatics of cross-cultural communication”. Applied
Tarasova, Yelena (1993). “Contrastive linguistics and the linguistic field theory”. Papers and
Studies in Contrastive Linguistics, 27, 67–77.
Teich, Elke (1999). “System-oriented and text-oriented comparative linguistic research: cross-
linguistic variation in translation”. Languages in Contrast, 2(2), 187–210.
Teich, Elke (2003). Cross-linguistic Variation in System and Text. A Methodology for the
Investigation of Translations and Comparable Texts. Berlin and New York: Mouton de
Gruyter.
Teubert, Wolfgang (1996). “Comparable or parallel corpora?” In J. Sinclair et al. (Eds.), Corpus
to Corpus: A Study of Translation Equivalence. Special issue of International Journal of
Lexicography, 9, 238–264.
On contrastive linguistics 
Teubert, Wolfgang (2002). “The role of parallel corpora in translation and multilingual
lexicography”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 189–214).
Tognini-Bonelli, Elena (2001). Corpus Linguistics at Work. Amsterdam and Philadelphia: John
Benjamins.
Tognini-Bonelli, Elena (2002). “Functionally complete units of meaning across English-Italian:
towards a corpus-driven approach”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast
(pp. 73–96). Amsterdam and Philadelphia: John Benjamins
Tops, Guy A. J., Betty Devriendt, & Steven Geukens (Eds.). (1999). Thinking English Grammar.
Leuven: Peeters.
Towell, Richard & Roger Hawkins (1994). Approaches to Second Language Acquisition. Clevedon,
Avon: Multilingual Matters.
Trager, George L. (1949). The Field of Contrastive Linguistics. Studies in Linguistics. Oklahoma:
Battenburg Press.
Turner, Ken (1996). “The principal principles of pragmatic inference: politeness”. Language
Teaching, 29, 1–13.
Twaddell, W. Freeman (1968). “The durability of ‘contrastive studies”’. In J. E. Alatis (Ed.),
Contrastive Linguistics and Its Pedagogical Implications: Report of the Nineteenth Annual
Round Table Meeting on Linguistics and Language Studies (pp. 195–201). Washington, DC:
Georgetown University Press.
van Benthem, Johan & Alice ter Meulen (Eds.). (1996). Handbook of Logic and Language.
Amsterdam: Elsevier Science B. V.
Van Buren, Paul (1974). “Contrastive analysis”. In J. P. Allen & S. Pit Corder (Eds.), Techniques
in Applied Linguistics (pp. 279–312). Oxford: Oxford University Press.
Van Buren, Paul (1976). “Review: Krzeszowski, T., Contrastive generative grammar.”
Interlanguage Studies Bulletin, 2/3, 250–329.
Veronis, Jean (Ed.). (2000). Parallel Text Processing: Alignment and Use of Translation Corpora.
Dordrecht, Boston and London: Kluwer Academic Publishers.
Viberg, Åke (1993). “Crosslinguistic perspectives on lexical organization and lexical
progression”. In K. Hyltenstam & A. Viberg (Eds.), Progression and Regression in Language.
Sociocultural, Neuropsychological and Linguistic Perspectives (pp. 340–385). Cambridge:
Viberg, Åke (1996a). “Cross-linguistic lexicology. The case of English go and Swedish gå”. In K.
Aijmer et al. (Eds.), Languages in Contrast. Papers from a Symposium on Text-based Cross-
linguistic Studies (pp. 151–182) [Lund Studies in English 88]. Lund: Lund University Press.
Viberg, Åke (1996b). “The meanings of Swedish dra ‘pull’: a case study of lexical polysemy”. In
M. Gellerstam, J. Jäborg, S.-G. Malmgren, K. Norén, L. Rogström, & C. Röjder Papmehl
(Eds.), Papers submitted to the Seventh EURALEX International Congress on Lexicography
in Göteborg Sweden (pp. 293–308). Göteborg: Department of Swedish, University of
Göteborg.
Viberg, Åke (1998). “Contrasts in polysemy and differentiation. Running and putting in English
and Swedish”. In S. Johansson & S. Oksefjell (Eds.), Corpora and Cross-linguistic Research
(pp. 343–376). Amsterdam and Atlanta: Rodopi.
Viberg, Åke (1999). “Polysemy and differentiation in the lexicon. Verbs of physical contact in
Swedish”. In Cognitive Semantics. Meaning and Cognition (pp. 87–129). Amsterdam: John
Benjamins.
 María de los Ángeles Gómez-González and Susana M. Doval-Suárez
Viberg, Åke (2002). “Polysemy and disambiguation cues across languages: the case of Swedish
få and English get”. In B. Altenberg & S. Granger (Eds.), Lexis in Contrast (pp. 119–150).
Vossen, Piek (Ed.). (1998). EuroWordNet: A Multilingual Database with Lexical Semantic
Networks. Dordrecht: Kluwer Academic.
Weigand, Edda (Ed.). (1998). Contrastive Lexical Semantics. Amsterdam and Philadelphia: John
Benjamins.
Weinreich, Uriel (1953). Languages in Contact. New York: Linguistic Circle of New York.
Wierzbicka, Anna (1980). Lingua Mentalis: The Semantics of Natural Language. Sydney:
Academic Press.
Wierzbicka, Anna (1991). Cross-Cultural Pragmatics: The Semantics of Human Interaction. Berlin
Wierzbicka, Anna (1992). Semantics, Culture and Cognition: Universal Human Concepts in
Culture-Specific Configurations. Oxford: Oxford University Press.
Resource list
CoLLaTE project homepage:

http://bank.rug.ac.be/contragram/collate.html
Complutense representative (Angela Downing) in Research Project Pragmatic Markers in
Contrast: http://www.ucm.es/info/fing1/downing.html
Complutense Research Group on Automatic Text Generation and Discourse Processing:
http://www.ucm.es/info/atg/
CONTRAGRAM homepage:
http://bank.rug.ac.be/contragram/
DELIC research group:
http://www.up.univ-mrs.fr/delic/publis-1995-1999.html
EAGLES results:
http://linguistlist.org/issues/7/7-1655.html
English-Spanish Contrastive Grammar Research Group homepage (Univ. of Huelva):
http://www.uhu.es/ogi/Memoria99/FILINGLESA.htm
ENPC project homepage:
http://www.hf.uio.no/iba/prosjekt
Joaquim Llisterri homepage:
http://liceu.uab.es/∼joaquim/home.html
LINGUA:
http://artsweb.bham.ac.uk/pking/multiconc/l_text.htm
Manuel Barbera’s Bibliography and Resources on Corpus Linguistics:
http://www.bmanuel.org/index.html
Michael Barlow homepage:
http://www.athel.com/mb.html
Natural Semantic Metalanguage:
http://www.une.edu.au/arts/LCL/disciplines/linguistics/nsmpage1.htm
SCIMITAR research group:
http://www.usc.es/scimitar/
On contrastive linguistics 
SEU Bibliography (Survey of English Usage Archives):

http://www.ucl.ac.uk/english-usage/archives/seu-biblio.htm
SPRIK project:
http://www.hf.uio.no/german/sprik/english
The ESPC homepage (English-Swedish Parallel Corpus):
http://www.englund.lu.se/research/corpus/corpus/espc.html
The Gateway to Corpus Linguistics bibliography:
http://www.corpus-linguistics.de/
The present book
Christopher S. Butler, María de los Ángeles Gómez-González

The papers collected together in the rest of this volume are grouped into five sections
according to the thematic area to which they belong.
. Form and function in a cognitive perspective
The book begins with two papers which strongly reflect the importance of cognitive
phenomena in the study of form and function in language.
Wallace Chafe’s paper tackles head-on the important but difficult question of the
relationship between grammar and thought, his treatment of this issue being funda
mentally functional in nature. His starting point, namely that thoughts are filtered
through the particular semantic resources permitted by the language, and that seman
tic structures are then converted to grammatical structures which are then symbolised
by sounds, is uncontroversial. What is essentially functional about Chafe’s position is
the contention that for an understanding of why grammar is as it is, linguists need
to pay more rather than less attention to grammar per se and more to semantics
and its relationship with thought, as well as to language change. Chafe postulates
four linked stages in the conversion of thoughts to semantic structures. Firstly, the
speaker must select which thoughts s/he wishes to verbalise. Secondly, these thoughts
must be categorised. Most thoughts consist of particular ideas relating to particular
people, things, events, times and places, and these ideas must be recast as instances
of more general and familiar categories. Thirdly, the speaker must select a complex
orientation consisting of values for time, place, epistemological factors and attitude.
Finally, categorisations and orientations must be combined. In explaining why the re
sulting semantic structure must be converted into a grammatical structure on its way
to representation in sound, Chafe points out that as languages change, the relationship
between the use of expressions and the thoughts they convey may become less direct,
through the historical processes of lexicalisation, relating to ideas, and grammaticali
sation, concerned with orientations. Chafe’s arguments are illustrated throughout by
reference to an extended piece of authentic conversational interaction.
 C. S. Butler, M. L. Á. Gómez-González and S. M. Doval-Suárez
The paper by Montserrat Martínez Vázquez discusses a specific semantico

grammatical area, that of verbs of communication, as seen from an English-Spanish
contrastive perspective. The author first presents an overall picture of linguistic ac
tions: the communicative situation itself; its participants (the person who says some
thing, the representation of what is said, and an optional recipient of the message);
and the process of communication, conceptualised as transfer of information. Here,
Martínez Vázquez introduces a cognitivist perspective, in that she points out that this
information transfer can be seen as involving the transmission of a ‘package’ of in
formation from sender to receiver, thus invoking the ‘conduit’ metaphor. The main
part of the paper is devoted to a discussion of the various ways in which linguis
tic actions can be reported in Spanish and English, covering not only verba dicendi
as such, but also the use of verbs of discharge (e.g. hurl (insults), spread (rumours),
lanzar (un mensaje); also linguistic uses of verbs such as cough/toser, spit/escupir), cog
nitive verbs (e.g. reason/razonar, speculate/especular), sound emission verbs which can
be extended by a metonymic process to convey manner of speaking, and which can
be subdivided into sounds emitted by human beings (e.g. murmur/murmurar, whis
per/susurrar, those emitted by animals (e.g. roar/rugir), and by inanimate objects (e.g.
thunder/tronar); instrumental verbs (e.g. (tele)phone/telefonear); and finally verbs of
gesture (nod (approval, agreement), smile (appreciation), and much more limitedly
verbs such as cabezear (nod), sonreír (smile) in Spanish). The analysis throughout
is of a cognitivist/constructivist nature, emphasising the metonymic processes which
give rise to the various uses. The paper is richly illustrated with examples from corpus
materials in both English and Spanish.
. Information structure
The second section contains four papers, each dealing with some aspect of the way in
which a speaker or writer chooses to present the information contained in a message.
Central to this area are notions such as topic, focus, theme and rheme, which have been
much studied within functional approaches to language. We may also include under
the broad heading of information structure other areas such as voice alternations and
the choice of more or less explicit realisations of a given conceptual content. The last
two papers are also contrastive in nature, showing how particular linguistic strategies
are put to use in English and Spanish discourse.
Lachlan Mackenzie’s paper is concerned with degrees of explicitness of realisa
tion, seen from the perspective of Incremental Functional Grammar (IFG). As we
have seen, this is a variant of Functional Discourse Grammar (FDG), which is in turn
derived from Dikkian Functional Grammar. IFG aims to demonstrate how the on
line mental construction of utterances affects their grammatical structure, and stresses
the importance, not only of fully formed clauses and their combinations, but also of
clause fragments and holophrastic expressions which are by their very nature more
condensed, less explicit forms of realisation. Mackenzie’s topic in this paper is the re
The present book 
lationship, in the language of football commentary, between the choice of expression

form (sentential, clausal, phrasal, holophrastic) and the time pressure the commen
tator is under at a particular point in the reporting of the game. Having sketched the
basic principles of FG and FDG, the author shows how IFG models utterance pro
duction, taking as its basis the minimal, holophrastic expression, which must always
bear focal information, and seeing this as the starting point for the elaboration of
more complex structures. Production is modelled in terms of an interpersonal level
of grammatical structure which, like the phonological level, is seen as dynamic in na
ture, operating in real time, while the representational and structural levels of the
grammar act as constraints on the operation of the two dynamic levels. Mackenzie
illustrates these principles from a commentary on televised highlights of a football
match, and shows that there is indeed a correlation between degree of time pressure
and choice of realisation, medium and high pressure leading largely to holophrastic
and phrasal expression, while low time pressure allows the production of more clausal
and sentential forms.
Michael Cummings’ paper is concerned with the role of information structure
in the organisation of written texts, analysed through the categories of Theme and
Rheme as defined within Systemic Functional Linguistics. Having summarised the SFL
account of Theme/Rheme, methods of development within a text, and the ‘point’, or
informational goal, of a text, Cummings presents a quantitative method for assessing
the differences between method of development, as signalled in the Themes, and point,
as realised mainly in the final clause element, or N-Rheme. He then applies this method
to the study of short narrative and expository texts, in order to contrast their use of
methods of development and point, and to explain stylistic differences between them.
Finally, he examines two texts which, for very good reasons, do not conform to the
norms for distribution of features within Themes and N-Rhemes.
In the paper by Gómez-González & Gonzálvez-García the area under focus is
it-clefts in English and their counterparts in Spanish. The account given is strongly
functional in its orientation, seeking to explain the behaviour of these clefts in terms of
a number of competing motivations of a structural, semantic/pragmatic and discour
sal nature. The study is based on material from corpora of both languages. The authors
first present an account of the syntactic properties of the clefts in their corpus sample,
concentrating on the features of the element in focus. They show that the quantitative
distribution of the various types, also the (non)-acceptability of particular instances
of clefts, can be explained in terms of the specific, identifying function of it-clefts and
their Spanish counterparts. Gómez-González & Gonzálvez-García go on to analyse the
semantic properties of these constructions, characterising them as identifying, as en
tailing an existential presupposition carried by the free relative element, and as having
an implicature of exhaustiveness. Finally, they examine the discourse properties of it-
clefts, under three main headings. Firstly, these constructions are shown to have an
inherent thematic flexibility which gives them great versatility, for example in specify
ing the current discourse topic, reactivating a lapsed topic, correcting an old topic, or
realising emphasis. Secondly, the material in focus in an it-cleft is usually perceived as
 C. S. Butler, M. L. Á. Gómez-González and S. M. Doval-Suárez
newsworthy, having what the authors call a ‘newness-orientation’. Thirdly, these clefts
often have a strongly interpersonal flavour, in encoding various aspects of the speaker
or writer’s attitude towards the proposition.
In the final paper of this section, Maite Toboada tackles one of the fundamental
problems of discourse and text analysis: the factors which influence the choice of a
particular type of anaphoric term at a particular point in a discourse. The theoreti
cal framework within which the analysis is conducted, known as Centering Theory, is
functional in its orientation inasmuch as it aims to relate discourse structure to the
intentions, purposes and attentional properties of the participants in the discourse.
For each utterance, the theory establishes a list of entities which have been mentioned
or evoked (the ‘forward-looking center list’) ranked for salience, usually in terms of
grammatical function. The top member of this list is the ‘preferred center’, and some
other member of the list, the highest ranking entity which has also been mentioned in
the previous utterance, is the ‘backward-looking center’. Relations among these various
types of center define classes of ‘transition’, representing the different ways in which the
discourse may proceed. Taboada illustrates these concepts and discusses their applica
tion to spoken language, and to Spanish as well as English. She then presents the results
of a contrastive study of English and Spanish whose aim was to examine the choices
of backward-looking center in a small corpus of telephone conversations in the two
languages. The study confirms, for both languages, the prediction that certain partic
ular types of transition will be more common than others. For each type of transition
found, quantitative patterns of anaphor realisations are documented and interpreted.
. Collocations and formulaic language
The two papers in this section take up a theme which has been particularly produc
tive in recent years, that of the co-occurrence relationships between words, which may
vary from purely statistical collocational associations, through varying degrees of ‘for
mulaicity’, to fixed phrases. Both contributions adopt a contrastive, cross-linguistic
perspective, and both pay attention to the interactional and cognitive functions of
formulaic sequences.
Christopher Butler’s paper examines formulaic language from two distinct but
complementary perspectives, that of the corpus linguist and that of the psycholinguist.
Butler first reviews the literature on recurrent continuous sequences of words in cor
pora of English and Spanish, pointing out that such sequences are common (especially
in spoken language), that they can be classified both structurally and functionally,
that they often consist of a core with optional extensions, and that they frequently
overlap. Functionally, recurrent sequences in English and Spanish are shown to have
many basic similarities, with some differences. Recurrent discontinuous sequences, or
‘collocational frameworks’, in the two languages are also discussed. The scope is then
widened to include more flexible syntagmatic patterns in English, Spanish and Italian,
and Sinclair’s concept of an ‘extended unit of meaning’, with preferences for partic
The present book 
ular semantic areas and often an additional ‘semantic prosody’, is illustrated. Overall
conclusions from corpus work are drawn in terms of Sinclair’s ‘idiom principle’. But
ler then deals more briefly with the psycholinguistically-oriented model of formulaic
language proposed by Wray on the basis of detailed studies of formulaic sequences in
the language of adults, children learning their native language, learners of a second
or foreign language, and aphasics. Finally, some implications of both perspectives for
language teaching and learning are outlined.
The paper by László Komlósi & Elisabeth Knipf presents a cross-linguistic study of
variable-sized lexical units. The authors begin with a distinction between rule-driven
(regular, predictable, highly compositional) and frequency-driven (gestalt-like) lexical
structures. They go on to discuss the interaction between grammatical meaning, lexical
meaning and lexical form, the contribution of semantics and pragmatics to meaning,
and the concepts of meaning extension and conceptual integration, all in relation to
lexical units of variable size. They relate these areas to the psycholinguistic concept of
entrenchment. Having proposed various levels of compositionality, they suggest that
three features jointly characterise formulaic expressions and help in understanding
their role in the grammar: the level of compositionality, the degree of productivity and
the type of processing involved. By way of illustration, the paper ends with a cross-
linguistic analysis and classification of expressions in English, German, Russian and
Hungarian, all of which involve some type of iteration (e.g. boys will be boys, a hush
hush project, by hook or by crook).
. Language learning
This section brings together three papers on the learning of languages, all written from
a cross-linguistic perspective.
Paul Meara begins by describing the three-dimensional model of vocabulary skills
developed by him and his colleagues at the University of Wales Swansea, and the tests
which have been developed to measure skills on each of these dimensions. X_Lex is
a computer program which asks learners to say whether or not they know a particu
lar word, the words presented being taken from different frequency bands, and some
being non-existent, though possible words of the language. This program allows accu
rate estimates of how many words learners know in particular frequency bands. The
V_Links program measures vocabulary organisation by asking learners to recognise
links between words. Q_Lex assesses access to the vocabulary by measuring reaction
times in a task in which learners have to identify words hidden in a longer string of
letters. Meara then discusses the problems of extending these vocabulary tests from
English, for which they were first developed, to a range of other languages. These prob
lems are not only practical (e.g. the non-availability of word frequency data for some
languages), but also raise the issue of the comparability of data from different lan
guages. For instance, Meara points to various properties of languages which will affect
the normal size of the working vocabulary (e.g. compounding, the existence of paral
 C. S. Butler, M. L. Á. Gómez-González and S. M. Doval-Suárez
lel families of words such as those of native and Classical origin in English). Similarly,
there are problems with tests of vocabulary organisation: for instance, many English
word forms can act as more than one part of speech, and so are likely to generate
more associations than Spanish words, which are usually unambiguous as to part of
speech. Finally, the extension to other languages of tests of access to vocabulary may be
complicated by differences in the way speakers of different languages recognise words.
Francisco Gutiérrez Díez’s paper examines some rhythmic and timing aspects of
the Spanish produced by a group of English learners, as compared with control data
from native English and Spanish speakers producing their own language. Comparison
of native English and Spanish shows statistically significant differences in the dura
tion of stressed syllables and of the rhythmic foot, ictus and remiss elements, as well
as in the degree of isochrony of syllables. The production of English learners of Span
ish reveals differences in tempo for feet, remiss elements and unstressed syllables, as
compared with native Spanish speakers. There are also differences in the lengths of
stressed syllables, and the ratio of duration in stressed and unstressed syllables. The
author relates these results to hypotheses which have been proposed in the literature
on timing.
Rafael Monroy’s paper presents an empirical study of the role of intonation in
signalling attitude in yes/no and wh questions in Spanish and English. A group of post-
intermediate students of English in a Spanish university listened to a set of questions in
English, representing a range of tones, and were asked to rate the intonation patterns
for familiarity, and to mark the most prominent (focused) word in each question. The
data are analysed quantitatively, and conclusions are drawn regarding perceptions of
similarity and difference between Spanish and English intonation systems.
. Discourse and culture
The papers in this section are all concerned with aspects of discourse/text. The first two
fall under the heading of comparative discourse-pragmatic studies, are both concerned
specifically with narrative discourse, and both adopt a broadly functional stance. The
final paper adopts a cross-cultural perspective.
Takashi & Wilkerson present an empirical study of the retelling of the Snow White
story by groups of male and female American and Japanese informants. The anal
ysis is rooted in ‘Place of Negotiation Theory’, according to which communication
involves the exploitation of a meaning-negotiating space with three dimensions: cog
nitive (centred on propositional meaning), emotional (foregrounding the speaker or
writer’s emotional attitudes) and interactional (involving the creation and manage
ment of an interactional social atmosphere). Takashi & Wilkerson study quantitatively
the expression of a number of features on each dimension: choice of characters and
relationships among them, narrative events focused upon and narrative voice (all in
‘cognitive place’); terms used in reference, attributes of characters, use of descriptive
phrases, and the attitude of the narrator towards the story and its retelling (‘emo
The present book 
tional place’); the inclusion of dialogue, use of fairy tale conventions, and indications
of empathy with the listener (‘interactional place’). On each of these dimensions, males
are contrasted with females and Americans with Japanese. The results of the analysis
do not support traditionally postulated distinctions between men’s and women’s dis
course; nevertheless, some interesting differences are found with respect to both the
gender and nationality/cultural variables. The analysis is clearly functional in the sense
that it attempts to explain linguistic features of discourse in terms of the interplay of a
set of social and cultural factors.
Guerrero Medina’s paper, on cardinal transitivity in English and Spanish, is
functionally oriented in a rather different sense: here, the aim is to link semantico
grammatical properties of transitivity to discourse-pragmatic function. More specifi
cally, Guerrero Medina first refines the account of cardinal transitivity in earlier work
by Hopper & Thompson, and then examines the correlation, in a corpus of narra
tive texts in English and their translations into Spanish, between a set of semantico
grammatical features (number of participants, agentivity, object affectedness, object
individuation, telicity, boundedness) and the discourse-pragmatic dimension of fore-
grounding/backgrounding in discourse. The results demonstrate that the manifesta
tions of high and low transitivity do not always reflect the foreground/background
distinction, and that the correlation between subordination and backgrounding is less
straightforward than has been postulated in the literature.
Finally, Paloma Tejada Caller’s paper analyses the language of a corpus of 19th
century scholarly writing in Spanish, from the viewpoint of images of English and
of Englishness. In particular, she addresses issues such as the reasons for allusions to
English, the events and people focused on, the conceptualisation of English in a non-
British culture, the ideological implications of the images created, and the possible role
of English in the creation of a Spanish cultural identity. The analysis reveals an image
of English as a civilised, literary language reflecting elements of education, tolerance
and dynamism in the people who speak it.
. Concluding remarks
The foregoing brief introduction to the papers in this volume has shown that many of
them bring together the two themes which characterise this collection: the contrastive
study of parallel phenomena in two or more languages, and a functional approach to
the analysis of language and its use. Furthermore, the majority of the studies reported
here are empirical in nature, many making use of corpus materials in the language(s)
under investigation. It is our contention that this combination of features is a very
powerful one, able to bring to light hitherto unexplored or under-explored aspects of
language in use. We hope and believe that this collection of papers has helped to vali
date this claim, and constitutes a modest advance in the fascinating task of unravelling
the complex dynamics of language use.
P II
Form and function in a cognitive perspective

The relation of grammar to thought
Wallace Chafe
Language associates thoughts with sounds, but those phenomena are so different
in nature that several kinds of adjustments are necessary before such an
association is possible. Among those adjustments are selecting what will be
verbalized from the realm of thought; categorizing the elements selected;
orienting them in time, space, epistemology, and other dimensions of thought;
and combining those choices within established patterns. Thoughts are thus
organized into semantic structures, which would be amenable to representation
by sounds, were it not for the historical processes of lexicalization and
grammaticalization. Those processes modify semantic structures to produce
grammatical structures, which constitute the input to phonological
representations. This way of viewing language is illustrated with an extended
example from an English conversation.
. Introduction
Whatever else people may do when they talk, a very basic thing they do is to asso
ciate thoughts with sounds. As they produce language they are experiencing thoughts,
and language somehow enables them to use sounds to represent those thoughts. The
sounds then pass through the air and strike the ears of listeners, who are able to in
terpret them in terms of thoughts that bear at least some resemblance to the thoughts
of the speaker. The thoughts of listeners will never be identical with the thoughts of
speakers because two minds are never the same, and new thoughts that enter anyone’s
mind will always be interpreted in terms of thoughts that are already there. But at least
language allows separate individuals to have a relatively rich access to what others are
thinking, whatever use they may make of such knowledge.
Much of linguistics is based on the assumption that between the thoughts and
the sounds there is an intermediate stage that might be called grammar or syntax,
and this is the stage on which many linguists focus their attention. I would like to
suggest a different perspective, putting grammar in its place as just one stage on the
way from thoughts to sounds (Chafe 2002). Thoughts and sounds differ in many ways,
some of which will be explored here. Between the two there is what I like to call a
basic unconformity. Because of their intrinsic natures there cannot possibly be a direct
correspondence between phenomena of the one kind and phenomena of the other.
 Wallace Chafe
Thoughts, therefore, must be filtered or adjusted before any association with sounds
can be achieved. But what are those adjustments, and just why are they necessary?
In the most general terms, thoughts are first organized into what may be called se
mantic structures. Each language accomplishes that in its own way, employing its own
unique semantic resources. Semantic structures are then converted into grammatical
structures, again in accordance with the unique resources of each language. It is likely
that the differences between languages are greatest in the grammatical area, and thus
in a sense linguists have been looking for universals in the wrong place. There may
be more universals of thought than of semantics, and more semantic universals than
can be found in grammar. However that may be, it is the grammatical structures that
proceed to be symbolized by sounds.
It would be far too simple, however, to suppose that we first have thoughts that
are independent of language, that we then have semantic structures, then grammatical
structures, then phonological structures and sounds. The overlap between adjacent
stages can be considerable, and it is important to take them into account. Thoughts are
already shaped to a significant degree by the semantic resources of a language, and, as
will be emphasized below, there is a considerable overlap between semantic structures
and grammatical structures. None of these stages – thoughts, semantics, grammar –
exists in isolation. In the final analysis we would like to know what is happening in
the brain, and the brain does not process things in fully isolated boxes. It is useful
nonetheless to be able to discuss thoughts, semantics, and grammar each on its own
terms, because each has its own properties and its own reasons for existence. Those
reasons for existence, however, can best be understood in terms of processes that lead
from one stage to the next.
It would be impossible to overemphasize the point that thoughts are where the ac
tion is (Chafe 2001). Thoughts determine the manner in which language flows through
time, the force that drives language forward. The rest of language exists in service to the
thoughts. What speakers are conscious of as they talk is the flow of their thoughts, and
what listeners are conscious of as language proceeds is the flow of their own thoughts.
Thoughts are where everything begins and ends. After the particular language that
was chosen on a particular occasion has been forgotten, it is the thoughts behind
that language that remain in memory. Some of the language that was used to express
them on a particular occasion may remain as well, but it will not be in all respects
the same language, as is shown by studies of people talking about the same thing on
different occasions (e.g. Chafe 1998; Norrick 1998). People do not repeat the same
language verbatim, even though they may be expressing what they would regard as the
same thoughts.
It is instructive to consider what happens when one listens to a language with
which one is unfamiliar. The sounds are all one can be conscious of. Suppose the sound
is one that might be written jiyæh. For all but a small number of people that would be
just a sound. To be sure, if it was produced by the vocal organs it might be recognized
as the sound of some language, but that is all. If a speaker of the Native American
language called Seneca heard that sound, however, he or she would not pay attention
The relation of grammar to thought 
to it as a sound at all, but would be thinking of a dog. The biggest step in learning a
new language is to hear sounds but experience thoughts. Until one is able to do that,
one does not really know the language.
The nature of thought is a frustrating puzzle. What can language tell us about it?
One might at first suppose that language could tell us only how thought is organized
by language itself, and nothing about thought that lies outside of language. But the fact
is that looking closely at how people talk can show us several things about the nature
of thought that go beyond its purely linguistic aspects. In (1) I have suggested what
some of these things are (see Chafe 1994 for further discussion).
(1) Some Things That Language Shows About Thought
1. Thought is dynamic.
2. Thought is segmented into foci of consciousness.
3. Thought is constructed of ideas of events and states and their partici
pants.
4. These ideas are oriented in various dimensions, including time, space,
epistemology, emotions, interaction, and context.
First, language is obviously dynamic, constantly changing through time (James
1890: 229–237). To the extent that language expresses thoughts, the thoughts must be
constantly changing as well, and introspection suggests that thoughts are constantly
changing even when they are not being verbalized. One of the problems with most
grammatical studies is that they tend to focus on static, isolated sentences that ignore
this dynamic flow of language.
Second, thoughts are segmented into relatively brief foci of consciousness, each fully
active for only a second or two before it is replaced by another. These foci appear
in language as prosodic phrases, or intonation units. Again, thought appears to be
segmented in this way even when it is not being verbalized.
Third, within these foci of consciousness thought is built up of ideas, to employ
that word as a technical term. They may be ideas of events (things that happen), of
states (the way things are), or of people and objects that participate in events and
states, ideas that constitute what may be called referents.
Fourth, these ideas are located within a complex web of orientations, involving at
least time, space, epistemology, emotions, social interaction, and context. I will provide
examples of all these properties of thought, but first we can return to the basic picture
of language organization that was outlined above, as summarized in (2).
(2) Principal Components of Language
Thoughts → Semantics → Grammar → Phonology → Sounds
 Wallace Chafe
. From thoughts to semantics
Why, to begin with, must thoughts must be organized into semantic structures? There
appear to be at least four reasons why they must be adjusted in this way before they
can ultimately be associated with sounds. First, speakers must select which of their
thoughts to verbalize, which thoughts to turn into words. Second, speakers must de
cide how to categorize their thoughts. Third, speakers must decide how to orient them.
And finally, speakers must be able to combine these various elements into patterns that
are already familiar to their listeners. These four processes are summarized in (3).
(3) Processes That Adjust Thoughts to Sounds
Selection
Categorization
Orientation
Combination
. Selection
To begin with selection, one can never say everything one is thinking. Thoughts always
contain more than can be verbalized, so the first step in associating thoughts with
sounds is to select which of them will be expressed. This process takes place at many
levels. At the most inclusive level it involves choices of topics.
To illustrate that level of selection it is necessary to refer to some data, and in so
doing we face a problem that arises repeatedly in discussions of actual discourse. To put
it simply, the examples one might use are necessarily long. In discussions of morphol
ogy or syntax one can use illustrations that consist of words, phrases, or sentences. It
is hardly possible to cite conversations or narratives in the same way. Even a single ex
cerpt from a conversation may occupy considerable space. General principles may thus
need to be taken on faith, simply because it is impractical to present enough examples
to illustrate convincingly what is being suggested. Here I will fall back on a particular
example to illustrate points I believe are quite general, although demonstrating that
lies beyond the constraints of this chapter.
The conversation from which this excerpt was taken was recorded as part of what
is known as the Santa Barbara Corpus of Spoken American English (Chafe, DuBois, &
Thompson 1991). At this stage in the conversation two people were speaking, although
only one of them did most of the talking. She is identified as Marilyn, the other as
Pete. Marilyn’s husband Roy was also present, but was silent at this point. The three of
them were preparing a meal in Marilyn and Roy’s kitchen. Marilyn had just picked up
a lemon, and that act reminded her of an incident that had occurred shortly before,
when she and her husband returned from a conference in Laguna Beach (a place in
California). The idea of the lemon can be said to have triggered this topic. The excerpt
is presented in (4) as a series of subtopics, each provided with a label in italics. Each
The relation of grammar to thought 
line represents a prosodic phrase, the sequences of dots represent pauses of varying
lengths, and @ is a pulse of laughter.
(4) Transcript of the Lemons Topic
Marilyn and Roy’s Weekend
Marilyn: .. You know,
. . . we came back from a,
. . . we had to go to,
. . . the Ritz Carlton in Laguna,
. . . for a . . . event,
. . . and . . . we were gone for like four days.
. . . You know,
.. really relaxing weekend,
. . . you know,
. . . it was really great.
.. And I’d had this,
. . . I’d had a particularly stupendous time.
Because I had to . . . be a wife most of the time,
but part of the time I got to be a worker,
and do the really fun work,
so it was really [great].
Pete: [Unhunh,]
Marilyn: I had a great time.
Discovery of Girl
. . . And uh,
. . . we’re pulling up,
. . . and I see this girl.
. . . Who I’d never seen before sort of,
. . . dart out of our driveway.
Pete: . . . Un[hunh],
Marilyn: [And,]
. . . stand there,
and watch us pull in,
and she goes like this.
. . . Like,
. . . oh my God,
I’m gonna get caught.
Conversation with Girl
I said,
.. Hi,
can I help you?
Pete: @@[@@,]
 Wallace Chafe
Marilyn: [You know,]

.. And she goes,
.. and it,
.. you know,
of course it’s this long drive so I,
I probably look like . . . total hell.
Pete: .. Right.
Marilyn: And she goes,
.. Oh,
um,
I was just getting .... some lemons.
Pete: @@@[@@,]
Marilyn: [I said],
oh yeah?
.. Who are you.
Pete: .. @[@@@,]
Marilyn: [And she] goes,
.. oh,
I’m your next door neighbor.
Permission from Kenneth
Pete: .. @@@[@@]
Marilyn: [No she] said,
first she said,
Kenneth said I could have some.
Who’s this guy,
.. old guy across the street.
Pete: .. Unhunh.
Marilyn: So Roy had said . . . to him,
.. if you ever want lemons,
you know,
go ahead and .. and [have em,]
Pete: [Right,]
Marilyn: then he said,
.. she said,
oh Kenneth said I could have some lemons.
Pete: . . . Right.
Conversation with Girl Continued
Marilyn: .. I said,
. . . he did.
Pete: @[@@@]
The relation of grammar to thought 
Marilyn: [She goes,]

. . . Yeah.
. . . And I said oh,
and she goes,
do you,
you don’t mind do you?
Well yeah,
in fact I do mind.
Cause I thought the lemon tree was dying.
I didn’t see any lemons on it.
Pete: .. Unhunh.
Marilyn: Like because she said,
. . . you wouldn’t mind if I came back and got a
whole bag full would you?
Pete: Right.
Marilyn: I said
. . . yeah.
[@@@
Pete: [@@@@@,]
Problems Getting Lemons
Marilyn: There’s like .. like] one lemon left on this
[tree] that I can reach.
Pete: [Right.]
Marilyn: You know?
Pete: .. Unhunh.
Marilyn: .... Roy’s about to go to Africa,
and so I’d have,
you know,
I’d have to get . . . a step ladder out,
to pick my own lemons,
. . . give me a break.
When Marilyn, in the kitchen, picked up a lemon that had come from her tree, she
selected this topic from her memory as something worth communicating to Pete,
assuming that he would find it interesting. The topic as a whole formed a coher
ent element within the larger conversation. Within it one can identify the subtopics
given the italicized labels. The first was Marilyn and Roy’s Weekend. At the beginning
of it one can see that Marilyn changed her mind during the selection process. She
started to tell about coming home and finding the girl, but then she changed her mind
and decided to provide first some background information about where she and her
husband had been.
 Wallace Chafe
(5) A Shift in Subtopic Selection
The entire first subtopic is repeated in (6). One may notice throughout the larger se
lection how Pete’s brief responses related to subtopic boundaries. With his unhunh
toward the end of (6) Pete showed his understanding that something coherent had
been concluded at that point. Marilyn essentially concluded this subtopic by saying so
it was really great, Pete uttered his response, and Marilyn repeated her evaluation with
the coda, I had a great time.
(6) The First Subtopic with Pete’s Response
.. for a . . . event,
. . . and .. we were gone for like four days.
. . . You know,
you know,
it was really great.
.... I’d had a particularly stupendous time.
Pete: [Unhunh,]
Marilyn was now ready to turn to the discovery of the girl, shifting from the time of the
weekend to the time when they returned home, as was reflected in her language with
the shift from the past tense in I had a great time to the present tense in we’re pulling up.
From this point on, the subtopics she selected were guided largely by the chronological
sequence of the events she remembered. Marilyn showed another change of mind in
selecting what to talk about when she first quoted the girl as saying I’m your next door
neighbor, but then remembered that the girl had first mentioned receiving permission
from Kenneth, as shown in (7).
The relation of grammar to thought 
(7) Another Switch of Subtopic
Marilyn: [And she] goes,
.. oh,
I’m your next door neighbor.
Pete: .. @@@[@@]
Marilyn: [No she] said,
first she said,
Kenneth said I could have some.
These were a few examples of the ways Marilyn selected what to talk about in terms
of topics and subtopics. But selection takes place at finer levels too. In (8) one can see
how Marilyn chose to verbalize the experience of discovering the girl.
(8) The Choice of Particular Ideas to Verbalize
Marilyn: . . . And uh,
.. we’re pulling up,
.. dart out of our driveway.
Marilyn: [And,]
. . . stand there,
.... Like,
.. oh my God,
I’m gonna get caught.
We can suppose that Marilyn’s thoughts at this point included a variety of images.
Probably she remembered their house, their driveway, their yard, the lemon tree, per
haps other trees, neighboring houses, and so on, and probably she remembered how
she and her husband were positioned inside the car. She did not choose to say who was
driving, where they parked, or where she was when she talked with the girl. Perhaps
she talked through the car window. All these things and more may have been included
in her thoughts, but she selected only some of them, presumably those she judged most
worth telling about.
In short, when thoughts are associated with sounds, speakers must select which of
their thoughts to express, being unable to verbalize everything they are thinking. It is a
process of many to few. This selection process can be regarded as the first step toward
the creation of a semantic structure.
 Wallace Chafe
. Categorization
The second process listed in (3) was categorization. Thoughts consist for the most part
of ideas that are particular: ideas of particular events, particular people, and particular
objects associated with particular times and places. It would obviously be impossible
for each of these particular ideas to be associated with a different sound. Franz Boas
(Boas 1963 [1911]) understood this well at the beginning of the last century:
Since the total range of personal experience which language serves to express is
infinitely varied, and its whole scope must be expressed by a limited number of
phonetic groups, it is obvious that an extended classification of experiences must
underlie all articulate speech. (Boas 1963 [1911]: 18)
Not only is the number of particular ideas that might be expressed in language vast,
there is no way a listener could know which particular idea was associated with which
particular sound, even if such an association were possible. We cope with this problem
by categorizing each particular idea as an instance of something already known. We
assign particular ideas to general and familiar categories.
A category serves, in fact, two functions. For one thing, it provides expectations
about a particular idea, expectations that can be associated with any instance of the
category. Perhaps we can lift it, perhaps we can eat it, perhaps we can sit in it, or
whatever. But a category also provides a phonological representation that is available
to be associated with any instance of it, allowing the particular idea can be expressed
with a particular sound.
Ideas vary in their amenability to categorization. Many years ago Roger Brown
wrote about ideas having different degrees of codability (Brown 1958). Some are pro
totypical instances of a category and thus easy to categorize, while others are less
prototypical and there is more difficulty in their categorization. In the above exam
ple the ideas expressed as see, girl, and driveway were evidently highly codable. It was
easy to interpret them as instances of those categories. On the other hand the so-called
hedge sort of, together with the fact that it was followed by a hesitation, are evidence
that the idea expressed as dart was less easy to categorize, less codable. This element
of Marilyn’s thought did not fit as readily with the semantic resources provided by the
English language.
. Orientation
The third process that leads from thoughts to a semantic structure is orientation. In
the area of thought, ideas are positioned within a complex web of orientations that
include at least an association with a particular time, a particular space, and an epis
temology (how we know about them, whether they are real or fictional, how reliable
they are). They may also be oriented in terms of emotions and attitudes, as exemplified
by Marilyn’s negative attitude toward the girl. In a conversation ideas are necessarily
oriented to the ongoing interaction with others, as exemplified by Marilyn’s frequent
The relation of grammar to thought 
use of the phrase you know. Finally, they are necessarily oriented to the ongoing con
text as well. Every idea is linked to other ideas that have already been activated, or that
the speaker expects to activate.
It would be impossible to express the full range of orientations that might exist in
thought in any particular use of language. Different languages, in fact, make certain
orientations more or less easy to express, and particular languages often force their
speakers to express certain orientations while they ignore others. English likes to orient
events in time, especially time with relation to the time the language is being created.
We call it tense marking, which is obligatory in English to a greater degree than in a
number of other languages. Many Native American languages, for example, are more
concerned with epistemology than with tense.
(9) Uses of the Past Tense
.. for a . . . event,
. . . and .. we were gone for like four days.
. . . You know,
you know,
it was really great.
.... I’d had a particularly stupendous time.
Pete: [Unhunh,]
It is interesting to observe the flow of temporal orientations in Marilyn’s narrative dur

ing the subtopic repeated in (9). Before she introduced the major topic of the lemon
tree, Marilyn and Pete were using the present tense as they talked about what they were
doing in the kitchen. They were in what I have called the immediate mode, discussing
things that were right in front of them. But now, as can be seen in (9), Marilyn began
talking about things she remembered from the past, switching to the displaced mode
(Chafe 1994: 196–201), and to the past tense with the word came as she remembered
the time they arrived home from the conference. But then she quickly changed from
that time to an earlier one: not even the time of their weekend but the time before that,
the time when they had to go there. At this point she had used the past tense for two
different times: the time when they arrived home and the time before that, when they
 Wallace Chafe
left for the weekend. But then, with were and was, she used the same past tense to cover
the entire extended period of their weekend.
The past tense had thus been used for three different temporal orientations, all
having in common the fact that they were prior to the time she was talking. But sud
denly, while she was talking about the same period of time (the extended period of
their weekend), she suddenly switched to the past perfect with the words I’d had. The
effect of this switch was to position the time of the weekend before the time they ar
rived home, the time represented by the past tense in the first line. For just this brief
moment in her narrative she distinguished linguistically between those two different
times in the past. But it was only for a moment, and immediately after that she returned
to treating the time of the weekend with the simple past in I had to be a wife.
(10) A Switch to the Present Tense
Marilyn: [And,]
. . . stand there,
But that was not the last change in temporal orientation. As Marilyn moved to the next
subtopic, the discovery of the girl, she switched to the present tense, as shown in (10).
This change established a clear boundary between the two subtopics, but it was also
a use of the so-called historical present, a pretense that the time of the events she was
talking about was the same as the time she was talking about them (Chafe 1994: 207–
210). It is a device sometimes used in conversational narratives to express a special
kind of vividness, a pretense that as we hear the language we are really experiencing
the events unfold.
Within this present tense orientation Marilyn began with the progressive aspect
(we’re pulling up) to cover the time occupied by their arrival, and then the simple
present (I see this girl) to focus on her sudden discovery of the girl within that time. But
then she jumped to the past perfect when she said who I’d never seen before to convey
the fact that she didn’t know the girl. At this point the linguistically expressed temporal
relations were more finely graded, capturing two times in the past. But the departure
from the historical present was only brief. The girl’s reaction was oriented again with
the present tense in she goes like this, illustrating a typical use of the historical present
for the attribution of reported speech.
The relation of grammar to thought 
. Combination
The last process listed in (3) as necessary in the adjustment of thoughts to seman
tic structures was combination. So far we have seen that speakers select what to talk
about, they categorize the ideas they have selected, and they orient those ideas in ways
that are favored by whatever language they are speaking. But the categorizations and
orientations must somehow be combined. As a basis for discussion we can look at the
segment shown in (11).
(11) Combining Ideas and Orientations

First of all, how much can we say about the cluster of thoughts that lay behind this
language? It is an especially difficult question to answer, because as soon as we use lan
guage to talk about thoughts we have already verbalized them. The question involves
preverbalized or only partially verbalized thoughts. The discussion would be easier if
there were a way to represent thoughts without using language to do it. Although the
thoughts themselves may already include certain ways of verbalizing them, they may
also include imagery, affect, and perhaps other modes of consciousness.
One thing that can be said is that the thoughts in (11) included the idea of an event
that we might try to characterize in the most general terms as awareness of a person:
not those words, but the idea. There was certainly more: an awareness of what the
person looked like, where she was located, Marilyn’s emotional reaction to her, and so
on. As mentioned earlier, the fact that thoughts contain more than any language that
might be used to express them becomes clear when we find someone verbalizing the
same thoughts on different occasions. Such data are not available in this case, but we
can imagine that at some other time Marilyn might have said, not I see this girl, but
perhaps this girl appeared, or there was this young lady, or whatever.
How was this thought converted into the semantic structure behind the words I
see this girl? In spite of their importance we still do not know how to represent seman
tic structures, but we can at least speculate on selected aspects of what was involved in
this case. This structure was built on the idea of an event, itself an element of thought.
It was an idea that was easy to categorize; Marilyn evidently had no trouble interpret
ing it as an instance of the see category. She oriented it in two ways: as a new event
within the ongoing discourse (an idea she assumed she was activating in her listener’s
consciousness for the first time), and temporally as present, using the historical present
discussed above.
A significant result of categorizing this event as an instance of seeing was to shift
the new information from the seeing itself to the thing that was seen. The see category
has a presentative function. It does not itself carry the burden of new information, but
 Wallace Chafe
starting point for �

presentation of �
idea of a referent idea of an event idea of a referent

specified as categorized as categorized as
speaker see girl
oriented as oriented as oriented as
given present new
identifiable nonidentifiable
pointed to
Figure 1. A possible semantic structure for ‘I see this girl’
introduces what is seen as the locus of what is new. In this case it introduced the idea
of the girl, an idea that was of course categorized as an instance of the girl category.
That idea was oriented as new, and also as nonidentifiable; that is, the girl was assumed
at this point not to be an idea the listener could identify. Finally, it was also oriented as
something the speaker was, in a metaphorical sense, pointing to.
When a semantic structure includes two ideas, such as the idea of the seeing and
the idea of the girl, they must be related in some way. There are several ways to regard
the relation between these two ideas, but assuming that a good way to characterize it
is in terms of the presentative function, we can say that the seeing functions to present
the idea of the girl.
But the see category also includes the idea of the person who does the seeing. In this
case that person was the speaker. The idea of the speaker need not be interpreted as an
instance of a category. It is true that there are a vast number of possible speakers; almost
any human could play that role. The identity of this particular referent, however, was
known to the listeners from the speech situation and no categorization was required.
This idea was oriented in two ways. First, it was given (assumed to be already active in
the listeners’ consciousness because of the speech situation). Second, it was identifiable.
The speech situation allowed the listeners to identify who this was.
How was this idea of the speaker related to the presentation of the girl? Because the
seeing and the girl already formed a single presentational unit, the idea of the speaker
was related to the constituent composed of see and girl. What can we call this relation? I
have used the term starting point (Chafe 1994: 82–92). Marilyn had already been talk
ing about various things she did, and here the idea of Marilyn herself continued to
function as the starting point for the introduction of this new idea of seeing the girl.
Already active in consciousness, the idea of seeing the girl was added as another of
Marilyn’s activities.
Figure 1 is an attempt to show how Marilyn’s thoughts at this point were adjusted
to a semantic structure through processes of selection, categorization, orientation, and
combination. Whatever the most appropriate representation may be, there is some
thing important to notice here. What is shown in Figure 1 is not very different from
a grammatical structure. Like a grammatical structure, it contains various elements
The relation of grammar to thought 
that are related in certain ways. The idea of seeing resembles a verb, the idea of the
girl a noun, the idea of the speaker a pronoun. These elements are related as con
stituents in a structure that resembles a syntactic tree. If there is a difference from a
grammatical structure, it lies in the fact that all these elements and their relations are
directly related to thought. All of them are meaningful. They are elements of thought
that have been filtered and adjusted in accordance with the semantic resources of the
English language.
. Differentiating semantics from grammar
What is it, then, that happens to the kind of structure shown in Figure 1 that leads
to a grammatical structure? And why, in fact, should anything at all have intervened
before Marilyn’s thoughts, semantically adjusted, were associated with sounds? There
is a trivial reason and there is an important reason. The trivial reason is a matter of
terminology. Terms used for describing grammar have been introduced from a variety
of sources over several thousand years in different places and for different reasons.
Many did have some basis in semantics when they were first invented, but with the
passage of time they have come to be used as arbitrary labels for grammatical elements
and relations that are no longer directly related to thought.
This fact is apparent when we consider how some of the labels in Figure 1 might
be replaced by more grammar-sounding labels. For example, the semantic label start
ing point might be replaced with the grammatical label subject. The semantic label
presentation might be replaced with the grammatical label object. The semantic label
identifiable might be replaced with the grammatical label definite. The semantic com
bination of nonidentifiable with pointed to might be replaced with indefinite demon
strative. Words like subject, object, definite, indefinite, and demonstrative are only
marginally appropriate as semantic labels. Such changes make no substantive differ
ence. They are only terminological changes from a more semantic basis to something
more in line with grammatical tradition.
More significant is the manner in which these elements were combined, most
notably the realization of the indefinite demonstrative (semantically associated with
nonidentifiability and pointing) as the separate word this. The structure shown in Fig
ure 2 sketches these elements, still largely semantic in the sense of their direct relation
to thought, although they are combined in a manner that is partially arbitrary. It was
something of this nature that presented itself to phonological symbolization.
Before we leave this example, it can be of some interest to discuss its realization
in prosody: the variations in pitch, loudness, timing, and voice quality that are an es
sential part of language. Prosody tends to reflect the flow of thought directly, although
we will see below that grammatical structure can affect the way prosodic features are
distributed. Figure 3 is a display of fundamental frequency, perceived as pitch, in the
phrase and I see this girl. It is apparent that girl, the element labeled new in Figure 2,
received by far the highest pitch. It was also lengthened and pronounced with greater
 Wallace Chafe
subject of �
object of
modifies �
I see this girl

given present new
Figure 2. Sketch of a grammatical structure for ‘I see this girl’
and I see this girl
Figure 3. Fundamental frequency in ‘and I see this girl’
intensity. There was a smaller rise on the presentative element see. Also of interest is
the rise-fall contour assigned to the word girl, expressive of Marilyn’s attitude toward
this person. Marilyn did not experience the discovery of the girl as a routine event but
as something that elicited a negative attitude, characterizable perhaps as disdain. This
phrase ended with a falling pitch that was partially obscured in Figure 3 because of a
lapse into creaky voice. It expressed the kind of closure associated with the end of a
prosodic sentence.
. Lexicalization and grammaticalization
To return now to the question raised above, why grammar and semantics should be
distinct, and leaving terminological differences aside, the answer can be found in pro
cesses of language change, and more specifically in the twin historical processes of
lexicalization and grammaticalization. As a language changes, certain words, phrases,
and constructions come to be used in ways that no longer have a direct relation to
thought. They are, in other words, no longer directly semantic. With reference to the
distinction made above between ideas and their orientations, lexicalization involves
changes in the way ideas are expressed, while grammaticalization involves changes in
the expression of orientations (Chafe 2002).
The relation of grammar to thought 
starting point for �
idea of a referent idea of an event

specified as categorized as
speaker “pull up”
and other oriented as
oriented as new
given present
identifiable progressive
Figure 4. Sketch of a grammatical structure for ‘we’re pulling up’
The distinction between semantics and grammar becomes more interesting and
complex when we look at the phrase immediately preceding the one just discussed:
we’re pulling up. What can we say, first of all, about its semantic structure? The nucleus
of that structure is the idea of an event that Marilyn categorized as an instance of the
pull up category. In semantic terms the relation of this category to what would other
wise be conveyed by the words pull up was indirect. The meaning of pull up here bore
a resemblance to the meaning of arrive. For historical reasons this meaning had come
to be represented grammatically by the sequence pull up, which was only indirectly
related to what Marilyn was thinking.
The idea of pulling up, in the sense of arriving, was oriented as a new idea, but in
addition it was oriented as if it was happening at the same time that Marilyn was talking
(again the historical present). It was also oriented as progressive, an arbitrary label for
one of the aspectual options made available by the semantic resources of English. The
progressive aspect meant that this event occupied a span of time within which another
event took place, in this case seeing the girl. The event of pulling up surrounded in
time, as it were, the seeing event.
An idea that is categorized as an instance of pulling up must include the idea of
the person who does it. In this case it was the speaker and her husband, a referent that
was oriented as given and identifiable, and that functioned as the starting point for
this phrase. Figure 4 is an attempt to represent the semantic structure of this phrase.
The fact that pull up bears only an indirect relation to thought is captured by the
quotation marks.
How does this become a grammatical structure? Trivially, the terminology can
be changed to reflect traditional practices, with starting point changed to subject and
identifiable to definite. More important is the fact that “pull up”, a semantically unified
category, was expanded grammatically into the phrasal verb pull up: two words, one
of them a verb and the other a particle. This expansion reflects a lexical change in
the English language: the invention of a new category that came to be expressed by
a phrase that no longer had a direct or straightforward relation to thought. Marilyn
was not talking about pulling something up; she was talking about arriving home. It is
helpful to regard pull and up as elements that are quasi-semantic. They sound as if they
are directly meaningful, directly related to thought, but they are not. Grammar comes
 Wallace Chafe
to be distinguished from semantics through the creation of quasi-semantic elements,

the results of historical processes such as lexicalization.
It was mentioned earlier that these stages from thoughts to semantic structures to
grammatical structures and eventually to sound do not exist in isolation. For one thing
semantic structures influence thought, because language provides what is probably
the most important way we have of organizing thoughts. With this example we have
seen that grammatical structures arise historically from semantic structures, but in fact
there is often a leakage from grammatical structures back into meaning. The amount
of leakage can vary with each case, and perhaps even across individuals.
The idiom pull up, with a meaning similar to that of arrive, evidently arose in the
days when people rode in horse-drawn wagons and pulled up on the horse’s reins to
make him stop. At first, people must have experienced what I have called a shadow
meaning of such an experience, a literal interpretation that may have continued for
a time after a horse was no longer involved. People thought primarily about arriving,
but this shadow meaning of pulling up on a horse’s reins would have accompanied that
thought. It is worth noting also that the up in pull up occurs in other idioms where it
involves doing something to completion, as in clean up, fix up, and finish up, a shadow
meaning associated with up in these contexts. Lexicalization, in other words, does not
always create elements that are totally divorced from thought. Residual shadow mean
ings create a leakage from grammar back into semantics. But the important point here
is that Marilyn was thinking about arriving, and the idiom pull up was the way she
chose to categorize that idea.
The pull up category shows how a separation of grammatical structure from se
mantic structure can be created by lexicalization. But within the same phrase there is
an excellent example of grammaticalization, a historical process that is similar to lex
icalization except that it applies to orientations rather than ideas. Here it applies to
the progressive aspect, expanding the progressive meaning into the verb be, here plu
ralized as are, while at the same time it adds the suffix -ing to the verb. The semantic
element progressive as such disappears, and present tense comes to be associated with
the auxiliary are.
Something like Figure 5 may be appropriate as a grammatical representation of we
are pulling up. The constituent structure assigned here to the sequence are pulling up
seems clear, but I have not attempted to label the relations between these constituents.
The assignment of such relations is not obviously appropriate within words such as
pulling, and the same may be true of more extended sequences like this one.
But this representation can be taken one step further. It is unlikely that a modern
English speaker would actually say we are pulling up unless the word are was used
to signal a contrast with some denial of this event. Normally we are is contracted to
we’re, and that is what Marilyn said. So phonological change, which is no respecter of
semantic function, has converted Figure 5 into a structure in which the earlier auxiliary
are is attached to the pronoun we, yielding a single word with the peculiar property
of expressing both the idea of the people who were pulling up along with half the
progressive orientation of that event. It is an excellent example of the extent to which
The relation of grammar to thought 
subject of �
we are pull ing up

given present present new new
progressive
new
Figure 5. A possible semantic structure for ‘we’re pulling up’
we’re pull ing up
Figure 6. Fundamental frequency and intensity in ‘we’re pulling up’
a semantic structure can be radically distorted through both grammaticalization and

phonological change.
We can consider prosody here as well, referring to Figure 6, where fundamental
frequency is shown above the transcription and intensity, perceived as volume, below.
The word we’re reflected the given status of the idea of Marilyn and her husband, and
was thus pronounced with low pitch and volume. New information was expressed
with the words pulling up. The pitch rose sharply on the first syllable of pulling and
reached a peak on the second. The word up was prolonged, and during it the pitch fell,
though not with the complete fall (accompanied by creaky voice) we observed above
in and I see this girl. This partial fall conveyed an incomplete closure, greater than that
associated with a rise, but still with an expectation of more to come: the sighting of
the girl. It is of some interest that the intensity display in Figure 6 follows a different
 Wallace Chafe
pattern, showing an increase rather than a decline on the word up. Thus pitch and
volume played different roles. Pitch showed a partial but incomplete conclusion in
the flow of thought, but volume showed a more arbitrary, grammatical assignment of
maximum prominence to the final element of the idiom pull up.
. Concluding remarks: Grammar and thought
Language enables us to associate thoughts with sounds. Because thoughts and sounds
are so different, but also because thoughts are particular and the ways they are ex
pressed must be shared by speakers of a language who agree on shared conventions,
thoughts must be adjusted in several ways before any such association is possible. For
one thing, speakers must select from the rich content of their thoughts those ideas
they wish to verbalize. They must then choose ways of interpreting those chosen, par
ticular ideas as instances of general categories already familiar to their listeners, and
that provide words or phrases appropriate for their symbolization. They must also de
cide how to orient their ideas in time, space, epistemology, emotions, interaction, and
context. Each language favors certain orientations and tends to ignore others. Finally,
the selections, categorizations, and orientations must be combined in accordance with
whatever patterns the language makes available. The result of these four processes is a
semantic structure, a way of organizing thoughts within the resources provided by the
language being used.
If languages never changed, that would be the whole story. But languages do
change. Lexicalization can make the expression of categories indirect, creating a divide
between semantics and grammar. Grammaticalization can similarly affect the expres
sion of orientations. Both of these historical processes lead to quasi-semantic elements
that are no longer directly related to thought, although in some cases there may be se
mantic leakage from grammar back into thought. The result is a grammatical structure
in which some of the elements and relations are still meaningful, but some are not.
A grammatical structure is thus an arbitrary melange of semantic and nonsemantic
elements. It is this mixture that is expressed with sound.
Linguistics has, to its detriment, tended to overemphasize grammar at the expense
of semantics. If the perspective on language described here is valid, more attention
should be paid to the nature of semantic structures, to the relations of semantic struc
tures to thought, and to ways in which lexicalization and grammaticalization create the
divide between semantics and grammar. To put it briefly, in order to understand why
grammar is the way it is, one must understand both semantics and language change.
Otherwise the study of grammar is empty, because grammar is left unexplained.
All of this leads to one final point. The relation of grammar to thought has been
a controversial issue for a long time. It is often presented as the question of whether
people who speak different languages, and whose grammars are therefore necessarily
different, actually think differently. Often it is presented as the question of whether
Benjamin Lee Whorf (1956) was right or wrong, but that is far too simple a way to
The relation of grammar to thought 
frame the question. In the beginning there is thought, and the basic question has been
whether grammatical structure affects it.
But it follows from everything said above that a direct relation between thought
and grammar should never be expected. The relation of interest is that between
thought and semantic structure. If we approach the question in that light, it seems
clear that different languages provide their speakers with different semantic resources.
No one would deny that different languages organize sounds in different ways, and
surely they must differ at least as much in the ways they organize thoughts. It must
then at least be the case that when people are speaking, the ways in which they or
ganize their thoughts for that purpose are shaped by their language. “Thinking for
speaking,” to use Dan Slobin’s term (Slobin 1996), must necessarily be different across
languages.
When people are speaking, then, their thoughts are necessarily adjusted to the
semantic resources of their language. But if we go on to assume that silent thought
consists in part, though certainly not entirely, of inner speech, language must play
a pervasive role in that experience as well. To that extent even silent thought must
be affected by language. Whether the influence of language on silent thought extends
beyond inner speech is a more difficult question, but it is at least something to think
about.
References
Boas, Franz (1963 [1911]). Handbook of American Indian Languages. Bureau of American
Ethnology, Bulletin 10, Part 1 (1911). Reprinted by Georgetown University Press.
Brown, Roger (1958). Words and Things. Glencoe, IL: The Free Press.
Chafe, Wallace (1994). Discourse, Consciousness, and Time: The Flow and Displacement of
Conscious Experience in Speaking and Writing. Chicago: The University of Chicago Press.
Chafe, Wallace (1998). “Things we can learn from repeated tellings of the same experience”.
Narrative Inquiry, 8, 269–285.
Chafe, Wallace (2001). “The analysis of discourse flow”. In D. Schiffrin, D. Tannen, & H. E.
Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 673–687). Oxford: Blackwell.
Chafe, Wallace (2002). “Putting grammaticalization in its place”. In I. Wischer & G. Diewald
(Eds.), New Reflections on Grammaticalization: Proceedings of the International Symposium
on Grammaticalization, 17–19 June 1999, at Potsdam University, Germany (pp. 395–412).
Chafe, Wallace, John W. Du Bois, & Sandra A. Thompson (1991). “Toward a new corpus of
Spoken American English”. In K. Aijmer & B. Altenberg (Eds.), English Corpus Linguistics:
Studies in Honour of Jan Svartvik (pp. 64–82). London: Longman.
James, William (1890). The Principles of Psychology. New York: Henry Holt. [Reprinted 1950 by
Dover Publications].
Norrick, Neal R. (1998). “Retelling stories in spontaneous conversation”. Discourse Processes, 25,
75–97.
 Wallace Chafe
Slobin, Dan I. (1996). “From ‘thought and language’ to ‘thinking for speaking”’. In J. J.
Gumperz & S. C. Levinson (Eds.), Rethinking Linguistic Relativity (pp. 70–96). Cambridge:
Whorf, Benjamin Lee (1956). Language, Thought, and Reality: Selected Writings of Benjamin Lee
Whorf. Edited with an Introduction by John B. Carroll. Cambridge, MA: MIT Press.
Communicative constructions
in English and Spanish*
Montserrat Martínez Vázquez
This paper examines constructions used to report communicative acts in English

and Spanish, both from a verb-centered perspective and from a constructional
approach, as in Goldberg (1995). I review sentences with verba dicendi, and other
constructions where verbs not belonging to the communicative domain are
inserted in a ditransitive/communicative frame. The analysis of corpus examples
shows that this conflated communicative pattern is not limited in English to
conventional constructions of the type kiss goodnight. Novel uses reveal that this
grammatical metonymy is more widespread than is generally believed. In
Spanish it is not as productive as it is in English, but I claim that it is not banned
by syntax, as has been suggested.
. Introduction
Human beings have always felt an irresistible urge to talk. The etymological mean
ing of the verb communicate (from Latin communicare “to share”) reveals its social
nature. The importance we give to this interchange of information is reflected in lan
guage, since “verbs referring to speech constitute one of the most important areas of
the vocabulary of any language” (Wierzbicka 1987: 3). Most of our daily experiences
consist of speech events, either the transmission of our thoughts, feelings, etc. or the
report of that speech by a third person, probably the hearer of the original conversa
tion. Reported speech has been defined as a “universal of the language capacity”, which
is pervasive in our daily language use (Collins 2001: 1). Indeed this “talk about talk”
has to be recognized as a central role of communication, which has a significant pre
dominance in language activity (Collins 2001: 1). This central activity of human beings
has received a lot of attention from different angles. Most of the abundant literature on
the metarepresentational ability of human beings has focused on the “theory of mind”,
“the ability to explain and predict the behavior of others by attributing to them certain
beliefs, intentions and desires” (Noh 2000: 1).
Information may be transmitted linguistically or non-linguistically. Ideas or feel
ings may be exchanged by speech, writing, gestures, etc. Different extralinguistic tools
used to convey meaning may include sounds or gestures like smiling, weeping, cough
 Montserrat Martínez Vázquez
ing, nodding or shaking the head. I will be concerned here only with linguistic con
structions, more precisely with constructions used to report a communicative act,
what Rudzka-Ostyn (1988: 513) calls a “secondary speech situation”.
The report of a communicative act may show up in a wide variety of linguistic
constructions ranging from those used to express verbal communication, (1), to those
reporting non-verbal communication, as in (2).
(1) He said hello.
(2) He nodded hello.
In this paper I will examine these English constructions – and their counterparts
in Spanish – both from a verb-centered perspective and from a constructional ap
proach, as in Goldberg (1995). On this view, constructions are independent form-
meaning pairs with their own semantics, capable of contributing arguments. Thus,
non-subcategorized complements are viewed as licensed by the construction rather
than by the verb. By considering not just the verb but also the semantics of the sentence
we will enlarge considerably the number of communicative constructions. Besides the
sentences built with verbs of saying (verba dicendi) as in (1), other linguistic devices to
report an interchange of information, as, for example (2), will be included. The verb
does not belong to the class of speaking verbs, but the object clearly reveals a com
municative act. Exclusively verb-centered approaches to communicative constructions
are incapable of dealing with the entire phenomenon. So besides the extensive class
of verba dicendi, an analysis of the phenomenon should look at other verbs which are
compatible and may, therefore, eventually appear in a communicative construction.
Sentences like (1) have been extensively studied both in English and in Spanish
under the rubric “communicative verbs”.1 Constructions like (2), however, are usu
ally considered a peripheral phenomenon in English. I have not seen any reference
to counterparts of (2) in Spanish. In fact, what I have found are statements suggest
ing the impossibility of such constructions. Starting from such a state of affairs the
main purpose of this paper is to investigate the productivity of such constructions in
English, and find out if there are equivalent expressions in Spanish, or, as has been sug
gested, Spanish syntax does not permit them. In order to reach this goal extensive use
of corpora is required. I have extracted the English data from the British National Cor
pus (BNC), the Brown Corpus, the Wordbank of the Collins Cobuild on CD-ROM, and
the British Component of the International Corpus of English (ICE-GB). Spanish occur
rences have been extracted from the Corpus de Referencia del Español Actual (CREA).
Unless otherwise stated, the examples compiled from the CREA illustrate instances of
Spanish from Spain. Occasionally references to linguistic data from other sources will
be made, especially examples taken from novels and their translations,2 which will help
to find out more about equivalences in each language.
This paper is organized as follows. First, a brief analysis of the scene of linguistic
action is made in order to isolate the main elements of communicative constructions.
A description of constructions with verba dicendi follows. A third section will focus on
different constructions with other non-speaking verbs, which result from fusing two
Communicative constructions in English and Spanish 
processes into one. In order to better understand these constructions basic concepts of
metaphor theory will be applied. A tentative account of why Spanish fails to make an
extensive use of this grammatical metonymy is offered in the concluding section.
. The communicative process
An interchange of communication involves at least three nuclear elements: sender,

message and receiver. According to Jakobson (1960) the message has to be related to a
context, and must be put into a code common to the speaker and the receiver. Finally a
contact enables addresser and addressee to communicate. Dirven et al. (1982: 4) offer
a much more complex picture:
. . .a sender transfers information to a receiver: this information transfer is pre
sented as a message and/or topic or evaluation; the information is transferred via a
channel, put down in a code and possibly caught in a textual conveyor. The whole
process takes place in a given socio-cultural context, at a specific time and place, in
a certain manner and under certain circumstances; the sender may have a certain
purpose and the receiver may be affected by the result of the information transfer.
The linguistic situation is even more complex with reported speech, which involves
a double communicative process. The second speaker, or reporter, adds her personal
point of view on the first communicative act. Thus, she can focus on any aspect of
the communicative interchange by selecting the appropriate verb. If she wishes to
emphasize the channel, she may select a verb of instrument of communication like
telephone; she may prefer to highlight the mood of the speaker by choosing a manner
verb like yell, scream or bark, or manipulate the original message by presenting it with
her own words.
. The communicative construction
. The participants

In (1) we find a verb belonging to the class of verba dicendi, verbs used to report a
speech act.3 Verbs of saying such as say, tell, explain, etc. involve two participants:
one which is +human/+intentional, and a second one containing the information
transmitted, which may be sentential. Downing & Locke (1992) call them “sayer”
and “verbiage” respectively; the third, optional, participant which may appear is a
“recipient”.
Processes of saying and communicating are verbal processes. The participant who
communicates is the Sayer, and is typically human. That which is communicated
is the Verbiage and may be a reported statement, a reported question or a reported
directive. A Recipient may also be present is some verbal processes.
(Downing & Locke 1992: 136)
 Montserrat Martínez Vázquez
The subject is inherently human since, as mentioned above, we are referring to a

social activity, as Cano (1987: 207) points out: “. . .todos exigen sujeto [+Humano],
pues designan una de las actividades más específicas de los seres humanos” [‘all de
mand a [+Human] subject, since they designate one of the most specific activities
of human beings’]. Non-human subjects, which frequently appear with these verbs,
involve metaphor:
(3) In those days poems often told a story in verse. . . (Brown)
(4) La sangre de Abel vocea el delito de Caín. (DRAE)
‘Abel’s blood shouts Caín’s crime.’
The second participant, the dictum, will either reproduce what has been said, or sum
marize the content of the communicative act. This participant is inherent to many
intransitive communicative verbs like talk, chitchat, etc.; when we talk information is
exchanged, even though it may not appear as an argument of the construction. When
this argument is present in the syntax it may take the form of a quotation,4 a subordi
nate sentence, or a pro-dictum, i.e. a phrase that summarizes the content of a speech
act. As Cano (1987: 207) explains in his analysis of Spanish communicative patterns:
En general, podemos considerar dos tipos: la secuencia verbal efectivamente emi
tida por alguien, como en Pedro dijo que vendría, o, sobre todo en el ‘estilo di
recto’: Pedro dijo: “Iré a casa”; por otro lado el contenido o tema de un acto verbal,
pero no lo enunciado como tal: Juan contó el modo en que entró. (1987: 207)
‘In general, we can consider two types: the verbal sequence actually uttered by
someone, as in Peter said that he would come, or, especially in ‘direct style’: Peter
said: “I will go home”; on the other hand the content or topic of a verbal act, but
not the utterance as such: John recounted the way in which he entered’.
The quotation is the most neutral and direct way of reporting a speech event, as in (5).
Though, as Noh (2000: 8) observes, the essence of quotation is resemblance rather than
identity. The subordinate pattern found in indirect speech involves a manipulation of
the actual utterance, since it has to be integrated in the syntax of the main sentence by
shifting person and deictics, (6). Finally, the pro-dictum realization implies a higher
grammatical and cognitive effort since it involves a great reduction of the content of
the utterance, involving in many cases the use of metonymy. This phrase functions as a
pro-form which stands for the original exchange of information. The most basic type
of substitution is a metalinguistic word, like message, story, news, facts, report, etc., as
in (7). Vorlat (1982: 27) describes it as a “recitable entity”. A pro-dictum may stand for
a simple sentence quotation, a paragraph or a larger text. The rest of possible pro-dicta
range to cover much more complicated interchanges of information like the expression
of emotions, moods, illocutionary acts, and many other elements which may be part
of this transfer process, as in (8).
(5) She said “I have been fired.”
(6) She said that she had been fired.
Communicative constructions in English and Spanish 
(7) She told me the news (= that she had been fired).
(8) She expressed her dismay.
Dirven et al. (1982: 3) distinguish three types of transferred information: the message,
which comprises direct enunciation (She said to me “I am 20”), indirect enunciation
(She told me that she was 20) and synthesis (She told me her age); the topic and the
speaker’s evaluation of the speech act (He told us a pack of lies). The topic is usually
introduced by prepositions like on, about, of, while message and evaluation tend to
appear as direct objects. I will use the term pro-dictum to include their message in
its variant “synthesis”, their topic and their evaluation, since on many occasions the
differences among the three types would require an analysis of the first speech act
and other elements of the communicative process which fall outside the scope of this
analysis. This abbreviation of the first speech act, or pro-dictum, will show interesting
syntactic and cognitive features, as will be discussed in the following sections.
The third participant, the recipient, is characterized by Goldberg (1995: 143–147)
as a “beneficiary or willing recipient”, and is sometimes optional. However, this partici
pant is inherent to any communicative process; without it there cannot be an exchange
of information. Vorlat (1982: 13) distinguishes between the receiver as a “mere recep
tor”, or as an “interactor”. It is beyond the present purpose to investigate interaction.
Summing up, the argument structure of speaking verbs contains a speaker, a
dictum and an optional recipient: verba dicendi <speaker dictum recipient>.5
. The process: Communication as transfer
Communicative events have also been grouped within a more general class of trans
fer events; the object transferred is the message, the source is the speaker and the goal
the listener: verbs of transfer <source theme goal>. For example, Jackendoff (1990: 266)
conceives the theme argument of the verb say as an entity, belonging to the category in
formation, which moves from a speaker to a receiver. Amberber (1996: 6) schematizes
Jackendoff ’s analysis in the following Lexical-Conceptual Structure:
(9) say
[Event CAUSE ([Thing i] A, [Event GO ([Info] A, [ FROM ([Thing i]])])] [Path TO [Thing] <A>]
(Amberber (15))
the verb say denotes an event in which the first argument is a Thing, i.e. the Causer and
the second argument is the Event. The embedded sub-event, designated by GO, has
two arguments: the entity that moves and the trajectory it traverses or the Path. The A
in sub-script stands for arguments which will be linked into syntactic positions. The
end-point of the trajectory – the argument of TO – is optionally A-marked capturing
the fact that the goal argument is not always present in the syntax (Amberber 1996: 7).6
Goldberg (1995: 127–128) observes that communicative verbs should be classified
together with verbs of instrument of communication as metaphorical classes, since
 Montserrat Martínez Vázquez
Sem CAUSE RECEIVE < agt rec pat >
R: instance, PRED < >
means
Syn V SUBJ OBJ OBJ2
Figure 1. Ditransitive construction
the information transmitted is conceived as being packaged and exchanged between

speaker and listener (following Reddy’s (1979) “conduit metaphor”). Communica
tion is viewed metaphorically as “traveling across from the stimulus to the listener”,
who “understands the communication upon reception” (1995: 148). Goldberg (1995)
illustrates this metaphor with the following examples:
(10) She told Joe a fairy tale. (1995: 148 (28))
(11) She wired Jo a message. (1995: 148 (29))
(12) She quoted Jo a passage. (1995: 148 (30))
(13) She gave Jo her thoughts on the subject. (1995: 148 (31))
This transfer metaphor licenses the use of the ditransitive construction with the pat
tern in Figure 1.
However, as Pilar Ron (2003: 88, n. 6) points out, the patient of Goldberg’s
(1995: 112) ditransitive construction should be better labeled “theme” or “argument
which undergoes a change of state or location”. Notice, however, the special status of
this theme, which is different from the theme of other transfer events since when in
formation is transferred, the principle of exclusive location does not apply, as pointed
out by Jackendoff (1990: 27):
If Bill transfers information to Harry, by (21)7 we can infer that Harry ends up
having the information. But since information, unlike objects, can be in more
than one place at a time, Bill still may have the information too.
The idea of transfer is applied to both English and Spanish verbs of communication.
Gutiérrez Ordóñez (1999: 1876) gives the following list of Spanish “verbs of commu
nicative transfer” (verbos de transferencia comunicativa):
transmitir, decir, comunicar, anunciar, avisar, confiar, contar, exponer, enseñar,
narrar, dictar, recordar, solicitar, contestar, escribir, manifestar, notificar, repetir,
revelar, referir, sugerir, declarar, gritar, explicar, replicar. . .
Communicative constructions in English and Spanish 
Vázquez, Fernández, & Martí (2000) provide a wider list of communicative transfer
verbs for Spanish:
anunciar, comunicar, conferenciar, conferir, confesar, contar, contestar, conversar,
cotillear, cuchichear, chafardear, chismear, chismorrear, decir, declarar, dialogar,
dictar, discursear, discutir, divulgar, emitir, explicar, hablar, indicar, informar,
murmurar, musitar, narrar, notificar, predicar, pregonar, preguntar, clamar, re
latar, sermonear, susurrar, telefonear, telegrafiar, televisar, transferir, transmitir.
Vázquez, Fernández, & Martí (2000) list a group of verbs of manner of speaking
(cuchichear, murmurar, susurrar, etc.), whereas Gutierrez Ordóñez only includes one
member of this class, gritar. Besides, there is also space for the class of verbs of in
strument of communication (telefonear, telegrafiar, televisar), a class which is absent in
Gutierrez Ordóñez’s classification.
Suñer (2000: 569) considers three classes of verbs introducing direct quotes in
Spanish: verbs of saying, manner of speaking and thinking, which comprise a great
variety of subdivisions. Among verbs of communication she includes: preguntar, con
testar, decir, corregir, interrumpir, añadir, definir, aventurar, or confesar; and in her
class of manner of speaking verbs we find the following verbs: bramar, bufar, cacarear,
cantar, farfullar, gemir, gritar, gruñir, murmurar, recitar, rumiar, suspirar, susurrar and
vociferar.
Levin (1993) divides English verbs of communication into nine semantic classes:
1. VERBS OF TRANSFER OF A MESSAGE: ask, cite, pose, preach, quote, read, relay,
show, teach, tell, write.
2. TELL (neutral).
3. VERBS OF MANNER OF SPEAKING: babble, bark, bawl, bellow, bleat, boom,
bray, burble, cackle, call, carol, chant, chatter, chirp, cluck, coo, croak, croon, crow,
cry, drawl, drone, gabble, gibber, groan, growl, grumble, grunt, hiss, holler, hoot,
howl, jabber, lilt, lisp, moan, mumble, murmur, mutter, purr, rage, rasp, roar, rum
ble, scream, screech, shout, shriek, sing, snap, snarl, snuffle, splutter, squall, squeak,
squeal, squawk, stammer, stutter, thunder, tisk, trill, trumpet, twitter, wail, warble,
wheeze, whimper, whine, whisper, whistle, whoop, yammer, yap, yell, yelp, yodel.
4. VERBS OF INSTRUMENT OF COMMUNICATION: cable, e-mail, fax, modem,
netmail, phone, radio, relay, satellite, semaphore, sign, signal, telephone, telecast,
telegraph, telex, wire, wireless.
5. TALK VERBS: speak, talk.
6. CHITCHAT VERBS: argue, chat, chatter, chitchat, confer, converse, gab, gossip, rap,
schmooze, yak.
7. SAY VERBS: admit, allege, announce, articulate, assert, communicate, confess, con
vey, declare, mention, propose, recount, repeat, report, reveal, say, state, blab, blurt,
claim, confide, declare, note, observe, proclaim, reiterate, relate, remark, suggest.
8. COMPLAIN VERBS: boast, brag, complain, crab, gripe, grouch, grouse, grumble,
kvetch, object.
9. ADVISE VERBS: admonish, advise, alert, caution, counsel, instruct, warn.
 Montserrat Martínez Vázquez
As Levin (1993) observes in her introduction, these semantic classes described in sec
tion II of her book have emerged from her study of diathesis alternations in Section 1.
Therefore, even though they are “semantically coherent classes”, they are also grouped
together because their “members pattern in the same way with respect to diathesis
alternations and other properties” (1993: 17). But an exclusively verb-centered per
spective leads her to exclude from this chapter of “verbs relating to communication
and the transfer of ideas” (p. 202) certain communicative constructions with non-
communicative verbs, such as, for example, those grouped under the heading “reaction
object construction”, in Section 1, which are paraphrased as “express (a reaction) by
V-ing”, found with manner of speaking verbs and verbs of gestures and signs, as, for
example, She mumbled her adoration.
I will group seven of Levin’s (1993) nine classes under the rubric verba dicendi.
Verbs of instrument of communication and manner of speaking will be discussed in
the section of constructions without verba dicendi, since their communicative value is
metonymically derived from their original meaning.
. Constructions with verba dicendi
The most general verbs in Spanish are decir and contar. The former is considered by
Cano (1987: 207) as an “archilexeme” (“archilexema”).8 The recipient is introduced by
a, (14), but it may also be substituted by a clitic pronoun, (15). If the dictum is also
pronominal the recipient will appear as se, (16).
(14) Dijo la verdad a su madre.
‘(s/he) told the truth to her mother.’
(15) Le dijo la verdad.

(s/he) her told the truth
‘She told her the truth.’
(16) Se la dijo.
se it told
‘She told it to her.’
This clitic pronoun is optional when the lexical Recipient is in a postverbal position,
(17), but if it precedes the verb the clitic must be present, (18).
(17) (Le) dijo la verdad a su madre
(s/he) (her) told the truth to her mother
‘She told the truth to her mother.’
(18) A su madre le dijo la verdad.
(s/he) to her mother her told the truth
‘To her mother she told the truth.’
Communicative constructions in English and Spanish 
According to Vorlat (1982: 165), say and speak would probably be the verbs chosen by
informants to be the most basic verbs of communication in English (1982: 32), even
though the first is much more frequent than the latter. In their corpus-based study of
speak, talk, say and tell, Dirven et al. (1982: 165) describe the first two as verbs focusing
on the linguistic action itself, whereas say and tell have message focus.
The verb tell takes a recipient,9 while say is used with and without this optional
participant. This verb is most productive with sentential complements or quotations:
only four from one thousand sentences randomly extracted from the BNC contained
the verb say with a nominal non-sentential object. All of them are pronominal:
(20) Look, Dennis obviously ain’t said nothing to him about your family. . . (BNC)
(21) . . .as having anticipated elements of psycho-analysis in what he said on this.
(BNC)
(22) This was said so low as to be virtually inaudible. (BNC)
(23) The same source, it must be said, also believes that . . . (BNC)
The Spanish verb decir with an NP object also occurs with modest frequency. From 127
occurrences randomly taken from García Márquez’s Crónica de una muerte anunciada
(CMA), only 12 examples occur with an NP object (9.4%), as in examples (24)–(27).
Another random search, this time from the oral component of the CREA corpus,
shows similar results: from 100 occurences only 17 appear with NP objects, all of which
are pronominals of the type lo que dijo, lo dijo, se lo digo, etc.
(24) Tres personas me dijeron la misma cosa. (CMA)
three people me told the same thing
‘Three people told me the same thing.’
(25) No nos dijeron la verdad. (CMA)
‘(they) did not tell us the truth.’
(26) Cuando me dijeron la noticia. . . (CMA)
when (they) me told the news
‘When they told me the news. . .’
(27) A mí nunca me dijeron el nombre de mi enfermedad. . .(CMA)
(they) to me never me told the name of my illness
‘They never told me the name of my illness.’
The mobility of the participants in the clause permitted in Spanish and illustrated in
the last examples contrasts with the rigidity of the English SVO pattern. In the follow
ing example, the translator makes use of the passive in order to keep the same thematic
order as in the Spanish original:
(28) “Hablaba con el alma en la mano” me dijo el doctor Dionisio Iguarán, que
estaba jugando con ellos. (CMA)
 Montserrat Martínez Vázquez
(29) “He was speaking with his heart in his hand,” I was told by Dr. Dionisio
Iguarán, who was playing with them. (CDF)
With quotes the subject is postponed in Spanish, whereas it usually keeps its preverbal
position in English:
(30) “Nunca la dejaba cargada”, me dijo su madre. (CMA)
(31) “He never left it loaded,” his mother told me. (CDF)
(32) “No seas bárbara” le dijo él. (CMA)
(33) “Don’t be a savage,” he told her. (CDF)
(34) “Tenía ese color verde de los sueños”, le dijo Pura Vicario a mi madre. (CMA)
(35) “He had that green color of dreams,” Pura Vicario told my mother. (CDF)
However, as Suñer (2000) notes, English subjects may also be postverbal with quota
tions (36a), but they can never be free-inverted, (36b), as in Spanish, (36c).
(36) a. “I’d like some more”, requested John from his friend. (Suñer 2000 (11a))
b. *“I’d like some more”, requested from his friend John. (Suñer 2000 (11c))
c. Nosotros, que bajemos – le decía a Gumersindo el guardia
we, should go down – him told to Gumersino the guard
joven. (Suñer 2000 (10b))
young
‘We should go down – the young guard told Gumersindo.’
When the recipient is not expressed, say is preferred.

(37) “Tiene el nombre bien puesto,” dijo. (CMA)
(38) “She’s well-named,” he said. (CDF)
Both say and decir are frequently attested in impersonal expressions, especially in the
passive:
(39) Bayardo San Román se había hecho muy amigo nuestro, amigo de tragos,
como se decía entonces, y parecía muy a gusto en nuestra mesa. (CMA)
(40) Bayardo San Roman had become our very good friend, a friend of a few
drinks, as they said in those days, and he seemed very much at ease at our
table. (CDF)
(41) His condition was said to be, “fair”. (Brown)
(42) All the sins of ancient Rome are said to be collected into this three-hour film.
(Brown)
(43) It was said that he had had a vision. (Brown)
Even though say may take a recipient, it does not allow the dative alternation. In fact,
as Levin (1993: 46) points out, only verbs of transfer of a message and instrument of
Communicative constructions in English and Spanish 
communication take the dative alternation. According to Bresnan & Nikitina (2003)
the most frequent verb that appears in the dative construction is tell. In 99% of their
corpus examples it occurs in the dative NP construction, a syntactic pattern which is
associated with possessive semantics, more precisely, change of possession. The dative
PP construction, on the other hand, is related to allative semantics or motion of an
object to a goal. This would mean that the recipient does not necessarily understand
the dictum, therefore does not “possess” it, while in the dative NP construction the
recipient decodes, and therefore, “possesses” the dictum. Thus, Krifka (1999) explains
that a sentence like Beth told her sorrows to God “presupposes that God exists”, whereas
Beth told God her sorrows “could be uttered by an atheist”.
. Constructions without verba dicendi
. Discharge verbs
Another lexical domain which may express verbal communication comprises verbs
like spread, hurl or Spanish soltar, lanzar, etc., which involve an idea of motion through
space. The presence of a dictum makes us decode them as communicative processes:
(44) I never heard them spread any gossip about anybody. (Brown)
(45) Just as now anyone may hurl insults at a citizen of Mars. . . (Brown)
(46) Luego, más calmado, me soltó una larga perorata sobre sus creencias. (CREA)
‘Later, more calm, (he) let out a long speech about his beliefs’
(47) El presidente del Gobierno lanzó un mensaje a ETA. (CREA)
the president of the government threw a message to ETA
‘The president issued a message to ETA.’
Sometimes within this frame the source is profiled, and is viewed as a container –
the mouth – metonymically understood to discharge words. This class is related to
manner of speaking verbs, since it usually also involves sound. As Rudzka-Ostyn
(1988: 514) observes:
The domains of acoustics and space frequently interact when they extend
into speech acts. The interaction is brought on either by a spatial particle or
preposition. . ., or a verb denoting discharge of some substance:
(9) Come on, cough it up, we know you are guilty.
For example, verbs like cough, spit, gasp, or sigh may appear as speaking verbs:
(48) Rather coughed his orders out. (BNC)
(49) “You’re from MID,” the old man spat out his accusation. (BNC)
(50) Yeremi at least gasped out an apology to the body below him. (BNC)
(51) She sighed a dirty word and left. (Brown)
 Montserrat Martínez Vázquez
Likewise, in Spanish toser, escupir, resoplar and jadear include dicta in the following
examples:
(52) En el castigo hallará la penitencia – tosió. (CREA)
‘In punishment (s/he) will find penitence – (s/he) coughed.’
(53) Le escupió mil improperios. . . (CREA)
him (s/he) spat a thousand insults
‘She spat a thousand insults at him.’
(54) ¡Ya estamos otra vez! – resopló. (CREA)
‘Here we come again! – (s/he) snorted.’
(55) Nos van a ver, amor – jadeó la muchacha. (CREA: CHILE)
‘(They) are going to see us, love – panted the girl.’
. Cognitive verbs
A linguistic action may also be coded as a cognitive process with verbs like rea
son/razonar or speculate/especular. However, the use of this domain to express com
munication is quite limited. In English, Rudzka-Ostyn (1988) mentions muse, reflect,
speculate, and reason as verbs of speaking, but only 14 out of 700 examples of com
municative constructions belong to this cognition domain. Indeed the productivity of
these verbs as compared to other speaking verbs is poor. A quick look at the CREA
corpus reveals a similar distribution in Spanish. Thus, while there appear 100,746 oc
curences of dijo and 1,156 cases of murmuró; only 87 examples of razonó and 133 of
especuló have been attested. Razonar is closer to the verba dicendi class than especular.
Thus, the former appears in 60.9% of the extracted examples with quotations, while
the latter only introduces direct speech in 8.2% of the examples. Razonar may ap
pear with nominal objects (su voto, su decisión, su optimismo), but especular introduces
the dicta mainly through the preposition con (38.3%). Both of them may introduce
indirect speech:
(56) La Sala razonó que,. . ., el fiscal debía concretar. . . (CREA)
‘The courtroom reasoned that, . . ., the attorney should specify. . .’
(57) Antrás especuló que dicho bufete podría solicitar. . . (CREA)
‘Antrás speculated that the aforementioned lawyer’s office could ask for. . .’
. Manner of speaking verbs
A much more productive group of verbs in communicative constructions is the class of

sound emission verbs, which extend their meaning to express manner of speaking. In
a detailed analysis of these verbs, Zwicky (1971) proposed twenty properties to single
them out. However, Mufwene (1978) proved that these properties were not exclusive to
manner of speaking verbs. He observed that the three structures in (58)–(60) share the
Communicative constructions in English and Spanish 
incorporation of an adverbial component “WHILE SHRIEKing”, but their syntactic

and semantic differences derive from their different subordinate structures.
(58) Martin shrieked that there were cockroaches in the caviar.
(59) Martin shrieked how we could free him from the trap.
(60) Martin shrieked to Pierre to fetch a nurse.
When manner of speaking appears as the adverbial component of the verbal entry
they are used transitively. But this behavior is also shared by manner of motion verbs
like float, bounce, roll etc., a phenomenon analyzed under the heading of “confla
tion patterns” (see Talmy 1985). Mufwene (1978: 286) suggests the following rule of
English Grammar:
. . .incorporation of some semantic material corresponding to an existing verb

as the adverbial component of another lexical entry (may) impose(s) on the
lexical entry the morpheme commonly associated with the adverbial compo
nent. Although the statement remains subject to verification, it demonstrates at
least that ‘manner-of-speaking’ is not the responsible factor here, while ‘manner’
probably is.
This manner component of the manner of speaking verbs is included internally in the
Lexical-Conceptual Structure Amberber gives for the quotative verb say.
(61) say
V
[Event GOsay ([Thing] A, [I. MANNER] A, [Path] <A>)]) ] (Amberber (18))
Amberber claims that this verb has to be represented as a variety of the conceptual
function GOSAY with an unsaturated manner component (INNER MANNER). This
component is usually encoded lexically, but with the verb say it will be encoded syn
tactically by the quotation. This would explain why manner of speaking verbs, which
encode this INNER MANNER lexically, do not require the quotation.
According to Faber & Sánchez (1990) manner of speaking verbs should be placed
in a transition zone between speech sounds (sonidos articulados) and non speech
sounds (sonidos no articulados). They claim that the sound component in these verbs
is more central than their communicative value; the sound is “foregrounded” inside
the communicative frame. But it is precisely this adverbial component what differenti
ates the members of this class of verbs, the only common component is the underlying
verbal lexeme (speak, say something) (1990: 23–24).
Faber & Mairal (1999: 254) illustrate this “transition zone” between the domains
of SOUND and SPEECH, where they place manner of speaking verbs, with the
following examples:
(62) She screamed/yelled/screeched/shrieked when she saw the cockroaches in the
refrigerator. [SOUND] (Faber & Mairal 1999 (289a))
 Montserrat Martínez Vázquez
(63) She screamed/yelled/screeched/shrieked, “Look at the cockroaches in the re

frigerator!” [SPEECH] (Faber & Mairal 1999 (289b))
Under a constructional perspective, we may explain the fact that manner of speaking
verbs appear transitively by proposing that when the verb occurs in a communicative
construction it comes to denote the manner in which the action is performed. The
participant role of the verb (agent/sound maker) fuses with the agent argument of
the construction (speaker); the construction contributes the dictum argument. The
recipient is optionally expressed, as already noted, gained from the fusion with the
construction:
(64) I saw Humber whisper something to one of the junior sisters. . . (BNC)
The blending of sound emission verbs with the communicative construction involves
a metonymic process, which expands the possibilities of expression of different com
municative reports without enlarging our vocabulary of reporting verbs. Metonymy is
lexicalized in most of these sound-for-speech metonymies and results in polysemy.
Seto (1999) classifies this type of metonymy as a whole event-subevent metonymy,
which implies a whole-part type of reference of a temporal metonymy (“the referen
tial transfer between two contiguous temporal entities” (p. 107). He includes manners
and gestures also as source domains for this type of metonymy. We will deal with these
domains in Section 5.5.
Another semantic feature associated with manner of speaking verbs relates to a
syntactic phenomenon which has received a lot of attention: the dative alternation. In
clear contrast with verbs of transfer of a message and instrumental verbs, English man
ner of speaking verbs are found to be ungrammatical with the NP dative. As discussed
above, this syntactic difference would reveal a semantic different type of transfer. Com
municative constructions with the PP dative are viewed as the movement of an object
to a goal, the recipient; whereas in the NP dative construction an idea of change of pos
session is implied. Krifka (1999: 9–10) further argues that manner of speaking verbs
denote a homomorphism between speech production and transfer of information,
thus they refer to a non-movement event, and so do not allow the NP dative (*Ann
yelled Beth the news). Instrumental verbs are found in both constructions because
only the initial stage of the transfer is specified. However, Bresnan & Nikitina (2003: 7)
provide corpus examples of manner of speaking verbs in the NP dative construction:
(65) a. . . .she muttered him a hurried apology. . .
b. You just mumble him an answer.
c. . . .and whispered me the answer. . .
Bresnan & Nikitina (2003: 12) claim that these verbs express the same emitted sound
continuously accompanying the speech acts in both the NP and the PP dative con
texts, and that in the NP dative construction “they appear not to be grammatically
impossible, but just improbable” (p. 12). We will return to this point in Section 5.4.
Communicative constructions in English and Spanish 
.. Sound classes

In order to analyze this metonymical process that turns verbs of sounds into manner
of speaking verbs – the sound for speech metonymy – I will distinguish three classes
with differences in the degree of metaphorization.10
... Sounds emitted by human beings. A first level of conceptual abstraction which
expands the meaning of sound related verbs in both languages involves sounds pro
duced by human beings. Intransitive verbs like murmur and whisper – Spanish mur
murar and susurrar – involve an unexpressed dictum and may easily actualize it to
express manner of speaking. The only change in argument structure is the expression
of a participant, a dictum, which is already part of the process, “an elaboration of the
event itself ” (as Amberber 1996: 6 views the dictum of the quotative verb say, equating
it to a cognate object).
These verbs may take the three types of dicta defined in Section 3.1: quotation,
(66), subordinate clause (67) and NP pro-dictum, (68).
(66) ‘Get up!’ the Chetnik screamed. (Cobuild)
(67) One girl whispered that she couldn’t tell anyone at home. (BNC)
(68) . . . Juron murmured an oath. (BNC)
Faber & Sánchez (1990: 23) state that Spanish morphosyntax does not allow this con
flation process found in English:
Sin embargo el foco – el componente de “sonido” – es estable y traducible con

más o menos razonable facilidad, aunque sea necesario un cambio, obligado por
la morfosintaxis española, de la perspectiva semántica, e.g. “. . .shrieked that. . .”
“. . .dijo estridentemente que. . .”. (emphasis mine)
‘However the focus – the component of “sound” – is stable and reasonably easily
translatable, even though a change, dictated by Spanish morphosyntax, is neces
sary in the semantic perspective, e.g. “shrieked that. . .” “said stridently that. . .”’
In previous research (Martínez Vázquez 1998) I also noted the absence of this phe
nomenon in Spanish. Similar remarks about the impossibility of conflation of manner
and motion in Spanish have been made (cf. Martínez Vázquez 2001 and references).
However, the following occurrences show that this conflation of manner + commu
nication in the verb is not totally banned by Spanish syntax:
(69) a. Me susurró que necesitaba hablarme a solas. . . (CREA)
‘(s/he) whispered to me that (s/he) needed to talk to me alone.’
b. . . .gimió que presentía que iba a morir. (CREA)
‘(s/he) moaned that (s/he) felt that (s/he) was going to die.’
(70) . . . pero casi al mismo tiempo murmuró una disculpa (CREA)
‘. . .but almost at the same time (s/he) murmured an apology.’
 Montserrat Martínez Vázquez
There is indeed a rejection of these conflation processes in Spanish, but it seems to me

that it is of a cognitive rather than a syntactical nature. Spanish speakers avoid fusing
two predicates into one, but the conflation pattern is not syntactically prohibited (cf.
Martínez Vázquez 2001 for occurrences of conflation of manner and motion in Span
ish). Notice that even Faber & Sánchez (1990), who argue that illocutionary acts like
“order” or “ask” cannot be expressed through manner of speaking verbs in Spanish,
admit exceptions with gritar and susurrar.
. . .en español no es posible encapsular en un solo predicado o unidad significativa
el acto de habla de ordenar o solicitar junto con la manera de hablar (estridente,
gritando, etc.) (p. 22). . . de estos verbos sólo gritar(2), con valor estilístico fam.
(familiar), contiene en su estructura un acto de habla: mandar a alguien hacer
algo. . . (p. 27) . . . posiblemente para algunos hablantes susurrar (“Me susurró
que cerrara la puerta”).
‘. . . in Spanish it is not possible to encapsulate in a single predicate or unit of

meaning the speech act of ordering or requesting together with the manner of
speaking (strident, shouting, etc.) (p. 22). . . of these verbs only gritar(2), with
the stylistic value fam. (familiar), contains a speech act in its structure: ordering
someone to do something . . . (p. 27) . . . possibly for some speakers susurrar (“Me
susurró que cerrara la puerta”. ‘He/she whispered me to close the door.”
Indeed susurrar occurs in such uses, but I have also found instances of chillar and
murmurar:
(71) . . . le entregó una pequeña bolsa de cuero, susurrándole que no interviniese y
que repartiese aquellas monedas conmigo. (CREA)
‘(s/he) gave him a small leather bag, whispering to him that he should not act
and that he should share those coins with me.’
(72) . . .para acercarse a Mojarrita y susurrarle que volviera al día siguiente.
(CREA)
‘. . .to come closer to Mojarrita and whisper to her that she should come the
following day.’
(73) Pablo me chillaba que volviera al coche que lo dejara ya. (CREA)
‘Pablo screamed at me that I should come back to the car, that I should stop
it.’
(74) . . . murmuró que lo hiciera pasar. (CREA: ARGENTINA)
‘(s/he) murmured that I should make him come in.’
If the construction is attested, though with a modest frequency (only 11 occurences of
chillar, 1.02%, and 74 of murmurar, 0.33%, in the whole CREA) we cannot claim that
it is syntactically impossible, but just infrequent.
... Sounds emitted by animals. A second class of sound emission verbs which ex
tends metaphorically to include a sense of manner of speaking involves sounds uttered
Communicative constructions in English and Spanish 
by animals. English verbs like roar, howl, purr, bark or crow or Spanish bramar/rugir,
aullar, ronronear, ladrar or cantar/gorjear, which originally describe the sounds made
by lions, wolves, cats and cocks, respectively, have moved to a human domain: the
agent, an animal sound maker, thus, becomes a speaker. Contiguity seems to be the
process at the heart of this grammatical blending of the sound process – once trans
ferred to the human domain – with the communicative process: part of the action
(“roar”) is used to convey a broader sense (“communicate by roaring”). The dictum is
added to a process in which it was not originally involved.
In English the construction is mostly found with direct speech. In a random sam
ple of 158 occurrences of roar, 37 introduced quotations (23,4%) and only 6 (3,8%)
appeared with NP pro-dicta (his triumph, it, his rage, its approval, approval, protests) as
in (75). Another random sample of bark throws up the following results: 57 examples
with quotes and 2 with NPs, (76)–(77); and of 90 occurrences of purr, 34 (37%) were
manner of speaking verbs introducing direct speech, as in (78)–(79). No examples of
roar, howl, purr or bark as manner of speaking verbs with sentential complements were
found, and only two examples of such a use with crow were attested in the entire BNC,
(80) and (81).
(75) He stuck his head out of one of the windows of the coach, and roared protests
at the policemen who were desperately trying to cope with the traffic confu
sion. (BNC)
(76) The platoon commanders barked their orders to dismount. (BNC)
(77) The portiere barked some instructions at the boy. (BNC)
(78) “Now that would be marvelous,” purred Imelda, effusive, just as she always
was after a victory. (BNC)
(79) “Perfect,” crowed Mr. Yarrow.
(80) One week after our chat Jeff hit Ilona with divorce proceedings and everyone
in the art world crowed that they had always known it wouldn’t last.
(81) Tanjug, the Belgrade news agency which reflects official thinking in Serbia,
reacted to the Washington agreement by crowing that “Bosnia will go down
in history as a state that never existed.”
In Spanish this sound for speech metonymy is expressed in a similar way. As has
been observed in English, most occurrences appear with direct speech, as shown in
(82)–(83). Only two examples with bramar and one with ladrar appear with sentential
complements in the CREA corpus, (84)–(86), and occasional uses with NP pro-dicta
tend to be limited to metalinguistic words, as in (87)–(88), though I have found an
example with an object expressing mood, as illustrated in (89).
(82) ¡Al! ¡Albert! – rugió el Jefe. (CREA)
‘Al! Albert! roared the boss.’
 Montserrat Martínez Vázquez
(83) Y bramó: – ¡Largo de aquí, guarra!. . . (CREA)

‘And (s/he) roared – out of here, dirty pig.’
(84) . . .uno de ellos ladró que lo fusilarían si hablaba una palabra más. (CREA)
‘. . .one of them barked that (they) would execute him if he spoke one more
word.’
(85) Y si se empeña en bramar que no. . . (CREA: MÉXICO)
‘And if (s/he) insists on roaring that no, . . .’
(86) . . .hasta salir a la terraza bramando que no se fueran sin él. . . (CREA: CHILE)
‘. . .until he went out to the balcony roaring that (they) should not go away
without him.’
(87) . . .pasa su brazo en torno a mi cintura, ronronea palabras sin sentido. . .
(CREA)
‘. . . (he) puts his arm around my waist, purrs words without sense. . .’
(88) El pomposo obispo francés debió cantar y gorjear aquellas palabras tan
labradas y pulidas. (CREA)
‘The sumptuous French bishop must have sung and chirped those words so
embroidered and polished.’
(89) . . .mientras la gata ronroneaba su misteriosa felicidad en mi regazo. (CREA)
‘. . .while the cat purred her mysterious happiness in my lap.’
These conceptual mappings between sound emission and information transfer are ver
balized in similar ways in English and Spanish, as shown in the following analysis of
the different degrees of metaphorization with the verbs aullar/howl. The Spanish verb
aullar appears with its original sense in (90). Examples (91) and (92) illustrate both
metaphor (the conceptual association between a human and an animal domain) and
the sound for speech metonymy. In (93), we find a much more complex conceptual
process, which may be briefly explained in the following terms: the source domain
aullar – which suggests a sad sound, a wolf ’s lament – is mapped into the domain of
sickness through the use of a siren, which reveals the presence of an ambulance. This
vehicle, in turn, stands as a source domain for the patients driven in it.
(90) Un lobo aulló a lo lejos. (CREA)
‘A wolf howled in the distance.’
(91) ¡Estás loca! – aulló. (CREA)
‘You are crazy – (s/he) howled.’
(92) El Subjefe de la Policía aulló una orden. . . (CREA: MÉXICO)
‘The second-in-command of the police howled an order. . .’
(93) Una sirena aulló en la calle. (CREA)
‘A siren howled in the street.’
The same type of conceptual mappings give rise to similar metaphorical linguistic con
structions in English. The verb howl in (94) reveals the non-metaphorical sense of the
Communicative constructions in English and Spanish 
verb. In (95) the use of a metaphor transfers the animal sound to an inanimate domain;
(96) and (97) imply the use of the sound for speech metonymy, though in (96) we
should rather term it sound for manner of communication, since the agent is a
dog, capable of sending information, though not exactly by speaking.
(94) A dog howled in the distance. (BNC)
(95) Outside, the storm howled about the house as though a pack of wild wolves
were trying to get in. (BNC)
(96) So when I skirted sleeping farms where dogs howled alarm. . . (BNC)
(97) Amy had screamed for him to stop and, scooping up the kittens, had thrown
back her head and howled obscenities at the cruelty of human beings. (BNC)
Despite the fact that these conceptual processes are similarly encapsulated in English
and Spanish communicative constructions, we find that this transfer process seems to
be more regularized in English than in Spanish. This fact can be tested by looking at
the information supplied by dictionaries. Thus, while COBUILD includes senses of
roar, howl, purr, and bark as transitive verbs describing manner of speaking, neither
MOLINER nor DRAE acknowledge such a use with verbs like rugir, bramar, aullar,
ronronear, ladrar, or gorjear in Spanish. Curiously enough, however, there are transi
tive entries for the verbs rugir and bramar in the first corpus-based Spanish dictionary,
DEA (“decir [algo] rugiendo/bramando”). This could suggest that these transitive uses
are recent and have not yet been included in dictionaries which are not based on cor
pora. However, the DEA shows no entry for grajear, and ronronear does not show a
transitive use.
... Sounds emitted by inanimate beings. A third class of sound emission verbs
involves sounds produced by inanimate beings. A first correspondence is established
between an inanimate domain and a human one. For example, the sound made by
a trumpet – to trumpet/trompetear – in (98)–(100) is applied to a human domain; a
person making the sound of a trumpet, thus, speaking in a very loud voice. But the
verb is also associated to the idea of boasting, through a metonymic inference, since
trumpets are used to make announcements (instrument for action metonymy).
Finally, we find the metonymy manner for action (sound for speech), which is
common to all the manner of speaking verbs.
(98) A writer in the Town Planning Review trumpeted that train-sheds were now
obsolete. (BNC)
(99) . . .as Khrushchev set deadlines and trumpeted warnings in his attempt to
frighten the West. . . (BNC)
(100) . . . el que trompeteó con entusiasmo sus méritos literarios por los cuatro pun
tos cardinales de la ciudad de Guatemala. (CREA)
‘He who trumpeted with enthusiasm his literary merits throughout the four
cardinal points of the city of Guatemala.’
 Montserrat Martínez Vázquez
A similar metaphorical expression is attested in both languages with the use of thun
der/tronar, the sound of storms, used as a source domain to express communication.
In (101) and (102) the sound for speech metonymy gives rise to a direct speech con
struction. Metonymy is at the heart of the constructional meaning; metaphor makes
us understand this manner of speaking as one involving violence.
(101) “Donna Frizzell, I thought better of you!” she thundered. (BNC)
(102) – ¿Quién le ha sacudido a Prada? – tronó Poveda. (CREA)
‘Who has beaten Prada? – thundered Poveda.’
. Instrumental verbs
These verbs appear in a communicative construction through the activation of a dou

ble metonymic process. A first instrument for action metonymy will turn the noun
of the instrument into the verb expressing the action performed with it. The class of
instrumental verbs is larger in English than in Spanish. Morphology plays an impor
tant role in this difference. English instrumental verbs derive from the noun denoting
the instrument by a very flexible derivational process – conversion – which changes
the word class without changing form. This metonymical morphological process of
word formation is not very productive in Spanish. Thus, for example, the instru
ment for action metonymy, turns the noun teléfono into telefonear, but we do not
get *modemear or *satelitear from modem or satélite, just to mention some examples.
In fact, from the 18 English instrumental verbs of communication given by Levin (ca
ble, e-mail, fax, modem, netmail, phone, radio, relay, satellite, semaphore, sign, signal,
telephone, telecast, telegraph, telex, wire and wireless) only four have been attested in
the CREA corpus: faxear, telefonear, telegrafiar and radiar.
A second metonymical process will insert these verbs into a communicative con
struction, which contributes the dictum. This communicative construction is, however,
different from the one built with sound emission verbs. The former activates the
means for action metonymy, the latter the manner for action metonymy. This
difference is also perceived through their different use of the dative construction. As
stated above, while manner of speaking verbs tend to appear only with the NP dative
construction, the instrumental verbs may be inserted into both the NP and the PP da
tive syntax. Krifka explains that verbs of means of communication behave differently
from manner of speaking verbs since they do not express a homomorphism between
the causing event and the transfer of information; they only make reference to the
initial phase of the transfer process. He illustrates his point with the following example:
(103) Ann faxed Beth the results. Actually, Beth’s secretary got the fax, and he
phoned them to Beth. (Krifka (39a))
This would explain why instrumental verbs may appear also in the PP dative construc
tion (Ann faxed the news to Beth), whereas manner of speaking verbs, which fuse the
manner activity with the transfer event, cannot express the movement event that the
Communicative constructions in English and Spanish 
NP dative construction contributes (*Ann yelled Beth the news), and can only appear
in the PP dative construction (Ann yelled the news to Beth). Bresnan & Nikitina argue
that these instrumental verbs “almost always involve communication – that is transfers
of the possession of information” (2003: 15). According to them, since the most pro
totypical dative verb of communication, tell, is most frequently used in the dative NP
construction, and instrumental verbs usually appear in a communicative discourse, we
tend to assimilate them to tell and use them in both dative NP and dative PP syntax.
The means of communication which is closer to speaking verbs is telephone, since
this action involves direct communication between speaker and listener, guaranteeing,
thus, that the transfer of information would successfully reach the recipient. From 105
examples of telephone randomly extracted from the BNC, 35 examples – 33.3% – are
followed by a communicative process in a clause with a verb from the verba dicendi
class, as in (104) and (105). In 5 cases – 4.8% – the message follows in a reduced
phrase introduced by for (106) or with (107).
(104) Liz telephoned her parents and told them that she had been sick. (BNC)
(105) . . .she telephoned, saying “Thank you for your letter”. (BNC)
(106) . . .she telephoned for help from the M50 motorway. (BNC)
(107) The next day, Mr. Goodwin’s source telephoned him with information. . .
(BNC)
However, only 5 of the 105 examples with telephone appear in a communicative con
struction: 1 introduces a quotation, 3 take NP pro-dicta (a warning, congratulations
and his offer to recommend Ramsey to the Queen for Canterbury), and 2 are built with
subordinate infinitive clauses, (108) and (109).
(108) At last, in 1975, the wife of one of the prisoners, a former air force pilot,
telephoned the young woman’s mother to come to her house quickly. (BNC)
(109) However thirty-six hours before she died she telephoned for some of us to
visit. (BNC)
The other 94 examples do not show up in a communicative construction, that is,
even though telephones are used for communication purposes, the constructional
metonymy is not easily activated. When an instrument verb is inserted in a commu
nicative construction, the dicta contributed by the construction may take the form of
a pro-dictum (110)–(115), a subordinate clause, (116)–(117), or a quotation (118).
(110) She had seen neither Andrew nor Erica Pringle since Friday night and was
acutely aware that Moira Harris had never telephoned her traditional ‘thank
you’ for Friday’s disruptive dinner party, a discomforting lapse in Moira’s
impeccable etiquette that added to her uncertainty. (COBUILD)
(111) After some crucial minutes, nothing had happened. I began to get worried –
I was learning that collectivity had its drawbacks. Had the other man phoned
the warning? (COBUILD)
 Montserrat Martínez Vázquez
(112) You can telephone, fax or post your order using the enclosed order form.
(COBUILD)
(113) Two officers radioed his warning to control. (BNC)
(114) Then Linda rang and advised me not to Fax the apology to Africa . . . (ICE)
(115) Well then could it be to ring out a warning. (ICE)
(116) Miguel telephoned that you were on your way and that you were exhausted.
(BNC)
(117) “The man radioed that his engine broke,” Bellybutton grumbled. (BNC)
(118) A Welsh farmer telephone [sic] him: ‘It’s all there but the body is not up to
much. (COBUILD: MAGAZINE)
In Spanish the communicative construction with instrumental verbs is also attested,
but it seems to be even less conventionalized than in English. Thus, while COBUILD
introduces a communicative construction in the entry for telephone (He telephoned
a warning from London), there is no space for such constructions with telefonear in
DEA nor DRAE.
From the 830 cases of telefonear which appear in the CREA corpus, there is no
example with an NP pro-dictum; 6 examples are built with the subordinative pattern
(0.72%), as in (119)–(120); and 8 examples show direct speech (0.96%), as in (121).
(119) Me pareció bien pegarme un baño, telefonear a Nilda que la iría a buscar el
domingo. . . (CREA: ARGENTINA)
‘It seemed good to take a bath, telephone Nilda that I would fetch her on
Sunday.’
(120) A primeros de setiembre, Crucita telefoneó que llegaría en el rápido de Irún. . .
(CREA)
‘At the begining of september, Crucita telephoned that she would arrive in the
train from Irun.’
(121) Al día siguiente me telefonea: “Oye, he pensado que. . . ”
the following day (s/he) me telephones: “hear, I have thought that. . .
‘The following day s/he telephones me: “Look, I’ve been thinking that. . .”’
The examples under (122)–(124) show conventional uses of telegrafiar and radiar with
metalinguistic pro-dicta objects. In (125) the verb telegrafiar introduces direct speech.
The verb faxear is of recent incorporation, and there are only two occurrences of it in
the whole CREA corpus, one in a communicative construction, (126).
(122) . . .le dejó telegrafiar textos en código. (CREA)
‘. . .(s/he) telegraphed texts in code.’
(123) El Almirante Latorre radió un mensaje al final del día. . .
‘Admiral Latorre radioed a message at the end of the day. . .’ (CREA)
Communicative constructions in English and Spanish 
(124) Esa información se radió al mundo el 11-9-1973 por el Comando de Unidade.

‘This information was radioed to the world on 11-9-1973 by the Comando de
Unidade.’
(125) No obstante, aquel mismo día telegrafié a mamá: Me he comprometido con
Ena.
‘However, that same day I telegraphed mom: I have gotten engaged to Ena.’
(126) También le indicó que le iba a faxear la orden de cese. . . (CREA: PERÚ)
‘(He) also told him that he was going to fax him the cease warrant.’
The subordinative pattern seems to be, if not ungrammatical, at least less conventional,
and probably not accepted by most native speakers.
(127) . . . su antiguo marido (. . .) telegrafió que “la necesitaba”. (CREA)
‘. . .her former husband telegraphed that (he) “needed her’.”
(128) . . .a las 2245, radió que recibía fuego de artillería. (CREA)
‘. . .at 22:45, (he) radioed that he was receiving artillery fire.’
. Verbs of gesture
Gestures frequently accompany communicative acts, as illustrated in (129). This as

sociation between a gesture and a speech act facilitates the metonymic transfer which
extends the meaning of the verb to express communication, as in (130).
(129) The driver – I know now it was Keith – waved to say thanks and drove into
the car. (BNC)
(130) [he] pulled his bicycle to the ground, waved his thanks at the old farmer,
and. . . (BNC)
The gestures used to introduce dicta are related to the domain of the face/head or
hands, and they include processes like smiling, nodding, shaking the head or waving.
They may activate different varieties of the part for whole metonymy: means for
action or instrument for action. Metonymy has been described as an “abbrevia
tion device”, which “enables us to say things quicker, to shorten conceptual distances”
(Nerlich, Clarke, & Todd 1999: 362). It also requires a lot of linguistic, conceptual and
pragmatic “unpacking”. The shortening device that gives rise to communicative con
structions is permitted when there is enough extralinguistic information to decode the
linguistic expression unambiguously. For this reason, it is quite common to find verbs
of gestures introducing dicta which stand for the prototypical message associated to
the verbal process. For example, the act of nodding is directly related to an act of ap
proval, as seen in COBUILD: “If you nod, you move your head quickly down and up to
show that you are answering ‘yes’ to a question, or to show agreement, understanding,
or approval.” This might explain why expressions like to nod yes are often labeled as
formulae, and are, therefore, not given a proper grammatical analysis. However, this act
 Montserrat Martínez Vázquez
is not completely lexicalised, as the different forms it takes in the following examples
illustrate:
(131) Beach nodded his approval. (BNC)
(132) Ruth nodded her agreement. (BNC)
(133) Duncan nodded his affirmation. (BNC)
(134) Biff glanced at his Sergeant, who nodded confirmation. (BNC)
(135) Donna nodded assent. (BNC)
Besides, the verb nod is not only used with dicta expressing affirmation, (136)–(139),
neither is agreement transmitted only by nodding, (140)–(141).
(136) . . .a stranger stood up and nodded farewell. (Brown)
(137) She nodded her thanks and reciprocal sympathy, and then moved on. (BNC)
(138) Shamlou nodded his understanding. (BNC)
(139) He nodded recognition at Chantal. (BNC)
(140) and President of the Club, smiled his agreement. (BNC)
(141) . . . the dog yapping his agreement. (BNC)
Thanking is also a very brief communicative act, which, as has been observed with
the act of agreement, is an expected answer in certain types of situations. Smil
ing is a prototypical gesture that accompanies the expression of thanking, and may
metonymically encapsulate a verb of communication, as in (142)–(143), though other
less conventional ways of thanking can also be expressed through this construction, as
in (144)–(145).
(142) Sir John grinned his thanks. . . (BNC)
(143) Paige smiled her thanks and pulled out a chair, yet before she sat down. (BNC)
(144) . . .she cried, hugging Daddy and Mummy tightly and squeezing her thanks
right into them. (BNC)
(145) Tail wagging his thanks, he took the handle of the basket in his mouth. . .
(BNC)
Another conventional gesture is the act of waving in a farewell situation, (146), but it
is not linguistically limited to this expression; the gesture is also used to convey other
meanings, as in (147)–(151).
(146) Meredith waved him goodbye from the gate. . . (BNC)
(147) When he waved down a taxi, he saw that her hand. . . (BNC)
(148) . . .and waved a greeting at a couple by the bar. (BNC)
(149) Lillee waved a fisted salute. (BNC)
(150) Hayman waved the objection aside. (BNC)
Communicative constructions in English and Spanish 
(151) She waved her thanks to Joujou. (BNC)

The object inserted in these constructions acquires a complex status. We can no longer
choose between the three dicta we have found in other constructions – quotation, con
tent clause or nominal pro-dictum; only the latter form may be used with verbs of
gestures. As noted in Section 3.1, these nominal objects are the most complex type
of dictum, since they involve substitution of the exchanged information. However, it
is important to distinguish different degrees of recoverability of the original message,
ranging from pro-dicta which permit us to recover the exact original message, to more
complex pro-dicta, which simply give us an idea of the type of information transferred,
but which do not allow us to recover the precise message. For example, in a construc
tion like (152), we may easily deduce that the first speaker uttered the words thank you.
In (51), reproduced here for convenience as (153), the object is a metalinguistic word
which describes the type of message, but which no longer enables us to recover the
exact speech act. The antecedent of its ecstasy in (154) is even more difficult to replace;
we would have to imagine all the people in this legendary stadium uttering enthusi
astic expressions of all kinds, which are summarized with a common feature they all
share: a state of ecstasy. The relation between the pro-dictum and its antecedent be
comes even more relaxed when the means of communication is a gesture, as in (155).
We no longer have a first speech act but a paralinguistic act.
(152) Merrill murmured her thanks. (BNC)
(153) She sighed a dirty word and left. (Brown)
(154) Old Trafford roared its ecstasy. (ICE-GB)
(155) McCready waved his gratitude. (BNC)
Levin (1993: 98) groups the different objects in (152), (153) and (154) under the
rubric “reaction object” (see also Huddleston & Pullum’s 2002: 305 “object of conveyed
reaction”). She describes it in the following terms:
Certain intransitive verbs – particularly verbs of manner of speaking and verbs

of gestures and signs – take nonsubcategorized objects that express a reaction (an
emotion or disposition); possible objects include: approval, disapproval, assent,
admiration, disgust, yes, no. When these verbs take such objects they take on an
extended sense which might be paraphrased “express (a reaction) by V-ing,” where
“V” is the basic sense of the verb. For instance, She mumbled her adoration can
be paraphrased as “She expressed/signalled her adoration by mumbling.” Most
of the verbs that allow such reaction objects name activities that are associated
with particular emotions, and the action they name is performed to express the
associated emotion. (Levin 1993: 98)
Note, however, that objects like yes or no express linguistic acts rather than reactions.
Words and feelings often appear in a paradigmatic relation, as illustrated with the
following resultative constructions:
 Montserrat Martínez Vázquez
(156) . . .but the coroner just waved his words aside. (BNC)
(157) Oswin waved her worries aside. (BNC)
Since words are used to express both ideas and feelings, it becomes at times very diffi
cult to distinguish between the two. Thus, the objects in the following sentences may
stand for the expression of feelings or speech acts of the type “I appreciate” or “I agree”;
but, anyway, there is communication, or transfer of information in both cases.
(158) . . . and Emily smiled her appreciation.
(159) Levitt smiled polite agreement.
As has repeatedly been noted, Spanish tends to avoid the use of non-subcategorized
objects (see Martínez Vázquez 2001, and references). Instead of inserting a verb into a
foreign construction, Spanish prefers the use of two separate predicates. This seems to
be also the case in other languages. Brdar & Brdar-Szabó (2003: 261) show that Croat
ian and Hungarian, in clear contrast with English, also have a natural tendency to keep
predicates formally apart, which is facilitated by their richer morphosyntactic devices.
Dirven (1990) also arrived at similar results in his contrastive analysis of English and
German verbs of speaking; while German exploits its rich morphology to create new
lexemes, English makes use of syntactic and metaphorical devices.
In the following translations into Spanish, we find that the English verb nod either
splits into two verbs, asentir y soltar, neither of which denotes gestures, or becomes an
adverbial, con la cabeza, while a general verb is introduced, dar.
(160) I smiled and nodded a curt though not impolite good morning without lodg
ing a complaint. (AP)
(161) he sonreído y asentido y he soltado un breve aunque no inadecuado buenos
días, sin llegar a presentar la protesta. (AP)
(162) (I) nod my thanks. (AP)
(163) doy las gracias con la cabeza. (AP)
However, some writers make novel uses of this metonymical device, which somehow
shows that the construction is not totally incompatible with the Spanish syntax:
(164) César cerró la boca y cabeceó que sí, que orinaba estupendamente. (CREA)
‘Cesar shut his mouth and nodded that yes, that he urinated well.’
As I have suggested above, the constraint is not of a grammatical nature; there rather
seems to be a cognitive tendency to avoid such a blending in Spanish. We may claim
that this metonymy is not syntactically conventionalized, but we cannot deny that it is
a possible device for creative Spanish speakers, as seen in the following examples with
gestures (165)–(167), and with a sound emission verb, (168) (see also (89) above).
Communicative constructions in English and Spanish 
(165) Los pobres ejercieron su derecho democrático a protestar, llorar sus penas y
agitar su desesperación frente a las cámaras. (CREA: BOLIVIA)
‘The poor exercised their democratic right to protest, weep their sorrows and
agitate their despair in front of the cameras.’
(166) Miss Susan sonrió su aprobación y. . . (CREA: PUERTO RICO)
‘Miss Susan smiled her approval and. . .’
(167) ¡Ríen su alegría. . .! (CREA: VENEZUELA)
‘(They) smile their happiness. . .’
(168) Y Pola ladrando su bienvenida con aire desconfiado. (CREA)
‘And Pola barking her welcome with a distrustful air.’
I have analysed the way communicative events are encoded in English and Spanish.
Besides the so called verba dicendi other semantic domains introducing secondary
speech situations have been examined. General verbs denoting discharge or motion
through space (let/soltar) and cognitive processes (speculate/especular), appear in sim
ilar constructions in both languages. Both are transitive verbs; the first group fuses
with the ditransitive construction to gain the recipient/listener argument. Another do
main which blends with the ditransitive/communicative construction involves sound
emission verbs. The sound will metonymically express manner of speaking. We have
observed a metaphorical gradation in this construction formation process, ranging
from sounds belonging to a human domain to those produced by inanimate beings.
This sound for speech metonymy is coded in a parallel way in both languages,
though it seems to be more lexicalised in English. This point is reflected in the informa
tion supplied by dictionaries, where most English sound verbs show a communicative
meaning which is the result of the metonymy. Spanish verbs rarely show this extended
meaning in dictionaries, even though examples of the construction have been attested.
Instrumental verbs can also be inserted in communicative constructions both in
English and Spanish. The number of instrumental communicative verbs in Spanish
is quite restricted, probably due to morphological reasons. Paralanguage can also be
inserted in a communicative construction. English shows a variety of constructions
with verbs of gesture introducing dicta. This metonymical device is not limited to
conventional situations of the type kiss goodnight or nod yes. Novel uses extracted
from different corpora prove that it is a powerful abbreviation tool which is more
widespread than is generally believed. Communicative constructions with verbs of ges
tures and “reaction objects” are rare in Spanish, though it cannot be claimed that it is
a non-existent phenomenon in this language.
Summing up, the conflated communicative pattern, or in constructional terms,
these mismatches between the semantics of the verb and the semantics designated by
the construction, which are very productive in English, also occur, though with low oc
 Montserrat Martínez Vázquez
curence, in Spanish. I have tried to show that even though this grammatical metonymy
is not as productive in Spanish as it is in English, it is not banned by syntax, as has
been claimed. I would rather suggest that the low occurrence of this process in Spanish
reveals a cognitive preference of Spanish speakers to avoid syntactic metonymies.
Reddy points out that English conceptualizes communication in terms of the
conduit metaphor and that the thought process is biased towards this preferred frame
work. He claims that even though there are ways of avoiding these conduit “metapho
risms”, “this would still not free you from the framework” (1979: 299). We can likewise
claim that Spanish conceptualizes predicates individually, and has therefore a natural
tendency towards analytic frameworks. But this framework should not be seen as an
oppressive system from which it is difficult to escape; it is rather the easiest cognitive
and linguistic option for speakers. The Spanish blended predicates in communica
tive constructions which have been attested have to be explained as unconventional
uses produced by creative speakers, which may eventually become part of the general
framework.
Notes
* The research presented in this article is part of the projects “Sintaxis contrastiva inglés
español” (BFF 2000–1271) and “Metáfora y Metonimia en el Metalenguaje” (BFF 2003–04064)
funded by the Spanish Ministry of Science and Technology and the FEDER (Fondo Europeo de
Desarrollo Regional) Program.
. The following terms have also been used: “linguistic action verbs” (Verschueren 1985), “verbs
of saying/utterance/communication”, and “metapragmatic descriptors” (see Collins 2001: 304
n. 9). In Spanish I have come upon the following labels “verbos de transferencia communica
tiva” (Gutiérrez Ordóñez 1999), “verbos de comunicación verbal” (Cano 1987) and “verbos
dialogales” (Contreras 1988).
. The examples have been extracted from: Bret Easton Ellis’s American Psycho (AP) and its
translation into Spanish; and García Márquez’s Crónica de una muerte anunciada (CMA) and its
translation into English, Chronicle of a Death Foretold (CDF).
. I use the term “speech act” as synonymous with “communicative act”, which implies a locu
tionary act, without taking into account possible illocutionary forces encoded in the report.
. For a full account of the syntactic nature of quotations see Amberber (1996).
. The term “speaker” stands for both the first speaker and the reporter of the first speech act.
“Recipient” likewise refers to the receiver of either the first speech act or the reported speech.
. Amberber (1996: 6–9) further argues that the subject of the quotative verb say is an Agent,
which behaves as the argument of the unergative verbs (walk, dance, laugh); the event is
internally-caused, i.e. it can only be caused by the person engaged in it. But, unlike the unergative
verbs, it appears with another obligatory entity. This entity, however, is different from thematic
arguments, and is viewed rather as “an elaboration of the event itself ”, which will occupy the
same syntactic position as a cognate object (lower Spec VP).
. The inference rule referred to, (21), reads as follows:
Communicative constructions in English and Spanish 
At the termination of [EventGO ([X], [Path TO ([Y])])],

it is the case that [State BE ([X], [Place AT ([Y])])].
. “El más genérico de todos estos verbos [los verbos de comunicación verbal] es decir. Podría
considerársele el ‘archilexema’ de esta área semántica de verbos”. [“The most generic of all these
verbs [the verbs of verbal communication] is decir. It could be considered the ‘archilexeme’ of
those semantic area of verbs.”] (Cano 1987: 207).
. It is exceptionally used without recipient (19a–c), or even without dictum (19d):
(19) a. Mary-Ann, who was a bit of an innocent, told all. (BNC)
b. But his mother told the story over and over. . . (Brown)
c. But at the coroner’s inquest Delphine told a forthright story. (Brown)
d. “Have you told your editor?” (BNC)
. I use “metaphorization” as a hyperonym which involves the shifts of meaning produced both
by metaphor and metonymy. Goossens (1990: 328) comes to the conclusion that sound used as
a donor domain for linguistic action shows a “hybrid character”, since it gives rise to “metonyms
in some contexts, metaphors from metonymy in others and sometimes undecided between these
two interpretations in actual contexts.”
. There is an entry for crow as an intransitive manner of speaking verb.
References
Amberber, Mengistu (1996). “The transitivity of verbs of saying revisited”. In J. Camacho, L.

Choueiri, & M. Watanabe (Eds.), Proceedings of the Fourteenth West Coast Conference on
Formal Linguistics (pp. 1–15). Stanford: CSLI Publications.
Brdar, Mario & Rita Brdar-Szabó (2003). “Metonymic coding of linguistic action in English,
Croatian and Hungarian”. In K. U. Panther (Ed.), Metonymy and Pragmatic Inferencing (pp.
241–266). Amsterdam and Philadelphia: John Benjamins.
Bresnan, Joan & Tatiana Nikitina (2003). “On the gradience of the dative alternation”. Draft of
May 7, available at http://www-lfg.stanford.edu/bresnan/download.html
Cano Aguilar, Rafael (1987). Estructuras Sintácticas Transitivas en el Español Actual. Madrid:
Gredos.
COBUILD = Sinclair, John (Ed.). (1987). Collins Cobuild English Language Dictionary. London:
Collins.
Collins, Daniel E. (2001). Reanimated Voices. Speech Reporting in a Historical-Pragmatic
Perspective. Amsterdam and Philadelphia: John Benjamins.
Contreras, Hernán (1988). “Los verbos dialogales en inglés y en castellano”. Revista de Lingüística
Teórica y Aplicada, 26, 23–27.
DEA = Seco, Manuel (1999). Diccionario del Español Actual. Madrid: Aguilar.
DRAE = Real Academia Española (1970). Diccionario de la Real Academia Española. Madrid:
Espasa Calpe.
Dirven, René (1990). “Prototypical uses of grammatical resources in the expression of
linguistic action”. In S. I. Tsohatzidis (Ed.), Meanings and Prototypes. Studies in Linguistic
Categorization (pp. 267–284). London and New York: Routledge.
 Montserrat Martínez Vázquez
Dirven, René, Louis Goossens, Yvan Putseys, & Emma Vorlat (1982). The Scene of Linguistic
Action and its Perspectivization by Speak, Talk, Say and Tell. Amsterdam and Philadelphia:
John Benjamins.
Downing, Angela & Philip Locke (1992). A University Course in English Grammar. London:
Prentice Hall.
Faber, Pamela B. & Ricardo Mairal Usón (1999). Constructing a Lexicon of English Verbs. Berlin
Faber, Pamela B. & Jesús Sánchez (1990). “Semántica de prototipos: el campo semántico de
los verbos que expresan la manera de hablar frente al de los verbos de sonido en inglés y
español”. Revista Española de Lingüística Aplicada, 6, 19–29.
Goldberg, Adele E. (1995). Constructions. A Construction Grammar Approach to Argument
Structure. Chicago: The University of Chicago Press.
Goossens, Louis (1990). “Metaphtonymy: the interaction of metaphor and metonymy in
expressions for linguistic action”. Cognitive Linguistics, 1(3), 323–340.
Gutiérrez Ordóñez, Salvador (1999). “Los dativos”. In Ignacio Bosque & Violeta Demonte (Eds.),
Gramática descriptiva de la Lengua Española (pp. 1855–1930). Madrid: Espasa.
Huddleston, Rodney & Geoffrey K. Pullum (2002). The Cambridge Grammar of the English
Language. Cambridge: Cambridge University Press.
Jackendoff, Ray (1990). Semantic Structure. Cambridge, MA: MIT Press.
Jakobson, Roman (1960). “Linguistics and poetics”. In T. A. Sebeok (Ed.), Style in Language (pp.
350–377). Cambridge, MA: MIT Press.
Krifka, Manfred (1999). “Manner in dative alternation”. In S. Bird, A. Carnie, J. D. Haugen, & P.
Norquest (Eds.), Proceedings of the Eighteenth West Coast Conference on Formal Linguistics,
Chapter 21. CD-ROM edition. Stanford: CSLI Publications.
Levin, Beth (1993). English Verb Classes and Alternations. Chicago: The University of Chicago
Press.
Martínez Vázquez, Montserrat (1998). “Effected Objects in English and Spanish”. Languages in
Contrast, 1(2), 245–264.
Martínez Vázquez, Montserrat (2001). “Delimited events in English and Spanish”. Estudios
Ingleses de la Universidad Complutense 9, 31–59.
MOLINER = Moliner, María (1994). Diccionario de Uso del Español. Madrid: Gredos.
Mufwene, Salikoko S. (1978). “English manner-of-speaking verbs revisited”. In D. Farkas et
al. (Eds.), Papers from the Parasession on the Lexicon (pp. 278–289). Chicago: Chicago
Linguistic Society.
Nerlich, Brigitte, David D. Clarke, & Zazie Todd (1999). “Mummy, I like being a sandwich”. In
K. U. Panther (Ed.), Metonymy in Language and Thought (pp. 361–383). Amsterdam and
Noh, Eun Ju (2000). Metarepresentation in English. Amsterdam and Philadelphia: John
Benjamins.
Reddy, Michael J. (1979). “The conduit metaphor. A case of frame conflict in our language
about language”. In A. Ortony (Ed.), Metaphor and Thought (pp. 284–324). Cambridge:
Ron, Pilar (2003). “Los verbos de posesión en inglés y en español”. In M. Martínez (Ed.),
Gramática de Construcciones. Contrastes entre el inglés y el español (pp. 59–90). Huelva: G.I.
Gramática Contrastiva.
Rudzka-Ostyn, Brygida (1988). “Semantic extensions into the domain of verbal communi
cation”. In B. Rudzka-Ostyn (Ed.), Topics in Cognitive Linguistics (pp. 507–553). Amsterdam
Communicative constructions in English and Spanish 
Seto, Ken-ichi (1999). “Distinguishing metonymy from synecdoche”. In K. U. Panther &

G. Radden (Eds.), Metonymy in Language and Thought (pp. 91–120). Amsterdam and
Suñer, Margarita (2000). “The syntax of direct quotes with special reference to Spanish and
English”. Natural Language and Linguistic Theory, 18, 527–578.
Talmy, Leonard (1985). “Lexicalization patterns: semantic structure in lexical forms.” In T.
Shopen (Ed.), Language Typology and Syntactic Description, Vol. 3: Grammatical Categories
and the Lexicon (pp. 57–150). Cambridge: Cambridge University Press.
Vázquez, Gloria, Ana Fernández & M. Antònia Martí (2000). Clasificación Verbal. Alternancia
de Diátesis. Quaderns de Sintagma 3. Lleida: Universitat de Lleida.
Verschueren, Jef (Ed.). (1985). Linguistic Action: Some Empirical-Conceptual Studies. Norwood,
NJ: Ablex.
Vorlat, Emma (1982). “Framing the scene of linguistic action by means of speak”. In R.
Dirven, L. Goossens, Y. Putseys, & E. Vorlat (Eds.), The Scene of Linguistic Action and its
Perspectivization by Speak, Talk, Say and Tell (pp. 9–35). Amsterdam and Philadelphia: John
Benjamins.
Wierzbicka, Anna (1987). English Speech Act Verbs. A Semantic Dictionary. Sydney: Academic
Press.
Zwicky, Arnold (1971). “In a manner of speaking”. Linguistic Inquiry, 2, 223–232.
P III
Information structure
Incremental Functional Grammar
and the language of football commentary*
J. Lachlan Mackenzie
The article investigates the notion that the grammatical characteristics of spoken
utterances vary with the time pressure under which they are produced. Within
the framework of my Incremental Functional Grammar, which takes acts of
utterance to consist of one or more sequentially produced subacts (i.e. actions of
predication or of reference), the hypothesis arises that a TV football commentary
will display different degrees of grammatical complexity according to the amount
of time pressure felt by the speaker at different points in the broadcast. I take
time pressure to correlate with various subgenres of the commentary. The results
confirm the hypothesis but also reveal the role of formulaicity and
time-conditioned limited consultation of the semantic and structural modules of
the grammar.
. Introduction
Thinking takes time, and speech takes time. As simple introspection reveals, in most
situations we start speaking before we have completed thinking about what it is we
want to say. As a result we very often begin an utterance without being certain how it
will finish. Incremental Functional Grammar (IFG; Mackenzie 1998, 2000) is a vari
ant of Functional Discourse Grammar (FDG; Hengeveld 2004) which seeks to take
account of how the chronology of the mental composition of utterances impacts upon
their grammatical form. This article will examine, from the viewpoint of IFG, the
grammatical characteristics of English utterances produced under varying degrees of
time pressure. More specifically, it will consider the syntactic properties of a BBC com
mentary on televised highlights of an English Premier League football game; as we
will see, this genre was chosen because it reflects different amounts of time pressure
upon the commentator. In general terms, four different kinds of situation arise in this
context: (a) high time pressure, reflecting a rapid succession of incidents near either
goalmouth during which a goal is likely or actually scored; (b) moderate time pressure,
reflecting midfield play with no immediate threat to either goal; (c) low time pressure,
during action replays or the treatment of an injured player; and (d) no time pressure,
for example when the players are trotting onto the field before play commences. The
 J. Lachlan Mackenzie
hypothesis will be tested that there is a correlation between the grammatical complex
ity of commentators’ utterances and the degree of time pressure under which they
are operating at any specific juncture of the broadcast. Each utterance, we shall ar
gue, contains at least one ‘subact’ of reference or predication; time pressure will dictate
how much more is verbalized. It will be hypothesized that high time pressure corre
lates with holophrastic utterances; moderate time pressure with phrasal structures; low
time pressure with clausal structures; and no time pressure with clause combining. As
we shall see, the whole matter is rendered more intricate by the use of formulaic speech
(in the sense of Wray 2002) and by the presence of ungrammatical utterances whose
form is explicable in terms of the communicative exigencies of the moment.
. Functional Grammar
IFG and FDG are the most recent manifestations of Functional Grammar (FG), which
arose in the eighties and nineties as a model of the grammatical organization of human
languages. FG is fundamentally associated with the name of Simon C. Dik (1940–
1995), who developed the theory in a large body of work that appeared between 1977
and 1992. His major publications were Dik (1989) and Dik (1997); the latter two-
volume work appeared posthumously under the editorship of Kees Hengeveld, the
first volume being a minor revision of Dik (1989). Functional Grammar is functional in
positing that language is above all an instrument of communication between human
beings; Functional Grammar is a grammar in being concerned with explaining the
formal properties of languages. What is characteristic of FG (and indeed of related,
older approaches such as that of the Prague School) is that linguistic phenomena are
explained in terms of their instrumentality, i.e. the contribution they make to language
users’ attempts to influence one another by using language. The view is taken that the
structures we encounter in languages can be understood as having arisen from myriads
of communicative events. These structures have served communicators well, and for
this reason have sedimented into the grammars of those languages.
Dik’s FG (for fuller presentations of the theory see Siewierska 1991 and García
Velasco 2003) in practice if not in principle concerned itself with the sentence as the
largest unit of structure in the organization of a language. In Dik’s approach, each sen
tence encodes a meaning, which is represented in a precise formalism. This meaning
is not identical to the speaker’s intention; rather, the meaning is itself instrumental
in the speaker’s attempt to influence the addressee. The formalism combines vari
ous sorts of information in a single representation. This representation specifies the
major words in the sentence (the predicates), the semantic requirements of those pred
icates (the argument structure), the speaker’s perspective on the meaning of the whole
sentence (reflected in the syntactic functions Subject and Object), the speaker’s expec
tations concerning the addressee’s awareness of various segments of the information
imparted (reflected in the pragmatic functions Topic and Focus), and the various ways
in which the speaker can modify the various components of the meaning (operators of
Incremental Functional Grammar 
Tense, Aspect, Evidentiality, etc.). Even in the earliest versions of the theory, this led
to quite complex representations. For example, (1), in Dik’s (1978) model, would be
represented as in (2):
(1) The postman was bitten by the dog
(2) Past biteV (d1x1 : dogN (x1 ))AgFoc (d1x2 : postmanN (x2 ))PatSubjTop
where biteV , dog N and postmanN are predicates; the variables (x1 ) and (x2 ), with their
semantic functions Ag(ent) and Pat(ient), represent the argument structure of the
predicate bite; Subj(ect) is a syntactic function; Foc(us) and Top(ic) are pragmatic
functions; and Past, d and 1 are operators.
As the model evolved, so meaning representations became more and more com
plex, as researchers laid bare more and more meanings that had to be incorporated into
the structure. Dik (1989), for instance, introduced a huge range of operators needed to
account for the various semantic refinements found in noun phrases and prepositional
phrases (in FG, terms) across the languages of the world and proposes many subdivi
sions of the pragmatic functions Foc and Top, reflecting different shades of Givenness
and Newness. The most far-reaching innovation, however, was Hengeveld’s (1989) ar
gument that the underlying representation should consist of four nested layers. He
showed, for example, that adverbials (known more precisely as satellites in FG) apply
to different sections of a representation. Thus viciously in (3) applies only to the verb
biteV :
(3) The postman was bitten viciously by the dog
However, in the garden in (4) applies to the entire representation in (2).
(4) The postman was bitten by the dog in the garden
Then there are other satellites, for example reportedly, which do not apply to the event
represented in (2) but to the proposition entertained by the speaker: there must there
fore be a higher propositional layer to which such satellites can apply. Finally, satellites
such as Between you and me apply neither to the event nor to the proposition, but to
an even higher layer, that of the speech act. Hengeveld’s layering hypothesis (immedi
ately adopted in modified form by Dik 1989) has been shown by various scholars to
be relevant to many other phenomena (modality, subordination, nominalization, . . . )
and mutatis mutandis even within the term (Rijkhoff 2002). The result was a single
complex structure showing each sentence as a speech act, incorporating a proposition,
which itself contained a predication, the nucleus of which was a predicate; at each layer
a characteristic set of operators and satellites were located. This four-layered structure,
though complex, turned out to reveal many facts about the languages of the world and
to predict such tricky phenomena as adverb positioning and morpheme ordering in
complex words.
However satisfactorily the theory was developing, various linguists originally at
tracted to the functionalist principles of FG had difficulty with the fact that the model
 J. Lachlan Mackenzie
was concerning itself exclusively with artificial-sounding sentences such as (5), con
structed to make theoretical points.
(5) Between you and me, the postman was reportedly bitten viciously by the dog
in the garden
They began to call for a return to basics, to the study of the communicator-
with-language, who does not achieve his/her purposes through individual sentences
but through entire discourses. It was from this discussion (reported in Hannay &
Bolkestein 1998) that Functional Discourse Grammar (FDG) arose, again chiefly in
work by Hengeveld. His model (Hengeveld 2004) offers an account of the natural
language user that comprises four interacting components: grammatical, commu
nicative, contextual, and output. As becomes a linguist, Hengeveld focuses on the
first-mentioned of these, the grammatical component, which in turn contains four
levels (interpersonal, representational, structural and phonological), each of which is
internally characterized by the layering familiar from more traditional FG.
. Functional Discourse Grammar
Let us consider four respects in which FDG differs from orthodox FG. Firstly, whereas
FG started from the predicate and added structure in a bottom-up fashion, FDG has a
top-down organization, starting with the speaker’s intention and working its way down
to the articulation of the linguistic expression. Secondly, FDG takes not the sentence
but the discourse act as the basic unit of analysis, and as such can handle not only regu
lar clauses but also stretches of language larger than the clause as well as what have been
traditionally regarded as fragmentary utterances. Thirdly, whereas FG took a holistic
view, FDG assumes a tripartite model of the speaker, which reflects what psychologists
recognize as the three major steps in language production, namely conceptualization,
formulation and articulation. The first of these is modelled in a conceptual component,
which is not part of the FDG as such, but triggers its operation; the third is modelled
in an output component, which serves to turn the output of the FDG into audible,
graphic or signed form; but it is formulation that is the central concern of the FDG
itself, and this corresponds to the grammatical component of the overall model. These
three components, finally, all draw upon a contextual component, which is accessible to
both speaker and addressee and which contains their memory of the evolving discourse
world. Finally, FDG differs from FG, which was rightly criticized for concentrating too
much on semantics and pragmatics and paying too little attention to questions of syn
tax, morphology and phonology, in including a structural level and a phonological level
as part of the grammatical component. (For a critical but sympathetic discussion and
assessment of the development of FG into FDG, which also brings out the inadequacy
of traditional FG as a pragmatic theory, see Butler 2003.)
It will be clear that FDG owes a great debt to psychological work on language
production. The parallels with the findings of Levelt (1989) are plain to see. Levelt’s
Incremental Functional Grammar 
blueprint for the speaker, based on an extensive survey of the psycholinguistic lit
erature on speech production to which he has made a distinguished contribution,
recognizes the following “autonomous components” (1989: 27): a Conceptualizer that
generates preverbal messages; a Formulator that contains a Grammatical Encoder and
a Phonological Encoder; an Articulator that executes the phonetic plan, and a Speech-
Comprehension System that permits self-monitoring. This rapprochement with psy
cholinguistics is in keeping with one of FG’s long-standing self-imposed requirements,
that of psychological adequacy: “[a functional grammar] must relate as closely as pos
sible to psychological models of linguistic competence and linguistic behaviour” (Dik
1989: 13).
The relationship between grammar and the analysis of language production, how
ever, is a vexed one. Jackendoff (1997: 7–8) considers that there are three logically
possible stances: (a) one can claim that there is no relationship (a position he attributes
to generative grammar); (b) one can separate the processes of language production
and the grammar, but permit the former to ‘consult’ the latter; (c) one can hold that
language production and grammar are not distinct, with the grammar being embod
ied in processes of production (and presumably then no more than emergent). It is
the second position that I shall adopt in the present article. After all, neither of the
extreme positions (a) or (c), although each is logically defensible, is compatible with
the principles of Functional Grammar. Position (a) isolates structure from use and
therefore cannot contribute to clarifying how language helps speakers to achieve their
communicative purposes. Position (c) denies the permanence of grammar, either as an
autonomous entity (as in position (a)) or as a tool in communication (as in position
(b)), and therefore has nothing to say about the generalizations across different lan
guages that are so essential to the typological adequacy of Functional Grammar. The
version of FDG that I shall propose is known as Incremental Functional Grammar (IFG;
Mackenzie 1998, 2000; for related but distinct proposals in and around Functional
Grammar, see Gómez-González 2004 and Bakker & Siewierska 2004; for earlier func
tionalist work in this direction, see Brazil 1995; and for current formalist work with an
incrementalist orientation, see Kempen & Harbusch’s 2002 Performance Grammar).
. Incremental Functional Grammar
IFG shares its overall architecture with FDG. Accordingly, the grammatical component
comprises four levels: interpersonal, representational, structural and phonological.
What is distinctive of IFG is that the first and last of these are modelled as proce
dural, i.e. reflecting real-time sequencing. I shall assume that discourse is shown at
the interpersonal level as consisting minimally of a number of moves which occur in
chronological succession. These moves in turn consist of a succession of acts, which are
also sequenced in time. Many familiar phenomena, such as anaphora and cataphora,
topic-chaining, the iconicity of coordination, and so-called left and right dislocation
(although these are of course misnomers from an incrementalist standpoint; see the
 J. Lachlan Mackenzie
discussion of example (20) below), can only be understood if the model permits the
temporal sequencing of acts within temporally sequenced moves. IFG, as we shall
see, also applies the same principle of temporal sequencing to the subacts (to be de
fined below) that make up each act: each is seen as an increment upon the temporally
preceding one(s). At the phonological level, too, linear sequence is essential. Even in
so-called non-linear models of phonology the temporality of intonation patterns is
not denied (cf. the prosodic templates of Autosegmental Phonology or the prosodic
level of Metrical Phonology), and in segmental analysis, too, linear sequencing is es
sential to the understanding of assimilation, umlaut, etc. IFG thus preserves the FDG
interpersonal and phonological levels, but crucially adds the time dimension to their
operation.
In contrast with the interpersonal and phonological levels, the representational
and structural levels in IFG will form declarative, i.e. non-procedural or non-
incremental, levels of the grammar. These levels may be taken to exist in the language
user’s competence, ‘out of time’, as it were. This is in agreement with the traditional FG
position (cf. Dik 1989: 70) that the representational level is unordered with respect to
the ultimate linearization in expression. The knowledge modelled here is knowledge
of the relations that hold, within the predication, between the component predicates
and arguments. The structural level, too, will be modelled as declarative, indicating the
various templates, with left-to-right ordering, that the language offers for NPs, clauses,
etc. Note that Bakker & Siewierska (2004) go further and argue for an incremental view
of the structural level, to which they attribute a hierarchical structure.
The operation of an IFG may be sketched out as follows. When s/he is about to
produce language, the communicator performs a succession of mental activities which
will be modelled in the cognitive component. The results of these activities are sent,
as soon as possible, to the interpersonal level within the grammatical component. The
sequence of activation within this interpersonal level will thus reflect that of the men
tal activation that prompts the communication. Not only does the input to the level
arrive piecemeal, as the ideas emerge in the speaker’s cognition, but the interpersonal
level similarly, and in parallel, produces sequences of moves, acts and subacts, with
their respective operators. As soon as it is ready, a piece of the output of the inter
personal level can, as we shall see, be dispatched to the phonological level for rapid
articulation. This model mirrors our everyday experience, mentioned at the outset of
this article: we typically start to articulate our ideas before they are fully organized.
The two dynamic processes (thinking and speaking) are accordingly modelled as tak
ing place concurrently, but with thinking of course always having a head start over
speech: in other words, this is what lies behind the proposal that both the interper
sonal and phonological levels are built up gradually along the time dimension. Let us
consider this proposal in a little more depth.
In IFG, each discourse act consists of at least one subact, including – crucially –
the focal subact that represents the core of what the speaker intends to communicate.
In many situations, ranging from laconic interaction in a shop to calls for help in a
life-and-death emergency, human beings are satisfied with (or constrained by lack of
Incremental Functional Grammar 
time to limit themselves to) what is known as holophrastic formulation. Consider the
following imaginary exchange, not untypical of simple commercial encounters in my
experience:
(6) Salesperson: Morning.
Customer: Morning.
Salesperson: Yes?
Customer: 6 rolls.
Salesperson: White?
Customer: Brown, please.
Salesperson: Anything else?
Customer: No, thank you.
Salesperson: 1 euro 20.
Customer: Here.
Salesperson: Thank you.
Customer: Bye.
Salesperson: Bye.
Note that each move in (6) contains one or two acts (Brown, please, for example,
contains a major act followed by a subsidiary act of politeness), but that each act is
holophrastic in the sense of containing only one subact, which is necessarily focal to
that act. The focal subact will tend to be either one of predication (as in White?) or
one of reference (6 rolls), although it should be said that (6) displays several examples
of acts that are ritualized (Morning) or oriented solely to politeness (thank you). (IFG
assumes that all such fixed formulations of Expressive illocutions pass directly to the
phonological level; for a similar analysis of the exclamation Congratulations! in FDG,
see Hengeveld 2005.)
The strategy of IFG is to take the kind of minimal utterances that predominate in
(6) as the starting point for grammatical analysis. This angle of attack chimes of course
with the phylogenetic and ontogenetic origins of speech: it is assumed that prehistoric
speech was “constrained in word order, structured in small chunks, and embedded in a
richer nonspoken matrix of gestures and exaggerated intonation” (Deacon 1997: 363);
and it is well-known that the use of holophrastic utterances is typical of the first speech
of infants (and indeed of their caretakers; see Mackenzie 1998: 272–274). From this
starting point, we can go on to see longer utterances, in which more is offered than
just focal information, as embellishments of the basic holophrastic structure. A more
complex utterance will therefore contain the focal subact plus one or more subacts
with other communicative functions. Thus alongside the focal material that is acti
vated in cognition, there may also be topical information that is already active and
which presents itself for co-expression. As shown in Mackenzie (1998), the expan
sion of utterances beyond the kind of minimal formulations exemplified in (6) can be
understood functionally as arising from the desire, need or convention to be explicit
about such matters as topicality, anaphoricity, illocutionary force, discourse organi
zation, etc.; another reason for greater complexity may well be the desire to avoid
 J. Lachlan Mackenzie
ambiguity. The complex act (i.e. the utterance that is more than merely holophrastic)
is characterized by polyfunctionality: the speaker is using extra material, i.e. additional
subacts, to achieve other purposes than merely to reveal the focus of his/her thinking.
The interest for the grammarian then comes to reside in determining the conditions
under which these additional subacts occur and the linguistic consequences of their
inclusion. In holophrasis, the Focus stands (trivially) in utterance-initial position. In
more complex utterances, does the Focus remain in act-initial position, or is it rele
gated to a later (often act-final) position as it is overruled by other functions? We see
here a reflection of the debate between Givón’s principle of Task Urgency (1988: 275),
which calls for focal information to be clause-initial, and the expectation of clause-late
positioning of focal information derivable from the Principle of Functional Sentence
Perspective of the Prague School (for discussion, see Gómez-González 2001: 31–34).
Let me give an example from the corpus examined for this article. In (7) below,
we have a move that consists of two discourse acts. The first of these is holophrastic
in consisting of only a focused subact of reference (LJUNGberg; the focus is expressed
with nuclear accent, here represented by capital letters). The second contains two sub-
acts: the first, to judge by its intonational prominence, is a focused act of predication
(ROBBED), the second a non-focused act of reference (by Acuña). Having evoked
Ljungberg with the first subact, the speaker refrains from mentioning him again, for
example as a topic in the second subact, but goes straight to his Focus. An IFG repre
sentation of (7) at the interpersonal level is given in (8), where M = Move, A = Act,
R = Subact of Reference, and T = Subact of Predication; it should be read as repre
senting, from right to left, the time course of ‘thinking for speaking’ as reflected in
the grammar:
(7) LJUNGberg, ROBBED by Acuña
(8) (M1 : (A1 : (R1 : Ljungberg)Foc ), (A2 : (T1 : robbed)Foc (R2 : Acuña)))
The assumption behind this approach is that in each discourse act it is the focal sub-
act that is mentally activated first during production: as Butler (2003: 326) puts it, “it
is, after all, the getting across of the focused element which justifies the production
of the utterance”. The ordering of subacts at the interpersonal level is therefore taken
to reflect the sequence in cognition: in (7), for example, I assume that in the second
discourse act, the situation of a player (and thereby his team) losing control of the ball
is observed and cognized first, (rapidly) followed by recognition of the identity of the
player responsible. (7) would thus be an example of a move in which the order of ex
pression of the ideas corresponds directly to the order in which they probably will have
been activated mentally, and, as we shall see, this is not untypical of communication
under time pressure.
Does word order then always reflect thought order? Of course not: every language
has its own word-order preferences that are superimposed upon the natural sequence.
At the same time, there are reasons for believing that free word order languages are
those in which the sequencing of constituents follows the succession of subacts more
closely, especially in speech. It is generally recognized nowadays that free word order is
Incremental Functional Grammar 
a misnomer, reflecting an unwarranted bias towards the expectation that constituent

order should be fixed; and we now understand that the difference between free and
fixed is gradual rather than categorical. What is characteristic of such free word order
languages is that the order of constituents is more strongly influenced by the cogni
tive status of the pieces of information to be imparted and/or the speaker’s strategic
choice of starting point or finishing point. In IFG terms, the distinction thus reflects
the extent to which the structural level imposes its own order upon the sequence of
subacts. Recall that both the interpersonal and the phonological levels are modelled
as operating in real time, while the representational and structural levels, which ac
count for the semantic properties of propositional and entity-identifying expressions
and their organization as clauses and NPs respectively, work as a set of constraints
upon the two real-time levels, enforcing an arbitrary organization (e.g. SVO) upon the
flow of formulation and the flow of speech. Returning to (7), we can observe that the
first discourse act is holophrastic, and as such is modelled as being sent directly from
the interpersonal level to the phonological level for immediate expression; the second
discourse act is not holophrastic but in this article will be termed phrasal, consisting
as it does of more than one constituent but without clausal organization. Each subact
is sent off for immediate expression as soon as it is cognized. The first subact does not
pass through the representational and structural levels; the second, after the emission
of the Focus ROBBED, does consult the representational and structural levels for the
formation of the post-Focus phrase by Acuña, an Agent term at the representational
level and a prepositional phrase at the structural level.
. Application to data
Example (7) was drawn from a football commentary, a genre which is characterized
by time pressure because of the speed of the game that is being described. At certain
stages of the broadcast, the commentator will have no time, I assume, to consult the
declarative constraints of the representational and structural levels; as a result, his ut
terances will display a direct mirroring of the cognitive processes underlying speech.
In these special circumstances, there is simply no processing time in which he can
mould his utterances to the requirements of the two declarative levels. To check this
hypothesis and discover how the chronology of the composition of utterances impacts
on their grammatical form, I have inspected a transcription of a BBC commentary on
televised highlights of an English Premier League football (‘soccer’) game between Ar
senal and Newcastle United, held on 23 March 2002. There were two commentators:
the primary commentator was John Motson; the secondary commentator was Trevor
Brooking. The highlights presented about one-third of the entire game and included
(a) the beginning and ending of each half; (b) most major events in the game (at
tacks, corners, fouls, sendings-off, etc.); (c) action replays of many major events, with
a maximum of 3 each; and (d) images of the crowd and personalities in it.
 J. Lachlan Mackenzie
The genre in question (at least in its instantiation in the television programme
transcribed for analysis) can be divided into two major text types: event-related and
non-event-related. Event-related commentary occurs either during play (when it is as
signed to the primary commentator) or during dead-ball situations (assigned to either
commentator). During play, a distinction can be drawn between situations in which
there is a direct threat to either goal (attacks and corners) and midfield play (when
there is less immediate likelihood of a goal being scored). During dead-ball situa
tions, the commentary deals either with preceding events (e.g. the foul that has led
to a free kick) or with speculation about upcoming events (e.g. to whom the ball is li
able to go from the free kick). Non-event-related commentary is the exclusive province
of the primary commentator and takes the form either of background information or
of summaries. Background information here covers commentary on such matters as
the treatment of injured players (outside the field of play) or the behaviour of the
crowd, but also covers the introduction to the entire game, as the players enter the
field, with which the broadcast begins. The summaries are given after each half and
briefly enumerate the major conclusions to be drawn from what has transpired.
The preliminary assumption was made that the commentators would be under
different degrees of time pressure according to the (sub)type of commentary required
at different stages of the game. A commentator will be most pressured during actual
play: specifically, he will be under high time pressure when there is a rapid succession of
incidents near either goalmouth during which a goal is likely or is actually scored. He
will be under moderate time pressure during midfield play with no immediate threat
to either goal; after all, the ball is still being moved around rather rapidly. He will be
less pressured, operationalized as low time pressure, during dead-ball situations, and
when giving background information about injuries and the crowd. Finally, he will be
under no time pressure when introducing the programme, before play actually starts,
and summarizing at the end of each half. The hypothesis was tested that there would be
a correlation between the grammatical complexity, in the sense explained above, of the
commentators’ moves and acts and the degree of time pressure under which they were
operating at any specific juncture of the broadcast; that is, the extent to which speakers
can consult the representational and structural levels is dependent on how much time
they have available. If this hypothesis can be substantiated, it would provide evidence
for a model of the speaker in which consultation of the declarative portions of the
grammar is optional and dependent upon the availability of sufficient time.
Grammatical complexity was operationalized in terms of four categories. The cat
egory holophrastic was applied, in keeping with the discussion above, to discourse acts
consisting of only one subact, either of predication or of reference, e.g. (9) below –
henceforth all examples will be drawn from the transcription of the commentary.
Phrasal was applied to a succession of subacts (each expressed as a phrase) within one
discourse act without clausal organization, as in (10). Clausal was applied to discourse
acts in which the subacts were organized into clausal form, e.g. (11). And sentential was
applied to the result of clause combining, i.e. a move with more than one discourse act,
as in (12).
Incremental Functional Grammar 
(9) SHEARer
(10) A little overEAGer Ashley Cole
(11) Look at Pires running in from DEEP
(12) Well he was wide Open and finished by placing the ball precisely into the
corner for his THIRteenth goal of the season
The various subgenres of the commentary, reflecting different degrees of time pressure,
were distinguished as follows:
High time pressure
Attack/Corner
Moderate time pressure
Midfield
Low time pressure
Dead-ball (primary commentator)
Dead-ball (secondary commentator)
Background (injuries; crowd)
No time pressure
Summary
Introduction
The results of the analysis are summarized in Table 1.
. Discussion of the results
Generally speaking, the hypothesis was borne out, as can be seen by inspecting Table
1. The results in bold show the best-represented form of expression per text subtype.
Of the acts carried out under high time pressure (Attack/Corner), 75% displayed ei
ther holophrastic or phrasal expression. In IFG terms, these are analyzed as bypassing
the representational and structural levels. And where there was medium time pres-
Table 1. Results
Sentential Clausal Phrasal Holophrastic Total
Introduction 7 (41.2%) 4 (23.5%) 5 (29.4%) 1 (5.9%) 17

Background 2 (9.1%) 13 (59.1%) 5 (22.7%) 2 (9.1%) 22
Dead-ball (sec.) 9 (6.1%) 87 (58.8%) 36 (24.3%) 16 (10.8%) 148
Dead-ball (prim.) 6 (4.1%) 30 (46.2%) 24 (36.9%) 5 (7.7%) 65
Midfield 0 1 (7.1%) 8 (57.1%) 5 (35.7%) 14
Summary 0 2 (9.1%) 16 (72.8%) 4 (18.2%) 22
Attack/Corner 2 (2.1%) 22 (22.9%) 32 (33.3%) 40 (41.7%) 96
Total 26 (6.8%) 159 (41.4%) 126 (32.8%) 73 (21.0%) 384
 J. Lachlan Mackenzie
sure (Midfield), almost 93% showed less than clausal structure, with phrasal structure
being prevalent; here again we see a direct link between the interpersonal and phono
logical levels. Where there is low time pressure (Dead-ball and Background), it is the
clausal form that predominates; in IFG terms, the message involves consultation of
the representational and structural levels. And during the Introduction (no time pres
sure) we see that sentential and clausal expression are prevalent (64.7%); it may well
be that the text of the Introduction was prepared in advance for reading out, which
could explain the predominance of sentential expression.
Certain results in the table, however, require further discussion. The prediction
that summaries, which take place at the end of each half after the play has ceased, would
be produced under no time pressure is gainsaid by the results, which surprisingly show
that summaries do not differ markedly from Midfield play, 90.9% being either phrasal
or holophrastic. The explanation for this must be that the commentator is in fact still
under time pressure, not from the game (which is inactive) but from the producer of
the television programme, who needs to wrap up the broadcast quickly. This suggests
that the time pressure is not solely determined by external events (the course of the
football game), but also and perhaps above all by the nature of the genre ‘football
commentary’ itself. Indeed, the alternation of different degrees of syntactic complexity
during the game may be attributed to one of the main purposes of the commentary on
a televised game, namely to enhance the viewer’s sense of alternating excitement and
relaxation.
A second observation is that there is no subtype within the genre that lacks
holophrastic acts, although these are generally more frequent where time pressure is
an issue. This finding is boosted by the analytical decision to recognize formulaic units
(in the sense of Wray 2002) as unanalyzed discourse acts: each of these is a ready-made
sequence of words that can be used as a whole with a single function in the discourse.
Among the formulaic units in the corpus regarded as holophrastic acts were those
exemplified in (13):
(13) a. believe it or not
b. you know
c. thank you very much
(13c) was used twice to close off the commentary on the action replay of a goal; indeed
all of these formulaic units had some role to play at the interpersonal level. There were
also various formulaic expressions drawn from the jargon of football journalists, e.g.
those given in (14):
(14) a. acres of space
b. across the face
c. great bit of skill
There were also no subtypes within the genre that lacked what I have been calling
phrasal expression. Thus, the corpus contained examples of non-insertion of the cop
ula, as in (15):
Incremental Functional Grammar 
(15) Uriah RENNie in charge

(15) displays a sequence of NP and PP; note that FG has always held that copulas are
inserted by rule rather than being deleted or zero-realized in those languages, dialects
and genres in which they are absent (cf. Dik 1989: 166); in IFG terms, the level at which
such copula insertion would take place is simply bypassed. Another source of phrasal
expression was zero-expression of the existential particle there’s, as in (16):
(16) change of refeREE from the first match
FG assumes that existential there arises from the presence in underlying representation
of an unspecified locative predicate (Dik 1989: 178); from the viewpoint of IFG, that
must again be seen as a requirement of the representational level that is not consulted
in the production of an utterance such as (16). Finally, pro-drop, the non-expression
of the subject of finite verbs supposedly characteristic of languages like Spanish and
Italian (Canto) but not English (I sing), was also regularly encountered in the data,
and items containing this were accordingly analyzed as phrasal rather than clausal,
e.g. (17):
(17) spreads it WIDE
Examples were plentiful of utterances whose structure strongly suggested that they
were composed as the commentator was taking account of events occurring before his
eyes. These directly reflected the order of events, as can be seen clearly in (18):
(18) Here with a chance is Robert PiREEEEEEES
(where the repeated ‘E’ indicates exaggerated, excited lengthening of the final
syllable)
Pires was actually scoring as the commentator spoke the final word. Here word order
mirrors thought order: the attention-calling deictic here, the situation suggesting a
goal-scoring opportunity with a chance and then the identification of the player, who
indeed does score.
Time pressure occasionally appears to have prevented full consultation of the
representational and structural levels, even when the expression is clausal. Thus
(19) shows three subacts ordered according to cognitive activation, but not English
grammar:
(19) Near post goes in RoBERT
Much as with (18), we observe the commentator drawing the listeners’ attention to
the location (near post, i.e. the goalpost nearest to the current position of the ball);
then the movement (goes in) and finally the identification of the player. We also find
examples of what Dik (1978: 132ff.) called a ‘Theme Predication’ construction, where
an NP is uttered in one intonation group, followed by a clause in another, with at least
a loose relation of relevance between the two. Dik claims that in English the relation
of relevance is normally expressed through coreference as in (That guy, is he a friend of
 J. Lachlan Mackenzie
yours?), whereas in Japanese or Mandarin Chinese the loose relation (without coref
erence) “may even have a certain dominance” (1978: 133). However, (20) displays no
coreference:
(20) The first CROSS, Distin does really well
In IFG this is analyzed as a succession of two discourse acts forming a move: the act
The first CROSS is holophrastic, containing only a subact of reference; the act Distin
does really well is clausal. Determining the relevance is left to the spectators’ observa
tion that it was Distin who made the cross, i.e. the (typically high) pass in front of the
goalmouth. We may also observe how the commentator sometimes needs a few frac
tions of a second to recognize a player and (presumably for that reason) divides the
move into a two-act sequence like (21):
(21) He LANDS . . . first of all on the foot of DABizas
In the second act he breaks the rule of English according to which proper names occur
in the Saxon genitive (i.e. Dabizas’s foot), but as a result gains extra time without losing
fluency. As against these examples of ungrammaticality caused by time pressure, there
are also cases in the data in which, once the time pressure is off, the declarative levels
kick in to regularize the syntax. In examples (22) and (23) the ball is caught by the
Newcastle and Arsenal goalkeepers Given and Wright respectively; in both cases this
immediately defuses an attack, such that the pressure switches from high (attack) to
low (dead ball):
(22) GIVen! . . . DOES well
(23) Comfortably in the arms of Richard WRIGHT . . . who gets the role of David
SEAman
The first act in (22) is holophrastic, while the first act in (23) is phrasal, forms char
acteristic of speech under time pressure. With the pressure off, the commentator can
continue to complete the clause or to add a non-restrictive relative clause respectively.
. Conclusion
The conclusion seems justified, then, that time pressure has influenced the complexity
and structure of the utterances that make up the commentary examined in this arti
cle. The effects can be understood within an Incremental Functional Grammar that
permits two routes for speech production processes. Under medium and high time
pressure, lexical material activated at the interpersonal level is quickly sent off to the
phonological level for immediate articulation, with the representational and structural
levels being bypassed. Evidence for this was found in the incompleteness and sometimes
even ungrammaticality of many utterances produced under these circumstances. Un
der less time pressure, clausal and even sentential formulation was the norm: in IFG,
Incremental Functional Grammar 
this involves the output of the interpersonal level being dispatched to the representa
tional level, where it is fitted into one of the propositional forms recognized in English,
and then to the structural level for formulation as a clause or indeed a complex sen
tence. The present investigation has thus yielded some insight into how a linguistic
model, by incorporating sensitivity to the elapse of time, can provide understanding
of certain grammatical characteristics of speech.
Note
* The research for this article was conducted in the framework of the research project Discourse
Analysis in English: Aspects of cognition, typology and second language acquisition, funded by the
Spanish Ministry of Education (grant reference BFF2002-02441), with further support from the
Xunta de Galicia (XUGA, grant reference PGIDIT03PXIC20403PN). The article appeared in
shorter form in British and American Studies (2004). My thanks to Chris Butler for his critical
and helpful comments on an earlier version of this article.
References
Bakker, Dik & Anna Siewierska (2004). “Towards a speaker model of Functional Grammar”.
In J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional
Brazil, David (1995). A Grammar of Speech. Oxford: Oxford University Press.
Butler, Christopher S. (2003). Structure and Function: A Guide to Three Major Structural-
Functional Theories, Part 2. Amsterdam and Philadelphia: John Benjamins.
Deacon, Terrence (1997). The Symbolic Species: The Co-Evolution of Language and the Human
Brain. Harmondsworth: Penguin.
Dik, Simon C. (1978). Functional Grammar. Amsterdam, New York and Oxford: North-Holland.
Dik, Simon C. (1989). The Theory of Functional Grammar, Part 1. Dordrecht and Providence:
Foris.
Dik, Simon C. (1997). The Theory of Functional Grammar, 2 Parts [Ed. K. Hengeveld]. Berlin
García Velasco, Daniel (2003). Funcionalismo y Lingüística: la Grámatica Funcional de S. C. Dik.
Oviedo: Universidad de Oviedo.
Givón, Talmy (1988). “The pragmatics of word order: predictability, importance and attention”.
In M. Hammond, E. Moravcsik, & J. Wirth (Eds.), Studies in Syntactic Typology (pp. 243–
284). Amsterdam and Philadelphia: John Benjamins.
Gómez-González, María de los Ángeles (2004). “Functional Grammar and the dynamics of
discourse”. In J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for
Functional Grammar (pp. 211–242). Berlin and New York: Mouton de Gruyter.
Hannay, Mike & A. Machtelt Bolkestein (Eds.). (1998). Functional Grammar and Verbal
Interaction. Amsterdam and Philadelphia: John Benjamins.
 J. Lachlan Mackenzie
Hengeveld, Kees (1989). “Layers and operators in Functional Grammar”. Journal of Linguistics,
25, 127–157.
Hengeveld, Kees (2004). “The architecture of a Functional Discourse Grammar”. In J. L.
Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional Grammar
(pp. 1–21). Berlin and New York: Mouton de Gruyter.
Hengeveld, Kees (2005). “Dynamic expression in Functional Discourse Grammar”. In K.
Hengeveld & C. de Groot (Eds.), Morphosyntactic Expression in Functional Grammar (pp.
53–86). Berlin and New York: Mouton de Gruyter,.
Jackendoff, Ray (1997). The Architecture of the Language Faculty. Cambridge, MA: MIT Press.
Kempen, Gerard & Karin Harbusch (2002). “Performance Grammar: a declarative definition.”
In M. Theune, A. Nijholt, & H. Hondorp (Eds.), Computational Linguistics in the Nether
lands 2001: Selected Papers from the Twelfth CLIN Meeting (pp. 148–162). Amsterdam and
New York: Rodopi.
Levelt, Willem J. M. (1989). Speaking: From Intention to Articulation. Cambridge, MA: MIT
Press.
Mackenzie, J. Lachlan (1998). “The basis of syntax in the holophrase”. In M. Hannay & A. M.
Bolkestein (Eds.), Functional Grammar and Verbal Interaction (pp. 267–295). Amsterdam
Mackenzie, J. Lachlan & María de los Ángeles Gómez-González (Eds.). (2004). A New
Architecture for Functional Grammar. Berlin and New York: Mouton de Gruyter.
Rijkhoff, Jan (2002). The Noun Phrase. Oxford: Oxford University Press.
Siewierska, Anna (1991). Functional Grammar. London: Routledge.
Press.
The role of Theme and Rheme in contrasting
methods of organization in texts
Michael Cummings
The Theme/Rheme distinction in Systemic-Functional linguistics connects the

content of successive Themes with the method of development of the text, that is,
its framing, and the content of successive Rhemes with the point of the text. that
is, its informational goals. Theme accordingly shows a potential for language
features appropriate to the method of development, e.g., reference chains. Rheme
has a complementary potential for lexical richness and variety. A quantitative
method produces measurements of both of these potentials. For example,
successive Themes in a text usually have higher proportions and much higher
densities of reference chain elements than do their Rhemes. Text segments which
reverse these proportions in favour of the Rhemes for some stylistic purpose are
markedly contrastive.
. Background and aims
Systemic-Functional linguistics has inherited a Theme/Rheme distinction from Prague

School linguistics, but sees clause Theme as a semantic orientation to the rest of the
message rather than given information as such; and in the case of English clauses,
it perceives Theme to be invariably realized as the initial element(s). In Systemic-
Functional approaches to English discourse, clause Theme is perceived to play a part
in a text’s method of development, i.e., its framing concepts. A contrasting part of a
text is its point, i.e., its informational goal; the Rheme elements in English clauses, es
pecially the final elements, or N-rhemes, play a role in the realization of the point of
the discourse similar to that played by Themes in the method of development. Sec
tion 1 of this article introduces the Systemic-Functional theories of Theme/Rheme,
method of development and point. Sections 2.1–2.2 present a procedure for a quanti
tative interpretation of the differences in language between realization of the method
of development of a text in its clause Themes, and the realization of the point of the text
in its N-rhemes. In Sections 3.1–3.2, the procedure is applied to explain stylistic dif
ferences among short written texts which take different approaches to the language of
their methods of development and their points. Section 3.3 looks at short written texts
which, for good strategic reasons, run absolutely contrary to the norms for English in
the quantitative distribution of features within Themes and N-rhemes.1
 Michael Cummings
Early on in M. A. K. Halliday’s development of Systemic-Functional theory,

Theme is defined as “what is being talked about, the point of departure for the clause
as a message” (1967: 212). A distinction is carefully maintained between this meaning
of Theme, and the lexico-grammatical means by which Theme is realized in English
clauses. In declarative mood clauses, Theme is most typically realized as the gram
matical Subject element. In yes/no interrogatives, Theme is most typically the Finite
element and the Subject both. In wh-interrogatives, Theme is most typically the initial
wh-element, regardless of its grammatical function otherwise. In imperative mood,
Theme is most typically the Predicator element; but if the imperative clause contains a
Finite before the Predicator (Do be quiet!), Theme will include them both; if a Subject
precedes the Predicator (You be quiet!), then the Subject alone is Theme. The principle
that underlies these various assignments of the Theme function is that the realization
of Theme extends from the beginning of the clause through to whichever element has
a propositional (experiential) role to play: this element (distinguished as the topical
Theme) once accomplished exhausts any potential for further Theme/Rheme contrast
at the beginning of the clause (Halliday 2004: 64–79).
In less typical clause structures, on the basis of this principle, a marked Theme may
be realized. In declarative mood clauses, for example, experiential elements other than
the Subject may precede the Subject. The most marked effect will be achieved by the
prepositioning of a Complement (Bananas I like). In this case it is the Complement
only which is Theme, and the Subject belongs to the Rheme. A marked effect with
lesser force can be achieved by putting the Predicator, or more likely, some Adjunct
element with experiential reference before the Subject; again the Subject will now fall
into the Rheme. Such an Adjunct may undertake initial position in any of the mood
types referred to above, and will in all such cases be Theme to the exclusion of any
other element from that role (Halliday 2004: 64–79).
Of course there are various initial elements in clauses besides Finites in yes/no in
terrogatives and in imperatives which do not play a part in the propositional content of
the message. Since these have no experiential content, they belong to the Theme stretch
without exhausting its potential so long as they precede an element which does have
experiential content. This makes for multiple Theme elements in the Theme stretch.
Comment Adjuncts and Vocatives (Frankly, my dear . . . ), also Mood Adjuncts (Never
. . . ), and Finites are all referred to interpersonal meaning, not to experiential meaning,
and so may belong to a Theme stretch which terminates only with the first experi
ential reference realized, i.e., the topical Theme. Similarly Conjunctions, Conjunctive
Adjuncts (Nevertheless . . .), and Continuatives (Oh. . .) are referred to textual meaning,
and so also may belong to Theme without exhausting its potential (Halliday 2004: 79–
87). The example in Figure 1 quoted from Halliday (1994: 55) illustrates multiple
elements in a Theme stretch at the beginning of an English clause.
The Theme/Rheme distinction extends to units of grammar both more inclusive
and less inclusive than the clause. In particular, Theme is found within the structure
of the sentence, seen as a clause-complex, i.e., a concatenation of clauses. A clause-
complex with an initial dependent clause will have that clause as a marked Theme;
The role of Theme and Rheme 
Figure 1. Multiple Theme elements in an English clause (Halliday 1994: 55)
the rest is Rheme. Should the dependent clause follow the independent clause, the
Theme of the independent clause is considered both the Theme of the clause and
the unmarked Theme of the entire clause-complex (Fries [1981] 1983: 121; Halliday
2004: 392–394).
An alternative point of view on the realization of Theme in English has emerged in
Systemic-Functional linguistics. In cases where a marked experiential Theme precedes
the Subject, an increasing number of linguists see the Theme as extending through
the Subject. This effectively revises the basic principle of Theme realization, which is
thus interpreted to mean that Theme in English extends from the beginning of the
clause through to the Subject element. The one exception is when the Subject fol
lows rather than precedes the verb (cf. Berry 1989: 71, 1995: 64, 1996: 35–46; Downing
1991: 127; Ravelli 1995: 219–226). A rationale for this view is provided by the theory
that textual meaning within the clause is principally realized at the initial element, the
Theme, and at the New Information element, typically final (cf. Halliday 2004: 87–94).
Between these two focuses of textual prominence, the Theme effect diminishes gradu
ally towards the centre, and the typical new information effect rises gradually from the
centre. Textual meaning may thus be metaphorically modelled as a wave-shape with its
trough somewhere in the middle of the clause, and peaks typically on the ends. Fail
ing to recognize the thematicity of a preverbal Subject when it is preceded by another
experiential element is to impose a categorical and segmental model of Theme some
what too zealously on a graduated phenomenon. Extending the Theme stretch beyond
the first experiential element through to the Subject element is to recognize the signif
icance of some thematicity at a lesser degree (Halliday [1979] 2002: 207–211, [1982]
2002: 233–234; Matthiessen 1988: 164–166, 170-171, 1992: 38–52, 1995a: 513–519; cf.
Halliday 1994: 336–337; and Martin 1992: 10–12, 1995: 225–227).
The meaning of Theme for English has also evolved within Systemic-Functional
theory, far beyond Halliday’s original formula of “what is being talked about, the point
of departure for the clause as a message” (1967: 212). Fries (1995a: 55) has proposed
that this definition should be taken as a metaphor, and that what is more literally at
stake is the role of Themes in the structuring of text (cf. Hasan & Fries 1995: xix; Fries
1995b: 4). Within some local and relatively unitary stretch of text, successive clause
and sentence Themes convey framing concepts for the text, especially if the structure is
relatively simple, perhaps following an outline logic. This is to “construct” the method
of development (Fries [1981] 1983: 116, 119, 121, 125, 135, 1995b: 9, 1995c: 324). In so
doing, the individual clause and sentence Themes fulfill the Hallidayan definition of
 Michael Cummings
Theme by providing a framework to their own clause and sentence messages, or an

orientation to these messages (Fries 1994: 234, 1995c: 318, 326, 2002: 125–126).
The method of development is characterized both by elements which provide a
local consistency to the text, and by elements which signal variation and develop
ment. Particularly in a text whose organization is relatively simple, successive Themes
can represent a consistency in field, marked by lack of variation in experiential mean
ings, and by lack of sub-structuring into semantic sub-fields (Fries [1981] 1983: 149,
1995b: 9, 1995c: 323–324). On rhetorical terms, this may be perceived as recurring mo
tifs within Themes which thus identify the field of the text (Halliday 1993: 95). Themes
also show a high incidence of presuming reference, which – notoriously in narrative
genre – may take the form of long chains which maintain thematic consistency over
large stretches of text (Fries [1981] 1983: 124, 1994: 230–231, 2002: 122–123). Of course
chaining may also take the form of lexical strings, or of 1st- and 2nd-person pro
nouns which are just as interpersonal in their purport as they are experiential (Martin
1992: 434–448). The consistency aspect of the method of development accounts for
the general (but not invariable) correlation of Theme with Given Information (Fries
[1981] 1983: 116, 144).
Consistency within the method of development is mirrored in structures which
extend beyond a local text segment. First, a text segment’s consistency may be predicted
by its Topic sentence or Hyper-theme; second, a set of Hyper-themes may be predicted
in turn by a Macro-theme, whose wording may take the form of a whole text segment
in itself (Martin 1992: 434–448, 1993: 244–247, 249–251; cf. Halliday 1985: 367).
While the consistency aspect of the method of development relies heavily on ex
periential Themes, the variation aspect is just as likely worded in the form of textual
or interpersonal Themes. Obvious instances are Disjunctives, like conjunctive Ad
junct however or conjunction but (Martin 1995: 244–245, 247–253). More elaborate
procedures may involve experiential Themes anyway, perhaps functioning to supply
first-time information by which to interpret the remainder of the clause or sentence
message. This may recontextualize a concept otherwise consistent with previous word
ing in the text. Perhaps it clarifies the local temporal and spatial framework. It may
signify the beginning of an elaboration of the concept (Fries 1995a: 58–62). On con
temporary Systemic-Functional terms, Theme in clause or sentence can represent con
junctive expansion, as elaboration, i.e., more about the same; extension, i.e., something
different; or enhancement, i.e., qualification (Matthiessen 1992: 60–66, 1995b: 26–40;
cf. Halliday 2004: 395–441; for method of development in general, cf. Gómez-González
2001: 98–100).
In terms of the wave-shape metaphor, the English clause is characterized by one
peak of textual prominence at its beginning, and, typically, by one peak of textual
prominence at its end. Thematic meaning focussed at the beginning of the clause is
balanced by the focus of New Information at the end of the clause. Just as the method
of development in a text segment is constituted by clause and sentence Themes at their
beginnings, so also is the Rheme-based point of the text especially constituted (but not
exclusively) by successive N-rhemes, i.e., the clause elements in the default position of
The role of Theme and Rheme 
the New Information element. The point of a text is then the assembly of the new in
formation which the text offers. The information which the point contains is different
in kind from the information presented in the method of development. For example,
its field of information tends to be more varied and less concentrated; accordingly its
realization is more lexical, making lesser use of pro-forms (Fries [1981] 1983: 128–129,
135, 1992: 464, 478–479, 1993: 338–339, 1994: 232–234. cf. 1995c: 351).
Just as the method of development in its consistency may be characterized by
a constancy of motif, so also may the point of a text have its own pattern of mo
tifs, amounting to an additional and different contribution to the structuring of the
text (Halliday 1993: 95–104). Furthermore, just as the method of development may
be predicted by Hyper-theme and Macro-theme, so may the point of some local text
segment be summed up in the form of a proposition serving as Hyper-new; and suc
cessive Hyper-news may in turn be summed up by a whole text segment which serves
as Macro-new to a whole piece (Martin 1992: 453–460, 1993: 247–251).
. Towards a procedure for a quantitative interpretation

of method of development
. The method of development and the point in an expository text
The Systemic-Functional approach to Theme/Rheme, method of development and

point can be demonstrated by scrutinizing a single expository text. Figure 2 is an ex
cerpt of three consecutive paragraphs from Brian Greene’s The Elegant Universe which
discuss Einstein’s view of gravity. The text in Figure 2 has been tabulated by numbering
consecutively each clause which has a Theme/Rheme structure.2 The Theme of each
such clause is bolded, and the N-rheme of each non-embedded clause is italicized,
which will necessarily include the whole of any following clause embedded within the
N-rheme – as in the case of clauses 7–9 embedded within the N-rheme of clause 6.
(Italics in the original text are represented here by underlining). Sometimes the whole
N-rheme is an embedded clause or clause-complex, e.g., clause 11 as the N-rheme of
clause 10 and clause 14 as the N-rheme of clause 13. Marked clause-complex Themes
are within square brackets (23, 26, etc.). The clause Theme always includes all elements
up to and including a Subject, making for more than one topical Theme in various
clauses, e.g., 2 and 4. Themes of embedded clauses will not factor in the analysis (Fries
2002: 139–140; Halliday 2004: 100).
Each of the three separate paragraphs adopts a different strategy in deploying the
resources of the Theme and Rheme. In clause 1 of paragraph 1, the Theme and Rheme
parts respectively contrast the presence of mass with its associated gravitational force.
This starts a chain of reasoning which will end in clauses 41–44 with the reinterpre
tation of gravity as a warping of space. Gravitational force is locally introduced in the
Rheme of clause 1, and becomes a Theme in clause 3, which locally introduces accel
erated motion in its own Rheme. Accelerated motion is repeated within the Theme of
 Michael Cummings
1 ¶A massive body like the sun, and indeed any body, exerts a gravitational force on other
objects.
2 In the example of the terrorist bomb, we learned
3 that gravitational forces are indistinguishable from accelerated motion.
4 In the example of the Tornado ride, we learned
5 that a mathematical description of accelerated motion requires the relations of curved
space.
6 These links between gravity, accelerated motion, and curved space led Einstein to the
remarkable suggestion
7 that the presence of mass, such as the sun, causes
8 the fabric of space around it to warp,
9 as shown in Figure 3.4.
10 A useful, and oft-quoted, analogy is
11 that much like a rubber membrane
12 on which a bowling ball has been placed,
(11) the fabric of space becomes distorted due to the presence of a massive object like the sun.
13 According to this radical proposal, space is not merely
14 a passive forum providing an arena for the events of the universe;
15 rather, the shape of space responds to objects in the environment.
16 ¶This warping, in turn, affects other objects moving in the vicinity of the sun,
17 as they now must traverse the distorted spatial fabric.
[Using the rubber membrane-bowling ball analogy,

18 if we place a small ball-bearing on the membrane
19 and set it off with some initial velocity,]
20 the path
21 it will follow
(20) depends on
22 whether or not the bowling ball is sitting in the center.
[23 If the bowling ball is absent,]
24 the rubber membrane will be flat
25 and the ball bearing will travel along a straight line.
[26 If the bowling ball is present
27 and thereby warps the membrane,]
28 the ball bearing will travel along a curved path.
[In fact, ignoring friction,

29 if we set the ball bearing moving with just the right speed in just the right direction,]
30 it will continue to move in a recurring curved path around the bowling ball –
31 in effect it will “go into orbit.”
32 Our language presages the application of this analogy to gravity.
33 ¶The sun, like the bowling ball, warps the fabric of space surrounding it,
34 and the earth’s motion, like that of the ball bearing, is determined by the shape of the warp.
35 The earth, like the ball bearing, will move in orbit around the sun
36 if its speed and orientation have suitable values.
37 This effect on the motion of the earth is
38 what we normally would refer to as the gravitational influence of the sun,
39 and is illustrated in Figure 3.5.
Figure 2. Tabulation of expository text segment (Greene 2000: 68–70)

The role of Theme and Rheme 
40 The difference, now, is

41 that unlike Newton, Einstein has specified the mechanism
42 by which gravity is transmitted:
(41) the warping of space.
43 In Einstein’s view, the gravitational tether holding the earth in orbit is not some
mysterious instantaneous action of the sun;

44 rather, it is the warping of the spatial fabric caused by the sun’s presence.
Figure 2. (continued)
clause 5, which locally introduces curved space in its own Rheme. Then gravity, ac
celerated motion, and curved space are repeated within the Theme of clause 6 as the
ground for the Einstein theory of spatial warping, introduced in the Rheme (which
includes 7–9). One of the characteristics of this Rheme-to-Theme chain pattern (fa
mously observed by Daneš 1974: 118–119) here is that each successive logical outcome
in a Rheme becomes the logical basis in a Theme for the next logical outcome. This
recurring pattern of logical dependency actually began in clause 1, since gravitational
force there was represented as a consequence of the mass of the sun. The Themes of
clauses 2 and 4 are not just a repetitive counterpoint to the framework of clauses 1, 3,
5 and 6 – the initial marked Themes in each, with their presuming reference, offer a
contextual orientation for the reintroduction of accelerated motion and curved space.
The differently structured remainder of the paragraph (10–15) has just three Themes
outside of embedded clauses. The first two, in clauses 10 and 13, offer analogy and,
with more presuming reference, radical proposal as orienting meta-categories for the
propositions, embedded in Rhemes, which precede and are about to be repeated. Un
marked Theme space in 13, part of a long lexical string which begins in 5, is repeated
once more in the Theme of 15, alongside of the elaboration signal rather.
In paragraph 2, the method of development shows the consistency and variety
characteristic of an outline structure. Themes are consistent first in field, concentrat
ing repetitively on a limited collection of objects like they (objects), the bowling ball, the
ball bearing and the rubber membrane. By contrast the point of the passage shows var
ious geometrical references, with N-rheme invocations of center, absent, flat, straight,
present, curved. The method is also consistent in its calculated symmetry, chiefly pro
duced by parallel marked sentence Themes. Of the 6 sentences, each of sentences 2–5
(including numbered clauses 18–31) starts with a contextualizing initial dependent
clause. Further, the initial clauses of the two inner sentences (23, 26) are identical
except for the antonymy of absent and present. The initial clauses of the two outer
sentences (18, 29) are also made parallel by the wording “if we . . . ball bearing”, pre
ceded by a further contextualizing non-finite clause. Variation within the framework
is signalled mainly by these same marked sentence Themes. Each of them introduces
a new step in the laboratory demonstration, and the Themes of each of the following
main clauses contain the objects affected by the new step. The very first and very last
sentences frame the framework: the first connects with the previous paragraph; and
the second, as a Hyper-theme, anticipates the substance of the following paragraph.
 Michael Cummings
As for the point, the field of geometry which characterizes it, particularly in the N
rhemes, is not just lexical. It is also carried by prepositional phrases as spatial Adjuncts
in clauses: in the vicinity of the sun (16), on the membrane (18), with some initial velocity
(19), etc. At the same time, the collection of references to geometry has less repetition
and is more lexically varied than the collection of objects in the Themes.
The last paragraph has two connected parts, 33–39, and 40–44. In the first part,
the framework provided by each successive clause Theme methodically takes us to
each next step in the application of the analogy to the solar system: the sun (∼bowling
ball, 33), the earth (∼ball bearing, 34, 35), earth’s velocity (36), outcome (37). The N
rhemes in each of these clauses continue the geometrical referencing of the previous
paragraph, down to the last (all of embedded clause 38) which concludes the analogical
argument by connecting back to the gravitational force concept. Clauses 40–44 con
trast the Newtonian and Einstein models of gravity, and deny the Newtonian. The first
Theme in an unembedded clause is predictive: The difference (40). The other two in
clude the gravity concept, and their N-rhemes contrast the Newtonian and the Einstein
theories of its nature.
. Quantitative interpretation of the method of development and the point
As different sets of meanings, the method of development and the point represent two
principal components in the rhetorical strategy of continuous discourse. The one sup
plies the bases for the addition of relevant new information. The other is the set of
meanings which are introduced as new because they constitute the goal of the dis
course fabric. Inevitably these different sets of meanings are worded in ways which are
characteristic of their different functions. The differences in wording are not always
discernible in individual clauses, but can also be constituted by different proportions of
the relevant forms in all the Themes and Rhemes taken together. Conjunction, for ex
ample, worded in the form of subordinate and coordinate conjunctions, is restricted to
Theme stretches; but conjunction worded in the form of Conjunctive Adjuncts occurs
in Rheme to some degree. Lexical variation, for another example, is more a character
istic of Rheme than of Theme, but only to a degree. Systemic-Functional analysis of
the language differences between method of development and point has tended more
to the qualitative and intuitional rather than to the quantitative. However, degrees of
difference in language require a quantitative approach for their description. A quanti
tative analysis of these language differences will serve at least two important purposes.
The first is the determination of the degree of difference in the distribution of forms
and features between Theme and Rheme stretches, including the N-rheme. The second
is the demonstration of the differences in the language of texts which differ stylistically
with respect to their handling of Theme and Rheme.
To simplify matters, this discussion of a quantitative approach to the languages of
method of development and point will be limited to the consistency aspect of these
respective languages. To narrow the discussion still further, only one approach to con
sistency will be considered: presuming reference. Presuming reference, as defined by
The role of Theme and Rheme 
Martin (1992: 102–140), excludes certain other forms of reference such as exophoric
reference, homophoric reference, substitution and ellipsis, generic reference and ad
dition reference. It often exhibits itself in the form of chains of reference items which
maintain continuity over long stretches of text (Halliday & Hasan 1985: 84; Martin
1992: 140–157). It has long been recognized that reference chains are very characteris
tic of successive Theme elements in clauses (Francis 1989: 211–212, 1990: 64–66; Fries
1995c: 350–354; Martin 1992: 434–448). What a quantitative analysis of the distribu
tion of presuming reference will show is that reference chains are more characteristic
of Themes rather than of N-rhemes or even of Rhemes altogether.
The most obvious way of making such a measurement is to compare the number
of reference chain items in each of three areas of a text segment: the Themes, the N
rhemes, and the remaining parts of the Rhemes, termed by Fries (1992: 478) the Other.
The very first example in the text of Figure 2 is a chain of two items which doesn’t con
tribute to the cohesion of the text as such because it chains an item in an N-rheme back
to an item in the same clause: in clause 1, other objects chains back to “A massive body
like the sun, and indeed any body”, giving a count of one item in Theme and one item
in N-rheme. More characteristically, the we pronoun in the Themes of clauses 2 and 4
chain to one another and back to previous such references in the near vicinity (only the
introductory we in the whole book would be considered exclusively exophoric). The
count for the whole of the three paragraphs yields the following comparison: chain
items in Themes, 27 or 55%, chain items in N-rhemes, 12 or 24.5%; chain items in
the Other part of the Rheme, 10 or 20.5%. Comparison with other texts (below) sug
gests that the proportional favouring of Theme is characteristic of English texts, but
that the effect here is only modest in comparison with the more pronounced contrast
within other texts. The secondary prominence of reference chain items in N-rhemes,
as against Other, seems highly characteristic of English texts.3
This first measurement has attempted to establish the comparative significance of
Theme and Rheme as a locus for the reference chain. It starts with the chains and asks
how much of them goes into Theme or Rheme. The proportion derived marks Theme
as peculiarly significant to this form of cohesion. But another issue arises within the
opposite perspective, and it is the significance of the reference chain to the Theme or
the Rheme. This issue starts with the Theme, or the Rheme, and asks how much of it is
filled by elements of reference chains. Because reference chaining is confined to expe
riential elements, excluding, for example, textual and interpersonal Theme elements,
a refined form of the question is, how much of the experiential content of each part
of the clause is filled by chained elements? The number of chained elements in the
method of development divided by the number of experiential elements is a density
fraction: it yields the reference-chain density for that part of the clause. The virtue of
this form of the index is that it adjusts for the variation in experiential content across
the Themes, N-rhemes and Other. In the Elegant Universe text of Figure 2, for example,
countable chain elements in Themes amount to 27 groups and phrases out of a total
experiential content of 35 countable groups and phrases, for a chain element density of
77%. In the N-rhemes, the chain elements are 12, out of a total experiential content of
 Michael Cummings
Figure 3. Table of proportions for the Elegant Universe text
30 groups and phrases, for a density of 40%. The Other part shows 10 chain elements
over 44 groups and phrases for a density of only 23%. Characteristically this contrast
in proportions follows the contrast among the absolute number of chain elements in
each part, but this contrast is much sharper.
A third procedure with similar intent is to measure as a factor the role of long
chains in Themes and Rhemes. This actually involves three additional indices. The
first measures the proportion of long chain elements that go into Themes. The second
measures the proportion of chain elements in Themes which are from long chains.
Both these perspectives on the role of the long chains are important for comparing
different texts; so, to provide an index which takes both into account, the third index
is the multiplication product of the first two indices. In the Elegant Universe text of
Figure 2, there are 9 chains with more than 2 elements; but of these one has 12 ele
ments (“ball bearing”), and the next longest have but 4 elements each. Taking the “ball
bearing” chain as the only long chain implies that there are only 5 out of 12 long chain
elements in Themes, a proportion of 42%. Since it has been noted that there are 27
chain elements in Themes altogether, then the proportion of chain elements which are
from the single long chain is 5 out of 27, or 19%. The long-chain product is 0.077 –
on the basis of comparison with other texts (below) this is an extremely low index.
The implication is that in comparison with some other texts, the long chain in this
text is not a big factor in the method of development; rather it keeps coming back in
the Rheme sections of clauses, thus running contrary to the more general pattern of
reference chaining, even within the same text. All of the proportions so far discussed
are set out in the table of Figure 3.
The analysis of the single segment of text in this section has modestly supported
the idea that the method of development and the point of English texts have charac
teristic differences in their language potentials, and that one of these differences is the
skewing of the distribution of reference chains into the Theme sections of clauses. This
is not to say that individual text segments must always have a preponderance of ref
erence chaining in the Themes; in fact, Section 3.3 of this article makes the point that
some do not, and have very good reasons for this strategy. Some whole sub-genres, for
example, kitchen recipes, avoid doing so. The next step, however, is to examine more
texts which demonstrate the greater potential of reference chaining as a characteristic
of the method of development, in order to show how variations which remain within
this general parameter may nevertheless still reveal differences in stylistic intent.
The role of Theme and Rheme 
. Comparison of texts
The purpose of this section then is to demonstrate how texts with different rhetori
cal strategies contrast in their management of reference chains. Elements in reference
chains often map together with Given Information, but not necessarily – as in the case
of the initial chain element which is also New Information. Such elements also often
map together with Theme, but again not necessarily – as in the case where some or
even all elements are located in Rheme. The general tendency of written English seems
to be to locate repetitions of a referenced identity in the Theme stretch, but contrasting
types of text vary according to the proportions to which reference chains are thematic
or rhematic. Accordingly a marked distribution of reference-chain elements into the
Rheme suggests a rather different discourse plan from one which more closely – and
more conventionally – confines the elements to Theme.
. Textual analysis of narrative
The illustration of these points begins with Figure 4, a narrative text rather than
an expository text, consisting of the first two paragraphs of Chapter 31 in Somerset
Maugham’s Of Human Bondage. These two paragraphs are about relations between
two main participants, the protagonist, Philip, and his new friend, Hayward, who have
come together by chance in Heidelberg. The notation of Figure 4 is the same as in
Figure 2.
The method of development is unified by repeated thematizing of the two partic
ipants, once together as they (16). A notional division of the text into parts I through
IV is shown in the Figure by spaces and roman numerals. Each of these four parts is
distinguished by the consistent main-clause thematizing of a different participant: Part
I, Hayward; II, Philip; III, Hayward; IV, Philip. Each of the boundaries between parts is
signalled by some particularly salient type of Theme, marking a large-scale elaboration
of the psychological situation. Part I and Part II open with a thematic lexicalization of
the participant, thereafter maintained thematically with pronouns. Part III opens with
a Hyper-theme (16 “They corresponded.”), followed in the next clause by a renewed
lexicalization of Hayward. Part IV opens with a marked clause-complex Theme, fol
lowed in the next clause by the relexicalization of Philip. Additionally, the long part III
has three Sub-parts: the boundary between IIIA and IIIB is marked by the shift from
thematic Hayward to a thematizing of the Rome of the Popes and its disjunctive con
junction, and that between IIIB and IIIC by a marked topical Theme (With the spring)
which marks a significant temporal transition.
This method of development provides a framework for a unified episode whose
parts show different informational goals, i.e., have distinguishable points. The N
rhemes characteristically show a great lexical variety within various participants, at
tributes or circumstances. But an underlying structure mirrors the framework pro
vided by the method of development. The main-clause N-rhemes consistently indicate
that the main point of Part I is Hayward’s attitude towards south-German sentimen
 Michael Cummings
I 1 ¶Hayward,
2 after saying for a month
3 that he was going South next day
4 and delaying from week to week out of inability to make up his mind to the bother of
packing and the tedium of a journey,
(1) had at last been driven off just before Christmas by the preparations for that
festival.
5 He could not support the thought of a Teutonic merry-making.
6 It gave him goose-flesh to think of the season’s aggressive cheerfulness,
7 and in his desire to avoid the obvious he determined to travel on Christmas Eve.
II 8 ¶Philip was not sorry to see him off,
9 for he was a downright person
10 and it irritated him
11 that anybody should not know his own mind.
12 Though much under Hayward’s influence, he would not grant
13 that indecision pointed to a charming sensitiveness;
14 and he resented the shadow of a sneer
15 with which Hayward looked upon his straight ways.
IIIA 16 They corresponded.
17 Hayward was an admirable letter-writer,
18 and knowing his talent took pains with his letters.
19 His temperament was receptive to the beautiful influences
20 with which he came in contact,
21 and he was able in his letters from Rome to put a subtle fragrance of Italy.
22 He thought the city of the ancient Romans a little vulgar,
finding distinction only in the decadence of the Empire;
IIIB 23 but the Rome of the Popes appealed to his sympathy,
24 and in his chosen words, quite exquisitely, there appeared a rococo beauty.
25 He wrote of old church music and the Alban Hills, and of the languor of incense and the
charm of the streets by night, in the rain,
26 when the pavements shone
27 and the light of the street lamps was mysterious.
28 Perhaps he repeated these admirable letters to various friends.
29 He did not know
30 what a troubling effect they had upon Philip;
31 they seemed to make his life very humdrum.
IIIC 32 With the spring Hayward grew dithyrambic.
33 He proposed
34 that Philip should come down to Italy.
35 He was wasting his time at Heidelberg.
36 The Germans were gross
37 and life there was common;
38 how could the soul come to her own in that prim landscape?
Figure 4. Tabulation of narrative from Of Human Bondage, Chapter 31 (Maugham 1963/

1915: 128–129)
The role of Theme and Rheme 
39 In Tuscany the spring was scattering flowers through the land

40 and Philip was nineteen;
41 let him come
42 and they could wander through the mountain towns of Umbria.
43 Their names sang in Philip’s heart.
44 And Cäcilie too, with her lover, had gone to Italy.
IV 45 [When he thought of them]
46 Philip was seized with a restlessness
47 he could not account for.
48 He cursed his fate
49 because he had no money to travel,
50 and he knew
51 his uncle would not send him more than the fifteen pounds a month
52 which had been agreed upon.
53 He had not managed his allowance very well.
54 His pension and the price of his lessons left him very little over,
55 and he had found going about with Hayward expensive.
56 Hayward had often suggested excursions, a visit to the play, or a bottle of wine,
57 when Philip had come to the end of his month’s money;
58 and with the folly of his age he had been unwilling to confess
59 he could not afford an extravagance.
tality as manifested in Christmas. N-rhemes of II are consistently attitudinal and lex

icalize Philip’s reaction to Hayward’s affectedness. Part IIIA has N-rhemes which first
introduce Hayward’s letters, then lexicalize his sensitivity. IIIB N-rhemes lexicalize or
symbolize romantic longing, then vector it to Philip. The point of IIIC is carried in the
N-rhemes by symbols of beauty and contrasting ugliness. Parallel to II, the N-rhemes
of IV convey Philip’s reaction, and concentrate on money and anxiety.
Once again, a quantitative comparison between the languages of Theme and
Rheme will be limited to the roles of reference chains. In the Human Bondage text,
each of the two principal participants has his own reference chain. In each case these
chains include lexicalized first-mention or renewed mention, pronominal references,
and nominal groups with possessive reference. For the Hayward chain, 19 out of 22
elements occur in the method of development, and 3 in N-rhemes, for contrasting
proportions of 86% and 14%. For Philip, 20 out of 29 chain elements occur in the
method of development, and 9 in N-rheme or the Other part of the Rheme, for a
contrast of 69% and 31% respectively. For all chained elements representing whatever
participants in the whole text, the proportions are 61% in Theme stretches, 23% in N
rhemes and 16% in the Other. (See the first line of the table in Figure 5 for these and
the following percentages.) These proportions mark this text segment as another grain
of supporting evidence for the intuition that reference chaining is more a characteristic
of Theme language than of Rheme language.
 Michael Cummings
Figure 5. Table of proportions for three texts
The second measurement for the relative distributions of reference chains is the
density measurement (cf. Section 2.2). In the Human Bondage text, the Themes con
tain 63 experiential elements of which 48 are chained – the density is 76%. The same
measurement for N-rhemes yields 33% and that for Other is 21%. This result suggests
that for this text segment too, reference chain cohesion is even more a characteristic of
Theme content than Theme location is a characteristic of reference chain.
The third measurement involves relative length. It is well known that some kinds
of texts, especially narrative texts, are characterized by reference chains of many el
ements located in Themes (Martin & Peters 1985: 74). This is true of the Human
Bondage text: the Hayward chain with 22 elements and the Philip chain with 29 el
ements contrast with the next longest chains, those for letter and Italy with only 6
elements each. Again, two different measurements for long chains are relevant: the
first measures the number of elements in long chains that go into Themes, and the
second measures the number of chained elements in Themes that come from the long
chains. In this text, 76% of the long chains are distributed into Themes, and 81% of
chained Theme elements are from the long chains. The joint effect of these two factors
is assessed in the multiplication product of the two: 0.621 for this text.
All this information is restated in the table of Figure 5. The first row of this Table
represents the Human Bondage data. The other rows show contrasting proportions for
other texts, of which the first is another narrative passage, from David Copperfield. A
comparison of the figures for the two texts shows that they are similar to the extent
of suggesting that reference chains should predominate in Themes, that the second
greatest concentration is in the N-rhemes, that the density of chain elements among
experiential references is even more distinctively a feature of Themes, and that long
chains in narratives are proportionally a feature of Themes also. But the differences be
tween the two sets of figures all tend in the same direction: reference chain distribution
in Themes is less emphatically a characteristic of the Copperfield text. Here, for exam
ple, in the N-rhemes, there are proportionally more chain elements, and the chain
element density is higher also. The differences between the two sets of figures are just
sufficiently noticeable to suggest scrutinizing the two texts to see if they have any sig
nificance to their respective rhetorical strategies. The Copperfield passage is tabulated
in Figure 6, in the same notation as Figure 4.
The role of Theme and Rheme 
Ia 1 ¶ [When the chambermaid tapped at my door at eight o’clock,

2 and informed me
3 that my shaving-water was outside,]
4 I felt severely the having no occasion for it,
5 and blushed in my bed.
6 The suspicion
7 that she laughed too,
8 when she said it,
(6) preyed upon my mind all the time
9 I was dressing;
Ib 10 and gave me,

11 I was conscious,
(10) a sneaking and guilty air
12 when I passed her on the staircase,
13 as I was going down to breakfast.
14 I was so sensitively aware, indeed, of being younger
15 than I could have wished,
16 that for some time I could not make up my mind to pass her at all, under the ignoble
circumstances of the case;

17 but, hearing her there with a broom, stood peeping out of window at King Charles on
horseback, surrounded by a maze of hackney-coaches, and looking anything but regal in
a drizzling rain and a dark-brown fog,
18 until I was admonished by the waiter
II 19 that the gentleman was waiting for me.
20 ¶ [It was not in the coffee-room]
21 that I found
22 Steerforth expecting me,
(20) but in a snug private apartment, red-curtained and Turkey-carpeted,
23 where the fire burnt bright,
24 and a fine hot breakfast was set forth on a table covered with a clean cloth;
25 and a cheerful miniature of the room, the fire, the breakfast, Steerforth, and all, was
shining in the little round mirror over the sideboard.

26 I was rather bashful at first,
27 Steerforth being so self-possessed, and elegant, and superior to me in all respects
28 (age included);
29 but his easy patronage soon put that to rights,
30 and made me quite at home.
31 I could not enough admire the change
32 he had wrought in the Golden Cross;
33 or compare the dull forlorn state
34 I had held yesterday,
(33) with this morning’s comfort and this morning’s entertainment.
35 As to the waiter’s familiarity, it was quenched
36 as if it had never been.
Figure 6. Tabulation of narrative from David Copperfield, Chapter 20 (Dickens 1958/

1850: 278)
 Michael Cummings
37 He attended on us,
38 as I may say,
(37) in sackcloth and ashes.
There are many thematic and stylistic differences between the two texts, but both
are about an insecure protagonist who has fallen under the sway of a charismatic men
tor. Each passage is part of a crucial episode in this relationship. The topical Themes in
each passage are mainly the human participants: 70% in the Human Bondage text,
75% in the Copperfield. But the first text concerns, almost without exception, just
two human participants, Philip and Hayward. In the second text, there are four hu
man participants, the protagonist, represented by the first-person pronoun, and the
chambermaid, the waiter, and Steerforth. In each of these texts there is an interaction
between the participants; in text one it is simply between the two friends, but in text
two it is almost entirely between the protagonist, and each of the three other partici
pants. The method of development in the second passage reflects this difference. The
single most constant Theme element is the first-person pronoun. But it alternates in
the first 18 numbered clauses with the chambermaid, and in the remaining clauses,
19–38, with Steerforth, and a few times with the waiter. This makes for a division into
two parts, between clauses 18 and 19, which is just one clause short of the paragraph
division. Clause 19 is transitional. Each of the two paragraphs opens with a marked
thematic device, underscoring the division – the first is a marked sentence theme,
clauses 1–3, and the second is the predicated Theme of clause 20. The first part has
a further linear substructure which contrasts clauses 1–9 with clauses 10–18: the first
set features the thematized chambermaid, the second features the thematized protag
onist. (Contrary to the usual exclusion of embedded clauses, clauses 7–8 are treated
here as if a projection after the grammatical nominalization metaphor The suspicion.)
The alternation of thematized participants in the second paragraph is by contrast an
interleaving of the protagonist and Steerforth or the waiter.
The distribution of reference chains in the two texts is crucial to their differences.
Of the four participants who dominate the Themes in the Copperfield text, the protag
onist is represented by the single long chain of 20 elements. The chambermaid has only
4, the waiter 5, and Steerforth 4 elements. The shorter chains are mainly in Themes,
contributing to the thematic alternation with the protagonist. Only the protagonist
makes a significant contribution to the point of the text, within the Rheme structures.
Of the 20 protagonist references, 12 occur in Themes, 60%, and more than half again in
Rhemes, that is, 8 or 40%. The kind of interaction which takes place in this text seems
to require a significant displacement of the protagonist into the Rhemes. By compari
son, the two major participants in the Human Bondage text are 76% in Themes, 24%
in Rhemes.
The point of the text is substantially evaluative. In the first section, clauses 1–18,
which deals with social embarrassment, N-rhemes are frequently pejorative, either lex
The role of Theme and Rheme 
ically, as in sneaking and guilty (10), being younger than I could have wished (14–15),
or the ignoble circumstances (16); or symbolically, as in having no occasion for it (4),
in my bed (5), King Charles . . . surrounded by a maze of hackney coaches . . . in a driz
zling rain and a dark-brown fog (17). In the second section, which deals with social
amelioration, the evaluation is carried in mainly approving N-rhemes: a snug private
apartment (20), bright (23), a clean cloth (24), this morning’s comfort (33). Intermixed
are various time and place references, and, most important, a continued reference in
Rhemes to the protagonist in the first person, connecting him with the evaluation. The
typical pattern is that the evaluation is instigated by one of the other participants – the
chambermaid in part one and Steerforth or the waiter in part two – is felt by the pro
tagonist, and has the protagonist as its object. It is largely this frequent incorporation
of the Copperfield protagonist into the unfolding point of the discourse that makes
for the differing proportions in the distribution of chain elements between the two
texts. Although the Human Bondage text also involves an interaction between partic
ipants, it is of a rather different kind – there the participants do not relate directly,
but through a variety of intermediate objects, like letters or the details of the Roman
cityscape or the landscape of Umbria, and the chained participants are more confined
to the Themes.
. Textual analysis of exposition
The third row in the table of Figure 5 shows similar kinds of data for a text extracted
from Colin Tudge’s The Time before History (1997). This text crosses the genre bound
ary back to exposition, and serves to show contrasting proportions rather opposite to
those of the previous contrast. Chained elements are even more concentrated in the
Themes than in the Human Bondage narrative, and at a higher density. But the long-
chain distribution product is very low (0.175 vs. 0.621) because the passage completely
avoids chains much greater than 6 elements, thus conforming to the notion that narra
tive contrasts with exposition in a preference for long thematic reference chains. This
extract is tabulated in Figure 7.
Close examination of this text reveals that one reference chain barely qualifies as
long with 8 elements (hominid fossils/corresponding fragments, 5ff.), but there are four
other chains with from 3 to 5 elements. These together are the largest single factor in
the method of development as an outline structure. Each of these chains represents
either the introduction at a higher outline level or the continuation at a lower out
line level of some outline division. Thus each of the 5 dispersed elements of the I/we
chain serve to introduce some conjunctive variation in the development, while concen
trated chains like hominid fossils/corresponding fragments (5ff.) or some species (17ff.)
represent a local consistency in the development.
Rarely does any chain have an element in the Rhemes of clauses. The N-rhemes in
this passage very frequently consist of adjectival groups and indefinite nominal groups
serving as Attribute elements in literal or metaphorical relational process clauses. The
rhetorical strategy of this passage is thus very different from those already discussed. It
 Michael Cummings
I 1 ¶Here, I am following the ideas of Bernard Wood.

2 [As he points out,]
3 it is not at all easy to decide
4 how many species are represented by the various fossils.
IA 5 To begin with, hominid fossils are usually fragmented,
6 they come from widely different areas,
7 and those from the same area often represent several or many individuals.
8 [If corresponding fragments seem to come from creatures
9 that are very similar to each other,]
10 then it seems reasonable to conclude
11 that they do indeed come from the same species.
12 But
13 [if they are different,]
(12) then they may come from the same species,
14 or they may come from a single species
15 that is highly variable.
IB 16 For we know from observation of living animals

17 that some species are extremely variable.
IB1 18 Those
19 that live in different regions
(18) may differ in size and to some extent in shape,
20 though they may still be perfectly capable of mating successfully
21 if brought together,
22 and so would generally qualify as the same species.
IB2 23 We may also expect

24 that long-lasting species would vary over time.
IB3a 25 ¶We know, too,
26 that males may be very different from females – usually, in mammals, larger and
bulkier.
27 This is sexual dimorphism.
IB3b 28 Confusingly, some primates are extremely dimorphic (like gorillas and baboons)
29 and others hardly at all (like gibbons and marmosets).
30 This dichotomy has enormous social significance.
31 [When the males are much bigger than the females, as in gorillas or elephant seals,]
32 we can generally assume polygyny:
33 each alpha male has many wives.
34 [When the sexes are much of a muchness,]
35 simple pairs are more likely.
IB3c 36 But
37 [when a pile of fossils seems to fall into two size groups,]
(36) does this imply two coexistent species, or a marked difference between the sexes?
38 This, in practice, can be an extremely hard nut to crack.
Figure 7. Tabulation of exposition from The Time Before History (Tudge 1997: 202)
The role of Theme and Rheme 
II ¶ [To sort the fragmented and scattered hominid fossils into distinct types –
39 which may or may not in the fullness of time be accorded the status of species – ]
40 Bernard Wood has applied two lines of thought.
IIA 41 In general, like many other modern anatomists, he applies the methods of analysis
first formulated by the German biologist Willi Hennig in the 1960s: those of cladistics.
IIA1 42 The basic notion is to decide rigorously
43 which characters
44 that are shared by the various fossils
(43) are truly homologous –
45 that is, were inherited from a common ancestor –
46 and which are merely analogous,
47 and have become similar through convergent evolution.
IIA2 48 Next, the cladist (the analyst employing cladistic methods) must decide
49 whether the homologous features, or characters, truly reflect a close relationship
or not
consists of a rather complicated argument in which the underlying syllogisms depend

on attributing qualities or generic types to various classes of objects. The concentration
of short chains within the method of development which instantiates this procedure
results in the proportions which contrast so markedly with those of the other passages.
The data set out in Figure 3 for the Elegant Universe text of Figure 2 can now be
brought into the comparison. The chain element distribution for this text is much
more like that of the narrative texts than of the other expository text. The chain ele
ment density figures also belong in the same ranges as those of the narrative texts. It
is only in the long chain distribution figures that the two expository texts together are
distinguishable from the narrative texts. The rhetorical procedure of the Time before
History exposition which makes it an extreme case of reference chain concentration in
Themes has been alluded to; but its avoidance of the long-chain strategy is even ex
ceeded by that of the first exposition (product 0.175 vs. product 0.077). Thus all four
texts skew reference chains towards the Themes regardless of genre; but the long-chain
figures distinguish the two genres. On stylistic terms, the one modestly long chain in
the Time before History segment plays a conventional role in supporting the method of
development, but the modestly long chain in the Elegant Universe segment contrarily
belongs to a Rheme pattern – its text is more organized in terms of bowling ball and
rubber membrane, to which the ball-bearing is more often an added factor.
. Extremely variant texts
This section will try to show that any statistic which makes reference chaining more
characteristic of Themes than Rhemes will have to factor in some texts for which the
exact opposite is true, and true for good stylistic reasons. The first issue is texts which
 Michael Cummings
1 Birds do it,
2 bees do it
3 Even educated fleas do it
4 Let’s do it,
5 let’s fall in love
6 In Spain, the best upper sets do it
7 Lithuanians and Letts do it
8 Let’s do it,
10 The Dutch in old Amsterdam do it
Not to mention the Finns
11 Folks in Siam do it –
12 think of Siamese twins
13 Some Argentines, without means, do it
14 People say
15 in Boston even beans do it
16 Let’s do it,
Figure 8. From Cole Porter, ‘Let’s Do It (Let’s Fall In Love)’, from the show Paris, 1928
(Porter 1928: 13–15)
so vary from the norm that they invert the expected proportions of reference chain
distribution, and the second issue is the stylistic intention which motivates this kind
of written expression.
.. A text written to be performed musically

Figure 8 includes a very specialized kind of written English, that written to be per
formed musically. The text of Figure 8 is just an excerpt from a much longer version
which would have larger absolute numbers of the forms under discussion, but never
theless would keep the same proportions of these forms. The method of development,
represented as before by the bolded Themes, including Subject after a marked topical
Theme, has a very well-defined structural function. It shows an alternation between
sets of creatures who fall in love and a repeated exhortation, let’s – to some intended –
to do the same. The refrain corresponds to the end of stanzas (1–5, 6–9, 10–17). The
components of this method of development are always topical Themes, without vari
ation by any textual or interpersonal Theme elements. Some variation is provided by
stanza-initial geographical reference (In Spain, 6 and in old Amsterdam, 10), and one
projecting clause (People say, 14). The effect is both a listing, a conventional comic
aria technique, and monotonous, which correlates with the even greater monotony of
the point.
The point is even simpler in execution than the method of development. It consists
of a repeated do it within the Rheme of most clauses, anticipating the lexicalized fall
in love which terminates each stanza (5, 9, 17). The general language characteristics of
method of development and point are thus reversed – the method is more lexicalized
The role of Theme and Rheme 
Figure 9. Table of proportions for ‘Let’s do it’
1 I hunt black widow spiders.

[2 When I find one,]
3 I capture it.
4 I have found them in discarded car wheels and under railroad ties.
5 I have found them in house foundations and cellars, in automotive shops and toolsheds,
against fences and in cinder block walls.
6 As a boy I used to lift the iron lids
7 that guarded underground water meters,
8 and there in the darkness of the meter wells I would often see something round as a
flensed human skull, glinting like chipped obsidian, scarred with a pair of crimson triangles
9 that touched each other to form an hourglass: the widow
10 as she looks in shadow.
11 A quick stir with a stick would trap her for a few seconds in her own web, long enough
12 for me to catch her in a jar.
Figure 10. From Gordon Grice, The Red Hourglass (Grice 1998: 1)
than the point, and the longest reference chain occurs within successive N-rhemes,
not the Themes. It is the lexical Themes Birds (1) and bees (2) that slyly suggest an
identification for N-rheme it which is then comically frustrated by the chaste fall in
love wording.
The quantitative analysis of reference chain distribution in this passage confirms
its exceptional character. The Table of Figure 9 gives a breakdown on the proportions.
The absolute numbers of reference chain elements show that more than 62% occur
in the N-rhemes, only 25% in Themes. The chain element density in N-rhemes is
88%, only 32% in Themes. Only 29% of long chain elements (us, fall in love) occur in
Themes, although all the chained elements in Themes are from a long chain.4 Relative
proportions are thus exactly opposite to what more conventional texts have led us
to expect.
.. A semi-autobiographical exposition

The very specialized genre of the text in Figure 8 suggests it would be useful to illustrate
a similar variation from norms with a text from a more conventional genre. The text
in Figure 10 is simply expository, the opening paragraph of a semi-autobiographical
book of popular science. The method of development is dominated by the I chain
which occurs as a topical Theme in 7 of the 8 non-embedded clauses.
 Michael Cummings
Figure 11. Table of proportions for “black widow spiders”
This consistency marker is in some clauses accompanied by conjunction or Ad

junct markers of sub-structure variation: When (2) shifts from the universal present
time of clause 1 to the differentiated time, present and past, of the rest of the para
graph; As a boy (6) further differentiates the past time started in clause 4; There and
in the darkness of the meter wells (8), makes more specific a localization introduced
in the Rheme of clause 7. The organization of the method thus involves a progres
sive focussing in. It culminates in the topical Theme A quick stir with a stick, which
is nominalized so as to leave out the person of the narrator. It is a jarring variation
from the pattern set by the I chain – no doubt to provide a climax to the growing sense
of horror.
The point of the paragraph involves two sets of entities, mainly in the N-rhemes:
the black widow and her locations. The degree of lexical variation which is character
istic of the point is here exhibited only in the locations: car wheels, railroad ties (4), etc.
The black widow spider itself is lexicalized in only 3 of 9 different mentions (in clauses
1 and 8). One of the lexicalizations occurs as modification of the pronoun something
(8). The persistence of the long black widow chain(s), like that of the fall in love chain
of the previous text, is very uncharacteristic of point, and suggests that something is
happening in the N-rhemes which more usually happens in the Themes. The black
widow spider in the point of the paragraph is an equal focus to the protagonist’s I
self-reference in the method of development. The paragraph is about an antagonistic
relationship between two adversaries, and each of these adversaries occupies one of the
two principal points of textual prominence in most of the clauses.
The quantitative analysis of reference chain distribution in this passage confirms
its character as exceptional, though less so than the previous passage. The table of
Figure 11 gives a breakdown on the proportions. The absolute numbers of reference
chain elements show that more than 53% occur in the two parts of the Rhemes, less
than 47% in Themes. Admittedly the chain-element density in Themes significantly
exceeds that of the N-rhemes, but by comparison with those of Themes in the more
normative texts measured in the Tables of Figures 3 and 5, it is still a very low density;
and the density in the N-rhemes of this text is noticeably higher than that of any of
the texts attested in Figures 3 and 5. In the matter of the long-chain distribution per
centages for this passage, the long I chain of the Themes is exceeded in length by the
long black widow chain in the Rhemes, which show 53% percent of the long chain ele
ments, and whose chained elements are also 100% from a long chain, for a long-chain
distribution product of 0.533, exceeding that of the Themes.
The role of Theme and Rheme 
. Conclusion
In sum, the Systemic-Functional distinction between clause Theme and clause Rheme
is best understood by reference to its discourse context. Theme is a meaning which
is componential to the discourse meaning called method of development. Rheme is a
meaning which is componential to the discourse meaning called the point, particularly
in its major clause focus, the N-rheme. The contrast in meanings makes for a contrast
in wordings, as measured quantitatively across an appreciable length of text. One such
contrast is in the proportional distribution of reference chains, which, along various
axes of measurement, favour Themes. But specimen texts which show distributions
widely variant from the norms are very interesting just for that reason, and can show
a calculated stylistic effect.
Notes
. The issues in Sections 1–2.2 of this article are also considered in a condensed form in Cum
mings 2004. An earlier version of Sections 3.1–3.2 was presented at the 15th Euro-International
Systemic Functional Workshop, University of Leeds, Leeds, England, July 21, 2003.
. Some clauses do not have a Theme/Rheme structure. For example, many non-finite clauses
lack the Theme element in the form of either a conjunctive or a Subject (Halliday 2004: 99).
An example is the unnumbered “Using the rubber membrane-bowling ball analogy” before 18,
which with 18–19 constitutes a dependent clause-complex serving as marked Sentence Theme.
. What to include and what not to include in reference chains is an issue. Certain kinds of pre
suming reference are here considered irrelevant: esphoric reference, i.e., cataphoric reference of
a nominal group modifier to the nominal group postmodifier, and instantial reference, i.e., ref
erence worded as a relational process, or as apposition (cf. Martin 1992: 136–137). Because their
grammatical prominence is small, the following are also excluded from consideration: ellipted
Subjects, chain elements in embedded clauses or deeply embedded within group structures. In
cluded in reference chains are groups which realize the referenced participant in the form of a
possessive determiner (Martin 1992: 147), e.g. in our language (32), the earth’s motion (34), its
speed and orientation (36) – a principle which is extended to groups which realize the referenced
participant in the form of a prepositional complement within the post-modifier, as in the case
of that of the ball bearing (31). Embedding beyond this level excludes it from consideration.
However the nominal group part of prepositional phrases is treated as if not embedded in the
phrase’s structure; i.e., in this respect, prepositional phrases, e.g., on other objects (1) and groups
are treated as if the same (cf. Fries 2002: 139–140).
. In the analysis of the first four text segments of this article, a long chain is effectively any chain
of more than 6 elements. These four segments are roughly comparable in length, permitting a
consistent standard. Text segments 5 and 6 are much shorter, so it makes sense to count both
the fall in love chain with its 15 elements and the us chain with its 6 elements as long. In text
segment 6 there are only two chains, respectively of 7 and 8 elements in non-embedded clauses,
both counted long.
 Michael Cummings
References
Berry, Margaret (1989). “Thematic options and success in writing”. In C. S. Butler, R. A.

Cardwell, & J. A. Channell (Eds.), Language and Literature – Theory and Practice: A Tribute
to Walter Grauberg (pp. 62–80). Nottingham: University of Nottingham.
Berry, Margaret (1995). “Thematic options and success in writing” (revised version of Berry
1989). In M. Ghadessy (Ed.), Thematic Development in English Texts (pp. 55–84). London:
Pinter.
Berry, Margaret (1996). “What is Theme? – a(nother) personal view”. In M. Berry, C. Butler,
R. Fawcett, & G. Huang (Eds.), Meaning and Form: Systemic Functional Interpretations (pp.
1–64). Norwood, NJ: Ablex.
Cummings, Michael (2004). “Towards a statistical interpretation of Systemic-Functional
Theme/Rheme”. In G. Fulton, W. J. Sullivan, & A. R. Lommel (Eds.), LACUS Forum 30:
Language, Thought, and Reality (pp. 343–354). Houston, TX: The Linguistic Association of
Canada and the United States.
Daneš, František (1974). “Functional sentence perspective and the organization of the text”. In F.
Daneš (Ed.), Papers on Functional Sentence Perspective (pp. 106–128). The Hague: Mouton.
Dickens, Charles (1958/1850). David Copperfield. New York: Pocket Books.
Downing, Angela (1991). “An alternative approach to Theme: A Systemic Functional pers
pective”. Word, 42, 119–144.
Francis, Gill (1989). “Thematic selection and distribution in written discourse”. Word, 40, 201–
221.
Francis, Gill (1990). “Theme in the daily press”. Occasional Papers in Systemic Linguistics, 4,
51–87.
Fries, Peter H. ([1981] 1983). “On the status of Theme in English: arguments from discourse”.
Forum Linguisticum, 6(1), 1–38. Reprinted in J. Petöfi & E. Sözer (Eds.), Micro and Macro
Connexity of Texts (pp. 116–152) [Papers in Textlinguistics 45]. Hamburg: Helmut Buske
Verlag.
Fries, Peter H. (1992). “The structuring of information in written English text”. In M. A. K.
Halliday & F. C. C. Peng (Eds.), Current Research in Functional Grammar, Discourse, and
Computational Linguistics with a Foundation in Systemic Theory, Special issue of Language
Sciences, 14(4), 461–488.
Fries, Peter H. (1993). “Information flow in written advertising”. In J. Alatis (Ed.), Georgetown
University Round Table on Languages and Linguistics 1992: Language, Communication and
Social Meaning (pp. 336–352). Washington, DC: Georgetown University Press.
Fries, Peter H. (1994). “On Theme, Rheme and discourse goals”. In M. Coulthard (Ed.), Advances
in Written Text Analysis (pp. 229–249). London: Routledge.
Fries, Peter H. (1995a). “Patterns of information in initial position in English”. In P. H. Fries &
M. Gregory (Eds.), Discourse and Meaning in Society: Functional Perspectives (pp. 47–66).
Norwood, NJ: Ablex.
Fries, Peter H. (1995b). “A personal view of Theme”. In M. Ghadessy (Ed.), Thematic Develop
ment in English Texts (pp. 1–19). London: Pinter.
Fries, Peter H. (1995c). “Themes, methods of development and texts”. In R. Hasan & P. H.
Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. 317–359).
The role of Theme and Rheme 
Fries, Peter H. (2002). “The flow of information in a written text”. In P. H. Fries, M. Cummings,
D. Lockwood, & W. Spruiell (Eds.), Relations and Functions within and around Language
(pp. 117–155). London: Continuum.
Greene, Brian (2000). The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest for
the Ultimate Theory. New York: Vintage.
Grice, Gordon (1998). The Red Hourglass: Lives of the Predators. New York: Delacorte Press.
Halliday, M. A. K. (1967). “Notes on transitivity and Theme in English, part II”. Journal of
Halliday, M. A. K. ([1979] 2002). “Modes of meaning and modes of expression: types of
grammatical structure and their determination by different semantic functions”. In J. J.
Webster (Ed.), On Grammar (Vol. 1 in the Collected Works of M. A. K. Halliday) (pp. 196–
218). London: Continuum. [First published in D. J. Allerton, E. Carney, & D. Holdcroft
(Eds.), Function and Context in Linguistic Analysis: A Festschrift for William Haas (pp.
57–79). Cambridge: Cambridge University Press].
Halliday, M. A. K. ([1982] 2002). “Text semantics and clause grammar: how is a text like a
clause?”. In J. J. Webster (Ed.), On Grammar (Vol. 1 in the Collected Works of M. A. K.
Halliday) (pp. 219–260). London: Continuum. [First published (in part) as “Text semantics
and clause grammar: some patterns of realization”, in J. E. Copeland & P. W. Davies (Eds.),
The Seventh LACUS Forum 1980 (pp. 31–59). Columbia, SC: Hornbeam Press (1981) and
(in part) as “How is a text like a clause?”, in S. Allen (Ed.), Text Processing: Text Analysis and
Generation, Text Typology and Attrition (Proceedings of Nobel Symposium 51) (pp. 209–247).
Stockholm: Almqvist and Wiksell (1982)].
Halliday, M. A. K. (1985). An Introduction to Functional Grammar. London: Edward Arnold.
Halliday, M. A. K. (1993). “The construction of knowledge and value in the grammar of
scientific discourse: Charles Darwin’s The Origin of Species”. In M. A. K. Halliday & J.
R. Martin (Eds.), Writing Science: Literacy and Discursive Power (pp. 86–105). Pittsburgh:
University of Pittsburgh.
Halliday, M. A. K. (1994). An Introduction to Functional Grammar [2nd ed.]. London: Edward
Arnold.
Halliday, M. A. K. (2004). An Introduction to Functional Grammar [3rd ed., revised by Christian
M. I. M. Matthiessen]. London: Arnold.
Halliday, Michael A. K. & Ruqaiya Hasan (1985). Language, Context and Text: Aspects of
Language in a Social-Semiotic Perspective. Victoria: Deakin University.
Hasan, Ruqaiya & Peter H. Fries (1995). “Reflections on Subject and Theme: an introduction”.
In R. Hasan & P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective
(pp. xiii–xiv). Amsterdam and Philadelphia: John Benjamins.
Benjamins.
Martin, James R. (1993). “Life as a noun: arresting the universe in science and humanities”. In
M. A. K. Halliday & J. R. Martin (Eds.), Writing Science: Literacy and Discursive Power (pp.
221–267). Pittsburgh: University of Pittsburgh.
Martin, James R. (1995). “More than what the message is about: English Theme”. In M. Ghadessy
(Ed.), Thematic Development in English Texts (pp. 223–258). London: Pinter.
Martin, James R. & Pam Peters (1985). “On the analysis of exposition”. In R. Hasan (Ed.),
Discourse on Discourse (pp. 61–92). Wollongong: Applied Linguistics Association of
Australia.
 Michael Cummings
Matthiessen, Christian M. I. M. (1988). “Representational issues in Systemic Functional

Grammar”. In J. Benson & W. Greaves (Eds.), Systemic Functional Approaches to Discourse:
Selected Papers from the 12th International Systemic Workshop (pp. 136–175). Norwood, NJ:
Ablex.
Matthiessen, Christian M. I. M. (1992). “Interpreting the textual metafunction”. In M. Davies &
L. Ravelli (Eds.), Advances in Systemic Linguistics: Recent Theory and Practice (pp. 37–81).
London: Pinter.
Matthiessen, Christian M. I. M. (1995a). Lexicogrammatical Cartography: English Systems.
Tokyo: International Language Sciences.
Matthiessen, Christian M. I. M. (1995b). “Theme as an enabling resource in ideational
‘knowledge’ construction”. In Mohsen Ghadessy (Ed.), Thematic Development in English
Texts (pp. 20–54). London: Pinter.
Maugham, W. Somerset (1963/1915). Of Human Bondage. Harmondsworth: Penguin.
Porter, Cole (1928). “Let’s do it (Let’s fall in love)”. In The Cole Porter Years (pp. 13–15). Secaucus,
NJ: Warner Bros [no date].
Ravelli, Louise (1995). “A dynamic perspective: implications for metafunctional interaction
and an understanding of Theme”. In R. Hasan & P. H. Fries (Eds.), On Subject and
Theme: A Discourse Functional Perspective (pp. 187–234). Amsterdam and Philadelphia:
John Benjamins.
Tudge, Colin (1997). The Time before History: 5 Million Years of Human Impact. New York:
Simon and Schuster.
On clefting in English and Spanish*
María de los Ángeles Gómez-González and
Francisco Gonzálvez-García
Based on empirical evidence from corpora such as the International Corpus of

English-Great Britain, the Corpus de Referencia del Español Actual, and the Archivo
de Textos Hispánicos de la Universidad de Santiago de Compostela, this paper sets
out to shed light on some salient semantico-pragmatic and discourse properties
of cleft constructions in English and Spanish. Under the three-dimensional
approach invoked here, the choice of an Element in Focus (EIF henceforth) is
shown to arise out of the dynamic interaction of structural, semantico-pragmatic
and discourse-cognitive constraints, involving such competing motivations as
Topic, Focus (of information and of attention), the Given-New contrast, and the
staging of information or text-type. The main conclusion ensuing from our
analysis is that the core meaning of cleft constructions resides in highlighting the
EIF, which tends to be referential and specific.
. Aims and background
This paper offers a contrastive account of English it-clefts (ICs henceforth) and their
counterparts in Spanish, thereby recognizing the existence of the latter contra the
view advocated by other scholars such as Barcelona-Sánchez (1983) or Martínez-
Caro (1999), to mention but a few. We shall explore the syntactico-semantic make
up, the interpretations and the discourse motivations of these constructions in both
languages, drawing on evidence from three corpora, viz. the International Corpus of
English-Great Britain, the Corpus de Referencia del Español Actual, and the Archivo de
Textos Hispánicos de la Universidad de Santiago de Compostela, against the background
of previous investigations (cf. Biber et al. 1999; Delin & Oberlander 1995; Di Tulio
1990; Fawcett & Huang 1997; Fichtner 1993; Gómez-González 2001, 2002, 2004b;
Gundel 2002; Huddleston & Pullum 2002; Lambrecht 2001; R. Martínez-Vázquez
1997; Moreno-Cabrera 1987, 1999; Nelson 1997; Perzanowski & Gurney 1997; Prince
1978; Sornicola 1988; Weinert & Miller 1996).
The overarching claim endorsed in this paper is that ICs and their counterparts in
Spanish behave roughly in the same way, despite some differences detailed especially
in Section 3. Adopting a multi-dimensional approach, we shall provide empirical evi
dence that the choice of an IC arises out of the dynamic interaction of (at least) three
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
types of constraints, i.e. structural, semantic(o-pragmatic) and discourse-cognitive, in

combination with a number of competing motivations such as Topic, Focus (of infor
mation and of attention), the Given-New contrast, and the staging of information or
text-type. Furthermore, a second major claim that will be substantiated in this paper
is that the core meaning of ICs, rather than being contrastiveness and exhaustiveness,
as often claimed in the literature, is concerned with highlighting the identification of
a discourse element, the element in-Focus (EIF henceforth), which tends to be referen
tial and specific. This identification underscores the subjectivity of ICs, which broadly
can be explained as the expression of the speaker’s subjective stance about the relative
communicative importance of one part of a proposition versus another, which in turn
is relevant to the current Topic of discourse. In addition, by exploring the tendency of
ICs towards specificity in conjunction with other semantic factors such as factivity and
presupposition, some light will be shed on the otherwise puzzling restrictions exhib
ited by some finite and non-finite complement clauses as EIFs in these constructions
(cf. Borkin 1984; Huddleston 1984; Ward, Birner, & Huddleston 2002).
Traditionally and in keeping with Bolinger’s (1977) oft-quoted Principle of Non-
synonymy of Grammatical Forms (i.e. a difference in syntactic form spells out a differ
ence in meaning and/or pragmatic function), clefts1 – such as those in (1) below – have
been considered to be information-packaging constructions (such as e.g. preposings,
passives, dislocations, inversions or extrapositions) which differ from corresponding
more basic, frequent or canonical structures – those in (2) – syntactically, semantically,
and in the way informational content is presented, but which are otherwise equivalent
in terms of truth-conditions (Borkin 1984; Davidse 2000; Fernández-Leborans 2001;
Gómez-González 2001, 2004b; Halliday 1967a, b, 1994; R. Martínez-Vázquez 1997;
Moreno-Cabrera 1999; Quirk et al. 1985; Ward, Birner, & Huddleston 2002).
(1) a. It ’s a character defect2 that makes them go into politics.
(ICE-GB: S1B-024 #93:1:C)
b. Es mi hijo el que se me viene a la yema de los dedos.
(ARTHUS Corpus, Diego, 43, 14)
‘It is my son that pops up in my head’
(2) a. A character defect makes them go into politics.
b. Mi hijo se me viene a la yema de los dedos.
‘My son pops up in my head’
Broadly, clefting identifies a discourse strategy whereby information is packaged or
“cleft” into two units in order to fulfill a two-fold discourse effect, despite the differ
ences entailed in variations of this pattern across languages: (1) to set up a relationship
of identity of the “X is Y” specifying type between two units (e.g. a character defect =
that makes them go into politics; mi hijo ‘my son’ = el que se me viene a la yema de los
dedos ‘that pops up in my head’) and (2) to give discourse prominence to generally
(part of) one of the two units, the EIF (e.g. a character defect and mi hijo ‘my son’).
On clefting in English and Spanish 
As already noted and due to space limitations as well as to the relatively great
heterogeneity of cleft configurations, our analysis will be restricted to ICs and their
counterparts in Spanish, thus leaving aside constructions such as wh-clefts and other
cognate sequences – exemplified in (3) and (4) below – which differ in information
structure, function and discourse effects (e.g. Collins 1991; Delin & Oberlander 1995;
Gómez-González 2001, 2004b; Gundel 2002; Haugland 1992; Hedberg 2000; Prince
1978; Lambrecht 2001).
(3) a. What makes them go into politics is a character defect.
b. Quien se me viene a la yema de los dedos es mi hijo.
‘Who pops up in my head is my son’
(4) a. The (only) thing/reason that makes them go into politics is a character
defect.
b. El (único)/La única persona que se me viene a la yema de los dedos es mi
hijo.
‘The (only) one/person that pops up in my head is my son’
Furthermore, in line with construction-based, monostratal, “what-you-get-is-what
you-see” approaches (cf. Davidse 2000; Lambrecht 2001), we shall not discuss here
whether ICs are base-generated in the grammar or whether they are derived from other
constructions, because, among other reasons, not all ICs have non-IC counterparts,
as illustrated in (5) (cf. Akmajian 1970; Collins 1991; Delahunty 1984; Fernández-
Leborans 2001; Green 1971; Gundel 1977; Halvorsen 1978; Huddleston 1984).3
(5) a. It was with some relief that Mary discovered Mr. Smith ’s mistake but his
manner had significantly soured and continued to do so as the afternoon
wore on. (ICE-GB: W1B-021 #68:5)
a’. ?How Mary discovered Mr Smith ’s mistake was with some relief (. . . )
b. It is your brother who knew Peter. (ICE-GB: S1A-019 #291:1:D)
b’. *Who knew Peter was your brother.
c. Es con gran honor y placer como yo les presento a nuestro siguiente invi
tado. (Example taken from Fernández-Leborans 2001: 303)
‘It is with great honour and pleasure that I introduce our next guest to
you’
c’. ?Como yo les presento a nuestro siguiente invitado es con gran honor y
placer
‘How I introduce our next guest to you is with great honour and pleasure’
Our basic assumption is that, although admittedly, ICs are instantiations of so-called
“Focus constructions” (Dik 1997: 291; Martínez-Caro 1999: 129), these constructions
should not be explained exclusively in terms of notions like Focus, presupposition or
contrast. Supporting evidence for such a claim stems from recognition of the fact that
it is a range of factors, rather than a single one, that determines the suitability of an
IC in a given context, such as the following: certain discourse effects like background
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
ing or the known-fact effect (Prince 1978; Delin & Oberlander 1995); politeness in
general and hedging in particular (Borkin 1984); disambiguation and/or avoidance of
conflicting interpretations, or the expression of the speaker/writer’s attitude towards
a discourse Topic, element or proposition (Perzanowski & Gurney 1997), to mention
but a few. Accordingly, in order to arrive at a satisfactory descriptive and explanatory
account of ICs, we shall adopt a multi-dimensional approach (cf. Delin & Oberlander
1995: 470; Perzanowski & Gurney 1997: 209), showing that the choice of these con
structions basically depends on the dynamic interplay of three types of constraints,
i.e. structural (Section 3), semantic (Section 4) and discourse-cognitive (Section 5), in
combination with such competing motivations as:
1. Focus of attention (FA henceforth), that is, the camera angle of discourse that shifts
from one attentional window to another as discourse progresses, thereby evidenc
ing the speakers’ subjectivity and point of view (cf. Gómez-González 2001, 2002,
2004a, b; Langacker 2001a, b);
2. the semantico-pragmatic notion of Focus and the assumptions it leads to in terms
of presupposition and assertion (Drubig 2003; Gómez-González 2001, 2004a, b;
Lambrecht 1994, 2001; Vallduví & Engdahl 1996);
3. discourse Topic, entailing a relationship of “aboutness”, or what is at issue, at
roughly two possible levels, a global discourse level (global Topic) and a lo
cal discourse level (local Topic) (Downing 1991, 1997; Gómez-González 2001,
2002, 2004a, b);
4. the mappings of referentially Given and New information, as well as the cohesive
ties entailed by such mappings (cf. Clark & Haviland 1997; Gómez-González 2001,
2002, 2004a, b; Halliday & Hasan 1976; Martin 1992);
5. text-type, characterized by a specific communicative function and such contextual
variables as mode (the medium through which linguistic contact occurs, e.g. writ
ten, spoken), tenor (kind of speaker-addressee relationship, e.g. formal, informal)
and field (subcategories of style referring to the social processes in which language
plays a part) (Halliday 2004); and
6. the staging of information, referring to the fact that, given the linear character
of linguistic expression, speakers have to make real time decisions as to how to
initialize discourse (units) and how to build them up incrementally as discourse
unfolds, thereby directing the conceptual monitoring of the message at issue (cf.
Gómez-González 2001, 2002, 2004a, b; Mackenzie 2000, 2004).
. The corpora and data
As already stated, this study is based on the analysis of three corpora: (1) the British
component of the International Corpus of English (ICE-GB henceforth), (2) the Cor
pus de Referencia del Español Actual (CREA henceforth), and (3) the Archivo de Textos
Hispánicos de la Universidad de Santiago de Compostela (ARTHUS henceforth). Only
a cursory description of these corpora will be provided here, as further details can be
On clefting in English and Spanish 
found in, for example, their corresponding websites listed in the bibliography section
at the end of this paper.
ICE-GB contains 1 million words broken down into spoken language (300 texts)
and written (200 texts) language, with an average of 2000 words per text, comprising a
wide variety of texts well-suited for our purposes. The CREA Corpus, in turn, contains
more than 200 million words from written (90%) and spoken texts (10%), achiev
ing a balance in the proportion between texts of Spanish (50%) and Latin-American
(50%) origin.
Our English data consist of 422 tokens extracted from ICE-GB, representing all the
ICs found in ICE-GB after running ICE-CUPIII, the corpus utility program. For Span
ish, we have considered 45 tokens, 22 from CREA and 23 from ARTHUS, extracted
from a sampling of 6,349 CREA examples featuring relative pronouns such as el que
(‘the [masculine, singular]-one-who’), la que (‘the [feminine, singular]-one-who’), los
que (‘the [masculine, plural] ones who’), las que (‘the [feminine, plural] ones who’),
quien (‘who’ [singular]), quienes, (‘who’ [plural]), donde (‘where’), como (‘how’), and
cuando (‘when’) forms, in addition to 6,379 ARTHUS instances containing the verb
ser (‘be’).
From the aforementioned numbers the first significant finding emerges, that ICs
are used sparingly, especially in Spanish (0.34% in CREA and 0.36% in ARTHUS). As
suggested by Castañeda-Castro (1997: 3) and Perzanowski & Gurney (1997: 209, 211),
who also found very few ICs in their analysis of Spanish university written exami
nations and wire service messages (respectively), we would claim that the scarcity of
ICs may be due to their highly specific conditions of usage, namely, a high degree of
certainty and, therefore, also of commitment on the part of the speaker towards the
proposition entailed in the construction, which allows very specific statements to be
made (see Sections 4 and 5 of the present study) (cf. Biber et al. 1999: 963).
. Formal properties
. Do clefts exist in Spanish?
The first question to be addressed when exploring clefts in Spanish is whether

or not these constructions do actually exist in this language. Thus, Martínez-Caro
(1999: 31), in much the same vein as e.g. Barcelona-Sánchez (1983) and Moreno-
Cabrera (1987), discards the plausibility of accepting the existence of clefts in Spanish
on the following grounds:
Sentences like “Es Juan el que / quien vino” [Is – 3rd Person SG – Juan the
one who / who came], while being superficially similar to cleft sentences in
English and other languages, only differ from pseudo-cleft constructions in
terms of the order of constituents (Moreno Cabrera 1987: 86–87). Further
more, these configurations do not feature the neuter pronoun occurring at
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
the beginning of the equivalent constructions found in other languages (“it”

in English, “ce” in French) and, in addition, they lack the invariable third per
son singular form of a linking verb typical of clefts in other languages (cf. *es /
fue Juan y María quienes vinieron [*Is – 3rd Person SG Present – / Was – 3rd
Person SG Past – Juan and María – who – came Past]. Therefore, I consider
these sentences to be a constituent order variation with respect to pseudo-cleft
constructions (Martínez-Caro 1999: 132). [Our translation and glosses]
By contrast, in much the same spirit as the present paper, some other linguists have
recognized the existence of clefts in Spanish, thereby handling them as a special type
of grammaticalized identifying clauses, as will be further explicated in Section 4 of
this study (cf. Di Tulio 1990; D’Introno 1979: 240–261; Hernanz & Brucart 1987: 95;
Moreno-Cabrera 1999).4 Crucially, Moreno-Cabrera (1999: 4248) argues that:
Sentences such as “Es Juan quien / el que ha llegado tarde” [It is John
that/who/the one that arrived late] (Is – 3rd Person SG Present – Juan –
who / the one who – has arrived – late) can be considered as a special type
of copulative sentences.5 It is non-controversial that the copula serves here
the function of linking or connecting “Juan” with the free relative clause (“el
que / quien ha llegado tarde”). It is obvious that “Juan” specifies the person
being alluded to by the free relative through the relative pronoun. Therefore,
the above sentence is a grammaticalized form of expressing something like
this: There is exactly one person who has arrived late and that person is Juan.
[Our translation]
Furthermore, we would contend that, in general terms, the unmarked declarative
pattern of ICs is also valid for Spanish, as will be detailed immediately below.
. Clefts in English and Spanish
ICs are two-part constructions consisting of: (i) X, a Head clause (It’s a character de
fect) containing a copular or relational verb (usually be) and (ii) Y, a “subordinate” –
typically relative – clause (that makes them go into politics). Figure 1 represents the four
element prototypical or unmarked declarative pattern of these constructions in En
glish (in decreasing frequency in ICE-GB). Marked deviations from this pattern do also
exist, but they were only marginal in ICE-GB (5 tokens, representing 1.18%): all in
volved the fronting of the EIF (e.g. Simeone it is who chips the ball, in ICE-GB: S2A-010
#174:1:A) and most occurred in the same spoken text, a live sports commentary.
As already noted, in Spanish the pattern is essentially the same, namely COPULA
+ EIF + Y, except for the absence of it, which is non-existent in pro-drop languages.6
In English, by contrast, it is compulsory, and it has been interpreted in the literature
roughly in two different ways: (i) as a non-referential dummy element that acts as
syntactic filler of the Subject slot (e.g. Sornicola 1988: 358–359); and (ii) – the posi
tion endorsed here – as a referential element of some sort, usually analyzed as third
On clefting in English and Spanish 
C (complex clause)
‘cleft’, i.e. divided up, into two distinct parts
X Y
main clause dependent clause
– non-restrictive relative
(Huddleston 1984: 461–462)
– subordinated,
not rankshifted/embedded
– antecedent: 3
– can be subordinated
He claimed that it was/is... – human antecendents: agreement
properties of verb relative clause, and
reflexivization
– may be interrogative/exclamative It’s her parents who are to blame
Was it.../How early it was that... It was for himself that Tom wanted to pour sherry
– wide range of antecedents ( � restrictive relatives)
– it does not ‘restrict’ the meaning of 3
– an element may be fronted – it has its own intonation contour
Lois it was who rang. – wide range of non-wh relatives fulfilling
a wide range of functions
– S can be omitted ( � relative clauses)
It is a CHARACTER DEFECT that makes them go into politics
Es mi HIJO el que se me viene a la yema de los dedos
‘It’s my son that pops up in my head’

1 2 3 4
Subject COPULA(*) SC/EIF RELATIVE-LIKE DEPENDENT CLAUSE (RDC)-EIF
It be
pres./past NPs that/who/which/zero/when/where–RDC
PPs prep. + wh-clause (marginal)
AdvP non-finite clause with -ing
(non)finite (adverbial) clauses S
Zero A(djunct)
AdjP DO
PrepC
Zero
Figure 1. Unmarked declarative pattern of it-clefts
person singular neuter pronoun that refers cataphorically to the presupposition con
veyed by 4 in the general pattern, which in turn serves as a postposed modifier to it
(Akmajian 1970; Bolinger 1972; Borkin 1984; Davidse 2000; Gómez-González 2001,
2004b; Gundel 1977; Hedberg 2000).7
Although absent from ICE-GB, other (X: 1) realizations by such elements as that
or those have also been attested in the literature as illustrated in (7):
(7) a. (No,) that was the doctor I was speaking to. (Quirk et al. 1985: 1386)
b. Those are my feet you’re treading on. (ibid.)
However, we would agree with Lambrecht (2001, personal communication) and
Mackenzie (personal communication) that examples such as those in (7) differ in some
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Table 1. Distribution of tense and polarity in the COPULA of clefts in ICE-GB
ICE-GB Total
Positive Polarity Negative Polarity

Simple Present: 264 [62.55%] Simple Present: 14 [52.17%] 278 [65.87]
Simple Past: 122 [28.90%] Simple Past: 16 [17.39%] 138 [32.70]
Future: 3 [0.71%] 3 [0.71%]
Conditional: 2 [0.47%] 2 [0.47%]
Pres. Progres. 1 [4.54%] 1 [4.54%]
Total 392 [93.12%] 30 [7.10%] 422 [100%]
respects from ICs. First, there is no way sentences like (7) can be uncleft. And second,
their (X: 1) elements (that, those) seem to be more referential than the it of ICs. Only
the former can be used exophorically, thereby participating in an equation between
two truly referential denotata. Thus, in the case of (7a), for example, we could utter
that sequence pointing to someone in front of us meaning that ‘that was the doctor,
the person I was speaking to’. By contrast, the it of ICs cannot be used exophorically
and neither can it participate in an equation between two truly referential denotata
because the dependent clause (Y) does not denote an entity or referent, but rather an
open proposition, as will be shown in Section 4.2.
Constituent 2 (X: 2), on the other hand, the copular verb, is mostly be (99% of cases
in ICE-GB), while other realizations are infrequent (0.71 %) (e.g. look (like)), as in (8):
(8) It uh looks like Justin Channing who’s receiving treatment <,> (. . . )
(ICE-GB: S2A-003 #60:1:A)
(X: 2) displays a preference for affirmative polarity (93.12%) and simple tenses, more
precisely, the simple present (62.55%) or the simple past (28.9%). Negative polarity
and other tenses are relatively marginal (7.10% and 1.24%, respectively). The data sup
porting these structural variations in ICE-GB are displayed in Table 1 and exemplified
in (9) and (10) below. Incidentally, besides tense and polarity markers, constituent 2
may also house other optional elements – symbolized in Figure 1 above with (*) – such
as markers of aspect and modality, as well as focusing adverbs (only, just, etc.) – as in
(9a) and (10b), though subject to the constraint that only one occurrence of each kind
is possible (cf. Biber et al. 1999: 959; Ward, Birner, & Huddleston 2002: 1416).
(9) a. So, it’s not the people that are covered. (ICE-GB: S1B-058#94:1:E)
b. It was often not the governing class, but agrarian reformers or popular
leaders who carried out annexations. (ICE-GB: W2A-001 #34:1)
c. So <,> my theory is with these very different demographic patterns that
you see within one pastoral population <,> shows that it cannot be pas
toralism per se that is determining the demographic parameters here <,>
but that you’ve got something else. (ICE-GB: S2A-047 #132:1:A)8
(10) a. Simple Present: it’s also just the the mere fact that it was he that left <,>
you know not <,> not she that left with you. (ICE-GB: S1A-075 #26:1:B)
On clefting in English and Spanish 
Table 2. Distribution of tense and polarity in the COPULA of clefts in CREA and ARTHUS
CREA CORPUS ARTHUS CORPUS
Positive Polarity 19/22 [86.36%] Positive Polarity 22/23 [95.65%]

Present: 11/22 [50%] Present: 12/23 [52.17%]
Indefinite: 5/22 [22.72%] Indefinite: 4/23 [17.39%]
Present Perfect: 2/22 [22.72%] Imperfect: 4/23 [17.39%]
Imperfect: 1/22 [4.54%] Future: 1/23 [4.34%]
Subj. Pluperfect: 1/23 [4.34%]
Negative polarity 3/22 [13.63%] Negative polarity 1/23 [4.34%]
b. Simple Past: You just guessed that it was sort of a gradient that did that.
(ICE-GB: S1A-008 #85:1:A)
c. Future: It’ll be Neil Webb who goes across to take it with Brian McClair
waiting at the near post. (ICE-GB: S2A-003 #23:1:A)
d. Modal (Simple Past) + Perfect Infinitive: It occurred to me that it might
have been the contrast with Isabel that Barnsley had found attractive.
ICE-GB: W2F-011 #92:1)
e. (Semantically) Future construction (be going to): As they come into the
final bend now it’s going to be Chris Louis who takes the chequered flag.
(ICE-GB: S2A-012 #84:5:A)
A similar picture emerges in Spanish, as attested in Table 2. The copular verb (ser ‘be’)
occurs more likely than not in the simple present or the indefinite (cf. (11) below) –
the imperfect being somewhat marginal and of restricted use – featuring once more
typically positive, rather than negative polarity, as illustrated in (12).
(11) a. Present: Soy yo quien les queda agradecido por su compañía.
(ARTHUS Corpus, Laberinto, 115, 26)
‘It is me who is grateful to you for your company’
b. Indefinite Preterite: Fue la presión de Carter lo que hizo que se levantase
la censura (. . . ) (CREA Corpus, ABC, 14/01/1978: Nuevos brotes de vio
lencia en Nicaragua)
‘It was the pressure exerted by Carter that made it possible for censorship
to be lifted’
c. Imperfect Preterite: Sentí que era yo la que me iba a morir, (. . . )
(ARTHUS Corpus, Crónica, 27, 19)
‘I felt that it was me who was going to die’
d. Present Perfect: Como el mismo Julián Marías confiesa, ha sido su propia
vocación filosófica la que le enseñó que el hombre es libre: (. . . ) (CREA
Corpus, ABC, 21/11/1987: Julián Marías: “Verdad y libertad son dos con
ceptos que en todo momento van unidos . . .”)
‘As Julían Marías himself confesses, it has been his own philosophical voca
tion that taught him that man is free: (. . .)’
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
e. Future: ¿Será él o yo quien se expresa? (. . . )

(ARTHUS Corpus, Paisajes, 181, 14)
‘Will it be him or me who is expressing himself?’
f. Subjunctive Pluperfect: Quizá el hecho de que hubiera sido Marescu el
que le hubiera proporcionado las noticias
(ARTHUS Corpus, Carta, 177, 7)
‘Perhaps the fact that it had been Marescu who had given him some news’
(12) a. (. . . ) Es en la poesía donde habita el sentido. (. . . )
(CREA Corpus, ABC Cultural, 02/02/1996: La palabra inicial)
‘It is in poetry where sense dwells’
b. En cualquier caso, no es tanto aquí la autoría del Discurso lo que interesa
como el contenido del mismo.
(CREA Corpus, ABC, 17/04/1982: El liberalismo como legado)
‘In any case, it is not so much the authorship of Discourse that is the focus of
our interest, but rather its content’
The structural simplicity exhibited by the copula of ICs reflects its comparatively sec
ondary functional load: to set up a background of affirmative or less frequently nega
tive identification, against which the EIF is brought to the fore. The nature of the EIF, as
seen in our data, confirms the results of other investigations, that is, the preponderance
of NPs and pronouns, followed by PPs (e.g. Borkin 1984: 136; Givón 1990: 704; Gómez-
González 2001: 308–312, 2004b: 104–106; Ward, Birner, & Huddleston 2002: 1418).
As a matter of fact, as shown in Figure 1 above, as well as in Table 3 and the ex
amples in (13) below, the EIF in English offers a range of five possible realizations –
in decreasing order: (1) noun phrase (NP) (77.25%), (2) prepositional phrase (PP)
(15.40%), (3) adverbial phrase (AdvP) (4.26%), (4) clause (Cl) (2.59%), either finite
(2.36%) (FinCl), mostly adverbial clauses (of time (cf. 8c), purpose, reason, means,
and condition (with only if )), or non-finite (NFinCl) (0.23%); and (5) adjectival
phrase (AdjP), a realization not attested in ICE-GB, which presumably reinforces its
dialectally marked condition – see Note 10.9
Table 3. Class and function of EIFs in the subordinate clause of clefts in the ICE-GB
S A DO PrepC Zero Total
NP 273 8 25 20 0 326 (77.25%)

PP 0 61 0 4 0 65 (15.40%)
AdvP 0 18 0 0 0 18 (4.26%)
Fin Cl. 1 9 0 0 0 10 (2.36%)
Zero 0 0 0 0 2 2 (0.47%)
NFin Cl 1 0 0 0 0 1 (0.23%)
Total 275 96 25 24 2 422
(65.16%) (22.74%) (5.92%) (5.68%) (0.47%) (100%)
On clefting in English and Spanish 
(13) a. NP: It’s the writer that gets you so involved. (ICE-GB: S1A-016 #238:1:D)
b. PP: It was only in December last year that the European Council decided
to establish a new bank. (ICE-GB: S1B-054 #19:1:D)
c. AdvP: It is here that the democrats have made significant headway.
(ICE-GB: S2B-006 #18:1:C)
d. Fin Cl: It is when temples then became bigger and you stop using it that
you can actually reduce the angle. (ICE-GB: S2A-024 #86:1:A)
e. NFin Cl: If that policy is then adopted by a Chief Constable it is he adopt
ing that policy that is accountable for that policy that he adopts in his force
area. (ICE-GB: S1B-033 #80:1:C) (ICE-GB: W1B-003 #180:2)
f. AdjP: It isn’t obsessive that Bill is, just manic-depressive.
(Delahunty 1984: 76)
A comment is in order here regarding pronominal EIFs. These may appear either in
the nominative or in the accusative case depending on two main factors. One is style:
the nominative is normally used in formal styles, while the accusative occurs in more
informal ones. The other is the function of the relative word in the FRE. Thus, while
in both (14a) and (14b) below the pronominal EIF serves as antecedent of Subject who
in the FRE, the formality of the former, in the written mode, explains its nominative
form, while the informality of spoken discourse justifies its accusative expression in
the latter. The same case is selected in (14c), also in the spoken mode, but the choice is
further justified by function this time: him acts as Object of the main verb understood:
(14) a. It was he who recommended that CPRE should be contacted.
(ICE-GB: 1WB-020 #117:7)
b. It was me who didn’t want <„> the physical relationship with Jeremy <,>
but I didn’t know why. (ICE-GB: S1A-050 #89:1:B)
c. It was him that you understood to be saying the step had been missing
for a week. (ICE-GB: S1B-066 #112:1:A)
As regards syntactic functions, EIFs acting as Subjects (S) are overwhelmingly produc
tive (65.16%), followed by Adjuncts (A) (22.74%), Direct Objects (DO) (5.92%) and
Prepositional Complements (PrepC) (5.68%). EIFs playing the role of Indirect Ob
ject (IO), Subject Complement (SC) and Object Complement (OC) are also possible
but these are not attested in ICE-GB. Examples of the seven categories in question are
listed in (15) below:
(15) a. S: It’s probably only half the population that are covered.
(ICE-GB: S1B-058 #56:1)
b. A: It was in one of the office buildings that I discovered the letters.
(ICE-GB: S2B-023 #61:3:A)
c. DO: It was actually the study of architecture I really enjoyed.
(ICE-GB: S1A-034 #16:1:D)
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Table 4. Class and function of EIFs in the subordinate clause of clefts in CREA
S A DO NPComp Total
NP 11 0 2 0 13 (50.09%)
Pers. Pron. 2 0 0 0 2 (9.09%)
Proper N. 1 0 0 0 1 (4.54%)
S. Prep. 0 5 0 1 6 (27.27%)
Total 14 5 2 1 22
(63.63%) (22.72%) (9.09%) (4.54%) (100%)
Table 5. Class and function of EIFs in the subordinate clause of clefts in ARTHUS
S A DO NPComp Total
NP 8 0 2 0 10 (43.47%)
Pers. Pron. 8 0 0 0 8 (34.78%)
Proper N. 3 0 0 0 3 (13.04%)
S. Prep. 0 2 0 0 2 (8.69%)
Total 19 2 2 0 23
(82.60%) (8.69%) (8.69%) (0%) (100%)
d. PrepC: But it was in one of the office buildings that I discovered the letters
<,> thousands of them <,> addressed to His Excellency President Mo
hammed Siad Barre but all unopened <,>. (ICE-GB: S2B-023 #61:3:A)
e. SC: It is angry that he was. (Collins 1991: 64)10
f. OC: It’s dark green that we’ve painted the kitchen. (Quirk et al. 1985: 1385)
g. IO: It was to Terence (that) she made her appeal. (Collins 1991: 62)11
In the case of Spanish, Tables 4 and 5 – and the examples in (16) and (17) below – show
that the range of morphosyntactic realizations and syntactic functions of the EIF, as
well as their frequencies, reproduce (at least formally) the English pattern, except for
the zero realization. Thus, nominal EIFs rank highest (72.73% and 91.31% in CREA
and ARTHUS, respectively), followed by PPs (27.27% and 8.69%, respectively), while
Subjects (S) are again the most productive (63.6% and 82.60%), followed once more
by Adjuncts (A) (22.72% and 8.69%), and Direct Objects (DO) (9.09 and 8.69).
(16) a. NP: Esto, naturalmente, es el Santo Padre el que nombra los cardenales y
yo estoy muy contento con lo que él disponga de mí.
(CREA Corpus, Cambio 16, n. 982, 17/09/1990)
‘Well, naturally, it is the Holy Father who appoints cardinals and I am more
than happy with whatever he decides to do with me’
b. PP: José Luis Pinillos, premio Príncipe de Asturias de Ciencias Sociales, se
traslada con asiduidad a universidades extranjeras y fue en Harvard donde
descubrió casi seiscientos libros que analizaban el debate entre modernidad y
postmodernidad. (CREA Corpus, ABC Electrónico, 12/05/1997: Pinillos:
“La ciencia al servicio de intereses políticos y económicos. . .”)
On clefting in English and Spanish 
‘José Luis Pinillos, who was awarded the Principe de Asturias prize for
Social Sciences, frequently stays at foreign universities and it was in Har
vard where he came across almost six hundred books dealing with the debate
on modernity and post-modernity.’
(17) a. S: (. . . ) Es mi hijo el que se me viene a la yema de los dedos.
(ARTHUS Corpus, Diego, 43, 14)
‘It is my son that pops up in my head.’
b. DO: Bueno, pues en la India, en la cárcel, en el caso ¿Aunque sea un acci
dente?, aunque atropelle sin querer a una, sí. (. . .) Allí los seres humanos
valen poco. Y no es uno una vaca la que te encuentras, te encuentras, por
ejemplo en Nueva Deli, por mitad de las avenidas, pues unos atascos, pre
cisamente por rebaños y rebaños de vacas, que no se les puede tocar.
(CREA Corpus, Protagonistas, 07/95/97, Onda Cero, Oral, Tertulias)
‘Well, in India, in prison, in the case . . . / even it is an accident? / even
if one runs over one [a cow] by accident, it is [considered a crime] (. . .)
Human beings are not highly valued there. And it is not just one cow that
you run into, in New Delhi, for example, you may find yourself, half-way
along an avenue, caught in a traffic jam, just because of herds and herds
of cows, which are untouchable’
c. A: Es en la poesía donde habita el sentido.
(CREA Corpus, ABC Cultural, 02/02/1996: La palabra inicial)
‘It is in poetry where sense dwells.’
d. PrepC: Es de la universidad española de lo que voy a hablar hoy.
(Moreno-Cabrera 1999: 4298)
‘It is about the Spanish University that I am going to talk today.’
e. IO: Fue a Pedro a quien Juan regaló un libro.
(Example adapted from Moreno-Cabrera 1999: 4254)
‘It was to Peter to whom John gave a book as a present.’
f. SC: Es un ladrón y un asesino lo que Juan es.
(Fernández-Leborans 2001: 293)
‘It is a thief and an assassin that John is.’
g. OC: Es verde oscuro como hemos pintado la cocina.
(Example translated from Quirk et al. 1985: 1385)
‘It is dark green that we have painted the kitchen.’
The preponderance of (pro)nominal Subjects, followed by PPs as the most frequent
EIFs in both English and Spanish can be explained in terms of the specific, identi
fying make-up of IC constructions. Thus, NPs have more referential potential than
other categories, just as Subjects and Objects can be more easily construed as being
identificational than Adjuncts (these being in turn more identificational than Sub
juncts, Disjuncts and Conjuncts which – unless metatextually used – never occur as
the EIF). In this connection, Borkin (1984: 136) purports to demonstrate that con
stituents which lack referential plausibility (i.e. those which cannot be plausibly used
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
to refer in a loose sense) cannot be the EIF of an IC as they yield ungrammatical re
sults, in the case of the following examples (examples and acceptability judgments are
Borkin’s):12
(18) a. *It was a spy that Mary was.
b. *It was by all indications that it was a good year.
c. *It’s that she’s sneaky and underhanded that he says.
d. (?)It was that she was so sneaky about it that bothered me the most.
e. *It is destroy that letter that I did.
f. ?It is destroy that letter that’s the most important thing for us to do now.
g. *It was flat that they hammered it.
Likewise, Huddleston & Pullum (2002: 1418) observe that some content clauses cannot
be foregrounded, while others can, as illustrated in (19) below:
(19) a. *It’s that he did it deliberately that I’m inclined to think.
b. *It’s why no one told us that I am wondering.
c. It’s that he’s so self-satisfied that I find offputting.
d. It’s not whether you win or lose that matters, but how you play the game.
Although these authors do not furnish an explanation for these otherwise puzzling ac
ceptability results, they nevertheless point out that (19a)–(19b) would be acceptable
in a pseudo-cleft environment (What I am inclined to think is that he did it deliberately,
and so forth). However, we would suggest that referential plausibility is not exclusively
responsible in determining the felicity of the outcome with ICs involving clausal EIFs.
As Borkin (1984: 140) acknowledges, there is an intricate interplay of factors here, with
factuality and the ability of subordinate clauses to act as presuppositional clauses play
ing a major role in determining the acceptability of the outcome.13 Moreover, going by
our previous results, a plausible generalization seems to emerge that the less factive the
subordinate clause, the more acceptable the IC configuration will be, although other
factors such as contrastive negation may also ameliorate the acceptability of the out
come. By way of illustration, sentence (18a) above improves in acceptability if followed
by a contrastive negation (e.g. It was a spy that Mary was, not an assassin), a tendency
also observed in Spanish clefts (cf. Fernández-Leborans 2001).
In addition, the specific/referring nature of the EIF emerges as a key factor in de
termining the degree of acceptability of a given combination. Particularly illuminating
in this respect is Moreno-Cabrera’s (1999: 4259–4260) observation that:
“Sentences of the type “Es un hombre / algún hombre el que / quien viene”
[Is-a-man/some-man-the-one-who/who comes – ‘It is a man / some man (the
one) who comes’] are hardly likely to be interpreted as PdR [clefts] unless
they are produced in an adequate supporting context. The reason for this is
as follows. Generally speaking, it is the case that the CES [our FRE] identifies
the entity or person alluded to in the relative clause. If we say, for instance,
“Es el cartero el que viene” [“Is-the-postman-the-one-who-comes” – ‘It is
On clefting in English and Spanish 
the postman that comes’], what we mean is that there is speficially one per
son who comes and that person happens to be the postman. This shows us
that the CES [FRE] must be definite and identificational but not character
izing. NPs such as “un hombre” [‘a man’] or “un cartero” [‘a postman’] are
commonly used as characterizing, especially if they lack a complement. How
ever, “un cartero” [‘a postman’] could be interpreted as being identifying and
not characterizing in a context like the following: let us suppose that two po
licemen and a postman wear a similar uniform and, in seeing one person
approaching, we say “viene un policía” [“comes-a-policeman” – ‘there comes
a policeman’], to which somebody could answer “No, es un cartero el que
viene” [“No,-is-a-postman-the-one-who-comes” – ‘No, it is a postman that
comes’]. In this particular case, “un cartero” [‘a postman’] is used in an identi
ficational fashion, which explains why the resulting PdR [cleft] is acceptable.”
[Our translation and glosses]
In the light of the above, and while acknowledging the need to view the speci
ficity/identifying restriction as a continuum rather than as a cut-and-dried dichotomy,
we believe that some predictions can be made as to which kind(s) of constituents will
be more acceptable as EIFs. Thus, roughly speaking, characterizing NPs, PPs without a
specific reading, and factual-like finite and non-finite clauses are at best non-felicitous
as EIF constituents in Spanish, as shown in (21) below:
(21) a. *Es una espía quien/* como María era.
‘It is *(a) spy who/ *how Mary was.’
b. Es el médico / ?? un médico / # médico quien Juan es.
‘It is the doctor / ?? a doctor / doctor who Juan is.’
c. *Es según todos los indicios que éste ha sido un buen año [But: Fue debajo
de la cama donde ella encontró el dinero].
*‘It is by all indications that this has been a good year’ [But: It was under
the bed where she found the money].
d. ?Es que ella es vil y recelosa lo que él dice [But: Lo que él dice es que ella es
vil y recelosa].
‘It is that she’s sneaky and underhanded that he says.’
e. ?Es que ella fuese tan recelosa con respecto a ello lo que me más molestó.
‘It was that she was so sneaky about it that bothered me the most.’
f. *Es destruir esa carta lo que yo hice.
*‘It is destroy that letter that I did.’
g. ?Es destruir esa carta lo que constituye ahora una prioridad para nosotros.
‘It is destroy that letter that’s the most important thing for us to do now.’
h. ?Es rancio como el pan se ha puesto.
‘It is stale how the bread has gone.’
In addition, the aforementioned structural and functional variation exhibited by the
EIFs in both English and Spanish proves that these constructions allow for a thematic
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
flexibility that is paradigmatic in orientation. In other words, given that neither it nor
ser (‘be’) are tied to a particular Complement, ICs are capable of achieving maximum
textual prominence by schematizing and giving thematic – or initial – prominence
to, as well as by bringing into Focus, virtually any element that may occupy this
slot. Hence, Subject, the syntactic function which is least likely to carry intonational
prominence in ordinary English declaratives, gets structural and normally intonational
prominence, at the same time infusing this constituent with discourse salience and a
flavour of incompleteness. Similarly, EIFs realized by Adjuncts, Direct Objects, and
the other syntactic functions, which are least likely to be schematized, also receive the
matic and therefore additional discourse salience through the cleaving strategy, as we
shall see in Sections 4 and 5.
Lastly, turning to constituent 4 in the general pattern, if present,14 this is mostly
structured in ICE-GB as a free relative element (FRE henceforth) (98.35%), but it may
also be realized by a free -ing non-finite clause (1.65%), all in the same spoken text,
a sports commentary (e.g. And it’s Altaya leading, ICE-GB: S2A-006 #169:4:A; But it’s
Australia now spreading the ball to Belcher, ICE-GB: S2A-004 #360:1:A). Figure 1 above
represents the FRE as an immediate constituent of the IC sentence as a whole, de
pendent on the be-clause. The implication is that the referent of the EIF reveals the
identity of the relative pronoun of the FRE as a dominant coreferential entity (in a
hypotactic construct), rather than as being restricted in meaning by the relative post-
Modifier. Hence the FRE, resembling the non-restrictive type of relatives, elaborates
its antecedent as information that is separate from, and secondary to, the remainder of
the dominant clause.15 However, FREs are closer to restrictive relatives (R-Rels) in that
they can be introduced by a relative phrase (who, that, etc.) or by zero, but differ from
them in mainly four respects:16
1. FREs allow much more freedom in relativizing (e.g. PPs) than R-Rels;
2. the relative pronoun of FREs is omissible in informal style when the relative ele
ment functions as Subject; while in the same context, such a pronoun can never be
omitted in R-Rels (It’s his Mum falls in love with him, ICE-GB: S1A-006 #128:1:A
vs. *Your Mum was the person/one falls in love).17
3. both constructions can be distinguished prosodically, so that in unmarked condi
tions an utterance such as It was the vase that she broke would have the main stress
falling on vase and possibly also a subsidiary intonation pattern on broke in the IC
reading, and on broke in the R-Rel clause; and
4. unlike R-Rels, FREs do not need to agree in person with their antecedent (e.g. It
was only in December last year that the European Council decided to establish a new
bank, ICE-GB: S1B-054 #19:1:B).18
The picture outlined above for English can, broadly speaking, be taken to be duplicated
for Spanish. However, a number of important qualifications should be made regarding
features (2) and (4). First, the relative pronoun of FREs, unlike English, is not omissi
ble, under any circumstances, in Spanish (cf. Mas no son ellas las que se están buscando,
CREA Corpus, ABC, 24/12/1983: La historia sumergida [‘But it is not them that are
On clefting in English and Spanish 
currently being searched for’; *Mas no son ellas se están buscando]). Second, FREs
generally agree in person and number with their antecedent (*Es/son mis padres quienes
me han regalado la moto [example taken from Moreno-Cabrera 1999: 4255] ‘It is my
parents who have given me the motorbike as a present’). However, exceptions to the
above tendency may occur, especially in those configurations in which the antecedent
is non-human and plural (Es/Son las ambulancias lo que más ruido hace) [example
taken from Moreno-Cabrera 1999: 4255 ‘It is ambulances that are the noisiest’].
. Semantic properties
ICs are non-permutable grammaticalized identifying constructions – as opposed to the

attributive type19 – which entail an existential presupposition, carried by the FRE, and
an implicature of exclusiveness, associated with the construction as a whole. These three
features are analyzed in 4.1–4.3 below, respectively.
. Identifying constructions
ICs have been regarded as non-reversible identifying clauses, as illustrated in examples

(23) and (24) below, of a grammaticalized (Huddleston 1984) or metaphorical type
(Borkin 1984; R. Martínez-Vázquez 1997) as a result of the it + be predication that is
encoded in X, the Head clause. This predication fossilizes the X-Identifier (IR) – Iden
tified (ID) pattern, conferring upon the EIF/IR – usually also the locus of intonational
prominence – a special kind of discourse salience.
(23) a. It’s a character defect (IR) that makes them go into politics (ID).
b. *What makes them go into politics (ID) it ’s a character defect (IR).
(24) a. Es mi hijo (IR) el que se me viene a la yema de los dedos (ID).
‘It is my son that pops up in my head.’
b. El que se me viene a la yema de los dedos (ID) es mi hijo (IR).20

Furthermore, while some authors question the identificational nature of clefts (e.g.
Sornicola 1988: 358–359), pace the widely accepted view alluded to above that they do
imply an identifying (or equative) relation, we would concur with Borkin (1984: 122)
that the core meaning of clefts is precisely that of sanctioning a relationship of identi
fication in a given context, no matter what other communicative functions they may
serve (cf. Section 5 below). Our data, displayed in Table 6 below, have attested that this
relationship of identification may be of three different kinds:
1. intensive, with the meaning ‘X is Y’: This was expressed in 78.67% ICE-GB, 68% in
CREA and 91% in ARTHUS, and has been illustrated in the majority of examples
considered so far;
2. circumstantial, of the ‘X is + [adverbial meaning] Y’ type, in which the rela
tive constituent of the FRE acts as Adjunct – as in (27) below; this relationship
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
was present in 20.85% of the ICs in ICE-GB (88 items), mostly involving time
(38.63%), followed by place (35.22%), manner (15.90%), reason (7.95%), pur
pose (1.13%), and result (1.13%); by contrast, in CREA and ARTHUS we just
found place (100%) and time (100%), respectively. Examples from English and
Spanish follow in (27) and (28) below:
(27) a. Time: And it was only <„> then that it i it uh the the the wound opened.
(ICE-GB: S1B-066 #25:1:A)
b. Place: But it was in one of the office buildings that I discovered the letters
<,> (. . . ) (ICE-GB: S2B-023 #61:3:A)
c. Manner: It was as a Guardsman that he came to the Second Battalion
which now he commands and eventually became a lance sergeant instruc
tor at the Guards Depot. (ICE-GB: S2A-011 #124:1:A)
d. Reason: It is for that reason that I have taken the assumption similar to
that taken in many other areas of an oil price down to twenty-five dollars
by the end of nineteen ninety-one <,> (. . . ) (ICE-GB: S1B-052 #80:1:C)
e. Purpose: It is only with the aim of achieving the same speed as the brain
that new methods of implementation are sought.
(ICE-GB: W2A-032 #62:1)
f. Result: It was as a result of something sticking out a piece of metal or
whatever uh in the in the steps that’d caused that injury.
(ICE-GB: S1B-067 #156:1:A)
(28) a. Time: Y fue durante uno de ellos cuando me descubriste con Fernando.
(ARTHUS Corpus, SUR, 35, 3)
‘And it was during one of those that you discovered me with Fernando.’
b. Place: Es ciertamente contradictorio que si todo iusnaturalismo es una
teoría de la moral y de la conexión subordinada del Derecho positivo en
ese sistema, sea en ese ámbito donde se genere la distinción. (CREA Corpus,
1983, Peces-Barba, Gregorio; Introducción a la filosofía del derecho).
‘It is certainly contradictory that if every jusnaturalism is a theory of
moral principles and of the subordinate connection of positive Law
within that system, it is in this domain where the distinction is conceived.’
c. Manner: Es con mucha paciencia como estoy escribiendo mi nuevo libro.
‘It is with a lot of patience that I am writing my new book.’
3. possessive, encoding a relationship of the ‘X has Y’ kind – as in (29); this was

restricted to ICE-GB, where it occurred very restrictively (0.23%), and has the
peculiarity that the marker of possession may be located either in the EIF – as in
(29a) – or in the FRE – as in (29b).
(29) a. It’s Mrs. Thatcher’s unselfish determination that free enterprise should
not be held back by unwelcome exposure to public scrutinity that has
allowed the industry to develop. (ICE-GB: W2B-014 #6:1)
On clefting in English and Spanish 
b. It’s Mrs. Thatcher whose unselfish determination (. . . ) has allowed the

industry to develop.
Examples (29a) and (29b) above represent two presentational variants with slightly dif
ferent effects. In (29a), both possessor (Mrs. Thatcher) and possessed (unselfish deter
mination. . .) are presented as IR/EIF/FA within one and the same attentional window.
In other words, in (29a) the determination is in Focus. By contrast, in (29b) posses
sor and possessed are presented in two separate moves: the possessor (Mrs. Thatcher)
is in Focus of the IC, while the possessed, though informative, is mapped onto the
presupposed move of the construction as subsequent and subsidiary information.
. Existential presupposition
The denotatum of Y, the dependent clause of ICs, is an open proposition, not a refer
ential expression proper, in which information is presented as presupposed. In other
words, the information borne by Y is an open proposition that is true regardless of the
truth value of the sentence as a whole. By way of illustration, let us consider the IC
in (1a) above, now renumbered (30a). Here we would arguably have the open propo
sition in (30b) (cf. Lambrecht 2001) and the presupposition glossed as (30c), which
could be more formally represented as (30d) (cf. Delin 1990):
(30) a. It’s a character defect that makes them go into politics.
b. that makes them go into politics (x)
c. Something makes them go into politics
d. ∃x makes them go into politics (x)
Both the presupposition and the open proposition contain an existentially-quantified
variable (x)/“something”, whose identity is highlighted in (30a), a character defect, as a
participant in the process of going into politics. This follows simply from the fact that –
as already noted in Section 4.1 – ICs have a “specificational” (or identifying) meaning:
they specify a value for a variable. Thus (30) specifies the value a character defect for
Table 6. Type of cleft identification across ICE-GB, CREA and ARTHUS
ICE-GB CREA ARTHUS Total

Intensive 332 (78.67%) 17(68%) 21 (91%) 370 (79.22%)
Circumstantial 88 (20.85%) 5 (23%) 2 (9%) 95 (20.34%)
time 34 (38.63%) 0 (0%) 2 (100%) 36 (37.89%)
place 31 (35.22%) 5 (100%) 0 (0%) 36 (37.89%)
manner 14 (15.90%) 0 (0%) 0 (0%) 14 (14.78%)
reason 7 (7.95%) 0 (0%) 0 (0%) 7 (7.36%)
purpose 1 (1.13%) 0 (0%) 0 (0%) 1 (1.05%)
result 1 (1.13%) 0 (0%) 0 (0%) 1 (1.05%)
Possessive 2 (0.47%) 0 (0%) 0 (0%) 2 (2.10%)
Total 422 (100%) 22 (100%) 23 (100%) 467 (100%)
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
a variable in the open proposition ‘the x that makes them go into politics’, whose ex
istence is presupposed. This suggests that the content of an IC is not exhausted by
the statement of its presupposition, but also that it contains an assertion, to the ef
fect that the EIF serves to instantiate the variable contained in the presupposition. An
important function of ICs is therefore to separate presupposed and asserted material
syntactically, and, in addition, to indicate that an instantiation of the variable in the
presupposition is taking place.
To round off this section, it should be borne in mind that presupposition and
assertion are understood here as logico-semantic notions independently of their cog
nitive status, that is, irrespective of whether or not the hearer is thought by the speaker
to already be aware of the presupposed or asserted information. Thus, a proposition
may convey given information, but need not be presupposed (e.g. It can’t be my father
(you saw), as an answer to I saw the man, where it and my father stand as given infor
mation, but they are not presupposed). Or, vice versa, informationally new items may
occur within a presupposition (e.g. It is this last fact which makes marine pollution of
any type so serious, ICE-GB: W2B-029 #38:1, where the FRE is comparatively more in
formative than the EIF). Be that as it may, the fact remains that the EIF of clefts encodes
information to be accommodated by the hearer: as an antecedent for the presupposed
information in the background of the discourse context into which it is incorporated
(Karttunen 1974: 191).
. The exhaustiveness implicature
It is generally taken as a non-controversial fact in the literature that ICs (with the
exception of the there-subtype) convey a conventional implicature of exhaustiveness
or exclusiveness which is absent from uncleft constructions and which is often ex
plicitly asserted by such words as only, precisely, just, actually, etc., as illustrated in
(31) below (cf. e.g. Atlas & Levinson 1981; Davidse 2000: 1121; Declerck 1984: 271;
Halliday 1967b: 238; Halvorsen 1978: 15; Hedberg 2000; Ward, Birner, & Huddleston
2002: 1416):
(31) a. It needs <,> It does need the individual to set the whole <,> thing doesn’t
it because it’s only the individual who can do it on.
(ICE-GB: S1A-045 #16:1:A)
b. Is it just that the food is becoming boring. (ICE-GB: S1A-059 #105:1:A)
c. The other thing is that it was actually Linda who said you can get a lot for
a video in South Africa. (ICE-GB: W1B-007 #11:1)
Broadly, the exhaustiveness feature of ICs has been attributed to the mappings of asser
tion, New information and intonational prominence onto the IR/EIF, on the one hand,
and presupposition, Given information and lack of intonational prominence onto the
ID/FRE, on the other (in the unmarked instances). Thus consider the examples in (32),
as opposed to ‘John kissed Mary’, where no exhaustiveness implicature nor existential
presupposition is implied (adapted from Collins 1991: 70):
On clefting in English and Spanish 
(32) Assertion Implicature

a. It was John that Mary kissed. Positive positive
b. It wasn’t John that Mary kissed. negative positive
c. It was John that Mary didn’t kiss. positive negative
d. It wasn’t John that Mary didn’t kiss. negative negative
Unlike Mary kissed John, the constructions in (32) all specify that out of the set of
salient individuals that could have also been kissed by Mary (Peter, Tom, Paul, etc.)
it was ‘John who was kissed by Mary’. These four constructions, however, differ in
that those with identical assertions convey different implicatures, and those sharing
the same implicatures have different assertions. Thus, both (32a) and (32b) share the
implicature that ‘there is somebody that Mary kissed’, but while the former asserts that
‘Mary kissed John’, the latter asserts the opposite, namely that ‘Mary did not kiss John’.
Similarly, both (32c) and (32d) presuppose that ‘there is somebody that Mary didn’t
kiss’, but both assert opposite propositions: (32c) that ‘Mary didn’t kiss John’ and (32d)
that ‘Mary kissed John’.
In this connection, it seems appropriate to bring into focus Sornicola’s (1988: 368–
369) argument against the unicity entailment as a necessary inherent feature of ICs on
the grounds that “[i]t is doubtful whether the reply in (1) [our (33)] [r]eally means
that ‘It was John and only John’/‘It was John and nobody else who came”’.
(33) a. Who came?.
b. It was John.
In this respect, it must be highlighted that we have also found in our corpora tokens –
some of which are reproduced in (34) below – which would at least question, if not
refute, the unicity feature typically associated with the EIFs of ICs:
(34) a. It’s greed and self-interest or private interest that has caused the collapse.
(ICE-GB: S2A-057 #3:1:A)
b. What was it that attracted you to Merlin?
(ICE-GB: S1B-048 #191:1:A)
c. Whether it is a socially initiated kick that brings out technology, or tech
nology that responds to an environmental kick is <unclear-word> par
ticular to each example one may choose in relation to any society.
(ICE-GB: W1A-012 #69:2)
d. However, it is not merely economic differences between the two countries
that cause varying patterns of rural-urban migration.
(ICE-GB: W1A-013 #92:3)
e. It is domestic politicians who communicate about the threat in order to
mobilise public support for their own policy and power base. It is also the
mass media who communicate about it in order to mobilise an increased
audience. (ICE-GB: W2A-017 #28-31:1)
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
As things stand, we would argue that a strong reading of the unicity entailment of ICs
‘only X (EIF) and nobody/nothing else’ is present only when this meaning is made
explicit by, for example, such markers as only, precisely, exclusively, and the like. Oth
erwise, like Sornicola (ibid.: 236), we would subscribe to a weak reformulation of the
unicity entailment of ICs on the grounds that it is the topicality of the EIF that is at
issue, not its referential or semantic role in the propositional scheme of the construc
tion. Thus, in It was John that Mary kissed, we will have the implicature ‘John is the
FA at that discourse stage’, and not ‘John and only John was kissed by Mary’. In other
words, we would regard the unicity entailment in terms of speaker construal of the EIF
as being the FA at a given point in discourse out of a limited set of candidates to prop
erly complete the proposition at issue (cf. Borkin 1984: 126). A compelling piece of
evidence that the exclusiveness feature of ICs is textual – not ideational – stems from
the fact that the feature concerned is not affected by such grammatical processes as
negation or interrogation, which do affect ideational meaning such as the one asserted
in uncleft sequences by words like only (e.g. The duke didn’t only give my aunt that
teapot vs. It is not that teapot that the duke gave to my aunt, where in the former the
implicature is contradicted, but in the latter it is retained).
. Discourse-cognitive functions
We shall now take a look at the discourse functions of ICs. Due to limitations of space,
we will restrict our discussion to what have been profiled as the three most relevant
aspects of their functional-cognitive load: (i) their thematic flexibility and its side-
effects in terms of Topic (dis)continuity; (ii) their newness-orientation and (iii) their
interpersonal character.
. Thematic flexibility
As pointed by Halliday (2004: 97), ICs admit two levels of thematic analysis: (1) a
global or constructional pattern, in which the Head clause (X) is the Theme, or ini
tial constituent, of the construction as a whole, and the dependent clause (Y) is the
Rheme; and (2) a local pattern, in which both X and Y have their corresponding
Theme-Rheme arrays, that is, their corresponding clause initial participants, circum
stances, processes or attributes (Theme), and their rhematic continuations. As a result
of the interaction of both patterns, in unmarked or prototypical cases, the EIF receives
(at a local level) rhematic, syntactic and prosodic highlighting; while (at a global level)
the thematic predicative structure it is/was-es/era/fué + EIF profiles the construction
as a whole, (re)focusing the interactants’ attention on the EIF/IR which is affirma
tively or correctively identified with the presupposed (ID) encoded in the subsequent
or rhematic FRE.
On clefting in English and Spanish 
It is your BROther who knew Peter

Es Tu herMAno el que conocía a Peter
(i) local Theme-Rheme pattern Theme Rheme/Focus Theme Rheme
(ii) global Theme-Rheme pattern Theme/TK/IR Rheme/ID/VL
Figure 2. Thematic structure of clefts
Moreover, as seen in Section 3, the structural and functional variation exhibited

by the EIFs of ICs demonstrates that these constructions allow for a thematic flexibility
that is paradigmatic in orientation. In other words, ICs are capable of achieving max
imum textual prominence by giving thematic, or initial, prominence and by putting
in Focus virtually any element that may occupy this slot. Hence Subject, the syntactic
function which is least likely to carry intonational prominence in ordinary declara
tives, gets structural and normally intonational prominence, at the same time infusing
this constituent with discourse salience and a sense of incompleteness. Similarly, EIFs
realized by Adjuncts, Direct Objects, and the other syntactic functions, which are least
likely to be thematic and as result least liable to profile the subsequent discourse, also
receive through the clefting strategy thematic and so additional discourse salience.
Endowed with this thematic flexibility potential, ICs reveal themselves as very ver
satile cohesive devices capable of implementing four main discourse strategies (cf. e.g.
Alonso-Raya & Castañeda-Castro 1997; Borkin 1984; Moreno-Cabrera 1999):
1. Topical: This is the case when ICs specify the current (local or global) discourse
Topic (or some aspect of it) out of a number of alternatives, as in (35):
(35) a. And just as it is the development of this body of work that is the focus of
my interest, and not Coleridge’s contribution to, or conformation with, dis
tinct forms of intellectual practice, so this book is not a description of
Coleridge’s return to Christian orthodoxy. (ICE-GB: W2A-003 #10:1)
b. La expulsión es aquella medida que tiene como objetivo, también, que el
extranjero abandone el territorio del Estado, pero la medida se produce
después de que, durante un cierto tiempo, ese extranjero ha permanecido
regularmente dentro del territorio. Las razones de la expulsión tienen que
ver con el aseguramiento del orden público. Es una medida de carácter
preventivo de policía administrativa.
Por último la extradición es una medida de cooperación internacional
en la represión de la delincuencia. No es preventiva, pues, sino represiva.
Tiene por objeto evitar que una persona perseguida condenada por las
autoridades judiciales de un Estado pueda escapar a la persecución o a la
ejecución de la pena simplemente por el hecho de encontrarse en el ter
ritorio de otro Estado.
De estas tres medidas son las dos últimas las que pueden recaer sobre aquel
los que hayan obtenido el refugio o el asilo. (. . . )
(CREA Corpus, 1991, Diego López Garrido, El derecho de asilo)
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
‘Expulsion is the step also aimed at making the foreigner leave the terri
tory of the Country, but this procedure is effected after, over a period of
time, the foreigner in question has stayed regularly within that territory.
The reasons for expulsion have to do with securing law and order. It is a
preventive procedure for administrative police.
Finally, extradition is an international cooperation procedure for the re
pression of delinquency. It is, therefore, not preventive, but repressive. It
is aimed at preventing a convict prosecuted by the judicial authorities of a
country from avoiding prosecution or the enforcing of a sentence simply
by virtue of being in the territory of another Country.
Out of these three procedures, it is the last two that may be effected upon
those who have obtained refuge or asylum.’
2. Transitional: This strategy obtains when ICs (re)introduce new or deactivated Top
ics or when they provide new or deactivated spatio-temporal settings, thereby
opening up new discourse stages – as in (36) below:
(36) a. It’s at this point that I think I can return to this theme of the name of the
prophet. (ICE-GB: S2A-036 #77:1:A)
b. Una vez que se ha conseguido atravesar una primera fase de admisión,
aparece el problema jurídico capital, consistente en calificar a una persona
como poseedora de los requisitos que se exigen para adquirir la condición
de refugiado. Es en esta fase en donde hay un mayor desarrollo de las inter
pretaciones judiciales del Convenio.
(CREA Corpus, 1991, López Garrido, Diego; El derecho de asilo)
‘Once one has gone through the first admission phase, there arises the
fundamental juridical problem, viz. that of assessing a given person as
fulfilling the requirements necessary to acquire the condition of refugee.
It is at this stage where a greater development of the juridical interpretations
of the Agreement takes place.’
3. Corrective: As the name suggests ICs can reformulate or correct old/previous Top
ics, as illustrated in (37):
(37) a. Still, it was not Rome that was merged in Italy, but Italy that was absorbed
into the Roman citizen body. (ICE-GB: W2A-001 #22:1)
b. No ha entrado Juan, es Pedro quien ha entrado.
(Example taken from Moreno-Cabrera 1999: 4299)
‘John has not come in, it is Peter who has actually come in.’
On clefting in English and Spanish 
4. Emphatic: This label is applicable to ICs that specify a given conceptual do
main with an emphatic purpose, thus presenting information as unquestionable –
as in (38):
(38) a. Indeed I argue that it’s the very complexity and flexibility of the British
class system which has enabled it to resist the pressures which a more rigid
structure would have been unable to bear. (ICE-GB: S2B-036 #20:1:A)
b. ¡Pues claro que es Pedro quien ha entrado! (Moreno-Cabrera 1999: 4299)
‘Of course, it is Peter who has come in!’.
. Newness-orientation
Throughout this paper, we have made it clear that ICs represent focal build-up de
vices because, through the predicative mechanism, the EIF becomes the FA and also
tends to attract intonational prominence. By placing, in the unmarked cases, a nu
clear tone (normally rising) on the EIF element itself, speakers provide this constituent
with a contrastive character and, at the same time, emphasize that the identifying con
struction is incomplete until the rhematic ID-presupposition encoded in Y is uttered
(Halliday 1967a, b, 2004; Harries-Delisle 1978; Hornby 1971). As a result, the EIF (and
the information encoded by it) tends to be presented and perceived as newsworthy, a
tendency which explains why ICs have traditionally been said to be newness-oriented.
This orientation is ratified by the three patterns found in our corpora and also attested
in Collins (1991), the frequency of which is reported in Tables 7 and 8 below: (i) New
+ Given, (ii) Given + New, and (iii) New + New.
For reasons of space, we will not go into the nitty-gritty of the notions of Given and
New information and associated concepts as a detailed account has been provided else
where (Gómez-González 2001: 15–58; Gundel 1999, 2002). For our purposes it suffices
to say that we shall interpret the Given-New distinction referentially. Thus, following
Collins (1991), the label New will be used to refer to information that is presented
either as not previously mentioned or as Contrastive, the latter representing typically
recoverable, but to-be-attended-to information associated with a formulation of con
trast. In addition, Given designates information that has been previously mentioned
and is expressed through such backgrounding strategies as anaphora, deixis, repeti
tion, ellipsis, etc., or otherwise is indirectly retrievable via inference (e.g. relationships
of synonymy, hyponymy, hyperonymy, etc.).
Now going back to pattern 1, New + Given ICs have an EIF that carries either non
recoverable or contrastive information, and a FRE that is Given, as illustrated in the
variations of this pattern collected in (39) below:
(39) a. New-Given: Who communicates about the threat and for what purposes ?
The answer is extremely simple. It is domestic politicians who communicate
about the threat in order to mobilise public support for their own policy and
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
power base. It is also the mass media who communicate about it in order to
mobilise an increased audience. (ICE-GB: W2A-017 #28-31:1)
b. Contrastive-Given: If herd size is small and can’t sustain a group alter
native lifestyles must be resorted to. This leads to a completely sedentary
way of life. However, as we can see from <unclear-character> Swift ’s eco
nomic account of pastoralism, it is not just the poor whole who makes the
transition to sedentism. (ICE-GB: W1A-11 #24-26:1)
c. New-Inferrable: I suppose, I said that he had formed an attachment to
some young woman in London. (. . .) Natalie wasn’t at all the sort of girl
who dresses up in mink and mascara and gets her picture in the Sunday
newspapers. (. . .) He found her rather colourless, especially by compari
son with Isabel. It occurred to me that it might have been the contrast with
Isabel that Barnsley had found attractive. (ICE-GB: W2F-011 #81:1-92:1)
d. Contrastive-Inferrable: There are too many books in this house. (. . .)
Uhm I think I ’m a bit too intense about books. So on going back to
your to your childhood it was your mother wasn’t it who was the the driv
ing force behind all of this behind this sort of intellectual rigour.
(ICE-GB: S1B-046 #54:1:B-57:1:A)
In (39a) above there are two ICs in a series; both have the same informative pattern
with two New EIFs (domestic politicians and mass media) and two Given FREs directly
recoverable from previous cotext (who communicate. . .). In (39b) the EIF the poor is
contrastive with alternative group lifestyles, while the information encoded in the FRE
is clearly recoverable from the repetition of sedentism. In (39c) the contrast with Is
abel introduces new information: it provides an explanation for Barnsley’s attraction
towards Natalie. The information encoded in the FRE is in turn inferrable from the
previously mentioned attachment to some young woman in London. Finally, the EIF
of (39d), your mother, is contrastive with respect to other relatives that could have
influenced the speaker’s childhood, while the information in the FRE is inferrable
via the association between books or the idea of being too intense about books and
intellectual rigour.
Turning to pattern 2, Given + New, this applies to ICs which normally have a
short EIF which carries typically anaphoric information and which prepares the con
text for the presentation of Y in the general pattern, which in turn encodes the more
newsworthy part of the construction. Variations of this pattern are gathered in (40):
(40) a. Given-New: In the Herbarium at Kew, botanists study plants from all over
the world in order to identify and name them and to provide methods
of helping other people recognise the plants around them. It is here that
dried specimens end their journey. (ICE-GB: W2B-030 #8-9:1)
On clefting in English and Spanish 
b. Given-Contrastive: The high salinity and wide temperature range are

maintained because the Gulf has only a narrow connection to the Indian
Ocean, which results in a limited exchange of water. And it is this last fact
which makes marine pollution of any type so serious.
(ICE-GB: W2B-029 #37-38:1)
c. Inferrable-New: Ay<l>atollah Khomeini, for instance, often killed as
many of his own citizens in a single day as the Shah had killed in his
entire reign. In the long term, it is the replacement of Arab dictatorships by
democracies that will be the best guarantee of freedom, stability and peace
in the Middle East. (ICE-GB: W2E-001 #60-61:3)
d. Inferrable-Contrastive: Women have certainly not benefited from the
greater economic success of other Third World countries to the same ex
tent as men. (. . .) They have much lower access to education, to secondary
and/or wage earning possibilities, in all countries across the Third World,
although it is fair to say that it is in the least developed countries that they
have suffered the most. (ICE-GB: W1A-014 #30:1)
In (40a) above the EIF is a place deictic (here: the Herbarium at Kew) whose infor
mational status is Given; the FRE introduces the New idea of ‘dried specimens ending
their journey there’. The EIF in (40b) is Given because anaphoric (this last fact), while
in the FRE marine pollution is contrastively New, the contrast being with other kinds
of pollution. In the case of (40c), the addressee is expected to draw an inference re
lationship between two dictators (Khomeini and the Shah) and dictatorships in order
to decode the EIF (although neither of these gentlemen are Arabs, which could lead
to the implication that only Khomeini was a dictator); whereas the FRE conveys New
information, namely that the replacement of Arab dictatorships by democracies are a
guarantee of freedom. In the last example of pattern 2, (40d), the addressee is expected
to infer a link between countries across the Third World and least developed countries,
while the FRE singles out the places where women suffer most (the least developed) out
of the presupposed set (Third World countries).
In the last informational pattern of ICs, New + New, both the main and the depen
dent clause encode newsworthy information, as shown in the following combinations:
(41) a. New-New: It was in nineteen hundred and six that the Queen’s great-
grandfather King Edward the Seventh decreed that privates in the House
hold Cavalry should henceforth to be known as troopers.
(ICE-GB: S2A-011 #91:1:A)
b. New-Contrastive: And I would like to say that it’s the Labour Councillors
on the Greater London Arts Board who are more to blame than anybody.
(ICE-GB: S1B-022 #105:1:C)
c. Contrastive-New: I began this thesis with images of the positive, gen
erous sexuality of the ancient Turkish deities and I conclude, with the
inescapable fact that, bar the Tantric tradition, it is the sexuality of the
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Goddess, and consequently real women, that has suffered most in the pro
cess of transition that the Aryan heroes brought into the world’s symbolic art
forms so variously enshrined in each religious tradition.
(ICE-GB: W1A-008 #11:1)
d. Contrastive-Contrastive: In Kylie Tennant’s Tiburon (1935), some women
take part in a 1930s strike in which the unemployed refuse to ‘scab’ on
other unskilled labourers, and therefore do not take jobs offered at a lower
rate of pay than that agreed by the majority of such labourers. But it is the
men who strike, and the women who dole out soup to the unemployed who
are refused relief, that is, unemployment benefit.
(ICE-GB: W2B-009 #43-44:1)
In (41a) both the main (X) and the dependent clause (Y) of this IC represent fresh
information; whereas in (41b), only X conveys fresh news, while Y contains a con
trastively new item (more). In (41c) the EIF, the sexuality of the Goddess, is contrasted
with that of male divinities and real men, while the content of the FRE is New. In
(41d) in turn both the EIFs (labourers: men-women) and the FRE contain contrastive
information (strike-dole out. . .).
The frequencies of the aforementioned patterns in ICE-GB are displayed in Table
7 below. Our findings can be taken to point to pattern 1 as the most frequent, followed
by patterns 3 and 2 (51.65%, 29.62% and 18.72%, respectively), a tendency also re
ported in Collins (1991: 111ff.) (with the only difference that in the latter, pattern 2
outnumbers pattern 3). The fact that pattern 1 emerges as the unmarked one is proba
bly a side-effect of the tendency of the global Theme of ICs to encode in the EIF either
non-recoverable information or contrastive information (as in (39) and (40) above).
Table 7. Informational classification of clefts in ICE-GB

EIF RE Number %
Unmarked 1a New Given 144 34.12
1b Contrastive Given 36 8.53
1c New Inferrable 29 6.87
1d Contrastive Inferrable 9 2.13
Subtotal 218 51.65
Marked 2a Given New 62 14.69
2b Given Contrastive 5 1.18
2c Inferrable New 10 2.36
2d Inferrable Contrastive 2 0.47
Subtotal 79 18.72
3a New New 76 18.00
3b New Contrastive 13 3.08
3c Contrastive New 32 7.58
3d Contrastive Contrastive 4 0.94
Subtotal 125 29.62
On clefting in English and Spanish 
In the latter case the EIF establishes a contrast with respect to the members of a def
inite set, whether reinforced or not by contrastive focal stress. By contrast, the FRE
tends to be associated with givenness because of its hypotactic status: on being sub
ordinated to the impersonal it + be thematic predication, the FRE is imbued with a
hedged, reportive flavour that presents it as logically presupposed.
Turning now to Spanish, Moreno-Cabrera (1999: 4298–4300) has also provided
confirming evidence that cleft constructions are newness-oriented (“Focus-bearing”),
this being the reason why these, unlike pseudo-clefts, are not felicitous in such contexts
as discourse openings (cf. also Fernández-Leborans 2001: 299–304), as illustrated in
(42) below (examples and acceptability judgements are Moreno-Cabrera’s):
(42) a. ¡Buenos días! De lo que voy a hablar hoy es de la universidad española.
‘Good morning! What I am going to talk about today is the Spanish uni
versity.’
b. ¡Buenos días! De la universidad española es de lo que voy a hablar hoy.
‘Good morning! The Spanish university is what I am going to talk about
today.’
c. *¡Buenos días! Es de la universidad española de lo que voy a hablar hoy.
‘Good morning! It is about the Spanish university that I am going to talk
today.’
The reason for this acceptability difference, according to Moreno-Cabrera (ibid.), lies
in the fact that the first two sentences, unlike the third variant, imply a neutral specifi
cation belonging to a conceptual domain which is recoverable in terms of the extralin
guistic context. In other words, with regard to (42), when we attend a conference, we
expect the lecturer to deliver his/her conference on a particular Topic.
To round off this section, an important observation must be made – as evidenced
in the patterns just described – that contrast is frequently, but not necessarily overtly
expressed in ICs, much in the same way as the FRE tends to contain referentially Given
information, but need not do so. These are only tendencies, not inviolable principles
(cf. Borkin 1984: 126; Collins 1991: 155ff.; Gundel 2002).
. Interpersonal flavour
In order to attest the interpersonal flavour of ICs, that is, their ability to encode the
speaker/writer’s attitude underpinning a proposition (Delin & Oberlander 1995), we
need to take a look at their distribution across our corpora as displayed in Tables 8, 9
and 10a–b.
Surprisingly enough and contra such studies as Collins (1991) and Miller &
Weinert (1998: 197), which seem to associate ICs with the written mode, our ICE
GB results suggest that these constructions are considerably more numerous in speech
than in writing. The clues for this skewed distribution are to be found in the main
communicative properties of the construction, i.e. specificational meaning, thematic
flexibility and newness-orientation, which make them particularly suitable for ar
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Table 8. Frequencies of it-clefts across ICE-GB spoken texts
Text category No. of Clefts %

texts
Dialogue Telephone calls A: 45 4 1.4
S1 Direct conversations Private 45 42 14.68
Classroom lessons B: 20 18 6.3
Broadcast discussions Public 20 21 7.34 44.4
Broadcast interviews 10 13 4.54
Parliamentary debates 10 11 3.84
Legal-cross examinations 10 16 5.6
Business transactions 10 2 0.7
Monologue Spontaneous A: 20 53 18.53
S2 commentaries Unscripted 30 36 12.58
Unscripted speeches 10 8 2.8 52.45
Demonstrations B: 10 8 2.8
Legal presentations Scripted 20 25 8.74
Broadcast talks 10 20 7
Non-broadcast
speeches
Mixed Broadcast news 20 9 3.15 3.15
S2
Total 300 286 100 100
67.77%
gumentative or persuasive discourse. This is precisely the rhetorical function of the

spoken ICE-GB texts around which ICs mostly cluster: (i) Spontaneous Commentaries
(18.53%), (ii) Direct Conversations (14.68%), (iii) Unscripted Speeches (12.58%), (iv)
Broadcast Talks (8.74%), and Broadcast Discussions (7.34%). In all these textual cate
gories it is natural to find argumentative discourse aiming to discuss and/or to express
one’s point of view. Irrespective of their marked or unmarked status, ICs are suit
able tools for these (sub)genres in that they simultaneously express specificational
meaning. In addition, they associate the FI and New information with the Theme,
or even with the Rheme (FRE) of the construct (cf. patterns 2 and 3), thereby direct
ing the addressee’s FA into a particular interpretation of the information. This guiding
rhetorical purpose also explains the most frequent usage of ICs (mainly patterns 1 and
3) in descriptive or informative written ICE-GB texts such as Humanities (12.5%),
Student Examination Scripts (10.33%), Creative Writing (9.55%), Untimed Student
Essays (9.55%), and (Non-academic) Natural Science (8.82%).
In speech (Dialogues and Monologues), there is a sense of mutually negotiated
fields; for the communicative act to be successful, the discourses created by the differ
ent participants have to share the same fields in the case of dialogue and, with regard
to monologues, both speakers and addressees contribute to the negotiation of the do
main. The interpersonal flavour of ICs makes them well suited for this negotiating
On clefting in English and Spanish 
Table 9. Frequencies of it-clefts across ICE-GB written texts
Text category No. of texts Clefts %

Non printed Untimed student essays A: 10 13 9.55
W1 Student examinations Non professional 10 14 10.3 29.41
scripts 19.85%
Social letters B: 15 8 5.88
Business letters Correspondence 15 5 3.68
9.56%
Printed Humanities A: 10 17 12.5
W2 Social sciences Academic 26.48% 10 7 5.15
Natural sciences 10 5 3.68
Technology 10 7 5.15
Humanities B: 10 9 6.62
Social sciences Non-academic 10 10 7.35
Natural sciences 30.88% 10 12 8.82 70.59
Technology 10 5 3.68
Press news reports C: Press 20 6 4.41
Administrative D: Institutional 10 1 0.74
Skills/hobbies 10 – –
Persuasive writing E: Editorials 10 4 2.94
Creative writing F: Fiction 20 13 9.55
Total 200 136 100 100
32.22%
Table 10a. Frequencies of Spanish cleft constructions across text types in CREA
Text category Number of sentences Number of clefts %

Journals-Politics 648 2 0.30
Journals-Literature 284 0 0
Newspapers-History 1,569 2 0.12
Newspapers-Philosophy 1,290 5 0.38
Books-Law 1,028 5 0.48
Oral-Debates 903 8 0.88
Total 5,722 22 100
Table 10b. Frequencies of Spanish cleft constructions by text type in ARTHUS
Text category Number of clauses Number of clefts %

Essay 1134 0 0
Narrative 2373 20 86.95
Oral 1203 0 0
Press 734 0 0
Drama 935 3 13.04
Total 6379 23 100
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
process. In writing, by contrast, the sense of institutional focus must be achieved by the
writer. The lack of simultaneity in the creation and perception of the written message
prevents the reader from negotiating the field; s/he has to discover it instead. Clefts are
useful devices to help the reader achieve this discovering process: by specifically identi
fying the FA, they (re)direct the reader’s attention towards a particular path, following
the writer’s dictates.
In Spanish, however, the picture changes. Clefts show a sparse, though continuous
distribution in oral texts (less than 1%), and seem to be more numerous in writ
ing (ARTHUS narrative (86%)). In this connection Fernández-Leborans (2001: 301)
claims that these constructions may be very productive in oral and dialogic discourse,
but they may also be the most effective way to encode emphasis in written discourse.
In addition, our data confirm that there exists a direct proportional relationship
between the frequency of ICs and degree of formality of the communicative situa
tion, that is, the extent to which discourse is pre-planned or not. Our data reveal ICs
to be typically associated with conditions of informality. More precisely, they tend to
cluster with spontaneous discourse in the spoken mode (e.g. Direct Conversations,
Commentaries, Unscripted Speeches, Broadcast Discussions) and with Non-academic
and Non-professional Writing (within the printed and non-printed categories, re
spectively).
In closing this section, let us consider some tokens of ICs which evidence their
opinion-giving nature, to which Perzanowski & Gurney (1997: 221–222) attribute ei
ther a slightly more accusatory or a slightly more laudatory tone than their non-cleft
SVO counterparts. First, we shall illustrate ICs with a laudatory tone – in (43) and
(44) – and then we shall turn to two examples of the accusatory type in (45) and (46):
(43) The Labour Party’s 1983 election manifesto, which committed it to a non
nuclear defence policy and, be it remembered, to withdrawal from the Eu
ropean Community, became known as the longest suicide note in history. A
Conservative Party manifesto hostile to Britain’s closer integration into the
European Community might qualify as Death Wish II. First and foremost,
such a campaign would bitterly divide the party. The long debate over full
participation in the Eu<l>ropean Monetary System revealed the rifts over
economic and monetary union. To fight an election on the issue that most
divides the party would be madness. For many Conservatives, it is an article of
personal and political faith that Britain must play a full part in European uni
fication. During the long years of Labour’s hostility to the EC, it was the Tories
who kept the European flame burning in Britain. For the roles to be reversed
would be a tragedy for many Conservative MPs and voters.
(ICE-GB: W2E-004 #48:3-64:3)
In example (43) above, the journalist is writing in favour of Britain’s integration into
the European Community and uses two ICs in a sequence to stress the positive role
of the Conservative Party over the years in this process, as opposed to the “suicidal”
position against that process endorsed by Labour Party. The laudatory tone is also
On clefting in English and Spanish 
present in (44) below, where the pressure exerted by Carter is considered to be the ma
jor driving force to free Nicaragua from censorship. As the text moves on, the American
support is again pinpointed as an essential element to preserve order in Nicaragua:
(44) Los especialistas en cuestiones iberoamericanas, sin embargo, no creen que la
cosa quede ahí. Después de haber dominado el país durante cuarenta años, la
familia Somoza encuentra hoy que le falta el mayor apoyo: el de Washington.
Fue la presión de Carter lo que hizo que se levantase la censura y lo que trajo
las críticas de Chamorro. Sin el respaldo norteamericano, esa dictadura hered
itaria no va a ser fácil de mantener. Claro que a los Somoza les queda todavía,
como estamos viendo, la Guardia Nacional. – J. M. C.
(CREA Corpus, ABC, 14/01/1978: Nuevos brotes de violencia en Nicaragua)
‘Experts on Latin American issues, however, do not believe that it will all stop
there. After having dominated the country for forty years, the Somoza family
finds their best support gone: that of Washington. It was the pressure exerted
by Carter that made it possible for censorship to be lifted and that also paved
the way for criticisms on the part of Chamorro. Without the American support,
that hereditary dictatorship will not be easy to keep. However, it is certainly
true that, as we can see, the Somozas can still count on the National Guard.’
Now we move on to the excerpts in (45) and (46), the two instances of accusatory ICs:
(45) But Glenda Jackson politics is an essential part of this [the crisis of arts and
entertainment industries in London] [. . .] Now you ’re a <,> Labour candi
date <,> Are you not uh <,> uh honest enough to say that there is a good
deal of of party politics involved in this at the moment. Yes but it was a Con
servative government that has systematically rate-capped so many Labour-led
boroughs in and around London (ICE-GB: S1B-022 #53-58:1:D) [. . .] Now let
John John Mortimer come in there. [. . .] And I would like to say that i it’s the
Labour councillors o o on the Greater London Arts Board who are who are more
to blame than anybody for for cutting their subsidies to the arts.
(ICE-GB: S1B-022 #106:1:C)
(46) Los felipistas buscan una salida airosa a Alfonso Guerra que deja la Vicese
cretaría. Mientras tanto, un nuevo misterio aparece en nuestra vida, saber
quiénes han asaltado y robado el despacho de don José Barrionuevo y para
qué. Y aparte de la explicación que nos ha dado José Luis Martín Prieto, que
probablemente fue Pedro jota quien se llevó la la caja fuerte, ¿no hay ninguna
explicación seria en los periódicos? No, aparte de esas de Martín Prieto, que
yo haya visto, no hay no hay ninguna otra.
(CREA Corpus, Protagonistas, 05/05/97, Onda Cero, Oral, Tertulias)
‘The followers of Felipe González are searching for an elegant way-out for
Alfonso Guerra, who is leaving the vice-secretary. In the meanwhile, a new
mistery emerges in our life, knowing who has broken into and burgled Mr.
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
José Barrionuevo’s office and why. And besides the explanation furnished by
José Luis Martín Prieto, that it was probably Pedro Jota who stole the safe, isn’t
there any sound explanation in the newspapers? No, besides those provided
by Martín Prieto, as far as I know, there is no other.’
In (45) the issue of the crisis in the London arts, culture and entertainment industry is
being debated in a broadcast discussion. It is interesting to see how the two instances
of ICs are used by different speakers with opposite views to put the blame for the situa
tion on the two antagonistic parties: Glenda Jackson on the Conservative government
and John Mortimer on the Labour councillors. In much the same vein, the Spanish ex
cerpt in (46) shows that the speaker (journalist Luis del Olmo) seems to be well aware
of the condemnatory colouring of the cleft construction and, in order to preserve the
desirable objective atmosphere of a debate, he decides to skilfully preface the cleft con
struction with probablemente [‘probably’]. In this way, the speaker shifts away his own
commitment to, and responsibility for, the claim encoded in the cleft construction.
. Conclusion
Throughout the preceding pages, we hope to have been able to provide satisfactory
supporting evidence for the major claims of this paper. Our main premise is that,
although a number of features such as contrastiveness, emphasis/Focus, and exhaus
tiveness are commonly identified as intrinsic features of IC constructions, their core
features are rather their specificational meaning, their thematic and cohesive flexibil
ity, their newness orientation and their interpersonal nature. We have concluded that
the specificational meaning of ICs derives from their constructional pattern in both
English and Spanish, namely ‘X (EIF) – or some aspect of it – is (positively or nega
tively) identified with Y (FRE)’, or put another way, ‘the X in the open proposition in
the FRE equals the EIF’, which in turn becomes the FA of discourse. It has been argued
that this specific identification is of a textual nature and is best seen as implying a weak
unicity entailment of the type ‘X (EIF) is identified as the FA at given point in dis
course out of a number of alternatives’. Furthermore, the specific identifying function
of ICs has been adduced to explain why the EIF is more likely than not referential and
specific (rather than characterizing-like). Besides, by exploring the interaction of the
specificity tendency of ICs in conjunction with other semantic factors such as factivity
and presupposition, we have shed some light on the otherwise puzzling restrictions ex
hibited by some finite and non-finite complement clauses as cleft constituents in cleft
constructions.
In addition, we have observed that, owing to their thematic flexibility, ICs can
be profiled as versatile cohesive devices which may display four main usages: (1)
topical (when specifying the current Topic); (2) transitional (when (re) introducing
New or deactivated Topics or spatio-temporal settings); (3) corrective (when refor
On clefting in English and Spanish 
mulating or correcting old/previous Topics); and (4) emphatic (when used with an
emphatic purpose).
Regarding the newness orientation of ICs, our data have shown that New + Given
and New + New are the most productive patterns. In this connection, we have also
emphasized that, although they are suitable for expressing contrast, or rather, the re
markableness of a contrast, the EIF does not always entail an overt contrast, much in
the same way as the FRE need not be always Given. Therefore, these should be best
handled as tendencies rather than as inviolable principles, which in any case should
always be balanced against the background of a context.
Lastly, we have shown that ICs are useful tools to express the speaker’s/writer’s
view of, or involvement towards, the content presented in the proposition. This inter
personal or subjective orientation motivates some of the differences between ICs and
their non-cleft counterparts, thus helping to elucidate the conditions of usage of the
former in general as well as their relatively high occurrence in informal or spontaneous
opinion-giving spoken discourse in particular.
Notes
* The authors wish to acknowledge financial support from two Research Projects: (i) Discourse
Analysis in English: Aspects of cognition, typology and L2 acquisition, awarded to the SCIMITAR
research group (http://ietsil.usc.es/scimitar) and sponsored by the Spanish Ministry of Edu
cation, the FEDER funds and the Xunta de Galicia (XUGA) (grant numbers BFF2002-02441,
PGIDIT03PXIC20403PN); and (ii) Sintaxis Contrastiva Inglés-Español, also financed by the
Spanish Ministry of Education and the FEDER funds (grant number MCYT BFF2000-1271).
We would also like to express our gratitude to Chris Butler, Peter Collins, Knud Lambrecht,
Lachlan Mackenzie and Francis Cornish for their perspicuous comments on and illuminating
discussions of previous versions of this paper. Needless to say, any remaining errors are our sole
responsibility.
. The term cleft (construction) was introduced by Jespersen (1965 [1927]: 251) to designate
those structures with the peculiarity of being split or divided into two separate parts, while
the label it-cleft became widespread in the 70’s to refer to that subtype (Hankamer 1974;
Prince 1978).
. Block capitals mark the locus of intonational Focus. Due to the lack of suprasegmental in
formation in our corpora, prosodic information will be neither reflected nor discussed in the
examples used in this paper. For further discussion of the prosodic dimension of clefts in inter
action with other formal and functional factors, the reader is referred to Lambrecht (2001) and
the references therein.
. This issue has been a continuous preoccupation among generativist linguists. Thus, Akmajian
(1970), for example, proposes that ICs be derived from wh-clefts (adducing such criteria as
reflexivization or person and number agreement), while Higgins (1979) argues for their base-
generation and Hankamer (1974) admits both possibilities. Conversely, Gundel (1977) analy
ses ICs as reduced forms of right dislocated pseudo-clefts, whereas Wirth (1978) claims they
are derived from extraposed relative clauses in copular structures. In the Spanish literature,
Moreno-Cabrera (1999) also invokes transformations of expansion and reduction as heuristics
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
to determine cleft constructions, while Fernández-Leborans (2001) discards transformations
altogether.
. Although these scholars do agree on the existence of cleft constructions in Spanish, there
exist, however, some terminological differences among them. Thus, Hernanz & Brucart (1987)
refer to these configurations as oraciones escindidas, the closest Spanish equivalent to the English
term cleft, while D’Introno (1979), Di Tulio (1990), and Fernández-Domínguez (2001) opt for
the term oraciones hendidas.
. Configurations like these are referred to in the literature as estructuras ecuacionales (‘equa
tional structures’) (Alarcos-Llorach 1980; Martínez 1994; Gutiérrez-Ordóñez 1986), as pointed
out by Moreno-Cabrera (1999: 4248).
. Very close to this pattern is Moreno-Cabrera’s (1999: 4250) PdR (perífrasis de relativo ‘relative
periphrastic form’) of the type COP-CES, that is, configurations in which the form of the linking
verb occurs in the first place (COP) followed by the EIF (CES). However, the closest counterpart
to English ICs is a special type of periphrastic forms commonly attested in some dialects of South
American Spanish. Moreno-Cabrera (1999: 4281) calls them perífrasis conjuntivas (‘conjunctive
periphrastic forms’), in which Y in the general pattern is realized by a sentence or a verb phrase
introduced by the conjunction que (‘that’), as illustrated in (6) below (from Moreno-Cabrera
1999: 4281; quoting Bolinger’s 1972: 96, 102, 103, 109, 111 examples):
(6) a. It was in the garden that I found John.

lit. ‘Fué en el jardín que yo encontré a John.’
b. It was yesterday that he came.
lit. ‘Fué ayer que él llegó.’
c. It’s on a desert that I live.
lit. ‘Es en un desierto que yo vivo.’
d. It was money that he stole.
lit. ‘Fue dinero que él robó.’
. There remain substantial differences among linguists as to the real locus of the definite feature
of ICs. Two cases in point are Hedberg (2000: 906) and Davidse (2000: 1124): for the former, the
definite description contained in ICs is effected by the semantic combination of it and Y, the
dependent clause; whereas for Davidse the locus of such definite meaning stems from the matrix
clause (it + be + EIF).
. Note that with negatives (10), modal auxiliaries ((9d), (9e), (10c)), and certain adjuncts (e.g.
It’s very often the able people coming who have to let go of certain questions, ICE-GB: S1A-001
#78:1:B), speakers have the choice of either putting such elements in the head clause – as in the
examples above – or locating them in the subordinate clause, thus yielding alternatives such as
the following: So, it’s the people that aren’t covered (10a’), It’s Neil Webb who will go across it (9e’),
It’s the able people coming who very often have to let go of certain questions.
. Delahunty (1984: 198–194) also mentions the possibility of particles and quantifier phrases
occurring as EIF (e.g. It wasn’t on that he pulled his boots, it was off ). In much the same vein,
Collins (1991: 42) refers to the possessive phrase as a further realization of the EIF (It’s Collin
Haddock’s signature I have, ICE-GB: S1B-061 #75:1:B). However, in this study the possessive
type is considered as a subtype of NP EIF.
. ICs with a SC NP/AdjP highlighted element are dialectally marked (e.g. Celtic-based vari
eties of English, as in Jespersen’s 1965 [1927] examples It is angry that he was or It’s an angel
you are to forgive me). Collins (1991: 64) notes that this restriction seems to be parallel to that
On clefting in English and Spanish 
needed to account for the unacceptability of non-clefts with a thematized predicative comple
ment (? The chairman he was, ? It was the chairman that he was, ? Poor he was, ? It was poor that
he was).
. Collins – personal communication – now disallows genuine IO EIF on the grounds that,
using the same example, we cannot say *It was Terence she gave a present (cf. Huddleston 1984).
. Chris Butler’s intuitions – personal communication – differ from Borkin’s in two cases: a.
is “more or less grammatical” for him, especially if followed by a contrastive not . . ., while d. is
perfectly acceptable, though perhaps more usual with the fact that (see Note 13).
. In this respect, Borkin (1984: 140), following a personal communication by Ellen Prince,
accepts that factivity may also be at work in determining varying degrees of acceptability in cleft
constructions, as illustrated in (20) below:
(20) a. ??It’s that he shot her that he remembered.
b. It is that he shot her that seemed to be the most likely possibility.
In much the same vein, for some native speakers, sentence (18d) above improves in acceptability
if that is expanded as the fact that.
. There is a subtype of cleft, the so-called inferential sentence or sentential Focus cleft, which
lacks constituent 4 in the general pattern and has a sentential EIF, such as It is not only that
our criminal laws are being broken more profligately than ever (ICE-GB: S2A-039 #55-58:1:A).
Reasons of space preclude independent treatment of this subtype, for which the reader is referred
to Delahunty (2001) and the references therein.
. While on balance it seems more plausible to analyse the EIF as the antecedent for the wh
word of the RE, there is some evidence that points to it as fulfilling such a function, to note: (1)
proper nouns may be the EIF of ICs, but they cannot serve as antecedent to a relative clause,
unless accompanied by a determiner; and (2) very often, especially in informal usage, the verb
in the relative clause agrees with it, rather than with the EIF (which is likely to be accusative in
form even if the FRE functions as Subject) (It’s you/me who’s in trouble, not them).
. For a summary of the main differences between cleft constructions and relative clauses,
the reader is referred to e.g. Ball (1994), Gómez-González (2001: 305–306), Huddleston (1984:
460ff.) and Sornicola (1988: 346).
. In our data, a configuration has also been recorded in which both the relative element and
the verb form of the FRE are elided, since they are retrievable from the context:
(22) a. It’s only going to be the Donaldson part of the library open probably.
(ICE-GB: S1A-069 #213:2:A)
b. But it’s Great Britain still in possession. (ICE-GB: S2A-004 #379:1:A)
. A possible explanation for this lack of agreement could be that the presupposition conveyed
in the FRE generally has third person reference, because the information it conveys is presented
as being about something or someone that is far from the speaker/writer’s perspective and that
therefore needs to be identified. As already noted, this is the function fulfilled by the referent
of the EIF, viz. to precisely identify the actual referent of the variable entailed in the relative
presupposition, which may, but need not, be a third person referent.
. Several other terms have been used to contrast these structures. Thus, in transformational
linguistics the distinction is between specificational and predicational structures (Akmajian 1970;
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Higgins 1979), whereas Kuno and Wongkhomthong (1981) speak about identificational and
characterizational constructions. Halliday (1967a, b, 2004), in turn, call clefts thematic equatives.
. An important observation is in order concerning the felicitous reversibility of (24a) above.
We believe this is due to the fact that the configuration in question accepts a cleft reading as
well as one typical of a copulative sentence followed by a restrictive relative clause (a situa
tion by no means specific to Spanish, but also attested in English, e.g. It is the dog that scared
me (Greenbaum & Quirk 1990: 412). Thus, in concert with Moreno-Cabrera (1999: 429ff.), the
configuration in question could be interpreted as follows:
(25) a. “Mi hijo se me viene a la yema de los dedos” (Cleft interpretation, PdR interpretation
in Moreno-Cabrera’s terminology)
b. “La persona que se me viene a la yema de los dedos es mi hijo” (Copulative sentence
with a relative clause, CoRL interpretation in Moreno-Cabrera’s terminology)
The important point to note is that while (25a) is two-way ambiguous, (25b) is no longer likely
to be interpreted as a cleft construction, but only as a copulative sentence with a relative clause.
According to Moreno-Cabrera, the ambiguity in question is favoured by the copying of the
temporal reference in both the main and the “subordinate” clause. If the identity in tempo
ral reference condition (or at least the occurrence of simple present temporal reference in the
main clause) is not met, the cleft reading will not be felicitous, as in (26) below:
(26) a. Ha sido mi hijo el que se me viene a la yema de los dedos.
‘It has been my son that pops up in my head.’
b. Era mi hijo el que se me viene a la yema de los dedos.
References
Akmajian, Adrian (1970). “On deriving cleft sentences from pseudo-cleft sentences”. Linguistic
Inquiry, 1, 149–168.
Alarcos-Llorach, Emilio (1980). Estudios de Gramática Funcional del Español. Madrid: Gredos.
Alonso-Raya, Rosario & Alejandro Castañeda-Castro (1997). “De la focalización con fórmulas
perifrásticas de relativo”. In J. A. De Molina Redondo & J. de Dios Luque Durán (Eds.),
Estudios de Lingüística General (II) (Trabajos presentados en el II Congreso Nacional de
Lingüística General, Granada 25–27 Marzo 1996) (pp. 1–9). Granada: Método Ediciones.
Archivo de Textos Hispánicos de la Universidad de Santiago (ARTHUS). http://www.bds.usc.es
Atlas, Jay D. & Stephen C. Levinson (1981). “It-clefts, informativeness and logical form”. In P.
Cole (Ed.), Radical Pragmatics (pp. 1–61). New York: Academic Press.
Ball, Catherine N. (1994). “Relative pronouns in it-clefts: the last seven centuries”. Language
Variation and Language Change, 6(2), 179–200.
Barcelona-Sánchez, Antonio (1983). “El Orden de los Constituyentes en Inglés y Español.”
Unpublished Ph.D. thesis. University of Granada.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, & Edward Finegan (1999). The
Longman Grammar of Spoken and Written English. London: Longman.
Bolinger, Dwight (1972). “A look at equations and cleft sentences”. In E. Scherabon Firchow et
al. (Eds.), Studies for Einar Haugen Presented by his Friends and Colleagues (pp. 96–114).
The Hague: Mouton.
Bolinger, Dwight (1977). Meaning and Form. London: Longman.
On clefting in English and Spanish 
Borkin, Ann (1984). Problems in Form and Function. Norwood, NJ: Ablex.
Castañeda-Castro, Alejandro (1997). Aspectos cognitivos en el aprendizaje de una lengua
extranjera. Granada: Método Ediciones.
Clark, Herb H. & Susan E. Haviland (1977). “Comprehension and the Given-New contract”.
In R. Freedle (Ed.), Discourse Production and Comprehension (pp. 1–40). Hillside, NJ:
Erlbaum.
Collins, Peter C. (1985). “Th-clefts and all-clefts”. In J. E. Clark (Ed.), The Cultivated Australian:
Festschrift in Honour of Arthur Delbridge (pp. 45–53). Hamburg: Buske.
Collins, Peter C. (1991). Cleft and Pseudocleft Constructions in English. London: Routledge.
Davidse, Kristin (2000). “A constructional approach to clefts”. Linguistics, 38(6), 1101–1131.
Declerk, Renaat (1983). “Predicational clefts”. Lingua, 61, 9–45.
Declerk, Renaat (1984). “The pragmatics of it-clefts and wh-clefts”. Lingua, 64, 251–289.
Declerk, Renaat (1994). “The taxonomy and interpretation of clefts and pseudo-clefts. Review
article on P. C. Collins (1991) Cleft and Pseudo-cleft Constructions in English”. Lingua, 93,
183–220.
Declerck, Renaat & Shigeki Seki (1990). “Premodified it-clefts”. Lingua, 82, 15–51.
Delahunty, Gerald P. (1984). “The analysis of English cleft sentences”. Linguistic Analysis, 13,
63–113.
Delahunty, Gerald P. (2001). “Discourse functions of inferential sentences”. Linguistics, 39(3),
517–545.
Delin, Judy (1990). Accounting for Cleft-Constructions in Discourse. Edinburgh: University of
Edinburgh Publications.
Delin, Judy & John Oberlander (1995). “Syntactic constraints on discourse structure: The case
of it-clefts”. Linguistics, 33, 465–500.
Dik, Simon C. (1997). The Theory of Functional Grammar. Parts I and II [edited by Kees
Hengeveld]. Berlin and New York: Mouton de Gruyter.
D’Introno, Francesco (1979). Sintaxis Transformacional del Español. Madrid: Cátedra.
Di Tulio, Ángela (1990). “Sobre hendidas y pseudohendidas”. Revista de Lengua y Literatura, 7,
3–16.
Downing, Angela (1991). “An alternative approach to Theme: a Systemic-Functional perspec
tive”. Word, 42(2), 119–143.
Downing, Angela (1997). “Encapsulating discourse topics”. Estudios ingleses de la Universidad
Complutense, 5, 147–168.
Drubig, H. Bernhard (2003). “Toward a typology of focus and focus constructions”. Linguistics,
41(1), 1–50.
Fawcett, Robin & Guowen Huang (1997). Enhanced Theme in English: Towards a Functional
Explanation of the ‘IC’ Construction. London: Cassell.
Fernández-Domínguez, Joaquín José (2001). “Aspectos de la Sintaxis de las Oraciones Hendidas
en Inglés Actual.” Unpublished MA Thesis, University of Seville.
Fernández-Leborans, María Jesús (2001). “Sobre formas de ambigüedad de las oraciones
escindidas: sintaxis y discurso”. Estudios de Lingüística de la Universidad de Alicante, 15,
285–305.
Fichtner, Edward G. (1993). “Cleft sentences in English: a comprehensive view”. Word, 44(1),
1–30.
Givón, Talmy (1990). Syntax. A Functional-Typological Introduction, Vol. 2. Amsterdam and
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Gómez-González, María de los Ángeles (2002). “First mention orientation in discourse”. In L.

Rábade, S. M. Doval, M. L. A. Gómez et al. (Eds.), Studies in Contrastive Linguistics (pp.
37–48). Santiago de Compostela: Servicio de Publicaciones de la Universidad de Santiago
de Compostela.
Gómez-González, María de los Ángeles (2004a). “Functional Grammar and the dynamics of
discourse”. In J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for
Functional Grammar (pp. 211–242). Berlin and New York: Mouton de Gruyter.
Gómez-González, María de los Ángeles (2004b). “A multi-factor approach to it-clefts in
discourse”. South West Journal of Linguistics, 23(2), 1–40.
Green, Georgia M. (1971). “Notes on clefts and pseudo-clefts and related matters”. Studies in
Linguistic Sciences, 1–7.
Greenbaum, Sidney & Randolph Quirk (1990). A Student’s Grammar of the English Language.
London: Longman.
Grice, H. Paul (1975). “Logic and conversation”. In P. Cole & J. L. Morgan (Eds.), Studies in
Syntax 3: Speech Acts (pp. 41–58). New York: Seminar Press.
Gundel, Jeannette K. (1977). “Where do cleft sentences come from?” Language, 53, 543–559.
Gundel, Jeannette K. (1999). “On three kinds of focus”. In P. Bosch & R. van der Sandt (Eds.),
Focus. Linguistic, Cognitive and Computational Perspectives (pp. 293–305). Cambridge:
Gundel, Jeannette K. (2002). “Information structure and the use of cleft sentences in English
and Norwegian”. In H. Hasselgård, S. Johansson, B. Behrens, & C. Fabricius-Hansen (Eds.),
Information Structure in a Cross-linguistic Perspective (pp. 113–128). Amsterdam and New
York: Rodopi.
Gutiérrez-Ordóñez, Salvador (1986). “Estructuras ecuativas y ecuacionales”. In S. Gutiérrez-
Ordóñez (Ed.), Variaciones sobre la Atribución (pp. 45–85). León: Universidad de León.
Halliday, Michael A. K. (1967a). “Notes on transitivity and Theme in English”. Journal of
Linguistics, 3(1), 37–81.
Halliday, Michael A. K. (1967b). “Notes on transitivity and Theme in English”. Journal of
Linguistics, 3(2), 199–244.
Halliday, Michael A. K. (2004). An Introduction to Functional Grammar [3rd ed. revised by
Christian M. I. M. Matthiessen]. London: Arnold.
Halliday, Michael A. K. & Ruqaiya Hasan (1976). Cohesion in English. London: Longman.
Halvorsen, Per-Kristian (1978). “The syntax and semantics of cleft constructions”. Texas
Linguistic Forum, 11, 1–99.
Hankamer, Jorge (1974). “On the non-cyclic nature of Wh-clefting”. In Papers from the Tenth
Regional Meeting of the Chicago Linguistic Society (pp. 221–232). University of Chicago:
Chicago Linguistic Society.
Harries-Delisle, Helga (1978). “Contrastive emphasis and cleft sentences”. In J. H. Greenberg
(Ed.), Universals of Human Language. Syntax, Vol. 4 (pp. 420–486). Stanford: University
Press.
Haugland, Kari E. (1992). “On the use of cleft and pseudo-cleft sentences in English”. ICAME
Journal, 16, 51–69.
Hedberg, Nancy (2000). “The referential status of clefts”. Language, 76(4), 891–920.
Hernanz, Ma Lluïsa & José María Brucart (1987). La Sintaxis. 1. Principios Teóricos. La Oración
Simple. Barcelona: Editorial Crítica.
Higgins, Francis R. (1979). The Pseudo-Cleft Construction in English. New York: Garland.
On clefting in English and Spanish 
Hornby, Peter A. (1971). “The role of Topic-Comment in the recall of cleft and pseudo-cleft
sentences”. In Papers from the Seventh Regional Meeting of the Chicago Linguistic Society (pp.
445–453). University of Chicago, Chicago Linguistic Society.
Huddleston, Rodney (1984). Introduction to the Grammar of English. Cambridge: Cambridge
University Press.
Huddleston, Rodney D. & Geoffrey K. Pullum (2002). The Cambridge Grammar of the English
Language. Cambridge: Cambridge University Press.
International Corpus of English (ICE). http://www.ucl.ac.uk/english-usage/ice
Jespersen, Otto (1965 [1927]). A Modern English Grammar on Historical Principles, Part III.
London: Allen and Unwin.
Jespersen, Otto (1969 [1937]). Analytic Syntax. Transatlantic Series in Linguistics. New York:
Holt, Rinehart and Winston.
Karttunen, Lauri (1974). “Presupposition and linguistic context”. Theoretical Linguistics, 1, 181–
194.
Kuno, Susumu & Preeya Wongkhomthong (1981). “Characterizational and identificational
sentences in Thai”. Studies in Language, 5(1), 65–109.
Lambrecht, Knud (1988). “Presentational cleft constructions in Spoken French”. In T. Haiman
& S. Thompson (Eds.), Clause Combining in Grammar and Discourse (pp. 135–180).
Lambrecht, Knud (1994). Information Structure and Sentence Form (Topic, Focus and the Mental
Representations of Discourse Referents). Cambridge: Cambridge University Press.
Lambrecht, Knud (2001). “A framework for the analysis of cleft constructions”. Linguistics, 39(3),
463–516.
Langacker, Ronald W. (2001a). “Dynamicity in grammar”. Axiomathes, 12, 7–33.
Langacker, Ronald W. (2001b). “Discourse in cognitive grammar”. Cognitive Linguistics, 12, 143–
188.
Mackenzie, J. Lachlan (2004). “Functional Discourse Grammar and language production”. In
J. L. Mackenzie & M. L. A. Gómez-González (Eds.), A New Architecture for Functional
Benjamins.
Martínez, José Antonio (1994). Cuestiones Marginadas de Gramática Española. Madrid: Istmo.
Martínez-Caro, Elena (1999). Gramática del Discurso: Foco y Énfasis en Inglés y Español. Barce
lona: PPU.
Martínez-Vázquez, Rafael (1997). “Metáfora oracional: incidencia en la estructura pragmática
de la oración”. In M. Martínez-Vázquez (Ed.), Gramática y Pragmática (pp. 77–111).
Huelva: Grupo de Gramática Contrastiva.
Miller, Jim & Regina Weinert (1998). Spontaneous Spoken Discourse. Oxford: Blackwell.
Moreno-Cabrera, Juan Carlos (1987). Fundamentos de Sintaxis General. Madrid: Síntesis.
Moreno-Cabrera, Juan Carlos (1999). “Las funciones informativas: las perífrasis de relativo y
otras construcciones perifrásticas”. In I. Bosque & V. Demonte (Eds.), Gramática Descriptiva
de la Lengua Española, Vol. 3: Entre la oración y el discurso – Morfología (pp. 4245–4302).
Madrid: Espasa Calpe.
Nelson, Gerald (1997). “Cleft constructions in spoken and written English”. Journal of English
Linguistics, 25(4), 340–348.
 María de los Ángeles Gómez-González and Francisco Gonzálvez-García
Perzanowski, Dennos & John Gurney (1997). “The functionality of it-clefts in selected
discourses: the message in the medium”. Word, 48(2), 207–236.
Prince, Ellen F. (1978). “A comparison of wh-clefts and it-clefts in discourse”. Language, 54,
883–906.
Quirk, Randolph et al. (1973). A University Grammar of English. London: Longman.
Quirk, Randolph et al. (1985). A Comprehensive Grammar of English. London: Longman.
Real Academia Española: Banco de datos (CREA) [en línea]. Corpus de referencia del español
actual. <http://www.rae.es> [Accessed August 2003].
Sinclair, John (Ed.). (1990). Collins Cobuild English Grammar. London: William Collins Sons
and Co.
Sornicola, Rosanna (1988). “It-clefts and wh-clefts: two awkward sentence types”. Journal of
The Survey of English Usage (ICE). http://www.ucl.ac.uk/english-usage/
Vallduví, Enric & Elisabet Engdahl (1996). “The linguistic realization of information packag
ing”. Linguistics, 34, 459–519.
Ward, Gregory, Betty Birner, & Rodney Huddleston (2002). “Information packaging”. In R.
Huddleston & G. K. Pullum (Eds.), The Cambridge Grammar of the English Language (pp.
1363–1448). Cambridge: Cambridge University Press.
Weinert, Regina & Jim Miller (1996). “Cleft constructions in spoken language”. Journal of Prag
matics, 25, 173–206.
Wirth, Jessica (1978). “The derivation of cleft sentences in English”. Glossa, 12, 58–82.
Anaphoric terms and focus of attention
in English and Spanish*
Maite Taboada
The choice of one anaphoric term (pronoun, clitic, noun phrase) over another is
related to what is in the focus of attention at each point in the discourse (the
entities and relations mentioned by the participants in a conversation). In this
paper, I explore the relation between choice of anaphoric expression and focus of
attention by applying a theory of local focus in discourse, Centering Theory
(Grosz, Joshi, & Weinstein 1995). I examined spoken language corpora in English
and Spanish, and determined the relationship between the focus of attention and
the type of anaphoric term used for the topic of each utterance. Results show that
focus is not the only factor involved; in conversation, other factors, such as
turn-taking, seem to have a role in the type of anaphoric term chosen.
A basic question in discourse analysis, and in the study of anaphoric terms is: what
is the anaphoric term chosen at any particular point in the discourse? The question
encompasses a number of other issues that relate the choice of a particular anaphoric
term (referring expression, discourse anaphora) to the context of situation and the
preceding text. A speaker chooses a pronoun, a proper name, a definite noun phrase,
an indefinite noun phrase, or a number of other possible types of noun phrase, de
pending on different factors: whether the referent for the anaphoric term has been
mentioned before in the discourse; whether it was mentioned as a Subject or an Ob
ject; whether there are competing referents in the discourse; or what is most salient at
that particular point in the discourse. The list could be much longer. The difficulty lies
precisely in determining what are the factors that affect the choice of anaphoric term.
In this paper, I apply a theory of focus of attention in discourse, Centering Theory
(Grosz, Joshi, & Weinstein 1995), to the problem of determining what anaphoric term
will be chosen for a particular referent. A number of other explanations have been pro
posed, based on a hierarchy of givenness of referents (Gundel, Hedberg, & Zacharski
1993), similar to a scale of accessibility of the referents encoded by different referring
expressions (Ariel 1990), or based on the syntactic function of the referring expression
(Gordon et al. 1999). Givón (1983) also correlates the form of the referring expression
with topicality (what is in the focus of attention) in his topic hierarchy. For instance,
 Maite Taboada
most accessible topics are encoded as zero anaphora, whereas least accessible topics
are represented as indefinite noun phrases. In general, the encoding and presentation
(or ‘packaging’) of different types of discourse referents is the object of a number of
studies (Chafe 1976; Daneš 1974; Kuno 1972; Lambrecht 1994; Prince 1981; Vallduví
1990; Vallduví & Engdahl 1996).
Centering Theory builds a hierarchy of entities in the discourse, for each discourse
segment. The entities are ranked according to different criteria, grammatical function
being the preferred (for English at least). For each discourse segment, one entity is
considered to be the topic of the discourse segment. This paper explores the choice
of anaphoric term for that particular topic, which is always an entity that has been
mentioned earlier in the discourse.
This study considers choices in two languages, English and Spanish. English has
been widely studied in general theories of anaphora, and within Centering in particu
lar (Brennan 1995; Brennan, Friedman, & Pollard 1987; Cornish 1999; Gordon, Grosz,
& Gilliom 1993; Hudson-D’Zmura & Tanenhaus 1998; Hurewitz 1998). A number of
other languages have been studied using Centering as a framework: Italian (Di Eugenio
1998), Japanese (Iida 1998; Kameyama 1985; Walker, Iida, & Cote 1994), Turkish
(Turan 1995), Greek (Miltsakaki 2001), German (Strube & Hahn 1996), Hindi (Prasad
& Strube 2000). Spanish, however, has not received enough attention. The comparison
between English and Spanish is interesting, since Spanish is different in a number of
parameters (presence of pro-drop and clitics, more flexible word order). I examined
spoken language corpora, which have often been neglected in discourse-based studies.
Spoken language will provide more interesting insights for phenomena relating to the
focus of attention, since attention issues are much more pressing in spoken language.
The corpus chosen for this study is the CallHome corpus, for English and Span
ish. CallHome was an effort by the Linguistic Data Consortium to collect spontaneous
telephone conversations. Participants were given thirty minutes of long-distance call
ing time, to call relatives or friends, provided they agreed to being recorded. For this
particular study, I chose five conversations in English and five in Spanish.
The paper is organized as follows: in Section 2, I provide a brief overview of
Centering Theory, the framework of the study. Since I am applying Centering to two
relatively new areas, I discuss the process of applying Centering to spoken language in
Section 3, and extending Centering to analyze Spanish in Section 4. Section 5 presents
the main results of the study: the relationship between transition type and anaphoric
term used for the topic of the sentence.1 The findings are presented in terms of ten
dencies: what anaphoric term tends to be used, given a particular transition. Speakers
often choose a term that does not follow those general tendencies. I discuss some of
those cases also in Section 5, which is followed by conclusions in Section 6.
Anaphoric terms and focus of attention 
. Centering Theory
Centering (Grosz, Joshi, & Weinstein 1995; Walker, Joshi, & Prince 1998) was devel
oped within a theory of discourse structure (Grosz & Sidner 1986) that considers the
interaction between (i) the intentions, or purposes, of the discourse participants, (ii)
the attention of the participants and (iii) the structure of the discourse. Centering
is concerned with the participants’ attention and how the global and local struc
tures of the discourse affect the referring expressions and the overall coherence of
the discourse. It models the structure of local foci in discourse, i.e., foci within a
discourse segment.
Centers are semantic entities that are part of the discourse model of each utter
ance in the segment. For each utterance, Centering establishes a ranked list of entities
mentioned or evoked, the forward-looking center list (Cf). The list is ranked according
to salience, defined most often in terms of grammatical relations (see Section 3). The
first member in the Cf list is the preferred center (Cp). Additionally, one of the mem
bers of the Cf list is a backward-looking center (Cb), the highest-ranked entity from the
previous utterance that is realized in the current utterance.
Example (1) below illustrates these concepts. Let us assume that the three ut
terances in the example constitute a discourse segment. In the first utterance of the
segment, (1a), there are three centers: Bilbo, Frodo and a ring. In (1b), the centers
are still the same, now all realized as pronouns (he=Bilbo, him=Frodo, it=ring). The
list of forward-looking centers for (1a) includes all three centers, ranked according to
grammatical function: Subject > Indirect Object > Direct Object.2 The Cf list for (1a)
is: Bilbo, Frodo, ring.3 The preferred center in that utterance is the highest-ranked
member of the Cf list, i.e., Bilbo. Similarly, in (1b), the Cf list contains, in this order:
Bilbo, Frodo, ring. The Cp is still Bilbo. Notice that (1a) does not have a backward-
looking center (the center is empty), because this is the first utterance in the segment.
The backward-looking center of (1b) is Bilbo, the highest-ranked entity from (1a) that
is also realized in (1b). In (2), we see the Cf, Cp and Cb for each of the three utterances
in the segment.
(1) a. Bilbo gave Frodo a ring.
b. He told him to look after it very well.
c. Frodo put the ring in his pocket.
(2) a. Cf: Bilbo, Frodo, ring – Cp: Bilbo – Cb: Ø
b. Cf: Bilbo, Frodo, ring – Cp: Bilbo – Cb: Bilbo
c. Cf: Frodo, ring, pocket – Cp: Frodo – Cb: Frodo
In addition to the three types of centers, Centering proposes different types of tran
sitions, based on the relationship between the backward-looking centers of any given
pair of utterances, and the relationship of the Cb and Cp of each utterance in the pair.
Transitions, shown in Table 1, capture the introduction and continuation of new top
ics. Cb(Ui ) and Cp(Ui ) refer to the centers in the current utterance. Cb(Ui–1 ) refers to
 Maite Taboada
Table 1. Transition types
Cb(Ui ) = Cb(Ui–1 ) or Cb(Ui–1 ) = ∅ Cb(Ui–1 )

Cb(Ui ) =
Cb(Ui ) = Cp(Ui ) continue smooth shift

Cp(Ui )
Cb(Ui ) = retain rough shift
the backward-looking center of the previous utterance. Thus, a continue occurs when
the Cb and Cp of the current utterance are the same and the Cb of the current utter
ance is the same as the Cb of the previous utterance. Transitions capture the different
types of ways in which discourse can progress: to give just two examples at opposite
ends of the spectrum, an utterance may refer to a previous topic, the Cb(Ui–1 ), and still
be concerned with that topic, the Cp(Ui ) in a continue; or it may be not linked to the
previous topic, in a rough shift. Transitions are one explanation4 for how coherence
is achieved: a text that maintains the same centers is perceived as more coherent.
In example (1), the first utterance has no Cb, because it is segment-initial, and
therefore it has no transition (also called a zero-Cb transition). The transition between
(1a) and (1b) is a continue, because the Cb of (1a) is empty, and the Cp and Cb of
(1b) are the same, Bilbo. Finally, the transition between (1b) and (1c) is a smooth
shift, since the Cb has changed. This signals a shift, the change to a new Cb (Frodo),
possibly the referent that is going to be continued, since it is also the Cp. Because
transitions capture topic shifts in the conversation, they are ranked according to the
demands they pose on the reader. The ranking is: continue > retain > smooth shift
> rough shift. This transition ranking is often referred to as Rule 2 in the Centering
paradigm. Centering predicts that continue will be preferred to retain, and retain
to shifts, all other things being equal. The preference applies both to single transitions
and to sequences of transitions.
Rule 1 captures the preference for pronouns when the same topic of discourse is
continued. The formulation of Rule 1 is as follows:
For each Ui in a discourse segment D consisting of utterances U1 , . . . , Um ,
if some element of Cf(Ui–1 , D) is realized as a pronoun in Ui , then so is
Cb(Ui , D).
Rule 1 is sometimes referred to as the Pronoun Rule. It captures the fact that a topic
that is continued from a previous utterance does not need to be signalled by more ex
plicit means than a pronoun (or a zero pronoun, in languages that allow those). Other
pronouns are of course allowed in the same utterance, but the most salient entity must
be realized by the least marked referring expression. In (1b), the backward-looking
center, Bilbo, is realized as a pronoun, following Rule 1, since a pronoun is also used
to refer to Frodo, which is not the backward-looking center.
There is a relationship between the type of transition that occurs between any two
utterances, and the type of anaphoric term chosen to realize entities in the focus of
attention. Centering predicts that entities higher up in the focus (the Cf list) will be
realized through zero pronouns, full pronouns or clitics, depending on the options
Anaphoric terms and focus of attention 
available in the language. Continue transitions, because they keep the same center,
will likely encode the Cb or the Subject of the sentence as a zero pronoun or a full
pronoun (Di Eugenio 1998). Shifts (smooth or rough) will result in less pronominal
ization. We will see that these relationships are quite complex, and different factors
come into play in the choice of anaphoric term.
. Applying Centering to spoken language
Centering has been, for the most part, applied to written language. A number of issues
arise as we extend it to cover spoken language. These have to do with the interpreta
tion of discourse segment or utterance; with spoken language phenomena such as false
starts, repetition, overlapping and backchannel signals; and with the treatment of first
and second person pronouns, which are not usually considered as part of the study of
discourse anaphora, but which are prevalent in spoken conversation.
Segmentation of spoken discourse is often fraught with difficulties. Speech does
not happen in clear, period-separated sentences, but in smaller intonation units, some
of which are not grammatically complete. Schiffrin (1994) argues for abandoning the
sentence as a unit of analysis, and considering ‘utterance’ as the basic unit of orga
nization in spoken language. However, the definition of utterance has not been fully
spelled out. Even within Centering Theory, regardless of whether the language is writ
ten or spoken, some argue for a main clause with all its embedded clauses (Miltsakaki
2002);5 some argue for the finite clause as the unit of analysis (Kameyama 1998); and
some consider that most clauses, finite or not, should be units, or utterances, in the
Centering sense (Poesio 2000).
We are in the process of evaluating different segmentation methods, and have
already proposed a method (Hadic Zabala & Taboada 2004). Our current approach
follows Kameyama (1998). It considers the tensed clause as the unit of analysis (the
utterance, in Centering terms): conjoined clauses are separated as two (or more)
units; main clauses are separated from any embedded clauses, provided the embed
ded clauses are finite; non-finite clauses form a unit with their matrix clause. In (3),
two utterances are present, since (3b) is a tensed embedded clause. There is a third
clause, irme a ver a mi hermana (‘to go see my sister’), which is not an utterance, since
the verb is non-finite.
(3) a. No compro nada, no nada, nada
‘(I) don’t buy anything’
b. porque quiero irme a ver a mi hermana
‘because (I) want to go see my sister.’
The next issue in spoken language is the presence of false starts, repetitions, backchan
nels and overlapping. Since Centering processes sentences in a linear order, these are
important aspects to consider. False starts are utterances that are abandoned in favour
of a different wording, or that may not be completed at all. Since some false starts can
 Maite Taboada
be units in themselves, we decided to consider as units only those that had entities in
them, i.e., nominal expressions with reference (Eckert & Strube 1999). Centering is
concerned with how entities are managed in the focus of discourse. As a consequence,
we need to consider only those units that contain entities. Repetitions are treated simi
larly: only repeated units that contain entities are considered. Other repetitions are not
included in the Centering analysis.
Backchannels (or backchannel signals) are expressions by the hearer that indi
cate to the speaker that the channel is still open, and that the communication is being
successful (Yngve 1970). Backchannels are especially important in telephone conversa
tions, which are the types of conversations analyzed. As with false starts, backchannels
are discarded when they do not contain any entities. In example (4), speaker B pro
vides a backchannel (oh yeah) to what speaker A is saying. The only units considered
for the analysis are (4a) and (4c).
(4) A: a. and we got one tree left over by the garbage can
B: b. oh yeah
A: c. yeah {laugh} so you got to leave one you know
Overlapping is particularly difficult, given the linear constraints in Centering. As a gen
eral rule, we have maintained the linearity in time reflected in the CallHome corpus
time stamps. Each utterance in the corpus transcripts has a time stamp at the begin
ning, indicating how long into the conversation it was produced. We used those to
order utterances, and also listened to the recordings. In some cases, it is clear to the
researcher that speakers do not produce speech in a perfect linear fashion. Moreover,
it is clear that even speech produced linearly has intricate connections. In example
(5), Speaker A describes when he goes to buy dinner. He says, however, ‘lunch’, for
food that he buys at 7 p.m. Speaker B corrects him, and asks another question about
friends. Speaker A accepts the correction, and answers the question at the same time. If
we look at the time stamps (the time interval when the utterance started and finished),
we can see that Speaker B started to say la cena (‘dinner’) at 201.01. Speaker A started
to accept that correction at 201.72. But while Speaker A was saying Sí, la cena, claro,
Speaker B was already asking the next question. From a discourse anaphora point of
view, we would want to be able to relate the entities in utterances (5a), (5b), and (5d)
on the one hand; and (5c) and (5e), on the other. From a strictly linear point of view,
however, utterances (5c) to (5e) have no entities connecting them.
(5) 193.92–200.51 A: a. No, a las siete, a las siete, siete y pico, suelo ir con
un japonés y con un francés a comprar la comida allá
enfrente.
‘No, at seven, at seven, at seven something, (I) usually
go with a Japanese (man) and a French (man) to buy
lunch across there.’
Anaphoric terms and focus of attention 
201.01–204.43 B: b. La cena.
‘Dinner.’
c. ¿Qué tal los amiguetes?

‘How about friends?’
201.72–211.02 A: d. Sí, la cena, claro.
‘Yes, dinner, of course.’
e. Bien, hay bastante gente de España,

‘Good, there are quite a few people from Spain,’
Centering, like most theories of anaphora, is concerned with third person reference.
First and second person pronouns are deictic, referring to the speaker and hearer in
the present situation, and not to a series of entities brought into the conversation
by the speakers. Theoretically, then, first and second person pronouns should not be
part of a Centering analysis. Byron & Stent (1998) devised an experiment, also using
part of the English CallHome corpus, to test the parameters of Centering. They found
that a configuration of Centering that included first and second person pronouns per
formed better than an analysis using exclusively third person pronouns. The measures
in an optimal configuration of Centering include: the number of empty backward-
looking centers (Cb) in utterances;6 whether the Cb coincided with the sentence topic;
and whether the method in general produced more cheap or more expensive transi
tions. Cheap and expensive refer to the inference load placed on the hearer (Strube &
Hahn 1999), based on whether the preferred center of an utterance (CpU ), which is
the expected Cb of the next utterance, is actually realized as CbU+1 . We also decided to
include first and second person pronouns in our analysis.
In summary, the approach taken here to spoken discourse uses strict linearity to
relate utterances, disregards utterances that do not contain any entities (false starts,
backchannels), and includes first and second person pronouns in the Cf list, although
strictly speaking they are not anaphoric, but exophoric or deictic.
. Ranking entities in the focus of attention
Centering Theory, which derives from a theory of focus of attention and discourse
structure (Grosz & Sidner 1986), attempts to describe the management of entities and
topics in the discourse. It is concerned with which entities are part of the focus of the
discourse, and which of those entities is more salient. Using that knowledge, one could
predict facts about reference and pronominalization. For these purposes, Centering
has created a number of constructs.
Traditionally, the most salient entity in the discourse has been called the discourse
topic, with the sentence topic being roughly the counterpart at the sentence level.
Topic, however, is a widely used term, often with very different interpretations: what
the sentence is about (Gundel & Fretheim 2004); the presupposed part of the sentence
(Jackendoff 1972); the first ideational (‘contentful’) element in the sentence, referred
 Maite Taboada
to as topical Theme (Halliday 1967), or only those ideational elements that refer to en
tities (Downing 1991). For an excellent overview of the terms and their meaning, see
Gómez-González (2001).
Given the impreciseness of the term ‘topic’, the creators of Centering Theory de
cided to use the term ‘center’ for all entities in the discourse, and ‘backward-looking
center’ for the particular entity in the current utterance that was most salient in the
previous utterance (Grosz & Sidner 1998; Joshi & Weinstein 1998). In order to deter
mine which is the most salient entity, a ranked list of entities needs to be created, the
list of forward-looking centers (Cf list).
The criteria for ranking the Cf list have long been debated. Originally, in a formu
lation that considered only English, the proposed ranking was based on grammatical
relations: Subjects are more salient than Objects (Grosz, Joshi, & Weinstein 1995). As
other languages were studied, language-specific characteristics forced a revision of the
criteria. For Japanese, topic markers and empathy became part of the ranking (Walker,
Iida, & Cote 1994). In German, Strube & Hahn (1999) incorporated discourse sta
tus (Prince 1981). In Italian (Di Eugenio 1998) and Turkish (Turan 1995), empathy
was also considered important. The analysis described here uses the most common
template for English, represented in (6). Subjects are always ranked higher, except
for impersonal and arbitrary Subjects (described below). The category Other includes
mostly Adverbials.
(6) Subject > Object(s) > Other
The starting point for ordering the Cf list in Spanish is the same as for English: Subjects
are ranked higher than Objects. Since Spanish is a pro-drop language, null or zero
Subjects are included in the ranking, and also ranked higher than Objects. Exceptions
are arbitrary and impersonal Subjects, in sentences such as “People like it here” or
“One/you like(s) it here”, which are ranked lowest in the scale. Arbitrary pronouns
(Jaeggli 1986; Suñer 1983) refer to person(s) that need not be defined or filled by a
specific referent, or that are actually not defined in the context (“They’re knocking
on the door”). Suñer (1983) showed that the null pronoun in Spanish may have an
arbitrary interpretation (Dicen que va a nevar, ‘(They) say that it is going to snow’).
In Spanish, besides plural third person and singular second person, arbitrary pro
nouns are represented in impersonal sentences with the pronoun se (García 1975): Se
vive mejor en España (‘One lives better in Spain’). In example (7), the speaker tells the
interlocutor that he sounds well (he had been sick), using an impersonal form: one
can hear that you sound well. This can be considered an impersonal in middle voice,
according to Mendikoetxea (1999).7
(7) Ya se te oye muy bien.
already se cl:2sg hear:3sg.pres very well
‘You already sound very well.’ Cf: b (interlocutor, te), impersonal (se)
Two other criteria are taken into account for ranking: empathy and animacy. Follow
ing Di Eugenio (1998), we take empathy with the speaker or hearer over strict word
Anaphoric terms and focus of attention 
order as a ranking criterion. Empathy, as defined by Kuno (1987: 206), “is the speaker’s
identification, which may vary in degree, with a person/thing that participates in the
event or state that he describes in a sentence.”
Empathy in Spanish is reflected in the Experiencer in psychological verbs, where
the point of view taken is that of the Experiencer, regardless of whether it is the Subject
or not (e.g., ‘it seems to me’, ‘I think’, and similar verbs). In example (8), the clitic me
refers to the speaker, for whom Thursday is a better date. A number of verbs in Spanish
follow this pattern (me conviene ‘it’s good for me’, me viene mejor ‘it’s better for me’,
se me hace que ‘it seems to me’). With those verbs, the thematic role of Experiencer is
ranked higher than the grammatical function of Subject.
(8) me vien-e mejor el jueves,
cl:1sg come-3sg.pres better the Thursday
‘Thursday is better for me,’ Cf: I (me), it (the meeting, null), thursday
Animacy also affects the ranking: animate entities are ranked higher. Animacy interacts
with empathy, since the Experiencer in psychological verbs always carries animacy. An
imacy also interacts with linear order, since when both Indirect and Direct Object are
expressed through clitics, in Spanish the Indirect Object (typically animate) is placed
first. This hierarchy is already present in Givón’s (1983) studies on topic, which include
semantic roles (and their grammatical function correlates) and animacy in calculations
of topicality. As an example of animacy, we can see in (9b) that the indirect clitic se ‘to
her’ precedes the direct lo ‘it’, which refers to a scholarship for a program that was
given to the speaker’s sister. The null Subject is arbitrary (it has no definite referent),
and thus ranked last.
(9) a. Mi hermana solicitó un programa de arqueología y antropología en Gre
cia.
‘My sister applied to a program in archaeology and anthropology in
Greece.’
b. ¡Y que se lo dan!
and that cl:3sg:dat cl:3sg:masc:acc give:3pl:pres
‘And they gave it to her!’
Cf: sister (se, ‘to her’), program (lo, ‘it’), they (null)
A more detailed description of all the phenomena pertaining to ranking the Cf list can
be found in Hadic Zabala & Taboada (2004) and in Taboada (forthcoming). The order
for Spanish is summarized below (10).
(10) Experiencer > Subj > Animate IObj > DObj > Other > Impersonal/Arbitrary
pronouns
 Maite Taboada
Table 2. Number of words and utterances in each corpus
English Spanish
Words 11,457 8,694
Utterances 1,174 1,198
. Choice of anaphoric term for the most salient entity
The purpose of this paper is to investigate what referring expression is chosen most
often for the backward-looking center (Cb) in any given utterance, comparing English
and Spanish. The hypothesis is that the choice of referring expression will depend on
the type of transition holding between current and previous utterance. Some of the
research in Centering has established certain trends. For instance, Di Eugenio (1998)
showed that, typically, when the transition between U1 and U2 is a continue transi
tion, the Subject of U2 is a pronoun, or a null pronoun in languages that allow it. When
the transition is a retain, regular or stressed pronouns are used instead. In this section,
I show some of those tendencies for a corpus of Spanish and English conversations, and
also what factors are involved when the general trend is not followed.
The study was conducted on five conversations in English and five in Spanish. The
conversations were selected from the corresponding version of the CallHome corpus,
distributed through the Linguistic Data Consortium. The conversations were selected
to be all of similar length. They are telephone conversations between two speakers who
know each other. Occasionally, more than two speakers participate, when one of the
speakers passes the phone to another person in the same household.
Table 2 presents a summary of the number of words and utterances in each lan
guage. Utterance is defined in Section 3, and in general it refers to a finite clause. The
rest of this section presents the results obtained for each of the two languages, and how
they compare to each other.
. Referring expressions and transitions
The first step in the analysis is to identify what transitions hold between utterances.
For that purpose, Centering structures are created for each utterance: list of forward-
looking centers (Cf), backward-looking center (Cb), and preferred center, the first
entity in Cf (Cp). Then, we compare each pair of utterances according to Table 1.
This yields four different types of transitions: continue, retain, smooth shift, and
rough shift.
Some utterances do not have a backward-looking center. This may be due to a
number of reasons: they start the conversation; they start a new topic; or they contain
no entities that can be clearly related to the previous utterance. A total of 395 utter
ances had an empty Cb in English, and 408 in Spanish. These utterances cannot be
considered in the analysis, since Centering relies on the presence of a Cb. Some other
Anaphoric terms and focus of attention 
Table 3. Transition types per language
Continue Retain Smooth shift Rough shift
English 65.6% 17% 14.4% 3%

(n = 779)
Spanish 65% 16% 15% 4%
(n = 790)
Table 4. Referring expressions for the Cb of each utterance; percentages are with respect to
the transition type in each language
Continue Retain Smooth shift Rough shift
English Spanish English Spanish English Spanish English Spanish
Zero pr. 5% 49.5% – 26% 4.5% 49% – 16.7%

Pronoun 80% 9% 54% 6% 67% 6% 58% 7%
Clitic – 21% – 33% – 20% – 43%
NP 12% 16% 24% 18% 23% 19% 17% 27%
Dem. pr. 2% 2.5% 4% 2% 4.5% 3% 4% 3%
Other 2% 2% 18% 14% 1% 3% 21% 3%
n 511 515 132 129 112 116 24 30
utterances are discarded, because they contained no entities. After removing these two
subsets, we came to a total of 779 utterances in English, and 790 in Spanish with a
clearly defined transition. The transitions were classified into the four types, as shown
in Table 3.
The results are interesting in that they confirm the prediction that certain tran
sition types are preferred over others, in the following order: continue > retain >
smooth shift > rough shift. The prediction holds for both languages. Another
study of different conversations in Spanish (task-oriented dialogues) showed a similar
distribution of transitions (Taboada 2002). It is clear that, when the topic is not being
changed (which results in an empty Cb, and is outside the results in Table 3), speakers
prefer to continue talking about exactly the same entity about 65% of the time.
The next step in the analysis consists of relating the transition type to the
anaphoric term used for the backward-looking center in the utterance, which can
be considered the topic of that utterance. Each Cb was coded, and divided into one
of the following categories: zero pronoun, pronoun, clitic, noun phrase, demonstra
tive pronoun, and other. At first we also coded possessive determiners, wh-words and
adjectives that contain some reference (e.g., gender and number in Spanish). The
numbers for each of those were so small that we decided to group them under the
“Other” category. Examples (11)–(17) provide some representative instances of each
type of Cb, with the Cb in question in italics. Some categories, of course, only ap
ply to Spanish (clitic and, to a certain extent, zero pronoun). Table 4 presents the
results of the analysis, and the following sections examine the distribution within each
transition type.
 Maite Taboada
(11) Zero pronoun

A: a. Elisa, ¿cómo anda?
‘Elisa, how’s she?’
B: b. Bien, __ está bien, está sigue laburando donde laburaba.

‘Well, (she) is well, she’s she continues to work where she worked.’
(12) Pronoun
B: a. so I might I’m going to my sister’s tomorrow
b. she needs a babysitter on Thursday
(13) Clitic
B: a. yo, yo, yo laburo ahora con la máquina del tipo éste, viste
‘I, I, I work now with this guy’s computer, you know.’
A: b. sí, ¿qué tiene?
‘Yeah, what does (he) have?’
B: c. lo hice comprar una de éstas. . .
‘(I) made him buy one of these. . . ’
(14) Noun phrase

A: a. and she said to ask about the food what kind of food do you have
B: b. oh the food is fine they’re they’re uh
(15) Demonstrative pronoun
A: a. So now what is carnival?
b. Does that have to do with Easter or not? No.
(16) Other – Possessive determiner
B: a. I want to tell you
b. that when I got to Chicago
c. my sister Agnes was right there to meet me
(17) Other – Possessive pronoun8
B: a. Sólo recibiste esas cuatro.
‘You only got those four [letters].’
A: b. Y una de Marta.
‘And one from Marta.’
B: c. Cinco.
‘Five.’
A: d. Sí.
‘Yes.’
B: e. Bueno, mía, mañana seguro que ya te llega por lo más tardar,

‘Well, mine, I’m sure you’ll get tomorrow at the latest,’
Anaphoric terms and focus of attention 
. Continue transitions
In the continue transitions, most Cbs are realized as pronouns in English (80% of
the total number of Cbs in English), and as zero pronouns in Spanish (49.5%). That is
to be expected: if the topic is continued from the previous utterance, the least marked
anaphoric term available in the language is used – pronoun in English, zero pronoun in
Spanish (Bentivoglio 1983; Givón 1983). However, the percentages are quite different:
pronoun is used much more often in English than in Spanish. That is because Spanish
has yet another type of anaphoric term available: clitics. These account for 21% of the
continue transition Cbs. In addition, Spanish can also resort to a Subject pronoun,
just as English does.
An interesting aspect in English is the presence of some zero pronouns, as in ex
ample (18). The speakers are discussing the rate for a long-distance phone call. Speaker
B called somebody else, and explains that the rate, ‘it’ in (18a), was the medium price.
Then, after A’s backchannel, Speaker B refers back to the rate, but does not use a pro
noun at all. Subject ellipsis in English is, although generally considered rare, quite
common in certain conversational contexts (Nariyama 2004).
(18) B: a. it was like the medium price
A: b. yeah
B: c. __ wasn’t the cheapest
Adding up zero and pronouns in English; and zero, pronoun and clitics for Spanish, we
find very similar percentages: 85% versus 79.5%. The next question is: how does Span
ish choose among those three types? It is obvious that the choice between pronoun and
clitic is determined by syntactic function: only Objects can be expressed as clitics. But
why does a speaker choose an overt Subject pronoun over a zero pronoun? The answer
to that question is, in part, related to dialectal variation in Spanish (Cameron 1992;
Enríquez 1984), with some dialects using pro-drop more often than others. There are
also pragmatic constraints on the presence of a Subject pronoun, such as claiming the
floor for an extended period of time, especially in the case of first and second person
pronouns (Davidson 1996). Stewart (1999) proposes that the use of the first person
singular pronoun is a politeness resource, which helps contrast the speaker with other
individuals or groups. Luján (1999) also points out the contrastive character of first
and second person pronouns. In example (19), the speaker seems to be emphasizing
his role as an actor in the narrative he is constructing. He is discussing how he came to
realize that he had a rash. In (19a), he is himself the backward-looking center. In (19b),
he is also the backward-looking center. However, in (19b), he uses yo (‘I’) to describe
his own thought processes. This utterance could be translated as ‘I thought, maybe it’s
something [serious]’.
(19) a. de repente me salió un granito en la ingle,
‘all of a sudden I got a zit in my groin,’
 Maite Taboada
b. yo dije, uy, yayay, no vaya a ser algo –

‘I said, uh-oh, maybe it’s something. . . ’
Third person pronouns are also expressed in Spanish, even when they are clear from
context, for emphasis. In (20), the speakers are discussing B’s friend, who applied for a
scholarship and got it. The friend is the Cb in (20a), because the Subject is impersonal,
and therefore lower in the Cf list. In (20c), the friend is referred to with a pronoun,
because it is emphasized.
(20) B. a. y se la dieron
‘and they gave it [=the scholarship] to her’
A: b. Ah, ¿sí? {laughter}
‘Really?’
B: c. No, no, ni ella lo podía creer,
‘No, no, not even she herself could believe it,’
Another interesting question that affects both languages is the choice of full noun
phrases. Both use an NP to encode the Cb, even in cases where the Cb is the same
as in the previous utterance (of all the continue transitions, 12% in English; 16% in
Spanish). We included proper names together with NPs, and many of the repetitions
have to do with the use of proper names. Proper name repetition might be a device
to establish common ground between the interlocutors. Downing (1996) points out
that proper names are used very often in conversation: to introduce individuals in
the conversation, as the most easily identifiable form of reference; and to refer again
to those individuals, as a marker of true familiarity with the referent denoted by the
proper name. In example (21), Johnny (probably speaker A’s partner, friend or child)
has been part of the discourse, but never before mentioned by name. Speaker A talks
about camping and biking, but she always uses the first person plural pronoun we (‘we
camped there’, ‘we rented a bike’). Speaker B mentions the name Johnny, part of the
previous ‘we’, for the first time here, and speaker A repeats the name. It is difficult to
explain why the proper name is repeated. B already knows that A is familiar with the
referent, since Johnny was in A’s party. It is possible that A repeats it to convey that
she understood who B was referring to.
(21) B: a. so did Johnny stay in front?
A: b. Johnny was basically in front
. Retain transitions
The percentage of pronoun forms (whether zero, pronoun, or clitic) in the retain
transitions decreases considerably, as compared to the continue transitions. In a re
tain there is a slight change of topic: the Cb of the current utterance is the same as the
Cb of the previous utterance. However, the current utterance has seemingly moved
the focus from the Cb to another entity, which is now the highest-ranked entity in the
Anaphoric terms and focus of attention 
Cf list, typically the grammatical Subject. Example (22) shows the shift from public
beach to I, the speaker. Since public beach is the link between (22a) and (22c), it is
the Cb (the backchannel in (22b) is ignored for analysis purposes). But in (22c) a new
entity, the Subject, is the Cp, the highest ranked member of the Cf list.
(22) A: a. the public beach is kind of hard to find
Cf: public beach – Cb: public beach
B: b. oh really
A: c. I mean I didn’t I didn’t know where it was
Cf: I, public beach – Cb: public beach
The Cb may, in retain transitions, still be expressed through a pronoun. According
to Di Eugenio (1998), this is possible especially in cases where the pronominalized
(Subject) Cb unequivocally refers to an entity other than the previous Cb. This un
equivocal reference can be achieved through syntactic features such as person and
number. In the previous example, person differentiates the pronominalized Subject
I from the pronominalized Cb it.
In cases where the reference is not clear, speakers use other anaphoric terms to
encode the Cb of a retain transition, which accounts for the decrease in pronoun
forms from the continue transition. In example (23), there are three different entities:
she, a third person female encoded in (23a) as her, mother, and cousin. In (23b), the
focus shifts from her cousin, the Subject in (23a), to she, the Subject in the new
utterance. Because at least two of these entities (she, her mother), and maybe even
her cousin are female, the use of a pronoun to refer to the backward-looking center
her mother would have been infelicitous. The speaker avoids confusion by using a
full noun phrase.
(23) a. First, you know, her mother her cousin is a manager of a travel agency.
b. And she invited her mother and her fr- uh and the cousin to come on out.
Some of the uses of anaphoric terms other than pronouns can be related to spoken lan
guage phenomena, such as holding the turn. In example (24c), which shows a retain
transition in relation to (24b), Speaker A repeats mi hermana (‘my sister’), the Cb, as
a full noun phrase. It seems that he is holding the turn while he thinks about what
he can tell his interlocutor. Notice that between (24a) and (24b) there is a continue
transition, with the Cb my sister also encoded as an NP.9
(24) B: a. ¿Y tu hermana?
‘And your sister?’ Cf: A’s sister – Cb: Ø
A: b. Mi hermana está bie
‘My sister is fine’ Cf: A’s sister – Cb: A’s sister
c. No, hombre, qué te platico de mi hermana,
‘Well, what can (I) tell you about my sister,’
Cf: A (null), B (te), A’s sister (mi hermana) – Cb: A’s sister
 Maite Taboada
. Smooth shift transitions
Smooth shifts represent the cases in which speakers are preparing to shift topics:
the previous topic is still present in the utterance, but a new entity has gained promi
nence, with the Cb of the current utterance being different from the Cb of the previous
one. However, in our data, the types of pronouns used do not differ very much from
those in continue transitions. The three pronoun types (zero, pronoun and clitic,
where applicable) still account for the majority of Cb expressions. The only significant
change, for both languages, is that few Cbs fall under the “Other” category. This could
be because it is necessary to encode the Cb with a form that makes the reference un
ambiguous. Some of the “Other” forms, such as demonstrative pronouns or adverbials
might make this difficult.
Example (25) shows a case of smooth shift with a clitic (me, ‘me’). In this case,
the reference is clear because of the switch between first and third person (‘I told her’,
‘She told me’).
(25) a. Yo le dije, sabes qué, maestra, estoy enfermísimo.
‘I told her, (you) know what, teacher, (I)’m really sick.’
b. Me dijo, no me bronca.
‘(She) told me, (it) doesn’t matter to me.’
In example (26), Speaker B is discussing her father’s dizziness. The entities in (26a)
are it (the dizziness) and heat. Speaker A picks up on the least prominent of those
utterances, heat, and elaborates on it. The shift from dizziness to heat requires that
Speaker A use a full noun phrase.
(26) B: a. and maybe it was because of the was the heat or something
Cf: dizziness (it), heat (the heat) – Cb: dizziness
A: b. maybe the heat it’s so God damn hot here. . .
Cf: heat, here – Cb: heat
. Rough shift transitions
Finally, rough shifts represent a change in the topic, while still preserving the same
entities from one utterance to the next (a radical change of topic results in an empty
Cb; those utterances were not considered in the analysis). Again, we see that the major
ity of Cbs are encoded using one of the three pronoun categories, with English using
exclusively the full pronoun, and Spanish again distributing most of the Cbs into either
zero pronouns or clitics, depending on their grammatical function. The most signif
icant difference between the two languages in rough shifts is under “Other”, used
much more often in English. The numbers of rough shifts, are, however, smaller,
which makes generalizations more difficult.
In some cases, a pronoun is the only possibility, because the pronouns are used
to refer to the speakers themselves. In (27c), the Cb is the speaker, the highest-ranked
Anaphoric terms and focus of attention 
entity from (27b). But since in (27c) the interlocutor is part of the Cf list (the ad
dressee of the imperative), the speaker is no longer the highest-ranked member of the
Cf list, which results in a rough shift. The pronoun me is enough to understand the
reference. Furthermore, no other form could have been used. It is also interesting to
point out that this example conforms to Rule 1 of Centering Theory: the Cb of (27c)
is a pronoun, and no other pronoun is used in the example.
(27) a. there’s no nice blessed mother over here {laugh}
Cf: blessed mother, here
b. and I would really like it for my bedroom
Cf: B (I), blessed mother (it), bedroom – Cb: blessed mother
c. just send me the blessed mother not the cross
Cf: A (imperative), B (me), blessed mother, cross – Cb: B
Similarly, in (28), the only possible form of the Cb is the noun phrase el treinta y uno
(‘the thirty-first’, New Year’s Eve), which is part of a prepositional phrase. Speaker B
needs to make clear which one of the dates she is picking, and the repetition of the
noun phrase seems necessary.
(28) A: a. Y no tienes todavía programa ni para el veinticuatro ni para el treinta
y uno.
‘And you have no plans yet for the twenty-fourth or for the thirty-
first.’
B: b. Para el treinta y uno estamos viendo
‘For the thirty-first we’re checking. . . ’
. Conclusions
Speakers manage topics and referents in conversation with incredible ease. The pur
pose of this paper has been to explore some of the management strategies used by
speakers in English and Spanish. Based on a corpus analysis of five English and five
Spanish conversations, I have discussed some of the factors involved in the choice of
anaphoric term for a given referent.
The framework of the analysis is Centering Theory, which provides clear defini
tions of ‘topic’ progression, and the relations between pairs of utterances in discourse.
Using Centering, the analysis shows some tendencies in what anaphoric term is cho
sen for the backward-looking center (Cb) of a given utterance, what could loosely be
characterized as the topic of the utterance. The choice of anaphoric term for the Cb is
related to whether the topic is continued, slightly changed or abruptly changed from
the utterance immediately before.
The results show that speakers tend to use pronominal forms when the topic is
continued (continue transitions), and fewer pronominal forms when there is a slight
change in topic (retain transitions). This was not surprising at all; more interest
 Maite Taboada
ing are the cases where speakers do not choose a pronoun (although one could be
chosen without creating ambiguity); or where speakers use a pronoun instead of a
more explicit form. In most cases, we can relate the choices against expectation to spo
ken language phenomena: speakers use a ‘heavier’ form than expected (e.g., full noun
phrase instead of pronoun) when they want to emphasize the referent. They also tend
to use heavier forms across turns, to signal familiarity with the referent, or that the
referent is clearly established for both interlocutors. On the other hand, they use less
explicit forms (e.g., pronoun when a noun would have been expected) if the pronoun
carries enough grammatical information to make the referent clear.
Future directions of this research include a more thorough examination of proper
name use in spoken conversation, and clitic doubling in Spanish (Suñer 1988). We
are also investigating optimal utterance segmentation methods, taking into account
Centering structures, but also agreement among coders. Another area to explore is
the relationship between the local focus of discourse, which Centering is supposed to
handle, and global discourse structure (turn-taking, side sequences, overall discourse
segments).
Notes
* A previous version of this paper was presented at the Third International Contrastive Linguis
tics Conference held in Santiago de Compostela (Spain) in September 2003. I would like to thank
Loreley Hadic Zabala for helping in the coding of the data. This work was supported by the Min
istry of Education of Spain, and the Xunta de Galicia, under project MCYT-FEDER BFF2002
02441/XUGA-PGIDIT03PXIC20403PN, and by Simon Fraser University, under a SSHRC grant
and a Discovery Parks grant.
. I italicize ‘topic’, because in this paper it refers exclusively to the backward-looking center,
which is a well-defined term (see Section 2).
. The order of the two Objects is variable in the literature. In the analyses carried out so far, we
have ranked Indirect before Direct. This also corresponds to the surface order of the sentence in
the example.
. Small capitals indicate that the list contains entities, not their linguistic realization. In Cor
nish’s (1999) terms, it contains referents, not antecedents or anaphors.
. Centering transitions are one possible explanation for coherence. A text can be coherent
without repeating or referring to the same entities (Brown & Yule 1983; Poesio et al. 2000).
. Similar to the T-unit used in developmental studies and in text analysis (Hunt 1977).
. A higher number of empty backward-looking centers is undesirable, since it indicates that
Centering structures found few connections between consecutive pairs of utterances. It is as
sumed that most utterances will have a connection with the preceding speech.
. Spanish examples are glossed only when word order or grammatical information are con
sidered relevant. Otherwise, they are simply translated. Abbreviations used in the glosses:
1/2/3 (‘first/second/third person’); cl (‘clitic’); nom (‘nominative’); acc (‘accusative’); dat (‘da
Anaphoric terms and focus of attention 
tive’); sg (‘singular’); pl (‘plural’); fem (‘feminine’); masc (‘masculine’); pres (‘present’); pret
(‘preterite’).
. In example (17e), mía (‘mine’) is the backward-looking center. It refers to cinco (‘five [let
ters]’) in (17c). (17d) was not part of the analysis, because it is a backchannel that contains
no entities.
. Another interpretation of this example would break the noun phrase mi hermana (‘my sis
ter’), and extract two entities from it: I (mi) and sister (hermana). I have taken that approach
elsewhere, but will ignore that distinction for illustration purposes in this example.
References
Ariel, Mira (1990). Accessing Noun Phrase Antecedents. London: Routledge.

Bentivoglio, Paola (1983). “Topic continuity and discontinuity in discourse: a study of spoken
Latin-American Spanish”. In T. Givón (Ed.), Topic Continuity in Discourse: A Quantitative
Cross-Language Study (pp. 255–311). Amsterdam and Philadelphia: John Benjamins.
Brennan, Susan E. (1995). “Centering attention in discourse”. Language and Cognitive Processes,
10, 137–167.
Brennan, Susan E., Marilyn W. Friedman, & Carl J. Pollard (1987). “A Centering approach to
pronouns”. In Proceedings of the 25th Annual Meeting of the Association for Computational
Linguistics (ACL-87) (pp. 155–162). Stanford, CA, USA.
Brown, Gillian & George Yule (1983). Discourse Analysis. Cambridge: Cambridge University
Press.
Byron, Donna K. & Amanda Stent (1998). “A preliminary model of Centering in dialog”. In
Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics
(ACL-98) (pp. 1475–1477). Montreal, Canada.
Cameron, Richard (1992). “Pronominal and Null Subject Variation in Spanish: Constraints,
Dialects, and Functional Compensation.” Ph.D. dissertation. University of Pennsylvania.
Chafe, Wallace (1976). “Givenness, contrastiveness, definiteness, subjects, topics, and point of
view”. In C. Li (Ed.), Subjects and Topics (pp. 26–56). New York: Academic Press.
Cornish, Francis (1999). Anaphora, Discourse, and Understanding: Evidence from English and
French. Oxford: Clarendon Press.
Daneš, František (1974). “Functional Sentence Perspective and the organization of the text”. In
F. Daneš (Ed.), Papers on Functional Sentence Perspective (pp. 106–128). Prague: Academia.
Davidson, Brad (1996). “‘Pragmatic weight’ and Spanish subject pronouns: the pragmatic and
discourse uses of ‘tú’ and ‘yo’ in spoken Madrid Spanish”. Journal of Pragmatics, 26, 543–
565.
Di Eugenio, Barbara (1998). “Centering in Italian”. In M. A. Walker, A. K. Joshi, & E. F. Prince
(Eds.), Centering Theory in Discourse (pp. 115–137). Oxford: Clarendon Press
Downing, Angela (1991). “An alternative approach to Theme: a systemic-functional perspec
tive”. Word, 42, 119–143.
Downing, Pamela A. (1996). “Proper names as a referential option in English conversation”.
In B. A. Fox (Ed.), Studies in Anaphora (pp. 95–143). Amsterdam and Philadelphia: John
Benjamins.
 Maite Taboada
Eckert, Miriam & Michael Strube (1999). “Resolving discourse deictic anaphora in dialogues”.
In Proceedings of the 9th Conference of the European Chapter of the Association for Compu–
tational Linguistics (EACL-99) (pp. 37–44). Bergen, Norway.
Enríquez, Emilia (1984). El pronombre personal sujeto en la lengua española hablada en Madrid.
Madrid: Consejo Superior de Investigaciones Científicas.
García, Erica (1975). The Role of Theory in Linguistic Analysis: The Spanish Pronoun System.
Amsterdam: North-Holland.
Givón, Talmy (1983). “Topic continuity in discourse: an introduction”. In T. Givón (Ed.), Topic
Continuity in Discourse: A Quantitative Cross-Language Study (pp. 1–41). Amsterdam and
Gordon, Peter C., Barbara J. Grosz, & Laura A. Gilliom (1993). “Pronouns, names, and the
Centering of attention in discourse”. Cognitive Science, 17, 311–347.
Gordon, Peter C., Randall Hendrick, Kerry Ledoux, & Chin Lung Yang (1999). “Processing of
reference and the structure of language: an analysis of complex noun phrases”. Language
and Cognitive Processes, 14, 353–379.
Grosz, Barbara J., Aravind K. Joshi, & Scott Weinstein (1995). “Centering: a framework for
modelling the local coherence of discourse”. Computational Linguistics, 21, 203–225.
Grosz, Barbara J. & Candace L. Sidner (1986). “Attention, intentions, and the structure of
discourse”. Computational Linguistics, 12, 175–204.
Grosz, Barbara J. & Candace L. Sidner (1998). “Lost intuitions and forgotten intentions”. In A.
Walker, A. K. Joshi, & E. F. Prince (Eds.), Centering Theory in Discourse (pp. 39–51). Oxford:
Clarendon Press.
Gundel, Jeanette K. & Thorstein Fretheim (2004). “Topic and focus”. In L. Horn & G. Ward
(Eds.), The Handbook of Pragmatics (pp. 175–196). Oxford: Blackwell.
Gundel, Jeanette K., Nancy Hedberg, & Ron Zacharski (1993). “Cognitive status and the form
of the referring expressions in discourse”. Language, 69, 274–307.
Hadic Zabala, Loreley & Maite Taboada (2004). “Centering Theory in Spanish: Coding Manual”.
Manuscript. Simon Fraser University.
Halliday, Michael A. K. (1967). “Notes on Transitivity and Theme in English. Part II”. Journal of
Hudson-D’Zmura, Susan & Michael K. Tanenhaus (1998). “Assigning antecedents to ambiguous
pronouns: the role of the center of attention as the default assignment”. In M. A. Walker,
A. K. Joshi, & E. F. Prince (Eds.), Centering Theory in Discourse (pp. 199–226). Oxford:
Clarendon Press.
Hunt, Kellogg W. (1977). “Early blooming and late blooming syntactic structures”. In C. R.
Cooper & L. Odell (Eds.), Evaluating Writing: Describing, Measuring and Judging (pp.
91–104). Urbana, IL: NCTE.
Hurewitz, Felicia (1998). “A quantitative look at discourse coherence”. In M. A. Walker, A. K.
Joshi, & E. F. Prince (Eds.), Centering Theory in Discourse (pp. 273–291). Oxford: Clarendon
Press.
Iida, Masayo (1998). “Discourse coherence and shifting centers in Japanese texts”. In M. A.
Walker, A. K. Joshi, & E. F. Prince (Eds.), Centering Theory in Discourse (pp. 161–180).
Oxford: Clarendon Press.
Jackendoff, Ray (1972). Semantic Interpretation in Generative Grammar. Cambridge: MIT Press.
Jaeggli, Osvaldo A. (1986). “Arbitrary plural pronominals”. Natural Language and Linguistic
Theory, 4, 43–76.
Anaphoric terms and focus of attention 
Joshi, Aravind K. & Scott Weinstein (1998). “Formal systems for complexity and control of
inference: a reprise and some hints”. In M. A. Walker, A. K. Joshi, & E. F. Prince (Eds.),
Centering Theory in Discourse (pp. 31–38). Oxford: Clarendon Press.
Kameyama, Megumi (1985). “Zero Anaphora: The Case of Japanese”. Ph.D. dissertation.
Stanford University.
Kameyama, Megumi (1998). “Intrasentential Centering: a case study”. In M. A. Walker, A. K.
Joshi, & E. F. Prince (Eds.), Centering Theory in Discourse (pp. 89–112). Oxford: Clarendon
Press.
Kuno, Susumu (1972). “Functional Sentence Perspective: a case study from Japanese and
English”. Linguistic Inquiry, 3, 269–320.
Kuno, Susumu (1987). Functional Syntax: Anaphora, Discourse and Empathy. Chicago:
Lambrecht, Knud (1994). Information Structure and Sentence Form: Topic, Focus, and the Mental
Representation of Discourse Referents. Cambridge: Cambridge University Press.
Luján, Marta (1999). “Expresión y omisión del pronombre personal”. In I. Bosque & V. Demonte
(Eds.), Gramática descriptiva de la lengua española (pp. 1275–1315). Madrid: Espasa.
Mendikoetxea, Amaya (1999). “Construcciones con se: medias, pasivas e impersonales”. In I.
Bosque & V. Demonte (Eds.), Gramática descriptiva de la lengua española (pp. 1631–1722).
Madrid: Espasa.
Miltsakaki, Eleni (2001). “Centering in Greek”. Paper presented at the 15th Symposium on
Theoretical and Applied Linguistics. Thessaloniki, Greece.
Miltsakaki, Eleni (2002). “Toward an aposynthesis of topic continuity and intrasentential
anaphora”. Computational Linguistics, 28, 319–355.
Nariyama, Shigeko (2004). “Subject ellipsis in English”. Journal of Pragmatics, 36, 237–264.
Poesio, Massimo (2000). “The GNOME annotation scheme manual”. Technical Report.
University of Edinburgh. (Available from: http://www.cogsci.ed.ac.uk/∼gnome/)
Poesio, Massimo, Hua Cheng, Renate Henschel, Janet Hitzeman, Roger Kibble, & Rosemary
Stevenson (2000). “Specifying the parameters of Centering Theory: a corpus-based
evaluation using text from application-oriented domains”. In Proceedings of the 38th Annual
Meeting of the Association for Computational Linguistics (ACL-2000) (pp. 400–407). Hong
Kong.
Prasad, Rashmi & Michael Strube (2000). “Discourse salience and pronoun resolution in Hindi”.
Penn Working Papers in Linguistics, 6, 189–208.
Prince, Ellen F. (1981). “Towards a taxonomy of given-new information”. In P. Cole (Ed.),
Radical Pragmatics (pp. 223–255). New York: Academic Press.
Schiffrin, Deborah (1994). Approaches to Discourse. Malden, MA: Blackwell.
Stewart, Miranda (1999). “Hedging your bets: the use of yo in face-to-face interaction”. Web
Journal of Modern Language Linguistics, 4–5, http://wjmll.ncl.ac.uk/issue04-05/stewart.htm
Strube, Michael & Udo Hahn (1996). “Functional Centering”. In Proceedings of the 34th Annual
Meeting of the Association for Computational Linguistics (ACL-96) (pp. 270–277). Santa
Cruz, CA, USA.
Strube, Michael & Udo Hahn (1999). “Functional Centering – Grounding referential coherence
in information structure”. Computational Linguistics, 25, 309–344.
Suñer, Margarita (1983). “proarb ”. Linguistic Inquiry, 14, 188–191.
Suñer, Margarita (1988). “The role of agreement in clitic-doubled constructions”. Natural
Language and Linguistic Theory, 6, 391–434.
 Maite Taboada
Taboada, Maite (2002). “Centering and pronominal reference: in dialogue, in Spanish”. In J.

Bos, M. E. Foster, & C. Matheson (Eds.), Proceedings 6th Workshop on the Semantics and
Pragmatics of Dialog, EDILOG (pp. 177–184).
Taboada, Maite (forthcoming). “Reference, centers and transitions in spoken Spanish”. In J.
Gundel & N. Hedberg (Eds.), Reference and Reference Processing. Oxford: Oxford University
Press.
Turan, Ümit Deniz (1995). “Null vs. Overt Subjects in Turkish Discourse: A Centering Analysis”.
Ph.D. dissertation. University of Pennsylvania.
Vallduví, Enric (1990). “The Informational Component”. Ph.D. dissertation. University of
Pennsylvania.
Vallduví, Enric & Elisabet Engdahl (1996). “The linguistic realization of information packaging”.
Walker, Marilyn A., Masayo Iida, & Sharon Cote (1994). “Japanese discourse and the process of
Centering”. Computational Linguistics, 20, 193–232.
Walker, Marilyn A., Aravind K. Joshi, & Ellen F. Prince (1998). “Centering in naturally occurring
discourse: an overview”. In M. A. Walker, A. K. Joshi, & E. F. Prince (Eds.), Centering Theory
in Discourse (pp. 1–28). Oxford: Clarendon Press.
Yngve, Victor H. (1970). “On getting a word in edgewise”. In Papers from the Sixth Regional
Meeting of the Chicago Linguistics Society (pp. 567–577). Chicago: University of Chicago.
P IV
Collocations and formulaic language

Formulaic language
An overview with particular reference
to the cross-linguistic perspective*
This article discusses, from a cross-linguistic perspective, two approaches to the

study of formulaic language, those of the corpus linguist and psycholinguist. I
first review corpus-based work, including my own, on recurrent continuous or
discontinuous sequences in English and Spanish. I then broaden the discussion
to include further work, on English, Spanish and Italian, which demonstrates the
complex interaction of grammar and lexis in the syntagmatic patterns found in
corpora and validates the concepts of semantic preference and semantic prosody.
I then go on to outline briefly Wray’s psycho-socially based model of formulaic
language and other psycholinguistic evidence for the importance of
(semi-)prefabricated sequences. Finally, I point out some implications of both
corpus-based and psycholinguistic work for language teaching and learning.
. Introduction
Fixed or semi-fixed sequences of words have long been of interest not only to theo
retical and descriptive linguists, but also to teachers and learners of foreign languages,
in the guise of the idioms which are so important for a native-like command of a lan
guage. In recent years, however, a considerable volume of research has demonstrated
that the influence of such sequences is much more pervasive than might be suggested
by the focus on idioms, in the sense of multi-word items whose meaning cannot be
predicted from those of their component words. This research has been of two rather
different but complementary kinds: studies showing the frequency of repeated syntag
matic patterns in corpora of texts, and work done from a psycholinguistic viewpoint.
In this paper I shall examine the corpus approach in some detail, and the psycholin
guistic approach rather more briefly, with an emphasis on the relevance of this work
for language comparison, translation and the teaching and learning of a second or for
eign language. The paper will conclude with a brief discussion of some implications
for language teaching and learning.
The term formulaic language, used in the title of this paper, is but one of many
which can be found in the literature. It is intended to suggest that certain strings in
 Christopher S. Butler
language may be treated by the language user, not as sequences of individual items
(for now, let us say words, though we shall see later that at least one approach has
a rather broader view of the items which may be involved), but as single entities, or
formulas, stored and retrieved as wholes. Another commonly used term is phraseology,
emphasising the multi-word nature of such entities. A further term which I have used
in my own work is recurrent multi-word sequence, which brings in the idea that such
strings are frequent in the natural use of language. These terms, and others to be found
in the literature, should not be seen as synonymous, since, as I have pointed out, they
emphasise different aspects of a complex phenomenon (see Wray 2002: 8–10 for more
detailed discussion). I shall use the term formulaicity in what follows as a general label,
though we shall see that different approaches focus on different features, and so tend to
use different terms. Part of the problem here is the difficulty of developing operational
procedures for the isolation of formulaic strings. I shall start with corpus-oriented
work, which takes a practical and quite pragmatic approach to this problem.
. Corpus-oriented approaches to formulaic language
In recent years there has been considerable interest in phraseology from corpus lin
guists, and good overviews can be found in, for example, Cowie (1998) and Moon
(1998). Here, in view of my focus on cross-linguistic aspects, I shall not attempt an
overall survey, but shall concentrate on a small number of studies in English, Spanish
and Italian, which will reveal both the similarities and the differences which can be
observed across these three languages. First I shall review studies of English and Span
ish concerned with what I referred to above as recurrent multi-word sequences, that
is sequences of words (continuous or discontinuous) which are repeated, in exactly
the same form, in some body of text under investigation. I then broaden the scope of
the inquiry to include studies of other, more flexible types of recurrent syntagmatic
patterning in text.
. Recurrent continuous sequences in English and Spanish
Studies of recurrent continuous multi-word sequences in the London-Lund Corpus of

spoken English (0.5 million words) have been made by Altenberg and his colleagues
(Eeg-Olofsson & Altenberg 1996; Altenberg & Eeg-Olofsson 1997; Altenberg 1998).
Biber et al. (1999) have investigated such sequences in the 40 million word Longman
Spoken and Written English Corpus, containing samples from British and American
conversation, fiction, news, academic prose and general prose, together with some
British non-conversational speech. Stubbs (2002) investigates recurrent continuous
sequences in a 2.5 million word corpus derived from the London-Lund Corpus, the
Lancaster-Oslo-Bergen (LOB) Corpus (written British English, 1961, about 1 million
words) and the Freiburg-LOB (FLOB) Corpus (written British English, 1991, about 1
Formulaic language 
million words) and more limitedly in the British National Corpus (100 million words
of British English, 90% written, 10% spoken). Stubbs & Barth (2003) examine recur
rent sequences in the LOB and FLOB Corpora and also in their American equivalents,
the Brown and Frown corpora. Work on continuous sequences in Spanish (Butler
1997) used a selection of small individual corpora covering spoken and newspaper
texts, together with edited transcripts of interviews for magazines, with a total of about
1.5 million words. Comparison of these various studies will show clearly that there are
some tendencies which are valid across corpora of different types, both within and be
tween languages, and others which are more prevalent in either in spoken or written
corpora, or in one language rather than another.
An initial significant finding is that recurrent sequences are not just a marginal
phenomenon, but rather are common in the language we use. One important prob
lem here is that different researchers have used different cut-off criteria for regarding
a sequence as recurrent, so making comparison difficult. Altenberg (1998: 102), who
defines a recurrent sequence as any string occurring more than once in the corpus,
presents a graph showing that 3-word combinations occur more than 60,000 times
in the London-Lund Corpus, 4-word combinations about 20,000 times and 5-word
strings about 4,000 times; he also estimates that more than 80% of the words in the
corpus occur within recurrent word combinations. Biber et al. (1999), in their study
of what they call ‘lexical bundles’ in English, defined as “sequences of word forms that
commonly go together in natural discourse” (p. 992), report that 3-word bundles with
a frequency of 10 or more in the corpus occur between 60,000 and 80,000 times per
million words, depending on the register involved, while 4-word bundles of F ≥ 10 oc
cur between 5,000 and 8,500 times per million words (p. 994). Butler (1997: 65), used
a much more restrictive cut-off frequency of 10 in the smaller Spanish corpora used
(consisting of between 125,000 and 153,000 words) and 30 in the much larger cor
pus (950,000 words), and correspondingly found much lower frequencies per million
words of between 810 and 2,570 for 3-word strings, 90-270 for 4-word strings. Despite
the difficulties in comparing across studies, it is clear that recurrent word strings are a
quantitatively important part of the language we use.
A further finding which is valid for both English and Spanish is that spoken lan
guage contains more recurrent strings than written language. Altenberg (1998: 101)
reports that “spoken English is very rich in such recurrent strings”, while Biber et al.
(1999: 994) find that the overall frequencies of lexical bundles are greater in conversa
tion than in academic prose. In Spanish, many more strings of F ≥ 10 occurred in a
corpus of spoken language than in a newspaper language corpus of similar size, while
edited transcripts of interviews in magazines showed intermediate frequencies (Butler
1997: 65).
It turns out that most of the sequences found in both English and Spanish are not
semantically opaque, as with the expressions normally classified as ‘idioms’: for most of
them, the meaning of the whole is derivable from that of the parts. Furthermore, most
of them could be generated by the rules of the normal grammar acting in an on-line
 Christopher S. Butler
productive mode, though those which are structurally incomplete (e.g. on top of the)
would, of course, need a suitable following string in order to complete the structure.
Altenberg (1998) shows that for English, the recurrent sequences can be classi
fied in terms of their structural characteristics, as independent clauses (e.g. that’s all
right, oh I see), dependent clauses (e.g. as you know, if I may), multiple clause con
stituents (e.g. and you know, do you know), single clause constituents (e.g. that sort of
thing, a lot more), and incomplete phrases (e.g. a couple of, in order to). The multiple
clause constituent type can be classified further in terms of the structural elements in
cluded and their position in the clause, in a scheme inspired by the work of Halliday
on theme and rheme: ‘frames’ (e.g. and you know, well I mean, well of course) occur in
pre-subject position; ‘onsets’ consist of elements, including the subject, which occur
before the finite verb; ‘stems’ (e.g. I think/thought (that), I suppose (that), so I said, do
you think, I don’t mind) contain the subject and finite verb, plus any pre-subject ele
ments; the ‘medial’ type of multiple constituent sequence consists of any post-subject
elements up to and including the verb; ‘rhemes’ contain the verb and any following
object, complement and/or adverbial; ‘tails’ consist only of object, complement and
adverbial elements; ‘transitions’ comprise any adverbial and discourse items at the end
of the clause. The most common types were stems, followed by onsets and frames. As
Altenberg points out, these various types of sequence overlap, and he regards them as
“successive building blocks that speakers use again and again to compose utterances in
on-going discourse” (1998: 111).
Altenberg’s classification can also be applied to recurrent sequences in Spanish,
with the exception that full clauses appear to be rather rare. Examples would include:
yo creo que sí ‘I think so’ (full independent clause), y por lo tanto (frame type of mul
tiple clause constituent sequence), la verdad es que lit. ‘the truth is that’ (stem type
of multiple clause constituent sequence), una serie de cosas ‘a series of things’ (single
clause constituent), en el caso de ‘in the case of ’ (incomplete phrase).
In both languages, recurrent sequences often consist of an obligatory core with
possible extensions: an example from Altenberg’s data is I don’t know, which frequently
occurs alone, but may also be preceded by oh or well, while in Spanish we have se
quences such as yo creo que ‘I think that’, which may occur by itself or prefaced by
conjunctions such as pero ‘but’ porque ‘because’, y ‘and’, or que ‘that’.
A further finding for both languages is that sequences frequently overlap. Al
tenberg gives, as an example of a plausible string made up entirely of overlapping
recurrent sequences, and you know we’ve got to do so you know really, and Biber et
al. (1999: 995) report that the final word in a lexical bundle is often the start of a new
structural unit. In Spanish too, we have overlaps such as that of yo creo que with que es
mejor to give yo creo que es mejor ‘I think that it’s better’.
The similarities across languages are not only structural, but also functional,
though there are some differences of detail between English and Spanish here. Butler
(1997) shows that in both languages, recurrent sequences can be classified according to
the terminology of Halliday (1994), as being experiential in function (concerned with
the representational content of what is being said), interpersonal (giving the speaker
Formulaic language 
or writer’s own angle on the content, through expressions of personal knowledge,

possibility, desirability and the like) or textual (concerned with the resources for ma
nipulating information structure in order to construct a cohesive text).1 An important
finding was that many sequences in both English and Spanish are interpersonal in
nature, and another large group is textual in function, while relatively few are experi
ential, though written language tends to contain more of these than spoken language.
In part, the relative paucity of frequently repeated experiential strings can probably be
accounted for by the large range of things we might want to talk or write about in any
given text; nevertheless, the special importance of recurrent strings to the interpersonal
and textual functions of language is a finding of some importance.
Comparison of the more detailed functioning of recurrent strings in English
and Spanish shows both similarities and differences (see Butler 1998a). In both lan
guages, the (relatively few) experiential strings were concerned either with the subject
matter of the text (e.g. in the House of Commons, el Congreso de los Diputados ‘the
Chamber of Deputies’) or with the expression of location in time or space (at the
end/beginning/bottom of the, in the middle of the, and at the same time, a lo largo de
‘along, throughout’, a la hora de ‘when it comes to’, lit. ‘at the time of ’). Interpersonal
expressions of thanks and appreciation were frequent in English (e.g. thank you very
much indeed, that would be very nice), while speaker agreement or disagreement was
expressed in Spanish by means of repetitions of sí or no. Both languages had sequences
expressing degrees of probability (it seems to me that, as far as I know, as a matter of
fact, (y) la verdad es que (no) ‘and the truth is that’, with or without negative, a mí me
parece que ‘it seems to me that’, various sequences with yo creo que ‘I think/believe that’,
etc.). On the other hand, frequent multi-word expressions of obligation were found in
Spanish (lo que hay que hacer (es) ‘what needs to be done (is)’, hay que tener en cuenta
‘it’s necessary to take into account’), whereas such expressions were not reported by
Altenberg for English. Both languages use recurrent sequences to express point of view
(from the point of view of, desde el punto de vista de (with the same translation)), and
also for other expressions of speaker comment such as and I said well I or vamos a ver
si ‘let’s see/we’ll see if/whether’.
As far as the textual meanings are concerned, English uses recurrent strings to
indicate that additional information which could have been given is being assimilated
to that already provided (and that sort of thing, and so on and so forth, and all the rest
of it), while textually oriented sequences in Spanish focalise particular elements of the
message (lo que pasa/occurre (es) que ‘the thing is that’, lit. ‘what happens is that’, (y)
eso es lo que ‘(and) that’s what . . .’, que es lo que ‘which is what . . .’, qué es lo que . . .?
‘what is it that . . .?/what it is that . . .’), reformulate previous information (en el sentido
de que ‘in the sense that’), or act as connectives between clauses or sentences (y por lo
tanto ‘and therefore’).
An additional important conclusion from these studies is that in both English and
Spanish the sequences tend to come very early in the clause, providing an interpersonal
and textual background to the later experiential content. Altenberg (1998: 112–113)
notes that the type of sequence which he calls a ‘frame’, occurring before the subject
 Christopher S. Butler
in a clause, usually consists of conjunctions, response items such as well, discourse

items such as you know, and modal adverbials such as of course, occurring in a quite
rigid ordering pattern. Furthermore, he points out that the sequences which he dubs
stems, which contain at least the clause subject and a verb, “form the springboard of
utterances leading up to the communicatively most important – and lexically most
variable – element” (1998: 113). Biber et al. (1999: 1003) likewise note that the most
common lexical bundles in conversational material, consisting of a subject pronoun
(usually I) and a verb phrase, sometimes followed by the beginning of a complement
clause (e.g. I don’t know what), are used utterance-initially, and serve as ‘utterance
launchers’ giving a personal angle on the following material.
Finally, Stubbs & Barth (2003) show, for English, that recurrent sequences can
discriminate accurately between text types (fiction, belles lettres, academic writing), in
terms of both the actual chains which occur and the degree of repetition in a text type.
As mentioned earlier, differences in the frequency of repeated sequences have also been
found between spoken and written texts, for both English and Spanish.
. Recurrent discontinuous sequences in English and Spanish
So far I have discussed only continuous multi-word sequences. Sinclair & Renouf
(1988) also showed that corpora of English contain very frequent discontinuous pat
terns, referred to as collocational frameworks, in which two closed class words enclose
an open class item. Renouf & Sinclair (1991) studied this phenomenon in more detail,
concentrating on the patterns a __ of, an __ of, be __ to, too __ to, for __ of, had __ of
and many __ of in samples of spoken and written British English from the Birming
ham Collection of English Text (later to form the basis of the Bank of English). They
were able to show that there was a very high type-token ratio for each of the frame
works (i.e. a high recurrence rate in the corpus), showing that the frameworks are very
selective in terms of the words they enclose as collocates.
A detailed study of the most frequent collocates for a/an __ of (which unsurpris
ingly are all nouns) in the written part of the corpus revealed that the frequency order
did not correspond to the ranking of the individual words in the corpus: for instance,
lot was the most common collocate of the framework, but only 9th in the ranking of
the top 20 collocates as individual items in the corpus. For some collocates (couple,
series, pair, lot), occurrences within the a __ of framework accounted for more than
half of the total occurrences in the corpus data. The framework an __ of was shown
to be important in the area of nominalisation: an extension of, an indication of, an el
ement of, an instrument of all account for a considerable proportion (15–21%) of the
total occurrences of these words in the corpus. In a more extended context these com
binations form part of a construction of the type an N1 of (det) (adj) N2 , and a close
examination of the N1 items showed that they fell into two broad classes: in examples
such as an ounce of jam it is argued that the second noun is the functional head of the
whole, while in combinations such as an agent of destruction neither noun appears to
be dominant over the other. In the first of these groups, N1 can indicate measurement
Formulaic language 
of N2 (e.g. ounce), focus on a part or attribute of N2 (e.g. edge, index), or support (e.g.
example). In the second group, the two nouns are in a propositional relationship, for
instance through nominalisation (e.g. extension, explanation).
Both be __ to and too __ to attract adjectival collocates, but only one item (ready) is
shared between the lists of the top 20 collocates for the two frameworks. Furthermore,
while several collocates of be __ to are verbal (e.g. allowed, expected, said), only one of
those for too __ to belongs in this category (tired).
Renouf & Sinclair also discuss the collocates for the frameworks many/had/for
__ of, the overlap between the sets of collocates in written and spoken corpora,
and the balance of literal and figurative meanings of triplets such as an injection of
(mainly literal), an army of (equally literal and figurative) and an avalanche of (mainly
figurative).
Further work on this area has been carried out by Altenberg and his colleagues.
Eeg-Olofsson & Altenberg (1994) point out that an investigation of discontinuous
frameworks can bring to light collocational patterns, such as as far as X BE concerned,
which would be missed if only continuous sequences were isolated. These authors set
out to discover a number of features of collocational frameworks: which frameworks
are particularly common in spoken English (as indicated by a study of the London-
Lund Corpus); what kinds of item tend to occur in such frames; how frames vary in
their productivity, internal variability and statistical distinctiveness; and what differ
ences can be found when framework patterns are extended to include more than one
word sandwiched between initial and final items.
In order to capture the maximum number of generalisations regarding the prop
erties of discontinuous sequences, Eeg-Olofsson & Altenberg isolated all ‘X1Y’ triplets
and ‘X2Y’ quadruplets from the corpus, and then applied three measures to the 1,000
most frequent frames of each type: type/token ratio (TTR), as a measure of the inter
nal variability of frames; a measure known as entropy, which indicates the degree of
predictability of the filler; and the mutual information (MI) score, which compares
the frequency of a frame with that predicted by a random distribution, and so acts as
a measure of statistical distinctiveness.
X1Y frames typically consisted of closed class, grammatical items with an open
class word between them. Their internal variability showed a wide range, from a
type/token ratio of 0.98 for highly variable frames such as the __ well to 0.005 for some
frames, such as going __ be, which had only one type of filler (in this case to). The
entropy values again showed a considerable range, those with high entropy usually
consisting of grammatical words with a wide range of open-class fillers (e.g. the __ of ),
while frames with unique fillers had zero entropy. Among the low entropy, high fre
quency frames are some which constitute single phrasal units (e.g. quite [a] lot, in [the]
middle, and others which often occur as ‘onsets’ in the clause (see §2.1), such as and
[of ] course, but [I] mean. Frames with high MI scores tend to be either unique occur
rences or fixed combinations such as Alice [in] Wonderland, suffers [from] constipation.
Among the most frequent frames, those with high MI values tend to form complete
constituents (e.g. point [of ] view, at [the] moment, very [much] indeed. Frames with
 Christopher S. Butler
high MI scores and low entropy are statistically distinctive, relatively fixed sequences,
while those with high MI but low entropy are distinctive but variable.
Eeg-Olofsson & Altenberg find that X2Y frames are, as expected, much more vari
able than X1Y. In the top 10 combinations, all the framing words are closed class items.
Again, a combination of MI and entropy scores permits a useful classification. High
MI, low entropy frames (i.e. distinctive frames with relatively predictable fillers) in
clude, for example, thank __ __ much, on __ __ hand, one __ __ things, and a noticeable
feature is that the framing words are often open class items, including general nouns
such as way, thing, side, time. Frames with high MI and high entropy (i.e. distinctive
and variable) include a number in which the framing words are prepositions (e.g. in
__ __ of, on __ __ of ), and the full sequence can often be seen as a complex preposition
with spatial or temporal meaning (e.g. in the middle of, in the beginning of ).
Given the wealth of information available on discontinuous sequences in English,
it would be of great interest to investigate this type of phenomenon in a range of other
languages, preferably belonging to a range of typological families. The results of a pre
liminary investigation of corpora of Spanish are reported in Butler (1998b). Five small
corpora were used for this study, two consisting of spoken Spanish, one of written
(newspaper) language, and two of interviews transcribed for a magazine. 28 frame
works were investigated, consisting of pairs drawn from a set of words comprising the
articles un, una, el, la and the frequent prepositions a, de, en and por. For each frame,
three statistics were calculated: the token/type ratio (i.e. the inverse of the type/token
ratio used in the Renouf & Sinclair study); the ratio of the observed frequency of the
frame to the frequency expected for a random distribution in the corpus (a measure
which is related in a simple mathematical way to the MI score); the T-score, which
measures the absolute frequency of collocations.2
It was found that the ratio of observed to expected frequencies was much greater
than unity only for the four frameworks consisting of an article as the first framing
word and the preposition de (or the form del in which it is contracted with the article
el) as the second. The use of correlational techniques showed a high degree of agree
ment across the different corpora with respect to the observed/expected ratio, though
clear sensitivity to the spoken/written dimension was shown by the fact that the high
est correlations were for the two spoken corpora, and also the two interview corpora,
while the lowest was between one of the spoken corpora and the written corpus. An
almost identical pattern of correlation was obtained for the token/type ratios. Overall,
T-values were found to correlate very highly with observed/expected values, and less,
though still very significantly, with the token/type ratio.
A detailed study of the ‘article __ de[l]’ frameworks revealed very strong similar
ities to the patterns found for English. As in Renouf & Sinclair’s study, constructions
of the form ‘article N1 de N2 ’ were of two types, one in which the second noun was
semantically dominant and another in which the two nouns shared semantic salience.
Again as in the English study, the frequent collocates could be classified into a number
of semantic groups: for un __ de[l] these were measure (with subgroupings of quantity
(e.g. millón ‘million’, poco ‘little’), increase (e.g. incremento, aumento ‘increase’, cre
Formulaic language 
cimiento ‘growth’), collectivity (e.g. grupo ‘group’, equipo ‘team’), and time (e.g. período
‘period’)), kind (e.g. tipo ‘type, kind’), process (e.g. proceso ‘process’, programa ‘pro
gramme’), matter (e.g. problema ‘problem’, tema ‘topic’), human (e.g. hombre ‘man’,
presidente ‘president’), together with a small residue which could not easily be fitted
into these groups. The una __ de[l] framework showed a similar set of groupings.
The frames consisting of a definite article and de[l], on the other hand, selected
a wider range of collocates, falling into not only the groupings found for the indefi
nite article frames, but also additional categories such as organisations (e.g. ministe
rio ‘ministry’, banco ‘bank’, partido ‘(political) party’), locations (e.g. centro ‘centre’),
modal concepts (e.g. posibilidad ‘possibility’, necesidad ‘necessity’) and existence or the
lack of it (e.g. presencia ‘presence’, falta ‘lack’). There was a clear tendency for the
collocates in such frameworks to refer to features of everyday life (e.g. puerta ‘door’,
iglesia ‘church’, calle ‘street’) in the spoken corpora, but to abstractions (e.g. obligación
‘obligation’, creación ‘creation’) in the written and interview corpora. There was also a
noticeable tendency for the abstract nouns to represent nominalisations of processes
(e.g. llegada ‘arrival’, descripción ‘description’, instalación ‘installation’, evolución ‘evo
lution, change’). These nominalisations were found in both the newspaper corpus and
the transcribed interviews, but not in the spoken corpora, and were mostly formed
from transitive verbs. Further, they occurred overwhelmingly with definite rather than
indefinite articles as the first word of the frame.
One clear contrast between English and Spanish is the very frequent use of ‘N1
de N2 ’ constructions with N1 as the semantic as well as the syntactic head. This sit
uation arises not only because of differences between of and de, but also because
Spanish grammar does not permit the compression of information into noun phrases
by means of the modification of one noun by another: compare a press conference with
una conferencia de prensa, or a market economy with un mercado de economía.
Frameworks consisting of ‘article __ a[l]’ and ‘de/a __ article’ were found to be
unimportant in terms of both their frequency and their selectivity. Of the frame
works consisting of two prepositions, five showed medium to high selectivity, and
clear semantic groupings could be discerned for some of these. For instance, for the
framework en __ de, we can recognise categories of replacement (en vez de, en lugar de,
‘instead of ’), support or opposition (en favor de ‘in favour of ’, en contra de ‘against’)
and matter (en cuestión de ‘in a matter of ’, en materia de ‘with regard to’, en caso de ‘in
(the) case of ’). For por __ de we find collocates concerned with measure (por ciento de
‘per cent of ’), behalf (por parte de ‘on behalf of ’), vertical position (por encima de ‘over
the top of, above’), por debajo de ‘below’) negative features (por falta de ‘through lack
of ’, por culpa de ‘through the fault of ’).
. Widening the scope of the study: More flexible syntagmatic patterns
in English, Spanish and Italian
The approach discussed so far is based on purely automatic extraction of recurrent

strings, whether continuous or discontinuous. This method has the advantage that
 Christopher S. Butler
it is quite independent of the fallibilities of human searching, though we have seen

that subjective decisions are involved in deciding on cut-off points for frequency. The
completely automated technique, however, suffers from a serious disadvantage, in that
it brings to light only those strings which are repeated in exactly the same form in
the corpus materials. In other words, it does not reveal the large number of recurrent
syntagmatic patterns which show an often complex mixture of fixity and variability.3
In order to isolate these, we need a combination of automatic and manual analysis:
concordances are produced for particular ‘node’ items, sorted in ways which tend to
reveal syntagmatic patterns, and then inspected manually. This technique, in turn, suf
fers from the disadvantage that we can only examine one (set of) item(s) at a time, and
must decide in advance what items we are going to look at. This introduces a clear el
ement of subjectivity, and also means that we focus on just one quite small part of the
lexicogrammar at a time, with no guarantee that the sum total of various investigations
will add up to a coherent overall picture.
Investigations of recurrent syntagmatic patterns in corpora have focused on the
phenomena of collocation (lexical co-occurrence) and colligation (originally defined by
Firth (in Palmer 1968: 183) in terms of co-occurrence of grammatical categories, but
nowadays often interpreted as relationships between such categories and particular
items), and have led to the postulation of extended units of meaning (see Sinclair 1996,
1998; Tognini-Bonelli 2001).
We may illustrate this concept by means of a particular type of patterning into
which the word form untrained enters in English. This word form occurs 143 times
in the British National Corpus, and if we look at the pattern of collocates provided by
the corpus search program WordSmith Tools, we see that the most frequent open class
collocate is eye, occurring 11 times, 10 of these occurrences being at the position to
the right of the node untrained, one at the next position to the right. Examination of
the concordance for untrained allows us to see further detail of the patterning involved
(see Figure 1).
8 of the occurrences are in the form to the untrained eye, where the definite article
is being used generically. The other 2 are with my untrained eye and by an untrained
eye. The final occurrence has the untrained eye in subject position in a clause. There
is also an instance with the plural eyes: to her untrained eyes. There is normally some
indication, in the immediate context, of the difficulty of perceiving what something is
really like (e.g. it is not clear, especially to the untrained eye; the untrained eye from off the
street can not apply the classification . . .; not visible to the untrained eye; not immediately
noticed by an untrained eye; they are few and far between to the untrained eye for most
of the best features are hidden . . .). Two of the examples use the verb look, as in the two
products look remarkably similar, suggesting that things are not necessarily what they
seem to be. Finally, we may note that there are also two examples where the noun after
untrained is ears rather than eye(s), and that both of these also contain the verb sound
(the results sound pretty good to my untrained ears; the recording sounded terrible to his
untrained ears), which is the auditory equivalent of visual look.
Formulaic language 
We may thus postulate a general schema, of which particular groups of forms are
more specific sub-types. The general pattern is shown in (1):
(1) perceivability + preposition + article/possessive + untrained + sense organ
(visual/auditory) + difficulty of perceiving accurately
The core of the extended item is untrained eye(s)/ears, and is itself semantically trans
parent. The semantic preference is for words indicating perceivability, either visual or
auditory, and there is also a semantic prosody which colours the meaning of the whole
item, in this case the meaning of ‘difficulty’. The most frequent pattern is concerned
with visibility and uses the preposition to and the definite article: to the untrained eye.
The whole extended unit is very similar in its behaviour to that which has as its core
the combination naked eye, which differs from untrained eye in that its meaning is
not predictable from the meanings of its components, and for which Sinclair (1996)
proposes the pattern in (2), again with an additional semantic prosody of difficulty:
(2) visibility + preposition + the + naked + eye
Figure 1. Part of concordance for untrained in the British National Corpus

 Christopher S. Butler
Many more examples of semantic preferences and prosodies can be found in the afore
mentioned publications by Sinclair and Tognini-Bonelli, and also in discussions by
Stubbs (1995, 1996, 2001), who develops these concepts in terms of the light they can
shed on ‘keywords’ in different cultures. Such extended units of meaning can cause
considerable problems for translators and for learners of a language as L2. For in
stance, there is, to my knowledge, no lexeme or extended item in Spanish which is
even roughly equivalent to to the untrained/naked eye, with its semantic preference
and prosody. The Collins Spanish dictionary suggests invisible a simple vista as a trans
Formulaic language 
lation of invisible to the naked eye, but examination of the phrase a simple vista in texts
quickly reveals that it is of much wider application than the English expression, often
being better translated by some circumlocution such as just by looking at it/him/etc.,
as is recognised in the Spanish-English section of the dictionary. For to the untrained
eye, the dictionary is forced back on the paraphrase para (el ojo de) quien no es experto
(literally, ‘for (the eye of) someone who is not an expert’), or para un lego en la materia
(‘for a novice in the subject’).
Tognini-Bonelli (2001: Chapter 7, 2002) looks at similarities and differences be
tween extended (or, using her label functionally complete) units of meaning in English
and Italian. For some units, the two languages show much the same kinds of pattern:
for example, both in case of and nel caso di collocate with items referring to disasters
of some kind, and show an overall negative semantic prosody which can be charac
terised as ‘provision for disaster’. In other cases, however, the correspondences are by
no means complete. To take just one example, the “focusing” function of the English
adjective real, in the construction “the/possessive + real + noun”, indicates a value judg
ment about other instances of the entity referred to by the noun, implying that they are
less important in some way: the real problem is in contrast to some other problem(s)
deemed less worthy of attention. This is also true of the Italian adjective vero when
used prenominally. However, the English word shows a semantic preference for items
concerned with discourse (problem, question, reason, etc.), while the Italian word does
not show this preference, collocating with items such as nome (name), artista (artist),
centro (centre), etc.
. Overall conclusions from corpus work: The idiom principle
It is clear that corpus studies have demonstrated beyond all reasonable doubt the quan
titative importance of recurrent syntagmatic patterns in language. Indeed, Sinclair
(1991) concludes that such patterns are so prevalent that they necessitate reassess
ment of the normal assumption of grammarians, namely that language, by and large,
obeys the “open choice” principle according to which lexical items occur in slots pro
vided by the grammar, choice being determined by syntactic properties and by such
constraints as semantic selection restrictions. Such a view ignores the importance of
co-selection of items in syntagmatic patterns, a phenomenon which is recognised in the
idiom principle:
The principle of idiom is that a language user has available to him or her a large
number of semi-preconstructed phrases that constitute single choices even though
they might appear to be analysable into segments. (Sinclair 1991: 110)
Note that we are not talking here just about idioms in the sense of combinations
of words whose meaning is not predictable from the meanings of the components.
Sinclair’s contention is that many ‘semi-preconstructed phrases’ are capable of being
generated by the normal grammar, but that in our actual use of language, they are not
generated afresh every time, but are available for use as single items. Furthermore, pro
 Christopher S. Butler
cessing according to the idiom principle is seen as the norm, rather than as a peripheral
phenomenon.
Sinclair contends that the producer of a text alternates between the idiom princi
ple and the open choice principle, the former predominating. Erman & Warren (2000)
take this alternation as the topic of an article in which they aim to assess “the impact
that prefabricated language has on the production and structure of the text, whether
written or spoken” (p. 30). The main contribution of the study, according to the au
thors themselves, is to put forward a scheme for textual analysis “with the objective
of making the alternation between prefabricated and non-prefabricated combinations
manifest” (p. 30). Prefabs are defined in this study as follows:
A prefab is a combination of at least two words favored by native speakers in pref

erence to an alternative combination which could have been equivalent had there
been no conventionalization. (p. 31)
The criterion most used in the study to identify prefabs is thus that of “restricted ex
changeability”, by which is meant “that at least one member of the prefab cannot be
replaced by a synonymous item without causing a change of meaning, of function
and/or idiomaticity” (p. 32). Clearly, this criterion involves a great deal of subjective
decision-making on the part of the analyst, and the authors candidly admit that “[t]he
identification of prefabs is difficult” (p. 33, emphasis in original), not least because
conventionalisation is a gradual rather than an all-or-none process, so that what counts
as a prefab in the usage of some members of a linguistic community may not count as
such for other members.
Erman & Warren analyse nineteen text extracts differing in length from 100 to
800 words, taken from the London-Lund Corpus of spoken British English and the
Lancaster-Oslo-Bergen Corpus of written British English, together with a portion of
each of two versions of a children’s story. They develop a simple visual notation to
indicate which words in a text belong to prefabs and which do not, so indicating the
transitions between parts of the text selected according to the idiom principle and
those created on the basis of the open choice principle. They conclude that on average
about 55% of text is composed of prefabs, the proportion being rather higher for spo
ken language (59%) than for written (52%). This latter finding is in accordance with
the generalisation made in §2.1 of the present article, as are the claims that prefabs
often have a core with optional extensions, and that they can combine into compos
ite prefabs. There is, however, a clear difference in the distribution of functional types
of prefab across the spoken and written modes: in written English, prefabs with a pre
dominantly lexical, referential function are in the great majority (71.5% of all prefabs),
with smaller proportions of those with grammatical (16.9%) and pragmatic (2.4%)
functions, and of “reducibles” such as it’s, I’m (9.2%); in the spoken language, on the
other hand, the distribution is rather more even, with just 38.8% of lexical prefabs,
16.9% of the grammatical type, and much higher proportions of pragmatic prefabs
(16.7%) and reducibles (24.0%) than in the written corpus.4
Formulaic language 
. A psycholinguistic approach to formulaicity
We now turn, more briefly, to a rather different angle on formulaicity. Wray (1999,
2000, 2002; Wray & Perkins 2000) has developed a model of formulaic language
which is derived from detailed searches of the literature on both ‘normal’ and apha
sic adult language, and on the learning of first and second languages. The following
discussion is based on Wray (2002), and is intended to highlight the similarities and
differences between her approach and the corpus-oriented work of Sinclair and others
discussed earlier.
Wray begins by attempting to get a grip on what formulaic sequences are. She
develops an account which, like that of Sinclair, proposes a dual-systems approach to
the handling of linguistic material by speakers and hearers: analytical processing uses
grammatical rules to create or decode language which is (at least potentially) novel,
whereas holistic processing involves the retrieval of prefabricated combinations stored
whole in memory.5 Again like Sinclair, Wray stresses that holistic processing is not
concerned only with those strings traditionally regarded as idioms, but with many
combinations which could be produced or understood by the analytical grammar. The
advantage of holistic processing is that it reduces the processing effort required, while
analytical processing allows us to produce and understand novel sequences when this
is necessary.
Wray defines a formulaic sequence as follows:
a sequence, continuous or discontinuous, of words or other elements, which is, or
appears to be, prefabricated: that is, stored and retrieved whole from memory at
the time of use, rather than being subject to generation or analysis by the language
grammar. (Wray 2002: 9)
Note that the definition is a psycholinguistic one, emphasising processing aspects,

but that in other respects it strongly resembles Sinclair’s statements about the idiom
principle, derived from corpus analysis.
Wray also devotes considerable attention to the important question of how we
might try to recognise formulaic sequences in texts. This is not an issue which comes
to the fore in corpus-based approaches, which tend to assume that any frequently re
peated word sequence is formulaic, in the sense of being stored and retrieved whole.
While recognising what has been achieved through corpus studies, Wray points to
some of the problems with such approaches, in particular the heavy reliance on fre
quency as a criterion for deciding on the importance of putatively formulaic strings,
and the subjective decisions which have to be made in terms of cut-off points. Equally,
there are problems with the intuitive identification of formulaic sequences by in
formants. Wray also surveys attempts to identify such sequences by means of their
internal form and phonological structure, and concludes that no single criterion is ro
bust enough to offer a reliable indicator of formulaicity. Superimposition of continua
based on the form, function, meaning and provenance of strings will not generally
lead to convergence, since strings which are formally similar may have quite different
 Christopher S. Butler
functions, and may also be dissimilar in terms of their degree of semantic composi
tionality. Wray’s approach therefore accepts that “formulaic sequences are not a single
and unified phenomenon” (2002: 66).
In her quest for a model which does justice to the diverse properties of putatively
formulaic strings as well as to their usefulness, Wray surveys the patterns of formulaic
ity in a range of text types taken from ‘normal’ adult language. A number of roles for
formulaic language are identified: formulaic sequences enable the speaker not only to
reduce the processing load and achieve greater fluency, but also to express his or her
individual identity and group membership, so performing an important social func
tion. This analysis allows Wray to develop a model of normal adult language in which
formulaicity is central:
it is the accessing of large prefabricated chunks, and not the formulation and anal
ysis of novel strings, that predominates in normal language processing.
(Wray 2002: 101)
In other words, analytic processing comes into play only when the formulaic system
fails to cope: it is used on a “needs only” basis. In this, as in many other respects,
Wray’s model is in agreement with the corpus-based model of Sinclair. Her approach
is, however, much more geared to the psycholinguistic and sociolinguistic advantages
conferred by formulaic language. In order to explain the distribution of novel and
prefabricated language in actual text, Wray proposes that the various roles played by
formulaic sequences can be reduced to three basic functions: reduction of the pro
cessing burden on the speaker, manipulating the hearer and his or her picture of the
speaker’s identity, and the management of the structure of the discourse. All three
functions can be seen as aspects of the speaker’s need for self-promotion: even the
function of formulaic language in easing the hearer’s comprehension burden is ulti
mately beneficial to the speaker.6 In other words, formulaic sequences are unified, not
in terms of their linguistic form or meaning, but in terms of the psychological and
social advantages they confer. Wray points out that when viewed in this way, formu
laic sequences would not be expected to be a fixed, homogeneous store for a given
individual, but rather would vary dynamically with differing needs.
With this model of adult formulaicity established, Wray goes on to examine in
detail the evidence for the role of formulaic sequences in first language acquisition and
to develop a model based on this information, which emphasises the restricted nature
of analytical processing and prioritises the principle that such processing occurs on
a needs-only basis. The changing balance of holistic and analytical processing during
language development is also addressed.
A further issue addressed in Wray’s account is second language learning by chil
dren, teenagers and adults. This is clearly an area of particular interest from the cross-
linguistic perspective, and I shall say a little more about it in §4.4. The model developed
emphasises the relationship between language processing systems and the needs and
goals of learners at different stages of life. Young children learning a second language
are in a similar position, in many respects, to those learning their native language: both
Formulaic language 
Table 1. Wray’s heteromorphic distributed lexicon model
LEXICON 1 LEXICON II LEXICON III LEXICON IV LEXICON V
Grammatical Referential Interactional Memorised Reflexive items

items items (routine) items
items
Formulaic in order to, on half past + great to see multiplication goodness
word account of NUM (1–12), you, the most tables, songs gracious! what
strings give NP to NP important the!
thing is
Formulaic because, established, alright, maybe acronyms goodness!
words towards kindness
Morphemes the, -able, un- help, match hey! hallo! blast! bother!
need to manage their own environment and establish their own roles within it. On the
other hand, the post-childhood language learner has a more complex, even conflicting
set of priorities, among which is the desire to achieve control over the language ma
terial they manipulate, often leading to a predilection for analytical processing. Thus
the teenage or adult second language learner’s expectations regarding the most impor
tant size of unit for processing differ from that of the native speaker: the former tend
to operate with smaller, analysed segments, while the latter favour larger, holistically
stored units.
The final part of Wray’s data analysis is concerned with aphasic language. In order
to account for the properties of such disordered language, Wray postulates a dis
tributed lexicon model with five functionally distinct components, each with different
linguistic and communicative effects. Each of the five components stores, in a holistic
fashion, linguistic units of any size, from morpheme to word string. Two components
are associated with left hemisphere brain function, two with the right hemisphere, and
the last with subcortical processing.
This Heteromorphic Distributed Lexicon forms a central part of the integrated
model of formulaicity which Wray proposes on the basis of all her models of specific
types of language. The summary in Table 1 gives examples of the kinds of item found
in each component, and is based on Wray’s Figure 14.1 (2002: 263). Note that even
morphemic combinations can be formulaic, according to Wray’s model. Note also that
certain strings may be available both formulaically, as wholes, and as sequences gener
ated by the open choice grammar (e.g. mind your tongue! is formulaic when it means
‘be careful what you say’, but not if it is used by, for example, a dentist to warn someone
to keep his or her tongue out of the way).
In summary, then, Wray’s overall model can be characterised as psycho-socially
based: the fluency and reduction of processing cost associated with formulaic language
are beneficial to both speaker and addressee, and ultimately serve the speaker’s self-
promotion. The overall function of formulaicity is thus that it
 Christopher S. Butler
bridges the gap between novelty and routine, and makes it possible for us to pro
tect our own interests by producing language that is fluent and easily understood.
(Wray 2002: 281)
Further psycholinguistic evidence for the holistic storage of prefabs, recognised accord
ing to the criteria made explicit in Erman & Warren (2000), is adduced by Erman (in
preparation), who finds that in spoken English, pausing is rare between the compo
nent parts of such sequences. Additional support is provided in recent work by Wray
herself. Wray (2004) reports the results of a study based on a S4C (Welsh Channel 4)
television programme, Welsh in a Week, in which learners with little or no previous
knowledge of Welsh have to communicate in the language, in a specific situation, af
ter only four days of tuition. Because of the extremely short time available, the input
is largely in the form of relevant strings of items, presented as wholes, which are to
be memorised and used by the learner in their final task. In the study of one specific
learner, Wray was able to show that both pauses and errors occurred much more fre
quently at boundaries between memorised strings than at non-boundary positions,
the differences being highly significant statistically.
. Some implications for language teaching and learning
Finally, I shall look very briefly at what the work surveyed here can tell us in relation
to the teaching and learning of second and foreign languages.
. The need to take syntagmatic patterns into account when thinking about
vocabulary acquisition
One extremely important lesson to be learned from the phenomena reported on here
is that vocabulary acquisition should not be viewed solely, or even principally, in terms
of the learning of individual lexemes. We have shown that words very frequently enter
into ‘extended units of meaning’, and this entails that language users do not select
words individually, but rather in the context of the extended sequences into which
they enter. Furthermore, it is an important over-simplification to think in terms of the
acquisition of lexemes, as if all forms of a lexeme behaved the same way – they clearly
do not. An interesting study of formulaic sequences in the written language of Swedish
learners of English can be found in Wiktorsson (2002).
. The complex interplay of grammar and lexis
It is equally dangerous to see the acquisition of vocabulary as quite different from the
acquisition of grammar. We have seen that lexis and grammar mingle very closely in
the specification of the extended units of meaning which are arguably the main vehicles
of our expression and understanding.
Formulaic language 
. Differences in collocational behaviour of apparent translation equivalents

across languages
It has been shown that expressions which might be thought of as translation equivalents
across languages may well differ when the subtleties of collocational behaviour are
included in the picture. It is clearly important, for native-like fluency in a language,
that such distinctions be learned correctly.
. Formulaic language in relation to psychological and social factors

in language learning
As demonstrated by Wray (1999, 2000, 2002, 2004), formulaic sequences do indeed of
fer support in the initial stages of L2 acquisition for most types of learner, by increasing
fluency and facilitating comprehension by the hearer, so contributing to the individ
ual’s promotion of self. This is particularly clear in the case of young children, who
readily imitate the formulaic material they hear around them, in order to carry out
speech acts which manipulate others and establish their own identity. The position is
rather different for adults learning the L2 naturalistically, who show considerable vari
ation in the use of formulaic material; furthermore, formulaic sequences do not appear
to facilitate the learning of grammatical forms. Indeed, adults may, in responding to
the urge to communicate, fossilise any useful string of items, whether native-like or
not. When teenage or adult learners of L2 are taught in the classroom environment,
they are able to apply to holistically learned sequences the analytical techniques they
learn in class. At an early stage in classroom-based learning, formulaic sequences are
found useful in enabling communication and building confidence, but later on they
tend to lag behind expectations. Although the balance between formulaicity and cre
ativity is a crucial one for native-like performance, taught adult learners tend to use
too much language which is created afresh.
. Conclusion
In this article I have attempted to summarise some important recent work on formu
laic language, with a particular slant towards cross-linguistic studies. Corpus-based
studies have demonstrated very clearly the importance of multi-item strings in large
bodies of texts, and have also shown that different types of text vary both quantita
tively, in the overall frequencies of such strings, and qualitatively, in terms of the kinds
of strings used. Particularly important, in my view, is work which has isolated extended
units of meaning from texts, characterised in terms of the concepts of semantic pref
erence and semantic/discourse prosody. Especially crucial from the point of view of
the contrastive perspective adopted in this paper is the need to establish interlinguistic
relationships in relation to such extended units rather than to individual words, and
 Christopher S. Butler
to recognise the often subtle but important differences between languages in terms of
semantic preference and prosody.
While corpus linguists tend to assume that frequently recurring combinations of
items are stored and retrieved holistically (Sinclair’s idiom principle), linguists with a
more psycholinguistic orientation have been concerned to demonstrate this principle
at work in the actual processing of language, in normal adult and aphasic communi
cation and during the process of language acquisition and L2 learning.
The important consequence of all this work is that it necessitates a revision of our
usual ways of thinking about languages, both individually and comparatively. It is no
longer realistic to operate with a neat distinction between grammar and lexis, or with a
model which prioritises single-word lexemes and the way they fit into separately spec
ified constituent structures. Rather, we must recognise that extended units of meaning
can involve semantically related sets of words, individual lexical items, grammati
cal patterns, and semantic/pragmatic ‘colouring’ across the whole unit. In processing
terms too, we must be able to accommodate a holistic mode which supplements, and
frequently predominates over, analytical processing.
Finally, these conclusions have important implications not only for cross-linguistic
studies as such but also for the teaching and learning of languages.
Notes
* The work reported here was supported by the project Análisis del discurso en lengua inglesa:
aspectos cognitivos, contrastivos y de adquisición (‘Discourse Analysis in English: Aspects of cog
nition, typology and (L2) acquisition’), financed by the Spanish Ministry of Education, the
Fondo Europeo de Desarrollo Regional (FEDER), and the Xunta de Galicia (grant numbers
BFF2002-02441, PGIDIT03PXIC20403PN).
. For other works in which Halliday’s categories are used to classify multi-word units, see Moon
(1994, 1998), Culpeper & Kytö (2002).
. For a discussion of various measures of collocational attraction see Stubbs (1995).
. If the sequences are discontinuous, there can, as we have seen, be considerable variation in
the intervening items. Nevertheless, the items forming the outer limits of the string are still fixed.
. For details of the various functional types and their possible subclassification, together with
ample exemplification of each, the original article should be consulted.
. As Alison Wray (personal communication) has pointed out to me, a slightly different way of
looking at this is that language processing always involves the combination of units by rule, but
that the sizes of the units differ. Furthermore, some of the units are internally complex, and may
involve gaps, so that some of the rules are oriented towards the filling of these gaps.
. The function of easing the comprehension burden on the hearer could perhaps be seen as a
specific manifestation of the conversational maxim of manner. I am grateful to an anonymous
reviewer for this suggestion.
Formulaic language 
References
Altenberg, Bengt (1998). “On the phraseology of spoken English: the evidence of recurrent
word combinations”. In A. P. Cowie (Ed.), Phraseology (pp. 101–122) [Oxford Studies in
Lexicography and Lexicology]. Oxford: Oxford University Press.
Altenberg, Bengt & Mats Eeg-Olofsson (1997). “The collocability of nouns in the London-Lund
Corpus”. In U. Fries, V. Müller, & P. Schneider (Eds.), From Ælfric to the New York Times:
Studies in English Corpus Linguistics (pp. 79–91) [Language and Computers 19]. Amsterdam
and Atlanta: Rodopi.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, & Edward Finegan (1999).
Longman Grammar of Spoken and Written English. London: Pearson Education.
Butler, Christopher S. (1997). “Repeated word combinations in spoken and written texts: some
implications for Functional Grammar”. In C. S. Butler, J. H. Connolly, R. A. Gatward, &
R. M. Vismans (Eds.), A Fund of Ideas: Recent Developments in Functional Grammar (pp.
60–77) [Studies in Language and Language Use 31]. Amsterdam: IFOTT.
Butler, Christopher S. (1998a). “Multi-word lexical phenomena in Functional Grammar”. In F.
J. Cortés Rodríguez (Ed.), Lexical Studies towards the Year 2000 (pp. 13–36) [Monographic
section of Revista Canaria de Estudios Ingleses 36]. La Laguna, Tenerife: University of La
Laguna.
Butler, Christopher S. (1998b). “Collocational frameworks in Spanish”. International Journal of
Corpus Linguistics, 3(1), 1–32.
Cowie, Anthony P. (Ed.). (1998). Phraseology [Oxford Studies in Lexicography and Lexicology].
Culpeper, Jonathan & Merja Kytö (2002). “Lexical bundles in Early Modern English dialogues.
A window into the speech-related language of the past”. In T. Fanego, B. Méndez-Naya, & E.
Seoane (Eds.), Sounds, Words, Texts and Change. (Selected Papers from 11 ICEHL. Santiago
de Compostela, 7–11 September 2000) (pp. 45–63). Amsterdam and Philadelphia: John
Benjamins.
Eeg-Olofsson, Mats & Bengt Altenberg (1994). “Discontinuous recurrent word combinations in
the London-Lund Corpus”. In U. Fries, G. Tottie, & P. Schneider (Eds.), Creating and Using
English Language Corpora: Papers from the Fourteenth International Conference on English
Language Research on Computerized Corpora (Zürich 1993) (pp. 63–77). Amsterdam and
Atlanta: Rodopi.
Eeg-Olofsson, Mats & Bengt Altenberg (1996). “Recurrent word combinations in the London-
Lund Corpus: coverage and use for word-class tagging”. In C. E. Percy, C. F. Meyer, & I.
Lancashire (Eds.), Synchronic Corpus Linguistics: Papers from the Sixteenth International
Conference on English Language Research on Computerized Corpora (ICAME 16) (pp. 97–
107) [Language and Computers 16]. Amsterdam: Rodopi.
Erman, Britt (in preparation). “Pauses as evidence of cognitive effort in prefabricated and non-
prefabricated structures”.
Erman, Britt & Beatrice Warren (2000). “The idiom principle and the open choice principle”.
Text, 20(1), 29–62.
Halliday, Michael A. K. (1994). An Introduction to Functional Grammar (2nd ed.). London:
Edward Arnold.
Moon, Rosamund (1994). “The analysis of fixed expressions in text”. In M. Coulthard (Ed.),
Advances in Written Text Analysis (pp. 117–135). London and New York: Routledge.
 Christopher S. Butler
Moon, Rosamund (1998). Fixed Expressions and Idioms in English: A Corpus-based Approach
[Oxford Studies in Lexicography and Lexicology]. Oxford: Oxford University Press.
Palmer, Frank R. (1968). Selected Papers of J. R. Firth 1952–1959. London and Harlow: Longman.
Renouf, Antoinette & John M. Sinclair (1991). “Collocational frameworks in English”. In K.
Aijmer & B. Altenberg (Eds.), English Corpus Linguistics: Studies in Honour of Jan Svartvik
(pp. 128–143). London and New York: Longman.
Sinclair, John M. (1991). Corpus, Concordance, Collocation [Describing English Language
Series]. Oxford: Oxford University Press.
Sinclair, John M. (1996). “The search for units of meaning”. Textus, IX(1), 75–106. [Reprinted
in G. Corpas (2000) (Ed.), Las Lenguas de Europa: Estudios de Fraseología, Fraseografía y
Traducción (pp. 7–38). Granada: Comares].
Sinclair, John M. (1998). “The lexical item”. In E. Weigand (Ed.), Contrastive Lexical Semantics
Sinclair, John M. & Antoinette Renouf (1988). “A lexical syllabus for language learning”. In R. A.
Carter & M. McCarthy (Eds.), Vocabulary and Language Learning (pp. 140–160). Harlow:
Longman.
Stubbs, Michael (1995). “Collocations and semantic profiles: on the cause of the trouble with
quantitative studies”. Functions of Language, 2(1), 23–55.
Stubbs, Michael (1996). Text and Corpus Analysis: Computer-Assisted Studies of Language and
Culture. Oxford: Blackwell.
Stubbs, Michael (2001). Words and Phrases: Corpus Studies of Lexical Semantics. Oxford:
Blackwell.
Stubbs, Michael (2002). “Two quantitative methods of studying phraseology in English”.
International Journal of Corpus Linguistics, 7(2), 215–244.
Stubbs, Michael & Isabel Barth (2003). “Using recurrent phrases as text-type discriminators: a
quantitative method and some findings”. Functions of Language, 10(1), 61–104.
Tognini-Bonelli, Elena (2001). Corpus Linguistics at Work [Studies in Corpus Linguistics 6].
Tognini-Bonelli, Elena (2002). “Functionally complete units of meaning across English and
Italian: towards a corpus-driven approach”. In B. Altenberg & S. Granger (Eds.), Lexis
in Contrast: Corpus-Based Approaches (pp. 73–95). Amsterdam and Philadelphia: John
Benjamins.
Wiktorsson, Maria (2002). “Learning Idiomaticity: A Corpus-Based Study of Idiomatic
Expressions in Learners’ Written Production.” Ph.D. dissertation. Department of English,
Lund University.
Wray, Alison (1999). “Formulaic language in learners and native speakers”. Language Teaching,
32(1), 213–231.
Wray, Alison (2000). “Formulaic sequences in second language teaching: principle and practice”.
Applied Linguistics, 21(4), 463–489.
Press.
Wray, Alison (2004). “‘Here’s something I prepared earlier’: formulaic language learning on
television”. In N. Schmitt (Ed.), The Acquisition and Use of Formulaic Sequences (pp. 249–
268). Amsterdam and Philadelphia: John Benjamins.
Wray, Alison & Michael R. Perkins (2000). “The functions of formulaic language: an integrated
model”. Language and Communication, 20, 1–28.
A contrastive analysis of entrenchment
and collocational force in variable-sized
lexical units*
László I. Komlósi and Elisabeth Knipf
In this paper, we offer a unified treatment of free, semi-fixed and fixed

expressions by positing unitized representations for variable-sized lexical units.
We distinguish between rule-driven versus frequency-driven lexical structure and
default and open processing. In frequency-driven lexical structure the retrieval of
entrenched features is decisive, prompted by encoded, highly context-bound
parameters. This approach allows us to see formulaic expressions as entrenched
lexical units, stored in the lexicon as triples of phonological, syntactic and
conceptual structures <PS, SS, CS> often encoding additional pragmatic and
contextual information. The paper reports on semantic and pragmatic meanings
in collocations, constructions, clichés, formulaic expressions, idioms,
pragmatemes, figures and other metaphorical expressions. We show
cross-correlation between compositionality, productivity and the types of
processing in a parametrization scheme, illustrated by cross-linguistic examples of
iterative structures.
There are too many idioms and other fixed expressions for us to simply disregard
them as phenomena on the ‘margin of language’. (Jackendoff 1995: 156)
A strict separation of lexicon and phrasal grammar, like a strict separation of word
lexicon and idiom lists, may prove to be yet another methodological prejudice. I
find this an intriguing question for future research. (Jackendoff 1997: 176)
Our paper offers a cross-linguistic analysis of variable-sized lexical units residing in

the mental lexicon. On the basis of linguistic data from English, German, Russian and
Hungarian we propose to define lexical structure with the help of the diverse mani
festations of lexical form and its unitized representations involving expressions at the
collocational level, the idiom level and the conceptual-integration level. The objective
of the analysis is to justify and support a unified treatment of free phrases and phrasal
lexical items, more specifically that of semi-fixed and fixed expressions by acknowledg
 László I. Komlósi and Elisabeth Knipf
ing the relevance of the theory of lexical licensing for these expressions regarded in a
unified fashion as phrasemes or phrasal lexical items.
The starting point in the analysis is the observation of the distinction between
rule-driven and frequency-driven lexical structure. The former category is character
ized by the classical notion of Fregean compositionality, high degree of regularity and
predictability together with on-line processing. The latter category is characterized by
gestalt-like perception (as opposed to mere pattern recognition) where the retrieval
of entrenched features of unitized lexical items is prompted by the activation of en
coded, highly context-bound parameters together with partly default and partly open
processing. The other crucial distinction we observe in our analysis concerns the mean
ings encoded in and prompted by lexical form. We draw attention to the nature of the
contribution of semantic and pragmatic meaning in phrasemes such as collocations,
constructions, clichés, formulaic expressions, idioms, pragmatemes, figures and other
metaphorical expressions. A cross-correlation between compositionality, productivity
and the type of processing is proposed in the form of a parametrization scheme and a fur
ther elaboration on the correlation is given with the help of cross-linguistic examples
of iterative structures.
The results of cognitively oriented linguistics research suggest that we ought to
view the lexicon of any natural language as a necessary mental component used in cog
nition that comprises a list of the arbitrary matches between sound and meaning with
syntactic properties added. The lexicon is a collection of stored associations among
fragments of disparate representations. Thus, a word and its computable combinations
(see Busse 2002) is a way of associating units from distinct levels of representation,
namely from the phonological and the conceptual representational levels. Jackendoff
(1997: 107–108) argues that in the cases of diverse mental components of perception
(visual, auditory, tactile, olfactory, gustatory, etc.) such associations or mappings be
tween distinct levels of representation, and even between different perceptual modali
ties, help the processing agent to overcome extreme arbitrariness for the sake of more
effective cognition. The brain stores – in its long-term memory – large numbers of
pre-digested and pre-assembled mappings to use as shortcuts. The role of a lexicon in
processing lies in its functioning as an interface between two or more representations
it connects. Jackendoff ’s startling conclusion is that however unique the lexicon might
be in its size and arbitrariness, it shares a similar formal character with other mental
components.
In the present article we argue that linguistic structure, and lexical form specifi
cally, has ample means of encoding very different types of information – phonologi
cal, phonotactic, morphosyntacic, semantic, conceptual, contextual and pragmatic –
only to prompt the activation of appropriate representations, and through them, the
activation of meaning assignment via active meaning construction and conceptual
integration.
There has been a substantial shift in research focus concerning the relationship be
tween syntax and the lexicon in the past fifteen years. According to the received view of
autonomous syntax, syntactic structure is simply to be complemented by lexical inser
A contrastive analysis of entrenchment 
tion rules to take care of the place and function of lexical items in grammar. This view,
however, has been widely challenged (see e.g. Fillmore & Kay 1997; Fillmore, Kay, &
O’Connor 1988; Nunberg, Sag, & Wasow 1994; Jackendoff 1995, 1997; Goldberg 1995;
Kay 1997). Breaking away from the generative tradition, people began to give credit
to the fact that fixed expressions are ubiquitous in everyday discourse: there are many
thousands of formulaic phrases and expressions and a vast number of memorized fixed
expressions in our natural languages which by no means should be treated as marginal
in the overall system of grammar. These expressions need to be seen as lexical units
stored in the lexicon as triples of phonological, syntactic and conceptual structures
<PS, SS, CS> often encoding additional pragmatic information as well.
Jackendoff (1995, 1997) advances a novel treatment of phrasal lexical items (or
multi-word constructions) together with a theory of lexical licensing. He proposes a
unified treatment of both simple lexical entries and set expressions. He takes it as a
triviality that lexical entries come in a variety of sizes, thus lexically listed units are
often larger than words. The observation itself, of course, is not all that new. The phe
nomenon has been known in traditional descriptive grammars for a very long time;
however, the unified theoretical status and the subsequent theoretical incorporation
of these diverse types of constructs, i.e. of variable-sized lexical units, has been de
nied or avoided. The novel stance advocated by Cognitive Linguistics, Construction
Grammar and Corpus Linguistics attempts to regard lexical structure as motivated in
the light of the complexity of different types of information encoded in it and to see,
at the same time, a continuum rather than a sharp demarcation line between phrasal
grammar and the lexicon.
. Grammatical meaning, lexical meaning and lexical form
Regardless of its size and its components, lexical structure is characterized by uni
tized representation for its meaning. Despite the unitary meaning, however, lexical
semantics distinguishes two types of meanings derived from two different sources:
1. grammatical meaning is the sum total of the meanings of the constituent words in
a complex expression and the result of the way the constituents are combined in
the literal meaning (see the notion syntactic phrase discussed below); while
2. lexical meaning is a matter of listing standard meanings for simple lexical entries
or particular meaning properties for fixed expressions (see the notion of lexical
form below).
This is how we obtain compositional meaning structures at the syntactic level in the
case of syntactic phrases (cf. the notion of syntactically transparent semantic composition
in Jackendoff 1997: 48) and noncompositional meaning structure at some nonsyntac
tic level, such as the lexical level in the case of lexical form (cf. Frawley 2002: 228–230)
or the conceptual level in the case of enriched composition (cf. Jackendoff 1997: 49)
underlying fixed expressions.
 László I. Komlósi and Elisabeth Knipf
What, then, is the relationship between grammatical meaning and lexical mean
ing? Do they compete (since lexical listing of fixed expressions must override in whole
or in part the literal meanings of the constituent words and the way they are com
bined) or do they complement each other according to some principle of the division
of linguistic labor? Frawley (1992, and specifically 2002: 229) introduces the princi
ple of computational efficiency to explain the job of the two meaning components
discerned above as complementary.
(1) The principle of computational efficiency
“List the idiosyncrasy once and compute the derivations.”
The analysis of the interplay of the complementary functions of grammatical mean

ing and lexical meaning has given rise to different explanations as illustrated by the
well-tried examples of English compounds postman, garbageman, snowman below. In
the analysis of Croft (2000: 258, 262) the mismatch between lexical and grammatical
meaning and their formal correlates and the fuzziness of the distinction between lexical
and grammatical meaning gets explained by grammaticalization: a diachronic process
by which lexical meanings shift to grammatical meanings. In his example -man in post
man and garbageman can be both a root in the compound or a derivational affix. In the
analysis of Jackendoff (1997: 164) an instance of a productive lexical rule is witnessed
in snowman and garbageman: -man indeed behaves almost like a productive deriva
tional affix (see also -berry in blueberry, blackberry, strawberry, cranberry), however,
the difference between snowman and garbageman remains unpredictable until further
semantic specification is added.
(2) snowman → ‘a replica of a man made of snow’
(3) garbageman → ‘a man who removes garbage’
Jackendoff provides for the instructions: the italicized parts in the paraphrases above
must be listed in the lexical entry, above and beyond the meanings of the constituent
nouns. Examples (2)–(4) show the restricted scope of lexical rules: on the basis of the
English N-N compounding rule an interpretation given in (4) could be licensed and
acceptable too, since one can easily imagine a context where such an interpretation
would be plausible:
(4) snowman →‘a man who removes snow’
However, under normal circumstances the interpretation of (4) would be recognized
by speakers of English as a joke or a neologism since the standard meaning of snow
man in English as given in (2) is already assigned to it: the standard meaning is lexically
listed. “A lexical rule creates possibilities, but one must learn form by form which of
these actually exist” (Jackendoff 1997: 237). It should be noted here that the notion of
lexical complementation as discussed above involves both (i) lexical listing of idiosyn
crasies or idiosyncratic properties and (ii) applying lexical rules, exactly in the spirit of
the principle of computational efficiency. It should also be observed in this context that
A contrastive analysis of entrenchment 
lexical rules are dynamic elements of any rule-based (or rule-driven) grammar, while
lexical listing is a static element of a frequency-based (frequency driven) grammar.
From a cognitive point of view, lexical listings need to be learnt, stored and recalled
easily for smooth processing. Frequency in terms of frequent use of both formal and
conceptual associations and co-occurrences seems to be a corner-stone for memoriz
ing properties that are crucial for discerning and maintaining distinctiveness encoded
in linguistic structure. Frequency in use leads to entrenchment. We will argue that a
highly sophisticated feature of the language faculty is how the retrieval of entrenched
features of unitized lexical items gets prompted by language-specific and context-
sensitive parameters encoded in linguistic structure. We will elaborate in detail on such
specific aspects and the psycholinguistic consequences of entrenchment below.
Grammatical meanings are often claimed to represent a conceptualization or con
strual of the content of experience represented by the lexical meaning (cf. Langacker
1987; Talmy 2000). According to the conceptualization hypothesis meanings do not
directly reflect an objective reality, but rather establish and view reality from a par
ticular perspective. On the other hand, construal plays a significant role in lexical
choice as well: it is often the case that different speakers use different lexical items
to describe the same event, thus providing different conceptualizations of the event
experienced. We can easily see that both grammatical meaning and lexical meaning
involve conceptualization of the experience represented by linguistic means.
According to the encyclopedic view of lexical items, lexical meanings are encyclo
pedic in character in virtue of the lack of a sharp dividing line between the part of an
expression’s meaning that is strictly linguistic and the part that conveys nonlinguistic
knowledge of a concept (cf. Croft 2000; Langacker 1987). If the meaning of a lexical
item is taken to denote a concept, the characterization of lexical meaning has to be con
nected with the context provided by the background knowledge configuration, referred
to as a domain. A semantic domain (cf. Fillmore 1985; Langacker 1987) provides the
frame supporting the denotation of the concept in question.
The notion of regularity makes sense in both grammatical structure and lexical
structure. Although general statements about a grammar come in the form of rules,
particular statements about subregularities in the grammar come in the form of lists
(cf. Kiefer 2000). Kiefer points out that regularity/irregularity is not a dichotomy but
rather a scalar phenomenon. He claims that in natural languages there seems to be a
balance between regular, predictable forms (produced according to regular patterns)
and irregular, unpredictable forms (stored in the mental lexicon). This is likely to be
due to two conflicting principles: the principle of least effort striving toward grammat
ical regularity and the principle of iconicity striving toward easily recognizable forms
(cf. Kiefer 2000: 300). The degree and strength of recognizability greatly hinges on fre
quency. Thus, we can easily acknowledge that frequency and recognizability show a
close correlation with entrenchment.
At this point, however, we want to reflect on the theory of lexical licensing for
phrasal lexical items by way of a brief summary of the discussion so far. When we talk
about variable-sized lexical items that cover semi-fixed and fixed expressions in the lex
 László I. Komlósi and Elisabeth Knipf
icon, we regard them both as syntactic constructs compatible with grammatical rules
and as lexical items compatible with lexical rules. The latter are complete only with a
substantial amount of lexical listings: meaning properties of semantic, conceptual and
contextual-indexical kinds complement literal meanings. We come to appreciate the
nature of the lexicon as a mental component in cognition, a storehouse of associations
between disparate representations manifested in the form of syntactic structure, and
to understand that the lexicon fulfils its function as a conceptual necessity by serving
the needs of the computational mind at a surprising level of efficiency. It is the ex
istence of the lexically listed meaning-idiosyncrasies that licenses syntactic phrases. In
other words, syntactic phrases are enriched by additional specifications of a nonsyn
tactic nature provided by individual listing. It is with the help of such a licensing that
phrasal lexical items get ready to be incorporated in the grammar just as full syntactic
phrases do.
. Semantic meaning and pragmatic meaning
The effects and consequences of the semantic and the encyclopedic views on lexical
entries can be attested in the structuralist view of a language’s lexicon. According to
this idea a lexicon is a vast calculus of language internal grammatical and semantic
relations together with language external relations as reflected by speaker’s conceptu
alizations of the extralinguistic reality (cf. Taylor 2002: 297). If this is so we can only
agree to recent proposals concerning the nature of concept structure: many concepts
are seen to have a flexible and even temporary nature since the way an entity is to be
categorized on any specific occasion is very much a function of the concerns of the
speaker, the purpose of the communication, and the conceptual model constructed
and established by negotiation in prior linguistic acts. While some aspects of the mean
ing of linguistic expressions in a language are stable and exist as entrenched parts of a
person’s linguistic knowledge, the different uses of linguistic expressions activate dif
ferent aspects of domain-based knowledge decisive in dynamic meaning construction
(cf. Taylor 2002: 299).
Lexical semantics deals with the meanings of word-like expressions, though first
it has to identify the forms that carry the meanings. The widening research into the
nature of linguistic expressions has shown that a wide range of forms from simple
words to variable-sized, multi-word lexical units are eligible for being accounted for in
lexical semantics as they are formally discrete and semantically unified. For the purposes
of our analysis we have adopted the use the term lexical form on the basis of Jackendoff
(1995) and Frawley (2002), as discussed in detail in the previous section.
It is important to emphasize in the characterization of lexical form, however, that
(i) it is a unified triple of phonetic, syntactic and conceptual information and (ii) the
meaning carried by lexical form has to have some identifiable relation to the world.
The aboutness relation can be of two kinds: (i) it may describe states of affairs that
correspond to the real world, hence it will convey truth and (ii) it may construe worlds
A contrastive analysis of entrenchment 
by virtue of linguistic expression and hence it will convey force of action. Truth is the
domain of semantics, force is the domain of pragmatics. Let us recapitulate an example
from Frawley (2002: 229–230) which proves to be highly illuminating in bringing out
the crucial difference in the accessibility and activation of the two types of information.
The expression fire implies ‘light and heat from combustion’ to be associated with
some state of affairs in the world. Whenever there is fire, there is a necessary infer
ence to ‘light and heat from combustion’. This implication is technically known as an
entailment. Entailment is a necessary truth-conditional inference, established by the
given lexical form such that it reveals a literal information structure explicitly depict
ing the world irrespective of any particular context. The expression “Fire!” uttered in
a crowded public place, however, will bring about the force of a command in virtue of
construing a world in which some action is forced. We infer ‘danger’ from the expres
sion fire, however this inference is triggered irrespective of truth. The pragmatic infer
ence from fire to ‘danger’ is a non-truth-conditional inference called implicature. The
lexical form itself reveals a non-literal information structure implicitly prompting some
interpretation of projected mental states. An implicature is highly context-sensitive. It
is important to note here that lexical form can convey both semantic and pragmatic
meaning via different types of implication relations: entailments, presuppositions and
implicatures, respectively.
It seems that lexically encoded information provides for a vast potential in the
way of interpretability of linguistic expressions, however the activation of appropriate
pieces of information depends exclusively on licensing or triggering mechanisms. On
a general plane of discussion, it should be obvious that some inferential mechanism
must decide on the selection of standard (default) or intended (particular) meanings.
However, the selection of certain chunks or aspects of potential information often
precedes the actual implementation of inference making. If we acknowledge that im
plication relations in general depend on selectional choices for potential information,
we can understand better the mechanisms underlying the phenomena of conversa
tional implicatures, metaphorization, blending and conceptual integration. In order
to illustrate the mechanisms at hand, we first bring into the discussion a phenomenon
from a specific realm involved in the knowledge of language: presumptive inferencing
in linguistic pragmatics.
The question of getting intended meanings across in conversation is best exempli
fied by the dichotomy of generalized conversational implicatures (GCIs) and particu
larized conversational implicatures (PCIs) elaborated in Levinson (2000). Generalized
implicatures can be drawn with very little inside or local knowledge. They are closely
connected to the degree of informativeness that we normally expect a speaker’s utter
ance to provide due to the standard meanings of the linguistic expressions used in the
utterance. However, particularized implicatures require not only general knowledge
but also knowledge which is particular or local to the speaker and the hearer, and of
ten to the physical context of the utterance itself. Let us look at two examples from
Stilwell-Peccei (1999):
 László I. Komlósi and Elisabeth Knipf
(5) a. Annie: Was the dessert any good?

Mike: Annie, cherry pie is cherry pie. (No, the dessert was pretty boring.)
b. Extra information: Mike loves cherry pie. As far as he is concerned, no
one can ruin a cherry pie, and Annie knows this.
Annie: Was the dessert any good?
Mike: Annie, cherry pie is cherry pie. (Yes, as good as any other time.)
(6) a. Maggie: Coffee?
James: It would keep me awake all night. (I won’t have any coffee.)
b. Extra information: James has to stay up all night to study for an exam and
Maggie knows this.
Maggie: Coffee?
James: It would keep me awake all night. (Sure, I badly need it now.)
It is not infrequent that the implicational force is borne out by the linguistic context
which, consequently, can be responsible for completely different interpretations, as
in (7a–c):
(7) a. A: Do I make good coffee?
B: You make great coffee.
b. A: Do you think I’m a good cook?
c. A: It’s your turn to make the coffee.
In this paper our objective is to analyze variable-sized lexical units in the lexicon. We
believe, however, that the understanding of reasoning and inferencing mechanisms
underlying conversational implicatures may throw light on the mechanisms we ap
ply when we process the semi-compositional or non-compositional fixed expressions
which so often surface in discourse. We suggest that a question, complex as it might be,
that lies at the heart of inquiry for pragmatic linguistics when striving to find out the
meanings of highly contextualized utterances, seems to be relevant to the analysis of
variable-sized lexical units as well: Which pieces of linguistically relevant information
coded in linguistic expressions actually get activated by which prompts? It is very sug
gestive that lexical licensing of phrasal lexical items and presumptive inferencing have a
lot in common. One common feature is, no doubt, their high information-sensitivity.
The concluding thoughts in Levinson (2000: 371) seem to support our argument for
lexical licensing.
Generalized Conversational Implicature theory does suppose that there is a body

of knowledge and practice concerned with the use of language. This knowledge
crucially involves metalinguistic knowledge about the structure of the lexicon –
specifically, knowledge about the structuring of semantic fields, the availability
of alternate expressions, subjective assessments of frequency and markedness of
specific expressions, knowledge about the stereotypical associations of linguistic
A contrastive analysis of entrenchment 
concepts in the speech community, mutual assumptions of principles for resolving

conflicts between inferences, and so on.
We believe that the theory of lexical licensing of phrasal lexical items can reasonably
identify the role of this specific body of knowledge and practice involved in the use
aspects of our linguistic knowledge. It can acknowledge lexical items above the word
level to license syntactic structure which in turn incorporates different types of fixed
expressions such as collocations, compounds, clichés, idioms, constructions, figures
and metaphorical blends which all contribute in crucial ways to prompting conceptual
representations.
. Meaning extension and conceptual integration
We now proceed further with our analysis of the various aspects of lexical units. An
important aspect to discuss is their combinatorial potentials, which license meaning
integration processes (metonyms, metaphors, blends and other types of conceptual in
tegration (cf. e.g. Fauconnier & Turner 2002; Grady, Oakley, & Coulson 1999; Radden
2002; Talmy 2000; Turner & Fauconnier 1995)).
We claim that actualized interpretations of lexical entries are to be seen as results
of (a) retrieving stored meaning properties associated with lexical form and (b) in
ferring from diverse linguistic contexts with the help of conceptual (re)-organization.
Two current ideas concerning conceptual creativity triggered by the limited inventory
of grammatical and lexical forms will be drawn upon in the discussion to follow. The
first is Leonard Talmy’s idea of conceptual organization (Talmy 2000), the second one
is the idea of conceptual integration, the most explicit and elaborate version of which
to date has been proposed in Fauconnier & Turner (2002). The underlying idea of lin
guistic creativity shared by both Talmy’s mechanisms of concept structuring and the
mechanisms for conceptual integration proposed by Fauconnier & Turner is that our
capacity for language also depends on our ability to use a relatively limited inventory
of grammatical and lexical forms to prompt for virtually unlimited ranges of cognitive
representations. While lexical listing makes use of entrenched storage of idiosyncratic
meaning properties in long-term memory, conceptual integration makes use of the
combinatorial potentials of lexical items prompting selective projections of mental
contents into novel conceptual structures. Let us advance two claims in (8) and (9)
that are extracted from the two approaches to conceptual integration:
(8) Claim 1:
Humans are claimed to use a relatively limited inventory of grammatical and
lexical forms to prompt for virtually unlimited ranges of cognitive repre
sentations.
(9) Claim 2:
Linguistic expressions with an encoded set of diverse types of information
 László I. Komlósi and Elisabeth Knipf
prompt listeners to construct cognitive representations that integrate cogni

tive structures and cognitive content into unified cognitive representations.
Leonard Talmy’s project of Cognitive Semantics proposes the view that semantics is
intrinsically cognitive and grammars of natural languages reveal conceptual structures
since linguistic expressions prompt for conceptual arrays rather than carry fixed mean
ings on their sleeves. Thus, linguistics is the method for discovering the way we think.
Talmy takes the view that human language consists of two subsystems, the grammati
cal and the lexical. The lexical subsystem consists of the open classes of linguistic forms,
while the grammatical subsystem consists of the closed classes of linguistic forms that
include grammatical categories, grammatical relations, word order patterns, grammat
ical complexes such as constructions, syntactic structures and complement structures.
Linguistic expressions prompt listeners to construct cognitive representations. The
grammatical subsystem provides cues for their structure, the lexical subsystem for their
content. Our language capacity depends on our ability to integrate diverse concep
tual structures and conceptual contents to create unified cognitive representations. We
should observe here that the emphasis in Talmy’s conceptual organization approach
is on the processes of concept structuring rather than on the classical view of semantic
compositionality.
Recent trends in the research field of meaning extension and conceptual integra
tion have streamlined two major interests (see especially Fauconnier & Turner 2002;
Grady, Oakley, & Coulson 1999; Radden 2002; Turner 1997, Turner & Fauconnier
1995):
1. metaphorical and metonymical projection of meaning properties from the physi
cal domain into the abstract domain;
2. selective projection of meaning properties into blended spaces.
It is not the aim of the present paper to discuss these phenomena in detail (for details
see e.g. Komlósi 2003). The relevance of the techniques of non-transparent meaning
construction, however, is that they concern the notion of semantic compositionality.
We call attention here to the fact that the specificities of phrasemes, including fixed
lexical expressions in the lexicon (as discussed in Section 2 above), and the different
varieties of conceptual integration call into question the classical notion of semantic
compositionality and call for substantial revisions in this respect. We must acknowl
edge that these techniques of non-transparent meaning construction crucially depend
on the existence of intrinsic means to prompt conceptual integration, i.e. the creation of
unified cognitive representations out of the most diverse – linguistically coded and
conceptually relevant – types of information. In the case of selective projection of
meaning properties to bring about conceptual integration, some relevance-theoretic
principle has to be at work: contextual effect and processing effort need to achieve a
critical trade-off for unique selection. This is exactly what lexical listing (in the form
of constrained lexical rules) guarantees for standard lexical meanings by overriding
unintended (unconstrained syntactic rule generated) meanings.
A contrastive analysis of entrenchment 
In (10) we make a preliminary attempt to visualize the correlation between ex

pression types, levels of transparency and types of compositionality in the case of the
elements of the syntactic continuum from single words to syntactic phrases, to phrasal
lexical items, to compounds and idioms, to metaphors, blends and implicatures.
(10) lexical units free syntactic metaphors/blends
structures
(selective) projections
of meaning properties
semi-fixed and fixed syntactic phrases metaphors/nonce expressions
expressions (collocations, implicatures
idioms)
non-compositional compositional semi-compositional
non-transparent transparent non-transparent
An illuminating example of the falsehood of the intuitive assumption of composi
tionality in two-word formal expressions is the set of seemingly analogous compound
constructions of dolphin-safe, shark-safe and child-safe, originally analyzed in Turner &
Fauconnier (1995). They point out that we get quite different conceptual integrations
in each of these cases despite the same formal constructions underlying these com
pounds. Their claim about the potentials of differing conceptual integrations resulting
from identical underlying formal construction of two-word formal expressions is rel
evant and valid, even if we are made to realize on closer scrutiny that both the formal
and the conceptual environments of these formal expressions are actually more fine-
grained than we can tell at first sight. One soon realizes that in order to construct the
intended meanings one needs to recruit additional material and reconstruct additional
contextual parameters, such as in (11)–(12)–(13):
(11) dolphin-safe canned tuna
‘No dolphins fell victim to the harvest of the tuna’
(12) shark-safe bay
‘No swimmers fall victim to cruising sharks’
(13) child-safe bay
‘No swimming child falls victim to cruising sharks or is exposed to unforeseen
dangers’
In a similar fashion, with appropriate parameter and perspective selection, the stan
dard and nonce interpretations for (14a–c) can easily be computed:
(14) a. child-safe room
‘The room conceals no unforeseen dangers for children’
We realize how quickly we adjust to an inevitably different conceptual integration
generated by the same expression in (14b) where this expression is meant to depict
 László I. Komlósi and Elisabeth Knipf
a guestroom in a protected quarter of one’s house such that the room is safe for one’s
visitors against any possible attack by the family’s small children.
(14) b. child-safe room
‘The room conceals no unforeseen dangers of child attacks for visitors’
Further, we can envisage a guest-room (in a thought experiment) in which some
strange earth-radiation prevents any attempt for a female egg to be fertilized, thus
female guests are safe/fortunate/?unfortunate (as the case may be) not to become preg
nant in that very room. Thus, the expression child-safe room would have yet another
interpretation:
(14) c. child-safe room
‘The room conceals no unforeseen danger of unwanted conception for a
fertile woman making love with a fertile man in that very room’
As a pragmatic consequence, we acknowledge that such a room would then provide
for child-safe love. It is not difficult to see that this expression could again have several
different meanings, one of which would evoke a meaning similar to the one in (14b),
mutatis mutandis.
This line of research seems to be very promising as it asks further questions
about the correlation between lexical and grammatical meaning, between linguistic
constructions and conceptual structures, and aims at exploring the psycholinguistic
processes of meaning extension and conceptual integration.
. Entrenchment: Collocational force and idiomaticity
In relationship to fixed expressions, one observes that these expressions exhibit unified
cognitive structures while being stored as unitized lexical items in the lexicon. Most of
them are semi-transparent and non-compositional. The phenomenon of a fixed asso
ciation between lexical form (usually of a non-productive or semi-productive syntactic
structure) and non-compositional meaning structure is addressed in psycholinguistic
research under the notion of entrenchment. Let us review the most pronounced charac
teristics of entrenchment and entrenched expressions in (15) as listed in Harris (1998):
(15) Characteristics of linguistic entrenchment
a. an entrenched expression is familiar to speakers from frequent use,
b. every use of a structure increases its degree of entrenchment in cognitive
organization (it strengthens sequential cohesiveness),
c. cognitive structures are stored in memory often as entrenched items
(which in turn strengthens conceptual cohesiveness),
d. entrenched expressions are often units larger than single words but
smaller than sentences, e.g. common word combinations, collocations or
idioms figuring as variable-sized units.
A contrastive analysis of entrenchment 
In (16a–c) we can find some examples for entrenched expressions:

(16) a. common word combinations or collocations
last chance, good job, faced with, a nice cup of tea
Germ. eingefleischter Junggeselle; Eng. a confirmed bachelor
b. fossilized word combinations
goldfish, goldberg, iceberg, nightclub, sandbar
c. multi-word collocations
great minds think alike, shop until you drop
When entrenched items (such as collocations, idioms, clichés, names, constructions,
dead metaphors, etc.) are being processed they activate information bundled together
in unified cognitive structures. From a lexical point of view, unified cognitive structures
enjoy unitized storage: sequential cohesiveness (based on the frequent use of relatively
stable word combinations or fixed expressions) and conceptual cohesiveness (based on
entrenched meaning properties) add to their unitized status. Thus, entrenchment is
to be seen as a fundamental and intrinsic feature of cognitive organization. As for the
processing of entrenched expressions, psycholinguistic research suggests that the fol
lowing mechanisms are likely to take place (cf. Harris 1998; Komlósi 2002). Conceptual
cohesiveness manifests itself in the semantic integration of words with their contexts
which is greatly facilitated by sequential cohesiveness. In the cognitive organization
both structural entrenchment and conceptual entrenchment play an equal role: spread
ing activation operates on both types of entrenchment. As Harris (1998) argues, these
entrenchment mechanisms can best be reconstructed and exemplified by running elic
itation tasks with subjects. In Harris (1998: 62) we find interesting examples for idiom
completion by word elicitation:
(17) a. First two words best at eliciting the idiom:
nothing ventured nothing gained, long time no see, breathing down my neck,
cry your eyes out
b. Middle two words best at eliciting the idiom:
little red riding hood, much ado about nothing, let sleeping dogs lie, let them
eat cake, all hot and bothered, beggars can’t be choosers
As elicitation tasks show a high degree of invariance and consistency across subjects,
we assume that some reliable, maybe even identifiable binding effect should be at work
in such entrenched expressions. We will term this binding effect a collocational force
exerting a unitizing effect that is manifest in spreading activation, semantic integra
tion, sequential and conceptual cohesiveness. The unit status of an expression in the
lexicon entails entrenchment prompting default processing of some unified cognitive
structure (cf. Harris 1998). This is a crucial finding since processing type as a param
eter thus can be appreciated as indicative in the delineation of entrenched expressions
of noncompositional structure from nonce expressions of noncompositional structure
which are a result of meaning extension and conceptual integration. As we shall see be
 László I. Komlósi and Elisabeth Knipf
low, nonce expressions tend to prompt open processing (cf. Fauconnier & Turner 2002;
Gibbs 1995; Turner 1997).
Entrenchment plays a central role in processing constructions as well. The exam
ples in (16) will show that the units of lexical representation in a lexicon of variable-
sized units are not part of a fixed architecture, they rather emerge through the extrac
tion of co-occurrence regularities. “One implication of this idea is that unit-status,
and the size of units, may be a matter of degree” (cf. Harris 1998: 65). Let us exam
ine the mechanism of entrenching and productively spreading collocational force or
co-occurrence regularities in constructions based on the examples in Harris (1998):
(18) a. overtly occurring expressions
lose sight, never mind
b. productive generalization over overtly occurring expressions
lose sight, lose touch, lose track, lose count
c. independent units
(i) relatively free ones:
lose one’s purse, lose one’s passport
(ii) with bonds of co-occurrence:
lose one’s temper, lose track of sg or sy, lose sight of sg or sy, lose touch of
reality
Constructions can be very peculiar in many ways. It is often the case that beside
morpho-syntactic and conceptual information, pragmatic information also gets en
coded in constructions. There are entrenched expressions – such as the set of examples
in (19) for example – whose literal meaning is related to:
1. a linguistically encoded illocutionary force or
2. the use of second-order speech acts (perlocutions or indirect illocutions) or
3. hedges.
(19) A mixed set of pragmatic or other text-organizing constructions:

after all, anyway, at any rate, besides, be that as it may, by the way, not that I care,
first of all, finally, frankly, furthermore, if you want my opinion, in conclusion,
indeed, in other words, now that you mention it, on the other hand, otherwise,
speaking for myself, strictly speaking, to begin with, to oversimplify, to put it
mildly, etc.
Let us take a closer look at some of these expressions.
(20) Not that I care but (p).
The utterance in (20) visibly exhibits non-truth-functional semantics: a pragmatic pre
supposition and a second-order speech act are involved in it. The presupposition is that
S is by no means involved or engaged, let alone interested in the issue at hand and
as such he/she should not attempt to express his/her opinion concerning the issue at
stake. However, it comes into conflict with the indirect illocution of giving an expla
A contrastive analysis of entrenchment 
nation for S’s maintaining (p), which qualifies as an act of expressing S’s implicit view
and S’s claim of the right to express his/her view on a given matter. (‘Despite S’s alleged
non-involvement, it is – for some hidden but intriguing reason – important for S to
express his/her opinion on the issue.’) Further, the utterance in (20) carries a perlocu
tionary effect as well to the effect that S should make H believe that the presupposition
should hold despite the identifiable (though indirect) illocutionary force.
(21) Now that you mention it, (p).
Again we experience an indirect illocutionary act in virtue of which S manages to
establish a justification of a claim (p) advanced and maintained by him/her. The
perlocutionary effect might actually function here as an acceptance of some external
contextual support for claiming and maintaining (p).
(22) If you want my opinion, (p).
Under normal conversational conditions, people do want to know their partners’ views
and opinions in complying with one of the basic conversational assumptions. This
compliance is the token of sincerity and cooperation. The utterance in (22) introduces
a condition coded at the meta-level of communication. It informs the conversational
partner implicitly of the practice of double standards: sometimes you are ready to
speak up for your views, but not always. S calls attention to changing the meta-rules,
admitting having flouted some conversational maxim so far.
(23) (P), to put it mildly.
Again, the utterance constitutes a tension between the propositional meaning and
the illocutionary force established by (p) – whatever (p) should be – and the meta
level conversational evaluation of (p). The intended perlocutionary effect overrides
the illocutionary force. It is a sort of hedging as it shifts the validity conditions of the
‘conversational category’ of evaluation from the default level to the intended meta
level.
. The levels of compositionality
We believe that we have been able to direct attention to some distinctive features of
fixed and semi-fixed expressions, on the basis of which we can conceive of their op
position to free syntactic structures in an independently motivated way. The most
outstanding characteristic feature, we assume, is the discrepancy between the variabil
ity in syntactic-semantic compositionality and the regularity and reliability of unified
cognitive structures. We have come to the understanding that the unit status of many
fixed and semi-fixed expressions in the lexicon is granted to them not so much by for
mal criteria but by psycholinguistically relevant mental constructs that are stored in
the lexicon as variable-sized lexical units with unitized lexical representations showing
a high degree of entrenchment.
 László I. Komlósi and Elisabeth Knipf
In this section we propose that it seems desirable and useful to draw up cat
egories that comprise expressions belonging to different levels of compositionality
when surveying and registering a wide range of fixed expressions. Beside the param
eter of compositionality, we will also take productivity and processing features into
consideration.
. The collocational-level
non-compositional structure
non-productive or semi-productive structures
default processing
A. Proper names evoking metaphorical collocations
(24) a. The ambassador of charm Maurice Chevalier (1888–1972) French actor

and chanson singer
b. Doctor Mirabilis Roger Bacon (1220–1292) English theologian and
philosopher in the Franciscan order
B. Common word combinations
(25) a. Germ. eingefleischter Junggeselle
Eng. a confirmed bachelor
b. Germ. frisch gestrichen
Eng. wet paint
c. a nice cup of tea
. The level of idiomaticity
non-compositional structure
non-productive or semi-productive structures
default processing
(26) a. Germ. von der Hand in den Mund leben
‘to lead a rather modest life’
b. Germ. auf der Palme sein
‘to be angry/raging’
c. Germ. mit heiler Haut davon kommen
‘to get out of a dangerous situation safely’
d. Germ. etwas auf dem Gewissen haben
‘to have a bad conscience’
A contrastive analysis of entrenchment 
d’. Germ. Er hat etwas auf dem Gewissen, darum kann er nicht schlafen.
‘He has a bad conscience, that’s why he cannot sleep at night.’
e. Germ. jemanden auf dem Gewissen haben
‘to be responsible for somebody’s death’
e’. Germ. Die leichtsinnige Touristengruppe hatte den Bergführer auf dem
Gewissen.
‘The light-hearted tourist group was responsible for the death of the
mountain-guide.’
. The level of constructions
non-compositional or semi-compositional structure

semi-productive or productive structures
default processing
(27) a. Germ. etwas zur Aufführung bringen

‘to perform something on stage’
a’. Germ. Die Gruppe hat das neue Drama des jungen Autors zur Aufführung
gebracht.
‘The theater-group has staged the play of the young author.’
(28) Never will I leave you. Long may you prosper! Bless you! Am I tired! Be back in
a minute. It satisfied my every wish. Not that I care. Looks like a soup, eats like a
meal. It’s time you got married. He talked his way out of it. Why go to the store. I
didn’t make it to Paris, let alone Berlin. The more you think about it, the less you
understand. (cf. Cacciari et al. 1997; Kay 1997; Fillmore et al. 1988)
. The level of metaphorical organization
semi-compositional structure
semi-productive structure
open processing
(29) a. Washington has declared war.

b. The school is on an outing today.
c. He is the Finnish Baudelaire.
 László I. Komlósi and Elisabeth Knipf
. The level of conceptual integration
semi-compositional structure
semi-productive or productive structures
open processing
(30) a. This surgeon is a butcher (selective projection into blended mental space)
b. dolphin-safe, shark-safe, child-safe (see examples (11)–(14) above)
. Cross-correlation between compositionality, productivity and processing
On the basis of the observations above, we suggest that there exists an important
cooperation between certain parameters that are able to jointly characterize fixed ex
pressions and help understand their role in grammar through the operation of lexical
licensing. The three interplaying parameters are: (i) the level of compositionality, (ii)
the degree of productivity and (c) the type of processing. Instead of the traditional
static notion of semantic compositionality, a dynamic notion of concept structuring pro
cesses is acknowledged with the help of a scalar organization of linguistic structure and
the corresponding types of processing:
(31) The Parametrization Scheme
Compositionality Processing type
(i) collocational-level
(ii) level of idiomaticity
(iii) level of constructions Default processing for (i–iii)
(iv) level of metaphorical organization
(v) level of conceptual integration Open processing for (iv–v)
Productivity
non-productive
semi-productive
productive
. Comparative linguistic data concerning entrenchment facilitated

through iteration
The first results of our ‘comparative-iterative project’, which looks at the particular
features of fixed expressions, have been published in Komlósi (2002). Some corpus
material is available in different contexts within linguistics (see e.g. The Wheel of For
tune Corpus in Jackendoff 1995 and 1997; Sinclair 1991; Turner 1997), although much
less exists in the way of comparative research methodology (however, see Dobrovolskij
2002; Feilke 1996; Mel’čuk 1995). With our study we make an attempt to expand the
comparative analysis to constructs (mostly fixed expressions) in which a certain de
A contrastive analysis of entrenchment 
gree or some specific aspect of iteration can be observed to play an important role in
entrenchment. Entrenched expressions are fixed or semi-fixed, often semi-productive
and pre-assembled, lexically tagged, unitized lexical items.
The investigation underlying the present paper concerns a specific set of linguis
tic constructions that is characterized by semantic interpretations similar to those
of idiomatic expressions, thus qualifying as entrenched constructions from the point
of view of language processing and lexical retrieval. Many of them, however, defy
compositionally-based morpho-syntactic (syntagmatic) analysis. Instead, they are ob
served to be lexemically unitized and are the result of an interesting interplay between
the combination of language-specific phonotactic and suprasegmental (prosodic)
phonological patterns on the one hand and non-compositional, semi-productive se
mantic structures on the other. We have found ample support in our investigation
for the claim that the underlying mechanism of iterativity, i.e. full or partial (modi
fied) recurrence of some elements of the phonological structures, contributes in crucial
ways to the entrenched nature and the psycholinguistic scope of the linguistic expres
sions under investigation. The distinction between lexeme-driven constructions (e.g. eat
away, eat crow, eat dirt, eat one’s heart out, eat one’s words) and phonology-driven con
structions (hem and haw, a hush-hush project, yakity-yak, zip one’s lip, hurdy-gurdy) will
be crucial for identifying the specific set of entrenched linguistic constructions under
discussion.
As we stated in Sections 2 and 5 (especially (15) above), an entrenched expression
is familiar to speakers from frequent use. It is widely accepted that every perceptually
enhanced and contextually bound use of a linguistic structure is thought to increase its
degree of entrenchment in cognitive organization. We can also witness an increasing
research interest in idioms, collocations, figures, constructions, i.e. in fixed linguistic
expressions that are lexical units larger than single words but smaller than sentences.
We want to show with our examples that the set of expressions generated with the help
of iteration or iteration-like processes cannot easily be assigned to either (i) idiom-
level representations, or (ii) collocation-level representations. We claim that partial
conceptual mapping and non-compositional unifying processes prompted by phono
logical structure are greatly enhanced by the dynamic and forceful phenomenon of
iteration. Iteration phenomena range from “duplication” to “truncated, resemblance-
evoking iterations” which are analyzed in a sample of English, German, Russian and
Hungarian examples of linguistic constructions. Such iteration-based constructions
may include set phrases, collocations, idioms, figures, sayings, proverbs, clichés, quo
tations, names, and metaphorical or metonymical constructions. To complement our
interest in the specificity of Hungarian iterative constructions (especially attributed to
vowel harmony, phonology-driven permutational root semantics and complex agglu
tinative affixational structures), we have chosen further English, German and Russian
examples. It is important to emphasize that the list of expressions cannot strive to be
exhaustive in any sense. We present examples in groups that exhibit some common
features in their behavior. However, at this stage of presentation, we propose gen
 László I. Komlósi and Elisabeth Knipf
eral features for group selection. The features we find decisive may need thorough
reconsiderations in future research.
(32) Examples of general, everyday iterational entrenchment
a. Boys will be boys
b. day by day
c. from step to step
(33) Examples of phonological entrenchment
a. hush-hush project
b. Hung. éjek éje
‘the night of nights’
(34) Examples of phonological entrenchment enhanced with root-and-affix analogies

a. Hung. szegről-végről case iteration with identical case forms
from distant quarter-from periphery
‘[to be] a distant relative’
b. Hung. körös-körül identical root-forms together with
alternative suffixal forms
‘round and round’
c. Hung. várva-várt identical root-forms with meaning iteration
waiting awaited via alternating inflectional forms
‘highly expected’
(35) Examples with phonotactic and suprasegmental features in focus
a. have bats in the belfry
b. have ants in one’s pants
c. be betwixt and between
d. by hook or by crook
e. one is full of fuss and feathers
f. hale and hearty
g. hem and haw
h. pit-a-pat
i. pig in a poke
j. zip one’s lip
k. hurdy-gurdy
(36) Examples with different types of root alteration
a. Ger. gang und gäbe
‘widely accepted’
b. Ger. ohne Sang und Klang
‘without fuss’
c. Hung. hébe-hóba
‘very infrequently’
A contrastive analysis of entrenchment 
d. Hung. hetet-havat összehord

‘incoherent speech’
e. Hung. immel-ámmal
‘reluctantly’
f. Hung. csere-bere
‘playful transaction’
g. Russ. mal’chik s pal’chik
‘Tom Thumb’
h. Russ. Sam ne gam i drugim ne dam
‘selfish and greedy’
i. Russ. Vz’jals’ja za guzh, ne govori, shto ne d’juzh
‘Once you have started something, accomplish it’
. A proposed formal classification of the linguistic data
Finally, we propose a classification on purely formal grounds: the expressions we have

collected follow a certain natural line of grouping as they are built up from constituents
which show in some cases full iteration, in other cases partial or truncated iteration.
The theoretical consequences of such a classification are not yet clear, however it tends
to reflect the depth of entrenchment carried by these fixed expressions.
(37) pure iteration
a. He walked, and walked, and walked.
b. Hung. Csak mondta és mondta és mondta.
‘He just spoke, and spoke and spoke’
c. Russ. davno davno
‘a very long time ago’
d. Russ. On jele-jele id’ot.
‘He walks aimlessly and reluctantly’
(38) full iteration (reduplication)
a. to lead a catch-as-catch-can life
b. boys will be boys
c. let bygones be bygones (literal)
d. let’s call a spade a spade (non-literal)
e. a hush-hush project
f. side by side
g. from X to X (from age to age; from time to time)
h. shoulder to shoulder
i. if worst comes to worst
j. Ger. Schlag auf Schlag
‘swiftly’
 László I. Komlósi and Elisabeth Knipf
k. Ger. Schritt für Schritt
‘gradually’
l. Hung. napról napra
‘from day to day’
(39) partial iteration of modified elements (modified reduplication)

a. day in day out
b. Hung. várva várt esemény
‘a long-awaited event’
c. Hung. körös-körül
‘ubiquitously all about, everywhere’
d. Hung. gólya-gólya gilice
ref. to ‘stork’ (child language)
e. Hung. meg-megáll
‘makes frequent stops’ (verbal prefix iterated)
f. Hung. be-benéz
‘comes by habitually’ (verbal prefix iterated)
g. Hung. éjek éje
the night of nights
‘the darkest of nights’
h. Hung. csak ment, csak mendegélt
‘he wandered aimlessly’ (aspectual alteration)
i. Russ. davnim davno
‘a very long time ago’
j. Russ. durak durakom
‘an idiot’
k. Russ. zhd’jot ne zhd’jots’ja
‘a long awaited event’
(40) truncated iteration (resemblance-evoking reoccurrence of trigger elements as

modified phonological patterns)
a. one has bats in the belfry
b. one has ants in one’s pants
c. one is betwixt and between
d. by hook or by crook
e. one is full of fuss and feathers
f. hale and hearty
g. hem and haw
h. pit-a-pat
i. pig in a poke
j. zip one’s lip
k. hurdy-gurdy
A contrastive analysis of entrenchment 
l. Ger. gang und gäbe
‘widely accepted’
m. Ger. ohne Sang und Klang
‘without fuss’
n. Hung. hébe-hóba
‘very infrequently’
o. Hung. hetet-havat összehord
‘incoherent speech’
p. Hung. szegről-végről rokon
‘distant relative’
r. Hung. immel-ámmal
‘reluctantly’
q. Hung. csere-bere
‘playful transaction’
s. Russ. mal’chik s pal’chik
‘Tom Thumb’
t. Russ. Sam ne gam i drugim ne dam
‘selfish and greedy’
u. Russ. Vz’jals’ja za guzh, ne govori, shto ne d’juzh
‘Once you have started something, accomplish it’
. Concluding remarks
Our paper is a report on an attempt to redefine lexical structure on the basis of the
various manifestations and the combinatorial outcomes of lexical form and unitized
representations. This is a study of the nature of variable-sized lexical expressions that
behave as free or semi-fixed or fixed lexical entries in the mental lexicon, as the case
might be. For the sake of unified treatment, semi-fixed and fixed lexical items are
treated as phrasemes.
Special attention has been paid to the understanding of the role of entrenched fea
tures of variable-sized lexical units. On the basis of the analysis of entrenchment man
ifested in phrasemes, we have found that the retrieval of entrenched features is greatly
facilitated by the activation of encoded, highly context-bound, information-sensitive
parameters. In the examples we have analyzed for semi-fixed or fixed expressions, we
find ample evidence to point to the specific combinations of semantic and pragmatic
meanings that vary according to the types of phrasemes. In the broad and liberal
set of examples for phrasal lexical items, we have included collocations, construc
tions, formulaic expressions, idioms, pragmatemes, clichés, figures, metaphorical and
metonymical expressions, and other linguistic constructs that are a result of meaning
construction, meaning extension and conceptual integration.
We have illustrated with the examples how collocational force leads to entrench
ment, which in turn proves to be a fundamental and intrinsic feature of cognitive
 László I. Komlósi and Elisabeth Knipf
organization. Our comparative linguistic data have concentrated on the particular phe
nomenon of iterativity: it has been shown that different types and extents of iteration
contribute to special effects evoked and activated by different types of encoded in
formation – phonological, phonotactic, morphological, morpho-syntactic, syntactic,
semantic, conceptual, contextual and pragmatic. It has been argued that these spe
cial effects prompt meaning assignment via active meaning construction, meaning
extension and conceptual integration.
Note
* The paper presents some of the results of a cognitive linguistics project supported by the
Hungarian Scientific Research Fund (OTKA T038142) under the title “Cognitive and Affective
Aspects of the Narrative Behavior Reflected in Situated Discourse: Linguistic Construction, Id
iomaticity and Conceptual Integration” undertaken by a research team at the Department of
English Linguistics, Faculty of Humanities, University of Pécs, Hungary. An earlier formula
tion of some issues raised and some examples discussed in the present paper are contained in
a publication by the two authors to appear in A. S. da Silva, A. Torres, & M. Gonçalves (Eds.):
Linguagem, Cultura e Cognição: Estudios de Linguística Cognitiva, Coimbra: Almedina, 2004.
References
Busse, Dietrich (2002). “Wortkombinationen”. In A. D. Cruse, F. Hundsnurscher, M. Job, & P. R.

Lutzeier (Eds.), Lexicology. An International Handbook on the Nature and Structure of Words
and Vocabularies, Vol. 1 (pp. 408–415). Berlin and New York: Mouton de Gruyter.
Cacciari, Cristina, Raymond W. Gibbs Jr., Albert Katz, & Mark Turner (1997). Figurative
Language and Thought. Oxford and New York: Oxford University Press.
Croft, William (2000). “Lexical and grammatical meaning”. In G. Booij, Ch. Lehmann, & J.
Mugdan (Eds.), Morphology. An International Handbook on Inflection and Word-Formation,
Vol. 1 (pp. 257–263). Berlin and New York: Mouton de Gruyter.
Dobrovolskij, Dimitrij (2002). “Phraseologismen in kontrastiver Sicht”. In A. D. Cruse, F.
Hundsnurscher, M. Job, & P. R. Lutzeier (Eds.), Lexicology. An International Handbook on
the Nature and Structure of Words and Vocabularies, Vol. 1 (pp. 442–451). Berlin and New
Fauconnier, Gilles & Mark Turner (2002). The Way We Think: Conceptual Blending and the
Mind’s Hidden Complexities. New York: Basic Books.
Feilke, Helmuth (1996). Sprache als soziale Gestalt. Ausdruck, Prägung und die Ordnung der
sprachlichen Typik. Frankfurt am Main: Suhrkamp Verlag.
Fillmore, Charles J. (1985). “Frames and the semantics of understanding”. Quaderni di
semantica, 6, 222–254.
Fillmore, Charles J. & Paul Kay (1997). Berkeley Construction Grammar. http://www.icsi.
berkeley.edu/∼kay/bcg/ConGram.html
Fillmore, Charles J., Paul Kay, & Mary Catherine O’Connor (1988). “Regularity and idiomaticity
in grammar: the case of Let alone”. Language, 64, 501–538.
Frawley, William (1992). Linguistic Semantics. Hillsdale, NJ: Lawrence Erlbaum.
A contrastive analysis of entrenchment 
Frawley, William (2002). “Fundamental issues in lexical semantics”. In A. D Cruse, F.

Gibbs, Raymond W. Jr. (1995). “Idiomaticity and human cognition”. In M. Everaert, E.-J. van der
Linden, A. Schenk, & R. Schreuder (Eds.), Idioms: Structural and Psychological Perspectives
(pp. 97–116). Hillsdale, NJ: Lawrence Erlbaum Associates.
Goldberg, Adele E. (1995). Constructions: A Construction Grammar Approach to Argument
Structure. Chicago: Chicago University Press.
Grady, Joseph, Todd Oakley, & Seana Coulson (1999). “Blending and metaphor”. In R. W. Gibbs,
Jr. & G. J. Steen (Eds.), Metaphor in Cognitive Linguistics (pp. 101–124). Amsterdam and
Philaldelphia: John Benjamins.
Harris, Catherine L. (1998). “Psychological studies of entrenchment”. In J.-P. Koenig (Ed.),
Discourse and Cognition: Bridging the Gap (pp. 55–70). Stanford: CSLI Publications.
Jackendoff, Ray S. (1995). “The Boundaries of the Lexicon”. In M. Everaert, E.-J. van der Linden,
A. Schenk, & R. Schreuder (Eds.), Idioms: Structural and Psychological Perspectives (pp. 133–
165). Hillsdale, NJ: Lawrence Erlbaum Associates.
Jackendoff, Ray S. (1997). The Architecture of the Language Faculty. Cambridge, MA: The MIT
Press.
Kay, P. (1997). Words and the Grammar of Context. Stanford, CA: CSLI Publications.
Kiefer, Ferenc (2000). “Regularity”. In G. Booij, Ch. Lehmann, & J. Mugdan (Eds.), Morphology.
An International Handbook on Inflection and Word-Formation, Vol. 1 (pp. 296–302). Berlin
Komlósi, László I. (2002). “Entrenchment facilitated through iteration in blended construc
tions in Hungarian”. In A. Hougaard & S. Nordahl Lund (Eds.), The Way We Think, Vol.
1 (pp. 337–339). Odense: Suddanske Universiteit.
Komlósi, László I. (2003). “In quest of cultural and conceptual universals for situated
discursive practice”. In L. I. Komlósi, P. Houtlosser, & M. Leezenberg (Eds.), Communi
cation and Culture: Argumentative, Cognitive and Linguistic Perspectives (pp. 159–176).
Amsterdam: SIC SAT.
Langacker, Ronald W. (1987). Foundations of Cognitive Grammar, Vol. I: Theoretical Prerequisites.
Stanford: Stanford University Press.
Levinson, Stephen C. (2000). Presumptive Meanings: The Theory of Generalized Conversational
Implicature. Cambridge, MA: The MIT Press.
Mel’čuk, Igor (1995). “Phrasemes in language and phraseology in linguistics”. In M. Everaert,
E.-J. van der Linden, A. Schenk, & R. Schreuder (Eds.), Idioms: Structural and Psychological
Perspectives (pp. 167–232). Hillsdale, NJ: Lawrence Erlbaum Associates.
Nunberg, Geoffrey, Ivan A. Sag, & Thomas Wasow (1994). “Idioms”. Language, 70(3), 491–538.
Radden, Günter (2002). “How metonymic are metaphors?” In R. Dirven & R. Pörings (Eds.),
Metaphor and Metonymy in Comparison and Contrast (pp. 407–434). Berlin and New York:
Mouton de Gruyter.
Sinclair, John (1991). “Collocation” (Chapter 8). In Corpus, Concordance, Collocation (pp. 109–
121). Oxford: OUP.
Stilwell Peccei, Jean (1999). Pragmatics. London: Routledge.
Talmy, Leonard (2000). Toward a Cognitive Semantics, Vol. 1: Concept Structuring Systems; Vol.
2: Typology and Process in Concept Structuring. Cambridge, MA: The MIT Press.
 László I. Komlósi and Elisabeth Knipf
Taylor, John R. (2002). “Conceptual approaches V: concepts and domains”. In A. D. Cruse, F.

Turner, Mark (1997). “Figure”. http://www.wam.umd.edu/∼mturn/WWW/figure.WWW/fc.
html
Turner, Mark & Gilles Fauconnier (1995). “Conceptual integration and formal expression”.
Journal of Metaphor and Symbolic Activity, 10(3), 183–204. http://philosophy.uoregon.edu/
metaphor/turner.html/
P V
Language learning
Designing vocabulary tests for English,
Spanish and other languages
Paul Meara
This paper describes an approach to lexical competence in an L2. The model

suggests that it might be possible to describe lexical competence in terms of three
dimensions: vocabulary size, depth of vocabulary knowledge, and the
accessibility of core lexical items. Some formal tests which assess these
subcomponents of lexical competence are described. The paper then goes on to
argue that it is not straightforward to adapt these tests for languages other than
English. “Typical vocabulary size” appears to be a flexible concept, which varies
from one language to another. This means that a statement such as “X has a
vocabulary of N words” may have different practical implications in different
languages. Similar arguments apply to measures of vocabulary depth and to
measures of accessibility of the core vocabulary. The paper argues that the
relationship between these three subcomponents of lexical competence may vary
from one language to another, and that “the same” data culled from “the same”
test in different languages may need to be interpreted in different ways.
. Three vocabulary tests1
For some years now, my colleagues and I at Swansea University have been develop
ing vocabulary tests and assessment instruments. Our work in this area has largely
focused on English, but we are also interested in other languages, particularly Span
ish. Our working model is that vocabulary skills can be accounted for in terms of a
three-dimensional model whose axes are vocabulary size, vocabulary organisation and
vocabulary accessibility. The first of these dimensions accounts for how many words
you know in the target language. The second dimension is concerned with how the
words you know are linked to each other in a complex network structure. The third of
the dimensions is concerned with how easily you can manipulate the words you know.
This model is rather more complex than the traditional vocabulary continuum model,
which simply sees words as moving along a receptive/productive continuum, and when
it is fully elaborated, and empirically tested, we believe that the model will enable us
to make powerful claims about the way vocabularies grow and develop during second
language learning. Of course, in order for us to do this, we need to have good tests,
which reliably measure the dimensions we are interested in, and much of our work
 Paul Meara
over the last 10 years has been concerned with developing tests which have the nec
essary characteristics. For each of these dimensions we now have good working tests,
which provide us with interesting insights about the way vocabularies grow, and the
way the relationship between these dimensions changes as learners get more proficient.
All the tests share a number of features, which gives our work its unique character. All
our tests are administered by computer, and are fully self-scoring. Our tests are quick
to administer – typically they require only 10–15 minutes for administration, and this
means that they can be used as part of a routine testing program, and do not put large
burdens on the test-takers. All our tests can be used with learner groups which vary
greatly in terms of proficiency. They provide scores which can be meaningfully com
pared with scores produced by native speakers on the same tests. And they generate
scores which have transparent and easily interpretable measurement characteristics.
The three main tests in our vocabulary suite are described in the sections that
follow.
. Vocabulary size
Our vocabulary size test is a simple YES/NO test (cf. Meara & Buxton 1987). In this
test, test-takers are presented with a large number of single words out of context, and
they have to indicate whether they know the meaning of each of these words or not.
They do this by pressing an appropriately labelled key. Typically, we use words from
different frequency bands of English, and this allows us to construct a vocabulary pro
file for each test-taker. The words tested include a number of non-existent, imaginary
words, which the testees should claim that they do not know, and these words are used
by the test format as a way of checking the reliability of the test-taker’s YES answers to
real words. Obviously, a test-taker who answers YES to most of the real words and also
answers YES to a large fraction of the imaginary words cannot be fully trusted. Scores
of this type are recognised by the program and adjusted downwards.
Two brief examples of YES/NO questions are shown in Table 1. In the first set of
words we have a number of real high frequency words, and a small number of imagi
nary words. Most readers of this paper will have little difficulty in sorting the real words
Table 1. Examples of Yes/No tests
If you know the meaning of any of these words, draw a circle round it.
Word set 1: Easy Picture Contord
Morlon Partint Teach
Wet Bartle Like
Door Store Vibrade
Word set 2: Patiful Rangue Mascule
Aliver Orduad Peneplain
Killick Prunella Mittimus
Leat Rickwall Masquin
Designing vocabulary tests 
Figure 1. Screen shot from X_Lex
from the imaginary words in this first set, and we would normally conclude from this
that the reader’s knowledge of high frequency vocabulary in English is extensive. The
second set is much harder, however. This set also includes a number of real words to
gether with a number of imaginary words, but the real words are very low-frequency
items. Most readers, including most native speaker readers, will probably find it hard
to identify the real words in the set, and we can conclude from this that their knowl
edge of low-frequency words is limited. With more extensive item sets, we can come
up with reasonably accurate estimates of how many words people know in each of a
number of different frequency bands.
Our standard test, X_Lex, is the computer program that instantiates this basic idea
(Meara & Milton 2003). Words are presented one at a time on a computer screen and
test-takers are asked to indicate whether they know each of these words by clicking
on an appropriate on-screen button. The test presents 150 items, from five different
frequency bands and typically takes only 10 to 15 minutes to complete – each response
requires only a couple of seconds (see Figure 1).
Because X_Lex tests a relatively large number of words, we can easily test a sig
nificant proportion of all the words a learner knows – especially at low levels of
proficiency – and this allows us to extrapolate from our test results to an estimate of
overall vocabulary size with some degree of confidence. It turns out that the estimates
provided by the X_Lex program can often be used as a surrogate for more extensive
testing. In my university, for example, we use the test as a placement test for assigning
students to a class of an appropriate level, and as a way of screening candidates for
public examinations.
 Paul Meara
. Vocabulary organisation
Our second standard test is a test of how far the basic vocabulary of the target language
is organised into a coherent lexical structure. Most linguists agree that vocabulary
is not just a list of words that students learn by rote. Rather, words form semantic
networks with many links joining each word to other words in the network. This
metaphor has been extensively developed by McCarthy (1990) and by Bogaards (1994),
for example. One implication of this idea is that we might expect the lexicons of
advanced learners to be more densely connected than the lexicons of less advanced
learners. In terms of McCarthy’s cobweb metaphor, we interpret this to mean that
there are more connecting links between words in advanced L2 lexicons than there are
in the L2 lexicons of less advanced learners. We have tried a number of different ways
to turn this idea into a viable testing tool (cf. for example Wilks & Meara 2002) and
the most successful approach so far is one in which we give the test-takers small sets of
words, selected more or less at random from the first 1000 words of English, and ask
them to identify any connections they can find between the words in the set.
This idea is instantiated in a program called V_Links, which presents sets of 8–12
words on screen, and allows test-takers to mark any word pairs which are associated
together by drawing a line from one word to the other (see Figure 2). As an example,
consider the set of words: dog, read, bark, song, tree, sun, page, play, book, note, ticket,
theatre. All these words come from the first 1000 words in English, so we can reason
ably expect most learners, even learners at elementary levels, to recognise them. We
would also expect native speakers of English to recognise that there are many associa
tive links between the items in this set: dog∼bark, tree∼bark, book∼page, book∼ticket,
and so on. Advanced learners typically recognise a large proportion of items such as
these. Less advanced learners, on the other hand, are often not aware of the secondary
meanings of words, or their usual collocations, and so they fail to see links that are
obvious to more advanced speakers. For instance, if you know only the verbal mean
ing of play, and you don’t know that a play is also a kind of drama then you may
not see the connection between play and theatre, even though this would be an ob
vious linkage for a native speaker test-taker. Similarly, if you know only the nominal
meaning of book, then you may not recognise the collocational connection between
book and ticket. V_Links uses a small database of word sets which we know perform
well, and for which we have reliable native speaker data. We score the test by allowing
one point for each association that the test-taker identifies which is also found in the
native speaker database. This control gets round the problem of idiosyncratic associa
tions, which many low-level learners produce in large numbers, and allows us to weed
out responses of this type.
Our current version of V_Links provides a set of twenty 12-word sets, and again,
the test can typically be completed in 20 minutes. All words used in the V_links sets are
taken from the first thousand most frequent words in English, and this means that the
test can be taken by most learners, from beginners to very advanced. The mathematical
models that underpin V_Links allow us to make some complex inferences about the
Designing vocabulary tests 
Figure 2. Screen shot from V_Links
overall structure of the core lexicon in the test-taker’s L2, and we can use this data
to develop interesting models of the way that lexical structure develops as learners
become more proficient in their L2.
. Accessing the lexicon
Our test of how accessible vocabulary is makes use of the finding that native speakers
are typically able to recognise words much faster than second language learners can
(cf. de Groot & Kroll 1995). Unfortunately, most of this work comes from laboratory
studies, which rely heavily on very sophisticated measuring equipment, which is not
normally available outside laboratories. Even with this equipment, work of this sort
typically produces differences which are technically “significant” but are very small in
real terms. Mean reaction time differences in the region of 50 to 100 milliseconds are
not uncommon in this work, but this difference is far too small to measure in most
 Paul Meara
Table 2. Examples illustrating the items used in the Q_Lex program
Identify the words hidden in these letter strings

ADLGMIRROREFCD
OPEFRIDGEESDTM
HGSWITCHSESDM
LGFALHAMMERKER
ERKSPIDERETMMO
TDESCHOOLESPES
KREGUITARTELMR
TDMCAHEIGHTGPO
DWHEELSREMONY
practical testing situations, and this means that these experimental methods are of
limited practical use. This has led us to develop an unorthodox word recognition test
which produces bigger and more robust differences, and is reliable enough to be used
in non-laboratory settings.
In our test, called Q_Lex, we present sets of items in which a target language word
is hidden in a string of letters. The basic idea is illustrated in Table 2. In this test format,
the test-taker’s task is to identify the hidden word as quickly as possible. The Q_Lex
program records the time needed to complete the identification, and compares this
time to the time native speakers typically take on that item. The entire test consists of
50 items of this sort. The reaction times generated by this task are very much slower
than those generated by laboratory lexical decision tasks – our RTs are typically in
the region of one second for native speaker Ss, and about 2.5 seconds for non-native
speakers. This difference is relatively large, and you can measure it reliably on standard
computer equipment available in ordinary classrooms. Any correct response which is
made in a time that is close to the native speaker reaction time for that item (the mean
RT plus two standard deviations) is deemed to be normal and scores one point. Like
the tests described earlier, Q_Lex uses words which are all taken from the core 1000
words of English, and this means that it can be used across a wide range of proficiency
levels, from beginner to advanced. And like the other tests described so far in this
section, Q_Lex takes about ten minutes to run. (See Figure 3.)
. The problems
The work described in the previous section has largely been carried out with English
test materials and learners of English as test subjects. On the face of it, however, there
is no reason why the format should not also be adopted for other languages, and our
recent work at Swansea has begun to do this. We have, for example, adapted the X_Lex
model to cover all the main languages spoken in the European Union at the time of
writing – French, German, Spanish, Italian, Portuguese, Greek, Finnish, Swedish, Dan
ish, Dutch and Irish, as well as Icelandic, Norwegian, Catalan, Galego, and we intend
Designing vocabulary tests 
Figure 3. Screen shot from Q_Lex
to extend this list to cover Polish, Czech, Hungarian, the Baltic languages, Slovene and
Maltese, which became EU languages in the last round of Community enlargement.
We also expect to be able to develop versions of V_Links and Q_Lex for at least some
of these languages. There were many practical problems to be overcome before we
could complete this impressive list. These problems mainly arose because of the lack
of reliable frequency counts for some of the languages, but in the end we were able to
overcome these problems by using alternative approaches to data sources. However,
the technical, resource-based problems have turned out to be much less complex than
unexpected difficulties of another kind, which raise some awkward questions about
the comparability of the data we collect from the different languages.
. Vocabulary size
Consider first the question of vocabulary size. Table 3 contains two short YES/NO tests,
one in Spanish, the other in Japanese, which show that the principles work straight
forwardly, despite the obvious differences between the two languages. Some readers of
this paper will have little difficulty in identifying the Spanish words, but the Japanese
words should be almost impossible for most people. The principle here, of course, is
that if you don’t know any Japanese, then you simply have no idea whether the target
 Paul Meara
Table 3. Examples of Yes/No tests in Spanish and Japanese
Spanish test Tratar Quedar Mariscar

Zanjar Vendar Sembrar
Hombrer Cejalar Mezclar
Seguir Manguear Volver
Falar Vaciar Extrar
Ordenar Nadar Nejocar
Japanese test Hohoemu Nageru Kiru
Naku Taberu Egaku
Kiku Warau Miru
Osu Nomu Kowasu
Hokabu Matsu Yomu
Tsukuru Kaku Horu
items exist or not, and any YES responses you make will simply be guesses. As long as
the imaginary words are well-chosen, then it should be impossible to identify the real
words unless you have some knowledge of them. This suggests that the YES/NO test
model will work easily across a wide range of languages. True, there are some minor
problems which need to be addressed in order to make the transfer from English to
other languages work smoothly, but these problems mostly concern the construction
of imaginary words – it turns out that some languages have a more tolerant attitude
to word forms than others do. The selection of real words is generally straightforward
and not problematical, as long as we have reasonably good frequency counts for the
relevant languages, or as long as we can elicit frequency estimates from reliable native
speakers in other ways.
The real difficulty, however, is not the production of the tests, but how we inter
pret the scores that the tests generate. For English, interpreting the scores is easy. We
have a great deal of experience with YES/NO tests, and as a result of this experience it
is relatively easy for us to convert test scores into behaviours that we will expect learn
ers to show. We know from experience, for example, that a vocabulary size score of
about 4500 on our basic test corresponds to what we expect to find in a learner who is
successful in the Cambridge First Certificate Examination. We know that learners with
a score of 1500 words on our tests are usually unable to operate effectively outside a
classroom environment. And we can make similar assessments of learners at other lev
els as well. It is much harder for us to come up with similar statements for the other
languages that we are interested in.
There are several reasons for this, but probably the most important is that English
is unusual in that it makes particularly heavy vocabulary demands on its speakers.
For historical reasons, English typically has two or three word families where other
languages often make do with a single family and its related forms. English for ex
ample has the WRITE and SCRIPT family of words whereas Spanish makes do with
ESCRIBIR and its related forms, and this means that English often ends up with many
pairs of words that have subtly different meanings. In English, a manuscript is not nec
Designing vocabulary tests 
essarily handwritten. We even find scriptwriters – people who write the text of a play or
a film. This sort of process operates quite widely in English, and contributes strongly
to the size of a working English vocabulary. In practical terms, it means that fluent
English speakers probably need to develop much bigger vocabularies than speakers of
some other languages do if they want to be effective. There may not be much difference
between what a person with a vocabulary of 1000 words in Spanish and a person with
a vocabulary of 1000 words in English can do, but it is not at all clear whether we can
make the same claim for larger vocabularies. My hunch is that a 4000 word vocabulary
in Spanish allows you to do a lot more than a 4000 word vocabulary in English, and
that an 8000 word vocabulary in Spanish might be considered large, whereas an 8000
word vocabulary in English is fairly modest.
A further distinguishing feature of English is that it tends not to use compounding
as frequently as some other languages do. In English for instance, we have refrigerator,
or fridge for the thing that in German is called a Kuhlschrank. The meaning of the
English word is not immediately apparent, despite its Latin pseudo-etymology. The
German word, on the other hand, is made up of two transparent and highly frequent
morphemes: kuhl meaning ‘cold’, and Schrank meaning ‘cupboard’. The meaning of the
compound is therefore obvious to anyone who understands the meaning of these two
components. Similarly, in Swedish, we have sjöfylleri, which seems to mean something
like being drunk while in charge of a boat. This word is not one you will find in most
Swedish dictionaries, but it does “exist” in the sense that I have seen it in newspaper
reports, (Styrman åtalad för sjöfylleri) and it is obvious from the context that Swedish
speakers would be expected to be able to generate a meaning for this word, even if they
had never seen it before. What we have here is a fundamental difference in the way the
vocabulary of English and the Romance languages on the one hand, and German or
Swedish and other Germanic languages on the other hand are structured. If you want
to acquire a large vocabulary in English, then you need to learn a very large number of
vocabulary items. On the other hand, if you want to acquire a very large vocabulary in
German or Swedish, then you need to learn a much smaller number of items, but you
also need to learn the system which allows you to exploit the combinatorial possibilities
inherent in German or Swedish vocabulary (see Ringbom 1983 for a longer discussion
of this idea). Again, this suggests that it is not possible to make simple one-to-one
matches between vocabulary size scores in English and equivalent scores in languages
like German and Swedish. A relatively small basic German vocabulary combined with
an understanding of the way that this vocabulary can be used productively, might be
functionally equivalent to a much larger vocabulary in English.
The implications of this deceptively simple idea are profound, and we will come
back to them in the final part of this paper.
. Tests of lexical structure
Our tests of lexical structure and lexical organisation are less well-developed than our
tests of vocabulary size, and so we have a lot less experience of the kinds of problem
 Paul Meara
that arise when we try to develop parallel tests in two or more languages. On the face
of it, we might expect that association-based tests of this sort would transfer relatively
easily, but here too we rapidly run into problems as soon as we try to implement the
idea in practice.
The first problem is a research culture one: many languages simply do not have the
long tradition of word association research that exists for English. In English, serious
word association research goes back to the early part of the 20th Century, and a vast
number of publications appeared in the 1950s and 1960s, exploring word associations
in some considerable depth. For most other languages, this background information is
just not available. What data there is, e.g. for French (Rosenzweig 1961, 1964) suggests
that the association patterns we find differ subtly from those of English, and that they
are affected by extraneous sociolinguistic factors.
For Spanish, the tradition of word association research is short, and relatively un
developed. Nevertheless, there are several reasons why we might expect a tool like
V_Links to produce data in Spanish which looks very different from what we might
expect for English. Some reasons of this sort have already been touched upon in the
previous section. The other main factor is that English words are not formally marked
for part of speech, and this makes English word association stimuli ambiguous in a way
that Spanish stimuli are not. Consider, for example, the word bank: this word can func
tion as a noun, a verb or an adjective, and this means that it can associate with a very
wide range of words. Even if we restrict ourselves to the financial meaning of bank, and
ignore meanings like river bank, we have in English bank∼account and bank∼holiday,
where bank forms part of a complex Noun Phrase, bank∼rate and bank∼clerk, where
the function of bank is more adjectival, and bank∼winnings or bank∼takings, where
bank functions as a verb. This means that English financial bank probably produces, at
least potentially, a much wider range of responses than the equivalent term in Spanish.
A further complication here is that words in English – particularly common words –
often have multiple meanings, and again, the lack of morphological markings in En
glish means that the word association format does not always specify which of these
meanings is intended. English, for example has bank∼to turn an aeroplane round, and
bank∼to depend on, both verbs, which in certain contexts could generate additional
associations that would not occur in Spanish. I don’t know of any research which has
systematically compared multiple meanings of words in English and Spanish, but my
hunch is that English makes more use of this feature than Spanish does. Again, this
would tend to reduce the range of responses produced on a Spanish word-association
test, and produce systematically lower scores on a test like V_Links. Responses in Span
ish are also constrained by word order considerations. For example, in English, it
would not be unusual for the stimulus word white to generate butterfly as an associate.
My guess is that blanco in Spanish would never generate mariposa as a response: in
Spanish, adjectives follow their nouns, rather than preceding them as is the case in En
glish, so this would rule out many of the obvious syntagmatic responses to blanco, but
in any case, the masculine -o ending on the stimulus word would probably preclude
the generation of a feminine word such as mariposa as a response to this stimulus.
Designing vocabulary tests 
Arguments of this sort suggest that Spanish word association data might be rather
more tightly constrained than data generated in English association tests, and this has
some implications for a test like V_Links. Specifically, it means that we might expect
a test like V_Links to generate systematically lower scores in Spanish than it does in
English. For example, a test-taker faced with a set of words that included bank, money,
fly, and depend might recognise associations between all of these words, while a test-
taker faced with the equivalent words in Spanish certainly would not be able to do so. I
don’t think this is just a one-off example: as far as I can see we might expect any set of
N Spanish words to produce fewer associated pairs on average than a set of N English
words. Again, this may look like a relatively minor problem, but, like the differences in
size that we discussed in the previous section, it has some serious implications which
we will pick up on in the final section of this paper.
. Tests of lexical availability
Our final test, Q_Lex, is a test of visual recognition for words, and we might expect
that this aspect of word knowledge would be the one which showed least variation
from one language to another. All speakers need to recognise words, and there is no
obvious reason to suggest that reading words in Spanish should be all that different
from reading words in English. Indeed, many writers lump English and Spanish to
gether under the heading of alphabetic languages, in contrast to syllabic languages,
like Japanese, or logographic languages like Chinese.
Surprisingly, however, there is some evidence to suggest that the languages that use
alphabets are not all “the same” in respect of the way their speakers read their words.
A few studies (e.g. Meara 1986) have suggested that Spanish speakers seem to recog
nise words more slowly than you would expect English speakers to recognise English
words, and there are even some reports where L1_English learners of Spanish appear
to recognise Spanish words faster than native speakers of Spanish do. Spanish speakers
appear to search letter strings in a way that is subtly different from the searches carried
out by L1_English speakers (Estébanez 1997). It is difficult to evaluate these claims,
because there is only a small tradition of experimental studies of word recognition in
Spanish (de Vega & Cuetos 1999), and an even smaller tradition looking at learners
of Spanish as a foreign language. One possible explanation for these results might be
that Spanish speakers rely much more heavily on phonological word recognition than
English speakers do. For English, phonological rules do not reliably predict the pro
nunciation of a word, (few, new, dew, jew and hew but sew, or cave, Dave, gave, rave
and save but have) and for this reason, English speakers are believed to develop both
a set of phonological procedures and a set of orthographical procedures for accessing
their mental lexicons (Coltheart 1996). The question this raises is why Spanish speak
ers should ever need to develop a set of specifically orthographical procedures. Since
Spanish orthography is almost entirely regular, a separate set of orthographic proce
dures would only serve to duplicate what was already possible via a phonological access
route, and there is no obvious reason to invest in redundancy of this sort. Some rather
 Paul Meara
Table 4. Lexical neighbourhoods for leer in English and Spanish
Leer
Beer Deer Jeer Peer Seer Veer
Lier
Lear
Leek Lees
Leer
Lees Leen
tentative support for this view is found in a few cases of L1_Spanish speakers who are
dyslexic in English, but appear to function more or less normally in Spanish (Meara,
Coltheart, & Masterson 1985). The interpretation of these cases is that they rely en
tirely on phonological access in Spanish, and have never developed an orthographic
access for their L1.
Even if we do not accept these radical interpretations of the Spanish word recog
nition data, there are other reasons for arguing that we need to interpret with caution
data culled from tests like Q_Lex. One factor which seems to affect word recognition
speed is the similarity of target words to other words in the language – commonly
called the lexical neighbourhood effect (Luce & Pisoni 1998). Again, I don’t know of
any systematic comparison of lexical neighbourhoods in English and Spanish, but my
hunch is that there are some critical differences that might affect performance on a test
like Q_Lex. Consider the data shown in Table 4. Here we have a single ambiguous word
which exists in both languages: leer. We can assess the size of the immediate lexical
neighbourhood of leer in English and Spanish by counting the number of other four
letter words that differ from leer by just one letter. The figure for English leer is ten. The
equivalent figure for Spanish is only two, and both of these items are morphological
variants of the original word. This simple example suggests that lexical neighbour
hoods in English and Spanish might be very different, and we might expect this to
impact on the speed with which words are recognised.
A further factor is that the stimulus strings in Q_Lex are made up of two parts,
the target words, and the letter strings that they are embedded in. The letter strings for
the English items are actually “second order approximations” to English (Miller 1963)
which reflect the frequency of occurrence of letter pairs in English. It turns out that
second order approximations to Spanish have different properties from the “equiva
lent” strings in English – for example, they contain a much larger proportion of vowels
than second order approximations to English do – and this has a direct effect on how
easy it is to identify a real word hidden in the letter strings. Spanish Q_Lex items seem
to be easier than English items because their syllabic structure is easier to identify.
Again, this means that we cannot directly compare data culled from a Spanish version
of Q_Lex with an English version of the same test format.
Designing vocabulary tests 
. Concluding remarks
The theme which runs through this paper is that creating equivalent tests in different
languages is not as easy as it looks. Differences between languages, some superficial
and obvious, some more subtle and harder to identify, mean that simple adaptations
of tests will not always work in quite the way that we expect. There are two reasons
why this matters.
The first reason is that most theoretical work on vocabulary development is largely
driven by experimental work in English. For example, most of the Spanish research on
vocabulary development is heavily influenced by Paul Nation’s work on English (e.g.
Nation 2001) and its derivatives (e.g. Schmitt 2000). There is an increasing tendency
for people to adopt test formats developed as part of this work without fully thinking
through the implications of a change of language. Nation’s Vocabulary Levels Test, for
example, contains a selection of vocabulary items taken from the 2K, 3K, 5K and 10K
lists for English, and items from an Academic Word List. It is by no means obvious
that these distinctions are the best ones to make for learners of Spanish. Similarly, the
current version of V_Links uses small sets of target words, because we know that this
number generates good data for English, but it isn’t obvious that we would get good
data for Spanish with sets of this size. As test development becomes more sophisticated,
and as more experience in using these test formats accumulates, there is huge pressure
on researchers to adopt the “standard” test formats for their research instruments, and
in this way, anglocentric assumptions become part of our way of doing things. It is not
at all obvious that this is a good thing.
Secondly, and more importantly, perhaps, because we all use the same test formats,
it is increasingly tempting to collect results from two versions of a test and assume that
the tests produce data that is immediately comparable. Suppose, for example, that we
collected two sets of word recognition data from L1_Spanish speakers using English
and Spanish stimuli. We might then feel that we could use the scores to make simple
inferences about the relative strength of the test-takers’ ability to handle words in their
L1 and their L2. Our expectation would be that L2 words would be more difficult to
handle than L1 words, and we could test this by comparing the raw scores for the
Spanish data with the raw scores for the English data. (This approach is the one taken
in some early work on bilingualism, for example Lambert, Havelka, & Gardner 1959.)
If the English data produced slower reaction times than the Spanish data, then we
might feel justified in arguing that this was because the L1 Spanish speakers were less
fluent with the English stimuli than they were with the L1 stimuli. But really, we can
only safely make this inference if we have extensively evaluated the test formats and
if we know what sorts of scores different languages typically produce. Without this
background research, we could be getting results which look like a “significant” L2
deficit which are in fact nothing of the sort.
The reason for developing these tests is that, in the longer term, I am interested
in looking at how different aspects of lexical knowledge interact with each other, and
developing a proper account of how vocabularies grow and how this growth affects the
 Paul Meara
way people can use their vocabulary resources. I suspect that this relationship is very
non-linear, and not at all straightforward. The danger is that tools like the ones I have
described may become extensively used for English, but their results are then gener
alised quite inappropriately to other languages as well. Researchers may inadvertently
treat scores produced by L1_English speakers learning Spanish in the same terms as
they would treat scores produced by L1_Spanish speakers learning English, simply be
cause the test methodology looks similar. As a result we may develop theories about
vocabulary growth and development which really apply only to English, and we may
end up forcing what we know about other languages into a straightjacket that is de
fined by the tests. My hunch is that vocabulary development, and the relationships
between the dimensions that I am trying to measure with these test formats, may vary
strikingly from one language to another. I hope that this preliminary skirmish with
these issues will encourage people to take a much more critical approach towards the
tools that we are developing, and to ask searching questions about the real meaning of
the data that they generate.
Note
. The current versions of X_Lex, Q_Lex and V_Links can be downloaded at: http://www.swan.
ac.uk/cals/calsres/lognostics/
References
Bogaards, Paul (1994). Le vocabulaire dans l’apprentissage des langues étrangères. Paris: Hatier/
Didier.
Coltheart, Max (Ed.). (1996). Phonological Dyslexia. Hove: Lawrence Erlbaum Associates.
de Groot, Annette M. B. & Judith Kroll (Eds.) (1995). Tutorials in Bilingualism: Psycholinguistic
Perspectives. Hillsdale, NJ: Lawrence Erlbaum.
de Vega, Manuel & Fernando Cuetos (1999). Psicolingüística del español. Madrid: Editorial
Trotta.
Estébanez, Salvador (1997). “Specific Word Recognition Strategies Used by Spanish, English and
Bilingual Subjects.” Ph.D. thesis. University of Wales Swansea.
Lambert, Wallace, J. Havelka, & Richard C. Gardner (1959). “Linguistic manifestations of
bilingualism”. American Journal of Psychology, 72, 77–82.
Luce, Paul A. & David B. Pisoni (1998). “Recognising spoken words: the neighbourhood
activation model”. Ear & Hearing, 19, 1–36.
McCarthy, Michael (1990). Vocabulary. Oxford: Oxford University Press.
Meara, Paul M. (1986). “Psycholinguistic relativity – is Spanish different?” In V. J. Cook (Ed.),
Experimental Approaches to Second Language Learning (pp. 151–156). Oxford: Pergamon
Press.
Meara, Paul M. & Barbara Buxton (1987). “An alternative to multiple choice vocabulary tests”.
Language Testing, 4(2), 142–154.
Designing vocabulary tests 
Meara, Paul M., Max Coltheart, & Jackie Masterson (1985). “Hidden reading problems in ESL
learners”. TESL-Canada Journal, 3(1), 69–79.
Meara, Paul M. & James L. Milton (2003). X_Lex: The Swansea Vocabulary Levels Tests. Newbury:
Express.
Miller, George A. (1963). Language and Communication. New York: McGraw-Hill Book
Company, Inc.
Nation, I. S. Paul (2001). Learning Vocabulary in Another Language. Cambridge: Cambridge
University Press.
Ringbom, Håkan (1983). “On the distinctions of item learning vs. system learning, and receptive
competence vs. productive competence in relation to the role of Ll in foreign language
learning”. In H. Ringbom (Ed.), Psycholinguistics and Foreign Language Learning (pp. 163–
173). Åbo: Åbo Akademi.
Rosenzweig, Mark (1961). “Comparisons among word association responses in English, French,
German and Italian”. American Journal of Psychology, 74(3), 347–360.
Rosenzweig, Mark (1964). “Word associations of French workmen: comparisons of associations
of French students and American workmen and students”. Journal of Verbal Learning and
Verbal Behavior, 3, 57–69.
Schmitt, Norbert (2000). Vocabulary in Language Teaching. Cambridge: Cambridge University
Press.
Wilks, Clarissa & Paul M. Meara (2002). “Untangling word webs: graph theory and the notion
of density in second language word association networks”. Second Language Research, 18(4),
303–324.
Timing in English and Spanish
An empirical study of the learning of Spanish
timing by Anglophone learners*
Francisco Gutiérrez Díez
This study deals with contrastive timing (English-Spanish) and the learning of
Spanish timing by a group of native English speakers with the purpose of
establishing the profile of timing errors made by them. We focus on the timing of
several prosodic units (rhythmic foot, ictus – i.e. the tonic syllable of the foot,
and remiss – i.e. the non-tonic syllables of the foot) and of stressed and
unstressed syllables, as well as several related notions, such as stressed/unstressed
syllable ratio, isosyllabicity and foot isochrony. As for the interlanguage of the
anglophone learners of Spanish, we comment on the presence of several
interference, developmental and hypercorrection errors. We atempt to explain
such errors by having recourse to the linguistic background of the learners, the
antagonistic nature of isosyllabicity and isochrony and in terms of two learning
hypotheses: Flege’s Perceptual Target Approach and Gutiérrez’s Contrastive
Perception of Timing.
. Introduction
The work reported in the present article is part of a broader empirical research project
on timing and rhythm in Spanish and English as well as on the rhythmic interlanguage
of two groups of learners: a group of native Spanish speakers who were learners of En
glish and a group of native English speakers who were learners of Spanish. In Gutiérrez
(2001) an account was given of several aspects of the timing and rhythm of English as
produced by a group of native Spanish speakers who were learning English.
In the present paper we intend to report on some timing and rhythmic aspects re
flected in a Spanish corpus as produced by a group of native English speakers who were
learners of Spanish. Specifically, several timing units that are present in the interlan
guage of our Spanish learners will be considered. Since by definition, interlanguage
contains learning errors, among other things, our main aim is to detect the pres
ence of such errors in the students’ timing and rhythmic features. Error detection,
especially interference errors, cannot be worked out without recourse to a contrastive
background in which both the native and the target language of the learner are com
 Francisco Gutiérrez Díez
pared. It should be kept in mind, though, that at present contrastive analysis is not
considered to have a predictive value but rather an explicative one. In Section 3 we
shall refer to the bearing of contrastive analysis on error detection and explanation.
Accordingly, we shall use a comparative background in the description of some
aspects of the timing and rhythmic interlanguage of our learners. For the part on
contrastive analysis, we shall present and expand information first given in Gutiérrez
(1996, 1998–1999). Some of what will be said in the present article regarding the ac
quisition of Spanish timing and rhythm by native English speakers was touched upon
in those works but the bulk of it is presented here for the first time. In accordance with
the scope restriction referred to above, we shall only present here those aspects of the
overall study – samples, procedures, corpus, results and conclusions – pertaining to
the objective singled out for the present report.
We will also examine whether the data obtained have any bearing on what in pre
vious work we have termed contrastive perception of syllable timing within the more
general hypothesis of contrastive perception of linguistic rhythm (Gutiérrez 1998–1999).
In Sections 2 and 3 we will refer to some literature on syllable length and phonological
learning, respectively, which we think is relevant for the purpose of the present study,
and we will advance some of the findings of our work.
. On syllable length
It is a known fact that the syllable is a conspicuous timing unit both in isolation and as
a constituent of other timing units such as the rhythmic foot, the ictus and the remiss
(Abercrombie 1967). As a unit, the syllable is usually considered to be the bearer of
the perceptual attributes of length, pitch and loudness. In both Spanish and English
the three attributes share the role of being responsible for various kinds of prominence
in a complex trading relationship. Our main concern is with the timing of several
prosodic units of which the syllable is a constituent, and this leads us to the question
of syllable length.
Syllabic length has two main roles in Spanish and English. The first one is to act as
a fundamental ingredient in the organization of rhythm and rhythmicality (Crystal
1969). The second one, shared with both pitch and loudness, is to act as a corre
late of linguistic stress. Regarding its status as a stress correlate, syllable duration has
been given different degrees of importance in the two languages by different authors –
always dependent on the other two competing stress correlates, pitch and loudness.
In the competition for ordered priority within a scale of stress correlates, authors
seem almost unanimous in ranking loudness as the least important factor. Regard
ing English, pitch seems to be slightly ahead of duration in most reports (Adams
1979; Bolinger 1958; Couper-Kuhlen 1986; Fry 1955; Kreidler 1989). In Spanish, opin
ions seem more divided on the issue: pitch is considered as the main stress correlate
by Bello (1960), Figueras & Santiago (1993), Monroy (1980), Real Academia Es
pañola (1959) and Solé (1985), while syllable duration would come first according
Timing in English and Spanish 
to Bolinger & Hodapp (1961), Contreras (1963), Gili Gaya (1975) and Ríos, Newman,
& Fragoso (1988).
The second role of syllable duration has to do with the organization of rhythm.
Within the temporal view of linguistic rhythm (Abercrombie 1967; Pike 1945) sylla
ble length is central in the structuring of isochrony, be it of stressed-timed units in
stressed-timed languages or of syllable-timed units in syllable-timed languages. As for
the non-temporal view of rhythm (Faure, Hirst, & Chafcouloff 1980), which sees it
simply in terms of the alternation of stressed and unstressed syllables, syllable duration
is indirectly relevant, since it is one of the three correlates of stress, clearly superseding
loudness and in close competition with pitch, as pointed out above.
According to Jassem, Hill, & Witten (1984), with the exception of monosyllabic
feet, in which the stressed syllable is longer than in polysyllabic feet, in the latter type
of foot both stressed and unstressed syllables have equal duration:
Individual syllables within a multisyllable NRU [or ‘narrow rhythmic unit’, which
is the equivalent of Abercrombie’s rhythmic foot] tend to be of equal length, i.e.,
the complete length of a polysyllabic NRU tends to be somewhat equally divided
among the constituent syllables. (1984: 206)
Although syllable duration in English is closely related to the question of foot

isochrony (i.e. a tendency of all the feet towards equal duration in spoken English)
and syllable compression (i.e. shortening of syllable duration as the number of sylla
bles within the foot increases) as a means for achieving it, both questions fall outside
the scope of the present report. Regarding Spanish, however, when linguists talk about
its syllable-timed rhythm or syllabic isochrony, they are basing their rhythmic stand
on the assumption that both stressed and unstressed syllables have equal duration or
nearly so. Hoequist (1983), O’Connor (1968), Olsen (1972) and several others seek
to solve the skewing between postulated isosyllabism and the inevitable variability of
physical syllable duration found in several corpora by stating that isosyllabism is ul
timately a perceptual construct. Hoequist (1983) contends that durational variability
due to the presence/absence of stress is specific to each language. Regarding the dura
tion of stressed and unstressed syllables, Olsen offers a stressed/unstressed durational
ratio of 1.16 for a half hour talk by a Mexican speaker. Gili Gaya (1940) found a ratio
of 1.39 in phonic-group non-final position for a brief prose passage read aloud by a
Castilian speaker. Delattre (1966) analysed 5 minutes of spontaneous speech (presum
ably produced by South American speakers) and found a ratio of 1.23 in non-final
position within the phonic group. Cuenca (1997) offers the following ratios: 2.79 for
an English prose passage read aloud and 1.71 for a stretch of English spontaneous
conversation; 1.22 for a Spanish prose passage read aloud and 1.10 for a stretch of
Spanish spontaneous conversation. These findings seem to point to two conclusions:
(i) physical isosyllabism is not present in the acoustic signal and (ii) the difference in
the duration of stressed and unstressed syllables is greater in English than in Spanish.
We can advance that our findings confirm conclusions (i) and (ii) mentioned
above, i.e. the lack of isosyllabism in the acoustic signal corresponding to the corpus
 Francisco Gutiérrez Díez
produced by native speakers of Spanish and a greater durational difference between

stressed and unstressed syllables in English than in Spanish. In addition, our data show
an equal duration of Spanish and English unstressed syllables.
. On phonological learning
The supporters of the strong version of the Contrastive Analysis Hypothesis (Lado 1957)
thought that points of dissimilarity between L1 and L2 gave rise to errors that could
be traced back to L1 and explained in terms of interference or negative transfer of L1
features to the learner’s interlanguage. According to the weak version of the Contrastive
Analysis Hypothesis (Wardhaugh 1970), errors relatable to points of dissimilarity be
tween L1 and L2 could not be predicted, but, once they occurred, part of them could
be explained in terms of interference. Errors came to constitute interlanguage fea
tures (Selinker 1972). Both interference and developmental errors (Archibald 1993;
Leather 1999; Major 1987) have survived as broad categories throughout protracted
discussions on the nature of language learning. Numerous learning hypotheses have
been offered to account for phonological learning. Their explanations are based on
dichotomies such as difference/similarity between target and mother tongue items,
marked/unmarked character of the items to be learned, and rate of learning of the ele
ments specified in the afore-mentioned dichotomies through different learning stages.
Related to the dichotomy difference/similarity we have Flege’s (1981) Perceptual
Target Approach, Kuhl’s (1991) Native Language Magnet (cf. Iverson & Kuhl 1995),
and Major’s (1987) Ontogeny Model. Based on the marked/unmarked dichotomy
are Eckman’s Marked Differential Hypothesis (Eckman 1977), later modified as the
Structural Conformity Hypothesis (cf. Eckman 1991), and Carlisle’s (1988) Intralin
gual Marked Hypothesis. A combination of the dichotomies similarity/difference and
marked/unmarked with rate of acquisition through learning stages underlies Major &
Kim’s (1999) Similarity Differential Rate Hypothesis.
In practice, the application of hypotheses related to phonological learning has
been restricted to segmental phonological learning, and, to my knowledge, none of
the previously mentioned hypotheses have been used to account for timing errors,
which fall within the scope of the present report. The lack of application of such hy
potheses to account for the learning of rhythm and timing is probably due to the fact
that concepts such as similarity, dissimilarity, marked and unmarked are easier to apply
to discrete units (segments, syllables, etc.) than to non-discrete ones.
Rhythmic learning can best be explained in terms of greater or lesser values of fea
tures (like stress-timing and syllable-timing) which are increasingly viewed as scalar.
We could safely say that prosodic universals are far from established; let alone the re
lated notions of similarity and markedness at the prosodic level. Whatever one has
to say about learning rate (and, consequently, about hypotheses concerned with this
learning variable) must be based on longitudinal studies (i.e. studies focusing on the
development of the learner’s interlanguage between two points in time), and ours
Timing in English and Spanish 
is cross-sectional (i.e. it focuses on a single moment of the learner’s interlanguage),

involving a group of advanced SL learners. Therefore, we shall be content with de
tecting and typifying the timing errors present in our Spanish learners’ interlanguage
as either interference or developmental errors (Major 1987): the former are errors
induced by features present in the learner’s mother tongue when such features are
(negatively) transferred to the target language contents being learned; the latter can
not be accounted for by interference since they are produced regardless of the learner’s
L1 and are of the same nature as those found in first language acquisition. Interfer
ence errors (also named transfer errors) are not the result of developmental processes,
while developmental errors stem from developmental processes which are part of
universal grammar.
For a tentative explanation of some of the errors detected in our study we shall ap
peal to two hypotheses: Flege’s Perceptual Target Approach and Gutiérrez’s Contrastive
Perception of Timing. According to Flege’s Perceptual Target Approach the learner sets
up perceptual objectives throughout the development of his/her interlanguage by
combining features of both the target and the mother tongue in varying ways. The
objectives change as the interlanguage progresses towards the target language: more
L2 features and fewer L1 features are set up as targets as learning increases. According
to the Contrastive Perception of Timing, when L2 timing is in contrast with L1 timing,
the contrast is perceptually magnified and L2 timing is distorted accordingly by the
learner. In Section 6.2 below we shall attempt to explain some of the errors found in
terms of these two hypotheses.
We will show how our data have allowed us not only to detect timing errors made
by Anglophone learners of Spanish but also to claim that some of the learning er
rors reported in the present investigation can be accounted for by the two hypotheses
mentioned in the previous paragraph.
. Objectives
The objectives we shall focus on in the present report are as follows:

1. Analysis of syllable timing in the interlanguage of a group of English speakers who
are learners of Spanish. A consideration of the pedagogical implications for the
teaching/learning of Spanish to/by native English speakers.
2. As a background subsidiary to the objective stated in (1), we shall present a brief
account of the comparison of syllable timing in Spanish and English. It is our
intention to find out to what extent there are significant intralinguistic and inter-
linguistic durational differences between stressed and unstressed syllables in the
two languages, and how they can influence the learners’ timing errors.
 Francisco Gutiérrez Díez
. The study
Three corpora will be used in this study: (1) Spanish by native speakers (G-1), (2)
English by native speakers (G-2), and (3) Spanish by English speakers who are learning
Spanish (G-3).
. Samples
Seven Spanish speakers, all of them students in their last year of studies in English
Philology at the University of Murcia, Spain, were used in the study as members of G
1 (that is, as native readers of a Spanish text). The informants were randomly chosen
from the Murcia region; they are all educated speakers of Standard Spanish, in some
cases with a light aspiration of [s] in coda position, i.e. in postnuclear position within
the syllable.
Seven educated native speakers of English, from Great Britain, were chosen to
form group G-2 (that is, as native readers of an English text). The same speakers were
also used as members of G-3 (that is, as non-native readers of a Spanish text). Three of
the British speakers were students at Salford University and the remaining four studied
at Essex University; they were all RP speakers in their final year of studies, and were
randomly chosen for the study.
. Instrument
Two texts were used, a Spanish text and English text, each consisting of the transcript
of a combination of extracts from various dialogues on the television (80% of the total)
or radio (20%), and which were illustrative of colloquial speech constrained only by
the presence of cameras or microphones during their production. Both texts are shown
in Appendix 1.
. Procedure
The Spanish text was read aloud by both the G-1 informants and the G-3 informants.
The English text was read aloud by the G-3 informants. The reading output for each
of the 3 groups of informants lasted some twenty minutes. Previous to the reading
aloud stage, the informants were allowed to read the text silently in order to get fa
miliarised with its content and thus minimise the number of false starts and pauses
during the reading-aloud process. Also previous to their reading aloud, the informants
were instructed to read at normal speed, that is, with the speed of somebody speaking
spontaneously in public.1 The readings of the G-1 speakers were recorded in a ‘Ra
dio Nacional’ recording studio in Murcia, using an AEQ mixing deck with a REVOX
open-reel master recorder and an AKG-190 unidirectional and cardioid microphone.
Timing in English and Spanish 
The recordings were subsequently transferred on to a cassette tape (using a TASCAM

122K cassette recorder) for use in a phonetics laboratory.
Three members of both G-2 and G-3 were recorded in a recording studio at Salford
University using a SONY F-30 michrophone and a DTC 1000ES mixing deck with a
SONY open-reel master recorder the contents of which were transferred later on to
a cassette tape (using a TAMBERG AT-771) for use in the phonetics laboratory. The
other four components of both G-2 and G-3 were recorded at the University of Essex
using a SONY-DTC-57ES cassette recorder and a 7-730 unidirectional and cardioid
microphone.
Two groups of judges were used in order to determine the actual stress placement
in the recorded texts by the 3 groups of readers. One group of judges was formed by
nine educated native speakers of Spanish who had to determine stress placement as
produced by the G-1 members (Spanish by native readers) and G-3 members (Span
ish by non-native readers). The other group of judges was formed by nine educated
native speakers of English who had to determine stress placement as produced by G-2
(English by native readers). Since the judges were linguistically naive, they were asked
to tick the syllables which they heard as “prominent” in the speech chain as they lis
tened to the short utterances into which the text had been divided. Each utterance was
sounded only twice for the judges in order to minimise their expectancy of stress (ac
cording to the instructions “only those syllables heard as prominent should be marked,
not those the judges thought ought to be prominent”). After the stresses were adjudi
cated, only those syllables judged as stressed by two thirds of the judges were computed
as such by the researcher.
The three corpora thus obtained (one for G-1, another for G-2 and a third for G
3) were divided into tone units in order to discard from our counts the syllables falling
under the tonic segments (or nuclear tones) of the tone units. This decision was taken
because the constraints on the rhythmic organization of the pretonic segment (or head)
are different from those operating at the tonic segment. In the latter, syllable duration
tends to be longer than in the pretonic segment and this holds true for both stressed
and unstressed syllables. Furthermore, in the tonic segment durational differences be
tween stressed and unstressed syllables are easily levelled out. The unstressed syllables
can be noticeably longer than the (stressed) nuclear syllable due to the slurring effect
determined, among other things, by their final position within the tone unit, by the
nuclear-tone type, and by the number of nuclear tone-bearing syllables. This fact al
lows the researcher, at least from an operational standpoint, to restrict the analysis of
linguistic rhythm to the pretonic segment, where syllable timing and rhythmic organi
zation can be said to be, if not totally independent from intonation, at least not affected
by the regular lengthening of nuclear tones.
For the division of the data into tone units the same criteria were used as in
Gutiérrez (1983, 1995); we shall simply list them here. We claim that the criteria for
the division of the stream of speech into intonational chunks hold equally good for
English and Spanish:
 Francisco Gutiérrez Díez
1. Jump to a high pitch level after a falling tone.

2. Jump to a low pitch level after a rising tone.
3. Jump to a high or a low level after a level tone.
4. Extra-length of nuclear tone-bearing syllables.
5. Anacrusis, that is, a sequence of unstressed syllables which occur after the final
unstressed syllables of a tone unit and are noticeably shortened, thus signalling
the beginning of a new tone unit.
We also omitted from our account syllables belonging to rhythmic feet which had been
fragmented by the readers’ false starts or hesitation pauses. The rest of the syllables,
that is, those which had not been affected by the constraints mentioned above, were
used in durational measurements. Recordings were digitalised by means of an AD con
vertor of the type CED-1401, and syllable duration was measured on the oscillographic
display using the Waterfall program (Cambridge University).
The following criteria were adopted regarding durational measurements:
1. All stressed syllables, that is, those at rhythmical ictus position were measured.
2. The remiss of all rhythmic feet, that is, the stretch of unstressed syllables in each
foot, was also measured. In this way we avoided or minimised the problems in
volved in setting boundaries between every single pair of phones in an utterance
and – to a great extent – between syllables. It was only necessary to establish the
boundaries between ictus and remiss. Dividing total duration of each remiss by the
number of unstressed syllables in it, we obtained average durations for unstressed
syllables.
3. Regarding plosive consonants, their measurement started at the plosion stage
when they occurred word-initially (i.e., after silence), and at the closing point
when followed by silence. In utterance mid-position we followed Wells’ (1990)
criteria for syllable delimitation in English; in Spanish, plosives always occur at
syllable heads and were thus computed as part of such heads.
4. The results of the measurements are expressed in milliseconds.
. Discussion of results
In this section we shall discuss the results related to the duration of stressed and un
stressed syllables and of the rhythmic units foot, ictus and remiss. In Section 6.1 a
comparison will be made of the results found in the data produced by group G-1 with
those found in the data produced by group G-2. In Section 6.2 we shall discuss the
results found in the data produced by group G-3.
. Results pertaining to groups G-1 and G-2
We have compared Spanish stressed and unstressed syllables (as produced by native
speakers) with their English counterparts (again as produced by native speakers). Us
Timing in English and Spanish 
Table 1. T-test to compare the mean duration of stressed and unstressed syllables and the
stressed/unstressed syllable ratios of groups G-1 and G-2
UNIT GROUP N MEAN STANDARD df t-VALUE PROB

DEVIATION
Stressed syllable G-1 509 158.58 57.26 923 –12.33 0.000
G-2 416 204.80 56.00
Unstressed syllable G-1 492 116.05 33.00 724.36 1.31 0.188
G-2 394 112.60 42.82
Stressed/unstressed G-1 486 1.52 1.11 842.54 –7.48 0.000
Ratio G-2 391 2.08 1.11
ing a t-test to compare the mean difference in durational values for the non-related
groups G-1 (i.e. native readers of a Spanish text) and G-2 (i.e. native readers of an
English text), the following significant results were found:
1. The mean duration of stressed syllables is smaller in Spanish (158.58 ms) than in
English (204.80 ms);
2. The mean duration of unstressed syllables is the same in both languages. The
details of the comparison appear in Table 1.
Table 1 above shows that the mean duration of stressed syllables is significantly shorter
for G-1 than for G-2. The mean duration of unstressed syllables of G-1 is not signifi
cantly different from the mean duration of unstressed syllables of group G-2.
Among the above results, the one relating to the equal mean duration of unstressed
syllables in both languages is particularly striking and is counter to the common belief
among Spanish teachers of English that unstressed syllables in English are shorter than
in Spanish. In Gutiérrez (2001) reference is made to the pedagogical implications of
that finding for the teaching of English to Spanish speakers.
A comparison through use of a t-test of the mean durational ratio for stressed/un
stressed syllables in English with the same type of ratio in Spanish shows that the ratio
for Spanish (1.52 ms) is significantly smaller than the same ratio for English (2.08 ms).
The terms of the comparison appear in Table 1 above: The ratios (R) of the two groups
are significantly different (P < 0,05 %). Since H0 : R for G-1 = R for G-2, and H1 : R for
G-1 = R for G-2, we accept H1 : R for G-1 < R for G-2.
The extent of the difference between the stressed/unstressed ratios in English and
Spanish found in our data seems to confirm Hoequist’s statement (Hoequist 1983)
that, from among the many factors determining syllable duration, only the pres
ence/absence of stress can determine language-specific durational differences between
stressed and unstressed syllables.
Let us now turn to the units rhythmic foot, ictus (the stressed syllable in each foot)
and remiss (sequence of unstressed syllables in the foot). We take up these units as de
fined by Abercrombie (1967) and Halliday (1967): each foot begins with an ictus (a
salient part normally filled with a stressed syllable) and may or may not have a remiss
(a non-salient part filled with unstressed syllables). The interlinguistic comparison of
 Francisco Gutiérrez Díez
these units by using t-tests yields the following results: the mean duration of the rhyth
mic foot is significantly greater in Spanish (477.52 ms) than in English (443.37 ms).
Since the duration of the ictus happens to be the same as that of the stressed syllable,
what was said above about the latter holds good for the former: the mean duration of
the ictus is significantly greater in English (204.80 ms) than in Spanish (158.58 ms).
As for the remiss, its mean duration is greater in Spanish (318.93 ms) than in English
(238.56 ms). These results are shown in Table 2 below.
. Results relative to group G-3
A comparison was made of foot, ictus and remiss mean durations in the output of G-3
(Spanish by non-native speakers) and G-1 (Spanish by natives) informants. The result
derived from the t-test displayed in Table 3 below shows that mean durational values
of rhythmic feet and remiss are smaller for G-1 than for G-3 (477.52 ms and 318.93 ms
vs. 551.76 ms and 369.19 ms, respectively).
Since the duration of both the rhythmic foot and the remiss of group G-3 is greater
than the duration of the same units in both Spanish by natives and English by natives,
the result obtained for Spanish learners would be related to their speech tempo: the
tempo of the learners, produced under the same production conditions as for group
G-1, happens to be slower than the tempo of Spanish native speakers. This can only be
interpreted in terms of a developmental tempo error.
Table 2. T-test to compare the mean durations of the rhythmic foot, ictus and remiss of
groups G-1 and G-2

DEVIATION
Foot G-1 509 477.51 199.91 922.29 2.82 0.005
G-2 416 443.37 167.94
Ictus G-1 509 158.58 57.26 923 –12.33 0.000
G-2 416 204.80 56.00
Remiss G-1 509 318.93 188.14 919.67 6.95 0.000
G-2 416 238.56 163.26
Table 3. T-test to compare the mean durational values of the rhythmic feet and remiss of
groups G-1 and G-3

DURATION DEVIATION
Rhythmic foot G-1 509 477.52 199.91 755.61 –5.17 0.000
G-3 370 551.76 217.08
Remiss G-1 509 318.93 188.15 754.21 –3.71 0.000
G-3 370 369.19 204.88
Timing in English and Spanish 
Table 4. T-test to compare the mean durations of the stressed and unstressed syllables and
of the stressed/unstressed syllable ratios of groups G-1 and G-3

DURATION DEVIATION
Stressed syllable G-1 509 158.58 57.26 877 –6.01 0.000
G-3 370 182.57 60.06
Unstressed syllable G-1 492 116.05 32.10 856 –8.69 0.000
G-3 366 136.50 35.51
Stressed/unstressed G-1 492 1.52 1.11 856 1.118 0.268
Ratio G-3 366 1.45 0.70
By definition, the duration of the ictus is the same as the duration of the stressed
syllable. Inter-group comparison of ictus by a t-test yields the following result: the
mean durational value of ictus is greater for G-3 (Spanish by non-natives) than for G
1 (Spanish by natives). By use of a t-test, this difference (158.58 ms for G-1 and 182.57
ms for G-3) has proved to be significant. The results are displayed in Table 4, where we
can see that the mean duration of stressed syllables of G-3 is significantly greater than
that of group G-1.
Since the ictus duration for G-3 (182.57 ms) is mid way between the ictus duration
for G-2 (204.80 ms) and the ictus duration for G-1 (158.58 ms), that is, intermediate
between the mother tongue and the foreign tongue of the learners, this result can be
interpreted in terms of an interference (or negative transfer) error. As pointed out
in Section 3 above, according to Flege’s Perceptual Target Approach, the learner sets
up perceptual objectives during the development of his/her interlanguage by combin
ing features of both the target and the mother tongue in varying ways. The objec
tives change as the interlanguage progresses towards the target language. Our learners
would fall short of attaining target duration because of interference from their mother
tongue ictus duration. They would make, so to speak, a blend of durational features
of Spanish and English regarding ictus duration. Flege’s hypothesis would thus receive
support from our data. Likewise, Table 4 above establishes as a statistically significant
difference that the mean duration of stressed syllables is greater for G-3 than for G-1
(182.57 ms vs. 158.58 ms, respectively). In this case, the interpretation is the same as
that made for ictus production in the previous paragraph.
In addition, use of a t-test to compare the mean durational value of unstressed
syllables for groups G-1 and G-3 yielded the following result: the mean duration of
the unstressed syllables of group G-3 (136.5 ms) is significantly greater than that of
group G-1 (116.05 ms). The results are shown in Table 4. Again, as in the cases of
foot and remiss, we would have a developmental error produced by the G-3 learners,
whose unstressed syllables have a mean duration of 136.50 ms, which is greater than
that obtained in both their mother and foreign tongues.
Comparison of the mean stressed/unstressed durational ratio of G-3 (1.45 ms)
with the ratio that obtained for G-1 (1.52 ms) using a t-test showed that they are not
significantly different from each other. These results are displayed in Table 4 above: the
 Francisco Gutiérrez Díez
ISOSYLLABICITY
ISOCHRONY
1 2 3 4 5 6 7 n
Figure 1. Functional relationship between rhythmic foot duration (d) and the number of
syllables per foot (n)
ratios (R) of both groups are not significantly different (P > 0.05%). Since H0 : R in
G-1 = R in G-3, and H1 : R in G-1 = R in G-3, we assume H0 : R in G-1 = R in G-3.
The interesting thing about the learners’ ratio (1.45 ms) is that it is smaller (even if
non-significantly so) than the ratio for Spanish by native speakers (1.52 ms), which is
the learners’ target ratio, and much smaller than the ratio for English by native speakers
(2.08 ms), which is the departure ratio of the learners. Taking into account both the
departure and the target ratio it seems that the learners went past the target in their
performance of the Spanish stresssed/unstressed ratio. What emerges here is a hint of
a greater degree of isosyllabicity, although it is not statistically great enough to call it a
hypercorrection error. Use of regression lines, however, will produce evidence of such
hypercorrection error related to isosyllabicity.
Regression lines were used to compare the degree of isosyllabicity in the speech
of G-3 informants with that of G-1 and G-2. We can say that in both Spanish and
English foot duration is a function of the number of syllables per foot. This functional
relationship is shown in Figure 1.
What happens is that in English the increase in foot duration, as the number
of syllables per foot increases, is not linear but modulated by simultaneous syllable
compression, which in turn favours foot isochrony (perfect ishocrony is represented
by a horizontal regression line in Figure 1). By contrast, what happens in Spanish is
that, given the tendency towards isosyllabicity, foot duration increases linearly as the
Timing in English and Spanish 
12
10
4
ANISOCHRONY/ISOSYLLABICITY
2 G-3
G-2
0 G-1
0 2 4 6 8 10 12
(Isosyllabicity of G-3 > isosyllabicity of G-1 > isosyllabicity of G-2)
Figure 2. Comparison of perfect anisochrony/isosyllabicity line with the isosyllabicity lines

for G-1 (Spanish by natives), G-2 (English by natives) and G-3 (Spanish by Anglophone
learners)
number of syllables per foot increases. Perfect isosyllabicity is represented by the angle
bisectrix in Figure 1 above.
Now in Figure 2 above, the slopes of the isosyllabicity lines for groups G-1, G-2
and G-3 as well as the slope of the perfect isosyllabicity line can be appreciated. The
slope for G-1 (Spanish by natives) is greater than for G-2 (English by natives), which
means that the feet of the former are less isochronous than those of the latter and,
conversely, the Spanish of G-1 is more isosyllabic than the English of G-2. The slope
of the line corresponding to group G-3 is greater than the slope for G-1, which means
that the Spanish of our learners is more isosyllabic than the Spanish of native speakers.
Through an ANCOVA test, all the differences between the line slopes compared were
shown to be significant, as shown in Table 5 and Table 6 below. This result, which seems
to run counter to expectation, since the native language of the speakers is still less iso
syllabic than their target language, can only be interpreted in terms of hypercorrection
relating to the degree of isosyllabicity.
As contended elswhere (Gutiérrez 1996, 1998–1999), I would like to stress the fact
that, in Spanish, isosyllabicity is a timing issue, not a rhythmical one, since isosyl
labicity runs counter to rhythmicality: we ought to remember that isosyllabicity and
foot isochrony are in antagonistic relationship and in inverse proportion to one an
other. We have empirically shown elswhere (Gutiérrez 1996, 1998–1999) that perfect
isosyllabicity is lacking in Spanish, and all we can say is that it is more isosyllabic than
 Francisco Gutiérrez Díez
Table 5. Analysis of covariance (ANCOVA) to compare the slopes of the regression lines
corresponding to the rhythmic feet of the 3 groups of informants (G-1, G-2 and G-3)*
COMPARISON OF SLOPES df F PROB
Isosyllabicity line with line for G-1, G-2 and G-3 3 8.5 0.000
There is no equality of slopes
Isosyllabicity line with line for G-1 1 5.0 0.000
Isosyllabicity slope > slope for G-1
Isosyllabicity line with line for G-2 1 3.3 0.000
Anisochrony line with line for G-3 1 1.3 0.000
Line for G-1 with line for G-2 1 6.6 0.000
Slope for G-1 > slope for G-2
Slope for G-1 < slope for G-3
Slope for G-2 < slope for G-3
* Dependent variable: Rhythmic foot (mean duration of rhythmic foot). Fixed factor: Informant
group. Covariable: Number of syllables per rhythmic foot.
Table 6. Regression lines of the rhythmic feet of groups G-1, G-2 and G-3 and square values
of the regression lines the slopes of which are compared in Table 5
Regression lines of rhythmic feet:
For G-1: 18.44 + 123.05*NSILP (where 18.44 is the point intersecting with Y (ordinate), in this
case 0; 123.05 is the line slope; * = product; and NSILP = number of syllables per rhythmic
foot).
For G-2: 113.66 + 105.59*NSILP.
For G-3: 21.59 + 143.82*NSILP.
Square values of the regression lines (the slopes of which are being compared):
For G-1, R2 = 73%; for G-2, R2 = 61.5%; for G-3, R2 = 77%.
English, and conversely, the latter, since it has syllable (and foot) compression, is less
isosyllabic, has greater foot isochrony and hence more rhythmicality than Spanish.
It is possible that the hypercorrection error relating to isosyllabicity is due to what
many Anglophone authors call the staccato effect, i.e. the perceptual impression that
Anglophones get from what they interpret as equal duration of Spanish stressed and
unstressed syllables. The staccato (others refer to the same effect as the machine-gun fire
effect) would actually underlie the frequent assumption that Spanish is more isosyllabic
than it actually is (cf. Gutiérrez 1996). This subjective perception of Spanish timing by
native speakers of English is related to what Gutiérrez (1998–1999) has proposed as
the hypothesis of ‘contrastive perception of isosyllabism’, and could be renamed for
the present purpose as ‘contrastive perception of timing’: when L2 timing is in con
Timing in English and Spanish 
trast with L1 timing, the contrast is perceptually magnified and L2 timing is distorted
accordingly by the learner. If we place English and Spanish along a timing scale with
isosyllabicity and non-isosyllabicity each at one extreme of the scale, we can say that
Spanish is nearer to the isosyllabicity end, and English nearer to the non-isosyllabicity
end. The Anglophone learner would perceive that difference or contrast in timing as
sharper than it actually appears in the accoustic signal as produced by native speak
ers, i.e. Spanish is perceived as more isosyllabic than it actually is, and that perceptual
distortion would be triggered by the mere existence of the referred contrast in the
accoustic signal.
Flege’s hypothesis could be slightly modified to make it account for our learners’
hypecorrection error. To that effect, we would like to contend that both the contrastive
perception of the timing of Spanish by the anglophone learners, and its reinforcement
by regular references to Spanish isosyllabicity through terms like staccato and machine-
gun fire contribute to set up perfect isosylabicity – which is lacking in Spanish – as
the learner’s final perceptual objective or target. If we allowed for Flege’s hypothesis
to encompass also perceptually distorted FL features (such as perfect isosyllabicity)
in the perceptual targets that the student sets up in the final stages of learning, such a
hypothesis could explain the learners’ hypercorrection errors, and thus receive support
from our data.
. Conclusions
We will conclude by summarising the results and their interpretation together with
some additional observations. In Section 7.1 we offer the conclusions relating to timing
contrasts between Spanish and English. In Section 7.2 conclusions regarding timing
errors made by anglophone learners of Spanish are offered.
. The timing of prosodic units in English and Spanish

1. The mean duration of stressed syllables is significantly greater in English than
in Spanish. The mean duration of unstressed syllables is the same in both lan
guages. As a consequence, the mean stressed/unstressed durational ratio is greater
in English than in Spanish.
2. The mean duration of the rhythmic foot is significantly greater in Spanish than
in English. The mean duration of the unit ictus is significantly greater in English
than in Spanish. The remiss mean duration is greater in Spanish than in English.
3. Spanish is more isosyllabic than English; the rhythmic foot is more isochronous
in English than in Spanish.
4. Our data show a clear contrast between Spanish and English regarding most of the
prosodic units compared.
 Francisco Gutiérrez Díez
. The timing of Spanish prosodic units by Anglophone learners of that

language
1. The Spanish of the learners shows a slower tempo than the Spanish of native
speakers in the production of rhythmic feet and remiss, thus causing two de
velopmental errors.
2. A further developmental error can be seen in the learners’ production of Spanish
unstressed syllables, which show a slower tempo than that used by native speakers
of that language.
3. The learners show an interference error in the production of ictuses and hence of
stressed syllables, which they make longer than native speakers of Spanish do.
4. The learners also make Spanish stressed and unstressed syllables more equal in du-
rational terms than they actually are by native speakers’ standards, thus producing
a hypercorrection error.
Needless to say that the interpretation of the errors reported is made possible by
looking at the linguistic background of the learners. Flege’s hypothesis would receive
support by accounting for the errors mentioned in (3) and (4). The ‘contrastive per
ception of timing’ hypothesis receives support from our data by accounting for the
hypercorrection error pointed out in (4).
By way of a final observation, it is obvious that much more research is needed in
the field of prosodic learning to test current competing hypotheses relating to phono
logical learning. Both cross-sectional and longitudinal studies would be needed to
carry out such testing. Psychoacoustic experimentation is also needed to test our hy
pothesis of ‘contrastive perception of timing’. The errors related to speech tempo point
to the need to conduct experiments with different speaking rates, and involving differ
ent speaking styles. The latter would allow us to see what the differences are between
the styles of reading aloud and spontaneous speech.
Notes
* The content of the present paper is part of a broader research project financed by the Spanish
Ministry of Education and Science (Ref. BE91.198).
. An advantage of our data is the naturalness of the language in comparison with the use of
distorted or non linguistic materials in other studies (such as isolated words or non-linguistic
stimuli inserted in carrier sentences).
Timing in English and Spanish 
Appendix 1
English text
In this talk I want to consider these questions in connection with the particular passage of the
investigation.
A: How do you actually recommend people to relax? What’s a good exercise for that?
B: Well, I think the thing is you can’t relax until you recognise tension.You’ve got to know
when your neck is beginning to ache because you’ve looked down too long.
A: What would you recommend?
B: I believe in sensible eating. I think that so much has been written and talked about it that
most of us know about food values and about the things that make us fat. The right sort of
pattern leads to health vitality, fitness and so on. And once you’ve established the right kind
of habits, you can forget about all these boring things; about having to do this and avoiding
eating that and so on. You just live a natural life.
It’s very easy to criticise family doctors for the way they prescribe. But the people who do
criticise, especially in the media, often just haven’t the faintest notion, you know.
There was a little boy walking along the road with his father and he looked in the field and
he said to his dad, ‘Dad, is it ten cows over there?’
It is very probable, almost certain in fact, that an officer of the Local Education Authority
will be sitting with the managers and governors in their selection committee, and, as I have said,
it is not within the power of managers and governors to make the final appointment.
A: Would it bother you if other people read your letters if you are not a Cabinet Minister?
B: Frankly, I don’t think it would. They are listening to my telephone conversations. That’s
never bothered me.
I’ve discovered that it’s dangerous, and that the chemical process it goes through is as risky as
the chemical processes that have been blowing up all over Europe.
In nineteen sixty three, and referring to the east of the United States the late president
Lyndon Johnson stated that surely every major river was then polluted.
Spanish text
La mayoría de las gasolineras del país, a excepción de las afectadas por servicios mínimos, se
encuentran también sin gasolina Súper, ya que el suministro sólo se garantiza para las de guardia.
Hace unos años la mayor ambición de cualquier madre con hija casadera era que su hija
contrajera matrimonio con un ingeniero de caminos, canales y puertos.
Entre las medidas urgentes, las fundamentales son ahora mismo construir viviendas de
protección oficial y, especialmente, las que vayan destinadas a aquellas personas que no tienen
capital inicial.
Yo no estoy de acuerdo con usted, porque, ¡mire usted!, hay ayuntamientos que han clasi
ficado mucho suelo y otros han clasificado poco; en todos por igual ha aumentado el precio de
la vivienda y ha aumentado el precio del suelo. El mercado del suelo tiene sus características
particulares como casi todos los mercados.
 Francisco Gutiérrez Díez
Veinticuatro salas en total exhiben las películas de las distintas secciones del festival. Vein
ticuatro salas de características muy distintas: algunas con capacidad para tres mil espectadores
y otras, con no más de sesenta y cinco butacas.
A: ¿Cómo celebran los gallegos el día de las letras gallegas?
B: Celebramos los gallegos el día de hoy como una jornada de afirmación de la personalidad
histórica de Galicia.
Aquí se recogen escritos de todos los grandes nombres de nuestras letras y de nuestro peri
odismo. ¿Cuáles podrían ser los que ocupasen la cabeza de la exaltación?
Él nos mandó esperar hasta los tres años y estuvimos esperando simplemente a que pasara
el tiempo.
A: Se ha dicho siempre que en España la justicia era lenta, cara e insegura; ¿sigue siendo así ?
B: Yo pienso que la justicia es lenta; es cierto que es lenta. No es quizás más lenta en España
que en otros países europeos. Yo esto siempre lo he dicho.
A: ¿Cómo observa la Presidenta de la Audiencia Provincial de Barcelona la puesta en fun
cionamiento del jurado popular?
B: Bueno, yo creo que en principio el jurado, como usted decía está previsto en la Constitu
ción. Por lo tanto, hay que darle un desarrollo normativo. Y yo creo que es bueno que exista
el jurado. Y creo que es bueno, porque es una forma de participación del ciudadano en la
administración de justicia.
References
Abercrombie, David (1967). Elements of General Phonetics. Edinburgh: Edinburgh University

Press.
Adams, Corinne (1979). English Speech Rhythm and the Foreign Learner. The Hague: Mouton.
Archibald, John (1993). Language Learning and L2 phonology. London: Kluwer.
Bello, Andrés (1960). Gramática de la Lengua Castellana. Buenos Aires: Sopena Argentina.
Bolinger, Dwight (1958). Intonation and Its Parts: Melody in Spoken English. London: Edward
Arnold.
Bolinger, Dwight & Marion Hodapp (1961). “Acento melódico, acento de intensidad”. Boletín de
Filología de la Universidad de Chile, 13, 33–48.
Carlisle, Robert (1988). “The effect of markedness on epenthesis in Spanish/English inter-
language phonology”. Issues and Developments in English and Applied Linguistics, 3, 15–23.
Contreras, Heles (1963). “Sobre el acento en español”. Boletín de Filología de la Universidad de
Chile, 15, 223–237.
Couper-Kuhlen, Elizabeth (1986). An Introduction to English Prosody. London: Edward Arnold.
Crystal, David (1969). Prosodic Systems and Intonation in English. Cambridge: Cambridge
University Press.
Cuenca, Heliodora (1997). “Estudio Comparativo del Ritmo Inglés y Español.” Unpublished
Doctoral Dissertation: Universidad de Sevilla.
Delattre, Pierre (1966). “A comparison of syllable length conditioning among languages”.
International Review of Applied Linguistics, 4(3), 183–188.
Eckman, Fred (1977). “Markedness and the contrastive analysis hypothesis”. Language Learning,
27, 315–330.
Timing in English and Spanish 
Eckman, Fred (1991). “The structural conformity hypothesis and the acquisition of consonant
clusters in the interlanguage of ESL learners”. Studies in Second Language Acquisition, 13(1),
23–42.
Faure, George, Daniel Hirst, & Michael Chafcouloff (1980). “Rhythm in English: isochronism,
pitch, and perceived stress”. In L. Waugh & C. van Schooneveld (Eds.), The Melody of
Language: Intonation and Prosody (pp. 71–80). Baltimore: University Park Press.
Figueras, Carolina & María Luisa Santiago (1993). “Producción del rasgo acentual mediante
síntesis de voz”. Estudios de Fonética Experimental, 5, 115–128.
Flege, James (1981). “The phonological basis of foreign accent: a hypothesis”. TESOL Quarterly,
15(4), 443–445.
Fry, Dennis (1955). “Duration and intensity as physical correlates of linguistic stress”. Journal of
the Acoustical Society of America, 27, 765–768.
Gili Gaya, Samuel (1940). “La cantidad silábica en la frase”. Castilla (Valladolid), 1, 287–298.
Gili Gaya, Samuel (1975). Elementos de Fonética General. Madrid: Gredos.
Gutiérrez, Francisco (1983). “Aspectos lingüísticos de la segmentación del tono en inglés,
castellano y catalán”. In AESLA (Ed.), Actas del Ier Congreso Nacional de AESLA (Universidad
de Murcia, 1982) (pp. 179–190). Madrid: SGEL.
Gutiérrez, Francisco (1995). La Función Demarcativa de la Entonación en Inglés, Castellano y
Catalán. Murcia: Servicio de Publicaciones de la Universidad de Murcia.
Gutiérrez, Francisco (1996). “Aprendizaje del español por anglohablantes: distorsión rítmica y
timing”. In F. Gutiérrez (Ed.), Actas del Ier Congreso Internacional de AESLA (Universidad
de Granada, 1992) (pp. 267–276). Universidad de Granada: Servicio de Publicaciones.
Gutiérrez, Francisco (1998–1999). “Aprendizaje de la pronunciación del español por
anglohablantes: distorsión rítmica y timing”. Revista Española de Lingüística Aplicada, 13,
7–26.
Gutiérrez, Francisco (2001). “The acquisition of English syllable timing by native Spanish
speakers learners of English. An empirical study”. International Journal of English Studies, 1,
93–113.
Halliday, Michael A. K. (1967). Intonation and Grammar in British English. The Hague: Mouton.
Hoequist, Charles (1983). “Durational correlates of linguistic rhythm categories”. Phonetica, 40,
19–31.
Iverson, Paul & Patricia Kuhl (1995). “Mapping the perceptual magnet effect for speech using
signal detection theory and multidimensional scaling”. Journal of the Acoustical Society of
America, 97, 553–562.
Jassem, Wiktor, David R. Hill, & Ian H. Witten (1984). “Isochrony in English speech”. In D.
Gibbon & H. Richter (Eds.), Intonation, Accent and Rhythm (pp. 303–325). Berlin: de
Gruyter.
Kreidler, Charles (1989). The Pronunciation of English. Oxford: Blackwell.
Kuhl, Patricia (1991). “Human adults and human infants show a ‘perceptual magnet effect’ for
the prototypes of speech categories. Monkeys do not”. Perception and Psychophysics, 50,
93–107.
Lado, Robert (1957). Linguistics across Cultures. Ann Arbor: University of Michigan Press.
Leather, Jonathan (Ed.). (1999). Phonological Issues in Language Learning. Oxford: Blackwell.
Major, Roy (1987). “A model for interlanguage phonology”. In G. Ioup & S. Weinberger (Eds.),
Interlanguage Phonology (pp. 101–124). Cambridge, MA: Newbury House.
Major, Roy & Eunyi Kim (1999). “The similarity differential rate hypothesis”. In J. Leather (Ed.),
Phonological Issues in Language Learning (pp. 151–184). London: Blackwell.
Monroy, Rafael (1980). Aspectos Fonéticos de las Vocales Españolas. Madrid: SGEL.
 Francisco Gutiérrez Díez
O’Connor, Joseph D. (1968). “The duration of the foot in relation to the number of component
sound segments”. Progress Report, Phon. Lab., Univ. College, London, June 1–6.
Olsen, Carroll Lee (1972). “Rhythmical patterns and syllabic features of the Spanish sense
group”. In A. Rigault & R. Charbonneau (Eds.), Proceedings of the 7th International Congress
of Phonetic Sciences (Montreal 1971) (pp. 990–996). The Hague: Mouton.
Pike, Kenneth (1945). The Intonation of American English. Ann Arbor, MI: University of
Michigan Press.
Real Academia Española (1959). Gramática de la Lengua Española. Madrid.
Ríos, Antonio, Michael Newman, & Jesús Fragoso (1988). “Los correlatos acústicos del
acento del español en la sílaba (en habla de laboratorio)”. Laboratorio de Fonética de la
Universidad Autónoma de Barcelona [Unpublished].
Selinker, Larry (1972). “Interlanguage”. International Review of Applied Linguistics, 10, 209–231.
Solé, María José (1984). “Experimentos sobre la percepción del acento”. In E. Martínez Celdrán
& M. J. Solé (Eds.), Estudios de Fonética Experimental (pp. 135–142). Barcelona: PPU.
Wardhaugh, Ronald (1970). “The contrastive analysis hypothesis”. TESOL, 4(2), 123–130.
Wells, John (1990). Pronunciation Dictionary. London: Longman.
Spanish and English intonation patterns
A perceptual approach to attitudinal meaning*
Rafael Monroy
In this paper the attitudinal function of wh- and yes/no questions is studied
taking O’Connor & Arnold’s description of English as a backdrop against which
the corresponding Spanish intonation patterns are discussed. Perceptual
similarities between English and Spanish were evident in the case of simple
nuclear tones, but less so in compound tones, particularly in the rise-fall.
Pre-nuclear patterns seem to play an important role in the identification of a
tone unit contour. Fall, rise and level tones were more easily identified when
preceded by high heads. Overall, the fall-rise ranked top of the wh- list whereas
the first pattern (low fall) was first in the yes/no questions. Focus placement was
correctly perceived in those cases characterised by default tonicity, but it turned
out to be problematic in patterns with marked tonicity.
. Introduction
In this paper the attitudinal function of wh- and yes/no questions is studied from a
contrastive viewpoint. Our starting point is O’Connor & Arnold’s statement that ‘The
pitch patterns or tunes of English are not necessarily the same in form as those of
other languages, nor do they necessarily produce the same effect as they would in other
languages’ (1973: 1). Indeed when one compares the intonation patterns of English and
Spanish,1 one finds phonetic similarities between the two systems beyond the falling
and rising basic tunes. But there are differences, as we shall see, in pitch range at nuclear
level as well as in pre-nuclear positions. It is no wonder that a foreign accent is usually
attached to the use of a tune in the target language which is typical of the speaker’s
native language. The obvious outcome is a misunderstanding on the part of the native,
although possibly to a lesser degree than one might surmise due to the fact that so-
called foreign talk, far from being particularly emotional, is characterised by the use of
a stereotyped, simplified intonation pattern. Native speakers, however, seem to make
less allowance for mistaken tunes than for faulty segmental pronunciation according
to O’Connor & Arnold (1973: 2).
Despite its importance from a pedagogical standpoint, intonation has always
lagged far behind other areas in phonology. There are several reasons for this cur
 Rafael Monroy
rent state of affairs. Undoubtedly, one of the most important ones is the difficulty of
disentangling phonological categories or intonational lexicons (Liberman 1979) out of
the phonetic complexity underlying intonation. In the case of British English, Halliday
(1967, 1970), for instance, has established five categories, giving fourteen patterns to
capture attitudinal meaning, whereas Crystal has posited just seven tones (1975), and
O’Connor & Arnold (1973) twenty patterns – ten non-emphatic and ten emphatic.
There are also differences among authors regarding the pitch movement of the
nucleus. While Halliday (1967, 1970) envisages eight nuclear movements (high fall,
high rise, low fall, low rise, fall rise, rise fall, low fall rise and low rise fall), Crystal
(1975) does away with the low fall rise and the low rise fall, introducing a level tone
instead. O’Connor & Arnold (1973) bring in a complex tone (fall plus rise) to the tones
used by Crystal and they are more specific about the level tone which they consider to
be mid level.
Greater discrepancies are apparent in the interpretation of the attitudinal mean
ing conveyed by the tones. Halliday (1967, 1970) introduces the concept of key which
is realised by secondary tones as opposed to neutral tones. This, by the way, raises the
question of how a tone can be neutral with a certain clause type (e.g. wh-question)
when in fact it carries attitudinal force if applied to another type of clause (e.g. a
yes/no question). If we take for instance the rise fall, a fairly common tone, and com
pare the attitudinal load attached to it by different authors, we find a whole gamut
of interpretations. While Gimson (1980) sees in it “enthusiasm, doubt, horror, sar
casm, indignation, etc.”, and Carr (1999) attaches to it “a sense of strong agreement
or disagreement”, Halliday (1970) considers that the speaker is “committed, insistent,
asserting”, which is a long way from O’Connor & Arnold’s view of this tone as show
ing that the speaker is “impressed, challenging, antagonistic” (1973: 214), a definition
that Crystal accepts with the qualification “. . .or the reverse, depending on kinesic
accompaniment” (1975: 38).
There is a general agreement that intonation is not the only means available to the
speaker in order to convey the speaker’s inner feelings. Kinesics, lexical choice, gram
matical construction, range of tone, pitch level, loudness, tempo and other prosodic
and paralinguistic features play an important role in this respect. But from a purely
linguistic standpoint, intonation seems to be the basic means to convey attitudinal
meaning, an opinion held by Bolinger (1986), Couper-Kuhlen (1986), Crystal (1969),
O’Connor & Arnold (1973) or Schubiger (1958), who see attitude as the primary
function of intonation. This view echoes in a way Bally’s (1935) dialectics of verbal
interaction according to which language in general reflects two basic functions, the
logical and the affective one. Both go hand in hand and constantly strive to win one
over the other. This means that all expressions in language consist of a logical side plus
an affective one in varying degree, to such an extent that, strictly speaking, no utterance
in a language is entirely logical nor purely emotional. There may be utterances where
the logical side may prevail, as when they are used in a referential function, as there
are expressions that can be heavily tinged with emotion, but this does not preclude the
presence of both elements.
Spanish and English intonation patterns 
These comments of a speculative nature are not shared by all linguists; the matter
has, of course, a lot to do with how we define attitude and how we perceive a partic
ular message. Roach, for one, draws a clear distinction between normal and emotional
language (1991: 137). Crystal, too, talks about attitudinally neutral tones as if this was
an inherent feature of certain tones. Tench shares this view too when he writes that
“the expression of attitude is an optional element” (1996: 108), “. . .mainly to be found
in the extent of a fall or rise and in variations of pitch in the head and pre-head”
(1996: 20). He claims that there are “neutral tones” (falling, rising and falling rising)
vs. non-neutral or “secondary tones” used for attitudinal purposes which are varia
tions in degree of the neutral ones. The former are not envisaged as attitudinal carriers
and are, therefore, the only ones present in the ideational, interpersonal and textual
functions of language if emotions are not present. While there is reason to doubt this
interpretation, there is a point on which I would sympathize with Tench’s quotation,
which echoes Halliday’s as well as O’Connor & Arnold’s view on this matter: that lin
guistically speaking, attitude depends not just on nuclear variation, but also on the
pitch variation of heads and pre-heads as we shall discuss below.
. Objectives
Our analysis will focus on the attitudinal component of the English and the Spanish
intonation systems as used with wh- and yes/no questions. Our interest in questions
derives from the fact that, unlike statements which are characterised by a falling tone
(“the vast majority being low falling in type”, according to Crystal 1975: 33), ques
tions display an array of tone patterns which makes them particularly attractive from
the standpoint of a study of attitudinal meanings. As for the choice of their attitudinal
function as an area of study, there are two reasons for this: firstly, the attitudinal mean
ing will probably be most familiar to our readers – in fact, it is considered by some as
the main function of intonation as pointed out above – and secondly, it has been the
most thoroughly described of all the intonation functions. English (the British vari
ant) will be our target language which shall be approached from the perception of a
Spanish speaking audience. Firstly, we shall consider the two systems in order to see
the degree of similarity between their nuclear tones. Secondly, we shall discuss the role
of the pre-nuclear elements in order to see whether they perform a fundamental or
an ancillary role in the overall perception of a given attitude. Thirdly, a comparison
will be made of all English intonation patterns in order to rank them from the easiest
to the most difficult for a Spanish learner of English. Finally, we shall see how focus
placement in English is perceived by Spanish speakers.
 Rafael Monroy
. Methodology
. Participants
Thirty four Spanish students of English Philology took part in this experiment. Their
level of English could be described as post-intermediate. All of them had followed
an introductory course on English Phonetics, but none had specific training in the
intonation system of English. They all come from Murcia, an area characterised by a
Spanish intonation system with a narrow pitch range.2
. Instruments
O’Connor & Arnold’s (1973) description of English intonation was used on the
grounds of it being internationally the best known textbook on British intonation,
and also because, despite its deficiencies, it is the most comprehensive description of
attitudinal meaning in English. Ten tone units constitute the axis of their system, giv
ing rise to two sets: ten patterns characterised by ten pitch movements of the nucleus,
preceded by low pre-head in all of them, and four types of head (high, low, falling and
rising), and a second set labelled “emphatic” involving high pre-head, plus stepping,
sliding and climbing heads before the ten nuclear tones. Our questionnaire confined
itself to the first set as it is the one thoroughly covered in their Intonation of Collo
quial English. A Practical Handbook (second edition, 1973). O’Connor & Arnold attach
all their intonational lexicon to four basic grammatical patterns: statements, wh- and
yes/no questions, commands, and interjections.3 As pointed out above, we decided to
take wh- and yes/no questions as the focus of our research; the nucleus was preceded in
all of them by either a head or an unmarked pre-head. In order to see the degree of vari
ability in pattern perception, we introduced twenty questions – ten of the wh- and ten
of the yes/no type – repeating each intonation pattern twice. We also devised a second
questionnaire to test the Spanish perception of focus in English (see Appendix 1).4
. Procedure
The first task consisted of listening three times to each of the twenty questions, all taken
from O’Connor & Arnold’s textbook (1973), so that the students could decide which
intonation contour was involved. After this, our informants were asked to mark on a
five-point scale their level of familiarity with the intonation pattern heard, as follows:
1. A, if the utterance was perceived as fully familiar;
2. B, if it was fairly familiar to them;
3. C, if it was occasionally heard and occasionally used;
4. D, if it was rarely heard and used; and
5. E, if the pattern was completely unfamiliar to them.
Spanish and English intonation patterns 
After completion of this task, they heard each expression twice again, but this time
they had to mark the most prominent word in each sentence, associated with the focus
of information and, therefore, with new or important information. Percentage values
were obtained by converting the nominal to ordinal data as follows: A: 10; B: 7.5; C: 5;
D: 2.5; E: 0 (see Table 1 in Section 4).
. Analysis and results
. Tone group 1: Low drop
Pattern similarity between the first English tone (low drop) and the Spanish equivalent
tone is high (50%) or very high (23.5%) in the wh-question. No one reported that
the tone was totally unknown to him/her. As seen in Table 1 below, there is a low
percentage of students who perceived it as an occasional tone (17.6%) and 8.8% who
considered it rarely heard. This is surprising as the low drop is a very common tone in
Spanish, even in detached or unsympathetic types of wh-questions.
A possible explanation for this lack of agreement among our informants could be
the pre-nuclear pattern used – a high, sustained head, which is not as uniform, at least
in the Murciano accent, as Navarro Tomás (1948: 48) surmises.5 The yes/no question
with identical tone pattern (num. 11)6 confirms this impression. We notice that the
nucleus here is a descending rather than an ascending movement, which is the normal
tone used with Spanish questions of a yes/no type; this may account for the lower
percentage (44.2%) of students who recognise the descending pitch as a fairly or very
common tone. The remaining informants judge it as occasionally familiar (20.6%) if
not downright strange (14.7%). The unexpected tone seems to have, however, more
to do with the interplay of the pre-head and the head than with the nucleus. A low
pre-head in “Can I. . .” followed by a high head beginning with see and continuing
Table 1. Percentage values of wh-questions (1 to 10) and yes/no questions (11–20); maxi
mum values in bold type
Wh- 1 2 3 4 5 6 7 8 9 10
%A 23.5% 17.6% 17.6% 44.1% 32.4% 26.5% 5.9% 8.8% 5.9% 5.9%
%B 50.0% 50.0% 23.5% 26.5% 20.6% 23.5% 41.2% 26.5% 26.5% 29.4%
%C 17.6% 26.5% 20.6% 20.6% 29.4% 23.5% 23.5% 26.5% 41.2% 29.4%
%D 8.8% 5.9% 29.4% 8.8% 17.6% 20.6% 11.8% 29.4% 17.6% 26.5%
%E 0.0% 0.0% 8.8% 0.0% 0.0% 5.9% 14.7% 5.9% 8.8% 8.8%
Y/N 11 12 13 14 15 16 17 18 19 20
%A 11.8% 35.3% 17.6% 17.6% 14.7% 11.8% 26.5% 17.6% 58.8% 26.5%
%B 32.4% 35.3% 41.2% 26.5% 23.5% 38.2% 38.2% 47.1% 23.5% 50.0%
%C 20.6% 14.7% 17.6% 41.2% 44.1% 26.5% 23.5% 11.8% 8.8% 17.6%
%D 20.6% 14.7% 23.5% 14.7% 14.7% 14.7% 8.8% 17.6% 5.9% 0.0%
%E 14.7% 0.0% 0.0% 0.0% 2.9% 8.8% 0.0% 2.9% 0.0% 0.0%
 Rafael Monroy
Table 2. Pattern ranking wh-/yes/no questions
Wh-Q Mean Yes/No-Q Mean
4 7.65 19 8.48
1 7.21 20 7.73
2 6.99 12 7.28
5 6.69 17 7.12
6 6.10 18 6.52
7 5.30 13 6.32
3 5.29 14 6.18
8 5.08 15 5.81
9 5.07 16 5.74
10 4.93 11 5.15
uniform until the low fall of the nucleus later is indeed unusual in Spanish. One has
only to translate the English expression to realise that the first and the second element
(Can I. . .) could never be pronounced in our language the way they are in English.
This explains why in terms of pattern ranking (see Table 2), the wh-sentence has a
mean of 7.21% (the second from the top) and the yes/no sentence appears bottom of
the list (2.06% mean difference, see Table 4 below) despite the fact that both share an
identical tone.
If we consider focus placement (see Table 3 below), we notice that in sentence
number one an overwhelming majority (76%) place it rightly on do, which is inter
esting as it means a departure from the Spanish tendency to favour the last strongly
stressed syllable as a focal point (Ortiz Lira 1994). Those who depart from this, place
it on either why (the beginning of the head) or on did (12% in both cases). However,
in sentence eleven, with an identical tone pattern, 38% misplace the nucleus on see vs.
26% who rightly placed it on the last strong word later.
. Tone group 2: High drop
This second tone group is characterised by a high fall on the nucleus, a common tune
used in Spanish with emotional intonation (Navarro Tomás 1944). In wh-questions,
half of the sample identified it as a fairly familiar and 17.6% as a very familiar tone.
It was an occasional pattern for 26.5% and no one considered it a strange one. In
the yes/no sentence with identical tone (see Table 1 above), familiarity increased: 71%
identified it as a fairly or very familiar pattern; and, as with the wh-tune, no one re
garded it as an unfamiliar one. In both examples the first word was the beginning of
the head and, despite the fact that it was a high, sustained head, it was perceived as
familiar, probably due to its short length. Both types of questions occupy identical
ranking (see Table 2 above), with a mean difference of –0.29% (see Table 4).
Focus was identified on the right words (England in the wh- and tomorrow in the
yes/no type) by 62% and 59% respectively, followed at a large distance by arrive (17%)
and free (26%) (see Table 3). A small percentage associated the nucleus with the first
Spanish and English intonation patterns 
Table 3. Focal variability in the ten tone patterns (wh- and yes/no questions)
EXPRESSIONS FOCUS OTHER FOCI N = 34
1. ÁWhy did you do such a stupid thing? 76% 12% 12%
do why did
2. ÁWhen did he arrive in England? 62% 9% 12% 17%
England when did arrive
3. At what time did you say? 15% 18% 12% 55%
what at what time
4. And Áwhat can I do for you? 41% 24% 15% 20%
you what can o
5. How far to ˝Heathrow? 79% 15% 6%
Heath. how far
6. What on earth are you doing here? 68% 20% 12%
doing earth here
7. He’s how tall? 9% 91%
how tall
8. Who else is there to do it? 79% 9% 12%
else who do
9. How can you be so hard-hearted 76% can 6% 3%
father? 15% father hard how
10. If you Ádon’t mind my →asking, where’s 50% 35% 15%
the money come from? asking don’t mind
11. Can I Ásee him if I come back later? 26% 12% 38% 24%
later can see him
12. ÁAre you free tomorrow night? 59% 9% 26% 6%
tomor. are free night
13. Is it my fault you are stupid? 0% 9% 59% 32%
my it fault stupid
14. Is it as Ápopular as all that? 35% 62% 3%
that popul. all
15. Have I had e ˝nough? 82% 15% 3%
enough have had
16. Yes but where can we get hold on you? 91% 6% 3%
hold can on
17. ÁDo the others like it? 29% 12% 47% 12%
others do like it
18. Can I count on Peter? 94% 6%
count can
19. Can I have another piece of cake? 29% can 15% 15%
41% cake have piece
20. If Áit is con→venient, Ácan we start 94% 6%
earlier? conven. it
 Rafael Monroy
Table 4. Comparison between means of identical intonation patterns
Tone patterns WH- Q Mean YES/NO Q Mean Mean Diff.

1. LOW DROP 1 7.21 11 5.15 2.06
2. HIGH DROP 2 6.99 12 7.28 –0.29
3. TAKE OFF 3 5.29 13 6.32 –1.03
4. LOW BOUNCE 4 7.65 14 6.18 1.47
5. SWITCHBACK 5 6.69 15 5.81 0.88
6. LONG JUMP 6 6.10 16 5.74 0.37
7. HIGH BOUNCE 7 5.30 17 7.12 –1.82
8. JACKKNIFE 8 5.08 18 6.52 –1.44
9. HIGH DIVE 9 5.07 19 8.48 –3.41
10. TERRACE 10 4.93 20 7.73 –2.81
elements of the tone units (when, 9%; are, 9%) due to their prominence as heads of
their respective units.
. Tone group 3: Take off
Pattern three is characterised by a low rise movement as typical of English as it is

of Spanish in wh- as well as in yes/no questions. And yet 29.4% of our informants
defined it as scarcely used in Spanish, and 8.8% as a tone unknown in our language; in
sharp contrast, 17.6% defined it as a very common tone (see Table 1 above). This wide
range of discrepancy can be explained in terms of the way people may apprehend the
intonation contour of this pattern. The pitch movement of the nucleus – a low rise –
is quite common in Spanish (Navarro Tomás’ semianticadencia ‘low or mid rise’ 1944),
which explains why 41% consider it a fairly or a very common tone. On the other
hand, the jump from a neutral pre-head (At. . ./Is it. . .) down to the beginning of the
low rise movement of the nucleus is certainly unusual to Spanish ears, which probably
justifies the choice of those who perceive it as a pattern of low or null occurrence in
our language. Notice, though, that this only applies to the wh-question; in the case
of the yes/no question, 23.5% considered that it was a contour rarely used in Spanish,
but no one considered it wholly unusual. In fact, both ranked similarly occupying the
seventh and sixth position respectively (see Table 2). Interestingly, when looking at the
focal elements in both sentences, of all tone patterns, this was the one with the lowest
score: only 15% answered correctly in sentence 3 (what) and no one in sentence 13
(my) (see Table 3 above). This seems to show that Spaniards have greater sensibility
to prominence linked to a higher than to a lower pitch, and that such prominence is
better perceived if the nucleus falls on a wh-element than if it does on other word,
particularly if it is not a content item.
Spanish and English intonation patterns 
Pattern identification 1–10

10,00
8,00
Scores 6,00
4,00
2,00
0,00
1 2 3 4 5 6 7 8 9 10
Intonation patterns
Figure 1. Pattern identification (wh-questions)
. Tone group 4: Low bounce
O’Connor & Arnold (1973) define tone four as a pattern with a low rise nucleus pre
ceded by a high head. English speakers resort to this tone whenever they want to convey
puzzlement or, in the case of yes/no questions, genuine interest. It is quite a common
tone in English as well as in Spanish where it can alternate allotonically with a mid
rise preceded by a mid head (e.g. ¿Es que vienen? ‘Are they coming?’). This explains
why 44% of our informants (Table 1 above) described it as very familiar in a wh-type
question (sentence four). The number declined dramatically, however, in the corre
sponding yes/no type expression (sentence fourteen) where 41% perceived it simply as
a pattern occasionally heard or used, and 14.7% judged it an unfamiliar tone; no one,
however, viewed it as a foreign pattern. The big difference in terms or ranking (first
position in the case of the wh-sentence vs. the seventh in the yes/no question), as illus
trated in Table 3 above and Figures 1 and 2 below, may be due to the influence that
the pre-head exerts on the pattern as a whole: the jump from the neutral pre-head (Is
it as. . .) to the beginning of the following high head is something unusual in Spanish.
The students’ reactions to focal prominence (Table 2 above) show this in a way. In the
yes/no question, 35% assigned the nucleus correctly against 62% who opted for popu
lar – the beginning of the head. In the wh-question, 42% rightly identified the focus of
information (you) vs. 24% who took the beginning of the head (what) as nuclear.
. Tone group 5: Switchback
This tone occurs in English whenever the speaker is “interested and concerned as well
as surprised” (O’Connor & Arnold 1973: 170). Roach comments that this intonation
pattern is “used a lot in English” (1991: 138), being characterised by a fall rise nucleus
preceded by a descending head. This nuclear pitch movement is used in Spanish to
show reluctance or disagreement, which is also a characteristic of the English tone. As
shown in Table 1 above, 53% of the participants agreed that the tone they heard in sen
 Rafael Monroy
Pattern identification 11–20

10,00
8,00
6,00
Scores
4,00
2,00
0,00
11 12 13 14 15 16 17 18 19 20
Intonation patterns
Figure 2. Pattern identification (yes/no questions)
tence five (wh-question) was a fairly or very familiar one, and only 17.6% thought that
it was a tone which was hardly used in the language. This percentage decreased, how
ever, in the yes/no question: 38.2% found the pattern fairly or very familiar vs. 14.7%
who considered it not very familiar. In between, 44.% described it as an occasional
pattern, rarely used or heard while 2.9% deemed it to be a totally foreign pattern.
Arguably, the fall rise constitutes part of the Spanish tonal system, but it has a
restricted use due partly to a narrower range of uses than in English, and also because,
more often than not, it is replaced in ordinary language by a high mid / mid high or
low mid tone, if disagreement is involved (e.g. ¿Guapa? ¡Qué va a ser guapa! ‘Pretty, she
is not pretty!’), or a mid high / high mid if the idea of continuity or emphasis is present
(e.g. Y si quiere entrar, que me pida la llave – ‘If she wants to go in, she should ask me
for the key’). The mean difference (0.88%, Table 4 above) between the two sentences
exemplifying this tone shows that the pattern is perceived differently when occurring
in a wh-question than when occurring in a yes/no question. The former occupies the
fourth position from the top in terms of level of familiarity whereas the latter is eighth
down the scale (see Table 2 above). My impressionistic view is that the reason for this
may lie not so much in the type of head used (descending) as in the fall rise tone as
such: its three pitch movements (high mid high – this last lower than the first) conform
well to a three-syllable word, but the drawling of one of the syllables when affecting a
two-syllable word may produce an effect of foreignness to Spanish ears. Because of
the shortness of both wh- and yes/no questions (only one content word in each, see
Table 3 above), focal prominence presented no difficulty to our informants: the nuclear
element was assigned correctly by 79% and 82% of the informants respectively.
. Tone group 6: Long jump
This pattern is linked to protesting in the two types of questions. The pitch move
ment of the nucleus is a high fall – like tone two – and likewise it is a common tone
Spanish and English intonation patterns 
in Spanish, corresponding to Navarro Tomás’ (1944) cadencia ‘high fall’. As such it is

recognised by 50% of the participants in both the wh- and the yes/no questions. Unlike
pattern two, however, the head presents an ascending movement from a low pitch, re
sulting in a pattern used in Spanish to show surprise rather than protest (e.g. ¿Cuándo
has venido? (high prominence on ni, ‘When have you arrived?). But it is the neutral
pre-head what produces an unusual tune as in sentences six (What on. . .) and sixteen
(Yes but..) (see Table 3 above). This seemingly accounts for the 26% who reckoned it
was a uncommon pattern (14.7% in the case of the yes/no question) or who consid
ered it a foreign intonation contour (5.9% in the wh-type structure and 8.8% in the
yes/no type).
As for their ranking, (see Table 2 and Figures 1 and 2 above), the wh-question
occupies mid position down the scale whereas the yes/no question with identical into
nation tune turns out to be the last but one pattern in terms of familiarity. There is no
apparent reason for this behaviour unless we consider the role played by the head: in
sentence six, earth initiates the head, and its prominence is felt as nuclear by 20%; in
sentence sixteen, however, no one noticed the presence of a head on where, thus result
ing in an anomalous pattern to a Spanish speaker (the corresponding Spanish adverb
is always stressed, i.e. it has high prominence attached to it).
Focus identification was no problem for the great majority of the informants: 68%
placed it correctly on doing (wh-sentence), and 91% targeted it rightly on hold (yes/no
type). Unlike pattern three, characterised by a low pitch movement, here the focus fell
in both sentences on those words with higher pitch (see Table 3 above), thus making
nuclear identification easier.
. Tone group 7: High bounce
The high bounce is a tone whose nuclear movement, a high rise from a mid low or
mid position, is very common in standard Spanish: it is the typical pattern of yes/no
questions as well as of most echo questions. Not surprisingly, 46.7% of our partici
pants corroborated this fact. What is more surprising though, is that only a modest
5.5% would consider it a very familiar pattern and that, in the case of the wh-question,
14.7% did not recognise it as a Spanish tone at all (the highest percentage of non
recognition of all intonation patterns, see Table 1 above). A likely explanation for such
behaviour might be, once again, the effect the head might have on the overall tune:
the English pattern is produced with a sustained high head followed by a jump down
in order to initiate the beginning of the rise in the nucleus. But this is a common tone
in Spanish too (e.g. ¿Trabaja en qué? – He works in what? – qué with a lower ascend
ing pitch than the pre-nuclear component or, alternatively, with a higher level tone).
Another tentative explanation could be the impact that the Spanish rendering of the
English utterance might exert on our listeners: the intonational contour of the English
expression would not fit the one that characterises all the structural components of the
Spanish translation. Thus, in sentence seven (He’s how tall?), the nucleus starts on how
with a mid low ascending pitch, whereas in the equivalent Spanish version (¿Cómo es
 Rafael Monroy
de alto?) cómo is usually produced with a higher pitch. This is confirmed by the way
our participants reacted to focus placement (see Table 3 above). In the expression un
der consideration, only 9% answered it correctly, placing the nucleus on how, whereas
the rest ticked tall as the focal point of the utterance.
The level of familiarity increases ostensibly in the yes/no question (64.7%) – not
surprisingly it is the tone more commonly attached in Spanish to yes/no questions –
followed by 23.5% who considered it an occasional pattern and 8.8% who perceived
it as a rarely used tune (no one saw it a strange pattern). This is manifest too in the
ranking position of the two types of questions: the yes/no type is higher up the scale
(i.e. it is recognised as more familiar) than the wh-type (see Table 2 above). As for fo
cus identification, the students favoured once again the most prominent content word
(like, 47%) at the expense of others (29%), the nuclear element, as shown in Table 3.
. Tone group 8: Jackknife
Tone eight is defined by O’Connor & Arnold (1973: 214) as “challenging and an
tagonistic” when applied to wh- or yes/no questions. Roach (1991: 139–140), who
nonetheless includes it in his five basic intonation patterns, remarks that “it is not
usually considered to be an important tone for foreign learners to acquire”, perhaps
considering rightly that foreigners’ talk is rarely challenging. The nucleus presents an
ascending followed by a descending movement (rise fall) preceded by a sustained high
head. This pattern has a restricted use in standard European Spanish:7 it only occurs
with certain types of statements and exclamations, but scarcely ever with imperatives
(e.g. ¡Pues, claro que sí! ‘Yes, of course!’ – with a lengthening and rising on cla-) nor with
any type of questions. It is not so uncommon, however, in other Spanish accents (e.g.
Cuban or Mexican Spanish) or even in Peninsular local accents (leonés or gallego, for
instance). Our informants’ reactions show that in the case of the wh-question, B and
C options (fairly familiar/familiar respectively) are surpassed by a modest difference
(29.4%, see Table 1 above). The data reveal that this tone is little known (for 5.9%
totally unknown) despite the fact that the final movement of the nucleus (a fall) fully
coincides with the tone that characterises wh-questions in Spanish.
On the other hand, this tune is not common in standard Spanish with yes/no ques
tions. It was this final fall which probably served as a clue for our students to recognise
the pattern as fairly familiar (47%) and even very familiar (17.6%), although a modest
2.9% considered it foreign (see Table 1 above). This raises the question of how a tone
uncommon in the system can be regarded as a familiar one. Admittedly, the tone as
such has a low frequency of occurrence in our language, but there is another pattern,
consisting of a high fall nucleus preceded by a descending (sliding) head, not included
by O’Connor & Arnold within their primary ten tone patterns, which is widely used
in Spanish and which sounds to Spanish speakers very much like the rise fall move
ment of the jackknife. This could explain why the English pattern seems familiar to
Spanish speakers. As shown by their mean difference (–1.44, see Table 4), there is a
noticeable difference between the two types of questions, the yes/no type ranks higher
Spanish and English intonation patterns 
(mid position in the scale) than the wh-type (eighth position). Focal information was
no problem this time: 79% in the case of the wh-sentence, rising to 94% in the yes/no
type, placed it on the right words, which shows once again that the greater prominence
of the rise fall, as was also the case with the fall rise, was unmistakeably perceived by
the large majority of the group.
. Tone group 9: High dive
Unlike the 1961 edition of Intonation of Colloquial English, where O’Connor & Arnold
discussed three compound nuclei (high fall + low rise, high fall + high fall, and high
fall + fall-rise), in the 1973 edition they did away with two of them, keeping the first,
the high dive, consisting of a primary nuclear element followed in a different word by
a secondary rising nuclear tone. This compound tune is in a sense a combination of
patterns two and three. It also has a close resemblance to pattern five – the fall rise.
In the first edition, our authors discuss at length the differences between the two tone
patterns, which they summarise in five main points. They acknowledge, however, that
“all these differences [between the fall rise and the fall + rise] may not be operating”
(1961: 28). As regards the occurrence of a high dive pattern, they point out that this
tone “with questions of any kind is unusual” (1973: 87). It is perhaps this characteristic
that moved the authors to recommend foreign learners not to use it (1973: 87). From a
contrastive standpoint, however, this tone is quite common in Spanish and, unlike En
glish where it conveys considerable emotion taking the form of “plaintiveness, despair
or . . . gushing emotion” (ibid.), the Spanish tone has an inquisitive, casual questioning
import as seen in wh- (e.g. ¿Cuándo dijo que se iba?, rising on i- ‘When did he say he
was leaving?’) as well as in yes/no questions (e.g ¿Quieres que hable con su padre?, rising
on pa- ‘Do you want me to talk to his/her father?’).
What is surprising is that such a familiar tone should be perceived as an occasional
pattern by 41.2% of the students. 32.4% considered it a fairly or very familiar pattern
vs. 26.4 who deemed it to be an unusual tune. This contrasts with their reaction to the
yes/no question. Here there was a large majority (58.8%) who described it as a very
familiar tune and 23.5% as a fairly familiar one, with a mere 5.9% finding it unusual.
For this reason, the yes/no question ranked top of the scale in familiarity (see Table 2
above), whereas the wh-type occupied the last but one position (–3.41 mean difference,
see Table 4 above). As for the focal point, most students (76%) rightly placed it on can
(sentence 9) and 15% on father. In the case of the yes/no question, a smaller percentage
was obtained, 29% choosing can and 41% opting for the final content word cake (see
Table 3 above).
. Tone group 10: Terrace
This tone is unusual in several respects. To begin with, it does not easily fit the defi
nition of tone as “pitch change” (Brazil 1975) on the tonic syllable. Unlike falls, rises,
rise falls, etc., where pitch alteration is self-evident, here we are confronted with a mid
 Rafael Monroy
level pitch with no upward or downward movement. Not surprisingly, not all authors
(Schubiger 1958; or Halliday 1970 for instance) recognise this as a proper tone. Sec
ondly, in O’Connor & Arnold’s typology (1973), it is an unfinished tone which cannot
exist on its own independent of another tone unit. In this respect, it is not possible
to include this pattern in the intonational lexicon of English for those who equate a
tone unit with a sentence. Thirdly, it is a tone almost devoid of an attitudinal import –
unless we take it that “marking non-finality”, because the utterance is inconclusive, is
attitudinal in itself. In fact, non-finality can be accomplished by other tones so long as
a pause is introduced between high or low heads and the nucleus. And yet, level tones
are common in English and they have been recognised as such by some intonologists.
Jassem (1952), for instance, accounts for two level tones in his tonal system: the high
level and the low level, and provides examples of them occurring in autonomous tone
units as in My lady? (high level) or That’s a pity (low level) where no sense of incon
clusiveness is involved. Crystal (1975) includes a level tone in his system which, when
final in a sentence, implies absence of emotional involvement, and when non-final,
conveys a sense of routineness. In his scale of affective involvement this tone “takes the
functions of either fall or rise” (1975: 38).
Very often Spanish too resorts to a level tone, particularly in the speech of the
South Eastern part of Spain.8 But, unlike in English, it is not just a single mid level
tone which is involved: the level tone occurs frequently with different keys (low, mid
or high) whenever the nucleus falls on the final syllable of a tone unit or a monosyllabic
word. One has to compare an expression like ¿Quién es? (‘Who is it?’) with ¿Quién vino?
(‘Who’s arrived?’) to realise that in the former, the nucleus, be it pronounced in a low,
mid or high key, is basically a level tone (alternating allotonically with a low or mid
rise), whereas the latter, due to its disyllabic character, is uttered with a low fall/low
rise/mid high tone, the accented syllable indicating the beginning of the pitch move
ment and the second the end of it. This explains why 75.% of our students find this
tone fairly or very familiar in the yes/no question; in the wh-question, unexpectedly,
64.7% of the sample consider this pitch movement not very common or unfamiliar,
perhaps due to the influence of the long jump pattern of the following tone unit. This
is counterbalanced in the case of focal prominence where the nuclear element in the
yes/no sentence was ticked by 94% – here the shortness of the tone unit might have
played a role; also, the fact that with neutral tonicity, as in this case, the nucleus falls
on the last content word. The longer structure of the wh-question might account for
a greater variation in focal choice. Thus, while 50% of the sample rightly assigned
prominence on asking (see Table 3 above), 35% took the beginning of the head as the
focus of the tone unit.
. Conclusions
Various aspects of intonation have been dealt with from a contrastive perspective. Re
garding nuclear tones, similarities between English and Spanish are clear in the case of
Spanish and English intonation patterns 
simple tones (high fall (high drop and long jump), low fall (low drop), high rise (high
bounce), and low rise (take off, low bounce)). There is a close similarity too as regards
the complex tone (high dive). Compound tones (fall rise (switchback) and rise fall
(jackknife)) are not very common in Spanish, although the fall rise is acknowledged as
very familiar owing to its close likeness to a high mid tone, typical of echo questions in
Spanish. There were mixed reactions as regards the high fall: in wh-questions the rarely
used option prevailed over the rest, whereas in the yes/no question the fairly familiar
one was the more common; perhaps due to its perceptual resemblance to the high fall
tone also common in Spanish. Pre-nuclear patterns are determining in the overall con
tour of an intonation unit. Tones consisting of high heads followed by a fall, a rise, or
a level nucleus were the ones more easily recognised; compound tones trailed behind
but in varying degrees: while the descending head of pattern five was no problem in
the wh-question, it was not clearly identified with the Spanish intonation contour in
the yes/no question, just the opposite of what happened with pattern eight (fall rise).
In terms of pattern ranking, it was tone four (the switchback) which, unexpectedly
(wh-questions favour a low fall in Spanish rather than a low rise), ranked top of the
wh-sentence list, pattern ten being the last one. Surprisingly, the last two tones in the
wh-questions were the first two in the yes/no type, pattern one ranking bottom of the
list. As for focus placement, it was rightly recognised by at least 50% of the students in
the majority of expressions, the exceptions being tones three and eight where, because
of the focus falling neither at the end nor on a content word, it was not identified by
most participants.
Notes
* This is a revised version of a paper read at the XIX AESLA Conference in Barcelona (2001).
My thanks to Antonio Serna Ballester for his help with data extraction from the questionnaire.
. See, among others, Bolinger (1954); Bowen (1956); Cárdenas (1960); Delattre, Olsen, &
Poenack (1962); Stockwell & Bowen (1965); Kvavik (1982); Fant (1984); Cid-Uribe (1989);
Garrido (1991); García Lecumberri (1995); Hidalgo Navarro (1997), etc.
. Monroy Casas (2002).
. O’Connor & Arnold were aware that clause types and communicative functions do not always
match. “What is true [they write]. . . is that some sentence types are more likely to be said with
one tone group than with any other” (1973: 46).
. We consider focus to be the main information carrier of the tone unit corresponding to
Halliday’s tonicity (1970). It is associated with the speaker’s decision to attach greater promi
nence to a given element – which becomes nuclear within the tone unit – in order to treat
something as new information. This does not preclude the fact that all accented words may sig
nal important or new information as held by Schubiger (1958), Bolinger (1972) or Brazil (1981)
among others.
 Rafael Monroy
. According to Ortiz Lira (in García Lecumberri 1995: 195) “there exists, at least in Chilean
Spanish, an intonation pattern consisting of a relatively level (possibly slightly rising) high-
pitched pre-nuclear accent followed by a relatively low-falling nuclear accent”.
. The questionnaire consists of 20 sentences (see Appendix I); the first ten wh-questions corre
spond to O’Connor & Arnold’s ten non-emphatic patterns, the remaining are yes/no questions
repeating correlatively the first ten patterns.
. According to Navarro Tomás, it is used mainly for the expression of affective meanings,
particularly in exclamatives as a substitute for semicadencias (high mid tones) and semianti
cadencias (low or mid rises) in utterance internal melodic units (1948: 159–160).
. We had to postulate up to five level tones in our analysis of the Murciano accent. See Note 2
above.
Appendix I
ENTONACIÓN INGLESA: TEST DE PERCEPCIÓN (Perception test)

1. Escucha las frases siguientes. En la hoja de respuesta has de indicar si el patrón entonativo
que oyes te resulta. (Listen to the following utterances. Tick in your answer sheet the option
that seems to you)
A. Totalmente familiar (lo utilizas mucho y lo oyes también mucho). (Wholly familiar.
You use it and hear it a lot)
B. Bastante familiar (lo utilizas bastante y lo oyes a menudo). (Fairly familiar. You use it
and hear it quite often)
C. Familiar (lo usas algo, y crees que otros pueden usarlo alguna vez). (Familiar. You
occasionally hear it and use it)
D. Poco familiar (no estás seguro de usarlo o de que lo usen otros). (Scarcely used. You are
not sure whether you use it or people may use it)
E. Nada familiar (ni lo usas ni crees que lo usen los hablantes de español). (Totally
unfamiliar. Neither you use it nor you think other people do it)
2. Escucha las expresiones siguientes y subraya la palabra que crees es más prominente en
cada oración. (Listen to the following expressions and underline the word you think is more
prominent in each sentence)
1. Why did you do such a stupid thing?
2. When did he arrive in England?
3. At what time did you say?
4. And what can I do for you?
5. How far to Heathrow?
6. What on earth are you doing here?
7. He’s how tall?
8. Who else is there to do it?
9. How can you be so hard-hearted father?
10. If you don’t mind my asking, where’s the money come from?
11. Can I see him if I come back later?
12. Are you free tomorrow night?
Spanish and English intonation patterns 
13. Is it my fault you are stupid?

14. Is it as popular as all that?
15. Have I had enough?
16. Yes but where can we get hold on you?
17. Do the others like it?
18. Can I count on Peter?
19. Can I have another piece of cake?
20. If it is convenient, can we start earlier?
References
Bally, Charles (1935). Le langage et la vie. Zürich.

Bolinger, Dwight L. (1954). “English prosodic stress and Spanish sentence order”. Hispania, 37,
152–156.
Bolinger, Dwight (Ed.). (1972). Intonation. Harmondsworth: Penguin.
Bolinger, Dwight L. (1986). Intonation and its Parts. Stanford: Stanford University Press.
Bowen, J. Donald (1956). “A comparison of the intonation systems of English and Spanish”.
Hispania, XXXIV, 30–35.
Brazil, David (1975). Discourse Intonation I. Birmingham: English Language Research Mono
graphs.
Brazil, David (1981). “The place of intonation in a discourse model”. In R. M. Coulthard & M.
Montgomery (Eds.), Studies in Discourse Analysis (pp. 146–157). London: Routledge and
Kegan Paul.
Cárdenas, Daniel (1960). Introducción a una comparación fonológica del español y del inglés.
Washington, DC.
Carr, Philip (1999). English Phonetics and Phonology. An Introduction. Oxford: Blackwell.
Cid-Uribe, Miriam E. (1989). “Contrastive Analysis of English and Spanish Intonation Using
Computer Corpora – A Preliminary Study.” Unpublished Ph.D. dissertation. University of
Leeds.
Couper-Kuhlen, Elizabeth (1986). An Introduction to English Prosody. London: Edward Arnold.
Crystal, David (1969). Prosodic Systems and Intonation in English. Cambridge: Cambridge
University Press.
Crystal, David (1975). The English Tone of Voice. London: Edward Arnold.
Delattre, Pierre, Carl Olsen, & Elmer Poenack (1962). “A comparative study of declarative
intonation in American English and Spanish”. Hispania, XLV, 233–241.
Fant, Lars (1984). Estructura Informativa del Español. Estudio Sintáctico y Entonativo. Uppsala:
Uppsala Universitet.
García Lecumberri, Maria Luisa (1995). Intonational Signalling of Information Structure in
English and Spanish. A Comparative Study. Vitoria: Universidad del País Vasco.
Garrido, Juan M. (1991). Modelización de Patrones Melódicos del Español para Síntesis y
Reconocimiento de Habla. Barcelona: Universidad Autónoma de Barcelona.
Gimson, Alfred C. (1980). An Introduction to the Pronunciation of English (3rd ed.). London:
Edward Arnold.
Halliday, Michael A. K. (1967). Intonation and Grammar in British English. The Hague: Mouton.
Halliday, Michael A. K. (1970). A Course in Spoken English: Intonation. London: Oxford
University Press.
 Rafael Monroy
Hidalgo Navarro, Antonio (1997). La entonación coloquial. Función demarcativa y unidades de

habla. Valencia: Universitat de Valencia.
Jassem, Wiktor (1952). Intonation of Conversational English (Educated Southern British). Prace
Wroclawskiego Towarzystwa Naukowego, Seria A. Nr 45: Wroclaw.
Kvavik K. H. (1982). “Spanish multiaccent intonations and discourse functions”. In P. Lantolf
& G. B. Stone (Eds.), Current Research in Romance Languages (pp. 46–62). Bloomington:
Indiana Univ. Ling. Club.
Liberman, Mark (1979). “The Intonational System of English.” MIT Ph.D. Thesis. New York and
London: Garland.
Monroy Casas, Rafael (2002). “El sistema entonativo del español murciano coloquial. Aspectos
comunicativos y actitudinales”. Estudios Filológicos, 37, 77–101.
Navarro Tomás, Tomás (1944). Manual de entonación española. New York: Hispanic Institute of
the United States.
Navarro Tomás, Tomás (1948). Manual de entonación española. Madrid: Guadarrama.
O’Connor, Joseph D. & Gordon F. Arnold (1973). Intonation of Colloquial English (1st ed. 1961).
London: Longman.
Ortiz Lira, Héctor (1994). “A Contrastive Analysis of English and Spanish Sentence Accentu
ation.” Unpublished Ph.D. thesis. University of Manchester.
Roach, Peter (1991). English Phonetics and Phonology (1st ed. 1983). Cambridge: Cambridge
University Press.
Schubiger, Maria (1958). English Intonation. Its Form and Function. Tübingen: Max Niemeyer
Verlag.
Stockwell, Robert P. & J. Donald Bowen (1965). The Sounds of English and Spanish. Chicago:
Tench, Paul (1996). The Intonation Systems of English. London: Cassell.
P VI
Discourse and culture

Emotivity in narrative discourse
Cross-cultural and cross-gender perspectives
Kyoko Takashi and Douglas Wilkerson
The authors analyzed narratives written by American and Japanese subjects, male
and female, to investigate differences related to culture and gender. Maynard
(2002) suggests that meaning is negotiated in a space defined by cognitive,
emotive, and interactional dimensions. While the cognitive features of the
discourse studied were broadly similar for all subjects, the gender and culture of
the subjects showed a high correlation with distinctive strategies of emotivity and
interaction. Differences in the use of terms of reference and of story-telling
conventions were quite marked. The results also suggest that narratives by
females were not, as often claimed, more emotional, but did exhibit a distinctive
emotivity.
. Introduction
It is often reported that “objective” retellings of simple narratives by subjects of dif

ferent gender demonstrate striking differences in presentation. For example, Makino
(1996: 127) reports of Japanese university students’ versions of Cinderella that, while
all of the nine women used conventional opening lines (the equivalent of “Once upon
a time, there was a . . .”) and wrote “just as though they were speaking to children”,
only seven out of eleven men began their tales with such opening lines. He also claims
that the males wrote in a less empathetic manner, abruptly introducing topics without
the familiarizing devices usually associated with children’s stories.
Turning to the reception of such narrations, research has shown that language
users generally hold stereotypical images of what characteristics distinguish male from
female language production in their own (and often in other) cultures. Takashi &
Wilkerson (2003) asked university students in America and Japan (twenty males and
twenty females in each culture) to guess the sex of the author of four anonymous
versions of the same tale, Snow White, written by college students in their native lan
guages, and to give reasons for their choices. The responses of the subjects, regardless
of sex or culture, were remarkably similar. There was a general assumption that women
would be more likely to remember the details of childhood stories, and also more likely
 Kyoko Takashi and Douglas Wilkerson
to make the effort to retell a story in great detail. Stories told in a matter-of-fact man
ner, briefly and with little elaboration, were usually thought to be written by males.
These attitudes evidence a tendency to polarize thinking about the sexes, view
ing males as logical, objective, and referential (information oriented) and females as
emotional, subjective, and representational (oriented toward empathetic creativity).
Such simplistic, two-dimensional analyses tend to concentrate entirely on metaphor
ical usages and what are commonly called emotion words, the province of cognitive
semanticists, but fail to appreciate the role of other expressions of emotivity. This study
examines the sex-related (male versus female) and cultural (Japanese versus Ameri
can) dimensions of some of the other strategies of emotivity deployed in the corpus
studied. Our analysis of the factors which influence the choice and deployment of
these strategies draws on Place of Negotiation theory (Maynard 2002) to put them
in perspective.
. Corpus and methodology
The corpus used in this study consisted of sixty-nine narrations by native speakers
of Japanese (twenty-three male and forty-six female) and forty narrations by native
speakers of American English (twenty of each sex). The subjects were asked to spend
twenty to thirty minutes retelling in writing the story of Snow White (chosen because
most subjects could be expected to be familiar with some version of the story) from
memory, without previously consulting with any sources, whether books or people,
about the story. A black-and-white photocopy of two scenes from the story (Snow
White smiling at the seven dwarves, and the witch with an apple in her hand) was
distributed beforehand in order to stimulate the memory of those few who might have
forgotten the basic plot. In order to examine dimensions of emotive meaning beyond
those mentioned above, we chose several linguistic elements for further analysis. The
nature of the selection process and criteria are explained below.
Maynard (2002: 53ff.) suggests that in Place of Negotiation theory meaning is ap
proached as a result of negotiation:
The Place of Negotiation . . . is bounded and defined as a meaning-negotiating
space. Upon this space, three different dimensions of place are projected, i.e.
cognitive, emotive, and interactional. . .. Cognitive place enables participants to
recognize objects and to construct propositions accordingly.
In <cognitive place> primary significance is assumed by propositional meaning. In our

case, the choice of which characters appear in the retelling and which do not, of who
those characters that do appear are and what they do, belong to the dimension of <cog
nitive place>. <Emotive place>, on the other hand, foregrounds the speaker’s broad
emotional attitudes. This is the space primarily concerned with the psychological and
emotional aspects of communication. What is relatively important is the speaker who
expresses emotional attitude and feelings as he or she incorporates social as well as
Emotivity in narrative discourse 
personal emotions. That is to say, issues in the <emotive place> include emotional
attitudes toward objects and persons, aroused emotional responses, a broad range of
one’s general feelings, as well as cultural sentiment. For example, referring to Snow
White as “S W”, or in Japanese as o-hime-sama (as opposed to hime) reveal subjective
attitudes toward and valuations of the story.
Within <interactional place> “an interactional social atmosphere is created, coor
dinated, and managed while incorporating personal interests” (54). In this dimension
special attention is paid to partnership in speech events, the manner in which speaker
and other participants express, understand, and manage interpersonal relations among
themselves. “Such relations are critical for the negotiation of meaning” (54). In this
sense then, the use of binding strategies such as dialogic structures, expressions of
empathy, or other story-telling conventions would all be part of this dimension.
Figure 1 below shows the six linguistic functions which Maynard associates with
these three dimensions of place, and the linguistic elements we have chosen as relevant
to our analysis. These functional categories are used for the convenience of our analysis
and should by no means be interpreted as mutually exclusive. Many linguistic elements
participate in more than one dimension. As Maynard points out, it is only through the
synergism of all dimensions that language takes on meaning.
In this study of emotivity we will first analyze the terms of reference for the charac
ters which appear in these retellings, in order to examine the attitudes of the narrators.
The extent to which dialogue between characters is included, and the tenor of style
Dimensions of place (type Language functions Aspects of retelling chosen for

of negotiation) analysis
Recognition of objects Choice of characters
Cognitive place
(Informational negotiation) Construction of Relationships among characters
propositions Narrative events emphasized
Narrative voice, point of view
Expression of Terms of reference used
Emotive place emotional attitude Attributes of characters
(Emotive negotiation) Descriptive phrases (adj., adv., etc.)
Communication of Attitude of narrator toward story
attitudes to others Attitude of narrator toward task of
retelling
Management of
Interactional place participatory action Inclusion of dialogue
(Interactional negotiation) Use of fairy tale conventions
Coordination of joint Empathy with listener
utterances
* The column on the far right lists the points of analysis by Takashi & Wilkerson
Figure 1. Dimensions of place and points of analysis
 Kyoko Takashi and Douglas Wilkerson
reveal much about attitudes toward the task assigned to the participants. By looking
at whether or not the narrators exhibit empathy with the listener and use the conven
tions expected in fairy tales, we can learn much about attitudes toward the telling of
such stories, and toward the intended (or implied) audience. These factors cross the
boundaries between <emotive place> and <interactive place> to shape the emotivity
of the narrative.
. Analysis and discussion
Widely held expectations about the differences which males and females are likely
to exhibit in language production are by no means infallible; they can even be ma
nipulated by clever writers to mask their sexual identity, should that be desirable.
Nevertheless, they do provide important clues as to where such differences, if they
really are consistent and widespread within a culture, might be found. Respondents
in our earlier research consistently pointed to two types of evidence upon which they
relied to make gender judgments about language samples. The first seems to be related
to socio-psychology, the second to discourse style.
The first type of evidence is based upon assumptions about the ways in which so
cially influenced (or determined) psychological traits, presumably different for male
and female, would evidence themselves in particular language tasks. For the most part,
this evidence is related to narrator attitudes toward the content of communication.
Presumably based on the belief that males are trained to pay attention to “important”
things and leave “insignificant”, childish affairs to women, many respondents to our
earlier study indicated a belief that women were more likely to remember the details
of childhood stories, and therefore would be better able, and more likely, to include
greater detail in their own narrations. Thus the degree of elaboration was cited as a
clue to gender. Several respondents also reported a belief that females would be more
likely to identify with a female protagonist, such as Snow White, and the romantic in
terest in her story, and would therefore be more likely to refer to the high social status
of the main character (using princess or its equivalent, rather than girl or some other
generic term which does not indicate social status). This assumption, as we shall see, is
not justified by the writers in the corpus which we studied, but is nevertheless reveal
ing of the assumptions about socio-psychological attitudes toward narrative characters
widely held by ordinary readers. An analysis of the terms of reference for the various
characters in the narrative yielded intriguing results; since these were for the most part
related to “the psychological and emotional aspects of communication”, we consider
them as belonging to <emotive place>, and examine them further below.
The second type of evidence concerns discourse style, especially as that is related
to attitudes which the narrator might have toward the act of communication itself.
There was, for example, a widely held assumption that female writers would respond
more positively to what might be seen as an insignificant (to the writer) task, and cer
tainly one from which personal benefit would not likely derive, regardless of how it is
Emotivity in narrative discourse 
performed. (The male is presumed to be more pragmatic and practical, as well as more
selfish and calculating.) Many Japanese feel that the use of the more polite verb mor
phology typified by the use of desu and masu (the desu/masu form) is less harsh and
grating, and shows a greater degree of empathy with the listener. The use of formulaic
elements, especially those typical of the fairy-tale style (suggesting that women are less
individualistic than men and more likely to follow convention), and a greater use of
dialogue, are also associated with styles typical of female language. Again, our earlier
research indicated that many Americans believe that abbreviated names for charac
ters (S. W.) and the use of contemporary slang (e.g. news flash, chow down, break out
as a transitive phrasal verb, psycho-linguistic essence) were more typical of male writ
ers. The enumeration of possible outcomes (necessary because the narrator did not
consider the content sufficiently important to remember the course of the plot) was
seen as showing impatience with the story-telling process, a typically male trait. The
insertion of extraneous comments (What is the correct spelling?, Let’s see if I remem
ber) was also associated with the male narrator, more likely to distance himself from
a childish story and display a cynical attitude toward both the story and the task of
retelling it. This evidence was most intimately associated with the interactional social
atmosphere created by the narrator, and therefore can be seen as belonging to <in
teractional place>. Below we further analyze some of the more prominent features of
<emotive> and <interactional place>.
. <Emotive Place>: Terms of reference
Let us begin our discussion of the characteristics of <emotive place> by examining the
ways in which the narrators refer to the characters in the story, beginning with the pro
tagonist, Snow White. In the previously mentioned study (Takashi & Wilkerson 2003),
twenty percent of Japanese males, and forty percent of Japanese females cited terms of
reference as the basis for their assumptions about the sex of the writer. Respondents
reported that the term hime (‘princess’, with no honorifics) was seen as being rather
impolite (to the character, Snow White), so its use was attributed to male writers; o
hime-sama (‘princess’, with honorific prefix o- and highly honorific suffix -sama) was
thought to evidence idealization of the Snow White character, and thus attributed to
female writers. What of the writing samples themselves?
Japanese narrators used a wide variety of terms to refer to Snow White. How
ever, the proportion of female writers (JF) who chose Shirayuki-hime (‘Princess Snow
White’, the title commonly used for the tale Snow White in Japanese, as well as for the
character herself) as their only term of reference, or as the main term to employ, did
not exceed the proportion of male writers (JM) who did so, as can be seen in the re
sults compiled in Table 1. The same Table also shows the surprising fact that the rather
“rude” hime was used primarily or exclusively by a greater percentage of women, fif
teen percent, than of men, nine percent. A look at the writers who employed hime at
least one time in their compositions (summarized in Table 2) also fails to confirm that
female writers are less likely to employ this more abrupt term. Turning to the idealistic
 Kyoko Takashi and Douglas Wilkerson
Table 1. Terms of reference used for “Snow White” in Japanese
Shirayuki- Mixed reference (Shirayuki-hime, Hime, O-hime-sama, Hime Other

hime only Shirayuki, etc.) only
Mostly Shira- Mostly O-hime- Mostly Shira- Mostly
yuki-hime sama yuki Hime
JF* 23/46 12/46 0/46 2/46 5/46 2/46 2/46
50% 26.1% 0% 4.3% 10.9% 4.3% 4.3%
JM* 13/23 7/23 0/23 0/23 1/23 1/23 1/23
56.5% 30.4% 0% 0% 4.3% 4.3% 4.3%
* JF = Japanese females, JM = Japanese males
Table 2. Proportion of writers employing reference terms for “Snow White”
Shirayuki-hime Shirayuki O-hime-sama Hime Kanojo (“she”)

JF 43/46 2/46 7/46 11/46 7/46
93.5% 4.3% 15.2% 23.9% 15.2%
JM 21/23 1/23 2/23 4/23 1/23
91.3% 4.3% 8.7% 17.4% 4.3%
o-hime-sama, the evidence shows little more than that females (15.2%) were some
what more likely to use the term at least once in their compositions than were males
(8.7%).1
While there were no less than five different titles used in Japanese for Snow White,
each with a distinctly separate level of politeness and formality, the English versions
used only Snow White and the occasional abbreviation S. W. A direct comparison of
these titles for Snow White in English and Japanese would therefore be difficult, and
probably not very productive. However, personal pronouns and other terms of refer
ence may prove to be more worthy of examination. When asked to guess the sex of
the anonymous versions of the tale, seven of twenty females (35%) and six of twenty
males (30%) cited the use of the abbreviation S. W. in English as the clue to the writer’s
male identity. One male respondent indicated that the use of “little girl” in one writer’s
opening, “Once upon a time there was a little girl with cheeks were [sic] as white as
snow, hair black as ebony and lips red as blood”, indicated that the writer was female.
Responding to another composition, one respondent noted that the use of the lady in
the sentence, The stepmother abandoned the lady in the forest, seemed to indicate that
the writer was male.
It is interesting that lexical choice should be chosen by several respondents as an
indicator of the writer’s sex. Aside from more nuanced studies, such as Deborah Tan
nen’s You Just Don’t Understand: Women and Men in Conversation (1990) and Talking
from 9 to 5: Women and Men in the Workplace: Language, Sex, and Power (1994), which
discuss the tendency of the two sexes to use different types of metaphors and different
vocabulary (Different Words, Different Worlds), it is probably apparent to most speakers
of English and Japanese that a small but important number of lexical items are marked
Emotivity in narrative discourse 
for sex. Obvious examples in English would include, for men, a variety of expletives
and crude references to items related to sexual reproduction, and, for women, mild
interjections such as (My) Gracious! or Oh my!. Ordinarily, one would not include the
items cited by these respondents, lady and little girl, in such a list of marked words.
Looking at the English writing samples, there is virtually no difference between male
and female writers in the frequency of such terms as woman (Males (M), five times;
Females (F), four times), lady (M 2; F 1), maid(en) (M 2; F 2), baby (M 2; F 1), or child
(M 3; F 2). However, the word girl (often combined with baby, little, or young, as in
six of the fifteen occurrences here) is used three times as frequently by female writers
(fifteen times) as by male (five times). The question is whether the lexical item girl by
itself is marked, or only its use in certain contexts or collocations. The answer to this is
suggested by the citation, not of girl, but of little girl as being typical of female speech,
and by another curious difference exhibited in the samples, a point to which we shall
return later.
It goes without saying that words like girl, woman, and lady have both denotative
and connotative differences. The latter distinctions are complex and, after looking at
the data given above based on the corpus used in this study (indicating little difference
between male and female usage), for the most part beyond the scope of this paper.
The denotative differences are simpler: girl usually denotes ‘female’ and ‘young(er)’
or ‘immature’, while woman (and lady, in most cases) usually denotes ‘female’ and
‘mature’. As the Snow White character is portrayed as being of different ages at different
stages in the story, this denotative difference accounts for most of the variations. The
story of Snow White as recounted by our writers can be divided into three stages or
acts. The first concerns the birth and infancy of Snow White, and often includes the
desire of her parents to have a child, the wish made by her mother before her birth
(Would that my little daughter may be as white as that snow, as red as that blood, and
as black as this ebony windowframe!), the fruition of this wish, and, frequently, the
death of Snow White’s mother. The second act tells of the command of the queen that
Snow White be killed, and her abandonment in the forest. The final act relates Snow
White’s discovery of the dwarves’ house, her reception there, attempts by the queen to
kill Snow White through the use of magic (poison apple, magic corset, comb), Snow
White’s apparent death, and dramatic recovery. The acts are in chronological order,
though the length of time which elapses between them varies considerably. As the first
act relates Snow White’s birth and infancy, it is natural to use terms such as baby, child,
and girl to indicate her age here. In the final act she is full grown, though still young,
so woman, lady, or even maid(en) are natural choices. Her age in the second act, in
which she is abandoned in the forest, varies from narrative to narrative, depending on
this second act’s temporal proximity with acts one and three: in some versions she is
abandoned while still a baby, in others she is already a young woman at the beginning
of the story, and thus a woman in this scene as well.
Turning to the narrations then, we find that those writers who include the first act
(birth and infancy) in their narrative are most likely to use words such as baby, child,
and girl. In the final scenes, Snow White is mature enough for marriage, so variously
 Kyoko Takashi and Douglas Wilkerson
young woman, beautiful lady, and other combinations frequently appear. The inclusion
of the modifier young often indicates a transition from an earlier act in which she was
described as immature, and simultaneously a temporal proximity. But in the second
act (abandonment in the forest) her age seems to be ambiguous across the corpus;
there is no agreement about the degree of maturity of Snow White at this point in the
story.2
The reason for the queen’s antagonism in most versions of the tale is Snow White’s
beauty. Traditional versions usually emphasize the notion that this is the beauty of a
mature woman by allowing the “truth telling” mirror to answer that the queen is “the
fairest in all the land” as long as Snow White has not grown to womanhood. Only
upon her attaining the full maturity of womanhood, and presumably the perfection of
beauty which comes with it, does the mirror inform the queen that another has out
shone the queen’s beauty. It would, therefore, not be expected that attributes such as
beautiful would play an important part in descriptions of Snow White as a girl. Never
theless, we do find that three female writers combine beautiful with girl: the beautiful
young girl, the most beautiful girl, and a beautiful, kind, and caring girl. Not a single
male writer used any descriptive related to beauty when speaking of Snow White as
a girl. (An equal number of male and female writers made explicit mention of Snow
White’s beauty, so this is probably not a result of any greater tendency among females
to speak of this attribute, though some people believe that females are conditioned to
be more aware of physical appearances, and therefore more likely to mention them.) It
would appear that the use of girl in this scene, in which Snow White’s age and maturity
are not clearly fixed by other events in the story, serves to emphasize not so much the
denotative meaning of ‘immature’, as the connotative meaning, ‘vulnerable’. It is Snow
White’s beauty that threatens the powerful queen, and Snow White’s youth (and purity
of heart) which prevents her from protecting herself. This seems to be borne out by the
fact that another writer (male) describes the huntsman taking the poor girl out to be
killed, and then after a change of heart taking pity on the girl (emphasis by authors).
The girl Snow White is defenseless before the mighty queen, and therefore deserving
of pity. Two of the female writers mentioned above who combined beautiful with girl
included the scene of Snow White’s abandonment, and referred to her consistently as
a girl who was threatened with murder. Even after Snow White’s escape into the woods
and rescue by the dwarves, her life continues to be threatened by the queen. In some
versions the dwarves realize this, and warn her not to go out or allow anyone into the
house lest she expose herself to this threat. When she does so in spite of their warn
ings, one narrator has the dwarves say, “Snow White, you foolish girl. . .” and go on to
wonder whether it was worth their trouble to attempt to save “such a stupid girl”. In
this scene too, the use of girl emphasizes her vulnerability to the queen’s threat.
Perhaps the most revealing use of girl was in a narration written by a male, a
portion of which narrative we include here [emphasis added by authors].
In a fit of jealous rage, the queen summoned her huntsman and ordered him
to take the girl deep into the woods, slay her, and then return with her heart so
Emotivity in narrative discourse 
that the queen might eat it. The hunter went to the woods, but, unable to kill
the girl, abandoned her instead to meet her fate in the wilderness. . .. Snow White,
meanwhile, wandered lost and alone through the woods. Eventually, exhausted,
she came upon a cabin, and as soon as she had entered, she fell asleep there. The
inhabitants of the cabin were seven dwarfs who worked in a nearby mine digging
for jewels. Upon their return and discovery of a strange woman in their house, they
decided that she had come to take their jewels, and that they would have to kill her.
Only her quick thinking saved her, as she offered to do all of their housework for
them while she stayed. They accepted her offer, but warned her against leaving the
house if she valued her life.
This writer uses four different terms for Snow White, a different one for each stage
of the story. At the beginning he makes her the daughter of a kindly king and queen,
and then the stepdaughter of the evil queen/witch. When taken out by the huntsman,
who threatens her life at the command of the queen/witch, she becomes the girl. After
her abandonment, she discovers the house of the dwarves, and falls asleep in (one of)
their beds, at which point she is spoken of as a woman. Later she becomes “the most
beautiful lady in the kingdom”. There is nothing unusual in the fact that she begins life
as a daughter and ends the story as a lady. But it is interesting that she is abandoned
as a girl and found asleep as a woman. According to most versions of the story, the
huntsman was commanded to kill Snow White, but took pity on her (or could not
bring himself to do it), and abandoned her in the forest instead, thinking that this
would accomplish her death just as well. She is saved from death only by discovering
the house of the dwarves. Thus the length of time elapsing between her abandonment
in, or flight to, the forest, and her entry into the house would be as short as a few hours
or one night, or possibly as long as a week or, at most, a month. Thus the switch from
girl to woman is not justified by the passing of narrative time, and the choice of these
words is due to their connotative, not denotative meanings. When her life is threatened
by the huntsman or queen, she is a girl; but when she appears in someone else’s bed,
she is a woman. Girl connotes vulnerability; but what should we make of woman?
A comparison of the terms used to designate Snow White when she is found asleep
by the dwarves is also revealing. Female narrators consistently put a girl in the beds of
the dwarves: upon returning home they look about, “and in one of the beds was a
girl”; or the dwarves “return home to find the sleeping girl”, or are “surprised to see
a girl laying [sic] in their beds”. Hungry, frightened, and so exhausted (from cleaning
house and cooking, according to some of the narrations) that she cannot keep awake,
Snow White is at the mercy of these little men (who in the traditional versions of the
tale were rather threatening and not so cute as Walt Disney later made them), just as
she was at the mercy of the huntsman; females consistently see her as a girl in this
scene. Moreover, as mentioned above, in this tale of a female who loses her mother
at an early age, who becomes the object of a powerful queen’s jealousy, whose life is
threatened many times, and who is driven from home to seek what shelter she can
find in the wild forest, female narrators used the term girl three times as frequently as
did male narrators. The male narrators do not seem to be as sensitively attuned to the
 Kyoko Takashi and Douglas Wilkerson
vulnerability and perils of this poor creature; in fact, when any term other than Snow
White or the neutral she is used in the scene of her discovery by the dwarves, it is Snow
White who appears to threaten the dwarves, who come home to find a strange woman
in their beds. One male writer even asserts that both the hunter (huntsman) and the
dwarves are “seduced by Snow White’s innocent beauty”. (When she reveals her naivety
by allowing the witch into the house, to the dwarves she is once again a foolish girl.)
The evidence from our corpus does not support the assertion that the use of girl is
evidence of female authorship, or that the use of lady is likely to indicate male author
ship, as suggested by some of the respondents. However, it does appear that there is a
clearly marked tendency to use the connotations of these terms to explicate the sex
ual politics implicit in the story, at least from the point of view of twenty-first century
male and female narrators.
So far we have discussed only terms which reflect the age or maturity of a character.
It was also suggested that female narrators would be more likely to use a title like
princess to indicate the high social status of a female character with whom they would
presumably be more likely to identify. We found no evidence to support this. On the
contrary, there were only three direct or indirect (daughter of a king, etc.) references
to Snow White as princess among the narratives written by females. In contrast, it was
a much more common theme among male narrators, who mentioned it eleven times.
There was little in the narratives that might clearly indicate a reason for this, but one
might speculate that, in so far as male writers could be said to “identify” with any
element of the story, they might identify with the prince who awakens Snow White
with a kiss and marries her at the end of the story. The pleasure of the prince would
presumably be the greatest if his consort was not only beautiful, but also powerful
and potentially wealthy, by virtue of her birth and social position. Female narrators,
on the other hand, were much more likely, as we shall see, to situate Snow White in
a network of inter-personal relationships, rather than abstract societal relationships,
which provide motives for actions by the various characters.
Next, our analysis will focus on the terms used for the prince and for the antag
onist, the latter referred to variously as queen, witch, step-mother, and old woman. In
Japanese the prince was referred to most frequently by the terms ouji and ouji-sama,
roughly equivalent to ‘prince’ and ‘his Highness the Prince’. Our research showed that
many Japanese readers felt that the term ‘his Highness the Prince’ (ouji-sama) was
more romantic and feminine. We examined the narrations to see whether or not it
was in fact used more frequently by female narrators. As you can see in Table 3, the
proportion of narrators who used ouji (‘prince’) Primarily or Only was just over thirty
percent (30%) for both males and females. By contrast, the proportion of females who
used ouji-sama (‘his Highness the Prince’), either principally or exclusively, was signif
icantly higher than the proportion of males. (More than a quarter of the males, 26%,
used other, often cynical or facetious, terms for the prince). The term ouji-sama (‘his
Highness the Prince’) is frequently encountered in children’s stories, and imparts a
softer, more empathetic tone to the telling, but we can probably also relate its use to
a tendency to romanticize and idealize the prince who rescues Snow White from her
Emotivity in narrative discourse 
Table 3. Terms of reference used for “Prince” in Japanese
Ouji Both terms employed Ouji-sama

only Primarily Equal Primarily only
Ouji proportions Ouji-sama
JF 14/46 1/46 5/46 1/46 24/46
30.4% 2.2% 10.1% 2.2% 52.2%
JM 6/23 1/23 1/23 1/23 8/23
26.1% 4.3% 4.3% 4.3% 34.8%
predicament. This can also be seen from the fact that all of the female writers (eight,
or 17.4% of the total) who used modifiers such as kakkoii (‘good-looking’), sutekina
(‘handsome’), hakuba ni notta (‘riding on a white horse’), continued with the term
ouji-sama. The proportion of male writers who employed such phrases was actually
somewhat higher than the proportion of females (26.1% versus 17.4%), but only one
single male among the six using such modifiers went on to write ouji-sama.
In the English narrations as well there was evidence of a tendency among females
to include romanticizing phrases such as handsome prince or riding on a white horse
to emphasize the essential role of the prince and his desirability as a marriage partner
(35% of women, vs. 5% for men). On the other hand, the collocation Prince Charming
was used almost equally by men and women (five females and six males). This is in
contrast to the opinion held by our male subjects, who believed that men would not
be likely to use a phrase such as Prince Charming.
Just as the references to the prince seem to evidence something of the narrators’
attitudes towards this tale of rescue and a happy marriage, terms for the queen also
reveal attitudes toward the tale of this antagonist and her dastardly deeds. But the
narrator’s choice of whether or not to use terms for this character which posit a famil
ial relationship with the protagonist Snow White (mother, step-mother, aunt, etc.), or
simply to refer to her as a witch or sorceress and mention no relationship, is also im
portant in establishing a point of view and in setting the voice of narration. Thus these
terms must be seen as negotiating in <emotive place> as well as in <cognitive place>.
Though seemingly simple and brief, these terms of address serve to define interper
sonal relationships and attitudes toward other characters essential to the story, as well
as expressing emotional attitudes toward the propositions being made. Their function
in discourse modality cannot be overemphasized.
Table 4 shows the results of our examination of the terms used for the antagonist.
We include here only the terms used for the queen before she transforms herself into
an old woman in order to trick Snow White and eliminate her by poison or strangula
tion. We have chosen five of the most common terms, and show here the number and
proportion of narrators who used those terms at least once. Of particular interest is the
high proportion of female narrators who used the terms jouou (‘queen’) or Shirayuki
Hime no [mama]haha (‘Snow White’s (step)mother’); among the American narrators,
females (AF) were twice as likely to use these terms as males (AM), while in Japan fe
males (JF) were more than three times as likely to use them as Japanese males (JM). In
 Kyoko Takashi and Douglas Wilkerson
Table 4. Terms of reference for “Queen” (number of writers who used each term at least
once)
JF JM AF* AM*
Queen 17/46 2/23 15/20 7/20
37.0% 8.7% 75% 35%
(jealous/wicked) 12/46 2/23 8/20 3/20
(Step-) Mother (of Snow White) 26.1% 8.7% 40% 15%
(evil/old) 19/46 11/23 6/20 14/20
Witch/Conjuror/Sorceress 41.3% 47.8% 30% 70%
(mean) 4/46 2/23 0/20 0/20
Old lady 8.7% 8.7% 0% 0%
(vain, jealous) 1/46 0/23 6/20 4/20
Woman 2.2% 0% 30% 20%
* AF = American females, AM = American males
Table 5. Mention of jealousy and allusion to “Queen” or “(Step)mother”
JF JM AF AM
Those using Queen or (Step)mother who mention jealousy 27/27 10/10 16/16 9/10
100% 100% 100% 90%
Those not using Queen or (Step)mother who mention jealousy 10/19 2/13 2/4 5/9
52.6% 15.4% 50% 55.6%
Total respondents who mention jealousy 37/46 12/23 18/20 14/19
80.4% 52.2% 90% 73.7%
contrast, while about 70% of American males used the term ‘witch’/majo, only forty
to fifty percent of Japanese narrators used it, and only 30% of American females.
There also seems to be an interesting connection between the use of terms indi
cating a, usually, familial relationship to Snow White, and the mention of a rivalry
between the queen and Snow White to be the most beautiful woman in the land. Table
5 shows the proportions of those mentioning the rivalry among those who allude to
the relationship and those who do not.
It is evident that nine out of ten American females, and eight out of ten Japanese
females, attribute the threat against Snow White’s life to jealousy because of her supe
rior beauty and purity; only seven out of ten American males mention this motivation,
while among Japanese males the proportion drops to one in two. It is often suggested
that in many cultures women are taught to be more concerned about their personal
appearance than men, and thereby are trained to be more sensitive to questions in
volving personal attractiveness; when there is a tendency to see one’s future success or
happiness (whether in a career, a marriage, or as a mother) as influenced by this physi
cal attractiveness, there is also a tendency to compare one’s physical appearance to that
of other women, and consequently to be more interested when such comparisons play
a part in a story or other situation. While almost all women remembered and com
mented on this aspect of the story, it would appear that about a third of the American
Emotivity in narrative discourse 
male respondents did not remember that female rivalry was the cause of the queen’s
hatred for Snow White, and half of the Japanese men either forgot or failed to men
tion this essential element of the story. While virtually all of those who remembered
some sort of connection between Snow White and the queen attributed the queen’s
hatred to this cause, those who portrayed the two as unrelated were equally likely to
attribute Snow White’s predicament to some other cause (the queen’s evil nature, a
falling out between Snow White and her father, etc.) or, particularly among Japanese
males (presumably the least sensitive to issues of concern to women) to no cause at all.
Even when using similar terms of reference for the queen, Japanese males and
females tended to use different forms of the term. While male narrators were more
likely to use the plain forms haha (‘mother’), mamahaha (‘step-mother’), ouhi (‘em
press’), etc., women used the polite forms of these terms, okaasan (‘mother’), okaasama
(‘honored mother’), okisakisama (‘queen’), in most cases adding the honorifics o- and
-san/-sama. In the case of the prince, the use of honorifics served to elevate the prince,
putting him on a pedestal as it were; in the case of the queen, these same honorifics
work in a very different way. In contrast to the more objective ouhi used by the Japanese
men, many of whom did not mention the jealousy of the queen, the okisakisama used
by the women emphasizes the ruthless power of the queen and the many forces she
had at her disposal to carry out her evil plans. Even a slight inflection conveys a great
deal about the subjective emotions of the narrator and repays well the effort put into
analyzing the modality of these terms of reference.
. Use of story-telling conventions
In order to determine the degree of empathy demonstrated by the narrator, and the
extent to which traditional story-telling conventions were followed, we looked at the
following features:
1. whether or not conventional opening and closing phrases were used;
2. whether or not the narrator prepared the listener adequately by introducing the
main characters;
3. use of dialogic structures; and
4. other aspects of style employed.
Table 6 shows the proportion of respondents who began their narrative with the con
ventional phrase “Once upon a time there was a . . .” (or in Japanese Mukashi, mukashi
aru tokoro ni). Included as Relatively Traditional are those openings that use any one
of the elements of this opening, such as “Once upon a time” (mukashi mukashi or
mukashi) or “There once was a . . .” (aru. . . ni). It might appear that (Japanese) male re
spondents were more traditional in their openings than females, but the totals for both
Traditional and Relatively Traditional show little difference between male and female
in either culture. However, a look at Table 7 will reveal that, among those respondents
who did not use one of these conventional openings, males of both cultures were much
more likely to begin in media res than were females. Even when not employing a rela
 Kyoko Takashi and Douglas Wilkerson
Table 6. Opening and closing passages
(Relatively) traditional opening Traditional closing

Traditional Relatively traditional Conclusive closing Conventional
ending phrase
“Once upon a “Once upon a time” “They lived “The End”,
time there was “There once was. . .” happily ever after”, etc.
a. . .” etc.
JF 7/43 (16.3%) 18/43 (41.9%) 42/46 (91.3%) 11/46 (23.9%)
Total 25/43 (58.1%)
JM 6/23 (26.1%) 6/23 (26.1%) 19/23 (82.6%) 3/23 (13.0%)
Total 12/23 (52.2%)
AF 2/20 (10.0%) 7/20 (35.0%) 18/20 (90.0%) 4/20(20.0%)
Total 9/20 (45.0%)
AM 1/20 (5.0%) 7/20 (35.0%) 11/20 (55.0%) 3/20 (15.0%)
Total 8/20 (40.0%)
Table 7. Non-traditional openings and degree of empathy
Non-traditional opening, empathetic style Non-traditional opening, abrupt style
“There was a girl whose name was. . .” etc. “Snow White was walking. . .” etc.
JF 6/18 (33.3%) 12/18 (66.7%)
JM 0/11 (0%) 11/11 (100%)
AF 3/11 (27.3%) 8/11 (72.7%)
AM 0/12 (0%) 12/12 (100%)
tively conventional opening, at least a quarter of the females in both Japan and America
began by in some way introducing an important character, using phrases such as “A
girl named Snow White. . .”. By contrast, all of the males, without exception, who did
not make use of a traditional opening jumped into the story with phrases like “Snow
White’s step-mother was. . .” or “Snow White was walking. . .”, apparently assuming
that the listener would already be familiar with the character of Snow White. Further
we find that nearly half of the Japanese women (47%) and a third of the American
women (35%) relate the origin of Snow White’s name to her skin color or beauty, fur
ther familiarizing her character to the intended audience, while only 15% of American
males, and not a single Japanese male employed this strategy.
With regard to ending the story (also shown in Table 6), we find that women in
both cultures were significantly more likely to add a conclusive ending (“And they lived
happily ever after.”) than were men (for American narrations, 90.0% of women com
pared to 55% of men). Women were also somewhat more likely to add conventional
ending phrases (“The End”) than were men. Thus we can see that in both America and
Japan there is a tendency for female narrators to exhibit greater concern for the audi
ence by ushering them into and out of the world of the narrative, through conventional
openings, explanatory passages, and formal closure.
Emotivity in narrative discourse 
Table 8. Employment of dialogue (direct and indirect quotations)
Not Used 1 Used 2 Used 3 Used 4 Used 5 Used 6 Used 6+ Usage

used time times times times times times times totals
JF 25/46 4/46 13/46 1/46 2/46 0/46 1/46 0/46 21/46
54.3% 8.7% 28.3% 2.2% 4.3% 0% 2.2% 0% 45.7%
JM 15/23 0/23 6/23 1/23 1/23 0/23 0/23 0/23 8/23
65.2% 0% 26.1% 4.3% 4.3% 0% 0% 0% 34.8%
AF 2/20 1/20 2/20 2/20 4/20 3/20 1/20 5/20 18/20
10.0% 5.0% 10.0% 10.0% 20.0% 15.0% 5.0% 25.0% 90.0%
AM 7/20 3/20 3/20 2/20 2/20 0/20 2/20 1/20 13/20
35.0% 15.0% 15.0% 10.0% 10.0% 0% 10.0% 5.0% 65.0%
Moving on to the use of dialogue in the narratives, the results of our examina
tion are presented in Table 8, which shows the number of respondents who employed
dialogue one or more times. It can readily be seen that females were more likely to
use dialogue in their story telling than males, and that Americans were more likely
to use dialogue than Japanese. Not only were female and American narrators more
likely to use dialogue at least once, they also exhibited a greater tendency to employ
it repeatedly. The proportion of Japanese respondents who employed it three or more
times was less than nine percent (8.7%), while American males were four times more
likely to employ it frequently (35%), and American females twice as likely as Ameri
can males (75%). Figures for the mean and maximum number of instances for these
groups shows the same pattern: American females used dialogue on the average nearly
five times (4.9) as opposed to the male average of two (2.1), the respective maximum
usages being fifteen and seven. For Japanese narrators, the averages were 1.02 (female)
and 0.83 (male) and the maximums six and four respectively. Table 9 shows the por
tions of the story in which dialogue was most likely to be employed. For all groups,
the queen’s interaction with her magic mirror (Table 9, A through D) was the episode
most likely to be graced with dialogue. The poetic quality of the phrase used by the
queen, “Mirror, mirror on the wall,/ Who is the fairest one of all?” undoubtedly con
tributed greatly to this result. The relatively large number of Japanese respondents who
included dialogue between the queen, disguised as an old woman, and Snow White
(Table 9, G, H) may have been influenced by the photocopies which were distributed
to stimulate the memory, one of which depicted this very scene. That American fe
males employed dialogue most frequently, and used it in a greater variety of situations,
even some not mentioned at all by any of the other groups (Table 9, F), would seem
to indicate that they either remembered the story more clearly, or that they were more
likely to elaborate on even the less well remembered scenes (or both). Related to this,
it is also interesting to note that, while none of the Japanese narrators mentioned the
names of the dwarves (as given in the Disney movie), 30% of the American males and
60% of the American females included at least some, if not all, of the names of the
dwarves in their retelling.
 Kyoko Takashi and Douglas Wilkerson
Table 9. Nature of dialogue used
Spoken by>Addressed to, “Quotation” JF JM AF AM
A. Queen>Mirror, “Mirror, mirror on the wall. . .” 3/46 1/23 15/20 10/20

(First time) 6.5% 4.3% 75.0% 50.0%
B. Mirror>Queen, “You, O Queen, are. . .” 3/46 1/23 8/20 7/20
6.5% 4.3% 40.0% 35.0%
C. Queen>Mirror, “Mirror, mirror on the wall. . .” 16/46 7/23 7/20 4/20
(Second time) 34.8% 30.4% 35.0% 20.0%
D. Mirror>Queen, “Snow White is. . .” 13/46 7/23 16/20 11/20
28.3% 30.4% 80.0% 55.0%
E. Queen>Huntsman 3/46 0/23 4/20 3/20
6.5% 0% 20.0% 15.0%
F. Huntsman>Snow White 0/46 0/23 8/20 0/20
0% 0% 40.0% 0%
G. Queen>Snow White, “Would you like a lovely 7/46 2/23 2/20 0/20
apple my dear,. . .” 15.2% 8.7% 10.0% 0%
H. Snow White>Queen (Old Woman) 3/46 1/23 2/20 0/20
6.5% 4.3% 10.0% 0%
I. Prince>Dwarves 1/46 0/23 3/20 0/20
2.2% 0% 15.0% 0%
J. Snow White>Dwarves 0/46 0/23 4/20 1/20
0% 0% 20.0% 5%
K. Dwarves>Snow White 0/46 0/23 5/20 2/20
0% 0% 25.0% 10.0%
L. Dwarves>Prince 2/46 0/23 1/20 0/20
4.3% 0% 5.0% 0%
M. Other 2/46 0/23 2/20 1/20
4.3% 0% 10.0% 5.0%
Turning now to style, let us first say something about conjugations in Japanese,
and their importance in determining style. In modern Japanese, conjugations are usu
ally divided into two groups, those following the da paradigm (plain), and those
following the desu/masu paradigm (polite). Clancy (1984: 56) initially suggests that
the choice of one of these paradigms “often seem[s] to be mandated by the medium
of communication rather than any characteristic of the speaker, hearer, or topic”. She
does, however, come to realize that the choice is not so simple, especially in the case of
narration:
It is apparent that narrators did not always make the same choices in a given
medium; some writers used polite forms, while certain speakers used plain
forms. . ..Thus the level of politeness is something that is negotiated between
speakers and hearers during the course of an interaction. (Clancy 1984: 59)
Concerning factors which may influence this negotiation, Maynard (1991: 578) notes
that an “author’s motivation for the [choice between the] da and desu/masu verb mor
phology is in fact founded in the Japanese speaker’s sensitivity toward ‘thou’. The da
Emotivity in narrative discourse 
Table 10. Style used in narration of “Snow White” (Japanese)
Mostly polite Mixed Mostly plain Other
JF 27/46 (58.7%) 3/46 (6.5%) 15/46 (32.6%) 1/46 (2.2%)
JM 11/23 (47.8%) 0/23 (0%) 9/23 (39.1%) 3/23 (13%)
Table 11. Characteristics of polite style users (Japanese)
Conventional opening Conventional closing Use of dialogue

JF (N = 27) 26/27 (96.3%) 27/27(100%) 19/27 (70.4%)
JM (N = 11) 11/11 (100%) 10/11 (90.9%) 4/11 (36.4%)
Table 12. Characteristics of plain style users (Japanese)
Conventional opening Conventional closing Use of dialogue
JF (N = 15) 4/15 (26.7%) 13/15(86.7%) 8/15 (53.3%)
JM (N = 9) 3/9 (33.3%) 5/9 (55.6%) 3/9 (33.3%)
ending represents the speaker’s perspective internal to the narrative setting”. The au
thor “is there and then” and expresses feelings and makes comments (including those
of interior monologues) immediately. The desu/masu ending takes the speaker’s per
spective and is not necessarily internal to the scene, but rather takes the position of
self-awareness of talking to the ‘thou,’ as in reading a story to an audience (Maynard
1991: 580). Clancy also remarks that “[t]he switch to polite endings . . . supports the
view that morphological politeness in Japanese is strongly responsive to the dimension
of ‘involvement”’ (1984: 60). Maynard also adds that “in predominantly desu/masu
discourse, the da ending often marks backgrounded information subordinate to the
overall structure of discourse.”
Based on the expectation that the desu/masu form will predominate in situa
tions requiring polite patterns of speech, the Japanese narrations can be divided into
three categories, depending on whether they employed desu/masu (Mostly Polite) or da
(Mostly Plain), or mixed the two more or less evenly (Mixed). Included in the “Mostly
Polite” category were samples which used da once, or included occasional sentence
fragments (and thus had no predicate to conjugate). Similarly, the “Mostly Plain” cat
egory also includes samples which use the desu/masu or some other form once. The
results in Table 10 show that nearly sixty percent of women (58.7%) used the polite
form, while less than fifty percent of men (47.8%) did so. In order to see whether or
not this polite style could be considered the standard style for story telling in Japanese,
we also looked at the correlation between its employment and the use of other story
telling conventions. As seen in Tables 11 and 12, more than ninety percent of all those
who used conventional opening and closing lines also employed this style, regardless
of sex, but only a third of those using the plain form began with a conventional open
ing. This is consistent with Clancy’s (1984: 61) claim that by using polite verb endings
 Kyoko Takashi and Douglas Wilkerson
(desu/masu), writers can affect a special jidoo bungaku (‘children’s literature’) style,
typical of the writing found in storybooks meant to be read to very young children, or
primers written for elementary school children. This study also looked into connec
tions between style (Polite versus Plain) and the use of dialogue, but failed to find any
significant correlation between the two.
Pronouncements about style in English were considerably more difficult to make.
There was no simple test, such as a choice of conjugation paradigms, which could serve
definitively to distinguish which narrators strive to engage the listener in a community
of negotiation and which do not. Wallace Chafe suggests the broader term involve
ment, describing an awareness “of an obligation to communicate what he or she has in
mind in a way that reflects the richness of his or her thought . . . to enrich it with the
complex details of real experiences”, which we might use to measure several strategies
of engagement, and use as an index of style in English. In his article “Integration and
Involvement in Speaking, Writing, and Oral Literature”, Chafe examines some of the
characteristics which generally distinguish written language from spoken language in
literate cultures. Aside from a lower degree of language integration, a factor which is
beyond the scope of the present paper, he mentions “‘involvement’ with the audience”
as being typical of a speaker, and “detachment” from the audience as typical of a writer
(1984: 45). Stylistic features which give evidence of involvement include “references to
a speaker’s own mental processes” (46), “particles expressing enthusiastic involvement
in what is being said, like just and really” (47), “vagueness and hedges” (what he calls
“fuzziness”, demonstrated by phrases such as and so on, something like, sort of ), and di
rect quotations (48). Conversely, a higher degree of language integration (“the packing
of more information into an idea unit”) and detachment characterize written language,
and, interestingly enough, in cultures without a literary tradition, ritual language.
In the context of this study it might be interesting to ask what characteristics
written story telling, which seems to fall between ordinary colloquial language and
written/ritual language, might be expected to exhibit, and what stances story telling is
likely to take in <interactional place>. Of ritual language Chafe writes:
[t]he same oral ritual is presented again and again, not verbatim, to be sure, but
with a content, style, and formulaic structure which remain constant from perfor
mance to performance. A piece of ritual language is something which is valued,
and which is repeated because of its value. (49)
In so far as traditional stories are “presented again and again, not verbatim, to be sure”,
but with substantially similar “content, style, and formulaic structure”, they may be
considered a type of ritual language, and could be expected to share the characteristics
of ritual (and written) language. The narrations in our corpus which seemed to come
closest to the traditional story-telling style avoided “references to a speaker’s own men
tal processes”, “particles expressing enthusiastic involvement”, and “fuzziness”, all cited
as evidence of involvement, yet were marked by the inclusion of “direct quotations”
(in Japanese this was paralleled by the use of the desu/masu verb morphology). They
exhibit, then, a highly selective use of the range of potential strategies of involvement
Emotivity in narrative discourse 
to negotiate the <interactional place>. Moreover, we may divide the strategies above
into those which serve to engage the speaker in this <interactional place>, namely ref
erences to the speaker’s mental processes, enthusiastic interjections, and fuzziness, and
those which serve to engage the listener in this <interactional place>, namely direct
quotation, and, in Japanese, the polite desu/masu forms. While English and Japanese
narrators tended to rely on different strategies, with English speakers using direct
quotation (in the form of dialogue) and Japanese speakers employing the polite verb
morphology (a choice not really available in English), females in both cultures exhib
ited a markedly stronger tendency to use only those strategies which engage the listener
while carefully avoiding those which engage the speaker; male narrators, again in both
cultures, were more likely to use those strategies which engage the speaker, sometimes
neglecting those which engage the listener. The two groups of narrations in English
exhibited roughly the same inclination to employ strategies which engage the listener
(about 45% of each group did so consistently), but in the differential deployment
of speaker-engaging and listener-engaging strategies of involvement male and female
differences in English became abundantly apparent. Only one female (5%) deployed
strategies of speaker engagement, while six male narrators (30%) did so, frequently in
terrupting the story with references to their own lapses of memory, uncertainty about
the nature of various characters and their actions, and other comments irrelevant to
the progress of the story.
. Conclusion
The discussion above points to conclusions in three areas: (i) the passive use of con
ventional phrases, (ii) the use of emotive words and phrases, and (iii) the static nature
of earlier analyses.
Makino’s (1996) earlier study seemed to conclude that the use of conventional
phrases, such as opening a story with “Once upon a time there was a. . .”, is more typical
of Japanese female language production and could be used to distinguish it from male
language production. Let us refer to this repetition of set expressions, whether opening
and closing phrases or other clichés such as Prince Charming, as a “passive use” of
convention. Directly contradicting widely held assumptions, the women in our study,
regardless of culture, were no more likely than men to make passive use of traditional
opening phrases. If anything, men were occasionally slightly more likely to do this than
women. On the other hand, female narrators were more likely to use a wider range of
strategies of listener engagement, including explanatory introductions of characters,
devices for formal closure, dialogue, and polite verb forms.
The language of women has frequently been criticized as being more emotional
and subjective than that of men. The basis for such claims usually comes from the per
ception that women are more likely to include in their speech words which connote
enthusiasm or convey positive feelings or attitudes, but which add little to the propo
sitional content of their discourse, elements of style which we have called speaker, as
 Kyoko Takashi and Douglas Wilkerson
opposed to listener, engagement. But using this as a standard for judgment, the narra
tions written by females were not significantly different from those written by males;
females were no more likely to include trite phrases like beautiful, charming, hand
some, or sweet than were male writers. If anything, our research might suggest the
opposite. What did mark female style was a greater degree of consistency and even
ness, remarkable, not for its enthusiastic outbursts or abundance of emotively charged
words, but for restraint. The male style, on the other hand, was far more likely to en
gage in strategies of speaker engagement, frequently punctuating the narration with
words, phrases, and whole lines of commentary expressive of the speaker’s own neg
ative emotion, rather than employing strategies of listener engagement which might
elicit or express sympathy or share interest. Our earlier research has also shown that
this style is by far the most readily identified (with male writers) and easily described
by readers attempting to determine the sex of an unknown writer.
This is not to deny the emotionally charged and subjective nature of much of
women’s discourse. Rather, we conclude that this is more clearly expressed in ways that
involve the listener in <interactional place>. Female writers were considerably more
likely to locate the motive force of the story in familial and interpersonal relations,
relating Snow White’s predicament to her relationships to the queen and the queen’s
jealousy of her beauty. Clancy (1984), based on a study of narratives in which young
Japanese were asked to tell what they had seen in a short film, has noted that some
female narrators who used polite verb forms in their narrations, thereby acknowl
edging a personal relationship between narrator and listener, went so far as to invent
personal relationships between characters in the story where none had before existed.
No narrators who did not use polite verb forms exhibited this tendency, suggesting
a connection between narrator-listener relationships and interpersonal relationships
among characters within a narration (1984: 61). Male narrators, more likely to flout
the expectations of <interactional place>, were far more likely to ignore the interper
sonal complexities of the story, and reduce the witch’s motivation to her evil nature,
universally directed toward the innocent and weak. The specific strategies of choice
for the managing of this <interactional place> also reflect most strongly the cultural
difference between America and Japan. Although deprived of any immediate feedback,
females in both cultures were more likely to use interactional strategies of engagement.
American females elaborated a larger number of scenes, and with more dialogue, than
did males. In contrast, even without a clearly defined relationship to their “listeners”,
female narrators from Japanese culture, in which people are held to be highly sensi
tive to the different demands imposed by specific interactional contexts, turned to the
polite style of speech (polite verb morphology) with its recognition of the physically
absent thou. New methods of analysis often point out problems with the older emo
tional/rational distinction by highlighting the fact that not only female discourse but
also much male discourse is emotional and subjective in nature as well. We cannot
support the traditional distinctions between male and female discourse on the basis of
our analysis, but do point out several other important distinctions which suggest that
Emotivity in narrative discourse 
there is still a clear qualitative, if not quantitative, difference between male and female
voices in industrialized cultures at the beginning of the twenty-first century.
Notes
. To support their arguments, the authors have made an effort to isolate elements of style and
language use which are easily quantifiable, as summarized in the tables, rather than relying on
primarily subjective criteria, and believe that their conclusions accurately reflect actual differ
ences in the samples gathered for this research. At the same time, it should be pointed out that
this discussion of stylistic and cultural differences examines relatively fine distinctions and barely
visible trends in largely similar productions, and aims to elucidate qualitative rather than quan
titative differences. The statistical significance of these conclusions awaits confirmation in future
research.
. In the animated Disney version, upon which many of the writers based their retellings, Snow
White is already mature, a young woman, so the prince can fall in love with her before she departs
from the castle, to be abandoned in the forest; in the Grimm version, though only seven, “she
was as bright as the day, and fairer than the queen herself ”.
References
Chafe, Wallace (1984). “Integration and involvement in speaking, writing, and oral literature”. In
D. Tannen (Ed.), Spoken and Written Language: Exploring Orality and Literacy (pp. 35–53).
Norwood, NJ: Ablex.
Clancy, Patricia M. (1984). “Written and spoken style in Japanese narratives”. In D. Tannen (Ed.),
Spoken and Written Language: Exploring Orality and Literacy (pp. 55–76). Norwood, NJ:
Ablex.
Makino, Seiichi (1996). Uchi to soto no gengobunkagaku (Sociolinguistics of uchi and soto).
Tokyo: Arc Press.
Maynard, Senko K. (1991). “Pragmatics of discourse modality: a case of da and desu/masu forms
in Japanese”. Journal of Pragmatics, 15, 551–582.
Maynard, Senko K. (2002). Linguistic Emotivity. Amsterdam and Philadelphia: John Benjamins.
Takashi, Kyoko & Douglas Wilkerson (2003). “Characteristics of young adult male and female
narratives: investigation of images held by Japanese and American college students”.
Unpublished manuscript.
Tannen, Deborah (1990). You Just Don’t Understand: Women and Men in Conversation. New
York: William Morrow.
Tannen, Deborah (1994). Talking from 9 to 5: Women and Men in the Workplace: Language, Sex,
and Power. New York: Avon Books.
Cardinal Transitivity in foregrounded
discourse
A contrastive study in English and Spanish*
Pilar Guerrero Medina
This paper, concerned with the relationship between grounding and degree of
Transitivity, questions the validity of the foreground/background distinction as
the unitary pragmatic principle embracing the semantic and morphosyntactic
manifestations of the scalar notion of Transitivity. The author investigates how
the individual Transitivity components correlate with the “independent”
discourse notions of foregrounding and backgrounding as postulated by Hopper
& Thompson (1980). The analysis of the selected narrative passages in the two
languages studied shows that high Transitivity features do not always co-occur in
foregrounded discourse. The author concludes that there seems to be an inherent
circularity in the association between foregrounding and high Transitivity, on the
one hand, and in the postulated correlation between foregrounding and
temporal sequentiality, on the other.
. Introduction
Hopper & Thompson (1980) conceive of Transitivity as a semantic continuum, a scalar

notion determined by a number of independent components, which correlate in the
grammar of natural languages. As illustrated in Table 1 (where “A” stands for “Agent”
and “O” for “Object”), Transitivity is defined as a composite notion, consisting of ten
semantic-grammatical features.
These ten parameters, whose aggregate results in the cardinal (or prototypical)
Transitivity of a clause, are presented as having a unified discourse value, since they
contribute to construct the temporal structure of a discourse:1 high Transitivity is
correlated with foregrounding, while low Transitivity characterizes backgrounding.
Each of the ten parameters, as described in Hopper & Thompson, measures the
degree of effectiveness with which the action is transferred from one participant to
the other:
A. Number of participants involved
High transitivity is associated with the presence of two (or more) participants in
the clause.
 Pilar Guerrero Medina
Table 1. Hopper & Thompson’s parameters of Transitivity
High Transitivity Low Transitivity
A. PARTICIPANTS 2 or more participants 1 participant

B. KINESIS action non-action
C. ASPECT telic atelic
D. PUNCTUALITY punctual non-punctual
E. VOLITIONALITY volitional non-volitional
F. AFFIRMATION affirmative negative
G. MODE realis irrealis
H. AGENCY A high in potency A low in potency
I. AFFECTEDNESS OF O O totally affected O not affected
J. INDIVIDUATION OF O O highly individuated O non-individuated
B. Kinesis
The Kinesis parameter is defined in terms of the opposition action vs non-action.
Only physical actions (e.g. I hugged Sally) can be effectively transferred from one
participant to another.
C. Aspect
The feature of Aspect, defined in semantic terms, refers to the telicity (or perfec
tivity) of the action. Telic predicates (i.e. predicates which specify “an endpoint or
conceptual boundary”, p. 285) are higher in Transitivity than atelic predicates.
D. Punctuality
Hopper & Thompson distinguish between discourse-based Aspect, in the sense of
telicity/perfectivity, and lexical aspect (or “Aktionsart”), which “comprises those
manners of viewing an action which are predictable from the lexical meaning
of the verb, such as punctual and durative” (p. 271). Punctual actions (i.e. “ac
tions carried out with no obvious transitional phase between inception and com
pletion”, p. 252) are regarded as more effectively transferred to their patients
than non-punctual (or durative) actions (i.e. “actions which are inherently on
going”, p. 252).
E. & H. Volitionality & Agency
These two parameters concern the “degree of planned involvement of an A in the
activity of the verb” (p. 286). An action is more effectively transferred to the pa
tient if the agent is “volitional” (i.e. “presented as acting purposefully”, p. 252).
The presence or absence of the volitional element is associated with the “deliber
ateness or spontaneity of the Agent” (p. 264).
Hopper & Thompson (p. 265) associate Agency with the “the degree of con
trol” exercised over the activity of the verb. The authors (p. 273) hold that the
Transitivity of the clause is reduced when the Object is higher than the Agent
on the following hierarchy of what Silverstein (1976: 113) calls “inherent lexical
content” of noun phrases:
(1) 1st p(erson) < 2nd p. < 3rd p. < proper name < human < animate < inanimate
Cardinal Transitivity in foregrounded discourse 
Although the (E) and (H) factors frequently co-occur (volitional verbs usually
require agentive subjects), these two properties can also be separated, since highly
agentive subjects are not necessarily deliberate instigators of events (Hopper &
Thompson 1980: 286).
F. Affirmation
The Affirmation parameter is defined in terms of the opposition affirmative vs
negative: affirmative clauses are higher in Transitivity than negative clauses.
G. Mode
The Mode parameter refers to the distinction between the realis and irrealis coding
of events by means of indicative (i.e. realis) verb forms or other moods such as sub
junctive and optative. According to Hopper & Thompson (1980: 252), “an action
which did not occur, or which is presented as occurring in a non-real (contingent)
world, is obviously less effective than one whose occurrence is actually asserted as
corresponding directly with a real event”.
I. & J. Affectedness of the Object & Individuation of the Object
The (I) Affectedness and (J) Individuation parameters measure the degree of effec
tiveness with which the action is transferred to the Object. The notion of Individu
ation refers to the distinctness of the Object both from the Agent and from its own
background (p. 253). A highly individuated Object (i.e. proper, human/animate,
concrete, singular, count, definite/referential) is more likely to be regarded as totally
affected than a non-individuated Object. Hopper & Thompson (p. 287) observe
that the properties of definiteness and referentiality in the Object stand out from
the other Individuation features in terms of their correlation with other features of
high Transitivity. Thus, the authors (p. 257) provide evidence for the fact that, in a
variety of languages, verbs with indefinite and non-referential Objects are assigned
to the class of intransitive verbs.
The aim of this paper, concerned with the relationship between clause-level Transitiv
ity and discourse factors, is twofold. First, I intend to refine the notion of high Tran
sitivity in Hopper & Thompson’s theory. Second, I will identify significant similarities
and/or differences in the linguistic encoding of Transitivity in English and Spanish,
discussing the correlation between foregrounding and (prototypical) Transitivity in
both languages.
. Refining the notion of cardinal Transitivity
In my view, the account of cardinal Transitivity presented in Hopper & Thompson

(1980) is in need of some revision. As Tsunoda (1985: 393) rightly points out, the au
thors seem to regard the ten Transitivity parameters as equally important in terms of
their relevance to the morphosyntactic manifestations of high and low Transitivity.
However, it is obvious that some of these parameters are more central than others in
the definition of prototypical Transitivity. In (2) I have isolated the six features that I
 Pilar Guerrero Medina
consider to be the most objective in the measurement of cardinal Transitivity. The rel
evance of these properties in terms of their correlation with foregrounding in English
and Spanish will be investigated in §4.
(2) a. Two (or more) participants involved
b. Agentivity
c. Object affectedness
d. Object individuation
e. Telicity
f. Boundedness
I will first consider the association of transitivity and agentivity, which has posed par
ticular difficulties in linguistic discussions of the syntax of English and other languages,
as observed by authors like Cruse (1973) and DeLancey (1984).
Cruse (1973: 18) argues that there are four independent features that operate in the
characterization of agentivity, i.e. volitive, effective, initiative and agentive. The volitive
feature “is present when an act of will is stated or implied” (p. 18). The effective feature
“is present in a sentence which refers to something which exerts a force (literally or
metaphorically), not by virtue of an internal energy source, but because of its position,
motion, etc.” (p. 19). The meaning of the feature initiative is glossed as “initiation of
an action by giving a command” (p. 20). Finally, the agentive feature “is present in any
sentence referring to an action performed by an object which is regarded as using its
own energy in carrying out the action” (p. 21).
As DeLancey (1984: 181) points out, some of the semantic properties involved in
the definition of agentivity (like control, volition or animacy) can vary independently
of one another. This author, who conceives of agentivity as a scalar category closely
related to the notion of causation, characterizes the prototypical agent as a “volitional
causer” (p. 185).
Since both the (E) Volitionality and (H) Agency properties listed in Table 1
strongly co-occur in prototypically transitive clauses, these two features have been sub
sumed under one single parameter, Agentivity, a scalar notion characterized in terms
of volitionality and control.2 The prototypical agent could thus be defined as the vo
litional instigator (typically human), in control over the event denoted by the verb.
As the notion of Agentivity obviously entails a kinetic quality (see Givón 1993: 90),
the Kinesis parameter has not been included as a separate Transitivity component in
(2) above.
The use of the terminology concerning aspectuality distinctions in Hopper &
Thompson’s theory is also in need of revision. Whereas the authors use the terms
telic/atelic and perfective/imperfective interchangeably,3 I consider it necessary to dif
ferentiate between telicity (a semantic parameter determined to a great extent by
pragmatic factors) and perfectivity (a parameter which concerns the grammatical cod
ification of a particular state of affairs in a given clause).
Comrie’s (1976) approach to the subject may shed some light on the terminologi
cal problem in discussions of aspect. The aspectual opposition perfective/imperfective,
Cardinal Transitivity in foregrounded discourse 
which is frequently expressed as an inflectional distinction in the languages of the

world,4 is defined by Comrie (1976: 3) as an opposition concerning “the different ways
of viewing the internal temporal constituency of a situation”. According to the author
(p. 16), “perfectivity indicates the view of a situation as a single whole, without distinc
tion of the various separate phases that make up that situation; while the imperfective
pays essential attention to the internal structure of the situation.”
Givón (1993: 100) defines the prototypical transitive verb as a perfective verb that
codes a “bounded, terminated, fast-changing event that took place in real time.” In the
author’s words (p. 90), a “bounded” change of state is “construed as a change from a
distinct initial state to a distinct terminal state”. By contrast, an “unbounded” event is
“construed as an ongoing process without firm boundaries.”
As implied in Givón’s definition of a prototypical transitive verb, there seems to
be a strong correlation between punctuality and perfectivity.5 In the words of Comrie
(1976: 17–18), “while it is incorrect to say that the basic function of the perfective is
to represent an event as momentary or punctual, there is some truth in the view that
the perfective (. . . ) has the effect of reducing it to a single point.” As the author (p. 42)
observes, since a punctual situation does not last in time, punctuality and imperfectiv
ity will be incompatible in those languages with separate forms to refer to the internal
structure of a situation, as is the case in Spanish, where the perfective/imperfective
opposition has been grammaticalized in the past tense form.6
On the other hand, a telic situation is defined by Comrie (p. 45) as “one that
involves a process that leads up to a well-defined terminal point, beyond which the
process cannot continue”. Therefore, while John is singing describes an atelic situation,
John is singing a song codes a telic situation, with a well-defined terminal point. Simi
larly, the situation in John is singing songs is atelic, whereas the sentence John is singing
five songs again describes a telic situation.7
Interestingly enough, Telicity may also be determined by the properties of the first
argument in the predication, as illustrated by Dik’s (1997a: 109) examples in (3):
(3) a. The demonstrators were passing the station. [+telic]
b. Demonstrators were passing the station. [–telic]
As illustrated in (3a), a definite and referential entity occupying the first argument
slot may determine the telicity of the predication: the terminal point (i.e. “passing the
station”) will be reached only in (3a), where the first argument designates a specified
number of demonstrators. On the contrary, the (non-referring) plural in (3b) triggers
the atelic reading. The parameter of Telicity thus appears to be semantically (and prag
matically) determined by referentiality and definiteness, two areas closely related to the
notion of Individuation (see §1).8
In what follows, Telicity will thus be defined as a context-dependent parame
ter, co-determined by the semantics of verbs and arguments, while the terms perfec
tive/imperfective (or bounded/unbounded) will be restricted to refer to those aspectual
distinctions which are grammatically expressed in the languages studied.
 Pilar Guerrero Medina
Summing up, and as outlined in (2) above, the definition of a prototypically

transitive clause can be formulated as follows:
The prototypically transitive clause describes a telic state of affairs with a well-
defined inherent terminal point, where a controlling volitional agent initiates
an action (presented as bounded) which results in the total affectedness of a
highly individuated object (i.e. definite and referential).
. The foreground/background distinction
Before proceeding to the data analysis in §4, I will discuss the distinction of fore
ground/background as the unitary pragmatic principle that underlies the Transitivity
generalizations in Hopper & Thompson’s theory. The authors (1980: 280) formulate
the distinction in the following terms:
That part of a discourse which does not immediately and crucially contribute to
the speaker’s goal, but which merely assists, amplifies, or comments on it, is re
ferred to as BACKGROUND. By contrast, the material which supplies the main
points of the discourse is known as FOREGROUND.
Hopper & Thompson show that high Transitivity features, which “tend strongly to be
grammaticized in the morphosyntax of natural languages” (p. 292), predominate in
foregrounded clauses. The fact that there is a similar correlation between low Transi
tivity and backgrounding seems to be implicit in the previous assertion: backgrounded
clauses, associated with one participant, usually contain non-agentive subjects, atelic
predications or non-individuated objects.
Hopper (1979: 214), who conceives of the background/foreground distinction as a
universal of narrative discourse which can be formally realized in various ways, claims
that the difference between foregrounded and backgrounded clauses has to do with
sequentiality.9
The author, who uses the term event as synonymous with “state of affairs”, assumes
that narrative discourse has mainly a referential or denotative function, and that fore-
grounded clauses have the capacity to refer to extralinguistic reality, reproducing the
chronological order of events as they occur.10
Foregrounding is associated with perfectivity and punctuality, while durative and
imperfective verbs typically occur in backgrounded clauses (see Hopper 1979: 215).
Hopper & Thompson justify the “unquestionable affinity” of perfective predicates for
foregrounded clauses in the following terms (1980: 285–286):
Foregrounded clauses typically recount sequences of events which mimic the
chronological order of those events, as they are supposed to have occurred. Each
event in foregrounding is thus viewed in its entirety; from the viewpoint of the dis
course, it is bounded at its beginning by the termination of the preceding event,
and at its end by the initiation of the next event. The discourse thus imposes a
Cardinal Transitivity in foregrounded discourse 
perfective interpretation on foregrounded events. (. . .) In backgrounding, how

ever, events and situations are not bounded by the discourse: they are presented
as on-going, or repeated, or simultaneous with foregrounded events.
The semantic notion of referential iconicity and the pragmatic notion of “main line”
information are taken to be the main criteria to distinguish foregrounding and back-
grounding in the work of Hopper (1979) and Hopper & Thompson (1980). But the
relationship between language and reality is obviously more problematic. Fowler (1996
[1986]: 24), who acknowledges the active role of language in providing a classifica
tion of our experience, argues in this regard: “All we have to challenge is the common
but fallacious assumption that the world has a natural structure from which language
draws its meanings passively, by reflection, as it were.”
As stated by Thompson (1987: 436), there has been much discussion in the linguis
tic literature as to how the distinction of foreground/background should be defined.
In the words of Givón (1987: 185),
(. . . ) the foreground/background distinction is both useful and dangerous. It is

useful in carrying us the first step toward a function-based definition of an im
portant strand in the thematic coherence of discourse. It is dangerous if we wed
ourselves to it rigidly and do not eventually trade it in for more elaborate, more
specific, less circular and empirically better grounded notions.
DeLancey’s (1987) paper, also written as a response to the research by Hopper &
Thompson, is interesting in that it explores the cognitive correlates of the semantic
and morphosyntactic parameters which Hopper & Thompson associate with cardinal
(or prototypical) Transitivity, questioning the validity of the foreground/background
distinction as the unitary principle underlying the expression of high and low Tran
sitivity in narrative discourse. Along the lines suggested by Lakoff (1977), DeLancey
(1987: 54) argues that “it is easier to explain the discourse phenomena identified by
Hopper & Thompson in terms of a prototype semantic model rather than to explain
the semantic facts in terms of a discourse-functional theory of transitivity”. The author
claims that the experiential basis of the transitive prototype is “the universal human
understanding of the physical fact that events have causes” (p. 60). The prototypi
cal transitive clause is analyzed as a sequence of two events: “a volitional act on the
part of the agent, and a subsequent and consequent change of state on the part of
the patient” (p. 61). Both the morphosyntactic expression of transitivity and its be
haviour in discourse directly reflect aspects of cognitive categorization: it is the notion
of “psychological salience” of events close to the cognitive prototype which explains
the behaviour of grammatical transitivity in discourse.11
 Pilar Guerrero Medina
. Data analysis
In this section, I will study how the Transitivity features of Agentivity, Telicity, Bound
edness, Individuation and Affectedness of the Object correlate with the discourse
notions of foregrounding and backgrounding, as defined by Hopper & Thompson,
trying to show if the semantic and morphosyntactic manifestations of high Transitivity
always reflect the foreground/background distinction in the two languages studied.
Before focusing on the correlation between grounding and degree of Transitivity
in English and Spanish (§4.2), I will briefly discuss the morphosyntactic manifesta
tions of higher vs. lower Transitivity at the clause level in the two languages studied.12
. Morphosyntactic and semantic transitivity at the clause level
Hopper & Thompson (1980: 255–256) observe that the component features of Tran
sitivity, which can be manifested either semantically or morphosyntactically, system
atically correlate in the grammar of all the languages they analyze, and propose the
following hypothesis as a language universal:
If two clauses (a) and (b) in a language differ in that (a) is higher in Transitivity
according to any of the features 1A-J, then, if a concomitant grammatical or se
mantic difference appears elsewhere in the clause, that difference will also show
(a) to be higher in Transitivity.
Linguists such as Haiman (1985) and Dixon (1991), also concerned with the formal
expression of transitivity, draw attention to the fact that there is a correlation between
the case (or adpositional) marking of an object and the transitivity of the verb. Haiman
(1985: 135–136), who makes use of Hopper & Thompson’s global definition of Tran
sitivity, claims that the “conceptual distance” between a verb and its object (measured
by linguistic distance) is an iconic reflex of the degree of transitivity of the verb.13
Likewise, Dixon (1991: 279) considers the insertion of a preposition between a tran
sitive verb and its object NP as a reflection of the conceptual deviance from “an ideal
transitive event” in English.14 According to this author (p. 280), the inclusion of a
preposition before the object of a transitive verb could be interpreted as indicating
that “the emphasis is not on the effect of the activity on some specific object (. . . ) but
rather on the subject’s engaging in the activity”.
Consider the examples in (4) and (5) in this regard:
(4) [I drove and would you believe, my ball hit the marker and bounced back.
And Willie admonished me.]
“I told ye to hit at it, I didna’ tell ye to hit it,” he muttered. (BNC CBC 11404)
(5) Philip kicked at the pile of sticks he’d gathered for his trap, scattering them.
(BNC ABX 413)
The preposition at in (4) clearly triggers the non-resultative reading of the verb hit.15 In
(5), however, the inclusion of at does not seem to be associated to the lack of achieve
Cardinal Transitivity in foregrounded discourse 
ment of the action, since the object (the pile of sticks) is certainly presented as affected:
the sticks were kicked at, and scattered as a result.
Similarly, in Spanish, the insertion of a preposition before the object is not always
associated with conceptual deviance from the transitive prototype. As illustrated by
the examples in (6), taken from Hopper & Thompson (1980: 256), highly individu
ated objects that are human and referential must be marked with the preposition a
in Spanish:
(6) a. Celia quiere mirar un bailarín.
Celia wants to watch a ballet dancer
b. Celia quiere mirar a un bailarín.
Celia wants to watch prep a ballet dancer
‘Celia wants to watch a ballet dancer.’
The preposition in (6b) signals that the object is referential: un bailarín here means
“one ballet dancer in particular”. In (6a) the absence of a before the indefinite article
marks the object as non-referential. In cases like this, the preposition is not regarded as
case-marking element signalling absence of completion, but as a morphological reflex
of high transitivity.
The authors (p. 256) illustrate the non-human vs. human distinction with examples
like Busco mi sombrero (“I’m looking for my hat”) vs. Busco a (prep) mi amigo (“I’m
looking for my friend”). However, an example such as the one presented in (7), where
the preposition a is (optionally) inserted between a highly transitive verb (golpeó) and
a non-human object (un taxi), seems to contradict this point, since both versions are
perfectly acceptable in Spanish:
(7) Una camioneta golpeó (a) un taxi (. . . ). (CREA)
a van hit-past.3sg prep a cab
‘A van hit a cab.’
It should be noted, however, that in Una camioneta golpeó a un taxi, the preposi
tion a triggers the referential reading, as in (6a) above. The prepositionally-marked
object is therefore viewed as more highly individuated and more affected than the
unmarked object.
There are other cases where the preposition actually marks conceptual de
viance from the prototype. Consider the example in (8), where a acquires a di
rectional reading:
(8) [Aplaudió entusiasmado el gol de Javi]
y miró al cielo (. . . ). (CREA)
and look-past.3sg prep.art sky
‘[He enthusiastically applauded Javi’s goal] and looked up to the sky.’
As Givón (1993: 109) observes with regard to English, there is no one-to-one corre
spondence between morphosyntactic and semantic transitivity in this language. The
frequent assignment of a transitive syntactic structure to verbs that are semantically
 Pilar Guerrero Medina
less prototypical (e.g. verbs of emotion or perception) is considered by the author as

a metaphoric extension of the transitive prototype, a phenomenon which is either the
cause or the result of a characteristic feature of English grammar: the notion transitive
is much more syntactic than semantic. Consider the examples in (9) in this regard:
(9) a. Jay liked her eyebrows, straight and black but with a decided curve at the
end. (BNC AOL 302)
b. He liked the harbour, Edith liked the promenades, they both liked the
sands and the Pierrot shows. (BNC H8A 525)
It is significant to note, however, that the use of the syntactic transitive structure to
encode a mental state of emotion seems to attribute a certain degree of agentivity to
the experiencer subjects (Jay, he, Edith, etc.), which become quasi-controlling entities
in the examples above. According to Jacobsen (1985: 96), the notion of dominance at
tributed to the subject in transitive constructions which depart from the prototype “is
itself a reflex of an element of dominance present in the form of control exercised by
the agent over the patient in prototypical acts of manipulation”.
Spanish does not allow the transitive encoding of certain states of affairs which are
not inherently transitive. Thus, as illustrated in (10) below the experiencer argument
(a mi tío) is marked as an oblique participant:16
(10) A mi tío le gustó la prueba
to-prep my uncle him-dat17 like-past.3sg the audition
[y me propuso que tomara unas clases de voz (. . . )] (CREA)
‘My uncle liked the audition [and suggested that I should take some singing
lessons].’
By contrast, examples such as the ones in (9) above illustrate the metaphorical ex
tension undergone by the transitive construction in English, where the “agent-action
patient” schema is widely used to encode states of affairs which are not semantically
transitive.
As Givón points out (1993: 116), the converse process also occurs in English, where
many semantically transitive verbs are often used as syntactically intransitive when the
omitted object is “either stereotypical, habitual, predictable, non-referring – or simply
unimportant”. The English and Spanish examples in (11) and (12) below illustrate
the use of the (semantically) transitive verbs eat and comer without their (clearly
predictable) objects:
(11) We ate heartily every night, we could afford croissants at breakfast, and we
stopped scrounging for cigarettes. (BNC HOF 1496)
(12) En su viaje, el alcalde durmió en hoteles y comió
In his journey the mayor sleep-past.3sg in hotels and eat-past.3sg
en restaurantes. (CREA)
in restaurants
‘During his journey, the mayor slept in hotels and ate in restaurants.’
Cardinal Transitivity in foregrounded discourse 
The clauses in (7), (9), (11) and (12) above exemplify the lack of one-to-one cor
respondence between syntactic and semantic transitivity in English and Spanish. As
shown in (10), however, Spanish seems to be more restrictive than English, not allow
ing the NOM-ACC transitive pattern for the encoding of certain states of affairs, which
clearly deviate from the semantic prototype.
. Transitivity in discourse
For this contrastive study I have used ten narrative passages from two works by the
Post-Modernist American author Paul Auster (City of Glass and The Red Notebook)
and their translations into Spanish,18 trying to establish how the six Transitivity fea
tures presented in (2) above correlate with the discourse notions of foregrounding and
backgrounding in the different English and Spanish passages (reproduced in the Ap
pendix). Following the analysis in Hopper & Thompson (1980: 284ff.), each of these
Transitivity features ((a) Two (or more) participants involved; (b) Agentivity; (c) O-
Affectedness; (d) O-Individuation; (e) Telicity and (f) Boundedness) has been computed
separately for its occurrence in the foregrounded and backgrounded portions of the
English and Spanish texts.19 The results of this analysis are presented in Tables 2 and 3
below.20
The figures in Tables 2 and 3 show that, as suggested by Hopper & Thompson,
high Transitivity features predominate in foregrounded English and Spanish discourse.
Thus, of the foregrounded clauses, 66% (35/53) are agentive in both languages. In
English, 88.6% (47/53) are telic and 100% are bounded. In Spanish 92% (49/53) are
telic while 52 out of 53 foregrounded clauses have been classified as bounded. There
is a similar correlation between low Transitivity and backgrounding: the analysis has
revealed most of the backgrounded clauses to contain non-agentive subjects, atelic
and unbounded predications: e.g. “He tottered as he went, listing first to the right,
then to the left, his legs by turns buckling and locking” (Passage 6) / “Se tambaleaba al
Table 2. Transitivity features in English foregrounded and backgrounded discourse
Foreground Background
a b c d e a f b c d e f
Pass 1 9/14 12/14 3/14 7/14 12/14 14/14
4/15 3/15 2/15 3/15 5/15 6/15
Pass 2 2/5 2/5 0/5 1/5 3/5 5/5
0/2 0/2 0/2 0/2 1/2 2/2
Pass 3 2/4 3/4 0/4 0/5 4/4 4/4
3/4 3/4 0/3 0/3 3/4 1/4
Pass 4 1/3 3/3 0/3 0/3 3/3 3/3
5/11 3/11 0/11 3/11 5/11 4/11
Pass 5 3/8 3/8 1/8 2/8 7/8 8/8
0/2 1/2 0/2 0/2 0/2 0/2
Pass 6 2/4 4/4 0/4 1/4 4/4 4/4
0/9 1/9 0/9 0/9 2/9 2/9
Pass 7 1/4 3/4 0/4 0/4 4/4 4/4
4/10 6/10 0/10 3/10 5/10 3/10
Pass 8 2/6 1/6 0/6 0/6 3/6 1/6
Pass 9 2/5 3/5 1/5 2/5 4/5 5/5 3/11 3/11 0/11 0/11 5/11 3/11
Pass 10 3/6 2/6 3/6 3/6 6/6 6/6 0/6 0/6 0/6 0/6 2/6 2/6
Sum total 25/53 35/53 8/53 16/53 47/53 53/53 21/76 21/76 2/75 9/75 31/76 24/76
 Pilar Guerrero Medina
Table 3. Transitivity features in Spanish foregrounded and backgrounded discourse
Foreground Background
a b c d e a f b c d e f
Pass 1a 11/14 11/14 4/14 7/14 14/14 14/14
7/16 4/16 3/16 4/16 7/16 3/16
Pass 2a 1/4 2/4 0/4 0/4 3/4 0/34/4 1/3 0/3 1/3 1/3 0/3
Pass 3a 2/4 3/4 0/4 0/4 4/4 2/44/4 3/4 0/4 1/4 3/4 1/4
Pass 4a 1/3 3/3 0/3 0/3 3/3 3/3
4/10 3/10 0/10 1/10 5/10 3/10
Pass 5a 3/8 3/8 1/8 2/8 7/8 0/28/8 1/2 0/2 0/2 0/2 0/2
Pass 6a 3/4 4/4 0/4 2/4 4/4 0/74/4 0/7 0/7 1/7 1/7 1/7
Pass 7a 3/4 3/4 0/4 0/4 3/4 1/64/4 4/6 0/6 1/6 3/6 1/6
Pass 8a 3/7 1/7 0/7 1/7 2/7 1/7
Pass 9a 2/6 4/6 1/6 1/6 5/6 6/6 3/9 4/9 0/9 3/9 4/9 2/9
Pass 10a 3/6 2/6 3/6 3/6 6/6 5/6 2/5 2/5 0/5 2/5 3/5 2/5
Sum total 29/53 35/53 9/53 15/53 49/53 52/53 22/69 23/69 3/69 15/69 29/69 14/69
andar, ladeándose primero a la derecha y luego a la izquierda, sus piernas se doblaban

y bloqueaban alternativamente” (Passage 6a).
There seems to be, however, an inherent circularity in the association between
foregrounding and high Transitivity, since Hopper & Thompson rely on high Transi
tivity features to identify foregrounded discourse. For example, it seems obvious that,
if foregrounded clauses are usually associated with the events signalling the climax of a
story, and climactic events are usually actions, the subjects of foregrounded verbs will
be agentive and volitional.
Similarly, the correlation between telicity, punctuality and foregrounding can be
questioned as well, since these notions are not defined independently in Hopper &
Thompson’s theory (see §3). The aspectuality features of foregrounded clauses are
easily justified if we take into account the fact that foregrounded clauses, by defini
tion, code sequences of events reproducing the chronological order of these events.
Foregrounding is defined in terms of perfectivity and punctuality, which are in turn
two of the aspectuality features that characterize high Transitivity.
The analysis also reveals the lack of one-to-one correspondence between O-
Individuation and O-Affectedness in foregrounding. Thus, Objects such as “the table”
(Passage 1) or “his own dead son” (Passage 2) have been coded as individuated but
non-affected.21 The presence of a highly affected Object, usually considered as a crucial
parameter in the measurement of prototypical transitivity,22 has not proved a signif
icant property of foregrounded clauses: only 8 out of 53 of the English clauses are
classifiable as containing an affected object. The proportion of affected objects in Span
ish foregrounded discourse is 9/53. As for the Individuation feature, only 16 out of the
53 English foregrounded clauses contain highly individuated objects, with a similar
figure for Spanish: 15/53.
Finally, the number of two-participant clauses in foregrounding is surprisingly
small in the two languages: in English only 25 out of 53 clauses have been computed
as two-participant clauses, while the proportion in Spanish is 29/53. Interestingly
Cardinal Transitivity in foregrounded discourse 
enough, Thompson & Hopper (2001), who question the cardinality of high Transi
tivity as defined in Hopper & Thompson (1980), show that most of the clauses in
conversational discourse are “either one-participant clauses or two-participant clauses
with very low Transitivity” (p. 39).
. Final remarks
The analysis of written narrative discourse summarized in §4 shows that the seman
tic (and morphosyntactic) manifestations of high and low Transitivity do not always
reflect the foreground/background distinction. As the figures in Tables 2 and 3 reveal,
high Transitivity features such as Agentivity, O-Individuation and O-Affectedness do
not always correlate in foregrounded discourse.
On the other hand, the correlation between formal subordination and back-
grounding does not seem to be so straightforward as suggested in the linguistic lit
erature. As Thompson (1987: 436) rightly observes, writers can actually play with the
foreground-background relations such that a clause that is syntactically marked as de
pendent may be associated with foregrounding.23 This is the case of a sentence like
“Mr Sugar suddenly appeared, driving up to the house in a cloud of dust (. . . )” (Pas
sage 9), where the event presented in the -ing clause (“driving up to the house”), clearly
in sequence with the previous event (“Mr Sugar suddenly appeared”), cannot be au
tomatically assigned to background. Significantly enough, the -ing form “driving up”
has been translated by the Spanish finite verb form “llegó”, which I have regarded as
foregrounded.
As illustrated by the examples in (13) provided by Dik (1997b: 122–123), the
absolute validity of the postulated correlation main-foregrounded and subordinate
backgrounded should be questioned, since the information in the subordinate clause
is often the most important from a communicative perspective:
(13) a. John was quietly sitting in his office, when suddenly Peter came rushing in.
b. I very much regret that circumstances force me to fire you.
c. I honestly think you are a pompous fool.
Hopper & Thompson (1980: 285) admit that, in literary texts of a belletristic sort, sig
nificant events may be presented “in a grammatical form associated with subsidiary
events”. However, the distinction between literary language and other varieties of a
discourse is, I contend, highly implausible, since as Crystal (1985 [1971]: 30) rightly
observes, “literary language is a heightened, and more sensitive, use of language, but it
nonetheless has to use the same basic range of devices as any other use of language.”
As stated by Givón (1987), the correlation between foregrounding, main line of
the information, and temporal sequentiality remains highly problematic. On the one
hand, it is obvious that sequentiality is not a useful tool to determine the main line of
a story, especially if we accept that “the coherence structure of many discourse types
is not temporally determined” (p. 177), and that temporal sequentiality is to a great
extent constructed by narrative discourse itself. On the other hand, the notions of fore
 Pilar Guerrero Medina
ground and main line of the action cannot be postulated as independent of each other:
they rather seem to define each other in a circular way.
Notes
* I would like to thank Faber and Faber Ltd for permission to reproduce the extracts from The
Red Notebook and City of Glass (The NewYork Trilogy), by Paul Auster. I am also grateful to
the Spanish Publishing House Anagrama for allowing me to reproduce the excerpts from the
corresponding Spanish versions: El cuaderno rojo and Ciudad de cristal (La trilogía de Nueva
York). Special thanks to Mr. Paul Auster for granting the US rights permission.
. See also Hopper & Thompson (1982: 3–4).
. The presence of control is frequently considered as a necessary condition for “canonical tran
sitive encoding” (Jacobsen 1985: 93). Volitionality, however, is not always regarded as a necessary
feature of agentivity (see, for instance, Schlesinger 1989: 194). For further discussion of the
notion of agentivity, see also DeLancey (1985: 5–10), Fillmore (1968: 24) and Lakoff (1977: 244).
. Hopper & Thompson (1980: 270–271), aware of the imprecise use of the terms “telic/atelic”
and “perfective/imperfective” in the literature, are in favour of using the latter terminology,
regarded as “broader” and “safer” in discussions of aspect.
. See Bybee (1985: 141).
. See Hopper & Thompson (1980: 271). As there is also an obvious link between aspectual
boundedness and the realis and affirmation features on Hopper & Thompson’s list, the lat
ter have not been included as independent parameters on a par with the other Transitivity
components in (2).
. It is important to note, however, that the English progressive cannot be equated with the
imperfective aspect of other languages (Dik 1997a: 224). Comrie (1976: 7) observes that the pro
gressive/nonprogressive opposition, which has been grammaticalized in English, “is comparable
to the imperfective/perfective distinction only in relation to a limited set of verbs (nonstative
verbs), and then only if habitual meaning is excluded”.
. Comrie (1976: 45) speaks of “telic situations”, and not of “telic verbs”, since “situations are not
described by verbs alone, but rather by the verb together with its arguments”. As stated by the
author (pp. 45–46), it is difficult to find sentences that are unambiguously telic or atelic, since
many atelic sentences can be given a telic interpretation if the appropriate context is provided.
. According to Langacker (1991: 307), definiteness is a subjective (topicality) factor, since “it
does not pertain to the inherent nature of a participant, but rather to the highly extrinsic prop
erty of whether the speaker and hearer have succeeded in establishing mental contact with it”.
See also Givón (1993: Chapter 5) for a detailed account of the grammatical devices associated
with definiteness and reference.
. The functional foreground/background dichotomy is clearly reminiscent of the structuralist
distinction established by Roland Barthes between the “proairetic” and “hermeneutic” codes. In
S/Z, his study of Balzac’s Sarrasine, Barthes (1970) distinguishes between the proairetic code (i.e.
the code of temporally sequenced actions) and the hermeneutic code (i.e. the code of puzzles),
both of which interact and relate in various ways in narrative syntax (see Scholes 1974: 148–154).
. Similarly, Haiman (1980) sustains that grammatical structure directly reflects some aspect
of the structure of reality. The author (p. 516) argues that “other things being equal, the or
Cardinal Transitivity in foregrounded discourse 
der of statements in a narrative description corresponds to the order of events they describe.”
Haiman’s discussion is clearly in agreement with Jakobson’s (1966 [1963]: 270) assertions about
the iconicity of grammatical structure.
. See also Rice (1987: 422–423). This author, who turns to non-linguistic cognitive models
of “real world action and interaction” for her formulation of a transitive prototype, claims
that transitivity is not simply a grammatical category but a conceptual description that may
be subjectively imposed by the speaker.
. The examples that illustrate the discussion in §4.1 have been mainly obtained from two
textual corpora of English and Spanish: the British National Corpus (BNC) and the Corpus de
Referencia del Español Actual de la Real Academia Española (CREA).The examples in the BNC
are identified by means of a three-letter code and the sentence number within the text where the
hit was found.
. Haiman (1985: 106–107) defines the idea of “conceptual distance” in the following terms:
“Two ideas are conceptually close to the extent that they (a) share semantic features, properties,
or parts; (b) affect each other; (c) are factually inseparable; (d) are perceived as a unit, whether
factually inseparable or not.”
. Anderson (1976: 23) points out that this same distinction between a direct object and a
prepositionally marked object is recurrent in other languages from different families (Maori,
Walbiri, Finnish, etc.) and claims that “it possible in general to indicate that an object is incom
pletely, inconclusively, etc. affected, or that an action is incompletely, inconclusively, etc. carried
out by putting the object into an oblique case.”
. See Tsunoda (1985: 389).
. According to Hopper & Thompson, a clause like the one in (10), is a one-participant clause
where the patient is coded “in the same way as a single participant with an intransitive clause” (p.
254). Vázquez Rozas (forthcoming) offers an alternative view in her study of Spanish “gustar
type” verbs, arguing that the dative indirect object (acting as experiencer) is a “core function” in
a two-participant clause like Me gusta la cerveza (“I like beer”).
. The dative clitic le is co-referential with the experiencer argument.
. The following editions have been consulted: City of Glass. (The New York Trilogy). London:
Faber, 1987; The Red Notebook. London: Faber, 1995; El cuaderno rojo. [4th ed.]. Translated by J.
Navarro. Barcelona: Anagrama, 1997 [1994]; Ciudad de cristal. (La trilogía de Nueva York). [7th
ed.]. Translated by M. de Juan. Barcelona: Anagrama, 1998 [1996].
. Modal auxiliary constructions, verb idioms (be about to, be going to, ir a, estar a punto de,
etc.), catenative constructions (seem to, turn out to, etc.), and main verbs with finite and non-
finite clausal objects or subjects (expect, occur, parecer (yue), sentir (gue), ocurrir, etc.) have been
excluded from this analysis.
. The English-Spanish comparison is based on differences and similarities that result from
contrasting the English excerpts taken from the two selected novels by Auster and their corre
sponding translations, which are only possible Spanish renderings of the states of affairs depicted
in the original extracts. The results obtained, based on such a limited corpus, are therefore
presented as tentative.
. Concrete, definite and referential entities playing some role in the discourse (Hopper &
Thompson 1980: 258) have been regarded as individuated. Reflexive objects have been coded
as non-individuated. Only those entities undergoing a physical change of state or position as a
result of the action denoted by the verb have been coded as affected.
 Pilar Guerrero Medina
. See, for instance, Hopper (1985: 67) and Tsunoda (1985: 393).
. According to Haiman (1985: 197, 229), there is evidence that the morphological reduction
of -ing clauses in English (e.g. Leaving her children, she fled for safety) and their freedom from
tense iconicity are motivated by “conceptual closeness” with the main clause rather than by
conceptual backgrounding. See also Matthiessen & Thompson (1988: 312).
Appendix
(Foregrounded predicates, i.e. those regarded as being in temporal sequence, are in capital
letters, while those in backgrounded clauses are italicicized)
ENGLISH SPANISH
Passage 1 (City of Glass, p. 12) Passage 1a (Ciudad de Cristal, p. 19)
The next morning, Quinn WOKE UP earlier A la mañana siguiente Quinn SE DESPERTÓ
than he had in several weeks. As he drank his más temprano de lo que había hecho en
coffee, buttered his toast, and read through varias semanas. Mientras se bebía el café,
the baseball scores in the paper (the Mets untaba las tostadas con mantequilla y leía
had lost again, two to one, on a ninth inning los resultados de los partidos de béisbol en
error), it did not occur to him that he was el periódico (los Mets habían perdido otra
going to show up for his appointment. Even vez, dos a uno, por un error en la novena
that locution, his appointment [italics in the entrada), no se le ocurrió que fuera a acudir
original], seemed odd to him. It wasn’t his a su cita. Incluso esa expresión, su cita, le
appointment, it was Paul Auster’s. And who parecía extraña. No era su cita, era la cita de
that person was he had no idea. Paul Auster. Y él no tenía ni idea de quién era
Nevertheless, as time wore on he FOUND esa persona.
himself doing a good imitation of a man No obstante, a medida que pasaba el tiempo
preparing to go out. He CLEARED the table SE ENCONTRÓ haciendo una buena imi
of the breakfast dishes, TOSSED the newspa tación de un hombre que se prepara para
per on the couch, WENT into the bathroom, salir. RECOGIÓ las cosas del desayuno,
SHOWERED, SHAVED, WENT on to the TIRÓ el periódico sobre el sofá, FUE al
bedroom wrapped in two towels, OPENED cuarto de baño, SE DUCHÓ, SE AFEITÓ,
the closet, and PICKED OUT his clothes for ENTRÓ en el dormitorio envuelto en dos
the day. He FOUND himself tending toward toallas, ABRIÓ el armario y ELIGIÓ la
a jacket and a tie. Quinn had not worn a tie ropa que iba a ponerse ese día. SE DES-
since the funerals of his wife and son, and he CUBRIÓ buscando una chaqueta y una
could not even remember if he still owned corbata. Quinn no se había puesto una cor
one. But there it was, hanging amidst the bata desde el funeral de su esposa y su hijo y
debris of his wardrobe. He DISMISSED a ni siquiera recordaba si todavía tenía alguna.
white shirt as too formal, however, and in Pero allí estaba, colgando entre los restos
stead CHOSE a grey and red check affair to de su guardarropa. DESCARTÓ una camisa
go with the grey tie. He PUT them ON in a blanca por parecerle demasiado formal, sin
kind of trance. embargo, y en su lugar ESCOGIÓ una de
cuadro grises y rojos para que hiciera juego
con la corbata gris. SE las PUSO en una es
pecie de trance.
Cardinal Transitivity in foregrounded discourse 
Passage 2 (City of Glass, p. 14) Passage 2a (Ciudad de cristal, pp. 21–22)

He HEARD the sound of someone enter Oyó que alguien entraba en la habitación a
ing the room behind him. Quinn STOOD su espalda. Quinn SE LEVANTÓ del sofá y
UP from the sofa and TURNED AROUND, SE VOLVIÓ, esperando ver a la señora Still-
expecting to see Mrs Stillman. Instead, it man. En su lugar había un hombre joven,
was a young man, dressed entirely in white, vestido enteramente de blanco, con el pelo
with the white-blond hair of a child. Un rubio claro de un niño. Extrañamente, en
canningly, in that first moment, Quinn aquel primer momento Quinn PENSÓ en su
THOUGHT of his own dead son. Then, just propio hijo muerto. Luego, tan rápidamente
as suddenly as the thought had appeared, como había aparecido, el pensamiento SE
it VANISHED. DESVANECIÓ.
Passage 3 (City of Glass, pp. 14–15) Passage 3a (Ciudad de cristal, p. 22)

Peter Stillman WALKED into the room and Peter Stillman ENTRÓ en la habitación y
SAT DOWN in a red velvet armchair op SE SENTÓ en una butaca de terciopelo
posite Quinn. He SAID not a word as he rojo enfrente de Quinn. No DIJO una pal
made his way to his seat, nor DID he AC abra mientras se dirigía a su asiento ni RE-
KNOWLEDGE Quinn’s presence. The act of GISTRÓ la presencia de Quinn. El acto de
moving from one place to another seemed moverse de un sitio a otro parecía requerir
to require all his attention, as though not to toda su atención, como si no pensar en lo
think of what he was doing would reduce que estaba haciendo fuera a reducirle a la in
him to immobility. Quinn had never seen movilidad. Quinn nunca había visto a nadie
anyone move in such a manner, and he re moverse así y comprendió inmediatamente
alized at once that this was the same person que aquélla era la persona con quien había
he had spoken to on the phone. hablado por teléfono.
Passage 4 (City of Glass, p. 15) Passage 4a (Ciudad de cristal, pp. 22–23)

Stillman SETTLED slowly into his chair and Stillman SE ACOMODÓ lentamente en su
at last TURNED his attention to Quinn. asiento y al fin DIRIGIÓ su atención hacia
As their eyes met, Quinn suddenly felt that Quinn. Cuando sus ojos se encontraron,
Stillman had become invisible. He could see Quinn sintió repentinamente que Stillman
him sitting in the chair across from him, se había vuelto invisible. Podía verle sen
but at the same time it felt as though he was tado en la butaca frente a él, pero al mismo
not there. It occurred to Quinn that per tiempo tenía la sensación de que no es
haps Stillman was blind. But no, that did not taba allí. Se le ocurrió que quizá Stillman
seem possible. The man was looking at him, fuese ciego. Pero no, eso no parecía posi
even studying him, and if recognition did ble. El hombre le estaba mirando, incluso es
not flicker across his face, it still held some tudiándole, y aunque a su cara no asomaba
thing more than a blank stare. Quinn did el reconocimiento, había en ella algo más
not know what to do. He SAT there dumbly que una mirada vacía. Quinn no sabía qué
in his seat, looking back at Stillman. A long hacer. SE QUEDÓ allí sentado y mudo, de
time passed. volviéndole la mirada a Stillman. Pasó mucho
tiempo.
 Pilar Guerrero Medina
Passage 5 (City of Glass, p. 23) Passage 5a (Ciudad de cristal, p. 31)

The sound of stockinged legs moving across El sonido de unas piernas enfundadas en
the room finally BROKE the silence. There medias cruzando la habitación ROMPIÓ
WAS the metal click of a lamp switch, and finalmente el silencio. SE OYÓ el sonido
suddenly the room WAS FILLED with light. metálico del interruptor de una lámpara y de
Quinn’s eyes automatically TURNED to its pronto la habitación SE LLENÓ de luz. Los
source, and there, standing beside a table ojos de Quinn SE VOLVIERON automáti
lamp to the left of Peter’s chair, he SAW camente hacia su fuente, y allí, de pie al
Virginia Stillman. The young man was gaz lado de una lámpara de mesa a la izquierda
ing straight ahead, as if asleep with his eyes de la butaca de Peter, VIO a Virginia Still-
open. Mrs Stillman BENT OVER, PUT her man. El joven seguía mirando fijamente al
arm around Peter’s shoulder, and SPOKE frente, como si estuviera dormido con los
softly into his ear. ojos abiertos. La señora Stillman SE IN-
CLINÓ, RODEÓ los hombros de Peter con
un brazo y le HABLÓ suavemente al oído.
Passage 6 (City of Glass, p. 24) Passage 6a (Ciudad de cristal, p. 31)

Peter GAVE a little spastic wave of the hand Peter HIZO un pequeño movimiento es
and then slowly TURNED and WALKED pástico con la mano como despedida y
across the room. He tottered as he went, list luego SE VOLVIÓ lentamente y CRUZÓ la
ing first to the right, then to the left, his habitación. Se tambaleaba al andar, ladeán
legs by turns buckling and locking. At the dose primero a la derecha y luego a la
far end of the room, standing in a lighted izquierda, sus piernas se doblaban y bloque
doorway, was a middle-aged woman dressed aban alternativamente. Al otro extremo de
in a white nurse’s uniform. Quinn assumed la habitación, de pie en un umbral ilumi
it was Mrs Saavedra. He FOLLOWED Peter nado, había una mujer de mediana edad
Stillman with his eyes until the young man vestida con un uniforme blanco de enfer
disappeared through the door. mera. Quinn supuso que sería la señora
Saavedra. SIGUIÓ a Peter Stillman con los
ojos hasta que el joven desapareció por
la puerta.
Passage 7 (Red Notebook, p. 3) Passage 7a (Cuaderno rojo, p. 27)

In 1972 a close friend of mine RAN into En 1972 una íntima amiga mía TUVO pro
trouble with the law. She was in Ireland that blemas con la ley. Vivía aquel año en una
year, living in a small village not far from aldea de Irlanda, no muy lejos de la ciu
the town of Sligo. As it happened, I was visit dad de Sligo. Yo había ido a verla por aquel
ing on the day a plainclothes detective drove entonces, el día que un policía de paisano
up to her cottage and presented her with a se presentó en la casa con una citación del
summons to appear in court. The charges juzgado. Las acusaciones eran lo suficiente
were serious enough to require a lawyer. My mente serias como para requerir un abo
friend ASKED around and WAS GIVEN a gado. Mi amiga PIDIÓ información, le RE
name, and the next morning we BICYCLED COMENDARON un nombre, y a la mañana
into town to meet with this person and dis siguiente FUIMOS en bicicleta a la ciudad
cuss the case. To my astonishment, he worked para reunirnos y hablar del asunto con aque
for a firm called Argue and Phibbs. lla persona. Con gran asombro por mi parte,
trabajaba en un bufete de abogados llamado
Argue & Phibbs.
Cardinal Transitivity in foregrounded discourse 
Passage 8 (Red Notebook, p. 4) Passage 8a (Cuaderno Rojo, p. 29)

It turned out to be a curious year. On the Fue un curioso año. Por una parte, el lu
one hand, the place was beautiful: a large, gar era precioso: un caserón de piedra del
eighteenth-century stone house bordered by siglo XVIII, rodeado de viñas por uno de
vineyards on the one hand and a national sus flancos y, por el otro, por un parque
forest on the other. The nearest village was nacional. El pueblo más próximo estaba a
two kilometers away, but it was inhabited by dos kilómetros de distancia, y no lo habita
no more than forty people, none of whom ban más de cuarenta personas, ninguna de
was under sixty or seventy years old. It was menos de sesenta o setenta años. Era un sitio
an ideal spot for two young writers to spend ideal para que dos escritores jóvenes pasaran
a year, and L. and I both worked hard there, un año, y tanto L. como yo, trabajando de
accomplishing more in that house than either verdad, sacamos en aquella casa mucho más
one of us would have thought possible. fruto del que ninguno de los dos hubiera
creído posible.

That was at four o’clock in the afternoon. Eran las cuatro de la tarde. Menos de una
Less than an hour later, the errant Mr. hora después, el imprevisible señor Sugar
Sugar suddenly APPEARED, driving up to APARECIÓ inesperadamente. LLEGÓ hasta
the house in a cloud of dust, gravel and dirt la casa en medio de una nube de polvo:
crunching all around him. If I think about la tierra y la gravilla rechinaban bajo los
it hard enough, I can still see the naive and neumáticos. Si me concentro, todavía puedo
goofy smile on his face as he bounced out of ver la cara boba e ingenua con que bajó
the car and said hello. It was a miracle. It was del coche y nos saludó. Era un milagro.
a genuine miracle, and I was there to witness Era un verdadero milagro. Y yo estaba allí
it with my own eyes, to live it in my own para verlo con mis propios ojos, para vivirlo
flesh. Until that moment, I thought those en mi propia carne. Hasta aquel momento,
things happened only in books. yo pensaba que cosas así sólo ocurrían en
Sugar TREATED us to dinner that night in a los libros.
two-star restaurant. We ATE copiously and Sugar nos INVITÓ a cenar aquella noche en
well, we EMPTIED several bottles of wine, un restaurante de dos tenedores. COMIMOS
we LAUGHED our heads OFF. (. . . ) copiosamente bien, nos BEBIMOS varias
botellas de vino, nos REÍMOS como locos.
(. . . )
 Pilar Guerrero Medina

One afternoon many years ago, my father’s Una tarde de hace muchos años a mi padre
car STALLED at a red light. A terrible storm SE le CALÓ el coche en un semáforo en
was raging, and at the exact moment his en- rojo. Se había desencadenado una terrible
gine went dead, lightning STRUCK a large tormenta y, en el preciso momento en que el
tree by the side of the road. The trunk of the motor fallaba, un rayo ALCANZÓ un gran
tree SPLIT in two, and as my father strug- árbol de la calle. El tronco del árbol SE PAR
gled to get the car started again (unaware TIÓ en dos y, cuando mi padre se esfor
that the upper half of the tree was about to zaba en volver a arrancar el motor (sin darse
fall), the driver of the car behind him, seeing cuenta de que la mitad superior del árbol es-
what was about to happen, PUT his foot on taba a punto de desprenderse), el conductor
the accelerator and PUSHED my father’s car del coche que lo seguía, viendo lo que iba
through the intersection. An instant later, a suceder, PISÓ el acelerador y EMPUJÓ el
the tree CAME crashing to the ground, land- coche de mi padre más allá del cruce. Un
ing in the very spot where my father’s car instante después, el árbol SE ESTRELLABA
had just been. (. . . ) contra el suelo, en el sitio exacto que había
ocupado el coche de mi padre. (. . . )
References
Anderson, Stephen R. (1976). “On the notion of subject in ergative languages”. In Ch. N. Li
(Ed.), Subject and Topic (pp. 1–23). New York: Academic Press.
Barthes, Roland (1970). S/Z. Paris: Seuil.
Bybee, Joan L. (1985). Morphology. A Study of the Relation between Meaning and Form. Ams
terdam and Philadelphia: John Benjamins.
Comrie, Bernard (1976). Aspect. Cambridge: Cambridge University Press.
Cruse, D. Alan (1973). “Some thoughts on agentivity”. Journal of Linguistics, 9, 11–23.
Crystal, David (1985 [1971]). Linguistics (2nd ed.). London: Penguin.
DeLancey, Scott (1984). “Notes of agentivity and causation”. Studies in Language, 8(2), 181–213.
DeLancey, Scott (1985). “Agentivity and syntax”. Papers from the Regional Meetings of the Chicago
Linguistic Society, 21(2), 1–12.
DeLancey, Scott (1987). “Transitivity in grammar and cognition”. In R. S. Tomlin (Ed.), Coher
ence and Grounding in Discourse (pp. 53–68). Amsterdam and Philadelphia: John
Benjamins.
Dik, Simon C. (1997a). The Theory of Functional Grammar, Part 1: The Structure of the Clause
(Ed. K. Hengeveld). Berlin and New York: Mouton de Gruyter.
Dik, Simon C. (1997b). The Theory of Functional Grammar, Part 2: Complex Constructions (Ed.
K. Hengeveld). Berlin and New York: Mouton de Gruyter.
Dixon, Robert M. W. (1991). A New Approach to English Grammar, on Semantic Principles.
Oxford: Clarendon Press.
Fillmore, Charles J. (1968). “The case for case”. In E. Bach & R. T. Harms (Eds.), Universals in
Linguistic Theory (pp. 1–89). New York: Holt, Rinehart and Winston.
Fowler, Roger (1996 [1986]). Linguistic Criticism (2nd ed.). Oxford: Oxford University Press.
Givón, Talmy (1987). “Beyond foreground and background”. In R. S. Tomlin (Ed.), Coherence
and Grounding in Discourse (pp. 175–188). Amsterdam and Philadelphia: John Benjamins.
Cardinal Transitivity in foregrounded discourse 
Givón, Talmy (1993). English Grammar. A Function-Based Introduction, Vol. 1. Amsterdam and
Haiman, John (1980). “The iconicity of grammar: isomorphism and motivation”. Language,
56(3), 515–540.
Haiman, John (1985). Natural Syntax. Iconicity and Erosion. Cambridge: Cambridge University
Press.
Hopper, Paul J. (1979). “Aspect and foregrounding in discourse”. In T. Givón (Ed.), Discourse
and Syntax (pp. 213–242). New York: Academic Press.
Hopper, Paul J. (1985). “Causes and affects”. Papers from the Regional Meetings of the Chicago
Linguistic Society, 21(2), 67–88.
Hopper, Paul J. & Sandra A. Thompson (1980). “Transitivity in grammar and discourse.”
Language, 56, 251–259.
Hopper, Paul J. & Sandra A. Thompson (1982). “Introduction”. In P. J. Hopper & S. A.
Thompson (Eds.), Studies in Transitivity (pp. 1–5) [Syntax and Semantics 15]. New York:
Academic Press.
Jacobsen, Wesley M. (1985). “Morphosyntactic transitivity and semantic markedness”. Papers
from the Regional Meetings of the Chicago Linguistic Society, 21, 89–104.
Jakobson, Roman (1966 [1963]). “Implications of language universals for linguistics”. In J. H.
Greenberg (Ed.), Universals of Language (pp. 263–278) (2nd ed.). Cambridge, MA: MIT
Press.
Lakoff, George (1977). “Linguistic Gestalts”. Papers from the Regional Meeting of the Chicago
Linguistic Society, 13, 236–287.
Langacker, Ronald W. (1991). Foundations of Cognitive Grammar, Vol. 2: Descriptive Application.
Stanford: Stanford University Press.
Matthiessen, Christian & Sandra A. Thompson (1988). “The structure of discourse and
‘subordination”’. In J. Haiman & S. A. Thompson (Eds.), Clause Combining in Grammar
and Discourse (pp. 275–329). Amsterdam and Philadelphia: John Benjamins.
Rice, Sally (1987). “Towards a transitive prototype: evidence from some atypical English
passives”. Proceedings of the Annual Meeting of the Berkeley Linguistic Society, 13, 422–434.
Schlesinger, Izchak M. (1989). “Instruments as agents: on the nature of semantic relations”.
Journal of Linguistics, 25, 189–210.
Scholes, Robert (1974). Structuralism in Literature. An Introduction. New Haven and London:
Yale University Press.
Silverstein, Michael (1976). “Hierarchy of features and ergativity”. In R. M. W. Dixon (Ed.),
Grammatical Categories in Australian Languages (pp. 112–171). Canberra: Australian
Institute of Aboriginal Studies.
Thompson, Sandra A. (1987). “‘Subordination’ and narrative event structure”. In R. S. Tomlin
(Ed.), Coherence and Grounding in Discourse (pp. 435–454). Amsterdam and Philadelphia:
John Benjamins.
Thompson, Sandra A. & Paul J. Hopper (2001). “Transitivity, clause structure, and argument
structure: Evidence from conversation”. In J. L. Bybee & P. J. Hopper (Eds.), Frequency
and the Emergence of Linguistic Structure (pp. 27–60). Amsterdam and Philadelphia: John
Benjamins.
Tsunoda, Tasaku (1985). “Remarks on transitivity”. Journal of Linguistics, 21(2), 385–396.
Vázquez Rozas, Victoria (forthcoming). “Gustar-type verbs”. In J. C. Clements & J. Yoon (Eds.),
Functional approaches to Spanish Syntax. New York: Palgrave Macmillan.
English consciousness in 19th century Spain
Paloma Tejada Caller
Part of a wider project, this paper aims to define the identity of English and
Englishness as seen by particular foreign-language users. A closed corpus of 108
institutional lectures publicly read at the Spanish Royal Academy of the Language
(1847–1885) has been analysed to show that English is conventionally described
as a distinct, civilized, literary, value-laden language, reflecting the spirit of a
homogeneous, educated, tolerant, dynamic and thriving people. This selective
image of English serves two functions: to reinforce nation-imagining myths
(English as a standard of universal values), and to suggest a model of innovation
(English as a space for experimentation and learning). Both functions proved
useful to supporters of the bourgeois revolution of 1870s, the new liberal elite.
In (1996: 217–218) Leith suggested new roads for a critical linguistic history of English
texts. He wrote:
Any historian of English must decide which texts to select; and how to interpret
them. Inevitably, this will reflect his or her idea of what the history of English ac
tually is. . . [R]ecently, however, scholars from other disciplines, such as Literature,
History and Cultural studies, have shown an interest in the field. They have other
important questions to ask. So also do the many second-language users of English
in different parts of the world: does, for instance, a textual history of English also in
clude texts produced beyond England? These issues of selection and interpretation are
therefore central to opening up the field. [Emphasis added]
This paper, probably inspired by Leith’s words, is part of a wider project in which En
glish is analysed as a cultural object outside the traditional borders of ‘Englishness’.
It aims to contribute to the development of the cultural history of English, assuming
that this discipline should include the images this language has or has had beyond its
national or conventional borders. If this assumption is correct, the contact between
English and other languages outside its colonial sphere should be described. The study
of language in its socio-cultural dimension should be expanded to the analysis of im
ages of Englishness constructed by different groups of people in different cultural and
ideological contexts. For that purpose, 108 19th century scholarly Spanish texts (ca.
 Paloma Tejada Caller
4200 pages) have been selected and carefully analysed. The texts belong to a closed
corpus of institutional lectures publicly read at the Spanish Royal Academy of the Lan
guage during the period 1847–1885 and deal, therefore, with philological issues of
greater or lesser contemporary concern. Part of a wider and more ambitious project,
the purpose of the present contribution is to give an initial and necessarily partial an
swer to new questions, such as the following: What image of the English language did
19th c. Spanish intellectuals project? Why and for what purpose did Spanish linguistic
authorities refer to English? Which features, events or characters did they focus on?
How is English conceived of in a non-British cultural system? Does the constructed
image help to reinforce any particular ideology? Has English been used to construct
some Spanish linguistic identity? To what extent are taken-for-granted assumptions
on language shared cross-culturally?
The role played by language in the construction of identity has been the object
of study of a long and multifold tradition of research. Sociolinguistics, Cultural stud
ies, Critical Linguistics, Ideology and Public Discourse studies, represented by scholars
such as Bailey (1992, 1996), Blommaert & Verschueren (1998), Brown (1990), Burke
(1987), Crowley (1989, 1990, 1991, 1996), Corfield (1991), van Dijk (1991), Dirven
et al. (2001), Hawkins (2000), Joseph & Taylor (1990), Koerner (1989, 2000, 2001),
Kress (1990), Kress & Hodge (1979), Leith (1996), Rissanen et al. (1992), Schieffelin
(1998), Taylor (1990), Wodak (1989), Woolard (1998), to mention just a few, recognise
the imbrication between knowledge (power) and identity and warn about the institu
tional and ideological nature of linguistic discourse. However, very little research has
been conducted so far on the identity of English and Englishness as seen by partic
ular second-language or foreign-language users; and no attention has been paid to
the particular image of Englishness elaborated in Spain. As a starting point for this
new project, metalinguistic institutional discourses referring to the English language,
literature and culture, produced by members of the Spanish Royal Academy of the
Language during the 19th c. have been selected and analysed.1
Several hypotheses sustain our research. It is widely known that the meaning of
‘reality’ depends on the way it is defined, and that reality becomes accessible through
images constructed by specific agents; ie. representation conditions our experience of
otherness. There is also little doubt that the Spanish Royal Academy is an authorized
institution per se (Zamora 1999), and as such a powerful codifier of (linguistic) im
ages. Therefore it seems likely that a particular image of ‘English’ may arise from the
selected texts.
Bearing in mind the fact that the intervention of language in institutional debates
often aims at provoking particular effects (Crossick 1991, Crowley 1990, Dellinger
1995, Kay 1987, Philips 1998, Silverstein 1998, Ward 1990, Williams 1983), the image
of English we will presumably get will respond to more or less implicit national pur
poses, such as promoting particular attitudes and social values; generating a particular
image of the Spanish language or creating and reinforcing nation-imagining myths.
The constructed image of English and Englishness is expected to reflect the impor
tance of 19th c. movements in elaborating national images of languages; it may also be
English consciousness in 19th century Spain 
conditioned by social movements prevailing in mid 19th c. Spain; and mediated by the
contact some of the scholars and members of the Academy maintained with England.
Last, but not least, the image of English constructed by members of the authority-
exerting Spanish Academy may prove to be a leading image, which may have condi
tioned or controlled the more or less intuitive, popular, unscientific knowledge Span
ish people might have had or may still partially maintain about the English language
and culture.
This paper is restricted in its time-span to the period 1847–1885. It aims, therefore,
to define the image of English and Englishness arising from 108 Spanish public lectures
pronounced in the Royal Academy of the Language in the second half of the 19th c. It
is structured into six sections. After the introduction in 1, we provide a schematic
description of the selected corpus in 2. The next two sections (3 and 4) introduce a
summary of quantitative and qualitative results. Finally an analysis of the results and
some concluding remarks are provided in Sections 5 and 6, respectively.
. Description of the corpus
To test our hypotheses texts from a closed corpus of public lectures read at the Spanish
Royal Academy of the Language during the years 1847–1885 have been selected. Before
1847 sessions at the Academy were private and lack official records. After the 1880s the
Spanish political and ideological situation undergoes important changes, entering a
new stage of turmoil.2 Lectures were taken from the following collections kept in the
Academy Library in Madrid:
1. Real Academia Española. Discursos inaugurales (undated)
2. Real Academia Española. 1860. Discursos de recepción. Serie I. 3 Vols. Madrid:
Imprenta Nacional
3. Real Academia Española. Discursos de recepción Vol. 1: 1867–(undated)
4. Real Academia Española. Discursos de recepción. Vol . 2: 1872–1879 (undated)
5. Real Academia Española. 1947. Discursos de recepción. Serie Segunda. 4 Vols. 1881–
1898. Madrid: Gráficas Ultra.3
As shown in the Appendix the corpus includes two different types of texts. First, the
so-called Reception lectures (Discursos de Recepción), each read by a Spanish scholar,
newly admitted as a member of the institution. These lectures were usually “answered”
in the same session by an older member of the Academy to formally welcome the new
colleague (Answering Lecture or Discurso de Contestación). And second, the Opening
lectures (Discursos Inaugurales) read by outstanding members of the Academy at the
beginning of every new “year”. They opened up each new session and were, thus, unan
swered. No difference between these two types has been made for the present study,
since their formal distinction does not affect their contents.
The process of analysing the material has proved hard and time-consuming. Lec
tures are very often untitled; books – some of them collections of independent lectures
 Paloma Tejada Caller
bound up together – offer no table of contents or index of subjects, which demands a

thorough, patient reading in order to select those lines, paragraphs or pages referring
to English. Some 4200 pages were read for 108 lectures (97 were reception/answering
lectures and 11 were opening lectures) by 69 different authors. The Appendix offers a
precise relation of the 45 lectures selected as including some reference to English. It is
to be noted that the extension of references to English is not homogeneous and ranges
from name-dropping to the use of English as reinforcing evidence or as secondary
topic of discourse. As expected, there are no lectures wholly devoted to English, as
scholars are invited to become members of the Academy of the Spanish Language and
they conventionally select topics of mainly Spanish interest or tradition.
After this hand-made collation of references, they were broadly classified under
language, literature and culture labels to obtain a detailed description of the relevant
material from a thematic, as well as chronological and sociological point of view. This
leads us to the reconstruction of an institutional image of English, made in Spain by
the official, linguistic (and intellectual) authority in an age particularly revealing for
the elaboration of linguistic and cultural nationalism.
. Quantitative results
As reported in previous sections, 108 lectures were read, 46 of which included refer
ences to the English language or culture. This yields a surprisingly high percentage of
circa 42%. However, it is only 18 of these lectures, a modest 16.8%, that make more
extensive reference to English topics;4 that is, they comment on a particular topic or
develop a particular argument centred around English or the English. With regard to
the number of authors who select English as discourse topic, 31 were found to do so,
which means 47.8%; but only 15 of them may be said to devote longer and more com
plex thoughts to the subject. More specifically, the names of Alcalá Galiano, Alarcón,
Canalejas, Cueto, Cutanda, Escosura, Fernández, Monlau, de Mora, la Puente, Núñez
de Arce, Olózaga, Pascual, Saavedra and Valera must be mentioned.
From this quantitative perspective two further issues should be finally empha
sized. First, almost half of the lectures in which some reference to English is made deal
with literary topics. And, second, the more extensive references to English concentrate
around the period 1860–1877.
. A selection of qualitative results
A more refined qualitative analysis of the evidence indicates that Englishness is de
fined around three major nuclear topics, which can be schematically summarized as
follows:5
English consciousness in 19th century Spain 
1. Stars of glorious names: Shakespeare, Milton, Byron and at close distance, Walter
Scott.
2. The clearly distinct spirit of its people, in which money prevails. The pragmatic
nature of British citizens, their high level of education, their freedom of thought,
their vigorous institutions and their sea power also deserve the reader’s attention,
as well as their ability to engender false myths on Spanish history; and
3. Language, closely related to the English soul; value-laden; of Germanic origin and
unexpected influence on Spanish; of chaotic pronunciation and vocal flaws; of an
extremely plain verbal system; of analytic weakness and rich vocabulary.
Features (1) to (3) roughly outline the image of English elaborated in institutional 19th
c. Spain. But a more detailed description of each topic is required, in order to get the
complete picture.
At this second level of description, literary topics diversify. Members of the Span
ish Royal Academy distinguish between “the greatest authors” – their more recurrent
matter of interest – and “other poets of remarkable virtue”, as will be shown. However,
particular topics, such as “the devils’ chapel” (9-Cueto,6 on English drama); English
satire and epigram (44-Cutanda); popular and medieval legends (50-Valera); poetry
and political freedom (15-Martínez de la Rosa) or English women writers (94-Conde
de Casa Valencia), are also worth the scholars’ attention to a greater or lesser extent.
Shakespeare is by large the most frequent and preferred topic, developed along
several main threads. He is literally described as “a total genius”, “one of the best of all
nations and times” (2-Alcalá Galiano); “the summit of Anglo-Saxon Parnassus” (88
Núñez de Arce); the one “whose name fills the realms of Earth” (36-Tamayo); the one
“English people will always be proud of ” (50-Valera):
(1) Los grandes escritores son los que graban este sello, con delicado y fuerte
buril, en el oro y en las joyas de sus escritos. . . Por eso decia Carlysle, con mu
cho fundamento -que, si le dijeran que eligiese para su patria entre la pérdida
de Shakspeare ó la de las Indias Orientales, preferiria la segunda, porque tarde
ó temprano se han de perder aquellas colonias, mientras que el glorioso poeta
vivirá vida inmortal, y será leido en los más remotos ángulos de la tierra, por
donde la Gran Bretaña ha derramado á sus hijos, y cuando éstos se hallaren
separados políticamente de la metrópoli, no sólo en América, sino en Aus
tralia y en otras islas y regiones del Pacífico y del Atlántico, se jactarán al leer
á Shakspeare de ser ingleses. (50-Valera: 239)
“The greatest authors are those who make engravings on their golden and
treasured writings. . . with a delicate and sharp burin. That is why Carlysle
very sensibly explained his preferences. If he were told to choose between the
loss of Shakespeare or that of the East Indies, he would take the latter. Sooner
or later these colonies will be lost, but the glorious poet will live an immor
tal life and will be read in the most hidden places where England has poured
out her children. When they eventually become politically independent from
 Paloma Tejada Caller
the metropolis, not only in America, but also in Australia and in other Pa
cific islands and regions, they will all be proud of being English when reading
Shakespeare.”
Secondly, Shakespeare is also praised as a playwright. He represents the essence of
drama; is himself “complete drama” (9-Cueto); “the very passionate creator of mod
ern drama” (36-Tamayo); “the author of admirable and never-ending tragedies” (94
Conde de Casa Valencia); he is “the mirror of life and autopsy of conscience” (90
Alarcón); the “best portrait-painter of historic characters” (11-Escosura); the one who
“suceeded best in describing man at all ages”:
(2) Shakespeare, como de costumbre en sus dramas-crónicas, se ha inspirado
directa y resueltamente en la historia; y sin aspirar ni á desfigurarla ni á em
bellecerla, ha procurado, y conseguido á mi juicio, reproducirla tan rigorosa
como fielmente en la escena. Como Plutarco describe, el vate del siglo de Isa
bel de Inglaterra representa, digo mal, evoca los personajes que en su tragedia
intervienen. El espectador asiste á los acontecimientos que prepararon y pre
cedieron la catástrofe de los idus de Marzo; á la catástrofe misma; y á sus
consecuencias hasta la batalla de Filipos, en que los matadores de César pa
garon con sus propias vidas la que al Dictador habian arrebatado.
César y Antonio, Octavio y Lépido, Bruto y Casio, Senadores y Tribunos, ciu
dadanos y legionarios, todo es romano, todo histórico, todo retrato; pero
retrato de mano maestra, de esos que pintan más acaso el espíritu que el
cuerpo en que anima.
No hay allí más pasion que la política, más amores que al poder ó á la pa
tria,. . . ” (11-Escosura: 49)
“As he often does in his chronicle-plays, Shakespeare has found clear and di
rect inspiration in history. Trying neither to distort nor embellish it, his main
aim and success – as I see it – is to reproduce history on stage faithfully and
rigorously. As Plutarch described, the Elizabethan bard represents or rather,
evokes the characters who take part in his tragedies. The public attend the
events leading to the disaster of the ides of March; they stand in front of
the disaster itself; they become aware of its consequences to finally reach the
battle at Philippos, where Caesar’s assassins paid with their lives that of the
dictator’s.
Caesar, Anthony, Octavius and Lepidus, Brutus and Casius, senators and
tribunes, citizens and legionaries. Everything there is Roman, everything is
historic; everything a portrait, but a masterly portrait; one that renders the
spirit rather than the body it enlivens.
There no passion is found but that of politics, no love but that of power and
fatherland”.
(3) Recordad el mundo animado en la esfera del arte por el númen de Shake
speare. Allí la inagotable variedad de la naturaleza, distinguiéndose cada
English consciousness in 19th century Spain 
personaje entre los demas por una fisonomía propia; allí el humano sin en
miendas ni mutilaciones, causando mal para lástima y admiracion; allí los
más ocultos móviles de la voluntad, las más impenetrables operaciones de
la conciencia, los más hondos abismos de la mente y el corazon; allí Lady
Macbeth, Julieta, Desdémona, Shylock, Ricardo III, Macbeth, Otelo, Romeo,
Hamlet, Lear, haciendo creer que un alma verdadera los vivifica; allí la hu
manidad retratada al vivo bajo todas sus fases, en su actitud más imponente y
expresiva; y ésta es la causa de que el nombre de Shakespeare llene los ámbitos
de la tierra. (36-Tamayo: 278)
“Remember the world inspired by Shakespeare’s soul in the sphere of art.
There the unending variety of nature is found, each character being different
from the rest, showing their own appearance; there the human beings, por
trayed as they are, causing evil, for sorrow of praise; there the darkest springs
of action, the most unintelligible acts of conscience; the deepest abysses of
the human mind and heart. There lady Macbeth, Juliet, Desdemona, Shylock,
Richard III, Macbeth, Othelo, Romeo, Hamlet, Lear, making us believe they
are driven by a true spirit. There man at all ages, described in his most im
pressive and meaningful attitudes. This is why the name of Shakespeare fills
the realms of Earth.”
Thirdly, Shakespeare is said to be hidden and spontaneous beauty (61-Valera; 51
Alcalá Galiano; 56-Saavedra); a “little educated creator of exquisite beauty”; “a sim
ple stage-player of little instruction”, “unestimated by classical critics” (9-Cueto). As
for Milton, he is mentioned as “an eminent poet” (51-Alcalá Galiano, 35-Segovia,
61-Valera, etc.). His words are found to spring from “the purest fountains of the
Bible” (15-Martínez de la Rosa); Paradise Lost is rendered as the safest foundation of
Christianity (70-Fernández); or as a divine poem, often criticised by preceptists (56
Saavedra). To Cueto’s eyes, Milton’s Satan is “a beautifully literary monster” (9-Cueto);
“a risky, heterodox, abominable picture” for Antonio M. Segovia (35-Segovia). Finally,
Milton’s richness of vocabulary seems also worth noting (61-Valera).
(4) Cuando apareció en Inglaterra el Paraíso perdido de Milton, suscitáronse
acaloradas controversias entre los doctos sobre si era ó no esta obra poema
épico. Á los que este carácter le negaban respondieron los adversarios: “no
será épico, si os empeñais, pero será un poema divino”.
Y es que los preceptistas deducen sus reglas de lo conocido, y rara vez adivi
nan las nuevas manifestaciones que puede ofrecer el ideal de la belleza en
ulteriores revoluciones sociales. (56-Saavedra: 444 )
“When Milton’s Paradise Lost was published in England, scholars openly dis
cussed whether it was an epic poem or not. Those who denied the poem its
epic nature were replied that ‘it may not be an epic poem, if you will, but it
is divine’.
The thing is that theorists draw their rules from what is known and rarely
 Paloma Tejada Caller
guess the new guidelines that will be offered by the ideal of beauty in subse
quent social revolutions.”
Byron stands out as “the greatest and most excentric of the English poets” for 19th c.
Spanish scholars (in Escosura’s words); “a glorious poet” (61-Valera); “the most cel
ebrated among modern poets” (88-Núñez de Arce). He managed to “break up the
prudishness of tradition” (9-Cueto), and is presented as “a man of an exceptional
mind, causing agent of his own disgrace” (11-Escosura). “An ungovernable soul, prob
ably injured but never defeated” (56-Saavedra). Byron represents “unlimited spiritual
freedom, freedom of opinion, freedom of belief and behaviour”, whose “skeptical dis
dain”, “ironical bitterness” and “despise for both good and evil” are to be praised.
(90-)Alarcón points out Byron’s “moralizing morbid melancholy” and the singularity
and uniqueness of his character. Escosura and Saavedra refer to Byron’s influence on
Espronceda. Finally, in lecture 69 Cánovas claims the poet’s admission to the classical
canon. Texts 5 and 6 illustrate some of these arguments:
(5) Tuvo la desdicha para él, aunque para su patria fué gloria, de nacer en aquel
país, veintidos años antes que Esproceda en España, un hombre de ilustre
linaje, preclaro ingenio y excepcional númen poético; pero de tan excéntrico
carácter en todo, tan pródigo, tan sin miramientos sociales, y tan predispuesto
á la oposicion á todo, así en el cielo como en la tierra, que él mismo fué artí
fice de su desdicha, destructor de su fortuna, enemigo de su felicidad, y por
último, logró morir escándalo de su época.
Ese hombre era un grandísimo poeta, el autor del Sardanápalo y de Marino
de Faliero, del Giaour y de La Desposada de Abydos, del Corsario, de Lara, del
Don Juan y de La Peregrinacion de Child Harold.
Este hombre era, Señores, lord Byron. . . (11-Escosura: 88–89)
“Though nationally praised, he was unfortunate to be born in his country,
22 years before Espronceda. A man of glorious lineage, brilliant wit and out
standing poetic inspiration, his character proved so eccentric, so exuberant,
so much opposed to social conventions and so much inclined to any earthly
or heavenly fight, that he himself was a causing agent of his own disgrace, a
destroyer of his fortune, an enemy to his happiness, until he eventually died
as the scandal of his age.
This man was a great poet, the author of Sardanapalus, Marino Faliero, Giour,
The Bride of Abydos, The Corsair, Lara, Don Juan, Childe Harold. This man
was, Gentlemen, Lord Byron.”
(6) Pero en el entretanto Byron no tomaba piedras para modelar sus concep
tos, sino en las canteras que los clásicos de Inglaterra habian abierto: digno
admirador del correctísimo Pope, sus frases son elegantes y armoniosos sus
versos; y bien que se dejase llevar alguna vez de los defectos de tantos de sus
contemporáneos, ásperamente lo reconoció el mismo, exclamando que ‘No
English consciousness in 19th century Spain 
valia un solo ardite el nuevo bajel poético en que todos parecian embarcados.’
(60-Cánovas: 42)
“But in the meanwhile Byron’s building stones were taken from quarries al
ready discovered by the English classicists. As a worthy admirer of Pope’s, the
most refined of poets, his sentences turn elegant and his verses harmonious.
And though he was at times miscarried by the faults of so many of his con
temporaries, he bitterly recognised the fact, when he said that ‘the new poetic
vessel they all seemed to be on was not worth a dime’.”
Fourth in importance to 19th c. Spanish eyes, Walter Scott, is praised for his “ability to
brighten up history” (98-Menéndez Pelayo); the conservative Nocedal refers to Scott’s
achievement of “a brilliant naturalness” in lecture 40; (41-)Duque de Rivas introduces
Scott as “the unique model of historical romance”, and (94-)Conde de Casa Valencia
finds his work “better than history itself ”:
(7) Walter Scott, el inmortal Walter Scott, padre verdadero del romance histórico,
es tan eminente en conservar la índole de sus personajes y en pintar la escena
y el tiempo en que los coloca, que he oido decir várias veces al ilustradísimo
inglés Mr. H. Frére que no habia nunca comprendido bastante la historia de
Escocia hasta que leyó las novelas de Walter Scott. Este gran escritor, mode-
lo único en el importante género de los romances históricos, muy á menudo
presenta como protagonistas personajes de su invencion, y hasta personas os
curas y de ninguna importancia; pero les da una vida tan verdadera, los rodea
de figuras tan conocidas, y los hace moverse en una escena exactamente ajus
tada á la verdad histórica, que el lector se encuentra trasportado á los lugares
en que ocurren los sucesos, y ve, y oye, y trata á las personas, y vive con ellas
como su contemporáneo. (Duque de Rivas: 412–413)
“Walter Scott, the immortal Walter Scott, father to the authentic historical
romance, is so skilful at preserving his characters’ souls and at portraying
their local and temporal surroundings, that even Mr. Frere, a well-read En
glishman, has often claimed that he had been unable to fully understand
the history of Scotland until he read Walter Scott’s novels. This great writer,
unique model of the important genre of historical romance, very often se
lects imaginary, even dark or unimportant people as his main characters. But
he grants them a true and consistent life, they meet too well-known figures
and move around in an environment extremely close to the historic truth, so
that the reader feels carried away to the places where the events occurred. He
can see, listen to, even meet the people there and live among them as their
contemporary.”
But as indicated above, Spanish scholars7 also find other “poets” worth mention
ing. Among them: Addison, Bede, Chaucer, Coleridge, Dryden, Harlowe, Ben Jonson,
Moore, Poe, Pope, Shelley, Sheridan, Sterne, Tenison, Thompson or Walpole. Relevant
playwrights (Marlowe, Wester, Middleton, Heywood, Rowe, Beaumont and Fletcher,
 Paloma Tejada Caller
Wicherly, Farguhar, Congreve, Lillo, Steele, Colman, Ciber, Sheridan, Garrick) are also
selected by Cueto in his lecture on English drama (9-Cueto); and Owen, Butler, Byron
and Pope are also mentioned by (44-) Cutanda in his lecture on satire and epigram.
Moving on to our second major nucleus of references in the selected corpus of
19th c. Spanish texts, i.e. cultural features defining the English people and the essence
of Englishness, the most prominent idea is that of “time is money”, which will be re
covered when talking about the English language. The “seal of practical superiority”
(13-Pastor Díaz) left in most English enterprises is very much admired (cf. also 22-la
Puente and 15 Martínez de la Rosa); as is also the image of Britain as “a never-ending
fertile soil which never gives in” (88-Núñez de Arce). This endurance explains why
England occupies a prominent place in European civilization. It has “paved the path
of culture” for the world (88-Núñez de Arce), and its success emerges from its ac
ceptance of political and religious freedom and its secularization process; England is
said “to appreciate and devote scholarly attention to subjects disregarded in our land”
(69-Cánovas), thus “offering the world highly-educated people, well-known and solid
institutions and noble men”. Illustrative evidence for this assertion is found in the
“authority exhibited by the English press as an instrument of reason” (88-Núñez de
Arce); in linguistic research carried out in London, focussed on experimental pho
netics (4-Monlau; 77-Segovia); in the Royal Institution (4-Monlau); in the works of
philosophers and scholars such as Bacon,8 Hobbes, Locke, Newton, Herbert-Spencer
and Darwin; of historians, such as Gibbon, Goldsmith, Hume, Robertson, Hallam, the
“controversial Macaulay” and Carlyle; of economists, statesmen and orators. England’s
relevance is further supported by “the vigour of its institutions” and its “famous and
almost invisible legal code”: English people prove “legally independent and free” (44
Cutanda; 18-Mora). Finally, England’s maritime power is often recalled: “all the seas
answer to her voice” (58-Núñez de Arenas; 18-Mora).
(8) En Inglaterra, desde muy temprano fueron hombres de accion los hombres de
inteligencia. Por eso en aquella tierra de grandes políticos, los estudios pro
fundos y los grandes descubrimientos precedieron en más de un siglo al saber
de otros pueblos. Por eso en aquel país de profundos pensadores sus primeros
estadistas, desde Bacon hasta Canning, han sido literatos y filósofos. Por eso
sus historias y sus empresas, sus revoluciones y sus conquistas, sus consti
tuciones y su industria, su política y su filosofía llevan impreso desde muy
antiguo aquel sello de superioridad práctica, de solidez y de duracion que se
reproduce en su poder material, en su dilatada influencia sobre el mundo y
en la misma organizacion interior de sus gerarquías sociales.
(13-Pastor Díaz: 42–43)
“English men of action have always been men of wit, almost from the very
beginning. That is why deep research and great discoveries were carried out
more than a century earlier in that land of great politicians than in other
places. That is why the first and most important statesmen in that country of
deep thinkers, from Bacon to Canning, were writers or philosophers. That is
English consciousness in 19th century Spain 
why her history and her enterprises, her revolutions and conquests, her con
stitution and her industry, her politics and philosophy have been imprinted
for ages by the seal of practical superiority, of strength and permanence,
which mirrors her material power, her wide influence over the world and the
very inner structure of her social hierarchy.”
(9) ¿No os sorprende, señores, este estado de perpétua renovacion y florecimiento
al compararle con la estéril flaqueza á que llegamos en el siglo XVII, y de la
cual áun no hemos convalecido? Pues no busqueis su explicacion en recóndi
tas diferencias de raza, ni en desigualdades intelectuales que la sana crítica no
admite y la experiencia desmiente; buscadla sólo, y la encontrareis de fijo, en
un hecho asaz significativo que no se ha escapado á la penetracion de la his
toria. Mientras España rodaba con los estremecimientos de la agonía hasta el
fondo del abismo, y aferrada á sistemas opresores sentia helársele por grados
la sangre en sus venas, Inglaterra conservaba, y conserva todavía, la porten
tosa actividad de su espíritu, . . . porque supo á costa de inauditos esfuerzos,
tenaces luchas é incalculables sacrificios, recuperar, mantener y asegurar por
último el derecho de los ciudadanos cuando otros pueblos le abandonaban ó
perdian; siendo por esta causa quizás la primera nacion de Europa que se ha
valido, para avanzar en la senda de su cultura, de las dos irresistibles palancas
con que puede removerlo todo el entendimeinto humano: la libertad política
y el libre exámen. . . (88-Núñez de Arce: 17–21)
“Is this state of perpetual renovation and flourishing not surprising, Gen
tlemen, when compared with the meager staleness we arrived at in the 17th
c. and from which we are still suffering? Do not try to explain it because
of obscure race or mental differences, unacceptable to the honest critic and
denied by experience. The very reason is to be found in a significant fact,
not unrecorded by history. As Spain rolled down in torment into the deep
est abyss and felt her blood freezing, bound to oppressive systems, England
maintained, as she still does, an impressive spiritual activity. . . After unheard
of efforts, tough fights and endless sacrifices, she knew how to recover, keep
and preserve the rights of her citizens, in a time when other people lost them
or gave them up. That may be the reason why she has been the first nation in
Europe to use the strongest two levers of the human mind, capable of turning
everything around in order to advance through the path of culture: those of
political and moral freedom.”
However, not all features of Englishness selected by members of the Spanish Royal
Academy sound equally attractive. Though an assertion sustained by merely a few (cf.
89-Valera and 103-Nocedal), England is reported to have “engendered false myths”
regarding Spain. England has broadcast the idea that the “Spanish decline was due to
religion”, that the “Inquisition was an exclusive Spanish practice (without considering
the fact that these practices were more fierce and violent in England than in southern
Europe)”. Furthermore, Spanish authors want to underline that “religious freedom
 Paloma Tejada Caller
did not come with protestantism, as is often claimed”; and that “the alliance between
a theocratic system and the king’s power was not just a Spanish phenomenon”. For
them, the idea that “Bacon’s Novum Organum sets up the age of reason” must also be
discarded; and particular injuries or contradictions as those expounded by Ticknor or
Macaulay must be fought.
As said above, the third nucleus organizing Spanish references to English in our
corpus refers specifically to the English language. In the first place, English is described
as a language consistently related with the “English soul” (87-Canalejas; 61-Valera;
18-Mora; 38-Monlau). That means that English is a language “able to bluntly dissect
thoughts” (62-Canalejas); it is a “clear and precise tongue, not indulging in mistakes
and ambiguity” (72-Olózaga, etc.). For some, the English language also reflects the
cultural importance of money (according to 82-León Galindo “it is an admirable lan
guage for figures and calculation”) and the productive relationship between time and
money (English is “a concise, monosyllabic, proud and practical language” according
to the same author). Finally, we are said that English proves an “active and democratic”
language (22-Fermín de la Puente):
(10) El Inglés, orgulloso y práctico, para quien el tiempo es dinero; úsalo conciso,
cortado, esdrújulo, monosilábico. Un minuto que ahorre en hablar, puede
dedicarlo á ocupaciones que le produzcan un penique ó evitar que se retarde
el cumplimiento de una órden. ¡Admirable idioma para cálculos y cuentas y
arrogantes preceptos! (82-Galindo: 16)
“The Englishman, proud and pragmatic, thinks that time is money. He uses it
concise, clipped, proparoxitone, one-syllabled. The minute he saves in conver
sation, he may devote to occupations from which he may earn a penny; that
minute may also be needed for the fulfilment of orders. What an admirable
language for figures and calculation, as well as for arrogant rules!”
(11) Confieso que estos juicios téngolos por exactos, á lo ménos en su mayor parte,
y lo mismo el del Sr. Galindo, especialmente en cuanto á los ingleses, como
que aquella Nacion ha profesado por axioma que el tiempo es dinero, diferen
ciándose completamente en este punto (. . . ) de nuestra gente, que dice que
gana tiempo cuando le pierde, y que hace tiempo cuando lastimosamente le
deshace ó malgasta. (22-la Puente: 46–47)
“I confess I believe these ideas are true, or mostly so. The same I think about
Galindo’s assertion on the English axiom that time is money. Things are rad
ically different among us. We say we ‘earn time’ or even ‘make time’ when we
are sadly just wasting it.”
Secondly, English is seen as a value-laden language, as shown by the adjectives used
to describe it: “perfect” (2-Alcalá Galiano), “energetic” (2-Alcalá Galiano), “extremely
rich” (18-Mora); “independent from law”; “cautious” (58-Núñez Arenas); “clear and
precise” (72-Olózaga); “active and democratic, strange and unstructured”; “harsh and
rough” (88-Núñez de Arce); “phonetically-faulty” (76-Arnao); “weak” (as far as it is
English consciousness in 19th century Spain 
analytic, according to Valera). Thirdly, it is interesting to note that the Germanic ori
gins of English are also emphasized in our corpus. Authors like Canalejas (in lecture
62) introduce Grimm and his research on Germanic languages as an essential land
mark of the new philological science; (86-)Pascual argues on the “origins of the proud
and fierce Germanic peoples”; (2-)Alcalá Galiano reports the historical changes un
dergone by the English language; and (38-)Monlau and (87-)Canalejas focus on the
Germanic layer in the Castilian language as an unexpected influence on its evolution,
paying special attention to morphology and lexicon.
A further issue, of lesser weight, should be mentioned. Spanish academicians rec
ommend taking English into consideration because it is the language of Shakespeare
and Milton and a profitable instrument to fight French (a clearer enemy) and to im
prove our knowledge of Spanish. It should be noted that the number of warnings
against an English-biased use of Spanish words is very small in this period, as com
pared to what we get in the first half of the 20th c. Only some hints are given about
English borrowings through French; on syntactic anglicisms (72-Olózaga); and on two
“badly used” words: pauperismo (4-Monlau) and docks (22-la Puente).
More specific linguistic features of English are also focussed upon. From a grapho
phonological standpoint, (61-)Valera highlights the “lack of correspondence between
English pronunciation and etymology”. (76-)Arnao comments on what he calls “vocal
flaws of English”, whereas 77-Segovia presents English as a language rich in sounds. In
grammatical terms, English is described as an analytic language (which equates to a
poor system, according to 61-Valera), consisting of a fixed structure and a strikingly
plain verbal system as opposed to Spanish (18-Mora):
(12) Notorio es á todos que en la pronunciacion dificilísima de esta lengua (tan
varia y controvertida hasta el punto de decir algunos gramáticos que la mejor
regla para aprenderla es no dar ninguna respecto de ella) sobre haber fre
cuentes aspiraciones, figuran diez sonidos vocales, de muy oscuro eco al
gunos; hay consonantes de poco valor eufónico, y casi nunca se emite la voz
franca y abiertamente, sino por el contrario, sale envuelta en las palabras de la
boca medio cerrada, circunstancia totalmente opuesta al ore rotundo de Ho
racio, y á las ineludibles leyes del canto y la vocalizacion. Con tales defectos,
y con los medianos recursos de su métrica, el ingles puede conceptuarse en el
expresado punto de vista como el reverso del italiano, lo cual para el caso no
hace seguramente su panegírico. (76-Arnao: 36)
“We are all familiar with the extremely difficult pronunciation of this language
(so varied and confusing that the best rule to learn it is to forget about rules,
according to some grammarians). Leaving aspiration aside, in English there
are 10 vowel sounds, some of them with a very dark echo; there are consonants
with little euphonic value and sound seldom comes out openly. Very much on
the contrary, sound comes out of a half-opened mouth, wrapped in words, far
from Horatio’s ore rotundo and from the inevitable laws of singing and vocal
izing. Considering such flaws and the lack of metrical resources, the English
 Paloma Tejada Caller
language may be considered the reverse of Italian in this respect, which is not
probably very praising.”
(13) . . . las palabras se trasladan de un pueblo á otro; pero la gramática es una
propiedad inenajenable; es un elemento fijo que no se altera, ni se enriquece,
ni se perfecciona. Así es que las lenguas semíticas, que carecen de tiempos
compuestos, no los han adquirido, ni imitado de las otras lenguas que los
poseen; así es que las lenguas más ricas del Sur de Europa carecen de partici
pio de presente, sin embargo de hallarse en la lengua de donde han tomado
sus partes constitutivas y la mayor porcion de sus riquezas; así es como el
número dual es una propiedad exclusiva del griego; así es, finalmente, como
los ingleses no han admitido, ni el número, ni el género en sus adjetivos.
(18-de Mora: 146)
“Words fly from one place to the other, but grammar is inalienable. It con
stitutes a fixed element, unable to change, to be enlarged or bettered. That
explains the fact that the Semitic languages, which lack compound verbs, have
not adopted them from other languages; that explains the fact that the rich
languages in Southern Europe lack present participles, although this element
exists in the language from which they have taken their constituent elements
and most of their treasures. That explains the dual number as an exclusive
property of Greek; and finally, that explains why English has not adopted the
categories of number and gender for its adjectives.”
But the linguistic layer most amply discussed is vocabulary. First, the “outstanding
combination of Germanic and Latinized words” in English is recurrently praised, and
its “tolerance towards new words” is often referred to as admirable: English seems to
be “undisturbed by foreign and made-up terms”; “it never rejects lexical innovation”,
its vocabulary therefore “becomes richer and richer”:
(14) Entre las lenguas modernas se observa gran diferencia en la facilidad ó rigor
con que se prestan á la admision de voces extrañas. Las del Norte se dis
tinguen por la suma latitud que conceden á la innovacion; latitud que raya
en los límites de la anarquía, y de que se aprovechan, no sólo los escritores
que tratan materias científicas y recónditas, sino los que manejan los asuntos
más comunes y vulgares. La lengua inglesa no esquiva ningun neologismo,
cualquiera que sea su procedencia, con tal de que conserve toda su integri
dad y su terminacion nativa. Así es que, á pesar de no tener voces que acaben
en i vocal, han tomado la voz banditti del italiano, y careciendo del sonido
gutural de la j, y del que nosotros damos á la ll, han tomado del español
junta, guerrilla y camarilla. Por este medio han conseguido poseer una lengua
riquísima, y que cada dia aumenta su vocabulario. Los ingleses, tan amigos de
la legalidad, como independientes y libres en el círculo que ella les traza, no
reconocen autoridad constituida en materia de idioma. Cuando les acomoda
trasladar un sentido de la cosa á la acción, convierten el sustantivo en verbo;
English consciousness in 19th century Spain 
cuando quieren expresar en una sola palabra un sentido complicado, de dos

ó tres voces simples forman un adjetivo compuesto, y si una voz de cualquier
otro idioma les parece más oportuna, más expresiva ó más sonora que la que
poseen en el suyo, la adoptan sin reparo, y le conceden sin formalidad alguna
el derecho de ciudadanía. (18-de Mora: 150–151)
“Modern languages differ much in that they do not apply the same rules to the
admission of foreign or strange words. Northern welcoming to innovation is
well-known. Freedom in this respect comes close to anarchy and benefits both
scientific writers and those dealing with more common subjects. The English
do not reject any new word, wherever it may come from, as far as it keeps its
integrity and native ending. So, they have taken the word banditti from Italian,
though they lack words ending in -i; and junta, guerrilla and camarilla from
Spanish, even though their language lacks the guttural sound j and that of ll. In
this way they have acquired a very rich language, richer and richer every day.
The English, so prone to legality and so independent within its boundaries,
recognise no authority in linguistic matters. When they feel the need to ex
press the sense of an object as an action, they turn the noun into a verb; when
they wish to express a difficult idea in just one word, they create a compound
adjective out of two or three single words; when they find a foreign word
more adequate, more expressive or euphonic than theirs, they adopt it with
no objection and bestow English citizenship on it with no qualms whatever.”
Although many texts seem to abound in references to this idea, it is not the only feature
of English vocabulary which attracts the attention of Spanish scholars. The “notice
able imprint of Biblical language on English vocabulary” is also pointed out in (18-)de
Mora. And again, its ability to reflect English cultural peculiarities, which explains
English linguistic “export of nautical and maritime terms”. Finally it is to be noted
that Johnson’s dictionary is confirmed as “the book regulating the English language”
(2-Alcalá Galiano):
(15) Por aquellos días apareció en la palestra literaria Samuel Johnson, á cuyo ape
llido agregan siempre sus compatricios el título de doctor, porque lo era en
leyes; hombre que ejerció en su patria grandísimo influjo, y hasta cierta domi
nacion, la cual se avenia con su condicion adusta, imperiosa, impaciente y
despreciadora de todo miramiento y toda regla de cortesía. Formó este autor
por sí solo y publicó un diccionario de su lengua; obra en todos los demas
pueblos de varios individuos juntos en academias, y tuvo tanto acierto, que es
hoy mismo su trabajo, con pocas adiciones y correcciones, el libro regulador
de su idioma, en cuanto cabe ser un libro regulador obedecido de la turba de
escritores, de los cuales muchos no se sujetan á disciplina, ni acatan ni tienen
por legítima autoridad alguna en punto á lenguaje. (2-Alcalá Galiano: 13)
“Then Samuel Johnson turned up on the literary scene. He was often referred
to as Dr. Johnson, because he was Dr. in Law. This man exercised a great
 Paloma Tejada Caller
influence on his land, to the extent of domination, a fact fitting his stern,
overriding, impatient and strong nature, not indulging in courtesy or rules of
politeness. He was a self-educated man and published an English dictionary,
the type of work compiled elsewhere by several people. His talent and success
was so great that his work (with few additions and corrections) is still used to
regulate his language, in as far as a book may be said to regulate the activities
of a mob of writers, many of whom reject all rules and deny any legitimate
authority in matters of language.”
. An analysis of the results
The study conducted on the lectures read by members of the Spanish Royal Academy
of the Language supports the hypothesis that they give rise to an image of English. That
is, the selected texts may be said to produce a mental representation, which conven
tionally manages to substitute, interpret, translate or model reality. And like any other
image, the one arising from the texts analysed here exhibits a selective nature, serves
particular functions and is open to re-elaboration.
Considering the extension of references in absolute terms, the image we get from
19th century Spanish texts may seem poor or scanty; there are no complete discourses
referring to English, and opinions on the language usually serve as an argumentative
excuse for longer debates, but the image may be said to be well defined. This means
that the resulting image organizes its elements in a structured way, around the function
devised for the whole. In addition, the image may prove paradoxically contradictory.
In more definite terms, the resulting image focusses on eight major elements to at
tract the observer’s attention. One, English is a distinct language, deserving the reader’s
attention as a cultural sign. Despite the unconnected nature of many of the references
found, global percentages prove surprisingly high. As said above, English is mentioned
in more than 40% of the lectures and by circa 47% of the authors. Even considering just
the more extensive references, it is to be noted that approximately 25% of the authors
adopt English as a partial topic for their lectures and that some 16.5% of the lectures
make more extensive reference to English issues. That means that the Spanish Royal
Academy echoes the existence of a culturally distinct product and seems attracted by
the existence of (the English as) “others”.
Two, English essentially equates to its literary canon. Research indicated that more
than half of the selected lectures deal with literature. The adjectives used to describe
English writers refer to the universal and atemporal nature of their work and recog
nise the extent to which they have contributed to the construction of the national
character of English literature. Almost all of the British writers mentioned belong to
the conventional canon; and there are even lectures offering closed lists of authors
as a short history of English literature. As for those writers most often singled out, a
deeper refinement of analysis would be required. Shakespeare, Milton, Byron and Scott
undisputably belong to the British literary canon. However, the frequent reference to
English consciousness in 19th century Spain 
Shakespeare and the highlighted features of his work are justified not only because of
the universal nature of his art, but also because he contributes to an emphasis on the
Spanish Golden Age of literature and drama. As for Milton, leaving aside the universal
character of his art, he serves the religious interests of the most conservative members
of the Spanish Academy (cf. Cayetano Fernández). Byron and Scott represent inno
vation, to a greater or lesser extent: Byron is introduced as the utmost representative
of change; Scott, a quieter and less radical model, will become the preferred romantic
icon in more conservative times of later Spanish politics.
Three, the English language reveals the English mind and culture, expressing its peo
ple’s national singularity. The results confirm that apparently distinct categories are
frequently mixed up; i.e. ideas on language, spelling and grammar often become inter
twined with those of nation, power and tradition. Four, English prestige derives from
the civilized nature of its people. The evidence showed that writers recurrently point to
preferred signals of civilization: the high degree of English culture, education and free
dom; the strength of institutions, the pragmatic nature of its speakers, or the linear,
unbroken line of English history. Five, English is shown to be a prominent, relevant
and valued language within the so-called 19th c. language hierarchy, because it suits
19th c. conventional parameters for what a language must be. More particularly, the
constructed image of English selects features commonly used for the diagnosis of real
languages: their written form, their high degree of lexical elaboration,9 their precise
nature and the clarity of their linguistic system (cf. Swiggers 1990: 119).
Six, the English language is value-laden, in that it behaves as the symbolic expo
nent of the social, moral, intellectual or political character of its people, as already
seen above. Seven, the English language, as defined by members of the Spanish Royal
Academy, represents the community, which creates an illusion of homogeneity. The im
age of English avoids any reference to intrasocial linguistic factors, such as social class,
socioeconomic status or mobility differences affecting English speakers. And eight,
English is portrayed as a contradictory reality. Institutional images of languages seldom
offer a global picture. They may act as coherent systems of meaning, and still prove
internally contradictory. It should be remembered that English is seen both as a re
spectable and as a non-respectable language (against which one must fight); it is both
a clear language, and one of dark vowels; it proves both orderly and anarchical at the
same time.
The study conducted on the lectures read by members of the Spanish Royal
Academy of the Language also supports the hypothesis that the image of English
obtained from institutional texts is organized around a previously devised function.
This function may be said to be instrumental, of a symbolic (and ideological) nature,
though apparently descriptive. It seems clear that highlighted elements are determined
by their instrumental value. Results show that English is used as: (i) an implicit mech
anism to reinforce nation-imagining myths, and (ii) a reasonably explicit model of
innovation.
Not surprisingly, English is used as a model to defend and construct universal, per
manent and inalterable myths around Spanish as a national language, mainly referred
 Paloma Tejada Caller
to the purity and epics in its cultural history. Through explicit contrasts and opposi
tions, English acts as the standard against which Spanish is measured, as though the
recognition of others’ values would allow readers/listeners to identify something as
one’s own.10
But even more interesting than these explicit and contrastive references to uni
versal values are those in which English is merely described. As responsible codifiers
of linguistic images members of the Spanish Royal Academy confirm their belonging
to a larger community of discourse, sharing common interests, goals and beliefs with
similar institutions in neighbouring countries. English, defined as a civilized, literary,
moral, homogeneous language, as a language enjoying an uninterrupted linear history,
with a glorious past, implicitly serves as a model for what Spanish already is, must be
or must be considered to be, no matter how Spanish was previously conceived. From
now on Spanish is essentially its literature and a system of moral value, representing
the spirit of a clearly distinct and homogeneous people (any trace of social movements
such as associationism, cantonalism, anticlericalism, anarchism or socialism existing
in 1880 Spain is thus made unnecessary and is kept out of the picture).
But, as suggested above, the image of English elaborated by members of the Span
ish Academy performs a second function: that of being a model of innovation. Oth
erness, English in this case, is presented as a space for experimentation and learning.
And thus, it is offered as the guideline to act in particular circumstances; it is the right
instrument to define patterns of change. This is perhaps the most surprising discovery
of our analysis. Contrary to expectations, the Academy is not serving now the inter
ests of ultra-conservatives, but echoing the voice of innovators. Change – we are told –
must come from education, from the secularization of culture, from tolerance, from
economic dynamism and initiative. English becomes the model of individual freedom
and moral self-bettering; of individual autonomy against the ruins of confessional dog
matism governing the official education in Spain.11 In sum, the dream of modernity is
reflected in the image of the English language reconstructed in our analysis.
Furthermore, English is portrayed as the door to two new cultural trends: Ro
manticism and Historical Linguistics. It should be remembered that although Herder,
Müller, Grimm and others were not translated into Spanish until the 20th c. (Artola
1973: 330), in the texts analysed there is a constant reference to the abolition of aes
thetic laws, and of moralizing classicism (Valera, Alcalá Galiano, Saavedra, Cueto).
Byron, Shakespeare and even Milton are praised for their rebellion.
As for Historical Linguistics, the fact that Müller’s theoretical perspective, the ori
gins of Germanic, and the linguistic reconstruction of Spanish were selected as basic
topics of institutional lectures must be taken as a meaningful step against those who
still remained anchored in the study of Classical literature. This argument is further
supported by the fact that the most extensive references to English are concentrated
in lectures read in the 60s and 70s and are introduced mainly by intellectuals who be
longed to a new, more open-minded and liberal aristocracy. We may assume that the
two functions mentioned above, i.e. that of being a model of innovation and creator
of national myths, proved useful to supporters of the so-called bourgeois revolution
English consciousness in 19th century Spain 
of 1870s, who claimed the introduction of political reforms, greater freedom and in
stitutional weight, and rejected both the standpoint of the most conservative Church
and of proletarianism or other more radical movements eroding the state.12 This is
the sociological profile of authors such as Alcalá Galiano, Monlau, Cueto, Escosura, de
Mora, Cutanda, Valera, Saavedra, Canalejas, Fernández, Olózaga, la Puente, Pascual,
Núñez de Arce or Alarcón. They belonged to well-educated professional groups who
came easily into touch with European political and philosophical perspectives. Some –
Valera, Olózaga, Cueto and la Puente – had lived abroad as diplomats and others –
Alcalá Galiano, Escosura, de Mora or Pascual – found their temporary home in Eng
land as political exiles; and most of them maintained liberal or even radical ideas for
at least part of their lives.
To conclude this subsection it should be noted that as far as English serves as a
model, otherness is not felt as menacing. English does not represent the incompatible
other or opponent of later times. In the period analysed there is no mention of cultural
or linguistic invasions; no reference to the Spanish Armada or to competitive Empires.
That will come in later on.
The present study also confirms the idea that the image of English generated by
19th c. members of the Spanish Royal Academy of the Language exhibits a directive
nature, in that it is authorized and likely to exert social influence. Although discon
tinuous, the image we get from these texts is offered as technical and exhibits a very
well studied rhetoric of legitimacy, which (i) limits the proper object of study, (ii) nat
uralizes potentially controverted aspects or individual beliefs and opinions, and (iii)
makes the reader believe that this subjective framework of interpretation is objective
description. As seen above, from the long list of literary authors only some are chosen:
the unchallenged Shakespeare, the religious Milton, the lawless Byron, the quiet and
romantic Scott; from the cultural standpoint, only those aspects in which we Spanish
people would like to be mirrored are selected; and linguistically, those topics related
to the elaboration of nation-imagining myths are preferred. Individual opinions and
beliefs are presented as unanswerable and irrefutable truths, as though they were part
of universally accepted knowledge. The lack of discussion and argumentation, the fact
that pronouncements are made in a vacuum, and the almost absolute absence of schol
arly references make it difficult for not so powerful groups to counterbalance these
assertions or to evaluate their weight. English is what experts want it to be. Although
a more difficult issue to evaluate is the precise influence that this constructed image of
English actually exerted on the different groups of Spanish society,13 nevertheless, the
strength and resistance of nation-imagining myths implicitly sustained through the
image of English should not be disdained. And neither should the survival of cultural
clichés regarding Englishness. Though these issues deserve much further attention, it
seems clear that the institutional image of English, its representation, has conditioned
our experience of otherness.
 Paloma Tejada Caller
This study has provided evidence that the intercultural analysis of metalinguistic pub
lic lectures contributes to the discovery of new patterns of identity. In particular, a new
idea of English and Englishness has been shown to emerge from Spanish institutional
19th c. texts. Results indicate that English is conventionally described as a distinct, civ
ilized, literary, value-laden language, reflecting the spirit of a homogeneous, educated,
tolerant, dynamic and thriving people. This well-defined image responds to the inter
ests of a Spanish new liberal elite, who codify the idea of English as a model to reinforce
nation-imagining myths and as a guideline for social and political reform: English as a
standard of “universal” values and as a space for experimentation and learning.
Notes
. The whole project covers the much longer period 1847–2000, and aims to compare 19th and
20th c. Spanish institutional attitudes towards English. Results will be published as a book, with
complete texts and references, by P. Tejada and M. Nadales.
. See Artola (1973); Carr (1969); Kamen (1967); Llorens (1968).
. Only Vol. 1 (up to 1884) was considered for the research.
. Lectures number: 2, 4, 9, 11, 18, 38, 44, 50, 56, 62, 63, 70, 72, 83, 86, 88, 89, 90.
. Unfortunately, texts justifying all our assertions are too many and/or too long to be repro
duced here. We will only offer some illustrative examples. In all of them the original spelling is
maintained. A working version of the texts is provided for those unfamiliar with the Spanish
language, though it lacks the attractive 19th c. flavour of the original.
. The number accompanying the author’s name refers to the number of discourse as in the
Appendix.
. Cf. 2-Alcalá Galiano; 62-Canalejas; 69-Cánovas; 94-Conde Casa Valencia; 41-Duque de Ri
vas; 88-Núñez de Arce; 22-la Puente; 92-Saavedra; 35-Segovia; 50, 61-Valera.
. Cf. 2-Alcalá Galiano; 15-Martínez de la Rosa; 88-Núñez de Arce; 13-Pastor Díaz; 89-Valera.
. Note references devoted in different lectures to the clarity and precision of English (72
Olózaga) and to the richness of English vocabulary in general and that found in Shakespeare
or Milton in particular cf. 2-Alcalá Galiano; 18-de Mora; 63-Valera, etc.
. A certain tone of superiority and defence is sometimes observed, to vindicate the universal
appreciation of Spanish values (see the excerpt in (10) above). It should be remembered that
this kind of ethnocentric attitude often characterizes the approach to others and otherness.
. It may be worth remembering that the Institución Libre de Enseñanza was founded in 1876
following English patterns, illustrating the clear influence England exerted on some sections of
leftish liberalism.
. This probably also explains why texts only reproduce nation-imagining myths of a gen
eral nature, discarding other more specific ones the English were simultaneously forging about
themselves in England: English superiority as a universal language; the need to have an educated
language for social ascension; the original purity of Germanic vocabulary, the creation of the
English consciousness in 19th century Spain 
Oxford English Dictionary as a national monument, etc. (cf. Aarsleff 1983, Bailey 1996). The
fact that these arguments remain unmentioned may be due to a lack of knowledge or to the fact
that they proved unnecessary to the lecturers’ purposes of moderate regeneration.
. On the one hand, it is well known that the Royal Academy of the Language was the only
recognised cultural institution of the times, together with the so-called Economic Societies. But
on the other, we should not forget that its lectures remained unpublished. The influence of their
opinions could therefore only be indirect. It should also be borne in mind that most of the
Spanish population was illiterate; that the institution exerting the strongest influence was the
Church; and that topics like the ones selected here most probably only proved interesting to a
very reduced number of people.
Appendix
Opening and Reception Lectures read at the Spanish Royal Academy of the Language during the
period 1847–1884 with reference to English. Asterisks indicate a longer reference. Numbers have
been added for the sake of clarity.
DISCURSOS INAUGURALES (‘Opening lectures’): 1860–1870

From (i) Real Academia Española. Discursos inaugurales (undated)
*2 Alcalá Galiano, Antonio 29 sept. 1861 pp. 1–28
*4 Monlau, Pedro Felipe 27 sept. 1863 pp. 1–123
*9 Cueto, Leopoldo Augusto de ? 1868 pp. 1–49
*11 Escosura, Patricio de la ? 1870 pp. 1–144
DISCURSOS DE RECEPCIÓN (‘Reception lectures’)

(Lectures on the left hand side of the page are “Reception Lectures”. Those on the right hand
side, the corresponding “Answering Lectures”)
From (ii) Real Academia Española. 1860. Discursos de recepción. Serie I. 3 Vols. Madrid:
Imprenta Nacional
Volume I
13 Pastor Díaz, Nicomedes pp. 26–48 15 Martínez de la Rosa, Fco pp. 77–87
17 Martínez de la Rosa, Fco. pp. 23–134
*18 Mora, José Joaquín de pp. 135–158 19 Gil y Zárate, Antonio pp. 159–171
22 Puente y Apezechea, Fermín pp. 217–308 21 Fdez. De Velasco, Bernardino pp. 199–215
Volume II
32 Cueto, Leopoldo Augusto de pp. 133–175 35 Segovia Antonio María pp. 223–251
36 Tamayo y Baus, Manuel pp. 253–290
*38 Monlau, Pedro Felipe pp. 307–330
40 Nocedal, Cándido pp. 371–402 41 Duque de Rivas pp. 405–414
 Paloma Tejada Caller
Volume III
*44 Cutanda, Francisco 17 March 1861 pp.
3–37
*50 Valera, Juan 16 March 1862 pp. 223–259 51 Alcalá Galiano, A. pp. 261–281
54 González Bravo, Luis 1 March 1863 pp.
347–395
*56 Saavedra, Enrique 14 May 1863 pp.
433–467
58 Núñez Arenas, Isaac 13 December 1863 pp.
505–537
From (iii) Real Academia Española. Discursos de recepción, Vol. 1: 1867–(undated)
60 Cánovas, Antonio November 1867 pp. 1–62 61 Valera, Juan pp. 63–89
*62 Canalejas, F.de Paula 28 November 1869 *63 Valera, Juan pp. 73–116
pp. 1–72 65 Marqués de Molíns pp. 49–84
68 Silvela, Manuel 25 March 1871 pp. 1–54 69 Cánovas del Castillo, A. pp. 55–108
*70 Fernández, Cayetano 16 April 1871 pp.
1–44
*72 Olózaga, Salustiano 23 April 1871 pp. 1–2
From (iv) Real Academia Española. Discursos de recepción, Vol. 2: 1872–1879 (undated)
76 Arnao, Antonio 30 March 1873 pp. 1–41 77 Segovia, Antonio María pp. 43–79
*83 Lapuente, Fermín pp. 29–68
82 Galindo, León ? 1875 pp. 1–27 87 Canalejas, Fco. Paula pp. 109–132
*86 Pascual, Agustín 30 April 1876 pp. 1–107 *89 Valera, Juan pp. 46–80
*88 Núñez de Arce, G. 21 May 1876 pp. 1–44 95 Valera, Juan pp. 54–78
*90 Alarcón, Pedro A. 25 February 1877 pp.
1–44
92 Saavedra, Eduardo 29 December 1878 pp.
1–56
94 Conde Casa Valencia 30 March 1879 pp.
1–52
From (v) Real Academia Española. 1947. Discursos de recepción. Serie Segunda, 4 Vols. 1881–
1898. Madrid: Gráficas Ultra
Volumen I
98 Menéndez Pelayo, M. 6 March 1881 pp. 9–54 103 Nocedal, Cándido pp. 177–199
References
Aarsleff, Hans (1983). The Study of Language in England 1780–1860. London: Athlone Press.
Artola, Miguel (1973). La burguesía revolucionaria (1808–1869). Madrid: Alianza.
Bailey, Richard W. (1992). Images of English. A Cultural History of the Language. Cambridge:

English consciousness in 19th century Spain 
Bailey, Richard W. (1996). Nineteenth-century English. Ann Arbor: The University of Michigan
Press.
Blommaert, Jan & Jef Verschueren (1998). “The role of language in European nationalist
ideologies”. In Schieffelin et al. (Eds.), Language Ideologies. Practice and Theory (pp. 189–
209). New York and Oxford: Oxford University Press.
Brown, Gillian (1990). “Cultural values: the interpretation of discourse”. ELT Journal, 44(1),
11–17.
Burke, Peter (1987). “Introduction”. In P. Burke & R. Porter (Eds.), The Social History of
Language (pp. 1–20). Cambridge: Cambridge University Press.
Carr, Raymond (1969). España 1808–1939. Barcelona: Ariel.
Corfield, Penelope J. (Ed.). (1991). Language and History. Oxford: Basil Blackwell.
Crossick, Geoffrey (1991). “From gentlemen to the residuum: languages of social description in
Victorian Britain”. In P. J. Corfield (Ed.), Language and History (pp. 150–177). Oxford: Basil
Blackwell.
Crowley, Tony (1989). The Politics of Discourse. Houndmills: MacMillan.
Crowley, Tony (1990). “That obscure object of desire: a science of language”. In J. Joseph & T. J.
Taylor (Eds.), Ideologies of language (pp. 27–50). London: Routledge.
Crowley, Tony (1991). Proper English? London: Routledge.
Crowley, Tony (1996). Language in History. London and New York: Routledge.
Dellinger, Brett (1995). Critical Discourse Analysis. http://users.utu.fi/bredelli/cda.html
Dirven, René et al. (Eds.). (2001). Language and Ideology, Vol. I: Theoretical Cognitive Approaches.
Hawkins, Bruce (2000). “Incorporating tensions: on the treatment of ideology in cognitive
linguistics”. In R. Dirven et al. (Eds.), Language and Ideology. Vol. I: Theoretical Cognitive
Approaches (pp. 1–22). Amsterdam and Philadelphia: John Benjamins.
Joseph, John E. & Talbot J. Taylor (Eds.). (1990). “Introduction”. In J. Joseph & T. J. Taylor (Eds.),
Ideologies of Language (pp. 1–6). London: Routledge.
Kamen, Henry (1967). Nacimiento y desarrollo de la tolerancia en la Europa moderna. Madrid:
Alianza.
Kay, Paul (1987). “Linguistic competence and folk theories of language: two English hedges”.
In D. Holland & N. Quinn (Eds.), Cultural Models in Language and Thought (pp. 67–77).
Koerner, E. F. Konrad (1989). Practising Linguistic Historiography: Selected Essays. Amsterdam:
John Benjamins.
Koerner, E. F. Konrad (2000). “Ideology in 19th and 20th century study of language: a neglected
aspect of linguistic historiography”. Indogermanische Forschungen, 105, 1–26.
Koerner, E. F. Konrad (2001). “Linguistics and ideology in 19th and 20th century studies of
language”. In R. Dirven et al. (Eds.), Language and Ideology. Vol. I: Theoretical Cognitive
Approaches (pp. 253–276). Amsterdam and Philadelphia: John Benjamins.
Kress, Gunther (1990). “Critical Discourse Analysis”. In R. Kaplan (Ed.), Annual Review of
Applied Linguistics. Vol. 11, 3–9, 11–13.
Kress, Gunther & Robert Hodge (1979). Language as Ideology. London: Routledge.
Leith, Dick (1996). A Social History of English. London: Longman.
Llorens, Vicent (1968). Liberales y románticos. Una emigración española en Inglaterra (1823–
1834). Madrid: Castalia.
Philips, Susan U. (1998). “Language ideologies in institutions of power. A commentary”. In B.
Schieffelin et al. (Eds.), Language Ideologies. Practice and Theory (pp. 211–225). New York
and Oxford: Oxford University Press.
 Paloma Tejada Caller
Rissanen, Matti et al. (Eds.). (1992). History of Englishes. New York: Walter de Gruyter.
Schieffelin, Bambi B. et al. (Eds.). (1998). Language Ideologies. Practice and Theory. New York
and Oxford: Oxford University Press.
Silverstein, Michael (1998). “The uses and utility of ideology. A commentary”. In B. Schieffelin
et al. (Eds.), Language ideologies. Practice and Theory (pp. 123–145). New York and Oxford:
Swiggers, Pierre (1990). “Ideology and the ‘clarity’ of French”. In J. Joseph & T. J. Taylor (Eds.),
Ideologies of Language (pp. 112–130). London: Routledge.
Taylor, Talbot J. (1990). “Which is to be master? The institutionalization of authority in the
science of language”. In J. Joseph & T. J. Taylor (Eds.), Ideologies of Language (pp. 9–26).
London: Routledge.
van Dijk, Teun (1991). “Racism and the press”. In R. Miles (Ed.), Critical Studies in Racism and
Migration (pp. 44–67). New York: Routledge.
Ward, Michael T. (1990). “Celso Cittadini and the origin of the vernacular: the convergence
of science and subjectivity”. In J. Joseph & T. J. Taylor (Eds.), Ideologies of Language (pp.
97–111). London: Routledge.
Williams, Raymond (1983). Keywords. London: Fontana.
Wodak, Ruth (Ed.). (1989). Language, Power and Ideology: Studies in Political Discourse. Ams
terdam and Philadelphia: John Benjamins.
Woolard, Kathryn (1998). “Introduction: language ideology as a field of inquiry”. In B.
Schieffelin et al. (Eds.), Language Ideologies: Practice and Theory (pp. 3–47). New York and
Zamora Vicente, Alonso (1999). La Real Academia Española. Madrid: Espasa.
Language index
B G P
Baltic Languages  Ga(l)lego ,  Polish , 
German , , , , , , Portuguese 
C , , , , 
Catalan  Greek , , , 
Chinese , ,  R
Czech  H Russian , , 
Hindi 
D
Hungarian , , , , ,
Danish  S

Dutch , , ,  Seneca 
Dyirbal 
I Slovene 
E Icelandic  Spanish , –, , , –,
English , , , , , , , , Indo-European  –, –, , –,
–, , , , , , , , Irish  , , , , , , ,
, , , , , , –, Italian , , , , , , –, , , , ,
–, , –, , , , , , – –, , , ,
–, –, –, , –, , , –,
, –, , , , , J –, , , , , ,
, , –, , , , Japanese , , , , , , –, –, ,
, , , , , , , , –, , , –, –, , –,
, , –, , , , – –, –, –,
, , , , , , –
–, , –, L Swedish , , , , 
–, , –, –, Lakhota 
, , , , , , , Latin , , 
, , –, –, T
–, –, – M Tagalog 
Maltese  Turkish , 
F Maori 
Finnish , , , 
French , , , , , , N W
, ,  Norwegian , ,  Walbiri 
Scholars index
[Note: Some contributors to this volume have adopted the practice of hyphenating the two surnames of
Spanish scholars, others have not. For the sake of consistency, all such names are given without hyphen
ation in this index.]
A Barbera, Manuel 
Busse, Dietrich 
Aarsleff, Hans  Barcelona Sánchez, Antonio ,
Butler, Christopher S. , , , ,
Aarts, Flor ,   , , , , , , , , ,
Abercrombie, David , , Barlow, Michael , , ,  , , , , , –,
 Barth, Isabel ,  
Adams, Corinne  Barthes, Roland  Buxton, Barbara 
Agard, Frederick B. ,  Becka, Jan V.  Bybee, Joan L. , , , 
Ahrenberg, Lars  Bello, Andrés  Byron, Donna K. 
Aijmer, Karin , , ,  Benson, Cathy 
Akmajian, Adrian , , , Bentivoglio, Paola  C
 Berry, Margaret  Cacciari, Cristina 
Alarcos Llorach, Emilio  Biber, Douglas , , , Caffarel, Alice 
Alharbi, Lafi M.  –,  Calzolari, Nicoletta 
Alonso Raya, Rosario  Birnbaum, Henrik  Cameron, Richard 
Alsina, Victòria  Birner, Betty , , ,  Cano (Aguilar), Rafael , ,
Altenberg, Bengt , , , , Blommaert, Jan  , 
, , –, ,  Blum-Kulka, Shoshana  Cárdenas, Daniel 
Amberber, Mengistu , , , Boas, Franz  Cardey, Sylviane 
 Bogaards, Paul  Carlisle, Robert , 
Ammon, Ulrich , , ,  Boland, Annerieke  Carr, Philip 
Anderson, Stephen R.  Bolinger, Dwight , , , Carr, Raymond 
Andreeva, E. G.  , , ,  Castañeda Castro,
Archibald, John  Bolkestein, A. Machtelt ,  Alejandro , 
Ariel, Mira  Borin, Lars  Celle, Agnès 
Arnold, Gordon F. –, , Borkin, Ann , , , , Chafcouloff, Michael 
– , , , , , ,  Chafe, Wallace , , –, ,
Artola, Miguel ,  Botley, Simon  , , , , 
Astington, Eric ,  Bowen, J. Donald , ,  Chesterman, Andrew , 
Atkins, B. T. Susan ,  Bradford, William  Chomsky, Noam , , 
Atlas, Jay D.  Brazil, David , ,  Cid Uribe, Miriam E. 
Brdar, Mario ,  Clancy, Patricia M. , , 
B Brdar-Szabó, Rita  Clark, Herb H. 
Bailey, Richard W. ,  Brennan, Susan E.  Clarke, David D. 
Baker, Mona ,  Bresnan, Joan , ,  Clyne, Michael G. 
Bakker, Dik , , ,  Brown, Gillian ,  Colleman, Timothy , 
Bald, Wolf-Dietrich  Brown, Penelope  Collins, Daniel E. , 
Ball, Catherine N.  Brown, Roger  Collins, Peter C. , , ,
Ballard, Michel  Brucart, José María ,  , , , –
Bally, Charles  Burke, Peter  Coltheart, Max , 
 Scholars index
Comrie, Bernard , , , , Dirven, René , , , ,  Frawley, William , , ,
 Di Tulio, Ángela , ,  
Connor, Ulla  Dixon, Robert M. W.  Frege, Gottlob 
Contreras, Heles , ,  Dobrovolskij, Dimitrij  Fretheim, Thorstein 
Corfield, Penelope  Doval-Suárez, Susana , ,  Friedman, Marilyn W. 
Corness, Patrick  Downing, Angela , , ,  Fries, Charles C. , 
Cornish, Francis , ,  Downing, Pamela A.  Fries, Peter H. –, , 
Cote, Sharon ,  Drubig, H. Bernhard  Fries, Udo 
Coulson, Seana ,  DuBois, John W. ,  Fry, Dennis 
Couper-Kuhlen, Elizabeth , DuFon, Margaret A. 
 Duszak, Anna 
G
Cowie, Anthony P. 
García, Erica 
Croft, William –, ,  E García Lecumberri, María
Crossick, Geoffrey  Ebeling, Jarle , ,  Luisa , 
Crowley, Tony  Eckert, Miriam  García Velasco, Daniel 
Cruse, D. Alan  Eckman, Fred  Gardner, Richard C. 
Crystal, David , , , , Eeg-Olofsson, Mats , ,  Garrido, Juan Manuel , ,
 Eliasson, Stig ,  
Cuenca, Heliodora  Engdahl, Elisabet ,  Garrudo Carabias, Francisco ,
Cuetos, Fernando  Enríquez, Emilia  , 
Culpeper, Jonathan  Erman, Britt ,  Gass, Susan 
Cummings, Michael , ,  Estébanez, Salvador  Gellerstam, Martin , 
Geukens, Steven 
D F Gibbs, Raymond W. Jr. 
Daneš, František ,  Faber, Pamela B. –, ,  Gil, David 
Danielsson, Pernilla  Fabricius-Hansen, Catherine , Gili Gaya, Samuel 
Darnell, Michael   Gilliom, Laura A. 
da Silva, Augusto Soares  Fant, Lars  Gimson, Alfred C. 
Davidse, Kristin , , , , Fareh, Shehdeh  Givón, Talmy , , , , , ,
,  Fauconnier, Gilles –,  , , , , , , ,
Davidson, Brad  Faure, George  , , , 
Dawes, Elizabeth  Fawcett, Robin P. ,  Goddard, Cliff , 
Deacon, Terence  Feilke, Helmuth  Goldberg, Adele E. , , , ,
DeCesaris, Janet  Fente, Rafael  
Declerck, Renaat  Fernández, Ana  Gómez González, María de Los
Defrancq, Bart  Fernández Domínguez, Joaquín Ángeles , , , , , ,
de Groot, Annette M. B.  José  , , –, , , ,
Delahunty, Gerald P. , , Fernández Leborans, María , 
,  Jesús , , , , , Gonçalves, Miguel 
DeLancey, Scott , ,  ,  Gonzálvez García, Francisco ,
Delattre, Pierre ,  Fichtner, Edward G.  , 
Delin, Judy , , , ,  Figueras, Carolina  Goossens, Louis , , 
Dellinger, Brett  Filipović, Rudolf ,  Gordon, Peter C. , , 
de Vega, Manuel  Fillmore, Charles J. , , , Grady, Joseph , 
Devos, Filip  , , ,  Granger, Sylviane , , ,
Devriendt, Betty  Firbas, Jan  –
Di Eugenio, Barbara , , Firth, John R. ,  Green, Georgia M. 
, ,  Fisiak, Jacek , , ,  Greenbaum, Sidney 
Dik, Simon C. , , , , , Flege, James , , , , Greenfield, Peter 
, , , , , , , ,  Grimm, Wilhelm , , 
 Fowler, Roger  Grosz, Barbara J. –, ,
D’Introno, Francesco ,  Fragoso, Jesús  
Di Pietro, Robert J. , –,  Francis, Gill ,  Guerrero Medina, Pilar , , 
Scholars index 
Guillemin-Flescher, Hodge, Robert  Kastovsky, Dieter 

Jacqueline  Hoeks, Jimmy  Kay, Paul , , 
Gundel, Jeanette K. , , , Hoequist, Charles ,  Kefer, Michel 
, , , ,  Hoey, Michael ,  Keizer, M. Evelien , 
Gurney, John , , ,  Hofland, Knut ,  Kemmer, Suzanne , , 
Gutiérrez (Díez), Francisco , Hopper, Paul J. , , , , Kempen, Gerard 
, , , , , ,  –, –, – Kiefer, Ferenc 
Gutiérrez Ordóñez, Salvador , Horie, Kaoru  Kim, Eunyi 
, ,  Hornby, Peter A.  King, Philip 
Gvozdanovic, Jadranka  Houghton, Diane ,  Kittay, Eva F. 
House, Juliane  Knipf, Elisabeth , 
Huang, Guowen  Koerner, E. F. Conrad 
H
Huddleston, Rodney D. , Komlósi, László I. , , ,
Hadic Zabala, Loreley , ,
–, , , , , , , 

 Kraig, Oliver 
Haegeman, Liliane 
Hudson-D’Zmura, Susan  Kreidler, Charles 
Hahn, Udo , , 
Hundt, Marianne  Kress, Gunther 
Haiman, John , , –
Hunt, Kellogg W.  Kreuz, Heinz J. 
Halliday, Michael A. K. , –,
Hurewitz, Felicia  Krifka, Manfred , , 
–, , , , , ,
Kroll, Judith 
, , , , , ,
I Kroon, Caroline 
, , , , 
Iida, Masayo ,  Krzeszowski, Tomasz P. –,
Halvorsen, Per-Kristian , 
Ilson, Robert F.  –
Hankamer, Jorge 
Iverson, Paul  Kuhl, Patricia 
Hannay, Mike , 
Ivir, Vladimir  Kühlwein, Wolfgang , 
Hansen, Klaus 
Kuno, Susumu , , 
Harbusch, Karin 
Kvavik, K. H. 
Harries-Delisle, Helga  J
Jackendoff, Ray , , , , Kyto, Merja 
Harris, Catherine L. , –
Harrison, Claire  , –, , 
Hartmann, Reinhard R. K. ,  Jacobsen, Wesley M. ,  L
Hasan, Ruqaiya , ,  Jaeggli, Osvaldo A.  Lado, Robert , , , 
Hasselgård, Hilde  Jager, Sake  Lakoff, George , 
Haugen, Einar  Jakobson, Roman ,  Lambert, Wallace 
Haugland, Kari E.  James, Carl , , , ,  Lambrecht, Knud , , ,
Havelka, J.  James, William  , , , 
Haviland, Susan E.  Janicki, Karol ,  Lamiroy, Béatrice 
Hawkins, Bruce  Jassem, Wiktor ,  Langacker, Ronald W. , , ,
Hawkins, John A. ,  Jaszczolt, Katarzyna M. , , , 
Hawkins, Roger  ,  LaPolla, Randy J. , , , 
Hedberg, Nancy , , , Jespersen, Otto ,  Laufer, Batia 
,  Johansson, Mats ,  Laviosa, Sara 
Hellinger, Marlis , , ,  Johansson, Stig , , –, , Leather, Jonathan 
Hengeveld, Kees , , , ,  Lehrer, Adrienne 
–,  Johns, Timothy  Lehtonen, Jaako 
Herder, Johann G.  Joseph, John E.  Leith, Dick , 
Hernanz, Maria Lluïsa ,  Joshi, Aravind , ,  Lerot, Jacques –
Hickey, Leo  Levelt, Willem J. M. 
Hidalgo Navarro, Antonio  K Levenston, Edward A. 
Higgins, Francis R. ,  Kalisz, Roman ,  Levin, Beth , , , , , 
Hill, David R.  Kamen, Henry  Levinson, Stephen C. , ,
Hill, Sam  Kameyama, Megumi ,  , 
Hirst, Daniel  Karttunen, Lauri  Lewandowska-Tomaszczyk,
Hodapp, Marion  Kasper, Gabriele  Barbara 
 Scholars index
Liberman, Mark  Milton, James L.  Payne, Jonathan , 

Liebe-Harkort, Marie Louise , Miltsakaki, Elena ,  Pérez Hernández, Chantal , 
 Moder, Carol M.  Pérez Quintero, María Jesús 
Littlewood, William  Monroy (Casas), Rafael , , Perkins, Michael R. 
Llisterri, Joachim  , ,  Perzanowski, Dennos , ,
Llorens, Vicent  Moon, Rosamund ,  , 
Locke, Philip  Morciniec, Norbert ,  Petch-Tyson, Stephanie –
Løken, Berit ,  Moreno Cabrera, Juan Peters, Pam 
Luce, Paul A.  Carlos , , , , , Philips, Susan U. 
Luján, Marta  , , –, , , , Pike, Kenneth 
 Pisoni, David B. 
Moulton, William G.  Poenack, Elmer 
M
Mufwene, Salikoko S. ,  Poesio, Massimo , 
Mackenzie, J. Lachlan , –, ,
Mukattash, Lewis  Pollard, Carl J. 
, , , , , , , 
Müller, P. Herman  Prasad, Rashmi 
Mair, Christian , 
Müller, Viviane  Prince, Ellen F. , , , ,
Mairal (Usón), Ricardo , 
Müller-Gotama, Franz  , , , 
Major, Roy , 
Pullum, Geoffrey K. , , 
Makino, Seiichi , 
Malinowski, Bruno  N
Maniez, François  Nariyama, Shigeko  Q
Markkanen, Raija  Navarro Tomás, Tomás , , Quirk, Randolph , , ,
Markus, Manfred  , ,  , 
Martí, Antònia  Nelson, Gerald 
Martin, James R. , , –, Nerlich, Brigitte 
R
, , ,  Newman, Michael 
Radden, Günter , 
Martin, John W.  Newmeyer, Frederick J. 
Rauh, Gisa 
Martínez, José Antonio  Nickel, Gerhard , 
Ravelli, Louise 
Martínez Caro, Elena , , Nikitina, Tatiana , , 
Real Academia Española ,
,  Noël, Dirk 
, , , 
Martínez Vázquez, Noh, Eun Ju , 
Reddy, Michael J. , 
Montserrat , , , , , Norrick, Neal R. 
Rein, Kurt , , 
, , ,  Nunberg, Geoffrey 
Renouf, Antoinette –
Martínez Vázquez, Rafael , Nuyts, Jan , 
Rice, Sally 
, 
Ridings, Daniel 
Martinovic-Zic, Aida 
O Riegelhaupt, Florence 
Marton, Waldemar 
Oakley, Todd ,  Rijkhoff, Jan 
Masterson, Jackie 
Oberlander, John , , , Riley, Philip 
Mathesius, Vilém 
 Ringbom, Håkan , , 
Matthiessen, Christian M. I.
O’Connor, Joseph D. , Ríos, Antonio 
M. , –, , , , 
–, , – Risselada, Rodie 
Maynard, Senko K. –, ,
O’Connor, Mary C.  Roach, Peter , , 

Odlin, Terence  Ron, Pilar 
McCarthy, Michael 
Oksefjell, Signe  Rosenzweig, Mark 
McEnery, Anthony 
Oleksy, Wieslaw , ,  Rudolf, Elisabeth 
McNaught, John 
Olhstain, Elite  Rudzka-Ostyn, Brygida , ,
Meara, Paul M. , –, ,
Olsen, Carl  

Olsen, Carol Lee  Rusiecki, Ian 
Mel’ˇcuk, Igor 
Ortiz Lira, Héctor ,  Russ, Charles V. J. 
Melia, Patrick J. 
Mendikoetxea, Amaya 
Merkel, Magnus  P S
Miller, George A.  Paillard, Michel  Sag, Ivan 
Miller, Jim ,  Palmer, Frank R. ,  Sågvall Hein, Anna 
Scholars index 
Sánchez, Jesús , ,  T Viberg, Åke , , 

Sajavaara, Kari –, ,  Taboada, Maite , , , , Vorlat, Emma , , 
Salkie, Raphael , ,   Vossen, Piek 
Santiago, María Luisa  Taeldeman, Johan 
Scheibman, Joanne ,  Takashi, Kyoko , , , 
Talmy, Leonard , , ,  W
Schieffelin, Bambi 
Tanenhaus, Michael K.  Wagner, Karl-Heinz 
Schiffrin, Deborah 
Tannen, Deborah , ,  Walker, Marilyn , , 
Schlesinger, Izchak M. 
Schmitt, Norbert  Tarasova, Yelena  Ward, Gregory , , , 
Schneider, Peter  Taylor, John R.  Ward, Michael T. 
Scholes, Robert  Taylor, Talbot J.  Wardhaugh, Ronald 
Schubiger, Maria , ,  Teich, Elke ,  Warren, Beatrice , 
Selinker, Larry ,  Tejada Caller, Paloma , ,  Wasow, Thomas 
Seto, Ken-ichi  Tench, Paul  Weigand, Edda 
Sidner, Candace L. , ,  ter Meulen, Alice ,  Weinert, Regina , 
Siewierska, Anna , , , ,  Teubert, Wolfgang ,  Weinreich, Uriel 
Sifianou, María  Thompson, Sandra A. , , , , Weinstein, Scott , , 
Silverstein, Michael ,  , , –, –, Wekker, Herman 
Simon-Vandenbergen, –,  Wells, John 
Annemarie ,  Todd, Zazie  Whorf, Benjamin Lee 
Sinclair, John M. , , , , , Tognini-Bonelli, Elena , , , Wierzbicka, Anna , , 
, , –, –, , , ,  Wiktorsson, Maria 
 Tomasello, Michael  Wilkerson, Douglas , , ,
Slobin, Dan  Tops, Guy A.  
Smith, Larry E.  Torres, Amadeu  Wilks, Clarissa 
Solé, María José  Towell, Richard  Willems, Dominique 
Solfjeld, Kåre  Trager, George L. ,  Williams, Raymond 
Sornicola, Rosanna , , , Tsunoda, Tasaku , ,  Wilson, Andrew 
, ,  Tucker, Gordon  Wirth, Jessica 
Sridhar, Shikaripur N.  Turan, Ümit Deniz ,  Witten, Ian H. 
Steffens, Petra  Turner, Ken ,  Wodak, Ruth 
Steiner, Erich  Turner, Mark –, ,  Wongkhomthong, Preeya 
Stent, Amanda  Twaddell, W. Freeman  Wray, Alison , , , , ,
Stewart, Miranda  , –
Stilwell Peccei, Jean  V
Stockwell, Robert P. , ,  Vallduví, Enric ,  Y
Strangert, Eva  van Benthem, Johan ,  Yallop, Colin 
Strube, Michael , – Van Buren, Paul ,  Yngve, Victor H. 
Stubbs, Michael , , , van der Auwera, Johan  Young, Lynne , 
,  van Dijk, Teun  Yule, George 
Suñer, Margarita , , ,  Van Valin, Robert D. Jr. , , ,
Svartvik, Jan  , 
Sveen, Andreas  Vázquez, Gloria  Z
Svensson, Mikael ,  Vázquez Rozas, Victoria  Zacharski, Ron 
Swales, John M.  Verhagen, Arie  Zamora (Vicente), Alonso 
Swiggers, Pierre  Veronis, Jean  Zampolli, Antonio , 
Szwedek, Alexander  Verschueren, Jef ,  Zwicky, Arnold 
Subject index
A native American languages backward-looking center (Cb)

academic word list  ,  see center
accessibility ,  anaphoric reference see bilingual/multilingual
accessible topic see topic reference grammars see grammar
accessibility/availability of animacy/inanimacy , , , bottom-up 
vocabulary items , , , , – boundedness , , –,
, ,  antecedent , , , , , , 
scale  , ,  bourgeois revolution , 
acquisition see language aphasia , , ,  brain , , , 
acquisition Applied Contrastive Studies see
act contrastive
communicative act see arbitrariness –, , , , , C
communication , ,  Cambridge First Certificate
discourse act , , –, arbitrary pronoun ,  Examination 
 archilexeme ,  cardinal Transitivity/
cardinality see transitivity
illocutionary act see argument structure , , ,
illocution –, , , , , , , cataphoric reference see
speech act see speech , , , ,  reference
activation , , , , , articulation , –,  categorization (as cognitive
 process) , , , , , ,
aspect , , , , , ,
addition reference  , ,  , , , , , , 
Adjunct , , , –, causative construction 
perfective/imperfective
, , ,  –, ,  Cb see center
Comment  progressive , ,  center
Conjunctive , ,  assertion , –, , , backward-looking center
Mood see mood  (Cb) , , , ,
adverb(ial) , , , , , , , , , ,
assimilation 
, , , , , , , –
atelic see telicity
 attitude/attitudinal meaning ,
empty backward-looking
affect  center , , ,
, , , , , , , ,
affectedness , , , 
–
Object affectedness see forward-looking centre (Cf)
attribute (as function of clause
Object , , , 
element) , , 
agent(ivity) , , , , , preferred center (Cp) , ,
autonomous functionalism see
, – passim , 
function(al(ism/ist))
allative  Centering Theory , –
Autosegmental Phonology 
ambiguity , , , , , Cf see center
, , , , , ,  channel of communication ,
America(n) , , , , , B 
, , , , , backchannel (signal) –, cheap transition see transition
– passim, ,  , ,  clause combining , 
American functionalism see backgrounding (in discourse) clause-complex , , , ,
function(al(ism/ist)) , , , – passim 
 Subject index
cleft construction , , – framework see multi-word conceptualization , , , ,
cliché , , , , , , sequence: discontinuous , , , , , , ,
,  level of lexical structure , , , – passim,
cline ,  ,  , , 
clitic , , , , , combination (in relation to and construal of experience
–, ,  grammar and thought) , 
direct clitic  – concept structuring/
doubling  Comment Adjunct see Adjunct organization , , ,
indirect clitic  communication , , , , , , , , 
closing , , ,  , , –, , , , , conceptual cohesiveness ,
codability  –, , , , , 
code (in communicative process) , , , , , , conceptual component see
 , , –, ,  component
cognition/cognitive approach , communicative act –, conceptual deviance from
, , ,  , , ,  prototype , 
in relation to cleft communicative component conceptual integration ,
constructions , , , (of Functional Discourse , , , –,
,  Grammar) see component , , 
in relation to contrastive communicative construction level of lexical structure
linguistics , , ,  – , 
in relation to emotivity communicative equivalence conceptual mapping 
–,   conduit metaphor see metaphor
in relation to entrenchment communicative verb see verb conflation , , , 
, , , , , , comparable corpus see corpus conjunction , , , , ,
, , , , ,  Comparative Linguistics (as , , , 
in relation to formulaic term)  Conjunctive Adjunct see
sequences  competing Adjunct
in relation to functionalism motivations , , , ,  consciousness , , , , ,
–, – referent  –
in relation to grammar and complement(ation) , , , constituent , , , , ,
thought , – , , , , , , , , , , –, ,
in relation to Incremental , ,  , , , , , , ,
Functional Grammar Object Complement  , , , , 
–,  Predicative Complement  constraint , , , , , ,
in relation to Place of prepositional complement , , , , , , ,
Negotiation  ,  
in relation to transitivity , Subject Complement  Construction Grammar/
 component (in Incremental constructionist approach
in relation to verbs of
Functional Grammar/ , , , , – passim,
communication ,
Functional Discourse – passim,  see
–
Grammar) also cleft construction;
coherence (in discourse) , , cognitive component  communication
, , , , ,  communicative component (communicative
cohesion , , , , ,  construction); dative
,  conceptual component  construction/alternation;
cohesiveness in relation to contextual component  ditransitive construction;
entrenchment ,  grammatical component identifying clause/
COLLATE , ,  – construction; resultative
colligation  output component  construction
collocation(al) –, , compositionality see lexical constructions in relation to
–, , , , , structure entrenchment –
, , , , , , compounding , ,  passim
, ,  Computational Linguistics , context , , , , , , , ,
force –,  ,  –, , , , , , ,
Subject index 
, , , , , , passim, , , , , discharge, verbs of see verb
–, , , , , , , – passim, discontinuous multi-word
, , , , , , , , –, , , sequence see multi-word
, , , –, , , , , , , , sequence
, , , , , , , –, , ,  discourse
, , , , ,  comparable corpus  act see act
contextual component see multilingual corpus ,  analysis , , , , , ,
component parallel corpus ,  , , 
Continuative  translation corpus ,  anaphora see reference
continue (as transition type) co-selection  discourse/text segment ,
, , , , – Cp see center , , , –, ,
CONTRAGRAM , ,  creaky voice see voice , , , , ,
contrast(ive) critical linguistic history  –, 
Applied Contrastive Studies cross-linguistic perspective see institutional discourse ,
,  contrastive , –
Contrastive Analysis (CA)  culture , –, –, , ,  move , –, , , 
Contrastive Discourse cross-cultural perspective , topic see topic
Analysis (CDA) ,  , –,  Disjunctive (as element of
Contrastive Linguistics , cultural history of English structure) 
–, ,  – dislocation , 
Contrastive Phonetics/ sociocultural factors , , , Disney, Walt , , 
Phonology (CP)   displaced mode 
Contrastive Pragmatics (CP) dissimilarity see similarity/
,  D dissimilarity between L1 and
Contrastive Sociolinguistics dative construction/alternation L2
(CSL) ,  , , , ,  ditransitive construction , ,
Contrastive Studies (CS)  declarative 
perception of isosyllabism mood see mood dyslexia 
 vs. procedural modelling ,
perception of rhythm  , , ,  E
perception of timing , , default processing , , , education , , , , , ,
, – , – –
convention , , , , definite(ness) , , , , elaboration (in Systemic
–, –, , , , , –, , , , Linguistics) , 
–, , , , –, ,  ellipsis , , , 
, , , , , , , deixis/deictic , , , , , emergent, grammar as , , 
, ,   emotion , , , , , ,
conversation , , , –, demonstrative  , –, , ,  see
, , , , , , , pronoun , ,  also emotivity
, , , , , , detachment  emotivity –
 developmental error see error empathy , , , –,
conversational implicature see diachrony see language: change/ , , 
implication diachrony English see also language index
conversational narrative  dialectal variation , , , as model of innovation ,
telephone conversation , ,  , , , 
– passim dialogue , , , , , perceptions of English and
cooccurrence see collocation; , , , – Englishness , –
colligation dictum , , , , , , , foreign-language users ,
coordination  , , , –, ,  
copula , , –, ,  direct literary canon 
coreference see reference clitic see clitic enhancement (in Systemic
corpus , , , , , , , , enunciation  Linguistics) 
, –, , , – Object see Object entity, highest ranked , ,
passim, , , – speech see speech –
 Subject index
entrenchment , – FG see Functional Grammar free

entropy ,  field (in Systemic Linguistics) expression , , , ,
epistemology , , , , ,  , , , , , ,  
error figure of speech  word order , 
analysis  finite frequency
developmental error , , clause , , , , , frequency-driven lexical
, , , ,  ,  structure see lexical
hypercorrection error , Finite element of clause structure
– structure  fundamental frequency ,
interference error , , verb see verb , 
, ,  first-mention  word frequency in vocabulary
predicting learners’ errors , first person pronoun , , studies , –,
,  , , , , , ,  –
timing error , – fixed expression , , , full pronoun , , 
transfer error , , , , , – passim see also function
, ,  formulaic language/sequence; grammatical function see
esphoric reference see reference holistic vs. analytical grammar
European functionalism see processing; holophrase; in Functional Discourse
function(al(ism/ist)) idiom; multi-word sequence; Grammar, Incremental
evidentiality  prefab; semi-fixed expression Functional Grammar
exclamation , , ,  focus/focal , , , , , , pragmatic function , 
exclusiveness (in relation to cleft , , , –, , , semantic function 
constructions) , ,  , , –, , , syntactic function , 
exhaustiveness , , –, , , , , , , (meta)function in Systemic
 , , , , – Functional Grammar
existential passim experiential metafunction
existentially-quantified Element-in-Focus (EIF) , –, , , ,
variable 
– passim 
particle 
of attention , – interpersonal
presupposition see
of consciousness  metafunction , ,
presupposition foot (as rhythmic unit) , , , , , ,
exophoric reference see reference – passim 
expensive transition see football commentary, language textual metafunction
transition of , – –, , 
Experiencer , ,  foregrounding , , , , , functional(ism/ist))
experiential – passim American functionalism , ,
(meta)function see function formalist approach to language , , 
Theme see theme –, , ,  autonomous functionalism 
exposition (as genre) –, formulaic language/sequence , European functionalism , 
,  , , , , , –, external functionalism 
extended unit of meaning , –, , ,  see also extreme functionalism 
–, – holistic vs. analytical functional-cognitive space ,
extension (in Systemic processing; fixed expression; 
Linguistics)  holophrase; idiom; functional equivalence 
external functionalism see multi-word sequence; prefab; integrative functionalism 
function(al(ism/ist)) semi-fixed expression mixed formal/functional
extreme functionalism see formulation (in Incremental functionalism 
function(al(ism/ist)) Grammar/Functional typological functionalism ,
Discourse Grammar) , , 
, , ,  West Coast functionalism
F forward-looking center (Cf) see –, , 
false start –, ,  center Functional Discourse Grammar
FDG see Functional Discourse FPG see Functional Procedural (FDG) , , , , , , ,
Grammar Grammar –
Subject index 
Functional Grammar (FG) , universal grammar , ,  illocutionary act , 
–, –, , , , –, Grimm , ,  indirect illocution , 
 image , , 
Functional Procedural Grammar image codifiers , 
H
(FPG) ,  of English and Englishness
head
Functional Sentence Perspective
clause , , ,  , –, , –
 imaginary word see word
functional head of complex
functionally complete unit of
NP  immanence 
meaning see extended unit of
noun  immediate mode 
meaning imperative mood see mood
of tone unit , –
fundamental frequency see imperfective see aspect
passim
frequency impersonal Subject see Subject
syntactic head of complex NP
 implication
G Head Driven Phrase Structure implicational hierarchy 
gender ,  see also sex Grammar  implicature , , ,
as grammatical category , hedge , , ,  –, , 
 hesitation ,  conversational implicature
cross-gender perspective Heteromorphic Distributed , 
– Lexicon  implicitness , , , ,
generative grammar see historical linguistics ,  , –
grammar holistic vs. analytical processing Incremental Functional
generic reference see reference , – Grammar (IFG) , , , ,
genitive  holophrase , , , , , –
genre , , , , –, , –,  indefinite(ness)
, , , , ,  homophoric reference see article , 
given information/givenness , reference demonstrative 
, , , , , , , , hypercorrection error see error in noun phrase/nominal
, , , –, ,  Hyper-new see new group , , 
global information/newness Object 
discourse structure ,  Hyper-theme see theme indeterminacy  see also
discourse topic see topic non-discreteness; prototype
Theme/Rheme pattern see I indirect
theme iconicity , , , , , clitic see clitic
grammar  enunciation 
bilingual/multilingual ictus , , , –, , illocution see illocution
grammar   Object see Object
generative grammar , , idea , , , –, – speech see speech
,  see also formalist identifiability , , , , , individuation of Object see
approach to language  Object
grammatical complexity , identifying clause/construction inference , , , –,
, , ,  , , , –, , , –
grammatical component see , , ,  load 
component identity , , , , , , presumptive inferencing ,
grammatical function , , ,  
, , ,  see also ideology , , –,  innateness 
syntax/syntactic: function idiom(aticity) , , , , inner speech see speech
grammatical meaning , , , , , –, instantial reference see reference
–,  , , , , ,  institutional discourse see
grammatical structure in idiom principle , –, discourse
relation to thought , ,  integrative functionalism see
–, ,  IFG see Incremental Functional function(al(ism/ist))
grammaticalization , , , Grammar intensity (of speech sounds) ,
– illocution , , , ,  
 Subject index
intention , , , , , , universals , , , , , , lexicalization (as historical
,  , , , ,  see process) , , –, , ,
interference , ,  also semantic universals; 
error see error grammar: universal linearity/linearization in
interlanguage , , , , , grammar linguistic structure , ,
, ,  laughter  , –, 
interpersonal layering in linguistic structure Linguistic Data Consortium ,
function/metafunction see ,  
function
left-to-right ordering  listener , , , , , , ,
level see level
level, in Functional Discourse , , , , , , , ,
meaning of cleft
Grammar/Incremental
–, , , –, 
constructions , – Functional Grammar
local
interpreting , ,  focus structure , , 
interpersonal level , ,
interrogative mood see mood
–, , , , , topic see topic
intonation , –, , , locative 

, , , , , , logogenesis 
phonological level ,
,  see also tone
–, , ,  logographic language 
English-Spanish in contrast
representational level , loudness , , , 
, –
–, –
perceptual approach –
intransitive see transitivity structural level , –,
– M
involvement (in relation to machine translation see
emotivity) – lexical
translation
isochrony , , , – accessibility/availability see
Macro-new see new
isosyllabicity , , – accessibility
information/newness
iteration , , , – bundle , , 
macrorole 
licensing , , –,
Macro-theme see theme

J mapping , , , , 
Japan(ese) , , – neighbourhood 
marked Theme see theme
passim see also language representation , 
meaning integration 
index rule –, 
memory , , , , , ,
jargon  semantics , ,  , , , 
juxtaposition (as step in structure , –, , message , , , –, , ,
contrastive studies)  , , , – , , , , , ,
compositionality/non –, , , , , 
K compositionality , metalanguage , , , , ,
kinesis , ,  , –, , , , , , 
–, – metaphor , –, , –,
computational efficiency , –, , , ,
L
 , , , , –,
language
acquisition , , , , ,
frequency-driven structure , –, , , ,
, , ,  , , 
, , –, ,
rule-driven structure , conduit metaphor , , 
, , , 
, ,  metaphorisms 
acquisitional adequacy 
change/diachrony , , , , unit  metaphorization , , ,
, , , , –, , multi-word unit see 
,  multi-word sequence transfer metaphor 
integration  variable-sized unit , metarepresentation 
learning see second language – method of development ,
teaching/ learning Lexical-Conceptual Structure , – passim
planning   metonymy , – passim
teaching see second language Lexical Contrastive Linguistics instrument for action
teaching/ learning (LCL)  metonymy , , 
Subject index 
manner for action N object of conveyed reaction ,

metonymy ,  narrative/narrator , , , , , 
means for action , , , –, , ontogenesis , , , 
metonymy ,  , , , , , open processing , , ,
part for whole metonymy – passim, , , , 
 , , – opening , , , , ,
sound for manner of
narrative voice , , , , 
communication
 operator , , , 
metonymy  nation-imagining myths , orientation (in relation to
sound for speech , , ,  grammar and thought) ,
metonymy , , –, Natural Semantic Metalanguage , , , –, –, 
  orthography
Metrical Phonology see negative transfer see transfer orthographic access to lexicon
phonology new information/newness , , 
modal verb see modality –, , , –, , otherness , , –
modality , , , , , , , , , , , , output component see
, , , , , , , –, , , , , component
,  , , , ,  overlapping
modulation  Hyper-new  in spoken discourse , 
mood  Macro-new  of multi-word sequences ,
declarative mood , ,
nominalization , , , , 
, , 

,  polysemy 
imperative mood , , 

nonce expressions , , 
interrogative mood , 
non-discreteness , ,  see
wh- 
P
also indeterminacy, prototype
yes/no 
paralanguage , , 
non-finite verb see verb
Mood Adjunct 
parallel corpus see corpus
morphology , , , , , non-subcategorized Object see
Object parameter setting 
, , , , , ,
N-Rheme see rheme parametrization , , 
–, , ,  see
nuclear path (in Lexical-Conceptual
also morphosyntax
accent  Structures) , , 
morphological variant 
movement  Patient , , , , , 
morphosyntax , , , ,
pausing , , , , 
, , , –,  nucleus/nuclear element/
nuclear tone, of tone perfective see aspect
move see discourse
multilingual corpus see corpus group , , , performance analysis 
multilingualism  – passim phonology/-ical , , –,
multiple Theme see theme , , –, ,
multi-word sequence , , , –, , , , ,
, , , , , , O , , 
 see also fixed Object , , , , , , , access to lexicon , 
expression; formulaic , , , , , , change , 
language/sequence; , , , , , , learning , , , 
holistic vs. analytical –,  level in Incremental
processing; holophrase; affectedness , –, , Functional Grammar/
idiom; prefab; semi-fixed , , –,  Functional Discourse
expression Direct Object , , –, Grammar see level
continuous , –, , , , , , ,  Metrical Phonology 
, ,  Indirect Object –, , phrasal expression of content ,
discontinuous , , , , , ,  , –
–, ,  individuation , –, phraseme , , 
mutual , , –,  phraseology see formulaic
correspondence  non-subcategorized Object language/sequence
information (MI)   phylogenesis , , 
 Subject index
pitch , , , , , , , existential presupposition , R
– passim , ,  reaction object see object of
Place of Negotiation theory , pre-tonic segment  conveyed reaction
 privileged syntactic argument  realization (of Theme and point
point (of a text) , , , , procedural modelling of of text) –, , 
, , , , , , , language see declarative vs. receiver , , , 
– procedural modelling of receptive vocabulary see
politeness , , , , , language vocabulary
, , , –,  pro-dictum , , , , , , receptor 
polysemy ,   recipient , , , –, , ,
possessive , , , , , pro-drop language , , , –
,  ,  recitable entity 
determiner , ,  productive vocabulary see reference
pronoun  vocabulary anaphoric reference , ,
pragmateme , ,  progressive aspect see aspect , –, –
pragmatic(s) , , –, , , prominence , , , , , passim
–, , , , –, , , , , , , , , discourse anaphora ,
, , , , , , , , , , , , , 
, , –, –, –,  see also salience zero anaphora , ,
, , , , , pronominalization , ,  , , , , ,
– Pronoun Rule  
pragmatic function see proper name/noun , , , cataphoric reference , ,
function , , , ,  
Prague School , , ,  proposition , , , , , coreference , , , 
predicated Theme see theme , , , , , , , esphoric reference 
predication (act of, in –, , , , , ,
exophoric reference , ,
Incremental Functional , , , 

Grammar/Functional prosody/-ic , , , , ,
generic reference , 
Discourse Grammar) , , , , , , 
homophoric reference 
, ,  learning 
instantial reference 
Predicator (in Systemic phrase , 
presuming reference ,
Linguistics)  semantic prosody , ,
–, 
predicting learners’ errors see –, , 

reference chain –
error unit , , , 

passim
prefab ,  see also prototype , , , , , ,
referent (in relation to
formulaic language/sequence; , , –, –, 
grammar and thought)
fixed expression; holistic vs. psycholinguistics , , , ,
, , 
analytical processing; , , –, , , ,
tracking , 
holophrase; idiom; , , 
multi-word sequence; referring expression , ,
psychological
, –
semi-fixed expression adequacy , , 
preferred center (Cp) see center register , 
verb see verb
pre-nuclear pattern , , , relational
punctual , , , 
, ,  process (in Systemic
prepositional phrase/ Linguistics) , 
complement/object , , Q verb see verb
, , , , , ,  Q_Lex , , , , ,  relative clause , , , ,
presentative , ,  question , , ,  , , , –
presuming reference see wh- , – relative pronoun , , 
reference yes/no , , – relevance , , 
presupposition , –, , quotation , , , , , , remiss , , , –, ,
, –, , , , , , , , , , , , 
, , , , ,  , , , ,  renewal of connection 
Subject index 
report second person pronoun , , , , , , , , , ,
of communicative acts , , , ,  , , , , –, ,
–, ,  secondary speech situation , –, , , , ,
reported speech see speech  –
reporter ,  segmentation (in relation to sound emission verb see verb
representational level see level grammar and thought)  space (in relation to grammar
resultative construction  selection (in relation to grammar and thought) , , , 
retain (as transition type) , and thought) , –, , Spanish see also language index
, , , ,  ,  Royal Academy , ,
rheme/Rheme , , – self-containedness ,  – passim
passim, , , , , semantic(s) texts (19th century) –
 diversity 
speaking verbs see verb
N-Rheme – passim function see function
speech
rhetoric(al function) , , , network , 
act , –, , , , ,
, , , , ,  preference , , –,
, , , , , 
rhetoric of legitimacy  , 
direct speech , , , 
rhetorical transfer  prosody see prosody/-ic
event , , 
rhythm , – passim semantico-syntactic
equivalence  indirect speech , 
temporal vs. non-temporal
view of rhythm 

structures , , , , , inner speech 
Role and Reference Grammar ,
, , , –, 
reported speech , , , 
–, –,  universals , 
speech/language
Romanticism  vs. grammar , 
comprehension system ,
rough shift (as transition type) vs. pragmatics –
, , , 
, , , , ,  semi-compositional lexical speech/language production
RRG see Role and Reference structures see lexical system , , , , ,
Grammar structure: , , , , , 
rule compositionality/non tempo see tempo
equivalence  compositionality staccato effect , 
Rule 1 (in Centering Theory) semi-fixed expression , , staging (of information) , ,
 , , , ,  
rule-driven lexical
sender ,  starting point (in relation to
structure see lexical:
sentence grammar and thought) ,
structure
sentential choice of , 
expression , –,  stress , , , – passim
Theme see theme
stressed pronoun 
S topic see topic
timing see time/timing
salience see also prominence sequential cohesiveness , 
discourse salience , , , structure
sex , , –, , ,
, , , , , ,  see also gender grammatical structure see
, ,  grammar
SFG see Systemic Functional
phonetic/phonological Grammar/Linguistics lexical structure see lexical:
salience  shadow meaning  structure
psychological salience  Shakespeare –, , semantic structure see
semantic salience  – semantic(s): structure
satellite (in Functional side sequence  structural level see level
Grammar)  silent thought see thought subact , , –, , 
scale of accessibility see similarity/dissimilarity between Subject –, , , , , ,
accessibility L1 and L2  , , , , , ,
second language teaching/ smooth shift (as transition , , –, , ,
learning , , , , , type) , , , ,  , , , , –,
–, –, –, Snow White , – passim –, –, , ,
– social factors/sociolinguistics , , , –, 
second order approximation  , , , , , , , , , , impersonal Subject , 
 Subject index
subjectivity , , , , , telephone conversation see token/type ratio  see also
, , , , , , conversation type/token ratio
, –, , ,  telicity , , –, , , tolerance , , , , 
subordination , , , , ,  tone (phonological) , , ,
–, , , , , , tempo , , ,  , – passim see also
, , , , ,  temporal intonation
substantive equivalence  orientation ,  tonic segment/syllable , ,
substitution (as a form of sequencing of acts ,  
reference in Systemic system of verbs see tense top-down organization , 
Linguistics)  view of rhythm see rhythm topic(al(ity)) , , , , ,
subtopic see topic tense , , –, –, , , , –, , , ,
switch-function  , , , ,  , , , , , ,
switch-reference  terms of reference , – , , – passim,
syllable , , , ,  tertium comparationis (TC) 
compression , ,  – accessible topic 
length/duration , , text type , , , , , , chain 
– passim , ,  see also genre, continuity , 
stressed vs. unstressed register hierarchy 
syllable see stress textual marker 
syllabic language (writing (meta)function see function progression 
system)  prominence see prominence subtopic , –, , 
syllabic structure  theme/Theme topic sentence 
timing see time/timing clause Theme , , ,  topical Theme see theme
symbolization (in relation to experiential Theme –, transcendence 
grammar and thought) , ,  transfer
, ,  global Theme/Rheme pattern communication as transfer of
synchronic approach , , ,  , ,  information , –
syntax/syntactic , –, , , , Hyper-Theme , , ,  during language learning/
, , , , , , , , in argument structure ,  translation , , ,
, , , –, , , in Functional Grammar  , 
, , , , , , in Systemic Functional error see error
, , , , , , Grammar , , –, negative transfer , 
, –, , –, , , , , ,  metaphor see metaphor
, , , , , Macro-Theme ,  verb see verb
–, , ,  see marked Theme , , , transition (in Centering Theory)
also morphosyntax , , , ,  – passim
complexity  multiple Theme ,  cheap transition 
function , , , , , predicated Theme  expensive transition 
 see also function (in sentence Theme , , , type see continue, retain,
Incremental Functional ,  rough shift, smooth
Grammar/Functional topical Theme , , , shift, zero-Cb transition
Discourse Grammar); , – transitivity , , –, , ,
grammar: grammatical unmarked Theme ,  ,  see also
function third person pronoun , , ditransitive construction
tree  ,  cardinal transitivity ,
synthesis  thought , – –
system equivalence  silent thought  high transitivity , –
Systemic Functional Grammar/ thinking for speaking ,  passim
Linguistics (SFG/SFL) , time/timing – intransitive , , 
–, , , , –, ,  error see error low transitivity , –,
pressure in production of , , , 
T discourse , , , translation , –, , , ,
Task Urgency  – , , , , , , ,
TC see tertium comparationis stress timing  , , , 
Subject index 
corpus see corpus finite verb , ,  narrative voice see narrative
equivalence , ,  instrument of passive voice , , 
machine translation ,  communication verb , voice quality 
translatability ,  , , , ,  creaky voice , 
T-score  instrumental verb , –, volitionality –, , ,
T-unit   , 
turn-taking , ,  linguistic action verb 
type/token ratio – see manner of motion verb 
also token/type ratio manner of speaking verb , W
typology –, , , , , , , –, , ,  West Coast functionalism see
, ,  non-finite verb  function(al(ism/ist))
typological adequacy ,  of gesture , , – wh-question see question
typological functionalism see of sound emission , , , word
functional(ism/ist) , –, ,  association tests , , ,
psychological verb  
relational verb  frequency, in vocabulary
U studies see frequency
saying verbs (verba dicendi)
umlaut  imaginary word , , 
, –, , –, ,
unbounded see boundedness order –, , , ,
, 
unconformity  , , 
speaking verb , , , ,
unitized
,  recognition –, –
(representations of)
verb of transfer , , ,  WordSmith Tools 
variable-sized lexical units
verba dicendi see saying verbs
–, , , , ,
V_Links , , , , ,
,  X
, , 
storage in mental lexicon  X_Lex , , , 
vocabulary , , , , ,
universal grammar see grammar
, , , 
unmarked Theme see theme
accessibility  Y
usage-based approaches , –,
frequency see frequency yes/no interrogative see mood:

network ,  interrogative
organization ,  yes/no question see question
V productive vocabulary , Yes/No test , , 
verb 
chitchat verb ,  receptive vocabulary 
cognitive verb ,  size – Z
communicative verb , test , , – zero
– vocative  anaphora see reference
communicative transfer verb voice realization , , , ,
– alternations  
discharge verb , ,  middle voice  zero-Cb transition 
In the Pragmatics & Beyond New Series the following titles have been published thus far or are
scheduled for publication:
143 BAKER, Carolyn D., Michael EMMISON and Alan FIRTH (eds.): Calling for Help. Language and social
interaction in telephone helplines. xvii, 346 pp. + index. Expected November 2005
142 SIDNELL, Jack: Talk and Practical Epistemology. The social life of knowledge in a Caribbean community.
xx, 259 pp. + index. Expected October 2005
141 ZHU, Yunxia: Written Communication across Cultures. A sociocognitive perspective on business genres.
xviii, 208 pp. + index. Expected November 2005
140 BUTLER, Christopher S., María de los Ángeles GÓMEZ-GONZÁLEZ and Susana M. DOVAL-SUÁREZ
(eds.): The Dynamics of Language Use. Functional and contrastive perspectives. 2005. xiv, 413 pp.
139 LAKOFF, Robin T. and Sachiko IDE (eds.): Broadening the Horizon of Linguistic Politeness. x, 335 pp. + index.
Expected October 2005
138 MÜLLER, Simone: Discourse Markers in Native and Non-native English Discourse. ix, 282 pp. + index.
Expected November 2005
137 MORITA, Emi: Negotiation of Contingent Talk. The Japanese interactional particles ne and sa. 2005. xvi, 240 pp.
136 SASSEN, Claudia: Linguistic Dimensions of Crisis Talk. Formalising structures in a controlled language. 2005.
ix, 230 pp.
135 ARCHER, Dawn: Questions and Answers in the English Courtroom (1640–1760). A sociopragmatic analysis.
2005. xiv, 374 pp.
134 SKAFFARI, Janne, Matti PEIKOLA, Ruth CARROLL, Risto HILTUNEN and Brita WÅRVIK (eds.): Opening
Windows on Texts and Discourses of the Past. 2005. x, 418 pp.
133 MARNETTE, Sophie: Speech and Thought Presentation in French. Concepts and strategies. 2005. xiv, 379 pp.
132 ONODERA, Noriko O.: Japanese Discourse Markers. Synchronic and diachronic discourse analysis. 2004.
xiv, 253 pp.
131 JANOSCHKA, Anja: Web Advertising. New forms of communication on the Internet. 2004. xiv, 230 pp.
130 HALMARI, Helena and Tuija VIRTANEN (eds.): Persuasion Across Genres. A linguistic approach. 2005.
x, 257 pp.
129 TABOADA, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and Spanish.
2004. xvii, 264 pp.
128 CORDELLA, Marisa: The Dynamic Consultation. A discourse analytical study of doctor–patient
communication. 2004. xvi, 254 pp.
127 BRISARD, Frank, Michael MEEUWIS and Bart VANDENABEELE (eds.): Seduction, Community, Speech. A
Festschrift for Herman Parret. 2004. vi, 202 pp.
126 WU, Yi’an: Spatial Demonstratives in English and Chinese. Text and Cognition. 2004. xviii, 236 pp.
125 LERNER, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 302 pp.
124 VINE, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004.
x, 278 pp.
123 MÁRQUEZ REITER, Rosina and María Elena PLACENCIA (eds.): Current Trends in the Pragmatics of
Spanish. 2004. xvi, 383 pp.
122 GONZÁLEZ, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004.
xvi, 410 pp.
121 FETZER, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp.
120 AIJMER, Karin and Anna-Brita STENSTRÖM (eds.): Discourse Patterns in Spoken and Written Corpora.
2004. viii, 279 pp.
119 HILTUNEN, Risto and Janne SKAFFARI (eds.): Discourse Perspectives on English. Medieval to modern. 2003.
viii, 243 pp.
118 CHENG, Winnie: Intercultural Conversation. 2003. xii, 279 pp.
117 WU, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004. xvi, 260 pp.
116 GRANT, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003.
viii, 330 pp.
115 KÄRKKÄINEN, Elise: Epistemic Stance in English Conversation. A description of its interactional functions,
with a focus on I think. 2003. xii, 213 pp.
114 KÜHNLEIN, Peter, Hannes RIESER and Henk ZEEVAT (eds.): Perspectives on Dialogue in the New
Millennium. 2003. xii, 400 pp.
113 PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. 2003.
xii, 285 pp.
112 LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp.
111 ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp.
110 ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of
Youth Identities. 2003. viii, 343 pp.

109 MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and
America. 2003. xiv, 228 pp.
108 BARRON, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study
abroad context. 2003. xviii, 403 pp.
107 TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems.
2003. viii, 446 pp.
106 BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person
pronouns. 2002. xiv, 344 pp.
105 BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp.
104 BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp.
103 FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets
conversational interaction. 2002. vi, 300 pp.
102 LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and
objects. 2002. xii, 252 pp.
101 LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in
conversational structure across languages and cultures. 2002. x, 295 pp.
100 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003.
viii, 496 pp.
99 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003.
xii, 388 pp.
98 DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002.
viii, 522 pp.
97 MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology
of pathos in Japanese discourse. 2002. xiv, 481 pp.
96 HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp.
95 FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002.
viii, 263 pp.
94 McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp.
93 BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and
masculinity in ethnography and discourse. 2002. xxiv, 357 pp.
92 GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp.
91 GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp.
90 KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A
Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp.
89 ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second
language contexts. 2001. xviii, 231 pp.
88 BAYRAKTAROĞLU, Arın and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of
Greek and Turkish. 2001. xiv, 439 pp.
87 MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp.
86 IFANTIDOU, Elly: Evidentials and Relevance. 2001. xii, 225 pp.
85 COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001.
xx, 384 pp.
84 ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the
language of adolescents. 2001. ix, 352 pp.
83 MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and
apologies. 2000. xviii, 225 pp.
82 KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp.
81 DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of
intercultural situations. 2001. xvi, 341 pp.
80 UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000.
xiv, 286 pp.
79 ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000.
viii, 273 pp.
A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com

Butler P 36

Uploaded by

Copyright:

Available Formats

Butler P 36

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Butler P 36

Uploaded by

Copyright:

Available Formats

The Dynamics of Language Use

Pragmatics & Beyond New Series

Functional and contrastive perspectives

María de los Ángeles Gómez-González

John Benjamins Publishing Company

of American National Standard for Information Sciences – Permanence

Library of Congress Cataloging-in-Publication Data

International Contrastive Linguistics Conference (3rd : 2003 : Santiago de

© 2005 – John Benjamins B.V.

any other means, without written permission from the publisher.

Functional approaches to language 3

II. Form and function in a cognitive perspective

The relation of grammar to thought 57

III. Information structure

Incremental Functional Grammar and the language of football commentary 113

IV. Collocations and formulaic language

Formulaic language: An overview with particular reference

VI. Discourse and culture

Emotivity in narrative discourse: Cross-cultural and cross-gender perspectives 327

Language index 395

Michael Cummings is Professor of English Language and Literature, Department of

Susana M. Doval-Suárez is Lecturer in English Language and Linguistics in the De­

María de los Ángeles Gómez-González received her European doctorate in Linguis­

Francisco Gonzálvez-García is Lecturer in English Language and Linguistics in the

ratura y Traducción. Aproximaciones Teóricas. Madrid: Arrecife, pp. 205–225). E-mail:

Pilar Guerrero Medina is tenured lecturer in English Grammar at the University of

J. Lachlan Mackenzie is an Academic Consultant in Languages and Linguistics, hav­

grammar, in particular the incorporation of aspects of discourse analysis into Func­

Montserrat Martínez-Vázquez (PhD, University of Seville, 1990) is Professor of En­

Rafael Monroy is Professor of English and Linguistics at the Department of English

Maite Taboada is Assistant Professor in the Department of Linguistics at Simon Fraser

Kyoko Takashi (Wilkerson) received her PhD in Linguistics at Georgetown Univer­

Douglas Wilkerson is Associate Professor in the Department of Japanese Studies at

. Introduction: Functionalism within the spectrum of linguistic theories

. The main tenets of functionalism

functionalist theory. This collapsing of functionalism and cognitivism is perhaps natu­

. Some further characteristics of functional approaches

modelled in terms of the privileged syntactic argument of a construction. Systemic lin­

Functionalist approaches, which occupy an important place within the spectrum of

see Butler (2003b: 306–331).

Functional approaches to language 

and Meaning. London: Edward Arnold.

Hengeveld, Kees (2004a). “The architecture of a Functional Discourse Grammar”. In J. L.

Functional Grammar web page: http://www.functionalgrammar.com

María de los Ángeles Gómez-González and

. Some terminological issues

purpose of some application; whereas Krzeszowski proposes the opposition text-bound

. The revival of Contrastive Linguistics (CL)

Finally, it should be mentioned that most of the aforementioned contrastive re­

. Looking ahead: Challenges and problems

Especially in Europe, a large number of research projects, conferences and journals

der comparison – as already noted – various kinds of contrastive investigations can be

* We are grateful to Stig Johansson, AnneMarie Simon-Vandenbergen, Christopher Butler,

Aijmer, Karin (1999). “Epistemic possibility in an English-Swedish contrastive perspective”. In

Gellerstam, Martin (1996). “Translations as a source for cross-linguistic studies”. In K. Aijmer,

Hickey, Leo (Ed.). (2003). Politeness in Europe. Clevedon: Multilingual Matters.

Kraig, Olivier (2002). “Translation alignment and lexical correspondences: a methodological

Markkanen, Raija (1985). Cross-language Studies in Pragmatics. Jyväskylä Cross-Language

Riley, Philip (1980). “Towards a contrastive pragmalinguistics”. In J. Fisiak (Ed.), Theoretical

Susana M. Doval-Suárez is Lecturer in English Language and Linguistics in the De

María de los Ángeles Gómez-González received her European doctorate in Linguis

J. Lachlan Mackenzie is an Academic Consultant in Languages and Linguistics, hav

grammar, in particular the incorporation of aspects of discourse analysis into Func

Montserrat Martínez-Vázquez (PhD, University of Seville, 1990) is Professor of En

Kyoko Takashi (Wilkerson) received her PhD in Linguistics at Georgetown Univer

functionalist theory. This collapsing of functionalism and cognitivism is perhaps natu

modelled in terms of the privileged syntactic argument of a construction. Systemic lin

Finally, it should be mentioned that most of the aforementioned contrastive re

The paper by Montserrat Martínez Vázquez discusses a speciﬁc semantico

It is interesting to observe the ﬂow of temporal orientations in Marilyn’s narrative dur