Making Words in English PDF

Magnus Ljung
Making Words
in English
Copying prohibited
All rights reserved. No part of this publication may be reproduced or

transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher.
The papers and inks used in this product are eco-friendly.
Art. No 31008
eISBN 91-44-02481-9
© Magnus Ljung and Studentlitteratur 2003
Printed in Sweden
Studentlitteratur, Lund
Web-address: www.studentlitteratur.se
Printing/year 1 2 3 4 5 6 7 8 9 10 2007 06 05 04 03
Manuella korrigeringar
storlek på noter, not i kap 1 bort.
sidhänvisningar manuellt gjort.
Contents
Preface 7
1 Words 9
The linguistic sign 9
The structure of complex words 16
Defining ‘word’ 19
Reference, denotation, sense 25
The origins of English words 26
Exercises 29
2 The English Word Classes 31

Criteria for word class membership 33
The closed word classes 43
Exercises 46
3 Outline of English Word-formation 49

Defining word-formation 49
The output of word-formation rules 51
Exercises 59
4 Prefixation 61
Characteristics of prefixes 61
Common English prefixes 67
Exercises 77
5 Suffixation 79
General characteristics of suffixes 79
Noun-forming suffixes 85
Adjective-forming suffixes 96
Verb-forming suffixes 103
© Magnus Ljung and Studentlitteratur 2003 3

Adverb-forming suffixes 104
Exercises 106
6 Conversion 109
General characteristics of conversion 109
Productive English conversion patterns 111
Partial conversion 117
Exercises 119
7 Compounding 121
General characteristics of compounding 121
Noun compounds: general characteristics 129
Adjective compounds 134
Compound verbs 138
Compound adverbs 139
Exercises 140
8 Combining forms and

neo-classical compounds 143
General characteristics 143
Neo-classical compounding compared with affixation 145
Productivity of the neo-classical compounds 148
Common combining forms 149
Exercises 151
9 Irregular Word-formation 153

Characteristics and types of irregular word-formation 153
Borrowing 154
Meaning extension 156
Initialisms: abbreviations and acronyms 157
Clipping 159
Back-formation 161
Blends and blending 163
Reduplicative compounds 164
Rhyming slang 166
4 © Magnus Ljung and Studentlitteratur 2003

References 169
Notes 173
Chapter 1 173
Chapter 2 175
Chapter 3 177
Chapter 4 178
Chapter 5 179
Chapter 6 181
Chapter 7 182
Chapter 8 183
Chapter 9 184
Discussion of exercises 187
Index and lists 201

Index 201
List of prefixes 204
List of suffixes 206

Preface
Preface
This book provides an introduction to modern English word-forma-

tion, i.e. the principles that lie behind the explosive production of
new English words. It is primarily intended as a textbook for first-
year university students in English with little or no experience of
linguistics. However, it should also be rewarding reading for the
general public interested in the vocabulary of today’s English.
The book has a fairly long prehistory and has profited a great deal
from the advice and insights of friends, colleagues and students. I
would particularly like to thank Gunnel Melchers, Dave Minugh
and Philip Shaw, who read parts of the manuscript and offered
helpful criticism. I am grateful for comments and advice from Dieter
Kastovsky, Vincent Petti, Peter Sundkvist, Ann-Marie Vinde, the
members of the English seminars at the universities of Stockholm
and Göteborg, and my students at the universities of Stockholm and
Zurich in 2001 and 2002. Finally I would like to express my
gratitude to Kerstin, who went to the trouble of reading all the
versions of the manuscript and was a source of constant encourage-
ment.
Stockholm, May 2003
Magnus Ljung

Preface

1 Words
1 Words
The linguistic sign

Since the early 1970s, we have been living in the information age,
an era characterised by ‘the abundant publication, consumption
and manipulation of information, especially by computers and
computer networks’ according to a recently published dictionary of
American English.1 The dictionary is of course right: nobody would
dispute that there has been a marked increase in the speed and ease
with which humans exchange information in the course of the past
thirty- odd years. At the same time it is important to realise that the
need for information and information exchange is as old as human-
ity itself, and that both humans and animals have always been cru-
cially dependent on information of some kind.
In view of this, it is not surprising that humans and animals alike
have developed ‘information systems’ that allow them to send and
receive information. Information can only be conveyed when a cer-
tain meaning is associated with a certain physical representation
like for instance sound. When a dog growls, for example, the most
likely interpretation is that it is angry: the growl and the meaning
‘anger’ are linked in a way that is immediately understood by both
humans and other animals (especially if the dog bares its teeth at
the same time). Together, the meaning and the sound used to
express it make up a ‘chunk of information’ often called a sign.
In the case of the growling dog, the physical representation used
to express the meaning ‘anger’ is a type of sound, i.e. the dog’s
growl. The growl serves as a vehicle for the meaning. Meanings can
also be expressed in many other ways; in fact, there is no limit to
what can serve as a vehicle for meaning. Consequently, it is useful
to have a common term for all meaning representations: the term
normally used is form. Sound—like e.g. growling—is one type of
form, other types are gestures and facial expressions.

1 Words
In signs like the dog’s growl, there is a ‘natural’ relation between

meaning and form, because regardless of their linguistic or cultural
background, most people would regard the growl as a natural way
to express anger, particularly for animals. Such universally under-
stood signs are called motivated signs and can be described in the
following way:
The motivated sign
Meaning: ‘anger’
natural connection
Form: growl
There are also non-motivated or arbitrary signs. These are the signs
we find in more complicated information systems like human lan-
guages, typically—but not exclusively—represented by the words in
those languages. Such linguistic signs are not universally understood
and the meaning of each sign has to be learnt separately, because
there is no natural relation between form and meaning.2
There is for example no particular reason why the letter combina-
tions cow and dog and the corresponding sound sequences [kau]
and [dɒg] are tied to the meanings ‘a fully bred female animal of a
domesticated breed of ox, used a a source of milk or beef’ and ‘a
domesticated carnivorous mammal that typically has a long snout,
an acute sense of smell … and a barking … voice’ as one dictionary
puts it rather than the other way round.3
Unlike the dog’s growling, the words cow and dog mean nothing
to those who don’t already know the code. Everybody understands
the meaning of a growl, but in order to understand the meanings of
cow and dog, you have to know English. Non-motivated signs like
cow etc. must accordingly be described in the following way:
The non-motivated sign
Meaning: ‘cow’
no natural connection
Form: [kau]/c+o+w/

1 Words
A further consequence of the absence of a natural connection

between meaning and form in these signs is that it would be mean-
ingless to ask questions like ‘Why is a cow called cow?, Why is a dog
called dog?’ There simply are no answers to such questions.
It might seem that you can in fact provide answers to questions
like theses by providing the etymology for the words involved. The
etymology of a word explains its historical development in the lan-
guage. In the case of cow, for example, it is easy to show that
present-day cow can be traced back to Old English c ū.
We could also point out that cu has so-called cognates in other
languages, like for instance German, Dutch, Swedish, Danish, and
Norwegian. The term cognate is used about words in different but
related languages which are assumed to have the same origin in
some earlier common language, in this case the group of Germanic
languages.
It is possible to go back even further, and trace the Germanic
word for cow back to a postulated earlier common language called
Indo-European. But no matter how far back we trace the history of a
word, there is still no explanation why—at a certain point in time—
a certain meaning happened to be expressed by means of a certain
combination of sounds rather than another one4.
The absence of a natural connection between meaning and form
in words like cow, dog has led certain linguists to describe the vocab-
ulary of a language as an unsystematic list, and to contrast it with
the grammar or syntax of the language, which consists of a set of
rules. A typical syntactic rule for English, for instance, is the one reg-
ulating word order: if we change Tony will do it again into Will Tony
do it again, there is a predictable change in meaning from ‘state-
ment’ to ‘question’. In such a rule, the correspondence between
form and meaning is close to 100%.
Vocabulary typically lacks such predictability: in principle, every
sound:meaning combination is unique and must be learnt as an
individual item by the language learner. With a certain amount of
exaggeration we could say that grammar is the intelligent part of
language, while the vocabulary is the stupid part. And certainly, in
the experience of most language-learners, it is easier to learn a rule
than to memorize long lists of words.

1 Words
Apparent exceptions to the arbitrariness of the sign

There are two categories of words that seem to contradict the claim
that vocabulary items in English—and other human languages—are
non-motivated, or arbitrary. The words in the first of these categories
are sound-imitating: they imitate sounds connected with the phe-
nomena they are used to refer to. (Another name for the sound-imi-
tating words is the Greek term onomatopoeic [ɒnəmtə pik] words,
from onomatopoeia [ɒnəmtə piə], literally ‘sound-imitating’).
Typical sound-imitating words are imitations of the calls of cer-
tain animals, for example English miaow for the sound typically
made by cats, cock-a-doodle-do for cocks, bow-wow or arf arf for dogs,
moo for cows etc. There are also a number of other sound-imitating
words that are not connected with animal calls, but imitate other
sounds, like e.g. clash, roar, bang, plop, plod, buzz, hiss, murmur,
zoom. (Note also the written sound-imitating words used in cartoons
and comic strips like chomp, kadunk, ke-rack, ke-rash, sploosh, splash,
vr-o-o-m and the names of cereals like crunchies and puffs.)
The other category challenging the view that words are non-
motivated consists of words that are instances of so-called sound
symbolism. Sound symbolism rests on the assumption that certain
sounds or sound combinations are associated with certain—usually
rather vague—meanings.
A well-known English instance of sound symbolism is the word-
initial consonant combination sl- found in many words suggesting
something unpleasant, the unpleasantness often involving wetness
of some kind: slime, slush, sloppy, slaver, slither, slug, slobber.
It has also been suggested that there is a tendency for the English
vowel sound [] to occur in words associated with dullness or indis-
tinctness, like e.g. dull, thud, thunder, dusk and many others. A third
claim is that the vowels [] and [i:] are characteristically found in
many words associated with smallness, like e.g., bit, thin, little, wee,
teeny, lean, meagre.
What, then, is our final verdict on these exceptions from the tra-
ditional view of the nature of the linguistic sign? Starting with the
sound-imitating words, there is no denying that words like miaow,
bow wow, arf arf, etc. are intended as imitations of the typical calls of
certain animals. However, there are several reasons why the impor-
tance of these words should not be exaggerated.

1 Words
To begin with, the sound-imitating words are quite few in

number. Secondly, and more importantly, different languages rep-
resent what must be the same animal call in ways that are some-
times quite different. It is true that many languages represent the
sound made by cats in more or less the same way, i.e. as something
like miaow. But as the following list shows, the barking of dogs is
represented quite differently in many languages.5
Afrikaans: woef
Albanian: ham ham/hum hum
Arabic (Algeria): haw haw
Bengali: ghaue-ghaue
Catalan: bup bup
Chinese (Mandarin): wang wang
Croatian: vau-vau
Danish: vov
Dutch: woef
English: bow wow, arf, woof, ruff ruff
Esperanto: boj
Estonian: auh
Finnish: hau hau/vuh vuh
French: ouah ouah
German: wau wau, wuff wuff
Greek: gav
Hebrew: haw haw (/hav hav)
Hindi: bho-bho
Hungarian: vau-vau
Icelandic: voff
Indonesian: gonggong
Italian: bau bau
Japanese: wanwan, kyankyan
Korean: mung-mung (/wang-wang)
Norwegian: voff/vov-vov
Polish: hau hau
Portuguese (Portugal): au au au (nasal diphtong)
Portuguese (Brazil): au-au
Russian: gav-gav
Slovene: hov-hov
Spanish (Spain, Argentina): guau guau

1 Words
Swedish: vov vov

Thai: hoang hoang (with falling tone)
Turkish: hav hav
Ukrainian: haf-haf
Vietnamese: wau wau
We could of course try to explain away the differences in the list by

claiming that dogs sound different in different countries. However,
a more plausible explanation would seem to be that animal calls are
perceived and represented differently by speakers of different lan-
guages. This is an indication that even in words imitating animal
calls, the relation between meaning and form is less than ‘natural’.
Let us turn now to the category of sound symbolism. In the pres-
entation above, three well-known claims concerning sound sym-
bolism in English were mentioned, i.e.the claim that words with an
initial sound sequence sl- denote something unpleasant, the claim
that words that contain the vowel [] are associated with dullness
and indistinctness, and the claim that words containing the vowels
[] and [i:] have meanings to do with small size.
Claims like these cannot be rejected out of hand: there is no
doubt that there is a tendency for certain sounds and sequences of
sounds to turn up in words with certain meanings so often that it
cannot be due to mere chance. This is particularly true of the initial
sequence sl- as in slime, slush, sloppy etc.
However, in all these cases there are far too many counter-exam-
ples for the sound-meaning associations to be anything else than a
tendency. To take just a few examples: there is no suggestion of
dullness or indistinctness in the words but, putt, mutt, luck, duck, sun
and hundreds of other words with the same [] sound. Nor is there
anything particularly unpleasant about the words slow, slight, slim,
slope, sling, slender. As for the claim that the i-sounds suggest mean-
ings connected with small size, there are vast numbers of words
with that vowel sound whose meanings have nothing to do with
smallness, for instance finger, sing, thing, tip, nifty, dish, lift, sick,
inside, ill. Clearly the meanings of these words have no connection
with smallness, and at least one i-word has the opposite meaning,
viz. big.

1 Words
A real exception to the arbitrariness of the sign:

word-formation
In the preceding section we found that although there seems to be
some truth in the claims made for onomatopoeia and sound-sym-
bolism in English, these phenomena are so limited and inconsist-
ent that they cannot be said to pose a serious threat to the generally
accepted view of English words as non-motivated signs.
However, if we return for a moment to the words like e.g. cow,
dog, etc. used earlier as evidence of the non-motivated nature of
English words, we will find that they are all simple words: they can-
not be further analysed into smaller meaning-carrying chunks. Lin-
guistic forms that cannot be further analysed into smaller meaning-
carrying elements are called morphemes.6 Simple words are one type
of morphemes.
Simple words play a prominent part in the English vocabulary: a
1974 study7 showed that of the 10, 000 most frequent words in
written English, roughly half were simple. Of the remaining 50 per
cent, the majority were complex words created by the addition of
prefixes or suffixes to a simple word, as in e.g. rewrite and speaker.
The English vocabulary also contains many compound words like e.g.
blackbird and washing-machine, in which two shorter words are com-
bined to make up a longer one. Compounds and prefixed/suffixed
words are created by different processes of English word-formation
(cf. Chapter 3).
What has been said above about English words may be summa-
rized in the following manner: simple words consist of a single mor-
pheme, complex words contain a prefix or a suffix (or several pre-
fixes and/or suffixes), compound words are made up of two or more
other words. In what follows I will focus on the complex words.
Complex words are partly motivated, i.e. their form does offer a
certain amount of guidance to their meaning. This is because such
words have been formed in accordance with the principles of Eng-
lish word-formation. More precisely, they have been created by the
addition of affixes, either a prefix like e.g. un- or re- in unfair and
rewrite, or a suffix like -er or -y in speaker and snowy.The affixes are so
few that the speakers of the language know them all.
That knowledge allows them to make educated guesses about the
meaning of a complex word, even if they don’t know the meaning

1 Words
of the stem, i.e. the part of a complex word to which the affix is
added.
They know, for instance that words ending in e.g. -ous or -able
like dangerous or readable are in all likelihood adjectives with certain
characteristic meanings. They also know that words ending in e.g.
-er like runner and dish-washer tend to be nouns describing a person
or machine carrying out a certain activity, and that words that are
verbs and begin with un- or re- like untie and recount denote, respec-
tively, the reversal and the repetition of a process or an action. They
have similar detailed knowledge of most other affixes.
The structure of complex words

A study of the English affixes will show that the great majority can-
not be further analysed into smaller, meaningful units. This means
that most affixes are morphemes, linguistic forms that cannot be fur-
ther subdivided into meaningful units.
In a way, this makes affixes similar to simple words, words that
consist of a single morpheme. But there is a very important differ-
ence: forms that are both words and morphemes can occur on their
own in sentences: they are free morphemes or base morphemes.
Affixes, on the other hand, are bound morphemes: they can only be
used as parts of complex words, never on their own. Dog, eat, slow
are all free morphemes (base morphemes), but re-, pre-, in-, -able, -er,
-ish are all bound. It is important not to confuse morphemes with
syllables: although there are many base morphemes that contain a
single syllable, others may consist of several syllables, for instance
stupid, mahogany, karaoke, rendez-vous.
Base morphemes and affix morphemes are sharply different in
number: there are tens of thousands of base morphemes in English,
but less than a hundred affixes. They also differ in the kind of
meaning they have: base morphemes denote phenomena in the
real or some imagined world, like objects, ideas, substances, proper-
ties, qualities, actions and processes. Affix morphemes have a lim-
ited range of rather special meanings. The suffixes, which are the
largest group, have meanings/functions like ‘manner’ (-ly), ‘agent’
(-er), ‘nominalization’ (-ing), ‘verbalization’ (-ize). The prefixes have

1 Words
meanings like ‘repetition’ (re-), ‘opposition’ (anti-), ‘reversal’ (un-),

‘sequence’ (pre- and post-).
The majority of complex words consist of a single base mor-
pheme and an affix. At the top of the list we find words consisting
of a base morpheme and a single suffix, words like speaker, talking,
slowly. The next most common type is made up of prefix+base mor-
pheme combinations like rewrite, unfair and anti-American.
Complex words may also contain more than one bound mor-
pheme, as for example. unbelievable with the base believe, the prefix
morpheme un-, and the suffix morpheme -able, and the word
untruthfulness, which is made up of the base truth and the affixes un-
, -ful, and -ness. There is in theory no upper limit to the number of
morphemes a word may contain, witness the word antidisestablish-
mentarianism, allegedly the longest English word in terms of mor-
phemes (it contains at least six morphemes). But in actual practice
it is unusual to find English words made up of more than four or
possibly five morphemes.
Complex words have an inner structure or morphology deter-
mined by the order in which the morphemes have been put
together. In a word like unreadable, for example, the base read and
the suffix -able must have been combined to form the adjective
readable before the prefix un- could be added, so the analysis of the
word would procede as in the diagram below:
LEVEL 1 unreadable
LEVEL 2 unreadable
LEVEL 3 read -able
Such an analysis rests on the common linguistic assumption that

multimorphemic words (words containing more than one mor-
pheme) may be analysed on a binary basis. In such an analysis the
word is first divided into two and only two parts or constituents,
which in their turn may be split up into further pairs of constitu-
ents. As soon as there are more than two morphemes in the word
being analysed, the analysis takes place on different levels: on the
first level, the entire word is divided into two constituents. These
constituents are then further divided into two meaningful parts on

1 Words
the next lower level.8 The analysis goes on until the entire word has
been divided into morphemes, in this case un-, read and -able.
The two parts isolated on each level are called the immediate con-
stituents of the unit on the level immediately above. Thus the
immediate constituents of the word unreadable, as analysed above,
are the prefix un- and the adjective readable, and the immediate
constituents of readable are read and -able. But why is such an anal-
ysis of the word considered to be ‘the right one’? Why not analyse
unreadable as consisting of unread and able, and then analyse unread
into un- and read as in the figure below?
unreadable
unreadable
unread
The answer has to do with what we know about the meanings and
combining habits of the bits isolated by the analysis. We know that
there is a prefix un- combining with verbs, found in for instance
unwind, unzip, untie. We are also aware that there is another un- pre-
fix combining with adjectives, in for instance unkind, untrue,
unwise. The two un- prefixes have different meanings: the one
found with verbs means ‘reverse the activity described by the verb’,
while the prefix combining with adjectives simply has negative
meaning: it means ‘not’.
It follows that to unwind something is to reverse the result of a
previous winding process, to untie a knot is to reverse the previous
process of tying that resulted in a knot. This is not a meaning that
makes sense together with the verb read: to unread a book would
have to mean something like ‘to make a book one has previously
read unread’. Meanings like that are impossible in the world as we
know it (but cf. the comments on e.g. unmurder on p. 68) and for
that reason we must reject the analysis in the second figure.
In the analysis in the first figure, on the other hand, we postulate
an adjective readable to which we add negative un-, an operation
resulting in the word unreadable meaning ‘not readable’, a meaning
that makes perfect sense.

1 Words
Defining ‘word’
So far I have been using the term ‘word’ as if it were a well-defined
entity with a single meaning. But the word word is a slippery cus-
tomer and without wishing to complicate matters more than neces-
sary, I must now point to some of the different meanings of the
term. My discussion will focus on the written word and will avoid
as far as possible the problems raised by words in the stream of
speech.
Words as types and tokens9

Written words may be viewed in basically two ways: as physical
entities—for example on a printed page—and as more abstract enti-
ties. The physical entities that we see when looking at a printed
page or a hand-written letter are known as word-forms or ortho-
graphic words. Such words may be defined as unbroken combina-
tions of letters preceded and followed by empty spaces and linked
to a meaning. The very fact that words are bounded by spaces indi-
cates that they are free units, i.e. they can stand alone in a sentence.
This ability to occur on their own is one of the main characteristics
of words in English.
Let us now make use of the above definition of the word to deter-
mine the number of words in the following short text:
(1) The cook was a good cook as cooks go, and as cooks go, she
went10.
The total number of letter combinations surrounded by spaces

(and/or punctuation marks) in the text above is 15. If by ‘word’ we
mean ‘word form’ or ‘orthographic word’, there are 15 words in the
text. Some of them occur twice (cook, cooks, go, as) but since we are
counting actually occurring word-forms, that’s not relevant. We
count all occurrences, even if they are repetitions.
However, there is also another way of counting word-forms: we
may feel that each word-form should only be counted once, i.e.
that repetitions of the same letter-combination should not count.
In this approach to word-definition, what we wish to find out is

1 Words
how many different word-forms there are in the text above. Count-
ing like that, we find that four of the word-forms are used twice, i.e.
cook, cooks, go, and as. Since we are now interested in different word-
forms, we count each of these four word-forms only once. That
gives us a total of 11 words (= different word-forms) in (1).
The words found in the first count—the one in which all repeti-
tions were counted—are called tokens, while the ones found in the
second count—the different words—are known as types. The dis-
tinction between tokens and types is made use of in frequency
word-lists from different texts, like the list below. Such lists are
made up of two columns, one containing the word types, the other
indicating the number of tokens each type has in the text. The
number of tokens for each type is known as the frequency (of occur-
rence) of that type.
Tabell 1.1 Word frequency list for the text in (1).
Types Frequency (Tokens)

as 2
cook 2
cooks 2
go 2
a 1
and 1
good 1
she 1
the 1
was 1
went 1
In the list above, the types as, cook, cooks, go have two tokens each,
while the types a, and, good, she, the, was, went, only have one token
each. The total number of tokens (15) is higher than the number of
types (11). This is the situation in all normal texts, where the
number of tokens is usually much higher than the number of
types.11 There is a simple reason for that, of course, i.e. the fact that
some words—in particular ‘grammatical’ words like the, a, an and

1 Words
others in the ‘closed word-classes’ (see pp. 43–46)—have meanings

or functions that make it necessary to repeat them often. In fact, in
large texts, the most common word is the definite article the, which
makes up ca 6% of the total number of tokens.
The type: token distinction is sometimes used as an instrument to
measure how difficult a certain text is to read from the vocabulary
point of view. The underlying idea is that the greater the number of
individual types the reader has to know, the more difficult is the
text. The degree of difficulty is measured by computing the so-
called type:token ratio for the text, a ratio which is obtained by divid-
ing the number of types in the text by the number of tokens. The
higher the type:token ratio, the more difficult the text.
We can illustrate the way the type:token ratio works by assuming
that there are three texts A, B and C, each containing 100,000
tokens each. The number of types in A is 11, 253, in B 6998, and in
C 9541. The type:token ratios for the three texts will then be as fol-
lows: A = 0.11, B = 0.07, C = 0.09. According to this method of
measuring text difficulty, A is clearly the most difficult text, C the
next most difficult one, and B the easiest.
When used as an indicator of text difficulty, the type:token ratio
as presented here also has certain shortcomings, one of which is
that it is sensitive to text length: the longer the text, the lower its
type:token ratio. In order to overcome such weaknesses of the
method, several refinements have been introduced; there are also
alternative methods for measuring overall text difficulty. One well-
known alternative method measures the lexical density of the text.
This is obtained by dividing the number of content words like nouns,
verbs and adjectives by the number of function words like preposi-
tions, articles and pronouns. (The terms content word and function
word are defined and discussed in Chapter 2).
The word as an abstract entity—the lexeme

The identification of word-forms and the counts of tokens and
types described in the previous section has so far been quite uncom-
plicated. It is based on the assumption that a word-form can be
defined as an unbroken string of letters, bounded on both sides by
spaces and linked to a certain meaning. In such an approach it is

1 Words
the physical shape of the items that determines what is a word: if

two strings of letters are even minimally different, as in the case of
cook and cooks, they are by definition ‘different words’.
Such an approach has its uses, but there are cases where it seems
reasonable to argue that certain formal differences among words are
less important than others, in particular differences caused by the
choice of inflectional suffix, like the plural -s, the third person singu-
lar -s, the regular past tense -ed etc., which are added to words with
almost total regularity.
This regularity means that inflectional suffixes add nothing new
to the basic—or ‘lexical’—meaning of words. As a result, words that
differ only with regard to inflection have the same basic meaning,
and can be counted as instances of ‘one and the same word’ in a
more abstract sense of ‘word’. The term used for this abstract notion
of word is lexeme12. It is lexemes that we have in mind when we say
e.g. What’s the meaning of that word?, I don’t know that word or This is
a new word.
Most lexemes differ from other lexemes both in form and mean-
ing: that’s how pull differs from full, grouse from mouse, and wizard
from horse. However, words may be different lexemes but have the
same spelling and/or pronunciation. In such cases it is the differ-
ence in meaning that keeps them apart. Thus for instance there are
two lexemes bank, both of which are nouns but with quite different
meanings:bank (1) means ‘institution that handles money’, and
bank (2) means ‘riverside’. Lexemes with the same form are said to
be homonyms13.
The two lexemes bank above belong to the same word class. How-
ever, in many cases, homonymous lexemes belong to different
word classes. Thus the noun lexeme bank ‘institution handling
money’ has a parallel verb lexeme bank, meaning among other
things ‘deposit money in a bank’ and ‘regularly use as a bank’.
Although the meanings of the noun bank and the verb bank may be
said to be rather similar, the fact that they belong to different word
classes automatically makes them different lexemes. English has a
huge number of such formally identical noun:verb lexemes, a fact
to do with the facility with which the language creates verbs from
nouns and nouns from verbs by the process of conversion (cf. Chap-
ters 3 and 6).

1 Words
It is sometimes said that lexemes are the same as ‘dictionary

words’ i.e. the entries (or ‘headwords’) found in dictionaries. This is
correct to the extent that the ‘words’ found in dictionaries are all
lexemes—no dictionary worth its salt would have three different
entries for e.g. laugh, laughed and laughs. However, the reverse does
not hold: it is certainly not true that in order to be a lexeme, a word
has to be found in a dictionary. To begin with, not even the most
compendious dictionary lists all the words in a language. Secondly
the principles of English word-formation—the subject matter of the
present book—are constantly used to create new lexemes. Some of
these will eventually make their way to the dictionary, but for a
number of reasons, many of them will not14.
Naturally, a word count in which ‘word’ is defined as ‘lexeme’
will yield quite different results than our previous counts of types
and tokens. If for example we count the number of lexemes in text
(1), we will find that it contains only nine words ( = lexemes): the,
go, cook, be, a, good, as, and, she. Here cook stands for both the singu-
lar cook and the plural cooks, go stands for both go and went and be is
represented by was. As pointed out a few paragrahs ago, words from
different word-classes cannot represent the same lexeme:it is all
right to include both the singular cook and the plural cooks under
one noun lexeme cook, but although they are formally identical, we
can not include the inflected verb cooks and the noun plural cooks
under one and the same lexeme.
Lexemes and multi-word units

As we saw earlier, a lexeme is a kind of abstract word that incorpo-
rates all the inflected forms of ‘the same word’. There is a clear com-
monsensical basis for this definition of the lexeme: after all we feel
strongly that inflected forms are just some sort of variants of ‘the
same word’. There are also other good reasons to believe that the
notion of lexeme is a reasonable one, for instance the fact that
speakers seem to store their vocabulary knowledge in terms of
lexemes or something like them.15
But the notion of lexeme not only permits us to regard inflected
forms as variants of the same word, it also helps solve another lin-
guistic problem, i.e. what to do with so-called multi-word units.

1 Words
These are combinations of separately written words considered to

form semantic units, i.e. to represent somehow ‘a single idea’. The
notion of lexeme permits us to give these combinations lexeme sta-
tus, i.e. to count them as single words. As an exemple, consider the
following sentence:
(2) We looked at the ashtray in front of our wine glasses.

This text contains seven lexemes, four of which are multi-word
units, i.e. look at, ashtray, in front of and wine glass. The reason why
we regard these combinations as lexemes is above all their semantic
specialisation. The ashtray, for instance, is a particular object whose
only function is to provide a place where smokers may deposit the
ash from cigarettes, cigars and pipes.
The same logic applies to wine glass. Even if it is perfectly possible
to drink wine from a beer glass, that doesn’t make it a wine glass. A
wine glass has a characteristic shape that distinguishes it from a
beer glass, etc. Both combinations are non-motivated signs that
must be learnt in the same way as single words. Counting look at
and in front of as units is made easy by the fact that they may be
replaced by single words with the same meanings, i.e. regard and
before.
English is a language rich in multi-word units. It has thousands of
such combinations with lexeme status, for instance clichés, com-
plex prepositions, compounds, idioms, phrasal verbs, prepositional
verbs. The following list illustrates some of the variety found in this
area:
Clichés: call it a day, call a spade a spade, the devil take the hindmost,
famous last words, go through the roof, go missing, the show ain’t over
till the fat lady sings, beggars can’t be choosers, you are the cream in my
coffee!
Complex prepositions: because of, except for, owing to, in front of, in
back of, in place of, in spite of, as a result of
Compounds: blackbird, teapot16, power plant, steamboat, English

teacher (‘person who teaches English’); awe-inspiring, mind-boggling,
self-styled, sun-tanned, user-friendly; crash-land, downgrade

1 Words
Idioms: in the doghouse ‘in disgrace’, up shit creek without a paddle ‘in
serious trouble’ (US slang), come a cropper ‘fail’, kick the bucket ‘die’,
as sober as a judge, as pissed as a newt (British slang for ‘dead drunk’).
Phrasal verbs: touch down, fall out (‘quarrel’), give in, come in, go on,
set up, bring up, call off, take in (‘deceive’), turn off, switch on
Prepositional verbs: look at, care for, think of, call on, succumb to, die
of, suffer from, talk about, shudder at, ask for, believe in
The categories above are often difficult to tell apart, in particular

‘clichés’ and ‘idioms’ 17. Both belong to a more comprehensive class
that we may call set expressions or set phrases. Many of them involve
the phenomenon metaphor18 It is an open question how to treat cer-
tain other multiword-word units, for example many of the interjec-
tions like e.g. Over my dead body!, Fat chance! and the many obscene
and vulgar interjections found in swearing like Fuck you!, Up yours!
and Shit a brick! If they are lexemes, what are their meanings?
Reference, denotation, sense 19
Before the discussion of lexemes is brought to a close, something

needs to be said about lexemes and meaning. There are at least
three aspects of the word ‘meaning’ that need to be considered. To
begin with, it is obvious that lexemes are used in utterances to call
attention to individual phenomena in a real or imagined world.
Suppose Bruce says to Irvin ‘Sorry I’m late, but I was attacked by a
cobra.’ Regardless of whether he is right in classifying the attacking
animal as a cobra or not, he has used the word to ‘stand for’ or ‘rep-
resent’ an animal, i.e. a phenomenon in the ‘real world’ also known
as an extralinguistic phenomenon. In linguistic terms he has used
cobra to refer to that animal. The function that cobra has in his
utterance is known as reference and the animal itself is known as
the referent of the word.
Reference is linked to individual utterances. But in order for the
reference function to work, the word cobra must have denotation
(also called extension), i.e. there has to be a class consisting of all the
animals that the word cobra may be used about20 This raises the

1 Words
interesting question of how such a class is defined, a question to

which we return in Chapter 2.
There is also a third aspect of meaning known as sense. The sense
of a word is basically its definition in terms of other words in the
same language. The fact that lexemes have sense means that we can
relate them to each other by means of so-called sense relations, for
example the relation of synonymy: synonyms are words that have the
same sense21.
The origins of English words

In the discussion of the linguistic sign at the beginning of the
present chapter, there was a brief reference to the etymology of Eng-
lish words. English was categorised as a Germanic language, among
which we also find e.g. Dutch, German, and the Scandinavian lan-
guages. The Germanic languages were furthermore shown to be one
of the many subcategories of the Indo-European language family
(cf. end note 4 to this chapter).
However, the characterisation of English as a Germanic language
needs qualifying in several respects. In particular, it is important in
a study of word-formation to point to the special nature of the Eng-
lish vocabulary, which differs from the vocabularies of the other
Germanic languages in having a much larger proportion of words
in common with Latin, Greek and the Romance languages.
The history of English begins around 449, when Britain was
invaded by three Germanic tribes from the mainland, i.e. the
Angles, the Saxons, and the Jutes. The members of these tribes—
who eventually became known as the Anglo-Saxons—became set-
tlers and eventually took possession of most of what is now Eng-
land. The original Celtic inhabitants were either killed, assimilated
or forced to take refuge in the extreme west of Britain.
The invading tribes spoke Germanic dialects that appear to have
been quite similar and which eventually merged into a fairly uni-
form Germanic language called Englisc, a word derived from the
name of the Angles. The invasion and settling down of these tribes
mark the beginning of a stage in the development of English
known by linguists as the period of Old English. Although, like all
languages, Old English changed in the course of time and also

1 Words
developed a number of different dialects, it remained basically the

same language for 700 years—from 450 to 1150.
The Germanic nature of English was further strengthened by a
number of Scandinavian (Danish) invasions, which extended over
more than two centuries. Eventually large settlements of Scandina-
vians arose and the settlers became part of the population of Eng-
land. The languages of the invaders and of the native inhabitants
were obviously related and sometimes mutually intelligible.
From around 1150, both the grammar and the vocabulary of Old
English started to undergo important changes that led to the devel-
opment of what is known as Middle English, a second stage in the
development of English generally considered to have lasted from
1150 to about 1500. At the beginning of the16th century, new
changes ushered in the still ongoing period of Modern English. Nat-
urally these ‘periods’ are merely useful fictions: language change
does not procede in such regular fashion and not all changes take
place at the same time. Neverthless, the terms Old English, Middle
English and Modern English are useful labels for stages in the devel-
opment of English during which the language had certain identifi-
able characteristics.
Thus from the point of view of the vocabulary, Old English was a
clearly Germanic language with cognates in other Germanic lan-
guages. Many of the Old English words have since been replaced by
borrowings, but many—e.g. child, cow, live, man, wife—remained in
the language and now belong to the core vocabulary of Modern
English. Old English also had a powerful word-formation system,
and many of the Old English word-formation elements are still
used, for instance the suffixes -hood, -ness and -er and the prefix un-.
There was also a small but noticeable influx of loans from Latin
into Old English. Some of the loans—like e.g. wall, street, wine—had
been borrowed into the language of the invading tribes while they
were still living on the continent. Others were borrowed from Latin
into Old English, especially after the introduction of Christianity in
597. Many of the new words introduced after this date had to do
with the Church and with Church-related activities like education
and learning, for instance priest, martyr, mass and master, meter,
school. These Latin loans constitute the first of several foreign
enrichments of the English vocabulary.

1 Words
The second and much more important wave of borrowing into

English came about as the result of the conquest and occupation of
Britain by the Normans, who spoke a variety of French. The actual
conquest took place in 1066. For roughly 200 years after that,
French was the language of the court, the ruling classes, the law
courts, the military, etc. However, from 1250 onwards, English
started to come into its own again as the ruling classes gradually
adopted English as their language. Used as they were to speaking
French, they liberally sprinkled their English with French words, a
habit that led to a massive influx of French words into English.
The extent of borrowing from French was such that it had vast
and permanent consequences for English. It has been estimated
that by the end of the Middle English period, the total number of
French loans into English was approximately 10, 00022, most of
which are still in use, including words as different as office, paralytic,
debility, innumerable, incumbent, nervous, testify, reject. From a word-
formation perspective, the loans from French were also important
by bringing into the language a vast number of new suffixes like
-ice, -ic, -ity, -able, -ent, -ify, -ous and -ify and prefixes like de-, in- and
re-.
A third wave of borrowing took place during the Renaissance, i.e.
in the early Modern English period (1500-1650). This was a time of
linguistic experiment marked by a desire to embellish English by
the introduction of foreign—mostly Greek and Latin—words.
Much of the borrowing was not triggered by the need for new
terms, but by a desire to incorporate as many Latin words as possi-
ble into English. Among the permanent borrowings from this
period we find e.g. education, expensive, emancipate, harass, benefit.
The fourth and most recent wave of borrowing involves so-called
‘combining forms’ from both Greek and Latin, which are used to
build neo-classical compounds in learned, scientific and technical
language. Although this type of word-formation has been available
in English for centuries, the rapid rise in importance of the sciences
in the 19th century triggered a tremendous increase in the produc-
tivity of this word-formation pattern. It is now one of the most pro-
ductive English word-formation patterns in the language, although,
admittedly, much of its output is understood by only a limited
number of specialists in each field.

1 Words
As the discussion in this section has shown, English has been

exposed to four major infusions of non-Germanic vocabulary in the
course of its history, mainly from French and Latin, but to some
extent also from Greek23. The extent of this influence has been so
strong that from the point of view of its vocabulary, English now
occupies a position mid-way between the Germanic languages and
the Romance languages that developed out of Latin, like French,
Italian and Spanish.
Exercises
1 Explain the difference between motivated and non-motivated
(arbitrary) signs. In which category would you put e.g. flag sig-
nals, human sign language and and the way(s) many animals
mark their territory?
2 What is the difference between grammar/syntax and vocabu-

lary discussed on p. 11?
3 On pp. 12–14, there is a discussion of onomatepoiea and sound

synbolism. Explain what the difference is between these two cat-
egories. Can you suggest additional members of each category?
4 Why do you think is it claimed on p. 15 that word-formation is

an exception to the principle that words are non-motivated,
and why is word-formation said to be ‘partly motivated’?
5 Pp. 15–16: What is a morpheme? What is the difference between

morpheme and a stem? What different kinds of morphemes are
there?
6 Try to show the internal structure of the complex words below,

using the kind of diagrams found on pp. 17–18: representative,
recalling, untruthfulness.
7 Explain the type:token distinction and the meaning of the type-

token ratio.
8 Calculate the type-token ratios for the two texts below

1 Words
(A)
When Chomsky’s ideas spread across into the field of psychol-
ogy in the early 1960s, they made an immediate impact. Psy-
chologists at once started to test the relevance of a transforma-
tional grammar to the way we process sentences. Predictably,
their first instinct was to test whether there was a direct rela-
tionship between the two.
Jean Aitchison The Articulate Mammal p. 183.
(B)
Advice for delegates to next week’s United Nations World Sum-
mit on Sustainable Development in Johannisburg: if someone
approaches your car, pushes a gun in your face, and shouts
“Hijack!” don’t reply “My name’s not Jack.” The good news is
there will be 8000 extra police for protection at the biggest con-
ference ever held in South Africa’s biggest and baddest city.
TIME Magazine, Aug. 26, 2002, p. 29.
9 What are the lexemes in texts (A) and (B) above? Are there cases
in which several answers are possible?
10 What non-Germanic languages have had the greatest impact on

the vocabulary of English?

2 The English Word Classes
If you look up a word—i.e. a lexeme—in a dictionary of English,

you will find that in addition to having meaning, it is also assigned
to a certain word class. Word classes are made up of words sharing
certain important characteristics, about which more will be said
presently. An understanding of the word classes—in particular the
‘open’ classes—and the criteria for inclusion in them is essential for
an understanding of word-formation processes.
There is some variation among dictionaries and grammars in
what word classes they recognise, but by and large the following is
a representative selection1: nouns (minister, tree, idea, confusion),
verbs (walk, write, realise), adjectives (big, thoughtful, economic),
adverbs (slowly, moneywise, yesterday), auxiliaries (can, have, will),
conjunctions (and, because, while), prepositions (on, during, like),
determiners (the, a/an, some), pronouns (it, she, this). A tenth class
that is also often included is the interjections (Oh!, Really?!, Yuk!,
Yum-yum!).
In the rest of this book, the following abbreviations will be used
to indicate word class membership: (n) for noun, (v) for verb, (adj)
for adjective, (adv) for adverb, (aux) for auxiliary, (pron) for pro-
noun, (conj) for conjunction, (prep) for preposition, (det) for deter-
miner, and (int) for interjection.
The word classes above fall into two broad subcategories known
as content words and function words (the latter sometimes also called
grammatical words or form words.) The word classes containing con-
tent words are the nouns, the verbs, the adjectives, and—partly—
the adverbs. The word classes made up of function words are the
auxiliaries, the pronouns, the conjunctions, the prepositions and
the determiners. The interjections are sometimes included among
the function words, but are really a category of their own.

Content words describe the content or meaning of sentences and

are used to refer to what earlier we called extralinguistic phenomena
or ‘things’ in the ‘real world’; of course they can just as well be
‘things’ in an imagined world, as in the case of science fiction and
similar genres. The word classes they belong to are known as ‘open
classes’, since new words can be added to them by means of borrow-
ing or word-formation. In fact, new words are constantly being
added to these classes, which explains why it is impossible to give an
exact figure for the number of words in English or any other
language2. In the present chapter, the focus will be on the open
classes, i.e. on the content words.
The function words differ from the content words in not being
used to refer to phenomena in the ‘real world’ (or some imagined
world). Instead most of them serve chiefly to link content words to
each other in more or less intricate patterns. In fact we often find
the term ‘bricks and mortar’ applied to content words and function
words. The underlying metaphor here is the notion that the func-
tion words keep the content words in place in the same way as mor-
tar keeps the bricks in place.
The word classes made up of function words are ‘closed classes’ in
the sense that it is rare or extremely rare for new members to be
added to them. The closed word classes differ among themselves in
this respect, however. Some of them, like the prepositions and the
conjunctions, are not particularly hostile to the addition of new
members. The class of personal pronouns, on the other hand, has
been closed to new members for roughly 1000 years. The last time
new personal pronouns were added to English was during the Old
English period, when the pronouns they, their and them were bor-
rowed from Scandinavian and replaced the older forms.
It was hinted above that the interjections are really a category of
their own. They are neither content words nor function words, but
words used as conventional representations for the speaker’s or
writer’s feelings or reactions: Wow! is the conventional expression
for surprise and admiration, Hell! and many other swearwords
express irritation etc. Although the interjections are traditionally
included among the word classes, their status as words is uncertain
and it has been argued that they are really a (minor) type of
sentences3.

The occurrence of function words is often so predictable that

(with the exception of most prepositions and conjunctions) they
may be left out without loss of meaning. This happens especially in
texts using simplified language, like recipes, manuals, and newspaper
headlines. The space available for such texts is often limited and, on
account of that, the authors of such texts tend to keep only those
words that are necessary for the message to be understood. The
words they keep are the content words, and the words they omit are
function words. This can be seen in the following examples of sim-
plified English, the first one from a recipe, the second taken from a
computer manual, and the third from a newspaper headline.
RECIPE
If old, place spinach in 2 cups rapidly boiling water. Reduce heat and
simmer, covered, until tender, for about fifteen minutes.4
COMPUTER MANUAL
When ready to quit, click on Disconnect icon
HEADLINE
Prince to wed page three girl
Criteria for word class membership

As noted earlier, the words in a language fall into word classes
whose members share certain important characteristics. In order to
be a member of a certain word class, a word has to have all or most
of these characteristics: the characteristics function as criteria for
membership in the class. At least for the open classes, the most
obvious shared characteristic is probably meaning, and there have
been many attempts to define the word classes in terms of a seman-
tic criterion, i.e. in terms of the meanings of the members of the dif-
ferent classes.
Thus a time-honoured—but still popular—definition of ‘noun’
says that nouns are words used to identify ‘people, places and
things’. There is no doubt that many nouns will be captured by
such a semantic, or notional, definition, for example teacher (peo-
ple), city (place), book (thing). Presumably we can also add names
like e.g. Susan and Sydney to the noun class, since the first refers to
a person, the second to a place.

However, many words identified as nouns by grammars and dic-

tionaries will be left out of the noun class, if meaning is our only
criterion for membership. Examples of such words are easily found,
for instance idea, salt, anger, departure, none of which denotes
people, places or things. We may wish to stretch a point and argue
that idea can be squeezed into the ‘thing’ category, although it is a
somewhat peculiar ‘thing’, being an abstract noun. But it is difficult
to regard salt, anger and departure as being about ‘things’, let alone
about people and places.
Clearly, if we wish to set up a list of conditions that words have to
meet in order to be included in a certain word class, the semantic
criterion on its own is inadequate. We have to add other condi-
tions—or ‘criteria’—for inclusion in the different word-classes. The
following three additional criteria are often used, particularly for
the open classes:
(1) the functional criterion: words in a certain word class must be
able to have certain syntactic functions like e.g. the subject
function, the predicate function, the predicate complement
function, the modifier function, etc.
(2) the combining criterion: words in a certain word class must be

able to combine with certain function words like e.g. determiners
or auxiliaries (for which see Chapter 2).
(3) the grammatical category criterion: words in a certain word

class must be able to—and in most cases have to—express cer-
tain grammatical categories, like e.g. tense (for verbs) and
number (for nouns), comparison for adjectives and adverbs, etc.
If a word satisfies all the criteria for a certain word class, it is by def-
inition a member of that class. However, even a superficial study of
the English word classes reveals that words differ in this respect. For
every word class that we consider, there are words that meet all the
criteria demanded for inclusion in that class, there are others that
meet some but not all of them, and there are words that meet just
one of these criteria.
Word class membership is thus not an all-or-nothing affair but a
matter of degree: some words are much more typical representa-

tives of a certain word class than others. The words that meet all the
criteria for inclusion in a given word class are central members of
the class, while those meeting only one or a few criteria are marginal
members of that class.5
Criteria for the noun category

In the previous section we noted that, although the semantic defi-
nition of ‘noun’ went part of the way towards defining that class, it
still left out many words that were also felt to qualify as nouns and
that the semantic definition must be supplemented by definitions
using syntactic function, combining habits, and grammatical cate-
gories as criteria. Adding these supplementary criteria to the seman-
tic one, we arrive at a total of four criteria for the noun class6:
1 Function: a noun can function as subject.
2 Combining habits: A noun can be preceded by the and other
determiners and/or by an adjective.
3 Grammatical categories: A noun can express number i.e. be sin-
gular or plural.
4 Meaning: A noun denotes a person, place or thing.
There is also morphological evidence of noun status: English pos-

sesses a set of suffixes typically only found in nouns, for instance
-age, -ation, -tion, -sion, -ion, -er, -ity, -ment, -ness, -ist. Many nouns
also typically take the apostrophe genitive (‘s). The presence of one
of the above suffixes or the apostrophe genitive may be taken as
supportive evidence that a word is a noun. However, neither suffix-
ation nor genitive may be used as a general criterion for noun status
since there are words identified as nouns by all other criteria that
don’t take them.7
Let us now return to the nine words used in the discussion of the
semantic criterion on pp. 33–34 and see how they fare with regard
to the four criteria for noun status listed above. The results are pre-
sented in the table below:

Subject Function Combines with Number Meaning

determiners
teacher + + + +
city + + + +
book + + + +
idea + + + (+)
departure + + + +
anger + + – –
salt + + – –
Susan + – – +
Sydney + – – +
As the table shows, all nine words meet the subject function crite-
rion: they can all replace X in constructions like ‘X is interesting’,
‘X impressed us’, etc. However, the first five words have to be
accompanied by determiners like the, their: we cannot say e.g.
*Teacher (city) impressed us but have to say The teacher (city etc) or
Their teacher (city etc.) (Here and in the rest of the book the asterisk
sign * will be used to identify constructions that are impossible).
Anger and salt may occur both with and without determiners, but
cannot express number. Only Susan and Sydney refuse determiners
altogether and also lack the ability to express number.
The distribution of the different nouns in the table above shows
that the class of nouns is made up of several sub-categories. The first
column is not distinctive in this respect, since all nouns may have
subject function. However, if we move to the second column we
find a clear difference between teacher, city, book, idea, departure,
anger on the one hand, and Susan, Sydney on the other. All the
members of the first group take determiners but neither of the
members of the second group do.
That distinction is the most fundamental one for the noun class.
It divides it into common nouns—which all take determiners—and
proper nouns which do not. Proper nouns, whose special status is
marked in writing by the initial capital, are used as names and nor-
mally take neither articles nor the plural. On the other hand, proper
nouns are certainly used to identify people and places, respectively,
so they would seem to satisfy the semantic criterion in the same
way as the common nouns.
However, on closer inspection, the way proper nouns like Susan
and Sydney identify people and places turns out to be different from

the way common nouns like teacher, city, etc. identify the people,
places and things they are used to refer to. Proper nouns like Susan
or Sydney are used as names: they have no sense or denotation as
those notions were explained in Chapter 1, but simply refer directly
to a specific individual person, place or thing. A personal name like
Susan, for instance, identifies a single individual known to both
speaker and hearer. Accordingly, there is no need to further define
the referent of Susan by means of the definite article the, nor is it
possible to use the indefinite article or the plural.
Within the category of common nouns, a further distinction
needs to be drawn between on the one hand teacher, city, book, idea,
departure, which may occur in both the singular and the plural, and
on the other salt and anger, to which the number category doesn’t
apply. The members of the first group are called count(able) nouns or
countables for short. As the name indicates, these nouns denote
phenomena that can be counted—they denote entities. But salt and
anger are uncountable: salt denotes a type of substance or concrete
mass—nouns like this are often called mass nouns—while anger
denotes a state of mind.
We can sum up our discussion above of the different kinds of
noun as in the following figure:
NOUNS
PROPER NOUNS COMMON NOUNS

no definite article, take the definite article, denote
no indefinite article, classes of phenomena
no plural; unique
denotation
COUNTABLE NOUNS UNCOUNTABLE

NOUNS
take indefinite article, no indefinite article,
have plurals, denote class no plurals, denote
of entities non-entities
Susan, Sydney teacher, city, book, salt, anger

idea, departure

Recategorisation
The classification above needs qualifying in two respects. To begin
with, it is possible to find constructions in which anger and salt
have plurals or are preceded by the indefinite article, for instance:
On those occasions I felt quite different angers.
The mine produces two different salts.
I experienced an anger I had never felt before.
Such a salt has the following properties.
In these examples, we have ‘forced’ words normally belonging to

the uncountables to behave as if they belonged to the countables.
The point of doing that is that it allows us to express a special
meaning: when uncountables are forced to behave like countables,
they take on the meaning ‘type of -‘, ‘kind of - ‘. Therefore salts and
angers in the first two examples mean ‘types/kinds of salt/anger’,
and a salt and an anger in the last two examples mean ‘a type/kind
of anger/salt’. This recategorisation of uncountables as countables
is a useful device, saving us the trouble of writing out type of salt,
kind of anger.
The discussion above had to do with the re-categorisation of
uncountable nouns as countable nouns. Noun re-categorisation can
also go in the opposite direction, i.e. countable nouns—in particular
concrete ones—may be turned into uncountables, a process involv-
ing a semantic change from ‘entity’ to ‘mass’. Thus if for instance a
lot of candles (a count noun) are left burning on a table, the result
may be that there is candle (uncountable noun) all over the table.
A more special meaning change may accompany the recategori-
sation of count nouns as uncountables when the count nouns
denote living beings, above all animals. In such cases the new
uncountable tends to acquire the meaning ‘food’. This kind of
recategorisation and the semantic change it triggers is illustrated in
the examples below:
Would you like a duck?
Would you like a little/some duck?
I like the duck (in the cartoon)
I like duck
The first and third lines above contain the word duck referring to an
animal; in the second and fourth lines the interpretation is ‘food’.

In principle this semantic change is open to all animal-denoting

count nouns. However, different cultures have different views on
what animals are edible and what are not. Exchanges like
Would you like some more dog (horse)?
No thanks, but I’ll have a little more fox (snake).
are linguistically well-formed, but from the point of view of eating

habits, they would be strange in, say, Britain or the US, where dogs,
horses, foxes and snakes are not, as a rule, considered to be food. In
a culture with different eating habits, however, the two examples
above would be both linguistically and culturally perfectly normal.
In a few cases this correspondence between ‘animal’ count nouns
and ‘food’ uncountable nouns does not hold. Meat from calves,
pigs and sheep, for instance, is always referred to as veal, pork and
mutton: nobody says Could I have some calf (pig, sheep) please.
The second qualification of the proposed noun classification con-
cerns proper nouns. As we have seen, they differ from the common
nouns in taking neither articles nor plural. However, there are cir-
cumstances in which proper nouns are reclassified as common
nouns. When that happens, these words take both the articles and
the plural as in e.g.
This is not the Sydney I once knew.
The Smiths of this world have a tough time.
I know several Susans. Which Susan do you mean?
In the examples above, the normally proper nouns Sydney, Smith

and Susan have been reclassified as common count nouns and have
acquired new meanings in the process. Thus Sydney here means
‘city called Sydney’, Smiths means ‘people called Smith’, and Susan
means ‘a person called Susan’.
In a similar manner, family names in the plural preceded by the
definite article are used as common nouns to denote an entire fam-
ily:
Has anyone told the Wilsons?
We’re going to the Browns tonight.
On the last few pages we have discussed recategorisations within

the class of nouns that are quite regular and accompanied by a cer-

tain change in meaning. Thus uncountables may be turned into

countables with meanings like ‘type of’, countables can be turned
into uncountables, often with the meaning ‘food’, and proper
nouns like Susan and Sydney may be transformed into common
count nouns meaning ‘person called Susan’, ‘city called Sydney’.
These changes sound remarkably like the type of word-formation
known as conversion discussed in Chapter 6. However, in this
book– as in most other word-formation studies—conversion is
regarded as a process that is by definition word-class-changing. For
this reason, recategorisations within the same word-class like the
ones discussed above are regarded as belonging to syntax rather
than to word-formation.
Criteria for verbs

The verbs to be discussed here are the so-called lexical verbs (also
labelled full verbs or main verbs). There is also the closed class of aux-
iliary verbs which do not play a part in word-formation but will be
listed for recognition purposes in a later section. The lexical verbs
form an open class. In order to be included in the class, a word must
have the following characteristics:
1 Function: verbs are words able to function alone as predicates in
a clause as in e.g. They left
2 Grammatical categories: regular verbs have finite forms which
inflect for the present tense by adding the inflectional suffix -(e)s
and for past tense by adding -ed. They also have three non-finite
forms: the infintive (to chatter), the present participle (chattering),
and the past participle (chattered). Some 200 irregular verbs have
other tense forms.
3 Combining habits: lexical verbs fall into two main classes
depending on whether they can take a direct object or not. Those
that can (e.g. see, kill, contain) belong to the transitive class while
those that cannot (e.g. chatter, die, participate) belong to the
intransitive class. There is also a small number of lexical verbs that
are always followed by predicate complements (predicatives), for
example be, become, turn as in She was a genius, He became/turned
angry.

4 Meaning: the majority of the lexical verbs denote acts, actions,

activities, event and processes. A minority denote states and rela-
tions.
From a word-formation perspective, the most interesting features of

the verb are the participles in -ed and –ing, which have both adjec-
tival and verbal functions. Thus amusing and frightened are partici-
ples—and consequently have verbal function—in e.g. Jean was
amusing us with her imitations of the mad professor and James was
frightened by his friends, but function as adjectives in e.g. her perform-
ance was very amusing, it was an amusing performance, I was beginning
to feel extremely frightened, and There was a frightened silence.
Certain forms in -ed and -ing are mainly or exclusively used as
adjectives, for example interesting, appalling, concerned, and flabber-
gasted. When forms in -ed and -ing meet all criteria for adjective sta-
tus, they are regarded adjectives converted from verbs (cf. Chapter
6). A useful test to find out whether a form in -ed or- ing is adjectival
or participial is to insert an intensifier like very or extremely imme-
diately before it. If the insertion produces an acceptable result, then
the -ed and -ing forms in question are indeed adjectival.
In addition to the characteristics above, some verbs are also for-
mally distinctive by ending in certain suffixes like -ate, -en, -ify, -ise/
ize, e.g. originate, soften, gentrify, internalise/internalize. Note however
that the distinctive verb suffixes are far fewer than the distinctive
noun suffixes.
Another formal characteristic of English verbs is that the lan-
guage has a vast and growing number of multi-word verbs (cf. also
Chapter 1 p. 25). The majority of these are either phrasal verbs or
prepositional verbs, both of which involve verbs followed by parti-
cles. Phrasal verbs are lexemes in which a verb and a following
adverb form a unit. Phonologically, the phrasal verbs have a dis-
tinctive stress pattern with the main—so-called primary—stress on
the adverb: put ón, slow dówn, look úp. Phrasal verbs may be both
transitive and intransitive and often the same item belongs to both
categories, cf. e.g. The train slowed down (intransitive)—The bad
weather slowed down the train (transitive). Prepositional verbs are lex-
emes in which a verb and a preposition form a unit. In typical prep-
ositional verbs, the main stress is on the verb: lóok at, thínk of, applý
for.

Criteria for adjectives

There are two syntactic criteria for adjectives and one semantic or
notional one.
(1) Adjectives can function as so-called predicatives or predicate com-
plements i.e. after verbs like be, become, seem, turn and certain
others. The adjective angry is a predicate complement in e.g. The
crowd was/became/seemed/turned angry8.
(2) Adjectives can be used attributively i.e. they can function as so-
called premodifiers before nouns, as in the angry crowd.
(3) Adjectives usually denote states or qualities that are gradable.
Most adjectives meet the first two criteria. Among the exceptions
we find words like e.g. former, naval, utmost, sheer, which are used
only attributively, and words with initial a- like afraid, aghast,
aware, which are used only predicatively. The majority of the adjec-
tives also meet the third criterion, i.e. they have gradable meaning
and can accordingly be modified by intensifiers (cf. next section) like
e.g. very, incredibly, extremely, and be compared by means of the suf-
fixes -er and -est, or the adverbs more and most. Certain other adjec-
tives are non-gradable and normally take neither intensifiers nor
comparison, for example the colour adjectives and words like
atomic, linear, lunar, naval, nuclear.
The bulk of the adjectives may also be turned into adverbs in -ly,
for instance quick-quickly, incredible-incredibly, surprising-surprisingly.
In addition, quite a number of adjectives are morphologically
marked: it can be fairly safely assumed that words ending in e.g.
-able, -ed, -ful, -ic, -ish, -ive, -less, -y will also meet the other criteria
for adjectivehood. (However, note the comments on suffixation
and conversion in Chapter 6).
Criteria for adverbs

Adverbs are unlike the other open9 word-classes in that it is impos-
sible to set up a number of common characteristics shared by all
members of the class. Not surprisingly, the adverb category has
sometimes been called the grammarian’s waste paper basket. How-
ever, a certain amount of order can be brought into the adverb class

if we focus on functions and meanings. Such an approach allows us

to set up the following simplified classification of English adverbs.
ENGLISH ADVERBS
INTENSIFIERS NON-INTENSIFIERS
(Modifying function) (No modifying function)
very (large) MANNER LOCATION TIME SPEAKER

highly (intelligent) ATTITUDE
somewhat (old)
a little (boring)
angrily here, there, often Fortunately
slowly skywards seldom Regrettably
The intensifiers may be used to modify all gradable adjectives and

adverbs. They usually heighten the degree of the characteristic
denoted by the adjective/adverb: very large is larger than just large,
doing something terribly badly is decidely worse than just doing it
badly, etc. More rarely the intensifier has a degree-lowering effect, as
in somewhat old and a little tediously (a little is counted as a complex
adverb).
Among the non-intensifiers, the most interesting category for our
purposes is that of Manner. Manner adverbs are regularly formed by
means of word-formation: as pointed out in Chapter 5, we can form
a manner adverb from almost any adjective by adding the suffix -ly,
a suffix that may also be used to form adverbs expressing the
speaker’s attitude to what he/she is saying as in Fortunately (Regret-
tably), there was nobody else around. Other adverb suffixes are
described in Chapter 5.
The closed word classes

In comparison with the open word-classes, the closed classes have
little to offer the student of word-formation. However, some of the
arguments about word-formation in the book presuppose familiar-

ity with certain aspects of the closed classes. The following brief
comments will hopefully facilitate the reader’s understanding of
those passages.
Auxiliaries
Formally, the auxiliaries as a group differ from lexical verbs in three
important respects:
(1) They don’t take the do-construction in negative or interrogative
sentences, and as a consequence they are moved to the begin-
ning of the sentence in direct questions
(2) They contract with a following not (as in shouldn’t, can’t)
(3) They cannot stand alone as predicate verbs in a sentence
(except in cases like e.g. [Can she come?]—Yes, she can/No she
can’t, where the lexical verb come has been left out).
From a functional point of view it is worth noting that the term

‘auxiliary’ means ‘helping verb’, i.e. the auxiliaries are verbs which
‘help’ the lexical verbs form certain complex constructions. There
are two broad categories of auxiliaries: the modal auxiliaries or
modals, on the one hand, and the non-modals be, do and have, on
the other.
There are nine modals, or modal auxiliaries, i.e. can, could, may,
might, will, would, shall, should and must. In addition there are three
verbs that may be treated formally either as auxiliaries or as main
verbs: dare, need and ought to. The modals have only finite forms,
add no -s in third person singular form, and can only be combined
with lexical verbs in the infinitive (without to).
Among the non-modals, be and have express different kinds of
aspect. Be is followed by a present participle to express progressive
aspect as in You are/were/have been comparing apples and oranges, it
seems. Have is used to express perfect aspect and is followed by the
past participle form of a lexical verb as in e.g. She has/had promised to
attend. The third non-modal auxiliary—do—combines with a fol-
lowing to-less infinitive in direct questions and in negative clauses
containing the negative adverb not.
Auxiliaries occasionally take part in the word-formation process
known as conversion (cf. Chapter 6). A has-been is a person who is

no longer capable of doing what he/she used to be good at, a must

is something regarded as necessary, and a wanna-be (want to be) is
somebody who aspires to be somebody he/she is not.
Determiners10
As we saw in the previous discussion of the noun, one of the char-
acteristics of nouns in English is the fact that they may be preceded
by determiners. The determiners give information about definiteness
and indefiniteness, quantity and proportion. The basic determiners
are the definite and indefinite articles a/an and the, but certain
other words and constructions may also function as determiners,
chiefly possessive and demonstrative pronouns like my, your, this,
those, and quantifiers like some and any. Nouns with an apostrophe
genitive also function as determiners, as in Jessica’s coat, London’s
underground.
Prepositions
Prepositions are words that typically occur before a noun, like beside
in beside Henry, beside the building, beside the new building. Certain
prepositions are part of prepositional verbs, like e.g. on in decide on
something, while others are not, for example on in stay on the ground.
Prepositions are simple or complex. Common simple prepositions
are about, after, against, before, beside, between, by, in, near, on, to,
with. Certain simple prepositions have homonyms that are con-
junctions, for instance after and before which are prepositions in
after/before her departure but conjunctions in after/before she had left.
Many simple prepositions also have homonyms that are adverbs,
for example around and in: in We walked around the house, and They
dropped it in the water, around and in are prepositions, but in We
walked around, They just dropped in, they are adverbs.
Complex prepositions are multi-word constructions that have
fused to form units with prepositional function by means of a proc-
ess known as grammaticalization11, for instance according to, in addi-
tion to, in case of, out of, with regard to, in front of, (in) back of.
Occasionally new prepositions are also formed from single parti-
ciples, for example considering and given, both of which have prepo-

sitional force in constructions like e.g. I have nothing to add consid-

ering your request and Given his interest in expensive cars, it comes as
no surprise that he has bought a Mercedes SUV.
Conjunctions
Conjunctions are of two kinds: co-ordinating and subordinating. Co-
ordinators are used to link any two (occasionally more) units that
have the same syntactic status. There are three main co-ordinators:
and, or and but.
The subordinating conjunctions introduce subordinate clauses
with different syntactic functions. Many conjunctions consist of a
single word, but many are made up of several words, for instance as
if, so that, in order that. These complex conjunctions obviously have
their origin in the same grammaticalization process that was
responsible for the formation of complex prepositions.
Grammaticalization is like meaning extension (cf. p. 156) in
being a slow process that may take hundreds of years to be com-
pleted. As a result, speakers are usually unaware of the origin of com-
plex units and of the fact that on-going change is forever present in
the language of today. However, sometimes we can see the begin-
nings of the formation of a complex unit. A fairly recent change of
this type in English concerns the prepositional phrases on the basis
(of) and in terms of in sentences like He’s asked for special treatment on
the basis he’s been with the firm over twenty years and They’re a general
nuisance in terms of they harass people trying to enjoy the park12
Exercises
1 What is a word class? What kinds of criteria are used when
words are assigned to different word classes? Why is the seman-
tic—or ‘notional’- criterion a problem?
2 What is the difference between content words and function words?

Why is it possible to leave out function words in simplified lan-
guage? What would happen to a text if the content words were
left out, but the function words were kept?

3 In this chapter it is claimed that there is a difference between

the central and the peripheral members of a word class. If that is
correct, certain nouns should be ‘nounier’ than others. Can you
think of any examples of this?
4 Explain the difference between proper nouns and common nouns,

countables and uncountables.
5 In the table on p. 36, the nouns anger and salt both have plusses
in the column ‘Combines with determiners’. However, there is
one determiner they cannot combine with—which one?
6 What different recategorisations are found among the English

nouns? What regular meaning changes are they accompanied
by?
7 Try to think of arguments for and against counting recategorisa-

tion as a kind of word-formation.
8 Given the criteria for adjectives given here, how do you suggest
that we handle items like upper class (upper-class), Sydney and
1930s in phrases like e.g. a very upper-class accent, a Sydney street,
a 1930s atmosphere. How many of the criteria for adjectives do
they meet?
9 Why are the adverbs so difficult to characterise in terms of a set

of common characteristics? What different kinds of adverbs are
there?
10 Explain what is meant by grammaticalization. The examples

given in the text all concern phrases made up of prepositions
and content words which have ‘fused’ to become linguistic
units treated as prepositions and in some cases, conjunctions.
Can you think of cases where a content word on its own has
turned into a function word?


3 Outline of English Word-formation
3 Outline of English
Word-formation
Defining word-formation
As we noted in Chapter 1, many English words are simple, i.e. they
consist of a single base morpheme: dog, door, smile, mahogany.
Others contain two or more morphemes, for instance, uncertain,
rider, teapot. Of these three, the first two words are complex. Such
words are formed by the addition of an affix—a prefix or a suffix—
to a stem which is itself a word. In the first word, the prefix un- has
been added to the stem certain, and in the second the suffix -er has
been added to the stem ride. The third word—teapot—consists of
two words combined to form a third—a so-called compound.
The three words above have been formed in accordance with
present-day English word-formation rules, principles for the pro-
duction of new words. They represent three types of regular English
word-formation, i.e. prefixation, suffixation and compounding. There
are also a number of less regular types of English word-formation,
among them initialisms (FBI, asap), clippings (para, demo), blends
(Bollywood, edutainment), back-formations (to backpack, from back-
packer, to laze from lazy), rhyming slang and reduplicative forma-
tions (argy-bargy, higgedly-piggedly). These latter will be dealt with in
the final chapter of the book.
Although both prefixes and suffixes belong to the family of
affixes, they have different functions and produce different results.
Prefixes modify the meaning of a word from a certain word class, but
don’t normally change its word class membership: we may add un-
to the adjective certain to create the new word uncertain, but the
new word is still an adjective1. Suffixes, on the other hand, usually
change the word class of the word they are added to: ride is a verb,
but rider is a noun.

Suffixes and prefixes that can be used to form new words are
called derivational affixes, the words produced in this manner are
known as derived words, and the process of forming new words by
the addition of prefixes and suffixes is called derivation. Deriva-
tional suffixes should be distinguished from the inflectional suffixes
discussed in Chapter 1 (p. 22), for instance the -s and the -ed in She
lives here and She lived here two years ago and the -s in The dogs
barked. The inflectional suffixes don’t form new words, but merely
add grammatical categories like tense, number, person to existing
words2.
The inflectional suffixes also differ from the derivational ones in
being used with great regularity: there are few exceptions to the rule
that the past tense of verbs is formed by adding the suffix -ed, that
third person singular forms end in -s, and that plural nouns also
end in -s. The majority of the derivational suffixes do not show the
same regularity, although individual derivational suffixes may
show almost the same regularity as the inflectional ones, in partic-
ular the adverb-forming suffix -ly (cf. Chapter 7).
Compounding differs from affixation in involving the combina-
tion of (usually two) words to form a new word. However, com-
pounding has one thing in common with prefixation: the first ele-
ment is a modifier: it modifies the second element, also known as
the head. In a word like teapot, for example, tea modifies the head
pot, just as e.g. re- in rewrite modifies the meaning of write. In neither
case is there a change in the word-class of the head/stem: like pot,
teapot is still a noun, and like write, rewrite is still a verb.
Prefixation, suffixation and compounding can be characterised as
‘additive’ types of word-formation, since their output is the result of
the addition of something to an already existing word. English also
has another highly productive type of regular word-formation, one
that forms new words without formal change. This type of word-
formation is known as conversion (alternative names are functional
shift and zero derivation).
Conversion is very common in English, especially in the case of
the noun and verb categories; according to one source3, ‘there is no
English noun that can’t be verbed’, i.e. in principle all nouns can be
transformed into verbs. Thus we can carpet a room, doctor a drink and
cash a cheque. We can even steamroller others into accepting what
they don’t want, and we can audition for a role in a play.

In all these cases of conversion it is easy to see that it is the nouns

that are the basic forms from which the verbs have been derived. In
other cases the conversion has clearly worked in the other direc-
tion: a jump is the result of or an instance of jumping, etc. However,
there are also many words that seem to be equally at home in two
or more word classes, which makes it difficult to tell which is the
basic form and which the derived. Examples of such words are love,
hate and fear.
Affixation, compounding and conversion are all regular word-for-
mation processes in the sense that it is generally possible to write
rules for them and to predict the nature of the output of the rules
fairly adequately. It is not always possible, on the other hand, to
specify what restrictions have to be placed on the rules to prevent
them from producing non-acceptable output forms. Nor is it always
possible to predict what kind of prefix or suffix a given individual
output word will use.
We can sum up the above outline of regular modern English
word-formation like this:
REGULAR ENGLISH WORD-FORMATION
ADDITIVE NON-ADDITIVE
DERIVATION COMPOUNDING CONVERSION
PREFIX SUFFIX
uncertain composer teapot doctor(v), carpet(v), steamroller(v)
The output of word-formation rules

An important thing to grasp about word-formation rules is how
they relate to the words that make up their output. To illustrate this
point let us consider an example involving suffixation. I suggested
above that by adding the suffix -er to the verb ride, we produce ‘a
new noun rider’. This is a fairly accurate description of the workings
of the rule for creating so-called agent nouns in -er, i.e. nouns identi-

fying the ‘doer’ of a verbal action or process. (The -er suffix may also
be used to form words denoting instruments like e.g. washer ‘wash-
ing machine’ and silencer).
On the other hand, the noun rider is obviously not new in the
sense ‘never produced before’. It has been in the vocabulary of Eng-
lish for hundreds of years. What the rule does is to provide a
description of the internal structure of derived words that may or
may not have been formed already. If the word resulting from the
application of the rule has been stored in the dictionary, then the
rule can be used to analyse that word. Thus word-formation has
both a productive and an analytic function.
The productive and analytic functions of word-formation are of
equal importance; it might even be argued that we have occasion to
analyse existing complex and compound words more often than we
have occasion to create new words. At the same time analytic
knowledge helps us both to create and to understand new forma-
tions created on a familiar pattern. Once we know that e.g. refusal,
bewilderment and readable are formed by the addition of the suffixes
-al, -ment, and -able to the verbs refuse, bewilder and read, we also
know how to analyse other words containing these suffixes.
Word-formation vs. word manufacture

The presentation above emphasised the rule-governed nature of
word-formation. However, certain types of word-formation are
clearly more rule-governed than others. It is only fair to say that the
word-formation rules form a hierarchy, with some very productive
and reliable rules at the top, and certain rather unpredictable rules
at the bottom. What they all have in common, though, is that they
use already existing words as input and produce ‘new’ words as
their output.
This makes word-formation fundamentally different from word
manufacture. Word manufacture does not operate on already exist-
ing words, but invents new words from scratch. The only rules that
word manufacture has to follow are the principles for permitted
sound/letter combinations in English (sometimes called the phono-
tactic rules of the language)4.

Thus if you are planning to market some new technical gadget

and are wondering what to call it, you will be well advised to avoid
e.g. *srink, * thlang, and *mitr, which all contain phoneme combina-
tions that are not allowed in English, i.e. sr-, thl- and -tr. The alter-
natives srink, thang and mirt may not bring immediate economic
success, but at least they are phonotactically correct.
Probably a fair number of manufactured words have been and are
being created by speakers and writers all the time. But the sheer
amount of work involved in ‘marketing’ new words is so forbidding
that the majority never get into a dictionary. Below you will find a
few of the manufactured English words that did make it to the dic-
tionary.
Quite a few names of products are—or appear to be—manufac-
tured, for example Kodak, claimed to have been invented from
scratch by George Eastman in 18885, and Prozac from the 1990s.
Among the common nouns there is the mathematical term googol,
defined as ‘a cardinal number equivalent to ten raised to a hun-
dred’6. Slang has its fair share of manufactured words, for instance
the words dweeb and nerd used especially in US English to denote a
foolish or contemptible person.
Determining what is and what is not a manufactured word is not
always easy. In a few cases we know that someone expressly decided
to invent a new word. This is the case with the verb grok ‘under-
stand intuitively’ invented by the writer Robert Heinlein, who used
it (about a Martian) in his novel Strangers in a Strange Land, and the
scientific term quark ‘very small unit of matter’, which first
appeared in James Joyce’s Finnegan’s Wake.
In other cases it is hard distinguish between manufactured words
and different types of shortening, such as abbreviation and clipping
(cf. Chapter 9). A case in point is the recent word pharm ‘a place
where genetically modified plants or animals are grown or reared’,
which is perhaps best regarded as a case of back-clipping from phar-
macology. It is also an instance of rather clever word-play on the
established word farm—a place that is also connected with growing
and rearing.
British English has a long word-manufacturing tradition, particu-
larly manifested in so-called ‘nonsense verse’ made popular by
19th-century poets like Lewis Carroll and Edward Lear. One of the
best known nonsense poems is the Jabberwocky, which first

appeared in Carroll’s Through the Looking Glass in 1871. Here is the

beginning:
The Jabberwocky
Twas brillig and the slithy toves
Did gyre and gimble in the wabe
All mimsy were the borogroves
And the mome raths outgabe
In this poem, Carroll cuts a few corners by retaining function words

like was, were, and, the, and by using certain well-known English
affixes like out- and -y, but the rest of the vocabulary is all ‘non-
sense’, i.e. word manufacture: brillig, slithy, tove, gyre, gimble, wabe,
mimsy, borogrove, mome, rath, outgabe. (Twas is an older short form
for ‘it was’).
Word-formation rules and the dictionary

Before bringing this presentation of English word-formation to an
end, we need to take a look at the relation between word-formation
rules and the dictionary. Dictionaries contain many—but certainly
not all—the words in a language. Since the basic function of the
word-formation rules is the ‘production of new words’, it may seem
natural to view them exclusively as devices for adding new words to
the dictionary. But in actual fact it is often difficult to predict which
complex or compound words are going to end up in the dictionary
and which ones are not7.
However, it is possible to point to certain factors that seem to be
involved in such matters. To begin with, the question whether a
word should be listed in the dictionary or not obviously has some-
thing to do with how common that word is: by and large, a com-
plex or compound word that is frequently used is more likely to end
up in the dictionary than one that is used seldom.
That notwithstanding, frequency cannot be the only explana-
tion: manner adverbs like e.g. angrily, beautifully, slowly, nouns in
-ness formed from adjectives like e.g. slowness, gentleness, common-
ness, and compounds like food scare, story-telling, rebel-supported are
seldom if ever found as head words in dictionaries. Yet all three
types of word are common enough in all kinds of texts.

We can go some way towards explaining the absence of such

words from the dictionary by adding a second explanatory factor,
i.e. predictability of meaning. Complex words like angrily and slow-
ness, and compounds like food scare, whose meanings are predicta-
ble because they can be worked out from a knowledge of their com-
ponent parts, and of the word-formation rule involved, tend not to
be found in dictionaries. It would be uneconomical to put such
words in the dictionary, since once we know the meaning of their
parts, we can work out the meanings of the full complex/com-
pound words.8
On the other hand, complex/compound words whose meanings
we cannot figure out on the basis of their component parts and the
word-formation rules involved have to be entered as headwords in
dictionaries. Two examples among many of such words are propeller
and wheelchair.
As we know, there is an English verb propel meaning ‘drive for-
wards’, as in e.g. the canoe was propelled by a paddle. There is also a
suffix -er that can be added to verbs to form nouns denoting both
agents and instruments. On the basis of that knowledge we might
be tempted to conclude that a propeller is any device or person driv-
ing something forward.
But as the dictionary definitions of propeller in The New Oxford
Dictionary of English and The American Heritage Dictionary indicate,
that conclusion would be wrong. The word propeller is defined as ‘a
mechanical device for propelling a boat … consisting of a revolving
shaft with two or more broad, angled blades attached to it.’ In other
words, propeller is the label for a very special category of objects and
such a labelling function is characteristic of complex words entered
as headwords in dictionaries. There is no way we could have figured
out the meaning of such a word simply by looking at its parts.
The same lesson can be learnt from the compound wheelchair.
Just like propeller, wheelchair contains a lot of information that is not
available if all we know is the meanings of the parts of the complex
word. Given the meanings of wheel and chair and the common
compounding pattern used to form it, we might be forgiven for
believing that a wheelchair is just any chair on wheels, like e.g. an
office chair. But as we know, that is incorrect: a wheelchair is ‘a
chair built on wheels for an invalid or disabled person, either
pushed by another person or propelled by the occupant’

The non-predictability of meaning in words like propeller and

wheelchair is sometimes accounted for in terms of the concept of
lexicalization: words with non-predictable meanings are said to be
(semantically)lexicalized.9 However, the terms lexicalize and lexicali-
zation have been used with a number of different meanings and will
not be used in this book.
The distinction between those complex words that have predict-
able meaning and those that do not is complicated by the fact that
what appears to be the same word may turn up in either category. A
case in point is the agent (and instrument) nouns formed from
verbs by the addition of the suffix -er like e.g. walker, singer, climber.
Basically, these nouns mean simply ‘person who walks’, person who
sings’, ‘person who climbs’. These meanings are predictable: if we
know the meanings of walk, sing, climb and -er and the word-forma-
tion rule for agent nouns, we can figure out the meanings of the full
complex words without help from the dictionary.
However, the same nouns in -er may also have additional mean-
ings indicating habit, profession, hobby and the like. Thus if we say
e.g. She is a singer, the noun singer does not mean merely ‘person who
sings’ but ‘professional singer’, and in the same way He is a walker
means ‘He is in the habit of walking’, and They are climbers means
‘They climb (mountains/tall buildings) as a hobby/professionally’.
With such extra semantic features, the meanings of these nouns are
unpredictable, which leads to their being listed in dictionaries.
The distinction between words with predictable and unpredicta-
ble meanings is often accompanied by a distinction having to do
with function. Words with unpredictable meanings are by and large
used with the function of labels: they label categories of phenom-
ena that are regarded as essential in society, for instance propellers
and wheelchairs.
Words with predictable meanings, on the other hand, are often
used with the function of syntactic repackaging10 As such they do
not label categories at all, but are used to simplify texts by replacing
longer and more complex syntactic phrases, often a noun followed
by a relative clause. This function is best illustrated by an example.
Let us assume that there is a text that runs like this:
In the house next door, someone was singing. The person who was singing
was our neighbour’s daughter.

This text could be made simpler and easier to read by syntactically

repackaging the noun phrase the person who was singing in the sec-
ond sentence as the singer. As a result, the text would be changed to
In the house next door, someone was singing. The singer was our neigh-
bour’s daughter.
Most agent nouns in -er (and other agentive suffixes) can be used
both in labelling and in syntactic repackaging, for instance driver,
which is found both in He works as a driver (labelling) and The driver
of the car was Harry Smith (syntactic repackaging). However, some of
the agentive (instrumental) nouns can only be used in syntactic
repackaging, presumably because there is no natural category that
they could be used to label.
It is, for example, perfectly possible to turn the verbs assert, nod
and pat into the agentive nouns asserter, nodder and patter, but they
will never appear in the dictionary for the simple reason that we
have no use for them as category lables. There are no categories of
people that need to be labelled by such nouns: that is why the fol-
lowing exchange sounds extremely strange:
A: What does your sister do?
B: She is an asserter/a nodder/a patter
But as the three examples below indicate, these nouns work per-
fectly well in syntactic repackaging:
Some people asserted that the problem could easily be solved. The asserters
were all males from the south.
When I asked my question, several people in the audience nodded. The most
enthusiastic nodder was Peter Wright.
There used to be people patting my back after my talks. But this time there
wasn’t a single back-patter.
The three sentences above demonstrate the workings of syntactic

repackaging with agentive nouns in -er. Syntactic repackaging can
also be brought about out by means of other types of word-forma-
tion, for example the words in -ly and -ness mentioned earlier. Thus
manner adverbs in -ly, like calmly in e.g. He spoke calmly, are used
instead of longer constructions like He spoke in a calm manner.
Nouns in -ness, like e.g. stubbornness in e.g. Her stubbornness sur-

prised us, are used instead of longer expressions like The fact that she
was stubborn surprised us or The degree to which she was stubborn sur-
prised us.
The syntactic repackaging function is also open to words formed
by means of prefixation, compounding and conversion. Thus
preschool in e.g. preschool teacher refers to a particular stage in the
education of children and is an instance of labelling. But in e.g. pre-
school activities, the likely interpretation is that this is a case of syn-
tactic repackaging meaning for instance ‘activities that take place
before school starts’.
As for compounds, the noun show-stopper is common as a label
and is defined in The New Oxford Dictionary of English as ‘a perform-
ance or item receiving prolonged applause’ (referred to below as
show-stopper (1)). However, if on a given occasion somebody actu-
ally stops an on-going show, the syntactic repackaging show-stopper
(show-stopper (2)) could perfectly well be used with reference to that
person (The show-stopper turned out to be a tall, side-whiskered Austral-
ian youth’).
It is less easy to find convincing examples of the labelling-syntac-
tic repackaging contrast with regard to conversion. However, some-
thing like syntactic repackaging is found with a verb like carpet
(‘provide with a carpet’), for example in an exchange like The room’s
got a new carpet.– Really, who carpeted it? This should be contrasted
with the verb carpet meaning ‘reprimand severely’ as in Sidney was
severely carpeted by his boss, obviously a kind of labelling.
In the rather long discussion above, I have suggested that the
question whether a complex, compound or converted word has to
be memorised and accordingly listed in the dictionary or not, is
mainly decided by the word’s predictability of meaning and its
function (labelling vs. syntactic repackaging). The argument is
summed up in the following figure:

COMPLEX/COMPOUND/CONVERTED
WORDS IN ENGLISH
Meaning predictable Meaning unpredictable

Used in syntactic Used to label categories
repackaging
Usually not in the dictionary In the dictionary
Examples
Slowly, gentleness, singer propeller, wheelchair
(‘person who is singing’) singer (‘person who
asserter, nodder, patter sings professionally’)
pre-school (activities) preschool (teacher)
carpet(v) ‘provide with a carpet’ carpet(v) ‘reprimand’
Exercises
1 Which of the following words are complex and which are com-
pound?: reclassify, unbelievable, water crisis, Londoner, sun-tanned,
(the) Bushies, arms cache, Clintonian, Clintonomics, steamroller.
2 Conversion from noun to verb may affect a wide range of

nouns, even nouns with suffixes that are clearly noun-indicat-
ing, like steamroller and audition. What does that tell us about
the relative strength of conversion and suffixation?
3 Mention is made here of words like love, hate, fear for which it is
difficult to indicate which word class is the original one and
which the converted. Can you think of other words for which
this is also true?
4 p. 51: The tree diagram indicates that English has two kinds of
affixation: prefixation and suffixation. Logically, there is a third
type, i.e. infixation, the insertion of a morpheme inside a word.
Does English have any infixed words?
5 On p. 53 in this chapter there are examples of sound/letter

sequences that are not allowed in English words, like word-ini-
tial sr- and thl-, and word-final -tr. Can you think of other such

‘forbidden’ sequences? What would a rule for permitted word-

initial consonant combinations look like?
6 What are the ‘productive and analytic functions’ of word-for-

mation mentioned on p. 52?
7 Pp. 53–54: Do you know any other manufactured English words

besides the ones mentioned here? Does the Jabberwocky poem
suggest any particular thoughts or feelings to you?
8 What factors determine whether a ‘new word’ will turn up in

the dictionary or will be used and forgotten? Can you think of
other factors in addition to the ones mentioned here?
9 What is the point of discussion of propeller and wheelchair on

p. 55?
10 Explain the notions labelling and syntactic repackaging and try to

find additional examples of these phenomena.

4 Prefixation
4 Prefixation
Characteristics of prefixes
As their name indicates, prefixes are bound affix morphemes occur-
ring word-initially, i.e. at the beginning of words. In addition to
their initial position, prefixes have certain additional defining char-
acteristics. One of these is that they are semantically special,
expressing a fairly narrow range of clear meanings of which the
main ones are: NEGATION, REVERSAL, REMOVAL, MANNER, DEGREE/SIZE,
ATTITUDE, LOCATION/DIRECTION, and TIME/SEQUENCE.
Prefix meanings
With the noticeable exception of DEGREE/SIZE, each of the meaning
categories above is represented by a fairly limited number of pre-
fixes. In addition it is often the case that one or two of the prefixes
linked to each meaning are dominant and occur more or less freely
with stems with certain characteristics, while the remaining ones
are infrequent and of unpredictable occurrence. The following list
shows the distribution of prefixes across the prefix meanings listed
above.
Meanings Prefixes
NEGATION 5
REVERSAL 3
REMOVAL 3
MANNER 2
DEGREE/SIZE 14
ATTITUDE 2
LOCATION/DIRECTION 4
TIME/SEQUENCE 5

4 Prefixation
The table above does not claim to present the total number of pre-
fixes or prefix meanings in English. What it does contain, arguably,
are those prefixes and prefix meanings that are often found in texts
of a general kind. Texts of a more specialised nature like e.g. scien-
tific/ technical texts use many prefixes not accounted for here.
There are three important cases of homonymy among the Eng-
lish prefixes discussed here, i.e. cases where the same spelling and
pronunciation represents separate prefixes with clearly different
meanings, namely un- expressing the meanings negation, reversal
(of the action of the following verb), and removal, dis- with the
same three meanings, and de- which may express both reversal and
removal. The table below summarises this information and pro-
vides examples of how the prefixes are used.
un- dis- de-

Negation unkind disbelieve, dishonest –
Reversal untie disaffirm decentralise
Removal uncork discourage defrost
As the table indicates, negative un- combines only with adjectives,

negative dis- with both adjectives and verbs. (The prefix de- does
not express negative meaning). Reversal meaning is expressed by
both un-, de- and dis-. It is a meaning that is possible only in combi-
nation with a following verb (and nouns formed from such verbs,
like decentralisation). Verbs with such reversal prefixes express the
reverse of the process or action they normally denote. Thus, if you
untie a knot, disaffirm a previous decision, or decentralise the run-
ning of an office, you reverse the actions of tying, affirming and
centralising.
Prefixes that express removal are attached to a following noun
and indicate that the referent of the noun is removed: when you
uncork a bottle you remove the cork from it, when you defrost the
fridge you take away the ice in it. Dis- is less productive than the
other two prefixes but can be found in certain genuinely English
formations like e.g. disburden1

4 Prefixation
Prefix function
Prefixes typically have a modifying function: they serve to modify
the following stem rather than change it radically. Unlike the suf-
fixes (cf. Chapter 5), prefixes as a rule don’t change the word class of
the stem they are attached to, nor the general semantic category to
which it belongs. Thus if we add the prefix re- to the verb stem fill,
the resulting combination refill is still a verb, and still represents the
semantic category of filling.
There are certain exceptions to this rule, however, cases where
prefixes do not have purely modifying function, but actually
change the word class of their stems. Among these we find the
removal prefixes de- and un- discussed in the previous section,
which are added to nouns and turn them into verbs meaning
‘remove whatever the noun stem refers to’. Accordingly, defrost and
delouse mean ‘remove frost/lice from’ and unroof and unsaddle
mean, respectively, ‘remove the roof (from a house)’ and ‘remove
the saddle from’(usually a horse). Unsaddle may also mean ‘remove
(the rider) from the saddle’ and is usually used about horses that
throw their rider. Another exception is the prefix out- in combina-
tions like outclass, outdistance, outwit. Here the addition of out- to
the nouns class, distance and wit results in new words that are verbs.
The modifying function of most prefixes makes them similar to
the first element in the so-called neo-classical compounds like e.g.
biography, economy and technology (see Chapter 8 for a presentation
of these). Bio-, eco- and techno- are so-called initial combining forms,
i.e. bound Latin and Greek base morphemes that are unlike prefixes
in that they usually only combine with other bound Latin/Greek
morphemes, i.e. the so-called word-final combining forms like
-graph(y), -nomy and -logy. But in certain cases bio-, eco- and techno-
combine with ordinary English words, as in e.g. bio-terrorist, eco-
crime and technopeasant.
On account of this similarity in both function and combining
habits it is sometimes difficult to distinguish between initial com-
bining forms and prefixes: certain of the forms listed as prefixes
below may arguably also be regarded as initial combining forms.
However, it is normally possible to distinguish between prefixes
and initial combining forms in terms of meaning. As has already
been pointed out, the meanings of prefixes belong to a restricted set

4 Prefixation
to do with matters like negation, time, attitude, size, degree and a

few others. The initial combining forms, on the other hand, may be
about anything: life (bio-), the stars (astro-), the environment (eco-),
etc.
The neo-classical compounds are not the only constructions that
may complicate the delimitation of prefixes from other forms. On
occasion, it is also difficult to distinguish prefixed words from ordi-
nary English compounds in which the first element is a particle—a
preposition or adverb that has been preposed, i.e. placed in initial
position, as in verbs like back-date, download, input, outsource, over-
write, update, adjectives like ongoing and uplifting, and nouns like
downslide, input and uptake. Are these words derived words made up
of a prefix and a stem, or are they compounds (cf. Chapter 7) made
up of a particle and another word?
There is no obvious answer to that question. However, it has been
suggested2 that certain constructions with preposed particles are
more prefix-like than others, i.e. those in which the particles have
meanings far removed from their usual meanings. This happens in
particular with the particles out-, over- and under-, for example in
constructions like e.g. outrun, outgun ‘run faster than’, ‘have more
guns/troops than’, overeat and overachieve ‘eat too much’, ‘achieve
too much’ and underachieve, underpaid ‘achieve too little’, ‘paid too
little’. In this book, out-, over- and under- with such meanings have
been regarded as prefixes.
Prefixes in English and in Latin

The majority of the English prefixes come from Latin, sometimes as
direct borrowings, sometimes via the intermediary of Old or Middle
French3. Certain prefixes have been borrowed from Greek (and
some of the Latin forms ultimately represent borrowings from
Greek). Only a minority of the prefixes are native—i.e. Germanic—
in origin. The following list contains all the prefixes of foreign ori-
gin discussed in the present chapter: a-, anti-, arch-, co-, de-, dis-,
ex-, extra-, hyper-, in-, inter-, macro-, mal-, mini-, non-, post- pre-, pro-,
re-, semi-, sub-, super-, trans-, ultra. The list of native prefixes is quite
short and consists of mis- (as in e.g. mismanage), over-, out-, un-,
under- (Note that the lists above do not reflect the existence of

4 Prefixation
homonymous prefixes: un- stands for both negative, reversative and

removal un-, etc.).
Despite their foreign origin, the prefixes borrowed from Latin and
Greek are now part and parcel of English; they are bound English
morphemes with clear meanings and combine with stems that are
English words. It is important not to confuse these forms with
forms with the same spelling that only exist in Latin words
imported wholesale into English.
Usually the difference is clear from the form’s meaning and abil-
ity to combine with English words as stems: if we compare e.g.
amoral and anomaly, anti-American and antipathy, ex-emperor and
except in these two respects, it is obvious that only the first word in
each pair contains an English prefix: a- in amoral has the meaning
‘negative’ and combines with the English word moral, while a- in
anomaly has no clear meaning and does not combine with an Eng-
lish word. Likewise, anti- in anti-American and ex- in ex-emperor have
clear meanings and combine with the stems American and emperor,
both of which are words, while antipathy and except meet neither
the semantic nor the combining requirement.
In many cases the English prefixes and the Latin initial forms are
also distinguished by a third factor, i.e. a difference in pronuncia-
tion and stress. The distribution of stress in English is a highly com-
plicated matter, which will not be gone into here4. All we need to
know is that a distinction is made between the heaviest stress in a
word, known as primary stress and the next heaviest which is called
secondary stress.
Many of the English prefixes with Latin look-alikes are set apart
from the Latin forms by the fact that the the English prefixes carry
secondary or even primary stress, while the Latin look-alikes are
unstressed. Thus the forms a-, anti-, and ex- discussed above tend to
have secondary (sometimes even primary) stress when used as Eng-
lish prefixes and to be pronounced [e], [nt] ([nta]) and [eks],
as for example in amoral, anti-American and ex-emperor. The corre-
sponding Latin strings, on the other hand, are unstressed and are
pronounced [ə], [n t] and [ks] for instance in anomaly [ə nɒməli],
antipathy [n tpəθi], and except [k sept].
The same type of distinction can be made for e.g. de-, pre-, re-:
compare the pronunciation of e.g. delouse [di las], pre-fabricate
[pri fbrket] and refill [ri fl] with that of deceive [d siv] prepare

4 Prefixation
[pr peə], and regret [r ret]. In the first three examples—which con-
tain the English prefixes de-, pre- and re- —the letter e is pronounced
[i] and has secondary stress, but in the last three examples—which
do not contain English prefixes—e is unstressed and has the pro-
nunciation [] or even [ə] (as when regret is pronounced [rə ret]).
Degrees of productivity among prefixes

In the preceding discussion, certain characteristics of English pre-
fixes have been mentioned. In particular it was pointed out that in
order to be counted as a prefix, a word-initial form must have a
clear meaning from among a rather narrow set of meanings. Typi-
cally it should also have clearly modifying function with regard to
the following stem, which must be an English word i.e. not a bound
morpheme. Obviously, it must also recur with the same meaning
often enough to make us want to regard it as a prefix at all.
As we have seen, these criteria help us establish a substantial
number of English prefixes. However, nothing has been said so far
about the productivity of these prefixes. Defining ‘productivity’ is
not an easy matter. The approach adopted in the present work is
close to that recommended in the recent Cambridge Grammar of the
English Language, and operates in terms of relative degrees of pro-
ductivity like ‘highly productive’, ‘fairly productive’ and ‘of low
productivity’5
A highly productive prefix is one that combines more or less
freely with a large class of well-defined stems. Such classes of stems
are usually defined both in terms of word class and meaning, for
instance the class of stems to which we can add the prefixes anti-,
pro- and ex-. The first two may be added freely to nouns denoting
something to which it is possible to have a negative or positive atti-
tude (which is, basically, any noun). Obvious members of this
group are nouns like abortion, war, Bush, EU, prohibition. The prefix
ex- ‘former(ly)’ is also attached to nouns, but in this case the nouns
must denote a position, status or condition that may be terminated,
for example king, husband, wife, President, alcoholic, cancer patient.
Compare also the stylistic effect that can be achieved by violating
the rule, for instance the famous ex-parrot of Monty Python fame.6

4 Prefixation
The claim made for the highly productive prefixes is that they
may be attached to all or at least the majority of stems with certain
characteristics. However, for the majority of prefixes it is indeed
possible to characterise the stems they prefer, but it is impossible to
claim that they combine with all or even the better part of such
stems. Reversative de- is a good example of this. It combines fairly
freely with verbs from Latin, especially those ending in -ate, -ify, -ise
as for example in decentralise, decontaminate. But it certainly doesn’t
combine with all such verbs; thus we find neither *detranslate,
*deglorify or *desymbolise. Prefixes like reversative de- may be called
‘fairly productive’.
Although it is a distinction difficult if not impossible to maintain,
we may want to establish a further category of weakly productive pre-
fixes that combine on an occasional basis with a certain type of
stem. A likely candidate for this group would be negative in-, which
sometimes combines with ultimately Latin stems as in inaudible,
incomplete (and with variation impossible, irregular but are impossi-
ble in others. Thus only e.g. unable, will do, never *inable, despite
the fact that the corresponding noun is inability.
Common English prefixes

Negative prefixes: non-, un-, a-, in-, dis-
There are two highly productive negative prefixes, i.e. non- and
un-. The prefix non- combines freely with nouns, adjectives and
open-class adverbs: non-smoker, non-starter, non-Muslim; non-Swed-
ish, non-trivial, non-clinical; non-trivially, non-naturally, non-sexually.
It implies a distinction between phenomena that are members of a
class and those that are not. The use of the hyphen with non- varies;
thus we find both e.g. non-smoker and nonsmoker.
The negative prefix un- combines freely with adjectives and parti-
ciples (and nouns formed from them, e.g. uncertainty): uncertain,
uncommon, uneven, unfair, unkind; undamaged, undefeated, unfinished,
unrivalled, unstructured; unconvincing, unsmiling, unyielding.
Unlike non-, un- does not combine with nouns not derived from
adjectives or participles, and as a result, nouns like e.g. *unbuilding/

4 Prefixation
un-building, *undog/un-dog, and *un-starter /unstarter etc. are not

found.7
With adjectives, un- denotes the opposite of the stem meaning.
This makes it different from non-, which merely denies that the
meaning of the stem is present. The semantic difference between
the two prefixes may be illustrated by means of the pair un-Ameri-
can activities: non-American activities. The former means ‘activities
opposed to America and American interests’ while the latter means
simply ‘activities that are not American’. (Note that certain very
common adjectives have their own opposites and cannot be com-
bined with any of the negative prefixes. Among the most common
of these we find good:bad, deep:shallow, big:small, strong:weak)8.
There are a number of less productive negative prefixes that
sometimes compete with the two above, in particular with un-: the
relatively learned a- [ei] combines with adjectives (amoral), and so
does in- (variant forms il-, im-, ir-) as in illegal impossible, inaudi-
ble, irregular, and dis- as in dishonest, disloyal, dissimilar. Dis- also
combines with verbs and nouns as in disobey, disorder, disrespect,
distrust.
Reversative prefixes: de-, dis-, un-

In addition to having negative meaning, the prefixes dis- and un-
also express reversal of the action or process denoted by a following
verb; a third reversative prefix is de-. Un-is by far the most common
reversative prefix. It combines freely with verbs that denote change
of state, and signals the reversal of the action or process denoted by
the verb. Familiar examples are undo, untie, unwrap, unzip. The lim-
its to the productivity of reversative un- are normally set by our
experiences of what can and what cannot be reversed: it is normally
difficult to unread a book, for example. However, in genres like sci-
ence fiction, such restraint is no longer necessary: in such texts it
may turn out that time travel makes it possible both to unmurder,
unmarry and uncreate other people9.
The reversative prefix de- is claimed to be particularly productive
with verbs ending in the suffixes -ate, -ify, -ise/ize, like decentralise,
declassify, decontaminate. Reversative dis- is unevenly distributed

4 Prefixation
across non-Germanic stems as in e.g. disengage, disinherit, disestab-

lish.
Removal prefixes: de-, dis-, un-

In addition to expressing reversal, the prefixes de-, dis-, un- may also
express the removal of something. They then combine with nouns
to form transitive verbs expressing removal, i.e. they belong to the
limited group of prefixes that change the word-class of the stem
they are added to. Un- is as usual the most common prefix. With
removal meaning, this prefix is word-class-changing: it is added to
nouns to form transitive verbs meaning ‘remove what the noun
denotes’ as in unleash ‘remove the leash from (a dog)’, unstrap
‘remove the straps from’ unsaddle (a horse), unroof (a house) and
many others.
But complex words formed by the addition of un- to nouns may
also have the meaning ‘remove somebody/something from what
the noun denotes’, for instance in unhinge a door ‘remove a door
from its hinges’, unsaddle a rider ‘throw a rider’, and unseat the pres-
ident ‘remove the president from power’.
The prefix de- has removal meaning in a limited number of com-
binations with nouns, as in e.g. debark (‘remove bark from tree’),
defrost, de-ice, delouse, debug, descent (a skunk), detick (‘remove ice/
frost/lice/bugs/scent/ticks from refrigerators/animals’). We can
notice here one characteristic that all the ‘removal prefixes’ have in
common, i.e. that the nouns they are attached to should be some-
how naturally (inherently) possessed by their owners: trees naturally
have bark, animals naturally have lice, bugs, ticks (cf. discussion of
inherent possession in Chapter 5). In support of this we may also
note that debark can have the additional meaning ‘remove a dog’s
bark by disabling its vocal cords’.
Very occasionally, de- is found in combinations that are less easy
to interpret in that way, as in dedog the premises10 However, we may
note in support of our hypothesis above that it is found in a phrase
like degender a text, where it is in all likelihood assumed that all texts
have male gender bias by nature.
The removal prefix dis-, finally, is like de- in having limited pro-
ductivity; it is found in e.g. disambiguate, disinfest, dislodge.

4 Prefixation
Manner prefixes: mal-, mis-

There are two semantically distinct but not very productive manner
prefixes, i.e. mal- ‘badly’ and mis- ‘wrongly’, ‘astray’. They are
found in combinations with verbs, past participles and abstract
nouns: malcontent, malformed, malfunction, malodorous, malpractice;
misdirect, misfire, misconduct, misapprehension. Combinations with
mal- and mis- are usually not hyphenated.
Degree/size prefixes: arch-, extra-, hyper-, macro-,

mega-, micro-, mini-, out-, over-, semi-, sub-, super-,
ultra-, under-
Of the prefixes in this category, several like e.g. extra-, hyper-, super-
have homonyms that are so’called ‘initial combining forms’. As
explained in Chapter 8, prefixes and initial combining forms differ
both in terms of meaning and stress assignment. The prefixes have
clear meanings and take either primary or secondary stress in all
combinations. The initial combining forms lack clear meanings and
are stressed in accordance with the stress rules for non-Germanic
words (cf. p. 146). Compare for example extramural [ekstə mjυərəl],
which contains the prefix extra-, and extravagant [k strvəənt],
where extra- is an initial combining form. (The word extramural
‘outside the walls’ is a term used about university courses not aimed
at regular students).
arch- ‘supreme, out-and-out, of the worst kind’

Combines freely with negative meaning with usually human nouns
with meanings permitting of degrees: arch-hypocrite, arch-believer,
arch-capitalist, arch-champion arch-conservative, arch-dealer, arch disci-
ple, arch-enemy, arch-Fascist, arch-federalist, arch-fixer, arch right-
winger, arch survivor, arch-theatricality, arch-traditionalist, arch wet; a
wet is a ‘conservative with liberal tendencies’). The prefix arch- usu-
ally takes primary stress.
extra- ‘highly’, ‘unusually’

Combines fairly freely with both Latin and Germanic adjectives as
in extra-high, extra-late, extra-long, extra-sensitive. Occasionally it also

4 Prefixation
combines with nouns: extra-funding, extra-length. The use of the

hyphen appears to be optional. See also the section on locative pre-
fixes.
hyper- ‘extreme(ly), excessive(ly)’

Hyper- combines fairly freely with gradable adjectives, as in e.g.
hyperintelligent, hyperreal, hypersensitive. With the meaning ‘very
large, large-scale’, it also combines with nouns, as in hyperinflation,
hyperconsumption, hyperintelligence. As a rule combinations with
hyper- do not take a hyphen. The prefix hyper- seems to take second-
ary stress when attached to adjectives—as in hypersensitive, hypercrit-
ical—but has variable stress in combinations with a following noun.
Thus is e.g. hyperlink, hypermarket it has primary stress, but in forms
like hyperinflation and hypertension primary stress falls on the second
syllable from the end of the word.
macro- ‘large, large-scale’

Macro- is originally a combining form as in e.g. macrocosm (see p.
150) but is treated here as a prefix combining with nouns with a
certain amount of productivity. We find it both in technical lan-
guage (macro-economics, macro-climate) and to a certain extent in
ordinary language (macrochange, macrocontract, macroscale). As a
prefix, macro- has primary stress and variable use of hyphens.
mega-: ‘large, outstanding’; ‘very’

In technical language this prefix means ‘1 million’ for example in
megabyte, megahertz, megadeath. In ordinary (particularly informal)
English, mega- is a productive prefix meaning ‘(very) large, out-
standing’. It combines freely with nouns as in e.g. mega-acquisition,
mega-brains (‘very intelligent people’), mega-contributions, megaflop,
mega-hit, mega-merger, mega-musical, mega-restaurant. It also com-
bines with adjectives, for instance in mega-rich, mega-strong, mega-
sloshed (‘very drunk’). Most combinations with mega- are hyphen-
ated. Mega- carries primary stress. (Mega may also be used as an
independent word, for instance in mega takeover, mega losses and
even e.g.: The concert was mega ‘very good’.)

4 Prefixation
micro- ‘small, of reduced size’

Like macro-, micro- is originally a combining form (cf. p. 63) which
has come to be used also with independent forms, but micro- seems
to be more productive than macro-. It is found in e.g. micro-engineer-
ing, micro-opera, micro-organism, micro car, micro fault, microwave. Pri-
mary stress on the prefix and hyphenation seem to be the rule, but
there is variation in this respect.
mini- ‘little, minimal’ (informal)

combines freely with countable nouns: minibar, minicab, minicom-
puter, mini-conference, mini-skirt, mini conference. The use of hyphens
is variable and the prefix often has primary stress.
out- ‘better than’

This is quite a productive prefix that turns intransitive verbs into
transitive as in outdo (somebody), outperform (somebody), outrun
(somebody), out-think (somebody), outvote (somebody). It also com-
bines with nouns, turning the resulting complex word into a tran-
sitive verb taking a personal direct object as in e.g. to outwit some-
body, outclass somebody, outdistance somebody (‘leave far behind’).
Occasionally it combines with adjectives, as in outsmart (somebody)
‘get the better of somebody’.
over- ‘too (much)’

The prefix over- combines freely with verbs and adjectives: overeat,
overestimate, overreact, overemphasize, overdo; overconfident, over-ripe,
overanxious. As a rule, over- is not followed by a hyphen.
semi-: ‘partly’
combines fairly freely with adjectives with gradable meaning, for
instance in semi-conscious, semi-naked, semi-open, semi-permanent,
semi-predictable, semi-public, semi-rotted, semi-secure, semi-skilled,
semi-tough. The prefix is usually hyphenated. In scientific language
there is also non-hyphenated semi meaning ‘half’ as in e.g. semi-
vowel, semicircle.

4 Prefixation
sub- ‘below’; ‘secondary’; ‘below the norm’; ‘at a lower level’

The prefix sub- has a number of related meanings. It combines fairly
freely with nouns when it has one of the meanings ‘below’ (subsoil,
subway), ‘secondary’ (sub-editor, subdean), ‘subordinate part (of)’
(subcommittee, subcategory, subclass, subculture, sub-plot). It also com-
bines with adjectives with the meanings ‘below’ (subconscious) and
‘below the norm’ (subhuman, substandard). With nouns, sub- tends
to have primary stress, as in e.g. subassembly, subgroup, subsection,
subtotal. The prefix sub- also combines with verbs with the meaning
‘at a lower level’, as in e.g. sub-contract, sub-edit, sublet, sub-under-
write, subdivide, sublease. As the examples indicate, the use of
hyphenation with sub- varies.
super- ‘more than the norm’; ‘highly, extremely’; ‘very large, prom-
inent’
Super- is a very productive suffix that combines with adjectives. In
words like e.g. super-confident, super-elevated, super-natural, supersen-
sitive it indicates that the quality denoted by the adjectives is
present to a greater extent than is normally the case. In most cases,
however, this prefix simply means ‘highly, extremely’, for instance
in super-bland, super-civilised, super-clean, super-cool, super-fit, super-
strong, super-smart, super-talented.
In combination with nouns, super- has meanings like ‘very large,
unusually prominent, powerful, etc’ as in e.g. super-G, super-clerk,
super-nerd, super-tanker, super-state.(Super may also be used as a word
of its own meaning ‘excellent’, ‘very good’ as in It was a super con-
cert, The concert was super.)
ultra- ‘extreme(ly)’
combines fairly productively with adjectives (and certain nouns)
meaning ‘extreme(ly)’. Examples are ultra-conservative. ultra-fair,
ultra-marginal, ultra-modern, ultra-nationalist, ultra-right-winger, ultra-
slim, ultra-Tory. The use of hyphenation varies.
under- ‘insufficiently, too little’

This prefix combines freely with verbs and -ed participles for
instance in under-achieve, underbid, undercharge, undercook, underesti-
mate, underexpose, underplay, under-spend, under-staff, understeer,

4 Prefixation
underpaid, under-qualified. Combinations with under- without a

hyphen seem to be the most common.
Attitude prefixes: anti-, pro-

The only two productive attitude prefixes are anti- and pro-. They
combine freely with adjectives and nouns, typically names of per-
sons and organizations and nationality adjectives. They are usually
hyphenated and carry primary stress.
anti- [ nt], [ ntai] ‘against, opposed to’

The most common meaning of anti- is ‘against, opposed to’. With
that meaning, the prefix may be attached to adjective stems, as in
anti-clerical speech, anti-British attitudes, but it appears to be more
common in combination with noun stems, as in e.g. anti-abortion
laws, anti-boxing demonstration, anti-hooligan measures, anti-terrorist
forces, anti-war pronouncements. When attached to nouns as in the
second set of examples above, anti- changes the word class of the
stem from noun to adjective. Thus the examples in the second set
above may be used in typical adjectival positions, as in e.g. His atti-
tude was very anti-abortion, The mood of the country is highly anti-war.
All the anti- words above tend to place the heaviest stress on the
second element.11 Anti- may also be used on its own, as in She’s pro-
Bush, but he is very anti.
pro- ‘in favour of’, ‘for’:

combines freely with adjectives and nouns:, pro-boxing, pro-British,
pro-Bush, pro-European, pro-family, pro-German, pro-life, pro-Nazi, pro-
socialist, pro-war. Like anti-, pro- added to a noun stem changes its
word class to adjective, cf. very pro-family, extremely pro-war.
Location and direction prefixes: co-, extra-, inter-,

trans-
co- ‘joint(ly), on equal footing’
This prefix combines fairly freely with agent nouns and with verbs
as in e.g. co-pilot, co-drive(r), co-conspirator, co-write(r), cooperate/co-

4 Prefixation
operate, co-exist, cohabit. The use of hyphens with co- varies. It tends
to take primary stress before nouns.
extra- ‘outside’
The prefix extra- combines freely with Latin-based adjectives, partic-
ularly those ending in -al, -ial, -ic, -ar, like extracontractual, extracur-
ricular, extragalactic, extragovernmental, extralinguistic, extramarital,
extraterrestrial, extraterritorial.
inter- ‘between’, ‘among’

Inter- combines freely with Latin-based adjectives ending in -al, -ial,
-ic, -ar with or without a hyphen: inter-continental, international,
interurban, interstellar. In some cases inter- combines with nouns as
in intercity train traffic, interstate highways.
trans- ‘across’
The prefix trans- combines with certain adjectives of Latin origin
derived from geographical nouns: transatlantic, trans-Siberian, trans-
pacific, trans-pennine. The use of hyphens varies.
Time and sequence prefixes: ex-, post-, pre-, re-

ex- ‘former’
This prefix combines freely with nouns and noun phrases denoting
titles, office or status; the following are all recent attested examples:
ex-boy friend, ex-cavalry officer, ex-Chief Rabbi, ex-fiancé, ex-husband,
ex-king, ex-lover, ex-major of horse, ex-president, ex-wife. The prefix
may even occasionally be modified by an adverb, as in e.g. a newly
ex-girlfriend12. Ex is also a word in its own right meaning ‘former
husband/wife’. Most combinations with ex- use a hyphen. The pre-
fix often takes primary stress.
neo- ‘new’, ‘modern’

The prefix neo- is added freely to nouns and adjectives denoting
phenomena that may be regarded as variants of earlier phenomena.
It is usually combined with words denoting schools of thought and
political movements as for instance in neo-Nazism/Nazi, neo-coloni-

4 Prefixation
alism/colonialist, neo-Gothic, neo-conservative, neo-Napoleonic. An

exception from this is the term neo-natal ‘new-born’ found in med-
ical language, where it may be a combining-form (cf. Chapter 8).
post- ‘after (in time)’

post- combines freely with all kinds of nouns denoting a period or
point in time. These combinations are used attributively as in e.g.
post-Budget (discussions), post-Christmas (shopping), post-election
(period), post-match (analysis), post-privatisation (woes), post-war
(period) and also e.g. post-September 11 events, a new post-flotation low
for the Argentine currency13 In such combinations, post- is usually fol-
lowed by a hyphen. The prefix also combines with Latin-based
adjectives in -al, -ial, -ic, -ar: post-classical, post-doctoral, post-natal,
post-industrial, post-millenial, postimpressionistic.
pre- ‘before’; ‘in advance’

Like post-, pre- combines freely with nouns denoting a period of
time. The combinations are used attributively, as in pre-war (period),
preschool (training), pre-sale (offer), pre-tax (value), pre-cancer (treat-
ment). This prefix is also highly productive with the meaning ‘in
advance’ before verbs in technical English: prebuild, precook, prefab-
ricate, preheat, pre-publish, pre-stress. The prefix pre- takes heavy stress
in the nouns pre-school ‘nursery school’ and prefab ‘prefabricated
building’. Pre- also combines with adjectives of Latin origin as in
e.g. pre-colonial, pre-linguistic, pre-conscious, premarital. The use of
hyphens with pre- varies.
re- ‘once more, again, anew’

The prefix re- combines fairly freely with verbs: rewrite, rebuild, rean-
alyse, re-enter, re-engineer and with nouns from these: reanalysis,
rewriting, re-entry etc. The verbs in re- are usually transitive and/or
denote a process ending in a result of some kind. The prefix occa-
sionally has the meaning ‘back’ in e.g. recall. Like most prefixes, re-
does not take primary stress. However, verbs in re- may be con-
verted to nouns with accompanying stress shift: thus students are
usually given the opportunity to retáke (British English resít) an
exam. The corresponding nouns have primary stress on re- as in
When is the rétake (résit)?

4 Prefixation
Exercises
1 In what way are the ‘removal’ prefixes de- and un- different from
most other prefixes?
2 Explain how productivity in prefixes is defined. What character-

istics should a prefix have to be considered fully productive?
How do you define prefixes that are fairly productive and
weakly productive?
3 The largest class of prefixes with similar meanings is that of

degree/size. Many of these prefixes are highly productive.
Search a text—or even better a computer-based corpus of
present-day English like e.g. the British National Corpus (BNC)—
for instances of the degree/size prefixes mentioned here, and try
to form an idea of the productivity of the different prefixes in
this group.
4 What is the difference in meaning between negative un- and

non-? What would combinations like e.g. non-person, non-dog,
non-book mean?
5 What other adjectives besides big, deep, good, strong have oppo-
sites not formed by the addition of a prefix?
6 It is claimed in this chapter that the prefixes mega-, mini-, out-

and overall combine freely with certain classes of stems. Can
you think of cases in which such combinations between mega-
etc. and their respective stem classes would be improbable or
even impossible?
7 The prefixes super- and ultra- are both claimed to mean

‘extremely’. Does that mean that they can always replace each
other? If not, why not?
8 According to the rule given here, the prefix co- combines fairly
freely with agent nouns. However, there is obvious a difference
between the common co-pilot, co-driver, co-producer, the less
common but possible co-runner, co-singer, co-believer and the
highly unlikely *co-liker, co-sitter, co-snore. Try to account for
these differences.

4 Prefixation
9 The examples in Chapter 4 are all words containing a single pre-

fix added to a stem. Try to find examples of prefixed forms that
begin with two (or more) prefixes.
10 How do the meanings of prefixes and initial combining forms

differ?

5 Suffixation
5 Suffixation
General characteristics of suffixes

One of the major types of present-day English word-formation is
derivational suffixation, a process that makes it possible to create new
words from already existing ones by adding a derivational suffix (cf.
Chapter 3). Normally, the word class of the original word (the stem)
is changed in the process of suffixation, and the suffix tells us what
word class the new word belongs to. Thus there are noun-forming
suffixes like -er, -ness and -ity (runner, kindness, formality), adjective-
forming suffixes like -ish, -y and -like (foolish, dirty, doglike), verb-
forming suffixes like -ize and -ify, (Americanize, simplify), and basi-
cally just one adverb-forming suffix, i.e. -ly (slowly).
Derivational suffixes are also particular in their choice of stems:
with certain exceptions, a given suffix combines only with stems
belonging to a certain word-class. Thus a noun-forming suffix like
-er meaning ‘agent’ or ‘instrument’ can only be added to verb stems
(runner, washer), an adjective-forming suffix like -ish meaning ‘hav-
ing the characteristics of the stem’ may only be attached to noun
stems (foolish), and the adverb-forming suffix -ly accepts only adjec-
tives as stems (slowly). On the other hand, the verb-forming suffix
-ize combines both with adjective and noun stems, as the examples
legalize and Londonize show. (The suffix -ize has a variant spelling
-ise common in British English).
In addition to selecting stems along word-class lines, many deri-
vational suffixes are also sensitive to the origins of potential stems.
Thus there is a clear tendency for certain suffixes to select stems of
Germanic origin, while others equally clearly prefer words of classi-
cal origin, i.e. words coming originally from Latin or Greek. The
noun suffixes -dom, and -hood, for example, never combine with
classical words, nor would -ation, -ity and -ive select Germanic words

5 Suffixation
as stems. Other suffixes are less particular: -able, -er and -ness are
found with both kinds of stems.
Suffix meanings
Derivational suffixes usually have two kinds of meaning. To begin
with, they all signal that the word they are attached to belongs to a
certain word class: words ending in e.g. -hood, -ion or -ness are
bound to be nouns, words ending in -ize and -ify are verbs etc.
(However, as described in Chapter 6, the process of conversion may
sometimes override such word-class assignments, cf. verbs like to
commission, to position, and to station).
In addition to signalling word-class membership, derivational
suffixes have other, more or less clear meanings: the -er used to form
nouns from verbs as in e.g. runner, reader etc. has the meaning
‘agent’ or ‘instrument’, the -y added to nouns to form adjectives as
in e.g. dirty, sandy, snowy carries the meaning ‘full of/covered with’.
There is also a fair amount of homonymy among derivational suf-
fixes i.e. suffixes may be formally identical but have different mean-
ings (and different combining habits). Thus in addition to the
familiar agent/instrument suffix -er mentioned above that can only
be attached to verbs, there is another suffix -er which is added to
nouns and has the meaning ‘person associated with what the stem
denotes’ as for instance in New Yorker, jet-setter, villager, brat packer.
We also have to recognise several different suffixes of the form -y.
The y-suffix meaning ‘full of/covered with’ has already been men-
tioned. Another adjective-forming y-suffix attached to nouns has
the meaning ‘resembling’, ‘having the characteristics of’, for exam-
ple in bossy, catty, powdery. A third carries the meaning ‘character-
ised by’, as in e.g. ballsy, gutsy ‘characterised by courage’, trafficky
‘characterised by traffic’. (In addition there are -y/-ie suffixes used in
nouns like aunty/auntie, lefty/leftie; cf. p. 91 below).
Degrees of productivity among suffixes

As in the case of the prefixes (cf. p. 66) the most important charac-
teristic of a productive derivational suffix is the ability to combine

5 Suffixation
freely with stems of a certain, definable class. A 100 % productive

suffix would be one about which we could say that it combines
with all the stems of category X and with no other stems.
Unfortunately, there are no suffixes like that. One that comes
close is the -ly suffix added to adjectives to produce adverbs, as in
e.g. gladly, quickly and importantly, but as our discussion later in this
chapter will show, not even this suffix is totally dependable in this
respect. Other suffixes that come close to the ideal are the noun-
forming -ness and -er found in e.g. kindness and writer.
At the other end of the scale, we find word endings that are still
recognisable as suffixes, but which are dead when it comes to the
formation of new words. The form -th is an example of a dead suffix
in English. This originally Old English suffix—whose meaning is
roughly the same as that of modern -ness, —survives only in a
handful of words, for example breadth, depth, length, strength,
warmth, width. It cannot be added to new stems (although on occa-
sion playful forms like e.g. coolth occur).
Between these two extremes, there is a range of levels of produc-
tivity that are difficult to pinpoint exactly. As in the case of the pre-
fixes, suffix productivity will be described in terms of notions like
‘highly productive’, ‘fairly productive’ and ‘weakly productive’.
Members of the first group will be said to combine freely with certain
stems, those in the second to combine fairly freely with certain
stems, etc.
Suffixation and word stress

Derivational suffixation in English often affects the stress pattern
and the pronunciation of the new words. In this section, a brief
account of the interplay between suffixation and word stress will be
given1.
When said in isolation, each English word has a single heavy or
primary stress. In words with a single syllable—monosyllabic words—
the stress obviously falls on that syllable. In polysyllabic words—
words with more than one syllable—one of the syllables is selected
as the bearer of primary stress.
Since there are many types of polysyllabic words, it would be
highly useful to have a rule or a set of rules predicting the location

5 Suffixation
of primary stress in such words, and there have also been many
attempts to establish such rules. This has turned out to be a very
challenging task, however, and none of the attempts so far have
been entirely successful.
Nevertheless, if we restrict our attention to derivational suffixa-
tion, certain general principles for stress position may be found.
These principles operate in terms of the type of suffix added. Basi-
cally there are three types of suffixes in this respect: those that don’t
change the stress pattern at all, those that take primary stress them-
selves, and those that require primary stress to fall on the syllable
immediately preceding the suffix. Examples of these types and how
they are used are given below.2
(1) Suffixes that don’t affect the stress pattern (stress-neutral suf-
fixes)
The most important members of this group are -able, -dom, -er, -ess,
-ie, -ing, -ish, -ism, -ist, -ise/-ize, -like, -ly, -ment, -ness and adjective-
forming -y. Examples: understandable, officialdom, researcher, steward-
ess, auntie, panelling, monkeyish, Americanism, Africanist, computerise,
statesmanlike, evidently, arrangement, unfriendliness, spidery. In certain
words ending in -ess primary stress goes on the suffix. The word
stewardess, for instance, may be pronounced both [ stjυədəs] and
[stjυə des].
(2) Suffixes that are themselves stressed
Representatives of this type include -ation, -ition, -ution, -ee, -eer,
-esque, -ette, -ese, (note that -ation, -ition, -ution and -eer count as sin-
gle suffixes here), -itis. Examples of words with these suffixes are
confirmation, definition, resolution, referee, auctioneer, kitchenette,
Kiplingesque, Japanese, appendicitis (cf. also nonce formations like
creditcarditis). Words ending in -ee like e.g. divorcee sometimes take
primary stress on the syllable immediately before the suffix.
(3) Suffixes requiring primary stress on the preceding syllable
A rather small number of suffixes require that the primary stress go
on the syllable immediately preceding the suffix. In stems that
already have primary stress on their final syllable, the stress remains
where it is (as in e.g. intense: intensity); in other stems it is moved to
the syllable preceding the suffix as in rapid:rapidity.

5 Suffixation
The most common of the members of the third group are -ial,
-ian, -ic, -ical, -ify, -ity. Examples: atomic, sequential, Clintonian, per-
sonify and readability. In many—but not all—of its occurrences, the
adjectival suffix -al belongs here too, for instance in fundamental
and adjectival. Note that moving the primary stress in words like
these often causes radical pronunciation changes: compare Clinton
[ klntən] and Clintonian [kln təυnən], and note how -able [əbəl]
changes to -abil- [ə bl] before -ity, as in e.g. countability, dependabil-
ity, readability.
Although it is usually attached to bound rather than to free
stems, mention should also be made here of the suffix -ion. English
has many words ending in -ion, all of which place primary stress on
the syllable immediately preceding the suffix, for instance prohibi-
tion, invasion, permission, depletion, derision (For the spelling of the
words in -ion, cf. p. 84).
It should be emphasized again that the above account of the
interplay between suffixes and stress cannot claim to be compre-
hensive and is based on general patterns rather than absolute rules.
By and large, these principles provide valuable and reliable guide-
lines for the stress patterns of suffixed words. However, certain com-
plications have been left unaccounted for. In addition, there are
sometimes individual exceptions to the principles.
Suffixation and stem change

A number of English verbs containing bound Latin base mor-
phemes change their spelling and pronunciation before certain
Latin suffixes3. A verb like permit, for example, changes to permiss-
when followed by one of the suffixes -ible, -ion, or -ive. What we get
is the forms permissible, permission, permissive.
Permit is not the only verb subjected to these changes: they affect
all English verbs containing the bound Latin base morpheme -mit,
for example admit, omit, submit, transmit. On account of this, these
changes are best described in terms of the bound Latin bases
involved, as in the account below. (Note that not all verbs combine
with all suffixes, however. Note also that the combination of -ss-
and a following suffix -ion is pronounced [ʃən]).

5 Suffixation
(1) Verbs containing -mit change final t to ss before -ible, -ion, -ive,
Examples: admit, admissible, admission, admissive; permit, permissi-
ble, permission, permissive; omit, omissible, omission.
(2) Verbs containing -cede, -ceed, -fend, -hend, -pand, -pend. -spond,
-tend change final -d(e) to -(s)s before -ible, -ion, -ive and -or.
Examples: concede, concessive, concession; succeed, successive, succes-
sion, successor; defend, defensive, defensible; comprehend,
comprehensive, comprehension, expand, expansive, expan-
sion; expend, expensive; respond, responsible, responsive;
extend, extensive, extensible, extension.
(3) Verbs containing -cide, -clude, -lide, -lude, -plode, -ride, -suade,
-vade, -vide change the final -de in the spelling to s before -ion
and -ive.
Examples: decide, decisive, decision; include, inclusive, inclusion; col-
lide, collision; allude, allusive, allusion: explode, explosive,
explosion; deride, derisive, derision; persuade, persuasive,
persuasion; invade, invasion; divide, divisive, division.
Note that the letter sequence -sive in the words in (3) above is pro-
nounced [sv], but that -sion is pronounced [ən]. Decisive, inclusive
are accordingly pronounced [d sasv] and [n klusv], but decision,
inclusion are pronounced [d sən] and [n klυən].
(4) Verbs containing the stems -solve and -volve change these to
solut- and -volut before the suffix -ion. Final -t followed by -ion is
pronounced [ʃ]
Examples: dissolve, dissolution; resolve, resolution; revolve, revolution
(5) Verbs in -ceive change final -ve to -pt before -ible, -ion, -ive, -or.
Final -t followed by -ion is pronounced [ʃ]
Examples: deceive, deceptible, deceptive, deception, ; receive, receptive,
reception; conceive, conception.
(6) Verbs in -duce change -duce to -duct before -ible, -ion, -ive, -or. The
combination -duction is pronounced [ dkʃən]
Examples: reduce, reduction, reductive; produce, production, produc-
tive; conduct, conductor

5 Suffixation
(7) Verbs in -pel change -pel to -puls before -ive, -ion. The combina-
tion -pulsion is pronounced [ plʃən]
Examples: compel, compulsive, compulsion; repel, repulsive, repulsion
expel, expulsion
Changes also occur before the suffixes -ive and -ion in the words
repeat, compete, expose, inquire, oppose, recognize as in repetition, repet-
itive, competition, competitive, exposition, expositive; opposition; inquis-
itive, inquisition; recognition.
In addition, stems ending in the common verb suffix -ify changes
-ify to -ific- before the suffix -ation, as in classify:classification, iden-
tify:identification, etc.
Noun-forming suffixes
Nouns from nouns
A number of noun suffixes combine with noun stems, thus violat-
ing the principle that suffixation should always be word-class-
changing. At least seven types of meaning may be distinguished:
AMOUNT, COLLECTIVES, ACTIVITY CONNECTED WITH STEM, STATE OF BEING
WHAT THE STEM DENOTES, PERSON CONNECTED WITH STEM, SIZE/SEX/
STATUS, FAMILIARITY.
AMOUNT: -ful
The suffix -ful is added fully productively to nouns to express the
meaning ‘the amount contained in what the stem denotes’. It is
attached to nouns denoting containers as in They poured bucketfuls /
spoonfuls/glassfuls of water onto the table. While the nouns in these
words are all prototypical containers, the notion of what is a con-
tainer may be extended considerably as the following examples
show: fistful, handful, mouthful, plateful, sinkful, shovelful.
COLLECTIVES: -ing, -ry

Collectives—as defined here—are collections of things of the same
type. In English, such nouns are formed by attaching the suffix
-ing to the relevant noun stems. Thus from the nouns carpet, panel,

5 Suffixation
mat, scaffold and tube we form the collectives carpeting, panelling,

matting, scaffolding and tubing. This is a fairly productive suffixa-
tion rule.
On a much smaller scale, the suffix -ry may be added to nouns
denoting different technical gadgets as in gadgetry itself and also in
rocketry, weaponry. The productivity of this suffix is limited.
ACTIVITY/STATE CONNECTED WITH WHAT THE NOUN STEM DENOTES:

-ing, -age, -(e)ry, -ism, -itis
Activity nouns are usually formed from verbs (cf. p. 93–95). How-
ever, in certain case, they are formed from nouns, and are used to
denote activities somehow connected with the meaning of the
noun. Here the suffix -ing is fairly common, as in e.g. black-berrying,
, cricketing, tunnelling, back-packing, jet setting which denote, respec-
tively, the activities of playing cricket, digging tunnels, making
excursions wearing a back-pack, and being a typical member of the
jet set.
While the suffix -ing above is used to describe activities in general,
the suffix -ism has a more specialized meaning. This suffix is quite
productive in connection with nouns denoting doctrines of various
kinds: Buddhism, Marxism, cubism, dadaism, capitalism, socialism,
and even America-Firstism, me-too-ism, flat-Earthism. (Many such
nouns in -ism have a corresponding personal noun/adjective in -ist,
but the correspondence is not perfect (cf. discussion on p. 90). The
-ismsuffix has recently also come to be used to describe various
types of discrimination, as in ableism (‘discrimination against the
disabled’), ageism discrimination against the elderly’, racism, sexism.
The suffix -ism has also come to be used with a more concrete
meaning, i.e. ‘(typical) saying by a public figure’. With this mean-
ing, -ism can apparently be freely added to any personal name, as
for instance in Bushism, Blairism, Clintonism, Churchillism, Thatcher-
ism, Brownism, Hitlerism, etc.
The suffixes -age and -ery are often added not just to noun stems,
but also to verbs and occasionally to adjectives. I comment here
only on their use with noun stems.
The suffix -age may be added to a noun to denote activity,
amount or place. The activity meaning is predominant in e.g. bro-
kerage ‘the activity of acting as a broker, i.e. someone who sells

5 Suffixation
assets and goods for others’, parentage, patronage ‘the activity of

being a parent/a patron’. The amount meaning is found in dosage
‘the amount of medicine to be given to a patient’, mileage ‘the
number of miles travelled’, percentage ‘the amount or number in
each hundred’, and similarly footage, voltage, wattage. Place mean-
ing occurs in e.g. hermitage and orphanage denoting respectively the
dwelling of a hermit and a home for orphans.
The suffix -ery alternates in the spelling with -ry. Like -age it is
used to form nouns denoting activity and place. Examples of activ-
ity meaning are banditry, crookery, dentistry, slavery, thuggery ‘the
activity of being a bandit, a crook, a slave and a thug’, respectively.
Recent formations are back-stabbery ‘the activity of back-stabbing’
and tycoonery ‘the activity of being a tycoon’4.
There are a number of nouns in -(e)ry denoting the place where
certain animals are kept, for instance piggery, rookery, snakery.
The last item in this group of suffixes is -itis, a suffix with the
meaning ‘inflammation or disease of what the stem denotes’, as in
e.g. laryngitis ‘inflammation of the larynx’. As the example indi-
cates, this suffix belongs to the language of medicine, and is nor-
mally attached only to bound stems. However, the suffix -itis has
enjoyed considerable popularity in informal humorous construc-
tions. In such cases it is attached to ordinary English words. In these
cases, -itis has the meaning ‘excessive preoccupation with or use of
what the stem denotes’ for example in creditcarditis, deadline-itis, jar-
gonitis, televisionitis.
The suffix also occurs with related but slightly different mean-
ings: football teams have been accused of suffering from Wembley-
itis—nervousness caused by the fact that they are playing at Wem-
bley Stadium (probably in a Cup Final). According to one sports
writer in The Independent 1992, teams could even suffer from Man-
chesteruniteditis, defined as ‘the inability to score easy goals’ The
reader may wish to ponder what other attested examples like Eng-
landitis and Westminsteritis mean.5
THE STATE OF BEING WHAT THE STEM DENOTES: -dom, -hood, -ship
The suffixes -dom, -hood and -ship are all added to nouns to form
new nouns denoting the state of being what the stem noun
denotes. Often the very same nouns may also be used to express

5 Suffixation
‘collectivity’, i.e. to refer to the people who are in such a state.

Examples of state meaning in -dom words are apedom ‘the state of
being (as stupid as) an ape’, billionairedom, hippydom, martyrdom,
officialdom and the more unusual topsy turvydom ‘the state of being
topsy turvy’, i.e. in a mess. Hippydom and officialdom may also be
used with collectivity meanings, and so may e.g. (football) fandom
‘people who are football fans’. In dukedom and earldom the nouns in
-dom denote the rank or territory of a duke or an earl. Despite occur-
ring in certain recent words like fandom, the suffix -dom must be
said to have low productivity.
Like -dom, -hood expresses either ‘state’ or ‘collectivity’. The state
meaning is found in a number of established words, for instance
childhood, girlhood, manhood, widowhood. This suffix is also been
used in more recent formations, for example citizenhood, hermit-
hood, outsiderhood meaning ‘the state of being a citizen, a hermit, an
outsider’ It occurs with ‘collectivity’ meaning in e.g. anthood,
cousinhood, officerhood6 As is well known, the -hood suffix also occurs
in nouns denoting organized collectivity as when brotherhood
(Brotherhood) and sisterhood (Sisterhood) are used about organiza-
tions, religious orders and the like. In comparison with -dom, -hood
is fairly productive.
The suffix -ship is found in nouns like e.g. apprenticeship, citizen-
ship, dictatorship, sponsorship partisanship, leadership, friendship in
which the suffix has been attached to a noun to denote the state or
quality of being what the noun stem stands for. Thus apprenticeship
basically means ‘the state/quality of being an apprentice’, etc. How-
ever, most of these examples may also denote the activity of being
for instance an apprentice or a dictator, and also a period of time.
Sometimes, ship-words denote a particular skill, as in e.g. She was
praised for her leadership, and on occasion they have ‘collectivity’
meaning, as when we speak of a writer’s readership or a politician’s
followership. Finally, words in -ship may also denote status or title:
chairmanship, kingship, lordship. The suffix has a certain limited pro-
ductivity.
PERSON SOMEHOW ASSOCIATED WITH WHAT THE STEM DENOTES:

-er, -an/ian, -ese, -i, -ite, -ist, -y/ie

5 Suffixation
The suffix -er may be added to names of cities and other geographi-
cal names to refer to people who live in them. The stems may con-
sist of more than one word. Examples of such combinations are e.g.
Londoner, Dubliner, New Englander, New Yorker (but only e.g. Bosto-
nian, San Fransiscan, Roman). The same kind of suffixation may be
used with nouns denoting groups of people with certain aims and
beliefs, as for example in bratpacker, jetsetter, free marketer, America-
firster.
The -an/ian suffix is found in combination with many types of
stems forming both nouns and adjectives. In the past, it was fre-
quently used to name persons and things from a certain country,
but this type of formation can hardly be regarded as productive in
today’s English. In many cases, the suffix was not attached directly
to the stem, but alternates with final -a and -ia as in America:Ameri-
can, Algeria:Algerian (but note exceptions like Canada:Canadian).
Compare also the use of the suffix with names of cities as in San
Fransisco:San Fransiscan, Paris:Parisian, Boston:Bostonian. Note that
here and elsewhere, primary stress in -ian words falls on the syllable
immediately preceding the suffix, while in -an words it stays in its
original place.
In today’s English the productive use of -an/ian is largely
restricted to combinations with personal surnames and results in
words that may be both adjectives and nouns. While -an is found in
certain cases like Elisabethan and Lutheran ‘a person who lived at the
same time as Queen Elisabeth I’ and ‘person who believes in the
ideas of Luther’, the -ian suffix is much more common than -an in
combinations with names7.
Typical examples of nouns in -ian are Chomskyan and Darwinian
‘person who believes in the ideas of Chomsky or Darwin’, Faustian
‘person who is like Faust’, Machiavellian ‘a person who is like Mach-
iavelli in cunning and ruthlessness’. The -ian suffix seems to be
associated with a certain pompousness, which probably accounts
for its use for comic effect in e.g. the recent—obviously rule-break-
ing-formation Star Warians ‘people who like the Star Wars films by
G. Lucas’.8
The suffix -ese is found in many adjectives denoting nationality
or other origin such as Japanese, Milanese, etc. It is also a noun suffix
attached to names of countries, regions, and the like, to create other
nouns denoting the language spoken in that country or region: Do

5 Suffixation
you speak Japanese?, My Vietnamese is a bit rusty. This use of the suffix
-ese is only slightly productive. However, the suffix -ese is fairly
freely used as a noun suffix to denote the jargon of certain writers,
groups or types of text, as in e.g. Johnsonese ‘language typical of Dr
Johnson’, bureaucratese ‘the jargon of bureaucrats’, officialese ‘the
jargon of officials’, teacherese ‘the jargon of teachers’, headlinese,
newspaperese. (Cf. also journalese ‘the language of journalists’,
legalese ‘the language found in legal documents’, and translationese
‘the language found in translations’.)
Nationality is sometimes also expressed by means of the suffix -i,
as in e.g. Israeli, Pakistani, Bengali, Azerbajani, Iraqi, Kashmiri mean-
ing both ‘from Israel, Pakistan’, etc and ‘inhabitant of Israel, Paki-
stan’ etc. Some of these—like e.g. Bengali, Kashmiri—may also be
used with reference to the language spoken in the country. It has
been observed that this type of formation seems to be restricted to
countries in the East or Near-East. Since the emergence of new
nations is a fairly unusual event, it is difficult to have an opinion
about its productivity9.
The suffix -ist is very productive. It is used to denote people who
act in accordance with certain doctrines (Buddhist, socialist, flat-
Earthist) and people who discriminate against others (ableist, racist).
In cases like these, there is commonly a corresponding form in -ism:
a Buddhist adheres to Buddhism, an ableist is guilty of ableism, etc.
However, there are also many -ist nouns with no corresponding
nouns in -ism, for instance those denoting people playing a certain
musical instrument (pianist, violinist) or specialising in certain areas
of study (physicist, psychologist).
As for -ite, it may be found attached to the names of politicians to
denote their followers, as for instance in Blairite, Bushite, Clintonite,
Thatcherite, Reaganite and also to the names of people who are the
originators of a school of thought as in e.g. Chomskyite.
The suffix -y (-ie) is used in informal style and may be added to a
noun stem to denote somebody who is interested in and/or good at
what the noun stem denotes. Thus a foodie is interested in good
food and possibly a good cook, a junkie is a drug addict (junk ‘her-
oin’), a techie/techy is somebody who is an expert on or enthusiastic
about technology’, and a winie a person interested in/knowledgea-
ble about wine(s). The suffix can apparently also be used to refer to

5 Suffixation
persons working for a politician or other public figure as in e.g.

Bushie, Bushies (cf. also also discussion under FAMILIARITY).
SEX, SIZE, STATUS: -ee, -ess, -ette, -let

Suffixes denoting differences in sex, size and status are on the
whole rare in English, but they do exist and their use is sometimes
extended playfully to new noun stems.
In certain words, female sex is expressed by the addition of the
suffix -ess to a noun stem, as for instance in lioness, tigress, actress,
duchess. With the exception of actress, duchess and a few others, the
use of this suffix with stems denoting humans—as in e.g. author-
ess—is disfavoured and regarded as politically incorrect.
The remaining suffixes -ee, -ette and -let are diminutive suffixes, i.e.
they denote small size: a boatee is a small boat, a bootee a small boot,
a kitchenette is a small-sized kitchen, a booklet a small book. Like -ess,
these suffixes have low productivity: the -ette suffix recurs in e.g.
cellarette, luncheonette, tycoonette, undergraduette which all seem to be
the result of linguistic playfulness.
A recent formation in -ette is the informal British word ladette
‘young woman with the same characteristics as a lad’ i.e. a person
who drinks too much alcohol and behaves in a generally uncon-
trolled way’.
Well-known formations in -let besides booklet are piglet and starlet.
FAMILIARITY: -y/-ie
As shown above, the suffix -y/ie may be added to nouns in informal
style to denote people who are associated with what the stem
denotes; among the examples were foodie and winie. An identical
suffix is attached to nouns as a stylistic marker merely to express
familiarity, informality and close community. Thus unlike the pre-
vious -y/ie suffix, this one has no definite meaning besides indicat-
ing familiarity.
It may be added to noun stems as in e.g. doggy (dog), auntie
(aunt), froggy (frog), drinky (drink), girlie, and also in many first
names like e.g. Charlie, Johnny, Teddy. It often operates in conjunc-
tion with so-called clipping (cf. p. 159) as when breakfast becomes
brekkie/brekky, television becomes telly, goalkeeper becomes goalie, and
nightdress becomes nighty/nightie.

5 Suffixation
Nouns from adjectives: -ity, -ness, -ie, -o

There are two competing suffixes here, -ity and -ness, both meaning
roughly ‘the state of being what the stem denotes’as in e.g. falsity,
density, reality and commonness, kindness, toughness10. The suffix
-ness is one of the most productive English suffixes and basically
combines with all adjective stems. It is much more productive than
-ity11; in the examples above, -ness could have been used in all three
-ity examples, but the reverse is not true. This is because -ity gener-
ally favours stems with a Latin or French background, while -ness
has no such restrictions. The suffix -ity is particularly common after
certain other suffixes like -able/-ible/-uble: reliability, reversibility, vol-
ubility.
It has been pointed out12 that we often find pairs of words in -ness
and -ity attached to the same stem whose members differ in mean-
ing, for instance sincereness: sincerity, productiveness: productivity, and
that the forms in -ity have a more institutionalized meaning than
those in -ness. Note also the difference in pronunciation: the addi-
tion of -ness to an adjective stem does not change the position of
the main stress, while adding -ity moves the main word stress to the
syllable immediately before the suffix (cf discussion of suffixes and
stress pp. 82–83).
While noun to adjective suffixation with -ity and -ness is one of
the most important word-formation processes in modern English,
suffixation by means of -o is considerably less significant. It has
been included here since it is after all a word-formation process
with a certain amount of productivity.
The suffixes -ie and -o are attached to adjectives in informal Eng-
lish to form nouns meaning ‘person who has the characteristic
denoted by the stem’. Well-known examples are cutie, ‘cute person’,
smoothie ‘smooth person’, softie ‘emotional person’ dumbo ‘dumb
person’, saddo ‘sad person’, fatso ‘fat person’, weirdo, w(h)acko
‘someone who is weird or w(h)acky’. The suffix -o is occasionally
also found with noun stems, as in wino ‘wine alcoholic’. (Note that
the well-known bimbo is not a suffixed form, but a loan from Ital-
ian).

5 Suffixation
Nouns from verbs

The creation of nouns from verbs—verb nominalization—is a central
function of the noun suffixes. Although the only visible result of
verb nominalization is the addition of a suffix to a verb, such for-
mations may be regarded as abbreviated versions of a noun with a
following relative clause. It is therefore reasonable to paraphrase
the meaning of writer and writing syntactically in the following way:
writer: a person who writes/is writing
writing: activity performed by somebody who writes/is writing
Writer and writing represent the two main types of verb nominaliza-
tions: those involving the ‘doer’ of the action, i.e. the agent or
instrument (writer), and those denoting the activity itself and its
results (writing). There is also a third less common but still active
type, i.e. nominalizations denoting the ‘receiver of the action’—
sometimes referred to as the patient—as in e.g. employee.
The types writer and writing are also very common in combina-
tions like letter-writer ‘person who writes/is writing letters’ and letter-
writing ‘the activity involving the writing of letters’. The noun letter
obviously corresponds to the direct object in the syntactic para-
phrases of these nouns: a letter-writer is someone who writes (pred-
icate verb) letters (direct object), and letter-writing is the activity of
writing (predicate verb) letters (direct object).
Nouns like letter-writer and letter-writing may be analysed in either
of two ways On the one hand it could be argued that the suffixes
-er and -ing belong to the entire noun-verb combination. In such an
analyses we would assume that these words consist of a verb letter-
write and the suffix -er or the suffix -ing. The problem with such an
interpretation is obviously the fact that we have to postulate a non-
existing verb *letter-write.
The alternative analysis—which is the one adopted here—is to
regard letter-writer and letter-writing as noun+noun compounds in
which letter is combined with writer and writing, respectively. This is
a more plausible analysis of letter-writer and letter-writing in that all
the elements involved exist independently of each other: letter,
writer and writing are perfectly normal English words.
However, the compounding approach does not entirely manage
to avoid the difficulties of the first analysis. Thus in words like e.g.

5 Suffixation
grenade-thrower and grenade-throwing, the analysis forces us to

assume that there are nouns like thrower and throwing that can be
combined with other nouns like grenade. In actual fact such nouns
are difficult to imagine on their own: they need to be preceded by
nouns like grenade, javelin, rock etc, which denote objects that are
‘throwable’.
However, the second analysis is clearly preferable from the point
of view that it requires us to postulate far fever ‘odd’ words than the
first. Further discussion of constructions of the letter-writer and let-
ter-writing types will be postponed until Chapter 7 which deals with
compounding (but cf. the comments on sheep-stealer, book-stealer in
the next section.)
AGENT/INSTRUMENT NOMINALIZATIONS: -er, -or, -ant/ent

The typical and by far most common agent/instrument suffix is -er,
as in e.g. writer (agent) and lawn-mower (instrument). This suffix is
extremely productive and can be attached to most kinds of verb
stems. There are exceptions, however, for instance breathe, eat and
start: a breather is not a person who breathes but a brief pause, eater
would hardly ever be used on its own, and a starter is normally a
small dish eaten before the main meal.
Interestingly enough, agentive nouns like breather, eater, etc.
become possible when they are combined with an adjective describ-
ing the way somebody breathes, eats or starts. We can thus speak of
a heavy breather, a big eater and a slow starter. We are not talking here
of a breather who is heavy, an eater who is big etc., but about some-
one who breathes heavily, eats big meals, and starts doing things
slowly. Occasionally there is real ambiguity of meaning, for
instance in beautiful dancer, which could mean either ‘beautiful per-
son who dances’ or ‘person who dances beautifully’.
Occasionally, -er suffixation from verbs is impossible because Eng-
lish already has a word expressing the meaning that the formation
in -er would have had—a phenomenon known as blocking and also
mentioned elsewhere. A textbook example of this is the impossible
noun *stealer ‘person who steals’, whose meaning is already
expressed by the word thief. Note, however, that if we specify the
kind of goods a particular thief specializes in, the form stealer sud-
denly becomes possible, for example in compounds like sheep-

5 Suffixation
stealer and book-stealer. (See pp. 130–31 for a discussion of the prob-
lems this raises for the analysis of compounds).
Although -er is the dominant agentive suffix in English, there are
cases in which it faces competition from other suffixes with the
same meaning, i.e.-ant (assistant, attendant, defendant), -ent (solvent,
repellent, dependent) and -or (actor, escalator, governor). The compet-
ing suffixes are all from the Latin/French part of the vocabulary
and, as the examples indicate, they are generally only found with
stems of the same origin. In fact, most of the words in -ant and -ent
are early loans from French, and many words in –or are loans from
Latin.
ACTIVITY/RESULT NOUNS: -ing, -age, -al, -ation, -ion, -ment

There is considerable variety among the noun suffixes denoting ver-
bal activity and its results. The only truly productive suffix is -ing,
which can be attached to all verbs to form activity nouns, as in e.g.
the lowering of the prices, the company’s sacking of Mr Price. Deverbal
nouns in -ing also often denote the concrete result of an activity, as
in a building, a shooting and a painting but this is a considerably less
productive use of the -ing suffix.
The remaining suffixes above are by and large limited to Latin
and French stems, as in e.g. revival, classification, inhibition, confine-
ment. The only one among them that is fully productive is -ation,
which is regularly used with verbs ending in -ise/ize (realize-realiza-
tion) and also in -ify; the latter changes change -ify to -ific- before
-ation (classify-classification). A variant form of -ation is -ition which
turns up in certain words like inhibition, prohibition.
Neither -age nor -al are productive English suffixes. The verb
stems to which they are attached can in almost all cases also be
nominalized by means of the suffix -ing. However, there is often a
semantic difference between the ing-forms and the forms in -age
and -al. Thus a form like wastage denotes loss through decay, leak-
age, etc, while wasting denotes the activity of wasting in general.
Similarly the noun referral (from refer + -al) is often a medical term
denoting the referring of a patient to a specialist, while referring is
the more general activity term.
Like -age and -al, the suffix -ment can hardly be claimed to be pro-
ductive in modern English, although there are occasional new for-

5 Suffixation
mations like staggerment ‘the activity of being staggered13. In most

other cases, -ment occurs in established words of French origin:
astonishment, accomplishment, achievement, confinement, deferment.
RECEIVER OF THE ACTION: -ee

In comparison with the agent/instrument meaning and the activ-
ity/result meaning, this meaning is formed on a less regular basis. It
is also often stylistically deviant and considered artificial by certain
speakers. Nevertheless it represents the only other possibility for
nominalization and is used fairly frequently in certain written
genres.
The -ee suffix used to identify the ‘receiver of the action’ is found
in a number of established dictionary words like employee ‘person
who is employed’, evacuee ‘person who is evacuated’, lessee, nomi-
nee, and others. But it may also be fairly freely used to create new
words such as the unlikely—but attested—expellees ‘people who
have been expelled from a country’. Even far-fetched examples like
kissee ‘the receiver of a kiss’, murderee ‘a murder victim’, flirtee,
kickee, laughee have been attested.14
Somewhat confusingly, the suffix -ee may also be used with agen-
tive meaning as in absentee, ‘person who is absent’, returnee, referee,
escapee, licensee. According to one source15, this is presently its most
productive use.
Adjective-forming suffixes
Adjectives from nouns
Among the adjective suffixes, those combining with noun stems are
by far the largest group. The suffixes fall into six relatively clear
semantic groups: POSSESSING, LACKING, RESEMBLING, COVERED WITH/
FULL OF, CHARACTERIZED BY, ASSOCIATED WITH/HAVING THE PROPERTIES OF
There is a certain overlap among these groups; in particular it is
not always easy to distinguish between the last two groups. The first
five groups use Germanic suffixes while the final group uses non-
Germanic suffixes (cf. Chapter 1 for ‘Germanic’ and ‘non-Ger-
manic’).

5 Suffixation
POSSESSING: -ed16
There are two kinds of possession. In the first kind, what is pos-
sessed is a natural part of the possessor. This is the possession we are
talking about when we say that dogs have legs, humans have heads,
arms and minds, cars have wheels, etc.17 This kind of possession—
called inherent or inalienable possession—underlies a highly regular
and productive type of English word-formation, exemplified by flat-
footed (person), long-haired (girl), blue-eyed (boy), long-eared (donkey),
white-roofed (house), four-wheeled (tractor). The adjectives in this cat-
egory always contain two words joined by a hyphen, the first mod-
ifying the second. When uttered in isolation, such words often
have a distinctive stress pattern in that the primary stress falls on
the first syllable of the second element: flat-fóoted, blue-éyed, etc.)
In the second—non-inherent—type of possession, what is owned is
not part of the ‘owner’. This is the type of possession that is
involved when we say, for instance, that our neighbour has two cars
or that a student has three books. Possession of this kind can never
be expressed by adjectives of the flat-footed, blue-eyed etc. type: it is
simply not possible to speak of a *two-carred neighbour or of a *three-
booked student, simply because we do not regard cars or books as
natural parts of human beings.
The formation of adjectives expressing inherent possession is one
of the most regular and productive word-formation processes in
English. As the additional—authentic—examples below indicate,
the possessors do not have to be humans, and what is possessed
may be anything naturally associated with the possessor18:
two-legged (animal), different-haired, strong-willed, abstract-minded,
blue-shirted, white-gloved, blue-helmeted, bowler-hatted, different-
skinned, five-sided, three-wheeled (car), four-masted (ship), long-keeled
(sailboat), shoebox-sized (flat), duck-shaped (bottle).
LACKING: -less
Just as there are adjectives denoting the possession of something,
there are also adjectives denoting the lack of something. There is a
single suffix used in creating such adjectives from nouns, i.e. -less.
This suffix combines more or less freely with both concrete and
abstract nouns. The following examples19 illustrate some of the
range of the -less suffix: baseless (charge), breathless, boneless (beef),

5 Suffixation
cashless (society), childless (couple), clueless (‘without a clue’ i.e. lack-

ing knowledge/understanding), jobless, odourless, riderless (horse),
topless (swimming-suit), (scare somebody) shitless/witless, smokeless
(tobacco).
The suffix -less has competition from the form (-)free as in e.g.
hazard-free, internet-free, leadfree and many other words. In this
book, (-)free has been regarded as the second element in adjective
compounds (cf. Chapter 7), but like many other such elements, it
may be on its way to acquiring suffix status.
RESEMBLING: -ish, -like, -ly, -y

All four suffixes above are added to noun stems to create adjectives
meaning ‘resembling what the noun stem denotes’, but have differ-
ent productivity. The most regular among them is -like, which can
in principle be added to any noun denoting humans and animals
and to a good many other nouns as well. In words containing the
suffix -like, the suffix-stem combinations are quite transparent:
words like doglike, childlike, carlike, treelike, etc, may all mean liter-
ally ‘like a dog’, ‘like a child’ etc. Note however, that there is often
a less direct meaning involved, as in e.g. He looked at me with doglike
affection, which means ‘He looked at me with the kind of affection
usually found in dogs’.
The remaining suffixes in this group have less precise meanings
and often add additional features of meaning to the new adjectives.
Thus the meaning of -ish and -y is often said to be ‘somewhat like’,
‘having the characteristics of’, while -ly is said to mean ‘having the
qualities of’. In actual fact it is hard to distinguish these meanings
from one another.
The suffix -ish often combines with ‘negative’ stems as in dandy-
ish, fiendish, foolish, ghoulish, loutish, prudish, snobbish, sluggish, thug-
gish. When the stem nouns are not normally perceived as negative,
the addition of -ish may suggest that they are, as in e.g. mannish,
vicarish. But -ish may also have neutral meaning as in e.g. childish,
girlish, boyish. It also combines with non-human stems, as in clan-
nish, hellish, nightmarish.
The suffix -y combines fairly freely with nouns denoting both
humans, animals and things considered to possess characteristic
features, as in for instance: bossy, catty, foxy, froggy, horsy, mousy, spi-
dery; bricky, creamy, powdery, silvery, stony, tinny. In informal (mostly

5 Suffixation
spoken) language, adjectives in -y are often modified by phrases like

kind of, sort of: it looked kind of resiny/rubbery/coppery. The suffix -ly is
less productive, and appears chiefly in fixed combinations denoting
positive characteristics: brotherly, fatherly, manly, motherly, sisterly,
soldierly, womanly.
COVERED WITH/FULL OF: -y, -ful

A very productive adjective-forming rule in English is the one form-
ing adjectives meaning ‘covered with/full of what the stem denotes’
by adding the suffix -y to concrete nouns20 as in e.g. sandy, rocky,
stony ‘full of/covered with sand, rocks, stones’. The rule is very sen-
sitive to the nature of the nouns functioning as stems: it will oper-
ate only if the potential noun stem is a concrete mass noun like
sand, or an inanimate, concrete, natural object like rock and stone.
The rule will not operate with stems that denote man-made
objects like brick and screw, nor with stems that denote animals or
people (so-called animate nouns) like dog and people. Consequently,
there are no adjectives *peoply ‘full of people’ or *doggy ‘full of dogs’,
*bricky ‘full of bricks’ or *screwy ‘full of screws’.
As we have seen, ‘coverered with/full of’ adjectives in -y generally
have concrete stems. There is also a less regular type of ‘full of’
adjectivalization which attaches the suffix -ful to abstract noun
stems, as in e.g. joyful, painful, regretful, stressful, thoughtful.
CHARACTERIZED BY: -y
The adjective-forming suffix -y is very productive. In addition to the
meanings described earlier, it also appears with the somewhat
vague meaning ‘characterized by’ and ‘having the characteristics
of’. Admittedly, it is sometimes difficult to distinguish between
adjectives with ‘characterised by’ meanings and those with ‘full of/
covered with’ meanings. The stems to which -y is added are mostly
nouns, but it also turns up in combination with verbs, and occa-
sionally even with adjectives.
Combinations with noun stems are the most productive type,
and especially in informal spoken English, it is easy to get the
impression that almost any noun stem may be involved. However,
there seems to be a preponderance of short Germanic words Exam-
ples: ballsy ‘possessing balls’ i.e. ‘bold’ courageous’, brainy ‘charac-

5 Suffixation
terised by brains = intelligence’, classy ‘having/characterised by

class’, gutsy ‘having/ characterised by guts, i.e. courage’, panicky
‘characterised by panic’, smelly ‘having/characterised by a strong
smell’, trafficky ‘characterised by (lots of) traffic, touristy ‘having/
characterised by (lots of) tourists’.
HAVING THE PROPERTIES OF/ASSOCIATED WITH:

-al/ial, an/ian, -ese, -esque, -ic, -ite, -ous
All the suffixes above may be found in adjectives formed from
nouns having the rather vague meaning ‘having the properties of’
or the even vaguer meaning ‘associated with’. Stylistically, these
adjective suffixes are more formal than those discussed earlier and
they have a tendency to occur in combination with stems from
Latin and/or French. For the stress changes caused by these suffixes,
see pp. 81–83. Note that -al/ial, -an/ian, -ese and -ite have homo-
nyms used to form nouns (see pp. 88–89 in the present chapter).
The adjective suffix -al appears mainly in Latin and French adjec-
tives like accidental, dialectal, formal. It has the form -ial after the
noun suffix -or as in editorial, professorial and the form -ical when
alternating with a final -y suffix in word pairs like philosophy: philo-
sophical, psychology: psychological. Its productivity is uneven.
Adjectival -an and -ian are like their noun-forming counterparts
in most respects. Thus in the overall distribution of these forms,
-ian is much more common than -an, and if we disregard national-
ity adjectives like American, Brazilian etc, the productive use of the
suffix is largely restricted to combinations with personal surnames.
Since -ian is by far the commonest form, my examples will be lim-
ited to words with that suffix.Note that in words in -ian, primary
stress falls on the syllable immediately preceding the suffix. Thus a
word like e.g. Dickensian is pronounced [d kenziən].
Adjectival -ian is typically found in words expressing what is
characteristic of a certain person: Chomskian/Darwinian ideas are
ideas typical of Chomsky and Darwin and their schools of though,
a Clintonian approach to economics is one characteristic of Clinton.
Other examples of this quite productive type of word-formation are
Byronian, Caesarian, Churchillian, Cromwellian, Dickensian, Machia-
vellian, Orwellian, Stravinskian, Wordsworthian. When royalty is
involved, first names may also be used as stems, cf. Arthurian,
Edwardian, Victorian.

5 Suffixation
The -ian suffix seems to prefer stems of more than one syllable,
although there are examples with monosyllabic stems like e.g.
Bachian [bɑkiən] and Wellsian ‘in the manner/style of Bach/Wells’.
As pointed out on p. 89, the suffix -ese is fairly productive when
used to create nouns denoting the jargon of certain writers, groups
of people or types of text, as in Johnsonese, officialese, etc. The adjec-
tival suffix -ese is chiefly used to form adjectives linked to place-
names as in Japanese, Genoese, Burmese, Vietnamese, Taiwanese, Can-
tonese etc. Although the number of such adjectives is quite high,
this use of -ese can hardly be claimed to be productive.
The suffix -esque means roughly ‘in the manner/style of’ and is
used with a certain amount of freedom mainly together with names
of prominent artists: Kafkaesque, Daliesque, Danteesque, Kiplingesque,
Hornbyesque, but it has competition from many other suffixes,
mainly -ian, cf. e.g. Orwellian, Huxleyan, Wagnerian, Lodgian. The
suffix -esque is less common than its competitors: it is often used for
stylistic effect, as in the e.g.: ‘… something Ludlumesquely titled “The
Vienna Project”21.
The suffix -ic is common in adjectives of mostly Latin origin like
atomic, idyllic, poetic. Sometimes -ic may be replaced by -ical without
any change in meaning. However, in a few cases words in -ic and
-ical differ in meaning: economic ‘to do with the economy’: econom-
ical ‘money-saving’, historic ‘famous’: historical ‘on record in his-
tory’, classic ‘great, memorable’;: classical ‘classical music, classical
languages’, electric ‘powered by electricity’: electrical ‘relating to
electricity’. The -ic suffix is productive in combination with Latin
stems in scientific English.
The suffix -ite is more common as a noun suffix than as an adjec-
tive suffix: words like Blairite, Clintonite, Thatcherite, Trotskyite are
more likely to be nouns meaning ‘follower/supporter of Blair etc’
than adjectives. This suffix is less common than its competitor -an/
ian.
The suffix -ous is found in a number of words like e.g. adventurous,
envious, furious, humorous, venomous. These words look as if they
were combinations of present-day English elements (adventure +
-ous, etc.), but most, if not all, of these were borrowed from French
during the Middle English period. The suffix -ous is also common in
loans featuring bound stems, particularly in words taking -ion as a
noun ending. As a result, there are plenty of pairs like e.g. ambition:

5 Suffixation
ambitious, caution: cautious, oblivion: oblivious, religion: religious, sus-

picion: suspicious.
Although many of the -ous adjectives are examples of loans from
French, there are adjectives in -ous in which the suffix has been
attached to free English noun stems, for instance libel(l)ous ‘con-
taining/constituting a libel’ (a libel is a false claim damaging to a
person’s reputation), murderous and mountainous. On the whole,
however, the productivity of -ous in modern English must be
described as weak.
Adjectives from adjectives: -ish

There is one adjective suffix that combines productively with adjec-
tives, i.e. -ish meaning ‘somewhat’: brownish, blueish, coldish, tallish.
In this connection mention should be made of the use of -ish in
approximations of age and time such as sixtyish, forty-fivish ‘about
sixty/forty-five years old’ and sevenish, twelvish ‘about seven/twelve
o’clock’. Note also earlyish, latish ‘somewhat early/late’, which func-
tion both as adjectives and adverbs.
Adjectives from verbs: -able, -ive, -y

The only adjective-forming suffix combining freely with verbs is
-able. The meaning of -able is ‘that can be V-ed’ or ‘that should be
V-ed’ (where V stands for the actual verb stem involved). Thus if
something is doable, it can be done, and if a book is readable, it
should be read or is worth reading. Other examples are allowable,
answerable, manageable, predictable, watchable (TV programmes),
workable (policy). The suffix -able is occasionally also found in com-
bination with phrases: if something is said to be get-at-able it is
accessible, and if a book reviewer calls a book unputdownable it is
claimed to be so absorbing that it cannot be put down. (Obviously
unputdownable is a stylistically special word, but it shows what can
be done with -able by a determined writer of English).
Another common suffix forming adjectives from verbs is -ive. The
meaning of -ive is ‘that Vs’ or ‘that can V’: a connective word is a word
that connects or can connect, if you are supportive, you support.

5 Suffixation
However, it is rare for -ive to combine productively with verbs in

today’s English, and the relation between verb stem and suffix is
often less straightforward than in words in -able. A possessive per-
son, for instance, is not just somebody who possesses something,
but a person showing a strong desire to own things. The majority of
adjectives in -ive are earlier loans from French (abortive) and words
formed from Latin elements.
The difference between -able and -ive also extends to the com-
pany they keep: -able combines with both Germanic words and
words with other origins, while -ive cannot be attached to words of
Germanic origin. In addition, -ive triggers spelling and pronuncia-
tion changes in the stems to which it is added, while -able does not
(see pp. 82–84). Thus e.g. conclude and repel may combine with both
suffixes, but the final results are different: cf. concludable and repel-
lable vs. conclusive and repulsive.
The third and final adjective suffix attached to verbs is -y, mean-
ing ‘inclined to V’, ‘tending to V’.
Thus if somebody is chatty he/she is inclined to chat, a jumpy per-
son is nervous i.e. inclined to jump etc. Other examples are flirty,
gossipy, leaky, runny (as in a runny nose), sticky, twitchy, wobbly. This is
a quite productive suffix, characteristically found in informal Eng-
lish.
Verb-forming suffixes
Verbs from nouns and adjectives: -ify, -ise/ize
The formation of new verbs in English is mostly handled by means
of conversion (for which see Chapter 6) and back-formation (Chapter
9). However there are two common verb-forming suffixes in
present-day English, i.e. -ise (-ize) and -ify that are attached to both
adjective and noun stems
Verbs formed from adjectives by means of -ise/-ize all have the
meaning ‘turn something into what the stem denotes’, for example
Americanise, centralise, criminalise, equalise, formalise, legalise, mar-
ginalise, modernise, tenderise ‘make (meat) tender’. This is sometimes
also the meaning when the suffix is attached to nouns, for example

5 Suffixation
in vaporise and victimise. But in many cases the verbs formed from
nouns have less simple and predictable meanings. Thus symbolise
means ‘act as a symbol’, hospitalise mean ‘put into hospital’, and
dieselise means something like ‘convert to diesel-powered power’.
The recent formations computerise and itemise mean ‘convert to a
computer-operated system’ and ‘present data as a list of individual
items’.
The suffix -ise/-ize is also sometimes added to place-names, creat-
ing words meaning ‘make something like that place’ in Balkanise,
Israelise, Londonise, for instance. There are also more unorthodox
uses of -ise/-ize, as for instance in the monarchy has Diana-ised itself 22
recently found in a major British newspaper.
The suffix -ify has the same meanings as -ize, but is less common.
It is found with adjective stems in e.g. amplify, simplify and with
nouns in beautify, gasify, personify. On occasion, -ify is used with a
pejorative meaning, for example in speechify ‘make long and useless
speeches’.
Adverb-forming suffixes
Adverbs from adjectives: -ly
There are few adverb-forming suffixes in English, but there is little
doubt that one of them is the most productive English suffix of all,
i.e. the suffix -ly, which is added to adjectives, typically to form
manner adverbs. Thus adverbs in -ly are formed not only from mor-
phologically simple adjectives like brave, nice and slow, but also
from participial adjectives, for example in The professor was charm-
ingly inaccurate, The film is killingly funny, a jaw-droppingly ambi-
tious project, She was unashamedly frank about it, an outspokenly
honest account.23
There are few constraints on the use of adverb-forming -ly The
best known one is formal: as a rule, -ly is not attached to adjective
stems that themselves end in ly. For this reason it is rare to find
manner adverbs like kindlily (kindly+-ly) or friendlily (friendly + -ly).
Obviously there is also a semantic constraint: manner adverbs are
normally only formed from adjectives used to describe acts, actions

5 Suffixation
and processes. Adjectives with other meanings hardly ever form

manner adverbs: we are not likely to come across manner adverbs
like *bluely, *deadly or *navally (but note that there is an adjective
deadly)24.
Not all adverbs in -ly have ‘manner’ meaning. Such adverbs are
also formed fairly freely from all sorts of adjectives to express the
meaning ‘with regard to’, for instance in basically, economically
‘with regard to basic/economic matters’, and even navally ‘with
regard to naval matters’, ‘from a naval point of view’ as in e.g. Nav-
ally, that makes a lot of sense.
Compare also adverbs like honestly, fortunately, regretfully express-
ing different speaker attitudes, as in Honestly, I don’t find it very excit-
ing, I never replied, fortunately and Regretfully, he died last year.
Adverbs from nouns

A number of suffixes serve to create adverbs from nouns, chief
among them -fashion, -style, -wise and -ward(s).
IN THE MANNER OF: -fashion, -style, -wise

The suffixes -fashion, -style and -wise have the meaning ‘in the man-
ner of what the noun stem denotes’, as in e.g. She advanced towards
us crab-fashion (crabwise) and He dresses bank-manager style. The pro-
ductivity of these suffixes is limited.
IN THE DIRECTION OF: -ward(s)

The suffix -ward(s) has the meaning ‘in the direction of’ and is
found in a small number of adverbs like outward(s), inward(s), home-
ward(s). These are all established words and do not represent a pro-
ductive use of -ward(s), however. What productive use this suffix
has occurs in combination with noun stems, like earthwards, heav-
enwards, skywards.
WITH REGARD TO: -wise

When added to nouns, the suffix -wise is also used to express mean-
ings like ‘with regard to’ or ‘regarding’. This apparently quite pro-

5 Suffixation
ductive use of the suffix is found in e.g. It makes no difference, ener-

gywise, Teamwise, Germany were much less impressive than two years
previously. Have we solved the situation carwise?
Exercises
1 As Chapter 5 shows, there are many homonymous derivational
suffixes in English. How many are there in Chapter 5? Are cer-
tain forms more prone to be homonymous than others?
2 What are the characteristics of productive derivational suffixes?
On what grounds is e.g. -th declared to be ‘dead’?
3 Where does the main stress fall in the following pairs and tri-
plets of words and why does it fall where it does? Johnson: John-
sonesque, Eton: Etonian, racket: racketeer, author: authorize: author-
ity: authorization, history: historic.
4 What is the difference between the two suffixes -ful discussed in
this chapter? Can you find other examples of words with these
two suffixes?
5 What is the pronunciation of word-final -sion in derision, implo-
sion, propulsion, transmission? Can the pronunciation of -sion be
predicted?
6 Several types of -ese suffix are discussed in this chapter? Which
of them is productive?
7 The function of the -y/ie suffix in e.g. auntie, doggy, nightie is
rather different from the functions/meanings of most other
English derivational suffixes. Can you explain how -y/ie differs
from the others and can you find further examples of this suf-
fix?
8 How productive would you say that the suffixes -er and -ee really
are? Can you find instances of verbs which do not appear to
take these suffixes at all?

5 Suffixation
9 Explain the conditions that have to be met by nouns taking part

in hyphenated adjectivalization of the flat-footed and four-
wheeled kind.
10 Explain the conditions that have to be met by nouns forming

adjectives meaning ‘full of’, ‘covered with’ by adding the adjec-
tival -y suffix.

5 Suffixation

6 Conversion
6 Conversion
General characteristics of conversion

It is common knowledge that English nouns denoting instruments,
tools, and the like may be used as verbs to describe the process of
applying the instrument in typical fashion. Thus we phone, fax or
email other people, just as we hammer in a nail and saw a hole in a
wall. In the case of hammer and saw, it has been possible to use
them both as nouns and as verbs for more than 500 years.
But phones, faxes and emails have not been around for that long.
The fact that nouns denoting such recently invented instruments
may be used with the same verb meanings as hammer and saw indi-
cates that we are dealing with an established word-formation proc-
ess. The process in question is usually called conversion1, but the
terms zero derivation and functional shift may also be found.
The transformation of instrument nouns to so-called instrumental
verbs illustrated above is only one example of the use of conversion:
as will soon become clear, the conversion process may involve
many other meanings as well. What all instances of conversion
have in common is the fact that they change the word-class mem-
bership of a word without formal change.
Conversion operates on all word-classes, but is particularly com-
mon between nouns and verbs: as we saw in Chapter 3, it has been
suggested that ‘there is no noun in English that can’t be verbed2,
and even if that is an exaggeration, it is probably true that the
majority of (at least) the concrete nouns in English can also be used
as verbs.3
Although conversion typically operates on words from the open
word classes, i.e. nouns, adjectives and verbs, it sometimes also uses
words from the closed word classes as input, as for example when
the adverbs up and down are converted to verbs in e.g. They immedi-
ately upped the price, Several enemy aircraft have been downed, but by

6 Conversion
and large this is a less important aspect of conversion. There are also
instances of items from the open word classes that have been con-
verted to closed, for instance the prepositional use of present parti-
ciples like concerning, regarding, as in e.g. Do we have a problem con-
cerning/regarding money?
Conversion takes both simple, derived, and compound words as
input. In the case of derived words, it may overrule the word-class
identification provided by suffixes. Thus the presence of the noun
suffix -ion does not prevent words like e.g. caution and commission
from being used as verbs as in e.g. The prime minister cautioned
against tax increases, The new office building was commissioned yester-
day.
There are many examples of conversion from compounds, for
example the verbs to bottom-line and to cold-shoulder (other spellings
may be found) from the compound nouns bottom line and cold
shoulder discussed on p. 114 in this chapter. Occasionally, conver-
sion operates on parts of grammatical constructions as in He is an
also-ran, Don’t you my good man me.
Productivity in conversion
If the productivity of conversion is measured only in terms of word-
class change, this type of word-formation is close to 100% produc-
tive, in particular when it involves the open word classes. Thus
most nouns may be used as verbs, a great many verbs may be used
as nouns and adjectives can be turned into nouns and into verbs.
But in order to be considered truly productive, conversion must
also produce an output with predictable meaning. There are many
conversion patterns for which this is in fact the case, like e.g. the
conversion to verbs of nouns denoting instruments and tools.
These verbs always have one and the same meaning, i.e. ‘to use the
instrument or tool in typical fashion’: to fax is to use a fax to send
messages and to hammer is to use a hammer (or a hammer-like
instrument) to drive in nails.
In fact, conversion works only because it can be safely assumed
that speakers have a certain knowledge of the world that will help
him/her understand a new converted word. Thus even speakers
who have not previously encountered the verbs to fax or to vat will

6 Conversion
be able to interpret these new words correctly by applying his/her

knowledge of how (the referents of) the nouns fax and vat are nor-
mally used: faxes are typically used to send messages and vats are
typically used to store wine and other alcoholic beverages.
We may conclude from this that conversion must be described in
terms of certain broad categories of meaning reflecting the speakers’
perception of the world4. That is the basis of all descriptions of con-
version and it will be the basis of the outline of productive English
conversion offered below.
However, far from all instances of conversion produce words that
can be immediately understood. As in the case of derivation and
compounding, conversion may also be used to name or label new
phenomena. In such cases the meaning of the new word cannot be
predicted from our knowledge the world. As an example, consider
the noun to verb conversion from the noun carpet discussed in
Chapter 3. Given the knowledge we have about carpets, we are fully
prepared for the normal meaning of the verb i.e. ‘to provide (a
room) with a carpet’. What we are not prepared for is the meaning
it has in e.g. Hugh was carpeted by his boss meaning ‘Hugh was
severely reprimanded by his boss’. As the following sections will
show, conversion is frequently intentionally used to create such
new words with non-predictable meaning.
Productive English conversion patterns

Conversion from noun to verb
There is a fairly stable set of meanings expressed by verbs converted
from nouns, i.e. PUT IN/ON, PROVIDE WITH, REMOVE FROM, USE INSTRU-
MENT DENOTED BY STEM, TRAVEL BY/SEND BY, ACT AS/LIKE, CHANGE INTO.
PUT IN/ON (LOCATION, STORAGE)

Verbs with the meaning ‘put in/on’ are frequently formed from
nouns denoting containers: we bottle, can, tin, and vat food or wine
in order to preserve it or treat it, and we pot plants. We can also bag
and box something; in addition we can corner somebody, garage the
car and position a building. The meaning is sometimes figurative, as

6 Conversion
in e.g. file a complaint, market a project and dustbin (a plan or a

project).
Figurative meaning is especially common when the starting-
point for the conversion process is a compound: if we blacklist or
fastlane/fasttrack somebody, it is only figuratively that we put them
on the black list or in the fast lane/track.
PROVIDE WITH
Examples of ‘provide with’ meanings are e.g. carpet (a room), wall-
paper (a room), wall (off) an area, arm the people, fuel (up) a car, plas-
ter a wall, line a coat, floor a room and roof a house. Typically there
is an obvious and expected connection between the thing that is
provided and the space that is provided with it. To carpet or wallpa-
per a room is natural in most English-speaking cultures. It would be
considerably less natural to carpet or wallpaper a car or a safe.
REMOVE FROM
Noun to verb conversion with this meaning is only possible with
nouns denoting entities that are by definition part of a certain type
of ‘owner’ (a relation called inherent or inalienable possession in
Chapter 5). As examples of such inherently possessed nouns we may
mention bark (by definition part of a tree) and shell (by definition
part of e.g. nuts, peas, mussels). In exceptional cases, such inher-
ently possessed entities may be removed from their owners: we may
take away the bark from a tree or the shell from a nut. It is such
cases that the ‘remove from’ verbs are used to describe. If there is no
inherent possession involved, the meaning of the new verb will be
of the ‘provide with’ type.
Further examples of the ‘remove from’ type of conversion is
found in constructions like to core an apple ‘remove the core from an
apple’, bone a fish, skin an animal, top a tree, brain an animal. There is
a limited need for such verbs and the pattern can hardly be claimed
to be very productive. In addition, this conversion type has compe-
tition from prefixes like like un-, de- and dis- discussed in Chapter 4.
USE INSTRUMENT DENOTED BY STEM IN TYPICAL FASHION

Verbs with this meaning—also known as instrumental verbs—can be
created fairly freely. Most nouns denoting tools and instruments

6 Conversion
may serve as starting-points for such verbs. The meanings of the

verbs are always tied to the prototypical use of the instrument: thus
to hammer means ‘to hit with a hammer’ or ‘to hit as if with a ham-
mer’; if you use the hammer to, say, scratch you back, you cannot
be said to have hammered it. Other instrumental verbs are drill, saw,
chisel, sandpaper, file (‘make wood smooth by means of a file’).
Instrumental verbs have also been formed from a vast number of
nouns denoting more recent instruments, as in e.g. sledgehammer,
steamroller, keyboard, videotape, radio, telephone, brake. They have also
been formed from nouns denoting body parts as in e.g. He headed
the ball into the net, She elbowed her way to the front; cf. also the less
straightforward finger, stomach.
One peculiarity of verbs formed from nouns denoting tools and
instruments is the fact that the original tool/instrument need not
always be involved. It is thus perfectly possible to say e.g. He ham-
mered the table with his shoe or She sawed a hole in the wall with her
false teeth. To provide for such cases, we need to qualify the defini-
tion of instrumental verbs into something like ‘use instrument or
something like it in typical fashion’. (Cf. also Ljung 1976)
TRAVEL BY/SEND BY
The verbs in the TRAVEL BY/GO BY category are closely related to the
instrumental verbs. As their name indicates, they have to do with
means of transportation and include verbs like bicycle, canoe, motor
meaning ‘go by bicycle, etc.’, and also chopper ‘send/go by helicop-
ter’, mail ‘send by mail’. Sometimes the connection with the origi-
nal noun has been lost. A case in point is the verb ship, which
means simply ‘send goods to buyer by whatever means available’,
i.e. no actual ship need be involved. On account of this it is not
uncommon to encounter sentences like The article you ordered was
shipped on Feb. 22 by first class mail. There is also an intransitive verb
as in The new version will ship in a month or two.
ACT AS/LIKE
The meaning ‘act as/like’ is found in both transitive and intransi-
tive converted verbs. Thus you may bully (somebody), pilot (a
plane), host (a meeting)‘act as host for’, chair (a meeting) ‘act as
chairperson for’, police (a meeting), referee (a football match).

6 Conversion
Intransitive uses are found in e.g. to lord (it over the others) ‘act in a
superior and domineering manner’ and to swan (around) ‘move
about in a casual and relaxed way typically perceived as irresponsi-
ble or ostentatious by others.5 The notion of ‘acting as/like some-
body or something’ is a vague one and it is sometimes not easy to
understand just what it means, as in the cases of e.g. badger (some-
body) ‘repeatedly and annoyingly ask someone to do something’,
ferret (out something) ‘search tenaciously for something’. Although
not highly productive, the ‘act as/like’ pattern is still available for
new formations.
CHANGE INTO
Straightforward examples of converted verbs with ‘change into’
meaning are easy to find: cash (a cheque), wreck (a car), remainder
(books), bundle (one’s clothes), slice (a melon). Often a figurative
meaning creeps in as an alternative to the literal one. The verb
trash, for instance, may mean ‘change into trash’, ‘wreck’, but it can
also mean ‘criticize severely’ in both British and American English.
The British English verb rubbish, on the other hand, is used only
with the meaning ‘criticize severely’. The ‘change into’ pattern has
limited productivity and faces competition from suffixed verbs in
-ize/-ise like atomize and carbonise.
OTHER MEANINGS IN NOUN TO VERB CONVERSION

In addition to the categories of meaning above, many verbs con-
verted from nouns have meanings that are less predictable than
those found in most of the examples above. One example among
many is the verb to bottom-line, converted from the compound
noun bottom line, originally used to denote the final result of a bal-
ance sheet. The noun soon took on the additional meaning ‘out-
come’, ‘final result’. Using that meaning as input, conversion set in,
producing the verb to bottom-line meaning ‘explain/sum up the
final outcome of’. (Note that both the noun and the verb have pri-
mary stress on the second element.)
Other examples include the transitive cold-shoulder ‘be unfriendly
to someone’ (I was cold-shouldered by my old friends), exodus ‘force
many people to leave’ (They decided to exodus the city), fastlane/fast-
track ‘give priority to’ (We had to fasttrack the project), leaflet ‘distrib-

6 Conversion
ute leaflets’ (Different political groups were leafleting the neighbour-

hood), mainstream ‘bring into line with mainstream practice’ (Try to
mainstream your thinking), statement ‘send out information about
bank account’ (Service charges will be statemented every month) top
‘exceed’ (Losses are said to top £100 this year). Personal names may be
involved, as in Are you trying to Joan of Arc me?6 meaning ‘Are you
trying to kill me by setting fire to me’?
There are also many intransitive examples of noun to verb con-
version like audition ‘perform at an audition for a role in a play’ (I
once auditioned for a minor role in Hamlet), dialogue (with) ‘take part in
a conversation’ (We are not willing to dialogue with them), grandstand
‘attempt to attract favourable attention’ (a grandstanding actor),
interface with ‘interact with’, ‘establish contact with’ (We wish to
interface with the world), network ‘interact with people’ (They network
with other females), peak ‘reach a highest point’ (Interest rates will
peak at 7.5 per cent), platform ‘stop’ [of trains] (This train doesn’t plat-
form at Oxford’)
Conversion from verb to noun

Conversion from verb to noun is considerably less productive than
conversion going in the opposite direction. However, there are cer-
tain clear semantic trends, i.e. the tendency to create nouns denot-
ing ACTIVITY/INSTANCE/RESULT and AGENT/INSTRUMENT. (Note that
verb to noun conversion involving stress shift as in e.g. a túrn-off
from turn óff will be dealt with under partial conversion.)
ACTIVITY/INSTANCE/RESULT
The most productive type of verb to noun conversion is that pro-
ducing nouns denoting the activity described by the verb, a single
instance of the activity, or the result of that activity (it is often
impossible to distinguish between these): a fight may be the activity
of fighting (She was wounded during the fight), a cut may be both an
instance of cutting (He divided the loaf with a single cut of his knife)
and the result of cutting (Her arms had several cuts). Other examples
here are nouns like bite, bow, cut, fling, go, leak, look, pee, spin, try,
walk most often found in phrases like take a bite, give a bow, have a
fling, take a leak, take a look, have a pee, go for a spin (in the car), have

6 Conversion
a try/go, go for a walk. Other examples include (some) sleep/rest, an

attempt, a display, a show.
Many of the examples above are of long standing and have been
in the language for hundreds of years. However, verb to noun con-
version is an ongoing activity, and there are many recent examples
of the process. One of them is the noun take as in e.g. She has her
own take on things meaning ‘She has her own approach to things’.
Another somewhat more involved example of verb-to-noun con-
version is the noun spin, created by conversion from the homony-
mous verb. The sense of the verb is that found in sports texts, i.e. ‘to
cause a ball to rotate by giving it a twist’. As could be expected, this
verb was soon converted to a noun with the regular meaning ‘activ-
ity/instance/result of spinning’, as in e.g. The ball had a lot of spin.
The noun spin was subsequently borrowed into the language of pol-
itics with a slightly different meaning, i.e. ‘the (often favourable)
interpretation put on political events’. Spin-giving in this sense
soon developed into a flourishing business, with experts known as
spin doctors.
AGENT/INSTRUMENT
As shown previously, the most common way of forming agent and
instrument nouns is by adding the suffix -er (and sometimes other
suffixes like -ant, -ent and -or) to a verb: somebody who writes is a
writer, etc. In a small way, the same meaning may be expressed by
conversion as in a bore (‘a boring person’), a cheat (‘a person who
cheats’), a snitch (‘a person who informs on others’).
Other types of conversion

In addition to the noun-to-verb and verb-to-noun conversions dis-
cussed above, there is also conversion involving other word classes.
The most important of these is the conversion of participial
forms of verbs to adjectives in -ed and -ing creating forms like (very)
astonished, (highly) irritated, (quite) amusing, (absolutely) hair-raising.7
There are also less productive and regular types of conversion, i.e.
conversion of adjectives to nouns, of adjectives to verbs, conversion
involving closed word classes, and conversion based on phrases.

6 Conversion
Examples of adjective to noun conversion are: a bitter/two

bitters = pint(s) of bitter, a daily/two dailies, a weekly/two weeklies, a
friendly/two friendlies (non-competitive game or match), a white/two
whites, a nasty/two nasties (‘an unpleasant or dangerous thing’). Very
occasionally participial adjectives in -ed may be converted to
nouns. A recent example of this is found in the newspaper headline
US steps up efforts to win over ‘undecideds’8.
Conversion from adjective to verb is fairly common. It is found in
established examples like calm, dry, slow, empty, black, smooth, clean.
All these converted verbs have meanings that are predictable from
the meanings of the adjectives involved. This need not be the case,
however: consider the verb total, converted from the homonymous
adjective, for example. Total (v) has the predictable meanings ‘add
up to’ as in His earnings total £50, 000 and ‘add together’ as in total
one’s receipts. However, it may also mean ‘destroy a vehicle com-
pletely’ (particularly in American English) as in My car was totalled
on the highway yesterday.
Conversion from closed word-classes: on occasion prepositions/
adverbs are turned into verbs as in She upped and left, They downed
the aircraft. A more recent example is out ‘reveal the homosexuality
of somebody’ as in The mayor was outed in the local newspaper. There
is also at least one case of conversion from a modal auxiliary, i.e. A
visit to the royal torture chamber is a must.
Conversion from phrases includes e.g. I was just an also-ran and
now I’m a has-been, in the word wannabe (‘want to be’) used about
people who want to be famous and successful and others like e.g We
left a message on a post-it.
Partial conversion
At the beginning of this chapter, conversion was defined as a type
of word-formation involving no formal change in the word it oper-
ates on. Rather contradictorily, linguists sometimes also use the
term ‘partial conversion’ to describe word class change that is
accompanied by formal change. Partial conversion is used in three
such cases:

6 Conversion
In a very limited number of noun-to-verb changes, there is

accompanying pronunciation and/or spelling change: from the
noun shelf we form a verb shelve ‘(figuratively) put on the shelf’: the
-f is voiced to -v(e). In the same way, the verbs use, abuse, house have
a final voiced -s, while the nouns have final voiceless -s. In British
English there is also a difference in the spelling between the nouns
licence and practice, and the verbs license and practise. (In American
English, the spellings are license and practice for both noun and
verb).
Conversion from verb to noun may also be accompanied by
stress shift, especially in British English. Thus a limited number of
words have word-final stress as verbs, but initial stress as nouns, for
example addict (verb [ə dkt], noun [ dkt]), attribute (verb
[ə trbjut], noun [ trbjut]), import (verb [m pɔt], noun [ mpɔt]),
intercept (verb [ntə sept], noun [ ntəsept]), transfer (verb [trnsfə],
noun [ trnsfə] 9. The same difference is found in nouns converted
from verbs with the productive prefix re-‘again’, ‘anew’, like rewrite,
re-sit, re-take (an exam), which have nominalizations taking pri-
mary stress on the prefix re-. These shifts in stress placement are
found above all in British English, while American English some-
times places the stress on the first element of the word in both
nouns and verbs.
The only regularly productive type of partial conversion is the
one turning phrasal verbs into nouns (for phrasal verb cf. Chapter 2).
In this process, primary stress, which lies on the second element
(the particle) in phrasal verbs, is moved to the first element i.e. the
verb itself. Accordingly, when phrasal verbs like drop óut, push úp,
put dówn, take óut are converted to nouns, they change their stress
pattern to dróp-out, púsh-up, pút-down, and táke-out. There is also
often a spelling difference in that the nouns are hyphenated as in
push-up, or written as a single word, as in the case of slowdown.10
Sometimes the meaning of such nouns is fairly predictable; a per-
son who drops out of school—or some other organized activity—is
a dróp-out. A similar fairly regular correspondence between the verb
and noun meanings is also found in e.g. slow down:slowdown, pay
off: pay-off, cut out:cut-out.
But it is very common for verb-to-noun conversions accompa-
nied by stress shift to result in nouns with meanings not found in
the original phrasal verb. It is generally possible, for instance, to

6 Conversion
push up or put down a number of things, but the nouns púsh-up and
pút-down can only be used with reference to an exercise for
strengthening arm muscles and a critical and humiliating remark,
respectively. Likewise the verb melt down could be used about e.g. a
snowman or an ice statue, but the noun melt-down can only be used
about an accident in a nuclear reactor.
Exercises
1 What are the similarities and the differences between conver-
sion and derivational suffixation?
2 Why is it claimed that truly productive conversion is based on

assumptions about people’s perceptions of the world?
3 What is the labelling function of conversion and how does it dif-

fer from what was previously called ‘productive’ conversion?
4 What is the difference between the nouns taking part in ‘pro-

vide with’ conversion and ‘remove from’ conversion?
5 How does the conversion of instrumental nouns to verbs work?

Can you find exceptions to this pattern in the form of instru-
ment-denoting nouns that refuse to take part in this kind of
conversion?
6 What kind of conversion has affected the words printed in bold

in the following sentences? Who is paying for the eats? This pro-
gram will graph the tables. Bruce is a real swot. What you say
dovetails with what we already know. The demand for their prod-
ucts has almost flatlined. The unmicrowavability of these French
fries is well known11 Note that in addition to conversion, some of
the examples above have also been affected by other word-for-
mation processes.
7 Consider the underlined forms in the following examples: a

sandstorm, a Chicago mobster, a let-us-all-be-friends smile. Should
these items be regarded as adjectives formed by conversion? If
not, how do you suggest that the above forms are best
accounted for? (Cf. p. 200).

6 Conversion
8 In a white ‘a white person’, a Red ‘a communist’, the adjectives

white and red have been converted to nouns denoting people
with certain characteristics linked to the meaning of the adjec-
tive. Can you find other adjectives that may be converted to
nouns denoting persons?
9 The term partial conversion is really a contradiction in terms.

Can you think of other—and perhaps better—ways of account-
ing for the word-formation described under this heading.
10 If—as suggested in this chapter—participial adjectives like

astonished and amusing are regarded as cases of conversion from
verb to adjective, how do you suggest we explain the fact that a
form like e.g. painting is found both as a verb (She was painting
the wall) and as a noun (Hang the painting in the corner, His paint-
ing is getting worse and worse)?

7 Compounding
7 Compounding
General characteristics of compounding 1
In previous chapters we have discussed how words (lexemes) may

be turned into new words (lexemes) by the addition of prefixes and
suffixes, or simply by being converted to a new word class and mean-
ing without outward change. In the present chapter we will discuss
another, very productive, word-formation type known as com-
pounding, which combines two or more words into a new word
known as a compound.
The great majority of compounds belong to the open word
classes. The largest and most varied category is that of the noun
compounds, like for example textbook and blackboard, the second
largest contains adjective compounds like childproof, accident-prone
and breath-taking. There are few verb and adverb compounds. Most
verbs that seem to be compound have in fact been formed by back-
formation (see pp. 161–162) or conversion. The majority of apparent
compound adverbs are really suffixal derivations formed by attach-
ing the suffix -ly to compound adjectives of the breath-taking type.
Compounds belonging to the closed word classes are not as a rule
created productively by speakers.
As the examples in the previous paragraph show, compounds are
usually made up of two words, of which the second (called the
head) determines the word class and general category of the com-
pound as a whole, while the first (the modifier) explains what kind
of head we are talking about. Compounds of this type are said to be
endocentric, which means that semantically the compound as a
whole is an extension—or a sub-class—of the head: a textbook is a
book of a certain kind, a person who is accident-prone is prone to
have accidents, etc. The discussion in the rest of this chapter will
focus almost exclusively on the endocentric compounds.

7 Compounding
The tree-diagram below indicates how the compounds black-

board, textbook and accident-prone are analysed into modifier and
head.
Compound
Modifier Head
black board
text book
accident prone
The positions as modifier (first element) and head (second element)

are open to all words, i.e. it is not the case that certain words may be
used only as modifiers and certain others only as heads. In a com-
pound like textbook, for instance, text happens to be the modifier
and book the head, but nothing stops us from reversing that order
and producing the compound book text, if that is a meaning we
want to express.
However, it so happens that certain words have meanings that
make them particularly useful as heads, making them occur more
and often in that function. In certain cases, this may be the begin-
ning of a grammaticalization process with the end result that the
word in question is transformed into a suffix, i.e. a bound form. As
suffixes, the former words typically undergo semantic specialisation
and loss of stress. The history of English provides several examples
of such developments, for example the suffixes -hood, -ly and -dom,
as in childhood, friendly and kingdom. The suffix -hood goes back to
the Old English word had ‘state, rank’, -ly is a development of Old
English lic ‘body’, and -dom was originally a word meaning ‘judge-
ment, rule’.
Developments of this kind are by no means confined to history,
but are an integral part of the overall process of change forever
going on in all human languages. In today’s English, for instance,
certain nouns have begun to develop suffix-like qualities and may
be regarded as ‘pseudo-suffixes’. A well-known example of this pat-
tern, which goes as far back as the Old English period, is (-)man (as
in policeman, businessman, congressman, etc. In many but not all of

7 Compounding
its uses, -man is phonologically marked in that it has lost its word
stress and is pronounced [mən].
Semantically the meaning of -man is specialised in that it is used
to refer to people belonging to a certain profession or occupation
(denoted by the stem). For a long time it was taken for granted that
the representatives of these professions etc were all male, but in
recent years a distinction between the sexes has been made possible
by the introduction of forms like policewoman, congresswoman, etc.
A later development is the introduction of sex-neutral forms in
-person, sometimes providing a choice between three different
forms, two of them sex-explicit and one sex-neutral, for example
chairman/chairperson/chairwoman, salesman/salesperson/saleswoman,
spokesman/spokesperson/spokeswoman.
Two recent pseudo-suffixes are -babble and -speak. They are both
used to describe the kind of language typically used by certain
groups of people, but also indicate that the speaker takes a critical
view of both the group members and their language (especially in
the case of -babble). Among the examples of -babble and -speak for-
mations we find academic-babble, counselling-babble, hippie-babble,
agent-speak, business-school-speak, Bush-speak, City-speak, Foreign
Office-speak. In addition to being added to stems that are words,
both -babble and -speak are also often attached to bound classical
stems, as in e.g. eco-babble/eco-speak, Euro-babble/speak, psycho-bab-
ble, techno-babble used with reference to the jargons of ecologists,
Eurocrats, psychologists and experts on technology.
Compound versus phrase

In the opening paragraphs above, a compound was defined as con-
sisting of two words, of which the second (called the head) deter-
mines the word class and general category of the compound as a
whole, while the first (the modifier) explains what kind of head we
are talking about.
An important question raised by such an analysis is how to draw
the line between a compound—like blackboard—on the one hand
and a syntactic phrase—like black board—on the other. This is a
problem that arises in particular in the analysis of combinations in

7 Compounding
which the final word (the head) is a noun, not when the final ele-
ment is an adjective or a verb2.
Compounds and phrases are similar in that both consist of a
modifier (black) and a head (board). They also differ in significant
ways, however. In the opinion of many linguists, the most obvious
difference has to do with word stress: noun compounds have pri-
mary stress on the modifier (the first element), while noun phrases
have primary stress on the head (the second element)3. Thus not
only blackboard, but also other compounds like blackbird and dark-
room have the stress pattern [ - -], while the phrases black board,
black bird and dark room are stressed [- '-].
The difference in stress patterns has often been claimed to be
accompanied by a difference in meaning. According to this theory,
compounds are like simple words in being ‘labels’ (cf. p. 56) i.e. in
being used to refer only to special categories of ‘things’ in the (real
or imagined) world. Such an explanation is certainly true of many
compounds with noun heads, in particular adjective+noun com-
pounds like blackboard, blackbird, hothouse, etc:.a blackboard is not a
board that happens to be black, but a particular type of teaching aid
and parallel arguments can be made for blackbird and hothouse. In
fact in none of these combinations does the adjective have any
descriptive value: it is for example perfectly possible to speak of
green blackboards, albino blackbirds and cold hothouses.
The meanings of the syntactic phrases are quite different: black
boards, black birds and hot houses do not make up distinct catego-
ries of ‘things’. Furthermore, the adjectives involved have kept all
their descriptive force: black boards, black birds and hothouses
really are black and hot, respectively, which means that it would be
contradictory to speak of *green black boards, *albino black birds and
*cold hot houses.
As the discussion above has indicated, the combined meaning-
stress argument above works well for the adjective+noun combina-
tions: if they have primary stress on the adjective, they also tend
to have labelling meaning and may thus be called compounds. If
they have primary stress on the noun, they tend not to have label-
ling meaning and are accordingly syntactic phrases. Inevitably,
there are exceptions, for example combinations like white paper,
Yellow pages and small beer all of which have clearly labelling
meanings, but in which primary stress falls on the noun. (A white

7 Compounding
paper is an official report by a government, the Yellow Pages (trade-

mark) is book containing the telephone numbers of businesses,
and the informal small beer means ‘someone/something unimpor-
tant’).
Let us now turn to the noun+noun combinations. Like the adjec-
tive+noun combinations, they can be assigned to two groups with
regard to stress: those with primary stress on the first noun, and
those with primary stress on the second noun. But unlike what was
the case with the adjective+noun combinations, it is less easy to
link this stress-related distinction to that between labelling and
non-labelling meaning.
Compare for instance the noun+noun combinations smoke sig-
nal, jet aircraft, and climate change on the one hand, and London
policeman, Birmingham school, Sunday morning on the other. In the
first group, primary stress falls on the first word, while in the second
it falls on the final word. If noun+noun combinations work the
same way as adjective+noun combinations, this should mean that
the members of the first group have labelling meaning, while those
in the second group do not. However, there doesn’t seem to be
much of a difference between the two groups in this respect: argua-
bly, if smoke signals, jet aircraft, and climate change are ‘distinct
categories of things’ in need of labels, so are London policemen,
Birmingham schools and Sunday mornings.
A more promising approach, if we want to establish a semantic
difference between the differently stressed noun+noun combina-
tions, is to look at the ways the two nouns are related, or to be more
precise, how the phenomena referred to by these nouns are sup-
posed to interact ‘in the real world’. Although there are exceptions
to this, there is a strong tendency for noun-noun combinations
with primary stress on the first element to be typically associated
with certain such relations rather than with others.
One such relation is ‘function/ purpose’—the second element is
used for/in connection with the first—as in e.g. cough medicine,
space shuttle, mouse-trap, another is ‘production’—the second ele-
ment produces the first—as in e.g. power plant, gas works, oil well.
There is probably an infinite number of such relations, but as
argued above some of them seem to occur much more often than
others. A list of such frequently occurring noun-noun relations will
be found on p. 132.

7 Compounding
Two things need to be mentioned in this context. The first one is

that with the passage of time many of the referents of such noun-
noun combinations take on a certain character which comes to be
regarded as typical: there is, for instance, an infinite number of pos-
sible mouse-traps, but only one has proved successful and commer-
cially viable. As a result, mouse-trap has developed a labelling func-
tion over the years. The meanings of many other noun-noun
combinations have been affected in the same way.
The second matter that should be brought up here is that many if
not most noun+noun combinations are open to several interpreta-
tions: typically a mouse trap is a trap used to kill mice, but if it
turned out that mice were clever enough to set traps for other ani-
mals, mouse-trap would be a perfectly reasonable term for such a
contraption.
In the discussion above it has been suggested that by placing pri-
mary stress on the first element in a noun+noun combination we
can create noun+noun compounds, i.e. we indicate that the relation
between the ‘things’ referred to by the nouns is of a particular kind.
The fact that, with time, many of these combinations may develop
rather specialised meanings—a phenomenon known as lexicaliza-
tion—should not be allowed to obscure the highly productive
nature of noun+noun compounding as a device for forming new
English words with predictable meanings.4
Compounds and spelling

It is customary, in discussions of English compounds, to say some-
thing about the relation between compounds and spelling. It might
be thought that the compound:phrase distinction should be
reflected in the spelling. More precisely we might assume that word
combinations that are spelt as single words or are hyphenated—like
blackbird or sun-worship—would be compounds, while combina-
tions in which there is a space between the words (black bird, tight
rope, green card) would be non-compounds i.e. syntactic phrases.
However, such an assumption has fatal flaws. To begin with, con-
sider the adjective+noun combinations tight rope and green card.
They have two interpretations: one the hand they may be phrases
meaning ‘rope that is tight’ and ‘card that is green’ respectively, on
the other they may be compounds meaning ‘rope/wire stretched

7 Compounding
tight above the ground on which acrobats perform their feats’ and
‘permit allowing non-US citizen to remain and work indefinitely in
the US’.
As could be expected, the phrase interpretations of tight rope and
green card have primary stress on the second noun, while in the
compound interpretations, primary stress falls on the first noun: in
other words, the placement of primary stress tells us which combi-
nations are compounds and which are phrases. But the spelling pro-
vides no guidance whatever in this matter.
In addition, in compounding as elsewhere, English spelling turns
out to be quite inconsistent. Consider for instance the items busi-
nessman, business-man and business man. There are three different
spellings involved, but from the point of view of meaning all three
must count as one and the same word, an interpretation confirmed
by the stress pattern. Other examples of the same phenomenon are
e.g. girlfriend:girl-friend:girl friend and teapot:tea-pot:tea pot. Obvi-
ously we would not want to argue that businessman and business-
man, girlfriend and girl-friend, teapot and tea-pot are compounds, but
that business man, girl friend and tea pot are not.
From these examples we may conclude that if both semantic and
stress-related evidence indicate that a certain word combination is a
compound, then the way it is spelt is irrelevant. On the other hand,
spelling may be helpful when we wonder where to place the stress
in written forms: if a written form is made up of two shorter words
and is spelt as an uninterrupted word, then a good guess is that it is
a compound and should accordingly be stressed on the first ele-
ment.
Exocentric compounds
As we have seen, one of the main characteristics of endocentric
compounds is that the entire compound basically belongs to the
same general category as the rightmost word, the head: a textbook is
a kind of book, to headhunt is to hunt in a certain way, originally to
collect the heads of dead enemies, now to find a suitable person for
a business position.
There is another smaller group of compounds containing com-
pounds like hardback, paperback, for which this is no longer the

7 Compounding
case: these compounds do not, strictly speaking, have a head. Thus

neither hardback nor paperback denote a kind of back, but stand for
different types of books; a hardback is a book that has a hard back, a
paperback a book that has a back made of paper. Other examples are
hardtop, which not a top but a car with a hard top, and blackhead,
which is a pimple with a black head. Such headless compounds are
labelled exocentric compounds.
The above examples of exocentric compounds all denote
objects. However, many if not most exocentric compounds are
used with reference to people with certain characteristics. A red-
head, for instance, is a person with red hair, a blue-beard somebody
with a blue beard, and an egghead is a person with a high forehead
and hence assumed to be an intellectual. Similar meanings are
found in e.g. big-nose, big-foot used about people with big noses
and big feet.
Many exocentric compounds used about people have negative
meaning and are stylistically informal or vulgar, for example bird-
brain used about somebody considered stupid, loudmouth ‘person
who talks loudly and offensively’, the American badass for a tough
and aggressive person, and smartass/smart-ass (British smart-arse/
smart-alec) meaning a person who always has a clever answer. Other
examples of negative person-oriented exocentric compounds are
hard-ass ‘tough person’, American English tight-ass ‘inhibited or
conventional person’ sticky-fingers (used about a person who is
‘sticky-fingered’ i.e. a thief), Australian English sticky-beak meaning
‘an inquisitive person’, slimeball, scumbag, shitbag used about peo-
ple considered repulsive by the speaker, and the somewhat milder
slowcoach (BrE)/slowpoke (AmE), used to denote a slow person.
Another type of exocentric compound is the category of hyphen-
ated ‘coordinating’ constructions used to describe people having
two different functions or roles simultaneously, like actor-writer,
owner-occupier, philosopher-statesman, player-manager, secretary-treas-
urer, singer-conductor, writer-director, writer-producer. Most of these are
used as attributes in connection with proper names as in e.g. actor-
writer John Fields etc. In other uses they function as common nouns
and form plurals by adding a plural -s to the final word as in e.g. (tax
relief for) owner-occupiers.
There are also exocentric adjective compounds with ‘co-ordinat-
ing’ meaning. The most common type indicates that two character-

7 Compounding
istics are possessed in equal measure, as in e.g. a bitter-sweet smile,

the Swiss-German border, a French-Italian enterprise.
Noun compounds: general characteristics

Noun compounds are the most common type of compound and
exhibit a wide range of forms and meanings. In terms of the word-
classes involved, we can distinguish between noun compounds
consisting of a single common noun+single common noun (snow-
ball), -ing form+noun (washing machine), noun+ing-form (wine
drinking), proper noun+noun (Markov chain), single noun+agentive/
instrumental noun (tax-payer), adjective+noun (small talk),
verb+noun (push-button), preposition+noun (aftershock),
adverb+noun (off-chance). The majority of English noun com-
pounds are combinations of two (sometimes more) nouns and the
following comments deal almost exclusively with this type.
The list above does not include nouns like push-up, line-up, cut-
out—all with primary stress on the first word—which are not
regarded as compounds but are analysed as conversions from the
phrasal verbs push up, line up and cut out (cf. Chapter 6).
Noun + noun compounds

Syntactic noun+noun compounds5
There are two main kinds of noun+noun compounds—syntactic
compounds and others. The syntactic compounds have heads that
are nouns derived from verbs by the addition of the suffixes -er or
-ing, such as payer in tax-payer and drinking in wine-drinking.6 The
non-syntactic compounds, on the other hand, usually have heads
that are simple nouns, for example stain, rod and shuttle in blood-
stain, fishing-rod, space shuttle.
The fact that the syntactic noun+noun compounds have heads
derived from verbs makes it possible to describe their meanings in
syntactic terms.
Thus syntactic compounds with heads in -er, like tax-payer, car-
driver, smoke detector, all have the structure Direct Object + Subject: a

7 Compounding
tax-payer is a person (Subject) who pays taxes (Object), a car-driver is

a person (Subject) who drives a car (Object), and a smoke detector is
a device (Subject) that detects smoke (Object).
Syntactic compounds with heads in -ing such as wine-drinking,
cheese-eating, deer-hunting have equally straightforward meanings.
They have the structure Direct Object + Predicate’: wine-drinking is
the activity of drinking (Predicate) wine (Direct Object) and cheese-
eating is the activity of eating (Predicate) cheese (Object).
Compounds of the wine drinking and tax-payer types are very
common and constitute a highly productive source of new words
in English. The following list is a small selection of attested wine
drinking examples: dress-making, fault-finding, cigar-smoking, book-
writing, story-telling, bear-shooting, bass fishing, bird-watching, train-
spotting. An equally small selection of attested tax-payer examples
includes the following items: matter converter, stone-chucker, match-
maker, sheep-stealer, tabloid-owner, engine-driver, dog-owner, Channel-
swimmer, block-buster (‘bomb capable of busting (= destroying) an
entire block of buildings’, cassette recorder/player, war crimes trial
reporter.
The formal structure of wine drinking and tax-payer is also found in
compounds with slightly different meanings, such as day-dreaming,
sun-bathing, ocean-sailing on the one hand, and factory worker, cave
dweller, night fighter on the other. Here, the first noun (the modifier)
denotes the place where the activity is carried out rather than a
direct object: day-dreaming is ‘dreaming in the day(time)’, and a fac-
tory worker is somebody who works in a factory.
Before we leave this section, mention will be made of a problem
raised in Chapter 5 in connection with certain of the words
described as syntactic compounds above, for example tax-payer,
sheep-stealer, dress-making and train-spotting. Compounds have been
defined as combinations of stems that are themselves words in Eng-
lish and accordingly the four words above must be analysed as com-
binations of tax and payer, sheep and stealer, dress and making, and
train and spotting.
However, no dictionary lists payer, stealer, making and spotting as
entries and their status as English words is accordingly somewhat
problematic. The alternative to a compound analysis of tax-payer,
sheep-stealer etc. would be to regard them as instances of suffixation
involving the agentive suffix -er. But as we know, the suffix -er is

7 Compounding
added to verbs and since there are no verbs *to tax-pay, *to sheep-
steal, *to dress-make and *to train-spot, it seems wise to abandon this
line of reasoning altogether.
Whatever the solution to this problem may be, it is clear that
words like tax-payer etc constitute a challenge to the definition of a
compound as the combination of two or more words to form a new
word.
Non-syntactic noun + noun compounds

The previous section dealt with noun+noun compounds of the wine
drinking and tax-payer types, which have morphologically complex
heads ending in -er, -ing or synonymous suffixes. The remaining
noun+noun compounds have heads that are simple nouns, or suf-
fixed forms that do not carry suffixes of the -er and -ing type.
In such compounds, there is no clue telling us how the (referents
of the) two nouns are supposed to be related: all we can be sure of is
that the things denoted by the two nouns are somehow related to
each other. The exact nature of the relation will have to be deter-
mined by the speakers’ knowledge of the world and their opinions
about what is a reasonable connection between the referents of the
two nouns. Thus most of us will agree that the most likely interpre-
tations for e.g. power plant, space shuttle and blood stain are ‘plant
that produces power’, ‘shuttle used in space/for space travel’, and
‘stain caused by blood’.
In fact the examples above illustrate three common semantic pat-
terns in English non-syntactic compounds. Such meaning patterns
may be used to describe families of compounds with similar mean-
ings; accordingly we speak of a PRODUCING pattern for compounds
like power plant, of a PURPOSE pattern for compounds like space
shuttle, and of a CAUSE patterns for compounds like bloodstain.
Admittedly, a description in terms of meaning patterns has cer-
tain shortcomings. One of them is that one and the same
noun+noun combination may be used to express several of these
patterns. Furthermore, it is probably not possible to predict and list
all possible patterns.
However, even if it is impossible to list all the semantic patterns
open for the non-syntactic noun+noun compounds, it is obvious
that some of these patterns are much more likely to occur than

7 Compounding
others. So, while it is no doubt impossible to list all the possible

meaning patterns for English noun compounds, it is certainly pos-
sible to present a list of the most common and representative pat-
terns. That is the objective of the following list.
Common meanings in noun+noun compounds

In the following survey of (endocentric) English noun compounds,
only compounds made up of two words have been included. When
a distinction needs to be made between the first and the second
word, the first is called A, the second B.
B is an act by A: bee-sting, sunrise, sound change, population growth
B is an action/activity involving A: book review, haircut, word forma-
tion, population poll, book launch
B is an activity done in/on/at A: homework, moonwalk, city stroll,
barn dance
B is used for/connected with A: space shuttle, security officer, speed
bump, dustbin, fishing-rod, ashtray, washing-machine
B is powered by A: motorcycle, combustion engine, diesel engine, jet air-
craft, wind farm
B produces/causes A: toy factory, silk-worm, power plant, death blow,
cancer cell, wind machine
B is produced/caused by A: bloodstain, oil stain, saw dust, dog drop-
pings
B is part of A: doorknob, window sill, broomstick, floor board, window
pane
B is an A: girlfriend, killer whale, feeder bus, she-goat, he-goat
B resembles A: frogman, goldfish, bullet train, snail mail. In e.g. bear
hug and hyena smile, the hug and the smile obviously don’t
resemble a bear and a hyena, but a bear’s hug and a hyena’s
‘smile’
B operates by means of A: smoke signal, water jet, cable TV, ball bear-
ing
In addition to the above list of noun-noun compound meanings, in

which initial primary stress is the rule, there is a type of noun+noun

7 Compounding
combination in which the placement of primary stress is much

more difficult to predict, i.e. the combinations meaning ‘B is made
of/consists of A’.
Among the members of this category we find e.g. snowflake, rain-
drop, sand dune, all of which have primary stress on the first word.
However, we also find e.g. gold medal, chocolate biscuit, clay pigeon
with the same type of meaning relation between the two nouns but
with primary stress on the final element. To add to the confusion,
there are combinations with this meaning which would seem to be
exactly parallel, but whose stress patterns differ. Compare for
instance chocolate cake—with primary stress on the first word just
like e.g. snowflake—and chocolate pudding which carrries primary
stress on the second word, just like chocolate biscuit, gold medal and
clay pigeon.
Primary stress is also placed on the second word in killer whale.
Sometimes stress movement triggers a change in meaning: if pri-
mary stress in toy factory is moved from the first to the second word,
the resulting compound is likely to be understood to mean ‘factory
that is a toy’ rather than ‘factory that produces toys’.
The noun+noun examples in the list above are all made up of
common nouns. However, not all English noun+noun compounds
are of this type. One fairly common type of noun+noun compound
has a personal name as its first element, for instance Chomsky
adjunction, the Tourette syndrome (also known as Tourette’s syndrome),
Markov chain, Turing machine. Compounds of this type tend to be
highly specialised technical labels for processes associated with the
persons named by the modifier.
Finally, although the great majority of compounds made up of
common nouns consist of two (or more) singular nouns, there are
cases in which the modifier is a noun in the plural, for example
arms cache [kʃ] ‘place where arms have been hidden’, priorities
debate and weapons inspection.
Noun compounds with adjectival and other

modifiers
In addition to the noun compounds made up of two nouns, there is
a fair number of noun compounds consisting of an adjective modi-

7 Compounding
fier and a noun head. As the discussion on p. 124 showed, most of

these are labels, i.e. have institutionalized meanings and contrast
both in stress patterns and meanings with syntactic phrases with
predictable meanings. Compare for instance the compounds dark-
room, blackboard, whiteboard, madman, blackbird, tight rope with the
ordinary noun phrases dark room, black board, white board, mad man,
black bird and tight rope.
Noun compounds may also consist of combinations of locative
particles and a following noun, as in e.g. outhouse, back-water, out-
lier, in-breath, downpour and others. (For the formally similar type
upstream, upmarket, upscale etc., compare the section on compound
adverbs p. 139).
Historically English had a class of exocentric compounds consist-
ing of a verb followed by a noun, for instance pickpocket ‘person
who picks your pocket’ and cut-throat ‘person who cuts other peo-
ple’s throats’. This type has little if any productivity today. There
exists another type of verb+noun compound which is endocentric
and fairly productive, exemplified by sailboat, hovercraft (‘boat that
sails’, ‘craft that hovers’), swimsuit, talk-show (‘suit/show in which
you swim/talk’).
Adjective compounds 7
Participial adjective compounds

The most important types of compound adjectives have second ele-
ments that are participles in -ing and -ed like for instance breath-tak-
ing, money-making, far-fetched, deep-fried. Just like the nominal com-
pounds wine-drinking and tax-payer discussed earlier, these
adjectival compounds are syntactic, i.e. they are really abbreviations
of syntactic relations. Thus a wine-drinking headmaster is a headmas-
ter who (habitually) drinks wine, a money-making scheme is one that
makes (predicate) money (direct object), and a deep-fried dish has
been fried (predicate) deep (adverb) in oil.
The adjectival status of such combinations varies. While the
examples above contain -ing forms that have basically remained
participial, others have purely figurative meaning and are fully

7 Compounding
adjectival: they may be used both in attributive and predicate posi-

tion and have gradable meanings making it possible for them to be
modified by intensifiers like very, quite, extremely, etc. Thus an
experience may be extremely breath-taking, and an analogy may be
highly far-fetched. Other participial combinations are impossible or
at least odd in predicative position and when preceded by intensifi-
ers: it is doubtful whether you can say that a certain scheme is
money-making or that it is a very money-making scheme.
Type 1: breath-taking /money-making

In this highly productive compound adjective construction the
modifier is a noun and the head a present participle. It has syntactic
meaning in that the noun can be thought of as the direct object of
the verb underlying the present participle: if a business venture is
money-making, then it ‘makes money’. As a rule, primary stress falls
on the modifier, except in combinations with self-, like self-denying,
self-supporting, etc.
Many of these constructions have become fully adjectival and
can take intensifiers like very, quite, completely, extremely, etc., for
instance awe-inspiring, blood-curdling, breath-taking, hair-rising, mind-
blowing, mind-boggling, mirth-provoking, mouth-watering, thought-pro-
voking. These expressions are of course all metaphorical (cf. p. 25): no
breath-taking or blood-curdling experience actually takes your breath
away, nor does it literally curdle you blood (curdle is a technical term
describing what happens to milk if left too long). We are normally
so used to these expressions that we never think about their literal
meanings. But sometimes speakers/writers come up with fresh and
astonishing metaphors. A recent example is the (British) compound
adjective jaw-dropping, found in e.g. jaw-dropping remarks, a jaw-
dropping experience, literally remarks/an experience so surprising or
shocking that they make your jaw drop.8
Other participial combinations retain much of their literal mean-
ing and cannot as a rule take intensifiers, for instance: headline-grab-
bing (news), money-making (scheme), dollar-hurting (rate of exchange),
gravity-defying (pirouettes), money-swallowing (project), substance-
snorting (heir), bribe-taking (referee), banner-waving (supporters), life-
threatening (illness) self-supporting (person). Many of these and similar
adjectives are particularly common in certain types of texts like
newspaper and magazine texts.

7 Compounding
Type 2: far-reaching/hard-working
The type 2 compound adjectives consist of an adverb modifier and
a head that is a present participle. They have syntactic meaning: a
person who is hard-working is a person who works hard. Primary
stress as a rule falls on the modifier (cf. hard-working [ - -] but the
word stress pattern is more variable than in Type 1. Stress on the
head is often found in combinations like far-reaching and outgoing,
and often also in e.g. left-leaning, slow-moving.
Most compound adjectives of this type are gradable and may be
modified by intensifiers like very, extremely, quite, etc. as in e.g. very
far-reaching consequences, extremely hard-working people, incredibly
high-flying plans, highly left-leaning views, very rapidly-changing ideas.
The modifiers in these adjectives may be used in the comparative
and the superlative as the examples the biggest-selling book and a
higher-grossing product show.
Type 3: rain-soaked/Los Angeles-based/language-retarded

Adjective compounds belonging to this type are often stylistically
somewhat marked constructions and typical of the writing found in
daily newspapers, official documents and certain types of academic
and technical texts.
Depending on the nature of the verb underlying the -ed partici-
ple, these compounds fall into several different categories, corre-
sponding to (at least) three different syntactic constructions:
(1) The modifier is a common noun that can be regarded as the pas-
sive subject of the ed-participle: a rain-soaked body is a ‘body
soaked by rain’. Further examples: self-styled, self-appointed,
expert-tested (system), moth-eaten, suntanned, city-owned, rain-
enforced (break), rebel-supported (initiative), propeller-driven (Doug-
las DC-7). (We should also include here compounds like snow-
covered, sand-filled etc.). Sometimes the modifier is a nominal-
ized adjective, often indicating nationality as in British-led (oper-
ation), American-brokered (solution), French-controlled (territory),
but cf. also for instance (the) black-preferred (vote). The majority
of these compounds are non-gradable, but a few have gradable
meanings, for example suntanned and rain-soaked.
(2) The modifier is a geographical name or a nationality adjective
and indicates the base or origin of something. Examples: Bristol-

7 Compounding
based (company), Dallas-headquartered (organisation), Swedish-

born(comic), Devon-registered (fishing-boats), Winchester-educated
(Smith). Cf. also e.g. home-brewed, home-made, foreign-made.
(3) The modifier is a noun indicating the phenomenon that is
affected by the -ed participle like e.g. hearing-impaired and lan-
guage-retarded This is a minor compound category apparently
restricted to cases where the modifying noun is ‘naturally con-
nected’ with its owner (cf. the discussion in Chapter 5 of -ed
adjectives of the long-legged and blue-eyed type.)
Type 4: far-fetched/much-needed
The modifier in this very productive type of compounding is often
one of a rather limited number of adverbs expressing spread, dis-
tance and the like: examples are far-fetched, far-flung, wide-spread,
high-flown, deep-seated. The modifier may also be an adverb of fre-
quency or degree as in e.g. nearly-packed (stadium), almost-finished
(game), often-quoted (expression), understated, overblown, or an adverb
of manner as in neatly-folded (napkins), firmly-held (conviction). Note
that far-fetched, far-flung, high-flown, nearly-packed, almost-finished,
overblown, understated have primary stress on the second word.
Non-participial compound adjectives

In addition to the participial compounds, there are at least two
undoubtedly adjectival compound types, both of which combine
initial nouns with a following adjective.
Type 1: user-friendly/lead-free/work-shy
These compounds are made up of a noun modifier and an adjective
head. The meaning expressed by all such compounds is that some-
thing or somebody has the qualities of the adjective with regard to
the referent of the noun head: a user-friendly gadget is a gadget that
is ‘friendly’ with regard to users, a work-shy person is shy (reluctant)
when it comes to working.
Examples: accident-prone, lead-free (petrol), environment-conscious,
capital-intensive, childproof, crashworthy, machine readable, user-
friendly, carsick, trigger-happy, street-smart, work-shy. This is a very
productive pattern, and certain adjectives in particular enjoy great

7 Compounding
popularity as second word in these combinations, for instance

proof, worthy, free, and, in particular, friendly. Thus there is not
merely user-friendly, but also baby-friendly, customer-friendly, road-
friendly (car), wheelchair-friendly. For some reason, primary stress
goes on the second element in all combinations with -friendly.
Type 2: sea-green/red-hot
These compounds are made up of a noun modifier and an adjectival
head. Their meaning is basically comparative: sea-green means ‘as
green as the sea’, stone dead ‘as dead as (a) stone’, etc. Further exam-
ples of this type of compound are ash-blonde, sea-green, bottle green,
blood-red, snow white, pitch black, stone-cold, stone-deaf, stone-dead,
grass-green, ice-cold. These compounds all have primary stress on the
second word. There is a small set of informal compound adjectives
with similar form and meaning, but in which the first element
seems to have more of an intensifying force, for instance dirt-cheap,
piss-poor, red-hot (and possibly also a form like shit-scared).
Compound verbs
The term ‘compound verb’ suggests that there exists a compound-
ing process that combines two existing lexemes to form a com-
pound verb. However, the majority of verbs that look as if they were
compounds have not been created by combining two existing lex-
emes to form a third, but have been derived from compound
nouns, either by back-formation or by conversion.
When such verb are created by means of back-formation, the
starting-point is compound nouns ending in -ation, -ion, -ing, -er or
-or. The suffix is removed and the result is a new, back-formed verb:
from back-formation we get the verb to back-form, etc. Further exam-
ples will be found in Chapter 9.
For seemingly compound verbs that are actually cases of conver-
sion (cf. Chapter 6), the origins are compound nouns converted to
verbs. Examples of verbs formed in this way (and already men-
tioned in Chapter 6) are bottom-line, fast-lane, fast-track, keyboard,
mainstream, network, videotape, wallpaper. Other instances of such

7 Compounding
verbs are breath-test, brown bag ‘take a packed lunch to work’ or

‘bring one’s one liquor to a party’, carbon-copy and block-bust.
The verbs crash-land, double-book, fine-tune, soft-land may be gen-
uine combinations of noun plus verb and adjective/adverb plus
verb, but may also be either conversions or back-formations. Com-
binations of particles and verbs like e.g. backtrack, download and
upgrade appear to be genuine compound verbs.
Compound adverbs
With the exception of the examples in the section on exocentric
constructions (pp. 127–129), all compounds discussed so far have
been endocentric, i.e. the word class of the head word has also been
the word class of the finished compound. Thus textbook and green
card [ rinkɑd] are noun compounds because book and card are
nouns, and awe-inspiring and lead-free are adjective compounds
because inspiring and free are adjectives. Although there were few
directly formed verb compounds, the word class membership of the
compound verb agreed with that of its final element.
But in the case of adverbs, such correspondance is extremely hard
to find. With the exception of the informal British example double-
quick ‘very quickly’ there don’t seem to be any compound adverbs
whose final members are themselves adverbs.
There is no lack of adverbs that are combinations of words: a very
important type of adverb is the type exemplified by words like
mind-bendingly (boring), heart-breakingly (sad), jaw-droppingly9 (excit-
ing). But these are all derived forms created by adding the suffix -ly
to the adjective compounds mind-bending, heart-breaking and jaw-
dropping.
Another type of adverb created by the combining of words are
those consisting of a locative preposition and a following noun10.
Common prepositions in these cases are up and down as in e.g.
downmarket/down-market, upmarket/up-market, downstairs, upstairs,
mid-market, upmarket/up-market, upstream, downstream, upscale.
The majority of these combinations may be used both as adverbs
and as adjectives, but in most of these cases, the stress patterns are
different. As adverbs these compounds have primary stress on the

7 Compounding
second element, as adjectives they tend to take primary stress on

the first element; compare e.g. They live upstairs, The bus was rolling
downhill in which upstairs and downhill are adverbs and stressed on
the second element, and I’ll give you the upstairs room and It was a
downhill race in which upstairs and downhill are used as adjectives
and accordingly have primary stress on the first elements up- and
down-.11
Exercises
1 Try to find additional examples of formations in -man/-person/
-woman and -babble and -speak. Can you think of other forms
that should be regarded as pseudo-suffixes?
2 In addition to the forms -man, -babble and -speak used as exam-
ples of ‘pseudo-suffixes’, there is the form -gate, found in e.g.
Dianagate, Irangate, Iraqgate, Saddamgate. Try to find out what
common element of meaning these words have and why they
have this particular meaning (clue: cf. the Watergate scandal of
1972). Are words formed by the addition of -gate of the same
type as those formed by the addition of -man, -babble and
-speak?
3 The following word combinations all qualify as adjective+noun
compounds on semantic grounds, since their meanings are not
predictable from the meanings of the individual words that
make them up: black sheep, blue grass, blue jeans, bluejay, blue-
print, fine print, grey (gray) area, grey (gray) matter, small print,
smart bomb. To what extent is their status as compounds con-
firmed by their stress patterns, and what do they mean?
4 Using the criteria for noun+noun compounds discussed on pp.
123 and 125–126, which of the following noun-noun combina-
tions would you consider to be compounds?: bomb scare, bomb
disposal, chocolate biscuit, chocolate box, fishing-rod, fish farm, fish
finger, salmon-fishing, winning post, winning streak. Motivate your
answers.
5 Indicate how you would analyse compounds consisting of more
than two words, for example war crimes investigation, open class

7 Compounding
words, talk-show host murder, jet airline pilot strike, spare parts sup-
ply management system.
6 Try to create compounds describing the following phenomena:

(a) a shovel used to remove sand
(b) young people who drink wine
(c) people who (habitually) drink white wine
(d) a collection of records with only one song or piece of music
on each side
(e) an injury involving the ligament in the knee
(f) a gun powered by air
(g) a fear that people riding bikes are chronically under the
influence of alcohol
(h) grass reaching as high as a person’s knee
(i) legislation designed to be helpful to people who vote
(j) a task that is so difficult that it boggles the mind
7 Among the noun+noun compounds discussed in this chapter,

the most regular and productive types are the so called syntactic
compounds described on pp. 129–131, i.e. compounds like tax-
payer and wine-drinking. It is possible to challenge the com-
pound status of these words and to argue that they are not
formed by compounding, but in another way. Try to think of an
alternative explanation and what the consequences of accept-
ing it would be.
8 Although in general noun+noun compounds have more pre-

dictable meanings than the adjective+noun compounds, there
are noun+noun compounds with meanings that are impossible
to predict, for example the items listed below. Look them up in
a dictionary and find out what they mean and what their stress
patterns are. Also consider which of them may be given a more
straightforward interpretation and whether the difference in
meaning in such cases is reflected in the stress patterns: glass
ceiling, ballpark figure, pigeon-hole, holding pattern, guest worker,
toy boy, desktop, fish-finger, country house.

7 Compounding

8 Combining forms and neo-classical compounds
8 Combining forms and

neo-classical compounds
General characteristics
In our earlier discussion of word-formation mention has sometimes
been made of a phenomenon called neo-classical compounding and
of the so-called combining forms of which such compounds are com-
posed. In this chapter we will return to this highly productive type
of English word-formation in somewhat greater detail. However,
this is not the place to give a full account the neo-classical com-
pounds, and all the brief discussion below can claim to do is to
draw attention to their main characteristics and to point to some of
the ways in which they are similar to and different from affixation
and ordinary compounding.
The combining forms are bound base morphemes from Latin and
Greek. The ‘combining’ part of the name refers to the fact that such
forms are bound stems combining (in principle only) with each
other to create words like e.g. astronomy, biology, democracy, economy,
astronaut, democrat and others. Such combinations have certain
basic similarities with ordinary English compounds in the way the
elements they consist of are related. Both may be said to consist of
a modifier and a head: astronomy, biology and astronaut, for example,
are basically the same as ‘star science’, ‘life science’ and ‘star sailor’.
However, for all their semantic parallelism, the neo-classical com-
pounds also differ fundamentally from ordinary English com-
pounds in consisting of bound stems: it is true that both biology and
life science are similar in the way the meanings ‘life’ and ‘science’ are
combined, but the English compound consists of two words while
the neo-classical compound is made up of the two bound stems
bio- and -logy.

Another difference has to do with the number and order of ele-

ments serving as building blocks in the two kinds of compound. As
the discussion in Chapter 7 showed, English compounds may con-
sist of a very wide variety of word combinations, and there are no
restrictions on what may appear as first element and as second ele-
ment: house is the first element in the compound houseboat, but the
second in boathouse.
The combining forms, by contrast, are restricted in number1 and
subject to constraints regarding their position in the word. Some of
them—for example astro-, bio-, biblio-, demo-, ethno-, psycho-,
techno—may occur only as first elements in a neo-classical com-
pound and will be referred to as Initial Combining Forms or ICFs
for short. Others—like -crat/cracy, -gram/-graph/graphy, -naut, -logy,
-nomy—are Final Combining Forms (FCFs) that may be used only as
second elements. In principle, the two categories do not change
their position in the compound: initial forms remain initial, final
forms stay final.2
There are certain exceptions to this principle. A few combining
forms may be used both at the beginning and at the end of neo-
classical compounds, for example phon(o) ‘sound’ and phil(o) ‘love’,
‘friend’. Thus there are both e.g. phonology ‘sound study’ and Fran-
cophone ‘person who is a native speaker of French’, and both phi-
losophy ‘love of wisdom’ and Anglophile ‘person who loves things
English’. Another element appearing in both positions is the Greek
stem –cosm- ‘world, universe’, which is used as an ICF in cosmology
‘science of the origin of the universe’ and cosmonaut ‘navigator in
the universe’, but as an FCF in microcosm, macrocosm meaning ‘min-
iature universe’ and ‘the whole universe’ respectively.
There now remains one loose end to be taken care of in this pre-
liminary section: why are words like biology, astronaut, telegram, psy-
chology said to be ‘neo-classical’ compounds? Neo- is a Latin form
meaning ‘new’, and as the name ‘neo-classical compounding’ sug-
gests, this kind of compounding is new in the sense that combina-
tions like biology, astronaut etc. were not found in classical Latin or
Greek, but are later combinations of elements taken from Latin
and/or Greek. The difference between Greek and Latin now no
longer matters: the two languages are freely mixed as for instance in
compounds like television (Greek tele- ‘distant’ + Latin -vision ‘view-
ing’)

The neo-classical compounds are obviously also new in the sense

that most of the ‘things’ they refer to simply did not exist at the
time when classical Latin and Greek were still used. The great
majority of the neo-classical compounds belong to the language of
science and have been formed in relatively recent times: the words
telephone (‘distant sound’), telegraph (‘distant writing’) and helicopter
(‘spiral wing’) did not make their appearance until they had some-
thing to refer to.
However, the combinatorial possibilities of neo-classical com-
pounding have been available in English (and other languages) for
a very long time and the mere thought of possible future technical
developments may trigger the development of terms to describe all
manner of as-yet-unrealised inventions, cf. for example the attested
English word teleportation, which refers to a presently unavailable
technique of instantaneously moving matter from one point (in
space or time) to another (cf. ‘Beam me up Scotty’ from the scifi
movie Star Trek).
I end this section with a list of familiar neo-classical compounds
illustrating the use of some common combining forms: astronomy
(astr(o) ‘star’ + nomy ‘knowledge’), biology (bio- ‘life’ + -logy ‘study’),
democracy (dem(o)-‘people’ + -cracy ‘rule’), democrat ‘supporter of
democracy’, ethnography (ethn(o)- ‘people’ + -graphy ‘description’,
‘writing’), psychology (psych(o)- ‘soul’ + -logy), technology (techn(o)-
‘art’, ‘craft’ + -logy).
Neo-classical compounding compared

with affixation
In the account of the neo-classical compounds given in the previ-
ous section, these compounds were described as combinations of
initial and final combining forms, bound Latin and Greek stems
that can basically only combine with each other. This description
makes them clearly different from the other bound forms in English
that are restricted to word-initial and word-final position, namely
prefixes and suffixes: prefixes and suffixes are only attached to
stems that are words, not to bound stems.

There are also other differences between affixes and combining

forms. If we compare the initial combining forms—the ICFs—with
the prefixes, we note that, unlike the prefixes, the ICFs must end in
a vowel-sound, typically -o, as for instance in psychology, biology,
astronomy, astronaut but sometimes also in other vowel-sounds, as
for example in extravagant, hyperbole (‘exaggerated statement’),
omnipotence (‘all-powerfulness’).
However, the tendency for ICFs to end in -o is very strong. Just
how strong it is becomes apparent when—as happens—ordinary
English words are sometimes turned into ‘fake’ ICFs used in words
intended as humorous names for more or less improbably types of
scientific expertise. Two examples of such jokes are the words insec-
tology and kiddology with the intended meanings ‘the knowledge of/
the science of insects’ and ‘knowledge of/the science of children’
respectively.Note that these are merely two examples picked at ran-
dom: in principle it is possible to turn all English nouns into fake
ICFs in -o and it may also happen to phrases as in common sensology
recorded as early as 1903.
The joke may be pushed even further by combining kid, insect etc.
not only with -logy but also with other final combining forms, for
instance the -crat and -cracy found in democrat, democracy. The result
would be humorous forms like kiddocracy [k dɒkrəsi] and insectoc-
racy [nsek tɒkrəs] meaning ‘government by kids’ and ‘government
by insects’ respectively, and kiddocrats [ kdəkrts] and insectocrats
[n sektəkrts], which could be used about people (?) supporting or
involved in these types of government.
When native English words are turned into ICFs as in the exam-
ples above, another difference between ICFs and prefixes becomes
apparent. It has to do with the assignment of word stress. As
pointed out in Chapter 4, prefixes have secondary or even primary
stress. The ICFs on the other hand, follow a quite different principle
for stress assignment. This principle says that in neo-classical com-
pounds, primary stress is placed on the third syllable from the end
of the word, the so-called antepenult.
In accordance with this principle, neo-classical compounds con-
sisting of just three syllables have their heaviest stress on the first
syllable, i.e. the third syllable from the end of the word, cf. e.g.
astronaut, kiddocrat, microcosm. If a fourth syllable is added, the prin-
ciple of antepenult stress places the primary stress on the second

syllable from the beginning of the word (because that is now the
third syllable from the end): cf e.g. astronomy, psychology, kiddology,
democracy. Predictably, the addition of a fifth syllable moves the
stress one further syllable to the right as for example in aristocracy
literally ‘rule by the best’, a word formed from Greek elements in
Old French.
The workings of the antepenult stress rule are sometimes dis-
turbed by other factors, in particular by the presence of suffixes
affecting stress placement (cf. p. 81). Consider for instance the
adjective democratic formed from the same elements as democracy
and democrat. However, while the last two words are stressed in
accordance with the antepenult stress rule, democratic is not, but
has primary stress on the second syllable from the end of the word.
The reason for this is the presence of the suffix -ic ; like quite a few
other English suffixes (cf. pp. 81–83), -ic requires primary stress to
be placed on the syllable immediately preceding it.
As the discussion above has shown, there are several good reasons
to distinguish between prefixes and initial combining forms. How-
ever, there is no denying that certain linguistic forms seem to
belong to both camps, i.e. to be used both as prefixes and as ICFs,
for example the forms extra- and hyper-.
These forms quite clearly combine with stems that are words in
e.g. extra-territorial, extra-kind, hyper-correct, hyperinflation. Further-
more, they carry at least secondary stress in all combinations in
which they occur, they do not end in -o, and the words in which
they occur do not follow the antepenult stress rule. These are all
excellent reasons for regarding them as—quite productive—pre-
fixes.
However—as we have already observed—extra- and hyper- are just
as clearly ICFs in words like extravagant and hyperbole, where they
combine with bound stems, have somewhat indistinct meanings,
and are stressed in accordance with the antepenult principle
(extravagant is pronounced [k strvəənt] and hyperbole
[ha pəbəl]. The conclusion is that we have to accept the existence
of a fair amount of homonymy between prefixes and initial com-
bining forms.

Productivity of the neo-classical

compounds
Most of the examples given so far of neo-classical compounds have
been familiar words, words that we encounter so often that we no
longer think of them as compounds, for example astronomy, biology,
democracy, ethnography and psychology.
However, even in these few examples it is easy to see that the
combinatorial possibilities among the combining forms are far
from exhausted. As speakers, we are free to recombine the elements
in the examples above in whatever ways seem fit to us. Nothing pre-
vents our forming new words like for instance astrocracy ‘govern-
ment by the stars’, astrography ‘the description of the stars’, psychon-
omy ‘the study of the soul’, and ethnocracy ‘government by
(different) peoples’.
Today such recombining of bound Latin and Greek stems is one
of the most productive sources of new English words. Many of the
new words are known only to specialists in science and technology
and other areas of research. However, an increasing number of such
terms are now finding their way into the English in general use. A
recent instance of this is the new word kleptocracy [klep tɒkrəsi]
‘government by stealing/thieves’.3 The writer must have assumed
that the -cracy element was familiar from words like democracy etc.,
and that most of his readers would know the meaning of klepto-
from the words kleptomaniac and kleptomania.
Telescopic combining forms

In the previous discussion, word-initial forms like astro-, bio-,
psycho- were described as initial combining forms with the mean-
ings ‘star’, ‘life’, ‘soul’, which combine exclusively with final com-
bining forms like -logy, -nomy, and others to make up neo-classical
compounds like astronomy, biology, , psychology.
However, in recent scientific, technical and scholarly terminol-
ogy, such forms no longer represent bound Latin and Greek stems,
but are shortenings of old neo-classical compounds like astronomy,
biology, psychology and so on. (This particular type of shortening is

called back clipping and is described in Chapter 9). These new

forms—which have aptly been called ‘telescopic combining
forms’4—combine both with bound classical stems and with ordi-
nary English words, carry secondary or even primary stress. In addi-
tion, the words in which they occur are not stressed in accordance
with the antepenult principle.
The following examples illustrate the use of the telescopic forms:
astrophysics ‘the use of physics in astronomy’, bio-medicine ‘the com-
bination of biology and medicine’, ecocide ‘ecological murder’ bio-
terrorism ‘biological terrorism’ ecofeminism ‘ecological feminism’,
biohazard ‘biological risk to human health’, biodegradable ‘biologi-
cally degradable’, ecosystem ‘ecological system’, eco-fatigued ‘tired of
ecology’, technofear ‘fear of technology’, technobabble ‘technological
jargon’, technopeasant ‘farmer who makes use of modern technol-
ogy’.
Like the original initial combining forms, the telescopic forms have
a strong tendency to end in -o, but there is sometimes variation
between -i and -o for example in words containing the stem agr-:
the form used in telescopic combinations is sometimes agri- some-
times agro-, both meaning ‘agricultural’. Thus there is on the one
hand agribusiness, from which we get agriproducts, and on the other
agrobiology, agrochemical, and even agrobiotechnology.
The strength of the ‘final -o principle’ is also obvious in what may
be termed ‘fake telescopic forms’, i.e. bound forms in -o that have
not been created by clipping operations on neo-classical com-
pounds, but are—etymologically incorrect—reductions of other
words like e.g. globo-cop ‘police force of the entire world’ (formed
from global), robo-cop ‘policeman that is a robot’ (from robot) and
enviro-friendly ‘friendly towards the environment’.
Common combining forms

The lists on p. 150–151 show some of the common combining
forms.

Initial forms
Combining form Original meaning Example

aero air aeronaut(ics)
astro star astronaut
auto self autobiography
biblio book bibliophile
bio life biology
centi hundred centimetre
chrono time chronometer
crypto secret cryptography
demo people democracy/crat
dys bad, difficult dysfunction(al)
giga billion gigabyte
homo same homonym
kilo thousand kilobyte
macro large-scale macrocosm
micro small-scale microscope
mono (having)one monorail
multi many multi-ethnic
neo new neo-nazi
pal(a)eo ancient pal(a)eography
pedo/paedo child paediatrician
phono sound phonology
photo light photography
poly (having)many polysemy
porno prostitute pornography
proto original prototype
psycho soul psychology
techno art, craft technology
tele distant television

Final combining forms
Combining form Original meaning Example

-cide killing homicide
-cosm world microcosm
-cracy form of government democracy
-graphy writing pornography
-logy science biology
-mania madness bibliomania
-meter measure centimetre
-naut sailor, pilot astronaut
-nomy the laws of astronomy
-nym name homonym
-phile loving bibliophile
-phobe fearing technophobe
-phone sound telephone
-scope view telescope
-vision view television
Exercises
1 Look up the following neo-classical compounds in a large Eng-
lish-English dictionary that provides both pronunciation and
etymologies. Find the meanings of the elements that make up
the words and determine whether the stress pattern agrees with
the antepenult principle: pleistocene, oxymoron, pentagon, carbo-
hydrate, appendectomy.
2 Try to explain the variation in stress found in the following

series of words (Tip: take a look at Chapter 5, pp. 81–83):
automaton:automate:automatic;
pathology:pathologist:pathological;
cryptogram:cryptography—cryptographer—cryptographic

Anglophile:Anglophobe:Anglophobia
acid:acidic:acidophilus:acidophilic
3 Form as many meaningful combinations as you can from the

following ICFs and FCFs: ICFs: astro-, auto-, bio-, crypto-, dys-,
multi-, micro-, phono-, techno- tele- FCFs: -cide, -cosm, -cracy, -logy,
-mania, -naut, -phobia, -phone, -scope, -vision
4 The word helicopter is often analysed as consisting of heli- and

-copter. This is etymologically incorrect, since helicopter is a com-
bination of helico- ‘spiral’ and -pter ‘wing’. However, the form
heli- occurs in a number of combinations, such as helipad ‘land-
ing and take-off area for helicopters’, heliport ‘airport for heli-
copters’, and heliskiing ‘skiing in which the skier is taken up to
the mountain by helicopter’. How would you classify the form
heli- in helipad, heliport and heliskiing—as a prefix, an initial
combining form or the first element in a compound?
5 In Chapter 8 it was found that forms like astro-, bio-, etc. may be
both ordinary initial combining forms and so-called ‘telescopic’
forms. Can you think of final combining forms that may be
used as telescopic forms?

9 Irregular Word-formation
Characteristics and types of irregular

word-formation
The purpose of the present chapter is to briefly present a number of
processes that serve to extend the English vocabulary in different
ways, but which are clearly on the outskirts of traditional word-for-
mation. There are three main types of such irregular word-forma-
tion processes, i.e. back-formation, shortening with the sub-categories
initialisms and clippings, and blending i.e. word mixing. Some atten-
tion will also be given to reduplicative compounding and to a minor
type of irregular word-formation known as rhyming slang. In addi-
tion, the phenomena of borrowing and meaning extension will be
briefly considered.
Before proceeding to a presentation of the different ‘irregular’
processes above, it is worth considering just how ‘irregular’ word-
formation differs from its ‘regular’ cousin. What is it that the regu-
lar word-formation types have got that the irregular ones haven’t?
There are at least two major differences. One has to do with the
nature of the output of the two types of word-formation. While it is
true that an important function of many regular word-formation
rules is to enrich the vocabulary by creating ‘labels’ for new catego-
ries in need of naming, many of these rules are also used for ‘syn-
tactic repackaging’ (cf. p. 56), i.e. to create forms that are used and
then forgotten.
By contrast, two of the three main types of irregular word-forma-
tion are used mainly to create labels, i.e. to enrich the vocabulary:
both shortening and blending have this function. Back-formation is
different in this respect, and has been regarded by some scholars as
a kind of derivation operating in the same way as prefixation, suf-
fixation and compounding.1

The other difference concerns the predictability of the forma-

tions. Most regular word-formation processes create new words
from old in reasonably predictable ways and by predictable means:
we expect action verbs to be able to develop agent nouns in -er, we
expect most adjectives to form state/activity or extent nouns in
-ness, we expect conversion and compounding to follow certain
patterns, and by and large we are right to do so.
Most irregular word-formation lacks such predictability. Thus the
process of shortening known as ‘back-clipping’ (p. 159) gives us the
clipped forms celeb, demo and rep from the words celebrity, demon-
stration and representative, but not e.g. *calam, *explan or *init from
calamity, explanation and initiative. In addition, we can never be
quite sure what parts of a word will be removed by clipping or be
involved in blending. Who would have guessed, for instance, that a
word like weblog would be ‘fore-clipped’ to blog (cf. p. 161), and why
is it that the blending of fog and smoke yielded the form smog rather
than e.g. *foke?
Borrowing
Strictly speaking, borrowing has nothing to do with word-forma-
tion: it is a way to enrich the vocabulary resources of a language by
importing foreign lexical material. However, borrowing is not
entirely unrelated to word-formation: what is borrowed, in certain
cases, is not merely words or phrases, but also the patterns and
affixes underlying word-formation in a foreign language. One
example of pattern borrowing is the construction el --- o, borrowed
from Spanish and used in American English with the meaning ‘the
most …’, as in el cheapo (‘the cheapest’), el creepo (‘the greatest creep,
most like a creep’), el sleazo (‘the sleaziest’)2.
Another loan from Spanish is the informal savvy from Spanish
sabe usted ‘you know’. On its own, savvy is used both as an adjective
meaning ‘competent, knowledgeable’ and as a noun meaning
‘knowledge, know-how’. It is used in word-formation to form adjec-
tive and noun compounds like for example computer-savvy, E-savvy,
street savvy, tech savvy. Thus a person may possess computer-savvy, E-
savvy (E = ‘electronics’) or he/she may be computer-savvy, E-savvy.

A few instances of affixes or affix-like elements have been bor-

rowed from German. Thus echt ‘true, genuine, real’ functions as an
English prefix in e.g. echt-Massachusetter, echt-Viennese. Another
German prefix loan is uber- [υbə(r)] used with a meaning close to
that of super- i.e. ‘the ultimate form of …’, as in uber-boomer, uber-TV
show.3
There is also a suffix -nik borrowed from Russian and Yiddish,
which enjoyed considerable popularity in the days of the Sputnik.
It is found in words like beatnik and peacenik and also in more recent
words like refusenik, spacenik, waitnik, nogoodnik. As the examples
indicate, this suffix is used to form nouns denoting individuals. The
exact meaning of the suffix is hard to define; in fact there seem to
be several suffixes -nik with more or less different meanings.4
A more recent suffix loan is -bot5, a suffix derived from Czech
robota meaning ‘forced labour’. This suffix is used to create words
related to machines, computer programs and the like. Examples
include: knowbot ‘an automatic agent able to search for electronic
networks’, cancelbot ‘computer programs that automatically cancel
unsolicited advertising on the Internet (also known as SPAM); there
is also a verb to cancelbot. Other examples include mobot (mobile
devices capable of independent behaviour), microbot, nanobot (dif-
ferent types of miniaturized machines).
Mention should also be made of the suffix -ville, obviously bor-
rowed from French and sometimes used in American English in the
form -(s)ville6 to denote a place or situation having the characteris-
tics of the word preceding -sville as in e.g. dullsville, dragsville,
squaresville These are fairly dated examples, but as the following
quotes indicate, -(s)ville still has a certain amount of productivity:
‘This is shitsville’ said Mr Leiter.7
‘While many of the MBA gold-diggers high-tailed it back to old
Economy-ville, the people who matter in Silicon Valley—the
geeks—weren’t going anywhere.’8

Meaning extension
Not everybody would regard meaning extension or semantic
change, as it is also called, as a kind of word-formation. Yet undeni-
ably the extension of the meanings of existing words is a very pro-
ductive way of forming new words. In the present book it has been
regarded as a form of irregular word-formation, i.e. basically as
word-formation for which no principles may be established; others
take a different view9.
There is a certain superficial similarity between meaning exten-
sion and the type of regular word-formation called conversion. Like
conversion, meaning extension brings about a change of word
meaning without an accompanying change in form. But at that
point the similarity ends: meaning extension does not change the
word class of a word and it has none of the regularity and predicta-
bility of meaning found in conversion.
Meaning extension operates in several different ways. In one of
its manifestations, it is a continuous historical process of semantic
change that affects all languages and gradually changes the mean-
ings of their vocabulary items. This kind of change is very slow and
speakers often take a long time realising that the meaning of a word
has gradually changed.
It has often been observed10 that the gradual type of meaning
change mentioned above can be described in terms of two diamet-
rically opposed semantic tendencies. On the one hand, it may lead
to generalization or widening of meaning, as in the case of the Eng-
lish word box, which originally denoted only small containers
made of box-wood, but has gradually acquired its present meaning
of ‘any container with a flat base and sides, typically square and
having a lid’11 On the other hand, it may lead to specialization or
narrowing of meaning. An often used example of this is the word or
meat, which originally denoted food in general and only later took
on the meaning ‘edible flesh’.
But semantic change may take place much more quickly: the
word gay for instance used to mean ‘happy’, ‘carefree’ until the
1960s, but is now exclusively used with the sense of ‘homosexual’,
a change that started in American English in the 1960s12 The new
meaning of gay has now replaced the old one more or less com-
pletely.

Initialisms: abbreviations and acronyms

Several of the irregular word-formation patterns involve word
shortening. The most common type of shortening is represented by
the so-called initialisms, which reduce multi-word combinations—
usually names of organizations—to single words made up of the ini-
tial letters of the words involved in the combinations. Initialisms
are now a permanent feature of English and a rapidly growing cate-
gory that numbers hundreds of thousands of members. In fact, the
recent Acronyms, Initialisms and Abbreviations Dictionary (22nd edi-
tion, 1997) contains over half a million entries.
There are two kinds of initialisms. The larger group, known as
abbreviations, contains items in which each letter is pronounced
separately: CIA is pronounced [sia e], UN is pronounced [jυ en] A
smaller but still important group contains shortenings known as
acronyms, in which the combinations of initial letters are read out as
words, as for instance in AIDS [edz], and NATO [ netəυ].
There are also initialisms in which the two types are mixed, in
particular compound initialisms in which the first element consists
of one—sometimes two—letters, like e.g. e-mail/email, g-man, g-suit,
T-bone, X-ray/X ray, AK-47, CDROM. In addition there are certain
initialisms that may be treated as either abbreviations or acronyms,
for instance RAF (the Royal Air Force).
Abbreviations
In writing, abbreviations consist of strings of letters, which may or
may not be separated by full stops. Most of them are spelt with cap-
ital letters, although there are those that are not. Many are proper
names of institutions or places, for instance CIA (Central Intelli-
gence Agency), FBI (Federal Bureau of Investigation), NYC (New
York City), UN (the United Nations), ANC (African National Con-
gress), GOP (the ‘Grand Old Party’ = the American Republican
Party), GB (Great Britain). A few abbreviations refer to—usually
famous—people, for example FDR (Franklin Delano Roosewelt), JFK
(John Fitzgerald Kennedy), GWB (George Walker Bush).
Other abbreviations are non-names: DJ/dj (disk jockey), DNA
(deoxyribonucleic acid), TEFL (the teaching of English as a foreign

language), IOU (for ‘I owe you’, a signed document explaining that

you owe money to somebody), MC/emcee (master of ceremonies),
VIP (very important person), TV, CEO (Chief Executive Officer),
RPM/rpm (revolutions per minute), SF/sf (science fiction), PC/pc
(personal computer; political correctness), SUV (suburban vehicle).
As the example PC/pc shows, one and the same abbreviation may
have more than one denotation.
There is also a smaller category of abbreviations which stand for
common phrases, for instance PDQ/pdq (‘Pretty damn quick’), sob
(‘son of a bitch’), asap (‘as soon as possible’—also used as an acro-
nym pronounced [eisap] or [aesap], aka (‘also known as’ as in John
Smith aka John Stickyfingers), FAQ (frequently asked questions).
Acronyms
Turning now to the acronyms, we may begin by noting one differ-
ence between these forms and abbreviations: since acronyms have
to be pronounceable, they can only contain letter sequences also
permitted in ordinary English words (cf. the discussion of phonotac-
tic rules on p. 52).
Like the abbreviations, acronyms often denote institutions and
organizations:
NAFTA (North American Free Trade Association), NATO (North
Atlantic Treaty Organization), UNESCO (United Nations Educa-
tional, Scientific and Cultural Organization), WAC (Women’s Army
Corps). Certain acronyms naming organizations have obviously
been composed as a (more or less obvious) reminder of the purpose
of the group, like e.g. GASP (Group against Smokers’ Pollution) and
AIM (American Indian Movement), PUSH (People United to Serve
Humanity).
Among the acronyms that do not name institutions we find e.g.
AIDS/aids, SAD (Seasonal Affective Disorder)13, soho (small office
home office) RAM (random access memory), ROM (read only mem-
ory). The following denote groups of people with certain character-
istics: yuppie (young upwardly mobile), GLAM/glam (greying, lei-
sured, affluent, married) and WASP/wasp (white, Anglo-Saxon,
Protestant).

A number of acronyms have to all intents and purposes become

common nouns and are always spelt with small letters like e.g. laser,
radar, scuba, and the rather surprising noun snafu, defined in dic-
tionaries as ‘a chaotic or confusing situation’ and said to be a short-
ening of ‘situation normal, all fucked up’. Acronyms may also occa-
sionally stand for more or less natural phrases, for instance nimby
‘not in my backyard’, and Thif ‘thank God it’s Friday’.
Clipping
English has a number of ways to reduce word length. In the previ-
ous section we discussed one of these methods of word shortening,
i.e. the creating of initialisms. This technique makes it possible to
reduce multiword units like the names of organizations or products
to strings containing the initial letters of the words in the original
multiword unit.
But shortening processes may also operate on single words by
removing part of the word, an operation known as clipping. In the
most common type—back-clipping—the final part of the word is
removed, as for instance in lab(oratory), porn(ography) and
prof(essor). Other examples of back-clipping are ad(vertisement),
admin(istration), advert(isement), beaut(y), biog(raphy), bod(y),
celeb(rity), demo(nstration), decaf (‘decaffeinated coffee’), ref(eree),
info(rmation), intro(duction), perk (>perquisite ‘fringe benefit’),
perp(etrator ‘person who has committed a crime’), rep(resentative),
veg(etable), Jag(uar), Merc(edes), dorm(itory), cert(ainty), rev(olution),
tech(nology).
Note that back-clipping usually operates only on certain senses of
a word: the back-clipping rev (plural revs), for instance, can only be
used in the technical sense ‘an instance of revolving’; rev is not a
back-clipping of revolution ‘political upheaval’ and consequently we
cannot say e.g. *the Russian Rev meaning ‘the Russian Revolution’,
only the engine was turning over at 300 revs a minute. (There is also a
verb to rev, which means to make an engine revolve at a faster pace:
She was revving her engine). Cert for certainty is mostly used in the
phrase It’s a (dead) cert!. The clipping Brit is predominantly used to

denote a person who is British, rather than as an alternative to the

adjective British.
As the examples above show, the majority of back-clippings are
nouns. There are also back-clippings from other word classes, for
instance the adjectives brill(iant) and fab(ulous).
Most of the back-clippings above are easy to understand even if
you have not seen them before. Sometimes the connection with the
underlying word is less obvious. A recent example of such a non-
transparent back-clipping is dis/diss ‘speak in a disrespectful way
towards or about somebody’, as in He was dissed by the White House
(Newsweek April 29, 2002, p. 44). The word could in theory be an
abbreviation of any of the many verbs in dis-, but is a back-clipping
from the verb disrespect.
Back-clippings may also be dephrasal, i.e. shortened forms of
phrases, for instance nuke<nuclear bomb, prefab<prefabricated build-
ing, zoo<zoological garden, pub<public house (in the British sense i.e. a
bar open to the public).
In another type of clippings, known as fore-clippings, it is the first
part of the word that is removed, for example in (tele)phone,
(omni)bus, (aero)plane, (para)chute. In a third type, both the begin-
ning and the end of a word are removed, leaving just the middle
section: (in)flu(enza), (re)fridge(rator). Neither of these is as common
as back-clipping; the third type is in fact rare.
Words formed by means of clipping may contain a plural -s suf-
fix, for instance binocs ‘binoculars’, hols ‘holidays’, specs ‘spectacles’.
On occasion, suffix-like elements with no clear meaning or func-
tion are added, for example in preggers ‘pregnant’, soccer ‘Associa-
tion Football’, sparks ‘radio operator’, turps ‘turpentine’.
Other clippings have added the ‘familiarity’ suffix -y/-ie (cf. p. 91),
for instance the informal British English Aussie ‘an Australian’ bevvy
‘alcoholic beverage’, bicky/bikky ‘biscuit’, brickie ‘bricklayer’, telly
‘television’. Another example of a clipping followed by a suffix is
aggro ‘violence, trouble’, probably from aggression and the suffix -o.
In a few cases of clipping, the new shortened forms have ousted
the original forms more or less completely, for example in the case
of bra and car, which have practically killed brassière and motorcar.
But in most instances of clipping, both forms remain in the lan-
guage, but with a stylistic difference: the clipped form is usually
considerably less formal than the unclipped one. Not surprisingly,

therefore, clipping is an important source of informal expressions

and slang.
Clipping sometimes removes parts of words that are in them-
selves meaningful units, for example when aero- is removed from
aeroplane and when respect is removed from disrespect. But, as most
of the examples above show, most clippings do not show the same
respect for natural boundaries: cf. for instance the clipped forms
biog, info, demo from biography, information, and demonstration,
which would have been more naturally analysed into bio- and
-graphy, inform- and -ation and demon(str) and -ation.
A recent example of unlikely clipping is the word blog from
weblog14 ‘easily updated website’ (Newsweek May 27, 2002, p79),
and also the derived forms blogger and blogging. Of course nobody is
unaware of the correct analysis of weblog, i.e. as the combination of
web and log; the reason for the new formation must be a desire to
distinguish weblogs from all other logs.
Back-formation
Back-formation creates new words from old by removing a noun
suffix from words that contain typical noun endings like e.g. -ation,
-er, -or, -ing, -ism and -ist. It is a process that has been at work in Eng-
lish for hundreds of years, the classic example being the verb edit
formed from the borrowed Latin word editor. The first attested
occurrence of the noun is from 1649, while the verb makes its first
appearance only in 179115. However, back-formation is very much
alive today and is responsible for more recent formations like e.g.
the verbs (to) orientate from orientation, (to) obsess from obsession,
and (to) opt (for something) from option.
Back-formation bears a superficial resemblance to clipping in that
both serve to create shorter words from longer ones, but the two
processes differ both with regard to their output, to the elements
that are removed from the original word, and to their function. The
output of back-formation is a new lexeme with a different meaning
and word-class membership than the original. The output from
clipping is a word that is shorter than the original but belongs to
the same word class and has the same basic meaning. The output

words are synonymous with the longer original forms, but occupy a
quite different stylistic niche: clippings like e.g. fab and demo are
entirely appropriate in laid-back communication as when you’re
chatting with your friends, but only the full forms fabulous and
demonstration will do on more formal occasions.
A further difference is that clipping may remove parts of a word
that are not meaning-carrying as in the examples biog and info cited
above. As we have seen, back-formation operates in terms of real or
perceived suffix morphemes: it is a process that changes what is per-
ceived as a complex word (normally a noun) into a shorter and
more basic one, normally a verb.
The examples given of back-formation so far have all involved
single words: from orientation we derived orientate, from obsession
we got obsess. To these we can add further examples, like to liaise ‘to
establish a working relationship with’ from liaison, to laze from the
adjective lazy, to peddle from the noun peddlar and to televise from
television.
However, in present-day English, back-formations are much more
often formed from compounds than from single words, usually by
the removal of the nominal suffixes -er, -ation and -ing from the sec-
ond part of noun compounds. (As nouns in -er normally also are
found as -ing-forms, it is often impossible to determine exactly
which of the two was the original). The great majority of the new
back-formed verbs are hyphenated.
The following list contains examples of fairly recent verbs result-
ing from the application of back-formation to noun compounds:
baby-sit<baby-sitter, back-form<back-formation, bird-watch<bird-
watcher (or from bird-watching), blow-dry<blow-drying/blow-dryer,
bottle-feed<bottle-feeding, carbon-date<carbon-dating, co-drive<co-
driver, gift-wrap<gift-wrapping, headhunt<headhunter, house-
keep<house-keeper, hunger-strike (verb)<hunger strike (noun), knife-
murder<knife-murderer, shoplift<shoplifting, stage-manage<stage-man-
ager.
Sometimes there is phonological evidence that back-formation
has taken place, for instance in the verbs contracept, cohese, intercept
and self-destruct, which have obviously been created from contracep-
tion, interception, cohesion and self-destruction. If that had not been
the case, the forms of these verbs would be *contraceive, (the actu-
ally existing) interceive, cohere, and *self-destroy.

Blends and blending

Blending is a kind of word mixing in which elements from two sep-
arate words are fused, as in e.g. Oxbridge (Oxford+Cambridge)—used
when the two universities are regarded as a single phenomenon as in
an Oxbridge education, —sexploitation (sexual+exploitation) and Clin-
tonomics (Clinton+economics) i.e. Clinton’s approach to economics.
The result of the process of blending is called a blend. As the three
examples above show, blends exhibit three kinds of fusion. In the
first example, parts of the fused words (Ox- and -bridge) have simply
been joined together, in the second there is partial overlap between
sex and exploitation, and in the third the first part—Clinton—has
been kept as a full word, to which -onomics from economics has been
added. The overlapping type is by far the most common one.
When it comes to the meanings of blends, it turns out that in the
majority of such formations, the two words that are fused are not
treated as being equally important16. Instead, the second word is
given more importance than the first: formally, it tends contribute
a longer and more distinctive element, and semantically it is nor-
mally the head of the construction making up the definition of the
blend: sexploitation is ‘exploitation involving sex’, Clintonomics is
‘economics (economic policy) as practised by Clinton’.
Further examples of blending, most of which support the claim
about the importance of the second element, are given below:
Amtrak (for American track; a US railway line),
ballute (balloon and parachute),
breathalyser (analyser using breath),
carbecue (barbecue to which you go in a car),
chunnel (tunnel under the Channel),
cremains (cremated remains),
dawk (dove and hawk used figuratively),
eargasm (ear and orgasm),
edutainment/infotainment education/information as entertain-
ment,
elint (electronic intelligence),
floptical (a floppy disk using an optical track),
glasphalt (glass and asphalt),
Globowood (global and Hollywood),

guestimate (guess and estimate),

hazchem (hazardous chemicals),
hi-fi (high fidelity),
Intelsat (Intelligence satellite),
kidvid (kids’ video),
misper (‘missing person’),
sci-fi (science fiction),
slanguage (slang and language),
smog (smoke and fog),
stagflation (stagnation and inflation),
twigloo (‘igloo-like building made of twigs’),
On occasion, second elements in blends take on a suffix-like char-

acter. They develop a meaning of their own and are used with a cer-
tain amount of productivity in new combinations. In this category
we find e.g. the second element -(o)nomics from economics, first used
in Reaganomics (‘economics as practiced by President Reagan’) and
later also in Nixonomics Thatchernomics and Clintonomics; no doubt,
the time will soon be ripe for Bushonomics. Other examples of such
semi-suffixes are the -el in hotel, motel, boatel, floatel and -(o)holic,
originally from alcoholic but now also found in e.g. chocoholic,
shopoholic, workaholic. Compare also the second elements in com-
pounds that have developed suffix-like characteristics like -babble,
-gate, -man, -person, -speak discussed in Chapter 7.
Reduplicative compounds
Reduplicative compounds are sometimes more aptly called ‘echoic
words’17. These constructions are combinations of two words that
somehow echo each other in that the second word repeats the first
with a slight change in pronunciation/spelling. There are two cate-
gories. In one the change affects the vowel in the first word, while
the consonants are kept intact, as for instance in flip-flop, riff-raff,
shilly-shally, tick-tack/tick-tock, wishy-washy, zig-zag.
In a second, equally common type, the vowels stay the same, but
one of the consonants is changed, as for example in hanky-panky,
higgedly-piggedly, hobnob, hoity-toity, nitty-gritty, pitter-patter, roly-poly,
teeny-weeny/teensy-weensy.

The meanings of these words are often sound-imitating (ono-

matopoeic): flip-flop refers to a kind of light sandal and to the noise
caused by walking in such sandals; for some reason it also has the
figurative meaning ‘to change one’s mind abruptly’. Riff-raff is a
term for undesirable people (cf. Fawlty Towers), to shilly-shally is to
fail to make up one’s mind or to act resolutely. The basic meaning
of wishy-washy is ‘weak’, ‘watery’, but in present-day English it is
normally used to describe behaviour or a state of mind character-
ised by lack of resoluteness. Tick-tack and zig-zag are self-explana-
tory.
As for the second type, hanky-panky means ‘improper behaviour’,
often to do with money or sex, higgedly-piggedly means ‘in disorder’,
to hobnob with certain people is to be friendly with them or meet
them socially; there is more than a hint that the people in question
are socially superior to the ‘meeter’, a hoity-toity person is snobbish.
The nitty-gritty is the most essential and often difficult aspect of
some undertaking, pitter-patter is the sound of quick steps, a roly-
poly is a kind of pudding and by extension also a term used about
plump people. Teeny-weeny (in American English teensy-weensy)
means ‘very small’ or ‘at all’ as in e.g. Doesn’t he have any money?—
Not a teeny-weeny bit!
It would be easy to find additional examples of reduplicative
compounds and it appears that new compounds of this type are still
being formed, especially for informal use. While the examples
above mostly contain words that are not themselves recognised
English lexemes, there are also reduplicative constructions involv-
ing established English lexemes, for example brain-drain, flower
power, high-five, hip-hop, hotshot, jet set, mumbo jumbo, multiculti,
shock jock (a disk jockey who expresses opinions in an offensive
manner) sky-high as in blow the place sky-high.
A recent example of reduplicative compounds is Aga saga, defined
in The Oxford Dictionary of New Words as “popular novel involving
an Aga stove focussing on semi-rural middle’class family life’. From
same source we get happy-clappy/happyclappy, an ‘informal and
slightly disparaging term for a member of a Christian group whose
worship is marked by enthusiasm and spontaneity’.
Finally there is an interesting example of a reduplicative com-
pound based on both clipping, suffixation and reduplication, i.e.
argy-bargy ‘quarrel, argument’. We can distinguish three distinct

phases in the formation of argy-bargy: first argument has been back-

clipped to arg, then the ‘familiarizing’ informal suffix -y is attached
to arg and finally the resulting argy has been reduplicated as bargy,
the whole process yielding argy-bargy.
Rhyming slang
Rhyming slang is a type of slang that can be traced back to the 18th
century and is particularly associated with Cockney English. The
formation of expressions of this type of slang is a two-stage process.
The first stage involves finding a two-word phrase, the last word
of which rhymes with the target word, i.e. the word whose meaning
we want to express. (In actual fact a lot of such two-word phrases
have already been created and are unlikely to be replaced by other
phrases). The two-word phrase is then used instead of the target
word to create a kind of code in accordance with the description
below:
Phrase Target word

Cain and Able table
china plate mate
butcher’s hook look
apples and pears stairs
But the process above may then be taken one step further in that
the final word in the phrase—the one rhyming with the target
word—is dropped:
Phrase Target Modified phrase

Cain and able table Cain
china plate mate china
butcher’s hook look butcher’s
apples and pears stairs apples

The coding is now complete and users of rhyming slang can pro-
duce utterances supposedly only understood by their fellow users,
like e.g. Take a butcher’s at that car!, Bring your china!
One of the chief uses of rhyming slang has been to provide
euphemisms for offensive or taboo words. Thus Richard the Third is
used instead of turd (‘shit’), Bristol Cities for titties, Hampton Wick for
prick, Berkshire/Berkley Hunt for cunt18.
The above examples of rhyming slang are of long standing and
you sometimes get the impression that this type of word-formation
is no longer productive. However, as the following examples indi-
cate, it is clearly the case that—to a certain extent at least—rhyming
slang is still productive: Adam and Eve ‘leave’, bangers (and mash)
‘cash’, Britney Spears ‘beers’, do bird ‘do time’ (in prison), from
(bird)lime, which rhymes with time, John Major ‘pager’, Steve
McQueen’s ‘jeans’ and (whiskey and) soda ‘mobile phone’ (soda
rhymes with Voda as in Vodafone)19.


References
References
Valerie Adams. 1973. An Introduction to Modern English Word-forma-

tion. London:Longman.
Valerie Adams. 2001. Complex Words in English. London: Longman.
Jean Aitchison. 1987 Words in the Mind. Oxford: Blackwell.
Albert C. Baugh and Thomas Cable. 1978. A History of the English
Language. Third ed. London, Boston and Henley: Routledge.
Laurie Bauer. 1983. English Word-formation. Cambridge etc.: CUP.
Laurie Bauer. 1988. Introducing Linguistic Morphology. Edinburgh:
Edinburgh University Press.
Laurie Bauer 1994. Watching English Change. London: Longman.
Laurie Bauer. 2001. Morphological Productivity. Cambridge: Cam-
bridge University Press.
Robert Beard. 1996 in ‘Once More on the Analysis of ed-adjectives’
Journal of Linguistics 12:2, 155–157.
Leonard Bloomfield. 1933. Language. New York: Holt.
Bloomsbury Dictionary of Contemporary Slang. 1990. Tony Thorne.
London: Bloomsbury Publishing Ltd.
Herbert E. Brekle & Dieter Kastovsky (eds.) 1977. Perspektiven der
Wortbildungsforschung. Wuppertaler Schriftentreihe Linguistik 1.
Bonn: Bouvier.
G.L. Brook. A History of the English Language. Second impression.
1960. London, Reading: Andre Deutsch.
Neil Burton-Roberts. 1998. Analysing Sentences 3 ed. London: Long-
man.
Cambridge International Dictionary of Idioms. 1998:Cambridge: Cam-
bridge University Press.
J. Chapman. 1986. New Dictionary of American Slang. New York:
Harper & Row.
Herbert E. Clark & Eve V. Clark. 1979. ‘When Nouns Surface as
Verbs’. Language 55:4, 767–811.

References
The Cambridge Grammar of the English Language (CGEL). 2001. Ed.

by Rodney Huddleston & Geoffrey K. Pullum. Cambridge: CUP.
David Crystal. 1997. The Cambridge Encyclopedia of the English Lan-
guage. Cambridge: Cambridge University Press.
Dictionary of Clichés. The Facts on File Dictionary of Clichés. 2001.
Christine Amner. New York: Checkmark Books.
A.C. Gimson & Alan Cruttenden. 1994. Gimson’s Pronunciation of
English. 5th edition. Revised by Alan Cruttenden. Fifth edition.
London, New York: Edward Arnold.
Robert A. Hall. 1974. The Comic Style of P.G.Wodehouse. Hamden,
Conn.: Archon Books.
Paul J. Hopper & Elizabeth Closs Traugott. 1993. Grammaticaliza-
tion. Cambridge: CUP.
Richard Hudson. 1976. ‘Problems in the analysis of ed-adjectives’.
Journal of Linguistics 11:1:69–72.
Geoffrey Hughes. 2000. A History of English Words. Oxford: Black-
well.
Howard Jackson & Etienne Zé Amvela. 2000. Words, Meaning and
Vocabulary. London & New York: Cassell.
Otto Jespersen. 1942. A Modern English Grammar on Historical Princi-
ples. Part VI: Morphology. Copenhagen: Einar Munksgaard.
Francis Katamba. 1994. English Words. London & New York:
Routledge.
Dieter Kastovsky. 1986. ‘The problem of productivity in word for-
mation’ Linguistics 24, 585–600.
Dieter Kastovsky. 1995. ‘The Syntactic Aspects of Word-Formation:
Where Are We Today?’, In Gunnel Melchers & Beatrice Warren
(ed.s) Studies inAnglistics. Stockholm Studies in English LXXXV.
Stockholm: Almqvist &Wiksell International.
Dieter Kastovsky 1986. ‘Word-Formation and Pragmatics’. In A Year-
book of Studies in English Language and Literature.Pp. 63–78. Her-
ausgegeben von Otto Rauchbauer.
Göran Kjellmer 1984, ‘Why great:greatly but not big:*bigly‘?’ On
the formation of English adverbs in -ly, Studia Linguistica 38, 1–19.
Göran Kjellmer. 2000‚ ‘Potential Words‘. Word 51/2:208–28.
George Lakoff and Mark Johnson. 1980. Metaphors We Live By. Chi-
cago: University of Chicago Press.
Roger Lass. The Shape of English. History and Structure.1987. London,
Melbourne: J. L. Dent and Sons Ltd.

References
Magnus Ljung. 1970. English Denominal Adjectives. Gothenburg

Studies in English 21. Lund: Studentlitteratur.
Magnus Ljung, 1974. A Frequency Dictionary of English Morphemes.
Gothenburg Data Linguistica Series. Stockholm: Almqvist & Wik-
sell.
Magnus Ljung. 1976 a. ‘Instrumental Verbs’ in Brekle & Kastovsky
1976.
Magnus Ljung. 1976 b. ‘Ed adjectives revisited’. Journal of Linguistics
12:2:159–68.
John Lyons. 1995. Linguistic semantics. Cambridge: Cambridge Uni-
versity Press.
Tom MacArthur 1992. The Oxford Companion to the English
Language. Oxford, New York: Oxford University Press.
Hans Marchand. 1969. The Categories and Types of Present-Day Eng-
lish Word-Formation. Second ed. Munchen: Beck’sche Verlags-
buchhandlung.
W.J. Meys 1975 Compound adjectives in English and the ideal speaker-
listener. Amsterdam: North Holland.
David Minugh. 1991. On Pronounceable English. A Handbook of Eng-
lish Pronunciation. Stockholm Papers in English language and Lit-
erature. University of Stockholm.
Oxford Dictionary of Collocations. 2002. Oxford, New York etc.:
Oxford University Press.
The Oxford Dictionary of New Words. 1997 Elisabeth Knowles with
Julia Elliott. Oxford, New York: OUP.
The Oxford English Dictionary (OED).
Eric Partridge. 1961. A Dictionary of Slang and Unconventional Eng-
lish. New York: MacMillan.
Steven Pinker. 1995. The Language Instinct. New York: Harper Peren-
nial.
Randolph Quirk et al. 1985. A Comprehensive Grammar of the English
Language. London: Longman.
Stuart Robertson and Frederic G. Cassidy. 1954. The Development of
Modern English. Englewood Cliffs, N.J.: Prentice-Hall.
John Saeed. 1997. Semantics. Reprint 2001.Oxford:Blackwell.
Robert Stockwell & Donka Minkova. 2001. English Words.History
and Structure. Cambridge: CUP.

References
Elizabeth Closs-Traugott and Mary Louise Pratt. 1980. Linguistics for

Students of Literature. San Diego, New York, etc.: Harcourt Brace
Jovanovitch.
Beatrice Warren. 1978. Semantic Patterns of Noun-Noun Compounds.
Gothenburg Studies in English 41.
Beatrice Warren. 1992. Sense Developments. Stockholm Studies in
English LXXX. Stockholm: Almqvist & Wiksell International.
John Wells. 2000. The Longman Pronouncing Dictionary. 2nd ed. Lon-
don: Longman.

Notes
Notes
Chapter 1
1 The American Heritage Dictionary of English.
2 The notion of sign and the distinction between motivated and
non-motivated signs was first formulated by the Swiss linguist
Ferdinand de Saussure in the early 1900s (cf C.Bally and Albert
Sechehaye (eds.) Cours de Linguistique Générale, 1st ed. 1916,
Paris: Payot. Translated into English as Course in General Linguis-
tics New York: Duckworth.
3 These definitions are from The New Oxford Dictionary of English.
4 The place of English and the other Germanic languages in the
larger context of the Indo-European language family is tradi-
tionally illustrated by a family tree diagram like the one below
adapted from Robertsson-Cassidy 1954:33.
Indo-European
Armenian Hellenic Albanian Italic Celtic Germanic Balto– Tokharian

Slavonic
West North East

(Gothic)
West East
(Dutch, English, (Danish

Frisian, German) Swedish)
5 The data about the barking of dogs comes from the list Sounds of
the World’s Animals Copyright Catherine Ball at
www.georgetown.edu/cball/animals/cat.html). Another example
making the same point may be found in Pinker (1995:152), i.e.

Notes
the fact that pigs in English-speaking countries go oink but in

Japanese they say boo boo.
6 Definitions of the morpheme may be found in most introduc-
tions to linguistics. See also e.g. Katamba 1994:32–52, Ljung
1974:12–23, MacArthur 1992:670. For a classic definition of
‘morpheme’ cf. for example Bloomfield’s Language pp. 161–8.
7 Cf. Ljung 1974:197.
8 For more information about immediate constituents and the
analysis of complex words, see Bauer 1988:144-46, Pinker
1995:133–36.
9 For further reading about words as types and tokens, see e.g.
Ljung 1990:3–4, Lyons 1995:49, 53, 176.
10 This example has been taken from Bauer 1988:7.
11 Obviously the number of types can never be greater than the
number of tokens. In very short texts there may be the same
number of types and tokens.
12 Definitions and discussions of the lexeme may be found in many
books on semantics and vocabulary like e.g. Lyons 1995:51, 56–
7, Saeed 1997:55–9. Note that although inflectional suffixes are
ignored when forms are brought together as one lexeme, the so-
called derivational suffixes are not: bank and banker are different
lexemes, as are slow and slowly, kind and unkind (cf. Chapter 3
for a definition of derivation and derivational).
13 For homonym and homonymy see e.g.Alm-Arvius 1998:59–61,
Lyons 1995:54–60, Saeed 1997:63–4.
14 In addition, many dictionary entries contain different lexemes.
See for instance the entries for the word charge in The American
Heritage Dictionary of the English Language and The New Oxford
Dictionary of English.
15 See the discussion of speakers’ word storage in Aitchison
1987:108–117, Pinker 1995:146.
16 Teapot, ashtray, wine glass and other compounds vary in the way
they are spelt. Teapot may also be spelt tea-pot and tea pot, and
ashtray and wine glass are also found with the spellings ash-tray
and wineglass. Cf. Chapter 7 for more comments on the spelling
of compounds.

Notes
17 Although neither clichés nor idioms can be clearly and exhaus-

tively defined, there is no shortage of ‘dictionaries’ of both. Two
recent books of this kind are Ammer 2001 and the Cambridge
International Dictionary of Idioms 1998.
18 Metaphor is perhaps the most important figure of speech, for
which see e.g. Alm-Arvius 1998: 58, Crystal 1997:70, Saeed
1997:302–8, 336. For a different approach to metaphor, cf.
Lakoff and Johnson 1980.
19 Sense and reference are central linguistic terms for which see e.g.
Alm-Arvius 1998:34–7, Lyons 1995:204, 225, Saeed 1997:12,
23–6, 30–32.
20 See Saeed 1997:23–4 for the difference between denotation and
reference; as Saeed points out (p. 27) the term extension is some-
times used instead of denotation. Cf. also Alm-Arvius 1998:35,
Lyons 1995:78–80.
21 For more information about sense relations cf e.g. Alm-Arvius
199846-61, Lyons 1995:79–80, Saeed 1997:62–5, 65–71.
22 For this figure, see for instance Baugh and Cable 1978:177,
Katamba 1994:208.
23 There was little or no direct borrowing from Greek into Old or
Middle English, but many of the loans from Latin or French had
originally been borrowed into Latin from Greek. Many scientific
and learned Greek terms borrowed later have been Latinized in
various ways.
Chapter 2
1 Extensive discussions of word classes will be found in Quirk et.
al. 1985 and The Cambridge Grammar of the English Language. Cf.
also Crystal 1997:206–213, Lyons 1968:317–29, Pinker
1995:105–6.
2 Assessments of the total number of words in English (and other
languages) vary wildly, as do estimates of how many words dif-
ferent categories of speakers know. For some of these, see Crys-
tal 1997: 118–125 and Pinker 1995:150–51.

Notes
3 Crystal 1997:213.
4 This quote is from Rombauer-Becker Joy of Cooking 1962 Bobbs
Merrill, Indiananpolis, Indiana.
5 For a full and interesting discussion of different types of class
membership cf. Taylor Linguistic Classification 1995 2nd ed
Chapter 2–3
6 These and the other criteria in this chapter are based on the
word-class classification in Greenbaum 1996, but basically
reflect mainstream grammatical thinking on these matters.
7 Cf. also the discussion in Chapter 6, where it is pointed out that
the word-formation process known as conversion may override
suffixation: a noun like e.g. position may be used as a verb as in
e.g. They positioned the table between the chairs.
8 There are also verbs followed by a direct object and a so-called
objective predicate complement, for example call, consider, regard as
in e.g. She called me a fool, We consider him a genius, They regard
this as a terrible mistake.
9 Unlike the other word classes, the adverb class is divided
between the open and the closed type: there are sets of adverbs
that are closed like e.g. those denoting definite time (now, then)
and place (here, there). But there are also those that are clearly
open, in particular the class of manner adverbs formed by the
addition of the suffix -ly.
10 The definition of determiner used here is along traditional lines;
cf. e.g. Greenbaum 1995, Burton-Roberts Analysing Sentences
(3rd ed.) page 154, Quirk et.al.1985:253–7. Huddleston and Pul-
lum 2001 The Cambridge Grammar of the English Language.
11 The standard work on grammaticalization is the book Grammat-
icalization by Paul J. Hopper and Elizabeth ClossTraugott CUP
1993.
12 These examples are from Hopper-Traugott 177–178.

Notes
Chapter 3
1 As the discussion of words like defrost, unsaddle on pp. 63 in
Chapter 4 will show, there are also prefixes that change the
word class of the stems they are attached to.
2 Certain suffix forms turn up both in inflection and in deriva-
tion, in particular -ed and -ing, which are found both in inflec-
tional verb forms (They talked, They were talking) and in derived
nouns and adjectives like flat-footed, carpeting, etc. Adjectives
like amused and amusing (as in an amused smile, amusing stories)
and nouns like a painting, the painting of the walls pose a partic-
ular problem in this respect (for discussion of these questions cf.
Chapter 2 p. 41, Chapter 6 p. 116).
3 MacArthur 1992:263
4 All languages have rules—sometimes called phonotactic rules—
determining what sound and letter-sequences are permitted in
the language. For information about the phonotactic rules for
English, see Traugott-Pratt 1980:61–4 and Cruttenden-Gimson
1994:216–23.
5 Cf MacArthur The Oxford Companion to the English Language
p. 876
6 This definition is from The Oxford Dictionary of English.
7 A related but not identical question is under what circum-
stances a potential word may become an actual one. Cf. Kjellmer
2000 for a discussion of this.
8 According to Saeed 1997:71, modern dictionary makers do not
always follow this principle but simply list all derived forms
even if their meaning is predictable.
9 For more information about lexicalization, see e.g. Bauer
1983:48–61.
10 The distinction between syntactic repackaging and labelling
used here has been borrowed from Kastovsky 1986 (and else-
where). Kastovsky uses syntactic recategorization instead of syn-
tactic repackaging.

Notes
Chapter 4
1 Most verbs containing ‘removal’ dis- like for instance discourage,
dismantle, dismember were borrowed from French during the
Middle English period.
2 This point is made in Adams (2001:71) who has also supplied
the examples used here. Cf. also Bauer (1983:37–38)
3 For the historical background of English prefixes, see e.g.
Marchand 1969:129–34
4 Stress assignment in English is notoriously difficult to describe.
There are several treatments of English stress in general and
word stress in particular in the literature, for instance Bauer
1983 Chapter 5, Gimson 1994 Chapters 10–12, and Minugh
1991, Chapters 5, 7–9.
5 For extensive discussions of productivity in word-formation, cf.
e.g. Adams 2001 (7–10, 146–53), Bauer 1983, Chapter 4, Bauer
2001, Kastovsky 1986. See also The Cambridge Grammar of the
English Language p. 1630. In this book, productivity in word-for-
mation is regarded as a cline extending from highly productive
to weakly productive processes. No English word-formation rule
is 100% productive, but some, like the rules for adverb-forming
-ly and noun-forming -ness come close.
6 From one of John Cleese’s Monty Python sketches.
7 The principle that negative un- doesn’t combine with simple
nouns is (very occasionally) violated in formations like e.g. un-
person/unperson, which means roughly ‘a (usually politically)
impossible person’ i.e. someone who is not accepted by the
establishment.
8 In George Orwell’s pessimistic vision of a totalitarian Britain in
the novel 1984, one of the prominent features is the simplifica-
tion of English word-formation permitting among other things,
words like ungood and even double-plus-ungood meaning ‘very
bad’. See Crystal 1997:135 for an account of Orwell’s fascinating
‘Newspeak’.
9 Cf. Kastovsky 1986:594, who provides the example ‘The Time-
Patrol also had to unmurder Capistrano’s great-grand-mother,
unmarry him from the pasha’s daughter in 1600, and uncreate

Notes
those three kids’ from the science-fiction novel Up the Line by

Robert Silverberg 1969/1975, p. 152, London: Sphere Books. Cf.
also e.g. unbreak (as in Unbreak my heart from a pop song) and
the unusual unsay ‘withdraw, retract’ something previously
said.
10 Unexpected formations like dedog are one of the distinctive fea-
tures of style of P. G. Wodehouse. In addition to dedog—from
The Mating Season (1949)—there is also the verb detrouser from
Frozen Assets (1964).
11 There is another type of combination with anti-, which is used
to form nouns like e.g. anti-hero, anti-matter meaning ‘person
who is the opposite of a hero’, ‘something that is the opposite of
matter’. Here anti- carries primary stress.
12 Newsweek April 2002
13 Heard on BBC World News, March 13, 2002.
Chapter 5
1 For more detailed accounts of the interplay between suffixation
and word stress, cf. The Cambridge Grammar of the English Lan-
guage (CGEL pp. 1669–72 and Bauer 1983:116–20), Minugh
1991, Lass 1987:204–10.
2 The account given here is based on Bauer 1983:112–22.
3 Cf. Ljung 1974:26.
4 These examples are from Adams (2001:64)
5 The examples creditcarditis and televisionitis are from NODE and
AHD respectively. The rest of the examples have been taken
from The Independent on CDROM from 1992. It turns out that
this use of the -itis suffix goes back to at least the early 1900s:
according to the Supplement to the Oxford English Dictionary,
Asquith used the term fiscalitis as early as 1903, when writing in
The Westminster Gazette.
6 These examples are from Adams 2001:62.
7 Cf The Cambridge Grammar of the English Language p. 1691,
Bauer 1983:268, and Marchand 1969:245–7.

Notes
8 From Newsweek May 20, 2002, p.65.

9 The i-suffix is discussed i Marchand 1969:354–5 and in Bauer
1983:253–5.
10 The meaning of nouns in -ness and -ity is traditionally said to be
‘state, quality, condition’, as in e.g. Marchand 1969 pp. 312 and
334. However, these nouns commonly mean ‘fact’ or ‘extent’.
Thus e.g. His carelessness/sincerity amazed me will normally be
taken to mean either ‘The fact that he was careless/sincere
amazed me’ or ‘The extent to which he was careless/sincere
amazed me’.
11 The greater productivity of -ness is shown by the fact that it may
combine with rather unexpected stems, for example proper
nouns as in the following quote from Ford Madox Ford Parade’s
End (Penguin Classics, 1982 p. 54: It was perhaps partly her Edin-
burgh-ness.
12 Cf Bauer 1983:222, Adams 2001:32
13 This example is from Adams (2001:28)
14 For examples of the creative use of -ee, see Kastovsky 1986:598.
15 See Bauer 1994:40–47.
16 For a discussion of the possessive -ed suffix, see Beard 1976,
Hudson 1976, Ljung 1976.
17 We are talking about what is normally the case: we expect people
to have heads and hands, cars to have wheels, etc. Obviously, in
less normal cases, all such possessions may be lost.
18 What is inherently possessed may also be properties like size,
and shape, as the examples shoebox-sized and duck-shaped show.
19 From The Independent on CDROM 1996.
20 Cf. Ljung 1970: 64–8, 76–91.
21 Newsweek Aug 5, 2002, p. 54.
22 The Daily Telegraph, July 2, 2002
23 The jaw-droppingly example is from The Times Literary Supple-
ment, Nov.1, 2002, p.12. Like the earlier example unputdowna-
ble, jaw-dropping and jaw-droppingly are stylistically marked
words. However they seem to be on the increase in certain jour-
nalistic genres.

Notes
24 For further discussion of the limits of ly-suffixation, see Kjellmer

1984.
Chapter 6
1 Accounts of conversion will be found in e.g. Marchand 1969,
Quirk et.al. 1985: Clark and Clark 1979, Adams 2001 and the
Cambridge Grammar of Modern English
2 MacArthur 1992:263.
3 In most cases of conversion it is possible to determine its direc-
tion i.e. what is the input word and what is the output. Thus in
word pairs like hammer(n): hammer (v) and fax(n): fax(v) it rea-
sonable to assume that the verb has been formed from the
noun, not the other way round. But there are also many
instances of homonymous word pairs for which such decisions
are impossible, for example hate, love, count.
4 For discussion of such categories (cf. e.g. Adams 2001:22, Bauer
1983, Clark and Clark 1979).
5 Both definitions from The New Oxford Dictionary of English.
6 From le Carré Smiley’s People. 1980. London and Sydney:Pan
Books p. 172.
7 As mentioned in Chapter 2 (p. 41), verb to adjective conversion
is probably the best explanation for adjectives like astonished,
amusing, breath-taking, etc. As noted in The Cambridge Grammar
of the English Language, p.1644 this is an unusual case of conver-
sion using inflected forms as its point of departure.
8 From the Financial Times, March 11, 2003, p. 3.
9 For further examples of such verb:noun pairs, see Minugh
1991:227–8.
10 Note that partial conversion from phrasal verb to nouns with
‘agentive’ or instrumental meaning may take many different
forms. Thus in addition to the type drop-out, drop-in, there are
the types in-drop, dropper-in and dropper-inner (the final example
usually found only in informal spoken English). Cf. Bauer
1983:288–90.
11 This example has been borrowed from Pinker 1995:129.

Notes
Chapter 7
1 There is a fair amount of disagreement among linguists con-
cerning the proper definition of ‘compound’in English. For an
early treatment, cf. for example Jespersen 1942, A Modern Eng-
lish Grammar Part VI pp. 134–183 and Marchand’s classic Cate-
gories and Types of Present-Day English Word-Formation. 1969:11–
30. More recent treatments will be found in Bauer (1983:106–9),
Bauer (1988:33–7, 100–104), Quirk et.al. (1985:1330–32, 1567–
70), The Cambridge Grammar of the English Language (2001:449–
51, 1644–66), and Adams (2001: 4–5). Note also that the so-
called ‘reduplicative compounds’ like argy-bargy, riffraff etc. are
dealt with in Chapter 9.
2 For this point cf. The Cambridge Grammar of the English Language
p. 447.
3 There is no total agreement about the role of stress in noun com-
pounds. Thus Bauer 1983:104–105 and The Cambridge Grammar
of the English Language (p. 451) have no faith in the role of stress
placement as a predictor of compound status. The latter argues
that the compound:phrase distinction can only be established
on syntactic grounds (pp. 449–50). Others are in favour of
regarding stress as at least an important tool in the recognition
of noun+noun compounds, for instance Marchand (1969:28–
29), Quirk et. al (1985:1568–69) and Adams (2001:80).
4 All the examples of compounds given here consist of just two
words. However, it is clear that both compounds and phrases
may consist of three or more words, for example weapons inspec-
tion crisis, or small arms factory owner. Such compounds have to
be analysed on several levels, the first into weapons inspection
and crisis and then into weapons and inspection, the second into
small arms factory and owner, then into small arms and factory
and finally into small and arms. (Small arms ‘firearms that can
be carried in the hand’).
5 For syntactic compound (sometimes also called synthetic com-
pound), see Bauer 1988:36, Adams 2001:78–9, The Cambridge
Grammar of the English Language p. 1652n

Notes
6 They may also have other agentive or activity-denoting suffixes

like -ent or -or and -ance or -ion, as in insect repellent, cassette ejec-
tor, slum clearance and word deletion
7 For a comprehensive treatment of compound adjectives, see
Meys 1975.
8 In the early 1990s, jaw-dropping was quite an uncommon word:
in the 100 million word British National Corpus, there are four
instances: a jaw-dropping experience (twice), jaw-dropping remarks
and a jaw-dropping piece of strategic diplomacy. However, both the
adjective and the adverb jaw-droppingly are clearly on the
increase. Thus The Independent on CDROM from 1998 has 24
instances of jaw-dropping, and 5 of jaw-droppingly. Those figures
have gone up to 33 for the adjective and 9 for the adverb in the
Independent on CDROM from 2001.
9 (From the Times Literary Supplement, Nov 1, 2002, p. 12.
10 The fact that the word class of such adverb compounds is differ-
ent from that of their heads raises the question whether they
should be regarded as exocentric compounds (cf. pp. 127–129)
11 In addition to the compound adverbs mentioned in this sec-
tion, there are occasional rather surprising adverbs like e.g. drop-
dead used as an intensifier in drop-dead gorgeous (also used as a
premodifier in e.g. Chicago—the drop-dead musical!
Chapter 8
1 They are not restricted for linguistic reasons—classical Latin and
Greek had very rich vocabularies—but because latter-day speak-
ers of English cannot be assumed to know all the classical forms
that could in principle be used to form neo-classical com-
pounds.
2 I owe the terms ICF and FCF to Bauer 1983. The vast majority of
the combining forms are initial: of the 550 plus combining
forms listed in The New Oxford Dictionary of English, more than
400 are initial forms.
3 Kleptocracy was found in Newsweek.
4 This is the term used in MacArthur 1992:233

Notes
Chapter 9
1 For discussions of back-formation, cf. Adams 2001:136–8,
Hughes 2000:54
2 From the Bloomsbury Dictionary of Contemporary Slang
3 The example of uber has been taken from The Oxford Dictionary
of New Words
4 There is a very extensive discussion of the suffix(es) -nik in
Bauer 1983:259–66.
5 The information about the suffix -bot is from The Oxford Diction-
ary of New Words
6 The information about -(s)ville is from Chapman 1986 (New Dic-
tionary of American Slang: Harper & Row)
7 From The Independent News on CDROM 1992.
8 From NEWSWEEK March 25, 2002, p. 58.
9 For example Beatrice Warren in Warren 1992.
10 in e.g. G.L. Brook A History of the English Language pp. 177–180,
Bloomfield Language p. 426 to give just a few sources)
11 This is the definition of box in The New Oxford English Dictionary.
12 The change in meaning of gay from ‘carefree’ to ‘homosexual’
and ‘homosexual man’ became established in the 1960s accord-
ing to The New Oxford Dictionary, The American Heritage Diction-
ary of the English Language, Crystal 1997:138.
13 See the entry for SAD from in the Oxford Dictionary of New
Words.
14 The clipping blog and its derivatives blogger, blogging was first
found in Newsweek May 27, 2002, p. 79. Later issues of News-
week also had blogoshere and blogspeak.
15 According to the OED
16 Cf. Adams 2001:139.
17 This term is from MacArthur 1992:854-55. Discussions of redu-
plicative compounds will also be found in Bauer 1983:213,
Marchand 1969, $ 8.2.1, Adams 2001 127-29, Katamba 1994:
79, Jespersen 1942, part VI pp 173-183. Cf. also Hughes 2000:61

Notes
who mentions forms from Yiddish English (Yinglish) like Oedi-

pus-schmoedipus.
18 Cf Hughes 2000:48, Partridge 1961.
19 I am grateful to Tanya Fitzgibbon for the examples of modern
rhyming slang.

Notes

Discussion of exercises
Chapters 1–8
There is no discussion of exercises of a fact-finding nature. This is
indicated by means of a dash (–).
Chapter 1
1 In motivated signs there is a natural connection between form
and meaning while there is no such connection in non-moti-
vated (arbitrary) signs. Flag signals are usually non-motivated
and so are most human sign languages, although they also have
certain motivated characteristics. Animal territory marking uses
motivated signs.
2 Grammar is concerned with changes in form with predictable
effects on the meaning of an utterance. This makes it possible to
write rules for grammar. Vocabulary is concerned with the rela-
tion between the form and meaning of linguistic signs. In a
vocabulary consisting only of simple words, that relation is arbi-
trary and consequently not predictable, and no rules can be
written about it. For non-simple words see the answer to ques-
tion 4 below.
3 Onomatopoeic words are intended to imitate certain sounds
like e.g. animal calls. Sound symbolic words on the other hand
contain (sequences of) sounds supposed to be associated with
certain meanings.
4 Unlike simple words, words created by word-formation rules
have meanings that are partly predictable: we know, for

instance, that words ending in -ation are highly likely to be

nouns formed from verbs and to express the activity or result of
carrying out what the verb denotes. In the same way we form
informed guesses about words ending in -er (likely to be agent
nouns), -ly (probably manner adverbs), etc.
5 A morpheme is the smallest meaningful form in a language i.e.
a form that cannot be further cut up without becoming mean-
ingless. Most morphemes in English are free, i.e. they can appear
on their own as words (read). A limited number are bound, i.e.
they can only occur together with a stem that is itself a word.
Bound morphemes in English are either prefixes or suffixes or
so-called combining forms (see Chapters 4, 5 and 8).
6 representative recalling untruthfulness
represent ative recall ing untruthful ness
re call un truthful
truth ful
7 The type:token distinction is usually applied only to written

texts, in which words are defined as strings of letters separated
by empty spaces. If we count all such words in a text, we are
dealing with tokens. However, if we only count each individual
letter-combination once (i.e. if we don’t count repetitions) we
are dealing with types. By dividing the number of types by the
number of tokens we arrive at the type:token ratio of the text.
That ratio is often regarded as an indication of text difficulty:
the higher the type:token ratio, the more difficult the text.
8 There are 42 types in Text A and 52 in Text B. The number of
tokens is 53 in A and 60 in B. Dividing the types by the tokens
we arrive at the following TTRs (type:token ratios): A = 0.79, B =
0.87.
9 In Text A the number of lexemes is 41: a and an represent the
same lexeme. In Text B, the number of lexemes equals the
number of types minus 1: the forms be and is are included under

the same lexeme BE. In this count, plural suffixes, past tense suf-
fixes, apostrophe s suffixes and n’t as in Don’t and ‘s as in name’s
have been counted as inflectional endings. The numbers would
be different if we counted n’t as not and ‘s in My name’s not Jack
as an instance of be. We might also wish to count at once as a
single complex lexeme.
10 The non-Germanic languages that have had the greatest impact
on English vocabulary are French, Latin, Greek.
Chapter 2
1 A word class is a class of words sharing certain important char-
acteristics having to do with syntactic function, the grammati-
cal categories expressed, and meaning. These characteristics are
also used to define the different word classes. The semantic
(notional) criterion is problematic because it is vague and less
comprehensive than the others.
2 Content words have denotation. They denote phenomena
found in ‘the real world’ or some imagined world, like people,
objects, processes, actions, qualities and others. Function words
do not have denotation, but are mainly used to link content
words in different ways. Most function words are limited in
number and do not readily accept new members: they are said
to belong to the closed word classes. There is no upper limit to
the number of content words and the word classes they belong
to are accordingly known as open classes.
It is possible to leave out certain function words because they
are predictable: leaving them out does not deprive the text of
essential information. Leaving out the content words, on the
other hand, would make the text incomprehensible.
3 It may for instance be argued that count nouns – especially
those denoting concrete phenomena – are more central to the
noun category than uncountable nouns.
4 Proper nouns lack denotation and are used as names, i.e. they
refer directly to different phenomena. Unless they are recatego-
rised, they do not normally take articles or the plural. (Note that

not all names consist of proper nouns, however). Common

nouns have denotation and can as a rule always take the defi-
nite article. If they are countables they also take the indefinite
article and the plural, but if they are uncountables they accept
neither.
5 They cannot take the indefinite article unless they have been
recategorised as count nouns meaning ‘type of’, ‘kind of’.
6 As indicated above, uncountables may be recategorised as
countables and then get the meaning ‘type of’,’sort of’. In addi-
tion, countable nouns may be recategorised as uncountables
and then get the meaning ‘mass’. If the original countable in
these cases denotes an animal, the recategorised form is usually
assumed to have a meaning linked to food, as in e.g. Would you
like some turkey? – No thanks, but I’d like a little more rabbit. Proper
nouns may be recategorised as common count nouns, as in e.g.
This was a new Susan.
7 Recategorisation resembles conversion in that a new word is
created without formal change. But unlike conversion, recate-
gorisation doesn’t change the word class of the original word:
turkey is still a noun when it is has been turned into an uncount-
able referring to food.
8 These examples show how difficult it is to distinguish between
nouns and adjectives. As the examples make clear, both upper
class, Sydney and 1930s can be used attributively. They may also
be used both in predicative position and with intensifiers as in
Her accent was very upper class, The atmosphere is very Sydney, and
This music is very 1930s.
However, most linguists would still not accept Sydney and
1930s as adjectives for a number of reasons. Upper class is clearly
a phrase and as such cannot be assigned to any word class. Syd-
ney is a typical proper noun, and 1930s looks like a plural noun
phrase which may occur in typical noun phrase positions like
e.g. in the 1930s.
What these examples show is that the ability to occur in
attributive and predicative position is not confined to adjec-
tives, but is open to a vast number of other words and phrases.

We may conclude that three criteria commonly said to single

out adjectives as a word class are not doing a very good job.
9 The adverbs are difficult to characterise in terms of a set of com-
mon characteristics mainly because they fall into different sub-
groups with different functions. Thus some, like the intensifiers,
operate only at phrase level and serve to modify a following
adjective or adverb, others function as adverbials in clauses, and
some express the speakers attitude to what (s)he is saying.
10 Grammaticalization is a cover term for a number of different
types of linguistic change leading to the reanalysis of single
words or combinations of words. Grammaticalization may turn
a content word into a suffix as when Old English lic ‘body’, ‘like-
ness’ was transformed into the suffix -ly found in e.g. manly. It
may also turn a content word into a function word, for instance
in the case of the verb bar developing into the preposition bar
‘except for’. An often cited example of a combination of words
gradually turning into a single word is the development of the
word gonna from be going to.
Chapter 3
1 Complex words: reclassify, unbelievable, Londoner, Bushie(s), Clin-
tonian.
Compound words: water crisis, sun-tanned, arms cache, steam-
roller. In this book, the word Clintonomics is analysed as a blend
(cf. Chapter 9) combining Clinton and the segment -onomics
from the word economics. An alternative analysis is to treat Clin-
tonomics as a suffixed word consisting of Clinton and -onomics.
2 The fact that forms ending in noun suffixes like steamroller and
audition may be converted to verbs indicate that when suffixa-
tion and conversion clash, the latter wins.
3 The problem with noun-verb pairs like love(n):love(v),
fear(n):fear(v) is that it is difficult – without turning to the his-
torical evidence – to determine the direction of the conversion,
i.e. what should be regarded as the starting-point for the con-
version process and what as the result. Other words presenting

the same kind of difficulty are e.g. count, play and whistle and
many others. In such cases it is not worth the trouble to try to
establish the direction of the conversion.
4 Strictly speaking, English does not have the type of word-forma-
tion known as infixation, i.e. insertion of bound morphemes
inside words. However, it has often been pointed out that Eng-
lish has a similar construction, making possible the insertion of
‘swearwords’ like bloody and fucking into words with at least
three syllables, as for example in e.g. absobloodilutely, imfucking-
possible. As the examples show, the insertion place is immedi-
ately before the syllable with main stress in the original word.
The effect of this type of insertion is to add emphasis to certain
words.
5 The permitted consonant combinations – known as clusters – at
the beginning of English words are listed below. Note that the
symbols used are phonetic and that sometimes there are several
spellings corresponding to a single cluster. Note also that certain
rare combinations occurring in loanwords have been excluded.
Permitted word-initial two-consonant clusters: [bj], [bl], [br],
[dj], [dr], [dw], [fj], [fl ], [fr], [gl], [gr], [gw], [kj], [kl], [kr], [kw],
[pj], [pl], [pr], [sf], [sj], [sk], [sl], [sm], [sn], [sp], [st], [sw], [ʃr],
[ʃm], [θr], [tʃ]. Examples of written words with these clusters:
beautiful, black, bride, duke, drive, dwell, few, flat, from, gloat,
grade, Gwendolyn, cute, class, croak, quick, puke, pluck, prim,
sphinx, suit, school, slow, smart, snow, speak, stone, swallow, shrink,
schmuck, throw, cheat.
Permitted three-consonant clusters: [skl], [skr], [skw], [spl],
[spr], [str] as in sclerosis, scream, sprawl, squeamish, spleen, street.
6 We use our knowledge of word-formation productively when
we actually create a word that is new to us. We use it analytically
when we analyse an existing complex or compound word.
7 –
8 Chief among the factors that determine whether a derived,
compound or converted word is entered in the dictionaries are
the word’s frequency and the question whether its meaning is
predictable or not. Words that are very infrequent or have

meanings that are highly predictable are often left out. As

pointed out in note 8 to this chapter, however, today’s diction-
ary makers do not always follow these principles.
9 The point of the discussion of propeller and wheelchair is to show
that many words that are labels have unpredictable meanings
from the beginning: once the propeller and the wheelchair were
invented – or even conceptualised – they were given these
names. Labels may also develop over time – a process sometimes
referred to as lexicalization – as in the case of typewriter which
was originally used with reference both to the machine and the
person, but is now only used (if at all ) about the machine (cf
Bauer 1983:57).
10 Labelling and syntactic packaging refers to the two main func-
tions of words created by word-formation. Labels are used to
identify phenomena that need a name and in so doing they add
to the word-stock of the language. (Cf. the discussion above
concerning propeller and wheelchair). In syntactic repackaging,
word-formation is used to rephrase a syntactic construction, as
for instance when the person who (had) helped me is replaced by
my helper.
Chapter 4
1 The ‘removal’ prefixes de- and un- untypically change the word
class of the stem they are attached to. This is also true of certain
other prefixes like e.g. out-.
2 A fully productive prefix combines freely with all stems belong-
ing to a certain well-defined class of stems. If new members are
added to that class of stems, a fully productive prefix is able to
combine with all of them. Fairly productive prefixes are less reli-
able in this respect, and weakly productive prefixes even less so.
3 –
4 When negative un- is attached to an adjective it creates a word
expressing the opposite of the original stem meaning, while
words created by the addition of non- merely deny the existence

of the qualities denoted by the stem . When un- is non-typically

added to nouns like person, dog and book the combinations
mean ‘the very opposite of a person, dog, book’. (See also Bauer
1983:279–285)
5 For example new, quick, tall, wide.
6 A few suggestions: clearly stems combining with mega- must be
gradable: a word like *mega-nuclear would be odd because
nuclear is normally not gradable. Mini- seems to combine only
with stems that are countable nouns and accordingly *mini-
snow is an unlikely new formation. Out- only combines with
verbs denoting activities in which competition is possible: out-
die sounds odd (but no doubt a context could be dreamt up for
it).
7 –
8 The prefix co- is normally added to agent nouns denoting peo-
ple performing some specialised activity in which it is normal to
have one or several partners.The verbs pilot and produce (films,
plays, etc.) both denote such specialised activities. To a lesser
extent that is also true of believe (‘have certain religious beliefs’),
run and sing, but not of like, sit and snore. (Note however, that
there will always be writers/speakers who will produce words
like co-snorer in order to show originality).
9 –
10 The combining forms (for which see Chapter 8) are all bound
Latin or Greek stems, and as such they usually denote extralin-
guistic phenomena like the stars (astro-), the soul (psycho-) etc.
The English prefixes on the other hand have meanings like e.g.
degree/size, liking/disliking, negation, reversal, repetition.
Chapter 5
1 –
2 The definition of a fully productive derivational suffix is basi-
cally the same as that for prefixes, excepting the type of forma-

tion involved. Thus a productive derivational suffix combines

freely with all the stems belonging to a certain well-defined
class and will combine with new stems meeting the specifica-
tions for that class.
3 The words Johnson, Eton, racket, author, history all have main
stress on the first syllable. The main stress remains on the first
syllable in authorise because -ise is one of the suffixes that do not
affect stress placement, but is moved to the syllable immedi-
ately preceding the suffix in Etonian, authority, historic because
-(i)an, -ity and -ic move the main stress in a word to the syllable
immediately to their left. The suffixes -esque, -eer and -ation
belong to the class of suffixes requiring that main stress be
placed on the suffix itself, which explains the stress patterns of
Johnsonesque, racketeer and authorization.
4 One -ful suffix is attached to nouns to form other nouns denot-
ing amount as in a plateful of beans A second -ful suffix is added
to nouns to create adjectives meaning ‘full of what the stem
denotes’ as in truthful.
5 Word-final -sion is pronounced [ʃən] in propulsion and transmis-
sion, but [ən] in derision and implosion. The two different pro-
nunciations can be predicted from the form of the stems of the
verbs underlying the -sion forms: if the verb stem ends in -de, as
in deride and implode, -sion is pronounced [ən]. In all other
cases it is pronounced [ʃən].
6 The only productive -ese suffix is the one used to form nouns
meaning ‘jargon typical of the person(s), organization, text
etc.that the stem denotes’. The noun and adjective suffixes -ese
encountered in e.g. a Japanese, Can you speak Japanese and a Jap-
anese company often occur in combination with nouns of coun-
tries, but their productivity is limited by the limited number of
new country names.
7 The function of the -ie/-y suffix in words like auntie, doggie/doggy,
nightie is merely to express familiarity and informality on the
part of the speaker. Other suffixes have more definite meanings.
Note that this also applies to the (different) -ie/-y suffix found in
words like foodie. junkie, cutie etc., which changes the stems food,

junk, cute into words denoting persons having different

relations to the stem: foodie and junkie denote people who like
food, junk (drugs), a cutie is a person who is cute.
8 –
9 Nouns that take the suffix -ed in these constructions must be felt
to be inherently part of their ‘owner’: this explains why a car
with four wheels may be described as four-wheeled, but a person
who owns four wheels may not.
10 The only nouns that can be turned into adjectives meaning ‘full
of, covered with’ by means of the suffix -y are concrete mass
nouns like e.g. sand and dust, and concrete nouns denoting nat-
ural objects like rocks and stones.
Chapter 6
1 Conversion is like (most cases of) suffixation in changing the
word class of the stem it is applied to. Conversion and suffixa-
tion also express much the same meanings, and are often ‘in
competition’ with each other. But – with the exception of ‘par-
tial conversion’ – word-formation by means of conversion does
not involve any formal change in the word to which it is
applied. Suffixation always changes the form of the word it is
applied to.
2 Truly productive conversion results in words with predictable
meanings as when instrument-denoting nouns are converted to
instrumental verbs. Such conversion is obviously based on the
assumptions speakers have about the way(s) different instru-
ments are typically used.
3 Like other types of regular word-formation, conversion is used
to create linguistic labels when its output is no longer predicta-
ble.
4 A noun like carpet is converted to a ‘provide with’ verb in e.g. to
carpet a room, while the noun bark can only be given a ‘remove
from’ meaning when converted to a verb as in e.g. to bark a tree.
These different interpretations have to do with the fact that car-

pet and bark have different relations to their ‘owners’, i.e. rooms
and trees: a room does not necessarily have a carpet, while a tree
must have bark (possesses bark inherently).
5 The formation of instrumental verbs from nouns denoting
instruments and tools exploits our natural expectations about
the way(s) the referents of such nouns are typically used. How-
ever, the meaning of the potential instrumental verb may
already be expressed by another word – a phenomenon known
as blocking. Thus an instrument noun like spade is not normally
converted to a verb to spade, since the verb dig already exists.
(Note however that there is a verb spade but with a different
meaning. Compare also axe, pen).
6 Eats ‘food’ : verb to noun conversion with plural suffix added,
graph ‘show in graph form’: noun to verb conversion, swot ‘per-
son who studies hard’: verb to agent noun conversion, dovetail
‘fit in with, agree with’: compound noun to verb conversion,
flatline ‘die’: compound noun to verb conversion The word
unmicrowaveability is the result of several word-formation proc-
esses: the compound noun microwave oven is abbreviated to
microwave (n), which undergoes conversion to the verb (to)
microwave ‘prepare (food) in a microwave’. To this verb the suf-
fix -able is attached, producing the adjective microwaveable. The
negative prefix un- is added and finally the noun suffix -ity is
attached yielding unmicrowaveability.
7 As in the case of Sydney, 1930s and upper class in question 2 in
Chapter 2, sand and Chicago are best regarded as nouns used
with attributive function, and let us all be friends as a phrase
with the same function. Regarding such cases as instances of
conversion to adjectives would force us to assign adjective sta-
tus to all words and word combinations that may occur attribu-
tively and predicatively, including most nouns.
8 –
9 The only productive use of ‘partial conversion’ is when it is
applied to phrasal verbs like e.g. show off – with main stress on
off – to turn them into nouns like a show-off ‘person who shows
off’ with main stress on show. As an alternative to calling this

‘partial conversion’ we could set up a new type of word-forma-

tion operating in terms of stress change.
10 In this chapter it was suggested that forms in -ing like amusing
etc. are best analysed as adjectives converted from participles.
However, when it comes to clearly nominal forms in -ing like
painting in e.g. Hang the painting in the corner, the general opin-
ion among linguists seems to be that we are dealing with suffix-
ation rather than conversion, and that in such cases a noun-
forming suffix -ing has been attached to the verb stem.This is
also the view taken in the present book. For further discussion,
see Adams 2001:28, The New Cambridge Grammar of English p.
1702, Marchand 1969:302.
Chapter 7
1 –
2 The form -gate comes from the name of the Watergate hotel in
Washington D.C. connected with the 1972 scandal that eventu-
ally toppled President Nixon. The -gate element then came to be
used in names for other political scandals. In such names, the
-gate ‘suffix’ is added to the name of a person, place, etc.
involved in the scandal, as in e.g. Dianagate, Camillagate, Sadd-
amgate, Irangate, etc. The words in -gate differ from those in
-babble, -man, -speak in that the element -gate has nothing to do
with the noun gate, while the meanings of -babble, -man and
-speak are clearly related semantically to babble, man, speak.
3 –
4 The combinations bomb scare (‘B is produced by A’ ), bomb dis-
posal, salmon-fishing (‘B is activity involving A’), chocolate box,
fishing-rod, winning post (‘B is used for A’) and fish farm (‘B pro-
duces/causes A’) meet both the stress criterion and the semantic
criterion for noun+noun compounds. Chocolate biscuit and fish
finger belong to the problematic group of ‘B is made of/consists
of A’ and have main stress on the fist element, as does winning
streak.

5 The tree-diagrams below give the analyses for war crimes investi-
gation and talk show host murder:
war crimes investigation talk show host murder
war crimes investigation talk show host murder
war crimes talk show host
talk show
6 (a) sand shovel, (b) young wine-drinkers (c) white-wine-drinkers

(d) a singles collection, (e) knee ligament injury, (f) air gun, (g) bike
alcoholism fear, (h) knee-high grass, (i) voter-friendly legislation (j) a
mind-bogglingly difficult task
7 In the analysis proposed in Chapter 7, a word like tax-payer is
considered to be the result of a compounding process combin-
ing tax and payer. To this it might be objected that payer is sel-
dom if ever found as a word on it own. An alternative analysis is
to claim that tax-payer has been derived by suffixation from a
verb to tax-pay and that a noun like wine-drinking is derived in a
similar way from to wine-drink. The problem is that there are no
verbs tax-pay or wine-drink in English. Yet another explanation
would be to derive nouns like these from phrases like ‘person
who pays taxes’ and ‘the activity of drinking wine’.
8 –
Chapter 8
1 –
2 Neo-classical compounds normally have primary stress on the
third syllable from the end – also known as the antepenult – and
this is what we find in the majority of the examples given here,
i.e. acidophilus, Anglophile, Anglophobe, automaton, automate,
cryptogram, cryptographer, cryptography, pathological, pathologist,
pathology. In acidic, acidophilic, automatic and cryptographic on

the other hand, the main stress falls on the second syllable from
the end of the word, a position due to the influence of the suffix
-ic (cf. Chapter 5).
3 –
4 Heli- in these words obviously stands for the whole word heli-
copter: heli-skiing is helicopter-assisted skiing, a helipad is a land-
ing pad for helicopters, etc. Heli- in these words must accord-
ingly be counted as a telescopic initial combining form, just like
the form bio- in words like biodegradable, where bio- stands for
the whole word biological(ly).
5 Word-final telescopic forms are much less common than word-
initial ones, but they do exist. One example is the form -naut in
chimp(o)naut ‘astronaut who is a chimpanzee’. In this example
-naut stands for ‘astronaut’, but its normal meaning is ‘sailor’, as
in e.g, astronaut ‘sailor among the stars’. (The example
chimp(o)naut is from Bauer 1983:271).

Index and lists
Index
abbreviation 157–158 endocentric compounds 121
acronym 157–159 exocentric compounds 127–
adjective 129
definition 42 noun compounds 129–134
adjectives in -ed and -ing 41, 116 spelling 126–127
adverb syntactic compound 129–131
definition 42–43 verb compounds 138–139
‘direction’ 105 conversion
intensifier definition 109–110
manner 42–43, 104–105 ‘act as/like’ 113–114
‘with regard to’ 105–106 ‘activity/instance/result’ 115–
affix 15, 49–50 116
antepenult 146–147 agent 116
‘change into’ 114
back-formation 161–162 instrumental verb 112–113
barking 13 location/storage 111–112
blends 163–164 partial conversion 117–119
borrowing 154–155 ‘provide with’ 112
‘remove from’ 112
clipping ‘send/travel by’ 113
definition 159
back-clipping 159–160 denotation 25
fore-clipping 160 derivation(al) 50, 80–81
cognate 11
combining forms etymology 11
definition 143
word-final 144 generalization of meaning 156
word-initial 144 Germanic languages 11
telescopic 148–149 grammaticalization 45–46, 122,
compound 191
definition 121–122
adjective compounds 134–139 head 50, 121–122
adverb compounds 139–140 homonym 22, 45, 62, 70, 80,
compound vs. phrase 123–126 116

Index and lists
idiom 25 prefix
immediate constituent 18 definition/function 63–64
Indo-European 11 ‘attitude’ 74
inflection 22 ‘degree/size’ 70–74
initialisms 157 ‘location/direction’ 74–75
irregular word-formation 153–167 manner 70
‘negation’ 67–68
Jabberwocky 54 productivity 66
‘removal’ 69
label(ing) 55, 56, 59 ‘time/sequence’ 75–76
lexeme 21–24 ‘reversal’ 68–69
lexical density 21 preposition
lexicalization 56, 126 definition 45–46
loans complex prepositions 24
from French 28
from Greek 28–29 recategorisation 38–40
from Latin 27–29 reduplicative compound 164–166
reference 25
Middle English 27 rhyming slang 166–167
Modern English 28
modifier 51, 121–123 sense 25–26
morpheme sign
definition 15 definition 9
base morpheme 16 motivated 10
bound morpheme 16 non-motivated (arbitrary) 10
free morpheme 16 simplified language 33
morphology 17 sound symbolic words 12
multimorphemic word 17 sound-imitating words 12
multiword unit 23–24 stem 16
suffix
neo-classical compound 143–151 definition 49, 79–80
nominalizations 93–95 adjective-forming 96–103
noun adverb-forming 104–106
definition 35 ‘agent/instrument’ 94–95
common noun 37 ‘activity’ 86–87, 95–96
count(able) noun 37 ‘amount’ 85
mass noun 37 ‘associated with’ 100–102
proper noun 37 ‘characterised by’ 99–100
uncountable noun 37 ‘collective’ 85–86
‘covered with/by’ 99
Old English 26 direction 105
onomatepoeic/peia 12 familiarity 91

Index and lists
‘having properties of’ 100–102 type 19–21

‘lacking’ 97–98 type/token ratio 19–21
manner 104–105
noun-forming 85–96 verb
person associated with definition 40–41
stem 88–90, 94 phrasal 41
‘possessing’ 97 prepositional 41
‘receiver of action’ 96
‘resembling’ 98–99 word
‘result’ 95–96 definition 19–25
‘somewhat’ 102 complex 15–18
‘sex/size/status’ 91 compound 15, 121–123
‘state’ 87, 92 content word 31–32
‘that can/should be V-ed’ 102– function word 31–32
103 orthographic word 19
‘turn into’ 103–104 simple word 15
verb-forming 103–104 word-form 19
synonym(y) 26 word class
syntactic repackaging 56–59 definition 31–35
closed 32–33
telescopic forms 148–149 open 32–33
token 19–21 word manufacture 52–54

Index and lists
List of prefixes
Prefix Meaning Page

a- negation 68
anti- attitude 74
arch- degree/size 70
co- location/direction 74–75
de- reversal 68
de- removal 69
dis- negation 68
dis- removal 68
dis- reversal 69
ex- time/sequence 75
extra- degree/size 70–71
extra- location/direction 75
hyper- degree/size 71
in- negation 68
inter- location/direction 75
macro- degree/size 71
mal- manner 70
mega- degree/size 71
micro- degree/size 72
mini- degree/size 72
mis- manner 70
neo- time/sequence 75
non- negation 67–68
out- degree/size 72
over- degree/size 72
post- time/sequence 76
pre- time/sequence 76
pro- attitude 74
re- time/sequence 76
semi- degree/size 72

Index and lists
sub- degree/size 73
super- degree/size 73
trans- location/direction 75
ultra- degree/size 73
un- negation 67–68
un- reversal 68
un- removal 69
under- degree/size 73

Index and lists
List of suffixes
(WCF stands for ‘word class function of suffix, WCS for word class
of stem)
Suffix WCF WCS Meaning Page(s)

-able A V ‘that can/should be V-ed’ 102
-age N N ‘activity/state’ 86–87
-age N V ‘activity/result’ 95
-al N V ‘activity/result’ 95
-al/-ial A N ‘having properties of stem’ 100
-an/-ian N N ‘person associated with stem 89
-an/-ian A N ‘having properties of stem’ 100
-ant/-ent N V ‘agent/instrument’ 95
-ation N V ‘activity/result’ 95
-dom N N ‘state’ 87–88
-ee N N ‘small size’ 91
-ee N V ‘receiver of action’ 96
-er N N ‘person associated with stem’ 88–89
-er N V ‘agent/instrument’ 94–95
-ery/-ry N N ‘activity/state’ 86–87
-ese A N ‘having characteristics of 100–101
-ese N N ‘person associated with stem 89–90
-esque A N ‘having characteristics of stem’ 101
-ess N N ‘female’ 91
-ette N N ‘small size’, ‘female’ 91
-fashion ADV N ‘manner’ 105
-ful N N ‘amount’ 85
-ful A N ‘full of’ 99
-hood N N ‘state’ 87–88
-i N N ‘citizen of stem’ 90
-ic A N ‘having characteristics of 101
-ie N N ‘person having characteristics of stem’ 92
-ie/-y N N ‘person interested in/good at stem’ 90
-ie/-y N N ‘familiarity’ 91
-ify V N/A ‘turn into what stem denotes’ 103–104

Index and lists
-ing N N ‘collective’ 85-86

-ing N N ‘activity/state’ 86
-ing N V ‘activity/result’ 95
-ion N V ‘activity/result’ 95
-ise/-ize V N/A ‘turn into what stem denotes’ 103–104
-ish A N ‘resembling’ 98
-ish A A ‘somewhat’ 102
-ism N N ‘activity/state’ 86
-ist N N ‘person associated with stem’ 90
-ite N N ‘person associated with stem’ 90
-ite A N ‘having characteristics of stem’ 101
-itis N N ‘activity/state’ 87
-ity N N ‘state of being stem’ 92
-ive A V ‘that Vs’,’that can V’ 102–103
-less A N ‘without’,’lacking’ 97–98
-like A N ‘resembling’ 98
-ly ADV A ‘manner’ (and others) 104–105
-ly A N ‘resembling’ 98
-ment N V ‘activity/result’ 95–96
-ness N N ‘state of being stem’ 92
-o N N ‘person having characteristics of stem’ 92
-or N V ‘agent/instrument’ 95
-ous A N ‘having characteristics of stem’ 101–102
-ry N N ‘collective’ 86
-ship N N ‘state’ 88
-style ADV N ‘manner’ 105
-ward(s) ADV N ‘direction’ 105
-wise ADV N ‘with regard to’ 105–106
-wise ADV N ‘manner’ 105
-y A N ‘characterised by’ 99–100
-y A N ‘resembling’ 98–99
-y A N ‘full of, covered with’ 99
-y A V ‘inclined to V’ 103

Making Words in English PDF

Uploaded by

Copyright:

Available Formats

Making Words in English PDF

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

Making Words in English PDF

Uploaded by

Copyright:

Available Formats

Magnus Ljung

All rights reserved. No part of this publication may be reproduced or

The papers and inks used in this product are eco-friendly.

2 The English Word Classes 31

3 Outline of English Word-formation 49

© Magnus Ljung and Studentlitteratur 2003 3

8 Combining forms and

9 Irregular Word-formation 153

4 © Magnus Ljung and Studentlitteratur 2003

Discussion of exercises 187

Index and lists 201

© Magnus Ljung and Studentlitteratur 2003 5

This book provides an introduction to modern English word-forma-

© Magnus Ljung and Studentlitteratur 2003 7

8 © Magnus Ljung and Studentlitteratur 2003

The linguistic sign

© Magnus Ljung and Studentlitteratur 2003 9

In signs like the dog’s growl, there is a ‘natural’ relation between

The non-motivated sign

10 © Magnus Ljung and Studentlitteratur 2003

A further consequence of the absence of a natural connection

© Magnus Ljung and Studentlitteratur 2003 11

Apparent exceptions to the arbitrariness of the sign

12 © Magnus Ljung and Studentlitteratur 2003

To begin with, the sound-imitating words are quite few in

© Magnus Ljung and Studentlitteratur 2003 13

Swedish: vov vov

We could of course try to explain away the differences in the list by

14 © Magnus Ljung and Studentlitteratur 2003

A real exception to the arbitrariness of the sign:

© Magnus Ljung and Studentlitteratur 2003 15

The structure of complex words

16 © Magnus Ljung and Studentlitteratur 2003

meanings like ‘repetition’ (re-), ‘opposition’ (anti-), ‘reversal’ (un-),

LEVEL 2 un- readable

LEVEL 3 read -able

Such an analysis rests on the common linguistic assumption that

© Magnus Ljung and Studentlitteratur 2003 17

18 © Magnus Ljung and Studentlitteratur 2003

Words as types and tokens9

The total number of letter combinations surrounded by spaces

© Magnus Ljung and Studentlitteratur 2003 19

Tabell 1.1 Word frequency list for the text in (1).

Types Frequency (Tokens)

20 © Magnus Ljung and Studentlitteratur 2003

others in the ‘closed word-classes’ (see pp. 43–46)—have meanings

The word as an abstract entity—the lexeme

© Magnus Ljung and Studentlitteratur 2003 21

the physical shape of the items that determines what is a word: if

22 © Magnus Ljung and Studentlitteratur 2003

It is sometimes said that lexemes are the same as ‘dictionary

Lexemes and multi-word units

© Magnus Ljung and Studentlitteratur 2003 23

These are combinations of separately written words considered to

(2) We looked at the ashtray in front of our wine glasses.

Compounds: blackbird, teapot16, power plant, steamboat, English

24 © Magnus Ljung and Studentlitteratur 2003

The categories above are often difficult to tell apart, in particular

Reference, denotation, sense 19