Colm Hogan The Mind and Its Stories PDF

This page intentionally left blank
The Mind and Its Stories

Narrative Universals and Human Emotion
This book argues that there are profound, extensive, and surprising
universals in literature that are bound up with universals in emotion.
Professor Hogan maintains that debates over the cultural specificity
of emotion have been misdirected because they have largely ignored a
vast body of data that bear directly on the way different cultures imag-
ine and experience emotion – literature. This is the first empirically
and cognitively based isolation and discussion of narrative universals.
Professor Hogan argues that, to a remarkable degree, the stories peo-
ple admire in different cultures follow a limited number of patterns
and that these patterns are determined by cross-culturally constant
ideas about emotion. In formulating his argument, Professor Hogan
draws on his extensive reading in world literature, experimental re-
search treating emotion and emotion concepts, and methodological
principles from the contemporary linguistics and the philosophy of
science. He concludes with a discussion of the relations among narra-
tive, emotion concepts, and the biological and social components of
emotion.
Patrick Colm Hogan is currently a professor in English, Comparative

Literature, and Cognitive Science at the University of Connecticut. He
is the author of nine books, including Colonialism and Cultural Iden-
tity: Crises of Tradition in the Anglophone Literatures of India, Africa, and
the Caribbean (2000); The Culture of Conformism: Understanding Social
Consent (2001); and the forthcoming Cognitive Science, Literature, and
the Arts.
STUDIES IN EMOTION AND SOCIAL INTERACTION
Second Series
Series Editors
Keith Oatley
University of Toronto
Antony S. R. Manstead
University of Cambridge
This series is jointly published by Cambridge University Press

and the Editions de la Maison des Sciences de l’Homme, as part
of the joint publishing agreement established in 1977 between
the Foundation de la Maison des Sciences de l’Homme and the
Syndics of Cambridge University Press.
Cette publication est publiée co-édition par Cambridge Univer-

sity Press et les Editions de la Maison des Sciences de l’Homme.
Elle s’intègre dans le programme de co-édition établi en 1977
par la Foundation de la Maison des Sciences de l’Homme et les
Syndics de Cambridge University Press.
Titles published in the Second Series:

The Psychology of Facial Expression
James A. Russell and José Miguel Fernández-Dols
Emotion, the Social Bond, and Human Reality: Part/Whole Analysis
Thomas J. Scheff
Intersubjective Communication and Emotion in Early Ontogeny
Stein Bråten
The Social Context of Nonverbal Behavior
Pierre Philippot, Robert S. Feldman, and Erik J. Coats
Communicating Emotion: Social, Moral, and Cultural Processes
Sally Planalp
Emotions across Languages and Cultures: Diversity and Universals
Anna Wierzbicka
Gender and Emotion: Social Psychological Perspectives
Agneta H. Fischer
(continued on page following the Index)

PATRICK COLM HOGAN

University of Connecticut
  
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press

The Edinburgh Building, Cambridge  , United Kingdom
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521825276
© Cambridge University Press 2003
This book is in copyright. Subject to statutory exception and to the provision of

relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.
First published in print format 2003
-
isbn-13 978-0-511-06244-5 eBook (NetLibrary)
-
isbn-10 0-511-06244-3 eBook (NetLibrary)
-
isbn-13 978-0-521-82527-6 hardback
-  hardback
isbn-10 0-521-82527-X
Cambridge University Press has no responsibility for the persistence or accuracy of

s for external or third-party internet websites referred to in this book, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
For Lalita
Contents
Acknowledgments page xi
Introduction: Studying Narrative, Studying Emotion 1

1 Literary Universals 17
2 Emotion and Suggestion: Lexical Processes in Literary
Experience 45
3 Four Hypotheses on Emotion and Narrative 76
4 Writing Beyond the Ending: A Problem of Narrative,
Empathy, and Ethics 122
5 Extending the Theory: Emotion Prototypes, Narrative
Junctures, and Lyric Poetry 152
6 Testing, Revision, and the Program of Research in
Narrative Universals: Ainu Epic and the Plot
of Sacrifice 172
7 The Structure of Stories: Some General Principles of Plot 202
Afterword: From the Emotional Nature of Narrative
to the Narrative Nature of Emotion 239
Works Cited 265

Index 285
ix
Acknowledgments
An earlier version of Chapter 1 and a small part of the introduction

were published as “Literary Universals” in Poetics Today: Interna-
tional Journal for Theory and Analysis of Literature and Communica-
tion 18.2 (1997): 223–49. c 1997 The Porter Institute for Poetics
and Semiotics, Tel Aviv University. All rights reserved. An earlier
version of Chapter 2, including one paragraph now in the intro-
duction, was published as “Toward a Cognitive Science of Poetics:
Anandavardhana, Abhinavagupta, and the Theory of Literature” in
College Literature 23.1 (1996): 164–78. I am grateful to the editor, Kostas
Myrsiades, for permission to reprint. An earlier version of Chapter 4
appeared as “The Epilogue of Suffering” in SubStance 94/95, Vol. 30,
Nos. 1 & 2. c 2001. Reprinted by permission of the University of
Wisconsin Press. These chapters, and my general thinking about lit-
erature and cognition, have benefited greatly from the comments of
editors and referees at these journals.
An earlier version of Chapter 3 was presented at the main panel
of the forum on cognitive science at the 1998 Modern Language
Association (MLA) convention. An earlier version of Chapter 4 was
presented at the symposium on cognition, evolution, and the arts at
the University of California at Santa Barbara in 1999. An earlier ver-
sion of Chapter 5 was presented at the 1999 MLA convention. In each
case, members of the audience and fellow panelists stimulated and
challenged my thought on these issues.
xi
xii Acknowledgments
Keith Oatley and Norman Holland read the entire manuscript with
care and gave insightful comments and helpful suggestions. The same
may be said about the five anonymous referees for Cambridge Uni-
versity Press. Finally, I am grateful to Philip Laughlin, psychology
editor at Cambridge University Press, for his masterful handling of
this strange, cross-disciplinary manuscript, and (again) Keith Oatley
who, as editor for the series “Studies in Emotion and Social Interac-
tion,” made this book possible.
Introduction
Studying Narrative, Studying Emotion
When empirical researchers in the social sciences consider the nature

of emotions and emotion concepts, they most often conduct anthro-
pological interviews, send out surveys, analyze linguistic idioms, test
stimulus response times, and so on. They may move toward medi-
cal and biological study as well, giving injections to test subjects,
engaging in neuroimaging, and the like in order to gather as much
relevant data as possible. But, with only a few exceptions, they al-
most entirely ignore a vast body of existing data that bears directly
on feelings and ideas about feelings – literature, especially literary
narrative.1 Stories in every culture both depict and inspire emotion.
Indeed, the fact that some stories are highly esteemed in any given
culture suggests that those stories are particularly effective at both
tasks – representing the causes and effects of emotion as understood
or imagined in that society and giving rise to related emotions in
readers. Of course, one cannot assume that depictions of emotion ac-
curately represent those emotions. This is the common, and quite rea-
sonable, objection to treating literature as empirical data. However,
we have very good reason to assume that widely admired depictions
of emotions tell us something important about the way people in a
1 As just noted, there are exceptions here, especially among researchers influenced
by psychoanalytic work, for psychoanalytic theory has drawn on literature since the
time of Freud. A recent example is Labouvie-Vief, who combines developmental and
empirical psychology with a study of myth from several Mediterranean cultures in
order to discuss aging.
1
2 The Mind and Its Stories
given society think about emotions. In other words, we have a body

of commonly enjoyed, elaborate, narrative portrayals of emotion sce-
narios. At the very least, these would seem to tell us far more about
common emotion ideas than some verbal definition of an emotion
term. Moreover, emotional reactions to literary works – the sorrow,
anger, mirth felt and expressed by readers – clearly tell us something
about what moves people in a particular culture, what touches them
emotionally. Indeed, literary response is as close as we can usually
get to a wide range of genuine and spontaneous human emotions
that are most often concealed in private interactions.
In these ways, the celebrated stories of any given society form an
almost ideal body of data for research in emotion and emotion con-
cepts. The central contention of this book is that anyone who pays
attention to this body of literary data by examining it cross-culturally,
cannot help but be struck by the uniformity of narrative structures
and of the emotions and emotion ideas that are inseparable from
those structures. More exactly, there are extensive and detailed nar-
rative universals. These universals are the direct result of extensive
and detailed universals in ideas about emotions that are themselves
closely related to universals of emotion per se.
literature as a human act

One reason literature has played such a limited role in cognitive study
is that science seeks generalities while literature seems to be tied to
narrow particularity. In connection with this, even humanists have
been resistant to the idea that there are universal patterns in litera-
ture. The sharp contrast between literature and the stuff of empiri-
cal research seems to be one of the few things that most humanists
and scientists agree on. Both literary critics and readers from other
disciplines tend to think of literature in terms of nations and peri-
ods, genres, schools, and movements. Indeed, the tendency is much
more pronounced among professional students of literature. Literary
historians and interpreters categorize works of literature by groups,
opposing the groups to one another and scrutinizing these groups for
differences. What distinguishes Romanticism from Neo-Classicism?,
Post-Modernism from Modernism?, European drama from Sanskrit
theater?, western lyric from Chinese tz’u? These are the sorts of
Studying Narrative, Studying Emotion 3
questions asked in comparative literary study. Asking these ques-

tions leads one to find answers, and thus to find differences. When
finding specific differences, one tends to exaggerate group difference.
In other words, when one examines what separates this group from
that group or this body of literature from that body of literature, dis-
tinguishing features become salient, while commonalities fade into
the background. The result is a disproportionate sense of discrepancy
and opposition.
But in fact there are far more numerous, deeper, more pervasive
commonalities than there are differences. As Donald Davidson has
argued, even to understand and think about difference, we need to
presuppose a vast range of similarities. In Davidson’s words, “dis-
agreement and agreement alike are intelligible only against a back-
ground of massive agreement” (137). Put differently, literature – or,
more properly, verbal art – is not produced by nations, periods, and so
on. It is produced by people. And these people are incomparably more
alike than not. They share ideas, perceptions, desires, aspirations,
and – what is most important for our purposes – emotions. Verbal art
certainly has national, historical, and other inflections. The study of
such particularity is tremendously important. However, literature is,
first of all and most significantly, human. It is an activity engaged in
by all people at all times. As Paul Kiparsky put it, “literature is nei-
ther recent nor a historical invention. In fact no human community
lacks a literature”; no group is “so wretched that it does not express
its memories and desires in stories and poems” (“On Theory” 195–6).
More recently, Mark Turner has argued that, “literary criticism has
given us a concept of literature as the product of circumstances . . . not
as a product of the capacities of the human mind. We do not ask, what
is the human mind that it can create and understand a text? What is a
text that it can be created and understood by the human mind? These
questions are not at the center or even the periphery of our critical
inquiries” (Reading Minds 16). The professional division of literature
by nationality, ethnicity, and so on, tends to occlude this fundamen-
tal, human condition of verbal art. The following chapters address
literature, then, not as the expression of an ethnic Weltanschauung,
nor as evidence of an historical episteme, but rather as a human
activity – something people do, and always have done, in all parts of
the world, and at all times.
At one level, this is, then, a study of literary universals. It is also

a cognitive study.2 Indeed, it is my contention that literary univer-
sals are to a great extent the direct outcome of specifiable cognitive
structures and processes applied in particular domains and with par-
ticular purposes. In this way, the study of literary universals is largely
a subfield of cognitive research.3 Moreover, it is a crucial subfield for
cognitive science. Cognitive science can hardly claim to explain the
human mind if it fails to deal with such a ubiquitous and significant
aspect of human mental activity as literature. In this way, cogni-
tive science is not only important to the study of literary universals.
The study of literary universals is equally important for cognitive
science. Indeed, a small, but significant – and expanding – group
of cognitive scientists has come to recognize the necessity of incor-
porating literary study into their domain as work by Steven Pinker,
E. O. Wilson, John Tooby and Leda Cosmides, Howard Gardner, Keith
Oatley, Jeffrey Saver (see Young and Saver), and others attests.
But this is not a book on literary universals in general – which is, in
any case, too large a topic for a single work. In Chapter 1, I do intro-
duce general principles for the cognitive study of literary universals.
However, the bulk of the volume focuses on the relation between two
crucial elements of literature and the human mind – narrative and
emotion. Despite the recent cognitive interest in literature, this is an
area that has hardly been explored, leaving aside the work of one or
two researchers, most importantly Keith Oatley – and even Oatley
has not studied this nexus cross-culturally, in an attempt to isolate
universal structures. Again, the general absence of attention to this is
2 I am, of course, not alone in linking literary study with cognitive science. The last
decade has seen the growth of a significant movement in literary study based on
cognitive science. Work by Turner, as well as Norman Holland (Brain), Ellen Spolsky,
Paul Hernadi, Jerry Hobbs, Mary Crane, Alan Richardson, and a number of other
writers, has provided a valuable alternative to recently dominant approaches to
literature. This book is, to a certain extent, part of that movement. At the same time,
however, there are some differences between my views and mainstream cognitivism,
as I have discussed in On Interpretation. For the most part, these differences do not
bear on the topics discussed in the following pages. Thus I shall leave them aside,
except for a brief discussion in the Afterword.
3 One important qualification here is that some literary universals do not seem to be
a matter of psychology per se, but rather of social conditions, either changeable or
permanent. We shall discuss some examples in Chapter 1 when treating implicational
universals.
unfortunate, for it is an area of seemingly obvious value for cognitive

science.
Narrative is, of course, central to verbal art. Indeed, it is, as I shall
argue, even more definitive and widespread than is commonly rec-
ognized. For example, it lies concealed in such apparently nonnar-
rative works as lyric poems. What is crucial for our purposes is that
narrative is intimately bound up with emotion. Literary stories, es-
pecially the stories we most admire and appreciate, are structured
and animated by emotions. Any coherent sequence of events might
constitute a story. But the stories that engage us, the stories we cel-
ebrate and repeat – “paradigm” stories – are precisely stories that
move us, most often by portraying emotions or emotionally conse-
quential events. Conversely, the emotive impact of verbal art cannot
be discussed separately from its narrative structure. Indeed, even real
life emotion is bound up with narrative. As a number of writers have
pointed out, our affective response to a situation, real or fictional, is
not a response to an isolated moment, but to the entire sequence of
events in which that moment is located, whether explicitly or implic-
itly. Consider someone’s death. This is narratively embedded, first of
all, in the simple sense that we infer the person was alive and some
causal sequence led to his/her death. But it is narrative also and more
significantly in the sense that we respond to the death in terms of the
narrative details through which we understand the person’s life. Sup-
pose we learn that someone died in an automobile accident. We are
likely to respond one way if we learn that the person was in the final
stages of a terminal illness with only a few pain-filled weeks to live.
We are likely to respond differently if we learn that the person was
driving to his/her wedding. Skeptics might reply by arguing that
narrative in these cases is simply a matter of causal inference and
evaluative judgment, and thus is not narrative in any interesting or
substantive sense. It is certainly true that what I am describing is in
part ordinary causal analysis and evaluation. However, that is not all
there is to it. One argument of the following pages is that our ideas
about, evaluations of, and most crucially our emotive responses to
all sorts of things are guided and organized by a limited number of
standard narrative structures. Human thought, action, and feeling
are not simply a matter of rational inference. They are also a matter
of emplotment in a narrow, specifiable sense.
Before going on, it is worth pausing for a moment over the notion
of “paradigm” stories. Paradigm literary works are works that are
widely shared by writers and readers within a tradition and that serve
to establish evaluative standards and structural principles within a
tradition. In the following pages, I shall refer repeatedly to prototypi-
cal literary works. These are not the same as paradigm literary works,
though the latter are most often instances of the former. Specifically,
prototypical works are works that share all our standard criteria for
verbal art. They share all the properties we consider “normal” for
literature. Thus, romantic novels and epic poems – including many
that are unknown or even unpublished – are most often prototyp-
ical literary works. Riddles are not. Paradigm works usually share
these prototypical properties, but add our collective familiarity and
esteem.
By “esteem” here, I mean esteem as literature. We may admire a
work for many reasons. It may express courage in the face of political
oppression. It may teach moral lessons that we find valuable. It may
celebrate our national heritage. But we may admire a work for any
of these reasons and still consider it a poor work of literature. In the
following pages, I am concerned with works that are widely admired
as good works of literature.
Put differently, we tell and write stories every day. Some discus-
sions of narrative are concerned with all these stories. Accounts of
that sort are valuable. But they are different from an account that is
concerned with prototypes and paradigms. The following analyses
do not treat ephemeral stories (for example, what I tell my wife about
how I had to go to three shops to get a particular spice). Ephemeral
stories may be very engaging at the moment, but they are engaging
for idiosyncratic and contingent reasons. What is important here are
stories that have sustained interest within their respective traditions.
A story that has sustained interest is unlikely to have its appeal for
contingent reasons, due to the particular relationship of the speaker
and addressee, or due to some unusual circumstance. As such, a story
that has sustained interest is more likely to tell us something about
the human mind.
The following analyses, then, aim to begin the process of describing
and explaining the remarkably detailed, cross-culturally universal,
and interwoven patterns of our emotions, our ideas about emotions,
and our most enduring stories. One might refer to this project as
an anthropology of world literature, in which it turns out that emo-
tions are central – indeed, definitive, and formative. I undertake this
task in relation to an encompassing research program in cognitive
science.4
universalism and cultural study

As we have noted, a handful of writers in cognitive science have re-
cently become interested in literature. In some cases, this interest has
extended to literary universals. Indeed, there has been a surprising
increase in attention to the topic over the past few years, for the most
part among cognitively oriented literary and film theorists. Except for
the pioneering work of Roman Jakobson and Paul Kiparsky, for a long
while there was little serious discussion of literary universals – hardly
a mention, in fact. However, in the context of developing research
programs in cognitive science, some scholars and theorists have be-
gun taking the idea seriously. In addition to my own earlier efforts
(see “Literary,” “Possibility,” “Beauty,” “Shakespeare,” and 286–95
4 It should go without saying that this analysis does not treat every aspect of narra-
tive, not even every cognitive aspect. Thus it does not in any way preclude other
cognitive approaches. Perhaps the most obvious omission is what narratologists call
“discourse,” the mode of presentation of a “story.” The “story” is the events as they
happened according to a particular narrative. The discourse is the way in which
these events are presented. For example, in a murder narrative, the story begins
with the murderer plotting the murder. It moves to commission of the murder, then
the discovery of the crime, then the investigation. But murder narratives are not
typically told this way. They usually begin with the discovery of the crime, then
the investigation leads us to learn about the preceding events. Thus the discourse
presents the events of the story out of chronological sequence. Discourse is clearly
central to the emotional impact of a work. I have not discussed it simply because
it is another topic, and a huge one. Readers interested in this topic should consult
Brewer and Lichtenstein for empirical research and Tan for an extended and influ-
ential development in relation to film.
More generally, there are many very useful ways in which narrative may be stud-
ied cognitively – and, indeed, has been studied cognitively. A particularly valuable
cognitive treatment of narrative is David Bordwell’s Narration in the Fiction Film that
addresses the film viewer’s cognitive construction of the story out of the discourse.
The most influential cognitive examination of narrative in literary study is proba-
bly Mark Turner’s The Literary Mind that focuses on the mini-narratives of everyday
life and their relation to conceptual blending. These, too, do not at all exhaust the
possibilities.
of Philosophical), a number of other writers have taken up the topic.

These include well-established theorists, such as David Bordwell
(“Convention”) and, in a very different way, Wendy Doniger, as well
as younger critics, such as Alan Richardson and Joseph Carroll, and
independent or extraacademic writers such as Ellen Dissanayake.
In addition, cognitive scientists such as Steven Pinker, though they
do not directly address the issue, clearly presuppose literary univer-
sals in their work on cognition and literature. Indeed, recently, the
University of Palermo sponsored a website devoted to the topic of
literary universals (http://litup.unipa.it). The aim of this website –
the Literary Universals Project – is to bring together researchers from
different fields in order to advance a research program in the area.
With the continuing development of cognitive science, the study of
literary universals is likely to expand in both breadth and depth.
Still, literary universals remain a minority interest. Mainstream lit-
erary critics and theorists pay little attention to the topic. As we have
already noted, in professional literary study, the focus of both theory
and practice tends to be on difference, cultural and historical speci-
ficity, and so on. What Carl Plantinga said of film theorists applies
equally to literary theorists: They tend to seek “explicit ways to link”
literary phenomena “to particular historical conditions and to ideol-
ogy” (450). In keeping with this, a self-evaluation by the American
Comparative Literature Association worried that comparative liter-
ature “may well be left behind on the dustpile of academic history”
if it does not incorporate the current trends variously referred to as
“culture studies,” “cultural critique,” and “cultural theory.” Indeed,
the authors of this study insisted that all work in comparative liter-
ature “should take account of the ideological, cultural, and institu-
tional contexts in which . . . meanings are produced,” which amounts
to an insistence that all comparatist study be focused on historical
and cultural particularities (Bernheimer et al. 5,6). Again, this sort of
work is undeniably important. Indeed, my own work (for example,
Colonialism and Cultural Identity and The Culture of Conformism) has
been, to a great extent, located within the field of culture study. But
to say that such particularist study is valuable is not to say it is all
that is valuable.
When universalism is mentioned in humanistic writing, it is most
often denounced as a tool of oppression. For example, in their
influential introduction to postcolonial literary study, The Empire

Writes Back, Ashcroft, Griffiths, and Tiffin maintain that the notion
of universality is “a hegemonic European critical tool” (149). There
are exceptions, certainly, and not only among writers in cognitive
science. For example, the important Kenyan Marxist novelist, Ngũgı̃
wa Thiong’o, has proclaimed himself “an unrepentant universalist”
(xvii). However, there has been a general sense in literary study
that attention to or advocacy of universals is somehow politically
suspect.
There is a fairly straightforward case against such political claims.
After all, no racist ever justified the enslavement of Africans or colo-
nial rule in India on the basis of a claim that whites and nonwhites
share universal human properties or that their cultures share univer-
sal principles. On the other hand, in saying this, I do not want to
fall into the opposite error of claiming that everything that goes by
the name of “universalism” is politically good. Things are never that
simple. Indeed, one does not have to look far to see how universal-
ist claims have been used to support oppression. Typically, humanist
criticisms of universalism refer back to those universalist claims that
derive from and serve to further colonial, patriarchal, or other ide-
ologies supporting unjust domination. However, as Kwame Appiah
has noted, what anticolonial opponents of universals “are objecting
to” in these cases “is the posture that conceals [the] privileging of
one national (or racial) tradition against others in false talk of the
Human Condition” (58). In other words, they are objecting to false
and duplicitous claims of universalism, assertions of universality that
are untrue and are, in addition, offered in bad faith. Appiah con-
tinues, “antiuniversalists . . . use the term universalism as if it meant
pseudouniversalism, and the fact is that their complaint is not with uni-
versalism at all. What they truly object to – and who would not? –
is Eurocentric hegemony posing as universalism” (58; see also Lalita
Pandit 207–8).
It is important to stress that this conclusion in no way detracts from
the standard forms of particularist literary study. It responds, not to
their positive worth, but to their exclusivity. More exactly, proponents
of cultural and historical study sometimes seem to assume that the
study of universals is opposed to or contradictory of cultural study.
But to argue for the study of universals is not at all to argue against the
study of culture and history. All reasonable students of literature –

including those engaged in a universalist project – recognize that
particularist research and interpretation are extremely valuable. In-
deed, the study of universals and the study of cultural and historical
particularity are mutually necessary. Like laws of nature, cultural
universals are instantiated variously, particularized in specific cir-
cumstances (cf. Ngũgı̃ 26; see also King 33, 127, on the culturally
“rooted” universalism of Derek Walcott and Wole Soyinka). Thus,
to isolate universal patterns, we often require a good deal of cul-
tural and historical knowledge. At the same time, in order to gain
any understanding of cultural particularity, we necessarily presup-
pose a background of commonality (as, once again, Donald Davidson
has argued forcefully [183–98]; see also Brown 151–2). In short, the
study of universality and the study of cultural particularity are not
contradictory, but complementary.5
universality and narrative: research programs,

research methods
Needless to say, these general comments do not establish that there
actually are literary universals. They do not even indicate just how
one might go about isolating universals. The first chapter takes
up the nature of and criteria for universals. Specifically, Chapter 1
draws on work in linguistics – the field that has made the great-
est advances in the study of universals – in order to explain what
constitutes a literary universal and what counts as evidence for
the existence of a universal. The second chapter considers the is-
sue of literary emotion, drawing in particular on Indic literary the-
ory and cognitive research to present an account of why literature
moves us.
Chapters 1 and 2 are, in a sense, preliminary to my main project,
introducing basic principles about literary universals and literary
emotion. Chapter 3 presents and defends my claims about narra-
tive universals and their relation to emotion concepts. In that chapter,
5 Other writers have implicitly presented universalism and particularism as at least

compatible, if not complementary. Good examples would include Zhang and
Kövecses.
I isolate two universal narrative structures, heroic and romantic tragi-

comedy. (A number of Shakespeare’s histories and tragedies, such
as King Lear – with its treatment of political usurpation, external
invasion, and defeat of both the invaders and the usurpers – provide
instances of the former structure; many of Shakespeare’s romantic
comedies – with their conflicts between lovers and parents or society,
separation and exile of lovers, and so on – are good examples of the
latter.) I argue that these structures are not only to be found across
unrelated traditions; they are, in fact, the dominant structures across
traditions. In order to avoid misunderstanding, I should emphasize
at the outset that the universality and prominence of these forms does
not mean that they are the only forms of narrative. One can tell stories
about anything. There are narratives about many different topics and
with many different structures. However, other narrative structures
simply do not have anything like the central, cross-cultural impor-
tance of these two. In addition, I argue that the centrality of these
structures is due to another universal – a universal prototype for
happiness, or, rather, two contextually dependent universal proto-
types for happiness. My contention, then, is that our most prominent
stories are generated from the prototypical structure of our emotion
concepts.
One of the main concerns in the following pages is with estab-
lishing the study of literary universals as a research program. The
influential philosopher of science, Imre Lakatos, has argued that sci-
ence does not proceed either by verification (as one might think intu-
itively) nor by falsification (as Karl Popper argued). Rather, theories
encounter recalcitrant data all the time. But recalcitrant data do not
simply falsify the theory. Theorists reformulate the theory in order to
account for the data. This is what it means for a theory to be part of
a research program. Theorists do not simply seek out confirmatory
instances, cases that match their theory. Rather, they look for data that
might contradict the theory. However, when they find such data, they
do not simply toss the theory out the window. Rather, they try to re-
formulate the theory. Lakatos argues that theorists may reformulate
the theory in a way that is ad hoc or in a way that is not ad hoc. An
ad hoc reformulation merely isolates the recalcitrant data. It does not
expand the explanatory capacity of the theory. A non-ad hoc reformu-
lation makes predictions that go beyond the recalcitrant data. If these
predictions turn out to be correct, then we have a robust research

program. In other words, the recalcitrant data have not impeded, but
advanced the research program. More generally, a successful research
program is one that continually expands its explanatory scope and
precision. Often it will do this through encountering disconfirmatory
data and dealing with those data.
In Chapters 4 through 6, I try to follow these Lakatosian princi-
ples. In Chapter 4, I raise some troubling issues about heroic tragi-
comedies. There are cross-cultural patterns that do not appear to
fit the account of narrative and emotion presented in Chapter 3. I
argue that the seemingly anomalous pattern is in fact predictable,
given the principles of emotion put forth in Chapter 2. Moreover,
I argue that this further development of the preceding hypotheses
has very fruitful consequences for understanding the nature of eth-
ical concerns in literary narrative. Thus, it ultimately expands the
explanatory scope of the theory. Chapter 5 does not treat recalcitrant
data. However, it does take up a set of data that seemed to lie out-
side the theory: lyric poetry. In this chapter, I seek to expand the
scope of the theory by arguing that lyric poems most often imply
narratives. In technical terms, they treat junctural moments in im-
plied narratives. Moreover, the implied narratives are prominently
(though, again, not invariably) those isolated in the third chapter –
heroic and romantic tragi-comedy.
Methodologically, my approach to this point had been fairly
straightforward. The isolation of the prototypes for emotion con-
cepts drew on empirical research performed by experimental and
social psychologists over the past several decades. The data support-
ing claims about universals was gathered more slowly. For years, I
had been reading important canonical works in all the major writ-
ten traditions – European, Middle Eastern, South Asian, Chinese,
and Japanese. Though concentrating on premodern works, I did in-
clude works of modern literature and film as well. More recently, I
made a concerted effort to read highly regarded oral narratives from
a broad range of regions – sub-Saharan Africa, South America, North
America, Australia, and so on. I also did my best to read histories
of the written traditions and scholarly outlines of structures and pat-
terns to be found in these different traditions. In connection with this,
I worked to familiarize myself with the major non-Western traditions
of literary theory as well.6 The isolation of heroic and romantic tragi-

comedy grew out of this reading.7
Experimental psychologists are likely to want laboratory tests and
are likely to feel that one does not have a real research program with-
out them. Laboratory research is undoubtedly of great potential im-
portance here. However, my claims concern literature. Thus, the re-
search program bearing on those claims will most directly address
literature. In this respect, a research program in literary universals is
6 There are important traditions of literary theory associated with all the major writ-
ten traditions. Most often, these are ignored by Euro-American writers. Even when
mentioned, they are typically seen as outdated or as applicable solely to their own
literary traditions. For example, writers who mention Sanskrit literary theory at all
tend to see it as bearing only on Indian literature. To my mind, the value of non-
European literary theories is precisely in the degree to which they help point us in
the direction of universal literary principles and structures. Insofar as we are trained
in the theories of one literary tradition, we are likely to be more conscious of elements
that conform to those theories. In other words, insofar as a theory organizes and
guides our reading of and response to literary works, it will render salient certain
aspects of those works, while partially occluding others. Insofar as foreign literary
theories differ from those of our own familiar traditions, they will help bring to
our attention different features of literary works in all traditions, including our own.
Put simply, Aristotle will lead us to notice certain things about European, Indic,
Chinese, and other literary works. A great precolonial South Asian theorist, such
as Abhinavagupta, will lead us to notice other things. Moreover, as it turns out,
some of these non-European theories are remarkably congruent with recent work in
cognitive science. For this reason, I make particular use of such theories, primarily
those from South Asia, in the following pages.
7 In undertaking this project, I am indebted to my former teacher, Northrop Frye.
However, my debt to Frye is perhaps not as obvious as it might seem. Specifically,
some readers may be inclined to see my account of romantic tragi-comedy as Fryean.
However, at the level where they overlap, neither account is greatly original. To a
considerable extent, Frye took well-known facts about New Comedy and its progeny
and integrated these into a larger, typological framework. I have drawn on the same
well-known facts – along with less widely known, but no less well-established facts
about literary works in other traditions – to make claims about literary universals and
to integrate these into a quite different, explanatory framework. Leaving aside a few
details, my greatest debt to Frye, then, is not in his articulation of the romantic plot.
The influence of Frye’s work on this study is, rather, much broader. It is primarily
a matter of adopting an inductive approach aimed at isolating recurrent literary
structures through empirical study of actual literary works. On the other hand, it
is probably true that, without reading Frye, I would not have been as sensitive to
some of the specific structures discussed here. In connection with this, I have also
benefited from the work of writers such as Hayden White who have extended Frye’s
ideas in valuable ways, and from commentators on and critics of Frye’s theories,
such as Paul Hernadi (see Beyond Genre 131–51) and Tzvetan Todorov (see Chapter 1
of The Fantastic).
precisely parallel to a research program in linguistic universals. Lab-

oratory experiments are important to linguistic study. However, the
main research programs in linguistic universals have focused on the
investigation of unrelated languages. I hope that experimental psy-
chologists will take up the hypotheses articulated in the following
pages and integrate them with laboratory study, leading to revision
and improvement of the theory. For now, however, I would like to
establish the possibilities for systematic study of literary universals
in the way linguists established the possibilities for systematic study
of linguistic universals many years ago.
On the other hand, the experimental model does highlight the
fact that my reading in world literature was not designed to test hy-
potheses. In Chapter 6, I set out to change this. I set out to study
a body of literature precisely in order to evaluate my hypotheses.
In other words, I set out to encounter potentially recalcitrant data
in order to advance a research program. The first step was deter-
mining what would provide a good test case for the claims I had
been making. A good instance would be a collection of narratives
that are well preserved, highly esteemed within their culture, and
isolated from traditions that I had already studied. An ideal instance
would help compensate for two deficiencies in the previous data.
First, the work would be oral, rather than written. Though I tried to
read systematically in oral traditions, my coverage there was certainly
much less adequate than in the case of written traditions. Second, and
even more important, the work would be composed by women – not
women working in a largely male tradition (such instances are readily
available), but women working in a tradition that is not dominated
by men.
I came upon just such a body of work in the Ainu epics. They cer-
tainly included heroic and romantic tragi-comedies. However, there
were some interesting variations. These variations did not quite fit
the claims from Chapter 3. This led to a slight reformulation of the
initial hypotheses. However, it was easy to see that the same varia-
tions could be found in other traditions as well. I had simply missed
them. This was a small case of the sort of program discussed by
Lakatos. But that was not all. I began to notice that a number of the
Ainu epics treated sacrifice. Strictly speaking, this did not contradict
my previous claims. Again, there are many possibilities for narrative.
Moreover, the heroic and romantic plots were clearly prominent in

the Ainu material. However, reading the Ainu stories of sacrifice, I
realized that the same narrative structure recurs across all those other
traditions. Returning to those traditions, I found considerable mate-
rial of the same sort (for example, in the Christian story of sin, exile,
and redemption – the story of Adam, Eve, and Jesus). This led to
the formulation of a third prominent narrative structure, sacrificial
tragi-comedy, and to the related isolation of a third contextually de-
pendent prototype for happiness. In this way, the explanatory scope
of the initial theory came to be expanded by revision in response to
new data – for these literary works provide data, however complex,
in just the way natural languages do.
Faced with these three genres, one might reasonably wonder what
the connections among them might be. In the seventh chapter, hav-
ing already treated these genres individually, I take up their interre-
lations. The three are remarkably similar in their organization and
development. Chapter 7 sets out to isolate and explain the detailed
structure that they share.
The afterword returns us to the issue of emotion, now asking how
the preceding account of narrative structures and emotion prototypes
might relate to a broader theory of emotion per se. Specifically, it
considers how the biological givens of emotion come to be bound up
with social narratives – not only in fictional stories, but in our real
emotional lives as well.
In undertaking this study, my main hope is that its various descrip-

tive claims and explanatory hypotheses will contribute to our under-
standing not only of literature, but of the human mind, specifically
human emotion and the human conception of emotion (for exam-
ple, the human imagination of happiness), with all that this entails.
Again, I envision the following analyses as part of an ongoing re-
search program – not only in literature, but in cognitive science and
the psychology of emotion. They are not an application of cognitive
principles to literary works, but a development of cognitive princi-
ples through the study of literature.
At the same time, a work of this sort is not merely scientific. Un-
derstanding the breadth and depth of cultural universals – of literary
and emotive commonality – is not politically inconsequential. As we
have already noted, racial and cultural hierarchies are routinely and
necessarily justified by an appeal to putative racial and cultural dif-
ferences, even if these appeals are sometimes hidden behind univer-
salist rhetoric (much as unequal treatment and double standards are
often concealed behind rhetoric of equality and fairness). A research
program that succeeds in uncovering genuinely universal principles
of human feeling, expression, and interaction, principles that are not
relative to race or culture, runs contrary to racism and ethnocentrism.
Of course, we should not decide in favor of universalist hypotheses
simply because they appear to be politically beneficial. False univer-
sals can be deeply pernicious. But an excessive readiness to accept
universality seems an unlikely danger in the current intellectual cli-
mate, at least that of humanistic study, where a laudable emphasis on
the value of examining cultural particularity is all too often viewed
as incompatible with the study of universals.
In short, I hope that readers of this book will come to recognize
that universalism versus particularism is a false dichotomy. More
generally, I hope that, at the end of this book, readers will be more
inclined to follow Chomsky, Ngũgı̃, Marx (see Economic 114), Frantz
Fanon (10), Samir Amin (see especially the preface and final chapter),
Kwame Appiah (58, 152), Aijaz Ahmad (316 and elsewhere), Edward
Said (6 and elsewhere), and others, in recognizing both the intellectual
and political value of studying universals – in this particular case,
recognizing that our aspirations and emotions are fundamentally the
same, no matter where we were born or what we look like, and that
the stories we admire and preserve, stories about these aspirations
and emotions, are most often mere variations on a handful of shared
patterns.
1
Literary Universals
the structure of a theory of universals

The first important point about literary universals is that they are not
necessarily properties of all literary works. Indeed, such properties
are rare, and often trivial (that is, a mere residue of our definition of a
literary work). Rather, literary universals are properties and relations
found across a range of literary traditions.
What, then, constitutes “a range of literary traditions”? In linguis-
tics, one counts a shared property as evidence of a universal only if
the languages in question are genetically and areally distinct, which
is to say, only if they have distinct origins and have not influenced
one another with respect to this particular property. The basic idea
is straightforward. If a shared property is the result of a common
source – either because the languages in question have a common
ancestor or because the property has been borrowed by one language
from the other – then that property does not provide evidence of a
universal. French, Italian, and Spanish do not provide three sepa-
rate instances of a shared property, indicating that it occurs sponta-
neously in a range of languages. In all likelihood, they provide only
a single instance of that property, because in all three cases that prop-
erty probably derives from a common source. The same may be true
of a property shared by, say, Spanish and Basque. Though Spanish
and Basque are genetically distinct languages, there has been enough
interaction between speakers of these languages that the property
17
in question may well be the result of influence. In this case too, the
shared property would not count as evidence of a literary universal.
With some slight qualifications, the general principle of genetic
and areal distinctness serves as an appropriate criterion for the iso-
lation of literary universals as well.1
1 The qualifications bear on areal distinctness. It is rather difficult for one language
to influence another. Contact between the language communities must usually be
prolonged and intense. Moreover, linguistic influence is most often localized. For
example, it is common for one language to introduce vocabulary items to another
language, but it is rare for one language to introduce syntactic structures into another
language.
The problem with literature is that it is, in general, much easier to influence lit-
erary composition than to influence language. (For a sense of how areal influence
operates in literature, see, for example, Edmonson 6–23.) A single story, transported
across continents, could have a significant impact on a distinct tradition. Something
like this could never happen in language. Even the transportation of a vocabulary
item across continents will have only very local impact within the receiving language.
But, then, the ease of influence in literature cuts both ways. It presents a problem,
but at the same time suggests a partial solution. We need vocabulary items. Our
inclination is always to add more words whenever they make distinctions lacking in
our current system. For example, every time some group encounters a new species
of bird or fruit, they want a name for it. There is no similar imperative in literature. If
a single work, transported across continents, does indeed have a significant impact
on another tradition – in the absence of political or related pressures – this seems to
suggest that there was some sort of prior aesthetic propensity that this new work
satisfies. Put differently, there is a common view in linguistic theory that all aspects
of universal grammar are available to all speakers, even when those aspects of uni-
versal grammar are not instantiated in a given speaker’s language. For example,
Kiparsky points out that “because the category of syllable onset is defined in univer-
sal grammar, words with identical syllable onsets are recognised as an equivalence
class . . . whether or not the grammar of the language happens to contain rules refer-
ring to syllable onsets” (“On Theory” 192). In cases of literary influence where the
input from the influencing source is minimal, we might infer that something similar
is going on. The source literature has influenced the recipient literature because the
recipients were already sensitive to the properties of the source work, because those
properties were already universal. This is particularly likely in those cases where the
relevant properties are complex and incorporated unself-consciously into the second
tradition.
In treating literary data, then, the ordinary linguistic criteria of genetic and areal
distinctness seem inadequate. Specifically, they are not sufficiently fine grained. We
can solve this problem if we simply turn to the ideas that underlie these criteria
in linguistic study. To count as evidence for literary universals, literary properties
should be found in genetically distinct traditions. Areal influence, however, should
be further analyzed in relation to two criteria. First, we need to distinguish those
cases where areal contact was intense and long-standing – and especially those cases
where there was forced cultural hegemony – from cases where the contact was lim-
ited, sporadic, and unforced. If the areal influence was a matter of a single text
Literary Universals 19
But this does not fully determine what counts as “a range of lit-
erary traditions.” It gives us a criterion for distinguishing traditions,
but it does not tell us what constitutes an appropriate “range.” We
have already noted that a literary universal need not apply to all lit-
erary works. One might assume that it must apply to all traditions.
But this is not the case. Linguists use the term universal to refer to any
property or relation that occurs across (genetically and areally un-
related) languages with greater frequency than would be predicted
by chance alone (see, for example, Comrie 10–12, 19–22). An
absolute universal is merely a special case of this, a property or re-
lation that occurs across all traditions. Universals with a frequency
below 100% (but, of course, higher than chance) are referred to as
statistical universals.
In the following chapters, I will be concerned primarily with abso-
lute universals or near absolute universals. Some confusion can arise
from the fact that these absolute universals recur in all traditions,
but (as already noted) they need not recur in all works. Rather, their
presence in individual works only has to be greater than chance. In
other words, there is a statistical element even in absolute universals.
For example, I shall argue that romantic tragi-comedy is one of three
predominant genres in all traditions. Thus I shall argue that it is an
finding its way into a recipient culture, it seems overly stringent to discount sharped
properties on the basis of areal influence. In contrast, prolonged or coerced cultural
contact of the sort that one encounters in, say, colonial occupation, should lead us to
discount shared properties. This is why postcolonial literature cannot have a place
in a theory of literary universals. Finally, in addressing area influence, we must
consider the degree to which the properties or structures in question are open to
self-conscious choice. The simpler and more self-conscious the property in question,
the more doubtful it is as an instance of universality. For example, suppose we wish
to assert the universality of dramatic enactment, the portrayal of events by actors tak-
ing on the roles of literary characters. This is clearly something that is self-conscious.
This is the sort of thing one tradition can easily take over from another. Specific pat-
terns in background image pattern, however – or even the idea of using patterned
imagery as background in a narrative – is an unlikely candidate for casual transferal.
The former is much more akin to vocabulary; the latter is much more akin to syn-
tax. Thus, even limited contact between literary traditions may lead us to question
shared theatrical representation as evidence of a literary universal. However, limited
contact would be a poor reason to dismiss shared patterning of background imagery.
(By “background imagery,” I am referring to patterned imagery – of seasons, day
and night, color, whatever – that bears directly on the themes of the work but is
never explicitly connected with those themes, through for example simile. Rather,
it is simply present as an apparently incidental part of the scenery.)
absolute universal. However, I shall not argue that all works are ro-
mantic tragi-comedies. At minimum, I have to argue that it occurs
among paradigm works in every tradition with a frequency greater
than chance. Some readers may find it odd to refer to a universal as
absolute if it does not apply to every work. But in fact this is a perfectly
ordinary usage. Take a simple example. I suspect that homosexuality
is an absolute universal. In other words, I suspect that it recurs across
every culture. The absoluteness of this universal is unaffected by the
fact that most people in every culture are not gay or lesbian.2
More exactly, a theory of literary universals includes a repertoire of
techniques available to authors and a range of nontechnical correlations
(that is, correlations that are not techniques) derived from broad sta-
tistical patterns. Nontechnical correlations comprise universal prin-
ciples that are not devices we could use to make literature – though
they may define a range of or limits on such devices. For example,
standard line lengths appear to fall regularly between five and nine
words. Clearly, a range of standard line lengths across different tradi-
tions is not a technique available to authors. Rather, it is a broad corre-
lation across literatures. On the other hand, this universal correlation
does presumably indicate a constraint on the techniques available to
poets cross-culturally – or, if not a constraint, at least some sort of
default tendency.
Techniques include all universal matters of “form” and “content” –
including poetic meters, rhetorical devices, and so on – that an author
2 This is clearly just an issue of terminology. A more serious question concerns not the
words universal and absolute, but the possibility of calculating chance in the case of
both statistical universals and absolute universals based on statistical properties. It
is true that often one cannot calculate random probability with precision in literary
cases. However, one can calculate it adequately for the determination of universals.
Consider plot. There are uncountable topics and structures for stories in ordinary
speech. For example, I might tell a story about how my lights went out and I had to go
buy a fuse. Because of this, we cannot calculate the precise likelihood of romantic love
turning up as the topic of story after story – say, in a third to half of paradigm works
in tradition after tradition. But, given the huge number of topics for narrative, that
likelihood is clearly near zero. Moreover, the impossibility of precise calculation is not
a unique issue here. It is common in areas of psychological complexity. For example,
David Rubin notes that he cannot make precise predictions using his theory of cues
and constraints for memory in oral composition “because all possible alternatives
to the words” used in particular cases “could not be listed” (302). But this does not
affect the fact that the actual words used in poems undoubtedly conform to Rubin’s
rules at a far greater rate than would occur by chance.
may draw upon in composing a literary work. Most basic techniques

used in English literature appear to be universal. A partial list would
include symbolism and imagery, which are present in every tradition
of which I am aware; assonance – found not only in Indo-European
works, but in Japanese (Bownas lii), Babylonian (Sandars 17), and
other verse (on the related phenomenon of rhyme, see Kiparsky 10);
alliteration – important in Sanskrit (Miller 12), Japanese (Bownas lii),
Ainu (Philippi 29), and so on; verbal parallelism – found in
Tikopia (Firth with McLean 41), Igbo (Egudu and Nwoga through-
out), Basotho (Kunene 68), Kuna (Sherzer 105), Chinese (Cooper 92
and Chapter XXXV of Liu Hsieh), Babylonian (Sandars 17), Hebrew
(Sandars 17), and so on. There are also broader organizational de-
vices that appear to be universal as well. An obvious case is fore-
shadowing. It is necessarily at least as widespread as the belief in
omens, with which it is often connected (see, for example, Brewitt-
Taylor, vol. I, 1–2 or 85). Another good case is plot circularity, that
is, beginning and ending a plot in the same place or situation or in
closely analogous places or situations, often with the repetition of spe-
cific phrases concerning those places or situations (for a striking in-
stance of this, Kālidāsa’s Abhijñānaśākuntalam, see my “Beauty” 25–6).
A particularly interesting device is structural assimilation. This is the
explicit or implicit patterning of one work on the plot of another,
often culturally central work. Obvious instances would include the
common use of the story of the Fall as a model for Judeo-Christian
works. It is also widespread in Muslim traditions, as shown by
Ferdowsi’s implicit use of the story of Moses (Mūsā) in his treat-
ment of the infant Dārāb (set afloat “in a casket on the Euphrates
river” [221]) or his use of the story of Abraham (Ibrāhı̄m) and Ishmael
(Ismā’ı̄l) in presenting Rostám and Sohráb. In India, we find a promi-
nent case in Kālidāsa’s modeling of Abhijñānaśākuntalam on parts of
the Rāmāyan.a (for discussion, see my “Beauty,” 30).
These techniques are organized into (explicit or implicit) schemas
defining literary types and subtypes, such as “sonnet,” in English.
Within schemas, techniques may be obligatory or optional. Obligatory
techniques are techniques an author must use. For example, if a cer-
tain poetic genre requires the use of end rhyme, then end rhyme is
obligatory in the schema for that genre. Optional techniques, on the
other hand, are techniques that are available to an author, but are
not obligatory. For example, the sonnet does not require alliteration,
but it permits alliteration; thus, alliteration is an optional technique
available to a poet composing a sonnet. We may also distinguish
techniques that, while not strictly obligatory, are standard. In rela-
tion to alternatives, we could understand such standard techniques
as the highest or default cases within a schema. A standard technique
is employed unless the author makes an explicit choice to the contrary,
or some concurrently operating principle or schema prevents the
implementation of the default.
Perhaps the best way to conceive of these schemas is by reference
to cross-indexed entries in one’s mental lexicon. Some techniques are
specified directly in the schema, which of course has its own lexical
entry (for example, “sonnet”). Others are made accessible indirectly
by reference to distinct lexical entries. These other entries may simply
be coordinated (that is, fully distinct, though cross-indexed), but they
may also be superordinate (encompassing the entry in question) or
subordinate (encompassed by the entry in question). Thus the schema
for “sonnet” might be structured in the following way. It would
list features specific to the sonnet (number of lines, rhyme scheme),
then add some reference to the overarching category “poem” (we
could think of it as a “See ‘poem’” instruction). The entry for “poem”
would be a superordinate category to “sonnet” and would include
a list of techniques standard in poetry and available for use in son-
nets (for example, alliteration). The precise nature of each of these
techniques could be viewed as defined in coordinate categories, also
cross-indexed (for example, with a “See ‘alliteration’” instruction).
Organization of obligatory and optional techniques into schemas,
crossindexing of techniques, and so on, all appear to be universal as
well.
The most broadly encompassing schema, and an absolute univer-
sal, is the minimally specified schema of verbal art itself. As Kiparsky
has pointed out, all societies have verbal art (“On Theory” 195–6). This
may seem a mere triviality, but it is not. There is no logical necessity
in the existence of verbal art. In our own society, very few people
actually produce verbal art. Why, then, should we expect it to appear
in every society? As Chomsky has emphasized, one of the first tasks
for researchers in the study of universals is to overcome habituation
and recognize how surprising universals are. We often “lose sight
of the need for explanation when phenomena are too familiar and
‘obvious’” (Language and Mind 25). Once we have recognized that
our expectation of verbal art is a mere matter of habit, we come to see
that, far from being trivial, it is in fact highly surprising that verbal art
is produced in small nomadic groups and in vast, highly urbanized
nations.
Below this, in their most general forms, the three major genres of
European literature – poetry, prose fiction, and drama – appear to be
instances of larger universal categories as well. Thus it appears to
be a universal that all or almost all societies have verse, which is to
say a verbal art involving formalized cyclical organization of speech
based on fixed, recurring patterns of acoustic properties. Tale telling
also appears to be a literary universal. Probably in all societies, peo-
ple articulate causal sequences of nonbanal events involving human
agency (with banality defined relative to culturally specific expecta-
tions), and they do so at least in part for aesthetic enjoyment, itself
based on identification, the patterned variation of emotional inten-
sity, and so on. Finally, some form of enactment for such tales seems
to be universal as well, though more limitedly. In some societies, this
may be confined to brief episodes on festival occasions or in rituals.
However, other societies – including all the major written traditions –
have developed a form of extended drama. Thus, we find elaborated
theatrical works in Europe, India, Southeast Asia, China, Japan, and
the Middle East. (As the Middle East is often claimed to have had no
precolonial theater, it is perhaps worth referring the interested reader
to Moreh, Martinovitch, and Chapter 1 of Badawi.)
More specific schematic patterns are isolable as well. To take a case
we shall discuss in detail in Chapter 3, it appears that every tradition
tells tales of conflict in two areas – love and political power. More-
over, these tales involve a wide range of common character types
and motifs, fall into similar subgenres, and so on. Consider love
stories. A romantic comedy, in its most minimal form, typically in-
volves two lovers who are separated, then reunited after a period
of uncertainty. It is already surprising that this structure should be
found in drama from Greece, Rome, India, China, and Japan, and
in stories from other regions as well. More surprising still is the fact
that more particular patterns in this genre are also widely shared.
For instance, the separation is typically a result of the lovers’ conflict
with social expectations and structures, often manifest in a conflict

with parents. It is frequently resolved by some sort of recognition,
leading to a reversal, in the standard Aristotelian manner. Moreover,
this recognition not only reunites the lovers, it often involves the un-
expected reunion of parents and children. It includes such characters
as a hero, heroine, hero’s companion, hero’s parasite, heroine’s confi-
dant, and so on. Indeed, the character typologies set out by Sanskrit
literary theorists two millennia ago (for example, in Bharatamuni’s
Nāt.yaśāstra) and those drawn from Greek and Roman New Comedy
by Northrop Frye are similar to one another, and widely appli-
cable beyond their own traditions, because of this cross-cultural
consistency.
Perhaps the most cross-culturally widespread version of the love
plot is a particular variation on the comic love story. This varia-
tion, “romantic tragi-comedy,” in effect includes the tragic love story,
where the lovers are separated, typically by death, often with a sug-
gestion of literal or metaphorical reunion after death – as in Arabic
and Persian retellings of the Laylā and Majnūn story (for example,
Niz.āmı̄’s twelfth-century poem), the Rāmāyan.a (100 b.c.e.), the love
suicide plays of Chikamatsu (for example, Love Suicides in the Women’s
Temple of 1708 [in Keene Major Plays]), and so on.3 Specifically, in
romantic tragi-comedy, lovers are almost joined, then separated in
a way that suggests death, then reunited in a sort of resurrection.
The separation at least threatens not only to keep the lovers apart,
but to prevent them from ever hearing of one another again (as in
Zeami’s early fifteenth-century Lady Han or The Reed Cutter [both
in Keene Twenty Plays]). This sort of separation is already akin to
3 Some readers might question my use of the term “tragi-comedy” here, complaining
that tragi-comedy is necessarily a late development that synthesizes tragedy and
comedy. I use “tragi-comedy” to refer to plots that pass through or closely approach
an apparently tragic conclusion before resolving happily. Since these are comedies, I
could simply have referred to them as such. I use the term “tragi-comedy,” however,
to signal the complex structure and to emphasize the close relation of tragedy to
comedy. In Chapter 3, I shall argue that tragi-comedy, in this sense, is the fullest
and most widespread literary form cross-culturally. Moreover, I shall argue that
tragedy is not a component of tragi-comedy, but a derivative of tragi-comedy – in
effect, a shortening of tragi-comedy. Indeed, according to the analysis in Chapter 3,
tragedy is only possible as a failed comedy, for the nature of narrative development
is necessarily oriented toward comedy.
death, but the link with death is typically more direct. Thus, the sep-
aration often involves an apparent death – as in Charitōn’s early
Greek romance, Chaereās and Kallirrhŏē; Shakespeare’s The Winter’s
Tale; Bhāsa’s fourth-century Vision of Vāsavadattā (in Woolner and
Sarup); Bhavabhūti’s eighth-century Uttararāmacarita; and Manohra
(in Brandon, Traditional Asian Plays), the “earliest known drama in
Thailand” (Brandon, Cambridge 234). If there is no literal death or near
death, the separation is frequently represented in imagery closely as-
sociated with death (as in Śakuntalā’s assumption into the heavens
or Chien-nü’s “soul leaving her body” in Chêng’s fourteenth-century
play [in Liu Jung-en]). In any case, the link with death is clear, consis-
tent, and important in a wide range of literary traditions. In keeping
with this, there is a regular association of the separation with imagery
of seasonal demise (for example, winter) and the reunion with im-
agery of seasonal rebirth (for example, spring). Additionally, more
general comic universals – conflict with society, recognition, reversal,
reunion of separated parents and children, and so on – carry over as
well, giving this schema remarkably detailed cross-cultural consis-
tency. In Europe, this sort of story most obviously makes us think of
Shakespeare, but it has been a standard part of European literature
for millennia, with prominent instances including, for example, the
“Erōtici Graeci” of the early centuries c.e. Outside Europe, beyond
the works already mentioned, we could list the first-century Toy Cart
and the seventh-century Ratnāvalı̄ (in Lal) from India; the thirteenth-
century Chang Boils the Sea (in Liu Jung-en) and the roughly con-
temporary story of Qiu Hu (see Dolby 155–6) from China; the final
voyage of Sindbād and the story of Aladdin (‘Alā’ al-dı̄n) from the
Middle East (in Dawood, Tales, on dating see 8–9; see also Allen 176
and 168 on the provenance of these stories); Kan’ami’s fourteenth-
century Hanakatami (in Waley Nō Plays) and the eighteenth-century
Love Letter from the Licensed Quarter (in Brandon, Kabuki) from Japan –
to take just two examples from each region that has produced a major
written tradition.
It is important to point out that the universality of either a (gen-
eral) technique or a nontechnical correlation in no way implies the
universality of any specification or instantiation of that technique
or correlation, nor is putative universality falsified by differences in
such specifications or instantiations. For example, the patterning of
images seems to be an absolute universal. It is certainly found in the

genetically distinct, and in part areally distinct, written traditions of
Europe, the Middle East, India, China, and Japan; it is also found in the
unrelated oral poetries of the Southern African Basotho (see Chapter 7
of Kunene), the Polynesian Tikopia (see Firth with McLean 36–9), the
Yirrkalla of Arnhem Land (see Berndt 73–6), and so on. This does not,
in and of itself, imply that implementations of the technique share
further universal properties. For instance, it appears to be a universal
that love is commonly associated with images of birds. Yet, if so, it is
a distinct universal. Even if different cultures used widely different
image patterns – some linking love with birds, others with tubers,
others with types of fabric – the use of image patterns would still be
universal.
Note, in connection with this, that each level of a hierarchy of
abstraction serves as a partial explanatory generalization of all ele-
ments on lower levels. Thus, we may find that the use of bird images
for romantic love and the use of seasonal images for human life are
far more common than chance, but not absolute. We may then find
the more general use of image patterns to be absolute. In this case,
the more abstract universal of imagery would partially explain the
more specific avian and seasonal images. Other universals – literary
and nonliterary – in combination with more specific accidental cir-
cumstances (such as environmental conditions), would yield a fuller
explanation of these lower-level universals. For example, the cross-
cultural connection between romantic love and imagery of birds may
derive in part from a universal, metaphorical, but not specifically lit-
erary correlation between “positive” emotions and the direction up
(cf. Chapter 5 of Lakoff and Johnson).
The more abstract universals may also be explained further. For
example, the patterning of images may in turn be an instance of a
still broader universal according to which the limited patterning of
normal discourse is generalized to all levels of structure in literary
art (as we shall discuss below). At the highest level, the complex of
distinctively literary universals should indicate what is at the origin
of the development of literature, what defines the human urge to
make and experience verbal art. Of course, these overarching literary
universals are not the end of the story either, for they too should
be open to further explanation in terms of even more encompassing
universals, such as those of psychology, sociology, history, and so on.

(We shall return to this topic also in the following section.)
Hierarchies of universals are defined not only by a receding series
of explanatory abstractions, but also by a series of conditional rela-
tions. As a general methodological principle, linguists (like scientists
in any other field) seek to redefine universals in such a way as to
limit exceptions. Through repeated reformulation, they seek to bring
statistical universals closer to absolute universals. Again, this may be
done through abstraction, as when the (perhaps) absolute universal
of image patterning is derived from the (perhaps) merely statistically
universal patterns of bird imagery, season imagery, and so on. (In
fact, bird and season imagery are probably absolute or near absolute
universals. The point is illustrative.) Beyond this, however, the goal
of absolute universality may be pursued through the delineation of
specific conditions in which statistically universal techniques or cor-
relations occur. In other words, statistical universals of the form “q”
may be revised into implicational universals of the form “If p, then q.”
Ideally, this reformulation would yield an absolute universal (that
is, there would be no cases of p and -q). At least, it would limit the
number of exceptions, bringing the universal closer to a frequency of
100%.
Alliteration provides a good example of how universals may be
redefined to limit exceptions. It is an obligatory technique in certain
forms of poetry in certain societies – more than would be predicted
by chance, but with a great many exceptions. One may abstract from
this to some broader principle, such as the no doubt absolute uni-
versal that all poetry has some obligatory features relating to sound
pattern. However, one may equally seek to formulate an implica-
tional universal, determining conditions in which alliteration is or is
not obligatory. Kiparsky argues that alliteration “seems to be found
as an obligatory formal element only in languages where the stress
regularly falls on the same syllable in the word, which then must be
the alliterating syllable” (Kiparsky, “The Role” 9). While this does
not fully fix conditions under which alliteration is obligatory, it does
fully fix conditions under which it is not obligatory (that is, whenever
syllable stress varies in the language). Thus, it yields the absolute im-
plicational universal, “If syllable stress varies in a given language then
alliteration is not an obligatory feature of poetry in that language.”
It also yields a second, more complex absolute implicational univer-

sal: “If syllable stress does not vary in a language and if alliteration
is an obligatory feature of poetry in that language then the stressed
syllable is the alliterating syllable.”
Much as unconditional universals may be subsumed into hierar-
chies of abstraction, implicational universals may be organized into
typologies (on this and other aspects of universals in linguistic study,
see Comrie and Croft). A typology consists in a set of mutually exclu-
sive categories, each of which coordinates a number of implicational
universals, forming them into a pattern that is more informative than
any of the implications considered individually. Each type in a ty-
pology serves as a partial explanation of any given implication that
it subsumes. In addition, the typology as a whole should come close
to defining an absolute, disjunctive universal – that is, all, or almost
all, literatures should fit under one or another type.
The broad distinction between oral and literate composition is a
case in point (on this distinction, see Ong). In effect, it sets up a broad
implication of the form: “If a culture does not have writing, then
its verbal art will be marked by frequent use of epithets, formulaic
phrases, specific sorts of repetition, etc.,” or more generally “The de-
gree to which a body of verbal art is marked by epithets, etc., is a
function of the degree of literacy of the culture in which that ver-
bal art is produced.” As the second formulation indicates, there is a
spectrum of possibilities here with two general tendencies as end-
points, rather than a universal grid of discrete types; however, such a
spectrum serves the same organizational and explanatory function.
Though some writers have disputed the validity of this loose typolog-
ical distinction (see, for example, Sherzer and Woodbury 9–10), it has
been convincingly applied to a wide range of literatures: Slavic, Greek
(see Lord), Sanskrit, Tamil (see Kailasapathy; though see also Hart,
Poems), Ainu (see the introduction to Philippi), Xhosa (see Chapter 6
of Opland), and so on.4
4 William Brewer’s “The Story Schema: Universal and Culture-Specific Properties”

treats some typological universals of orature. Brewer’s essay is valuable and in-
sightful. At the same time, it illustrates some of the problems that occur in the study
of universals, primarily the difficulty of noticing universals, a difficulty stressed by
Chomsky, as we have already noted. For example, Brewer explains that “Conven-
tionalized story openings occur throughout the world” (179). This would seem to be
Indeed, this is a particularly interesting universal for our purposes

as it makes reference to historical conditions. Humanists tend to think
that the study of history and the study of universals are diametrically
opposed. In fact, they are not. Marx’s isolation of historical/economic
laws (Marx, Capital 8) is an obvious case in which historical change has
been understood in relation to universals. Moreover, historical lin-
guists have been no less inclined to isolate universals than have their
colleagues in other areas of linguistics (see, for example, Chapter 10
of Comrie and Chapters 8 and 9 of Croft). In this case, the proposed
typology not only involves historical contingency (the presence or
absence of writing), it also involves historical development and thus
specificity – though all in the context of universal principles. Another
instance of this type is Arnold Hauser’s hypothesis that realism in
literature is a function of social structure, the more realistic art be-
ing produced in urban, market economies, with feudal or tributary
economies fostering stylization (see Hauser 49).
Related to this, a particularly important distinction, relevant to his-
torical and cultural variation, is that between indexical and nonindexi-
cal universals. Indexical universals are universals that cannot be fully
particularized by reference to shared social or cognitive properties
alone (for example, limits on the capacity of human working mem-
ory or the type of economic system in which works are produced).
Rather, to be fully particularized, indexical universals must make
reference to individually variable, subjective conditions (such as per-
sonal memories). For example, it appears to be a psychological uni-
versal that one’s conception of oneself is structured into a hierarchy
of properties, such that properties important to one’s self-conception
(such as sex) are high in the hierarchy and properties less important
to that self-conception (such as ring size) are lower in the hierar-
chy. It seems that readers and auditors identify with a character on
the basis of shared high-level properties in their self-conception and
that they prefer works involving characters with whom they iden-
tify (see Klemenz-Belgardt 367–8 and citations).This (likely) literary
a clear case of a universal. However, Brewer places this in the category of “culture-
specific properties” (177) because the particular form of those openings varies. It
is, of course, important to recognize that the precise conventions differ (that is, all
cultures do not use “Once upon a time”). But it is at least as important to recognize
that conventionalization of openings is itself a universal.
universal is indexical because those high-level properties – and thus

the associated literary preferences – vary from person to person (by
sex, race, religion, and so on). One common preconception about
literary universals is that they entail universal agreement in mat-
ters of taste. The existence of indexical universals shows that this
preconception is false.
Finally, we shall see in subsequent chapters that many difficulties
in the study of universals may be overcome if we understand that
universals are more likely to be based in prototypes than in necessary
and sufficient conditions. Prototypes present us with standard cases.
It may seem that prototypes must be less universal than necessary
and sufficient conditions as they are more concrete or less abstract.
But, in fact, it turns out that standard cases are far more likely to be
universal than are limiting definitions. Put differently, cross-cultural
variation (and, indeed, individual variation) is most often a matter
of variation in marginal cases, limiting instances that bear on the
formulation of necessary and sufficient conditions; it is much less
frequently a matter of central or prototypical cases. For example, in
looking at narrative and emotion, we will be considering prototype
cases of narrative and prototype cases of emotion. We will not be
trying to formulate necessary and sufficient conditions for narrative
or emotion. Again, it is the former, not the latter, that appear most
universal. In this way, the more we look to prototypes, the more likely
we are to find absolute universals.
explaining literary universals: the nature of a

research program
One could draw further, consequential distinctions – for example,
between universals bearing on aesthetic experience and universals
bearing on evaluation outside aesthetic experience (for example, uni-
versals concerning canonization and dominant ideology). However,
it is beyond the scope of this chapter, or this book, to present a final
and complete listing of concepts germane to a theory of universals.
Indeed, an empirically based theory – unlike the speculative theories
common in the humanities – cannot arise in a fully developed and
final form, ready only to be “applied” in explications of individual
texts. Rather, an empirically based theory is always and necessarily
part of an ongoing, broadly collaborative research program – in this

case, a program involving the collaboration of a wide range of schol-
ars in different fields, with different areas of literary expertise. For
present purposes, the preceding analysis should provide an adequate
idea of how a theory of literary universals can be structured initially,
and what some parts of that theory might look like.
Needless to say, a research program of this sort necessarily
seeks not only descriptive adequacy, but explanatory adequacy (see
Chomsky, Aspects 24–6). Descriptive adequacy is achieved when one
formulates a generative account of the phenomena being studied
(for example, language). A generative account presents rules that
produce the phenomena. Suppose we are doing grammar. We dis-
cover that some languages place the object before the verb and some
place it after the verb. We then discover that some languages place
adpositions before the noun (making them prepositions), while oth-
ers place them after the noun (making them postpositions). We go
on discovering things of this sort. We then notice further patterns. A
certain adposition/noun order is regularly correlated with a certain
object/verb order and so on. We conclude that head/complement
order is usually constant within a language and that there are just
two options for head/complement order: head/complement, com-
plement/head. Each time we isolate a pattern and formulate that
pattern as a rule, we have a more descriptively adequate account of
the phenomena under consideration. Note that in physics this is usu-
ally all we have to do. In physics, once we can generate the patterns
from rules, we usually stop. But this is not true in linguistics or in
other areas of cognitive study. In linguistics, we need to formulate
a set of rules that generate the language, as in physics we formulate
a set of rules that generate observable physical phenomena. But, in
addition to this, we need to account for just how this set of rules
gets into people’s minds.5 How does a child acquire this complex
grammar? Linguistics and related fields of psychology must add this
second step to a theory. This second step is “explanatory adequacy.”
5 Of course, the rules might, for the most part, get there genetically. For example, in
Chomsky’s widely influential “principles and parameters” approach, the principles
of grammar are innate and experience serves simply to set parameters (see, for
example, Chapters 1 and 5 of Chomsky’s New Horizons). In this account, explanatory
adequacy is achieved largely by innatism.
In the present study, my aim is to isolate universal patterns of

story structure in such a way that we can see many specific stories
as instances of these patterns (descriptive adequacy), then to explain
these patterns by reference to well-defined cognitive principles (ex-
planatory adequacy). Thus, having treated some broadly structural
and descriptive aspects of a theory of literary universals in the pre-
ceding section, I should like, in this section, to consider the way in
which the explanatory part of a research program in universals might
proceed. To do this, I shall examine two specific cases, one concern-
ing a complex of universal formal techniques and a principle that
can be abstracted from those techniques, the other concerning a uni-
versal statistical correlation and its likely derivation from cognitive
structure.
As to the first, consider the list of formal devices that are used
in a wide range of genetically distinct traditions – assonance, allit-
eration, parallelism, and so on. Again, these are best thought of as
techniques available to writers in creating literary works. A first step
in an explanatory research program would be to abstract some sort
of principle from this list, a principle that indicates what these items
share, what pattern they form, and that does so in such a way as
to give this pattern some sort of function in verbal art. We could
call such a “function-congruent” abstraction from empirically ob-
servable patterns a “secondary principle.” Paul Kiparsky has noted
what is at issue in this particular case. He observes broadly that
“it appears . . . all literary traditions . . . utilize the same elements of
form” (“The Role” 11). Following Roman Jakobson, he goes on to
suggest a reason for this: “[L]anguage allows certain ways of orga-
nizing sounds, and . . . poetic form must draw on this organization”
(20). He concludes that the relations “between grammar and poetry
account, at least in part, for the universality of poetic form” (22).
The general connection between linguistic sounds and poetic
sounds is plausible, and Kiparsky advances the discussion by draw-
ing this link. However, this falls well short of an explanation, even
such a minimal explanation as is given in a secondary principle,
for it does not say anything about the specific use of the linguistic
phenomena in literature. Part of the point of the literary devices is
that they are different from the ordinary linguistic phenomena to
which Kiparsky reasonably relates them. For example, onsets (the
beginnings of syllables) are not used in the same way in ordinary

language and in literature. The two uses are continuous, but not iden-
tical. We do not seek alliteration (or rhyme or assonance) in ordinary
talk. We do in poetry. Relating the two may be part of an explanatory
account, but necessarily only a part.
What, then, is going on in literature? Literary theorists from dif-
ferent traditions regularly stress the unusual degree of structure and
relevance in literature (see, for example, the valuable discussion in
Bateson, especially 14). Presumably at least some of our enjoyment
of verbal art is related to this. All speech is patterned. Whenever we
speak, we try to make a coherent statement, present a coherent narra-
tive, and so on. We choose our words so that they have the right sort
of connotations. We try to avoid harsh or comic sequences of sounds.
However, in the creation of verbal art, we do more of this and we
do it more intensively. One could say that we seek to maximize this
sort of patterning. In other words, we seek to render the causal se-
quence of the plot more rigorous, reenforcing it with foreshadowing
and circularity. We seek to coordinate connotations and ambiguities
of the words and phrases – including purely graphic connotations
and ambiguities where these occur, as in Chinese (Cooper 68–72). We
also seek to pattern the sounds through rhythm, assonance, alliter-
ation, and so forth. One differentia of literature, then, would seem to
be a maximization of patterning, a maximization that makes a wide
range of features (narrative, semantic, phonetic, and so on) relevant
to our literary experience.6
6 Because I use the word “relevance” here, a term associated with Sperber and Wilson,
it is important to stress that my notion of relevance has no special relation to their
work. Gibbs gives a nice summary of the Sperber/Wilson approach: “Relevance is
defined in terms of contextual effects and processing effort. Contextual effects are
achieved when a speaker’s utterance strengthens, contradicts, or denies an existing
assumption or when a speaker’s utterance is combined with an existing assumption
to yield some new contextual implications” (230). My use of “relevance” has no par-
ticular bearing on contextual effects or processing effort. In my usage, maximizing
relevance is a matter of taking more and more properties and relations as germane to
the experience of the literary work. Of course, these properties and relations may be
germane in the sense that they “yield some new . . . implications.” But they need not
be. Indeed, a wide range of poetically relevant features rarely bear on implications.
More exactly, when I listen to a conversation, I simply do not attend to many
features of the speaker’s language, such as assonance or patterns in syllabic stress
(for example, iambs). Obviously, I pay attention to contrastive stress and other forms
But this is only a first approximation. The Jakobson/Kiparsky hy-

pothesis indicates that formal poetic devices are not merely a mat-
ter of maximizing relevance, but of maximizing a certain sort of
relevance – linguistically specified relevance, in their view. As
Kiparsky stresses, “certain patterns of considerable formal simplic-
ity are never utilized in the construction of verse. . . . For example, no
one thinks of filling in a stanzaic pattern on the principle that the
last words of certain lines must contain the same number of sounds”
(“The Role” 12–13). Why not? Kiparsky’s answer is that no linguistic
rule involves counting sounds in this way and thus no poetic rule will
do so. But this formulation is clearly too narrow. It covers the case
at hand, but it does not cover, for example, imagery, foreshadowing,
or other nonlinguistic patterns open to maximization. Many sorts of
maximization are not narrowly linguistic and thus cannot be covered
by Kiparsky’s principle. Moreover, it does not seem to provide an
adequate explanation of the limitations on maximization even in the
case of linguistic phenomena. One is left asking, “Why is there such
a linguistic constraint?,” even in the relevant instances.
So, Kiparsky is responding to a genuine problem, but he seems
to be responding to it too narrowly. One way of trying to resolve
this dilemma would be to consider what it is about Kiparsky’s ac-
count that allows it to solve the linguistic cases, then to see if this
aspect of the account can be generalized. Consider again Kiparsky’s
examples. One obvious and crucial difference between, say, onsets
(or beginnings) of syllables (used both in language rules and in po-
etry) and number of speech sounds (used neither in language rules
of semantically relevant emphasis, but that is different. Indeed, its difference is pre-
cisely the point. When I read poetry, I do attend to such features as assonance and
recurring patterns in syllabic stress. They become relevant to my experience. But,
unlike contrastive stress, assonance and patterns in syllabic stress do not necessarily,
nor even typically, yield implications. The point is obvious as soon as one thinks of
concrete examples. The presence of iambic pentameter is relevant to my experience
of a huge variety of poems in English. However, with rare exceptions, it does not
strengthen, contradict, or deny an existing assumption, nor does it yield any con-
textual implications, as is obvious from the fact that it is relevant to my experience
of eulogies and satires, poems of victory and of defeat, poems of love and of hate,
and so on. I do not find a poem of love contradicted by iambic pentameter – nor do
I find a poem of hate contradicted by it.
In short, the work of Sperber and Wilson, though important and insightful, just
has no special bearing on my concept of maximizing relevance in literature.
nor in poetry) is that we “hear” the former, but have to calculate the
latter. More generally, any linguistic feature that is part of a linguistic
rule is a feature we “hear” – not in the sense that we are conscious
of it (typically we are not), but in the sense that it makes a difference
to our experience. For the most part, we do not “hear” features not
included in linguistic rules, but can at best calculate them. Indeed,
when Kiparsky elaborates his hypothesis, he makes the point himself,
arguing that the “faculty of language . . . equips” us with “modes of
perceiving” certain features, but not others (Kiparsky, “On Theory”
191).
Insofar as this notion can be generalized, it would seem to solve the
problem we have been considering. And, as it turns out, the idea can
be generalized easily. Indeed, it is well known in cognitive science.
The linguistic “hearing” or “perceiving” of onsets (but not number of
speech sounds) is simply a specific case of a more general cognitive
mechanism, called “encoding.” Encoding, then, appears to be what
is crucial in all these cases, both those that fit Kiparsky’s model (for
example, alliteration) and those that do not (for example, foreshad-
owing). More exactly, whenever we perceive something, we do not
perceive every aspect and relation of the thing. That would be impos-
sible. Rather, we perceive some aspects and store them in memory
while others escape us. The aspects we perceive and store are the
aspects we “encode.” Holland, Holyoak, Nisbett, and Thagard give
the following example: “[Y]ounger children often cannot learn about
the rules underlying the behavior of balance beams” simply from ob-
servation because they “do not encode the distance of objects from
the fulcrum.” However, when the distance is pointed out to them,
they begin to encode the feature on new observations, and are able to
induce rules (Holland et al. 55). Linguistic rules are just a particular
case of cognitive principles that allow for encoding.
This leads us to reformulate our earlier principle. Now we would
say that a wide variety of formal literary techniques (alliteration, as-
sonance, circularity, foreshadowing, and so on) function to maximize
relevance or patterning across encoded properties or relations. These
literary techniques could be thought of as particularizations of a rel-
evance/patterning schema in which the properties or relations in
question are values of variables, with the variables necessarily con-
fined to the class of encodable values. Thus, when applied to onsets
(which are encodable), this schema yields alliteration. When app-

lied to speech rhythm (which is encodable), it yields meter and
related forms of organization. When applied to the background con-
stituents of a scene (for example, landscape), it yields a form of image
patterning.
Note also that encoding can be learned, though there are limits
to what can be encoded. This is important because it seems clear
that greater experience of and training in literature increases one’s
sensitivity to certain sorts of patterns. This is only to be expected
when the process is understood in terms of the general cognitive
process of encoding. It is less obvious that Kiparsky’s hypothesis
could accommodate this sort of development.
On the other hand, not all of our problems are yet solved. Our
revised formulation does not allow for limits on maximization; it
treats maximization as a good in and of itself. Literature, however,
does not absolutely maximize patterns (cf. Bateson 18). It does so
only to a certain point. One degree of alliteration is aesthetic, but
even a little more may be comic. This, too, is true cross-culturally.
Some traditions may employ more alliteration than others, but, with
rare exceptions, they do not equate more alliteration with a better
poem (other things being equal), as this formulation would appear
to imply. Rather, alliteration, rhyme, and so on reach a sort of ceiling,
after which they detract from aesthetic effect.
To explain this, we need to extend our analysis of perception.
Again, in perceiving any object, we fail to encode some features at
all; other features we do encode. Of those features we encode, some
have, so to speak, forced themselves on us, drawing our attention.
A well-known case of this concerns background conversation. Typi-
cally we are not paying attention to background conversation when
we are engaged in a conversation ourselves. However, if we hear
our name mentioned, our attention will suddenly shift to the back-
ground conversation, entirely independent of conscious decision (see
Johnson-Laird 148), and perhaps even against our will. Miller and
Johnson-Laird note that “surprising stimuli” have this obtrusive or
attention-forcing effect as well (133).
More generally, we might say that any perceptual feature has a cer-
tain degree of saliency for a particular perceiver in particular circum-
stances. (Features that are not encoded could be thought of as having
a saliency of zero.) Various qualities of the feature, the perceiver, and

the context determine that degree of saliency. As just noted, unex-
pectedness of the feature is one such quality. For temporally ordered
occurrences, frequency of repetition would be another. Once the de-
gree of saliency (resulting from these qualities) goes above a certain
threshold, it automatically draws our attention. We can refer to this
as the “threshold of forced attentional focus.” Consider the quality
of frequency. Suppose I use the phrase “of course” more than other
speakers. If I use it once every other paragraph, the attention of read-
ers is unlikely to be drawn to this usage (that is, they are unlikely to
“notice”). As frequency increases, however – suppose I use it in every
sentence – the usage will become obtrusive, “drawing attention to it-
self.” In other words, it will eventually cross the threshold of forced
attentional focus. This threshold probably varies somewhat both cul-
turally and individually. However, it is also no doubt governed by
broad cognitive constraints.
We may further refine our general principle, drawing on this
distinction as follows: A wide variety of formal literary techniques
(alliteration, assonance, foreshadowing, circularity, and so on) func-
tion to maximize relevance or patterning across encoded properties
or relations with a normative limit at the point where such maximization
would surpass the threshold of forced attentional focus.
There are still exceptions to this formulation. They typically in-
volve an extension of forced attentional focus to the point where
use of the feature comes to be seen as humorous or as manifesting
a sort of cadenza-like virtuosity on the part of the poet. The pre-
cise nature of these exceptions would be further examined in an
ongoing research program. However, this should adequately illus-
trate the abstraction of secondary principles and their refinement
in such a program. Therefore, I should now like to turn to another
aspect of such a research program, considering a nontechnical uni-
versal with a more direct relation to a particular aspect of cognitive
structure.
As we have already noted, standard line lengths for poetry in a
wide range of traditions tend to fall between five and nine words.
Standard line length is of course defined not in terms of words, but
in terms of some acoustic property. However, it typically puts the
number of words per line in this range – and that is what is crucial
for achieving descriptive adequacy in this case.7 Thus, in Chinese, a

monosyllabic language, one standard line length has only five syl-
lables, which equals five words, while another has seven syllables
and thus seven words (see Cooper 63). The Yirrkalla poems quoted
by Berndt (Appendices 1 and 2), the Dinka songs cited by Deng, the
Basotho verses quoted in Kunene, many of the Hawai’ian poems in
Pukui and Korn, and the Babylonian creation poems discussed by
Sandars (17) also fit this pattern reasonably well. As to European
literatures, the first twenty lines of The Canterbury Tales contain
144 words, about seven per line; the first twenty lines of Paradise
Lost average a bit under eight words per line; twenty lines taken at
random from the Aeneid have just under seven words per line; and
twenty from the Odyssey have almost exactly seven words per line.
Blake’s Songs of Innocence have unusually short lines, but the first
poem of the sequence (Kazin 83), which is in no obvious way formally
different from those that follow, has six words per line. French lines
tend to be unusually lengthy, but the first twenty lines of Racine’s
Phèdre include just under nine words per line and thus still do not
exceed the range.
Nonetheless, there are many exceptions. In considering these, we
need, first of all, to determine which counterexamples fall within the
range of phenomena we are seeking to characterize. Clearly, our con-
cern here is with literary forms defined by fixed, recurrent, phonetic
patterns, such as iambic pentameter. Noting this makes free verse
irrelevant and thus eliminates a large number of possible counterex-
amples. Moreover, we are, as already noted, speaking about standard
line lengths. Nonetheless, there remain a number of recalcitrant cases.
One option would be to say that the universal is statistical and lim-
ited in application – well above chance, certainly, but with many
7 Different traditions formulate their criteria for poetic line length in different ways.
The lines bear a much greater similarity than the stated criteria would indicate.
To achieve descriptive adequacy in such a case, we need to specify the similarity
among the lines as closely as possible. To do that, we have to set aside the official
formulations (for example, the definition of the standard English line as iambic
pentameter), because these do not reflect the cross-cultural pattern. In fact, they
obscure it. I have attempted to achieve descriptive adequacy, capturing the cross-
cultural similarity, by formulating the pattern by reference to the number of words.
This formulation also points directly to an explanatory account of the pattern in
terms of cognitive structure, as I discuss as follows.
exceptions. This is, of course, possible. But it is not a solution as we

must go on to determine what gives rise to the exceptions anyway.
Another option is to change the predicate of the universal. In every
exception of which I am aware, the standard line length is less than
five words, not more than nine. One way of dealing with these excep-
tions would be to revise the universal to say that all standard lines are
less than nine words. However, the five- to nine-word spread of line
lengths fits well with the structure of rehearsal memory or, equiva-
lently, working memory (see Garman 322; for a fuller discussion, see
Gathercole and Baddeley). Rehearsal memory allows us to “keep in
mind” strings of information, cycling them through our attention as
we try to complete some encompassing cognitive task. Specifically,
rehearsal memory is structured in such a way as to include a lim-
ited amount of information – typically, five to nine words – at any
given time. There are several reasons why this correlation is theoreti-
cally appealing. Most importantly, poetry in all traditions demands a
sort of plenary attention, for which rehearsal memory seems ideally
designed.
More exactly, as the great tenth-century Indian theorist,
Abhinavagupta, put it, we “savor” poetry: “Aesthetical experience
takes place . . . by virtue, as it were, of the squeezing out of the po-
etical word. Persons aesthetically sensitive, indeed, read and taste
many times over the same poem. In contradiction to practical means
of perception, that, their task being accomplished, are no more of
any use and must then be abandoned, a poem, indeed, does not lose
its value after it has been comprehended” (quoted in Gnoli xxxii).
Abhinavagupta’s view here is almost a necessary consequence of the
maximization of unobtrusive patterning. As more features become
relevant to our experience of a literary work, we are less and less
able to appreciate it without “savoring.” More features must be en-
coded in our experience of poetry than in our experience of ordinary
speech. The most obvious way of ensuring this is through rehearsal
memory, which does, in effect, allow us to “savor” segments of a
poem. It makes a great deal of cognitive sense that the unit of savor-
ing would be the poetic line. Or, rather, given this need for savoring,
it makes sense that the recurring unit of poetry would develop in
accordance with the structure of rehearsal memory. Moreover, this
is true not only receptively, but productively. A poet composing a
poem is generating short, repeatable, nonsyntactic units. He/she has

to revise and “polish” these units – to satisfy metrical and other con-
straints and the broader criterion of nonobtrusive maximization of
patterning. Given the structure of human cognition, one would ex-
pect that any such unit would almost necessarily be structured in
accordance with rehearsal memory. This is even more obvious when
one takes into account the oral, bardic composition which is at the ori-
gin of poetry (on the nature of oral poetic composition, see Lord, Ong,
and Rubin). Without the aid of writing, the recurring unit of poetic
form would almost necessarily be structured by rehearsal memory.
In short, rehearsal memory seems to provide a good explanation
of the universal as initially stated. Indeed, it not only accounts for the
five to nine words, but relates this to the maximization of relevance,
and to other aspects of the experience and creation of poetry. Thus,
we have broad theoretical reasons not to adopt a formulation of the
line length universal that would dissociate it from rehearsal memory.
Of course, if lines are only shorter and not longer, then it is still possi-
ble to link the poetic line with rehearsal memory. Longer lines would
not fit into the limited space available for rehearsal memory; shorter
lines fit, even though there is room left over. We do often use rehearsal
memory in this way. On the other hand, the nature of any link with
rehearsal memory is less clear if we accept the reformulated universal
(“less than nine words”). In this case, it would seem that rehearsal
memory does not structure the line, but merely limits its extent. Is
there a separate, structuring principle, then? The majority of standard
lines appear to fall within the five- to nine-word range and thus to
be structured by rehearsal memory. How are we to explain this?
One option that preserves the straightforward relation to rehearsal
memory is to reconsider the notion of the line. Indeed, there seems
no necessary reason to identify the unit at issue with what is, at least
to some extent, a convention of writing and printing (a similar point
has been made by Frederick Turner [73–4]). In the great majority
of cases, the printed line is indeed the unit we want. In the case
of the alexandrine or iambic pentameter or a wide range of other
patterns, the printed line rightly defines the recurring nonsyntactic
unit. In other cases, however, this is less clear. One of the most obvious
exceptions to the universal as initially formulated is the Japanese
haiku, a seventeen-syllable poem divided into sections of five, seven,
and five syllables. We typically conceive of this as a three-line poem.

One way of reconciling the haiku with our universal and with the
structure of rehearsal memory would be to conceive of the poem as
a single line with a single caesura. This fits perfectly well, yielding
roughly six to nine words per unit (see, for example, the haiku in
Keene, Anthology, 361–9). An alternative account that also preserves
the structural function of rehearsal memory, might incorporate silent
beats, akin to musical rests. This would fit with certain aspects of
Japanese aesthetic theory. Moreover, the general phenomenon is well
attested. As Rubin points out, in certain English poems, “beats must
be allowed to fall on rests. Although our conventional writing system
ignores rests . . . their presence in verse can easily be heard by trying
to recite” certain poems “without including the appropriate rests”
(252–3n.3).
On the other hand, a further source of exceptions may be found
in such languages as Kuna and Dyirbal, which tend to have between
two and four words per line (see Sherzer 107–10 and the poems in
Dixon and Duwell), and where the problem does not appear to be
one of defining the recurring rhythmic unit. (There is also no obvious
motivation for positing silent beats in these cases.) These counter-
examples rather lead us to a reconsideration of the nature of rehearsal
memory. Both Kuna and Dyirbal are highly morphologically complex
languages. It may be that rehearsal memory is not appropriately mea-
sured in terms of words at all, but in terms of, say, morphemes. Even
in European languages, it seems odd to count “habeo” (“I have”) as
one word, but “j’ai” (“I have”) as two, or “et” (“and”) as one word,
but enclitic “-que” (“and”) as none. Perhaps “et virum” (“and man”)
takes up two slots in rehearsal memory and “virumque” (“and man”)
takes up only one, but this seems counterintuitive.
In any case, at this point, the research project of literary universals
abuts the broader research project of cognitive psychology. Before
the literary project can proceed, the broader psychological study of
rehearsal memory has to proceed further. Moreover, as this example
indicates, aspects of this broader study can be inspired by literary
questions, data, and hypotheses. The problems of Kuna and Dyirbal
point to further areas for cognitive research – specifically, the exam-
ination of whether rehearsal memory is best understood in terms of
words, morphemes, or something else. Should it turn out that it is
best understood in terms of words, then the Kuna and Dyirbal data
will be difficult to account for and we may have to weaken or modify
our hypothesis. Should it turn out that it is morphemes, this will solve
the Kuna and Dyirbal problems. On the other hand, it may render
Chinese problematic. Should it turn out to be something else, this too
will probably produce anomalies.
Indeed, some research indicates that the relevant parameter, or one
relevant parameter, is something we have not considered – subvocal-
ization time. Rehearsal memory, in this view, should not be measured
in words (five to nine), but in seconds (about two; see Gathercole 20).8
If correct, this solves the problem of highly morphologically complex
languages, since the subvocalization time for each word in such a
language would be much longer than for a word in a language such
as English. It may suggest that we not treat a Haiku as a single line,
for an entire Haiku would seem to take longer than two seconds for
rehearsal. By the same token, however, it makes our alternative ac-
count of Haiku more plausible, for it indicates how pauses or silent
beats could occur in a cycle of syllabic verse. On the other hand, the
Chinese data now become more difficult to explain (assuming that the
subvocalization time of Chinese lines is as short as it appears). Once
again, there is no contradiction as the Chinese lines do not exceed
the limit. But we are still left with the question of why the Chinese
standard line would apparently be so short.
8 Six years after writing the first version of this argument, I came upon Frederick
Turner’s essay, “The Neural Lyre.” In that essay, Turner sets out to explain the uni-
versality of line length in poetry. His approach is based on articulation time, which
he finds to be about 3 seconds in most traditions – though his tabulation does in-
clude instances ranging from 2.2 seconds, for the Chinese four-syllable line, up to
3.9 seconds, for the Latin Alcaic strophe (76). As just indicated, Turner may have been
right to emphasize time rather than words or some other variable. But his temporal
account would need to be reworked in a research program, and not only in order
to treat the anomalous data just cited. First, he almost certainly needs to consider
subvocalization time rather than vocalization time. Second, he probably should ad-
dress the issue of standard or, equivalently, prototypical line lengths. Again, my own
hypotheses have concerned standard line lengths. It seems to me very unlikely that
rules governing line length can be formulated in terms of necessary and sufficient
conditions (that is, in such a way as to cover all line lengths, rather than standard –
thus prototypical – line lengths). Finally, Turner offers “the three-second present mo-
ment of the auditory information-processing system” as the source of universal line
length and connects this with the effects of poetry, which in his view operates “to im-
prove the memory, and to promote physiological and social harmony” (103). This at
least requires further clarification and development, not to mention empirical study.
Another option is to view the structure of working memory as

more complex than any of these accounts allows. Perhaps the num-
ber of morphemes, or the extent of phonological information (for
example, the use of tone in Chinese) combines or interacts with sub-
vocalization time in the structure of working memory. Indeed, we
know that rehearsal memory is affected by a number of variables, not
just one. As Kintsch, Healy, Hegarty, Pennington, and Salthouse, put
it, “there is no single, all-encompassing capacity-limiting factor for
working memory” (420; see also Miyake and Shah, “Toward” 447).
For example, features of sound (especially rhyme) affect rehearsal
memory, as does the lexical status of items in the rehearsal string
(roughly, words are easier to recall than nonwords). (For a discussion
of this research, see Gathercole.)
A further possibility, consistent with the final point, is to set aside
working memory as a distinct part of cognitive architecture and ac-
count for the relevant data – such as the five- to nine-word capacity
for rehearsal – by reference to other structures and processes. For
example, in Barnard’s theory, there are acoustic, articulatory, mor-
phonolexical, and other subsystems that bear on the phenomena of
“working memory.” Thus, the “properties of performance on work-
ing memory tasks,” including the five- to nine-word pattern, “must
be attributed to process-mediated interactions among multiple sub-
systems of cognition” (313). An account such as this opens the possi-
bility that several of the preceding suggestions have bearing on the
determination of standard line length, with morphology, phonology,
subvocalization time, and other factors interacting in complex ways.
Clearly, there is a great deal of work here – not only in one research
program, but in several competing programs.
The maximization of unobtrusive patterning and the relation be-
tween rehearsal memory, line length, and aesthetic experience nicely
illustrate the process of descriptive and explanatory study in a theory
of literary universals.9 However, as such, they remain mere starting
points for research, hypotheses to be modified, elaborated, replaced.
9 The importance of explanatory adequacy is not confined to the study of absolute

universals. An excellent illustration in the study of typological universals is provided
by David Rubin. Rubin articulates a powerful cognitive analysis of oral composition
in terms of cues for and constraints on memory. The development of epithets and
formulae is directly predicted by his theory, as is the typological difference between
oral and written composition.
Again, the study of literary universals, like the study of linguistic uni-
versals, is a project that can progress only through the cooperative
efforts of a broad range of researchers engaged in an ongoing process
of empirical reevaluation of theories and theoretical reorientation of
empirical research. As the concluding discussion in particular illus-
trates, such a program of research could be greatly valuable, not only
for our understanding of those cognitions and affections that gener-
ate and sustain literary art, but for our broader understanding of the
human mind as well.
conclusion
In sum, there are many types of universals – absolute, statistical,
implicational, typological, and so on. There are also well-defined cri-
teria for determining universality – recurrence across a higher per-
centage of genetically and areally distinct traditions than would be
predicted by chance. By these criteria, there appear to be many liter-
ary universals. These may be roughly organized into techniques and
nontechnical correlations. Many literary universals may be partially
understood through the abstraction of secondary principles, such as
the maximization of unobtrusive patterning. Others may have a more
direct relation to a cognitive (or other) structure or process, such as
the capacity of working memory. In each case, the study of literary
universals seeks both descriptive and explanatory adequacy through
the development of a research program in cross-cultural, compara-
tive literary study and the (mutually beneficial) integration of this
program with research programs in cognition, as well as history and
other related fields.
2
Emotion and Suggestion

Lexical Processes in Literary Experience
The study of non-European literary theories is particularly valu-

able for the isolation of literary universals. Such theories of-
ten carry insights into universal patterns that are occluded by
European theories, thus patterns that readers may miss, if their
attention and understanding are too guided by European ap-
proaches. Indian theories are unusually significant for the present
study as they are particularly compatible with recent develop-
ments in cognitive psychology. Perhaps this is due to the fact that
Indian literary theory developed partially in relation to Indian lin-
guistic theory, which is widely considered to have close connec-
tions with Chomskyan developments (see, for example, Kiparsky’s
Pān.ini).
In any case, starting more than two millennia ago, and extending
over a millennium, Sanskrit writers developed an elaborate theory
of poetics. This theory reached a culmination and a sort of theoreti-
cal impasse in the writings of Ānandavardhana and Abhinavagupta.
With the aid of cognitive science, we may recognize the value of
Sanskrit poetics and develop it beyond the impasse it had reached in
the eleventh century. I hope to show that, when redeveloped in the
context of cognitive science, the work of these writers provides us
with a plausible and productive cognitive theory of poetic feeling, a
45
theory that is not a conclusion, but the start of a larger research pro-
gram in the field.1
More exactly, I shall begin with a broad outline of Sanskrit poetics
from its beginnings through Ānandavardhana (for a more detailed
discussion, the reader may wish to consult a history of Sanskrit poet-
ics, for example that of Gerow). Roughly speaking, Ānandavardhana
sought to develop and systematize previous ideas in Sanskrit the-
ory in order to provide an adequate description of poetic effects.
Abhinavagupta took Ānandavardhana’s descriptive ideas and
sought to provide an explanatory framework for them. The sec-
ond section is consequently devoted to the explanatory views of
Abhinavagupta. In the third section, I turn to contemporary cognitive
science, first presenting what I take to be a plausible, partial theory of
the internal lexicon (or mental dictionary/encyclopedia), then going
on to reformulate Abhinavagupta’s views in terms of this theory. The
fourth section responds to some possible misunderstandings of the
resulting account, by distinguishing among the objects, causes, and
sources of emotion. The chapter concludes with a brief interpretive
illustration from Hamlet.
ānandavardhana and suggestion

Histories of Sanskrit poetics generally divide early Indic theories of
verbal art into two broad traditions: alam . kāra and rasa. “Alam . kāra”
or “ornamentation” refers to a range of poetic devices and rhetorical
figures from alliteration to metaphor. In itself, the alam . kāra tradi-
tion is of limited interest, for it hardly extends beyond taxonomy (for
example, listing and describing the varieties and subvarieties of fig-
ures such as simile). On the other hand, without this often tedious
program of analysis and categorization, including its rudimentary lit-
erary semantics, Ānandavardhana could never have formulated his
seminal theory of dhvani or “suggestion,” to which we shall turn in a
moment.
Ānandavardhana’s work is, however, a culmination and synthe-
sis, not only of alam . kāra analysis, but of rasa theory as well. Rasa
1 Indeed, since the first version of this chapter was published, such a project has
made at least a tentative beginning – most significantly, in Keith Oatley’s work (see
“Emotions”).
Lexical Processes in Literary Experience 47
is usually translated as “sentiment.” It is distinguished from bhāva

or “emotion,” to which it is, nonetheless, closely related. Specifically,
bhāva is what we feel in ordinary life – love, sorrow, happiness, anger,
and so on. Rasa, in contrast, is what we feel in experiencing a work
of art. It is akin to emotion, but not identical with it. (In the rest of
this chapter, I shall use the English word “emotion” in place of the
Sanskrit “bhāva” as the two are close enough in meaning for present
purposes; I shall, however, most often use “rasa” rather than “senti-
ment” as “rasa” is a technical term with no precise ordinary language
equivalent in English.) Specifically, when I watch a romantic play, I
do not actually love the hero or heroine (as I love my wife), but I
do experience some sort of feeling. Moreover, this feeling is related
to love in a way that it is not related to sorrow or anger. Thus the
Sanskrit theorists say that I am experiencing the “erotic rasa,” not the
emotion of love per se.
The earliest extant development of this theory is in the Nāt.yaśāstra
or Treatise on Dramaturgy attributed to Bharatamuni, but composed
by a number of authors between the second century b.c.e. and the
sixth century c.e. This foundational volume lists eight primary emo-
tions and corresponding rasas: love/the erotic, sorrow/the pathetic,
and so on. (Lest this seem too restrictive, they acknowledge a wide
range of ancillary feelings also.) Each literary work, in order to be
aesthetically effective, was required to have one dominant rasa. This
is not to say that other rasas could not enter; they not only could,
but must. However, any additional rasas should function to further
the dominant rasa. For example, suppose that the dominant rasa of a
work is the erotic. Then it makes perfect sense to bring in the pathetic
rasa. The pathetic may be part of the erotic, just as sorrow may be part
of love. (Indeed, another way of thinking about the relation between
rasa and bhāva is that the characters experience the bhāvas, such as
love and sorrow, while the readers/spectators experience the rasas,
such as the erotic and the pathetic.) But the pathetic rasa in the work
must operate to contribute to the erotic rasa. The same constraint
holds if the dominant rasa is the pathetic and the erotic is subsidiary.
If the subsidiary rasa does not contribute to the dominant rasa, the
experience of the reader/spectator will push in different directions
and the overall aesthetic experience will be weakened; it will not be
a satisfactory experience of the erotic or of the pathetic.
Again, Ānandavardhana combined the two strains of Sanskrit po-

etics, developing them into a unified theory of aesthetic response,
in part based on the concept of dhvani or “suggestion.” By the time
of Ānandavardhana, Sanskrit theorists of “figures of thought” had
isolated several varieties of meaning. At this level, the distinctions
with which they were operating were roughly the same as those op-
erative in western poetics/semantics today. They had concepts of
literal meaning, idiomatic meaning, and various metaphorical and
related meanings (encompassing simile and other figures); they had
also isolated a number of ways in which nonliteral meanings could be
manifested, recognized, and interpreted (for example, through some
explicit marker, such as English “like” and “as” in similes). We would
consider much of this fairly straightforward today. “Vārān.ası̄ is on
the Ganges” does not literally mean that Vārān.ası̄ is on the Ganges;
it means (idiomatically) that Vārān.ası̄ is on the bank of the Ganges.
“Moon-faced beauty” does not literally mean that the woman’s face
is the moon, but that it is rounded and fair, and so on. More interest-
ingly, Sanskrit theorists had isolated a variety of complex poetic and
other implications – or, rather, nonlogical “implicatures,” to use Paul
Grice’s term – both literal and nonliteral. For example, a message sent
by a woman to her lover, “The lion, they tell me, does not prowl at
the riverbank,” would involve a complex metaphor/implicature to
the effect that, if the woman and her lover meet at the riverbank, they
will not be caught.
Ānandavardhana systematized this work by distinguishing va-
rieties of dhvani that encompass these nonliteral meanings. Most
importantly for our concerns, however, he maintained that there
was one sort of meaning that is not part of this typology and is
most appropriately the referent of the term “dhvani.” This “dhvani
proper” is not an idiom or metaphor (or simile, metonymy, and so
on); nor is it an implicature (cf. Amaladass 92–3; unless otherwise
noted, I shall use “dhvani” to refer to this “dhvani proper”). Ex-
tending Ānandavardhana somewhat, we may say that it is, rather,
a nonparaphrasable suggestion of a word, phrase, sentence, topic,
or (linguistically constructed) situation (see Amaladass 105). To say
that the dhvani of a given utterance is nonparaphraseable is not to
say that one cannot say anything about it (as Ānandavardhana em-
phasizes [671]). Quite the contrary. One can say many things about
it. It is not unparaphraseable because of being ineffable. There are,

rather, three reasons why dhvani is unparaphraseable. First, it is in-
finitely ramified. We can never enumerate all the suggestions, even
all the relevant suggestions, of a given work (cf. Abhinavagupta,
Locana 206). This is not, however, a distinctive property of dhvani.
For example, some metaphors are, roughly, paraphraseable in the
sense that one can list all the relevant information conveyed by the
metaphor (especially those metaphors that are close to idioms). But
many metaphors are not fully restatable in this sense. And at least
for larger texts, even the relations among literal meanings cannot be
ennumerated exhaustively.
The second reason that dhvani is not paraphraseable is more im-
portant and more distinctive. Indeed, it indicates that, in a sense,
dhvani is not even partially paraphraseable. In the language of an-
alytic philosophy, the dhvani of a text cannot be substituted for the
text with a preservation of truth value (cf. Abhinavagupta, Locana 81).
Indeed, dhvani is usually not statable in the form of a proposition
that might have or not have a truth value. Consider a metaphor
partially explicit in Derek Walcott’s highly mythic play about Afro-
Caribbean identity, Dream on Monkey Mountain: “Makak is a lion.”
As a metaphor, this means that Makak is brave, ferocious to his foes,
and so on. (For the purposes of the example, we can assume that it
is not fully paraphraseable and thus that this list will not come to an
end.) But note that each of these interpretations is substitutable for
“a lion” with a preservation of truth-value. “Makak is brave” and
“Makak is ferocious to his foes” are just as true as “Makak is a lion.”
(I leave aside the issue of how to specify “metaphorical truth.” Clearly,
at least at a certain point in the play, this sentence is [ambiguously]
true in a way that, say, “Makak is a gazelle” or “Makak is a uni-
corn” is not.) In contrast, consider the standard example of “a village
on the Ganges.” This phrase, the Sanskrit commentators emphasize,
suggests holiness due to the holiness of the sacred river Ganges. How-
ever, there is no way in which the word “holiness” can be substituted
for the phrase or for part of the phrase. Take, for example, the begin-
ning of a story, “Rām was the headman of a village on the Ganges.”
The truth value of this statement is not preserved with the relevant
substitution: “Rām was headman of holiness.” Moreover, there is
no way that this suggestion can be turned into a proposition. For
example, the suggestion of holiness does not imply or implicate that

the village itself is holy or that any of the villagers (for instance, Rām)
is holy. Returning to Walcott, we might say that the name “Makak”
(or “macaque”) suggests a whole range of west African mythological
motifs concerning the divinity of the monkey, and at the same time
a range of contrary European attitudes in which “monkey” serves as
a racial insult. These are not substitutable for “Makak.” Nor do they
imply or implicate any propositions.
Even at this level, the notion of dhvani seems to be substantially
different from standard concepts of literary semantics in the western
tradition. For example, though “connotations” are not viewed as truth
preserving, they are typically conceived of as involving some sort
of assertion. Western theorists may well say that “a village on the
Ganges” connotes “holiness,” but I suspect that they would unreflec-
tively think of this as part of a (perhaps ironic) assertion – that the
village is or should be holy, that the villagers have a special duty to
be holy, and so on. For this reason, it would seem somewhat odd
to say that various Yoruba beliefs are a “connotation” of the name
“Makak,” for, again, none of these beliefs is asserted (or denied) by
this “connotation.”
But this is not all that distinguishes dhvani from common western
notions – and this brings us to the final and most important reason
why dhvani is not paraphraseable. Dhvani is not purely semantic. It is
affective as well. Specifically, in the strictest sense, Ānandavardhana
tells us, dhvani is rasadhvani, the dhvani of rasa – not the intel-
lectual implication of some sentiment, but the “suggestion” of a
rasa as an affective experience (see Abhinavagupta’s comments on
Ānandavardhana in Locana 70). In other words, dhvani is not para-
phraseable, most importantly because it is not solely, nor perhaps
even primarily, a meaning; as rasadhvani, the “truest” form of dhvani
is bound up with feeling. It is, in short, an experience of rasa. A liter-
ary portrait of a village on the Ganges involves a full rasadhvani of
holiness – or, more aptly, sacred peace – only if it gives us a feeling
of that holiness or sacred peace along with the more narrowly se-
mantic suggestion. Indeed, the feeling is what is crucial. The name
“Makak” communicates not only the idea of west African beliefs and
customs carried by slaves to the new world, but feeling as well –
perhaps sadness (the pathetic rasa) over the partial loss of these
beliefs and customs, or anger (the furious rasa) over their colonial
denigration.
In sum, aesthetic response is a matter of the experience of rasas.
These rasas are evoked in a reader by words, sentences, topics, and so
on, presented in a literary work. This is, of course, in part the result of
literal meanings. But it is also, and crucially, a function of the clouds
of nondenumerable, nonsubstitutable, nonpropositional suggestions
that surround these texts – what Keith Oatley has recently called the
“Suggestion Structure” of the work (“Emotions” 45, 51–9). Finally,
it is important to add that, in the overwhelming majority of cases,
these rasas are evoked without our having any explicit awareness of
suggested meaning. We do not, in other words, self-consciously infer
some semantic suggestion, then feel the rasa. Rather, we experience
the rasa as we watch the play or read the poem. Indeed, if we have to
stop and work through the suggestions, we may not experience the
rasa at all.
abhinavagupta, memory traces, and

aesthetical feeling
Abhinavagupta turned his attention away from the linguistic and
related abstractions that had preoccupied even Ānandavardhana,
focusing his attention instead on the human mind, specifically the
mind of the reader or viewer of a literary work. The first step in
Abhinavagupta’s project involved the, at least tacit, recognition that
the theory of rasadhvani could not be understood as a theory of ab-
stract linguistic structure. Rather, it only made sense as a theory of
the way people respond to literature. In other words, rasadhvani had
to be conceived in psychological terms.
Broadly speaking, Abhinavagupta was a transcendental realist
whose philosophy involved a strong empiricist component. (A Hindu
theologian himself, he was one of the harshest opponents of the
Vedāntist view that perceptual reality is illusion or māyā; see, for
example, B.N. Pandit 23.) His theory of mind was, consequently, a
realist theory. Unsurprisingly, it was broadly similar to other (eastern
and western) theories of mind in isolating perception, memory,
and other components or faculties. For our purposes, the most im-
portant part of Abhinavagupta’s theory of mind is his theory of
memory, both storage and recollection. For Abhinavagupta, all ex-

periences – perceptual, verbal, emotional – leave “traces” in the
mind (see Aesthetic Experience 79). Drawing out the implications of
Abhinavagupta’s analysis, we may understand these traces as hav-
ing two components: one representational, one emotive. The emotive
part is not the abstract recollection of one’s having had an emotion –
such a recollection would be representational. It is, rather, a sort of
reexperiencing of the emotion. It is not, in other words, remember-
ing that one was sad, happy, or frightened at a given time and place,
but actually feeling again, in some degree, that sadness, happiness,
or fright. The point is most obvious with respect to strong emotions.
For example, if one recalls a deceased friend or relative, one will have
certain visual and other impressions, one will remember certain facts,
and one will probably experience again, in a more or less attenuated
form, the sense of loss, the sorrow that one felt at his/her death.
These traces, Abhinavagupta tells us, are usually latent in our
minds. At times they are fully activated – that is, at times we recall
these memories and reexperience the emotions. This is all common-
sensical enough. But, Abhinavagupta continues, there are also times
when these traces are neither latent nor fully activated – and these
are the most crucial. In other words, there are times when these traces
are in some sense activated, but are not brought into self-reflective
consciousness. More exactly, there are times when we are not self-
reflectively aware of the representational content of the trace, and
yet feel some hint of associated affect. This is, I think, common in
our experience, though we may not immediately recognize the fact.
I enter a building and am suddenly sad; I ask myself why and then
recall an embarrassing or unpleasant event that occurred the last time
I was in the building. Clearly, the memory had in some sense been
activated (that is, it was not fully latent) and clearly the associated af-
fect had bled into my conscious experience (that is, I felt sad), but the
representational content of the memory was initially not conscious.
This sort of analysis is what allows Abhinavagupta to explain the
phenomena isolated and described by Ānandavardhana. Specifically,
Abhinavagupta indicates that rasadhvani operates in the following
manner. Through dhvani, the literary work activates memory traces
in the mind of the reader (see Abhinavagupta, Locana 81) without
bringing these memories into consciousness. Again, these traces may
be activated by words, phrases, topics, events, or whatever. Thus sto-

ries of suffering will activate memories of suffering, stories of roman-
tic love will activate memories of romantic love, and so on. Once these
traces are activated, the associated emotions seep into consciousness –
again, not as ideas, but directly as feelings. The experience of the rasa
of a literary work is precisely the experience of these feelings. More ex-
actly, developing Abhinavagupta’s ideas, we may say that all speech
involves the activation of memory traces with their associated affect.
However, most often, these traces are activated in a haphazard and
noncumulative manner, or else they are fully recalled. What makes
aesthetic experience distinctive is that such activations are not fully
recalled, but they are patterned; they are focused on traces of a spe-
cific type, which is to say, traces that bear to one another a certain
similarity in both representation and affect (for example, in being
memories of death and feelings of sorrow). While any given activa-
tion is likely to produce only very limited, indeed imperceptible and
fleeting experiences of affect, this sort of repeated, patterned activa-
tion should result in a more pronounced and continuous experience.
As Abhinavagupta puts it, “the basic emotion is put to use in the
process of relishing [a work of art]: through a succession of memory-
elements it adds together a thought-trend which one has already ex-
perienced in one’s own life” (Locana 117; see also 182 – “the relishing
of beauty arises in us from our memory bank of mental states which
are suitable” to the “basic emotions” of the characters – and Aesthetic
Experience 112).
Abhinavagupta extends this idea somewhat further when he ar-
gues that aesthetical pleasure results from the “generalization” of
emotion in rasa, which is to say, its removal from the self-interest that
is part of the link between the affect and the representational content
in memory traces. In other words, when we fully remember a trace,
the emotion that we experience is tied to self-interest (for example
to our own personal loss of the loved one whose death we are re-
calling). However, through literature, we experience a version of the
affect removed from its direct link with any particular (egocentric)
representation in memory, and thus at least partially removed from
self-interest (see Abhinavagupta, Aesthetic Experience 86–7 and 96–7).
In this way, rasa may be redefined as emotion isolated from such
self-interest and may even be compared, in Abhinavagupta’s view,
with the experience of religious enlightenment or moks.a, where such

self-interest is entirely extinguished (see, for example, Locana 226).
(Indeed, for Abhinavagupta, the dominant rasa of a successful work
always resolves itself into śāntarasa, the rasa of peace, a temporary
and partial version of the endless and perfect peace that accompanies
moks.a.) Another way of putting the general point is to say that the
feeling of rasa is empathic.
To some extent, I have explicated and developed Abhinavagupta’s
ideas in a manner coherent with and partially derivative of contempo-
rary cognitive science. However, I have not, I believe, fundamentally
altered his claims or basic concepts. Indeed, it is my view that current
theoretical and empirical work allows us to understand those claims
and concepts more thoroughly than could Abhinavagupta himself,
because he lacked a psychological theory adequate to his literary in-
sights. However, I have refrained thus far from giving a full blown,
reductive account of Abhinavagupta’s ideas within a broader theory
of cognition. In the following section, I shall attempt something along
these lines: a translation of Abhinavagupta’s theoretical premises
into one version of a theory of mind that I take to be plausible in
light of recent research in cognitive science. As I have already indi-
cated, cognitive science has little to add to Ānandavardhana’s and
Abhinavagupta’s ideas about literary response. My concern, then, is
to situate these ideas in a current theoretical framework that is part
of ongoing research programs.
cognition and lexical structures

Due to the nature of Abhinavagupta’s hypotheses, I shall be concen-
trating on the storage and accessing or recollection of representational
knowledge (that is knowledge about something, including memories,
and so on, as opposed to procedural knowledge, knowledge of how
to do something). I shall assume that all representational informa-
tion is stored in and accessible through a single long-term memory
“unit.” This unit includes a broad range of information which, in
everyday life, we would be inclined to divide between a dictionary
(meanings of words), an encyclopedia (general facts about things),
and a biography or personal archive (propositional and perceptual
memories of individual history). This is a controversial assumption.
However, the following analyses could be rewritten – with some loss

in elegance – in terms of two or three separate, but closely interre-
lated units. Thus, nothing in this analysis is contingent on there being
just one unit. In any case, following one standard usage, I shall re-
fer to this unit as the “lexicon.” I should emphasize that this term
should not be taken to imply a sort of unidirectional organization
and access: from dictionary/meaning through encyclopedia/belief
to personal archive/memory. In fact, lexical “entries” (for example,
that for “monkey”) are structured to allow access not only from words
to memories, but from perceptions (for example, seeing or hearing a
monkey) to words, from memories to beliefs, and so on. Indeed, just
as there are words for which we have no beliefs or memories, and
some for which we do not have even basic meanings (“I’ve heard
that word before; what does it mean?”), there are also uncategorized
perceptions and perceptual memories (“I’ve seen that before; what is
it?”). Thus the mental lexicon is no more a “lexicon,” in the common
sense of “dictionary,” than it is a “mnemonicon” or an “aistheticon”
(or an encyclopedia). Fortunately, however, these complications do
not directly affect our concerns, which are with texts, and thus
roughly fit the commonsense model of organization from “meanings”
to beliefs and memories. Were we to extend this study to aesthetic
response in, say, the visual arts, however, we would have to be more
careful not to allow the prejudicial nature of the term to mislead us.
In the present context, then, we may think of the lexicon as
structured into “lexical entries,” which is to say, meanings, percep-
tions, and so on, clustering around lexical items, such as “monkey,”
“dhvani,” “chant,” “compose,” “saffron,” and so on. These entries
are multiply cross-indexed such that each entry is part of a num-
ber of networks that allow access across entries. Thus, “monkey”
is linked with “ape” and “chimpanzee” in one network (of related
species), “Africa” and “India” in another network (of habitats), and
so forth. These networks may themselves be categorized by some
lexical item (for example, “primate”), but they need not be – though,
of course, one can always construct a category for a given network
(for example, “habitats of monkeys”) and, indeed, we often use such
ad hoc categories in actual cognition.
As this indicates, our mental lexicon – a structure that exists only
in individual people’s minds and thus varies somewhat from person
to person – is different from a dictionary or encyclopedia, not only

in content, but in structure. Perhaps most importantly, it does not
have a single strict organizing principle (such as alphabetical order).
Rather, our mental lexicon can be reordered in a variety of ways to
suit our accessing needs. For example, we may access words by first
consonant (as in a dictionary, more or less) or by some other phonetic
property, such as final syllable – if, say, we are writing an alliterative,
rhymed poem. In most contexts, however, we are likely to access items
by topic. One way of thinking about this is in terms of the networks
just mentioned. When discussing animals, we reorder our lexicon to
make, for example, “monkey” more directly or swiftly accessible than
when discussing computer technology (see Garman 293). Moreover,
this effect is cumulative. If we are discussing Africa and animals,
then “monkey” will be more accessible than if we are discussing
animals alone. This “extra” accessibility of related terms, concepts,
and so on, is standardly interpreted in terms of “priming.” In this
view, the introduction of one item “primes” cross-indexed items (for
example, “monkey” will prime the entries for “ape,” “primate,” and
so on). At this level, priming can be understood as a reordering of the
lexicon, such that the primed items are those that are placed highest
in the order of a lexical search. For example, when we read or hear
the syllable “mon” at the beginning of a word, we begin searching
the lexicon for a “fit.” When the topic is Wall Street, we will reach
“money” first; when the topic is animals related to apes, we will reach
“monkey” first. (In fact, the situation is somewhat more complicated
than this, but the point is adequately valid for present purposes.)
On the other hand, priming has complex effects not only on access,
but on other aspects of comprehension and response. In this way, it
is not simply a matter of reordering the lexicon. There seems to be a
change of status in the primed network. Primed items are, in effect,
brought out of long-term memory, though they are not accessed di-
rectly in consciousness. When speaking of primates, I will have part
of one lexical entry in consciousness, and parts of several lexical en-
tries in rehearsal memory, which may be understood as circulating
through consciousness. I also have a whole range of material stored
in long-term memory (for example, the meaning of “ambidextrous,”
visual images of my first day at school, and a vast range of other
currently irrelevant stuff). A network of primed lexical entries is in
a different mental state from either the conscious/rehearsal material

or the material stored in long-term memory. Consider the following
sentence: “Animals commonly associated with Africa would include
lions, tigers, elephants, and apes.” When I am directly conscious of
the word “ape,” the rest of the sentence is in rehearsal memory, and
various irrelevant entries (for example, “Plōtı̄nus”) are in long-term
memory. The lexical entry for “monkey,” however, is primed and
is therefore in a state different from “apes” and “lions,” on the one
hand, and “ Plōtı̄nus,” on the other. The primed entries are, we might
say, placed temporarily in a sort of buffer between long-term mem-
ory and consciousness (or between long-term memory and rehearsal
memory), a buffer that operates in part to allow access with minimal
search.
This gives us a basis for rearticulating and reunderstanding
Abhinavagupta’s notion that traces may be activated while not con-
sciously recalled. Specifically, we may conceive of dhvani as lexical
networks primed and stored temporarily in the memory buffer. Note
that these networks may “decay” rapidly, which is to say, drop quickly
out of the buffer if they are not repeatedly primed. (The rapid decay
of priming effects is well documented; see, for example, Garman 294.)
However, when repeatedly primed, they would yield a pattern of un-
stated suggestions of precisely the sort we discussed previously as
defining dhvani.
To get a more detailed idea of how rasadhvani might work, how-
ever, we need to consider not only the relations between lexical en-
tries, but their internal structures as well. Here, as elsewhere, cogni-
tive scientists are not unanimous in theories or even terminology. I
will use the terms “schema,” “prototype,” and “exemplum” to define
three types of substructure within a lexical item. By “schema,” I mean
a hierarchy of principles defining a lexical item. The hierarchy is based
on “definitiveness” or “centrality” of the properties; the most central
or definitive properties are at the top of the hierarchy, with increas-
ingly peripheral properties listed in descending order. For example,
being organic is more central to our conception of a human than is
having two arms – a person with one arm would count as a human,
but a statue with two arms would not. (This roughly recapitulates the
distinction between the dictionary meaning and encyclopedic factual
beliefs, but the distinction is one of degree, not of kind; some elements
are “more definitive” than others, rather than one set of elements be-
ing “the definition” and another being “the empirical beliefs.”) As
virtually all cognitive scientists emphasize, this hierarchy involves a
number of “default” options. These are properties or relations that
we assume unless we are told otherwise. Thus, “human” includes
“having two arms” unless we are given information about a birth
defect or amputation.
By “prototype,” I mean, first of all, a sort of concretization of the
schema with all default values in place, including those that are rel-
atively unimportant in our schematic hierarchy (cf. Johnson-Laird
and Wason 342). Again, these vary somewhat from person to per-
son, but probably all of us have a prototype of, say, “man” as having
two arms, two legs, and so forth. In addition, the prototype will have
some “average” properties as well, properties that we would not or-
dinarily think of as “default assumptions.” For example, one person’s
prototype human might be brown haired (that is, not bald and not
grey, blonde, or black haired), clean shaven, of medium height, and
so on. This is probably not a simple averaging, but a “weighted”
averaging, with certain instances counting more than others – in
part due to salience, but due to other factors as well. An impor-
tant part of weighting is contrast (cf. Tversky, Ortony “Beyond,” and
Barsalou 212). When forming our prototype of men, we tacitly weight
more heavily those individual men who contrast most strongly with
women. Thus our prototypical man will probably be more “manly”
than the average man. Likewise our prototypical case of sadness will
be sadder than the average. Related to this, certain prototypes have a
normative element, as several authors have noted (see, for example,
Kahneman and Miller 143). This is probably a function of the degree
to which the term itself is considered normative. For example, our
prototypical man or woman is probably better looking than the av-
erage man or woman and our prototypical surgeon is probably more
skillful than the average surgeon.2
2 For my purposes, it is not terribly important to determine just how prototypes

arise – what particular cognitive architecture they involve and just what cognitive
processes and structures constitute them. I tend to speak in a “representational”
idiom. Someone else might wish to adopt a purely connectionist idiom (for a con-
nectionist treatment of prototypes, see McLeod, Plunkett, and Rolls 84–8, summariz-
ing work by McClelland and Rumelhart; to be made consistent with the preceding
Finally, by “exemplum,” I mean any specific instance of a category.

Thus any man I know is an “exemplum” of man. The more common
term here is “exemplar.” However, use of this term has been confus-
ing. Hampton points out that “Exemplar models are . . . frequently
account, this particular connectionist network would have to be revised slightly to

allow some properties to weigh more heavily than others). Barsalou presents an inter-
mediate account that involves representationalist categories in a spreading activation
network. Specifically, Barsalou argues that prototypes are well-established networks
of connection among concepts, instances, and properties (212–13; this account too
may require some tinkering to incorporate ideals). For my purposes, choosing a par-
ticular account is not important at this point.
On the other hand, there are properties that such an account should have. Specifi-
cally, George Lakoff has stressed that experimental research gives us only “prototype
effects,” not prototypes per se. The point is, of course, just a generalization of all sci-
entific inference. Studies of falling objects, the tides, and planetary motion do not
give us some direct experience of gravity. Rather they give us data that we try to
explain. We might refer to these data as “gravitational effects.” In addition to pre-
viously established data or prototype effects, the present study adds cross-cultural
constancy in emotion concepts and in narrative genres. Thus any treatment of proto-
types – or any alternative explanations – will have to treat these “prototype effects”
as well. In the appendix, I suggest some ways in which these particular constancies
may be explained.
In addition, my description of prototypes indicates that they cannot be gener-
ated in any simple way (for example, by some sort of straightforward averaging).
Moreover, they are not fixed or univocal. For example, in the next chapter, I shall
discuss how they vary with context. I have argued elsewhere that they are affected
by a range of variables, including, for example, one’s mood – even to such an extent
that normative elements may shift radically, from positive to negative or the reverse
(see Culture 126–31; for obvious reasons, the point is particularly consequential with
ethnic, racial, and related prototypes). The purposes for which we invoke a proto-
type have bearing here as well. Barsalou argues that ad hoc categories are generated
due to purposes. We have aims (for example, to raise money in a garage sale) and we
generate categories (for example, “items that are appropriate to sell in a garage sale”)
in order to achieve those aims. I would argue that even our well-entrenched proto-
types are sensitive to purposes. That sensitivity, too, allows for contextual variation,
leading to contrast effects, weighting, and the incorporation of ideals depending on
the degree to which a given term is normative. Whatever the cognitive architecture,
an alogrithmic level account of prototypes should capture this complexity.
One account of “prototype effects” that does not cohere well with the present
discussion is that of Lakoff. With work, the two could probably be reconciled. How-
ever, I find Lakoff’s account implausible for other reasons. Lakoff sets out to explain
prototype effects by reference to a broad range of structures. Thus, despite his re-
peated return to the notion of prototypes, he in effect disputes the category. But the
structures he isolates do not seem to be of the same logical type, nor do they seem
clearly distinct in the way that theoretical categories need to be distinct. He begins
by isolating “radial structures” that have a central case and a gradient of less cen-
tral/more peripheral cases. He then isolates a series of “metonymic models” in which
“a subcategory . . . is used to stand for the entire category” (84). First, it is not clear
vague about just what an exemplar is. There is an ambiguity between

types and tokens” (98). Moreover, the term “exemplar” suggests a
“good” or “definitive” case of a category. In part, to avoid this final
problem, and the associations that have attached to “exemplar”
that this differentiates metonymic models from radial structures where the central
cases would seem to have just this function. In any case, metonymic models in-
clude “social stereotypes,” “typical examples,” “ideals,” “paragons,” “generators,”
“submodels,” and “salient examples.” First, “paragons,” and “salient examples” are
instances. (Despite the name, “typical examples” are not.) Thus, they do not fall un-
der the general prototype category to begin with. (For discussion of these, see the
following note.)
It is very difficult to see how social stereotypes differ from typical examples, ex-
cept that the term “social stereotypes” implies that the examples are both mistaken
and socially harmfull. It is certainly important to distinguish true from false proto-
types. And it is important to discuss the social consequences of prototypes. However,
it is not clear that a “typical example” picked up from seeing many birds, reading
about birds, hearing people talk about birds, and so on, is cognitively different from
a “social stereotype” picked up from seeing many representations of some ethnic
minority on television, hearing people talk about that minority, and so on, especially
when combined with contrast effects, the biasing effect of norms, and so forth. As to
ideals, these appear to be a function of the degree to which the relevant category is
normative. “Submodels” are a strange category that seems to include some typical
examples, some ideals, and some other things that do not involve prototype effects
at all. Lakoff’s example is body temperature. But this is an actual norm, and it in-
volves a continuous gradient of numbers. Not everything that involves a norm and
a gradient is a prototype effect. The same point seems to hold for generators.
I may simply have missed the organizing principle here. However, it seems that
these categories are partially overlapping and partially irrelevant to prototype ef-
fects. Lakoff is right to point to the varieties of prototype effect. However, these seem
best treated through an account that recognizes the complexities and functions of
prototype generation, as discussed previously.
Finally, I should reemphasize that I see the lexicon as having a complex structure
that is not reducible to prototypes (understood as distinct and stable representa-
tions, connectionist circuits, or whatever). Thus, the present account does not suffer
from the problems associated with purely prototype-based accounts of concepts. For
example, Fodor has argued that such accounts run aground on semantic composi-
tionality. Specifically, they have no way of explaining such simple concepts as “pet
fish” (see Fodor’s Chapter 5). This is not a problem for an account that includes
schemas and exempla along with prototypes, and that recognizes schematic com-
plexity involving defaults, alternatives, and so on. (A full account of a lexical entry
would include ideals, common beliefs, and other elements.) Indeed, compositional-
ity is itself highly complex, and involves its own prototype effects. It seems likely,
therefore, that a complex understanding of lexical entries, including prototypes, pro-
vides the best way of accounting for compositionality. For example, Fodor is right
that a goldfish is not a good case of either pet or fish, but it is the best case of a pet fish.
Of course, this means that “pet fish” has an associated prototype. But that prototype
is also related to the prototype for “pet” in that a goldfish is probably the best case
through various theories, I have adopted the term “exemplum”

(plural, “exempla”) for simple instances. (“Exemplar” may then
be reserved for “good” instances, instances conforming to the
prototype.)3
It is important to stress that all three substructures operate in cog-
nition.4 We do not merely interpret or respond to the world and to
other people in terms of abstract schemas. Indeed, we are more likely
to understand and respond by reference to prototypes or particularly
salient exampla (on the last, see, for example, Chapter 3 of Ross and
Nisbett; on the use of instances and larger, schematic abstractions,
see, for example, Klein, Loftus, Trafton, and Fuhrman). The cogni-
tive role of exempla is particularly important to literary response for
a number of reasons. Most significantly, it appears to be the case
that exempla – which are regularly primed and accessed along with
of a pet from among the set of fish. In accounting for compositionality in this case
we need to account not only for the fact that a pet trout would still count as a pet
fish. We also need to account for the fact that the prototype for a pet fish is related to
the prototype for pet and to the set of fish concepts, but not to the prototype for fish.
Here we return to the issue of prototype effects. In this case, such effects suggest the
value of a multilevel account of lexical entries, an account in which there is at least
some sort of hierarchical structure corresponding to a distinction among exempla,
prototypes, and schemas (again, however these are ultimately specified in terms of
cognitive architecture).
3 In this terminology, Lakoff’s “paragons” are exemplars of highly normative cate-
gories. His “metonymic model” of “salient examples” combines instances that con-
tribute disproportionately to the formation of prototypes (that is, highly weighted
exempla) with instances that are highly nonexemplary (that is, exempla that contra-
dict the prototype) and override the prototype in our thought. Thus when people
refuse to fly on a DC-10 after a crash (Lakoff’s example [89]), they are allowing
the specific case to substitute for their broader prototype of “airplane.” Here too it
seems that Lakoff’s categories are not well formulated. As with prototypes, how-
ever, Lakoff’s analysis does indicate the complexity of exempla and exemplars. This
complexity should be explicable by reference to weighting, contrast effects, the in-
corporation of ideals depending on the degree to which terms are normative, utility,
and perhaps other factors.
4 There is some disagreement as to whether it is necessary to have three lexically
distinct types of structure in order to account for the three distinct levels of structure.
Thus some theorists try to account for the prototype level by reference to schemas or
exempla (see Shanks). Of course, to be descriptively adequate, any such reduction
would have to incorporate all the complexities outlined previously. I am skeptical
of attempts at genuine eliminative reduction. Even if it turns out that prototypes
may be understood as relatively stable networks of exempla, then those networks
themselves still form the intermediate of three levels. In other words, they are the
prototypes, and they are not identical with any individual exemplum.
schemas and prototypes – often have considerable affective force.

Clearly, many such exempla (for example, the memory of a close rel-
ative’s death) cause one to feel strong emotion when one recollects
them consciously.
cognition, memory, and rasadhvani

This, then, gives us a way of translating and redeveloping
Abhinavagupta’s notion that the affective component of memory
traces is the source of rasa. Specifically, the dhvani of a text may now
be understood as the schemas, prototypes, and exempla primed or
placed in a buffer between long-term memory and rehearsal memory.
The exempla include not only representational content, but affective
force. As Schacter puts it, “we do not store judgment-free snapshots
of our past experiences but rather hold on to the meaning, sense, and
emotions these experiences provided us” (5). When an exemplum is
sustained in the buffer, its affective force should lead to precisely the
sorts of effect hypothesized by Abhinavagupta when he explained
rasa in terms of memory traces. Specifically, we have every reason
to expect that the affective force of an exemplum would bleed into
consciousness without our being aware of its associated representa-
tional content, which is to say, the perceptual or propositional aspect
of the exemplum. Or, rather, we have every reason to expect this
when a set of affectively and representationally related exempla (for
example, sorrowful exempla of love in separation) are maintained in
the buffer through repeated priming due to the patterned dhvani of
a text. (It is important that the exempla are related both affectively
and representationally because the representation, so to speak, gives
definition to the affect; for example, sorrow over a relative’s death is
not precisely the same as sorrow over personal failure.) Schacter puts
the general point simply, “someone might experience emotions and
affects that result from incidents that are not recollected explicitly,”
such that “implicit effects of past experiences . . . shape our emotional
reactions” to new situations (232–3), presumably including literary
situations.
The repeated priming to which we have referred a number of times
is the consequence of two factors. First, it is one direct result of the
maximization of relevance or patterning. We not only assume that
a literary work is maximally relevant, authors make works that are

maximally relevant. In doing this, they shape not only sound, but
suggestion and feeling as well. In other words, one thing an author
does is work through the precise phrasings, the exact formulation
of events, in order to keep the rasadhvani consistent, thus in order
to prime semantically and emotionally relevant memories. Note, by
the way, that an author does not need to have any idea whatsoever
that he/she is doing this. All that is necessary is that the author read
his/her own work, evaluating its emotive impact, and revising in
order to produce the desired impact. Being a person like his/her read-
ers, the author will experience some comparable priming of memo-
ries, and so on, whether he/she knows it or not.
But this rasadhvani effect is not solely the result of authorial efforts.
It is also the result of readers’ cognitive propensities. To understand
how this occurs, we might consider the discussions of social judgment
by Bower, Forgas, and others that present a virtual mirror image of
the analysis we have just set out. Specifically, rather than primed
memories leading to emotion, they discuss ways in which emotion
serves to prime memories – and thus to sustain the initial emotion.
The general point is well known. For example, Oatley explains that
“There is now substantial empirical evidence to indicate that when
happy, happy memories come to mind, and when sad, sad memories
come to mind” (Best Laid Schemes 201). But the “algorithmic level”
accounts by Bower and Forgas – the detailed, explicit, step-by-step
accounts of cognitive processing – fit the present analysis much
more specifically (on the different levels of cognitive scientific anal-
ysis, see Dawson 9; on the algorithmic level in particular, see his
Chapter 5). Bower argues that the “Activation of an emotion node,”
which is to say, the activation of some lexical cluster linked by an
emotion (for example, a cluster linked via romantic love, fear, or
anger), “spreads activation throughout the memory structures to
which it is connected, creating a subthreshold excitation at those event
nodes” (“Mood” 135). Put differently, when a particular “emotion
is aroused . . . activation will spread out along its connections, thus
priming and bringing into readiness . . . associated ideas and mem-
ories” (“Affect” 389). This not only produces priming of rele-
vant memories, but leads us to “focus on affect-consistent rather
than -inconsistent information” (Forgas, “Affect” 244), thus further
perpetuating the feeling. Bower discusses the ways in which partic-

ular emotions lead to selective attention and other mechanisms that
tend to be mood-reenforcing as they lead us to notice and remem-
ber particular aspects of an event or object (“Mood” 142–3). People
“selectively attend to and learn more about stimulus material that is
congruent with their feeling” (Bower, “Affect” 387). Though he dis-
putes the sufficiency of an account based on spreading activation
in this case, Mathews discusses “biases in emotional processing”
in just these terms as well. Specifically, “The nature of these pro-
cessing biases strongly suggests the potential for a circular relation-
ship between cognition and mood. Anxious or depressed mood acts
so as to give processing priority to the type of information that is
most likely to enhance or maintain that mood state” (300); moreover,
“Selective processing is specific to material that matches the content
of the individual’s current concerns” (303).
Though inverted, Bower’s treatment of “subthreshold excitation”
of memories is precisely the process we have been describing. More-
over, Bower’s, Forgas’, and Mathews’ work indicates that, once the
initial memories and emotions are triggered, the emotions will be to
a degree self-sustaining. This is because any given emotion will con-
tinue to prime relevant memories and to focus the reader’s attention
on textual details that themselves sustain the emotion and prime
those memories. Of course, this also means that it may become harder
to return a reader to a desired emotion once he/she has gotten “off
track.” Hence we find the emphasis of the Sanskrit writers on the
consistency of rasadhvani, and on avoiding certain “contradictory”
feelings, such as disgust in connection with romantic love.
A similar idea has been put forth by the philosopher and aestheti-
cian, Noël Carroll. Carroll has argued that fictional narratives are
“prefocused” so that “[c]ertain features or situations and characters
will be made salient through description or depiction.” The important
thing about these features, and so on, is that they govern “the iden-
tity of [our] emotional states” (202). This becomes self-perpetuating
as our emotional states themselves lead us to focus on particular
features that are relevant to those states, as the cognitive research
attests. Fear provides a commonsensical illustration. The emphasis
and consistency of the writer will draw our attention to aspects of an
initial situation that make us fear for the protagonist. Then, fearing
for the protagonist, we focus our attention on just those aspects of the
developing situation that are relevant to our fear.
Carroll’s discussion of works as “prefocused” suggests that there
are many ways in which our initial attention may be directed, and
thus many ways in which even our initial emotional response may
be guided. For example, much of Carroll’s work is on the horror
film. It seems clear that, when someone goes to a horror film, he/she
looks for certain things right from the beginning. This affects his/her
emotional response. More generally, genre often “prefocuses” our
attention, understanding, and response – including the priming of
memories and correlated generation of rasa. It would take us away
from the concerns of the present chapter to discuss this in detail.
However, in the following chapters, I shall argue that there are two
primary prototypical narrative structures, cross-culturally – romantic
and heroic tragi-comedy. These genres are remarkably consistent in
their elaboration of narrative goals and in their development of nar-
rative events. As such, they facilitate the consistent development of
rasa, both for the author and for the audience or reader. First, the idea
of the genre, then its unfolding structure helps to orient the priming
of a reader’s memories, and thus helps to produce a coherent emotive
effect.
Indeed, as we shall see, these genres are based on prototypical emo-
tion scenarios. Research by Conway and Bekerian indicates that just
such scenarios operate to prime or activate personal memories along
with their associated feelings. Though Conway’s and Bekerian’s
research was done outside a literary context, it seems directly
applicable to literary study. First, Conway and Bekerian emphasize
the importance of “situational determinants” to our understanding
of emotions. Though Conway and Bekerian do not conceive of them
in these terms, these situational determinants are, in effect, mini-
narratives. For example, one situational determinant of happiness,
derived from this research, is reunion with a loved one (189). This
implies an entire story of falling in love and being separated – and
simultaneously suggests how “situational determinants” are just
the sort of broad, emotion prototypes that guide the construction of
literary narrative, as I shall argue in the next chapter. In any case, the
general relation between narratives and situations such as this should
be clear. Conway and Bekerian tested subjects to discover just how
they came to imagine such situations concretely – something that we

do routinely in reading literature, as, for example, Roman Ingarden
emphasized (see, for example, his discussion of “schematized
aspects” in Chapters 8 and 9). They found that, the vast majority
of the time, people concretized these situations by reference to
their own “personal experiences.” Thus, “specific autobiographical
memories were cued” by “rather general . . . situations” (178), a
point that should hold equally for literary narratives that are based
on precisely the same sorts of situation, such as reunion with a
loved one. The point also fits well with research of Shaver, Schwartz,
Kirson, and O’Connor who found that personal memories of emotion
episodes tended to be much the same as “typical” emotion episodes,
a result that may be explained by the idea that “prototypes guide
memory . . . of personal experiences” (181). Even more significantly,
in the Conway and Bekerian study, the memories had direct emo-
tional effects on the test subjects, just as our analysis (following
Abhinavagupta) would predict. As Conway and Bekerian sum-
marize, “Mood state appeared to shift” in keeping with “personal
memories of experiences which mapped onto the specified target
situation” (179).
There is also direct empirical research on literary response that sup-
ports this analysis. Specifically, Uffe Seilman, Steen Larsen, László
Halász, János László, and others have conducted experiments de-
signed to determine the degree to which literary and nonliterary
texts spontaneously elicit personal memories from readers (see the
articles by Halász; Seilman and Larsen; and Larsen, László, and
Seilman).5 They have found that literary works trigger memories of
“more personally experienced, and therefore self-relevant, events”
(Larsen, László, and Seilman, “Across” 102). These memories appear
to govern a sense of “personal resonance to and engagement in” a
literary work (102). Oatley and his colleagues report further empir-
ical work that bears out these conclusions as well (see “Emotions”
and citations). Oatley cites experiments going back to the 1930s indi-
cating that “concepts” of “the suggestive (dhvani) . . . and sentiments
5 There is also a long history of less experimentally formalized work that suggests
similar conclusions. Norman Holland’s writings are a particularly good case of this
sort.
(rasas)” should be “of great importance for understanding the im-

pact of narrative.” Among the studies cited by Oatley is research by
F. C. Bartlett that examined the reaction of men and women to a story
about war. He found that the readers’ anxieties about or memories
of separation from their loved ones during war had a significant im-
pact on their response to the story: “Most of the male subjects had
been in the war or faced the prospect of going. For the women, losing
relatives and friends was an ever-present threat.” In consequence,
they responded most strongly to this aspect of the story. In research
done sixty years later, Elise Axelrad “found an effect comparable
to that of Bartlett. She had people record autobiographical memo-
ries that surfaced as they read [James Joyce’s short story] ‘Clay,’ and
found that pieces of these memories became part of what they re-
told when they reproduced the story” (Oatley, “Emotions” 52). In
keeping with this, Halász explains that, according to his research,
literary texts affect “the emotional memory network of the reader,”
activating (or priming) “relevant events and sources which are em-
bedded in an autobiographical context” (“Effect” 83). Halász specifi-
cally links this with the development of empathy for the protagonist
(83 – recall that rasa may be understood as an empathic version of an
emotion).
The general idea is also consistent with research in children’s re-
sponses to stories, research indicating that children often use stories
to deal much more directly with emotional concerns drawn directly
from their own lives. Miller, Hoogstra, Mintz, Fung, and Williams
point out that “a child might appropriate and use for his or her own
purposes someone else’s experience, someone else’s story. Framed in
this way, any story has the potential to be a personalized story” (91).
These researchers go on to discuss one young child’s rather remark-
able use of the Peter Rabbit story. Kurt, a two-year old, hears the
Peter Rabbit story and becomes fascinated with it. He listens to the
story repeatedly. He then retells the story, such that “real-life events
that Kurt had experienced . . . in the company of his mother and
grandmother . . . are attributed to Peter Rabbit and his mother” (103).
The fact that Kurt integrates his own memories into his retellings of
the story – his explicit “personalization” of these stories, as Miller
et al., put it – suggests that memories played a part in his enthusiastic
response to the story initially.
The theory of aesthetic response presented above suggests that

aesthetically successful works, while not entirely “personalized,”
do have a strong component of personalization. The personalized
stories discussed by Miller et al., involve the child self-consciously
taking over the story, integrating it with his/her own life. Aesthetic
experience involves in a sense the first stage of this process – the prim-
ing of congruent or semicongruent experiences of just the sort a child
might self-consciously integrate into the story. We could see aesthetic
response, then, as a development out of the childhood tendency to
appropriate a work entirely to one’s own autobiographical concerns.
The point is reenforced by another aspect of Kurt’s relation to the
story. As Miller et al., note, Kurt did not respond to the story simply
because of its broad thematic relevance to his developmental stage.
Rather, he responded to the story because it was linked with his own
experiences and memories in direct and detailed ways – for example,
“the fact that the story is set in a garden.” Kurt “had spent many en-
joyable hours in his grandparents’ gardens” and only a month before
hearing the story for the first time “his maternal grandparents had
told him about the mother rabbit who had a nest of babies in their
garden.” Moreover, some of Kurt’s earliest personal narratives, prior
to the story, treated this garden (96). Thus, despite Kurt’s young age,
the story related to a complex and emotionally significant set of per-
sonal memories. Miller et al., state that these experiences “prime his
interest in Peter Rabbit’s adventures” (96). They are speaking rather
loosely here. More technically, their analysis indicates that the story
primed the memories, leading to the interest, as he listened to the
story on the first reading, and subsequently.
Again, the obvious difference between adult aesthetic response
and Kurt’s retellings is that Kurt accessed the memories and inte-
grated them into his retellings. But perhaps this is not a difference
at all. Perhaps Kurt did not do this. Perhaps Kurt’s retellings inte-
grated the autobiographical material simply because, being primed,
it impacted his elaboration of details of the story, his “play” with
the story. Axelrad’s research indicates that this happens with adult
responders – as one might expect from our general account. Indeed,
it seems likely that this sort of thing occurs with authors all the time.
It seems likely that authors incorporate autobiographical material
even in entirely nonautobiographical works (for example, in filling
out characters in historical novels) through just such a process. Cer-

tainly, biographical criticism has given us many examples of this sort.
In this way, Kurt’s retellings support not only our account of aesthetic
response, but a parallel account of artistic creation.
objects, causes, and sources of emotion

The mention of Kurt’s revision of the Peter Rabbit story in terms of
his own life brings up a possible misunderstanding of the present
analysis. One might argue that, in responding to the story, Kurt is
actually responding to people, places, and situations in his own ex-
perience. Whether or not this is happening with Kurt, it is important
to emphasize that this is not, in my view, what happens in adult liter-
ary response. That would be akin to what Levinson calls “the shadow
object proposal,” that “objects of response” in fiction are not characters,
but “real individuals or phenomena from the subject’s life experience,
ones resembling the persons or events of the fiction, and of which the
fiction puts the subject covertly or indirectly in mind” (23). To say
that our memories are primed and thus their affect enters into our
response to a fictional work is not to say that the people in those mem-
ories (including ourselves) are the objects of our emotional response.
In relation to this, we might clarify the present analysis by drawing
a simple and common distinction between the object of an emotion,
the causes of an emotion, and the source or sources of an emotion.
The object of an emotion is always some intentional object. It is some
person, place, or thing as we understand or imagine it. Our relation to
the world is always a relation to the world as we see it. My response to
Smith is not a response to Smith as he really is, but as I imagine him
to be. In other words, I have an idea of Smith, an “intentional object,”
that may or may not be accurate. I may trust Smith; thus, my “intent-
ional Smith” is trustworthy, even though the actual Smith may be
duplicitous. The situation is much the same with respect to fictional
characters. I have an intentional Hamlet or Leopold Bloom. Moreover,
I build these up in just the same way that I build up intentional ver-
sions of real people, imputing motives and broad character traits on
the basis of the person’s/character’s actions, statements, and so on. It
is considered a great paradox that a reader might care about Hamlet
in the way he/she cares about Smith. (Indeed, this is Levinson’s
concern.) However, this does not seem to me to be much of an issue.

Yes, I know that Hamlet is not real, but the process of constituting
an intentional version of Hamlet is automatic or spontaneous. I do
not plan it out. It is just part of the way our minds work. Once an
intentional person is constituted, then he/she is open to the same
sorts of emotive response as anyone else. This is because our emotive
responses to persons are not based on a bare idea that they are or are
not real. Rather, they are based on the salience, detail, particularity,
and other aspects of the intentional object. This is why we may weep
over Hamlet and yet remain indifferent to the lives of millions of real
people. Our constitutions of those people are bare, while our con-
stitution of Hamlet is highly elaborated. In any case, the important
point here is that, when we weep for Hamlet, we do in fact weep for
Hamlet (as we understand him, Hamlet in our intentional version).
We do not weep for ourselves.
This leads us to the issues of the cause and the source of emotional
response. The cause of an emotional response is every condition that
gives rise to that response. In the most obvious cases, an emotion
directly concerns one’s own welfare. Thus, the cause involves some
personal concern and some relevant situation. I am in danger, so
I feel fear. But what about cases where someone else is in danger
and I feel fear? Sometimes, for practical reasons or due to affection,
one’s own welfare is bound up with that of the other person. We
may refer to these indirectly egocentric cases as “interested.” But
there are cases where there appears to be no egocentric involvement.
What about these? In such cases, it seems that we assimilate the other
person into a situation from our own experience that was in some
way congruent. For example, seeing someone who is suffering from a
particular illness, I may connect that person with my own experiences
of illness, partially transferring the associated feelings. This is, of
course, the empathic case, to which we have been referring.
This brings us, finally, to the source of an emotion. In egocentric
and interested cases, there is no source for the emotion outside the
current situation. However, in empathic cases, the source of the emo-
tion is external. The source itself is still egocentric or interested, even
though the current situation does not have direct or indirect ego-
centric significance. In these cases, a preexisting emotion enters into
the current situation. Specifically, in the empathic case, the source of
our emotion is our memories of emotional experiences in egocentric

or interested cases. For example, our experience of love comes into
play when we empathize with a character in love; our experience of
sorrow enters when the object of our empathic emotion is someone
suffering grief. Note that this is true whether we are speaking of real
life situations or stories. Our empathy with a real Romeo or Juliet is
much the same as our empathy with the fictional Romeo or Juliet.
Indeed, that is part of the value of the Sanskrit theory, as extended
through recent work in cognition. It suggests not only how we re-
spond to fiction, but also how we respond to life. (We shall return to
this topic in Chapter 4.)
a brief illustration
For the most part, my purpose in these analyses is not to produce
a theory of interpretation, but to produce a theory of literature and
emotion. In other words, my aim is not to present a theory that allows
us to generate more interpretations of literary works. Rather, my aim
is to isolate some of the recurrent features of literary structure, set out
some universals of literary narrative, and articule at least a prelimi-
nary explanatory account of these universals in relation to emotion.
Nonetheless, the theory just presented does have some implications
for the discussion of particular literary works. I shall conclude by
illustrating the point briefly.
In “Emotions and the Story Worlds of Fiction,” Keith Oatley con-
siders two lines from Hamlet. The lines are spoken by Horatio just
after Hamlet dies (5.2.385–6):
Now cracks a noble heart. Good night, sweet prince,
And flights of angels sing thee to they rest!
Oatley writes, “Commentators have remarked that the second of

these lines is the most beautiful in the whole of Shakespeare.” Not
everyone would go that far. Nonetheless, the lines are striking, and
lend themselves to analysis in terms of the principles presented in
the preceding sections of this chapter. A cognitive priming analysis
of rasadhvani should lead us to examine emotive/situational associ-
ations suggested by the lines, both general or common associations
and individual memories. I shall focus on the former and spare the
reader my personal memories. However, I believe it will be clear how
such memories arise for readers in a structured, nonrandom way and

how they bear on individual response.
Obviously, there is sadness in the entire scene. Hamlet has just
died. He died when still young, and he died unnecessarily. Certainly
part of what we feel has to do with the situation itself – a situation that
one could easily analyze in terms of suggestion, priming of memo-
ries, and so on. But obviously the effect of the lines is more than the
effect of the situation. Whether or not one would label them “best in
Shakespeare,” they are certainly more moving than obvious alterna-
tives – “Hamlet is dead,” “The state has suffered a great loss today,”
“No, no!,” “Don’t leave us, Hamlet!,” and so on. But why are they
more moving?
One of the most striking things about these lines, and certainly one
main reason for their effect, is that they implicitly present Hamlet
to us as a child. There is particular pathos in the death of a child.
First of all, in a sense, everyone who dies is like a child dying. If I
know someone for thirty years and he/she passes away, I see not
only the adult, but also the child I first met. Though Hamlet is not a
child now, he was a child – as we are reminded in his reminiscences
about Yorick (“He hath borne me on his back a thousand times”
[V.i.179–80]). His death is the death of that child, especially as it is a
premature death. Moreover, the sorrow of Horatio is presented almost
as if it were sorrow for a child. Indeed, that sorrow is what makes the
lines moving, at least for me; it is what triggers the most significant
memories and associated feelings. In this respect, the crucial line is
“Good night, sweet prince.” Though Hamlet is, of course, literally a
prince, no matter what his age, the term itself suggests youth. It recalls
the fact of his youth, of his infancy even, for he was a prince from
the moment he was born. Moreover, the term “prince” is something
like “seed” or “sapling.” It implies a further condition that is its
culmination – kingship. When a “prince” dies, that person has failed
to reach the culmination his very title implies.
This impression is only increased by the adjective “sweet.” Though
clearly used more broadly in the Renaissance than currently, it none-
theless seems a term that should apply most readily to children – or to
one’s beloved. “Sweet prince,” used as a form of address to the no-
bleman in line to inherit the kingship, shows the affection that one
could easily show to a child, but which would be more difficult – not
impossible, but more difficult – to show to an adult of that social

stature. Conversely, if it was easy for Horatio to show Hamlet that
degree of affection, it suggests that Hamlet retained a youthful open-
ness and a sort of friendly warmth characteristic of children.
“Good night” furthers the point. Of course, we say “good night”
to adults. But saying good night to children is particularly important.
The moments before they sleep are the moments when they can be
readily frightened in the dark, fearing goblins under the bed and
monsters in the closet. Horatio beside the prostrate body of Hamlet
is like an adult at the bedside of a boy. He wishes him good night and
tells him that there will be no goblins and ghosts, but rather “flights
of angels.” He is like a father, reassuring his son. I hardly need to say
in this context what personal memories might be primed for me or
how very similar personal memories might be primed in others, for
the general experience is common, as are the associated feelings.
But, there is a problem here, for Hamlet is not a boy going to sleep.
He has died and cannot hear Horatio’s assurances. Is Horatio simply
denying the fact here, as a bereaved father might deny the death
of his son, looking at his corpse and imagining that he only sleeps?
The ending of the second line contains both an acceptance of Hamlet’s
death and a sort of denial as well. Most obviously, Horatio adopts the
common euphemism for death, “rest” – a euphemism that itself both
asserts and occludes the death. More significantly, this word was part
of the last sentence spoken by Hamlet. Just before the death and this
lament, Hamlet spoke his famous dying words, “the rest is silence”
(5.2.384; I am grateful to Lalita Pandit for drawing my attention to this
line). It is as if Horatio is troubled by that word, that phrase. Hamlet
says that his future is an end to speech. The line implies despair. But
Horatio changes “the rest,” the nothingness that Hamlet suggests will
follow death, into “thy rest,” a peace. More importantly, he transforms
the silence of Hamlet’s death into something even greater and more
beautiful than speech, the song of angels. It is all both acceptance and
denial – something anyone experiences who has lost someone they
love.
This may seem excessive on Horatio’s part. He was, certainly,
Hamlet’s close friend. But would he really have responded to
Hamlet’s death in this way, with this degree of denial and anguish?
There are two points to make here. First, it does not matter. The
preceding analysis concerns the way a passage suggests ideas, scenes,

meanings, and associations, that would foster the priming of tender
and sorrowful memories, leading to feelings of tenderness and sor-
row over Hamlet’s death. It does not matter whether Horatio, the
particular vehicle for these feelings at this particular moment, re-
alistically might have all these emotions. In fact, in some ways the
phrasing itself is understated (contrast “Oh, Hamlet! Do not leave
us!”). It is only through an elaboration of rasa and dhvani that one
comes to see the emotion these lines involve and thus the source of
their power.
The second thing to remark is that this excess itself reminds a critic
that the source of these lines is no doubt Shakespeare’s own experi-
ences. It is well known that the character of Hamlet bears in some way
on Shakespeare’s son, Hamnet, who died, at the age of eleven, not
long before Shakespeare wrote this play (Wright and LaMar xxx). In
treating Hamlet’s death, Shakespeare no doubt drew directly on the
death of his own son. In other words, his own composition of these
lines was animated by the same associations and the same sorts of
personal memory that make the passage powerful for a reader. The
writing of the lines was, in short, a process of precisely the sort we
have been discussing with respect to the reading or hearing of the
lines – though, like Kurt’s retellings of Peter Rabbit, done in the op-
posite direction. In this way, the passage is exceptionally appropriate
for our purposes, for it indicates again that the preceding analysis
applies not only to the emotional response of readers, but to the emo-
tional expression of authors as well, and in precisely the same way.
In following through the development of plot and character, the au-
thor too feels emotions, due to the priming of his/her own memories.
These memories help to guide the author’s selection of further details,
his/her specification of conditions, characters, and events, in just the
way that they guide a reader’s concrete imagination of conditions,
characters, and events.
conclusion
In short, both the production and reception of literature appear to
be inseparable from rasadhvani, understood as the patterned, cumu-
lative priming of personal memories. These personal memories are
representationally congruent with the literary situations developed

in the course of the work and are sometimes cued initially by genre
expectations. The emotional components of these memories serve as
the primary source for our empathic emotional response to literary
situations. On the other hand, the situations – or the characters in
those situations – remain the (intentional) object of our emotional re-
sponse. Finally, the emotions are to a degree self-sustaining, as they
themselves further prime relevant memories and guide our attention
and interpretation while reading (or, in the case of the author, while
writing and rereading).
3
Four Hypotheses on Emotion and Narrative
defining emotions and narrative

Keith Oatley and Philip Johnson-Laird have argued persuasively that
emotion is the product of an agent’s evaluations of his/her success
or failure in achieving particular goals within what is, in effect, a
narrative structure. In connection with this, they argue that there are
five “basic” emotions: happiness, sadness, fear, anger, and disgust.
In part, Oatley and Johnson-Laird are drawing on the research of
Paul Ekman, who has demonstrated that the facial expressions for
these emotions are universal, as are the expressions for surprise and
“interest.” Ekman has maintained that all seven represent basic emo-
tions. Oatley and Johnson-Laird delete surprise and interest from
Ekman’s list on the grounds that they are not genuine emotions. Sur-
prise appears to be a mode of experiencing other emotions (Oatley 60).
For example, surprise is what makes fear into fright; in this case, it is
a mode of experiencing fear.
Anyone who approaches the theories of Ekman, Oatley, and
Johnson-Laird with a knowledge of classical Sanskrit dramatic theory
is likely to be struck by the remarkable similarity between these mod-
ern developments of biological and cognitive science, on the one
hand, and the ancient Indic ideas about narrative and emotion, pre-
sented in the Nāt.yaśāstra and elsewhere over two millennia earlier.
The Nāt.yaśāstra distinguishes eight “permanently dominant” emo-
tions or bhāvas. The list includes mirth, sorrow, fear, anger, and
76
Four Hypotheses on Emotion and Narrative 77
disgust, the five basic emotions of Oatley/Johnson-Laird. It also in-

cludes astonishment or wonder. This is roughly equivalent to “sur-
prise,” from Ekman’s list, but seems less open to Oatley’s objections.
Wonder is not simply what one feels when something unexpected
happens – something frightening, for example. Rather, it is what one
feels when something almost inconceivably great happens. It is, for
example, what one experiences in witnessing a miracle, or what one
takes to be a miracle. In other words, perhaps the relevant facial ex-
pression is ambiguous between surprise, which is indeed modal, and
wonder, which is, rather, an emotion, and Ekman merely chose the
wrong alternative.
I stress the universal facial expressions here in part because they
were themselves part of the Sanskrit theory. Bharatamuni, to whom
the Nāt.yaśāstra is attributed, develops the theory of emotion in
relation to a set of stereotyped facial expressions. Learning these ex-
pressions was an important part of the actor’s training in classical
Sanskrit drama, and continues to be a part of an actor’s training in
regional theater. Unsurprisingly, the set of plates in Ekman, illustrat-
ing the universal facial expressions, bears striking similarity to the
expressions learned by actors in Indian dance and theater traditions
continuous with the Nāt.yaśāstra (see, for example, the photographs
following page 96 in Singha and Massey).
Moreover, though Oatley relies on Aristotle for his understanding
of plot, Sanskrit narrative theory would have been far more rele-
vant. Aristotle, despite his well-known references to hamartia (error
or flaw), tends to treat plot in terms of a relatively agent-neutral
causal structure, discussing “necessary or probable” sequences of
events linking beginning, middle, and end. In contrast, the Sanskrit
theory is entirely agent-centered. Indeed, Oatley ends up modify-
ing the theory of the Poetics until it bears less similarity to its origi-
nal than to the theory of the Nāt.yaśāstra. More exactly, the Sanskrit
writers relate their eight emotions to a narrative theory that is ex-
plicitly structured around an agent’s pursuit of goals and his/her
periodic reevaluation of likely success or failure in achieving those
goals. This reevaluation is formalized in Sanskrit narrative theory
through the concept of the “pause” or “deliberation,” a critical ele-
ment of narrative structure, and one that involves “anger or passion
or temptation” (Dhanam . jaya 24). The standard translations refer to
this as a “juncture,” which is, coincidentally, just what Oatley terms

the moments of evaluation that give rise to emotion (see Oatley 25,
36). More exactly, like Aristotle, the Sanskrit theorists distinguish
three parts to a narrative. The Sanskrit writers have relatively little
to say about the beginning (ārambha, which implies the commence-
ment of an undertaking [see Monier-Williams]) and conclusion –
also termed “culmination” or “attainment of the object.” However,
they do analyze the “middle” – the “development,” “expansion,” or
“progression” as they call it – in some detail.
If we follow Abhinavagupta (Locana 438) and combine elements
from two different parts of this analysis, we may divide the middle
“progression” into three sections: 1) the hero’s initial effort; 2) his/her
“pause” to evaluate the prospects for success; and 3) the renewal of
effort, with full concentration. In Sanskrit theory, which is designed
for comedy, the third stage also carries the implication that the hero
is fully confident that he/she will succeed. Hence, this part is usu-
ally translated “certainty of success.” However, the Sanskrit term,
niyatapti, has a meaning related to constant or steady concentration
involving the control of passion. This is the element I have empha-
sized in defining the parts of the progression. I have dropped the
assumption of success in order to allow application outside of com-
edy. (I should say that this interpretation is far from uncontroversial.
For a different view, see Ingalls 441n.) In any case, the relevance of
this to Oatley’s and Johnson-Laird’s views should be clear.
Oatley at one point remarks that the sort of theory he is developing
“is in a sense, familiar to the European or American mind,” a mind
accustomed to the “idea of rational planning” (29). Oatley is prob-
ably just being cautious here. Afraid that someone will accuse him
of imperialistic universalism, he confines his claim to Europe and
America. But in fact his theory is much closer to Indian theories than
to Euro-American theories. This congruence serves to remind us that
much of potentially great intellectual value is lost when we remain
ignorant of other intellectual traditions, or assume that our own ideas
are necessarily more in keeping with those of “our own” tradition,
and necessarily incommensurable with those of “alien” traditions.
Ekman’s theory of basic emotions would have been advanced and
clarified by incorporation of the Sanskrit views. Much of Oatley’s pre-
liminary analysis regarding narrative and emotion would have been
unnecessary, and his work could have begun at a more advanced

point.1
More importantly, however, this congruence adds weight to
Oatley’s theory, by showing that its application is not confined to
Euro-American “rational planning.” Still, it is important to be clear
about the precise nature of this support. For example, does the simi-
larity between Oatley and the Sanskrit writers suggest that there are,
in fact, basic emotions? This is not clear. The whole issue of basic
emotions is highly controversial. It is clear that some emotion terms
are basic within certain limited domains. For example, it is clear that
“fear” is the basic term by reference to which we define, say, “terror”
or “fright.” In a similar way, “cold” and “hot” are basic terms by
reference to which we define, say, “scalding” or “frigid.” But it does
not seem to make sense to say that coldness itself – not the term, but
the property or the feeling – is basic, whereas scaldingness or frigid-
ness is derivative. The same point applies to “fear” and “terror.” Fur-
ther difficulties arise when Johnson-Laird and Oatley see jealousy
and hatred as forms of disgust (114, 116), an analysis that I suspect
most readers will find difficult to accept. (For a fuller discussion of
problems with theories of basic emotions, see Ortony and Clore and
citations. We shall return to the issue – or, rather, the related issue
of the biological givens of emotion – in the Afterword.) Oatley and
Johnson-Laird have made an argument regarding basic emotions that
must be taken seriously, and they have clarified the concept in a way
that few other theorists have managed to do. Though certainly worth
pursuing as one alternative in a research program, however, the idea
of basic emotions remains problematic. At the very least, we cannot
simply assume it as a default hypothesis.
1 I should note that, once I pointed out the connection with Sanskrit writing, Oatley
was delighted. He set himself the task of reading extensively in the field, and incorpo-
rating Sanskrit poetics into his own theoretical and interpretive work. Unfortunately,
most western humanists do not appear to be so open minded. Many have developed
such a thoroughgoing commitment to cultural difference that they are incapable of
recognizing similarity. A good example of this is found in the introduction to Hjort’s
and Laver’s fine collection, Emotion and the Arts. There, Hjort and Laver cite the “basic
emotions” of rasa theory as “specific to a cultural community” (15) and as evidence
of “radically different cultural self-understandings” (15). I should perhaps note that
the emotions that do not overlap with Ekman’s universal facial expressions are far
from exotic, as we shall discuss below. For example, one is romantic love.
But, if the idea of basic emotions is problematic, what is one to

make of the universality of some facial expressions and of the close
correspondence between the theories of the Sanskrit writers and those
of Ekman, Oatley, and Johnson-Laird? The first thing we need to do
here is separate the issue of universality from that of basicness. The
existence of a limited set of universal facial expressions of emotions
certainly suggests something, but what? The fact that these emotions
turn up in the Sanskrit theory indicates that they have some sort of
special importance for aesthetics, but it seems rather unlikely that all
societies developed the same facial expressions so that they could all
have theater. What, then, are we left with, beyond the hypothesis of
basic emotions? To answer this question, we need to concentrate for a
moment not on the emotions, but on the facial expressions. The most
obvious implication of the universality of such facial expressions is
not that the correlated emotions are more basic than other emotions,
but that these emotions are more directly involved with and germane
to empathic communication. After all, that is what facial expressions
do – communicate emotion. Moreover, this is consistent with other
data as well, such as the finding of Shaver et al., that the putatively
basic emotions are the most prototypical emotions. It would make
sense for prototypicality to be bound up with interpersonal saliency.
Finally, this view fits the Sanskrit case as well, for the communication
of emotion is obviously central to facial expression in the theater –
and to other, nonfacial aspects of literature.2
This leads us to a significant difference between the Sanskrit theory
and that of Oatley and Johnson-Laird. The Sanskrit theory does not
require that the emotions specified in the theory be basic. It requires
only that they have some sort of special place in aesthetic feeling.
Moreover, the reason for this place is suggested in the precise for-
mulation of the theory. Specifically, Sanskrit theorists do not focus
2 Oatley and Johnson-Laird stress the communicative function of emotions. However,

they are speaking of emotions themselves, not facial expressions, and their view
is that all emotions have communication as their primary function. I have only urged
that certain emotions may be more readily open to empathic communication – a very
different thing. Indeed, it is not even clear that the use of the word “communication”
is the same in the two contexts. For Oatley and Johnson-Laird, “communication” is
simply the transmission of information, not only between persons, but within one
person (see Oatley 44).
on emotion per se or “bhāva,” as we noted in the preceding chapter.

Rather, they focus on “rasa.” Once again, the term refers to an aes-
thetic feeling. That aesthetic feeling is, roughly, an empathic version
of an emotion or bhāva. When the hero is faced with a dragon, we
do not fear the dragon, experiencing the bhāva or emotion itself, and
therefore running from the theater. Rather, we fear for the hero. We
feel something clearly related to fear, but not identical with fear: We
feel empathic fear, which is to say, the “terrible” rasa. In the preceding
chapter, I simply used “emotion” to refer to “bhāva.” In this context,
however, “bhāva” refers most importantly to egocentric feelings, and
in that way it is narrower than “emotion.” “Rasa,” again, refers to em-
pathic feelings. From this point on, I shall use “emotion” to refer to
feelings of both sorts, reserving “ bhāva ” and “rasa” for the more
specialized meanings.
From this, we may conclude that, whatever we decide about basic
emotions, the congruence of the Sanskrit theory with the cognitive
theories of Oatley and others has to do with the particular empathic
salience of certain emotions and the centrality of empathic commu-
nication to literary feeling.
For the moment, I should like to turn away from the particular list
of empathically salient emotions and focus on the more general issue
of the empathic communication of emotion. This, in turn, will pro-
vide us with a way of reunderstanding the relation between emotion
and narrative. At the same time, it will help us to rearticulate, and
organize, the list of predominant empathic emotions or rasas.
empathy and emotion prototypes

Empathy involves, among other things, a conscious or unconscious
inference to what someone else is feeling. (It also involves an open-
ness to putting oneself in the place of the other person – see Davis
16–17; I shall not be considering this aspect of empathy here.) When
we identify an emotion, we access an internal lexical entry for the
emotion in question – linking, for example, some set of perceptions
with the entry for “fear,” “disgust,” or whatever. These entries in-
clude some sort of connection to the “phenomenological tone” of an
emotion – what it is like to feel fear, disgust, and so on. However, this
cannot be all there is to our lexical entry. Most importantly, this does
not give us a way of identifying when someone else has an emotion,

unless he/she tells us – for we have no access to his/her “raw feels.”
Moreover, his/her statement that he/she is experiencing a certain
emotion is of no value in this context, because if the only way to
identify someone else’s emotions is through verbal communication,
we have no way of learning emotion terms to begin with.
For this Wittgensteinian reason, Anna Wierzbicka has argued that
“emotions can be identified only by a reference to a standard situa-
tion” (“Emotions” 60) and thus that the meaning of emotion terms has
the general form, “X feels emotion1 = X feels as one does when . . .”
For example, “X feels sad = X feels as one does when one thinks
that what one has desired to happen has not happened and will not
happen” (61). Ronald de Sousa makes a similar point: “We are made
familiar with the vocabulary of emotion by association with paradigm
scenarios” (142). De Sousa goes on to distinguish two aspects of these
scenarios – roughly, causes and effects. Along the same lines, Oatley
distinguishes eliciting conditions (what leads one to feel an emotion)
from expressions or action consequences (what one does when one
feels the emotion). The Sanskrit theorists, once again, anticipated this,
distinguishing “vibhāvas” or determinants and “anubhāvas” or con-
sequents of bhāvas.
In keeping with these divisions, then, we might say that our lexical
entry for any given emotion term includes not only some link to the
feeling of the emotion (typically via memories, as discussed in the
preceding chapter). It also includes some account of the kinds of sit-
uation that give rise to the emotion and some account of the kinds of
expression and action that result from an emotion. Wierzbicka charac-
terizes these accounts in fairly abstract terms. She is no doubt correct
that we have, or are able to generate, abstract definitions of this sort.
However, I doubt that our actual cognitive processes surrounding
emotion – for example, our ordinary empathic responses – operate at
such an abstract level. Rather, I suspect that, like most of our lexical
entries, the entries for our emotion terms are more concrete. Empathic
and other forms of inference, imagination, projection of emotion are,
in other words, less a matter of subsuming instances under categories,
than of comparing instances with more fully specified types.
More exactly, a good deal of research suggests that our lexical en-
tries are, most often, structured around prototype cases, not necessary
and sufficient conditions (see Holland et al., 182ff. and citations). For
example, our use of the word “bird” is more a function of comparison
with a bird prototype or with exemplary instances of that proto-
type, such as robins, than of any strict definition – “warm-blooded,
oviparous vertebrate with alar forelimbs,” or whatever. In identifying
something as a bird, in discussing birds, and so on, we do not begin
with an abstract category and judge whether or not instances fit the
general conditions specified in that category, whether they fit the def-
inition “warm-blooded vertebrate,” and so on. Rather, we base our
inference, identification, or whatever, on similarity to prototypes.
first hypothesis: the prototype basis of

emotion terms
This leads to the first hypothesis I should like to put forth here: Emo-
tion terms are prototype-based in both eliciting conditions and expres-
sive/actional consequences. In other words, the “situations” identified
by Wierzbicka and “scenarios” discussed by de Sousa should be un-
derstood as specific in the way that the prototype for “bird” is specific.
Thus, when we judge someone to have a certain emotion, we do so by
comparing his/her situation with prototypical situations and his/her
response with prototypical responses. For example, our lexical entry
for “sad” would not include a Wierzbicka-like abstract equation, but
rather something along the following lines: “What you feel when
someone you love dies and what you express through weeping.”
(Or, rather, if it does include an abstract equation, that equation is
subordinate to the prototype in our usual cognitive processes.)
Here, it is valuable to return to the view shared by Oatley, Johnson-
Laird, and the Sanskritists that emotions are embedded in stories.
One way of putting the difference between Wierzbicka’s view and
the prototype account just proposed is to say that the latter defines
emotion concepts in terms of what are in effect mini-narratives, seeds
of stories (to use one of the Sanskrit writers’ favorite metaphors).
For example, suppose that the prototype – or one prototype – of
the eliciting conditions for sorrow is indeed the death of a loved
one. This is, implicitly, the outcome of a story, a story of personal
affiliation between the sorrowful person and the person who has
died, and the story of that person’s death. As the prototype becomes
more specific, the implicit narrative becomes more fully developed

as well. For example, suppose the prototype is “the premature death
of a loved one.” This further concretization of the prototype entails a
further concretization of the implied narrative.
Readers familiar with the literature on emotion will know that, in
influential work, George Lakoff and Zoltán Kövecses have proposed
a prototype-based account of emotion concepts as well. However, the
Lakoff/Kövecses view of emotion prototypes is quite different from
the view I shall be developing in the following pages. Before going
on to discuss my own account in more detail, it is worth indicating
what these differences are.
Most obviously, Lakoff and Kövecses treat much broader se-
quences of events, and they treat these events much more abstractly.
Kövecses explains that, in their view, the “prototypical cultural model
of anger” is the following “five-stage” sequence: “cause of anger, ex-
istence of anger, attempt at control, loss of control, retribution” (173).
“Cause of anger” involves five elements: “Wrongdoer offends self,”
“Wrongdoer is at fault,” and so on (143). Though these clearly do
not present necessary and sufficient conditions, they are far more
abstract than my prototypes – so abstract that I would refer to them
as “schemas” or, alternatively, “scripts.” (I shall return to this point
in a moment.) For example, in my view, “wrongdoer offends self” is
too broad. I am offended when someone cuts in front of me in line.
But I would hardly consider this a prototypical eliciting condition
for anger. Rather, in my usage, the prototype eliciting condition for
anger – or one such prototype, depending on context – would be
something more along the lines of “someone unexpectedly strikes
me” or perhaps “someone unexpectedly tries to knock me off
my feet.”
More importantly, perhaps, the Lakoff/Kövecses five-stage se-
quence collapses different levels of response and analysis. For
example, I do not believe that “Wrongdoer is at fault” is commonly
a prototypical eliciting condition for anger. Rather, it is a necessary con-
dition for the justification of anger. Thus, the Lakoff/Kövecses cate-
gory “Cause of anger” combines direct considerations of anger and
meta-level considerations of our response to or evaluation of anger.
Their next two categories operate at this meta-level as well. “Attempt
at control” and “loss of control” point toward actional/expressive
outcomes. However, the former bypasses those outcomes entirely
for a meta-level response to anger. The latter includes quite general

outcomes – such as “aggressive actions” (143) – but also incorpo-
rates the meta-level response to the motivating impulse of anger. The
final category is somewhat odd as it overlaps with “loss of control,”
repeating the actional outcome. In this case the outcome is a “retribu-
tive act against wrongdoer,” explained again in very general terms
as “usually angry behavior” (144). However, this category does not
lack its own meta-level addition, for an evaluative element enters,
determining that “The intensity of retribution balances the intensity
of offense” (144).
It is no doubt the case that our ideas about emotions include
not only the emotional sequence itself – eliciting conditions, phe-
nomenological tone, and actional/expressive outcomes – but also
(vague) principles of justification, (variable) norms bearing on the
self-conscious evaluation and control of the emotional impulses, and
so on. Indeed, we shall find many elements of just this sort in the
development of literary narratives in relation to emotion prototypes.
Nonetheless, I doubt that there are scripts of precisely the sort set out
by Lakoff and Kövecses. Rather, I suspect that their findings are a
residue of our separate ideas about the causes of emotions, the jus-
tifications of emotions, prudential or ethical responses to emotions
(for example, it is normative that one try to control one’s anger), and
so on. In any case, it should be clear that sequences of this sort are
not “emotion prototypes” in my sense of the phrase. Even if we do
have scripts of this kind with respect to different emotions, they are
not what I am concerned with here.3
3 This variance in usage of the term “prototype” is not confined to Lakoff and Kövecses.
It indicates a much broader difficulty regarding terminology among writers on cogni-
tion. James Russell points out that some “prototype” accounts focus on “remembered
individual, concrete experiences”; some treat “generalized schemata”; and some
focus on “average” cases (39). In the preceding chapter, I distinguished schemas,
prototypes, and exempla. But the quotation from Russell indicates that “schema”
(Russell’s “generalized schemata”) and “exemplar” (Russell’s “individual, concrete
experiences”) are often used as equivalent to “prototype.” In fact, for many writ-
ers, the only important conceptual division seems to be one between necessary and
sufficient conditions, on the one hand, and everything else, on the other hand. It
is undeniably important to distinguish between accounts based on necessary and
sufficient conditions and accounts not based on necessary and sufficient conditions.
Indeed, that distinction will figure prominently in the following pages. But that
is not the only consequential distinction in this area, hence my more systematic
usage.
In short, the sort of emotion scenarios described by Lakoff and

Kövecses, as well as those treated by James Russell and a number of
other writers – including, for example, Wierzbicka – are schemas, in
my sense, not prototypes. Consider, again, sorrow. I have suggested
that the prototype for sorrow is, roughly, “what you feel like when
someone you love dies and express through weeping.” This illustrates
the point nicely. It is not abstract, but concrete. It includes the default
values characteristic of schemas (for example, desiring something to
happen and it not happening). But it specifies and further concretizes
these with “average” properties and distinguishing characteristics.
As elsewhere, the “averaging” here occurs over highly salient in-
stances. The displays of sorrow that accompany death, along with
the elaboration of mourning in funerary rituals, make deaths highly
salient instances of sorrow. (The same point could be made about
marriage and joy.) Deaths are also highly distinguishing instances of
sorrow. Many things might make us feel sorrow. But death is distinc-
tive in its almost invariable association with sorrow and in the way it
intensifies sorrow-relevant properties (for example, irreversibility).
second hypothesis: the basis of prototypical stories

in emotion prototypes
As I have already emphasized, the concreteness of emotion proto-
types is related to the fact that they are bound up with implicit stories.
This leads us to our second concern, the nature of narrative. Here, I
shall propose a more complex hypothesis. First of all, the research
on prototypes suggests that we should understand narrative too in
terms of prototypicality. Like birds, there are some narratives we con-
sider to be “more standard” cases than others. For example, the story
of young lovers overcoming obstacles to be united is a more “stan-
dard,” which is to say, prototypical narrative, than the story of how
a furnace operates.4 There are no doubt many properties that tend
4 To get a sense of prototypicality, it is useful to have a nonprototypical example. Take

any standard love story – canonical or ephemeral, moving or banal – and contrast the
following story from Schank, which is highly nonprototypical: “Diane was trying to
figure out why hairdressers won’t take credit cards. She thought that maybe they
had a poor clientele but realized it was also true in Westport, a well-to-do area. She
never found an answer” (150).
to characterize prototypical narratives. For our purposes, the most

important of these is emotional interest. In other words, one dif-
ference between prototypical narratives (for example, stories about
lovers) and nonprototypical narratives (for example, stories about
the operation of appliances) is that the former engage our feelings, or
at least address and appeal to feelings.5 This part of the hypothesis
seems fairly uncontroversial.
But what does it mean to say that some narratives are prototypical
and others are not? It means that our mental lexicons include proto-
types for narratives and we judge narratives – including what is or is
not a narrative – by reference to these prototypes. (Again, just how we
understand the existence of these prototypes is not important for our
present purposes.) But the relevance of prototypes to literature does
not end there. We employ prototypes not only in identifying narra-
tives, but in understanding and creating them. Moreover, we employ
a wide range of prototypes – for genres (for example, epic or novel),
characters (for example, romantic heroine or action hero), scenes (for
example, bar scene or battle scene), and so on. Prototypes guide our
ongoing interpretation of stories, as we hear or read them, and our
ongoing generation of stories, as we speak or write them. The point is
merely a generalization of the research on prototype-based thought,
mentioned previously. Consider genre. When writing an epic or a
novel, an author is almost certainly guided far more by prototypes
of these genres than by necessary and sufficient conditions. He/she
is guided by the standard or typical opening, character structure,
and so on, for epic, not by some rigorous definition. Similarly, when
readers are reading an epic, their expectations are guided by their
prototypes for the genre. What they find surprising or innovative or
faulty is not a result of an abstract definition. Put differently, the sorts
of reader expectation isolated and examined by phenomenological
5 The research of Brewer and Lichtenstein clearly supports this contention, for they
found a close correlation between emotional effect and significance of narratives,
on the one hand, and judgments of storiness, on the other. However, Brewer and
Lichtenstein were not operating with a prototype model. Thus, they did not formu-
late their study nor examine their findings in these terms. Moreover, they treated
aspects of discourse order – another factor in prototypicality judgments – rather than
story structure per se (that is, story structure in the narrow, technical sense). Again,
discourse is an extremely important topic for understanding literature and emotion.
However, it is a large and complex topic, requiring treatment in a separate volume.
and response-oriented critics such as Wolfgang Iser and Hans Robert

Jauss are, to a considerable degree, a function of narrative prototypes.
Again, a wide range of prototypes enter into our generation of, or
response to, narrative. If indeed one standard goal of literary story-
telling is emotional appeal, then we would expect one crucial set of
prototypes to address this emotive aspect of narrative. This leads to a
complication of my second hypothesis. When the purpose of our sto-
rytelling is, in part, emotive – thus, in prototypical narratives – that
storytelling will be bound up with emotion prototypes. These emo-
tion prototypes will help guide our decisions as to what sort of story
is tellable, what is of interest, what is valid, and what is effective and
engaging. This is true whether the narrative in question is fictional,
biographical, or historical; set in the form of an epic, a drama, or a
novel. In each of these cases, due to the emotive purpose of the tale,
emotion prototypes will provide central structural principles for the
story, partially guiding its overall shape and outcome, its tone, and
so on.
To say that an emotion prototype guides storytelling is to say, ac-
cording to the preceding analysis, that both the eliciting conditions
and the actional consequences guide that storytelling. However, in-
sofar as prototypical narrative involves a sustained, empathic appeal
to readers or listeners, it almost necessarily involves a particular em-
phasis on eliciting conditions, themselves understood, once again, as
the conclusions of implicit narratives. We may, of course, feel badly
for a character who is weeping. But our response is given depth and
intensity only through an understanding of what has led to this ex-
pression of sorrow. Indeed, our knowledge of the actual eliciting con-
ditions may make weeping comic and laughter sorrowful. The point
should become clearer in the next section, when we distinguish dif-
ferent types of emotion and their different functions in prototypical
narratives. For now, however, it is important to emphasize the par-
ticularly consequential role of eliciting conditions in the generation
of narratives.
My second hypothesis, then, fully formulated, is the following: Pro-
totypical narratives – including literary narratives – are generated largely
from prototypes, prominently including the prototype eliciting conditions
for emotions. Put differently, our prototypical stories are, in their broad
structure, expansions of the micronarratives that define our emotion
terms. Moreover, we experience them in that way, implicitly linking

events in the story with their projected conclusions in prototype elic-
iting conditions for emotions – as when we experience Aristotelian
fear while watching a sequence of events unfold that we expect will
lead to prototypical conditions of sorrow for the hero or heroine.
If I am correct in this hypothesis, it should immediately be clear
that this coheres nicely with the empathic account of literary emotion
given above. For the proximity of a literary narrative to prototype
cases of eliciting conditions should facilitate the immediate, unself-
conscious lexical identification of the emotion, which is a crucial part
of empathic identification. More importantly, it is also likely to facili-
tate the priming of relevant personal memories, with their associated
phenomenological tone (as discussed in Chapter 2).
But is this to say that narratives are made from the prototype de-
terminants of any and every emotion? Well, in fact, there are nar-
ratives for every emotion. For example, we often tell stories simply
to communicate anger or disgust. Thus, if narratives are understood
in terms of necessary and sufficient conditions, we have to say that
no emotion is excluded from this narrative-generating use. However,
this is not true for our prototype narratives. Our prototype narratives
may include anger or disgust, but they are not defined by reference
to, and thus via prototypes for, anger and disgust. In other words,
while emotive narratives are more prototypical than nonemotive nar-
ratives, some emotive narratives are themselves more prototypical
than others. This prototypicality is, in turn, bound up with the typo-
logical and structural relations among emotions.
typological and structural relations

among emotions
Consider again the emotions discussed by Oatley, Johnson-Laird, and
the Sanskrit theorists. Whether we consider these to be basic, to be
particularly open to empathic communication, or neither, it should
be obvious almost immediately that there is a problem with treat-
ing these emotions as if they are all of the same type. In terms of
the broader narrative structures discussed by all these theorists – in
which narrative is goal-oriented action involving periodic evalua-
tion of one’s likely success – we could say that we never evaluate
happiness in terms of fear, anger, or disgust. Rather, we evaluate

fear, anger, and disgust – or, more exactly, fear-, anger-, and disgust-
relevant events – in relation to happiness and sorrow. For exam-
ple, fear is, precisely, fear of what will lead to sorrow or block
happiness.
Happiness and sorrow are, in many ways, general markers for pos-
itive and negative feeling, with “positive” here meaning something
like “the feeling we wish to achieve and sustain” and “negative”
meaning something like, “the feeling we wish to end and avoid.”
Their phenomenological tone is, roughly, the mental equivalent of
what in bodily sensation is termed “pleasure” and “pain.” Moreover,
anger, disgust, and fear are what might be called punctual emotions.
They are prototypically elicited by temporally thin events – an attack
by a wild animal, the smell of something putrescent – and are not
(prototypically) enduring. Happiness and sorrow, in contrast, pro-
totypically arise with respect to more temporally thick conditions.
Though the death of a loved one is, in and of itself, momentary, it
gives rise to an enduring condition, and that condition is what in-
spires the emotion. Moreover, they are themselves (prototypically)
more lasting. In actual fact, the experience of fear or anger may
be more lasting than the experience of happiness; but facts often have
limited bearing on our prototypes (due to salience – including the
salience of fictional instances – norms, and contrast effects, as dis-
cussed previously). For example, happiness may, in fact, last only a
few moments. But our prototype for happiness is well-expressed in
the formulaic ending for a romantic comedy: and they lived happily
ever after.
In connection with this, and in keeping with Oatley’s notion that
emotions occur at moments of periodic evaluation, the moments of
narrative “pause” in the Sanskrit scheme, we could distinguish be-
tween two types of emotion. On the one hand, there are emotions
that have their most important function at specific points within the
course of an encompassing narrative – whether the sort of literary
narrative treated by the Indian writers, or the sort of life narrative
treated by Oatley. These emotions are typically limited in their tem-
poral scope, and prominently include fear, disgust, and anger. To this
list, we might add wonder, as well as mirth, which is distinguishable
from happiness, understood as an enduring state, and is probably
closer to what is expressed in the universal facial gesture. This covers

the emotions that have universal facial expressions and thus may in-
clude the most important or frequent emotions in contexts, such as
literature, where empathy is of critical importance. In any case, I shall
refer to these, and any other emotions of the same general type, as
“junctural” emotions, for they typically define some sort of juncture,
some interruption or pause in the encompassing narratives. More ex-
actly, we might define a juncture as any implicit or explicit assessment
of an agent’s position in an imagined trajectory of actions and events
as they seem likely to play out relative to some goal.6 Note, then,
that junctures are not objective features of a causal sequence, but re-
sponsive features, features of an agent’s own realization, an author’s
reflection, or a reader’s concern and consideration. Of course, some
points in life and in narrative lend themselves to this sort of evalu-
ation. When someone is attacked by a pack of wolves, this is pretty
likely to count as a junctural moment for him/her, and for anyone
reading about it. Moreover, Labov notes that points of particular in-
tensity in a story, such as the beginning of a fight, often involve a
suspension of the action and even an expression of the protagonist’s
own assessment of the situation (374). Nonetheless, in principle, such
reflection can occur at any moment.
In contrast with junctural emotions, I shall refer to happiness and
sorrow as “outcome” emotions, for they define the enduring feelings
we prototypically consider the final evaluation points for junctural
emotions. Thus, we find the prototypical endings of comedy and
tragedy – the former, in which everyone lives “happily ever after”; the
latter, which carries the implication of a sorrow that is irreversible (for
example, due to a death). Of course, happiness and sorrow do operate
as junctural emotions as well as outcome emotions. It is particularly
frequent for sorrow to operate as a junctural emotion in narratives
with an outcome emotion of happiness. But their most important and
distinctive role in narratives is as outcome emotions. In actual life,
one might argue, happiness is never an outcome emotion, but only
6 In connection with the preceding discussion of emotion and memory, it is worth

remarking that, according to Schank, remindings occur with particular force and
importance at points of expectation failure. Junctural moments are often just such
points – hence points when memories, and their associated affects, are likely to be
particularly pressing.
a junctural emotion. In other words, in life, there is no such thing as

“happily ever after.” However, our prototype of happiness is, again,
enduring. Moreover, this is not at all confined to fiction; it is part of
our beliefs about real life as well. For example, one purpose of such
ideas as Heaven is to allow us to imagine enduring happiness as an
encompassing goal of real life.
This distinction between junctural and outcome emotions allows
me to clarify my second hypothesis. Insofar as an emotion operates
as an outcome emotion in a narrative, its eliciting conditions are most
crucial for the generation and reception of that narrative. It is precisely
these eliciting conditions that define the narrative conclusion. Indeed,
insofar as narratives recount pursuit of a goal – and prototypical nar-
ratives do just this – they are necessarily organized by reference to the
eliciting conditions for one outcome emotion, happiness, since happi-
ness is the aim of action. Moreover, we do not pursue undifferentiated
happiness. Our goals in pursuing happiness are necessarily far more
specific. The specificity of the goals, the specificity of happiness is of
course the specificity of eliciting conditions. To desire happiness is to
desire a particular set of circumstances that one takes to cause hap-
piness. These too are, most often, prototype-based. In other words,
when we imagine happiness, we tend to imagine prototypical elicit-
ing conditions for happiness (that imaginative inclination is, after all,
part of what makes such eliciting conditions prototypical). And, of
course, the goals we pursue (expect characters to pursue, and so on)
are just the ones we imagine will produce happiness – thus, again,
generally the prototypical ones.
Indeed, the account of emotions as defined by prototypes, rather
than by necessary and sufficient conditions, is perhaps nowhere more
obviously consequential than with respect to “happiness” and “sor-
row,” for these emotion terms are almost vacuous when defined by
necessary and sufficient conditions. As Wierzbicka puts it, “X feels
joyful = X feels as one does when one thinks that something has hap-
pened that one desires to happen” and “X feels sad = X feels as one
does when one thinks that what one has desired to happen has not
happened and will not happen” (“Emotions” 61). These are, again,
almost bare positivity and negativity. But clearly happiness and sorrow
mean something much more than that in our experience, inference,
action, and so on.
This leads us to the question of what these prototype eliciting

conditions might be. In order to think about this, I should like to
return once more to Sanskrit aesthetic theory. Consider again the
list of bhāvas. Most of the bhāvas overlap with the emotion lists of
Ekman, Oatley, and Johnson-Laird, and we have already included
the shared items – fear, anger, and so on – in one of our two emotion
categories. However, there are two bhāvas on the Sanskrit list that
have no correlates in Oatley et al., and which thus far have no place
in our anatomy of emotion types. These are “rati” and “utsāha.” Rati
is sexual love, but is prototypically represented as romantic love.
Utsāha is resolute perseverance, but is prototypically represented as
heroic devotion to a cause in the face of great odds. Clearly, they
are not outcome emotions. They are not emotions that arise at the
resolution of a series of events. But, at the same time, they are not
junctural emotions. They are not momentary responses to temporally
thin events. They are, rather, emotions that motivate us to engage in
the actions that make a plot. They are feelings that drive characters to
proceed through the various junctural emotions to a final outcome.
I shall refer to them as “sustaining” emotions.
Clearly, actions based on romantic love or heroic perseverance are
actions that one hopes will result in happiness. But, again, this is not
generic happiness. Rather, it is happiness of a specific sort. It is happi-
ness resulting from the achievement of a particular goal. In keeping
with this, the Sanskrit theorists maintain that every story is orga-
nized around one of four more particular goals, because these are the
four goals around which all human life is organized: kāma, artha,
moks.a, and dharma. “Moks.a” or spiritual release and “dharma” or
ethical duty are “higher” goals. Though they are important to a com-
plete treatment of our topic, I shall leave them aside for the moment.
They have less direct bearing on the topic of emotion and narrative.
Moreover, they are, to some degree, narratively parasitic on the other
goals. Thus, narrative treatments of moks.a are often modifications of
narratives based on kāma or romantic union; narrative treatments of
dharma are often integrated into stories of artha or political ascen-
dancy. (We shall return to these points in subsequent chapters.)
In this way, kāma and artha are the two fundamental or pri-
mary goals in the Sanskrit scheme. “Kāma” means “pleasure,” es-
pecially sexual pleasure, but its prototype is the romantic union of
lovers. To take a non-Indic example, Chikamatsu expressed the idea

in Love Letter from the Licensed Quarter: “The only happiness in this
broad world” is “True love to true love” (234). “Artha” refers to
material well-being, and is often translated as “prosperity.” However,
an examination of the classical texts on artha – such as the Arthaśāstra
and Nārāyan.a’s Hitopadeśa – indicates that the prototype for artha
is political power or practical social authority. As the Malian Epic of
Son-Jara puts it, “All people . . . seek to be men of power” (Fa-Digi/I.
1277–8). The two points are combined by Euripidēs’ Hecuba, speak-
ing to the dead Astyanax: “[I]f you had enjoyed youth and wedlock
and the royal power than makes men gods, then you would have
been happy” (Trojan Women 200). These two goals are obviously cor-
related with the two sustaining emotions peculiar to the Sanskrit list –
rati and utsāha. Rati (romantic love) is the emotion that inspires and
accompanies one’s pursuit of kāma (romantic union). Utsāha (heroic
commitment or perseverance) is the emotion that sustains one’s pur-
suit of artha (social and political power).
third hypothesis: romantic union and power as the

prototypes for happiness and the bases for
prototypical narratives
This suggests a solution to our quandary about the precise proto-
types operating in the formation of narratives from outcome emo-
tions. Thus, it leads to my third hypothesis: Romantic union and social
or political power (including material prosperity) are the two predominant
prototypes for the eliciting conditions of happiness. Thus, they are the proto-
typical outcomes from which our prototype narratives – including literary
narratives – are generated. Put differently, romantic union and social or po-
litical power are the goals sought by protagonists in prototypical narratives.
The corresponding prototypes for sorrow are the death of the beloved and
the complete loss of social or political power, typically through social and
political exclusion, either within society (through imprisonment) or outside
of society (through exile).
The position of romantic love is clearest. Oatley points out that
“being in love is, in our society, the very paradigm of happiness” (370;
see also 364) – though I think Oatley slightly misphrases the matter,
which is more properly “being united with the beloved.” When the
lovers are separated, being in love is part of the paradigm for sorrow.
More significantly, J. L. Freedman “conducted a large questionnaire
survey of 100,000 Americans” and found that the one thing “respon-
dents most closely associated with happiness,” which is to say, the
one condition they most spontaneously identified as the prototype
case of happiness, is “love in marriage” (Oatley 361), which is to say,
enduring union with the beloved.
As to sorrow, M. A. Conway and D. A. Bekerian did research ask-
ing subjects to list “the sorts of situations in which a person might
typically experience an emotion” (154) – a listing that is likely to rely
on prototypes. They then translated the most frequent attributes into
sentence pairs for further research. They do not report the initial data.
However, they do provide the sentence pairs (189–91). For the group
“Grief, Misery, Sadness,” over half the resulting sentences describe
the death of a loved one or some other permanent separation from
a loved one, and the majority of these treat the death of a romantic
beloved.
As to artha, a quarter of Conway’s and Bekerian’s “Grief, Misery,
Sadness” sentences concern poverty or professional failure. A further
sentence pair concerns permanently leaving one’s home – in effect,
exile. Indeed, only one sentence pair does not fit into one or the other
category. That pair concerns bad weather. It is no doubt prototyp-
ical, and relevant to literary study. But it bears most obviously on
universal image patterns, as they derive from our lexical entries for
emotions, rather than on narrative structures.7 Only one-sixth of the
corresponding positive group treats professional success. However,
that is because five-sixths of the positive group concerns or suggests
some form of love, most often romantic love.8
7 In fact, it does fit the third prototype narrative structure, discussed in Chapter 6.
I have left this and related observations aside in order to preserve the sense of a
research program advancing in the course of the book itself.
8 Another relevant aspect of the Conway and Bekerian study concerns the group-
ing of emotion terms. Conway and Bekerian found that some emotion terms tend
to cluster together. For the most part, the clustering is definitional – thus, “terror,”
“fear,” “panic,” “anxiety,” and “apprehension” are linked. However, in two impor-
tant cases, the groupings are not simply definitional. One group includes not only
“joy,” “pleasure,” and “happiness,” but also “love.” Another links “hate,” “anger,”
and “jealousy” (152). It seems clear that there is an implicit narrative linking the
terms in both cases – the prototype narrative of romantic tragi-comedy, as we shall
discuss in a moment.
Work by Shaver, Schwartz, Kirson, and O’Connor has direct bear-

ing here as well. This work is particularly relevant to the present
study, for Shaver et al., treat prototypicality. First, they stress that
some emotions (for example, fear) are viewed as more prototypi-
cal than others (for example, boredom). Second, and more impor-
tant for our purposes, they note that particular emotion terms have
prototypical features. Though they do not fully develop the implica-
tions of the second idea, their research further supports the preceding
hypotheses. Specifically, Shaver et al., had subjects record personal
or “typical” accounts of particular emotions. A group of “coders”
then abstracted “features” from these stories. The researchers note
that they “were coding antecedents at a fairly abstract level” (1082).
In tabulating their lists for sadness and happiness, then, these writers
included a number of general ideas that do not bear on our definition
of prototypes. However, they included a number of directly relevant
points as well.
Their category for sadness includes eight features of eliciting con-
ditions. The remaining features all bear on expressive or actional
outcomes. Of the eight eliciting condition features, three are general –
though it is worth remarking that they quite clearly involve evaluat-
ing the final point of a narrative (for example, the first is “Undesirable
outcome” [1074]). In this way, they lend support to the larger, narra-
tive account of emotion. The more concrete or prototype-like eliciting
condition features are the following: “Death of a loved one,” “Loss
of relationship; separation,” “Rejection, exclusion, disapproval,” and
“Discovering one is powerless, helpless” (1074). The first and second
are part of the romantic prototype; the fourth is part of the social
dominance prototype; the third, though somewhat abstract, suggests
both – the romantic prototype through “rejection” and the social dom-
inance prototype through “exclusion” and “disapproval.” (I have left
out one final concrete feature, “Empathy with someone who is sad,
hurt” [1074]. Its presence is extremely important – indeed, crucial
for literature – but the other person’s sadness or hurt is itself pre-
sumably imagined in terms of “death of a loved one,” and so on.
Thus, its inclusion does not affect our understanding of the relevant
prototypes.)
The happiness category is also directly supportive of the preced-
ing claims. However, it is somewhat more affected by the coders’
preference for abstraction. Six of the nine eliciting condition fea-

tures are too general – though, again, they do indicate the narra-
tive structure of emotion (“Task success, achievement,” “Desirable
outcome; getting what was wanted,” “Getting something striven
for,” and so on). The more concrete features are the follow-
ing: “Receiving esteem, respect, praise,” “Being accepted, belong-
ing,” and “Receiving love, liking, affection” (1075). The first is
clearly part of the social power prototype. The second is too,
for the power sought by the protagonist in narratives focused
on artha need not be individual; indeed, both in fiction and in
life, it is most often a matter of collective domination, the power
of one’s in-group. In this way, in-group definition is crucial to
such stories – as indicated by the place of social isolation in
the prototype eliciting conditions for sorrow. (We shall return to
this point in subsequent chapters.) The third feature is clearly
part of the romantic prototype. Interestingly, the link between
happiness and power seems more obvious here than that between
happiness and love. Evidently, however, this was not the case with the
stories. Shaver and his colleagues emphasize the very close relation
between love and happiness in their study (see 1078–9), going so far
as to assert that “love may be conceptualized as a personalized form
of joy” (1079).
But why are there two prototypes here, rather than one? In order
to answer this, we must note, first, that prototypes are in part context-
dependent. Our prototype for “bird” is something along the lines of
a robin. But if we are out at sea and I say that there are birds on
the prow, you will be surprised if you look and discover robins.
Kahneman and Miller make the same general point when they dis-
cuss interpretations of the sentence “Jane owns a small dog”: “The
same statement will yield different norms of size and different ideas
of the size of Jane’s dog if she is known to live in a New York apart-
ment or on a farm in Maine” (140). The different norms here are a
matter of different prototypes for “dog” in these different contexts.
The two prototypes for happiness and sorrow are, like all prototypes,
context-dependent. The romantic prototype bears on personal rela-
tions, one’s private or familial life. The power or domination proto-
type bears on relations with society at large, one’s public or social
life.
final hypothesis: the universality of romantic and

heroic tragi-comedy and their derivation from
personal and social prototypes for happiness
The first and second hypotheses presented previously are, prima
facie, good candidates for universals. Indeed, given what we know
about the human mind, it would be very odd for those hypotheses
to be true, but culturally relative. In other words, they may be false.
But, if true, it seems extremely likely that they are true for everybody.
For example, it would be very strange were it to turn out that funda-
mental structural principles of emotion entries in the mental lexicon
are culturally variable. However, the third hypothesis seems, on the
face of it, less likely to be a universal. Perhaps these particular proto-
types apply to modern America or Europe, but – one might argue –
surely Africans, Indians, and Chinese think differently; surely, the
prototypes apply only to a narrow cultural and historical area.
But, in fact, this does not appear to be the case. Rather, it appears
that these are cross-culturally dominant prototypes for happiness
and sorrow. First, contrary to most people’s intuitions, semantic pro-
totypes are often more likely to be universal than are necessary and
sufficient conditions. Color terms provide a striking instance of this.
The boundaries of color terms vary considerably across languages,
while prototype cases are relatively constant, or vary only within
specifiable limits (see Comrie 36–8). Second, the predominance of
these prototypes for happiness is particularly clear in literary nar-
ratives. Indeed, it seems unlikely that there would have been any
extended debate on, for example, the universality of romantic love
(see Jankowiak and Fischer) had researchers paid more attention to
literary representations of emotion across a range of cultures.
In any case, this leads to my final hypothesis: Cross-culturally, there
are two prominent structures of literary narrative, romantic and heroic tragi-
comedy, derived respectively from the personal and social prototypes for
happiness.
As noted in the introduction, for a number of years, I have read
extensively the literatures of the major written traditions and a wide
range of oral traditions as well. The evidence for the broad recur-
rence of these structures seems to me virtually indisputable. Some
colleagues have asked me if this means that I have put forth a theory
that is not falsifiable. No, it does not mean that at all. In general, I
doubt that any complex theory is open to definitive refutation. How-
ever, this theory is no less falsifiable than others. It makes predictions
that can certainly run up against contradictory data. Indeed, I discuss
some data of just that sort in Chapters 4 and 6. My point is only that
reading in various traditions reveals these structures over and over in
such clear forms that I do not see how anyone could read the material
and not find these structures.
The following sections outline the main features of romantic and
heroic tragi-comedy and present some of the evidence from a range
of traditions. (As it is somewhat repetitive, I have placed most of this
evidence in notes.) In presenting this evidence, I have, of course, cited
stories that are either romantic or heroic tragi-comedies. These are,
after all, what show the cross-cultural recurrence of my two narra-
tive structures. As I have already stressed, most literary universals
do not occur in every work of literature, just as most linguistic uni-
versals do not occur in every sentence. For example, any universal
linguistic principle bearing on plurals will not apply to sentences
with all singular nouns. No one would cite sentences with only sin-
gular nouns in discussing a universal bearing on plurals. Similarly,
I do not cite examples of stories that are neither romantic nor heroic
tragi-comedies. Nonetheless, this mode of presentation may give rise
to some misunderstandings. I should deal with those briefly before
going on.
As I have already noted, my claim here is not that all narratives are
of these two forms. Narratives may be about anything. I can tell a story
about how the post office has repeatedly failed to forward important
bills to my summer address. This is neither romantic nor heroic. In-
deed, canonical works of literature are not necessarily romantic or
heroic. For example, Waiting for Godot does not seem to be heroic or
romantic in structure (though a reading that emphasizes its relation
to the French Resistance may make it heroic for a given reader and
a reading that emphasizes its homoerotic implications may make it
romantic for a given reader). It is important to note that in some
cases these structures may be present, but not easy to recognize. For
example, one might argue that Alain Robbe-Grillet’s postmodern La
Jalousie is a variation on the romantic plot. Indeed, I would contend
that many postmodern works take up romantic and heroic stories, but
conceal them through radical manipulations of discourse or presen-

tation (for example, through rearranging scenes, cutting explanatory
links, shifting perspectives, and so on.). Nonetheless, there are many
works that are neither heroic nor romantic tragi-comedies.
Once again, my contention is that heroic and romantic tragi-
comedy are prominent among highly esteemed narratives – and that
they almost exclusively account for what we consider prototypical
narratives. Put differently, given the innumerable possible structures
that count as stories, the likelihood of any given structure recurring
cross-culturally with any degree of prominence is about zero. Heroic
and romantic tragi-comedies not only recur across all or nearly all
traditions, they constitute a high percentage of canonical stories in
those traditions (probably well over fifty percent, though it is not
entirely clear how one would come up with a precise figure here,
or even if that is a very productive way of approaching the issue).
Moreover, they constitute virtually the entire field of what we con-
sider good examples of storiness. Waiting for Godot is a canonical work
that is neither heroic nor romantic. But it is also very far from being
a prototypical narrative.
In connection with this, I should also stress that my claim about
these universals is not merely statistical. My contention is that the
prominence of these prototype narrative structures is an absolute or
at least near absolute universal. Some readers have found this point
confusing, so it is worth developing for a moment. Once again, an
absolute universal is a universal that recurs across all traditions. If
heroic and romantic tragi-comedy are the prominent forms of canon-
ical and popular narrative in all traditions, then their prominence
constitutes an absolute universal – even though there are many sto-
ries in every tradition that are neither heroic nor romantic tragi-
comedies. Again, many stories in each tradition do not conform to
these structures. But that is simply irrelevant to their status as abso-
lute universals.
I hope these points serve to clarify my differences from some other
writers who have sought to isolate universal narrative patterns. In
addition to basing my conclusions on more systematic reading in un-
related literary traditions, I have put forth more qualified claims. As
will become clear in the following chapters, I do believe that heroic
and romantic structures are very widespread outside of canonical
literature and that they are deeply consequential for our thought
and behavior. However, I do not believe that I have discovered the
key to all stories. This is not a monomyth (or a duomyth). Indeed,
this is far from the only structure one might isolate even from these
particular stories. Like everything else, each story combines many
different structures. The romantic and heroic structures have un-
usual importance – for narrative and for human thought. But they
are not everything. Moreover, the importance they do have is pri-
marily the result of the prototypes for happiness. In fact, these proto-
types are arguably far more significant than the story structures they
generate.
romantic tragi-comedy
The most common plot structure across different traditions is almost
certainly romantic tragi-comedy, the story of the union, separation,
and ultimate reunion of lovers. Note that what has to be explained
here is not merely the fact that separation and reunion of lovers is a
common narrative theme, but that it is evidently the most universal
literary topic. In addition, the precise characterization of the separa-
tion in terms of death and other recurrent details have to be explained
as well. More exactly, as noted in Chapter 1, the standard structure of
romantic tragi-comedy involves two lovers who cannot be united due
to some conflict between their love and social structure, typically rep-
resented by parental disapproval. This conflict commonly involves a
rival as well, a suitor preferred by the interfering parents. The lovers
are separated, frequently through exile and imprisonment. This sep-
aration often involves death or imagery of death. In the end, they are
reunited, sometimes following a direct conflict with and defeat of the
rival. It may happen that the reunion of the lovers takes place only
in the afterlife.
We are dealing here with a highly specific, highly complex, and
very widespread literary universal. We need to account for its speci-
ficity, complexity, and extent. At the same time, we are dealing with
a universal that does not determine all works of literature. Again,
there are plenty of stories that do not follow this structure. Thus, we
need a theory powerful enough to explain the predominance of this
structure, but not so powerful as to predict (falsely) that all literary
works will be of this sort. Put differently, we need to explain the pro-
totypical nature of this structure, for we are clearly not dealing with
necessary and/or sufficient conditions for narrative here. This is pre-
cisely what the preceding hypotheses do. They predict that the union
of lovers – the prototype eliciting conditions for personal happiness –
will define the outcome goal for a predominant narrative structure,
with romantic love as the sustaining emotion.
But this is not all that can be explained by the preceding
hypotheses. Again, the structure of romantic tragi-comedy is quite
consistent across cultures and historical periods. It not only involves
the ultimate union of lovers, but their prior separation, a separation
closely associated with death – sometimes the lovers die and are
reunited only in death; sometimes there is a rumor of death or an
apparent death; sometimes there is a death and resurrection; some-
times there is a reunion in a divine world that suggests death; some-
times there is a near death; sometimes there is extensive imagistic and
metaphorical reference to death. Instances would include Charitōn’s
Chaereās and Kallirrhŏē, where both lovers are believed dead, though
one is abducted and the other is sold into slavery; Shakespeare’s The
Winter’s Tale; Bhāsa’s Vision of Vāsavadattā (in Woolner and Sarup),
where the heroine’s death is faked in order to fool her husband, so
that he will marry someone else; the Thai folk drama, Manohra (in
Brandon Traditional Asian Plays), in which Manohra is to be sacrificed
by her father-in-law, though she manages to escape and travel to a
heavenlike sanctuary; Chêng’s The Soul of Ch’ien-Nü Leaves Her Body,
where the heroine nearly dies due to separation from her beloved;
the Rāmāyan.a, where the heroine contemplates suicide in separation,
and so on.
We can account for these data if we extend our third hypothesis only
slightly. Clearly, the construction of a plot leading to the prototype
eliciting conditions for happiness must necessarily develop through
a period when those conditions do not obtain. This period will con-
stitute the Aristotelian “middle” or the “progression,” in the Sanskrit
terminology. That seems fairly obvious. Here is the addition to our
hypothesis: this middle or progression is prototypically assimilated to the
correlated prototype eliciting conditions for sorrow. Thus, when lovers are
separated temporarily, we tend to assimilate that junctural separa-
tion to a prototypical outcome separation; we tend to identify the
junctural sorrow that will end with the outcome sorrow that is the
result of death.
This is plausible for two reasons outside of the data on roman-
tic tragi-comedy (and parallel data on heroic tragi-comedy, as we
shall see). First, we have a general cognitive tendency to choose alter-
natives from within a lexically defined semantic field. The eliciting
conditions for personal happiness and personal sorrow are stored
in our lexicons in a way that makes one readily accessible from the
other. Thus, when forced to devise a scenario in which the happiness
prototype does not apply, we have the sadness prototype ready to
hand. Even if we do not take it up literally, its cognitive salience is
likely to foster the use of relevant imagery, metaphors, and so on – in
this case, images and metaphors of death.
A second reason we might expect the “middle” or “progression”
to develop in this way is that happiness is intensified by at least
some degree of preceding sorrow. As Frijda has noted (Emotions 323),
we become accustomed to happiness, so that the joy resulting from
joyful conditions tends, in our actual experience, to fade, to be-
come the norm. The contrast between the sorrowful progression and
the joyful conclusion serves to prevent this fading and to intensify
the final joy. Especially given the salience of the sorrow prototype,
it is unsurprising that storytellers would discover this and that the
practice would become well established in different traditions.
This analysis also allows us to give at least a preliminary account
of romantic tragedy. Though not nearly as common as romantic tragi-
comedy, it is still widespread – ranging from such European dramas
as Romeo and Juliet, to the love suicide plays of Japanese Kabuki. In
the context of the present theory, romantic tragedies are still orga-
nized by reference to the prototype goal of happiness in romantic
union, but they stop at the point of prototypical romantic sorrow –
most often, the death of one or both lovers. Tragedies are, then, trun-
cated tragi-comedies – unsurprisingly, as the prototype goal toward
which the entire narrative aims, the goal pursued by the protago-
nists, is necessarily the comic goal of romantic union. This is particu-
larly clear in a play such as Romeo and Juliet that, with slight changes
(for example, had Juliet awakened from her deathlike slumber only
a few moments earlier), would be a paradigm case of romantic
tragi-comedy.
This account – which we shall extend to heroic plots in the next

section – gains support from some otherwise surprising facts about
tragedy. First, tragedy in general (not merely romantic tragedy) is
considerably less common cross-culturally than comedy – a fact that
makes much more sense if tragedy is understood as a transforma-
tion of comedy than if the two are seen as alternative genres of the
same type. Moreover, many prime instances of tragedy are integrated
into larger cycles that are, ultimately, comic. Consider, for example,
Aeschylus’ Oresteia and Sophocles’ Oedipus plays, or, to take a mod-
ern example, the two parts of Goethe’s Faust. Along the same lines,
the performance of Japanese Nō and Bunraku dramas was struc-
tured so that tragic plays came in the middle of the day, with comic
plays following, so that “all programs end happily” (Gerstle 60). All
of this suggests that comedy is, indeed, the primary form, and tragedy
operates as a shortened version of comedy.
In both romanic tragi-comedy and romantic tragedy, there is an-
other recurrent motif in the narrative middle or progression that fits
this analysis as well, if more complexly. In addition to the suggestion
or imagery of death, it often happens that one of the lovers is exiled or
imprisoned or threatened with exile or imprisonment. This motif is
found in works ranging from Romeo and Juliet to the Rāmāyan.a, where
Sı̄tā is imprisoned in Laṅka and, later, exiled from Ayodhyā; Niz.āmı̄’s
Laylā and Majnūn, where Majnūn is (self-)exiled; Chêng Teh-hui’s
The Soul of Ch’ien-Nü Leaves Her Body, where Wang Wên-Chü has to
spend three years away in the capital; and Ma Chih-yüan’s Autumn
in Han Palace, where Wang Chao-Chün is sent away from her beloved
Emperor Yüan to the court of Emperor Huhanya. Sometimes, there
is imagery of this sort, even when there is no literal exile or impris-
onment. Thus, for example, Niz.āmı̄ makes repeated use of prison
imagery to characterize the condition of both his hero and heroine.
This appears to be a matter of interference from the prototype cases
of sorrow for social life – exile and imprisonment. It is unsurpris-
ing that prototypes from different life-contexts should retain some
salience and have some consequences for perception, expectation,
narrative imagination, and so on, outside their usual domains.
Finally, as we have already suggested, the ultimate reunion of the
lovers is often not a simple repetition of the initial union. There is,
in most cases, some sacred or otherwise absolutized aspect to the
reunion. The clearest case of this is when the lovers are joined only
after death, in heaven. In other cases, there is imagery of heaven or
of some parallel condition. In part, this follows from the introduction
of the sorrow prototype in the narrative middle. But it is also the
fullest possible development of the prototypically enduring character
of happiness – just the sort of contrast effect we would expect, given
a prototype-based account.9, 10
9 An excellent example of the general structure may be found in Kālidāsa’s

Abhijñānaśākuntalam, generally considered the greatest masterpiece of Sanskrit
drama. Dus.yanta and Śakuntalā fall in love. They are united, then separated. Their
separation is not due to a social disapproval of their union. In fact, Śakuntalā’s father
entirely approves of the marriage. However, it is due to another sort of conflict with
society – a dereliction of social duties that results from their mutual infatuation. Due
to his affection for Śakuntalā, Dus.yanta does not adequately fulfill one of his obli-
gations toward his mother. More significantly, distracted by thoughts of Dus.yanta,
Śakuntalā fails to give appropriate honor to a religious visitor. He curses her, and the
curse leads to her separation from Dus.yanta. Śakuntalā is in effect exiled to another
world, in what is clearly a deathlike scenario, while Dus.yanta suffers misery in his
palace – with associated imagery of death (for example, spring flowers do not blos-
som [Kālidāsa, Theater 149]). Eventually, they are reunited in a heavenly hermitage.
In fact, it is well known that a romantic structure of this sort is pervasive in Sanskrit
drama. Other clear instances include Śūdraka’s Little Clay Cart, Bhāsa’s Dream of
Vāsavadattā, Hars.adeva’s Ratnāvalı̄, and Bhavabhūti’s Uttararāmacarita.
Though it is a very complex, and very long work, A Dream of Red Mansions man-
ifests much the same structure. A Dream is generally considered “the greatest novel
in the Chinese literary tradition” (Hsia, A Dream 262–3). What may be called its main
plot line concerns a young aristocrat named Magic Jade. He deeply loves a cousin
named Black Jade. But Black Jade is sickly, and from a less-affluent branch of the
family. His family disapproves of the relationship, favoring instead a marriage with
Precious Hairpin, another relative, a healthier girl, from one of the most powerful
families in the area, the Xues (vol. 1, 111–12). In addition, Black Jade’s father intends
to marry her to someone other than Magic Jade. Black Jade appeals to Magic Jade’s
family against her father’s decision, but, unsurprisingly, she has no success. When
she hears that Magic Jade is going to be married to someone else, she resolves to die.
Her health, always weak, slowly fails. Magic Jade begins to show signs of mental
derangement. Magic Jade’s health improves only when he mistakenly believes that
he is to marry Black Jade. Black Jade, aware of the actual marriage arrangements, dies
at “the very moment” Magic Jade unknowingly weds Precious Hairpin (vol. 4, 375).
Magic Jade goes through the ceremony with his veiled bride. In effect, Magic Jade
believes he is in a romantic tragi-comedy. He believes that, after familial conflict, and
near death experiences, the lovers have made it through the ordeal of mental and
physical sickness, and will now be reunited. When he discovers that he has married
Precious Hairpin, he too begins to waste away, approaching death. Unaware of Black
Jade’s fate, he asks to be moved to her room so that they will die together (vol. 4,
370). Here, too, Magic Jade imagines a romantic tragi-comedy, for he imagines that
he and his beloved will at last be joined in death – and thus, one may assume, in life
after death.
Footnote no. 9 (continued)
Eventually, Magic Jade does recover from his illness. But, once recovered, he is not
the same. He joins a wandering monk and renounces the world. Hsia contends that
“we should perhaps feel happy that he has finally gained wisdom and leaves this
world of suffering. . . . But we cannot help feeling that his spiritual wisdom is gained
at the expense of his most endearing trait – his active love and compassion.” In Hsia’s
view, Magic Jade is “the most tragic hero in all Chinese literature” (270). However,
I do not believe that this is a tragic ending. Indeed, I do not believe that it is tragic
even from the perspective of the love story. The ending is a version of the standard,
spiritualized reunion of the lovers, even if this is only implicit. Specifically, in his as-
sessment of the novel’s end (in Chapters 119 and 120), Hsia leaves out an important
point from Chapter 98. When he learns that Black Jade has died, Magic Jade begins
“howling” in sorrow until he passes out (vol. 4, 371). In his unconscious state, he
finds himself on the road to the Nether World, though it is not yet time for him to
die. A “stranger” on the road tells him that he must return to life, explaining that “if
you really want to find [Black Jade] you must cultivate your mind and strengthen
your spiritual nature. Then one day you will see her again” (vol. 4, 372). After this,
whenever he was tempted by thoughts of suicide, Magic Jade “remembered the
words of the stranger” (vol. 4, 374). Magic Jade’s final attainment of spiritual release
is, in part, a fulfillment of the stranger’s prophecy, a spiritualized reunion with Black
Jade. Of course, this may be a triviality, for perhaps the reunion is merely the same
reunion he experiences with all life. But the spiritualized love is not undermined
for being more spiritual. At the end of the novel, Zhen Shi-yin says “If the Fairy
Flower [that is, Black Jade] regained its true primordial state, then surely the Magic
Stone [that is, Magic Jade] should do likewise.” When Magic Jade’s “worldly karma
was complete” – when the effects of his past acts no longer bound him to the world,
his “substance had returned to the Great Unity” (vol. 5, 371). This is, in effect, an
absolutization of the romantic reunion.
The structure is quite common in Japan as well. Hanakatami, a Nō drama by
Kan’ami, revised by Zeami, concerns a prince and his beloved who are separated
when the prince becomes emperor and leaves for the capital. They are reunited when
they meet by chance and the emperor recognizes a flower basket that he gave to the
woman upon his departure (see Waley 263–5). The Reed Cutter, by Zeami, concerns
two lovers who are separated because of the man’s poverty. The wife travels to the
city to seek employment. Having succeeded in gaining a position, she returns to the
village to find her husband. After some slight complications, they are reunited and
“return to the city” together (Keene, Twenty Plays 162). These plays involve love,
separation, exile, and reunion. In both cases, the reason for the exile is unusual, but
nonetheless a variation on the standard structure. Specifically, there is no representa-
tive of society who forbids the union of the lovers. However, in each case the lovers
are separated due to some disruptive social condition. In the case of Hanakatami, the
lover is assuming a social role that inhibits his individual freedom. His role as em-
peror circumscribes his personal choice and action. In The Reed Cutter, it is poverty
that separates the lovers. Indeed, this play is remarkable not only for presenting
poverty as a cruel and unjustifiable impediment to love. It is equally remarkable for
making the woman the active figure who goes away to seek success before saving
her husband.
An in some ways, more standard narrative may be found in Zeami’s Lady Han.
In this story, a young man and a prostitute fall in love. The man leaves, promising
that he will return. The woman’s proprietress, unhappy to find the heroine mooning
over one client, dismisses her and sends her away. The lovers are eventually re-
united. Clearly, the proprietress is the interfering social figure here. But again there
is a difference, for in this case the objection to the lovers is that their love is decreasing
economic productivity. The woman is not making money for the proprietress. Here,
as elsewhere in Zeami, the conflict between the lovers and society or the represen-
tative of society is very practical. Of course, the point holds in other traditions as
well, where the parents’ choice for their child’s mate is usually someone of greater
wealth or social prestige. In these last two cases from Zeami, however, the systematic
injustice of social structure is perhaps more obvious than usual.
Of course, romantic plots in Japan are not at all confined to work by Zeami. Con-
sider, for example, the eighteenth-century play, Love Letter from the Licensed Quarter.
This work represents the relationship of Izaemon and Yūgiri, a relationship seen as
“typical” for its genre of Kabuki drama (Brandon Kabuki 6). Izaemon, a wealthy
young man, and Yūgiri, a prostitute, have fallen in love. However, Izaemon has
been disinherited by his family. Thus, he must leave the licensed quarter. Here, the
parental/social interference is presented in a distanced, but nonetheless effective
manner. The moment of exile is clear as well. So is the imagery of death, for Yūgiri
falls ill due to the separation (226), and even feels close to death (231). Izaemon re-
turns to visit Yūgiri. However, believing that she has been unfaithful, and has been
enjoying the company of a wealthy samurai, he spurns her when they meet. She
convinces him that he has no rival. They are reconciled and, by good fortune, his
family decides to reinstate him, sending him chests full of money, enough to pay his
debt and to purchase Yūgiri’s freedom (236).
A tragic version of the structure may be found in Chikamatsu’s Love Suicides
in the Women’s Temple. Kumenosuke is to be a priest and Oume is to be married to
Sakuemon. However, Kumenosuke and Sakuemon fall in love. The High Priest of the
temple denounces Kumenosuke and evicts him when he learns of the affair. Oume’s
parents push ahead with the marriage to Sakuemon, even when they learn of the
love between Kume and Oume. Oume’s father directly rejects a suggestion that this
couple marry (Chikamatsu, Major Plays 146). The lovers escape together and take
refuge in a women’s temple where, by coincidence, Kume’s sister has brought the
ashes of their recently deceased father. Kume conceals his identity. In the temple,
thus in a spiritually elevated place (reminiscent of the final reunion of lovers in full
comic versions), Kume and Oume realize that they have no place to go. Calling on
the Buddha to redeem them, they commit suicide. However, before this, Oume re-
calls the love she shared with her mother and Kume honors his father’s ashes. They
pray: “We shall be reborn on one lotus with our parents” (159). A reconciliation with
parents is often part of the comic conclusion in a romantic tragi-comedy. This prayer
for posthumous reconciliation in a tragic play is one of those striking variations on
standard structures that makes for a great and distinctive work.
Turning to the Persian and Arabic traditions, we have already noted the relevance
of Laylā and Majnūn. In this story, Laylā and Majnūn fall in love, but Laylā’s father
refuses to allow them to wed. Rather, he marries Laylā to another man. Majnūn
goes into self-imposed exile in the desert, while Laylā remains a sort of prisoner
in her home. (Though, in an unusual variation, Niz.āmı̄ presents Laylā’s husband
as a relatively admirable man who does not force himself on Laylā, but suffers the
grief of unrequited love, eventually dying from this illness [see 164].) In the end,
they both die and are finally united in heaven. An angel explains that “in the fabled
garden . . . they suffer grief no more. So it will be until eternity” (176).
Needless to say, this is not the only romantic tragi-comedy in the Arabic and
Persian tradition; indeed, Niz.āmı̄’s is not the only version of this particular story,
which has been widely rewritten (see Gelpke xi on the literally hundreds of versions
of Laylā and Majnūn). Arabic instances of the romantic plot would include such sto-
ries as the final voyage of Sindbād and the tale of Aladdin, from The Thousand and One
Nights. Restless Sindbād sets out on a journey where he marries and falls in love with
a beautiful young woman (the marriage precedes the love in this case). He settles
in the woman’s country, only to find that all the people there are demons – literally.
At one point, he praises Allah and is therefore abandoned “on the top of a high
mountain” (Dawood, Tales 160). With the help of God, he manages to make his way
back home and his wife is “overjoyed” (161). She explains that the people are all
devils and that her father had come there from a foreign land. They flee the satanic
community and return to Sindbād’s home. The story follows the romantic structure
point for point – romantic union, followed by separation, a sort of exile, and imagery
of death (the transportation by demons up into the sky). Moreover, the separation
is due to a conflict between the lovers and the larger community. The lovers are
reunited, and the reunion has the usual spiritual component. One thing that is re-
markable in this story is the treatment of society. The society that opposes romantic
love is almost always criticized in romantic plots, but it is rare for that society to be
demonized literally. On the other hand, this characterization is a simple extension of
the standard spiritualization of the lovers’ final union. If the final union of the lovers
is godly, then it would seem that their separation was demonic. “The Last Voyage of
Sindbad the Sailor” merely takes up this clear structural implication.
In the story of Aladdin, our hero falls in love with the beautiful Princess Badr-
al-Budur. His mother claims that they cannot marry because of the class difference:
“Your father was the poorest tailor in this city” (185). Aladdin insists that he will die
if they are not united. Through magic, he convinces the Sultan to agree to the mar-
riage. However, the Sultan later marries the princess to another man – thus acting
as the forbidding father and introducing the rival. Through magic, again, Aladdin
manages to separate the newlyweds, ultimately ending the marriage and wedding
the princess. The new couple is blissfully happy, but then another interfering char-
acter enters, a second rival, himself a magician. This magician abducts the princess,
taking her to Africa. When she is abducted, the Sultan threatens Aladdin with death
(thus introducing the standard elements of exile, imprisonment, and death). Aladdin
defeats this new rival, and returns home with the princess.
I take it that the ubiquity of this structure in European literature does not require
demonstration. It is the standard structure of New Comedy and its later progeny
(such as Shakespearean comedy), common in prose romances, such as Charitōn’s
Chaereās and Callirrhŏē, as already mentioned, and a staple of popular cinema. Though
it ends tragically, Romeo and Juliet, provides a very clear example. It begins with the
heroic tragi-comedy
Heroic tragi-comedy also exhibits a surprisingly specific cross-
cultural pattern, though there is somewhat greater variation in this
structure than in that of romantic tragi-comedy. The fullest version
begins when the rightful leader of a society is displaced from rule or
prevented from assuming rule, most often by a close relative. He/she
lovers separated due to the parents, who prefer a rival. The separation leads to the
exile of Romeo and the virtual imprisonment of Juliet. This is followed by the death-
like state of Juliet. In a standard tragi-comedy, Juliet would have awakened from
her deathlike sleep to be reunited with Romeo. However, Shakespeare delays her
waking just long enough to make it a tragedy rather than a comedy, though even
here the lovers are in effect united in death.
Romantic tragi-comedy is not confined to written traditions either. For instance,
many of the Native American tales recounted by Lévi-Strauss follow this pattern.
Lévi-Strauss’s “key myth” (The Raw and the Cooked 35–7), a Bororo story, has this
general structure, if in a peculiar form (a form that may suggest a psychoanalytic
provenance for the genre). The hero has sexual relations with his mother. Unsurpris-
ingly, his father disapproves of this union. He tries to kill the son, and eventually
succeeds in driving him away so that he must live in exile for some time. The boy
does eventually return and kill his father. With only slight changes, we have a very
standard romantic plot. The main difference is that the interfering father is fused
with the rival – often a figure whom the hero must overcome or even kill. In other
words, the two most common blocking characters are combined. This fits the social
conflict particularly well, for what could define a greater conflict with society than
a violation of the central rule of marriage – the prohibition on incest? It is particu-
larly noteworthy that the story, at least in Lévi-Strauss’s version, develops sympathy
with the boy. This is entirely in keeping with the tendency of the genre to support
the lovers over society in romantic conflict. However, one might have expected the
outcome to be different in a case where the social precept being violated is so central
to social structure.
Though many of Lévi-Strauss’s myths fit the pattern, a particularly touching in-
stance is from the Arawak. In this case, a Jaguar assumes human form and becomes
the wife of a hunter. Though obviously not realistic, the point is that the woman and
the man come from groups that are not supposed to intermarry. In general, the social
conflict that keeps the lovers apart is the result of some social taboo on marriage
across a dividing line – class, for example, or race, caste, region, or nationality. The
marriage of a man and a jaguar can serve to represent any taboo marriage of this sort.
The couple lives happily, for “She turned out to be an exceedingly good wife” (From
Honey to Ashes 257). One day, in keeping with her role as a good wife, she suggests
that they visit her in-laws, but on the condition that her husband not reveal her true
identity. The husband does tell his mother, who informs the village. Though Lévi-
Strauss does not indicate that the community took any action against the wife, she
feels “so ashamed” in front of the group that she flees back into the forest, returning
to her life as a jaguar. The story ends during the period of separation and exile: “Her
poor husband searched the bush in vain, calling out his wife’s name, but there never,
never came any reply” (257).
is exiled or imprisoned. This exile or imprisonment is linked with

death – imagery of death, the threat of death, and so on. While he/she
is in exile or imprisoned, the kingdom is threatened by some outside
force, typically a (demonized/bestial) invading army or, less often, a
demonic beast. The hero defeats the threat to the kingdom. He/she
then battles the usurper, and is restored to his/her proper place as
leader of his/her society.
It should be clear immediately that this pattern is precisely what
one would expect from our prototype hypotheses, and the corol-
laries introduced in connection with romantic tragi-comedy. It is a
plot based on achieving the prototype eliciting condition for social
happiness – social and political power. Moreover, the relevant pro-
totype eliciting conditions for sorrow define the narrative middle or
progression. These prototype conditions are, of course, complete loss
of social power, typically through exile or imprisonment, or through
the absolutization of such loss in death – primarily the death of the
hero, but also the death of the nation (threatened by invasion). The
complexity of this structure may appear baffling at first. But it is, in
fact, readily explained by the preceding hypotheses. Specifically, the
complexity of the heroic plot mirrors the complexity of the social pro-
totype for happiness. This prototype encompasses both individual
10 As I emphasized in the introduction, the relation between the study of univer-

sals and the study of culture should be seen as complementary, not contradictory.
In keeping with this, writers in culture studies have done very valuable analyses
of emotional particularity both in and outside literature. The studies in Lutz and
Abu-Lughod are a case in point. Yet, many of these studies appear to assume a
universality/particularity dichotomy. Abu-Lughod’s essay on the radical particu-
larity of Bedouin love poetry is especially relevant in the present context. She be-
gins by strongly supporting efforts “to deconstruct the concept of emotion through
showing its specific cultural meaning” (25) and urges researchers to undertake “the
important task of relativizing” (26). She then focuses on one particular example, pre-
sumably an example she takes to demonstrate profound cultural difference: Fathalla
and a young woman were in love, but “[T]he young woman’s father decided to
refuse to give his daughter to the young man. In despair . . . the man set off for
Libya. . . . Some time afterward, the girl’s father arranged to marry his daughter
to someone else. When Fathalla heard the news, he composed and recorded these
poems and sent the tape to the girl’s brother.” When she heard the poems, “Fifteen
days after the wedding,” the young woman “gasped for air, fainted, and then fell
over dead.” Abu-Lughod concludes, “This story tells us a great deal about the pol-
itics of emotion discourse in Bedouin society” (29–30). It does, as Abu-Lughod’s
analysis shows. But it does this in a way that his far from radically particular-
ist. It does so by repeating the most cross-culturally common plot structure in the
world.
and in-group domination. The usurpation story takes up the individ-

ual component. An individual aspires toward (deserved or rightful)
dominance in society. He/she is prevented from achieving that dom-
inance, and is even exiled from the society. However, he/she eventu-
ally succeeds. The external threat to the kingdom treats the in-group
component. The entire society aspires to (deserved or rightful) dom-
inance over other societies. This dominance is threatened by another
group, but ultimately the challenge is defeated. The double structure
of the heroic plot is in effect necessitated by the double structure of
the social prototype for happiness.11
11 A particularly clear example of this structure may be found in the Rāmāyan.a, where
Rāma is exiled from Ayodhyā due to a dispute over succession relating to his brother.
While in exile, he enters into conflict with the demonic ruler of Laṅka, Rāvān.a, who
poses a threat to all society. (Indeed, this threat is the reason that the god Vis.n.u be-
came incarnate as Rāma to begin with.) He defeats Rāvan.a in battle, saving not only
Ayodhyā, but the entire world. He then returns to Ayodhyā in triumph, to assume
his rightful rightful place as king. As his brother refused to accept the kingship
in Rāma’s absence, there is no need for a battle between them. The Mahābhārata,
too, involves a structure of this general type, if less perfectly. There, too, a familial
conflict leads to the exile of the true rulers, the Pān.d.avas, who also have to save
society in battle with soldiers or supernatural creatures. However, these battles do
not serve to reestablish them in their rightful place. That occurs only after a civil war
with their usurping cousins. Some Sanskrit drama follows this pattern as well, most
obviously works drawn from the epics, such as Śaktibhadra’s Āścaryacūd.aman.i, a
revision of the Rāmāyan.a. Indeed, variations on this structure are to be found in the
many regional, caste-based, and other versions of the Rāmāyan.a found throughout
south and southeast Asia (on the variety of Rāmāyan.as, see the essays in Richman
Many Rāmāyan.as).
The Tamil epic, Shilappadikaram (Cilappatikāram), attributed Ilangô Adigal, is an
interesting instance of the patterns we have been discussing. (Tamil is a south Indian
language, unrelated to Sanskrit.) The first and second parts are romantic, while the
third is heroic. In this part, Shenguttuvan, a great monarch, learns of a new goddess
and vows to have a suitable image made for her temple. In order to do this, he
has to defeat all kingdoms that oppose his project. The threat/defense sequence is
manifest here (even if the threat is rather limited and provoked by the ruler himself).
Moreover, the story clearly presents Shenguttuvan’s narrative-defining goal as the
social prototype for happiness. Though the overt purpose of the king’s action is
honor the goddess, the actual motivation – gaining social domination for both the
individual and the group – is not at all concealed. As the chief minister puts it, “No
one can stop you [Shenguttuvan] if you choose to impose Tamil rule over the whole
sea-encircled world” (160).
The Malian Epic of Son-Jara – a widely popular and frequently retold epic, with
versions across west Africa – manifests the heroic structure in a robust form. Son-
Jara is prevented from acceding to the throne, and his half-brother becomes monarch
instead. While Son-Jara wanders, homeless, seeking refuge and facing the danger
of death, a neighboring king, Sumamuru, threatens his homeland. Son-Jara defeats
Sumamuru, and becomes king.
But the heroic structure is not only complex, it is also variable.

There are two particularly common types of variation. The first is
shared with romantic tragi-comedy – truncation. As with romantic
tragi-comedy, heroic tragi-comedy may be shortened into a tragedy,
with the deposed ruler killed, or exiled without hope of return. More
significantly, perhaps, the heroic structure may be and often is short-
ened without becoming tragedy, for it is more complex than the
romantic structure; there are more elements that may be cut. Some
diminished versions do not delete the ending of the full structure, but
rather delete one of the two component substructures. Thus, many
versions of the heroic tragi-comedy involve the hero’s exile and sub-
sequent defeat of the usurper, but do not include the threat to the
The Nyanga epic of Mwindo varies the structure slightly. Mwindo is prevented
from assuming the chieftancy by his father. Indeed, his father makes numerous at-
tempts to kill him. This leads to a period of exile in which Mwindo is given support
by his maternal aunt, Lyangura. Ultimately, he defeats his father and becomes chief.
It is only after this that a supernatural menace threatens his people – a dragon that
he defeats.
As these examples suggest, the heroic plot is frequently expressed in epics,
especially epics of particular social or political significance. In Japanese literature,
the work that comes closest to national epic is The Tale of the Heike. This work devel-
ops a broad version of the plot, with the exiled Genji ousting the Taira warriors who
had usurped rule. The Taira had usurped rule overtly by forcing Emperor Takakura
to abdicate in favor of Emperor Antoku. More importantly, they had usurped rule
implicitly by taking over the operation of government. The Taira clan head does
“whatever he pleases . . . because the court has lost its authority” (43). Thus “Ranks
and offices were not conferred at the discretion of the retired and reigning sovereigns
in those days, nor yet by decision of the Regent, but solely as the Heike [that
is, Taira] saw fit” (46). It also recapitulates the heroic structure in more specific
stories.
An interesting case is the rebellion plotted by Narichika. The rebellion petered
out before it ever began: Narichika “was holding secret consultations and making
preparations of various kinds, but they were mere surface activity; there seemed
little chance of their leading to a successful revolt” (62). The Taira clan head discov-
ers the rebellion and the conspirators are killed. Narichika is killed in a way that
is particularly “cruel” (84). This is, in a sense, a threat/defense sequence, thus the
social part of the heroic plot. However, the threat is to the Taira, who are themselves
usurpers, and the threat is not much of a threat either. Moreover, the discovery
of the plot almost leads the Taira to attack the Retired Emperor – itself a classic
threat scenario. However, this, too, peters out before anything happens. In this way,
the Narichika section repeats the threat/defense sequence, but with an interesting
ambiguity regarding point of view. What is defense from one perspective is usurpa-
tion from another. As we shall see in the next chapter, this is not the only point of
ambiguity and ambivalence in the Heike.
kingdom. Others do not include a usurper, but do include a threat to

the kingdom.
The second common type of variation on this structure concerns
not the inclusion or exclusion of narrative sequences, but the precise
identity of the hero. I have been speaking as if the hero were invariably
the ruler. In many cases, he/she is the ruler. But not always. Some-
times, the hero is a loyal retainer of the ruler. One can see how this
sort of shift could result from a number of different factors. Taboos
surrounding the treatment of monarchs could easily lead to a focus
on retainers. At the same time, ordinary readers might find it easier to
identify with a retainer than with a hereditary ruler, easier to imagine
themselves in the place of the former than in the place of the latter.
Moreover, this substitution is possible in heroic tragi-comedy because
the retainer typically does achieve the same prototypical social goal as
his/her ruler, if to a lesser degree. The retainer’s power and authority
are a function of the power and authority of the ruler. In other cases,
the hero changes in the course of the story. An initial hero dies and is
replaced by a second hero, usually his/her child or grandchild. For
example, a story may begin with a deposed ruler, but conclude with a
descendant of the ruler who regains the kingdom. Neither variation
significantly affects the overall structural pattern.12, 13
12 Having noted these variations, we can point out some further instances of this struc-
ture, giving a clearer picture of its cross-cultural breadth. We have already cited three
prominent instances from India and two from Africa, all five being epics. We may
augment our single, epiclike example from Japan, The Tale of the Heike, by noting that
much Japanese drama fits this structure well, though often with the substitution of
a retainer for the king. For example, Chikamatsu Monzaemon’s most popular work
(Keene, Major Plays 195), The Battles of Coxinga, concerns an exiled minister who
returns with his son to defeat the Tartar invaders and reestablish his monarch. As
James Brandon explains, “In a very typical oiemono [history] play the young heir
to a feudal house is deposed by a scheming clan retainer or relative and forced to
escape in the disguise of a commoner. . . . At the end of the play, the evil retainer is
defeated by retainers loyal to the house and the young heir is reinstated” (9). More
generally, the broad tendencies of kabuki drama fit our hypothesis well, for in the
nineteenth century, it was commonplace among Japanese commentators that the
two main centers of play writing were associated with plot structures of the two
sorts we have been discussing – Kyoto plays being romantic; Edo plays being heroic
(see Brandon, Kabuki 13).
There are other prominent African examples too. The Ijō epic of Ozidi (see
Ōkabou) treats the murder of Ozidi’s father. It recounts Ozidi’s birth in exile, his
return to his homeland, and his defeat of his father’s murderers. The Bambara epic
about Da Monzon of Segou also involves many episodes of this sort, some treating
Da Monzon directly, and others treating the heroic warrior Bakari Dian. The latter
are particularly relevant, as he is entirely socially isolated when he goes to defend
Segou (see Sissoko III 4).
In the Middle East, Ferdowsi’s Persian epic, the Shâhnâme – “considered over the
centuries the greatest work of Persian literature and the strongest pillar of Persian
identity” (Yarshater, “Foreword” xv) – is also a case in point, manifesting a version
of the heroic pattern in many of its sections. (The work covers the entire period from
the creation of the world to the last pre-Islamic ruling dynasty. It cannot be consid-
ered a single narrative in the relevant sense. It is, rather, a series of narratives.) After
some prefatory material, the poem begins with the exemplary reign of Keyumars.
His son must battle the threatening Black Demon. The son is killed, but Keyumars’s
grandson, Hūshang, defeats the demon and becomes king. Though the throne is
not lost in this sequence, the rightful heir is killed, and the injustice of this is undone
in the next generation through the defeat of the threatening enemy. This is not only
an inaugural narrative in the story, but part of the beginning of civilization. The
next generation repeats the threat/defense sequence, when Shah Tahmuras defeats
the demons and is named “Demon-binder” for his success (9). These two conflicts
prepare for a third and fuller narrative in which the Arab Zah.h.āk, working with
lblis (or Satan), manages to take control of Iran. Jamshı̄d, the Shah, flees to “the sea
of China” – a standard instance of dethroning and exile. In line with the common
variation in which the restoration is deferred to the next generation, Farı̄dūn inher-
its the divine right to rule. Zah.h.āk searches for Farı̄dūn, whose mother flees with
him to the Alborz mountains, in another instance where the rightful ruler is denied
and sent into exile. Farı̄dūn eventually defeats Zah.h.āk and restores rightful rule to
Iran.
A further story introduces the element of familial conflict, completing the pro-
totypical structure. Farı̄dūn divides the world into three kingdoms, awarding one
to each of his three sons. Iraj is given Iran. In consequence, his two brothers become
jealous and kill him, after the threat of exile arises (32). The retribution is reserved for
Iraj’s grandson, Manuchehr. Years pass until the brothers send their armies against
Iran. This threat to the kingdom inspires Manuchehr, who defeats the invaders,
eventually assuming the throne.
The triumph of Manuchehr introduces the character of Sām, and thus a sequence
of stories related to warrior heroes rather than the monarch per se. In the first of
these, Sām rejects his son, Zāl, denying him his rightful place, and sends him into
exile (with danger of death). They are eventually reconciled, and Zāl himself has
a son, Rostám, probably the single most important hero in the poem. He enters
actively into the plot in defending Iran against another invasion and rescuing the
abducted Shah Kay Kavús.
Here as elsewhere there is no point in going through every story in the
book. However, it is worth considering the single most renowned story of the
Shâhnâme – the episode of Rostám and Sohráb. Sohráb is Rostám’s son. However,
he is living with his mother in Turán. He gathers an army together and invades Iran.
There is a conflict between the Shah and Rostám in which the Shah threatens to kill
Rostám, thus denying Rostám his rightful position. Rostám leaves briefly, in a sort
of mini-exile, swearing that he has cut off his relations with the Shah. The two are
reconciled as the threat to Iran is imminent. Rostám returns and, after some diffi-
culty, defeats the invading army. At the climax, he kills Sohráb, learning his identity
after the fact. Unsurprisingly, the result of all this is not the joyful celebration of
triumph one might otherwise expect. As we will see in the following chapter, this
sort of ambivalence is not uncommon in heroic plots, for specifiable reasons. In any
case, other than that final sorrow, the story is clearly a variation on the structure we
have been considering.
The ta’ziyeh of Iran comprises a set of dramatic reenactments of the massacre
of al-H. usayn and his followers at Karbalā and the events surrounding that mas-
sacre (see Chelkowski). The basic story of this paradigm narrative is of just the
sort we have been discussing. The leadership of the Muslim community has been
illegitimately taken from ‘Alı̄. Moreover, both ‘Alı̄ and his son al-H . asan have been
killed. Now al-H . usayn remains, and al-H . asan’s son Qasem. The narrative tells of
the battles of Qasem and al-H . usayn to defeat the usurper and restore the rightful
line of Caliphs. This story is a tragedy, and thus culminates in the final defeat of
al-H. usayn at Karbalā – though, of course, Shı̄’ah Muslims believe that, in the end,
they will triumph. Moreover, as often happens, episodes of the ta’ziyeh frequently
incorporate the romantic structure as well. Thus, Qasem is married to al-H . usayn’s
daughter, but they are separated on their wedding night as Qasem must go to battle
the armies of the usurper. He is defeated and dies after explaining to his beloved
that “our marital union shall occur at the Judgment Day” (Humayuni 14).
Another relevant Persian work is Gurgānı̄’s eleventh-century poem Vı̄s and
Rāmı̄n. The story of two lovers, Vı̄s and Rāmı̄n combines the two genres in an unusual
way. The narrative is a complicated one in which two brothers, Moubad and Rāmı̄n,
vie for the love of Vı̄s. Vı̄s is awarded to Moubad by her mother, but she falls in love
with Rāmı̄n. Moubad discovers their intrigue and exiles Vı̄s. As this standard ro-
mantic conflict is proceeding, there is suddenly a threat to the kingdom from Rome.
One might expect Rāmı̄n to repulse the invading army. However, Moubad defeats
this threat, having imprisoned Vı̄s while he was away at war. The lovers neverthe-
less manage to meet. After further complications, the lovers are estranged, Rāmı̄n
nearly dies, and finally they are again reconciled. At this point, the lovers plot a
coup against Moubad. They wrest control of the kingdom. As Morrison explains,
“Rāmı̄n is left supreme and is welcomed . . . as a liberator” (xvi). When the lovers die,
they are reunited in heaven (351). The romantic part of the story, though complex,
is perfectly in keeping with the usual structure. The heroic part, however, is very
strange. It includes both the usurpation by a family member (a younger brother, in
this case) and the defense of the society against an external threat. In other words,
it includes both parts of the full heroic structure. However, the defense against a
threat occurs first and is undertaken by Moubad, whereas the usurpation occurs
last and is accepted as a just outcome – a particularly striking and unusual, even
disconcerting, variation on the structure.
Arabic epics often fit here as well. Lichtenstadter summarizes their standard
format in a way that shows their conformity to the general pattern – “the out-
break of hostilities, the preparations made by the attacker, the fearful anticipation
of the intended victims,” and so on (30). A good example of the full pattern may be
found in the Sı̄rat ‘Antar or Life of ’Antar. ‘Antar is deprived of his rightful place in
society by his father, who treats him as a slave. Conflict with his father leads to his
exile. ‘Antar subsequently saves his father from captivity and saves his society from
conquest, thereby gaining his freedom. This story is interwoven with a romantic
tragi-comedy in which ‘Antar continually seeks union with his beloved ‘Abla. She
believes that ‘Antar has died and is self-exiled to the desert before their union. (On
the Sı̄rat ‘Antar, see Heath.)
The central Islamic story of the hijrah, the prophet’s flight from Mecca, is an-
other instance of this structure. The flight was, of course, a sort of exile, precip-
itated, in the Muslim view, by the society’s rejection of its rightful leader. This
rejection was largely led by members of Muhammad’s own tribe, the Quraysh,
and most importantly his uncle Abū Lahab (see, for example, Ibn Ish.āq 161). Fi-
nally, it culminated in his return from exile and his military victory. The point
is not affected by the historicity of the events. For, out of the vast array of de-
tails that made up the Prophet’s life, Muslims have focused on this particular
sequence of events, selected and shaped, in part, according to the prototypes
we have been considering. (The emplotedness of historiography has been recog-
nized at least since Hayden White’s pioneering work. The relation between his-
tory and narrative has been explored from the literary side as well, through the
consideration of historical fiction – see, for example, Paul Hernadi’s Interpreting
Events.)
The Turkish Book of Dede Korkut, widely “accepted as the national epic of the
Turkish people” (Clinton 1497), is a collection of thirteen short narratives. The great
majority of these stories clearly manifest the heroic structure, complete or partial.
For example, the first story concerns Boghach Khan and his father Dirse Khan, a
Turkish noble. Dirse is convinced by his evil retainers that Boghach is committing
dishonorable acts. Dirse therefore agrees to deceive his son on a hunting trip and
kill him. Dirse Khan wounds his son on Kazilik Mountain and leaves him for dead.
Thus, we have the familial conflict, Boghach’s deprivation of his rightful place in
the nobility, his near death and a sort of exile (on Kazilik Mountain). (As to the last-
named element, Chadwick and Zhirmunsky note that “the ‘expulsion and return’
of the unjustly accused hero” is “typical” in central Asian epics [298].) Boghach’s
mother manages to cure the boy, with the help of a wandering saint. However, he
remains in hiding – again, a form of isolation from society. At this point, Dirse’s
retainers capture Dirse in order to sell him to the “infidel.” This is a variation on
the danger to the kingdom. Boghach rescues his father, which is to say, fends off the
threat. This leads to the restoration of Boghach to his rightful place.
Most of the other stories focus on the second part of the heroic sequence, the
threat and defense – often in the form of an abduction and rescue. For example, in
the second story, Salur Kazan, one of the great heroes of the book, is away from his
camp. The evil “infidel beast,” King Shökli attacks and abducts several of the Turks.
Kazan returns and defeats Shokli before the captives are harmed. It is a great social
victory, for in the course of the fight “Twelve thousand unbelievers were put to the
sword” (57). There is no need to go through every instance. It is worth noting that
only a few stories do not fit the heroic format. Of these, two have romantic goals.
Though the epic is notoriously absent from China, military romance and his-
tory serve much the same function. C. T. Hsia points out that in Chinese military
romances the reader is frequently faced with a loyal retainer of the emperor who
suffers “imprisonment” or “exile” (“The Military Romance” 362), often pretending
to be dead until an “invasion at the frontier makes [his] services indispensable”
(363). There is also much work of this sort in historiography. Consider, for exam-
ple, one of the most important ancient Chinese histories, and one of the first to
develop an extended, fluid narrative – the Kuo yü or Conversations from the States.
As Burton Watson explains, the longest section of the work concerns, first of all, a
conflict between brothers – or, more exactly, between a prince, Shen-Sheng and his
step-mother, regarding the prince’s half-brother. This results in the disinheritance
of the prince and his death. This is followed by the exile and eventual accession
to the throne of another prince, who has, in effect, taken up Shen-sheng’s role and
completed the heroic narrative. Though history, it is emplotted history, and its plot
is one of heroic tragi-comedy.
Three Kingdoms, generally considered one of the five masterpieces of Chinese
prose fiction, fits the pattern also. This work is too long and complex to count
as treating a single plot, though it has a tighter and more prototypically unified
structure than a work such as the Shâhnâme. Specifically, Three Kingdoms treats the
disintegration of the Han dynasty, the period of its division into three kingdoms,
and the return to unity under a new dynasty, the Jin. On the whole, the novel treats
the denial of the legitimate emperor, various exiles and deaths of legitimate heirs,
and, finally, the reestablishment of legitimate succession in a later generation. While
some of the other works we have considered stress the middle part of the heroic
structure – the external threat and defeat of that threat – Three Kingdoms treats the
opening and conclusion of that structure. It concerns usurpation and the internal
strife that goes along with it, rather than invasion and the threat of conquest.
The novel begins with a rebellion, defeated by the hero, Liu Xuande, and the man
who will ultimately become the villain, Cao Cao. This is a sort of threat to the king-
dom. However, it is internal, not external. In any case, it puts us in mind of the
theme of usurpation and prepares us for the conflict that follows. Through a com-
plex series of actions, Cao Cao undermines the ruling dynasty, killing or otherwise
neutralizing members of the ruling family. After Cao Cao’s machinations, Liu is
clearly the only legitimate alternative to the new dynasty, illegitimately proclaimed
by Cao Cao’s son, Cao Pi. However, the conflict with Cao Cao has driven Liu away
from the capital. He has established himself in the Riverlands, organizing an alter-
native kingdom from which he can fight against the usurpers. Thus we have all the
major elements from the first part of the heroic structure – denial of rightful place
to the ruler, death, and exile. In the most straightforward fictional version of such a
story, either Liu himself or one of his descendants would ultimately defeat Cao Cao
(or one of his descendants), reuniting the empire under its rightful ruler – perhaps
after defeating some external threat from, say, the Mongols or another non-Chinese
group. Indeed, the novel develops in this direction, with Liu aided by an expert
advisor (Kongming), a standard character in such works. However, Three Kingdoms
has the task of writing a heroic story within the general confines of history. Histor-
ically, China was reunited. In this sense, the usurpation was overcome. However,
China was not reunited by a descendant of Liu or any one of the murdered heirs.
This difficulty is overcome by the doctrine of the Mandate of Heaven. The proper
conclusion
In sum, I have argued, first, that lexical entries for emotion terms, thus
our practical cognition about emotion, are most significantly defined
by prototypes. Second, prototypical narratives involve the different
types of emotion – junctural, outcome, and sustaining. But they are
structured primarily by reference to the prototype eliciting conditions
for the outcome emotion of happiness. In prototypical stories, the
complex process of narrative construction is guided and organized
by the expansion or elaboration of the micronarratives that define the
prototype eliciting conditions for happiness, whether or not those
ruler is the person who inherits the Mandate of Heaven, even if that person is not a
descendant of the prior ruler. Thus, in the end, the new, unifying dynasty is taken
to have inherited this mandate. In this way, the new dynasty serves as the heir of
the dethroned monarch and its rule operates as a restoration and an overcoming of
the usurpers.
I take it that European examples are too well-known to require elaboration.
However, we could mention The Song of Roland, Beowulf, King Lear, Macbeth, certain
plot sequences from such mixed heroic/romantic poems as Jerusalem Delivered (for
example, the story of Rinaldo) and the Niebelungenlied. Even Hamlet or The Tempest
could reasonably be understood as a variation on this structure, with magic replac-
ing military battle in the latter case – an unsurprising substitution in that magic
enters directly into military battle in many works of this sort. The list could be
extended almost indefinitely.
13 For reasons discussed in the first chapter, having to do with methodological re-
quirements for the determination of universals, I have drawn almost all preceding
examples from the period before the rise of global European colonialism. Some
readers have taken this to mean that the genres I am speaking of have long since
passed from this world. The impression has no doubt been made stronger by my
references to monarchs and warriors. In fact, however, these genres still continue to
dominate literature, and now film also. There are three ways in which this is true.
Even if no new works were written in these genres, a massive collection of romantic
and heroic tragi-comedies has been canonized in every tradition. These works are
remembered, read, acted on stage, filmed, taught in classes – Shakespeare being an
obvious example. These works are not merely historical stepping stones by which
we arrived at Modernism. They are a continuing part of the body of literature. They
are as much a part of the modern period as they were a part of their own peri-
ods. Indeed, these are the paradigm works of the various literary traditions. Thus,
many of them are more central to our contemporary experience and understanding
of literature than they ever were in the past. Indeed, they are more central to our
contemporary experience and understanding of literature than almost all contem-
porary works. Hamlet remains more widely read, seen, and interpreted than any
nouveau roman. In keeping with this, such works continue to instantiate our concept
of prototype narratives.
But, of course, that is not all. Romantic and heroic tragi-comedies do continue
to be written today. In fact, to all appearances, they dominate narrative production
today no less than they did in earlier centuries. Most obviously, these genres dom-
inate popular cinema. As I write, Titanic is the largest grossing film ever made and
it is a prototypical romantic tragi-comedy. Its development is not merely simple
and straightforward. In fact, its variations on the standard structure are sometimes
very sharp and effective. (A good example is the way Jack takes up the female role
and “saves” Rose from despair. Rose, then, takes up the male role and saves Jack
from physical danger.) However, the general outline is so clearly prototypical that
it hardly requires explanation – the lovers face parental opposition, based on social
class, with a parentally chosen rival, and so on. The second biggest U. S. box office
success ever is from the Star Wars series. The four movies of the series are all in
the box office top twenty. (For these figures, see http://www.the-movie-times.com.)
The trilogy makes up a straightforward, if rather minimal heroic tragi-comedy. The
opening, implicit in the first episode, is usurpation – the usurpation of the Republic
by the Empire and, simultaneously, of Obi-Wan Kenobi by Darth Vader. The focus
of the trilogy is on the struggle against usurpation. This struggle is waged by Han
Solo and, more importantly, by the siblings, Luke Skywalker and Princess Leia, a
standard generational displacement, if a somewhat unusual one in that they are the
biological children of Darth Vader, even if they are the spiritual children of Obi-Wan
Kenobi. This culminates in a final victory against usurpation. There is, in addition,
a standard, if simple, love story. We also find the standard elements of exile, and so
on. The Phantom Menace has a complete sequence of invasion, usurpation, exile of
the monarch, struggle, and final victory.
When adjusted for inflation, Gone With the Wind captures first place, by the mon-
etary measure of popularity. In the romantic plot, Ashley and Scarlett are in love.
Scarlett’s father opposes her, and in any case Ashley is engaged to another woman.
The love story works its way through the four hours of the movie, interwoven
with a shortened version of the heroic plot – where the home society is threatened
and then defeated by what is in effect a barbarian invasion. (The point that the
“barbarians” put an end to slavery is not relevant to the film’s characterization of
the national enemy, which follows standard heroic principles in this regard.) It also
takes up the standard elements of the death or near death of the beloved (all four
lovers almost die), exile (from the home plantations of Twelve Oaks and Tara – the
latter imagistically represented as a nation rather than a plantation), return from
exile, and the final death and/or separation of the lovers (as it is, like Titanic, a
tragedy, if a tragedy with a hopeful end). There are certainly complications here.
For example, the love story involves a fairly common variation in which the love
triangle is doubled (with Rhett Butler pursuing Scarlett). It includes an unusually
sympathetic portrayal of all the rivals. It resolves with a partial realignment of the
characters’ feelings in order to accommodate the union of two couples – but then
splits all the lovers in its tragic conclusion. Moreover, it presents the heroic plot
from the perspective of the women at home, rather than the men in battle. In short,
it varies standard structures in interesting ways. Nonetheless, the structures that it
varies are the prototype romantic and heroic plots.
It is no coincidence that the most popular or successful movies ever (Gone With
the Wind and Titanic) have prototypical romantic plots with, in the top case, a heroic
subplot, while their nearest competitors (the Star Wars movies) form a prototypical
heroic story with a romantic subplot.
Finally, the importance of these two genres is not confined to popular narrative,
but extends to many works of high Modernism and Post-Modernism. Of course, to
a great extent, both Modernists and Post-Modernists set out to break with canon-
ical and popular structures. Thus, it would hardly be surprising if their works,
self-consciously set against common practices, had little relation to the prototype
narrative genres. Indeed, it would have no bearing on the present argument. If a
work is designed to be nonprototypical, it cannot count against a theory of proto-
typical narratives. However, as it turns out, many of our paradigm Modern and
Post-Modern works do indeed treat prototypical stories. Romantic structures in
particular are clearly present in major writings by such canonical modernists as
D. H. Lawrence and Marcel Proust or even in such a paradigm Post-Modern work
as Alain Robbe-Grillet’s masterful nouveau roman, La jalousie. Moreover, the pres-
ence of these genres is not confined to such relatively obvious cases.
Consider, for example, James Joyce’s Ulysses. Ulysses is probably the paradigm
Modernist work. It is certainly not a work that anyone would think of as manifest-
ing heroic or romantic patterns. Indeed, there is not a great deal in the novel that
could be considered plot. But there is something. In fact, there are a couple of rough
plot sequences. The most straightforward plot sequence centers on Bloom. It is un-
equivocally a version of the romantic tragi-comedy. Bloom’s wife is beginning an
affair with another man that very day. Bloom returns home, considering a divorce.
This is a common part of the romantic plot – the entry of a rival and the possible
estrangement of the lovers. And, of course, the Blooms’ relationship is one of those
“abnormal” unions that are at the center of romantic tragi-comedy. Though not for-
bidden, they remain a socially suspect couple because of their different ethnic and
religious backgrounds. Moreover, the epic model for this part of the novel is itself
a version of a small part of the romantic plot – the exilic wanderings of Odysseus
and his eventual return to and reunion with his faithful beloved, Penelope.
The other, even more minimal plot concerns Stephen. Stephen sees himself as
the legitimate poet laureate of Ireland. But he is not recognized as such by anyone
else. Rather, Buck Mulligan seems to be receiving precisely the accolades that
Stephen (perhaps) deserves. It should be clear already that this is a version of the
heroic plot, the plot aimed at social domination. In this case, Joyce has primarily
taken up the usurpation part of the story, for Stephen has been set aside from his
rightful place of social esteem and authority. Indeed, Joyce explicitly models this
part of the novel on two heroic stories of the standard political/military sort, two
prototypical stories of usurpation – the Telemachus portions of the Odyssey (re-
call that Telemachus’s position as ruler of Ithaca is threatened by the suitors) and
Hamlet, where rule has been usurped by Claudius. The links with military heroism
are stressed at various points. For example, Stephen lives with Buck Mulligan in
a tower built for military purposes (Gifford and Seidman 1.542). He thinks of his
conflicts with Mulligan as a sword fight: “Parried again” (l. 152), he remarks, refer-
ring to one of Mulligan’s replies. He explains that Mulligan “fears the lancet of my
art as I fear that of his” (l. 152) – though a lancet is not a sword, the point is that
Mulligan, a doctor, fights with a lancet, while Stephen battles with “The cold steel
pen” (l. 152–53). At the end of the opening chapter, Stephen aptly names Mulligan
“Usurper” (l. 744). Joyce extends the heroic resonances still further, suggesting the
element of threat to the nation, by linking Mulligan with Haines, and thus implicitly
conditions are ultimately achieved in the narrative. In connection

with this, prototypical stories most often incorporate some version
of the prototype conditions for sorrow, placing them in the narrative
middle or progression. In some cases, the narrative ends with the per-
manent establishment of these sorrowful conditions. Rather than be-
ing an autonomous, sorrow-based structure, however, stories of this
sort are truncated forms of the happiness-based structure. In other
words, tragedy is a derivative of comedy. Third, the predominant
prototypes for happiness are (a) romantic union with one’s beloved
and (b) the achievement of political and social power, both by an
individual and by that individual’s in-group (for example, his/her
nation). These prototypes are triggered by different contexts – those
of personal and social happiness respectively. Finally, these happi-
ness prototypes define romantic and heroic tragi-comedy, which are
the most common and most prominent narrative structures cross-
culturally. Moreover, along with the related prototypes for sorrow
(and some well-established cognitive principles, such as priming),
they account for the complexity of these universal narrative struc-
tures in detail.
with the history of English colonialism, “The imperial British state” (l.643), one of
the “two masters” (l.638) to whom Stephen has been bound as a “servant” (l.638).
This colonial conflict culminates in Stephen’s encounter with the British soldiers
late in the novel.
In his self-conscious revision of European epic, Joyce takes up the Roman-
tic identification of hero and artist, putting the artist into the prototypical heroic
structure – for Buck usurps Stephen’s position as poet, not as king. (In some other
modern works – such as the highly popular and award winning play, How to Succeed
in Business Without Really Trying – the businessman is put into that structure. The
point is not confined to the west, as is clear from the Japanese salaryman who “even
today sees himself as the heir of the samurai” [Williams 272n. 13, citing Hiroshi
Yoshioka].) Other than that, Joyce’s changes in the prototype story are almost en-
tirely a matter of style and discourse – primarily a matter of radical reduction and
ellipsis, combined with a great elaboration of nonstory elements and an often heavy
use of irony. This is typical of Modernist and Post-Modernist writers.
4
Writing Beyond the Ending

A Problem of Narrative, Empathy, and Ethics
In the preceding chapter, I argued that prototype narratives ex-

hibit remarkable regularity, with detailed patterns of romantic and
heroic tragi-comedy recurring prominently across unrelated tradi-
tions. Moreover, I argued that this pattern may be accounted for by
reference to a cross-cultural constancy in personal and social proto-
types for happiness – that is, romantic union and social domination,
respectively. In connection with this argument, I noted that the heroic
structure is more variable than the romantic structure. For the most
part, this variability is not problematic and the various alternatives
are easy to accommodate into our general account. However, there is
one apparently anomalous variation in heroic tragi-comedies.
A surprising number of paradigm heroic tales include a peculiar
ending, a sort of negative postscript to the achievement of the heroic
goal. Moreover, this is common enough across genetically unrelated
traditions to count as a universal. This raises a serious explanatory
problem, for it is not at all clear why any story would, so to speak,
“go beyond the ending” in this way, continuing the tale after the
achievement of its guiding aim. After all, this is extremely uncom-
mon in romantic plots. Indeed, its uncommonness in romantic plots
suggests the basic reasonableness of our – or at least my – puzzlement
over the heroic “continued” or “excess” ending, as one might call it.
It also suggests that this odd ending may have the potential to ad-
vance significantly a research program in cognition, emotion, and
narrative.
122
A Problem of Narrative, Empathy, and Ethics 123
I shall first lay out the problem, noting some of the primary
cases from European, Middle-Eastern, African, Indian, and Japanese
literature. I shall then consider what I believe to be the two primary
elements in an account of this phenomenon – ethical evaluation and
empathy – elaborating on the function of these elements in literary
creation and response.
anomalies of heroism: the epilogue of suffering

The continued or excess ending of heroic tragi-comedy focuses on
pain following the achievement of social dominance. I shall refer to this
ending as the “epilogue of suffering.” As with the heroic plot more
generally, this is a variable structure. In its least developed form, it
simply involves a turn from the triumph of the hero to the sorrow
of those who have been defeated. A good example of this is to be
found in the Iliad. We are perhaps so accustomed to the ending of
this poem that it does not strike us as odd. But it is odd. The an-
tepenultimate book deals with Achillēs’ triumph over Hector. The
poem could have concluded here, for from this point the reader can
envision the Greeks’ ultimate triumph, a triumph predicted explicitly
by Prı̄am (see 257). In this way, it could end like the Aeneid, which
concludes with the victory of Aeneas over Turnus, modeled in part
on the victory of Achillēs over Hector.1 But the Iliad does not end this
way. In the next book, the Iliad continues to the funeral of Patroclos
and the games following this ceremony. This too could be seen as a
triumphal ending – though even here, the triumph is marred by
1 Of course, the Aeneid was incomplete (see Knight 11). We do not know how Virgil
would have continued his poem. Perhaps he too would have added an epilogue of
suffering. Virgil does establish a strict parallel between Turnus and Aeneas, between
the Greeks at the outset and the Trojans at the conclusion, as writers such as David
Quint have noted (see Quint 66–74). Indeed, Virgil does this to such an extent that
the poem seems to lead inevitably to an epilogue of suffering. The point is only re-
enforced when one recalls the suicide of Dı̄dō, which develops far greater sympathy
than the parallels with Cleopatra should have allowed, given the political conditions
at Virgil’s time (on Dı̄dō connection with Cleopatra, see Quint 28, 109, and 383n. 14).
Having referred to Quint’s Epic and Empire, I should perhaps note what some
readers will take to be obvious – that I disagree with a number of Quint’s conclusions
about “epics of the victors,” such as the Aeneid. These poems, quite a few of them
anyway, are far more equivocal and ambivalent than Quint usually allows (despite
occasional disclaimers).
sorrow. But the ultimate conclusion of the poem is Prı̄am’s acqui-

sition of Hector’s body, his return to Troy, and the laments of the
women (“grief intolerable came upon every heart” [296]), including
Andromachē’s prediction of the murder of Astyanax: “The boy is
only a baby . . . he I think will never grow up to manhood. . . . [W]ives
and little children . . . will soon be carried off in ships, and I with
them. . . . [S]ome enemy will catch [Astyanax] by the arm, and throw
[him] over the wall to a painful death” (296). In short, the final scene
of the poem is one of utter despair, focused in this case on those who
have been defeated.
This is not some idiosyncracy of Homer’s. The Japanese Tale of the
Heike has a similar, pathetic conclusion. The Heike is the major heroic
tragi-comedy in the Japanese canon, and as such it has a place in
Japanese literary culture comparable to that of the Iliad in the west.
It concerns the battle of the Minamoto and the Taira. The Minamoto
triumph. But the final chapter turns our attention to an Andromachē-
like figure, the Imperial Lady, Kenreimon’in. Her young son, a boy
of eight, was the Taira emperor. He and his grandmother committed
suicide to avoid being taken by the Minamoto soldiers who, like the
Greeks in Troy, killed all possible Taira heirs. Kenreimon’in retires to
a lonely hermitage in the mountains, praying to the compassionate
Buddha for the well-being of her dead child and relief of her own
suffering: “never, in all the lives to come, shall I forget the Former
Emperor’s [that is, her son’s] face. I try to forget, but forgetting is im-
possible; I try to control my grief, but that is also impossible. Nothing
causes such sorrow as parental affection: that is why I pray faithfully
for the Former Emperor’s enlightenment, morning and evening. I
believe my love for him will guide me to enlightenment, too” (434).
The misery is not confined to Kenreimon’in: “there was no house free
of disquieting winds, even inside jade blinds; there was no dwelling
where dust never rose, even beyond brushwood doors. Husbands
and wives who had slept on adjoining pillows were as remote from
one another as the sky; nurturing parents and their children were set
apart, neither knowing the whereabouts of the other” (437). The final
sentences of the story recount Kenreimon’in’s death, followed by a
statement regarding the sorrow of her two attendants, their eventual
enlightenment, and their rebirth in the Paradise of Buddha.
The ending of the Heike introduces some motifs that we find, usu-
ally presented from the other side, in the more fully developed or
“complex” epilogues of suffering. These include at least some ref-
erence to the misery of the defeated enemy (if usually less than in
the “simple” versions such as those just discussed), but they focus
on the triumphant heroes. In the most completely elaborated version
of this epilogue, the hero triumphs, which is to say, attains the goal
of social power. But this does not bring happiness. Rather, the hero
finds that the losses outweigh the gains, often vastly, and he/she is
plunged into remorse and despair. In consequence, he/she sets out
on a sort of spiritual journey – or is forced into exile – in which he/she
experiences great pain and anguish, explicitly or implicitly as pun-
ishment for past acts. This is followed by what is sometimes a sort
of spiritual transcendence, but sometimes little more than a glorified
suicide.
Again, not all these elements are necessarily present in any
given case. The most complete case I know of is the Mahābhārata –
unsurprisingly, perhaps, as it is the longest of the relevant texts. This
poem concerns the battle between the Pān.d.avas and the Kauravas.
The Pān.d.avas are the heroes. Of the Pān.d.avas, Arjuna is the most
heroic, in that he is most engaging as a warrior, but Yudhis.t.hira is
the most admired as a person, and he is the appropriate ruler as well.
When the Pān.d.avas triumph, Yudhis.t.hira weeps uncontrollably and
shouts “I am burning with grief, like a person thrown into a blazing
fire” (Vyāsa vol. 7, “Stree Parva” 43). Subsequently, for other reasons,
Arjuna too sinks into despair: “O best of men, it behoveth thee to
tell me what is good for me now, for I am now a wanderer with an
empty heart, despoiled of my kinsmen” (vol. 12, “Mausala Parva”
271). More importantly, at the end of the poem, though it follows
the triumph by a number of years, the Pān.d.ava brothers and their
common wife, Draupadı̄, abandon all earthly comforts, change their
clothes for the bark of trees, and begin to wander in the wilderness
“resolved to observe . . . Renunciation” (vol. 12, “Mahaprasthanika
Parva” 274). One by one, they drop dead, explicitly due to their own
spiritual weaknesses. When Yudhis.t.hira himself enters heaven, he
finds the Kauravas there, and the Pān.d.avas in hell. The situation
is quickly changed, and the Pān.d.avas ascend into heaven, but the
epilogue of suffering intervened – or, in light of this ultimate up-

liftment, perhaps I should call it the “epilogue of suffering and
transcendence.”
The Mwindo Epic of the Nyanga (Congo) is another very good
example, though the final conclusion is not so much a spiritual tran-
scendence as a sort of ethical reorientation. Mwindo has been de-
prived of his rightful place in society. When he finally succeeds in
gaining the chieftancy that was his due, he performs an act for which
he is explicitly punished by the god of lightning, Nkúbá. The pun-
ishment consists in a physical and spiritual journey through the sky,
which in this case operates as a sort of hell. In keeping with the wan-
dering motif, Mwindo has “No house!” but “live[s] there in mere no-
madism” (137–8). After a year, he is returned to his society, an ethically
reformed man.
The Rāmāyan.a also has an ending of this sort, in this case one
illustrating the way in which assertions of transcendence are often a
sort of gloss upon something that is much more dismal. The center of
the poem is Rāma’s battle with Rāvan.a to rid the world of demonic
terror and to be reunited with his beloved Sı̄tā. After his victory,
both Sı̄tā and Rāma suffer great misery for complex reasons. Sı̄tā
eventually calls on the earth to swallow her, which it does. Rāma
plunges into a river. Both are represented as transcendent moments of
union with divinity. However, there are clearly suggestions of suicide
in each case. Indeed, this may be one reason for Bhavabhūti’s revision
of the ending of the Rāmāyan.a in his Uttararāmacarita. In that play,
Bhavabhūti represents Sı̄tā as trying unsuccessfully to commit suicide
by jumping in a river – a combination of the two scenes from Vālmı̄ki’s
original.
Gilgamesh concludes with a failed quest for transcendence. Indeed,
it does so twice. Gilgamesh and Enkidu succeed in killing Humbaba
and the Bull of Heaven. But Enkidu dies, and Gilgamesh is deeply
distraught. He undertakes a journey to find Utnapishtim and learn
his wisdom. But Gilgamesh fails to fulfill Utnapishtim’s challenge; he
loses the plant of rejuvenation; and he returns to his kingdom know-
ing only that everything is impermanent. In an alternative ending,
Enkidu dies and returns to Gilgamesh, describing the existence of a
miserable spirit who has no one to remember him in all the living
world.
There are other cases that suggest problems of this sort as well.
The Turkish Book of Dede Korkut ends with a chapter treating civil war
and the disintegration of the very society that has been celebrated
in the preceding episodes. The Malian epic of Son-Jara points to the
great king’s demise – but the poet refuses to go into the details, and
even warns, “never try, wretch, to pierce the mystery which Mali
hides from you. Do not go and disturb the spirits in their eternal
rest. Do not ever go into the dead cities to question the past, for the
spirits never forgive. Do not seek to know what is not to be known”
(Kouyaté 84). The warning is general and need not refer to anything
specific about Son-Jara’s death. But it is certainly ominous and does
not suggest happy events at the end of this story.2
Moreover, this is not all there is to the matter. The epilogue of suf-
fering may appear as a part of the original work, or it may be added in
a subsequent revision or “reappropriation” of the work within a liter-
ary tradition. For example, one of the most famous revisions of a story
from The Tale of the Heike is Zeami’s great Nō drama, Atsumori. In “one
of the most cherished stories in the Heike” (Varley 305), Kumagae,
a Minamoto warrior, kills Atsumori, a young man of sixteen or
seventeen. Kumagae hesitates before killing him, and deeply regrets
his death. In Zeami’s play, Kumagae has left his home and become
a monk, wandering and “pray[ing] for the salvation of Atsumori’s
soul” (707). He meets the ghost of Atsumori, who runs to kill him with
a sword, “‘There is my enemy,’ he cries, and would strike,/But the
other is grown gentle/And calling on Buddha’s name/Has obtained
salvation for his foe;/So that they shall be re-born together/On one
lotus-seat” (712). This is, of course, precisely the structure I have
been examining – including the moment of transcendence – though
it occurs in a subsequent play, not in the main work.
Finally, paradigm dramas sometimes represent an epilogue of suf-
fering without relying on or referring to some commonly familiar
2 It is interesting to note that the oddity of the endings we are discussing is some-
times explicit in the tradition surrounding the work. The epic of Son-Jara seems to
suggest that there is or was such an epilogue, only to repudiate it; Johnson notes
that Son-Jara’s death is something that “bards . . . keep as a secret” (22). The final
chapters of The Tale of the Heike and the Rāmāyan.a are disputed (see Varley 306 on the
former; on the latter, see Dimock 73 and Narayan 171). The endings appear, then,
both appropriate and wrong.
source work. Sophoclēs’ Oedipus plays are a clear case. Taking place
after Oedipus has achieved social domination, the first play barely
sketches the story of Oedipus’ success. Moreover, in that sketch, it
focuses on a crime that initiates the epilogue of suffering. While the
first play begins the punishment for that crime, the second takes up
the exile and concludes with the spiritual elevation, completing the
usual structure.3
spiritual transcendence as romance

and reconciliation
The examples I have been considering suggest that the epilogue of
suffering involves ethical and what might be called spiritual or re-
ligious issues or dilemmas. These concerns may arise directly, as in
the fully elaborated form where the hero evidences remorse or even
despair (clearly an ethical feeling in these cases), suffers punishment,
and ultimately reaches some transcendent spiritual state. But both
ethical and spiritual issues are also suggested in the simpler versions
that adopt the victim perspective rather than the victor perspective.
In the Iliad, the lament of Andromachē implies guilt on the part of
the Greeks – a guilt that is made fully manifest in Euripidēs’ Trojan
Women, which takes up the events predicted by her. Moreover, the
concluding ceremony is Hector’s funeral, a ritual to benefit his soul.
The point is even clearer in The Tale of the Heike.
In analyzing the contexts for and types of happiness prototype, I
drew on the Sanskrit theory of the goals of human life, referring to
the goals of kāma and artha, romantic union and social domination.
These fit well with aspects of Sanskrit aesthetic theory, and, more
importantly, with cross-cultural literary patterns. However, for sim-
plicity, I left aside two other goals specified by the Sanskrit writers:
dharma, or ethical duty, and moks.a, or spiritual transcendence. These
are, of course, just the concerns that bear on the epilogue of suffering.
Before discussing the epilogue of suffering in more detail, and before
3 One could see an element of this in the second and third parts of the Oresteia as well.
It seems that the revenge plot is a variation on the heroic structure. Indeed, it appears
to be closely related to the epilogue of suffering, as the case of Atsumori suggests.
However, the precise relation between these narrative types is complex and requires
separate treatment.
trying to account for it, it is worth considering the goals of ethics and
transcendence in relation to literature more generally.
The preceding data and the Sanskrit division of goals indicate that
our earlier organization of happiness prototypes into two contexts is
incomplete. In addition to the default context of personal happiness
and the alternative context of social happiness, there seems to be a
third, also alternative context, that of divine or transcendent happi-
ness. These three could be schematized simply as happiness in the
home, happiness in the social world, and eternal happiness, happi-
ness beyond mundane life (for example, heaven). Put differently, our
lexical entries for outcome emotions seem to involve a distinction
along these lines.4
However, the three categories are not equal. Not only is personal,
romantic happiness the default case, divine happiness does not ap-
pear to have any distinct prototypical eliciting conditions. Like many
abstract concepts, we understand transcendent joy only by reference
to some more tangible, direct, mundane experience of joy, most often
romantic union. In other words, to understand religious happiness,
we invoke the default type of happiness – union with the beloved.
This is particularly true in literature. Indeed, this accounts for another
literary universal. The assimilation of spiritual bliss to requited love
not only occurs cross-culturally, it is often elaborated into an allegory
of God and humanity as lovers – an allegory found in writings from
the Song of Songs to the bhakti literature of Hinduism to the Persian and
Arabic ghazal. (We will return briefly to this idea in the next chapter.)
Perhaps more importantly for our purposes, the converse holds as
well. Not only does romance serve as a model for spiritual relations,
spiritual motifs pervade romance. Thus, there are often direct sugges-
tions of spiritual trials and achievements in straightforward romantic
tragi-comedies. Most obviously, the reunion of the separated lovers
frequently involves imagery of spiritual elevation, as I stressed in
Chapter 3.
In heroic works, transcendent concerns enter as a literal outcome
of the events in the epilogue. Somewhat surprisingly, they are often
4 I suspect that the operation of this division is independent of one’s self-conscious

religious beliefs; for example, I am fairly certain that it operates in my own lexicon,
despite my self-conscious atheism.
represented by or associated with some sort of union directly paral-

lel to the conclusion of the romantic plot. Indeed, sometimes there
is even a reunion of lover and beloved – as in the Rāmāyan.a, where
the two virtual suicides bring Rāma and Sı̄tā back together in their
divine forms of Vis.n.u and Laks.mı̄. With greater frequency, the tran-
scendental conclusion involves some other sort of familial reunion –
a reunion of brothers in the Mahābhārata, in Hell, then in Heaven; a
(possible) reunion of mother and child in Paradise in The Tale of the
Heike; the reunion of Mwindo with his father in the Nyanga epic.
Familial reunions in heroic stories are, however, somewhat peculiar
as prototypical instances of happiness for they often follow not only
separation, but severe antagonism. The reunion of Rāma and Sı̄tā fol-
lows their estrangement. The reunion of Mwindo and his father, Shé-
Mwindo, follows Shé-Mwindo’s attempt to kill Mwindo. The reunion
of the Pān.d.avas too fits here, for Karn.a is included in this heavenly
family, while in life he was mistakenly despised by his brothers as
low caste. Indeed, at the end, he was killed by one of these broth-
ers. There may even be a hint of this in the case of Kenreimon’in
and her son, for, when he died, she neither protected him nor died
with him.
Other instances take us outside the family to the strange union
of enemies as friends – or, indeed, as members of one family. Prı̄am
and Achillēs almost become father and son for a few hours when
Prı̄am visits Achillēs to retrieve Hector’s body. In keeping with this,
he introduces himself by saying, “Remember your own father, most
noble prince Achillēs, an old man like me near the end of his days”
(291). Another instance of reunited enemies, in this case with an im-
plication of eventual brotherhood, may be found in Atsumori, when
the ghost of Atsumori rushes at Kumagae to kill him, but achieves
enlightenment due to Kumagae’s prayers, and the two are united,
“re-born together/On one lotus-seat” (712). Gilgamesh presents a re-
union of friends, rather than enemies. In keeping with the dismal sec-
ond and even third endings of this poem, the reunion fails. Enkidu
returns from the dead, a mere shade: “The ghost of Enkidu issued
from the darkness like a dream./They tried to embrace, to kiss one
another./They traded words, groaning at one another” (263). But
here too there has been a sort of severing or alienation between the
friends. For in this episode (which contradicts earlier events in the
poem), Enkidu has died because he ignored the words of Gilgamesh

when entering the underworld.
In all these cases, the eliciting conditions for transcendent joy are
specified, literally or metaphorically, by reference to the eliciting con-
ditions for another, more concrete happiness – the joining, or rejoin-
ing, of two or more people. In romantic tragi-comedy, this joining
is, of course, romantic. In heroic tragi-comedy, it is often familial.
Moreover, even when it is not literally familial, there is often a strong
component of familial imagery associated with the reunion. Perhaps
more importantly, the reunion at issue in heroic tragi-comedy sug-
gests reparation, forgiveness, the reconciliation of antagonists.
Thus, the third context for happiness – spiritual, divine, or
religious – is bound up with the romantic and heroic contexts as well.
Once again, spiritual joy appears to lack its own prototype eliciting
conditions for happiness. In consequence, it borrows them from other
contexts. While one prominent cross-cultural prototype for spiritual
happiness is romantic union, another is the happiness that results
from forgiveness and reconciliation – particularly, a reconciliation
that moves from the sharpest conflict to the sharpest union, thus, a
familial reconciliation of enemies.
This reference to forgiveness and reconciliation returns us to the
issue of ethics.
ethical traumas of heroism

As noted previously, the Sanskrit writers distinguish four goals of
human life. I have just maintained that there are three lexical con-
texts for happiness, which is to say, three specifications of the goal
of happiness. In saying this, obviously I have left aside the fourth
goal mentioned previously, dharma or ethical duty. I have done this
because ethical goals appear to operate somewhat differently from
romantic, social, and spiritual goals. First of all, ethical happiness is a
happiness in facts, not in substances, to draw on an idea of Elizabeth
Anscombe’s. If I am ethically pleased by something I have done, that
is to say that I am pleased about the fact of having done it; whether
I enjoyed the act itself is irrelevant. One enjoys the substance of
union with one’s beloved or reunion with an estranged family mem-
ber, but one probably does not enjoy the substance of defending a
tenure candidate against one’s own professional self-interest, say;

nor does one enjoy the substance of the various “negative actions” or
restraints that we engage in every day – not deceiving someone for
self-advancement, and so on. We enjoy these last two types of action
only as facts. We enjoy the fact of having defended someone who de-
served defending, despite the prejudices of our colleagues. We enjoy
the fact of negative actions, or we avoid the factual pain – the guilt
and remorse – that we would have felt had we acted otherwise.
In connection with this, it is unlikely that ethics will define pri-
mary goals – the crucial eliciting conditions for happiness – in life
or in literature. Ethical enjoyment is a sort of second-order enjoy-
ment. We might reasonably expect ethical goals to be second-order –
“meta-goals” – as well. Certainly, it is possible to set out solely to be
moral. However, more typically, ethical excellence operates, not as a
final, motivating aim, but as a sort of norm against which we eval-
uate our goals and the actions that lead to the achievement of those
goals. This is, in any case, what we find in prototypical narratives. It
is rare to have a story structured entirely around the goal of achieving
ethical purity. However, ethical evaluation is part of most heroic and
romantic tragi-comedies. The point is suggested by Aristotle’s the-
ory of hamartia – which adheres closely to his ethical account of the
“kinds of injury between man and man” treated in the Nicomachean
Ethics (415; 1135b). In this view, ethical evaluation of the protago-
nist is centrally important to our aesthetic response. For example, in
Aristotle’s account of plot, “The change of fortune . . . should come
about as the result not of vice, but of some great error or frailty”
(47; 1453a). Clearly, we are not dealing with ethical goals here. We
are dealing, rather, with a sort of ongoing ethical evaluation of ac-
tions guided by other sorts of goal. For example, Macbeth’s tragic
hamartia is not a matter of failing in his pursuit of ethical happi-
ness. Rather, it is a matter of acting wrongly in the pursuit of social
happiness.
Vālmı̄ki’s Rāmāyan.a provides a striking instance, for we are re-
peatedly faced with moral explanations of Rāma’s actions, especially
those actions that appear suspect. There is a recurring ethical com-
mentary that serves to justify Rāma’s behavior – not always suc-
cessfully. Indeed, one thing that makes the Rāmāyan.a interesting to
nonorthodox readers is the fact that it also includes contradictory
commentaries, such as Sı̄tā’s famous speech on Rāma’s tendency

toward violence. This speech includes her judgment – so at odds
with the lifestyle of a warrior such as Rāma and with the moral ratio-
nalizations surrounding his actions – that the “mind . . . is perverted
by . . . the use of weapons” (II.17).
Drawing on this account of ethical goals, we can begin to see how
the epilogue of suffering might derive from ethical evaluations de-
veloped explicitly or implicitly in the course of a heroic plot. But
before continuing with this, we need to consider a few more gen-
eral structural issues relating to narrative. As I have already noted,
narrative is produced not simply by the unfolding of emotion pro-
totypes, nor even by their specification. There are other prototypes,
as well as schemas and exempla, that contribute, in necessarily com-
plex ways, to the generation of any story. One of the schemas appears
to involve the development of conflict in the narrative “middle” or
“progression.” This conflict almost always has an ethical component;
it is almost invariably involved with moral evaluation. I am not re-
ferring here to a conflict between those characters who obey moral
precepts and those who do not – the battle between moral heroes and
moral villains. Rather, I am referring to a conflict between the hero’s
or heroine’s actions, on the one hand, and some putatively overriding
moral principle, on the other – including the sort of conflict isolated
by Aristotle under the name hamartia. Though not a necessary con-
dition for narrative, it is extremely common in the most prominent
narratives across cultures. Indeed, this is fully in keeping with the
general narrative principle that the converse of a goal precedes the
achievement of the goal (for example, the prototype eliciting condi-
tions for sorrow precede the achievement of the prototype eliciting
conditions for happiness). In this case, moral conflict precedes (what
should be) morally acceptable success.
In romantic plots, the conflict is typically between the lovers, on
the one hand, and rules of social hierarchy on the other. Of course,
these rules do not operate autonomously. They are asserted by some-
one representing authority. Again, this authority is often parental,
as in the typical New Comedy plot. It may be legal or economic,
as in some Kabuki plays where the beloved is legally bound to a
brothel and subjected to “the owner’s permission . . . the elders of the
Quarter” and “the manager of the Quarter,” as one character explains
in Chikamatsu’s The Courier for Hell (774). It may be religious, as in

the conflict between the dogmatic shaikh and the lovers in the ghazal
tradition (see, for example, Ralph Russell 38). In any case, the author
of a romantic narrative most often treats the hierarchical authority
as invalid. Indeed, the overthrow of this authority in romantic tragi-
comedy is almost invariably something to celebrate (cf. Frye on New
Comedy 163ff.). In keeping with this opposition to hierarchical au-
thority, the lovers in these stories typically come together as equals,
at least equals in love.5
In heroic plots, the ethical conflict is in many ways similar. In both
cases, there is some sort of opposition between affiliation and per-
sonal feeling, on the one hand, and a socially hierarchical ethics on
the other. At the same time, however, there is a very significant differ-
ence that comes from the fact that the defining ends of the two genres
are distinct. While hierarchical principles are, prima facie, incompat-
ible with romantic love, they are of a piece with social domination.
Moreover, in heroic plots, it is often the hero him/herself who artic-
ulates the hierarchical principle – though it may also be a parental,
religious, or related figure (as when the sage Nārada tells Rāma that
he must kill a low caste person who is engaging in spiritual practices
permissible only for upper castes [Vālmı̄ki III.572]). While the hierar-
chical principles of authority tend to lose in romantic tragi-comedies,
they tend to win in heroic tragi-comedies. In other words, in the ro-
mantic plots, affiliation tends to overcome principles of ethical hier-
archy; in heroic plots, principles of ethical hierarchy tend to overcome
affiliation. This occurs most obviously and most strikingly when the
hero is bound by duty to kill a relative or someone like a relative in
battle (for example, Sohráb in the Shâhnâme or Atsumori in The Tale of
the Heike). There are also less extreme cases, where, for example, the
hero must abandon a loved one (for example, Sı̄tā in the Rāmāyan.a,
Dı̄dō in the Aeneid). What is perhaps particularly odd about heroic
tales is that the reader’s – and the author’s – sympathy may go either
to the hero following the principle, or to the victim of the principled
5 When this is not the case, when the lovers are not equal in love, the comedy tends
to become “problematic,” as in Shakespeare’s All’s Well that Ends Well and Measure
for Measure or Terence’s distressing The Mother-in-Law, in which the romantic union
is based on a prior rape.
action. This sort of affective uncertainty is extremely rare in romantic

tales.
The importance of these conflicts – especially in heroic works – is
not confined to the original stories in which they appear. Subsequent
literary reappropriations of paradigm heroic tragi-comedies – such
as the Iliad, The Tale of the Heike, the Rāmāyan.a, and so on – often focus
on just these points of moral tension and distress, what we might call
“traumas of heroism.” I have already referred to Atsumori. I could add
Namiki Sōsuke’s Chronicle of the Battle of Ichinotani, a very different
treatment of the same incident, in which Kumagae sacrifices his own
son. As Paula Richman and others have carefully documented, there
are numerous revisions of the Rāmāyan.a in which hero and villain,
victor and vanquished, change places. There are revisions in which
acts of apparent cruelty are justified or softened – as in Bhavabhūti’s
Uttararāmacharita, where Rāma’s killing of the unarmed (low-caste)
Śūdra is presented as the latter’s spiritual liberation, for which he
is deeply grateful. In Bhavabhūti’s version, Rāma blesses the Śūdra
with the following benediction: “May those luminous worlds, full of
bliss . . . be thine for ever” (20). In contrast, in Vālmı̄ki’s Rāmāyan.a,
Rāma kills the Śūdra and the gods congratulate him, declaring “Well
done, well done! . . . thanks to you . . . this Śūdra has been denied
heaven” (III.574). Other revisions heighten the conflict, often devel-
oping sympathy for the victim – for example, Bhāsa’s Broken Thighs,
which takes up the killing of the Kaurava leader, Duryodhana, at the
end of the Mahābhārata war.
On the other hand, these final examples suggest that our formu-
lation of the main issue is not quite right. For instance, Rāma has no
affective attachment to the Śūdra whom he kills. I have been speak-
ing as if the conflict in these cases is between ethics, on the one hand,
and feeling, on the other. But this is not accurate. Indeed, even in
cases where there is a conflict between ethics and feeling, that is not
necessarily the main problem. Consider, for example, Rāma’s aban-
donment of his pregnant wife, Sı̄tā. What is deeply disturbing about
this is not the fact that he has allowed the hierarchical ethics of king-
ship to overrule affiliation, which is to say, his personal preferences.
Rather, the problem is that he has allowed the hierarchical ethics of
kingship to overrule the ethics associated with affiliation. In this case –
and, indeed, in most cases – what we find is not a conflict between
preference and principle, and not even a conflict between different

specific principles, but rather a conflict between two different types of
ethics. In this way, we could see the ethical conflict in heroic tales
as fundamentally different from that in romantic tales. Ethical con-
flict in the romantic genre concerns the meta-ethical issue of just how
far principles of hierarchical authority may ethically be extended to
constrain individual preference. Ethical conflict in the heroic genre,
in contrast, concerns the meta-ethical issue of which of two moral
systems should prevail when they adjure contradictory behaviors.
Many ethicians and psychologists have sought to formulate some
basic opposition that structures ethical ideas. Most often, these divi-
sions seem inadequate. Part of the reason for this inadequacy, I be-
lieve, is that most writers assume our internal ethical principles are
rough ethical theories with something like necessary and sufficient
conditions. In fact, it seems much more likely that our ethical deci-
sions are, in the first place, prototype based. We judge actions good
to the degree that they approximate positive ethical prototypes; bad
to the degree that they approximate negative ethical prototypes.
There are, no doubt, many different prototypes for ethical and un-
ethical behavior. I should like to distinguish two. One is protecting
one’s group – nation, religion, family, and so on. This is an ethics of
defiance and bravery. It is an ethics of martyrdom, for one’s country
or one’s faith. It is of a piece with a politics of identity. In literature,
it is an ethics linked with the sustaining emotion of heroic energy. It
is in general imagined as a defiance of those who are threatening –
more numerous, hostile, powerful, and so on. The other prototype I
should like to isolate is that of comforting and sustaining the miser-
able. This is an ethics of gentleness and compassion. It is the ethics
of the healer or of the Samaritan, in the famous parable. It is of a
piece with egalitarian and, so to speak, universal politics (as opposed
to identity politics). In literature, it is an ethics associated with the
sustaining emotion of love – though parental/filial love more often
than romantic love. It is in general imagined as service to those who
are weak.6
6 Readers will notice a similarity between the second ethics named here and the “ethics
of care” associated with the work of Carol Gilligan. There are several crucial differ-
ences, however. First, it is not opposed to an “ethics of justice,” but to an ethics of
This general division is closely related to that between

ks.atriyadharma and sādhāran.adharma in Hindu thought.
Ks.atriyadharma is the duty of warriors, prominently including the
prosecution of war, both defensive and offensive – the “protection of
life” and “military occupation,” according to Kaut.ilya’s
Arthaśāstra (6). Sādhāran.adharma is the “universal” duty of all
persons, primarily the duty of ahim . sa, which is to say, the refusal to
cooperate with violence. It is clear that these two types of ethics are
sharply opposed. Much early Indic literature develops implicitly –
and, at times, even explicitly – out of the conflict between these
irreconcilable systems of moral thought.
As already noted, I do not believe that these are the only ethi-
cal prototypes we have. Rather, I assume that we have a number of
ethical prototypes, one or another of which is likely to be triggered
in any given context – much as different prototypes for happiness
are triggered in different contexts. Moreover, this is only a prelim-
inary formulation of even these prototypes. However, it does seem
clear that the conflict between these two prototypes forms the basis
for the traumas of heroism. This is unsurprising as the prototype of
group defense is the one most obviously relevant in the context of
a heroic narrative. Moreover, the prototype of compassion stands in
direct conflict with the prototype of group defense – and it too is
particularly likely to arise in “heroic” contexts, such as war.
All of this leads us back at last to the initial problem, for the con-
flict between these two ethics provides at least a preliminary account
of the anomalous second ending of heroic tragi-comedy. Specifically,
the despair and punishment found in the epilogue of suffering are re-
peatedly focused on or directly related to particular killings or other,
usually irreparably harmful acts. These acts were performed in the
group protection. Gilligan’s opposition of justice and care is, in my view, a false
dichotomy. Second, Gilligan’s division is not prototype-based. Third, my division
is not intended to be complete. I believe that there are numerous ethical prototypes
that may be triggered by context. Finally, my division is not in any way related to
gender divisions, except via stereotypes. One is not a male way of thinking about
ethics; one is not a female way of thinking about ethics. I have argued elsewhere
that Gilligan’s case for a gender-based ethical division is not plausible even on the
basis of the data she cites (“Some Prolegomena” 250–4). Indeed, the politics of iden-
tity implicit in Gilligan’s analysis falls into my first category, which Gilligan would
presumably label “male.”
name of an ethics of defense, but they were rendered traumatic by

their incompatibility with an ethics of compassion. The epilogue of
suffering is a sort of attempted moral rectification, a period of mourn-
ing, remorse, and punishment that serves to balance the ethical terrors
that preceded. For the hero, the author, and the reader are not unaf-
fected by those conflicts. These traumas can undermine the heroic
triumph; they cry out for reparation.
In some cases, this relation is straightforward, as when Kumagae
becomes a priest as a direct result of his murder of Atsumori. In
other cases, the link is less direct. Gilgamesh goes in search of
Utnapishtim in response to the death of Enkidu and the consequent
realization of his own mortality. Even this is not irrelevant, because
the warrior lifestyle chosen by Gilgamesh is ultimately responsible
for Enkidu’s death. But there is a more important and direct eth-
ical conflict here. Enkidu dies because he killed Humbaba. In the
Hittite version, Humbaba begged to be spared and Enkidu argued
against it (Gardner and Maier 11). Indeed, after a touching appeal
to Gilgamesh, Humbaba took Gilgamesh “by the hand and led him
to his house, so that the heart of Gilgamesh was moved with com-
passion” (Sandars 24). Enkidu’s death and Gilgamesh’s suffering
derive directly from this exemplary moment of ethical conflict be-
tween a murderous ethic of defense (“He will bar the mountain road
against you . . . this Humbaba must die” [Sandars 24]) and an ethic
of “compassion.”
Other instances do not involve explicit causal connections, but
more indirect links. The epic of Mwindo provides a good example.
Mwindo is punished by Nkúbá, putatively for killing a dragon who
was Nkúbá’s friend (137). This does not appear to involve any ethi-
cal element as the dragon was himself killing and eating people. So,
either this scene of suffering is nonethical or there is some other rea-
son for Mwindo’s painful sojourn through the sky. Biebuyck argues
that the punishment is the result of Mwindo’s own ethical failings,
particularly his boastfulness (145). This is to some extent true. How-
ever, Mwindo’s hamartia is more importantly a particular instance of
the sort of ethical contradiction we have been considering. Earlier in
the poem, Mwindo had himself called on Nkúbá for aid. Specifically,
as part of his heroic battle against usurpation, he had Nkúbá – that
is, lightning – descend from the sky and destroy a village, Tubondo,
his own natal village. This is in a sense recapitulated after Mwindo

kills the dragon. When he returns to the village with the corpse of
the dragon, several people say, “he who has killed this one cannot
fail to kill one of his relatives” (133) – just what Mwindo did in de-
stroying Tubondo. Mwindo responds by killing these people (134). It
seems clear that when Nkúbá descends to punish Mwindo, this is an
indirect retribution for Mwindo’s use of Nkúbá/lightning to destroy
the village, and his subsequent behavior toward those who reminded
him of the earlier violence.
To say that there is a conflict of prototypes in these cases is not, of
course, to say that the conflict is solely, or even primarily, conceptual.
It is a conflict that is felt – by the hero, by the author, by the audience. It
is a cognitive conflict pervaded by the fear and pride that animate the
ethics of group protection and by the compassion that animates the
ethics of comforting the miserable. Indeed, it is feeling that makes
the conflict traumatic; it is feeling that necessitates the mourning and
punishment. But there is a cognitive structure enabling emotional re-
sponse in these cases. The conflict arises when the antiheroic ethics of
aid become salient in the relevant episode, when we feel compassion
for the “enemy.” Insofar as the ethics of group protection go unchal-
lenged, even cruelty on the part of the hero will not prove traumatic.
Insofar as we respond to Rāma’s murder of the unarmed Śūdra boy
as a protection of society against the violation of social hierarchy, in-
sofar as the opposed ethical prototype and its associated compassion
are not triggered, we will simply accept Rāma’s act. But what is the
difference between this case and the case where the murder does con-
stitute a trauma, the case where compassion and its ethical prototype
trouble the fear and pride linked with the ethics of group protection?
This brings us to our final topic.
two types of empathy and the morally

conflicted world
In some ways, the answer to the preceding question is easy. An ethics
of compassion is triggered by empathy. Concerns of compassion be-
come salient when we can adopt the other person’s point of view,
when we can think what it means to be in his/her position. Adopting
another person’s point of view is a dangerous thing, at least given
the combative, usually military goals of heroic narratives. It creates

feelings and attachments. It leads us to think of this other person in
moral terms, as someone to whom we have obligations, as someone
who has rights – in short, as someone like ourselves. As Davis notes,
in a number of experimental studies, “instructions to imagine the
affective state of a target frequently trigger a process which ends in
the offering of help to that target” (145) or, depending upon the ex-
perimental circumstances, in “inhibit[ing] aggressiveness” (162). The
effects are weak, and limited in extent, but still significant.
Of course, the mere experience of empathy cannot be all there is
to the matter. If I adopt Rāma’s point of view when he is killing the
Śūdra boy, that empathy will not trigger the prototype of compas-
sion, but reenforce the ethics of group protection. More generally, not
all empathy is the same. In every case, empathy is based on some
sort of similarity. After all, to empathize with someone is to put one-
self in his/her place, and that substitution presupposes something
that is shared, something that enables the “fusion of horizons,” as
Hans-Georg Gadamer would put it. But there are different types of
similarity from which empathy may grow. I shall distinguish two,
roughly parallel with the two types of ethics. I shall refer to these as
categorial empathy and situational empathy.
Each of us has some sort of self-concept. We can conceive of the
self-concept as a lexical entry with a hierarchized list of features, in-
cluding character traits, physical properties, and so on – in short,
everything a person thinks is true of him/herself, ordered by impor-
tance or centrality to his/her self-understanding. For example, I am
bearded, a teacher, living in Connecticut, male, and so on. I see my-
self much more centrally as male and as a teacher than as bearded or
living in Connecticut. Put differently, I can far more readily imagine
myself shaving or moving to another state than changing profession
or sex. Thus, all these features are included in my lexical self-concept,
but “male” and “teacher” are placed higher in the lexical hierarchy
than “bearded” and “living in Connecticut.” Each feature in this self-
concept defines some group membership for me. “Man” defines me
as a member of the group, “men”; “right-handed” defines me as a
member of the group “right-handed people.” Some of these group-
defining features are particularly salient, particularly high in the
default hierarchy of the lexical entry, primarily because they have a
function in society. Sex is an obvious example – as are race, nationality,

ethnicity, and so on. It is well established that we oppose in- and out-
groups based on such category-defining traits, and the delimitation
of such groups has deeply significant consequences in terms of both
evaluative judgment and practical discrimination. When “subjects
are asked to allocate rewards (or punishments) between ingroup and
outgroup members, they do so in a manner that maximizes the dif-
ferential between ingroup and outgroup even though this may re-
duce the absolute benefits to the experimental subjects or even to the
ingroup” (Duckitt 68–9). Worse still, when people “are given the op-
portunity to discriminate” against out-groups, they “show increased
self-esteem” (85). It is just this sort of collective self-definition of an
in-group, and opposition to an out-group, that provides the basis for
the social prototype of happiness as group domination. In this pro-
totype, the in-group is, in effect, a version of oneself – a set of people
with whom one shares a definitive self-concept feature (for exam-
ple, “White,” “Black,” “Hindu,” “Christian,” or whatever). I shall
use the phrase “identity category” to refer to a self-concept feature
that serves to define such an in-group, which is to say, a group with
which one shares “categorial identity.” This sort of group definition
undergirds the ethics of group protection, for the group to be pro-
tected is, precisely, the in-group (national, religious, or whatever). I
shall use the phrase “categorial empathy” to refer to empathy based
on this sort of identification.
The lexical entries that record our self-concepts include not only a
set of group-defining traits, but also an archive of particular memories
and their various associated feelings. Indeed, these memories promi-
nently include experiences imbued with unusually strong emotions –
positive or negative, happy or sad, wondrous or terrible – for those
are the events we find most “memorable.” These memories, too, may
trigger identification through a sort of structural mapping. A com-
plex similarity of new events may activate or prime these memories.
As discussed in Chapter 2, that activation or priming will, in turn,
reproduce the original feeling, in a more or less attenuated form.
More exactly, a personal memory involves a structure with some
sort of perspectival position – my consciousness, feeling, aims, and so
on, at the moment of the memory (my tenth birthday, my wedding,
or whatever). Each current situation recalling such a memory also
involves a perspectival structure. The simplest activation or prim-

ing of a memory results from a current, “ego-focused” experience,
an experience in which the person in question maintains the same
perspectival position in the memory and in the current situation.
Doe, who was once run over by a car, suddenly hears a car screech
to a halt very close by. The screech of the car serves as perceptual
probe to activate elements of his/her procedural memory system –
most obviously, response schemas (“jump out of the way”) – and
to prime or activate relevant autobiographical memories, along with
their associated affects. (Procedural memory records knowledge of
how to do things, such as ride a bicycle [see Schacter 17].) In this case,
Doe is at the perspectival center of both the memories and current
experience.
In the cases we wish to understand, this sort of priming or
activation results from nonego-focused experiences. Specifically, a
triggering of autobiographical memories may involve a shift in the
structuring perspective that leads us to put ourselves in the place
of someone else and thus results in empathy, associating our own
feelings with those of the other person. Suppose Doe hears a car
screeching to a halt, but it is at a distance. He/she looks up and sees
someone else in the path of the car. The screech is still likely to serve
as a probe. It still activates relevant memories, including the memory
of being run over, along with its associated feelings. But, in this case,
the perspective will be shifted from Doe to the person in danger of
being hit by the car – or, as Gadamer would put it, their perspec-
tives are “fused” – leading to empathic fear. Doe tacitly maps his/her
memory onto the current situation, but his/her perspectival place in
the memory is now taken by the other person, who is in the path of
the car. This sort of mapping is no doubt facilitated by the fact that we
reconstruct our memories from partial and fragmentary information,
which is always to some extent integrated with aspects of current
experience, as Schacter has shown (see 8, 40, and 104–13). Indeed,
in this way, Gadamer’s term, “fusion,” may be more accurate than
“mapping.” In any case, I shall refer to this fusion or mapping with
a perspectival shift as “situational” empathy. It is situational empa-
thy that animates the ethics of compassion, and thus it is situational
empathy that is crucial to understanding traumas of heroism and the
epilogue of suffering.
Readers familiar with Turiel’s work on the development of

moral judgment in children will immediately see a connection here.
Specifically, situational empathy is closely related to Turiel’s view
that “the child will judge the presence of a victim as undesirable if he
has connected his own experience of pain with that of the victim” and
that “experience of [one’s] own pain” makes “the observed experi-
ence of the victim aversive” (Blair 7).7 Moreover, this is precisely the
sort of empathic identification discussed by Halász with respect to lit-
erature. Specifically, Halász explains “contact between the reader and
the protagonist’s emotions,” which is to say, a “reader’s empathy,”
by reference to “the emotional memory network” and the activation
(more precisely, priming) of “relevant events and sources which are
embedded in an autobiographical context” (“Effect” 83). The point
also fits the analysis of “personal resonance to and engagement in”
literature by Larsen, László, and Seilman (“Across” 102). As noted in
Chapter 2, these researchers found that such resonance and engage-
ment were in part a function of “remindings,” specifically remindings
of “personally experienced, and therefore self-relevant, events” (102).
But there is a problem here. In the case of Doe and the screeching
automobile, the generation of empathy seems clear enough. But what
about cases of empathy that do not derive from a shared particular
experience – as when Jones fears for the person in the path of the
screeching automobile, just as Doe does, even though Jones has never
been in an automobile accident. Or, stranger still, what about the
many literary cases where the work leads to empathic identification,
though the reader has never been, say, a deposed monarch or an
exiled knight?
The first thing to say here is that no two people have ever had
precisely the same experiences. Every situation is singular. My expe-
riences may be more or less similar to yours, but they cannot be one
with yours. Empathy, then, is not a transfer of feeling across identical
experiences. All empathy must be to some degree analogical. It must
be a matter of experiences that are similar – in structure, consequence,
intensity, and so on. Jones may empathize with Smith because they
have both had surgery for colon cancer or because they have both had
7 I am grateful to Tamar Szabo Gendler for drawing my attention to the work of Turiel
and Blair.
surgery for cancer of different sorts or because they both have had
surgery of some sort or because they both have been in the hospital
or because they both have experienced severe physical pain from
illness or have had reason to fear a premature death. How this ana-
logical mapping works is obviously a difficult issue. In most cases, it
probably involves multiple memories, interacting with one another
and with probes from the current situation in complex ways (all no
doubt underwritten by universal and innate propensities, perhaps of
the sort discussed by Blair). Clearly, a treatment of these intricacies
is beyond the scope of the present chapter. The crucial point here
is that this form of empathy is not based on categorial identification.
Rather, it is based on some tacit comparison of experience, a structure
of memories that is spontaneously mapped onto another person’s ex-
periences in such a way as to identify one’s own perspective in the
memory with the other person’s current perspective – or (following
Schacter’s research) even a structure of memories that is actually re-
constituted from fragments in such a way as to incorporate that other
person’s perspective as one tacitly imagines it.
But this is still not all there is to the matter, for not all compara-
ble situations trigger mapping and empathy. Sometimes mapping is
blocked. Most obviously, it is blocked whenever I find myself so dif-
ferent from the other person – in general, or in a particular case – that
a Gadamerian “fusion of horizons” appears impossible or repugnant.
Categorial opposition between an in-group and an out-group often
functions in this way, constraining empathy across race, religion, na-
tion, and so on, even for people who are not self-consciously bigoted.
Moral evaluation can serve this blocking function as well. We tend to
have compassion for those who are “more sinned against than sin-
ning,” but not for those who “receive their just deserts.” Indeed, part
of the meaning of these phrases concerns empathy; the former marks
its referents as deserving of empathy, the latter as undeserving. The
point was indicated by Aristotle, in the notion of hamartia, for he in-
sisted that the hero cannot be wholly evil because the downfall of
such a character does not inspire pity (45; 1453a). More generally,
no trauma of heroism will occur insofar as we judge the villain so
distant from ourselves – in categorial identity, in moral character, or
whatever – that any situational identification is blocked, presumably
because any even tacit imagination of his/her perspective is blocked.
Saying this, however, makes our initial dilemma appear only

worse; heroic tragi-comedy now seems still more anomalous. The
epilogue of suffering does appear to be a sort of reparation for the
traumas of heroism. Traumas of heroism, in turn, appear to result
from the conflict between the ethics of group protection and the ethics
of aiding the miserable, a conflict that arises in the course of the heroic
narrative along with the development of compassion. Finally, com-
passion is triggered by situational empathy for suffering. But why is
situational empathy there in the first place, given that it almost in-
variably concerns the enemy? Our empathy with such figures should
be blocked both by categorial identity and by their moral depravity
as villains.
Unsurprisingly, things do not work out that simply. The sorts of
self/other opposition that block empathy – including categorial and
moral irreconcilability – are typically based on a reification of the
other person. To the extent that this reification is undermined, the
empathy-blocking opposition may be undermined as well. More ex-
actly, the empathic adoption of someone else’s perspective involves,
among other things, a shift from what one might call “objective mind”
to “subjective mind.” What I mean here is that we may tacitly con-
ceive of the human mind as a sort of ghostly object or thing, de-
fined, in effect, by a list of specifiable properties, mechanisms, and
so on. In doing cognitive science or other sorts of psychology, we
often necessarily treat the generic human mind in this way. At times,
in daily life, we also treat individual minds this way – typically,
the minds of people we do not know well. The alternative to this
reified view of mind is, roughly, phenomenological. We may under-
stand the mind as pure subjectivity, ongoing experience, idea, feel-
ing, and memory. If objective mind is quasi-spatial, subjective mind
is purely temporal. If objective mind manifests a limited set of traits
and procedures, subjective mind is either almost traitless – a contin-
ual, flexible negotiation with changing circumstances (see Holland
et al., 222–4) – or it has a limitless variety of traits, all mixed-up
and contradictory, good and bad, cheerful and melancholy, smart
and stupid (see Davis 97–8). For ourselves, we are always subjective
mind. Others become subjective minds for us as we come to know
them, as we come to see them not as mechanically enacting fixed
character traits, but as engaging mutable conditions in variable and
open ways, or as combining an array of irreconcilably contradictory

traits.
Like all character predicates, ethical predicates, especially negative
ethical predicates – “bad,” “disloyal,” “selfish,” and so on – apply
most readily to objective mind (see Davis 97). The more subjective our
understanding of someone, the more difficult it is to detach ethical
predicates, especially negative predicates, from his/her particular
acts and apply them to his/her “character.” In literary terms, the more
we come to know someone, the more we understand and respond to
him/her as subjective rather than objective mind, the harder it is
for us to view him or her as a villain. A pure villain – or, for that
matter, a pure hero – is almost necessarily a simple list of objective
traits. Familiarity pushes toward the middle, setting limits on both
idealization and vilification, elevating the bad guys, and diminishing
the good guys. A parallel point holds for group identities. The more
we come to know someone, the more difficult it is for us to reduce
him/her to an instance of some collective category – racial, national,
and so on. As Holland et al., explain, “social-category . . . values are
typically suppressed in the face of individuating evidence” (221).
Literature, to create its effects in the first place, must develop em-
pathy with the hero. But, in developing the subjectivity of the hero,
a work not only creates the possibility of empathy – it may simulta-
neously limit the possibilities for heroism. More significantly, when
an author elaborates on the action or condition of an antagonist, a
villain, or an enemy, he/she almost necessarily begins to develop
a subjectivity for that antagonist, thus dulling the contrast with the
hero. In this way, the sorts of opposition that would ordinarily block
empathy tend to be undermined through ordinary literary exposition.
Villainy in particular is difficult to sustain. Each new discussion of the
villain renders him or her more subjective, and thus less villainous.
Georg Lukács’s well-known distinction between reifying reportage
and “dialectical” portrayal fits here. An author may report objective
facts about a character – hero or villain. But an author might equally
develop the subjectivity of a character, “portray” him/her, painting
“a picture with which we can empathize” (57). The point here is that,
at least in paradigm heroic plots, what begins as reportage regard-
ing the enemy often develops into an empathy-triggering portrayal,
perhaps even against the author’s intent. We might refer to this as
the spontaneous tendency to subjectivize, and thus undermine, villainy.

Many heroic tragi-comedies provide examples. To take one, H. Paul
Varley remarks that, “In the second half of the Heike the Taira, hereto-
fore despised, increasingly elicit our sympathy, as they suffer defeat
after defeat at the hands of the Minamoto” (301).
But our response in these cases is not solely a matter of greater
fullness and detail in characterization of the enemy. Varley stresses
the suffering of the Taira in explaining the reader’s change in atti-
tude, and he is right to do so. Given the preceding account of ethical
prototypes, it is to be expected that the portrayal of suffering in par-
ticular should tend to activate an ethics of compassion. The ethics of
compassion apply most naturally to situations in which one is faced
with a suffering and debilitated person – or, more generally, some-
one who is weaker, lesser in strength, weaponry, health, and means.
When the Other is represented as a threat, that representation tends
to activate our prototype ethics of defense. But when the Other is rep-
resented as weak and miserable, that tends to activate our prototype
ethics of compassion. Thus, it is precisely in those stories where the
hero triumphs that we are likely to feel the greatest conflict – for it is
in those stories that the “villain” experiences the greatest and most
acute suffering. This is almost a necessary result of the genre itself,
of the elaboration of the villain’s acts and condition, and of his/her
pain.
The point is illustrated with near allegorical precision in a work I
have not discussed in this context, a work that is not even primarily
heroic, but romantic: Niz.āmı̄’s Laylā and Majnūn. This poem includes
one heroic section in which Majnūn recruits an army to battle Laylā’s
tribe, so that he may be united with his beloved. But as soon as the
battle begins, Majnūn regrets it deeply. Niz.āmı̄ describes Majnūn’s
condition in words that (at least in the English translation) could ap-
ply equally to many writers: “While each warrior thought of nothing
but to kill the enemy and to defend himself, the poet was sharing
the sufferings of both sides” (58). The conflict between ethical sys-
tems is obvious, and unmitigated. It need not be that the poet is a
better person or even necessarily more empathic in general. But, in
imagining, detailing, elaborating the conflicts of heroic tragi-comedy,
in portraying the antagonists and their condition, he/she almost in-
evitably comes to feel the humanity of all those involved – including
those enemies who suffer irremediable pain at the hands of the

hero.
But this too is not all. In addition to this unintended develop-
ment of subjectivization and compassion, many authors of heroic
tragi-comedies, including authors of paradigm canonical texts, seem
to have set out to cultivate situational empathy with the enemy –
empathy that runs counter to the categorial empathy an original
reader might have been expected to share with the hero. A work
such as Euripidēs’ Trojan Women is a particularly obvious case. In this
play, Euripidēs systematically cultivates situational empathy with the
Trojans against what he could assume to be his audience’s sponta-
neous inclination to identify categorially with the Greeks. Indeed,
Euripidēs did this for particular political purposes – to condemn
the brutal treatment of Mēlos in the then-ongoing Peloponnesian
War, and to protest the Sicilian expedition and the jingoistic enthusi-
asm of his fellow Athenians (see Hadas 173), an enthusiasm based,
of course, on categorial identification and profoundly devoid of
compassion.
But Euripidēs is far from the only case. Here I should like to re-
turn to one final oddity found in the epilogue of suffering – the
“transcendental” conclusion with its motif of familial reunion. As
I noted, this reunion responds to a prior severing and opposition
which is part of the trauma of heroism. Why is this opposition and
reunion so consistently familial? This pattern results I believe, at least
in part, from authors’ attempts to cultivate empathy. I do not mean
that authors self-consciously decide to create empathy, make a list of
possible techniques, and so on. Rather, I take it that the authors set out
to achieve a certain sort of effect, an effect which includes empathy,
whether they explicitly recognize this or not. As they try different
techniques, some seem to work better than others. One frequently
successful technique is the identification of hero and enemy as mem-
bers of one family, literally or metaphorically, for this “familializing”
of enemies runs strongly contrary to any strict categorial division that
might block empathy.
Striking instances occur when the trauma involves the hero killing
a close relative. To protect the kingdom, Rostám kills his son Sohráb.
Mwindo calls down lightning on his natal village, killing every-
one, presumably including his own siblings (90–2). The Book of Dede
Korkut ends with Kazan’s killing of his uncle. The Mahābhārata is a

great war between cousins, and the horror of this familial slaughter
is a continual theme in the poem, as when the Pān.d.avas discover
that, in killing Karn.a, they have killed their eldest brother (see Vyāsa,
vol. 7, “Stree Parva” 45–6). This structure, opposing close relatives
across lines of battle, tends to limit the possibilities for dichotomizing
hero and enemy, and tends to enhance the subjectivization of that en-
emy by stressing his/her similarity with the hero. It also enhances our
empathy with the guilt and remorse of the hero in the epilogue of suf-
fering, for we all know the particular pain of hurting those to whom
we are closely related – brothers, sisters, parents. Finally, in keeping
with the preceding point, this familial structure sharpens the moral
conflict that generates the epilogue of suffering, for we generally
believe that people have stronger moral obligations to their family
than to strangers (as when we believe that we have a greater obli-
gation to feed our own children than to feed other people’s children
[see Sommers]).
The empathic function is even more obvious when the connection
is metaphorical or analogical, rather than literal and direct, for in
these cases the empathic aim is often explicit. Consider, for example,
Prı̄am’s appeal to Achillēs when they first meet after Hector’s death:
“Remember your own father, most noble prince Achilles, an old man
like me near the end of his days” (291). This is straightforwardly an
appeal for empathy. It is an appeal from one character to another, but it
operates also on the reader. Indeed, when I read this passage recently,
I thought of my own father. The situational similarity invoked by
Prı̄am served to subjectivize Prı̄am, not only for Achillēs, but for me.
Or consider the encounter of Gilgamesh and Humbaba. Humbaba
explains “I have never known a mother, no, nor a father who reared
me” and asks Gilgamesh to be his “lord.” Then he takes Gilgamesh
“by the hand” and leads him “to his house” – what should be a
familial place. Gilgamesh is thus assimilated to Humbaba’s father.
As a result, Gilgamesh is “moved with compassion” (Sandars 24).
Perhaps an even more striking case may be found in The Tale of the
Heike. When Kumagae sees Atsumori, he thinks of his own son, and
transfers the feeling to Atsumori’s parents, a transferal that makes the
situational empathy explicit: “When I think of how I grieved when
Kojirō suffered a minor wound, it is easy to imagine the sorrow of
this young lord’s father if he were to hear that the boy had been
slain” (317).
conclusion
In sum, heroic tragi-comedies often include an “epilogue of suffer-
ing” in which the story continues beyond its expected conclusion.
This epilogue is focused either on the misery of those who are van-
quished or on the anguish of the victorious hero, who surrenders the
domination he/she has won, suffering remorse or undergoing some
punishment. This anomalous “second ending,” following the heroic
triumph, involves ethical concerns. Specifically, heroic plots regu-
larly manifest a conflict between two ethical prototypes, one based
on group protection, one on individual compassion. Due to the mar-
tial circumstances of the story, the hero most often acts in accordance
with the former, violating the latter. The epilogue of suffering is, in
effect, a reparation for that choice. This reparation is not necessitated
simply by abstract ethical conflict, however. Rather, the conflict in-
volves us empathically. Here too a distinction is in order, in this case
between “categorial empathy,” which is based on shared group iden-
tities (for example, nation or race), and “situational empathy,” which
is based on shared experiences, especially experiences of suffering.
Ethical and empathic conflicts are unusually likely to arise in heroic
tragi-comedy because the violent group antagonism is apt to trigger
categorial and defense-oriented responses, while the violent resolu-
tion of that antagonism, including the defeat of the enemy, may trig-
ger situational and compassionate responses. This is particularly true
if the author develops a fuller portrayal of the enemy, for that tends
to humanize the enemy and thereby to draw our empathy when that
enemy suffers.
Thus, the anomalies outlined at the beginning of this chapter
seem quite comprehensible, given a fuller understanding of empa-
thy, ethics, and the development of literary plot and character. There
is, however, one further implication of this analysis that it is impor-
tant to draw out. The preceding discussion indicates that there is a
contradiction, or potential contradiction, at the center of heroic nar-
rative itself, because there is a contradiction implicit in the prototype
of social happiness. Horror, it seems, is an inevitable concomitant of
individual or group domination. The traumas of heroism are akin to

killing one’s own son or father, for they are a matter of killing some-
one’s son or father. But they are not the evil product of a particular
reified character. They result, rather, from a general condition. Specif-
ically, the traumas of heroism arise in a social world that allows, in-
deed drives, inexorably toward conflicts between our two sorts of
ethics: a social world created whenever powerful men and women
pursue a prototype of happiness as individual or group domination.
In “Moral Dilemmas and Consistency,” Ruth Marcus argues that one
of our most important ethical duties is “to conduct our lives and
arrange our institutions so as to minimize predicaments of moral
conflict” (121). The traumas of heroism and the epilogue of suffering
show over and over again that the pursuit of individual and group
domination violates that duty. In short, the apparent anomalies of
heroic tragi-comedy are not anomalous at all, but the direct result
of that genre’s origin in a faulty prototype for social happiness, a
prototype that creates a morally conflicted world.
5
Extending the Theory

Emotion Prototypes, Narrative Junctures,
and Lyric Poetry
The literary importance of the emotive and plot structures and princi-
ples discussed in the preceding chapters would appear to be confined
to explicitly narrative literature – stories, plays, novels, epics. This is
hardly a narrow scope. Narrative literature is vast and there is no rea-
son to expect that an account of explicit narratives could be extended
to another area of literature. On the other hand, such an extension
would give greater force to the initial account and would further its
research program in valuable ways. In the following pages, I shall ar-
gue that we can in fact extend these structures and principles to lyric
poetry. Indeed, in the conclusion to this chapter, I shall argue that the
preceding theory can be extended even beyond literature to aspects
of our ordinary lives, such as religious belief, and that the narrative
study of lyric poetry helps us to see this.
lyric poems as implicit plots: the narrative

hypothesis
More exactly, we tend to think of narrative and nonnarrative verbal
art as sharply distinct. However, the structure of plots and the uni-
versal features of lyric poetry indicate that this is a misconception.
In particular, lyric poems are commonly set off against narrative, op-
posed to it. For example, one standard handbook defines a lyric poem
as “A poem, brief and discontinuous, emphasizing sound and pic-
toral imagery rather than narrative” (Frye, Baker, and Perkins 268).
152
Emotion, Narrative Junctures, Lyric Poetry 153
However, I shall argue that lyric poems are not separate from nar-
rative. Rather, lyric poems are elaborations of junctural moments in
narratives and are governed by the same general cognitive principles
as stories. These narratives are, of course, implicit and necessarily to
a degree vague. However, they are crucial to the sense and conse-
quence of lyric poems. Put differently, a lyric poem prototypically
involves a focus on a particular emotion (see, for example, Padgett
111), paired with a particular event or brief sequence of events. But
a lyric poem is not thereby monadically encapsulated in a moment.
Rather, the event in a lyric poem draws its meaning and affective force
from its tacit relation to a prototype for an outcome emotion and its
location within a narrative structure implied by that prototype. More
exactly, prototypical lyric poems are tacitly located at junctural mo-
ments of heroic or romantic tragi-comedies and imply the emotion
prototypes for those genres.1 I shall refer to this as the “narrative hy-
pothesis.” My claim is that this relation of lyric poems to prototypical
narratives and emotion prototypes is an absolute universal. As with
1 The general relation between lyric poetry and narrative has not gone unremarked.
A few critics have commented explicitly, if very broadly, on the emplottedness of
poems. For example, Voigt notes that “Stephen Dobyns has said that every lyric
poem implies a narrative.” However, the point has hardly been developed, or it has
been developed differently. Thus Voigt explains Dobyns’ idea in the following terms:
“What he means is a sequence of past events, left out of the poem, that brought the
speaker to the present intensified moment in the poem. Assuming the speaker is the
poet, Dobyns also means autobiography” (725). The connection here is obviously
a very general one, little more than a matter of causal precedence, which does not
necessarily merit the title “narrative” and which is, in any case, not confined to pro-
totype structures.
Some other writers have suggested a link of this sort through more historically
and textually particularized studies. For example, Sylvia Huot has examined the
complexities involved in the use of motifs from Troubadour lyric poetry in Old
French narratives. Though this “transposition of lyric into narrative” (263) is differ-
ent from what we are considering, it is not unrelated. The fact that such transpo-
sitions were common suggests again that there was an implicit link between lyric
poetry and narrative to begin with, a link that enabled authors to undertake the
transpositions.
Finally, a few writers have drawn broader connections that could be taken to
imply a lyric poetry/narrative link. For example, Sarbin has emphasized the em-
beddedness of all imaginings in implicit narratives, and lyric poems are certainly
imaginings. On the other hand, the precise nature of Sarbin’s claim is somewhat dif-
ferent from mine. In particular, Sarbin’s use of “narrative” is both broader (“a script
with a beginning, a middle, and an end,” 21) and narrower (for example, in stressing
a person’s “self-narrative” as, say, a “survivor”).
explicit paradigm narratives, this is not to say that every individual

work is implicitly located within a heroic or romantic plot. Rather, it is
to say that, in every literary tradition, lyric poems are predominantly
of this sort.2
Once stated, the general point about the relation between lyric
poems and implicit narratives is almost obvious, at least in many
cases. Consider a poem by the eighth-century Chinese poet, Wang
Wei, “Written Crossing the Yellow River to Ch’ing-ho” (Mack et al.
1,304). The poem begins with an enigmatic reference, “The boat set
sail.” The presumption of familiarity here (as if we know just which
boat is in question) signals precisely an implicit narrative context. The
speaker goes on to describe how the water extends to the horizon.
Then, “suddenly,” he sees “a district capital.” He sees the market
and remarks on what goods he can discern. Then he turns “back
toward my homeland.” He can see nothing but water extending to
the horizon.
The poem is complex. It is highly allusive, as is standard in Chinese
poetry, and involves subtle Buddhist motifs. However, even with-
out understanding any of this, it is clear that the poem is a sin-
gle, junctural moment of beginning exile in an implicit narrative of
union, separation, and perhaps reunion. The speaker sees clearly the
homes of the capital, thousands of them – but none of them is his
own home. Looking back in the direction of his home, he can see
nothing but endless water. The poem thus recounts an isolated mo-
ment of alienation that has its effect by integration into an unstated
narrative.
Indeed, the poem foregrounds one common aspect of lyric poetry
that can give it a unique impact and force, distinct from most elab-
orated narratives: irresolution. The speaker of the poem is, in effect,
trapped in an isolated moment with no clear future. The narrative
structure itself implies that there are two and only two possible out-
comes: success or failure in returning home. But we have no indication
of which will be achieved.
2 My suspicion is that nearly all paradigm lyric poems are implicitly narrative. If
forced to give a rough estimate, I would say that two-thirds to three-quarters of
these poems imply one of the three prototype genres (including sacrificial tragi-
comedy, to be discussed in Chapter 6). That would be my rough estimate for the
percentage of explicit paradigm narratives in these genres as well.
Of course, not all lyric poems involve irresolution. Indeed, many

poems focus precisely on prototype outcomes – a point to which we
shall return. Others imply a resolution that has not yet been achieved.
Obvious cases of the last sort would include the psalms of lament in
the Old Testament. These regularly involve an implication that “the
Lord has heard the prayer” (Murphy 628) and thus that the narrative
conclusion will be positive. In other words, they are located in an
unspoken comic narrative – if a comic narrative that has not yet been
realized and thus has not yet reached resolution.
Indeed, the psalms fit the narrative hypothesis very well in
other ways also. The standard categorization of the psalms into
psalms of lament, thanksgiving, and praise, or their more recent
categorization into psalms of “orientation,” “disorientation,” and
“reorientation,” indicates that the narrative aura surrounding these
poems has been implicitly recognized by biblical scholars. These cat-
egories are, clearly, junctural moments in narratives. Moreover, the
relation of the psalms to prototype narratives is often straightforward.
Consider, for example, one of the best known psalms, number 137,
“Beside the streams of Babylon/we sat and wept” ( Jerusalem 920).
Like the poem by Wang Wei, this is an elaboration upon the begin-
ning of exile. Here, however, it is clear that the exile is the result of
a collective military defeat. It is the loss of group domination, and it
points toward the reestablishment of that domination at a future time.
Two further features of this poem are particularly significant in this
context. The first is that it clearly assimilates the relation between the
exile and his home or nation to the relation between a lover and his
beloved – for the exile takes a lover’s vows with respect to Jerusalem,
swearing that he will never forget her and will always count her as
his greatest joy. At a literal level, then, the poem is embedded in an
implicit heroic narrative. But it specifies the heroic juncture by ref-
erence to the other (romantic) prototype narrative. Second, the ulti-
mate group triumph toward which the poem points is a triumph that
explicitly includes the sort of violence that entails an epilogue of suf-
fering. Specifically, the poem ends with the distressing image of the
massacre of innocents: “a blessing on him who takes and dashes/your
babies against the rock!” One could interpret the poet as simply heart-
less. But one could also interpret the poem as part of a tacit story in
which the heroic triumph is, as usual, tarnished by cruelty. One could,
in other words, see the poem as much more ambivalent, or potentially

much more ambivalent – much more in keeping with a narrative that
includes a scene of sorrow after the final triumph.
lyric poems in narratives

The idea that lyric poems are embedded in narratives has been im-
plicitly recognized in some traditions through the ways lyric poems
have been joined and organized. The development of sonnet se-
quences in European literature is a case in point, for such sequences
often structure lyric poems into loosely plotted series of events –
most often romantic, as with Sidney’s Astrophel and Stella, Spenser’s
Amoretti (combined with the marital poem, Epithalamion), or
Browning’s Sonnets from the Portuguese. Some Sanskrit mahākāvyas
are of this sort as well. Though “mahākāvya” is often translated as
“epic,” these works are more akin to sequences of lyric poems with
narrative bridges, as many writers have pointed out. For example,
Dimock, Gerow, Naim, Ramanujan, Roadarmel, and van Buitenen
discuss the mahākāvya in the context of lyric poetry, not epic, and
emphasize that it is really a sequence of lyric poems (see 152–5;
Dimock et al. somewhat overstate their case, primarily because they
fail to see that lyric poetry and narrative are not opposed or mutually
exclusive). The renowned Gı̄tagovinda of Jayadeva is a case in point,
for it is largely a series of lyric poems treating moments in a story of
separated lovers.
Sequences of this sort are not the only places where we find lyric
poems incorporated explicitly into larger narratives. Such incorpo-
ration is perhaps particularly obvious in drama. Elite instances may
be found in Sanskrit drama, which regularly involves the utterance
of verses at key junctural moments in the plot – such as partings, re-
unions, and so on. When Śakuntalā leaves home to join her husband,
her adoptive father speaks the following lines, which constitute a
perfectly good free-standing lyric poem: “When I see the grains of
rice/sprout from offerings you made/at the door of your hut,/how
shall I calm my sorrow!” (Kālidāsa, Theater 132). Before sending his
brother to take Sı̄tā away and abandon her, Bhavabhūti’s Rāma recites
verses to the sleeping Sı̄tā, telling her that she should abandon him
for he has been made an untouchable by his “atrocious deeds”; he is
like a poisonous tree and she is like a vine that has used it for support,
though it will only prove deadly for her (14). The poem clearly elab-
orates on the sorrow and guilt felt by Rāma at this moment of eval-
uation, a moment that immediately precedes his separation from his
beloved and her exile due to a conflict with society. The first act of
Bhāsa’s Vision of Vāsavadattā establishes the heroine’s separation from
her husband in conjunction with her brief entrance into a hermitage.
The stop at the hermitage marks a clear point for Vāsavadattā’s and
our reflection on her loss of romantic union, her exile. The act ends
with a poem about the hermitage, explaining how “The birds have
returned to their nests” – precisely the sort of return home that is
now forbidden to Vāsavadattā – while “The hermits have plunged
in the stream,” and the smoke of the sacred fires “is spreading in the
penance-grove” (16).
The junctural placement of these verses is, of course, crucial. Again,
by the narrative hypothesis lyric poems, appearing alone, mark and
elaborate junctural moments in implicit plots, predominantly roman-
tic and heroic plots. In the cases just cited, the lyric verses perform
the same function, but the encompassing story is explicit.
Japanese Nō dramas also incorporate lyric poetry at such crucial
moments. Zeami’s plays, though short, often include many junc-
turally embedded lyric poems. In Lady Han, the main character’s
distress at being separated from her beloved is regularly expressed
through the quotation of poems from the Kokinshū (905 c.e.) and other
important early Japanese anthologies. The Reed Cutter is a particu-
larly striking case, for Zeami actually developed the play out of the
ambiguous narrative implicit in two poems from a tenth-century an-
thology, the Shūishū (see Keene, Twenty Plays 148).
Of course, the practice is not confined to such elite theaters. It is
equally obvious in Broadway and Hollywood musicals, where the
songs are lyric poems that punctuate the larger narrative. The same
point holds for the even more numerous musicals of Indian cinema.
Nor is this use of lyric poems confined to drama. A particularly
striking case of the incorporation of lyric poems into a nondramatic
narrative is found in Gurgānı̄’s Vı̄s and Rāmı̄n. This important Persian
work is itself in verse, but it incorporates “Songs of Rāmı̄n,” which
are “easily distinguishable” due to their distinctive use of rhyme
(Morrison xii). A good example of this is when Rāmı̄n is lying beside
his forbidden beloved, Vı̄s. He “dreads the break of day and sings of
the night” (xiii), a clear junctural moment. Similarly, a lyric poem en-
ters when Rāmı̄n weeps and bewails his separation from Vı̄s (Gurgānı̄
121); when he has just set off on a journey to meet her and is filled with
hope – proclaiming, for example, that “The way to union with you will
be short for me” (164); and when he finally sees her face again (171).
More salient instances occur in prose works. As Lewis points out,
the stories of The Book of Dede Korkut “are in prose interspersed with
rhythmic, alliterative, and assonant or rhyming passages” (14), which
is to say, lyric poems – or, rather, partial lyric poems, for not all of them
could be separated successfully from their context. A good example
occurs on the second page of the first story. A husband calls to his
wife, saying, “Come here, luck of my head, throne of my house,/Like
a cypress when you go out walking” (28). In the following five lines,
the speaker employs common conventions of love poetry in describ-
ing his beloved. It is a standard romantic poem, integrated into a
narrative – just at the junctural point when the speaker is consider-
ing the fact that he has no children. Another good example occurs
a few pages later in the praise poem concerning a character who
has just proven his strength and prowess (31). Subsequent stories in-
clude further types of lyric poetry, such as religious poems of worship
(137–8).
The practice is not absent from European literature. For example,
as Maureen Boulton has discussed, “lyric insertion” – the incorpo-
ration of lyric poems into other works – is “a technique present in a
substantial corpus of medieval French texts” (xiii). These poems pro-
duce a sort of “lyric ‘pause’” – again, a juncture – in the encompassing
“narrative movement” (xiv). Though not common in all European
Medieval narratives, this technique is also present in many Spanish
works (1n.1).
Chinese tradition makes extensive use of lyric poems in prose nar-
ratives, often via the convention of having characters compose poems
at key moments (a more mimetically plausible version of the com-
mon practice of having characters break into song or simply speak
poems). Alternatively, Chinese historical fiction incorporates poems
written or putatively written by contemporary poets involved in the
events or by later poets commemorating the relevant historical occur-
rences. For example, early in Three Kingdoms, an uprising is marked
by a “seditious song” – a four-line lyric poem. Subsequently, two of

the main heroes are celebrated by a “poet of later times” with a short
verse (Luo 10), as is the main hero, Liu Xuande (11). Not all lyric po-
ems are explained in these ways. Some are just inserted by the author.
For example, at the end of the first chapter, an interpolated lyric poem
addresses the reader’s evaluation of narrative movement. It tells us
that Zhang Fei will “pay out every ingrate what he’s due” just before
asking the question, “Did Zhang Fei kill the Imperial Corps comman-
der? Read on” (14; see also Brewitt-Taylor, vol. I, 10). Here we find an
explicitly marked case of a narrative juncture.
The Story of the Stone includes a number of authorial insertions of
this sort. For example, in one case, we encounter one of the two main
lovers in the novel. She is weeping and the author inserts a poem
that serves to elaborate on this junctural moment. He introduces the
verses simply by writing, “As the poet says” (Cao and Gao, vol. 4,
166). Of course, in keeping with standard Chinese literary practices,
characters also compose poems at such critical points. A good exam-
ple may be found later in the same volume. After his dear friend,
Skybright, passes away, Bao-yu (Magic Jade), feeling mournful, com-
poses a poem addressed to her. We read the poem, then he burns it
(201–2), the smoke from the paper rising to the heavens. More impor-
tantly, this novel makes explicit the connection between lyric poetry
and prose narrative. Early in the first chapter, one character com-
plains about the stereotyped romances that evidently plagued the
literary market in eighteenth-century China. The situation is reminis-
cent of North America today, with one difference – the prominence
of lyric poems in these narratives: “The trouble with this last kind
of romance is that it only gets written in the first place because the
author requires a framework in which to show off his love-poems.
He goes about constructing this framework quite mechanically,
beginning with the names of his pair of young lovers and invari-
ably adding a third character . . . to make mischief between them”
(50). Obviously this did not lead Cao or Gao to repudiate the integra-
tion of lyric poems into prose narratives. Quite the contrary. In fact,
this novel, like Three Kingdoms, concludes with a lyric poem (Cao and
Gao, vol. 5, 376; Brewitt-Taylor, vol. II, 622–3). Note that by our defi-
nition, the ending too is a juncture, for it is almost necessarily a point
of reflection and evaluation.
Many Japanese monogatari or “tales,” such as the Tale of Genji, also

integrate lyric poems directly into the narrative. Here too, characters
frequently compose poems at key junctures of particular emotional
intensity. For example, in the ninth-century Tale of Ise, we find Narihira
going out to inspect his estates and seeing two beautiful sisters. The
passion inspired by the sight of sexual beauty is a standard junc-
tural moment in a romantic narrative. Narihira immediately writes
a poem and sends it to the sisters (Keene, Anthology 64). We find
other examples of this sort in the Tosa Diary from the same period,
the Mirror of Increase, an historical romance from the fourteenth cen-
tury, and so on. Some striking cases are found in the works of Bashō,
the seventeenth-century master of the haiku. The Narrow Road of Oku,
for example, begins with Bashō setting out on a journey, a sort of
voluntary exile “that was perhaps to separate us forever.” To mark
this, he composes a poem in which birds and fish weep at the end of
spring. “I set out after composing this poem” (Keene, Anthology 348),
he explains.
Finally, it is not only authors who integrate lyric poems into larger
stories. At least in Japan, editors sometimes took on this task as well.
The love poems in the Kokinshū are integrated “into a longer nar-
rative sequence.” Specifically, the compilers “fashion[ed] a single
narrative out of the poems by many different authors.” In conse-
quence, “the poems chronicle the progress of a love affair” (Danly
2,076).
heroism and romance

It is crucial, of course, that the Kokinshū poems are integrated, not
into any old narrative, but specifically into a narrative of love. My
hypothesis is not simply that paradigm lyric poems are implicitly
narrative – though that is significant in itself – but that they are,
most often, tacitly a part of prototypical narratives. Thus, if the nar-
rative hypothesis is correct, then, cross-culturally, the largest number
of lyric poems – including, of course, the poems we consider most
prototypical – should fall into the romantic and heroic categories,
especially the former as that is, again, more basic and pervasive. This
does indeed appear to be the case, from the time of the earliest poems
written for emotive (rather than, say, magical) purposes.
The European tradition of lyric poetry extends back to two great

canonical figures who in effect embody this division: Pindar, the au-
thor of odes that are explicitly heroic in orientation, and Sapphō, the
author of renowned romantic poems. The prominence of romantic
poetry in the West hardly needs the be stressed. Interestingly, the
heroic genre was, if anything more prominent in ancient Greek writ-
ing, at least in what has been preserved. Our earliest extant poet,
Archilochos, was a soldier and proudly proclaims himself such in
the first fragment. Kallı̄nos wrote that “honor and glory bedeck the
man/who fights for his land” (Fowler 67). Tyrtaios tells us, “To die
is a fine thing when a noble man falls/fighting in the forefront of
battle on behalf/of his fatherland” (Fowler 77). This generic op-
position itself is marked nicely by some of Sapphō’s best known
lines, from fragment 16. She points out that some people find a
“host of horsemen,” “soldiery,” or “a fleet of ships” to be the most
beautiful thing on earth. But, for her, “it’s whomever one loves”
(Fowler 131).
Even more strikingly, classical Tamil poetry of South India, dating
from the first three centuries of the common era, is explicitly divided
into two sorts: first, “akam, or interior, poems that view life from inside
the family and concern the love between man and woman”; second,
“puram, or exterior, poems that view life from outside the family and
concern such topics as kings, heroism in battle,” and so on (Hart 3).
Just north of the Tamil writers, Sanskrit lyric poets focused their poetic
energies upon two junctural moments of the romantic tragi-comedy,
writing poems of sexual longing and union. Merwin’s and Masson’s
The Peacock’s Egg collects Sanskrit poems from 500 b.c.e to 1,000 c.e.
treating these junctures. The tradition extends well beyond that, from
such works as Bilhaṅa’s well known Fantasies of a Love Thief (eleventh-
century c.e.), to later works in Hindi and other Indic languages, such
as Bihārı̄’s Satasaı̄ (seventeenth century).
In pre-Islamic Arabic literature, there was a considerable amount
of poetry that fit implicitly within the heroic plot. Lichtenstadter ex-
plains that “Much of the surviving poetry has praise for the heroes
or scorn for the enemy as its main topic” (14). Indeed, the Arabic
qas.ı̄dah includes a particular section called “praise” that presents a
“picture of the ideal Arab hero,” “Virile, enduring, a valiant fighter”
(23). The marthiya is a distinct genre devoted to “mourning a fallen
hero” (23). Clearly, such celebrations of the hero and denunciations

of the enemy are an important feature of heroic tragi-comedies. One
need only think of the opening of Gilgamesh (“the man to whom all
things were known. . . . He was wise, he saw mysteries and knew
secret things”; he had “a perfect body,” and so on [13]) or of the
Rāmāyan.a (“a man . . . who has both virtue and power, who knows
the Law and has a lively sense of gratitude . . . who lives a blameless
life and has the good of all beings at heart” [Vālmı̄ki 1]). Indeed, in
some cases, segments of heroic poems are used directly as praise po-
ems. Speaking of the Epic of Son-Jara, Irele notes that “The many pas-
sages of praise poetry with which the epic is interspersed stand in a
close formal relationship to the formulaic plan of the narrative. . . . As
generally with praise poems in oral literature, they are composed of
strings of epithets, often hyperbolic, emphasizing the uncommon at-
tributes – the heroic essence as it were – of their subject.” Moreover,
they “are often performed as autonomous pieces” (2337). In short, the
praise and blame poetry of the Arabs and other peoples fall squarely
within an implicit heroic plot.
There is also a great deal of romantic poetry in Arabic, much of it
in the nası̄b which served as preface to the qas.ı̄dah. Arberry explains
that, while early Arabic poems emphasized “the martial virtues,”
the “lyric of love” came into prominence later (12). He goes on to
signal the narrative element, noting that Arab “poets developed a
kind of drama out of the pleasures and pains of love-making” (17).
Persian writers reworked the nası̄b into the ghazal, which focuses al-
most entirely on love. Levy explains that the term “ghazal” itself
“derived from an Arabic original meaning ‘lovers’ exchanges.’”
Indeed, as Ralph Russell points out, the ghazals typically involve
an implicit narrative in which two lovers are kept apart by the censo-
rious legalism of a Muslim shaikh, an “elder” with putative “moral
authority” who is the “implacable foe” of the “true lover” – a straight-
forward instance of the standard romantic plot of conflict between
the lovers and an interfering social representative. Levy notes that
Persian poetry “in general . . . concerns itself either with flattery or
love” (32–3), the flattery in question being largely political and, as
such, part of the heroic genre.
In keeping with this general pattern, ancient Egyptian lyric poetry
too prominently includes poems of praise – largely religious – and
love lyrics. The love poems treat moments of seduction, brief encoun-
ters with the beloved, and other junctures of the standard romantic
plot. For example, in one poem, “I think I’ll go home and lie very
still” (Mack et al. 59), the lover pretends that he is about to die, so
that everyone will come to visit, including his beloved – though she
will know it is all a ploy. This takes up the common motif of lovers
separated by death or near death, but twists it into a comic pretense,
understood and shared by lover and beloved.
Returning to China, we find poems of love scattered throughout,
from the Book of Songs to the writings of China’s renowned later poets,
such as Li Po (Li Bo). A uniquely well-known example among English
speakers is Li Po’s “Ballad of Ch’ang-Kan,” familiar in Ezra Pound’s
rendition, “The River-Merchant’s Wife: A Letter.” Though one could
name many relevant writers, a particularly apt case is Li Ch’ing-chao
(twelfth century), generally regarded as China’s greatest woman poet
(Ling 83). She has treated many junctures of the romantic plot, with
poems of falling in love, poems of union, poems of separation, and
poems on the death of her beloved – all collectively recounting a
romantic tragedy.
The heroic plot figures centrally in Chinese lyric poetry as well.
Indeed, Li Ch’ing-chao herself wrote poems in this mode, addressing
moments in the rightful emperor’s defeat and exile. In this context,
one might consider, for example, her “Poems on Yüen Chieh’s ‘Ode to
the Restoration of T’ang’ to Rhyme with Chang Wen-ch’ien’s Poem,”
one of which begins by referring to the reign of Ming Huang and “the
amazing story/Of . . . downfall and restoration” (57). The heroic plot
was taken up by the man widely viewed as China’s greatest male
poet as well – Tu Fu (Du Fu). Indeed, his most famous poem (Cooper
168) is a treatment of the aftermath of war, a lyric version of the
epilogue of suffering. In keeping with the structure of heroic tragi-
comedy, it begins with apparent enthusiasm, evidently celebrating
the noise of the advancing soldiers and their weapons. But it ends
quite differently, with “white bones” and “bitter” ghosts who “cry
aloud” (168).
Moreover, these are not isolated instances, as broader surveys
of Chinese literature show. James Liu distinguishes a range of im-
portant motifs in Chinese poetry: the sorrow of parting, the hor-
ror of war, nature, time, history, leisure, nostalgia, love, and wine
(Art 48–60). Love, parting, and war obviously fit our argument well.
The others do also. Liu’s example of time is, precisely, the passage of
time as experienced by separated lovers, while his example of nostal-
gia focuses on exile. The historical motif is, unsurprisingly, heroic. In
addition, Liu’s categories mark out just the sorts of junctural moment
we have been considering. Elsewhere, in a discussion of tz’u poetry,
Liu stresses “worlds in lyric poetry . . . dominated by romantic love”
(“Literary” 135). He also emphasizes “lyrics which embody heroic
sentiments” (144).
Waley organizes the Book of Songs into seventeen categories. Like
the items on Liu’s list, these are largely junctural. The first three are
“Courtship,” “Marriage,” and “Warriors and Battles.” Two further
sections address “Dynastic Songs” and “Dynastic Legends.” These
are often straightforwardly concerned with heroism – “Terrible in his
power was King Wu; None so mighty in glory” (230); “Oh, gloriously
did the king lead”; “He brought peace to myriad lands. . . ./Bold was
King Wu, Guarded and aided by his knights/He held his lands on ev-
ery side” (237). These are by far the longest sections and, collectively,
cover roughly two-thirds of the volume.
Bownas presents a catalogue of early Japanese poetic concerns,
showing the same tendencies: “Nature, love, partings, and time”
(lvii). In the important early anthology, the Kokinshū, the largest cate-
gory of poems focus on love, followed by poems on nature. (We will
consider the category of nature poems in the next section.) Together,
these comprise roughly two-thirds of the collection. The remainder
includes work devoted to such relevant topics – or junctures – as
“Felicitations,” “Parting,” and “Laments” (see Berggren 410). In his
preface to the Kokinshū, generally considered the inaugural essay
of Japanese literary theory, Ki no Tsurayuki comments that “Poetry
makes sweet the ties between men and women and comforts the
heart of the brave warrior” (quoted in Keene, Major Plays 201n.). The
statement is clearly designed to evoke the breadth of lyric poetry.
It tells us that lyric poems range precisely from the romantic to the
heroic. Brower and Miner quote many relevant poems. One romantic
instance, again marking the moment of separation, is Hitomaro’s “On
Parting from His Wife as He Set Out from Iwami for the Capital,” a
forty-nine line poem that expands this junctural moment, dwelling
on its pathos (116).
These patterns are not confined to Europe, Asia, and the Middle
East. George Cronyn begins American Indian Poetry with a sequence
of lyric poems from Abanaki, “The Parted Lovers.” The first line
reads, “My parents think they can separate me from the girl I love”
(3). Subsequently, he has love songs translated from Chippewa. The
first marks the separation of the lovers (21), while the last represents
reunion (24).
Trask’s The Unwritten Song includes two poems from the Trobriand
Islands. The first is a “Song of War,” covering “a fight, a death, and an
exile” (203), while the other is a “Love-Magic Spell” (205), thus a spec-
ification of the moment when the lover seeks to attract the beloved.
He includes two songs from the Kurelu (New Guinea). One is a “Dirge
for a Son Killed in Battle.” The other is a rather general lament of men
for girls they have lost. Of the next ten songs from New Guinea, one
concerns a courtship ceremony, another is the “Lament of a Widow
for Her Dead Husband,” and another six are straightforward songs
of longing for a beloved. Firth and McLean divide songs by the Poly-
nesian Tikopia into a small number of categories – songs of “travel,”
“eulogies and farewells,” “protest,” “erotic arousal,” “laments and
funeral dirges,” and “songs on historical and mythic themes.” A
number of the poems in various categories emphasize one or an-
other element in romantic and/or heroic stories (for example, exile –
see 129–30, or 182). The songs of “erotic arousal” clearly take up
the former. The most directly heroic poems are in the “historical”
section, which includes songs on “the midst of a battle” (257),
“Evading arrows” (261), “An attack in Vanikoro” (262), “Struggle
with invading Tongans” (264), and so on. The Yirrkalla poems
(in Berndt) treat sexual union almost exclusively, concentrating on
what is obviously a crucial juncture in the romantic plot, if a junc-
ture often passed over without explicit mention in more priggish
traditions.
Trask’s section on Africa begins with “Song After Defeat,” from the
Bambara. It includes poems of men and women who have just fallen
in love, poems of men and women who have been separated from
their beloved through exile, elegies and funeral songs, songs of exhor-
tation for battle, praise songs for rulers, and so on. Indeed, many ex-
amples could be cited from Africa. One particularly interesting – and
challenging – body of African work comes from the Dinka. Deng’s
collection of Dinka lyric poetry includes a range of heroic poems. War

songs are one of his main topical categories. Moreover, many of the
nonwar poems treat social domination. Thus, many poems that do
not strictly fit the heroic tragi-comic structure, nonetheless involve an
implicit or partially explicit narrative based in the social prototype
for happiness.
Even more interesting than the heroic songs of the Dinka are their
love songs, for these operate not only in a literary narrative, but
in life stories that are at least imagined to fit the standard struc-
ture of romantic tragi-comedy. Deng explains that poems “are of-
ten a basis for winning affection between young men and women,
and for imploring elders to respond to the demands of the in-
tending partners. They are also used by the partners themselves
to influence each other in order to strengthen their hands against
their opposing elders. Many are the cases in which marriages re-
sult from a song addressed to elders, the intended bride, or the
groom” (80).3
eulogy and nature: some complications of the

narrative hypothesis
Without actually admitting it, I have slightly modified the narrative
hypothesis from its original statement in order to admit praise and
blame poetry. Poems of this sort clearly fit into the heroic genre. How-
ever, they do so via character traits or physical properties rather than
narrative junctures per se. This addition is unsurprising, indeed pre-
dictable, for such elaboration of a character is an important part of
heroic plots, as we have noted, and it occurs at moments that are
almost necessarily junctural – for example, just before or just after a
battle. A second very important type of nonjunctural poetry is that
concerning nature. (Thus poetry focusing on the third component
of narrative, scene, rather than event or character.) This too is in
keeping with the spirit of the narrative hypothesis, though not its
3 A large number of Dinka poems concern cattle and thus do not clearly fit into either
the romantic or the heroic structure. This is not in itself a problem. Again, we are not
dealing with necessary and sufficient conditions here. However, we shall see in the
next chapter that there is a third prototype narrative, concerning hunger and food.
Many of the Dinka ox songs fit directly into that structure.
initial formulation, for patterns of nature imagery are a crucial part

of romantic tragi-comedy. A great deal of nature poetry is, in effect,
an elaboration of this imagery.
I mentioned above that Sanskrit drama often integrates lyric po-
ems into narrative sequences. Again, some excellent cases of this
are to be found in Kālidāsa’s Abhijñānaśākuntalam. Throughout the
play, key junctures are marked and expanded by verses. Some of
these make no direct reference to the story, focusing only on nature.
Thus, if they were printed separately, they would seem to have no
narrative element. However, in context, they are clearly bound up
with the encompassing story. For example, when the king is first
overcome with passion for Śakuntalā, there is a short poem about
a stampeding elephant. After a short separation, the lovers are re-
united, and the king recites a couplet: “Parched with thirst, the bird
has only to crave for water/and a shower from a fresh rain-cloud falls
into its mouth” (210). When Śakuntalā is leaving her parental home,
one of her friends utters another poem that consists only of three
images – a doe that does not eat, motionless peacocks, and faded
leaves (224).
Nature imagery is used in precisely the same way in much lyric
poetry. Li Po’s “Ballad of Ch’ang-Kan,” spoken by a young woman
separated from her beloved, makes reference to the thick mosses
by the door, suggesting that no one has walked through the en-
trance. She notes that butterflies are flying in pairs – the contrast
with herself walking alone is obvious. Without explicitly mentioning
the husband’s travel to distant places, such natural images operate
to suggest that absence, thus reenforcing the junctural element of
separation.
Of course, nature poetry is not always explicitly linked to narra-
tive. Nonetheless, these explicit connections suggest that there may
be implicit connections in other cases. My contention is that a great
deal of apparently “pure” nature poetry derives from and implies
an unspoken romantic narrative. Consider, for example, a poem by
Bashō, perhaps his most famous (Yuasa 149n.2). First, the Japanese
(quoted in Shirane 322):
furu ike ya
kawazu tobikomu
mizu no oto
Then a translation (Aston 295, altered):
An ancient pond.
A frog plunges in,
The sound from the water.
The implications of the poem are incomprehensible outside its liter-

ary context – the complex, well-defined patterns of topic and imagery
that governed the genre. Shirane explains that the frog or kawazu
was “almost exclusively associated with the blossoms of the yamabuki
(kerria), the bright yellow mountain rose, and with limpid mountain
streams.” Moreover, the “poetic essence” of the kawazu was under-
stood as its “beautiful voice.” The typical poem on the kawazu would
treat “the plaintive voice of the frog singing in the rapids or calling
out for his lover” (323). Thus, the context of Bashō’s poem, implied
by the conventions of the literary tradition, is one of exile from one’s
beloved. Note also, however, that the conventions themselves suggest
an eventual resolution or reunion. The voice of the kawazu is beauti-
ful. He is surrounded by blossoms. The moving water of the limpid
stream indicates advancement, at least movement toward some sort
of future. In contrast, Bashō’s poem is almost despondent. In place
of the moving water, and the future it suggests, there is a motionless
pond. In and of itself, this need not be a despairing image. For ponds
can be beautiful and peaceful. But this pond is not filled with lotuses.
The beautiful blossoms, expected due to convention, are nowhere to
be seen. Neither is the pond animated with the new and vigorous life.
Rather, it is old. Perhaps the frog is old too, for he does not sing to his
beloved. In context, it seems that the separation has become hope-
less. There is no longer any future. There is no murmur of a moving
stream, but only the isolated sound of the frog entering the still water.
This is a good example, for initially it appears to be a straightfor-
ward, uncomplicated nature poem, representing a pure natural expe-
rience, entirely separated from any narrative – a moment sealed off
from any story. But upon examination, it turns out to be thoroughly
embedded in a romantic plot. Many of Bashō’s poems are like this.
Indeed, it is worth noting in this context that Bashō formulated an
explicit poetics based on the concept of sabi or loneliness. In a sense,
the entire aesthetic of Bashō’s poetry is oriented toward the lover’s
period of isolation in romantic tragi-comedy.
Given this addition of much nature poetry, and the previous inclu-
sion of much eulogy and satire, we need to reformulate the narrative
hypothesis. Prototypical lyric poetry is not, as one might imagine,
opposed to or even independent of narrative. In fact, it is regularly
bound up with narrative in its production, meaning, and impact. In-
dividual poems typically imply encompassing narratives. More ex-
actly, though lyric poems may in principle treat any number of topics,
prototypical lyric poems are most often elaborations of emotively
distinctive constituents from romantic or heroic tragi-comedy. The
constituents in question are primarily narrative junctures, with their
particular junctural emotions. However, these constituents may also
be emotionally consequential features of character or natural images
associated with such junctures (as when imagery of spring is linked
with the reunion of lovers).
religious lyric poetry and the narrative

structure of spirituality
I should like to conclude by suggesting that this analysis of lyric
poetry has consequences that extend beyond our understanding of
literature and genre. Specifically, I should like briefly to consider a
particular aspect of religious belief, by way of a look at devotional
lyric poetry, an important type of poetry that we have not yet dis-
cussed. In keeping with the preceding analyses, there appear to be
two standard types of relation to God in lyric poetry. The first is a re-
lation of awe, expressed in poems of praise. The second is a relation
of longing, expressed in poems of separation and mystical union.
Lyric poems of the first sort are so widespread that they hardly
require isolation. This type ranges from ancient Egyptian poems
(“God . . . majestic, awesome, bedazzling, exalted. . . . Revealed like a
king in glory” [Mack et al. 45, 47]), to some of the Hebrew psalms
of praise, to some passages in Chinese poetry (“Mighty is God on
high,/Ruler of His people . . . Swift and terrible,” as the Book of Songs
puts it [Mack et al. 785]). Poems of this sort clearly relate to the heroic
genre. More significantly, they manifest a view of divinity based
largely on the prototypical features of the heroic genre. This view
includes the assimilation of God to a temporal monarch, a related
stress on God’s combative, protective, and punitive aspects, and an
emphasis on the rightful domination of society by the religious in-

group. Correlatively, this view of divinity as absolutized hero often
carries with it an enemy of absolutized evil, though this figure does
not often inspire lyric poetry. In any case, instances from Christianity
and Islam are too obvious to require mention. An instance from Hin-
duism may be found, outside of lyric poetry, in the case of Rāvan.a –
who, of course, makes his appearance in the heroic tragi-comedy of
the Rāmāyan.a.
Returning to lyric poetry, we find the devotional alternative al-
most equally well represented. Just as praise poems are part of the
heroic structure, devotional poems are most often part of the romantic
structure. Hindu bhakti poems and the ghazals of Muslim Sufis illus-
trate the point strikingly, for they are explicitly based on romantic
tragi-comedy. They directly assimilate the relation between a devo-
tee and God to the relation between lover and beloved – for example,
characterizing life on earth as exile from the beloved. Poems of this
sort manifest a view of divinity based on prototypical features of the
romantic genre, emphasizing mutual love, mutual devotion, the de-
spair of isolation and exile, and the ultimate union, even unity, of
God and devotee. The Song of Songs, as commonly interpreted in the
Jewish and Christian traditions, draws on this structure as well, as
do the poems of such Christian mystics as St. John of the Cross (see
Brenan for examples). Finally, the Nine Songs of the fourth- or third-
century b.c.e. are among the earliest extant lyric poems of the South
Chinese Ch’u culture. As Owen points out, “If the north Chinese writ-
ing was largely concerned with statecraft and sober social virtues, the
Nine Songs, by contrast, celebrate the senses and a passionate, often
erotic, relation between gods and mortals” (828). A good instance of
this romantic devotionalism is discussed by David Hawkes. This par-
ticular poem begins “The goddess comes not, she holds back shyly”
and goes on to treat the devotee’s relation to the goddess in terms of
frustrated romantic love – as in the lines, “The lady is sad, and sighs
for me;/And my tears run down over cheek and chin” (45).
The connection of all this with actual religious belief indicates that
the embedment of lyric poetry in prototypical narrative is not cogni-
tively special or isolated. Rather, our conception of and response to
divinity – more generally, our religious attitudes and thus much of
what goes along with those attitudes – are, to a great extent, shaped
by implicit prototype narratives, and not only when we write lyric

poems. Moreover, it seems unlikely that this relation between plot
and religion is itself unique or anomalous. Rather, it seems likely
that broad areas of nonliterary cognition are bound up with standard
narrative structures and emotion prototypes in much the same way.
conclusion
The “narrative hypothesis” asserts that prototypical lyric poems most
often treat emotionally distinctive elements (primarily junctural mo-
ments) in implicit heroic or romantic narratives. In this chapter, I
hope to have shown that the narrative hypothesis gives a highly
plausible account of some otherwise very surprising cross-cultural
features of lyric poetry. The final reflections on devotional lyrics and
religion suggest in addition that some form of the narrative hypothe-
sis may be extended, with equal plausibility, to a range of nonliterary
phenomena.
6
Testing, Revision, and the Program of Research

in Narrative Universals
Ainu Epic and the Plot of Sacrifice
testing and research programs in literary

narrative
As I have already stressed in the Introduction, the work of Imre
Lakatos suggests that the greatest value of a theory lies in the re-
search program that it enables. Specifically, Lakatos distinguishes
between “progressive” and “degenerating” research programs. Both
encounter anomalous data. Both uncover phenomena that contradict
their theoretical predictions. The difference is that a degenerating
research program deals with recalcitrant data through ad hoc formu-
lations. In other words, a degenerating research program formulates
principles that account for the recalcitrant data, but do not explain
anything beyond those data. In contrast, a progressive research pro-
gram is continually reformulating its constituent theories in such a
way as to expand the explanatory scope of those theories beyond
any recalcitrant data. Indeed, a progressive research program actu-
ally seeks out recalcitrant data as a way of increasing explanatory
reach and precision.
The most obvious way of generating new data – recalcitrant or not –
is laboratory experimentation. When engaging in experimentation,
we artificially create circumstances in which our hypotheses might
prove problematic. In other words, we create situations in which our
theory may not work, situations in which events may develop in a
way that contradicts predictions made by our theory. This sort of
172
Testing, Revision, and the Program of Research 173
experimentation is possible for certain areas of literary research, in-

cluding research in narrative. It is important and should be pursued
much more vigorously than has been done in the past. Unfortunately,
in many areas of literary theorization – including most of what we
have discussed in the preceding chapters – this option is not read-
ily available. For example, a colleague trained in laboratory research
suggested an experiment in which independent observers, unfamiliar
with the preceding theory, judge if Abhijñānaśākuntalam has the struc-
ture of a romantic tragi-comedy, The Epic of Son-Jara has the structure
of an heroic tragi-comedy, and so on. But I suspect that readers famil-
iar with these works will not find this to be something that requires
controlled study. It seems uncontroversial that Abhijñānaśākuntalam is
about two people who meet, fall in love, are separated due to a con-
flict with social authority, and so on. It seems uncontroversial that
Son-Jara is the rightful ruler, exiled from his kingdom, as a relative
takes the throne, that there is a threat to the kingdom, that Son-Jara
defeats that threat, and so on. In short, an experiment of this sort
seems unlikely to produce anything that would advance a research
program in narrative universals and emotion.
Here, it is worth returning to the example of linguistics that is,
again, the one field in which the study of universals has made sig-
nificant progress. In some respects, the situation in literary study
is not greatly different from that in linguistics. Many areas of lin-
guistic theorization are open to laboratory research. Many, how-
ever, are not. Of the latter, the isolation of linguistic universals is
a case in point. There is no doubt laboratory work that bears on
the study of linguistic universals. However, the isolation of the uni-
versals themselves, their full description and elaboration, is depen-
dent upon the study of natural languages. Fortunately, in this area
of linguistic study, there is a parallel to experimental research – the
examination of a previously unstudied language. We may follow
Lakatosian prescriptions and test hypotheses regarding linguistic
universals by examining languages that were not used in the initial
formulation of those universals. The same point holds for bodies of
literature. We may to some degree test hypotheses regarding literary
universals – in this case, narrative universals – by studying a body of
narratives that did not contribute to the initial formulation of those
hypotheses.
Broadly speaking, a new language will be most valuable to eval-

uating linguistic hypotheses if it is genetically and areally distinct
from languages previously studied and if it is widely attested. As to
the last, if we are considering a living language, this is not a prob-
lem, for we can simply recruit a native speaker. If it is a dead lan-
guage, however, it is valuable to have a large body of literature – not
simply a few lists of merchandise or the like. These general criteria
apply directly to literature as well. We discussed genetic and areal
distinctiveness in Chapter 1. As regards attestation, almost all liter-
ary traditions are, in effect, in the position of dead languages. Outside
oral traditions, it does not really make sense to refer to a native in-
formant. For example, it is not clear what he/she would inform us
about. Thus, we need a tradition that has an adequate body of extant
work.
These are not the only criteria for choosing a linguistic or liter-
ary test case. In addition, one should seek typological diversity. In
other words, a good test case would be a language that is of a dif-
ferent type from the majority of languages studied thus far – with
“type” here being simply a matter of some potentially theoretically
relevant variable. (Indeed, all these criteria are ultimately a matter of
controlling variables, in direct parallel with laboratory experimenta-
tion.) If a linguistic theory is well supported by research in synthetic
languages, then a researcher may be well-advised to look into ana-
lytic languages. The same point holds for universals of verbal art. The
most obvious uncontrolled variable in the present theory is authorial
sex. In order to limit areal contamination, I have focused on rela-
tively early writings in each major written tradition. Unfortunately,
one result of this historical focus is that most of the works I have con-
sidered were composed by men. Certainly, romantic tragi-comedy
figures prominently in the novels of Jane Austen or in the devo-
tional lyrics of the women bhakti poets such as Mahādēviyakka and
Mı̄rābāı̄. But the preceding pages include few references of this sort.
Moreover, some writers might object even to these references, argu-
ing that I have not considered women’s own narrative tendencies, but
only women’s writing that has thoroughly absorbed the preceding,
largely male traditions, has been canonized and preserved by men,
and so on. While it could be argued that, for example, the female
bhakti poets were not entirely determined by a literary tradition – or
that, more generally, women authors are not so entirely subservient to

male forebears – it would nonetheless be beneficial if we were able to
uncover a body of narrative dominated by women writers or oral
tellers. Indeed, oral tellers would be better, since I have relied dispro-
portionately on written literature (that is, literature proper) as well.
Though I have drawn on orature, it is relatively less well represented.
In part, this is necessitated by the material available. Most orature
has necessarily passed forever from the world, or been incorporated
into written literatures. However, though a degree of such under-
representation is unavoidable, further research in orature remains a
desideratum.
ainu epic
Ainu epic poetry constitutes an almost ideal body of work for our pur-
poses. First, Ainu is one of those anomalous languages that bear no
clear genetic relation to other languages. As Philippi points out, Ainu
“cannot be genetically connected with any other language groups in
the world” (7; for discussion, see Shibatani 5–8). This suggests that
Ainu traditions of narrative are genetically distinct as well, since story
telling appears to be coeval with full human language use. Moreover,
on the island of Hokkaido, the Ainu were relatively isolated. Philippi
argues that, for the most part, “the Ainu were living a way of life
rooted in the remote past, and Hokkaido remained untouched by the
main currents of Asian history” (6). There was interaction with the
Japanese. However, Japanese influence did not become strong un-
til the end of the seventeenth century (see Philippi 6). Philippi has
assembled and translated a set of Ainu narrative poems that reflect
“the social conditions and the cultural milieu” of the pre-Japanese
period (9). Specifically, “The ideas which are expressed in the epics
and which, in fact, formed the basis for the traditional Ainu way of
life, are extremely archaic and share many common features with
the ideology which was prevalent among the Paleoasiatic peoples
of northeast Asia until recent times. . . . The Ainu epic tradition, with
its extremely archaic mental patterns and modes of diction, is one
of the purest and most beautiful surviving examples of the oral lit-
eratures of the hunting and fishing peoples of northeast Asia” (50).
Philippi’s selection itself provides the relatively broad attestation we
require as well. Rather than one or two canonized epics, Philippi

gives us thirty-three narrative poems of different lengths, types,
purposes, and topics, poems recited by different bards in different
regions.
Thus, the Ainu epics satisfy the general criteria we set out initially –
genetic distinctness, areal isolation, and wide attestation. They also
satisfy the criteria that are more specific to our hypotheses regarding
narrative. First, as the preceding discussion suggests, the poems in
this collection were recited by oral bards working in an oral tradition.
Indeed, the tradition is unusually nonliterate; Philippi points out that
“The Ainu never developed any system of writing” (6). Perhaps the
most striking feature of the Ainu narrative poems, however, is that
“nearly all” of them “were recited by women” (Philippi 50). Thus,
we have a large selection of works of verbal art, evidently preserving
highly archaic elements, that derive from a tradition that is to a great
extent female. Indeed, this is particularly significant as “Ainu women
had a highly developed culture of their own which differed from that
of men and contained many elements which were kept strictly se-
cret from them” (50). Moreover, among the Ainu, poetry was closely
related to shamanistic traditions and “Among the Hokkaido Ainu, al-
most all the shamans are women” (3). In short, the Ainu epics provide
a nearly perfect test case.
What, then, do we find when we examine these epics? How do
they affect the ideas put forth in the preceding chapters? Beginning
with the variable of gender, we do find some gender related differ-
ences from the other works we have considered. But these differ-
ences are superficial and manifest underlying indexical universals,
such as a desire for autonomy, a particular concern with one’s own
in-group, and so on. For example, in the Ainu narratives, women
characters are more active, more independent, more likely to be the
focus of the events or to provide the point of view. Such differences
are consequential and worth further research, as well as cross-cultural
comparison. But they are not in any way contrary to the preceding hy-
potheses. Moreover, the preceding account of heroic tragi-comedy –
a genre one might expect to be affected by gender – fits the Ainu
material without significant alteration (though the Ainu poems do
highlight the variability of heroic narratives and the derivation of
political revenge plots from the usurpation sequence of the heroic

structure).1
On the other hand, this is not to say that the Ainu epics leave
us exactly where we started. In fact, these poems do lead to two
1 The final long poem, “The Epic of Kotan Utunnai,” provides a good example. (Here
and elsewhere, citations of Ainu epics refer to Philippi.) It concerns the great Ainu
hero Poiyaunpe. He is raised by an “enemy” woman in the enemy land. He learns
that his father “held sway over the upper/and the lower regions/of Shinutapka”
(368), but was killed by the enemy when abroad. On hearing this, Poiyaunpe sets
off on a series of battles. After defeating various enemies, he returns in triumph to
Shinutapka. Though the discourse or telling of the epic slightly alters the order, the
story or plot is straightforwardly that of heroic tragi-comedy, with the common vari-
ation that the usurpation of the rightful ruler is overcome by his son. Specifically,
the rightful ruler is killed; in consequence, the heir lives in exile. When he reaches
maturity, the heir battles various dangers to his society, then returns to his society
and assumes his rightful place.
The enemies faced by Poiyaunpe are particularly interesting here. Most are of
the general sort we considered previously. For example, at one point he battles
“Etu-rachichi,” a renowned warrior of the repunkur people, the traditional enemy
of the Ainu. At other points, he battles foreign rulers or armies from strange lands.
However, some of his antagonists are less traditionally realistic. For example, the last
battle fought by Poiyaunpe is against the “bad weather demon,” an obvious threat
to his people, but a different sort of threat than an enemy army or warrior. In fact,
many Ainu epics focus on threats from nonhuman antagonists.
Consider, for example, the third “Song of Aeoina-kanui.” In this story, the country
is threatened by an attack from Big Demon. The hero departs to defend his society
from the danger. After a long battle, both the hero and the Big Demon die. However,
the hero is revived by the gods and returns home. Here we have a clear case of the
central sequence from the heroic plot. A threat to the society is repulsed by the hero.
We also have death in the narrative middle. However, we are not faced with a tradi-
tional, military/national enemy. Moreover, this is not an unusual case. As Philippi
points out, Ainu “epics in which the culture hero” defeats “evil deities” are “very
numerous.” Other Ainu stories make animals into the threat. In “Song of an Evil
Bear,” for example, Okikurmi deceives and kills a bear who poses a threat to human
society. In the first “Song of Aeoina-kanui,” the enemy is a huge char.
In all these cases, the greatest theoretical interest of the variation is probably in the
way it highlights the common metaphorization of enemies as animals. In the heroic
tales considered earlier, the enemy is most likely to be another national group, not
a demon or an animal per se. However, in those stories, it is common to character-
ize the national enemy in demonic or bestial terms. (This is clearly the alternative
or contrary to the empathic humanization that occurs in the epilogue of suffering.)
That dehumanizing of the enemy is rendered more salient through the Ainu data.
Interestingly, almost all these cases manifest only the threat/defense sequence.
In part this is an instance of a universal pattern. Cross-culturally, the threat/defense
sequence appears more likely to occur on its own than the usurpation/exile se-
quence. In other words, the former is more autonomous than the latter. To some
alterations in the preceding hypotheses, one small, one quite big. The
first is a relatively minor change in the precise delimitation of the
romantic plot. The second involves the addition of an entirely new
prototypical narrative structure. As it turns out, these alterations are
extent, this is true for formal reasons. If there is a threat to the society, the hero must
defeat it, wherever he/she might be. His/her action does not in any way require
a prior displacement and exile. However, the displacement/exile sequence does to
some degree require the threat/defense sequence, for the hero proves the injustice
of his/her displacement and exile and regains his/her place in society precisely by
defending the society against threat. In the Ainu case, however, there is more to it.
The relative scarcity of usurpation plots in this case results also from the social and
political structure of Ainu society. Historically, the Ainu lived in minimal communi-
ties: “Generally speaking, Ainu settlements were small, consisting of perhaps one to
ten households” (Philippi 6). Moreover, “The Ainu never developed . . . any concept
of a political state” (6). The very small local groups of five to ten households did
have headmen (Watanabe 79, Ohnuki-Tierney 73–8). However, the larger collectives
of local groups operated “without any chief” (Watanabe 79; see also Chapter II of
Watanabe).
On the other hand, the usurpation sequence is not absent. We have seen one
poem in which the usurpation/exile sequence occurs – “The Epic of Kotan Utunnai.”
There is one other poem of this sort as well, “Song of the Young God Okikurmi.”
This song recounts a dream dreamt by Okikurmi’s son. He has been raised by a can-
nibal woman, who brings home great quantities of human flesh to eat. Clearly, she
is a threat to human society. In addition, she is raising Young Okikurmi because she
killed his father. Thus, we have the standard sequence of the hero being killed, while
the son is exiled (thus denied his rightful place). In this case, there is the additional
twist that the older hero’s murderer is the threat to human society and the young
hero’s exile is in the land of that murderer. The cannibal takes young Okikurmi out
to the place where she killed his father. However, young Okikurmi kills the woman.
He does not return to his home in triumph, for he wakes and realizes that it was a
dream. Nonetheless, the presence of the standard structure is clear.
These two direct instances – and an indirect instance in the romantic “Woman of
Shinutapka” (to which we shall turn below) – indicate clearly that the usurpation
sequence is part of the heroic structure among the Ainu. Given the formal and social
factors working against the manifestation of this structure, its appearance in three
poems (which, due to their length, comprise over a sixth of the collection) suggests
the psychological persistence of this structure and of the prototypes it follows. In
keeping with this, “The Epic of Kotan Utunnai” and the “Song of the Young God
Okikurmi,” the two poems directly containing the usurpation sequence, are also
among the few Ainu poems that point toward the positive social goal of domina-
tion, rather than the negative social goal of avoiding destruction.
The “Song of the Young God Okikurmi” shows us something else as well. In this
poem, the pursuit of domination takes a form that is consequential for our concerns.
That form is revenge. Of course, this is not unusual. Indeed, it is noteworthy precisely
because the motif of revenge is widespread cross-culturally. Usurpation generally
results from the actions of some heroic rival. Moreover, the final return of the hero,
ruler, or heir is regularly accompanied by the punishment ofthat rival. The political
not ad hoc, but apply to a wide range of literary works in the traditions
considered earlier, revealing patterns that had been left aside in the
earlier formulations. In other words, these reformulations do just
what Lakatos says they should do. They extend the explanatory scope
and precision of the theory.
The first reformulation is really just a matter of emphasis. In setting
out the structure of romantic tragi-comedy, I stressed that the separa-
tion of the lovers was primarily a matter of conflict with society, most
often conflict with parents. I noted in this context that there is often a
rival as well, frequently a rival chosen by the parents as a more suit-
able spouse for the lover or beloved. It also happens, however, that
the rival acts more independently – even to the point where a love tri-
angle occurs. In other words, the rival may be developed more fully
and autonomously, and the interfering parent may be deleted, though
this rival character does often retain the same function of standing for
social norms. For example, in opposition to the lover, the rival may
be of the right religion or social class and thus he/she may appeal
to the beloved for precisely the reasons he/she would appeal to the
beloved’s parents. Outside the Ainu material, an example of this gen-
eral sort may be found in Jane Austen’s Emma, where Harriet rejects
a loving farmer, Robert Martin, setting her sights on the more socially
elevated Mr. Elton – just the sort of thing a dominating parent would
demand in the more complete version of the romantic structure. Of
course, as often happens, this is not a pure case of a love triangle, for
revenge plot merely separates out this elementof the usurpation sequence – a point
well illustrated by, for example, Hamlet.
It is perhaps worth noting in conclusion that clichés about gender difference
would lead us to expect that female bards would craft their poems in such a way as to
undermine violence and warfare. Thus, given standard beliefs, we would expect rela-
tively little emphasis on battle, and much greater emphasis on remorse and penance.
In fact, we find almost the exact opposite. One striking feature of these poems is that
the women regularly join in battle, killing the enemy with great strength and vigor.
At the same time, we find less attention to the horrors of war and the remorse that
follows battle. The greater presence of female warriors is presumably due to the fact
that the bards are women. Following a general human tendency, they place them-
selves or people like themselves in more active and heroic roles. Indeed, this may
be one reason for the relative lack of remorse, and the virtual absence of an epilogue
of suffering. (We find it only in one poem, “Song of the Daughter of the Moun-
tain God.”) Insofar as the poems allow women to experience a sort of heroism that
is not ordinarily available to them, an emphasis on remorse and penitence would
undermine one important function of the work.
Emma herself has in effect taken up the role of interfering parent, if in

a diminished form, by opposing the marriage to Martin and urging
Elton as an appropriate alternative. Obviously, the love triangle is not
confined to Emma and Ainu epics. Despite its very noticeable recur-
rence, however, I had understated its importance in romantic plots.
The repetition of this general structure in the Ainu poems drew my
attention to it elsewhere as well. (The Ainu poems also suggest that
love triangle plots often incorporate elements of the heroic structure
to an unusual degree.)2
2 A number of Ainu epics fit the romantic structure quite well. Consider, for example,
the “Song of a Human Woman.” In this tragic story, a human woman marries a god.
The older brothers of the god find out about the marriage and force the husband
to leave the earth and return to heaven. The wife continues to long for her hus-
band until she dies. The story represents the separation of lovers due to a conflict
with family and, implicitly, with society (here, the society of the gods). This conflict
leads to separation and to the exile of the lover. The ascent of the husband into the
heavens carries suggestions of death as well, and, again, the wife literally dies. The
story ends tragically – but perhaps there is a hint of a possible spiritualized reunion
after the wife’s death.
A more complex case may be found in the “Song of the Younger Sister of the Owl
God.” In this story, the Owl God has followed the urging of the other gods and ar-
ranged for his younger sister to wed the dislikeable god of Poroshir. The great hero,
Ainurakkur, visits the Owl God’s home because he is “so exceedingly/lonely” (215).
Poroshir mistakenly believes that the Owl God intends to give the younger sister
to Ainurakkur. So he gathers his armies to “launch/a war of annihilation” (218).
Ainurakkur has to leave the land of the Owl God and fight the armies of Poroshir.
However, it seems that Poroshir’s apprehensions were not entirely misplaced, for
Ainurakkur takes the younger sister along with him, and she joins him in battle
against Poroshir. They eventually defeat Poroshir and are married. The story ends
with the younger sister explaining that “We lead/a magnificent married life” (228).
The story begins with an implicit conflict between the lovers and both a parental
figure – the Owl God, who raised the “younger sister” – and society more generally,
as represented by the other gods. There is a period of exile and imagery of death
(Ainurakkur actually describes what will happen if he dies in battle). The hero – and
the beloved – struggle against the rival in a section that borrows from the heroic
structure. Finally, the lovers are united.
This story presents a good instance of the gender differences noted previously.
Here as elsewhere, the women are far more likely to join directly in battle or other
struggles with the men. They are far less likely to be passive victims, merely saved
or protected by men. However, their activity does not change the basic structure.
This story also illustrates the importance of the rival suitor. I had earlier treated
the rival as a relatively passive character. Clearly, that is not the case in this story.
Indeed, here the main conflict is directly with the rival. He is the one who engages
in battle with the lover and the beloved. The theme is, of course, extremely common
cross-culturally. Obvious instances would include the Rāmāyan.a, in which the ro-
mantic part of the plot concerns the abduction of Sı̄tā and her rescue by Rāmā; the
The other alteration is a much deeper and more consequential the-

oretical revision. The presence of several Ainu stories treating famine,
sacrifice, and ritual indicate that we should add a third prototype
narrative, sacrificial tragi-comedy, which concerns a physical, rather
tragic story of Deirdre and Naoise in Irish epic tradition, where Naoise is killed by
Conchubar, who takes Deirdre for himself; the broader story of the lliad, involving
Helen and Paris, and so on. My previous formulation of the structure did note that
a complete romantic plot often involves the defeat of a rival. However, the Ainu
poems indicate the degree to which this part of the romantic plot may be split off
and emphasized. Moreover, there seems to be more to it than that. As these instances
suggest, particular stress on the love triangle often appears to involve the importa-
tion of elements from the heroic plot, even the joining of the two plots, with the rival
assimilated to the threatening enemy. Indeed, abduction can be the act of a rival in
a romantic plot or the act of a threatening enemy in a heroic plot.
A particularly striking instance of the love triangle version of the romantic plot
is found in the “Song of the Fire Goddess.” In this story, the Goddness of the Waters
abducts the husband of the Fire Goddess. The latter pursues her rival, only to find
the husband apparently accepting the situation quite willingly. In keeping with what
we have just noted about the incorporation of heroic structures, the Fire Goddess
then defeats the Goddess of the Waters in battle. However, she does not take her
husband back with her. Rather, the husband collects “ashimpe,” a compensation
for damages. He finally returns home with “many precious treasures,” apologizing
abjectly. “After that,/he remained [at home]” (74). Here too we have the familial
conflict, the abduction, and battle. We also have a strange epilogue, directly parallel
to the epilogue of suffering in heroic plots. After the reunion has been made pos-
sible, the husband goes off on a exilic journey of repentance, ultimately returning
home in the sort of somber and contrite mood that characterizes the conclusion of
the heroic epilogue. This case, then, shows a particularly extensive incorporation of
heroic elements into the romantic structure.
There is no point going through all the instances of the romantic plot structure
in the collection. However, there are two other instances worth particular mention.
These are two of the three longest poems in the collection. The first presents an inter-
esting variation on the abduction structure, and on the structure of familial conflict.
This is the “Woman’s Epic.” The speaker was abducted by Repunnot-un-kur when
she was still a child. Repunnot-un-kur raises her as a father, with the intention of
marrying her when she reaches maturity. The girl is warned by a bear cub, who
rescues her from Repunnot-un-kur, returning her to her home. When they arrive at
the home, the bear cub is ritually killed by the girl’s brothers. He returns to life as
a human being and marries her, explaining that his father disapproved of the mar-
riage. They live happily.
This version reshapes the standard structure remarkably. However, it remains
a clear variation of that structure. The conflict with the family is given an oedipal
twist when the father and the abducting suitor are, in effect, identified. The hero first
saves the beloved from the abduction, and is then killed (by the beloved’s family),
leading to separation. Finally, when they are married, we learn about his father’s
disapproval of the union. All the parts here are from the standard romantic story –
but with strange repetitions, combinations, and specifications. The result, though
than a personal, social, or transcendental goal – prototypically, food,

plenty of the primary means for maintaining life. Since the revision
of the romantic structure is minor, I shall concentrate on the new
prototype structure.
Before going on to this, however, I should return to the issue of
just what stories are covered by these three structures. In presenting
these ideas, I am sometimes asked if I believe these three structures
are “exhaustive.” The problem with this question is that it is ambigu-
ous. Most obviously, the question may refer to universal, prototype
clearly within the general structure of romantic tragi-comedy, is a strikingly innova-

tive story.
Certainly the most outstanding poem in the collection is “The Woman of
Poi-Soya,” a technical tour de force that uses different voices and perspectives in
an extremely powerful way. The story content is no less fascinating than the form.
The male hero is Otasam-un-kur. His parents wish for him to marry the Woman of
Poi-Soya, a militaristic woman, with whom Otasam-un-kur engages in battle. Dis-
gusted with Poi-Soya-un-mat, Otasam-un-kur pursues another woman only to find
that she has been unfaithful – though this is not her fault as she has been deceived by
a rival pretending to be Otasam-un-kur. He kills this woman, her lover, and himself.
The lover’s sister, Kunnepet-un-mat, revives Otasam-un-kur and the two of them
are married. This, too, is a complex revision of the standard structure. First, there is
the parental conflict in which Otasam-un-kur rejects his parents’ choice of a spouse.
Then there is the variation on the love triangle and abduction when the beloved is
deceived and seduced. Here, Otasam-un-kur follows the standard sequence in de-
feating the rival. Finally, we have the motif of death and rebirth, followed by union.
The big difference, of course, is that it is not union with the beloved, for she has been
killed by the “hero” and is not reborn. Rather, in an incorporation of an element from
the usurpation plot, her place is taken by her sister.
What is perhaps most striking about this poem is that it is so cynical about roman-
tic love – an attitude that is not absent from some of the other narratives also. After
Otasam-un-kur’s initial account of the evils of Poi-Soya-un-mat, we hear about many
of the same events from Poi-Soya-un-mat herself. She is humanized in the course of
this section, and Otasam-un-kur’s flaws begin to appear. The following sections
only make his character less and less sympathetic, until we can only pity kind and
generous shamaness, Kunnepet-un-mat, for marrying him. Indeed, Otasam-un-kur
virtually curses Kunnepet-un-mat when he thinks about their marriage, explaining
that it is only because she saved his life that he “allowed her to cook for me” (365).
Here, we seem to have the most extreme case of gender difference – a use of the
romantic structure to criticize male cruelty and bias. (Of course, this is once more
a superficial difference that manifests widespread commonalities – in the desire for
autonomy, respect, and so on.) In fact, the aim of this story appears to be the “dero-
manticization” of romantic tragi-comedy. But, once again, to do this, the poet remains
within the structure of that genre. The story repudiates the sexual politics of male
domination. But it preserves the story form in which romantic union is imagined,
however falsely, as a prototype for happiness and is sought by characters as such a
prototype. It varies, but nonetheless retains the universal romantic structure.
narrative structures, or it may refer to stories of any sort. In the latter

case, the answer is an emphatic “No, these genres are not exhaus-
tive!” Here I must emphasize once more that we can tell stories about
anything. There are infinitely many stories with infinitely many top-
ics, aims, and combinations of emotion. Again, that is what makes
the recurrence and prominence of these three types so noteworthy.
But these infinitely many stories are not prototypical stories – or they
have a significantly lower degree of prototypicality. Nor do they recur
universally and prominently.
More importantly, I am inclined to believe that these structures do
exhaust the range of universal, prototype narrative structures. My incli-
nation here is both empirical and theoretical. Empirically, these are
the structures that appear to recur consistently and prominently in
the cross-cultural data. Theoretically, it is difficult to imagine what
other common context there may be for happiness beyond physical,
personal, and social (given the preceding analysis of the spiritual
context as lacking a distinct prototype). Of course, I can hardly assert
this with complete certainty. After all, prior to reading the Ainu mate-
rial, I had thought there were only two universal, prototype narrative
structures. Who is to say that I or someone else will not discover a
fourth pattern and context in the future. Still, it seems likely that the
three structures are exhaustive at this level.
Here, a third way of interpreting exhaustiveness arises. Do the
three structures exhaust not only universal prototype structures for
happiness, but even culturally particular prototypes? Put differently,
are there culturally particular (thus nonuniversal) prototypes for hap-
piness? One’s first inclination is to say, “Of course there are.” This
seems true experientially. There is also theoretical reason to expect
this. As Barsalou points out, we generate new, ad hoc categories all
the time. Thus, we might speak about “the sort of thing you would sell
at a garage sale” or “the kind of gift someone might buy for Mother’s
Day.” Barsalou has shown that these categories exhibit the usual pro-
totype patterns. For example, perfume might be a more prototypical
Mother’s Day gift than, say, a set of coffee mugs. Barsalou also ex-
plains that these ad hoc categories might become well entrenched,
thus firmly established lexically. In some cases, these ad hoc cate-
gories are, in effect, contextually specified prototypes (for example,
“Mother’s Day” provides a context for specifying the gift prototype).
The point clearly applies to happiness. Prototypes for happiness can

and do come to be contextually specified for individuals, for small
groups, for larger cultures, in at least apparently nonuniversal ways.
But here one might ask whether many seemingly idiosyncratic pro-
totypes are not just more fully specified instances of the universal pro-
totypes. For example, my own evidently idiosyncratic prototype for
happiness involves a prestigious university position in a particular
region of the Midwest. This seems to be irremediably culturally par-
ticular. After all, historically, most cultures have not even had univer-
sities. Moreover, the geographical location is highly specified. But in
fact this is not, ultimately, culturally idiosyncratic at all. The presti-
gious university position is clearly a specification of achieving social
dominance, if of a rather moderate sort, within a group (in this case, a
professional group). The other element is proximity to the city where
my wife teaches. In this way, my apparently idiosyncratic and cultur-
ally relative prototype turns out not to be idiosyncratic and culturally
relative at all. Indeed, it turns out not to be a distinct prototype at all.
It is, rather, an instance of the social and personal prototypes.3
3 Of course, it may not be a very good instance, especially for the social prototype.
Exampla, too, exhibit “prototype effects,” such as grading. In other words, some ex-
empla are better cases of a prototype category than others. Indeed, despite attempts
by some linguists and psychologists to anchor prototypicality in a “basic level” of
human cognition, it seems clear that “schema,” “prototype,” and “exemplum” are
relative to topic. For example, the basic level is commonly defined as the highest
level that can be encompassed by a “single mental image,” the level at which “motor
programs” function, and so on (for a summary of criteria, see Lakoff 47). The idea is
that we can encompass “chair” in an image, but not “furniture.” However, this is not
true. One can easily argue that imagining “furniture” simply involves imagining a
variety of different pieces simultaneously – as in a living room or a furniture store.
Conversely, one can argue that we don’t actually imagine “chair” at all, but rather a
dinner chair or a comfy chair (itself understood as a variety of different parts, such
as legs, imagined simultaneously). This would seem to make both “furniture” and
“comfy chair” basic, but not “chair.” The reason for this ambiguity is that, again,
“basicness” varies with topic. While chairs are a highly prototypical sort of furni-
ture, “chair” is also a schema of which particular subtypes of chair (for example,
dinner table chairs) are more prototypical than others (for example, dentist chairs) –
a point that may be repeated at every level. Of course, there may be a “basic level”
in the sense of a default level, a level at which we tend to operate spontaneously. In
fact, that seems pretty well established. But that does not fix these levels in any the-
oretically rigorous way. It merely says that we will tend to think about the world at
a certain level, and thus, that our default use of “prototype” may refer to that level.
But we can and do shift levels, and our three sorts of categorization follow such
shifts.
On the other hand, whether or not all happiness prototypes are

reanalyzable in this way, it is clear that these three prototypes do
not exhaust all specific happiness-related goals. Moreover, it is clear
that all traditions include highly esteemed works that do not fit
any of the prototype structures. My own intuitive sense (for what-
ever it is worth) is that the three genres cover perhaps two-thirds
to three-quarters of canonical and popular narrative, with the re-
maining works not fitting into any particular cross-cultural pattern.
The proportion is toward the high end in popular narrative, perhaps
even surpassing this range; it may be closer to the two-thirds mark
in canonical narrative. Within these percentages, the proportions of
each genre seem to shift with the circumstances of the poets and
their societies. Hierarchical, militarized societies in which the poet is
patronized by a ruling elite tend to produce a higher percentage of
heroic plots. Societies in which narratives are produced for broader,
popular consumption particularly emphasize romantic plots. Soci-
eties in which famine is a constant threat have a higher percentage of
sacrificial plots. Of course, at this point, comments on exact propor-
tions are conjectural. Moreover, it is very difficult to say just how one
would determine such proportions rigorously. Canons do not have
strict boundaries. No one can give an exact count of all the mem-
bers. Finally, I could be mistaken about the minority of nonuniversal
stories. Perhaps they fit some universal pattern which I have simply
failed to isolate. In any case, the crucial point is that the three gen-
res may be exhaustive for universal, prototype narratives – as I am
hypothesizing – without in any way exhausting all the stories that
can be told, all highly esteemed narratives, or all happiness goals.
In short, stories come in countless varieties. However, these three
structures are, it seems, uniquely important. They form the core of
what we mean by prototypical narrative. They constitute the majority
of the stories that inspire our collective admiration. Finally, they help
guide not only story telling, but our reception of stories, and our
The points raise interesting issues for the present analysis, such as the degree
to which there are patterns to prototypicality at lower levels than that of story. In
other words, story prototypes appear to be of just three sorts – romantic, heroic, and
sacrificial tragi-comedy. Do subprototypes for, say, romantic tragi-comedy fall into
any significant patterns? Issues such as this could obviously be pursued as part of a
research program in this field.
thought about other matters (such as religion) as well. Once again,

I take their prominence to be an absolute or at least near absolute
universal. But they simply do not define all stories. Indeed, that is
in many ways the point of speaking about prototypes rather than
necessary and sufficient conditions. Only necessary and sufficient
conditions would define all stories.
With these qualifications, we may now turn to our third prototyp-
ical structure.
sacrificial tragi-comedy
As I just noted, a number of Ainu epics treat the genesis and overcom-
ing of famine. Some of these incorporate concerns about famine into
heroic plots such that the agent of famine is some enemy who must
be defeated by the hero. In itself, this could be viewed as a variation
on the heroic structure. However, there are several stories that treat
famine separately. This suggests that, in the heroic cases just men-
tioned, famine has been incorporated into the heroic structure in the
way romance is often incorporated into an heroic plot and heroism
is often incorporated into a romantic plot. In other words, it may be
that the famine motif is part of another recurrent structure, parallel
with romantic and heroic tragi-comedy.
This makes sense within our general explanatory framework. Just
as there are personal, social, and transcendental varieties of hap-
piness, there should be a form of bodily or physical happiness as
well. For most of humankind, throughout most of history, it is very
likely that the physical prototype for happiness has been plenty while
the correlated prototype for sorrow (which would define the middle
or progression in relevant narratives) has been famine. In connec-
tion with this, the sustaining emotion of a narrative generated from
this prototype would be hunger. (It may seem inappropriate to call
hunger an emotion. I shall return to this issue in the afterword.) This
may be less intuitively plausible than the romantic and heroic cases.
However, our intuition here is likely to be the result of habituation.
While hunger is still very common today, most people who can con-
cern themselves with cognitive science and literary theory simply
expect plenty as an ordinary condition of their lives. Thus it would
be surprising if the force of plenty as a prototype for happiness had
not become dulled. The point finds empirical support in research

by Harris, Olthof, Terwogt, and Hardman. These researchers stud-
ied children’s views about “the situations that provoke emotion,”
which is to say, the eliciting conditions for emotion. They investi-
gated a set of European children and a set of Nepalese children from
a remote area. They found that the children in general responded
similarly. However, the Nepalese children were considerably more
likely to bring agricultural “anxieties” into their accounts and to as-
sociate “pleasure” with “foods” (338). For example, they report only
two eliciting conditions for happiness from the Nepalese children.
The first one, translated from a seven year old child, is “With good
things to eat, people feel happy” (334). Needless to say, this research
is not the only empirical source of support for such a hypothesis.
Literature provides many cases, cross-culturally.
Returning to the Ainu tales, we find that, in the nonheroic famine
stories, famine is ended by the benevolent action of a deity, who
awards food to the hungry people after petition and/or some sort
of sacrifice. In each case, the award of food is followed by a ritual.
More exactly, the human community has committed some ethical or
spiritual violation that offends the gods – typically a violation in the
treatment of food. This has given rise to the famine. In consequence,
they have to make some sort of offering to the gods and/or send a
messenger, petitioning for divine intercession. The gods grant hu-
mans their food, introducing some ritual that will preserve their well
being.
For example, in “Song of the Goddess of the Waters,” the goddess
receives a goblet of wine as an offering from the humans. Through
the wine, Okikurmi speaks to the goddess, explaining that there is
famine and that he has used his “last grains” to make the wine for this
offering in the hope that she will help. The goddess invites a number
of deities to a feast and speaks to them about the famine. She learns
that the humans have treated various gods with disrespect through
their actions toward deer and fish. She restores food to the humans,
but simultaneously explains that they must practice specific rituals,
and they must make a propitiatory sacrifice of wine and ritual wands
to the offended gods. “Song of the Owl God” more or less follows the
same structure. It begins with a strange sequence of events in which
the Owl God asks a crow and a jay to deliver a message to heaven.
Each one falls asleep when listening to the message. The Owl God
reacts by killing them. Then the “Dipper Boy” agrees to take the mes-
sage. He is fully attentive, learns the message, and leaves for the land
of the gods. The message concerns a famine that is occurring among
the humans. The gods explain that the famine has resulted from the
insulting manner of the humans in their treatment of game and fish.
The Owl God teaches humans the proper rituals and the famine
ends. Other Ainu poems include at least parts of this sequence
as well.
Of course, it does not follow from these few cases that there is
a third cross-culturally prototypical narrative structure. One might
simply count this as an oddity of Ainu orature. After all, a certain per-
centage of stories in any collection will not fit the romantic and heroic
prototypes. There is no reason why some of the nonprototypical tales
might not form themselves into a culture-specific pattern. However,
in reading these stories, I was struck by the fact that this particular
structure recurs in culturally significant narratives elsewhere. For ex-
ample, one of the most important stories of the Yoruba people of west
Africa concerns a famine brought on by the greed of “Land” with re-
spect to a small bit of game. Land had been hunting with Heaven
and they killed a mouse. Land refused to give the mouse to Heaven.
In consequence, Heaven withheld rain. Due to the terrible drought
that ensued, people had to sacrifice to heaven. In Amos Tutuola’s
retelling of this story, they used a human “carrier,” a man who could
transport the offerings up to Heaven, asking that Heaven allow rain
to fall once again (301–2). The sacrifice is highly ritualized and of a
piece with actual Yoruba practices. Heaven accepts the sacrifice and
rain falls to earth, ending the drought. Though the story speaks of the
error as made by “Land,” it is clear that “Land” here represents the
human community. The greed of the human community, and its dis-
respect for “Heaven,” led to the famine. Only ritual sacrifice, which
allows the carrier to convey prayers and offerings from humans to the
gods, can restore food. (Tutuola’s novels actually include numerous
sacrificial stories, drawn from Yoruba sources. For an analysis of The
Palm-Wine Drinkard in these terms, see Chapter 3 of my Empire.)
The general structure of these “sacrificial” narratives appears to
be something along the following lines. There is a violation of divine
prescriptions, hence a divine/human conflict. The violation may be
individual, but its results are collective. It commonly concerns food

and results from overstepping one’s proper (human) bounds through
arrogance and/or greed. This violation leads to famine, the loss of
food. (We will see later that this can be represented as a sort of exile,
in keeping with the structure of romantic and heroic tragi-comedy.)
Death and imagery of death follow. The community must engage
in a sacrifice that serves to reconnect the human and divine worlds.
This is frequently associated with the institution of a ritual bearing on
food. The result of this sacrifice and ritual is the restoration of plenty
by a now benevolent deity. The sacrifice may be a “blood sacrifice”
or not; if a blood sacrifice, it may be the sacrifice of an animal, but, in
literature, it is often the sacrifice of a human.4
4 A structure bearing some of these features, especially that of sacrifice, has been
isolated by a number of writers, in very different theoretical contexts. No doubt
the most famous discussion of this sort is by René Girard. Readers interested in
pursuing the topic of sacrifice would certainly be well advised to consult Girard’s
seminal Violence and the Sacred. However, Girard’s fundamental propitiatory view
of sacrifice is entirely different from mine. Specifically, Girard takes the propitiatory
purposes of sacrifice to be a sort of surface appearance only. The real social function
of sacrifice is to release violent feelings that, if left uncatharted, would tear apart
the whole society. Early in his book, he maintains that the “theological basis of the
sacrifice” is a functional “misunderstanding,” for “The celebrants do not and must
not comprehend the true role of the sacrificial act” (7). Specifically, “The sacrifice
serves to protect the entire community from its own violence” (8).
I disagree with Girard on two points. First, it seems clear that all peoples seek
some ethical explanation for disaster. It is a human tendency to posit moral blame
for famine, plague, and so on, beyond simple physical causality. We explain normal
events by normal causal sequences. We explain extraordinary events by abnormal
casuality, which almost always has an ethical component. (Indeed, one source of
moral feeling probably is just this sense of bafflement when faced with abnormal
events, especially very painful events.) It is a human tendency to see extraordinary
suffering – for example, that produced by famine – as punishment for immorality.
Sacrifice arises almost predictably in this context as a reparation aimed at ending that
punishment. Thus, I do not see the “theological basis” as a mere manifest content or
ideology. Certainly any social act has many motives. For example, there may be an
enjoyment of cruetly that contributes to sacrifice. In actual human sacrifices, there
are probably political motives as well, individual jealousies, and so on. However,
the theological explanation is part of a genuine belief, a belief that is bound up with
important feelings and goals.
My second disagreement with Girard concerns catharisis. Girard’s assumption is
that one act of violence reduces our tendency to commit further acts of violence. But
this seems to be empirically mistaken. If Girard were correct about sacrifice, capital
punishment should have much the same effect. It should not only frighten potential
murderers. It should in general reduce the violent tendencies of the population. But,
in fact, just the opposite seems to be the case. In general, capital punishment shows
cross-cultural evidence for sacrificial narratives

But is this new category merely an ad hoc addition? It does cover
some instances from Yoruba as well as the Ainu material, but that is
hardly compelling. In fact, once one notices the sacrificial structure, it
is clear that there are literary instances in a wide range of cultures. In
other words, its addition is not ad hoc, but extends the explanatory
reach of the theory beyond that of the initial data.
Oedipus the King is a particularly prominent example of this
structure (though, in combination with its sequel, it is also part of
an epilogue of suffering, as noted previously). This play, “univer-
sally recognized as . . . the most brilliant example of theatrical plot”
(Knox 131), begins with a famine in Thebes: “Thebes is dying. A
blight on the fresh crops/and the rich pastures, cattle sicken and die”
(Sophoclēs 160). The fault is that someone killed King Laı̈us and the
land will not be fertile again until the murderer is punished – in this
case by exile. Individual expulsion from the community is a com-
mon variant, substituting for actual death in some instances of the
sacrificial structure. (This is distinct from, but perhaps related to the
communal exile that is sometimes associated with famine, as men-
tioned previously.) Specifically, the scapegoat is often seen as bearing
something desirable to the offended deity as a sort of appeasement.
However, he/she may also be understood as purging the society by
transporting the sin to another place. In the former case, the victim
is more likely to be killed, so that his/her spirit may ascend to the
deity; in the latter case, he/she is likely to be exiled. As the oracle
says, “Drive the corruption from the land/don’t harbor it . . . don’t
nurse it in your soil – root it out!” (164).
Euripidēs offers instances as well. Acts of human sacrifice often
appear in Euripidēs as part of heroic tragi-comedy. However, some
instances of sacrifice in his plays are more clearly part of the sacrificial
structure (even if they continue to draw on the heroic plot as well).
no deterrent effect. Indeed, it seems to encourage violence. As Amnesty International

points out, “The death penalty never has shown to benefit a society. In fact, there
are strong indications that it increases people’s tolerance of and tendency toward
violence.” Thus “The Bowers-Pierce study, analyzing executions between 1907 and
1963, concluded that an average of two additional homicides were committed in the
month after an execution took place.” In short, this basic premise of Girard’s analysis
of sacrifice seems mistaken.
One good case is the Phoenissae. In this play, Tı̄resias explains that
Cadmus “had killed the dragon born from the earth.” Ārēs demands
“the blood of a descendant” of the men who sprang up from the
dragon’s teeth (O’Connor-Visser 76). For complex reasons, the only
descendant who will satisfy this demand is Menoiceus. If Menoiceus
is not sacrificed, Thebes will be overrun by the invading army. The
explicit reason for the sacrifice, a reason borrowed from the heroic
structure, is not hunger. However, the link with the preceding, pro-
totypical structure, is not difficult to discern. The original violation is
a crime against the earth, and Menoiceus “had to be sacrificed to the
earth” (84). Indeed, J. P. Guépin has argued that the play involves a
detailed agricultural allegory (207–8).
Moving outside Europe, we find numerous instances in a wide
range of cultures. Consider, for example, The Injustice Done to Tou Ngo,
a drama by the most renowned Chinese dramatist, Kuan Han-ch’ing.
In this work, the death of Tou Ngo is a violation that leads to drought
and requires reparation before abundance is restored. A Japanese ex-
ample may be found in the Kabuki play, Narukami. In this play, a
recluse has been snubbed by the emperor. As an act of vengeance, he
captures the God of Rain and imprisons him. Halford and Halford
summarize the results: “No rain falls. The country is on the verge of
drought and starvation. . . . As a last resort, Taema-hime, a princess
of the Imperial house, offers to sacrifice herself” (231). Thus, we have
an initial violation, followed by drought and famine, rectified by a
sacrifice – or possible sacrifice. In the end, Taema-hime does not sacri-
fice herself, but overcomes Narukami’s power through her seductive
charms (for a fairly full version of the play, see Saint Narukami and the
God Fudō in Brandon); nonetheless, the basic structure is present.5
The structure appears to be particularly common in Native
American narratives, at least those that have come to be seen as
paradigm works. John Bierhorst picks out the Mayan Cuceb and the
Nahuatl/Aztec Quetzalcoatl as, each, “the leading example,” which
is to say the central paradigm work, “of the cultures that produced
5 Interestingly, the sacrificial structure is integrated into a heroic structure. Narukami

is in league with the emperor’s usurping brother and his actions are designed to
force the emperor to resign. Moreover, it is joined with a classic romantic plot as
well. Taema-hime is able to overcome the objections of her rival’s father and marry
her beloved precisely because she manages to restore rain to the kingdom.
them” (xi). Cuceb, a “ritual epic” (187), is a dense and complex piece.
But the general structure is clear – and it is a sacrificial narrative, par-
tially combined with a heroic narrative. The heroic section concern’s
the “people’s deliverance from an alien ruling class, the so-called Itzá”
(187). But, what is important here, is that this is combined with a sac-
rificial narrative in which the rule of the Itzá is a period of “drought”
and famine which is “brought about by the sexual sins of the people”
(224). The drought and famine are stressed repeatedly:“It will be then
that the water sources are dry”; “Their bread/will be snatched away,
the water snatched out of/their mouths”; “The earth/will burn, the
sky will burn” (196; the poem is cast as a prophecy, hence the future
tense). Subsequently, we read about “Days of hunger and days of
thirst” (215), “for great is the punishment of guilt” (216). This leads
to a partial exile as well when “The drought-stricken Maya, driven
from their towns, roam through the forest in search of food and
water” (230–1). However, eventually, the “ceremonial penances” sur-
rounding “the human victim,” who “is purged,” make “the heavens
open and the (fruitful) rains fall” (237). Specifically, “Bound are his
eyes . . . the poison is summoned . . . [a]nd he will rise up to a dif-
ferent world” (220–1). Though Bierhorst refers to a human victim,
the poem subsequently refers to “the death of . . . the deer, a/sudden
death” (221), the deer “stand[ing] for the human surrogate” (265),
in Bierhorst’s view. In any case, this is a straightforward sacrificial
narrative, with sin giving rise to famine as well as exile, and food
restored only after a ritual sacrifice.
As to Quetzalcoatl, Bierhorst explains that there are five extant
fragments of the poem. The first fragment concerns the origin of
food and the theft of food by the rain gods (20–1). As Bierhorst
notes, this “explains how agriculture became dependent upon the
rain gods” (71n.12), which is to say, how the satisfaction of hunger
became dependent upon rain and thus on the will of deities. This
clearly prepares for the sacrificial narrative. The second fragment is
not really part of the same sacrificial sequence. It is a brief heroic
plot in which Quetzalcoatl’s father, the Sun, has been murdered by
Quetzalcoatl’s uncles. There is a sort of feint, suggesting sacrificial
narrative, when Quetzalcoatl asks “What shall I sacrifice to the tem-
ple?” (22). However, the sacrifice is a trap by which he captures and
kills the usurping uncles. He goes on to conquer many lands. In a
very truncated version of the epilogue of suffering, he falls ill when he

reaches Tlapallan, “the mysterious land of the rising sun” (74), where
he dies quickly and is “consumed in flames” (24). The third fragment
is more complex. It focuses on Quetzalcoatl’s new religion, including
his opposition to human sacrifice and his practice of offering “only
snakes, birds” instead (28; “butterflies” are also added to the list). The
important point here is that Quetzalcoatl himself becomes the per-
fect sacrifice in this scheme, for he is the “plumed serpent” (4), both
snake and bird. Subsequently, Quetzalcoatl is seduced into drunk-
enness and perhaps incest (81n.45). He is “filled with remorse” (34).
This transgression (involving food – or, rather, drink) leads to his exile
and self-sacrifice. Specifically, he “traveled” and “ventured widely.”
Eventually, “he set fire to himself” (36). In this fragment, the sin and
reparation are presented as wholly individual. They do not appear
to have any social consequences. In the fourth fragment, however,
the social implications become clear – and they are precisely the
sort we would expect. Initially, during Quetzalcoatl’s time there was
“great abundance” of “foods” and of “all the crops” (39). His people
“never wanted. There was/nothing lacking in their houses, they were
never/hungry” (40). However, due to the fact that “Quetzalcoatl and
all the/Toltecs grew lax” (41), this period of plenty and ease ended.
Most importantly for our purposes, “Stones . . . rained down” (55).
Bierhorst explains that “A rain of stones means barrenness, presum-
ably brought about by drought” (88–9n.80). Bierhorst’s explanation
is confirmed by the following events, in which people are sacrificed
according to “the drought rituals held in Aztec times” (89n.81). The
sacrifices are only partially successful. There is food, but it is “bitter,”
so much so that “it could no longer be/placed on the lips” (56).
The idea relates to the Quetzalcoatl’s forbidding of human sacrifice
and substitution of birds and snakes – which is to say, himself. (The
metaphorical point of the food being inedibly bitter after the human
sacrifice seems too obvious to require spelling out.) The following
section of the poem concerns Quetzalcoatl’s exile and death – or,
rather, his implied death, for the fourth fragment does not fill in the
death itself. Thus, we must return to the third fragment to see this
as a ritual sacrifice. In the fifth fragment, which is clearly the end of
the poem, we are told that Quetzalcoatl has departed and that “the
variegated maize . . . came forth” (65).
Related to the cross-cultural repetition of this plot structure, in re-

searching lyric, I had come upon quite a few poems, especially from
oral cultures, that addressed agricultural fertility or livestock. For ex-
ample, Firth and McLean cite a number of Tikopia songs that deal
with great abundance of fish (134), or with herons depleting the fish
supply (133), with “a time of grave food shortage” (130), and so on.
Deng notes the centrality of cattle to Dinka life. They “provide an es-
sential part of their food supply” (30) and also serve “as the medium
of reconciliation between man confronting God and ancestral spirits
through sacrifice” (31). A very large number of Dinka lyrics concern
cattle (see 96–158 – over one-third of Deng’s collection). This distri-
bution makes sense, and fits the narrative hypothesis perfectly, given
our third prototype narrative. Otherwise, it would appear surprising.
In addition to these cattle songs, there are songs of “offerings,” sung
“during sickness, war, drought, famine” (238). One of these lyrics
makes the general point of sacrifice, “If I wrong Him/I make it right”
(239). Deng explains that this “making right” is accomplished “by
sacrificing or consecrating an animal to the offended spirit” (239n.6).
Numerous Native American poems address deities of corn or
call for rain. For example, Cronyn’s classic collection (see Lincoln)
presents many instances. From the Sia, he includes a “Prayer for
Rain,” a “Rain Song of the Giant Society,” and an “Invocation” that
adjures “Bring your showers and great rains . . . /I throw out to you
my sacred meal . . . /All come out and give us your showers and great
rains” (64–5). There is a Pueblo poem “announc[ing]” a “feast” and
calling on the clouds “that the planting may yield abundance” (81).
There is a Zuñi “Song of the Blue-Corn Dance,” also calling on the
clouds (82), and a “Corn-Grinding Song” (123). There are Hopi lyrics
about “the blossoming virgin corn” (124) and “Patches of beans in
flower” (125), a Teton Sioux “Song of the Famine,” and so on. A very
fitting sequence of lyrics may be found in the renowned Navajo Night
Chant. Bierhorst explains that “In its narrowest use,” the Night Chant
“serves as a form of therapy, conducted by a shaman . . . for the bene-
fit of a principal communicant” (281). The poems, directly integrated
into the loosely narrative structure of the ritual, recur continually to
rain and agricultural fertility. The “first morning prayer ritual” refers
to rain in the opening stanza (293). The communicant’s first speech
begins “Owl!/I have made your sacrifice” (294). After calling for a
restoration of his/her physical well-being, he/she goes on to call for

“abundant dark clouds,” “abundant showers,” and “abundant vege-
tation” (295). This mixing of the physical well-being of the individual
with agricultural plenty for the society indicates again that the pro-
totype for physical happiness is having plentiful food. Subsequently,
the various poems make repeated reference to rain, the “sacred blue
corn-seed” (313), “The white bean/And the great corn-plant” (314),
“The voice of the thunder . . . that beautifies the land” (318), and so
on. The Night Chant ends when the deity “takes up his sacrifice” (331;
see 346n.67) and “The corn comes up, the rain descends. . . . /The rain
descends, the corn comes up” (346n.65), and of course the well-being
of the communicant is assured.
As this suggests, lyrics are often explicitly embedded in sacrificial
narratives, including rituals. A striking case of this may be found in
the Kondh lyrics used in the ritual of human sacrifice. (The Kondh are
one of the “tribal” groups of India.) These poems punctuate the ritual,
which is itself an implicit plot. They include a song concerning the ac-
quisition of the scapegoat, the “first stab by the priest” (Mahapatra 1),
and, most revealingly, the “Invocation to Earth,” that begins “Let no
famine/Visit our land”; instead, allow “Our crops” to “flourish” in
“plenty” (2). (For a discussion of the sacrifice itself, which has not
been practiced since the nineteenth century, and its agricultural pur-
poses, see Mahapatra xviii–xlii.) Other ritual examples may be found
in Christian services and various Central Asian shamanistic rites (see,
for example, Chadwick and Zhirmunsky 248–9).
Perhaps even more interestingly, like heroic and romantic struc-
tures, sacrificial plots are also to be found in historiography. A
straightforward example is that of Tuñjı̄na in Kalhan.a’s twelfth-
century history of Kashmir, the Rājataraṅgin.ı̄. There is famine in
Kashmir. The king says to his queen, “it is through some trans-
gression of ours . . . that such a calamity . . . has befallen the inno-
cent people” (53). His solution is just what we would expect,
“Therefore . . . I shall now sacrifice my body in the blazing fire” (54).
As it turns out, this is unnecessary, for the queen is able to end
the famine, evidently through her “pious conduct” (55). This pious
conduct is centrally a matter of “Devotion to the husband” (55).
In keeping with this, many years after the famine, the queen com-
mits satı̄, which is to say, throws herself into the funeral pyre of her
husband (55) – thus, in effect, performing the sacrifice that he had not
performed at the time of the famine.
No doubt the most commonly known sacrificial narrative, world-
wide, is the Judeo-Christian story of the Fall and Redemption of
humankind. The story of Jesus might at first seem quite different
from the Yoruba story of Land and Heaven or the songs of the Water
Goddess and the Owl God. However, it is, in fact, closely related to
them. It, too, is a variation on the basic sacrificial structure. The full
story begins in “Genesis.” Adam and Eve offend God when, acting
from pride and greed, they violate his one prohibition, a prohibi-
tion regarding food: “of the fruit of the tree in the middle of the
garden God said, ‘You must not eat it, nor touch it, under pain of
death’“ (3.3). In keeping with the standard pattern, their individ-
ual violation has broad, communal consequences – indeed, universal
consequences, for it affects all of humankind. First, it “Brought Death
into the World,” as Milton had it (1.3). While the Ainu and Yoruba
stories concern particular deaths, the Judeo-Christian story concerns
death in general, death in all its forms. Moreover, this death is bound
up with hunger. When God created the garden of Paradise, the first
thing he did was produce food: “there was as yet no wild bush on
the earth nor had any wild plant yet sprung up, for Yahweh God had
not sent rain on the earth.” Then “Yahweh God planted a garden in
Eden. . . . Yahweh God caused to spring up from the soil every kind of
tree, enticing to look at and good to eat” (2.5, 8–9). God’s first words
to Adam and Eve concern food – along with sexual union and social
domination, thus covering our three prototypes: “Be fruitful, multi-
ply, fill the earth and conquer it. Be masters of the fish of the sea, the
birds of heaven and all living animals on the earth. . . . See, I give you
all the seed-bearing plants that are upon the whole earth, and all the
trees with seed-bearing fruit; this shall be your food” (1.28–30).
In keeping with the general structure of the primary narrative
modes, the way in which death enters the lives of Adam and Eve is,
first of all, through exile – exile from this garden, which is defined by
its abundance of food. The words in which “Genesis” expresses this
exile stress precisely the loss of this food: “So Yahweh God expelled
him from the garden of Eden, to till the soil” (3.23). Indeed, life itself
is identified with a particular food in this lost garden, the fruit of the
“tree of life,” from which humankind is particularly forbidden (3.22).
Finally, this exile involves two particular curses from God. The second
is the crucial one for our purposes. Though not a curse of famine, it is a
curse that allows famine: “Accursed be the soil because of you./With
suffering shall you get your food from it” (3.17). In Christian teaching,
this expulsion was so absolute that it continued not only in earthly life,
but after death as well. It was an absolute and all-encompassing con-
demnation. Despite this, eternal life could still be regained. God could
grant the gift of life, return us to the plenteous garden. But this could
occur only after a sacrifice.
The narrative continues millennia later with Jesus. Though God,
Jesus is incarnate. Thus, his sacrifice is a sacrifice on behalf of human-
ity. Typically the sacrifice itself has to come from humans, though it
is often bound up with the aid of a benevolent deity, as we have seen.
Jesus combines these two roles, taking divine benevolence one step
further by sacrificing himself. Like the Yoruba carrier, then, Jesus is
a sacrificial victim who communicates between earth and heaven,
putting an end to the curse of eternal death and exile from abun-
dance, just as the sacrifices of the Ainu and Yoruba stories put an
end to the more local and mundane death resulting from a particular
famine. As part of this sacrifice, Jesus institutes a ritual in which the
sacrifice itself is reenacted, specifically through the consumption of
food. (Note that in the Ainu stories, the ritual is initiated by the helper
deity – the Water Goddess, in one case, the Owl God, in the other.)
Images surrounding these events (for example, the characterization
of Jesus as a sacrificial lamb) reenforce the link with more ordinary
sacrificial stories treating famine, as do various peripheral elements,
such as the common belief that it rained when Jesus died. Finally,
Jesus’ resurrection serves as a concrete example of the way this sac-
rifice has overcome the lapsarian condition of death and led God to
give life and plenty back to the damned society – at least in offering
the possibility of eternal return to Paradise after death.6
6 Not all sacrificial narratives are ancient. A famous modern instance is William Butler
Yeats’s The Countess Cathleen. In this story, the peasants of Ireland are starving (“the
land is famine-struck” [2]). In consequence, they are selling themselves to the devil.
Cathleen agrees to sacrifice herself in order for the starving peasants to be saved,
both physically and spiritually. She sacrifices herself, is transported to heaven, and
the peasants are restored. Yeats uses this structure not only to communicate religious
points, but to make quite modern social and political points as well. For example,
a note on ethics and sacrificial tragi-comedy

As we saw in Chapter 4, each prototype narrative structure raises par-
ticular ethical issues. The personal prototypical narrative, romantic
tragi-comedy, raises the ethical issue of the degree to which society
can circumscribe personal freedom. The social prototypical narrative
the demons clearly represent the merchant classes and the capitalist order, while the
countess represents the aristocracy and – along with the peasants – the feudal order.
Thus Yeats uses the sacrificial structure to comment on capitalism, and on the place
of the aristocracy and the peasantry in a mercantile economy.
Needless to say, Yeats is not the only important modernist who draws on sacrifi-
cial tragi-comedy. For example, Kafka relies on this structure – though his treatment
is far more elliptical and ironic than that of Yeats, as one would expect.
A still more recent example may be found in Kamala Markandaya’s widely read
novel, Nectar in a Sieve. The story concerns Rukmani and Nathan, a poor peasant
couple in South India. At one level, their suffering is explained in straightforwardly
materialist terms. Specifically, it results from the fact that they do not own their land
and from the intrusion of heavy industry into the region. But the plot draws im-
plicitly on the sacrificial structure as well. Rukmani and Nathan begin well enough.
But Nathan kills a “sacred” cobra (20) among the ripening fruits and vegetables of
the garden (18). After this, their hardships begin – prominently including drought,
famine, and exile. The initial state of plenty is not restored for many years. Specif-
ically, Nathan dies, in a temple (thus in a religious context), during a rain storm
(recalling, and inverting, the drought that began their suffering long before). After
this, Rukmani’s situation improves quickly. She not only has enough food, but she
is able to return home as well. Nathan’s death is clearly an implicit sacrifice that
allows for her well-being – and indeed for the well-being of the larger community,
for when Rukmani returns home, she finds a hospital serving the poor people of the
area.
This is, of course, a modern version of the sacrificial structure. Though he dies in
a temple, Nathan is killed by the wage labor he takes up in the city; he is killed indi-
rectly by capitalists, not directly by a priest. Moreover, the new life brought about by
this sacrifice is not merely a matter of agricultural conditions (though there is rain, as
just noted). It is equally a matter of cooperative and rational work to help the poor.
Yet, despite all this, the family’s improved condition is still a sacrificial restoration.
It is a restoration of sustenance brought about by death in a temple. Moreover, this
death compensates for an earlier sin against nature, a sin linked with food. In making
her modern social and political points, Markandaya has not created a wholly new
and modern narrative structure. Rather, she has varied and specified the ancient and
universal structure of sacrificial tragi-comedy.
These are not isolated cases. As already noted, Amos Tutuola takes up sacrificial
narratives explicitly. This is true also for the Nigerian Nobel laureate, Wole Soyinka,
who draws on sacrificial motifs overtly in The Strong Breed and Death and the King’s
Horseman, implicitly in The Bacchae of Euripides, and still more subtly in The Swamp
Dwellers (on the mythic implications of the last play, which indicate its sacrificial
links, see my “Particular Myths”).
Contemporary films make use of this plot as well. Santosh Sivan’s award-winning
The Terrorist in effect sets out to undermine the militaristic use of the sacrificial
stages the contradiction between the ethics of defense and the ethics
of compassion. In short, they raise personal and social issues re-
spectively. The sacrificial plot most obviously involves obedience to
God. Simply put, defiance brings suffering. In a sense, then, it is a
straightforward narrative of moral conformity. However, there are
complexities here. Sacrificial plots often raise issues about the proper
treatment of nature, which is bound up with divinity in these nar-
ratives. They touch on what we would now see as ecological ques-
tions regarding humanity’s relation to the rest of the natural world.
Specifically, they address the conflict between the human need to
consume nature and the ethical imperative to respect or revere na-
ture. In the Judeo-Christian story, for example, God begins by as-
serting human supremacy over nature – but he immediately circum-
scribes that supremacy. The point is even more obvious in the Ainu
stories.
Prototypical narratives also define morally admirable acts that bal-
ance their particular ethical violations. Heroic plots value loyalty as
the reverse of usurping betrayal. Romantic plots value the assertion
and pursuit of one’s individual affiliative choice against overbearing
authority or social conformism. Sacrificial plots value self-denial as
reparation for greed and arrogance. Of course, this is not just any sort
of self-denial. Just as loyalty in heroic plots is prototypically a militant
or combative loyalty, self-denial here is prototypically self-denial with
respect to food. In other words, the ethics of sacrificial tragi-comedy
is in part an ethics of self-abnegation that focuses on one’s own body
and that, in effect, denies that body or, rather, makes it suffer for a
higher good, repudiating the flesh that comes from self-indulgence
in food. This ethical prototype manifests itself cross-culturally in
narrative by consistently inverting the plot and opposing sacrifice to fertility. In Luis
Puenzo’s The Plague, the plague is ambiguously linked with “collective punishment”
by a priest’s sermon. The priest subsequently sacrifices his own life among those who
have died. Almost immediately after his death, the cold weather arrives and puts
and end to the plague – a point made explicitly by the main character. Note that the
shift in weather is directly parallel to the shift from drought to rain, even though
the plot here has been varied from the prototype structure, with disease substituting
for famine. This variation would not necessarily suggest a weather-based solution
to the problem, were it not for the fact that the basic prototype is famine.
The list of modern and contemporary sacrificial narratives could be very much
extended – not as much as the lists for the other genres, but a great deal nonetheless.
a range of practices from Hindu asceticism and Jain self-starvation

(called “sallekhana” – for discussion, see Dundas 155–6), through
Christian and Muslim fasting, perhaps to the modern, secular obses-
sion with slimming, and even to certain aspects of anorexia.7
conclusion
Our consideration of Ainu epics has to a great extent provided cor-
roborating evidence for our previous hypotheses concerning narra-
tive structure. However, by rendering salient certain motifs and pat-
terns that are common cross-culturally, it has led to some changes
in emphasis and formulation – prominently a greater stress on and
elaboration of the love triangle in romantic tragi-comedy. More sig-
nificantly, it has suggested that there is a third prototype for happi-
ness, a physical prototype (plenteous food) and a correlated, phys-
ical prototype for sorrow (famine). This prototype, too, generates
a set of universally recurrent narratives. In these sacrificial tragi-
comedies, some act of human greed or arrogance, usually regard-
ing food, violates the sanctity of nature and offends some deity. In
punishment, the society in question suffers famine, perhaps along
with exile. The famine ends only when the society makes a sacri-
fice, often a human sacrifice. The human victim might be killed,
thus transporting the sacrifice to the heavens, or he/she might be
expelled from the society, symbolically carrying away the initial sin.
This structure appears autonomously and in combination with heroic
and romantic plots (as when martyrdom – a sacrificial idea – enters
heroic tragi-comedy). It raises distinct ethical issues regarding human
rights over nature and establishes repudiation of the flesh as a moral
value.
The isolation and definition of this third narrative prototype is, of
course, tentative – like the isolation and definition of the first and
second narrative prototypes. It is a step in a research program. It is
not a final resolution of problems, but a hypothesis set out to account
7 I am not, of course, trying to explain anorexia through sacrificial narrative. I am

simply noting that the ethical component of sacrificial narrative appears to bear on
compulsive slimming, insofar as the concern with thinness is driven by ethical con-
cerns, by a view of fleshiness not as unattractive, but as indecent, immoral, “obscene,”
as Jean-Paul Sartre puts it (see 401–2 of Being, 407 of L’être).
for some patterns. As such, it should lead to further questions, fur-

ther study, further testing, and so on. Indeed, this examination of
Ainu epic has been a success – if it has – not insofar as it has reached
definitive conclusions, but insofar as it has advanced a research pro-
gram in such a way that the theories encompassed by that program
can be further evaluated, criticized, revised, and extended in the
future.
7
The Structure of Stories

Some General Principles of Plot
One of the major contentions of the preceding chapters is that narra-

tives are generated, not by complexes of abstract conditions, but by
much more concrete prototypes. Obviously, there are exceptions to
this. If a playwright determines to follow a five-act structure or an
epicist accepts the notion that a heroic poem should begin in medias
res, his/her work will be organized in part by reference to abstract
conditions. However, the application of such abstract conditions is
very limited. Indeed, one might argue that even in these cases we are
mistaken in attributing a primary generative function to the general
principles (“Organize the play into five acts,” “Begin in the middle
of things”). For any author, these general principles are bound up
with standard types and with particular instances. If I write an epic
in twelve books, I have Milton and Virgil in mind. My conception of
the twelve-book structure is shaped by their poems. In other words,
here too my work is generated from more concrete precedents – in
this case, exempla.
Nonetheless, to say that narrative is most importantly derived
from prototypes (and exempla) is not to deny the possibility of mak-
ing general statements about narratives. Indeed, to say that narratives
are generated from emotion prototypes is one such general statement.
The purpose of this chapter is to articulate and develop some further
generalizations about narrative, as they derive from the preceding
analyses and bear on our understanding of the relation between nar-
rative and emotion. In fact, it turns out that there is a great deal that
202
Some General Principles of Plot 203
can be said about narrative at this level, a great deal that is, I believe,
both surprising and theoretically consequential.
Before going on, however, it is important to stress two points. First,
the following discussion is in no way intended to be a complete ac-
count of the abstract principles of narrative. Again, it is a treatment
of some significant generalizations that concern narrative and emo-
tion as analyzed in the preceding pages. Indeed, even on this topic,
it barely scratches the surface. Second, these generalizations are not
absolute laws giving strictly binding conditions. In fact, they remain
closely related to prototypes. Specifically, the following generaliza-
tions aim at defining tendencies of narrative production and reception
that move toward prototypicality. They do not set out to delimit what
narrative is, full stop. They isolate the properties, relations, and struc-
tures that make narratives more prototypical. It is perfectly possible
to have something that counts as a narrative, but does not fit these
generalizations.1 This is not to say that such generalizations do not
1 It will be clear to readers familiar with “classical narratology” that my approach

differs radically from that of “story grammarians,” such as Rumelhart and Prince.
The goal of a story grammar is well articulated by Thomas Pavel, who explains
that, “Properly constructed, an abstract narrative grammar should generate only
the correct narrative strings” (12). There are several problems with this approach.
First, it treats texts or utterances as if they are neatly divisible into two categories –
stories and nonstories, with stories following the strict rules of story grammar and
nonstories failing to follow those rules. This seems to me implausible. Rather, our
judgment of “storiness” is a matter of degree. Again, it involves approximation to
a prototype. Second, story grammarians regularly begin with the assumption that
a particular linguistic theory of grammar, with its particular theoretical organiza-
tion, applies directly to narrative. There is no reason to make this assumption. Yes,
we do tell stories in language. But we also do physics and logic in language and
we do not assume that we must use the same type of rules for language and log-
ical inference. Brewer notes that “When psychologists first came into contact with
generative-transformational linguistics, an attempt was made to convert it directly
into a psychological theory. . . . After about a decade of work it became evident that
linguistic theory could not be used in a simple fashion as a psychological theory”
(“Literary Theory” 222). Lerdahl and Jackendoff make much the same point. Though
they are speaking of music, the problem they identify is generalizable. As they put it,
“many . . . applications of linguistic methodology to music have foundered because
they attempt a literal translation of some aspect of linguistic theory into musical
terms” (5). The claim holds equally if we substitute “narrative” for “music” and
“musical” in this sentence. To make matters worse, Chomskyan linguistic theory
has changed radically since its early formulations, taken up in narratology. Thus,
the linguistic principles assumed by story grammarians are, most often, no longer
accepted within linguistics.
involve rules and principles. They certainly do. These are just rules
and principles with a tendential form (for example, one rule might
state that a particular property contributes to prototype status). The
crucial point is that these are not rules in the sense of statements that
give necessary and sufficient conditions for narrativity.
the telic nature of plot

The most obvious general point that can be drawn from our analyses
is that prototypical plots involve a person and a goal. This funda-
mental idea has been recognized by writers from the ancient Sanskrit
theorists to Keith Oatley. It has also been noted by more narrowly
empirical researchers. For example, Katherine Nelson maintains that
Admittedly, these objections apply more to some theorists than others. For ex-
ample, they apply less to Pavel than to Rumelhart. Nonetheless, the basic difference
in orientation remains. Put simply, story grammarians tend to seek necessary and
sufficient conditions for storiness and they tend to understand necessary and suffi-
cient conditions in terms of generative rules modeled on Chomskyan grammar. The
prototype approach was designed precisely against the assumption of necessary and
sufficient conditions.
This is not to say that I entirely reject the project of the story grammarians. First,
their attempt to examine story structure with precision is extremely valuable. Even
if one rejects the claim that they have defined a grammar, they undoubtedly isolate
some important patterns. For example, Rumelhart’s analysis of episodes in terms
of events and reactions (see 214–18) is insightful and clarifying, no matter what
one’s broader theoretical commitments. Second, I am not at all claiming that more
general structures do not enter into narratives. They most certainly do, as I have
noted repeatedly. For example, it is clear that we require scripts, of the sort famously
isolated by Schank and Abelson, to understand both everyday life and stories. Sup-
pose a story involves the statement, “I picked up the receiver. 785–4432. I waited.
Nothing.” Clearly, we require a script for telephone use if we are to understand this
passage. David Herman has made this case with particular force.
But, again, all this is something different from story structure itself. My con-
tention is that story structure most importantly derives, at least in its large contours,
from cognitive processes that are prototype based at every stage. One main pur-
pose of the preceding chapters, has been to outline those processes and support this
contention. In short, though I find script theory enormously valuable, and I find
many Chomskyan principles enormously valuable, I do not believe that the particu-
lar narrative phenomena considered here – the literary universals of tragi-comedy –
are best understood in terms of scripts or generative rules modeled on Chomskyan
transformational grammar. Rather, my argument is that they are best understood by
reference to particular emotion prototypes. Moreover, generalizations that bear on
these universals are themselves related to prototypes in isolating prototype-related
tendencies, not necessary and sufficient conditions, as already noted.
“At the most basic level, narratives consist of the report of a sequence
of actions by actors that are connected in some way, usually because
they are organized to achieve a goal or solve a problem” (5; in my
usage of “goal,” solving a problem is one sort of goal). If Nelson is
correct that the agent/goal structure is the most common structure
and the one found in children, it makes sense that this would be pro-
totypical. Of course, there is not only an agent/goal structure, but
also some sort of causal sequence – as Nelson indicates, and as vir-
tually every writer on narrative has noted for the last 2,500 years as
well.
In short, prototypical narratives have a telic structure including an
agent, a goal, and a causal sequence connecting the agent’s various actions
with the achievement or nonachievement of the goal.
We should consider a few complications of each component.
The Agent
There has been a good deal written on the nature of the protagonist.
For example, Bal points out that the hero tends to receive more de-
scriptive attention, to appear more frequently in the narrative, to have
relations with a larger number of other characters, and so on (132). At
the same time, Frye notes that heroes can be almost null, with all the
focus being on subsidiary characters – as in many New Comedies,
where it sometimes happens that we hardly know or see the lovers,
all our attention being drawn to the wily slave and the senex iratus
(167). These are valuable points, worth sorting out in their differ-
ent emphases and contexts. However, I should like to explore a few
aspects of the protagonist that bear more directly on the preceding
analyses.
Socially Representative Characters and Individualistic Characters

A number of writers have noted that we at least partially empathize
with the protagonist. Even if we disapprove of him/her, we must
in some sense share his/her feelings and interests if we are to be
engaged by his/her pursuit of a goal. Again, this empathy is likely
to be based on identity categories, such as nationality, or on shared
situations, such as being in love. As the preceding examples indi-
cate, romantic stories are likely to rely on situational empathy, while
heroic stories are more likely to draw on identity categories. In fact,

love stories often work against identity categories (such as economic
class or nation) by putting the lovers on two sides of some identity
conflict. Heroic stories, in contrast, are often formulated in the ser-
vice of identity categories – supporting the nation, the emperor, or a
particular ethnic group.
Or, rather, the construction of romantic plots relies from the begin-
ning on situational identification and that of heroic plots relies from
the beginning on categorial identification. However, situational iden-
tification enters into any engaging literary experience. As discussed
in Chapter 2, literature is structured in such a way that its emotional
effects are founded upon the priming of emotion-laden memories.
Our reading practices, as we approach literary works, reenforce this
tendency. In other words, the nature of literature and of the attitudes
we adopt to literature tend to stress the personal memories that foster
situational empathy. In this way, situational empathy is a likely prod-
uct of any literary reading. Indeed, this is no doubt one reason why
so many writers have felt that literature is humanizing. It humanizes
us in the sense that it tends to develop certain sorts of compassionate
identification. (Of course, it is not at all clear that this sort of iden-
tification extends beyond the literary work to the real world.) This
is particularly striking in the case of heroic tragi-comedies, for, once
again, in those cases the situational empathy very often runs con-
trary to the categorial identification that is central to the nationalist
or other propaganda purposes of such works. Indeed, the situational
empathy is often most significant with respect to the “enemy” and
not with respect to the hero, at least in those cases where the hero
achieves his/her goal, while the defeated enemy is left impotent and
miserable.
The point about the relation between hero and categorial identity
is worth dwelling on for a moment. To account for the different sorts
of protagonist in prototypical stories, we need to draw a distinction
between heroes – or, more generally, characters – who represent so-
cial categories and the norms they involve and heroes or characters
who represent individuality and personal affiliation especially against
social categories and the norms they involve. We may refer to these
simply as socially normative and individualist characters. The differ-
ence between individualist and socially normative characters is not
that one has individual characteristics and the other does not. Rather,
the difference is a matter of whether the character in question stands
for or against particular identity categories and the social hierarchies
these entail. The hero – or, rather, heroes – in romantic tragi-comedies
are most often individualist. We will be turning to other characters in
a moment, but even at this point it should be clear that blocking char-
acters, such as the forbidding father in New Comedy, are socially nor-
mative. Speaking of romantic comedy, Northrop Frye explains that
“The opponent to the hero’s wishes, when not the father, is generally
someone who partakes of the father’s closer relation to established
society” (164–5). Indeed, as just indicated, it is not merely a closer
relation. Characters of this sort represent the rules and judgments
of society. They stand for class, nation, race, or religion. Note that
the hero in romantic stories may be very nonindividualized (that is,
he/she may have few particular, “individualizing” characteristics),
but he/she is nonetheless individualistic in this sense. Conversely,
a blocking character may be drawn in considerable, idiosyncratic
detail, though he/she is socially normative.
Things are a little more complicated in the case of heroic tragi-
comedy. Again, the genre itself tends to be used – cross-culturally
and trans-historically – to support particular sorts of categorial iden-
tity. For this reason, the hero is almost always to some degree socially
normative. However, the hero in this genre often has individualistic
elements also. Indeed, sometimes, there is a split between the king or
emperor, on the one hand, and the warrior, on the other, as with Kay
Kavús and Rostám in the Shâhnâme. In these cases, the king is most
often a “pure” representative of the social order. The warrior is also
a socially normative character, but he may have consequential indi-
vidualistic elements as well. For example, Rostám clearly represents
Iranian nationality against the Turānian enemy. However, there are
sometimes very strong individualistic strains in his character and ac-
tion, as when he refuses to come at Kay Kavús’s bidding, then refuses
(however briefly) to serve the Shah in battle. A further complication
here is that the pure social representative – whether split off from
the partially individualistic warrior or not – is often characterized in
such a way as to indicate that he/she does not really merit that role.
For example, the inadequacy of Kay Kavús is made entirely clear by
Ferdowsi. He is no more worthy to make decisions for other people
than are most New Comedy fathers. Yet at the same time Ferdowsi
clearly supports loyalty to the Shah. Similar points could be made
about Rāma in the Rāmāyan.a, the Pān.d.avas in the Mahābhārata, and
so on. Indeed this inadequacy is part of the conflict between categorial
and situational empathy in heroic plots. This suggests that categorial/
situational conflict is more pervasive in the heroic genre than is at first
apparent. Though it only becomes obtrusive through the epilogue of
suffering, the conflict may be there, in a more submerged form, from
the outset.
In connection with this, an entire group of prototypical narratives
results from shifting the warrior protagonist in heroic tragi-comedy
to a fully individualistic character. This is, for example, what happens
in stories that fall under the broad category of Romantic Satanism. To
take the most famous instance, insofar as Satan is the hero of Paradise
Lost, we have a hero who is entirely individualistic. He rebels against
the socially normative God the Father in much the way a lover rebels
against an oppressive parent in romantic plots. Karn.a’s choice of
personal affiliation for Duryodhana over his birth brothers – and his
related war against these rightful rulers – is a case of this general sort
as well. Though Karn.a is not “the hero” of the Mahābhārata, his heroic
status is indicated at different points in the poem, including the end,
where he appears in Heaven with the Pān.d.avas. This status is also
presented in other works, such as Bhāsa’s early Sanskrit play, Karn.a’s
Burden, in which Karn.a is the tragic hero. More generally, Karn.a has
a status in Indic literature not entirely unlike that of Milton’s Satan
in European literature after Romanticism.
I should note that this structure is different from that found in
narratives of social rebellion, which provide another variation on
the heroic plot, still of very high prototypicality. In stories of social
rebellion, the author splits the hero and the emperor or other official
leader, making the hero socially normative, but of a different group –
a different nation, ethnicity, economic class, religion. Unsurprisingly,
narratives of this sort are not common in feudal societies, except in
cases where the official ruler is viewed as a usurper. An example
of this sort would be Chikamatsu’s The Battles of Coxinga, where the
emperor is a Tartar and the heroes, seeking to restore the rightful
(Chinese) heir, are Chinese. On the other hand, one might argue that
this is true even for, say, Marxist works, in which the bourgeoisie
holds social power through a sort of usurpation of the rights of the

workers.
These two variations on the heroic structure are particularly strik-
ing because parallel variations in the romantic structure are extremely
uncommon. Prototypical romantic plots are already in effect Satanic
and rebellious. Thus the only evident parallel to the rebellious vari-
ation (in which the roles of rightful ruler and rebel are reversed)
would involve making the parents, not the children, into the heroes.
The parallel to the Satanic variation (in which the – otherwise socially
normative – hero is made individualistic) would presumably involve
making the – otherwise individualistic – lovers socially normative.
As to the former, there are a few, modern South Asian works that
present a child’s desire for love marriage as a mistake and the social
preference for arranged marriage as preferable. But these are fairly
rare. The second variation, which shifts the lovers from individual-
istic to socially normative, does not seem to exist – at least not in a
prominent enough form to be noticeable.
This dissymmetry between heroic and romantic plots suggests that
narratives generally drift toward individualistic characterization of
heroes and they tend to favor individualistic over socially normative
characters. There is also a more limited preference for rebels over
rulers, as rebels may be heroes in either genre, but figures of so-
cial authority are heroes almost only in heroic plots. The tendency
to favor rebels may be simply the residue of the preference for in-
dividualism. But what leads to this preference? In part, this results
from the situational empathy discussed above. It is also related to
the general narrative tendency toward personalization – the trans-
formation of objectified characters into subjective persons (the main
reason it is difficult to sustain villainy, as we have noted). Perhaps
it is overly optimistic to say so, but this also seems to suggest a ten-
dency of narratives to drift toward expanding individual freedom
relative to social norms. This would appear to be an odd tendency in
and of itself. However, it is related to the increase in personalization
and empathy. As we develop greater emotional identification with
a character, we automatically come to view his/her preferences as
more human and valid. Thus, we come to see inhibitions on his/her
actions as less fair or reasonable – or, at least, as less unequivocally
good, even when necessary. In short, authors – especially authors of
heroic plots – may set out to create works that favor categorial iden-
tity and social hierarchy by celebrating socially normative characters
and repudiating individualistic characters. However, the develop-
ment of narrative tends to pull against this. The pull may be weak
in many cases and thus have little noticeable impact (that is, it may
not overcome an author’s intent of favoring categorial identity). But
it appears always to be a pull toward individualistic characters and
away from socially normative characters. In ethical terms, it is always
a pull toward individual freedom, not toward social constraint.
The situation becomes more complex when we consider the sac-
rificial plot in this context, for in the sacrificial plot these character
tendencies are often fused in surprising and complex ways. The cru-
cial characters in the sacrificial narrative are the violator, the offended
deity, and the scapegoat. The agent around whom the action is orga-
nized and who sets out to save the society may be the violator or the
scapegoat or someone else. Perhaps more importantly, the scapegoat
may or may not be the violator. If the scapegoat is not the violator,
then we have a, so to speak, “pure” sacrificial plot. In this case, the
scapegoat is a socially representative character, which is to say, he/she
stands for the entire community and purges all their sins. (Socially
normative characters are always socially representative, but a charac-
ter may be socially representative without thereby acting to enforce
the norms of the society.) The violator is most often an individual-
istic character. When the scapegoat and the violator are identified,
then the story is, so to speak, secularized. It is a story of retributive
justice, not of communal guilt.2 The story of Jesus and the Yoruba
myth of Earth and Heaven provide good examples of the pure sacri-
ficial plot. To take the former, Jesus is the scapegoat, while Adam and
Eve are the violators. Adam and Eve clearly acted individualistically,
2 It may be that the fullest version of the sacrificial plot incorporates both the pure and
the secularized versions, with the former leading to the latter. The influential 1922
F. W. Murnau film, Nosferatu, appears to manifest a structure of this sort. Nosferatu
is clearly an evil presence. He brings plague to the city of Bremen and may be
seen as the guilty violator. One young woman sacrifices herself to save her fellow
citizens. Once she dies, Nosferatu dies also. When these deaths occur, the plague
ends immediately. Thus, we have the sacrifice of an innocent scapegoat, in keeping
with the pure version, and the death of the violator, in keeping with the secular
version; conjunctively, they produce the restoration of physical well-being to the
entire society.
rebelling against God’s law. When sacrificed, Jesus was socially rep-
resentative, standing for all humanity. But he was not really socially
normative – or, rather, he was socially normative for divine society,
but an individualistic rebel for human society. (Recall the official rea-
son for his execution – “Jesus was condemned to death . . . as a rebel
against the Roman State in one of its subject provinces” [Cullman 6;
see also 42–3, and Pagels 6–7].)
The second sacrificial structure, in which the violator and the
scapegoat are the same, may be found in some of the most promi-
nent instances of this plot – for instance, the important Yüan drama,
The Injustice Done to Tou Ngo, and the paradigm European tragedy,
Oedipus the King. In the former, Tou Ngo is executed for a crime she
did not commit. The violators are, then, the people who participated
in the unjust execution. Before she dies, she prophesies that, due
to her murder, “the county of Ch’u shall suffer a drought” (Kuan
Han-ch’ing 143). The prophecy comes true. Three years later, as the
drought continues, a new magistrate arrives. After investigation, he
condemns all those who acted falsely in the prosecution of Tou Ngo
(157–8). Thus the scapegoats are the same as the initial violators. Their
punishment returns plenty to the land.
This play has remarkable complexity in the areas we are consid-
ering. Tou Ngo is socially normative in the content of her action, but
individualistic in its form. More exactly, Tou Ngo defends traditional
Chinese beliefs and practices. But, in doing so, she acts individualis-
tically against her mother and the court. In contrast, her antagonists
are socially normative in their formal position – as parent, magistrate,
and so on. But their motives are individualistic and their specific ac-
tions are contrary to the very social principles they are supposed to
represent and enforce. This is like the story of Jesus, then, in hav-
ing it both ways. The apparent conflict may be resolved by distin-
guishing between immanent social norms, which may be false, and
transcendent principles which are necessarily true and which may be
embodied in socially individualistic characters. Thus, the tendency
toward individualistic characterization and sympathy is connected
with a tendency toward separating ethical principle from common
social practice.
The cases we have been considering in this section suggest the
following broad narrative principles: A standard generative element
in prototype narratives is a conflict between individualistic and so-

cially normative characters. Ordinary narrative development tends
to foster sympathy for the former over the latter. Ethically this is
often allowed by an overt or tacit distinction between transcendent
and immanent norms, which is to say, ethical principles (or puta-
tively authentic cultural principles) and current cultural practices.
The transcendent norms regularly include an implicit maximization
of individual freedom, at least for subjectively defined or personal-
ized characters.
Empathic and Nonempathic Agents

There are, of course, many other ways in which one might organize
heroes. Any system of organization will be more or less valuable
depending on one’s purposes and the body of literature one is con-
sidering. The division between the individualistic and the socially
normative, however, seems broadly consequential – and, at the same
time, relatively underrecognized.
A second division that is consequential and largely unnoticed is,
in some ways, a version of Aristotle’s distinction among characters
who are like us, characters who are superior to us, and characters who
are inferior to us. Aristotle has touched on a significant principle of
differentiation here, but has not, I believe, drawn it precisely. The
important division has to do not with the intrinsic value of the char-
acters, but with empathy. It concerns characters with whom we can
in principle empathize situationally and characters who are, for some
reason, beyond situational empathic identification. This is related to
Aristotle’s division because our ability to identify with characters has
at least something to do with whether or not the characters are “like
us.” Indeed, perhaps the most obvious cases of characters beyond
empathic identification are absolutes – absolutes of good and evil,
power, knowledge, and so on. The initial impulse of much heroic liter-
ature is to put the enemy beyond identification in a demonic direction.
There is also an impulse (encouraged no doubt by royal patronage) to
put the king beyond situational identification in a divine direction.
In keeping with this, heroic tragi-comedies regularly link the ruler
with divinity and the enemy with Satanic or related powers of evil.
This division not only inhibits situational identification, but encour-
ages categorial identification (via the unifying, divinely chosen ruler)
along with a sense of group threat (from the demonized enemy). In

this way, it fosters an invigorated ethics of defense.
However, as we saw in Chapter 4, one striking thing about
paradigm literature cross-culturally is that it does not maintain this
division at all consistently. In fact, literary works routinely undermine
this division. In this regard, narratives regularly begin with an oppo-
sition between characters with whom we can identify and characters
with whom we cannot identify. But the development of narratives
itself conduces toward the extension of empathic identification to
increasing numbers and increasing varieties of characters. Put dif-
ferently, narratives might begin by setting up certain characters as
objects of awe or disdain. The development of narratives, however,
pushes away from both awe and disdain toward a sort of subjective
normalcy.
Another way of thinking about the issue is in terms of suffering. Sit-
uational empathy is most significantly triggered by suffering. There
is, of course, a difference between deserved and undeserved suffer-
ing. However, the fact that the victim of suffering deserves punish-
ment does not mean that situational empathy is impossible. Rather,
there are several variables governing when we feel situational empa-
thy for a character who merits punishment. First, there is the intensity
and extent of the punishment. It should be proportionate to the crime,
not excessive. Second, there is the degree of threat; insofar as the char-
acter is weak and unthreatening, we are more likely to feel situational
empathy, no matter how deserving the punishment. Third, there is
the presence or absence of remorse. A character’s feeling of remorse
seems to be a particularly powerful factor in inspiring our situational
empathy. Fourth, there is the salience of the pain. We are more likely
to empathize with suffering that we experience as salient – for ex-
ample, suffering that is represented in vivid and specific detail that
encourages visualization and recollection. Finally, there is the extent
of individualism. Again, there appears to be a pull toward individu-
alistic characters, whatever the initial intent of the work.
Most obviously, then, we are unlikely to have situational empathy
with those who do not suffer or who deserve unqualified suffering –
thus divine and demonic absolutes. But, again, the idea that there are
such characters is continually undermined by practical representa-
tions. As Romantic Satanism attests, this undermining occurs even in
myth, where literal gods and demons are in question. The point holds
still more obviously for cases where the deification and demoniza-
tion are metaphorical. The enemy soldier in pain looks just like the
comrade in pain. It is difficult to sustain a distinction between them.
The image tends to trigger personal memories and corresponding
identifications even when the enemy has been demonized.
In sum, a second important axis of character differentiation is
that between characters with whom we can identify situationally
and those with whom we cannot. This division contributes to many
common propagandistic functions of literary works (for example,
nationalistic functions). The nonempathic characters are often ren-
dered nonempathic by some sort of absolutization of their character
(for example, absolutization of their moral excellence or depravity).
However, as a result of individual particularity in characterization,
the course of narrative tends to undermine idealization and dehu-
manization, if these are established initially. Finally, we identify most
readily with characters who are suffering, especially when that suf-
fering is salient, excessive, and remorseful and the characters are
individualistic and unthreatening.
Ancillary Characters
In the preceding pages, I have made reference to characters other than
the hero, but unsystematically. In fact, the agent/goal structure im-
plies that there will be two types of ancillary character, types widely
identified by writers on narrative structure (see, for example, Greimas
205–7): helpers and opponents, characters who facilitate the hero’s
achievement of a goal and characters who inhibit that achievement.
There are many ways of organizing ancillary characters in a plot.
Again, these will be more or less valuable, depending upon one’s
purposes and depending upon the body of works under consider-
ation. However, helper and opponent are the only characters that
result directly from the agent/goal structure of prototype narrative.
Thus, they are the two types that have most obvious significance in
the present context.
These two types are relatively clear in themselves. Helping figures
seem particularly straightforward. To a certain extent, they are ver-
sions of the hero. In fact, sometimes it is difficult to distinguish one
from the other. In some heroic stories, the leader (for example, the
king) is the hero and any character who acts on his/her behalf (for
example, a warrior) is really a helper figure. In other stories, both
are heroes. In keeping with this, the variables mentioned previously
apply to helpers. Thus, helpers may be elevated above situational
empathy, most obviously when they are divine, but they may also
elicit pathos. They may be individualistic, but they may also be so-
cially normative. This does not often seem to be significant, for the
helper’s character traits tend to repeat those of the hero. It does some-
times happen that the helper is socially normative while the hero is
individualistic. But, in these cases, the helper typically stands for a
distinct set of social norms from those to which the hero is opposed.
This is probably most common in romantic works, but it occurs else-
where as well. One heroic example may be found in the Malian Epic
of Son-Jara, where Son-Jara is in conflict with his royal half-brother,
but receives aid from his aunt. While the half-brother represents the
male ruling hierarchy, the aunt represents a distinct, female social
organization, more ancient, more mystical, and more purely a mat-
ter of pre-Islamic Mande belief. (Of course, as is common in heroic
plots, Son-Jara himself is socially normative on the male side, though
the form of his relation to that hierarchy is individualistic. Here, as
elsewhere, we see a distinction between false, immanent practices
and true, transcendent principles.) Distinctions of this sort sometimes
have consequences for our understanding of a work’s political atti-
tudes and its ethical implications.
It is also worth noting that helping characters may be transferred
from one prototype structure to another or may take on characteristics
from different structures. For example, helpers in heroic plots may be
self-sacrificing. Consider an important scene from The Epic of Son-Jara.
Without the fruit of the shea tree, Son-Jara cannot perform the magic
necessary to succeed in achieving heroic greatness. He can find only
“one old dry Shea tree” (Fa-Digi l. 2418), which will not supply the
necessary fruit. Here we have a version of the famine scenario, writ
small, and integrated into a heroic plot. Son-Jara’s mother calls out,
“Before the break of day/That dried up shea tree here,/Let it bear
leaf and fruit” (ll. 2429–31) explaining that she will “change [her]
dwelling” before dawn (l. 2443). The next morning, she is dead and
the tree bears fruit. Thus, the sacrifice has produced food, via the
standard sacrificial sequence. However, in this case that sequence
has been integrated into a heroic narrative. This cross-genre use of

helping characters is obviously an instance of the more general cross-
genre borrowing of structural elements, motifs, and so on.
Blocking figures follow the same general patterns. As we have al-
ready discussed, blocking figures are regularly demonized or other-
wise set outside the realm of situational identification. However, once
again, this deempathization is regularly undermined in the course of
paradigm narratives. Thus, enemies often develop from nonempathic
to empathic characters.
The division between socially normative and individualistic char-
acters seems particularly consequential with respect to blocking char-
acters. Just as there are socially normative and individualistic heroes,
there are socially normative and individualistic opponents. In heroic
plots, the enemy king is socially normative. But the enemy warrior
is likely to be individualistic in just the way that the hero warrior
is individualistic. Indeed, the individualistic opponent is often more
an admired rival than a simple enemy. Moreover, typically, our situa-
tional empathy with the enemy in heroic plots is focused on characters
of two sorts: innocents, which is to say, characters who cannot be seen
as bearing the guilt of the enemy (most obviously, children), and in-
dividualistic characters, characters who are to some degree at odds
with the enemy, even though they are ultimately loyal. Hector in the
Iliad is a good example of the latter. It is fitting that the epilogue of
suffering should concern him, for he was always uneasy with Trojan
policy. Though he never betrayed his side, but rather fought with ex-
emplary bravery, he was nonetheless the dissenting voice. As such,
the individualistic enemy is in some ways more of a hero than the
individualistic hero, for the hero is ill at ease with the “right” side.
It is as if Hector simply happened to be born into the wrong camp,
and thus had no choice but to fight on its behalf, though his heart
and mind would have led to the right decisions. It is not clear that
Achillēs’ propensities are equally admirable. There is a similar divi-
sion in the Rostám and Sohráb episode of the Shâhnâme. There the
tragedy results in part from the fact that Sohráb is very young (thus
partially innocent), and very much at odds with the Turānian ruler
whose troops he leads into battle.
The same distinction between socially normative and individual-
istic applies perhaps even more obviously to the romantic antagonist.
The father in New Comedy, the censorious shaikh in the Persian

ghazal, the brothel-keeper (along with the parents) in certain kabuki
plays (see Halford and Halford 447), are all socially normative. The
rival, in contrast, may be either socially normative or individualistic.
Here too the socially normative characters are typically placed be-
yond empathic identification – in part because they do not suffer, or
suffer only the disappointment of not having controlled the lovers
and prevented their union. However, it is possible for an individ-
ualistic rival to inspire empathic identification, especially if he/she
is innocent of the lovers’ separation and his/her own feelings are
sincere (as with Laylā’s husband, in Niz.āmı̄’s poem).
In plots of sacrifice, the punitive deity is typically one opponent.
By the nature of the plot, this opponent cannot be defeated – hence
the need for appeasement through sacrifice. This deity is almost al-
ways socially normative. In this way, the punishing deity is a pure
enemy, in the sense of a socially normative opponent, though it is
of course theologically incorrect to say this. On the other hand, the
punishing god is not the only antagonist of hapless humans, such as
Adam and Eve. There is often an individualistic opponent as well.
This opponent is typically the figure who leads the violators (for ex-
ample, Adam and Eve) into transgression. The obvious example is
Satan – in Genesis, the serpent. However, the point is not confined to
the Judeo-Christian and Muslim traditions. In Yoruba belief, Ès.ù is a
deity whose function is to trick humans into committing violations of
divine prescriptions so that they have to make sacrifices to the gods
(see Awolalu 29). He, too, is an individualistic opponent or “enemy
rival.” Here again we have an indication of how Romantic Satanism
arises in its various forms. Romantic Satanism is the result of making
a demonized opponent individualistic (and making his/her suffer-
ing salient) while making the main deity a socially normative (and
nonsuffering) opponent. Whatever his intent, this is what Milton did
in Paradise Lost.
The Causal Sequence
A story involves not only agents but a series of events. The agent/goal
structure implies that the hero must at some point be lacking the
goal. Thus, in the comic form, the series of events involves, at the
very least, the general sequence: Hero lacks goal; hero pursues goal;
hero achieves goal. In some cases, the sequence is circular, with two
elements preceding those just mentioned: “hero has achieved goal;
hero loses goal,” or something along those lines. This sequence must
involve causal links. Or, rather, the later events and conditions must
follow plausibly from the earlier events and conditions. This is most
often a matter of causality in the real world sense. However, there are
other ways in which a series of events may be linked together. What
is important in stories is not precisely objective causal plausibility.
Rather, it is the subjective sense that new events are understandable
as the result of preceding events. Aristotle emphasized this when
he noted that narrative outcomes may appear plausible or convinc-
ing due to “design,” whatever their causal status. Aristotle’s exam-
ple concerns the death of Mitys’s murderer: “the statue of Mitys at
Argos . . . fell upon his murderer while he was a spectator at a festi-
val, and killed him” (39). As Aristotle noted, we accept this sort of
sequence – “poetic justice,” as it is called in English. Even though
it hardly makes sense in terms of objective causality, it strikes us as
plausible. The point is not confined to cases of poetic justice per se.
In connection with this, we may distinguish three types of causal-
ity that operate significantly in narrative. The first is the ordinary,
broadly predictable causality of the macroscopic world. Not only the
causality of objects operating on objects, but the predictable causal-
ity of human experience and action – certain objects cause fear and
lead to flight, and so on. Within this normal or “general” causality,
we need to distinguish the ordinary causality that applies to things,
“efficient” causality, from the ordinary causality that applies to per-
sons, “final” causality. Final causality is a causality of means and
ends. A final causal explanation involves naming a goal and a set of
beliefs: Why did Smith go to the store? He wanted to buy some milk.
Efficient causality is a causality of objectal events and their effects,
prototypically events and effects related to one another by physical
laws – unsupported objects fall to the ground; fire burns human flesh;
lack of water causes death. It is, roughly, “billiard ball” causality.
Perhaps the most important point to make about causality in pro-
totypical narrative is that it pushes inevitably toward final causality.
In keeping with its emotive orientation, prototypical narrative tends
to maximize explanation in terms of intent. Such narrative not only
presents the main events in terms of an agent’s goal and actions to

achieve that goal. It also tends to transform as many causal sequences
as possible into final causal sequences. We may refer to this as the
personalization of causal factors. Prototypical narratives, then, tend to
personify and individualize conflicts with nature or with broad social
structures. The tendency of prototypical narratives is to treat as many
events as possible in intentional terms, in terms of human thoughts,
feelings, propensities, attitudes, affiliations, and aims. The point is
particularly obvious in sacrificial narratives where forces of nature
are personalized, such that a famine is the withdrawal of rain by
a god. The way to end the famine is to appease the angered deity,
compensating for the sin that caused this punishment.
The second type of causality is a sort of augmentation of ordinary
causality to include common beliefs about causal relations in the real
world. In a society that believes in witchcraft, witchcraft may figure
in the causal sequences of the narrative. Indeed, once a belief of this
sort exists in a society, and people become aware of it, it is almost
always available for incorporation into fictional narratives, even nar-
ratives outside the culture where the belief is accepted. For example,
witchcraft appears commonly in narratives written and read by peo-
ple who do not believe in witchcraft. One might think that the prob-
lem with, say, literary witchcraft is explaining why people would
accept such a thing in a literary work when they reject it in real life.
But, in fact, it is really not all that surprising that people who do
not believe in witchcraft would be able to accept it in fiction, or that
they would be able to accept other objectively implausible types of
causality. After all, we accept new events, persons, conditions that we
know do not exist, why not counterfactual causal principles? Rather,
the problem with witchcraft as a causal device is just the opposite – its
limited use. The introduction of witchcraft opens the door for a wide
variety of causal developments and thus, at a certain point, can render
any causal sequence implausible. Once witchcraft or related nonnor-
mal causalities are introduced, their nonuse has to be explained. For
example, if witchcraft saves the hero in the end, one might reason-
ably ask why it was not used to defeat the enemy right at the outset.
For this reason, nonnormal causalities operate within an often rather
strict set of operational constraints. They form a set of “restricted”
causalities, as opposed to “general” or normal causality.
Perhaps the most theoretically interesting sort of causality is,

roughly, what Aristotle isolated with the concept of “design.” In terms
of objective properties, it is often based on thematic affinities (to bor-
row a phrase from Adolf Grünbaum) – similarities (or related asso-
ciations) that allow us to link two events distinctively. To return to
Aristotle’s example, there is a thematic affinity between Mitys’s mur-
der and Mitys’s statue falling on the murderer. The distinctive sim-
ilarity is straightforward. Someone killed Mitys. Now “Mitys” kills
that person. The same point would apply if the murderer simply died
by some random accident on the anniversary of Mitys’s murder. Here,
the similarity would be that Mitys died on this day, then his murderer
died on this day. In both cases, the similarity leads us to think of the
two events as forming a mutually relevant temporal sequence, which
we then understand as a specifically causal sequence.
Of course, this is hardly a full explanation. We see many events as
similar or thematically related without thereby concluding that they
form a causal sequence. In the cases we are considering, we infer a
causal sequence in part because the similarity and mutual relevance
are integrated into a larger final causal framework. Thematic asso-
ciation becomes causal precisely by being part of a “design,” which
is to say through its operation in the intent of a designer. Ultimately,
this form of causality is bound up with providence. It takes part in
the causality of a divine plan. Often, the intervention of divinity is
explicit in narrative. In romantic plots, such as Niz.āmı̄’s Laylā and
Majnūn or Kālidāsa’s Abhijñānaśākuntalam, a deity takes a direct and
active role in uniting the lovers. The importance of divine interven-
tion in classical European drama is suggested by the fact that our
term for a plot-resolving coincidence is “deus ex machina.” This phrase
refers to “the practice of some Greek playwrights . . . to end a drama
with a god who was lowered to the stage by a mechanical appara-
tus and, by his judgment and commands, solved the problems of the
human characters” (Abrams 42). The causality of Aristotelian design
is merely the same sort of providential intervention, now rendered
implicit – or, rather, the same sort of providential and punitive in-
tervention. Divine shaping of life is always conceived of as a mat-
ter of both reward and chastisement. Moreover, this causality of de-
sign is of a piece with the ethical causality that underlies sacrificial
narrative – the explanation of extreme conditions (such as drought)
not by ordinary causal principles but by principles of reward and,

more commonly, punishment. Of course, it is characteristic of the
other two prototype genres as well. For example, in heroic works,
God is continually intervening on the side of the in-group to defeat
the demonic enemy.
This causality of design is, then, an extension of final causality
to distinctively associated (usually distinctively similar) sequential
events. In that way its importance is unsurprising, for it is another
instance of the maximization of final causality. Indeed, the same is
true for restricted causalities such as witchcraft. All such nonnormal
causal principles operate to personalize sequences of cause and effect.
Goals
The final component of our preliminary definition is goals. In many

ways, the idea of goals in the agent/goal structure is straightforward.
A goal is anything an agent might strive to achieve. As we have al-
ready discussed, what an agent strives to achieve is happiness. This is
trivially true. It is part of what we mean by “happiness.” Even when
a masochist strives for the experience of pain, he/she is striving for
happiness. Thus, technically, an agent’s goals are always imagined to
be eliciting conditions for happiness, or means to the eliciting condi-
tions for happiness.
However, this already brings us to a complication. The very na-
ture of final causality is, in a sense, paradoxical. The pursuit of a goal
sustains the hero, not only in narrative, but in life. Achieving the goal
leaves one in a sort of quandary as to what one should do next. Put
differently, life is, to a great extent, decision and action aimed toward
the achievement of goals. Our daily lives are animated by a sense
of future possibilities. We move through the seeming trivialities of
ordinary existence in the hope of reaching some more encompassing
goal in the longer term. In a sense, the very lack of happiness is what
keeps us going. Sustaining emotions sustain us toward the achieve-
ment of some goal. Deprivation, combined with a sense of possibility,
is what gives us hope for the future. Once the goal is achieved, how-
ever, we risk losing a sense of direction, and possibility. Though it
seems counterintuitive, success itself can bring with it a form of de-
spair, for it leaves an agent without a goal, and thus with nothing to
motivate and sustain him/her. Moreover, ordinary human goals are,

as they say, not everything they are cracked up to be. We have already
seen that individual and group domination are highly dubious aspi-
rations, likely to bring sorrow and remorse even to those who achieve
them. Though romantic plots almost always treat romantic union as
an unequivocal good, the actual situation of married people – as well
as some of the Ainu poems – suggest that this is not true, even in
societies where marriages based on love and individual choice are
the norm.
Of course, one might respond that this all may bear on human
life, but it does not really have any consequences for stories. Despair
from success does contribute to the epilogue of suffering, in the way
already discussed, but its narrative importance is, it might seem, con-
fined to that one area. However, this is not true. The ramifications of
despair are much more general, and not only because we explicitly or
implicitly understand our actual lives in terms of narrative sequences.
In order to clarify this relation between plot structure and the
despair of success, we need to draw a few distinctions. First, we
need to distinguish between short-term aims and more encompass-
ing projects. Projects are the long-term goals that we see as fostering
happiness. In other words, projects are the eliciting conditions for
happiness – or, rather, for a particular type of happiness, which itself
may be called “continuing” or “condition” happiness. Specifically,
we may distinguish between happiness as an experience, which may
be momentary, and happiness as a condition, which is, by definition,
long lasting and continuous. The latter is what the Greeks called “eu-
daimonia,” the good life. The two are not unrelated. Sexual union is
one of our primary prototypes for experiential happiness, or enjoy-
ment. It is not, and cannot be, a prototype for condition happiness
or eudaimonia. However, it is related to, and part of, one prototype
for eudaimonia – romantic union. A similar point could be made
about various sorts of limited success (for example, in games, exams,
battles) and long-term social standing.
Of course, we may pursue enjoyment, just as we may pursue eu-
daimonia. Thus, at this point, we may distinguish three types of
goal: short-term aims that serve as means, intrinsically pleasurable
pursuits (such as sexual union) that yield enjoyment, and projects
that, when achieved, define condition happiness or eudaimonia.
Short-term aims may conduce toward pleasurable pursuits or toward

long-term projects. Note that projects are limited. Indeed, they are
largely limited to the prototype eliciting conditions for the outcome
emotion of happiness as discussed above. In other words, prototyp-
ical happiness, as treated in chapter three, is eudaimonia, not simple
enjoyment. The happiness of enjoyment may also be an outcome
emotion for a given narrative (fictional or lived), but it cannot be pro-
totypical for happiness due to its ephemeral quality. Moreover, inso-
far as it contributes to eudaimonia – that is, insofar as it is subsumed
under a project – it is junctural in the way that means are junctural.
In these terms, the problem of despair may be seen as taking two
forms. The obvious form is a loss of hope regarding the achievement
of a particular project (for example, the loss of hope that one will ever
marry one’s particular beloved). The second form is a loss of projects –
the sense that there is no project worth pursuing and thus nothing
that would define one’s short term aims and organize one’s tempo-
rary pleasures. Again, this can occur through the achievement of a
project. Specifically, once one accomplishes a project, one is likely to
be faced with several problems. First, there is the problem that one has
been deprived of the goal that animated one’s activities, gave mean-
ing to one’s aims. The sense and purpose of short-term aims seems
lost. Second, the joy we experience in a long term condition will fade.
This results from a general psychological principle, the “law of ha-
bituation,” as Frijda calls it (“Laws” 277). We become accustomed to
positive emotional experiences that are continuous or predictably re-
curring. In other words, the intensity of happiness wanes as the state
of romantic union or social domination becomes habitual. Finally, it is
simply a fact about human life that everyone’s future holds physical
decline – thus decrease in the success and joy of project-related enjoy-
ments (such as sex) – and, ultimately, death. Again, the problem with
accomplishing a project is that one must then ask, “What now?” In
addition, the answer is not terribly cheering – “Well, there’s sickness,
decrepitude, and death.”
The problem recurs with any achievable project one might under-
take. This is the familiar problem of ambition and greed. They are, in
principle, boundless in that each success proves unsatisfactory. As a
matter of logic, there is only one type of goal that would solve this
problem. It would have to encompass and render complete all normal
actions. It would have to absolutize the enjoyments of our pursuits

and the eudaimonia of our condition while, at the same time, render-
ing those enjoyments and that eudaimonia impervious to the effects
of habituation, as if they were always new. Finally, it would have
to prevent both decline and death. In other words, it would have to
be a form of outcome happiness that actually does conform to the
enduring and all-encompassing qualities of its prototype.
Unsurprisingly, most societies have imagined just such a goal into
existence. In Christianity, it is Heaven. In Islam, it is the Garden. In
Vedāntic Hinduism, it is moks.a, release from rebirth and reunification
with divinity. In each case, it is a spiritual state of absolute bliss that is
undecaying and eternal. One function of such an imagined goal – not
simply a project, but a culminating telos of human life – is to provide
something beyond projects, a principle that motivates and gives sense
and direction to all ordinary life. Of course, as an imagined goal, it is
vague, difficult to specify. In consequence, it is routinely understood
in terms of more concrete, prototypical projects – primarily romantic
union, as we have already stressed, but also including plenty and
even in-group domination (insofar as one’s in-group is, precisely, the
saved).
This notion of a transcendental goal or telos, modeled on pro-
totypical projects, but at the same time extending and intensifying
them, allows us to explain an otherwise peculiar feature of prototype
narratives. We have noted repeatedly that both romantic and heroic
tragi-comedies tend to culminate in a transcendental or spiritualized
version of their respective goals. Thus, Laylā and Majnūn are joined
in the divine Garden; Śakuntalā and Dus.yanta meet again in a heav-
enly hermitage; Rāma leaves his kingdom to return to his place with
the gods; the heroes of Gilgamesh and The Mwindo Epic return to their
kingdoms with spiritual wisdom and a sense of religious resignation.
This transcendentalizing of the conclusion seems to derive from the tacit
incorporation of the transcendental goal or telos into the realized
project. Specifically, in our understanding of real causal sequences,
we tend, explicitly or implicitly, to imagine a telos beyond even our
projects. This telos serves to structure and give sense to our projects
in the way that our projects serve to structure and give sense to our
short-term aims. In that way, it is always there, beyond the projects.
(This seems true, at least in some degree, even for those of us who
are atheists. In general, our minds operate with a wide range of be-
liefs, expectations, etc., that we would self-consciously repudiate.)
In consequence, the telos – itself imagined on a romantic, sacrificial,
and/or heroic model – tends to interfere or merge with the project in
a narrative, for the telos is what makes the project worth achieving.
Without the telos, the project risks becoming hollow. In short, it risks
despair. One result is the transcendental or spiritualized conclusion –
which may, in fact, recognize and respond to that despair directly, as
in the epilogue of suffering, some of the Ainu romantic poems, and
such sacrificial narratives as the story of Jesus.
The transcendentalizing of the project in narrative is related to
two other tendencies as well – the social generalization of the project
and the idealization of the primary object. Social generalization is found
prominently in heroic narratives. As we have seen, these stories
standardly aim toward group domination, especially national dom-
ination. There are, of course, dominant individuals in these stories.
But their accession to power regularly occurs in the context and ser-
vice of group domination. The triumph of the hero is a triumph of
the entire nation. The condition of the individual is generalized to the
group. The same sort of social generalization is obvious in sacrificial
stories as well, for the way to overcome individual hunger is through
overcoming collective famine. The sacrifice that restores food to the
individual does so insofar as it restores food to the group as a whole.
Romantic plots may seem to be an exception here. However, there
is an impulse toward social generalization in romantic plots also,
most obviously in marriage. Thus we sometimes find romantic plots
including not only the marriage of the hero and heroine, but the ro-
mantic union of various ancillary characters as well. These multiple
marriages extend personal happiness beyond the main couple to a
broader society. Cases from Shakespeare come to mind immediately,
but the same generalization occurs elsewhere. In the Sanskrit com-
edy, The Little Clay Cart, Cārudatta and Vasantasenā, the main lovers,
are able to wed, but so are Śarvilaka and Madanikā (Vasantasenā’s
servant), a couple nearly cheated of their chance for happiness due
to poverty and Madanikā’s servitude. When Rāma wins the hand
of Sı̄tā, they are not married alone; Rāma’s brothers receive spouses
as well and all are wed in a collective ceremony. The story of Bamsi
Beyrek, from the Turkish Book of Dede Korkut, involves a still more
obvious social generalization. When Beyrek marries, he and Prince

Kazan arrange brides for his thirty-nine warriors also, and the forty
couples share an extravagant, forty day wedding feast.
Finally, the prototype conditions for happiness standardly include
not only a situation, but an object or a person as well (a number of
writers have emphasized the importance of the “sought for” object or
person; see, for example, Greimas 202–3 and citations). In itself, the
point is obvious and seemingly inconsequential. The sacrificial plot
involves food; the heroic plot concerns the kingdom or nation; the
romantic plot aims toward the beloved – the prototypical physical,
social, and personal object, respectively. Everyone is aware that the
beloved is idealized in romantic plots, in the lyric poems that derive
from such plots, and so on. In fact, all three objects are regularly
idealized. Of course, this idealization is likely to be limited by other
tendencies, such as that toward empathy-producing humanization
in the case of the beloved. Nonetheless, it is clear that idealization of
the object is important to prototype narratives of all three sorts.
Thus we find that heroic tragi-comedies regularly establish the
kingdom as a utopia. The Rāmāyan.a informs us that “When Rāma
ruled, nowhere was the widow’s wail heard . . . and no one was over-
taken by calamities. It did not fall to the lot of old men to do the ob-
sequies of the young. Gladness was everywhere and every one was
keen on Dharma [ethical duty]” (Vālmı̄ki, vol. III, 371). When Son-Jara
reigned, “peace and happiness entered” society; “justice prevailed
everywhere” and “the villages knew prosperity again” (Kouyaté 81).
When the society is not established as a utopia, it is nonetheless often
elevated as a place of particular prosperity, virtue, and valor. Again,
it is regularly a society special to God. As to sacrificial narratives,
the result of sacrifice is often abundance, sometimes marked by a
sort of perfect food. Consider again the story of Jesus. The salvation
brought by his crucifixion is eternal bliss – and it is commemorated by
an unparalleled spiritual food, the Eucharist. (I take it that examples
of idealizing the beloved in romantic tragi-comedy are too extensive
and evident to require citation.)
There are several sources for this idealization. Most obviously, we
tend to idealize what we want and do not have, whether it is a person,
a political position, or what someone at the next table is eating in a
restaurant. Unsurprisingly, this broad human tendency turns up in
stories. Of course, in stories, the object really is idealized; it is not, most

often, presented as idealized only in the imagination of the agent.
This brings us to a second reason. Objective idealization fosters an
appropriate response on the part of the reader, for it tends to repeat,
and thus prime, the reader’s own idealizations – most obviously, in
the case of the romantic plot, his/her idealizations of some beloved
person at the time he/she was most in love.
Finally, as the preceding examples suggest, idealization is bound
up with spiritualization. The sacrificial story not only results in food,
but in heaven. The kingdom not only wins, it is selected by God.
The hero who rules does so with the dispensation of God and, after
the epilogue of suffering, with divine wisdom. The beloved is not
only beautiful and charming, but morally and spiritually elevated –
so much so that the final union of lover and beloved may take place
only in heaven. Thus, idealization of the object is, among other things,
a consequence of the spiritualization of projects via the telic trajectory
in which they are implicitly embedded.
a note on emotional intensification

Both idealization and spiritualization are in part a matter of strength-
ening characteristics that foster an emotional response. In other
words, both are in part a matter of intensification, which is to say,
the enhancement of some aspect of a work that produces emotional
impact. Intensification bears not only on agents (through idealiza-
tion) and goals (through telic spiritualization), but also on causal
sequences. We have already noted that the initial conflict in proto-
typical narratives is often familial – parents forbid the lovers’ union;
a brother usurps the hero’s throne. This familializing of the conflict is
clearly a way of intensifying that conflict, both ethically and emotion-
ally. In part, this is the result of “role expectations.” As Ortony, Clore,
and Collins point out, it is more shocking to discover that a nun has
murdered someone than to discover that a gangster has done so (see
79–81). The same point holds for near relatives.
More broadly, the minimum requirement for the agent/goal struc-
ture is that the agent lacks the goal, strives for it, and (in the full comic
version) achieves the goal. In prototypical narratives, however, lack
of the goal is intensified into the opposite of the goal. The prototype
conditions for happiness are preceded by the prototype conditions for

sorrow. The opposite of spousal union is the death of the beloved. The
opposite of social domination and esteem is ostracism and disgrace;
the opposite of group domination is defeat and conquest. The oppo-
site of being well fed is not simply waiting for dinner; it is famine.
Note that this also intensifies the goal itself, in part by increasing the
effort necessary to achieve the goal. As Ortony, Clore, and Collins
point out “increases in effort tend to increase the degree to which
goals are positively valued” (73).
In addition, while each prototype narrative draws primarily on
emotion prototypes of its own category (physical, personal, or so-
cial), the different prototypes for both sorrow and happiness may be
conjoined as well. This occurs most obviously when elements from
different prototypical and narrative structures are combined into a
single story – when the Rāmāyan.a intertwines a romantic plot with
a heroic plot or when Oedipus the King mixes the epilogue of suffer-
ing with a sacrificial narrative. It can also happen, however, that one
type of story draws on another for structural relations, imagery, or
plot elements, without the main concerns of the other narrative be-
ing literally incorporated as a plot line. Lovers in Sanskrit romantic
tragi-comedies waste away like famine victims. Exiled princes long
for their people like lovers separated from their beloveds, and so
on. Conversely, Son-Jara’s reign, once he is rightfully placed on the
throne, is marked by plentiful harvests: “Vast fields of millet, rice, cot-
ton, indigo and fonio surrounded the villages” (Kouyaté 81). Jesus’
resurrection is frequently represented as a military triumph over Sa-
tan. Combining prototypes for happiness and sorrow from different
prototype stories is another way of intensifying the event sequence
of the plot.
Indeed, a number of the patterns discussed previously (for exam-
ple, the personalization of causality) could be understood as forms of
intensification as well – indicating once again the close interrelation
between prototypical narrative and emotion.
place, time, and social order

Despite the apparent implications of the discussion thus far, character,
goal, and causal sequence are not the only constituents of prototyp-
ical plots, nor are they the only elements that are subjected to such
principles as intensification. Three other constituents appear particu-

larly significant. Labov notes that, in most stories, “At the outset, it is
necessary to identify in some way the time, place, persons, and their
activity or the situation.” “Persons” and “activity” or “situation” re-
fer to the elements discussed previously. Labov adds time and place
to the list. These may appear to be mere incidentals, necessary though
not intrinsically consequential aspects of narrative. However, this is
not true. An understanding of time and place is, in fact, crucial to our
understanding of the general principles of narrative.
First of all, both time and place in narrative should not to be under-
stood objectively, but subjectively. Like so much else in prototypical
narrative, they are personalized, moved away from efficient causal-
ity toward final causality. Time is not, most importantly, the time of
chronometers. It is lived time. On the one hand, it is the psycholog-
ical time of individual expectation, remembrance, enjoyment, fear,
and so on. On the other hand, it is the naturally organized time that
structures human action. The latter may be divided into cyclical and
noncyclical structures. Cyclical structures would include, most ob-
viously, days and seasons, both of which figure prominently in the
literal development of plots and, perhaps even more, in their asso-
ciated imagery. Noncyclical time would most obviously include the
time of aging.
Space, too, is lived space. Not mere objective location, but that lo-
cation as it relates to one’s experience of life – aspiration, feeling, ac-
tion, memory, and so on. The most crucial spatial division is between
what is “home” and what is not – with “home” understood contex-
tually to refer to a particular house, town, region, country, or what-
ever. The spatial structure of narratives is, then, most significantly
a structure of home and away. “Home” in romantic plots is a per-
manent place with the beloved. “Home” in heroic plots is the nation.
“Home” in sacrificial plots is a paradise of natural comfort and plenty.
It is this personalized spatial structure that makes exile (that is, expul-
sion and exclusion from home) a central element in all three sorts of
narrative.
Put differently, in stories, time and space are, first of all, emotional
time and space.
The intensifications of space and time in prototypical narratives
are probably clear from what has already been said about the telic
conclusions of these narratives. Time is intensified into eternity in
the spiritualized endings of all three prototype stories (e.g., when the
lovers live “happily ever after”) and home is intensified into heaven
or some comparable spiritual ideal. Conversely, those who are pun-
ished may be sent to an intensified exile – for example, in hell.
In addition to these two elements, named by Labov, it is important
to add social structure. Social structure provides the social context for
the action of the narrative in a way that directly parallels the spatial
and temporal structure. However, the social structure is even more
crucial, for the conflicts and developments in prototypical narratives
continually involve social hierarchies. The lovers cannot marry be-
cause of class differences or some internecine conflict in the society;
the rightful heir is dethroned and exiled by a usurper; a sin by some
members of the community condemns the entire society to hunger. In
all three prototypical genres, social structure is of vital consequence.
Here, the most important aspect of the social structure seems to be
its hierarchical organization. Again, in romantic plots, the lovers run
directly against this hierarchy, for that is what causes their separation.
Sacrificial plots are most often straightforward treatments of just how
a violation of hierarchy leads to massive suffering. In heroic works,
the dispossession of the hero is typically a case of some villainous
disruption of the hierarchy.
The intensification of social structure is found perhaps most ob-
viously in the common tendency to make the heroes of works into
nobles or other figures high in the social hierarchy. In some cases,
such as narratives of social rebellion, the intensification may push
in the opposite direction giving us heroes of little social prestige or
position. In either case, the characterization operates to enhance the
emotive impact of the work – depending of course on the individ-
ual reader (with his/her particular categorial identities, and personal
experiences) – because social hierarchy is crucial to our sense of cate-
gorial identity and to our personal experience of social happiness or
suffering. (The point obviously bears on the ideological functions of
literary works as well.)
the common form of prototypical narratives

The tendencies outlined thus far set out some of the broad principles
that apply across prototypical narratives. However, there are much
more specific properties shared by these narratives as well. Indeed,

the three story prototypes have a common structure that is remark-
ably detailed.
Specifically, each of the genres begins with some sort of social
conflict. This is standardly presented as a conflict between the hero
and a socially normative opponent, abetted by a rival, who may be
individualistic. (For a summary of this analysis, see Table 7.1.) In
romantic stories, this is commonly a conflict between the lovers and
their parents, who prefer a rival suitor. In the case of heroic plots,
we find a usurper, who is often a rival character and thus frequently
individualistic, but may be socially normative as well (as when a
king tries to exclude his son from kingship). In sacrificial plots, we
often find a character “like us” – not precisely a hero – violating some
divine (thus socially normative) prescription, perhaps at the urging
of an individualistic opponent (such as Satan in the Judeo-Christian
story).
This is followed by deprivation of the object (the beloved, the
kingdom, food) and exile, which is to say, social expulsion of the
hero or his/her imprisonment (thus, in either case, his/her exclusion
from home). In romantic stories, the lover is often sent away; the
beloved may be confined by the father or abducted by the rival. In
heroic plots, the hero is often exiled to another nation. The object –
which is to say, the kingdom – is taken away through usurpation. In
sacrificial plots, food is withheld in obvious ways – and the trans-
gressors may be exiled from a place of plenty, as occurs with Adam
and Eve.
This deprivation and exile are associated with death – either literal
death, a deathlike state, or simply death imagery. This point is obvious
enough in all three structures that it does not require elaboration.
The next common element (though it need not occur at just this
point) is a social extension, in this case a social extension of threat
and suffering. The point is obvious in the heroic structure with its
threat/defense sequence. In sacrificial narratives, the point is obvi-
ous as well, for the initial fault, though individually committed, af-
fects everyone. Romantic stories appear to be an exception, for by
their nature they focus on individualistic characters. However, to
a remarkable extent, romantic stories do involve this sort of social
generalization.
table 7.1. The Common Form of Prototypical Narrative Genres
General structure Social conflict, Deprivation of Death or death Extension of threat Dissension and Reacquisition of Spiritualization
especially with a object; exile or imagery/ and/or pain reconciliation of the object, return
socially imprisonment of deathlike state from individual heroes or hero from exile,
representative hero/heroine to society and helper; familial reunion,
character, and/or nature defeat and imagery of
abetted by some punishment of rebirth,
individualistic the antagonist sanctifying
antagonist ritual
Romantic Social conflict of Separation of Death of lovers Generalization of Division and Return of hero Spiritualization of
tragi-comedy lovers with lovers: exile of and/or death conflict from reconciliation of from exile, lovers (union
socially lover; imagery lovers’ social lovers; defeat of reunion of after death or in
232
representative confinement or difference (e.g., enemy (e.g., lovers, a heavenly place)
characters (e.g., abduction of class) or parents), parent/child
parents), often beloved generalization of exclusion or reunion,
abetted by a lovers’ pain and punishment of imagery of
rival lover mourning rival rebirth, marriage
Heroic Social conflict of Usurpation Death of hero Danger to society Division and Return of hero, Epilogue of
tragi-comedy hero with (= deprivation and/or death from invasion reconciliation of restoration of suffering, with
(socially of the kingdom imagery (or civil strife) heroes (e.g., king kingship or final
representative) or other rightful and warrior); rightful place, spiritualization
ruler and/or position), exile initial hero or familial reunion, (see below)
rival/usurper or imprisonment heir defeats and imagery of
of hero punishes the rebirth (perhaps
socially birth of an heir),
representative perhaps
enemy and/or coronation or
rival/usurper related ritual
Sacrificial Violation of divine Withholding of Death of people Danger to society Conflict between Return of food, Reconciliation of
tragi-comedy prescription food, exile from and/or death from starvation society and return to land of society with
(with the deity land of plenty imagery scapegoat; plenty, perhaps God/gods,
as a socially sacrifice resurrection and possibility of
representative punishes society reunion of heavenly life
character), often (pure version) or families,
at the urging of violator (secular institution of
some version) and ritual
individualistic defeats the
tempter rival/tempter,
who is excluded
or punished
Epilogue of Hero, as socially Hero is deprived of Death imagery Society loses ruler The hero, tacitly The kingdom is Achievement of
suffering representative, is his/her position or other placed in the returned, divine home
responsible for and exiled, dominant figure position of the imagery of through
the death of either due to enemy or rival, rebirth, perhaps apotheosis or
innocents his/her own is punished or a coronation or suicide or
233
remorse or due self-punished; other ritual somber return
to actions of separation from to human home
some deity helper with spiritual
wisdom
One obvious way in which this negative or sorrowful social ex-

tension can occur in romantic plots is through an elaboration of the
social prejudices that are preventing the marriage. Modern works
that treat interracial love are an obvious case of this, for they almost
invariably take up the problems faced by the lovers as symptomatic
of larger social problems of racism. The point is not confined to the
modern period. For example, Terence, himself a former slave, ad-
dresses the problem of slavery in several of his plays, sometimes
explicitly. In a New Comedy, we are often faced with two lovers who
are unable to wed because it seems that the woman is a slave. How-
ever, it turns out that the “slave” was actually free born, and thus the
marriage is permitted. In Eunūchus, Terence takes up this standard
sequence. However, in this case, the two lovers are actually a young
man and the woman he raped. The man excuses the rape by explain-
ing that he thought she was a slave. Moreover, he vows that “I did
it for love of her” (207) – though, after the act, the girl “does noth-
ing but cry and can’t bring herself to answer any questions” (194).
Of course, this is a “problem comedy.” We can hardly be happy that
these “lovers” are united. Nonetheless, the social generalization is
clear.
In other cases, broader social problems, conflicts, or conditions
may grow out of the initial separation of lovers, even if this is not
based on any systematic social division, such as race or slavery. In
Abhijñānaśākuntalam, the king forbids the communal festivities of
spring, and, as a result, even the processes of natural renewal are
stopped (Kālidāsa, Theater 148–9). Moreover, he partially neglects
his duties as king, assigning his minister to make judgments in his
place (150–1). In Laylā and Majnūn, Majnūn convinces one Bedouin
leader to wage war against Laylā’s family, thus destroying the peace
of Bedouin society. In each case, then, we see the pain of the lovers
extended to the larger society.
This leads us to the defeat and punishment of the antagonist, of-
ten aided by some helper. The lovers are reunited, thus overcoming
the “enemy.” The rival may be killed or banished in the process. At
the very least he/she is typically removed from the comic commu-
nity. In the heroic plot, the usurper is defeated, often killed in battle.
The point is equally clear in sacrificial tales. In secularized versions,
the violator is punished directly. In the pure or religious version, the
(guilty) community is punished through the representative scape-

goat. In connection with this, the tempter is typically defeated as well,
often through exclusion from the final comic society (like a romantic
rival). Satan, again, is an obvious example.
In some instances, the defeat of the rival is preceded by a split
between, then reconciliation of, the heroes or the hero and a prime
helper. Though I did not discuss the point earlier, it happens with
some frequency that the lovers are estranged at some point in the
story. Cases range from Gurgānı̄’s Vı̄s and Rāmı̄n to the Japanese
Love Letter from the Licensed Quarter to Abhijñānaśākuntalam. We some-
times find a comparable conflict in heroic plots, as when Rostám and
the Shah divide briefly before the battle with Sohráb. Finally, sac-
rificial narratives may involve some dispute between the sacrificial
victim and the social representatives who are there to enact the sac-
rifice. For example, in the Kondh sacrifice, one segment of the ritual,
involves “a very long and somewhat sentimental, but at the same
time argumentative, conversation between the priest and the Meriah
[sacrificial victim], the object of the one being to show that the victim
must calmly submit to suffering in a cause so greatly to the benefit
of mankind . . . and of the other, to prove that he had been cruelly
deceived” (E. T. Dalton, quoted in Mahapatra xix–xx). Jesus’ “agony
in the garden,” where he questions his role and asks not to have to
make the sacrifice, provides another instance.
The defeat of the rival and enemy leads to the reacquisition of the
object and the return home. The lovers are reunited and retrieved
from exile. The rightful ruler comes home to the throne. Food is re-
stored to the people, and they are allowed to enter once again the land
of plenty, at least potentially. It is also common for other sorts of fam-
ily reunion to occur in these three structures. In romantic plots, the
parents and children, formerly divided over their romance, may be
reconciled once the lovers are joined. Moreover, the lovers may have
a child from whom they had been separated. The lovers/parents may
be reunited with this child as well – as in Abhijñānaśākuntalam, or the
seminal Japanese story of Izaemon and Yūgiri. In the latter, the im-
poverished lovers can at first neither wed nor raise their own child,
but in the end they are united and their child is returned to them (for
a summary of this story, and a note on its historical significance, see
Halford and Halford 210–11).
Heroic plots may involve the reuniting of families separated dur-

ing war, especially the royal family, with the rightful heir restored.
An obvious case of this is the reunion of the various family members
in Ayodhyā, when Rāma returns to rule the kingdom after he has
defeated Rāvan.a and rescued Sı̄tā. In the Mwindo Epic, Mwindo’s tri-
umph is accompanied by a reunion with his aunt, who had worried
that he would die during his final journey, and with his father –
though there is the peculiar twist here that Mwindo’s father had
fled after trying to kill Mwindo, then had to be tracked down by
him, hardly our prototype case of family reunion. The ending of the
Odyssey fits here as well. Though not, as a whole, a strict heroic plot,
the part of the story that takes place in Ithaca is clearly a version
of the heroic plot, with the exiled ruler’s kingdom threatened by
usurpers. (There is a version of the romantic plot operating in this
part as well.) Chikamatsu’s Battles of Coxinga brings a father and son
together at the final point of heroic triumph. The Chinese Yüan play,
A Stratagem of Interlocking Rings, includes a reunion as well, though an
unusual one – a reunion of separated lovers, clearly borrowed from
the romantic plot. The enormously popular movie, Independence Day,
a straightforward heroic tragi-comedy, concludes with the reunion of
parents with children and also the reunion of lovers.
In sacrificial narratives, it sometimes happens that the dead are
resurrected or otherwise spiritually freed (e.g., for Heaven), in a sce-
nario that allows either a wordly or spiritual reunion for those who
have been separated. We see this in the general salvation produced
by Jesus’ sacrifice in the Christian story. There are also more mun-
dane reunions, such as the reunion of mother and children after the
sacrificial death of Nathan in Markandaya’s Nectar in a Sieve.
As the sacrificial case suggests, this is all typically accompanied
by imagery of rebirth, paralleling the imagery of death in the first
half of the structure. In all three prototype structures, this may be
marked by a sanctifying ceremony. The most obvious case of this is
the sacrificial narrative, for it frequently includes an account of the
origin of a standard ritual (for example, the Eucharist in the story of
Jesus). But the romantic story too culminates, explicitly or implicitly,
in a ceremony – marriage. And the heroic narrative often leads up to
a coronation or some similar ritual, as, for example, in The Mwindo
Epic.
Finally, prototypical narratives end in some sort of social general-

ization and spiritualization, as discussed previously. The only oddity
here is that the heroic narrative often goes through an entire further
plot sequence before this occurs – the epilogue of suffering.
But the epilogue of suffering is not as anomalous in this respect
as it may seem, for it follows the common structure as well, more
or less running through the sequence a second time. The epilogue
begins with a social conflict in which the hero is the socially nor-
mative character responsible for the death of some innocent. This
leads to deprivation of the object and exile – voluntary or forced
abandonment of home – as the hero gives up the kingship or other
rewards of victory and goes in search of knowledge (as in the case of
Gilgamesh) or atonement (as in the case of Kumagae) or is deprived
of the kingship and exiled in punishment (as in the case of Mwindo).
This involves considerable imagery of death, and is directly conse-
quential for the society (thus involving social extension). While there
is not really an issue of defeat here, it is clear that the entire sequence of
events serves to punish the hero as if he/she were the antagonist. This
sequence also frequently involves a graphic separation of the hero
from a prime helper (as with Gilgamesh and Enkidu or Yudhis.t.hira
and his brothers in the Mahābhārata). All this is often followed by
the return of the object (for example, the return of the kingship to
Gilgamesh or to Mwindo), imagery of rebirth (as in Zeami’s Atsumori,
where Kumagae and Atsumori “shall be re-born together/On one
lotus-seat” [73]), and/or ritual (as in the case of Mwindo [see Rureke
143], where the ritual is strikingly close to the rituals of sacrificial
narratives). As we have repeatedly noted, this final return, even if
it is a return in defeat or loss (as with Gilgamesh), implicitly in-
cludes a sense of spiritual elevation, if in a somewhat somber way.
In addition, some plots of this sort have a final moment when the
hero – perhaps without returning to the kingship – achieves a di-
vine home through apotheosis (as in the case of Oedipus) or suicide
(Rāma).
This outline of common structures makes one more thing salient
about the organization of prototypical narratives – they are symmet-
rical. The initial conflict is paired with the final defeat of the enemy
and punishment of the rival; the deprivation of the object and the
exile are paired with the restoration of the object and the return of
the hero; the social generalization of the danger is paired with the
concluding social generalization of happiness.
conclusion
In sum, though the generation and reception of paradigm narratives
is a matter of prototypes, the various prototype structures have many
things in common. They share a set of strong tendencies – toward
personalization and final causality, individualistic characterization,
social generalization of both the sorrow and the happiness, conclud-
ing spiritualization, idealization of the object, intensification, and so
on. They also share a remarkably detailed and complex structure, go-
ing well beyond the minimal pattern implied by agent, goal, and
causal sequence of events. These common features are bound up
with the emotional premises, components, and consequences of nar-
ratives. But only part of this is explained by emotion prototypes.
There is clearly a great deal of material here, not only for further
descriptive specification, but even more so for further psychological
explanation – in short, for a continuing research program in narrative,
emotion, and cognition.
Afterword
From the Emotional Nature of Narrative to the
Narrative Nature of Emotion
Up to now, we have treated emotion fairly narrowly, focusing in par-

ticular on emotion prototypes as generative of stories, though also
treating emotive response to literature. In these final pages, I should
like to consider emotion itself, outside literature. Specifically, I should
like to reverse the direction of study and examine the ways in which
emotion is a function of narrative – not emotion in literary response,
but emotion as we experience and understand it in ordinary life. First,
I shall discuss some of the ways in which the eliciting conditions, ac-
tional outcomes, and even phenomenological tone of emotions are
shaped and oriented by stories and lexical prototypes. In connection
with this, I shall touch on some political consequences and ideological
functions of prototypical narratives in their various social specifica-
tions. Second, I shall consider the biological givens of emotion, just
how these are reshaped socially, and how both the biological and
social components relate to narrative, including the prototypical nar-
ratives we have been considering.
convention, feeling, and social function

Our ordinary view of emotions is that they preexist prototypes and
stories. Our lexical items name them. Our stories recount and, when
successful, trigger them. In part, this is no doubt true. It is simply a
pattern of nature that people run from packs of wolves, scream, con-
tort their facial features in certain ways, and so on. In other words, in
239
240 Afterword
part our prototype for “fear” simply captures a natural phenomenon.

But there is more to it than that.
First, as discussed briefly in Chapter 3, we do not simply have
some direct intuitive way of identifying emotions. The point is obvi-
ous with other people. We do not have access to anyone’s experiential
subjectivity. We cannot feel anyone else’s emotions. Thus, to identify
their emotions, we must rely on clues from what we can experience.
Prototypes are extremely helpful in this regard. How can we tell if
someone is in love? Well, largely the same way we decide whether
something is a bird. In the case of a bird, we compare it with a pro-
totype. In the case of romantic love, we compare the person’s actions
and expressions with our prototype for actional/expressive outcomes
of love. Indeed, this is how we come to decide that sometimes people
are mistaken about being in love or not being in love. We compare
one part of their behavior (the verbal expression of love or lack of
love) with other parts, using our lexical prototype as a guide. What
is most important for our present concerns is that this sort of com-
parative judgment applies not only to other people, but to oneself.
Indeed, even where there is no question of my being mistaken, it
is nonetheless just these prototypes that allow me to judge whether
I have particular emotions. Since I do not have direct access to the
“qualia” of another person, the “raw feel” of his/her emotions, how
can I judge that the two of us are feeling the same thing? For exam-
ple, how can a child judge that he/she is feeling something that is
aptly referred to as “jealousy” or “suspicion”? In part, it is a matter
of parental attribution. The parent compares the child’s situation and
actions with prototypes and says, “Oh, Johnny is jealous.” In part it
is also a matter of the child forming such prototypes and compar-
ing his/her own situation and responses with them. In both cases,
the learning process is the result of inference from prototype elicit-
ing conditions and outcomes. (This general point was argued most
famously by Wittgenstein.)
This may seem to be a fairly simple and psychologically incon-
sequential process of acquiring vocabulary. But it is not. Emotional
inference is not just a matter of learning the right labels for expe-
rientially self-evident and fully pregiven emotions. Some studies
suggest that we must infer what we are feeling even after we have
learned emotion terms. Daryl Bem’s self-perception theory contends
Afterword 241
that “Individuals come to ‘know’ their own attitudes, emotions, and

other internal states partially by inferring them from observations
of their own overt behavior and/or the circumstances in which this
behavior occurs” (222), which is to say by inference from expressive/
actional outcomes and eliciting conditions. Stanley Schachter’s work
is particularly relevant here. Schachter has argued that an emotion
involves some state of physiological arousal plus some cognitive in-
terpretation of that state. Thus, given “a state of arousal . . . one la-
bels, interprets, and identifies this stirred-up state in terms of the
characteristics of the precipitating situation and one’s appercep-
tive mass. . . . It is the cognition which determines whether the state
of physiological arousal will be labeled ‘anger,’ ‘joy,’ or whatever”
(402–3). In keeping with this, Nisbett and Ross argue that we judge
our own inner states by much the same process of reasoning through
which we judge other people’s inner states: “knowledge of the self is
produced by the same strategies as knowledge of other social objects,
and is thus prone to essentially the same sorts of bias and error” (195).
With respect to emotions, they argue that “Once the individual be-
comes aware of his own state of physiological arousal, the labeling of
that state – and the subjective experiences, self-reports, and emotion-
ally relevant behavior that accompany such labeling – is the result of
a search for a plausible cause of the arousal” (199). Indeed, “People’s
labeling of their emotional states . . . depends on an analysis of evi-
dence conducted in the light of preconceived theories about which
antecedents produce which states and which states are the product
of which antecedents” (200).
I suspect that these conclusions are overstated. Indeed, there are
problems with much of Schachter’s research. In one famous study
(see section II of Schachter), he injected subjects with a stimulant.
He then placed the subjects, one by one, in a room with a confeder-
ate who behaved either angry or euphoric. Some test subjects were
told that they would feel a degree of physical arousal due to the
injection. These subjects did not judge themselves either angry or eu-
phoric after the session with the confederate. Other subjects were not
told about the effects of the injection. They tended to judge them-
selves euphoric after some time with the euphoric confederate and
angry after some time with the angry confederate. (There was also
a placebo group.) Prima facie, this may seem to suggest that people
242 Afterword
interpret their emotions and do not simply feel them. However, there
are other possible explanations. For example, it may be that members
of the “informed” group judged the feelings of the confederate to be
false, euphoria or anger resulting from the injection only. Thus, they
distanced themselves from the confederate’s emotions. In contrast,
members of the uninformed group were unguarded in their response
to the confederate and thus tended to experience a sort of emo-
tional contagion (a well-known and widely attested phenomenon).
By this interpretation of the data, the difference in the groups is that
the explanation actually changed the eliciting conditions and thus
changed what emotions the test subjects felt. The informed group
really did not feel anger or euphoria. The uninformed group did. It
was not a matter of their interpretation of the physiological arousal
produced by the injections, but of the test subjects’ actual emotive
state.
Nonetheless, if Schachter’s conclusions appear to be somewhat
exaggerated, they are not entirely mistaken. As Ross and Nisbett
explain, “While many contemporary theorists would challenge
Schachter and Singer’s ideas about the lack of physiological speci-
ficity in emotional experience, few now would deny that we can be
led to mislabel our feelings and to reach erroneous conclusions about
the source of such feelings” (80). In short, there is an interpretive
component to emotional experience, and that component is at least
in part a matter of relating a raw feel to eliciting conditions and
actional/expressive outcomes.
On the other hand, putting the matter this way may understate
the case. This formulation appears to return us to a basically pre-
Schachterian position that the emotions are fixed in themselves. It
only adds the qualification that they are not self-evident. But, in fact,
emotions do not appear to be entirely fixed, even if they are also not
entirely malleable. Again, Nisbett and Ross note that our inferential
reasoning about our own emotions is not simply a matter of labeling,
but of “the subjective experiences . . . and emotionally relevant be-
havior that accompany such labeling” (199). Extending Nisbett and
Ross, we might say that emotions are not only identified by reference
to emotion prototypes. Specific emotional experiences and outcomes
are also, to some degree, shaped by our ideas about emotion, partic-
ularly by our emotion prototypes and our prototype narratives. Most
Afterword 243
obviously, we construe our experiences in light of eliciting conditions

for emotions and thus in light of stories, in a process that reacts back
on those experiences. We do not merely see a particular event (for
example, a spouse’s behavior) in its own terms. We see it as approxi-
mating or not approximating relevant prototypes and our emotional
response is guided by this. Moreover, we place ourselves in standard-
ized narrative sequences when we act on emotions (for example, in
expressing romantic love).
This returns us to Keith Oatley’s point that emotions are embedded
in narratives and to our development of Oatley’s ideas in Chapter 3.
Once again, we have short-term aims and long-term projects. We
continually evaluate events in our lives in relation to these aims and
projects. Our emotions are, largely, responses to those evaluations.1
As discussed in the third chapter, these may be divided into junctural,
sustaining, and outcome emotions, which are narrative categories.
Moreover, our ultimate evaluative criteria, projects, are nothing
other than micro-narrative eliciting conditions for happiness–or,
rather, they are nothing other than the micronarratives we take to be
eliciting conditions for happiness.2 As this final point suggests, the
relation of emotion to narrative is perhaps most obvious with regard
to eliciting conditions for happiness and our ideas about these elicit-
ing conditions. We might, then, begin with eliciting conditions, and
turn subsequently to phenomenological tone and expressive/actional
outcomes.
1 I am leaving aside emotions that operate spontaneously, without significant cognitive

mediation. As, for example, LeDoux points out, some perceptions directly excite
fear – or do so without any higher cortical involvement, thus presumably without
the sort of narrative appraisal I am discussing (see Emotional 164). For an attempt
to synthesize appraisal with an account of emotions in terms of perceptual triggers,
see Chapter 7 of my Cognitive Science.
2 Indeed, emotions are narrative not only in their direct experience, but in retrospect
as well, and the retrospective narratives may be quite different from the ongoing or
direct, experiential narratives. Tomkins points out that someone may have a terrible
experience, leaving him/her with a series of bad memories. “But what of the more
permanent effect of these bad scenes on the quality of [a person’s] life?,” he asks.
This effect “will depend critically on the degree of magnification which follows” the
experience. Specifically, “rehearsals” of the bad memories “co-assemble them in
such an order and with such spacing that they are experienced as magnifying or
attenuating the negative affects” (215). In other words, the long-term emotional
effects of our experiences appear to be, at least in part, a function of the narratives
through which we organize and understand these experiences post facto.
244 Afterword
We do not really have very good ways to determine what will

bring happiness, especially insofar as happiness is understood as an
enduring condition. Indeed, we have only limited ways of determin-
ing what will bring us simple pleasure or pain. How do we know
even what kind of food we will like – not to mind what kind of food
will make us healthy in the long term? For the most part, we learn
these things by trial and error, or by accepting the judgments of oth-
ers. If we do not even have a good way of determining what foods we
might enjoy, how can we possibly have a good way of determining
what will make us happy for the rest of our lives, what will give us
eudaimonia? We don’t. Well, we have some clues. We have friends
and like spending time with them. This suggests to us that spending
time with people we like will be a good thing. But there are many
forms that this can take, many details to be developed. Even with
regard to new foods, we always know beforehand that eating is bet-
ter than starving. Moreover, we have certain general ideas about our
preferences. But this does not give us some algorithm for determin-
ing specifics.
Our goals, plans, and projects, then, are based only in part on our
own experiences of happiness. To a great extent, they are based on
what we learn from others about happiness. In forming our ideas
about happiness, we form prototypical eliciting conditions that op-
erate not only for literature, but for our lives. This is true both at the
universal level and at the culturally particular level. Most if not all
societies appear to establish romantic union as a – indeed, the pro-
totype eliciting condition for happiness. This is no doubt partially
based on a biological given, the sex drive. But it is also the result of
human society. After all, cross-culturally, romantic union is far more
than sexuality. It is a matter of living together, of sharing intimacy
and affection, and of having children. It is, in other words, a very
complex aspiration, and one that arises socially, even though it arises
universally. In all or almost all societies, this sort of relation comes to
be prized as an ideal, as the paradigm of personal happiness – even
in societies where people have little say regarding their own mar-
riage, where individual choice in marriage is officially condemned,
and where romantic love appears to be denigrated or ignored (un-
til one takes a look at the literature of the place). Indeed, there are
probably many properties and general structures or functions in this
Afterword 245
area that are cross-culturally constant or near constant. For example,

probably all societies envision some sense of home associated with
romantic union, some connection with social work, some organized
system of child-rearing, and so on. (There is a common view that
universals must be biologically innate. This is simply untrue. For ex-
ample, universals may arise from group dynamics. We shall return
to this issue.)
Of course, the precise nature of that home or work, the precise
relation to the children, will vary at least to some degree from culture
to culture. In fact, it is here that we can see where cultural specificity
enters. Accidents of geography, economy, political history, and so on,
have led to different instantiations of the romantic structure. Thus,
in traditional Igbo society, the man will farm yams, the woman will
grow other crops and perhaps market some goods, and the couple
will, ideally, have many sons; in America in the 1950s the man will
have a white collar job, the woman will stay at home with the children
and the dog, and the home will be a house in the suburbs, perhaps
with a picket fence. Indeed, cultural definition goes further still. We
even learn what sorts of men or women we should fall in love with – in
one standard cliché, men who are tall and have good careers, women
who are thin and cook well. This is not to say that we do actually fall
in love with people according to social criteria of this sort. However,
it is to say that we very often determine whether or not someone is a
candidate for romantic involvement on the basis of such criteria (that
is, on the basis of his/her similarity to the relevant paradigms). We
do not allow ourselves the possibility of romantic love with some
people; we open ourselves to the possibility with others. Indeed,
love is not love of a person anyway, but love of an idea of a per-
son. One’s idea of a person is based in part on the actual personality,
intellect, aspiration, feeling, of the man or woman in question. But it
is also built up out of various prototypes and different exempla. Thus,
we come to love or not to love people in part on the basis of social
ideas.
Though not accounted for via an adequate psychology, the point
has been recognized for quite some time. For example, this is precisely
the dilemma of Emma Bovary. Indeed, Emma Bovary’s situation is
particularly relevant to our discussion, for Emma comes to define her
goals, and to feel romantic love, on the basis of social principles that
246 Afterword
she has learned from reading romances, which is to say, from literary
narratives – and, indeed, prototypical literary narratives.
Of course, Emma’s case also suggests that the eliciting conditions
for happiness cannot be entirely shaped by social prototypes. There
is a difference between the real eliciting conditions for happiness
and the conditions defined by society. At least to some degree, there
are things that genuinely make us happy and other things that do
not. There is a joke about how you should never wish too hard for
anything, because you might just get it. Very often, we find that the
achievement of our goals is disappointing. Part of the reason for this is
simply formal. Without the goal to organize our lives, ordinary activ-
ities seem to wither into insignificance. However, part of the problem
is in the nature of the goals themselves. Often, socially determined
eliciting conditions for happiness simply do not bring happiness.
Moreover the social specification of prototype eliciting conditions
is not confined to happiness. One might imagine that some emotions
are simply not open to such alteration, that some are simply too direct,
too biologically fixed. But this does not seem to be the case. If there
were ever a prime candidate for an evolutionarily fixed emotion, that
would be fear. Fear is what leads us to flee dangerous predators, thus
surviving and having more offspring. But fear too is easily manipu-
lable. Take, for example, the United States in the 1950s, where many
ordinary people clearly came to fear the Soviet Union as if it were the
proverbial wolf at the door – or the United States today where a South
Asian woman can be reported by a frightened passenger, taken away
by the police, and questioned for five hours because she remarked
enthusiastically on the sights while looking out an airplane window
(see Joseph). One of Margaret Atwood’s characters put the general
point well: “what to feel was like what to wear, you watched the
others and memorized it” (131). Indeed, contrary to one’s initial ex-
pectations, the case of fear suggests that social definitions may have
their greatest impact on junctural emotions. Junctural emotions of-
ten rely on particular beliefs and expectations that may vary a great
deal across cultures or historical periods. In connection with this,
one might argue that social definitions have systematically differ-
ent consequences for different types of emotion, as follows: junctural
emotions are deeply affected by social specifications; sustaining emo-
tions are affected primarily in altering our openness to certain goals
Afterword 247
and objects (for example, our openness to certain people as possible

partners in romantic love); and outcome emotions are affected only
superficially.
These points bear not only on eliciting conditions, but on the rema-
ining aspects of emotion – phenomenological tone and expressive/
actional outcomes – as well.
The phenomenological tone of an emotion is “what it feels like” to
have the emotion. It is the part of an emotion that is entirely inacces-
sible to other people. It is usually seen as both certain and definitive.
However, as we have already noted, we do not simply identify our
emotions with a sort of introspective directness. We not only have
to determine what caused a particular “raw feel.” We often have to
determine just what that feel is. This indicates that phenomenological
tone too is not entirely fixed – or, at least, it may be vague, ambiguous.
In any case, our understanding of it is not as absolute as we usually
imagine. In consequence, the role of that raw feel in our ideas and
actions is open to considerable variability. Part of this variability re-
sults from the way our society organizes ideas about emotions. For
example, in the United States today, we tend to think of romantic
love as an emotion not only different from, but opposed to, the feel-
ing of friendship – hence the proverbially shattering utterance from
one’s beloved, “I like you . . . as a friend.” This has all sorts of conse-
quences. Most obviously, we may tend to segregate off people who
fit our “profile” for friends, separating them from people whom we
might consider possible romantic objects.
In keeping with this, Forgas and Dobosz argue that “individuals
from a given culture have fairly clear implicit cognitive representa-
tions of the range of relationship prototypes available in their mi-
lieu, and are likely to define their emerging contacts with others in
terms of the relationship repertoire at their disposal” (290). Thus, a
developing relationship between John and Jane will be understood
by both parties at least partially by reference to possible relationships
in their society – friendship, marriage, cohabitation, a “fling,” and
so on. The relevant prototypical possibilities will help guide their
feelings, their understanding of those feelings, and their behavior.
For example, these prototypes serve to establish expectations, and
thus, to define parameters for interpretation and evaluation of other-
wise neutral events or nonevents (for example, if a goodnight kiss is
248 Afterword
expected, then its absence is reason for disappointment and perhaps

for a reevaluation of the relationship). As Forgas and Dobosz put
it, “relationship scenarios . . . have a crucial influence on how part-
ners come to perceive and define their relationship and its expected
progress” (291).
Our cognitive organization – including our tendency to oppose
feelings of love and friendship – has other effects as well. Most im-
portantly, it may lead to differences in our understanding not only
of other people and of our possible relations with them, but of our
own emotions themselves. A particular raw feel may not clearly fit
friendship or romantic love. We may interpret it as one or the other
and act accordingly based on various socially prototypical concerns –
the nature of the object (does he/she fit the romantic profile), broader
aspects of the situation, and so on. Once we have understood our am-
biguous feeling in a certain way, and especially after we have begun
to act on that interpretation, our experience of the feeling itself will
alter (though not necessarily in any simple, predictable way).
Indeed, feelings are in all likelihood not only vague, but complex.
We have a range of different feelings at any give time. Many are not
clearly attached to one another or to any particular object. For social
and other reasons, we cluster some feelings together, separating them
from putatively different clusters of feelings (for example, separat-
ing sexual desire from friendship), and linking them with particular
objects. This interpretive clustering and object orientation almost in-
variably reacts back on the feelings themselves – intensifying, diffus-
ing, or otherwise altering them. Suppose I am tired and frustrated.
Initially, both feelings may be tinged with sadness. Perhaps the tired-
ness has a physiological cause (for example, I am coming down with
a cold) and the frustration relates to work. But I cluster the two to-
gether and attribute both to the behavior and habits of my spouse.
This changes my experience almost immediately. I begin to feel the
tiredness and frustration as a single emotion, now tinged with anger
and resentment. In contrast, if I think there is a physiological source
for the fatigue, I separate the tiredness from other raw feels, experi-
encing it as tinged with apprehension. The frustration, now ignored,
may fade to insignificance. Clearly, I will behave very differently in
each case as well (for example, if I blame my spouse, I am unlikely to
respond by gargling).
Afterword 249
This brings us to our final component of emotive structure –

expressive/actional outcomes. As with phenomenological tone and
eliciting conditions, these are in part simply given biologically. Fear
leads to flight or paralysis; mirth tends to manifest itself in laughter.
But, also like raw feels and eliciting conditions, outcomes are open to
a wide range of social specifications. Consider, again, romantic love.
All that it clearly involves biologically is a desire for sexual union.
Social conditions appear to give rise to a range of other desires and
behaviors, which are themselves universal as well. These include a
generalized desire for continuing proximity or living together. There
is probably a biological/evolutionary kernel to this also, as I shall dis-
cuss in the next section. But the desire to live together, share a home,
and so on, is almost certainly a case of a (universal) social specifica-
tion. This aiming toward a shared life leads to behaviors that conduce
toward living together – life choices that lead to continued proxim-
ity, and so on. As part of this desire for a shared life, there is com-
monly a desire for reciprocity and exclusiveness. These are bound
up with outcomes such as a tendency toward jealousy (though jeal-
ousy may have other sources as well). Another part of the desire for a
shared life is a mutual identification of interests. The goals and gen-
eral well-being of the beloved become, to some extent, part of one’s
own goals. This sharing of goals leads to a sharing of means toward
the achievement of those goals – for example, a sharing of material
resources.
Needless to say, things do not always work out this way in prac-
tice. For example, patriarchy may be understood as, in part, a sys-
tematic skewing of these social desiderata. This leads us to the level
of cultural particularity (as opposed to universal social specification).
Clearly, the desire to share a life is particularized differently in differ-
ent cultures. The same point holds for other actional outcomes. For
example, in a capitalist economy, one standard way of expressing
love is through purchasing commodities, such that more love seems
broadly to imply more purchases, as well as more expensive pur-
chases, and vice versa (as Cohen and Rogers have in part suggested
[70]). This may seem to be some wholly new idea, unique to late cap-
italism. In fact it is merely a specification of the socially universal
sharing of means. Nonetheless, it is a culturally specific version of
that universal, with distinct, culturally specific, consequences.
250 Afterword
In all these cases, we find that social specifications – universal or

culturally relative – are bound up with plots. They are part of implicit
or explicit prototype narratives – in this case, a narrative of romantic
love leading to enduring romantic union at “home,” all duly speci-
fied. Moreover, one of the ways in which these social specifications are
communicated and internalized is through explicit stories, literature.
Of course, another way is through observation, seeing one’s parents,
and so on. However, while observation of one’s parents is certainly
important and influential, it need not indicate that a particular life
defines the eliciting conditions of happiness – for example, their mar-
riage may not be ideal; it may not instantiate romantic love. Indeed,
given the vacillations, uncertainties, contradictions, and general im-
perfectness of ordinary life, one might say that fictional narratives,
imagined stories, are necessary to organize and specify our emotions.
In Chapter 3, I argued that stories are connected with, indeed,
formed out of emotions and emotion ideas. Here we see that emotions
and emotion ideas are connected with, indeed, formed out of stories
as well. More exactly, we saw previously that prototype stories are
in part formed out of emotion prototypes. Those emotion prototypes
are in part based in actual emotions. There are facts about emotions.
However, at the same time, emotions are not fixed and fully definite.
Rather, emotions – specific feelings, as experienced and enacted at a
particular time and place – appear to be formed in part from emotion
prototypes and from the narratives these prototypes define. More-
over, to a great extent we internalize these prototypes through the
stories they have shaped.
This brings us to a final topic in this first section – the relation of
emotion to politics and ideology, a topic clearly raised by the preced-
ing discussion, though one we can only touch upon in the present
context. Much of the preceding discussion may seem to suggest that
cultural specification is nothing more than an innocuous process of
filling in some missing details in otherwise universal patterns. To
a great extent, this is true. The problem, however, comes with the
word “innocuous.” Our references to commodification and patri-
archy indicate that the process of social specification may be far from
innocent and inconsequential. In fact, emotions and standard plot
structures are particularized in socially functional ways. Specifically,
they tend to be defined and organized to preserve social structure.
Afterword 251
The specification of romantic love in a market economy is a specifica-

tion that commodifies as many conditions and outcomes as possible.
It turns the raw feel of love and the associated universal structures
to the purposes of the market. Patriarchal societies develop dissym-
metries throughout romantic plots, dissymmetries that operate to
extend, preserve, and justify gender hierarchies.3
Of course, this is true generally, not just with respect to emo-
tion and narrative. It is the nature of ideology to coopt whatever
it can to its own purposes. Moreover, as we have repeatedly noted,
there are elements in literary narrative that work against ideological
conformity – the rebellion of the lovers against the parental genera-
tion with its traditional strictures of class, caste, or race, the epilogue
of suffering with its humanization of and empathy for the enemy.
Nonetheless, the prototype narratives, despite their dissident ele-
ments, tend to be particularized in ways that preserve social hier-
archies. The very complexes we have discussed, the very structure of
romantic love that develops with romantic tragi-comedy, the socially
universal principles of a shared life, and so on, are broadly socially
functional. They are part of a cross-cultural structure of marriage,
which is a central organizing principle of human society. Indeed,
one could argue that romantic and heroic tragi-comedy are proto-
typical precisely because they treat the most crucial aspects of social
structure – reproduction and group power, the family and the nation
(tribe, village, or whatever). Narratives of sacrifice, too, treat a cru-
cial functional aspect of society – agricultural production. Of course,
their functionality is not due to the agricultural efficacy of sacrifice.
It seems, rather, to be bound up with the social hierarchies of scape-
goating and state-authorized killing.
In this way, not only cultural particulars, but cultural universals
are imbued with ideology from the start. The division of emotions, the
definition of emotion prototypes (including eliciting conditions and
expressive/actional outcomes, even to a certain degree the raw feel
of the emotion itself), and finally the prototype narratives we have
been discussing, are the product, not only of cognitive structure (and
human biology), but of socially functional categories and relations.
3 They also develop dissymmetries in the conceptualization and normative character-

ization of emotions. For a discussion of this, see Lutz.
252 Afterword
This ideology is not simple. It is continually troubled by contradictory

ethical and empathic impulses. But it is nonetheless functional and
pervasive.
biological bases of emotion

The socially based account of emotion prototypes sketched in the
preceding section is clearly a departure from the cognitive focus of
earlier chapters. It may even seem contradictory with that earlier fo-
cus. However, it is not contradictory. It may seem so due to the com-
mon assumption that a cultural universal is, necessarily, biological –
something “hard wired” into our brains. Moreover, in the standard
view, this biological endowment is the result of evolution – and we
can trace the evolutionary path largely by imagining ourselves in the
place of our hunter/gatherer ancestors and guessing what sorts of
things we would do in different hunter/gatherer situations. I sup-
pose the tone of the last comment indicates my attitude toward this
approach. Evolutionary accounts should, in fact, be far more com-
plex, as such writers as Stephen Jay Gould and Noam Chomsky have
argued (for a lucid treatment of the topic, see Jenkins). Despite my
skepticism about current approaches, however, it is clear that there
is some biological component to emotion, and some evolutionary
background to this biological component.
Often, writers on the biology of emotion begin with the idea of ba-
sic emotions. In Chapter 3, I expressed some skepticism about basic
emotions. Specifically, I do not believe that our emotions are struc-
tured in so discrete a manner as to allow strict division into emotion
categories, followed by some sort of hierarchization in which some of
these well-defined emotions form the basis of other emotions. It is, I
believe, a mistake to understand emotion on the model of, say, atomic
physics or molecular chemistry. Unlike atoms, emotions do not begin
with well-defined boundaries and fully specified constituents.
On the other hand, it is clear that there are clusters of feeling,
behavior, and idea, that are in effect biological givens of emotion –
“proto-emotions,” as we might call them. A subset of these acquire
prototype status because they are biologically salient – through, for
example, spontaneous facial expressions – presumably for functional
(thus evolutionary) reasons.
Afterword 253
However, having granted that, I want to stress that what I take to be

biological is extremely limited. There is very little that is deterministic
in my view of the physiological givens of emotion. Evolution leaves
us with broad, partially inchoate, complexes of feeling and action.
It does not determine what we do with those complexes. Indeed,
the “endowment” itself is partially contradictory, at times adjuring
incompatible behaviors and feelings. Moreover, societies can develop
in such a way as to socialize emotion in ways that seem directly
contrary to the biological givens. Of course, there are limits to this.
However, it is my contention that the endowments are much more
flexible than they might seem. Or, more properly, the givens of our
emotional life – givens that we can assume to be correlated with some
sort of evolutionary/genetic inheritance – are not rigidly constraining,
and bear on the majority of emotional universals only in limited ways.
In this section, I should like to consider what I take to be the likely
candidates for our evolutionary endowment in terms of emotion, or
proto-emotion. But, at the same time, I wish to stress the limits of that
endowment.4
In order to begin an analysis of proto-emotions, we need to draw
a few simple distinctions regarding emotions. Perhaps the most cru-
cial division concerns the nature of their object. Specifically, we need
to separate dispositional or pseudo-dispositional entities, on the one
hand, and nondispositional things, on the other. A dispositional entity
is anything to which one attributes intents. People are obviously dis-
positional entities, as are animals. But storms and sicknesses can also
be understood dispositionally. Indeed, they regularly are understood
in this way, at least in many societies, as their widespread deification
attests. Put differently, the division into dispositional entities and
nondispositional things is a division between objects that act on us
(dispositional entities) and objects on which we act (nondispositional
4 My account is in some ways related to the “componential approach.” As Panksepp

explains, in this view, “emotions are . . . constructed during early social develop-
ment from more elemental units of visceral-autonomic experiences that accompany
certain behavior patterns.” Thus, “componentialists suggest that biologically given
subunits are compiled into full-blown emotional systems via cognitive appraisals
and learning” (45). Of course, in my view, familial and social interactions and prac-
tices result in emotional systems that are themselves to a great extent universal.
Moreover, I do not see the “components” as fully specified and constant atoms of
emotion.
254 Afterword
things). Leaving aside happiness and sadness, the proto-emotions, at

least those that are most salient, may be divided productively accord-
ing to whether their object is understood as dispositional or not. It is
worth noting that this division is salient from early childhood. For
example, Steven Pinker points out that “Infants divide the world into
the animate and the inert early in life” (322; Pinker divides emotions
in a parallel fashion, more roughly distinguishing between “emotions
about things” and “emotions about people” [374]).
The most important division among dispositional objects is that
between malevolent and benevolent agents. We determine malev-
olence and benevolence relative to our own pursuits, goals, and
projects. Malevolent agents inhibit our achievement of goals; benevo-
lent agents further our achievement of goals. Among malevolent acts,
we may distinguish those that are threatening from those that are in-
hibitory. Threat is the eliciting condition for fear (or proto-fear) and
flight is the primary outcome. Inhibition – “thwarting and frustra-
tions,” as Panksepp puts it (52) – is the eliciting condition for anger (or
proto-anger) and attack is the primary outcome. Panksepp notes that
anger is a response to frustration that projects an inhibiting agent –
an opponent or enemy – even when there is no such agent behind
the frustration. As Panksepp puts it, “frustration emerges from the
ability of . . . cognitive systems to monitor the probability of forth-
coming rewards. If an expected reward is not registered, the higher
cell assemblies send out opponent process messages” (197). Benevolent
agents negate such malevolent actions. Thus, they protect us against
threat and/or aid us in fulfilling goals. There is a range of emotive
responses to benevolent behaviors. When limited, they form the elic-
iting conditions for gratitude (or proto-gratitude). When extended,
they form the eliciting conditions for affection (more properly, proto-
affection). The outcome of affection is following and joining with the
benevolent agent – in other words, the reverse of flight and/or attack.
(The limited/extended variable may affect other proto-emotions as
well. For example, extended inhibition may provide eliciting condi-
tions for hate.)
Anger and fear are two of the emotions standardly seen as “basic.”
This account adds affection (sometimes called “attachment”). On the
other hand, all lists include sorrow or sadness, which is separate from
this list. This is because, by the preceding account, both sadness and
Afterword 255
happiness are specified by way of these other emotions. Consider,

for example, the prototype case of sadness – the loss of a person for
whom one has affection. Here, sorrow is clearly reliant on another
emotion, specifically affection. (Note that “reliant on” does not mean
“composed of.”) This is not to say that sadness (or proto-sadness) is
not part of our biological endowment. Happiness and sadness are
certainly part of that endowment. But they operate only by reference
to these other emotions. Indeed, affection itself is often – perhaps al-
ways – reliant on fear and anger. This is broadly in keeping with our
analysis of these emotions in relation to narrative. Anger and fear are
purely junctural; love, as a sustaining emotion, is more general and
encompassing; and sadness or sorrow is one of the outcome emotions,
thus still more inclusive, for it must take in the entire preceding narra-
tive. Indeed, what I have just said about these proto-emotions is itself
very storylike. Although some proto-emotions are biological givens,
their relations are quickly defined by a social narrative, for the sub-
ject here is a protagonist seeking a goal and his/her emotions bear
on helpers and opponents. Once again, the origins and development
of emotion and narrative appear inseparably intertwined.
Turning to nondispositional things, we may draw a distinction
parallel to that regarding agents. There are things that appeal to us
and things that repel us. The prototype of repulsion is decay. The
decayed object provides eliciting conditions for disgust, leading to
the outcome of avoidance. There is a detailed parallel with fear. The
feared object is an active threat, whereas the repulsive object is a
passive threat. Flight is, in effect, an intense version of avoidance,
appropriate to the difference in object.
Objects to which we are drawn are most obviously of two sorts –
edible and sexed objects. Edible objects to which we are drawn pro-
vide partial eliciting conditions for the desire to consume and lead
to the outcome of eating. Sexed objects to which we are drawn –
attractive objects – provide partial eliciting conditions for lust and
lead to the outcome of pursuing sexual relations.
In the last two cases, I have spoken of partial eliciting conditions
because lust and hunger arise spontaneously from an internal bodily
cycle. Thus the desire for sexual relations and the desire to eat are not
solely dependent on their object. Indeed, this is one reason why lust
and hunger are often not considered to be emotions. It seems clear,
256 Afterword
however, that they belong in the same proto-emotion category as fear,

and so on. After all, the relation of lust and hunger to their objects is
parallel to that of disgust and the entire set shows obvious parallels
with the more uncontroversial emotions of the dispositional category.
Moreover, it is clear that lust and hunger – or, rather, the desire to
eat – may be inspired by a particular object. Panksepp explains that
“animals take large meals if their food is especially tasty but become
finicky nibblers if it is not” (169). On the other hand, one hardly
needs to do esoteric laboratory research to discover people saying,
“I’m stuffed. But I can’t resist that ice cream.” Moreover, the inclusion
of lust and hunger among emotions makes neurobiological sense. As
Joseph LeDoux explains, “the brain has a number of emotion systems,
including networks involved in identifying sexual partners and food
sources, as well as detecting and defending against danger” (Synaptic
321; see also 221, 230, and Brothers 47).
In part, the difficulty here is just terminological. First, the common
term “emotion” (with the details of folk psychology that it presup-
poses) does not fully fit the scientifically relevant concept. Second,
“hunger” does not really fit the phenomenon. Again, it is more prop-
erly called “desire to eat.”5 It is also worth mentioning that other
5 Anna Wierzbicka is probably the writer who has spent the most time cautioning re-
searchers against linguistic biases in work on emotion (see Emotions). Cases such as
this show the value of her cautions. On the other hand, Wierzbicka’s arguments go
well beyond urging researchers not to over generalize from linguistic peculiarities.
Indeed, there is a way in which Wierzbicka herself over generalizes from such pecu-
liarities. For example, her criticism of Ekman (in Chapter 4 and elsewhere) involves
looking at the oddities of English emotion terms and using these as an argument that
Ekman’s assertions about universal facial expressions cannot be correct. In treating
anger, she contends that the peculiarities of the English word “anger” are not present
in all languages. Thus there is no cross-cultural concept of anger. Thus Ekman is
wrong to assert that a particular facial expression is a universal expression of anger.
But why should one assume that Ekman’s claims involve all the cultural unique-
ness of a particular lexical item? If followed consistently, Wierzbicka’s arguments
would disallow such statements as “Eating food is a human universal.” “Food” in
English includes McDonald’s hamburgers. Most languages (for example, Sanskrit)
have not had a term that includes McDonald’s hamburgers. Thus, there really is no
equivalent for “food” in other languages. Therefore, eating food cannot be a hu-
man universal. Indeed, Wierzbicka’s reasoning would even disallow statements of
physics, such as “Objects fall to the earth’s surface at a constant rate of acceleration.”
After all, how many languages include precisely the semantic range of contempo-
rary English“fall”?
It seems that a more fair approach would urge caution on the part of researchers,
as noted previously, but at the same time would urge readers to maintain their
Afterword 257
emotions are not so fully object dependent as we might think. Even

such emotions as anger and fear are in part dependent on our own
prior state when encountering the frustration or threat.
However, this is not to say that the differentiae of lust and hunger
are unimportant. After all, we become hungry when we have not
eaten in a while. But we do not become fearful because we have not
run away in a while. In fact, these differentiating characteristics are
quite important. Lust and hunger provide us with the only emotions
on our list that define independent goals. Hunger defines the goal of
eating. Lust defines the goal of sexual union. Moreover, in both cases,
this goal is itself understood as defining happiness. These are not the
only outcomes that define happiness, of course. However, they are
the two that define happiness independently, without reference to
prior emotions or contingent circumstances.
Before going on to this last point, however, we should briefly
review the foregoing account of proto-emotions. The following list
summarizes what has been said thus far. (The form of the following
strings is eliciting conditions → emotion/phenomenological tone →
expressive/actional outcome.)
ordinary interpretive generosity. In any given case, Wierzbicka may be correct that
a researcher has been misled by linguistic idiosyncracies. And, of course, she can
always urge greater precision and explicitness in a theorist’s definitions. But one
cannot merely assume ethnocentric error. Showing that a concept has peculiarities
is not the same as showing that those peculiarities have guided a researcher’s rea-
soning and led to false conclusions.
On the other hand, insofar as Wierzbicka does isolate genuine problems, I hope to
have avoided these by linking emotion concepts with relatively concrete prototypes,
by distinguishing between emotions and proto-emotions, and by considering the
place of culture in the specification of emotion concepts and even in the definition,
development, and experience of emotions themselves.
In connection with the last point, I should perhaps mention that I am far from
being alone in discussing the effects of culture on emotion. The general idea has
been central to the work of social constructionists for many years – indeed, that is
what makes them social constructionists (see, for example, the essays in Gergen and
Davis or Harré and Parrott). That said, I should note that my treatment of cultural ef-
fects differs significantly from that of most social constructionists. First, I emphasize
specific biological givens in emotion. Second, I argue that the social component of
emotions develops along largely universal lines and that cultural differences, even in
the “socially constructed” part of emotion, are relatively superficial. Finally, I have
maintained elsewhere that the idea of social construction, far from being a theoreti-
cal advance on prior theories of socialization and ideology, is in fact a theoretically
faulty notion that confuses some very different phenomena (see my “Ideological
Critique”).
258 Afterword
Dispositional Objects:
threatening object → fear → flight (or freezing)
inhibiting object → anger, then hate → attack
protecting/aiding object → gratitude, then affection → joining (opposite
of flight and attack)
Nondispositional Objects:
decaying object → disgust → avoidance
edible/tasty object (and hunger cycle) → desire to eat → consumption
sexed/attractive object (and sexual cycle) → lust → courting or related
sexual pursuit
Again, sexual relations and the consumption of food are the two inde-
pendent goals yielded by this analysis. This suggests origins for two
of the prototype conditions for happiness and their associated nar-
ratives.6 However, while food is as such the prototype for physical
happiness (with the slight change that, as “plenty,” it is made ex-
tensive and enduring), the prototype for personal happiness is more
complex. To understand this, we need to consider the relation of the
other proto-emotions to happiness. We would not ordinarily say that
successful flight from a threat (due to fear) or successful avoidance
of decay (due to disgust) leads to happiness. Success in these ac-
tional outcomes is purely negative – unlike success in the outcomes
of eating or courting. Joining with a protector/helper, however, is
positive, a sort of proto-eliciting condition for happiness. Moreover,
it is the most enduring of the resultant conditions. In the personal
prototype for happiness, and thus in romantic stories, we find a
synthesis of the independent goal of sexual union with this endur-
ing, though partially derivative, goal of joining with a protector/
helper.
6 One colleague, on reading an earlier version of this argument, expressed concern

that I treat emotion in relation to narrative, but children lack narratives even though
they have emotions. I hope the preceding discussion of proto-emotions indicates
why this objection does not have force. Specifically, it conflates an implicit narrative
organization of experience with the act of verbal story telling. Children certainly
experience the world in terms of agents and nonagents (for example, parents and
food), as well as benevolence and malevolence (for example, feeding and spanking);
children certainly have goals (for example, to get food), encounter and recognize
impediments to goals (for example, an empty bottle), and so on. That is all the
narrative structure required at this point.
Afterword 259
This leaves only anger/hate. Successful attack against some malev-

olent agent is, it seems, not merely a matter of relief, but of more pos-
itive feelings as well. This is clearly one ancestor of the prototype for
happiness as social domination and for the associated prototype
narrative of heroic tragi-comedy.
Thus, given this analysis, we might have predicted four prototype
eliciting conditions for happiness and four corresponding narratives,
for lust, hunger, affection, and anger all give rise to positive feelings
when their actional outcomes are successful. This is more or less what
we find, with only the slight difference that the sexual and affection-
ate goals have been joined. This joining is one of those universals that
derives, in part, from the working out of human social relations, in
this case due to the social function of sex and affection in marriage
systems. The general point is suggested by Panksepp, when he writes
that “our cultural evolution has sought to bind our desire for sex and
our need for social bonding together in an inextricable whole . . . of
marriage” (226). Once this conjunction comes about, one would ex-
pect the resultant prototype (romantic love) and the corresponding
genre (romantic tragi-comedy) to be exceptionally prominent, for it
combines two sets of prototype eliciting conditions for happiness.
This is, again, precisely what we find.7
The structuring of anger/hate into domination is complex as
well. It is inseparable from the – highly socially functional, and also
universal – development of categorial identity and in-group/out-
group relations, which are distinct from the biological givens of
emotion per se. Here, as with romantic love – and, indeed, with the
sacrificial element in narratives of hunger – societies develop the
proto-emotions, both conceptually and practically, by extending and
specifying them in ways that contribute to the preservation of the
society. Indeed, we arguably have a sort of social evolution here.
Societies that develop in-group/out-group definitions, specifying
7 Of course, the eliciting conditions for affection and lust do define narratives sepa-
rately as well. In keeping with the fact that lust reaches its full development much
later than the other proto-emotions, this narrative division is age related. I realized
this when Keith Oatley pointed out to me that many Disney films (for example,
Bambi) are based on nonromantic “attachment” themes. Such films take up the goals
defined by affection without combining them with lust. Complementary to this,
“adult” narratives, in the euphemistic sense, focus on the goals of lust, with relative
indifference to affection/attachment.
260 Afterword
proto-emotions in relation to these definitions, may be more likely

to survive as coherent cultures, especially when faced with external
threat. (They are probably more likely to pose a threat to other soci-
eties as well.) In this way, the universality of the complex emotion
prototypes, and the resultant complex, narrative universals, may be
seen in part as outcomes of selection pressures on groups rather than
individuals. (In this way, a tendency to value, for example, in-group
domination could be universal without being based in some unalter-
able, biological propensity of all – or even any – individual people.)
All this helps us to understand why literature takes particular
forms. I hope that it also helps us to understand the human mind
more broadly. But at this point one might ask what sort of emotive
interest might lead to the development of literature initially. Why
would we want to read literature or listen to a story in the first place?
The reasons are certainly emotive. But they do not obviously derive
from any emotions discussed thus far. Here, we need to return to our
set of proto-emotions. The division of emotions based on type of ob-
ject is clearly incomplete, for it leaves out emotions that do not bear
on objects, but on ambient conditions. Indeed, these are perhaps the
most fundamental in our experience – for the entire sequence is hier-
archical. Every dispositional object is also a nondispositional object.
Every nondispositional object is also experienced as an alteration in
ambient conditions. Our response to ambient conditions is, then, in
some sense the most definitive.
The first proto-emotion in this category has as its eliciting con-
ditions an excess of stimulation; its actional response is withdrawal
from stimulation. We might refer to this as “sensitivity.” This ob-
viously parallels the withdrawal from nondispositional objects in
disgust and the withdrawal from dispositional objects in fear. Tem-
perature regulation appears to be our prototype case for this. The
fundamental place of ambient condition emotions is suggested by the
fact that thermal regulation is probably the most common source do-
main for metaphors about more standard or prototypical emotions –
anger and lust are heat; affection is warmth; fear is cold.
The second proto-emotion of this sort is the reverse – the pursuit
of stimulation, parallel to the pursuit of food or aggressive approach.
This is roughly what Panksepp sees as a sort of generalized urge
to act. He calls it “seeking,” but since there is no particular object
Afterword 261
being sought, this is a somewhat misleading term. In fact, the proto-

emotion he isolates is more or less what we ordinarily refer to as
“boredom,” or “boredom” plus “curiosity.” The eliciting conditions
for boredom/curiosity are a lack of stimulation – an obvious parallel
to the lack of food, and a less obvious parallel to the success depriva-
tion that triggers anger. The actional outcome of boredom/curiosity
is simply seeking stimulation.
In the proto-emotion of boredom/curiosity, we have one source of
interest in literature. Literature provides stimulation – and precisely
the sort of stimulation we lack in ordinary life. Stories about indi-
vidual and in-group domination, stories about love – in other words,
prototype stories – provide us with precisely the sort of stimulation
we might seek, but they do so in an innocuous form. In real battles,
after all, one can really be killed. The stimulation of an imaginary
battle is less than the stimulation of a real battle. But the risks are
reduced as well. Note that this division also suggests why we might
find some types of movie or novel unpleasant and why we might
turn away – put the book down, leave the theater. Relative to one’s
individual sensibility, the stimulation provided by, say, a particular
horror film might be excessive; for someone else, this may not be the
case at all.
On the other hand, the pursuit of stimulation is not all there is to
aesthetic experience. Moreover, sensitivity and boredom/curiosity
should not be the only proto-emotions bearing on ambient condi-
tions. I shall end this section with one last feeling that I take to be part
of our emotional endowment. From a very early age, children delight
in patterned sounds and images. For example, as to the former, from
about seven months, babies begin to engage in spontaneous rhythmic
babbling (Locke 283). This undoubtedly has a function in aiding lan-
guage acquisition, as Locke and others have argued. However, from
the baby’s point of view, that is irrelevant, at least initially. Infants
engage in rhythmic babbling at least in part out of simple enjoyment.
In short, they are not entirely hard-headed pragmatists setting out
to practice their language skills. They are also playful children hav-
ing fun. Propensities to enjoy verbal and visual symmetries, rhythms,
and so on, develop as we grow, yielding an emotive relation to objects
that involves attraction, but is not a matter of affection or lust. This
is a sort of proto-aesthetic sense. Its eliciting conditions are what we
262 Afterword
call “beauty.” Its associated feeling is a proto-form or ancestor of what

we awkwardly refer to as aesthetical pleasure, but is perhaps more
aptly characterized (following the Sanskrit theorists) as wonder or
wonder/delight. Its associated actional outcome is contemplation –
just looking or just listening. When successful, its result is not pre-
cisely happiness, but rather what Abhinavagupta said is the ultimate
goal of all art – peace. The point is no doubt related to Gordon Orians’
idea that we find just those landscapes beautiful where we feel safe,
for the feeling of safety is a necessary part of peace, or proto-peace
(on Orians’ work, see Pinker 376–7 and citations).
Wonder completes the triad of ambient condition proto-emotions.
But, of course, it is not confined to ambient conditions. Just as we may
have a nondispositional emotion directed toward a dispositional ob-
ject (for example, disgust with the body of another person), we may
have wonder directed toward an object, nondispositional or dispo-
sitional (for example, wondrous delight at the beloved). Again, the
proto-emotions are hierarchized such that the more “fundamental”
proto-emotions apply at all the higher levels. Moreover, like lust,
wonder is bound up with affection in complex ways. Even when the
object of our contemplation is not dispositional per se, even when
we are contemplating nature or some artifact, we regularly trans-
fer feelings of affection to some associated dispositional entity – the
person behind the beautiful face, the creator of the beautiful thing
(for discussion, see Ong “Jinnee,” and my “Beautiful”; the point has
implications for attitudes ranging from fascination with the lives of
authors to religious sentiment, as Ong has indicated). The point cer-
tainly bears on the personalizing tendency of arts discussed in the
preceding chapter, as well as literary empathy and related aspects of
literary development and experience.
The addition of wonder here suggests that our enjoyment of art,
including verbal art, derives straightforwardly from the biological
givens with which we start our social and personal lives. Indeed, our
delight in art results from the goals of two proto-emotions, goals that
are no less independent and positive than eating and sex – the goals
defined by boredom/curiosity and wonder. (Perhaps a similar point
could be made about science, which is not unrelated to curiosity and
wonder as well.) Of course, as in the other cases, the proto-emotion
of wonder is vague, inchoate, open to alteration, redevelopment, and
Afterword 263
specification through society and individual biography. But it is part

of what we all begin with – and no less universal than other emotions,
not only in its biology, but in much of its social development as well.
I imagine that most people who would accept this idea would
nonetheless say that the propensity to wonder is not functional; it
is not a development that directly serves our adaptive needs, but is
rather a by-product of some other, adaptive development (for exam-
ple, delight in rhythmic sounds is merely a by-product of commit-
ment to language learning). Literary critics sometimes take offense
when evolutionary theorists claim that literary properties did not de-
velop from the pragmatic survival of the fittest. Personally, I rather
like the idea that literature is different from big business and the mili-
tary, that we are magnificently untainted by the survival of the fittest,
with its unpleasant suggestions of large male brutes pummeling their
competitors into a bloody mush, then taking all the food and all the
women – perhaps with an occasional large female brute doing more
or less the same. On the other hand, I do wonder if the ability to con-
template beauty is not adaptive. Once our mental lives are filled with
fear, anger, lust, disgust, and so on, the contemplative peace brought
by a sense of beauty seems as if it would not be wholly lacking in
practical value, even for our ancient ancestors.8
conclusion
In sum, it appears that our emotions are much more malleable than we
are inclined to believe. They are formed from biologically given proto-
emotions – proto-fear, proto-anger, proto-affection, and so on – that
bear on ambient conditions, nondispositional objects, or dispositional
objects. These proto-emotions are specified by socially functional
practices and ideas that affect eliciting conditions, phenomenological
tone, and actional/expressive outcomes. These specifications are, in
part, culturally distinctive. But universals are not confined to biology.
Social specifications are to a great extent universal as well. This should
not come as a surprise, for the intimate interactions of individuals and
8 Tooby and Cosmides present a very different, adaptational account of aesthetic en-
joyment in babbling and elsewhere (16–17). For a discussion of some problems with
this account, see Chapter 8 of my Cognitive Science.
264 Afterword
the social relations of groups are patterned in common ways. They

are not random and wildly diverse, but quite consistent.
Noam Chomsky has repeatedly argued that differences among
human languages, though highly salient, are structurally minuscule.
As he put the point in one recent essay, a “Martian scientist” who
has come to study Earth “might reasonably conclude that there is a
single human language, with differences only at the margins” (New
Horizons 7). In my view, the same point holds for culture in general,
including broad areas of culture that are not simply biologically deter-
mined but result from social life as well. Among these are the almost
inseparable phenomena of narrative and emotion – prominently in-
cluding romantic, heroic, and sacrificial tragi-comedy, along with the
happiness prototypes they specify and the junctrual and sustaining
emotions they elaborate. These derive from universal social devel-
opments of biologically given proto-emotions – social developments
and proto-emotions that also account for our engagement with verbal
art and its unique, powerful, and universal satisfactions.
Works Cited
Abhinavagupta. The Aesthetic Experience According to Abhinavagupta. 2nd ed.

Ed. and trans. Raniero Gnoli. Varanasi, India: Chokhamba Sanskrit Series,
1968.
. “Locana.” In The Dhvanyaloka of Ānandavardhana with the Locana
of Abhinavagupta. Trans. Daniel Ingalls, Jeffrey Masson, and M.V.
Patwardhan. Cambridge, MA: Harvard University Press, 1990.
Abrams, M. H. A Glossary of Literary Terms. 5th ed. Fort Worth, TX: Holt,
Rinehart and Winston, 1988.
Abu-Lughod, Lila. “Shifting Politics in Bedouin Love Poetry.” In Language
and the Politics of Emotion. Ed. Catherine A. Lutz and Lila Abu-Lughod.
Cambridge: Cambridge University Press and Paris: Editions de la Maison
des Sciences de l’Homme, 1990.
Ahmad, Aijaz. In Theory: Classes, Nations, Literatures. New York: Verso,
1992.
Amaladass, Anand. Philosophical Implications of Dhvani: Experience of Sym-
bol Language in Indian Aesthetics. Vienna: De Nobili Research Library,
1984.
Amin, Samir. Eurocentrism. Trans. Russell Moore. New York: Monthly Review
Press, 1989.
Amnesty International. “Deterrence.” http://www.aiusa.org/abolish/
deterrence.html.
Ānandavardhana. “Dhvanyaloka.” In The Dhvanyaloka of Anandavardhana
with the Locana of Abhinavagupta. Trans. Daniel Ingalls, Jeffrey Masson, and
M. V. Patwardhan. Cambridge, MA: Harvard University Press, 1990.
Anscombe, G. E. M. “On the Grammar of ‘Enjoy.’” In Metaphysics and
the Philosophy of Mind: Collected Philosophical Papers. Vol. 2. Minneapolis:
University of Minnesota Press, 1981.
Appiah, Kwame Anthony. In My Father’s House: Africa in the Philosophy of
Culture. New York: Oxford University Press, 1992.
265
266 Works Cited
Aristotle. “Nicomachean Ethics.” Trans. W. D. Ross. In Introduction to Aristotle.

Ed. Richard McKeon. New York: Modern Library, 1947.
. “Peri Poietikes.” In Aristotle’s Theory of Poetry and Fine Art. 4th ed.
Ed. and trans. S. H. Butcher. New York: Dover, 1951.
Ashcroft, Bill, Gareth Griffiths, and Helen Tiffin. The Empire Writes Back:
Theory and Practice in Post-Colonial Literatures. New York: Routledge, 1989.
Aston, W. G. A History of Japanese Literature. London: William Heinemann,
1899.
Atwood, Margaret. Surfacing. New York: Fawcett Crest, 1972.
Awolalu, J. Omosade. Yoruba Beliefs and Sacrificial Rites. Harlow, United
Kingdom: Longman, 1979.
Badawi, M. M. Early Arabic Drama. Cambridge: Cambridge University Press,
1988.
Bal, Mieke. Narratology: Introduction to the Theory of Narrative. 2nd ed. Toronto:
University of Toronto Press, 1997.
Barnard, Philip J. “Interacting Cognitive Subsystems: Modeling Working
Memory Phenomena Within a Multiprocessor Architecture.” In Models of
Working Memory: Mechanisms of Active Maintenance and Executive Control.
Ed. Akira Miyake and Priti Shah. Cambridge: Cambridge University Press,
1999.
Barsalou, Lawrence W. “Ad Hoc Categories.” Memory and Cognition 11.3
(1983).
Bateson, Mary Catherine. 1970. Structural Continuity in Poetry: A Linguistic
Study in Five PreIslamic Arabic Odes. The Hague: Mouton, 1970.
Bem, Daryl J. “Self-Perception Theory.” In Cognitive Theories in Social Psychol-
ogy. Ed. Leonard Berkowitz. New York: Academic Press, 1978.
Bernheimer, Charles, et al. “A Statement of Purpose: Comparative Literature
at the Turn of the Century.” Report circulated in October 1993 to members
of the American Comparative Literature Association.
Berggren, Paula. Teaching with the Norton Anthology of World Masterpieces,
Expanded Edition: A Guide for Instructors. New York: W. W. Norton, 1995.
Berndt, Ronald M. Love Songs of Arnhem Land. Chicago: University of Chicago
Press, 1976.
Bharatamuni. The Nāt.ya Śāstra. Delhi, India: Sri Satguru Publications, n.d.
Bhāsa. “The Broken Thighs.” In Thirteen Plays of Bhāsa. Vol. II. Ed. and trans.
A. C. Woolner and Lakshman Sarup. Delhi: Motilal Banarsidass, 1985.
. “The Minister’s Vows.” In Thirteen Plays of Bhāsa. Vol. I. Ed. and trans.
A. C. Woolner and Lakshman Sarup. Delhi: Motilal Banarsidass, 1985.
. “The Vision of Vāsavadattā.” In Thirteen Plays of Bhāsa. Vol. I. Ed. and
trans. A. C. Woolner and Lakshman Sarup. Delhi: Motilal Banarsidass,
1985.
Bhavabhūti. The Uttararāmacharita of Bhavabhūti. Ed. and trans. M. R. Kale.
Delhi: Motilal Banarsidass, 1988.
Biebuyck, Daniel. Notes to The Mwindo Epic. In The Mwindo Epic. Shé-
Kárisi Rureke. Ed. and trans. Daniel Biebuyck and Kahombo C. Mateene.
Berkeley: University of California Press, 1971.
Works Cited 267
Bierhorst, John, ed. and trans. Four Masterworks of American Indian Litera-
ture: Quetzalcoatl/The Ritual of Condolence/Cuceb/The Night Chant. New York:
Farrar, Straus and Giroux, 1974.
Blair, R. J. R. “A Cognitive Developmental Approach To Morality: Investi-
gating the Psychopath.” Cognition 57 (1995).
The Book of Dede Korkut: A Turkish Epic. Trans. and ed. Faruk Sumer,
Ahmet E. Uysal, and Warren S. Walker. Austin: University of Texas Press,
1972.
Bordwell, David. “Convention, Construction, and Cinematic Vision.” In
Post-Theory: Reconstructing Film Studies. Ed. David Bordwell and Noel
Carroll. Madison: University of Wisconsin Press, 1996.
. Narration in the Fiction Film. Madison: University of Wisconsin Press,
1985.
Boulton, Maureen and Barry McCann. The Song in the Story: Lyric Inser-
tions in French Narrative Fiction, 1200–1400. Philadelphia: University of
Pennsylvania Press, 1993.
Bower, G. H. “Affect and Cognition.” Philosophical Transactions of the Royal
Society of London, Series B 302 (1983).
. “Mood and Memory.” American Psychologist 36 (1981).
Bownas, Geoffrey. “Introduction.” In The Penguin Book of Japanese Verse.
Ed. Geoffery Bownas and Anthony Thwaite. London: Penguin, 1967.
and Thwaite, Anthony, eds. The Penguin Book of Japanese Verse. London:
Penguin, 1964.
Brandon, James R. The Cambridge Guide to Asian Theatre. Cambridge:
Cambridge University Press, 1993.
. “Introduction.” In Kabuki: Five Classic Plays. Ed. James R. Brandon.
Honolulu: University of Hawaii Press, 1992.
, ed. and trans. Kabuki: Five Classic Plays. Honolulu: University of
Hawaii Press, 1992.
, ed. Traditional Asian Plays. New York: Hill and Wang, 1972.
Brenan, Gerald. St. John of the Cross: His Life and Poetry. Trans. of his poetry
by Lynda Nicholson. Cambridge: Cambridge University Press, 1973.
Brewer, William F. “Literary Theory, Rhetoric, and Stylistics: Implications
for Psychology.” In Theoretical Issues in Reading Comprehension: Perspectives
from Cognitive Psychology, Linguistics, Artificial Intelligence, and Education.
Ed. Rand J. Spiro, Bertram C. Bruce, and William F. Brewer. Hillsdale, NJ:
Lawrence Erlbaum, 1980.
. “The Story Schema: Universal and Culture-Specific Properties.” In
Literacy, Language, and Learning: The Nature and Consequences of Reading
and Writing. Ed. David R. Olson, Nancy Torrance, and Angela Hildyard.
Cambridge: Cambridge University Press, 1985.
and Edward H. Lichtenstein. “Event Schemas, Story Schemas, and
Story Grammars.” In Attention and Performance IX. Ed. John Long and Alan
Baddeley. Hillsdale, NJ: Lawrence Erlbaum, 1981.
Brewitt-Taylor, C. H., trans. Romance of the Three Kingdoms (San Guo Yan Yi).
By Luo Guan Zhong. 2 vols. Singapore: Graham Brash, 1991.
268 Works Cited
Brothers, Leslie. Friday’s Footprint: How Society Shapes the Human Mind.
Oxford: Oxford University Press, 1997.
Brower, Robert H. and Earl Miner. Japanese Court Poetry. Stanford, CA:
Stanford University Press, 1961.
Brown, Donald. Human Universals. Philadelphia, PA: Temple University
Press, 1991.
Cameron, James, dir. Titanic. Script by James Cameron. Photography by
Russell Carpenter. Edited by Conrad Buff, James Cameron, and Richard
A. Harris. Paramount Pictures, 1997.
Cao Xueqin. The Story of the Stone. Vol. 1. Trans. David Hawkes. New York:
Penguin, 1973.
Cao Xueqin and Gao E. The Story of the Stone. Vol. 4. Trans. John Minford.
New York: Penguin, 1982.
. The Story of the Stone. Vol. 5. Trans. John Minford. New York: Penguin,
1986.
Carroll, Joseph. “Universals in Literary Study.” In Universalien und Konstruk-
tivismus. Ed. Peter Hejl. Frankfurt am Main: Suhrkamp, forthcoming.
Carroll, Noël. “Art, Narrative, and Emotion.” In Emotion and the Arts. Mette
Hjort and Sue Laver, eds. New York: Oxford University Press, 1997.
Chadwick, Nora K. and Victor Zhirmunsky. Oral Epics of Central Asia.
Charitōn. Chaereas and Callirrhoe. Trans. B. P. Reardon. In Collected Ancient
Greek Novels. Berkeley: University of California Press, 1989.
Chaucer, Geoffrey. The Canterbury Tales. Ed. John Halverson. Indianapolis,
IN: Bobbs-Merrill, 1971.
Chelkowski, Peter, ed. Ta’ziyeh: Ritual and Drama in Iran. New York: New York
University Press, 1979.
Chêng Teh-hui. “The Soul of Ch’ien-Nü Leaves Her Body.” In Six Yüan Plays.
Ed. and trans. Liu Jung-en. New York: Penguin, 1972.
Chikamatsu Monzaemon. “The Courier for Hell.” Trans. Donald Keene. In
Masterpieces of the Orient. Ed. G. L. Anderson. Expanded Edition, New York:
W. W. Norton, 1977.
. Major Plays of Chikamatsu. Ed. and trans. Donald Keene. New York:
Columbia University Press, 1990.
Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press,
1965.
. Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger,
1986.
. Language and Mind. New York: Harcourt, Brace, Jovanovich, 1972.
. New Horizons in the Study of Language and Mind. Cambridge:
Chung, Ling. “Biography of Li Ch’ing-chao.” In Complete Poems. Li Ch’ing-
Chao. Trans. and ed. Kenneth Rexroth and Ling Chung. New York: New
Directions, 1979.
Clinton, Jerome. “The Book of Dede Korkut.” In The Norton Anthology of World
Masterpieces, Volume 1: Beginnings to 1650. Expanded ed. Ed. Maynard Mack
et al. New York: W. W. Norton, 1995.
Works Cited 269
Cohen, Joshua and Joel Rogers. On Democracy: Toward a Transformation of

American Society. New York: Penguin, 1983.
Comrie, Bernard. Language Universals and Linguistic Typology: Syntax and
Morphology. 2nd ed. Chicago: University of Chicago Press, 1989.
Conway, M. A. and D. A. Bekerian. “Situational Knowledge and Emotions.”
Cognition and Emotion 1.2 (1987).
Cooper, Arthur, ed. and trans. Li Po and Tu Fu. London: Penguin, 1973.
Crane, Mary. Shakespeare’s Brain: Reading with Cognitive Theory. Princeton, NJ:
Princeton University Press, 2000.
Croft, William. Typology and Universals. Cambridge: Cambridge University
Press, 1990.
Cronyn, George W., ed. American Indian Poetry: An Anthology of Songs and
Chants. New York: Fawcett Columbine, 1991.
Cullmann, Oscar. The State in the New Testament. London: SCM P, 1957.
Danly, Robert Lyons. “The Kokinshū.” In The Norton Anthology of World
Masterpieces, Volume 1: Beginnings to 1650. Expanded ed. Ed. Maynard Mack
et al. New York: W. W. Norton, 1995.
Davidson, Donald. Inquiries into Truth and Interpretation. Oxford: Clarendon
Press, 1985.
Davis, Mark H. Empathy: A Social Psychological Approach. Madison, WI: Brown
and Benchmark, 1994.
Dawson, Michael R. W. Understanding Cognitive Science. Oxford: Blackwell,
1998.
Dawood, N. J., trans. Tales from the Thousand and One Nights. New York:
Penguin, 1961.
Deng, Francis Mading. The Dinka and their Songs. Oxford: Clarendon Press,
1973.
de Sousa, Ronald. “The Rationality of Emotions.” In Explaining Emotions.
Ed. Amelie Oksenberg Rorty. Berkeley: University of California Press, 1980.
Dhanam . jaya. The Daśarūpa: A Treatise on Hindu Dramaturgy. Trans. George
C. O. Haas. New York: AMS P, 1965.
Dimock, Edward C., Jr. “The Rāmāyan.a Story.” In The Literatures of India:
An Introduction. Dimock, Edwin Gerow, C. M. Naim, A. K. Ramanujan,
Gordan Roadarmel, and J. A. B. van Buitenen. Chicago, IL: University of
Chicago Press, 1978.
, Edwin Gerow, C. M. Naim, A. K. Ramanujan, Gordon Roadarmel,
and J. A. B. van Buitenen. The Literatures of India: An Introduction. Chicago,
IL: University of Chicago Press, 1978.
Dissanayake, Ellen. Homo Aesheticus: Where Art Comes From and Why. New
York: Free Press, 1992.
Dixon, R. M. W. and Duwell, Martin, eds. The Honey-Ant Men’s Love Song and
Other Aboriginal Song Poems. Melbourne: University of Queensland Press,
1990.
Dolby, William, ed. and trans. Eight Chinese Plays From the Thirteenth Century
to the Present. New York: Columbia University Press, 1978.
Doniger, Wendy. The Implied Spider: Politics and Theology in Myth. New York:
270 Works Cited
Duckitt, J. H. The Social Psychology of Prejudice. New York: Praeger, 1992.

Dundas, Paul. The Jains. New York: Routledge, 1992.
Edmonson, Munro S. Lore: An Introduction to the Science of Folklore and Litera-
ture. New York: Holt, Rinehart and Winston, 1971.
Egudu, Romanus and Donatus Nwoga. Igbo Traditional Verse. London:
Heinemann, 1973.
Ekman, Paul. “Cross-Cultural Studies in Facial Expression.” In Darwin and
Facial Expression: A Century of Research in Review. Ed. Ekman. New York:
Academic Press, 1973.
Euripides. The Trojan Women. In Euripides: Ten Plays. Trans. and ed. Moses
Hadas and John McLean. New York: Bantam Books, 1960.
Fa-Digi Sisòkò. The Epic of Son-Jara. In The Epic of Son-Jara: A West African Tradi-
tion. Ed. and trans. John William Johnson. Bloomington: Indiana University
Press, 1992.
Fanon, Frantz. Black Skin, White Masks. Trans. Constance Farrington. New
York: Grove Press, 1963.
Ferdowsi, Abolqasem. The Epic of the Kings: Shāh-Nāma. An abridged ver-
sion. Trans. and ed. Reuben Levy. Costa Mesa, CA: Mazda Publishers in
association with Bibliotheca Persica, 1996.
. The Tragedy of Sohráb and Rostám: From the Persian National Epic,
the Shâhnâme of Abol-Qasem Ferdowsi. Trans. Jerome W. Clinton. Seattle:
University of Washington Press, 1987.
Firth, Raymond with Mervyn McLean. Tikopia Songs: Poetic and Musical Art of
a Polynesian People of the Solomon Islands. Cambridge: Cambridge University
Press, 1990.
Fodor, Jerry A. Concepts: Where Cognitive Science Went Wrong. Oxford:
Clarendon Press, 1998.
Forgas, Joseph Paul. “Affect in Social Judgments and Decisions: A Mul-
tiprocess Model.” In Advances in Experimental Social Psychology. Vol. 25.
Ed. Mark P. Zanna. San Diego, CA: Academic Press, 1972.
. “The Effects of Prototypicality and Cultural Salience on Perceptions
of People.” Journal of Research in Personality 17 (1983).
. “Mood and the Perception of Unusual People: Affective Asymmetry
in Memory and Social Judgments.” European Journal of Social Psychology 22
(1992).
and Barbara Dobosz. “Dimensions of Romantic Involvement: To-
wards a Taxonomy of Heterosexual Relationships.” Social Psychology Quar-
terly 43 (1980).
and Stephanie J. Moylan. “Affective Influences on Stereotype Judge-
ments.” Cognition and Emotion 5 (1991).
Fowler, Barbara Hughes, ed. and trans. Archaic Greek Poetry: An Anthology.
Madison: University of Wisconsin Press, 1992.
Frijda, Nico. H. “Comment on Oatley and Johnson-Laird’s ‘Towards a Cog-
nitive Theory of Emotions.’” Cognition and Emotion 1.1 (1987).
. The Emotions. Cambridge: Cambridge University Press and Paris:
Editions de la Maison des Sciences de l’Homme, 1986.
Works Cited 271
. “The Laws of Emotion.” In Human Emotions: A Reader. Ed. Jennifer

M. Jenkins, Keith Oatley, and Nancy L Stein. Oxford: Blackwell, 1998.
Frye, Northrop. Anatomy of Criticism: Four Essays. Princeton, NJ: Princeton
, Sheridan Baker, and George Perkins. The Harper Handbook to Litera-
ture. New York: Harper and Row, 1985.
Gadamer, Hans-Georg. Truth and Method. 2nd ed. Trans. Joel Weinsheimer
and Donald Marshall. New York: Crossroad, 1989.
Gardner, Howard. Art, Mind, and Brain: A Cognitive Approach to Creativity.
New York: Basic Books, 1982.
Gardner, John and John Maier, eds. and trans. (with the assistance of Richard
A. Henshaw). Gilgamesh. (Sin-Leqi-Unninni version.) New York: Vintage
Books, 1985.
Garman, Michael. Psycholinguistics. Cambridge: Cambridge University Press,
1990.
Gathercole, Susan. “Models of Verbal Short-Term Memory.” In Cognitive
Models of Memory. Ed. Martin A. Conway. Cambridge, MA: MIT Press,
1997.
and Alan Baddeley. Working Memory and Language. Hove, East Sussex:
Gelpke, Rudolf. “Translator’s Preface.” In The Story of Layla and Majnun.
Nizami. Ed. and trans. Rudolf Gelpke. New Lebanon, NY: Omega Publi-
cations, 1997.
Gergen, Kenneth J. and Keith E. Davis, eds. The Social Construction of the
Person. New York: Springer-Verlag, 1985.
Gerow, Edwin. Indian Poetics. Wiesbaden: Otto Harrassowitz, 1977.
Gerstle, C. Andrew. “The Concept of Tragedy in Japanese Drama.” Japan
Review 1 (1990).
Gibbs, Raymond W. The Poetics of Mind: Figurative Thought, Language, and
Understanding. Cambridge: Cambridge University Press, 1994.
Gifford, Don with Robert J. Seidman. Ulysses Annotated: Notes for James Joyce’s
Ulysses. Revised and expanded ed. Berkeley: University of California Press,
1988.
Gillegan, Carol. In a Different Voice: Psychological Theory and Women’s Develop-
ment. Cambridge: Harvard University Press, 1982.
Girard, René. Violence and the Sacred. Trans. Patrick Gregory. Baltimore, MD:
Johns Hopkins University Press, 1977.
Gnoli, Raniero. The Aesthetic Experience According to Abhinavagupta. Varanasi:
Chowkhamba Sanskrit Series Office, 1968.
Greimas, A. J. Structural Semantics: An Attempt at a Method. Trans. Daniele
McDowell, Ronald Schleifer, and Alan Velie. Lincoln: University of
Nebraska Press, 1983.
Grice, Paul. “Logic and Conversation.” In Studies in the Way of Words.
Cambridge, MA: Harvard University Press, 1989.
Grünbaum, Adolf. The Foundations of Psychoanalysis: A Philosophical Critique.
Berkeley: University of California Press, 1984.
272 Works Cited
Guépin, J. P. The Tragic Paradox: Myth and Ritual in Greek Tragedy. Amsterdam:
A. M. Hakkert, 1968.
Gurgānı̄, Fakhr ud-Din. Vı̄s and Rāmı̄n. Trans. George Morrison. New York:
Hadas, Moses. Introduction to The Trojan Women. In Euripides: Ten Plays.
Trans. and ed. Moses Hadas and John McLean. New York: Bantam Books,
1960.
Halász, László. “Effect and reminding in literary text processing.” In Empiri-
cal Studies of Literature: Proceedings of the Second IGEL Conference, Amsterdam
1989. Elrud Ibsch, Dick Schram, and Gerard Steen, eds. Amsterdam:
Rodopi, 1991.
. “Emotional Effect and Reminding in Literary Processing.” Poetics 20
(1991).
Halford, Aubrey S. and Giovanna M. Halford. The Kabuki Handbook. Rutland
VT: Charles E. Tuttle, 1956.
Hampton, James A. “Psychological Representation of Concepts.” In Cogni-
tive Models of Memory. Ed. Martin A. Conway. Cambridge, MA: MIT Press,
1997.
Harré, R. and W. G. Parrott, eds. The Emotions: Social, Cultural, and Biological
Dimensions. London: Sage, 1996.
Harris, Paul, Tjeert Olthof, Mark Terwogt, and Charlotte Hardman. “Chil-
dren’s Knowledge of the Situations that Provoke Emotion.” International
Journal of Behavioral Development 10 (1987).
Hart, George. The Poems of Ancient Tamil. Berkeley: University of California
Press, 1975.
. Poets of the Tamil Anthologies: Ancient Poems of Love and War. Princeton,
NJ: Princeton University Press, 1979.
Hauser, Arnold. The Social History of Art (Volume I): Prehistoric, Ancient-
Oriental, Greece and Rome; Middle Ages. Trans. Stanley Godman. New York:
Vintage Books, 1957.
Hawkes, David. “The Quest of the Goddess.” In Studies in Chinese Literary
Genres. Ed. Cyril Birch. Berkeley: University of California Press, 1974.
Heath, Peter. The Thirsty Sword: Sı̄rat ‘Antar and the Arabic Popular Epic. Salt
Lake City: University of Utah Press, 1996.
Herman, David. “Scripts, Sequences, and Stories: Elements of a Postclassical
Narratology.” PMLA 112. 5 (1997).
Hernadi, Paul. Beyond Genre: Tragicomedies of History on the Modern Stage.
Ithaca, NY: Cornell University Press, 1972.
. Interpreting Events: Tragicomedies of History on the Modern Stage. Ithaca,
NY: Cornell University Press, 1985.
Hjort, Mette and Sue Laver. “Introduction.” In Emotion and the Arts. Ed. Mette
Hjort and Sue Laver. New York: Oxford University Press, 1997.
Hobbs, Jerry. Literature and Cognition. Stanford, CA: Center for the Study of
Language and Information, 1990.
Hogan, Patrick Colm. “The Beautiful and the Merely Pleasing: Love, Art,
and the Jinnee in the Well Wrought Urn.” Time, Memory and the Verbal Arts:
Works Cited 273
Essays on the Thought of Walter Ong. Ed. Dennis Weeks. Selinsgrove, PA:
Susquehanna University Press, 1998.
. “Beauty, Politics, and Cultural Otherness: The Bias of Literary Dif-
ference.” In Literary India: Comparative Studies in Aesthetics, Colonialism, and
Culture. Ed. Patrick Hogan and Lalita Pandit. Albany: State University of
New York Press, 1995.
. Cognitive Science, Literature, and the Arts: A Guide for Humanists. New
York: Routledge, 2003.
. Colonialism and Cultural Identity: Crises of Tradition in the Anglophone
Literatures of India, Africa, and The Caribbean. Albany: State University of
. The Culture of Conformism: Understanding Social Consent. Durham, NC:
Duke University Press, 2001.
. Empire and Poetic Voice: Cognitive and Cultural Studies of Literary Tra-
dition and Colonialism. Albany, NY: State University of New York Press,
forthcoming.
. “Literary Universals.” Poetics Today 18 (1997).
. “Ideological Critique and the Profession of Literature.” The Review of
Education/Pedagogy/Cultural Studies 19.1 (1997).
. On Interpretation: Meaning and Inference in Law, Psychoanalysis, and
Literature. Athens: University of Georgia Press, 1995.
. “Particular Myths, Universal Ethics: Wole Soyinka’s The Swamp
Dwellers in the New Nigeria.” Modern Drama 41.4 (1999).
. Philosophical Approaches to the Study of Literature. Gainesville: Univer-
sity Press of Florida, 2000.
. “The Possibility of Aesthetics.” British Journal of Aesthetics 34.4 (1994).
. “Shakespeare, Eastern Theatre, and Literary Universals: Drama in
the Context of Cognitive Science.” In Shakespeare East and West. Ed. Minoru
Fujita. Richmond, Surrey: Japan Library of Curzon Press, 1996.
. “Some Prolegomena to the Study of Literary Difference.” Poetics 22
(1994).
. “Toward a Cognitive Science of Poetics: Ānandavardhana,
Abhinavagupta, and the Theory of Literature.” College Literature 23.1
(1996).
Holland, John, Keith Holyoak, Richard Nisbett, and Paul Thagard. Induction:
Processes of Inference, Learning, and Discovery. Cambridge, MA: MIT Press,
1986.
Holland, Norman. The Brain of Robert Frost. New York: Routledge, 1988.
. 5 Readers Reading. New Haven, CT: Yale University Press, 1975.
. Poems in Persons: An Introduction to the Psychoanalysis of Literature.
New York: W. W. Norton, 1975.
and Murray Schwartz. “The Delphi Seminar.” College English 36 (1975).
Homer. The Iliad. Trans. W. H. D. Rouse. New York: New American Library,
1938.
. The Odyssey. Trans. A. T. Murray and rev. George E. Dimock.
Cambridge, MA: Harvard University Press, 1995.
274 Works Cited
Hsia, C. T. “The Military Romance.” In Studies in Chinese Literary Genres.

Ed. Cyril Birch. Berkeley: University of California Press, 1974.
. “A Dream of Red Mansions.” In Eastern Canons: Approaches to the
Asian Classics. Ed. William Theodore de Bary and Irene Bloom. New York:
Humayuni, Sadeq. “An Analysis of the Ta’zayih of Qasem.” In Ta’ziyeh: Ritual
and Drama in Iran. Ed. Peter Chelkowski. New York: New York University
Press, 1979.
Huot, Sylvia. “Troubadour Lyric and Old French Narrative.” In The
Troubadours: An Introduction. Ed. Simon Gaunt and Sarah Kay. Cambridge:
Ibn Ishāq. The Life of Muhammad: A Translation of Ishāq’s Sı̄rat Rasūl Allāh.
Trans. and ed. A. Guillaume. London: Oxford University Press, 1955.
Ilangô Adigal. Shilappadikaram. Trans. Alain Daniélou. New York: New
Directions, 1965.
Ingalls, Daniel H. H., ed. The Dhvanyaloka of Ānandavardhana with the
Locana of Abhinavagupta. Trans. Daniel Ingalls, Jeffrey Masson, and M. V.
Patwardhan. Cambridge, MA: Harvard University Press, 1990.
Ingarden, Roman. The Literary Work of Art: An Investigation on the Borderlines
of Ontology, Logic, and the Theory of Literature. Trans. George Grabowicz.
Evanston, IL: Northwestern University Press, 1973.
Irele, F. Abiola. “Africa: The Mali Epic of Son-Jara.” In The Norton Anthology of
World Masterpieces, Volume 1: Beginnings to 1650. Expanded ed. Ed. Maynard
Mack et al. New York: W. W. Norton, 1995.
Jakobson, Roman. Language in Literature. Ed. Krystyna Pomorska and Stephen
Rudy. Cambridge, MA: Harvard University Press, 1987.
Jankowiak, W. R. and E. F. Fischer. “A Cross-Cultural Perspective on Roman-
tic Love.” In Human Emotions: A Reader. Ed. Jennifer Jenkins, Keith Oatley
and Nancy Stein. Oxford: Basil Blackwell, 1998.
Jauss, Hans Robert. Toward an Aesthetic of Reception. Trans. Timothy Bahti.
Minneapolis: University of Minnesota Press, 1982.
Jenkins, Jennifer M., Keith Oatley, and Nancy L. Stein, eds. Human Emotions:
A Reader. Oxford: Basil Blackwell, 1998.
Jenkins, Lyle. Biolinguistics: Exploring the Biology of Language. Cambridge:
Jerusalem Bible. Gen. Ed. Alexander Jones. Garden City, NY: Doubleday, 1966.
Johnson, John William. “The Epic of Son-Jara.” In Oral Epics from Africa:
Vibrant Voices from a Vast Continent. Ed. John William Johnson, Thomas
A. Hale, and Stephen Belcher. Bloomington: Indiana University Press,
1997.
Johnson-Laird, Philip N. The Computer and the Mind: An Introduction to
Cognitive Science. Cambridge: Harvard University Press, 1988.
and Keith Oatley. “The Language of Emotions: An Analysis of a
Semantic Field.” Cognition and Emotion 3.2 (1989).
and P. C. Wason, eds. Thinking: Readings in Cognitive Science. New
York: Cambridge University Press, 1977.
Works Cited 275
Joseph, George. “Indian Actress Creates Flutter On Board Plane.” India Abroad
(July 26, 2002).
Joyce, James. Ulysses. Ed. Hans Walter Gabler with Wolfard Steppe and Claus
Melchior. New York: Vintage Books, 1986.
Kahneman, Daniel and Dale T. Miller. “Norm Theory: Comparing Reality to
Its Alternatives.” Psychological Review 93.2 (1986).
Kailasapathy, K. Tamil Heroic Poetry. Oxford: Clarendon Press, 1968.
Kalhan.a. Rājataraṅgin.ı̄: The Saga of the Kings of Kaśmı̄r. Trans. and ed. Ranjit
Sitaram Pandit. New Delhi: Sahitya Akademi, 1968.
Kālidāsa. “Abhijñānaśākuntalam (The Recognition of Śakuntalā).” In
Kālidāsa: The Loom of Time. Trans. and ed. Chandra Rajan. New Delhi:
Penguin, 1989.
. Theater of Memory: The Plays of Kālidāsa. Ed. Barbara Stoler Miller.
Trans. Edwin Gerow, David Gitomer, and Barbara Stoler Miller. New York:
Kaut.ilya. Arthaśāstra. 6th ed. Trans. R. Shamasastry. Mysore: Mysore Printing
and Publishing, 1960.
Kazin, Alfred, ed. The Portable Blake. New York: Viking, 1968.
Keene, Donald, ed. Anthology of Japanese Literature to the Nineteenth Century.
, ed. and trans. Major Plays of Chikamatsu. New York: Columbia
, ed. and trans. Twenty Plays of the Nō Theatre. New York: Columbia
King, Bruce. Derek Walcott: A Caribbean Life. Oxford: Oxford University Press,
2000.
Kintsch, Walter, Alice Healy, Mary Hegarty, Bruce Pennington, and Timothy
Salthouse. “Models of Working Memory: Eight Questions and Some Gen-
eral Issues.” In Models of Working Memory: Mechanisms of Active Mainte-
nance and Executive Control. Ed. Akira Miyake and Priti Shah. Cambridge:
Kiparsky, Paul. Pān.ini as a Variationist. Ed. S. D. Joshi. Cambridge, MA: MIT
Press, 1979.
. “The Role of Linguistics in a Theory of Poetry.” In Essays in Modern
Stylistics. Ed. Donald C. Freeman. London: Methuen, 1981.
. “On Theory and Interpretation.” In The Linguistics of Writing: Argu-
ments Between Language and Literature. Ed. Nigel Fabb, Derek Attridge, Alan
Durant, and Colin MacCabe. New York: Methuen, 1987.
Klein, Stanley, Judith Loftus, J. Gregory Trafton, and Robert W. Fuhrman.
“Use of Exemplars and Abstractions in Trait Judgments: A Model of
Trait Knowledge About Self and Others.” Journal of Personality and Social
Psychology 63 (1992).
Klemenz-Belgardt, Edith. “American Research on Response to Literature:
The Empirical Studies.” Poetics 10.4/5 (1981).
Knight, W. F. Jackson. “Introduction.” In The Aenid. Virgil. Trans. W. F. Jackson
Knight. New York: Penguin, 1958.
276 Works Cited
Knox, Bernard. Introduction and Notes. In The Three Theban Plays. Sophocles.
Trans. Robert Fagles. Intro. and notes Bernard Knox. New York: Penguin,
1984.
Kouyaté, Djeli Mamoudou. Sundiata: An Epic of Old Mali. Ed. D. T. Niane.
Trans. G. D. Pickett. New York: Longman, 1965.
Kövecses, Zoltán. Metaphor and Emotion: Language, Culture, and Body in Human
Feeling. Cambridge: Cambridge University Press and Paris: Editions de la
Maison des Sciences de l’Homme, 2000.
Kuan Han-ch’ing. The Injustice Done to Tou Ngo. In Six Yüan Plays. Ed. and
trans. Liu Jung-en. New York: Penguin, 1972.
Kunene, Daniel P. Heroic Poetry of the Basotho. Oxford: Clarendon Press, 1971.
Labouvie-Vief, Gisela. Psyche and Eros: Mind and Gender in the Life Course.
Labov, William. Language in the Inner City: Studies in the Black English Vernac-
ular. Philadelphia: University of Pennsylvania Press, 1972.
Lakatos, Imre. “Falsification and the Methodology of Scientific Research Pro-
grammes.” In Criticism and the Growth of Knowledge. Ed. Imre Lakatos and
Alan Musgrave. Cambridge: Cambridge University Press, 1970.
Lakoff, George. Women, Fire, and Dangerous Things: What Categories Reveal
about the Mind. Chicago, IL: University of Chicago Press, 1987.
and Mark Johnson. Metaphors We Live By. Chicago, IL: University of
Chicago Press, 1980.
Lal, P. Great Sanskrit Plays. New York: New Directions, 1964.
Lameer, Joep. Al-Fārābı̄ and Aristotelian Syllogistics: Greek Theory and Islamic
Practice. Leiden: E. J. Brill, 1994.
Larsen, Steen, János László, and Uffe Seilman. “Across Time and Place:
Cultural-Historical Knowledge and Personal Experience in Appreciation
of Literature.” In Empirical Studies of Literature: Proceedings of the Second
IGEL Conference, Amsterdam 1989. Elrud Ibsch, Dick Schram, and Gerard
Steen, eds. Amsterdam: Rodopi, 1991.
LeDoux, Joseph. The Emotional Brain: The Mysterious Underpinnings of Emo-
tional Life. New York: Touchstone, 1996.
. Synaptic Self: How Our Brains Become Who We Are. New York: Viking,
2002.
Lerdahl, Fred and Ray Jackendoff. A Generative Theory of Tonal Music.
Cambridge, MA: MIT Press, 1983.
Levinson, Jerrold. “Emotion in Response to Art: A Survey of the Terrain.”
In Emotion and the Arts. Ed. Mette Hjort and Sue Laver. New York: Oxford
Lévi-Strauss, Claude. The Raw and the Cooked (Introduction to a Science of
Mythology, Vol. 1). Trans. John and Doreen Weightman. New York: Harper
and Row, 1969.
. From Honey to Ashes (Introduction to a Science of Mythology, Vol. 2).
Trans. John and Doreen Weightman. New York: Harper and Row, 1973.
Levy, Reuben. An Introduction to Persian Literature. New York: Columbia
Works Cited 277
Lewis, Geoffrey, ed. and trans. The Book of Dede Korkut. New York: Penguin,
1974.
Lichtenstadter, Ilse. Introduction to Classical Arabic Literature. New York:
Schocken Books, 1976.
Li Ch’ing-Chao. Complete Poems. Trans. and ed. Kenneth Rexroth and Ling
Chung. New York: New Directions, 1979.
Lincoln, Kenneth. “Foreword.” In American Indian Poetry: An Anthology of
Songs and Chants. Ed. George W. Cronyn. New York: Fawcett Columbine,
1991.
Ling Chung, “Biography of Li Ch’ing-Chao.” In Li Ch’ing-Chao. Complete
Poems. Trans. and ed. Kenneth Rexroth and Ling Chung. New York: New
Directions, 1979.
Liu Hsieh. The Literary Mind and the Carving of Dragons. Trans. Vincent Yu-
chung Shih. New York: Columbia University Press, 1959.
Liu, James J. Y. The Art of Chinese Poetry. Chicago, IL: University of Chicago
Press, 1962.
. “Some Literary Qualities of the Lyric (Tz’u).” In Studies in Chinese
Literary Genres. Ed. Cyril Birch. Berkeley: University of California Press,
1974.
Liu Jung-en, ed. and trans. Six Yüan Plays. New York: Penguin, 1972.
Locke, John L. “Development of the Capacity for Spoken Language.” In
The Handbook of Child Language. Ed. Paul Fletcher and Brian MacWhinney.
Oxford: Basil Blackwell, 1995.
Lord, Albert. The Singer of Tales. Cambridge, MA: Harvard University Press,
1960.
Lukács, Georg. Essays on Realism. Ed. Rodney Livingstone. Trans. David
Fernbach. Cambridge, MA: MIT Press, 1981.
Luo Guanzhong. Three Kingdoms: A Historical Novel. Ed. and trans. Moss
Roberts. Beijing: Foreign Language Press and Berkeley: University of
California Press, 1999.
Lutz, Catherine A. “Engendered Emotion: Gender, Power, and the Rhetoric
of Emotional Control in American Discourse.” In Language and the Poli-
tics of Emotion. Ed. Catherine A. Lutz and Lila Abu-Lughod. Cambridge:
Cambridge University Press and Paris: Editions de la Maison des Science
de l’Homme, 1990.
Lutz, Catherine A. and Lila Abu-Lughod, eds. Language and the Politics of
Emotion. Cambridge: Cambridge University Press and Paris: Editions de
la Maison des Science de l’Homme, 1990.
Ma Chih-yüan. Autumn in Han Palace. In Six Yuan Plays. Ed. and trans. Liu
Jung-en. New York: Penguin, 1972.
Mack, Maynard, et al., eds. The Norton Anthology of World Masterpieces,
Volume 1: Beginnings to 1650. Expanded ed. New York: W. W. Norton, 1995.
Mahapatra, Sitakant, ed. and trans. Staying is Nowhere: An Anthology of Kondh
and Paraja Poetry. Calcutta, India: Writers Workshop, n.d.
Marcus, Ruth Barcan. “Moral Dilemmas and Consistency.” The Journal of
Philosophy 77.3 (March 1980).
278 Works Cited
Markandaya, Kamala. Nectar in a Sieve. New York: New American Library,

1982.
Martinovitch, Nicholas N. The Turkish Theatre. New York: Theatre Arts, 1933.
Marx, Karl. Economic and Philosophic Manuscripts of 1844. Ed. Dirk Struik.
Trans. Martin Milligan. New York: International Publishers, 1964.
. Capital, Vol. 1: A Critical Analysis of Capitalist Production. Ed.
Frederick Engels. Trans. Samuel Moore and Edward Aveling. New York:
International Publishers, 1967.
Mathews, A. “Biases in Emotional Processing.” In Human Emotions: A Reader.
Ed. Jennifer M. Jenkins, Keith Oatley, and Nancy L. Stein. Oxford: Basil
Blackwell, 1998.
McClelland, J. and D. Rumelhart. “Distributed Memory and the Representa-
tion of General and Specific Information.” Journal of Experimental Psychol-
ogy: General 114 (1985).
. “A Distributed Model of Human Learning and Memory.” In Par-
allel Distributed Processing, Vol. 2. Ed. J. McClelland and D. Rumelhart.
Cambridge, MA: MIT Press, 1986.
McLeod, Peter, Kim Plunkett, and Edmund T. Rolls. Introduction to Connec-
tionist Modelling of Cognitive Processes. Oxford: Oxford University Press,
1998.
Miller, Barbara Stoler. “Introduction.” In Love Song of the Dark Lord: Jayadeva’s
Gı̄tagovinda. Ed. and Trans., Brabara Stoley Miller. New York: Columbia
Miller, George and Philip Johnson-Laird. Language and Perception. Cambridge,
MA: Harvard University Press, 1976.
Miller, Peggy J., Lisa Hoogstra, Judith Mintz, Heidi Fung, and Kimberly
Williams. “Troubles in the Garden and How They Get Resolved: A Young
Child’s Transformation of His Favorite Story.” In Memory and Affect in De-
velopment: The Minnesota Symposia on Child Psychology. Vol. 26. Ed. Charles
A. Nelson. Hillsdale, NJ: Lawrence Erlbaum, 1993.
Milton, John. Paradise Lost. In Paradise Lost, Paradise Regained, Samson
Agonistes. Intro. Harold Bloom. New York: Collier, 1962.
Miyake, Akira and Priti Shah. “Toward Unified Theories of Working Mem-
ory: Emerging General Consensus, Unresolved Theoretical Issues, and
Future Research Directions.” In Models of Working Memory: Mechanisms of
Active Maintenance and Executive Control. Ed. Akira Miyake and Priti Shah.
Monier-Williams, Monier. A Sanskrit-English Dictionary. New ed. Delhi:
Motilal Banarsidass, 1997.
Moreh, Shmuel. Live Theatre and Dramatic Literature in the Medieval Arab World.
New York: New York University Press, 1992.
Morrison, George. “Introduction.” In Vı̄s and Rāmı̄n. Fakhr ud-Din
Gurgānı̄. Trans. George Morrison. New York: Columbia University Press,
1972.
Murnaw, F. W., dir. Nosferatu, A Symphony of Horror. Screenplay by Henrik
Galeen. Cinematography by Günther Krampf and Fritz Arno Wagner.
1922.
Works Cited 279
Murphy, Roland E. “Psalms, Book of.” In The Oxford Companion to the Bible. Ed.
Bruce M. Metzger and Michael D. Coogan. New York: Oxford University
Press, 1993.
Namiki Sōsuke with Namiki Shōzō I, Namiki Geiji, and Asada Itchō. Chronicle
of the Battle of Ichinotani. In Kabuki: Five Classic Plays. Ed. and trans. James
R. Brandon. Honolulu: University of Hawaii Press, 1992.
Narayan, R. K. The Ramayana. New York: Penguin, 1972.
Nārāyan.a. Hitopadeśa. Trans. and ed. M. R. Kale. Delhi, India: Motilal
Banarsidass, 1967.
Nelson, Katherine. “Events, Narratives, Memory: What Develops?” In Mem-
ory and Affect in Development: The Minnesota Symposia on Child Psychology
(Volume 26). Ed. Charles A. Nelson. Hillsdale, NJ: Lawrence Erlbaum,
1993.
Ngũgı̃ wa Thiong’o. Moving the Centre: The Struggle for Cultural Freedoms.
London: James Curry, 1993.
Nisbett, Richard and Lee Ross. Human Inference: Strategies and Shortcomings
of Social Judgment. Englewood Cliffs, NJ: Prentice-Hall, 1980.
Nizami. The Story of Layla and Majnun. Ed. and trans. Rudolf Gelpke in col-
laboration with E. Mattin and G. Hill. Final chapter trans. Zia Inayat Khan
and Omid Safi. New Lebanon, NY: Omega Publications, 1997.
Oatley, Keith. Best Laid Schemes: The Psychology of Emotions. Cambridge:
Cambridge University Press and Paris: Editions de la Maison des Sciences
de l’Homme, 1992.
. “Emotions and the Story Worlds of Fiction.” In Narrative Impact.
Ed. Melanie Green, Jeffrey Strange, and Timothy Brock. Mahwah, NJ:
and P. N. Johnson-Laird. “Toward a Cognitive Theory of Emotions.”
Cognition and Emotion 1.1 (1987).
O’Connor-Visser, E. A. M. E. Aspects of Human Sacrifice in the Tragedies of
Euripides. Amsterdam: B. R. Gruner Publishing, 1987.
Ohnuki-Tierney, Emiko. The Ainu of the Northwest Coast of Southern Sakhalin.
New York: Holt, Rinehart and Winston, 1974.
Ōkabou Ojobolo. The Ozidi Saga. Ed. and trans. J. P. Clark. Ibadan, Nigeria:
Ibadan University Press and Oxford University Press, 1977.
Ong, Walter J., S.J. “The Jinnee in the Well-Wrought Urn.” In The Barbarian
Within and Other Fugitive Essays and Studies. New York: Macmillan,
1962.
. Orality and Literacy: The Technologizing of the Word. New York:
Methuen, 1982.
Opland, Jeff. Xhosa Oral Poetry: Aspects of a Black South African Tradition.
Ortony, Andrew “Beyond Literal Similarity.” Psychological Review 86 (1979).
and Gerald L. Clore. “Emotions, Moods, and Conscious Awareness:
Comment on Johnson-Laird and Oatley’s ‘The Language of Emotions: An
Analysis of a Semantic Field.’” Cognition and Emotion 3.2 (1989).
, Gerald Clore, and Allan Collins. The Cognitive Structure of Emotions.
280 Works Cited
Owen, Stephen. “The Nine Songs.” In The Norton Anthology of World Master-
pieces, Volume 1: Beginnings to 1650. Expanded ed. Ed. Maynard Mack et al.
New York : W. W. Norton, 1995.
Padgett, Ron, ed. The Teachers & Writers Handbook of Poetic Forms. New York:
Teachers & Writers Collaborative, 1987.
Pagels, Elaine. The Origin of Satan. New York: Vintage, 1995.
Pandit, B. N. Essence of the Exact Reality or Paramārthasāra of Abhinavagupta.
Delhi: Munshiram Manoharlal, 1991.
Pandit, Lalita. “Caste, Race, and Nation: Dialectic in Rabindranath Tagore’s
Gora.” In Literary India: Comparative Studies in Aesthetics, Colonialism, and
Culture. Ed. Patrick Colm Hogan and Pandit. Albany: State University of
Panksepp, Jaak. Affective Neuroscience: The Foundations of Human and Animal
Emotions. New York: Oxford University Press, 1998.
Pavel, Thomas G. The Poetics of Plot: The Case of English Renaissance Drama.
Minneapolis: University of Minnesota Press, 1985.
Philippi, Donald L. Songs of Gods, Songs of Humans: The Epic Tradition of the
Ainu. San Francisco: North Point Press, 1982.
Pinker, Steven. How the Mind Works. New York: W. W. Norton, 1997.
Plantinga, Carl. “Film Theory and Aesthetics: Notes on a Schism.” The Journal
of Aesthetics and Art Criticism 51 (1993).
Prince, Gerald. A Grammar of Stories. The Hague: Mouton, 1973.
. Narratology: The Form and Functioning of Narrative. Berlin: Mouton,
1982.
Puenzo, Luis, dir. The Plague. Screenplay by Luis Puenzo. Based on La Peste by
Albert Camus. Photography by Félix Monti. Edited by Juan Carlos Macı́as.
Companie Française Cinématographique, The Pepper-Prince Company,
Oscar Kramer Producciones SA, 1992.
Pukui, Mary Kawena and Alfons L. Korn. The Echo of Our Song: Chants and
Poems of the Hawaiians. Honolulu: University Press of Hawaii, 1973.
Quint, David. Epic and Empire: Politics and Generic Form from Virgil to Milton.
Princeton, NJ: Princeton University Press, 1993.
Racine, Jean. Phèdre. In Théâtre Complet: II. Ed. André Stegmann. Paris:
Flammarion, 1965.
Richardson, Alan. “Rethinking Romantic Incest: Human Universals, Literary
Representation, and the Biology of Mind.” New Literary History 31 (2000).
Richman, Paula. “E. V. Ramasami’s Reading of the Rāmāyan.a.” In Many
Rāmāyan.as. Ed. Paula Richman, 1991.
. “Introduction: The Diversity of the Rāmāyan.a Tradition.” In Many
Rāmāyan.as, Ed. Paula Richman, 1991.
, ed. Many Rāmāyan.as: The Diversity of a Narrative Tradition in South
Asia. Berkeley: University of California Press, 1991.
Ross, Lee and Richard E. Nisbett. The Person and the Situation: Perspectives of
Social Psychology. Philadelphia, PA: Temple University Press, 1991.
Rubin, David C. Memory in Oral Traditions: The Cognitive Psychology of Epic,
Ballads, and Counting-out Rhymes. New York: Oxford University Press, 1995.
Works Cited 281
Rumelhart, David E. “Notes on a Schema for Stories.” In Representation and

Understanding: Studies in Cognitive Science. Ed. Daniel G. Bobrow and Allan
Collins. New York: Academic Press, 1975.
Rureke, Shé-kárisi. The Mwindo Epic. Ed. and trans. Daniel Biebuyck and
Kahombo C. Mateene. Berkeley: University of California Press, 1971.
Russell, James. “In Defense of a Prototype Approach to Emotion Concepts.”
Journal of Personality and Social Psychology 60.1 (1991).
Russell, Ralph. The Pursuit of Urdu Literature: A Select History. London: Zed
Books, 1992.
Said, Edward. “Bookless in Gaza.” The Nation 263.8 (September 23, 1996).
Śaktibhadra. Āścaryacūd.aman.i. Trans. V. Raghavan. In The Wondrous Crest-
Jewel in Performance: Text and Translation of the Āścaryacūd.aman.i of Śaktibhadra
with the Production Manual from the Tradition of Kūt.iyāt..tam Sanskrit Drama.
Ed. Clifford Reis Jones. Delhi, India: Oxford University Press, 1984.
Sandars, N. K. “Gilgamesh.” In The Norton Anthology of World Literature.
Expanded ed. Gen. ed. Maynard Mack. New York: W. W. Norton,
1997.
Sarbin, Theodore R. “Believed-In Imaginings: A Narrative Approach.” In
Believed-In Imaginings: The Narrative Construction of Reality. Ed. Joseph De
Rivera and Theodore R. Sarbin. 1998.
Sartre, Jean-Paul. Being and Nothingness: A Phenomenological Essay on Ontol-
ogy. Trans. Hazel Barnes. New York: Washington Square, 1966.
Sartne, Jean-Paul.. L’être et le néant, essai d’ontologie phénoménologuque. Paris:
Gallimard, 1943.
Schachter, Stanley. “The Interaction of Cognitive and Physiological Deter-
minants of Emotional State.” In Cognitive Theories in Social Psychology.
Ed. Leonard Berkowitz. New York: Academic Press, 1978.
Schacter, Daniel L. Searching for Memory: The Brain, the Mind, and the Past.
New York: Basic Books, 1996.
Schank, Roger C. “Explanation: A First Pass.” In Experience, Memory, and
Reasoning. Ed. Janet L. Kolodner and Christopher K. Riesbeck. Hillsdale,
NJ: Lawrence Erlbaum, 1986.
Schank, Roger C. and Robert P. Abelson. Scripts, Plans, Goals, and Understand-
ing: An Inquiry into Human Knowledge Structures. Hillsdale, NJ: Lawrence
Erlbaum, 1977.
Schwarz, Norbert. “Social Cognition: Information Accessibility and Use in
Social Judgment.” In Thinking: An Invitation to Cognitive Science. Vol. 3. 2nd
ed. Cambridge, MA: MIT Press.
Seilman, S. and S. F. Larsen. “Personal Resonance to Literature: A Study of
Remindings While Reading.” Poetics 18 (1989).
Shaver, Phillip, Judith Schwartz, Donald Kirson, and Cary O’Connor. “Emo-
tion Knowledge: Further Exploration of a Prototype Approach.”Journal of
Personality and Social Psychology 52.6 (1987).
Sherzer, Joel. “Poetic Structuring of Kuna Discourse: The Line.” In Native
American Discourse: Poetics and Rhetoric. Ed. Sherzer and Anthony
Woodbury. Cambridge: Cambridge University Press, 1987.
282 Works Cited
Sherzer, Joel and Anthony Woodbury. “Introduction.” In Native American

Discourse: Poetics and Rhetoric. Ed. Sherzer and Woodbury. Cambridge:
Shakespeare, William. Hamlet. Ed. Louis B. Wright and Virginia A. LaMar.
New York: Washington Square Press, 1959.
. Macbeth. Ed. Louis B. Wright and Virginia A. LaMar. New York:
Washington Square Press, 1959.
. The Tragedy of Romeo and Juliet. In The Riverside Shakespeare. Ed.
G. Blakemore Evans. Boston, MA: Houghton Mifflin, 1974.
Shanks, David R. “Representation of Categories and Concepts in Memory.”
In Cognitive Models of Memory. Ed. Martin A. Conway. Cambridge, MA:
MIT Press, 1997.
Shibatani, Masayoshi. The Languages of Japan. Cambridge: Cambridge
Shirane, Haruo. “The Poetry of Matsuo Basho.” In Eastern Canons: Approaches
to the Asian Classics. Ed. William Theodore de Bary and Irene Bloom.
New York: Columbia University Press, 1990.
Singha, Rina and Reginald Massey. Indian Dances: Their History and Growth.
New York: G. Braziller, 1967.
Sissoko Kabine, et al. Da Monzon de Segou: Épopée Bambara. 4 vols. Ed. and
trans. Lilyan Kesteloot with A. Traore, J.-B. Traore, and Amadou Hampate
Ba. Paris: Fernand Nathan Editeur, 1972.
Sivan, Santosh, dir. The Terrorist. Story by Santosh Sivan. Screenplay by
Santosh Sivan, Ravi Deshpande, and Vijay Deveshwar. Photography
by Santosh Sivan. Edited by Sreekar Prasad. Moderne Gallerie Motion
Picture – Wonderfilms, 1998.
Sommers, Christina. “Filial Morality.” The Journal of Philosophy 83.8 (August
1986).
Sophocles. The Three Theban Plays. Trans. Robert Fagles. Intro. and notes
Bernard Knox. New York: Penguin, 1984.
Soyinka, Wole. “The Bacchae of Euripides.” In Collected Plays 1. Oxford:
Oxford University Press, 1973.
. Death and the King’s Horseman. New York: Hill and Wang, 1987.
. “The Strong Breed.” In Collected Plays 1. Oxford: Oxford University
Press, 1973.
. “The Swamp Dwellers.” In Collected Plays 1. Oxford: Oxford Univer-
sity Press, 1973.
Sperber, D. and D. Wilson. Relevance: Communication and Cognition. Oxford:
Basil Blackewell, 1986.
Spolsky, Ellen. Gaps in Nature: Literary Interpretation and the Modular Mind.
Albany: State University of New York Press, 1993.
“Stratagem of Interlocking Rings.” In Six Yüan Plays. Ed. and trans. Liu
Jung-en. New York: Penguin, 1972.
Suso, Bamba. In Sunjata. Bamba Suso. Trans. and annotated Gordon Innes.
and Banna Kanute. Sunjata. Trans. and annotated by Gordon Innes.
Ed. and intro. Lucy Durán and Graham Furniss. New York: Penguin, 1999.
Works Cited 283
The Tale of the Heike. Trans. Helen Craig McCullough. Stanford, CA: Stanford
Tan, Ed S. Emotion and the Structure of Narrative Film: Film as an Emotion
Machine. Trans. Barbara Fasting. Mahwah, NJ: Lawrence Erlbaum, 1996.
Terence. The Comedies. Ed. and trans. Betty Radice. New York: Penguin,
1976.
Todorov, Tzvetan. The Fantastic: A Structural Approach to a Literary Genre.
Trans. Richard Howard. Cleveland, OH: Case Western Reserve University
Press, 1973.
Tomkins, S. S. “Script Theory: Differential Magnification of Affects.” In
Human Emotions: A Reader. Ed. Jennifer M. Jenkins, Keith Oatley, and Nancy
L. Stein. Oxford: Basil Blackwell, 1998.
Tooby, John and Leda Cosmides. “Does Beauty Build Adapted Minds?
Toward an Evolutionary Theory of Aesthetics, Fiction, and the Arts.”
SubStance: A Review of Theory and Literary Criticism 94/95 (2001).
Trask, Willard R., ed. The Unwritten Song: Poetry of the Primitive and Tra-
ditional Peoples of the World, Volume I: The Far North/Africa/Indonesia/
Melanesia/Australia. New York: Macmillan, 1966.
Turner, Frederick. “The Neural Lyre: Poetic Meter, The Brain, and Time.”
In Natural Classicism: Essays on Literature and Science. New York: Paragon
House Publishers, 1985.
Turner, Mark. The Literary Mind. New York: Oxford University Press, 1996.
. Reading Minds: The Study of English in the Age of Cognitive Science.
Princeton, NJ: Princeton University Press, 1991.
Tutuola, Amos. The Palm-Wine Drinkard and My Life in the Bush of Ghosts.
New York: Grove Press, 1994.
Tversky, Amos. “Features of Similarity.” Psychological Review 84 (1977).
Vālmı̄ki. Srimad Vālmı̄ki Rāmāyan.a. 3 vols. Trans. N. Ragunathan. Madras,
India: Vighneswara Publishing, 1981.
Varley, H. Paul. “The Tale of the Heike.” In Eastern Canons: Approaches to the
Asian Classics. Ed. William Theodore de Bary and Irene Bloom. New York:
Virgil. The Aeneid. Trans. W. F. Jackson Knight. Revised ed. New York: Penguin
Books, 1958.
. P. Vergili Maronis: Opera. Ed. R. A. B. Mynors. Oxford: Clarendon
Press, 1969.
Viśākhadatta. Mudrārākshasa. 6th ed. Ed. and trans. M. R. Kale. Delhi: Motilal
Banarsidass, 1976.
Voight, Ellen Bryant. “Narrative and Lyric: Structural Corruption.” The South-
ern Review 30 (1994).
Vyāsa. Mahabharata of Krishna-Dwaipayana Vyasa. 4 vols. Trans. Kisari Mohan
Ganguli. New Delhi: Munshiram Manoharlal, 1970.
Waley, Arthur, ed. and trans. The Book of Songs. Boston: Houghton Mifflin,
1937.
, ed. and trans. The Nō Plays of Japan. London: Unwin, 1988.
Watson, Burton. Early Chinese Literature. New York: Columbia University
Press, 1962.
284 Works Cited
Watanabe, Hitoshi. The Ainu Ecosystem: Environment and Group Structure.

Seattle: University of Washington Press, 1972.
White, Hayden. Metahistory: The Historical Imagination in Nineteenth-Century
Europe. Baltimore, MD: Johns Hopkins University Press, 1973.
Wierzbicka, Anna. “Emotions.” In Semantic Primitives. Trans. Anna
Wierzbicka and John Besemeres. Frankfurt am Main, West Germany:
Athenaum Verlag, 1972.
. Emotions Across Languages and Cultures: Diversity and Universals.
Williams, Mark. “Mansfield in Maoriland: Biculturalism, Agency and
Misreading.” In Modernism and Empire. Ed. J. Howard Booth and Nigel
Rigby. Manchester: Manchester University Press, 2000.
Wilson, Edward O. Consilience: The Unity of Knowledge. New York: Vintage,
1998.
Wittgenstein, Ludwig. Philosophical Investigations. 3rd ed. Trans G. E. M.
Anscombe. New York: Macmillan, 1958.
Woolner, A. C. and Lakshman Sarup, eds. and trans. Thirteen Plays of Bhāsa.
Delhi: Motilal Banarsidass, 1985.
Wright, Louis B. and Virginia A. LaMar, eds. Hamlet. In Shakespeare, Hamlet.
, eds. Macbeth. In Shakespeare, Macbeth.
Yarshater, Ehsan. “The Development of Iranian Literatures.” In Persian
Literature. Ed. Yarshater. N.C.: Bibliotheca Persica, 1988.
. “Foreword.” In The Epic of the Kings: Shah-nama by Ferdowsi. Ed. and
trans. Reuben Levy. Costa Mesa, CA: Mazda Publishers with Bibliotheca
Persica, 1996.
Yeats, William Butler. The Collected Plays of W. B. Yeats. New ed. New York:
Macmillan, 1965.
Young, Kay and Jeffrey L. Saver. “The Neurology of Narrative.” SubStance:
A Review of Theory and Literary Criticism 94/95 (2001).
Yuasa, Nobuyuki, ed. and trans. Bashō: The Narrow Road to the Deep North and
Other Travel Sketches. Baltimore, MD: Penguin, 1966.
Zeami Motokiyo. “Atsumori.” Trans. Arthur Waley. In Masterpieces of the
Orient. Expanded ed. Ed. G. L. Anderson. New York: Norton, 1977.
Zhang Longxi. Mighty Opposites: From Dichotomies to Differences in the Com-
parative Study of China. Stanford, CA: Stanford University Press, 1998.
Index
Abanaki tradition 165 55, 68, 69, 80, 81, 93, 128,
Abduction 102, 108, 114, 116, 132, 168, 261–3
180–2, 231, 232 Affection 70, 72, 73, 97, 105, 124,
Abelson, Robert P. 204 166, 244, 254, 255, 258–62
Abhijñānaśākuntalam 21, 105, 167, Agency 23, 76, 77, 91, 186, 205,
173, 220, 234, 235 210, 212, 214, 217, 219,
Abhinavagupta xi, 13, 39, 45, 221, 227, 238, 254, 255,
46, 49–54, 57, 62, 66, 78, 258, 259
262 Aggression 85, 162, 260
Abraham (Ibrāhı̄m) 21 Ahim . sā 137
Abrams, M. H. 220 Ahmad, Aijaz 16
Absolutization 104, 106, 110, Ainu epic 14, 15, 21, 28, 172,
170, 214, 224 175–83, 186–90, 196–201,
Abstraction 26–8, 32, 37, 44, 51, 222, 225
61, 96, 97 Akam poems 161
Abū Lahab 116 Aladdin (‘Alā’ al-dı̄n) 25, 108
Abu-Lughod, Lila 110 Alam . kāra 46
Abundance 191, 193–7, 226 Algorithms 59, 63, 244
Achillēs 123, 130, 149, 216 Al-H . asan 115
Adam 15, 196, 210, 217, 231 Al-H . usayn 115
Adaptation 263 ‘Alı̄ 115
Ad hoc categories 55, 59, 183 Allegory 129, 147, 191
Aeneid 38, 123, 134 Alliteration 21, 22, 27, 28, 32, 33,
Aeschylus 104 35–7, 46, 56, 158
Aesthetic experience 18, 23, 30, Allusion 154
36, 39, 41, 43, 47, 48, 51–3, Amaladass, Anand 48
285
286 Index
Amin, Samir 16 Atwood, Margaret 246

Amnesty International 190 Austen, Jane 174, 179
Ānandavardhana xi, 45, 46, 48, Avoidance 255, 258
50–2, 54 Awe 169, 213
Andromachē 124, 128 Awolalu, J. Omosade 217
Anger 2, 47, 51, 63, 76, 77, 84, 85, Axelrad, Elise 67, 68
89, 90, 93, 95, 219, 241, Aztec tradition 191, 193
242, 248, 254–61, 263
Anorexia 200 Babbling 261, 263
Anscombe, G. E. M. 131 Babylonian tradition 21, 38
Antagonist 131, 146, 147, 177, Badawi, M. M. 23
211, 216, 217, 232, 234, 237 Baddeley, Alan 39
‘Antar 115, 116 Baker, Sheridan 152
Anubhāva 82 Bal, Mieke 205
Anxiety 67, 95, 187 Bambara tradition 113, 165
Appiah, Kwame Anthony 9, 16 Bambi 259
Arabic tradition 24, 107, 108, Banality 23
114–16, 129, 161, 162 Barnard, Philip J. 43
Ārambha 78 Barsalou, Lawrence W. 58, 59,
Arawak tradition 109 183
Arberry, A. J. 162 Bartlett, F. C. 67
Archilochos 161 Bashō, Matsuo 160, 167, 168
Architecture, cognitive 43, 58, Basicness 184
59, 61 Basotho poetry 21, 26, 38
Areal distinctness 17–19, 26, 44, Bateson, Mary Catherine 33, 36
174, 176 Beauty 48, 53, 160, 262, 263
Ārēs 191 Bedouin poetry 110, 234
Aristotle 13, 24, 77, 78, 89, 102, Bekerian, D. A. 65, 66, 95
132, 133, 144, 212, 218, 220 Bem, Daryl J. 240
Arjuna 125 Beowulf 118
Arnhem Land 26 Berggren, Paula 164
Artha 93–5, 97, 128 Berndt, Ronald M. 26, 38, 165
Arthaśāstra 94, 137 Bernheimer, Charles 8
Asceticism 200 Bhakti 129, 170, 174
Ashcroft, Bill 9 Bharatamuni 24, 47, 77
Assonance 21, 32–5, 37, 158 Bhāsa 25, 102, 105, 135, 157, 208
Aston, W. G. 168 Bhāva 47, 76, 81, 82, 93
Astyanax 94, 124, 194 Bhavabhūti 25, 105, 126, 135, 156
Atonement 237 Bible 155
Atsumori 127, 128, 130, 134, 135, Biebuyck, Daniel 138
138, 149, 237 Bierhorst, John 191–4
Attention 36, 37, 39, 64, 65, 75 Bihārı̄ 161
Index 287
Bilhan.a 161 Ceremonies 105, 123, 128, 165,

Blair, R. J. R. 143, 144 225, 236
Blake, William 38 Chadwick, Nora K. 116, 195
Blending Chang Wen-ch’ien 163
see Conceptual blending Character 19, 24, 29, 47, 53, 64,
Bordwell, David 7, 8 69, 74, 75, 87, 88, 92, 93,
Boredom 96, 261, 262 109, 117, 119, 133, 144–6,
Bororo 109 149–51, 157–60, 166, 169,
Boulton, Maureen 158 176, 179, 180, 182, 205–17,
Bower, G. H. 63, 64 220, 225, 228, 231–3, 237,
Bownas, Geoffrey 21, 164 246
Brandon, James R. 25, 102, 107, objectification of 145, 209
113, 191 subjectivity in 145–9, 209
Brenan, Gerald 170 Character types 23, 24, 87
Brewer, William F. 7, 28, 29, 87, empathic 212–14, 216
203 individualistic 205, 207–17,
Brewitt-Taylor, C. H. 21, 159 231–3, 238
Broadway musicals 157 socially normative 206–12,
Brothers, Leslie 256 215–17, 231, 237
Brower, Robert H. 164 socially representative 205,
Brown, Donald 10 210, 211, 232, 233
Bunraku 104 Charitōn 25, 102, 108
Businessman 121 Chaucer, Geoffrey 38
By-products, evolutionary Chelkowski, Peter 115
263 Chêng Teh-hui 25, 102, 104
Chikamatsu Monzaemon 24, 94,
Cadmus 191 107, 113, 134, 208, 236
Cao Xueqin 159 Chinese tradition 2, 12, 13, 21,
Capitalism 198, 249 23, 25, 26, 33, 38, 42, 43,
Capital punishment 189 98, 105, 106, 114, 116, 117,
Care 136, 137 154, 158, 159, 163, 169,
Carroll, Joseph 8 170, 191, 208, 211, 236
Carroll, Noël 64, 65 Chippewa 165
Catharsis 189 Chomsky, Noam 16, 22, 23, 28,
Causality 5, 23, 33, 77, 91, 138, 31, 45, 203, 204, 252, 264
153, 189, 205, 217–21, 224, Ch’u tradition 170
227–9, 238 Cilappatikāram 111
efficient 218, 229 Circularity
final 218–21, 229, 238 see Plot circularity
general 218, 219 Cleopatra 123
restricted 219, 221 Clinton, Jerome 116
see also Design Clore, Gerald L. 79, 227, 228
288 Index
Cohen, Joshua 249 Dalton, E. T. 235

Collins, Allan 227, 228 Da Monzon de Segou 113, 114
Colonialism 8, 9, 19, 23, 51, 118, Danger 16, 70, 111, 114, 116, 119,
121 139, 142, 177, 238, 246, 256
Color terms 98 Danly, Robert Lyons 160
Comedy 11, 13, 23–5, 78, 90, 98, Dārāb 21
101–4, 107–9, 110, 112, Davidson, Donald 3, 10
113, 116, 121, 133, 134, Davis, Keith E. 257
155, 205, 207, 208, 217, Davis, Mark H. 81, 140, 145, 146
225, 227, 234, 235 Dawood, N. J. 25, 108
Commodification 250, 251 Dawson, Michael R. W. 63
Compassion 106, 124, 136–50, Death 5, 24, 25, 52, 53, 62, 71–4,
199, 206 83, 84, 86, 90, 91, 94–6,
Comrie, Bernard 19, 28, 29, 101–11, 113–17, 119, 124,
98 126, 127, 130, 131, 138,
Conceptual blending 7 144, 149, 157, 161, 163,
Conchubar 181 165, 174, 177, 180, 182,
Concretization 58, 66, 84, 86 189–93, 196–9, 210, 211,
Congolese poetry 126 215, 218, 220, 223, 224,
Connectionism 58–60 228, 231–3, 236, 237
Connotation 33, 50 Decay 255, 258
Consciousness 13, 35, 36, 52, 53, Deconstruction 110
56, 57, 62, 81, 89, 141 Dede Korkut, The Book of 116, 127,
Contemplation 262, 263 145, 158, 225
Conway, Martin A. 65, 66, 95 Defaults 20, 22, 58, 60, 79, 86,
Cooper, Arthur 21, 33, 38, 163 129, 140, 184
Cosmides, Leda 4, 263 Defense 112, 115, 116, 137, 138,
Crane, Mary 4 147, 150, 199, 213
Croft, William 28, 29 Dehumanization 177, 214
Cronyn, George W. 165, 194 Deification 214, 253
Crucifixion 226 Deirdre 181
Cuceb 191, 192 Demonization 108, 110, 111, 114,
Cues and constraints 20, 43 126, 177, 198, 212–17, 221
Cullmann, Oscar 211 Deng, Francis Mading 38, 165,
Culture 1, 2, 7–10, 14–16, 18–21, 166, 194
23–6, 28, 29, 36, 37, 59, 79, Design 218, 220, 221
84, 98, 102, 104, 109, 110, De Sousa, Ronald 82, 83
113, 124, 133, 170, 175–7, Despair 73, 110, 119, 124, 125,
183, 184, 188, 190, 191, 128, 137, 168, 170, 221–3,
194, 212, 213, 219, 244–7, 225
249–52, 256, 257, 259, 260, Dhanam . jaya 77
263, 264 Dharma 93, 128, 131, 226
Index 289
Dhvani 24, 45, 46, 48–52, 55, 57, Egocentrism 53, 70, 71
62, 66, 71, 72, 74 Egudu, Romanus 21
see also Rasadhvani Egyptian literature 162, 169
Dı̄dō 123, 134 Ekman, Paul 76–80, 93, 256
Dieting 200 Emotion
Dimock, Edward C., Jr. 127, 156 actional outcome of 84, 85,
Dinka poetry 38, 165, 166, 194 88, 96, 239, 240, 242, 249,
Discourse 7, 87, 100, 121, 177 258–63
Disgust 76, 77, 79, 81, 89, 90, 255, and ambient conditions
256, 258, 260, 262, 263 260–3
Dispositional objects 253–4, 256, basic 53, 76–81, 89, 252, 254
258, 260, 262, 263 categorization of 247–8
Dissanayake, Ellen 8 cause of 1, 46, 62, 69, 70, 82,
Dixon, R. M. W. 41 84, 85, 92, 124, 218, 241,
Dobosz, Barbara 247, 248 247, 248
Dobyns, Stephen 153 clustering of 248
Dolby, William 25 and communication 80–2,
Domination 9, 96, 97, 111, 120, 89, 242
122, 123, 128, 134, 141, componential approach 253
150, 151, 155, 156, 170, eliciting conditions for
178, 182, 184, 196, 222–5, 82–5, 88–90, 92–7, 102,
228, 259–61 103, 110, 118, 129, 131–3,
Doniger, Wendy 8 147, 187, 215, 221–3,
Drama 2, 9, 19, 23–5, 47, 76, 77, 239–51, 254–63
88, 102–7, 111, 113, 115, expressive outcomes of
127, 128, 133, 152, 156, 16, 76, 77, 79, 80, 82, 83,
157, 162, 164, 167, 190, 88, 91, 96, 240, 241, 243,
191, 202, 211, 217, 220, 234 247, 249, 251, 252, 256,
Draupadı̄ 125 257
Duckitt, J. H. 141 factual versus substantial
Dundas, Paul 200 131–2
Duryodhana 135, 208 labeling of 240–2
Dus.yanta 105, 224 malleability of 241–52, 263
Duwell, Martin 41 object of 69–71, 78, 226, 227,
Dyirbal tradition 41, 42 231, 235, 237, 238, 248,
253–62
Eat, desire to 255–8, 262 phenomenological tone of
Ecology 199 81, 82, 85, 87, 89, 90, 109,
Eden 196 239, 240, 242, 243, 247–9,
Edible object 255, 258 251, 263
Edmonson, Munro S. 18 source of 46, 62, 67, 69, 70,
Edo kabuki 113 75, 189, 242
290 Index
Emotion prototypes 15, 65, Epilogue of suffering 123, 125–9,

84–98, 133, 171, 202, 204, 133, 137, 138, 142, 145,
228, 238, 239, 242, 250–2, 148–51, 155, 163, 177, 179,
260 181, 190, 193, 208, 216,
personal 94–8, 102, 103, 121, 222, 225, 227, 228, 233,
122, 129, 182–6, 198, 199, 237, 251
225, 226, 228, 244, 258 Epithets 28, 43, 162
physical 183, 186, 195, 200, Erōtici Graeci 25
226, 228, 258 Ès.ù 217
social 94–8, 101, 104, 110–13, Ethical conflict 133–9, 150, 151,
121–5, 128–34, 139, 141, 198–200
150, 151, 166, 178, 182–6, Ethics 12, 85, 93, 122, 123,
196, 226, 228, 230, 246, 259 126–42, 145–7, 150, 151,
transcendental 125–31 187, 189, 198–200, 210–13,
Emotion terms 2, 79, 82, 92, 95, 215, 220, 227, 252
96, 256 of care 136, 137
Emotion types of compassion 136–47, 199
junctural 91–3, 102, 103, of defense 136–47, 199, 213
169, 246, 255, 264 of justice 136, 137
outcome 91–4, 102, 103, 118, of self-abnegation 199, 200
129, 153, 223, 224, 243, Ethnocentrism 257
247, 255 Eucharist 226, 236
sustaining 93, 94, 102, 118, Eudaimonia 222–4, 244
136, 186, 221, 243, 246, Eulogy 34, 158, 162, 165, 166,
255, 264 169, 170
Empathy 54, 67, 70, 71, 75, 80–2, Euripidēs 94, 128, 148, 190,
88–91, 96, 122, 123, 198
139–50, 177, 205, 206, 208, Eurocentrism 9
209, 212–17, 226, 251, 252, Eve 15, 196, 210, 217, 231
262 Evil 113, 116, 144, 151, 170, 177,
categorial 140, 141, 148, 150, 182, 210, 212
208 Evolution xi, 246, 249, 252, 253,
situational 140–50, 205–9, 259, 263
212–16 Exempla 57, 59, 60–2, 85, 133,
see also Identification 184, 202, 245
Emplotment 5, 116, 117, 153 Exemplars 59–61, 83, 85
Encoding 35–9 Exile 11, 15, 94, 95, 101, 104–20,
Epic 6, 14, 87, 88, 94, 111–16, 125, 128, 143, 154, 155,
120–3, 126, 127, 130, 138, 157, 160, 163–5, 168, 170,
152, 156, 162, 172, 173, 173, 177, 178, 180, 181,
175–81, 186, 192, 200–2, 189, 190, 192, 193, 196–8,
215, 224, 236 228–37
Index 291
Explanatory adequacy 31, 32, 43, Funeral 86, 123, 128, 165, 195
44 Fung, Heidi 67
Fusion
Facial expression 76–80, 91, 252, of horizons 140, 142, 144
256 of narrative structures 228
Fa-Digi Sisòkò 94, 215
Fall, the 21, 196 Gadamer, Hans-Georg 140, 142,
Familializing conflict 148, 149, 144
227 Gao E. 159
Famine 181, 185–200, 215, 219, Gardner, Howard 4
225, 228 Gardner, John 138
Fanon, Frantz 16 Garman, Michael 39, 56, 57
Fasting 200 Gathercole, Susan 39, 42, 43
Fear 63–5, 70, 73, 76, 79, 81, 89, Gelpke, Rudolf 108
90, 93, 95, 96, 115, 120, Gender 137, 176, 179, 180, 182,
139, 142–4, 218, 229, 240, 251
243, 246, 249, 254–60, 263 Gendler, Tamar 143
Ferdowsi, Abolqasem (Firdausi) Generalization, social 225, 226,
21, 114, 207, 208 231, 234, 237, 238
Feudal societies 29, 113, 198, 208 Genetic distinctness 17–19, 26,
Firth, Raymond 21, 26, 165, 194 32, 44, 122, 174–6
Fischer, E. F. 98 Genji 112, 160
Flaubert, Gustave 245 Genre 2, 13, 15, 19, 21, 23, 59, 65,
Flight 218, 246, 249, 254, 255, 258 75, 87, 104, 107, 109, 115,
Fodor, Jerry A. 60 118–20, 134, 136, 147, 151,
Food 86, 166, 182, 187–200, 215, 153, 154, 161, 162, 166,
225–7, 231–3, 244, 245, 168–70, 176, 185, 199,
256, 258, 260, 261, 263 207–9, 230, 231, 259
Foreshadowing 21, 33, 34, 35, 37 Gergen, Kenneth J. 257
Forgas, Joseph Paul 63, 64, 247, Gerow, Edwin 46, 156
248 Gerstle, C. Andrew 104
Formulae 28, 43 Ghazal 129, 134, 162, 170, 217
Fowler, Barbara Hughes 161 Gibbs, Raymond W. 33
Freedman, J. L. 95 Gifford, Don 120
Freud, Sigmund 1 Gilgamesh 126, 130, 131, 138, 149,
Friendship 247, 248 162, 224, 237
Fright 52, 73, 76, 77, 79, 189, 246 Gilligan, Carol 136, 137
Frijda, Nico H. 103, 223 Girard, René 189, 190
Frustration 170, 248, 254, 257 Gnoli, Raniero 39
Frye, Northrop 13, 24, 134, 152, Goals 65, 76, 77, 88, 89, 91–4,
205, 207 102, 103, 111, 113, 116,
Fuhrman, Robert W. 61 122, 125, 129–33, 140, 178,
292 Index
Goals (cont.) Harris, Paul 187

182, 185, 189, 203–6, 214, Hars.adeva 105
217–24, 227, 228, 244–6, Hart, George 28, 161
249, 254, 255, 257–9, 262 Hate 34, 79, 95, 254, 258
short-term 222–4, 243 Hauser, Arnold 29
transcendental 182, 224 Hawai’ian tradition 38
Goethe, Johann Wolfgang von Hawkes, David 170
104 Healy, Alice 43
Gould, Stephen Jay 252 Heath, Peter 116
Gratitude 162, 254, 258 Hebrew tradition 21, 169
Greimas, A. J. 214, 226 Hector 123, 124, 128, 130, 149,
Grice, Paul 48 216
Grief 71, 95, 108, 124, 125, 149 Hecuba 94
Griffiths, Gareth 9 Hegarty, Mary 43
Grünbaum, Adolf 220 Heike, Tale of the 112, 113, 124,
Guépin, J. P. 191 125, 127, 128, 130, 134,
Guilt 128, 132, 149, 157, 192, 210, 135, 147, 149
216 Herman, David 204
Guinean tradition 165 Hernadi, Paul 4, 13, 116
Gurgānı̄, Fakhr ud-Din 115, 157, Hijrah 116
158, 235 Hinduism 51, 129, 137, 141, 170,
200, 224
Habituation 22, 186, 223, 224 Historiography 116, 117, 195
Hadas, Moses 148 Hitomaro 164
Haiku 40–2, 160 Hitopadeśa 94
Halász, László 66, 67, 143 Hittite tradition 138
Halford, Aubrey S. 191, 217, 235 Hjort, Mette 79
Halford, Giovanna M. 191, 217, Hobbs, Jerry 4
235 Holland, John 35, 83, 145, 146
Hamartia 77, 132, 133, 138, 144 Holland, Norman xii, 4, 66
Hamlet 46, 69–74, 118, 120, 179 Hollywood musicals 157
Hampton, James A. 59 Holyoak, Keith 35
Happiness 11, 15, 47, 52, 63, 65, Home 95, 108, 111, 113, 119, 120,
76, 90–8, 101–3, 105, 106, 127, 129, 154–7, 163, 167,
108, 110, 111, 121, 122, 177, 178, 180, 181, 198,
125, 127–33, 137, 141, 150, 229–31, 235, 237, 245, 249,
151, 166, 182–7, 195, 200, 250
221–30, 234, 238, 243, 244, Homer 124
246, 250, 254, 255, 257–9, Homosexuality 20
262, 264 Hoogstra, Lisa 67
Hardman, Charlotte 187 Hopi tradition 94
Harré, R. 257 Horror films 65, 261
Index 293
Hsia, C. T. 105, 106, 116 Influence, literary 17–19

Humanization 150, 177, 182, 206, Ingalls, Daniel H. H. 78
226, 251 Ingarden, Roman 66
Humayuni, Sadeq 115 In-group 97, 111, 121, 141, 144,
Humbaba 126, 138, 149 170, 176, 221, 224, 259–61
Hunger 166, 186, 187, 191–3, 196, Inhibition 254, 258
225, 230, 255–9 Innatism 31
Huot, Sylvia 153 Intensification 23, 86, 88, 103,
153, 160, 223, 224, 227–30,
Iblis 114 238, 248
Ibn Ishāq 116 Intentional object 69, 70, 75
Idealization 146, 214, 225–7, 238 Intertwining
Identitification of narrative structures 228
categorial 141, 144, 145, 148, Iranian tradition 114, 115, 207
206, 207, 210, 212, 230, 259 Irele, F. Abiola 162
situational 144, 206, 212, Ise, Tale of 160
214, 216 Iser, Wolfgang 88
see also Empathy Ishmael (Ismā’ı̄l) 21
Ideology 8, 9, 30, 175, 189, 230, Islam 116, 170, 224
239, 250–2, 257
Ijo tradition 113 Jackendoff, Ray 203
Ilangô Adigal 111 Jainism 200
Iliad 123, 124, 128, 135, 181, 216 Jakobson, Roman 7, 32, 34
Imagery, 19, 21, 25–7, 34, 36, 95, Jankowiak, W. R. 98
101–5, 107, 108, 110, 119, Japanese tradition 12, 21, 23, 25,
129, 131, 152, 155, 167–9, 26, 40, 41, 103, 104, 106,
180, 184, 189, 197, 214, 107, 112, 113, 121, 123,
228, 229, 231–3, 236, 237, 124, 157, 160, 164, 167,
261 175, 191, 235
Imagination 1, 15, 66, 69, 73, 74, Jauss, Hans Robert 88
82, 91, 92, 96, 104, 105, Jayadeva 156
113, 136, 140, 144, 147, Jealousy 79, 95, 114, 189, 240, 249
149, 153, 166, 169, 182–4, Jenkins, Lyle 252
221, 224, 225, 227, 246, Jerusalem Delivered 118
247, 250, 252, 261, 263 Jesus 15, 196, 197, 210, 211, 225,
Immanent norms 211, 212, 215 226, 228, 235, 236
Implicature 48 John of the Cross, Saint 170
Indian cinema 157 Johnson, John William 127
Indian tradition 9, 10, 21, 23, 25, Johnson, Mark 26
26, 39, 45, 46, 55, 76–8, 90, Johnson-Laird, Philip N. 76–80,
111, 113, 123, 137, 161, 83, 89, 93
165, 195, 198, 208 Joseph, George 246
294 Index
Joyce, James 67, 120, 121 Kuna tradition 21, 41, 42

Judaism 170 Kunene, Daniel P. 21, 26, 38
Judeo-Christian tradition 21, Kuo yü 117
196, 199, 217, 231 Kurelu tradition 165
Juncture 12, 78, 91–3, 102, 103, Kyoto kabuki 113
152–71, 223, 243, 246, 255,
264 Labouvie-Vief, Gisela 1
Labov, William 91, 229, 230
Kabuki 25, 103, 107, 113, 133, Lakatos, Imre 11, 12, 14, 172, 173,
191, 217 179
Kafka, Franz 198 Lakoff, George 26, 59–61, 84–6,
Kahneman, Daniel 58, 97 184
Kailasapathy, K. 28 Laks.mı̄ 130
Kalhan.a 195 Lal, P. 25
Kālidāsa 21, 56, 105, 167, 220, 234 LaMar, Virginia A. 74
Kallı̄nos 161 Larsen, Steen 66, 143
Kāma 93, 94, 128 László, János 66, 143
Kan’ami 25, 106 Laughlin, Philip xii
Karn.a 130, 149, 208 Laver, Sue 29
Kaut.ilya 137 Lawrence, D. H. 120
Kazin, Alfred 38 Laylā 24, 104, 107, 108, 147, 217,
Keene, Donald 24, 41, 106, 113, 220, 224, 234
157, 160, 164 LeDoux, Joseph 243, 256
King, Bruce 10 Lerdahl, Fred 203
King Lear 11, 118 Lesbianism 20
Ki no Tsurayuki 164 Levinson, Jerrold 69
Kintsch, Walter 43 Lévi-Strauss, Claude 109
Kiparsky, Paul 3, 7, 18, 21, 22, 27, Levy, Reuben 162
32, 34–6, 45 Lewis, Geoffrey 158
Kirson, Donald 66, 96 Lexical context 131
Klein, Stanley 61 Lexical entry 22, 55–7, 60, 61,
Klemenz-Belgardt, Edith 29 81–3, 95, 118, 129, 140,
Knight, W. F. Jackson 123 141
Knox, Bernard 190 Lexical identification 89
Kokinshū 157, 160, 164 Lexical item 43, 55, 57, 239, 256
Kondh poetry 195, 235 Lexical links 63
Korn, Alfons L. 38 Lexical networks 57, 63, 103
Kouyaté, Djeli Mamoudou 127, Lexical processes
226, 228 access 22, 54–7, 61, 68, 81,
Kövecses, Zoltán 10, 84–6 82, 103
Ks.atriyadharma 137 activation 52, 53, 57, 59,
Kuan Han-ch’ing 191, 211 63–5, 67, 141–3, 147
Index 295
cross-indexing 22, 55, 56 Macbeth 118, 132

priming 56, 57, 61–5, 67–9, Ma Chih-yüan 104
71–5, 89, 104, 121, 141–3, Mack, Maynard 154, 163, 169
206, 227, 235, 237, 246 Mahābhārata 111, 125, 130, 135,
reordering 56 149, 208, 237
Lexical structure 54, 61, 140, 183, Mahādēviyakka 174
239, 240 Mahākāvya 156
Lexicon 22, 46, 55, 56, 60, 87, 98, Mahapatra, Sitakant 195, 235
103, 129 Maier, John 138
Li Ch’ing-chao 163 Majnūn 24, 104, 107, 108, 147,
Lichtenstadter, Ilse 115, 161 220, 224, 234
Lichtenstein, Edward H. 7, 87 Malian tradition 94, 111, 127, 215
Lincoln, Kenneth 194 Mapping 66, 141, 142, 144
Ling Ching 163 Manohra 25, 102
Li Po (Li Bo) 163, 167 Marcus, Ruth Barcan 151
Literary universals website 8 Markandaya, Kamala 198, 236
Liu, James J. Y. 163, 164 Marriage 86, 95, 102, 105,
Liu Hsieh 21 107–10, 115, 164, 180–2,
Liu Jung-en 21, 25 191, 209, 222, 223, 225,
Locke, John L. 261 226, 230, 232, 234, 236,
Loftus, Judith 61 244, 247, 250, 251, 259
Lord, Albert 28, 40 Marthiya 161
Love 11, 20, 23–6, 34, 47, 48, 53, Martinovitch, Nicholas N. 23
62–7, 71, 73, 79, 83, 84, 86, Martyrdom 136, 200
87, 90, 93–8, 101–10, 115, Marx, Karl 9, 16, 29, 208
119, 120, 124, 129, 130, Massey, Reginald 77
133, 134, 136, 155, 156, Masson, Jeffrey 161
158–70, 173, 179, 180, 181, Mathews, A. 64
182, 200, 205, 206, 208, Maximal relevance
209, 217, 220, 222, 225, see Patterning
227, 228, 230–2, 234–6, Maximization of patterning
240, 243, 244, 245, 247–51, see Patterning
255, 259, 261 Māyā 51, 192
triangle 179–82, 200 Mayan tradition 191
Loyalty 113, 117, 199, 208, McClelland, J. 58
216 McLean, Mervyn 21, 26, 165, 194
Lukács, Georg 146 McLeod, Peter 58
Luo Guanzhong 159 Memory 3, 20, 29, 35, 39–44,
Lutz, Catherine A. 110, 251 51–7, 62–75, 82, 89, 91,
Lyric poetry 2, 5, 12, 152–71, 174, 141–5, 206, 214, 229, 243,
194, 195, 226 246
devotional 169–71, 174 buffers 57, 62
296 Index
Memory (cont.) Moral conflict 133–9, 145, 147,

long term 54, 56, 57, 62 149–51, 199, 227, 252
personal 29, 54, 55, 65, 66, Moreh, Shmuel 23
68, 71, 73, 74, 89, 141, 143, Morrison, George 115, 157
206, 214, 230 Moses (Mūsā) 21
perspectival center in 141, Muh.ammad 116
142, 144 Murnau, F. W. 210
rehearsal 39–43, 56, 57, 62, Murphy, Roland E. 155
243 Muslim tradition 21, 115, 116,
working 39, 43, 44 162, 170, 200, 217
Meriah 235 Mwindo Epic 112, 126, 130, 138,
Merwin, W. S. 161 139, 148, 224, 236, 237
Meta-ethical issues 136
Micronarratives 65, 88, 118, 243 Nahuatl tradition 191
Military Romance, Chinese 116, Naim, C. M. 150
117 Namiki Sōsuke 135
Miller, Barbara Stoler 21 Naoise 181
Miller, Dale T. 58, 97 Nārada 134
Miller, George 36 Narayan, R. K 127
Miller, Peggy J. 67, 68 Nārāyan.a 94
Milton, John 196, 202, 208, 217 Narrative Hypothesis 152–7,
Mind 160, 166, 169, 171, 194
objectal view/reification of Narratology 203
145, 146, 151, 209 Narukami and the God Fudo 191
qualia in 240 Nası̄b 162
subjective view of 145–9, Nāt.ya Śāstra 24, 47, 76, 77
209, 212, 240 Navaho tradition 194
Miner, Earl 164 Necessary and sufficient
Mintz, Judith 67 conditions 30, 42, 82–5,
Mı̄rābāı̄ 174 87, 89, 92, 98, 102, 133,
Mirth 2, 76, 90, 249 136, 166, 186, 204
Miyake, Akira 43 Nelson, Katherine 204, 205
Modeling Nepal 187
of narrative structures 21, New Comedy 13, 24, 108, 133,
120, 123 134, 205, 207, 208, 217, 234
Models 14, 21, 35, 55, 59–61, 84, Ngũgı̃ wa Thiong’o 9, 10, 16
87, 120, 123, 129, 204, 224, Niebelungenlied 118
225, 252 Night Chant 194, 195
Modernism 2, 118, 120, 121, 198 Nisbett, Richard E. 35, 61, 241,
Moks.a 54, 93, 128, 224 242
Monier-Williams, Monier 78 Niyatapti 78
Monogatari 160 Niz.āmı̄ 24, 104, 108, 147, 217, 220
Mood 59, 63, 64, 66, 181 Nō 25, 104, 106, 127, 157
Index 297
Nondispositional objects 253, Pandit, Lalita 9, 73

255, 258, 260, 262, 263 Pān.ini 45
Nontechnical correlations 20, 25, Panksepp, Jaak 253, 254, 256,
44 259, 260
Nosferatu 210 Paradigms 5, 6, 20, 82, 94, 95,
Nouveau roman 118, 120 103, 115, 118, 120, 122,
Nwoga, Donatus 21 127, 135, 146, 148, 154,
Nyanga tradition 112, 126, 130 160, 191, 211, 213, 216,
238, 244, 245
Oatley, Keith xii, 4, 46, 51, 63, 66, Parallelism 21, 32, 123
67, 71, 76–83, 89, 90, 93–5, Parrott, W. G. 257
204, 243, 259 Particularism 8, 9, 10, 16, 110
Obedience 199 Pathos 47, 50, 72, 124, 164, 215
Obligatory techniques 21, 22 Patriarchy 9, 249–51
Obtrusiveness 36, 37, 40 Patterning, maximization of
O’Connor, Cary 66, 96 33–40, 43, 44, 63
O’Connor-Visser, E. A. M. E. 191 see also Relevance
Odyssey 38, 120, 236 Pavel, Thomas G. 203, 204
Oedipus Complex 181 Penelope 120
Oedipus plays 104, 128, 190, 211, Pennington, Bruce 43
228, 237 Perkins, George 152
Ohnuki-Tierney, Emiko 178 Persian tradition 24, 107, 108,
Oiemono plays 113 114, 115, 129, 157, 162, 217
Ōkabou Ojobolo 113 Personalization 67, 68, 97, 209,
Olthof, Tjeert 187 212, 219, 221, 228, 229,
Omens 21 238, 262
Ong, Walter J., S.J. 28, 40, 262 Philippi, Donald L. 21, 28, 175–8
Opland, Jeff 28 Pindar 161
Optional techniques 21, 22 Pinker, Steven 4, 8, 254, 262
Orality 12, 14, 20, 26, 28, 40, 43, Pity 144, 182
98, 162, 174–6, 194 Plague, The 199
Orature 28, 175, 188 Plantinga, Carl 8
Oresteia 104, 128 Plot circularity 21, 33, 35, 37, 218
Orians, Gordon 262 Plunkett, Kim 58
Ortony, Andrew 58, 79, 227, 228 Poetic line 20, 37–44
Ostracism 228 Poetics 45, 46, 48, 77, 79, 168
Out-group 141, 144 Poetry 3, 5, 6, 12, 20–3, 26–8,
Owen, Stephen 170 32–42, 45, 46, 48, 110, 118,
Ozidi Saga 113 120, 121, 127, 147, 152–71,
174–82, 185, 188, 194, 195,
Padgett, Ron 153 202, 218, 222, 225, 226
Pagels, Elaine 211 Point of view 112, 139, 140, 176
Pandit, B. N. 51 Polynesian tradition 26, 165
298 Index
Popper, Karl 11 Rāmāyan.a 21, 24, 102, 104, 111,

Postmodernism 2, 99, 120, 126, 127, 130, 132, 134,
121 135, 162, 170, 180, 208,
Praise poem 226, 228
see Eulogy Rasa 46, 47, 50, 51, 53, 54, 62, 65,
Prefocus 64, 65 67, 74, 79, 81
Priam 123, 124, 130, 149 Rasadhvani 50–2, 57, 62–4, 71, 74
Priming Rati 93, 94
see Lexical processes Rāvān.a 111, 126, 170, 236
Prince, Gerald 203 Realism 29
Probes 142, 144 Reception of literature 39, 74, 92,
Project 223–5 185
Protection 254, 258 Relevance 33–5, 37, 39, 40, 62, 63
Proto-emotions 253–64 see also Patterning
Prototypes 6, 11, 12, 15, 30, 42, Religion 30, 54, 105, 120, 128,
57–62, 65, 66, 80–98, 129, 131, 134, 136, 141,
100–5, 110, 111, 113, 114, 144, 152, 158, 162, 169,
116–22, 128–33, 136, 137, 170, 171, 179, 186, 193,
139–41, 147, 150–5, 160, 197, 198, 207, 208, 224,
166, 169–71, 178, 181–6, 234, 262
188, 191, 194–6, 198–200, Remindings 91, 143
202–6, 208, 209, 212, 214, Remorse 125, 128, 132, 138, 149,
215, 218, 219, 221–4, 150, 179, 193, 213, 214,
226–31, 236–40, 242–8, 222, 232, 233
250–2, 255, 257–61, 264 Reparation 131, 138, 145, 150,
Psalms, Book of 155, 169 189, 191, 193, 199
Pseudouniversals 9 Repentance 181
Psychoanalysis 1, 109 Research programs 7, 8, 10–16,
Puenzo, Luis 199 30–2, 37, 41–3, 44, 54, 79,
Pukui, Mary Kawena 38 95, 122, 152, 172, 173, 185,
Puram poems 161 200, 201, 238
Pursuits 222–4, 254, 258 Resurrection 24, 102, 197, 228,
232
Qas.ı̄dah 161, 162 Revenge 128, 177–9
Quetzalcoatl 191–3 Rhyme 21, 22, 33, 36, 43, 56–8,
Quint, David 123 163
Richardson, Alan 4, 8
Racine, Jean 38 Richman, Paula 111, 135
Racism 9, 16, 234 Ritual 23, 86, 128, 181, 187–9,
Rāma 111, 126, 130, 132–5, 139, 192–7, 232, 235–7
140, 156, 157, 180, 208, Rival 101, 107–9, 119, 120,
224–6, 236, 237 178–82, 191, 216, 217, 231,
Ramanujan, A. K. 156 232, 234, 235, 237
Index 299
Roadarmel, Gordon 156 Satanism 208, 213, 217

Robbe-Grillet, Alain 99, 120 Satı̄ 195
Rogers, Joel 249 Saver, Jeffrey L. 4
Rolls, Edmund T. 58 Savoring 39, 53
Romanticism 2, 208 Scapegoat 190, 195, 210, 211, 235,
Romantic union 93, 94, 103, 106, 251
108, 121, 122, 128, 129, Schachter, Stanley 241, 242
131, 134, 157, 182, 222–5, Schacter, Daniel L. 62, 142, 144
244, 245, 250 Schank, Roger C. 86, 91, 204
Romeo and Juliet 71, 103, 104, 108, Schemas 21–3, 25, 28, 35, 36, 54,
109 57, 58, 60–2, 66, 84–6, 129,
Ross, Lee 61, 241, 242 133, 142, 184
Rostám 21, 114, 115, 148, 207, Schwartz, Judith 66, 96
216, 235 Scripts 84, 85, 153, 204
Rubin, David C. 20, 40, 41, 43 Secondary principles 32, 37, 44
Rumelhart, David E. 58, 203, 204 Seidman, Robert J. 120
Rureke, Shé-kárisi 237 Seilman, Uffe 66, 143
Russell, James 85, 86 Self-concept 29, 140, 141
Russell, Ralph 134, 162 Self-denial 199
Self-perception theory 240
Sabi 168 Senex iratus 205
Sādhāran.adharma 137 Sexuality 29, 30, 93, 109, 140,
Said, Edward 16 141, 160, 161, 165, 174,
Śaktibhadra 111 182, 192, 196, 222, 223,
Sakuemon 107 244, 248, 249, 255–9,
Śakuntalā 105, 156, 167, 224 262
Sallekhana 200 Shah, Priti 43
Salthouse, Timothy 43 Shâhnâme 114, 117, 134, 207, 216
Samurai 107, 121 Shakespeare, William 7, 11, 25,
Sandars, N. K. 21, 38, 138, 149 71, 72, 74, 102, 108, 109,
Sanskrit tradition 2, 13, 21, 24, 118, 134, 225
28, 45–9, 64, 71, 76–83, 89, Shanks, David R. 61
90, 93, 94, 102, 105, 111, Shaver, Phillip 66, 80, 96, 97
128, 129, 131, 156, 161, Sherzer, Joel 21, 28, 41
167, 204, 208, 225, 228, Shı̄’ah 115
256, 262 Shibatani, Masayoshi 175
Śāntarasa 54 Shilappadikaram (Cilappatikāram)
Sapphō 161 111
Sarbin, Theodore R. 153 Shirane, Haruo 167, 168
Sartre, Jean-Paul 200 Shūishū 157
Sarup, Lakshman 25, 102 Sia tradition 194
Satan 114, 208, 217, 228, 231, Sidney, Philip 156
235 Sindbād 25, 108
300 Index
Singha, Rina 77 211, 213, 214, 216, 217,

Sioux tradition 194 222, 225, 227, 228, 230,
Sı̄rat ‘Antar 115, 116 231, 233, 235, 237, 251
Sissoko Kabine 114 Sufis 170
Sı̄tā 104, 126, 130, 133–5, 156, Suggestion
180, 225, 236 see Dhvani, Rasadhvani
Sivan, Santosh 198 Suicide 24, 102, 103, 106, 107,
Social constructionism 257 123–6, 130, 237
Sohráb 21, 114, 115, 134, 148, 216, Sundiata
235 see Son-Jara
Sommers, Christina 149 Surprise 76, 77, 263
Song of Roland 118 Sympathy 109, 123, 134, 135,
Son-Jara, Epic of 94, 111, 127, 162, 147, 211, 212
173, 215, 226, 228
Sophoclēs 104, 128, 190 Ta’ziyeh 115
Sorrow 2, 47, 52, 53, 62, 71, 72, Tamil poetry 28, 111, 161
74, 76, 83, 86–98, 102–6, Tan, Ed S. 7
110, 115, 121, 123, 124, Tasso, Torquato 118
133, 149, 156, 157, 163, Techniques, repertoire of 20–2,
186, 200, 222, 228, 234, 25–7, 32, 35, 37, 44, 148,
238, 254, 255 158
Sōsuke 135 Telemachus 120
Soyinka, Wole 10, 198 Telic structure 204, 205, 227, 229
Spatiality 228–30 Tellability 88
Spenser, Edmund 156 Telos 224, 225
Sperber, D. 33, 34 Tempest, The 118
Spiritualization 106, 108, 180, Temporality 228–30
224, 225, 227, 230, 237, 238 Terence 134, 234
Spolsky, Ellen 4 Terwogt, Mark 187
Stereotypes 60, 77, 137, 159 Thagard, Paul 35
Stimulation 260, 261 Thai tradition 25, 102
Story grammars 203, 204 Thousand and One Nights, Tales
Story of the Stone 106, 159 from the 108
Stratagem of Interlocking Rings, Threat 24, 67, 104, 108, 110–20,
A 236 136, 147, 173, 177, 178,
Subvocalization 42, 43 181, 185, 213, 231, 236,
Śūdra 135, 139, 140 254, 255, 257, 258, 260
Śūdraka 105 Threat/defense sequence 111,
Suffering 53, 60, 70–2, 105, 106, 112, 114, 177, 178, 231
108, 117, 123–8, 133, 137, Three Kingdoms 117, 158, 159
138, 142, 145, 147–51, 155, Thresholds, attentional 37
163, 177, 179, 181, 189, Tiffin, Helen 9
190, 193, 197–200, 208, Tikopia tradition 21, 26, 165, 194
Index 301
Tiresias 191 Transcendentalization 224,

Titanic 119 225
Todorov, Tzvetan 13 Trask, Willard R. 165
Tomkins, S. S. 243 Trauma of heroism 131, 135,
Tooby, John 4, 263 137–9, 142, 144, 145, 148,
Tosa Diary 160 151
Tradition 6, 9, 11–26, 32, 33, Trobriand Islands 165
36–9, 42, 44, 46, 50, 77, 78, Troubadours, French 153
98–103, 105, 107–9, 118, Tu Fu (Du Fu) 163
122, 127, 134, 154, 156, Turiel, E. 143
158, 161, 165, 168, 170, Turkish tradition 116, 127, 225
174–7, 179, 181, 185, 211, Turner, Frederick 40, 42
217, 245, 251 Turner, Mark 3, 4, 7
Trafton, J. Gregory 61 Tutuola, Amos 188, 198
Tragedy 11, 24, 91, 103, 104, 109, Tversky, Amos 58
112, 115, 119, 121, 163, Tyrtaios 161
211, 216 Tz’u 2, 164
Tragi-comedy 11–15, 19, 20, 24,
65, 95, 98–105, 107–10, Universalism 7–10, 16, 78
112, 113, 116–24, 129, 131, Universals 2, 4, 6–23, 25–32,
132, 134, 135, 137, 145, 37–45, 71, 76, 77, 79, 80,
147, 148, 150, 151, 153, 91, 95, 98–101, 110, 118,
154, 161–3, 166–70, 173, 121, 122, 129, 136, 137,
174, 176, 177, 179, 181, 144, 152, 153, 172–7,
182, 185, 186, 189, 190, 182–6, 190, 198, 200, 204,
198–200, 204, 206–8, 212, 244, 245, 249–53, 256, 257,
224, 226, 228, 232, 233, 259, 260, 263, 264
236, 251, 259, 264 absolute 19, 20, 22, 26–8, 30,
heroic 11–15, 65, 98–104, 43, 100, 153, 186
109–24, 128–40, 145–57, biological 252–63
161–6, 169–73, 176–82, cultural 9, 10, 15, 250–2,
185–92, 195, 199, 200, 202, 258–60, 263
206–16, 221, 224–31, historical 28, 29
234–7, 251, 259, 264 implicational 4, 27, 28, 44
romantic 11–20, 24, 65, 87, indexical 29, 30, 176
95–122, 129–36, 153–74, statistical 19, 20, 27, 32, 38,
178–82, 185–9, 191, 195, 44, 100
198–200, 205–9, 215, 220, typological 28, 43, 44
222–36, 245, 251, 258, 259, Usurpation 11, 110–21, 138,
264 177–9, 182, 191, 192, 199,
sacrificial 15, 154, 181, 208, 209, 227, 230–2, 234,
185–201, 210, 211, 215, 236
219, 220, 225–37, 259, 264 Utsāha 93, 94
302 Index
Vālmı̄ki 126, 132, 134, 135, 162, Wilson, D. 33, 34

226 Wilson, Edward O. 4
van Buitenen, J. A. B. 156 Winter’s Tale, The 25, 102
Varley, H. Paul 127, 147 Wittgenstein, Ludwig 82, 240
Vedāntism 51, 224 Wonder 15, 77, 90, 141, 262, 263
Vibhāva 82 Woodbury, Anthony 28
Victim perspective 128 Woolner, A. C. 25, 102
Victor perspective 128 Wright, Louis B. 74
Villainy 133, 135, 144–7, 209
Virgil 123, 202 Xhosa tradition 28
Vis.n.u 111, 130
Voight, Ellen Bryant 153 Yarshater, Ehsan 114
Vyāsa, Krishna-Dwaipayana Yeats, William Butler 197, 198
125, 149 Yirkalla tradition 26, 38, 165
Yoruba tradition 50, 188, 190,
Walcott, Derek 10, 49, 50 196, 197, 210, 217
Waley, Arthur 25, 106, 164 Young, Kay 4
Wang Wei 155 Yüan drama 104, 211, 236
Wason, P. C. 58 Yuasa, Nobuyuki 167
Watanabe, Hitoshi 178 Yudhis.t.hira 125, 237
Watson, Burton 117
White, Hayden 13, 116 Zeami Motokiyo 24, 106, 107,
Wierzbicka, Anna 82, 83, 86, 92, 127, 157, 237
256, 257 Zhang Longxi 10, 159
Williams, Kimberly 67 Zhirmunsky, Victor 116, 195
Williams, Mark 121 Zuñi tradition 194
Other books in the series (continued from front of book)
Metaphor and Emotion: Language, Culture, and Body in Human Feeling

Zoltán Kövecses
Feeling and Thinking: The Role of Affect in Social Cognition
Joseph P. Forgas
The Causes and Consequences of Feelings
Leonard Berkowitz
Emotions and Beliefs: How Feelings Influence Thoughts
Nico H. Frijda, Antony S. R. Manstead, and Sacha Bem
Identity and Emotion: Development Through Self-Organization
Harke A. Bosma and E. Saskia Kunnen
Speaking from the Heart: Gender and the Social Meaning of Emotion
Stephanie A. Shields
The Hidden Genius of Emotion: Lifespan Transformation of Personality
Carol Magai and Jeannette Haviland-Jones

Colm Hogan The Mind and Its Stories PDF

Uploaded by

Copyright:

Available Formats

Colm Hogan The Mind and Its Stories 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

Colm Hogan The Mind and Its Stories PDF

Uploaded by

Copyright:

Available Formats

What is the main argument of the book?

What is the main argument of the book?

What evidence does the author use to support their claims?

What evidence does the author use to support their claims?

This page intentionally left blank

The Mind and Its Stories

Patrick Colm Hogan is currently a professor in English, Comparative

This series is jointly published by Cambridge University Press

Cette publication est publiée co-édition par Cambridge Univer-

Titles published in the Second Series:

(continued on page following the Index)

PATRICK COLM HOGAN

Cambridge University Press

© Cambridge University Press 2003

This book is in copyright. Subject to statutory exception and to the provision of

Cambridge University Press has no responsibility for the persistence or accuracy of

Introduction: Studying Narrative, Studying Emotion 1

Works Cited 265

An earlier version of Chapter 1 and a small part of the introduction

When empirical researchers in the social sciences consider the nature

given society think about emotions. In other words, we have a body

literature as a human act

questions asked in comparative literary study. Asking these ques-

At one level, this is, then, a study of literary universals. It is also

unfortunate, for it is an area of seemingly obvious value for cognitive

universalism and cultural study

of Philosophical), a number of other writers have taken up the topic.

influential introduction to postcolonial literary study, The Empire

study of culture and history. All reasonable students of literature –

universality and narrative: research programs,

5 Other writers have implicitly presented universalism and particularism as at least

I isolate two universal narrative structures, heroic and romantic tragi-

predictions turn out to be correct, then we have a robust research

of literary theory as well.6 The isolation of heroic and romantic tragi-

precisely parallel to a research program in linguistic universals. Lab-

Moreover, the heroic and romantic plots were clearly prominent in

In undertaking this study, my main hope is that its various descrip-

the structure of a theory of universals

may draw upon in composing a literary work. Most basic techniques

with social expectations and structures, often manifest in a conflict

images seems to be an absolute universal. It is certainly found in the

universals, such as those of psychology, sociology, history, and so on.

It also yields a second, more complex absolute implicational univer-

4 William Brewer’s “The Story Schema: Universal and Culture-Specific Properties”

Indeed, this is a particularly interesting universal for our purposes

universal is indexical because those high-level properties – and thus

explaining literary universals: the nature of a

part of an ongoing, broadly collaborative research program – in this

In the present study, my aim is to isolate universal patterns of

beginnings of syllables) are not used in the same way in ordinary

But this is only a first approximation. The Jakobson/Kiparsky hy-

(which are encodable), this schema yields alliteration. When app-

a saliency of zero.) Various qualities of the feature, the perceiver, and

for achieving descriptive adequacy in this case.7 Thus, in Chinese, a

exceptions. This is, of course, possible. But it is not a solution as we

poem is generating short, repeatable, nonsyntactic units. He/she has

and five syllables. We typically conceive of this as a three-line poem.

Another option is to view the structure of working memory as

9 The importance of explanatory adequacy is not confined to the study of absolute