Untitled

Narrative
‘An excellent critical survey of recent research in narratology and linguis-

tic approaches to narrative . . . Eminently readable, clear, always insightful
and of consistently high standard throughout . . . An invaluable textbook
for all students of narrative.’
Jean Jacques Weber, Journal of Literary Semantics
‘The book’s clarity, comprehensiveness, suggestions for other reading, and
applications make it useful to all those who have some grounding in liter-
ary theory . . . Toolan is very well read in current theory, which is intelli-
gently presented.’
Q. Grigg, Choice
‘A valuable introduction to a wide range of analytical theories and narra-
tive types.’
Cynthia Bernstein, Southern Humanities Review
Narrative explores a range of written, spoken, literary and non-literary
narratives. It shows what systematic attention to language can reveal
about the narratives themselves, their tellers, and those to whom they are
addressed. Topics examined include plot structure, time manipulation,
point of view, oral narratives and children’s stories.
This classic text has been substantially rewritten to incorporate recent
developments in theory and new technologies, and to make it more usable
as a course book. New material includes sections on film, surprise and sus-
pense, and online news stories. The section on children’s narrative has
been updated, and the discussion of newspaper stories incorporates
contemporary examples. There are new exercises which relate closely to
chapter content and new sections on further reading.
Michael Toolan is Professor of Applied English Linguistics at the Univer-

sity of Birmingham. His previous books include Language in Literature
(1998), Language, Text and Context: Essays in contextualized stylistics
(1993) and The Stylistics of Fiction (1990).
The INTERFACE Series
Language, Literature and Critical Feminist Stylistics

Practice Sara Mills
Ways of analysing text
Twentieth-century Fiction
David Birch
From text to context
Literature, Language and Change Peter Verdonk and Jean Jacques Weber
Ruth Waterhouse and John Stephens
Variety in Written English
Literary Studies in Action Texts in society: societies in text
Alan Durant and Nigel Fabb Tony Box
Language in Popular Fiction English in Speech and Writing
Walter Nash Investigating language and literature
Language, Text and Context Rebecca Hughes
Essays in stylistics
Language through Literature
Edited by Michael Toolan
An introduction
The Language of Jokes Paul Simpson
Analysing verbal play
Delia Chiaro Patterns in Language
An introduction to language and
Language, Ideology and Point of View literary style
Paul Simpson Joanna Thornborrow and
A Linguistic History of English Poetry Shân Wareing
Richard Bradford
Exploring the Language of Drama
Literature about Language From text to context
Valerie Shepherd Edited by Jonathan Culpeper,
Twentieth-century Poetry Mick Short and Peter Verdonk
From text to context The Discourse of Advertising
Edited by Peter Verdonk Second edition
Textual Intervention Guy Cook
Critical and creative strategies for
literary studies
Rob Pope
The Series Editor

Ronald Carter is Professor of Modern English Language at the University of
Nottingham and was National Coordinator of the ‘Language in the National
Curriculum’ Project (LINC) from 1989 to 1992.
Narrative
A critical linguistic introduction
Second Edition
Michael Toolan
London and New York

In loving memory of
Margaret ‘Mac’ McAloren
who told such stories,
and never told on us.

First published 1988
Reprinted 1991, 1992, 1994, 1995, 1997
Second edition first published 2001
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Simultaneously published in the USA and Canada
by Routledge
270 Madison Ave, New York NY 10016
Transferred to Digital Printing 2005
Routledge is an imprint of the Taylor & Francis Group
© 1988, 2001 Michael Toolan
Typeset in Times by Wearset, Boldon, Tyne and Wear
All rights reserved. No part of this book may be reprinted or
reproduced or utilised in any form or by any electronic, mechanical,
or other means, now known or hereafter invented, including
photocopying and recording, or in any information storage or
retrieval system, without permission in writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from The British
Library
Library of Congress Cataloging in Publication Data
Toolan, Michael J.
Narrative : a critical linguistic introduction / Michael Toolan. –
2nd ed.
p. cm. – (Interface)
Includes bibliographical references and indexes.
1. Discourse analysis. Narrative. I. Title. II. Interface (London,
England)
P302.7 .T66 2001
401⬘.41–dc21
2001019304
ISBN 0–415–23174–4 (hbk)
ISBN 0–415–23175–2 (pbk)
Contents
Preface viii
Acknowledgements xi
1 Preliminary orientations 1
1.1 Teller, tale, addressee 1
1.2 Typical characteristics of narratives 4
1.3 Narratives and non-narratives 8
1.4 Story – text – narration 10
Further reading 12
Notes and exercises 13
2 Basic story structure 15

2.1 Story/fabula/histoire 15
2.2 Propp’s morphology of the Russian fairytale 17
2.3 Barthes on narrative 22
2.4 Plot-summarizing: modelling intuitions 29
2.5 In search of the grammaticization of plot structure 31
Further reading 37
3 The articulation of narrative text I: time, focalization,

narration 41
3.1 Narrative text: a single level of analysis 41
3.2 Text and time 42
3.3 Temporal refractions in text: Nabokov’s Pnin 54
3.4 Focalization 59
3.5 Perceptual focalization as primary 63
3.6 Narrators and narration 64
3.7 Simpson’s typology of narratorial modes 68
Further reading 77
vi Contents
4 The articulation of narrative text II: character, setting,
suspense, film 80
4.1 Character 80
4.2 Greimas’ actant model 82
4.3 Character traits and attributes 86
4.4 Distinctive feature characterology 88
4.5 Setting 91
4.6 Character and setting in ‘The Dead’ 94
4.7 Creating surprise and suspense in narratives 99
4.8 From prose to film: radical translation 103
4.9 The grammaticization of character and situation 107
Further reading 112
5 The articulation of narrative text III: representing

character discourse 116
5.1 Achieving immediacy in the narration of thoughts 116
5.2 Modes of speech and thought presentation 119
5.3 Differences between Direct and Indirect Discourse 125
5.4 Different again: free Indirect Discourse 130
5.5 Who speaks, who thinks? 133
5.6 FID: functions and effects 134
Further reading 140
6 Narrative as socially situated: the sociolinguistic

approach 143
6.1 Labov and narrative structure 143
6.2 Fixed narrative clauses, free evaluative clauses 145
6.3 Abstracts and orientations 149
6.4 Evaluation 151
6.5 Doing and saying 153
6.6 Internal evaluation 155
6.7 Coda 157
6.8 Stories in societies 159
6.9 Narrative performance 160
6.10 Dispersed, embedded, and group oral narratives 162
6.11 From Labov to literature 167
Further reading 172
Contents vii
7 Children’s narratives 178
7.1 Stories for, by, and with children 178
7.2 Storytelling and emergent literacy 180
7.3 Differing styles, differing orientations 182
7.4 Children’s narrative development 185
7.5 Children’s narratives and the development of
registers and genres: the systemic-linguistic
approach 189
7.6 The systemic-linguistic account of story genres 193
7.7 Stories for and with children 197
Further reading 203
8 Narrative as political action 206

8.1 The contexts of narratives 206
8.2 Hard news stories in the newspaper 206
8.3 Political narratives in the news 208
8.4 The unfolding news story: a contemporary
example 212
8.5 The linguistic apparatus of political construal:
notes on key resources 221
8.6 News stories online 230
8.7 Stories of class and gender 233
8.8 Prejudice in ethnic narratives 234
8.9 Stories in court 235
Further reading 239
Bibliography 242
Index 255
Preface
Narratives are everywhere, performing countless different functions in

human interaction; therefore the area of inquiry of this book must be delim-
ited rather strictly. As the subtitle indicates, it is intended as a critical intro-
duction, and I hope to be genuinely critical and genuinely introductory.
More narrowly still, this critical introduction is specifically concerned with
language-oriented or linguistically-minded perspectives on narrative: ways
of looking at narrative that attend systematically to the language of stories,
and models of narrative-analysis that focus on the linguistic form of narra-
tives or their linguistically-describable structure. The basic rationale for such
an emphasis is the conviction that systematic analytical attention to the logic
and dynamics of language behaviour can shed light on any sub-domain or
mode of language behaviour. The mode spotlighted here is narrative.
What is it about narrative that makes it such a pervasive and fascinating
phenomenon? And how can one begin to answer such a question without
entering into a narrative of one’s own? The fact is, as my opening sentence
announces, narratives are everywhere. Or are potentially so. Everything we
do, from making the bed to making breakfast to taking a shower (and notice
how these combined – in any order – make a multi-episode narrative), can
be seen, cast, and recounted as a narrative – a narrative with a middle and
end, characters, setting, drama (difficulties resolved), suspense, enigma,
‘human interest’, and a moral. (The moral of the story of my making break-
fast this morning could be stated as ‘Don’t try to clean the toaster while
cooking porridge’.) From such narratives, major and minor, we learn more
about ourselves and the world around us. Making, apprehending, and then
not forgetting a narrative is making-sense of things which may also help
make sense of other things.
Just how pervasive and important oral and written narratives are to our
lives becomes startlingly clear if we stop to think of the forms of narrative
we depend on as props and inspirations: biographies and autobiographies;
historical texts; news stories and news features in many media; personal
letters and diaries; novels, thrillers and romances; medical case histories;
school records; curricula vitae; police reports of ‘incidents’; annual perfor-
mance reviews; and, often most crucially, the stories we tell about our-
selves and others – stories of triumphs and disasters, pleasure and pain – in
Preface ix
the course of our everyday lives. These are only some of the materials
shaping our lives that are palpably narrative in form and function.
But we might also consider many other preoccupations which, as a
means of assisting comprehension, we ‘narrativize’. Law students strug-
gling to grasp and retain the ramifications of the law concerning theft may
well, as a sense-making procedure, cast the law(s) as a developing story
shaped by attendances to and departures from precedent, and by statutory
revisions. And the criminal law in its entirety can be seen as a revisable
story: the story is about socially impermissible conduct and the means of
redress available when such conduct is exposed. This all-embracing ‘story
of the law’ subsumes an infinite number of more specific episodes (actual
and hypothetical), with probable but contestable outcomes: if you do this,
in those circumstances, then you may be liable to such and such penalties.
Science, too, may at first glance look very different from narrative. We
often think of it as an expanding storehouse of incontestable facts, the hal-
lowed repository of objective knowledge of how things in the world work:
a rich but static description, quite remote from ‘storytelling’. But that
turns out to be mistaken in both theory and practice. In theory, the
emphasis on scientific enquiry as an ongoing revisable narrative (with revi-
sions made on the rational grounds that the revised account brings
enhanced descriptive or explanatory power, and greater generalizability)
is now commonplace. And in practice, too, one has only to think of
how science is taught in schools to see the centrality of narrative to under-
standing.
For instance, the concepts of fuel, energy and work might be taught in
the primary school by telling stories about eating breakfast before running
around, and putting fuel in the car before going on a long trip. If the child
doesn’t get the point of these stories, and see the logical connections
between the stages within each story as well as the analogical parallels
across the stories, they won’t begin to understand the concepts involved.
At secondary school the presentation may be less informal and more
theorized, but narrative methods persist. Any laboratory exercise in
physics, chemistry or biology, for example, is a planned and guided story
in which the child is an essential participant. Testing for the hydrogen that
is released when copper filings are added to sulphuric acid is, for teacher
and lab assistant, an old, old story (ah, they don’t make them like that any
more!). But it’s a new story, a narrative of enforced personal experience if
you like, for the child, the moral of which is to be learned. And afterwards,
in the passive voice style that tries to keep human interest out of the
picture, they must ‘write up’ the experiment.
If the above is a reminder that narrative is a mode that, directly or more
indirectly, may inform almost every aspect of human activity, I must now
stress that the following chapters are concerned almost entirely with narra-
tives in a narrower sense: literary narratives, folktales, stories by and for
children, conversationally-embedded spoken narratives, and news stories
in the media. There are linguistic similarities between these types of
x Preface
stories which I hope, rather than leading to a boring sameness, will be
thought-provoking, and linguistic differences, too, which are yet not so
great as to make for unmanageable heterogeneity.
For this second edition, I have made many minor revisions and rephras-
ings, some cuts and several additions or expansions. The expansions have
been kept in check by the need to keep the book to a manageable length,
and the requirement that something linguistic and introductory could be
relevantly said about each topic. The substantially new sections include
ones on narrativity (1.3), modes of narration (3.8), surprise and suspense
(4.7), film narration (4.8), Labov applied to literature (6.11), systemic story
genres (7.6), the structure and analysis of hard news stories in print and
online (8.2–8.4, 8.6), and gender (8.7). All the Further Reading and Notes
and Exercises sections, at the end of each chapter, have been radically
revised and brought up to date.
On the other hand, I have sometimes retained from the first edition
particular demonstration analyses even where these use approaches that
may have developed further in very recent years. The more recent work
often builds on the earlier work, so that the latter remains both important
and truly introductory.
In the following chapters, particularly the earlier and more literary-
minded ones, presentation and discussion of models and theories often
involves detailed reference to one or more of a few celebrated literary
texts which I have taken as exemplary. So the best way to read these chap-
ters is with those narratives both firmly in memory and close to hand for
direct consultation. This special collection comprises the following stories:
James Joyce’s ‘Eveline’ and ‘The Dead’, from Dubliners; William
Faulkner’s ‘That Evening Sun’ and ‘Barn Burning’; Katherine Mansfield’s
‘Bliss’; and Vladimir Nabokov’s novel, Pnin. Many other narratives, short
and long, oral and written, literary and non-literary, will be discussed in
the course of the book. But those six are especially relevant to the first five
chapters.
A linguistic introduction can hardly avoid the occasional use of more
technical terms that may at first seem off-putting to those who have taken
no linguistics courses. I have tried to keep specialist jargon to a minimum,
explaining terms as the discussion proceeds.
Acknowledgements
Nearly all the work on the first edition of this book was done while I was a
member of the Department of English Language and Literature of the
National University of Singapore, while any better thoughts that have
emerged in the second edition arose at my next academic home, the
English department of the University of Washington, Seattle, or my
present one, the Department of English of the University of Birmingham.
Much of the material presented here has been used on courses in Stylistics
or Narrative at all three institutions, and all kinds of small debts are owed
to students on those courses. I still owe thanks to all those listed in the first
edition, friends, colleagues and students, together with a goodly number of
scholars who have one way or another influenced or unwittingly con-
tributed to the second edition: Anneliese Kramer-Dahl, Betty Samraj,
Brian Ridge, Carmen-Rosa Caldas-Coulthard, Carol Marley, Charles
Owen, Chris Heffer, David Birch, David Butt, Gail Stygall, George Dillon,
George Wolf, Hayley Davis, Heidi Riggenbach, Jim Martin, K. P.
Mohanan, Malcolm Coulthard, Michael Halliday, Michael Hoey, Monika
Fludernik, Nigel Love, Norman Macleod, Paul Hopper, Paul Simpson,
Peter Verdonk, Peter White, Phil Gaines, Roy Harris, Ruth Page, Sandy
Silberstein, Talbot Taylor, Thara Mohanan, Thiru Kandiah, Tony Hung,
and Victor Li. Thanks also to Ms Gouri Uppal for permitting me to repro-
duce conversational data from her National University of Singapore MA
thesis (1984).
Special thanks remain due to Rukmini Bhaya Nair, with whom I first
co-taught courses on Narrative Structure.
Very special thanks to Ronald Carter, general editor of the Interface
series, for entrusting this project to me in the first place and for being a
powerful advocate for literary linguistics these twenty years and more; and
to Louisa Semlyen at Routledge who was unfailingly supportive when the
second edition turned out to be harder and longer work than I had
expected.
Mega special thanks to Julianne Statham, who has again gone above
and beyond the proverbial in reading through chunks of this in draft,
helping me absent the most turgid bits. And a verbal hug, too, for Roisin,
Patrick, and Miriam, teenagers with attitude; Chapter 7 was first written
xii Acknowledgements
when only Roisin was old enough to enjoy Burglar Bill, and Miriam’s story
had not even begun. Now all three are old enough and smart enough to
explain the new narratives to me.
The author and the publisher also wish to thank the copyright holders for
their permission to reproduce the following material in this book:
‘Hague links Labour with murder rise’, by Philip Webster and Tom
Baldwin, © Times Newspapers Limited, 19 December 2000.
‘Race and Policing: Hague’s defiance inflames the anger’, by Paul Waugh
and Andrew Grice, 19 December 2000, © The Independent/Syndication.
‘Hague race jibe angers ministers’, by Nicholas White and Nick Hopkins,
19 December 2000, © The Guardian.
‘Tory Leader ‘‘won’t be gagged on crime’’’, by George Jones, 19 Decem-
ber 2000, © The Telegraph.
‘Crazy Hague defies Dami dad’s plea’, 19 December 2000, © The Mirror.
1 Preliminary orientations
1.1 Teller, tale, addressee

What is narrative? What do we mean by ‘narrative structure’? Where
does a linguistic approach come in, and how helpful can it really be?
The following are introductory notes on these and other basic issues,
which should at least indicate the terrain to be covered, and why it is
significant.
Commentators sometimes begin by stating the truism that any tale
involves a teller, and that, therefore, narrative study must analyse two
basic components: the tale and the teller. But as much could be said of
every speech event: there is always inherently a speaker, separable from
what is spoken. What makes narratives different, especially literary or
extended spoken ones, is that the teller is often particularly noticeable.
Tellers of long narratives can be surprisingly present and perceptible even
as they unfold a tale that ostensibly draws all our attention, as readers or
listeners, to other individuals who are within the tale. As a result we may
feel that we are dividing our attention between two objects of interest:
the individuals and events in the story itself, and the individual telling us
about these. Thus when we read Coleridge’s ‘Rime of the Ancient
Mariner’ or Bronte’s Wuthering Heights or listen to the rambling anec-
dote of a friend, part of the experience is the activity of ‘reading’ or scru-
tinizing the character of the teller: the returned mariner, Lockwood, the
friend. Already the two literary examples cited involve an enriching com-
plication. In both texts mentioned, there is more than one teller: besides
the mariner, for instance, is a ‘higher’ teller who writes, ‘It is an ancient
Mariner/And he stoppeth one of three’. But we can address such compli-
cations later, and should concentrate here on narrative’s dual essential
foci, teller and tale.
The possibility of achieving this effect of divided attention exploits a
basic characteristic of narrative. Narrative typically is a recounting of things
spatiotemporally distant: here’s the present teller, seemingly close to the
addressee (reader or listener), and there at a distance is the tale and its
topic. This selection of effects of closeness and distance can be represented
graphically:
TELLER
TALE
ADDRESSEE
But since the present teller is the sole access to the distant topic, there is a
sense, too, in which narrative entails making what is distant and absent
uncommonly present: a three-way merging rather than a division. Dia-
grammatically this merging-and-immediacy can be represented as:
TELLER
TALE ADDRESSEE
However, since tellers can become intensely absorbed in their self-

generated sense of the distant topic they are relating, addressees some-
times have the impression that the teller has withdrawn from them, has
taken leave, so as to be more fully involved in the removed scene. This
third type of relation between tale, teller, and addressee (a withdrawing
and merging) might be cast thus:
TELLER
ADDRESSEE
TALE
In short, narratives always involve a Tale, a Teller, and an Addressee, and

these can be ‘placed’, notionally, at different degrees of mutual proximity
or distance. Hawthorn (1985) broaches these same issues, taking a painting
by Millais, The Boyhood of Raleigh, as capturing something central to nar-
rative. In that painting an old seaman, with his back to the viewer, appears
to be addressing two young boys who are evidently fascinated and
absorbed by what he tells them. The old man is using his whole arm to
point out to the sea, visible in the distance. But the boys’ eyes are on the
man and his gesturing arm, not any distant scene he may be designating.
Narrative focusses our attention on to a story, a sequence of events,

through the direct mediation of a ‘telling’ which we both stare at and
Preliminary orientations 3
through, which is at once central and peripheral to the experience of
the story, both absent and present in the consciousness of those being
told the story. Like the two young boys we stare at the ‘telling’ while
our minds are fixed upon what that telling points towards. We look at
the pointing arm but our minds are fixed upon what is pointed at.
(Hawthorn, 1985: vii)
One of the distinctive characteristics of narrative concerns its necessary

source, the narrator. We stare at the narrator rather than interacting with
him as we would if we were in conversation; at the same time, in literary
narratives especially, that narrator is often ‘impersonalized’, and attended
to as a disembodied voice.
Thus there is a teller in every tale to a far greater degree than there is a
speaker in any ordinary turn at talk. Because narratives are, relative to
ordinary turns of talk, long texts and personalized or evaluated texts, there
is a way in which, while your conversational remarks reflect who you are
(your identity and values), in the course of any narrative the narrator’s
text describes that narrator. In brief snatches of conversation, a person
may be able, through accent-mimicry for example, to ‘pass’ for someone of
a different class or gender or ethnic identity; but to take on another’s iden-
tity in a sustained fashion, across a number of personal narratives, is ordi-
narily very difficult, and may even imply disabling confusion or a
personality disorder. The reflection/description contrast may be chiefly a
matter of degree, but it is arguably an important contrast with far-reaching
consequences – e.g. even for assessments of mental health or illness.
This brings us to another important asset of narrators: narrators are
typically trusted by their addressees. In at least implicitly seeking and
being granted rights to a lengthy verbal contribution, narrators assert their
authority to tell, to take up the role of knower, or entertainer, or producer,
in relation to the addressees’ adopted role of learner or consumer. To
narrate is to bid for a kind of power. Sometimes the narratives told cru-
cially affect our lives: those told by journalists, politicians, colleagues,
employers assessing our performance in annual reviews, as well as those of
friends, acquaintances, enemies, parents, siblings, children – in short, all
those which originate from those who have power, authority or influence
over us. Any narrator then is ordinarily granted, as a rebuttable presump-
tion, a level of trust and authority which is also a granting or asserting of
power. But this trust, power and authority can be exploited or abused, as is
reflected in literary critical discussion of ‘unreliable narration’. Narrative
misrepresentation is a complex process, difficult to unravel. One exemplifi-
cation of it arises far from literature: in criminal cases of serious fraud.
Where, after having pleaded not guilty, a defendant is found guilty, the
sentencing judge often refers to the obfuscating detailed deception that
has been uncovered as ‘a complex tissue of systematic distortion and fabri-
cation’, or uses a similar revealing description.
Even before we attempt a working definition of narratives, it is clear
that these are typically ‘cut off’ in some respects from surrounding co-
text and context (their verbal and non-verbal environment, respectively:
the former comprises any language that precedes or follows the narra-
tive, the latter includes anything non-verbal of relevance, including the
situation and the identities of teller and addressee). Narratives often
appear to stand alone, not embedded in a larger frame, without any
accompanying information about the author or the intended audience:
they’re just ‘there’, it seems, like pots or paintings, and you can take
them or leave them. They differ, at least in degree, from more transac-
tional uses of language, as when someone asks you a question, or makes
a request or a promise or warning: in such cases there is strong expecta-
tion that the addressee will respond or act in predictable ways. So some
of the normal constraints on how we make sense of discourse seem to be
suspended. And it seems we do not always have to relate narratives
directly and immediately to their authors, or socio-historical back-
grounds.
1.2 Typical characteristics of narratives

We can begin to define narrative by noting and inspecting some of its
typical characteristics:
1 A degree of artificial fabrication or constructedness not usually appar-

ent in spontaneous conversation. Narrative is ‘worked upon’.
Sequence, emphasis and pace are usually planned (even in oral narra-
tive, when there has been some rehearsal – previous performance – of
it). But then as much could be said of, for example, elaborate descrip-
tions of things, prayers, scholarly articles.
2 A degree of prefabrication. In other words, narratives often seem to
have bits we have seen or heard, or think we have seen or heard,
before (recurrent chunks far larger than the recurrent chunks we call
words). One Mills and Boon heroine or hero seems much like another
– and some degree of typicality seems to apply to heroes and heroines
in more elevated fictions too, such as nineteenth-century British
novels. Major characters in the novels of Dickens, Eliot, Hardy, etc.,
seem to be thwarted (for a time at least) in roughly comparable ways.
And the kinds of things people do in narratives seem to repeat them-
selves over and over again – with important variations, of course.
Again, prefabrication seems common in various types of writing and
visual spectacle besides narrative, although the kinds of things men-
tioned above seem particularly to be prefabricated units of narrative.
3 Narratives typically seem to have a ‘trajectory’. They usually go some-
where, and are expected to go somewhere, with some sort of develop-
ment and even a resolution or conclusion provided. We expect them
to have beginnings, middles, and ends (as Aristotle stipulated in his
Poetics). Consider the concluding words of children’s stories:
And they all lived happily ever after;
since then, the dragon has never been seen again.
and notice the finality and permanence conveyed by the ever/never

pair. Or consider the common story-reader’s exit-line:
And that is the end of the story.
which has near-identical counterparts in the closing sequences of radio

and television news bulletins. All these examples mark this attention
to the expectation of closure and finality, itself just one aspect of the
broader underlying expectation of narrative trajectory. Relatedly, the
addressee is usually given to understand, and does so assume, that
even embarking on their story the teller knows how it ends up (not the
precise wording, but the event-based or situational gist). Exceptions to
this might include Dickens’s serialized novels, and the bedtime story
that a parent makes up for a child, impromptu. Even these, along with
more planned-outcome narratives, can be distinguished from both
diaries and live commentaries, in that in the latter new intervening
acts, beyond the control of the witness/reporter, can dictate the shape
and content of the report. In true narratives, arguably, the teller is
always in full control although, like Dickens and the bedtime story-
teller, they may not fully foresee at the outset that material which they
will control.
4 Narratives have to have a teller, and that teller, no matter how back-
grounded or ‘invisible’, is always important. In this respect, despite its
special characteristics, narrative is language communication like any
other, requiring a speaker and some sort of addressee.
5 Narratives are richly exploitative of that design feature of language
called displacement (the ability of human language to be used to refer
to things or events that are removed, in space or time, from either
speaker or addressee). In this respect they contrast sharply with such
modes as commentary or description. Arguably there has to be some
removal or absence, in space or time, for a discourse to count as a nar-
rative. Thus if I listen in my car to a radio running commentary on a
simultaneously-occurring football match or funeral, this approaches
the status of narrative by virtue of spatial displacement (it is not a nar-
rative at all if I am at the football match directly witnessing, and listen-
ing to the radio commentary). But live commentaries, like real diaries,
breach characteristic 3 above, and are arguably not narratives at all.
More borderline are edited TV highlights of sports and other events
(interestingly, one rarely finds edited highlights of matches and events
on radio).
6 Narratives involve the recall of happenings that may be not merely
spatially, but, more crucially, temporally remote from the teller and
his audience. Compare our practices with those of the honeybee,
whose tail-wagging dance overcomes spatial displacement, in that it
communicates about distant sources of nectar, but cannot encompass
temporal displacement. Thus it can only signal to its chums back in
the hive immediately upon its return from the nectar-source.
Accordingly, the honeybee’s tail-wagging is no proper narrative in
our sense, but merely a kind of reflex observation. As Roy Harris
has remarked:
Bees do not regale one another with reminiscences of the nectar

they found last week, nor discuss together the nectar they might
find tomorrow.
(Harris, 1981: 158)
This is a lovely image (or narrative), partly because in fact it is some-

thing that (as far as we know) we humans alone do, and no other
animals – even those with simple language systems.
A first attempt at a minimalist definition of narrative might be:
a perceived sequence of non-randomly connected events
This definition recognizes that a narrative is a sequence of events. But

‘event’ itself is really a complex term, presupposing that there is some
recognized state or set of conditions, and that something happens, causing
a change to that state. The emphasis on ‘non-random connectedness’
means that a pure collage of described events, even given in sequence,
does not count as a narrative. For example, if each member of a group in
turn supplies a one-paragraph description of something or other, and
these paragraphs are then pasted together, they will not count as a narra-
tive unless someone comes to perceive a non-random connection. And by
‘non-random connection’ is meant a connectedness that is taken to be
motivated and significant. This curious transitional area between sequen-
tial description and consequential description is one of the bases for the
fun of a familiar party game in which people around a table take turns to
write a line of a ‘story’, the other lines of which are supplied, in secret, by
the other participants.
The important role of ‘change of state’ has been celebrated in the more
linguistic term transformation by the structuralist Tzvetan Todorov (1977:
233):
The simple relation of successive facts does not constitute a narrative:

these facts must be organized, which is to say, ultimately, that they
must have elements in common. But if all the elements are in
common, there is no longer a narrative, for there is no longer anything
to recount. Now, transformation represents precisely a synthesis of
differences and resemblance, it links two facts without their being able
to be identified.
The one-line definition of narrative also suggests that consequence is not so

much ‘given’ as ‘perceived’: narrative depends on the addressee seeing it as
narrative – the circularity here seems inescapable. While most would agree
that the traditional novel is a narrative, there can be legitimate disagreement
as to the status, as narrative, of less familiar and complex structures. For
example, imagine you enter a cartoonist’s studio and find three frames, on
separate pieces of paper, on his desk. They have quite different characters,
settings, furniture, etc., and seem to be about quite unrelated topics. They
seem to be rough drafts, because on the corner of one is a coffee-ring, where
the cartoonist has carelessly left a cup, and on the second one there’s a food
stain in the top corner, while the third has some cigarette ash and a coffee
ring on it! You think you see a narrative before you – though the cartoonist
hotly denies this.
That example, however whimsical, tries to touch on a fundamental but
problematic feature of narrative study. Perceiving non-random connected-
ness in a sequence of events is the prerogative of the addressee: it is idle
for anyone else (e.g. a teller) to insist that here is a narrative if the
addressee just does not see it as one (see 7.3). In this respect at least, the
ultimate authority for ratifying a text as a narrative rests not with the teller
but with the perceiver or addressee.
But, in practice, we expect and demand much more complex connected-
ness, non-randomness, and sequentiality in the events of narratives. In the
terms first highlighted by Aristotle, we expect ends as well as beginnings
and middles (something not commented upon in the quotation from
Todorov above). In more twentieth-century terminology, we expect
complex motivations and resolutions – even in quite ‘simple’ tales such as
folktales. Thus Benjamin Colby has written:
Folk narrative in its simplest form is the verbal description of one or

more concern-causing events and of the way in which the concern is
eliminated or diminished.
(Colby, 1970: 177)
As we shall see in the next chapter, this definition is similar to that of the
pioneer Russian narratologist Vladimir Propp. Propp studied the over-
arching structure of the Russian fairytale, identifying it as one in which an
initial state of equilibrium is disturbed by various forces of turbulence.
This turbulence brings disequilibrium and upheaval before some sort of
action (perhaps an intervention) leads to the restoration of a modified
version of the original equilibrium.
My skeletal definition emphasizes the role of the perceiver, but not that
of an independent teller. This is because the two roles are not entirely
separate. I have already suggested that the activity of perceiving
consequential relatedness of states is the enabling condition for narrative:
we might in addition speculate that it is an activity necessarily to be per-
formed by all tellers and addressees, at least if they intend to be tellers or
addressees.
1.3 Narratives and non-narratives

If a slightly less minimal definition were proposed, it might run something
like this:
A narrative is a perceived sequence of non-randomly connected

events, typically involving, as the experiencing agonist, humans or
quasi-humans, or other sentient beings, from whose experience we
humans can ‘learn’.
This definition introduces (1) one or more foregrounded individuals as

experiencers, and (2) the idea of addressees ‘learning’ from narratives –
i.e. the factor of human interest without which intended recipients might
well not be willing listeners. And if one wished to elaborate this one step
further, it might be to specify that our preference is often for the sequence
of connected events to take shape around a state or period of turbulence
or crisis, subsequently resolved. That is, while a sequence of events entails
some sort of change of state, a sequence containing a resolved crisis or
problem entails a pronounced change of state.
All these more contentious elaborations of the initial minimalist defini-
tion are useful when one attempts to distinguish narratives from all kinds of
non-narrative texts. For if a narrative is a sequence of logically and chrono-
logically related events, bound together by a recurrent focus (which may not
be constant or exclusive) on one or more individuals (‘characters’) in whom
the reader becomes interested (positively or negatively); and if the narrative
is in addition that kind of chronological sequence in which a period of turbu-
lence, crisis or uncertainty is superseded by a later stage of calm, solution or
closure; and if these are indeed defining criteria of narratives; then texts
which display few or none of the above features will not be narratives. And
texts which have some of the above features, or all of the above but only
weakly or barely perceptibly, should feel ‘marginal’, or ‘semi-narrative’. We
can summarize those three chief or defining features as:
• sequenced and interrelated events;

• foregrounded individuals;
• crisis to resolution progression.
Now consider a literary text such as Michelle Roberts’ novella – if that is

what it is – entitled Une Glossaire/A Glossary. This comprises a series of
alphabetized glosssary entries, with French headword and English-
language commentary, which offer brief (typically one-page) glimpses or
recollections of the rural Norman childhood that the narrator, with
French mother and English father, had access to. Cumulatively the text
amounts to an act of recovery, of beachcombing for fragments, and of
memorializing, all made the more poignant since the narrator’s cherished
French aunt, Brigitte, is dying of cancer as this glossary-compilation
proceeds.
Is Une Glossaire/A Glossary (as a whole, not in its separate entries) a
narrative, by the foregoing criteria? Certainly a conventional glossary
would lack all three core desiderata listed above. And initially, at least,
Une Glossaire/A Glossary appears to lack them too – unlike indisputably
narrative texts, like ‘Barn Burning’ and Pnin, where the reader can almost
immediately find indicators of the onset of an event-sequence, the begin-
ning of a foregrounding of one or more individuals, and even the first
sketching of conditions that may lead to a crisis. But is this ‘glossary’ in
fact more a journal than a glossary? If it were genuinely a journal then,
like a genuine diary, it would lack the possibility of mapping-out a before
and after, one already known and past to the teller, who is now in a con-
trolling position from which to recount them.
But the very first section of Une Glossaire, titled Absence, ends by
asserting that in writing a sort of geography the writer aims ‘to reclaim the
past’; and s/he ‘set(s) myself to remembering’. Both these suggest some
activity of narrative recall may be what is going to be undertaken. Even
this, however, can be removed from orthodox narrative if a number of dis-
tinct past scenes, objects, feelings, are recalled (just as an old person relat-
ing all their wartime memories may not integrate these into a unified
story). Now consider Roberts’ second section, Artichauts. There is very
little in the way of narrative in this brief section: it is quasi-encyclopaedic,
with only two sentences about how ‘we’ and Grandpère eat artichokes,
individualizing the text. However, in subsequent sections, we do gradually
learn more of particularly cherished or focussed-upon individuals: Grand-
père and Brigitte and, indirectly but perhaps most crucially, the maturing
narrator herself. A further feature suggestive of narrative status is the fact
that the glossary terminates at the letter P (Plage). We can take this as an
indication that the text is not really or merely a true glossary at all. The
plage or beach of course is a place of termination, of profound transition,
the end of one situation or status and the beginning of a very different
one. So, an ‘objective correlative’ for narrative change, or resolution. And
on the beach at low tide the narrator finds a kind of glossary of ‘semiotic’
fragments. And when the narrator says, ‘I must put them [the fragments,
the memories?] into some sort of order. Make a list,’ we are reminded of
course that she has done just that, in the foregoing text. An alphabetized
list is sharply non-narrative. But since this is such an annotated and
embellished list, one in which the whole becomes clearer as the later items
on the list are reached, we still have grounds for saying that here is a nar-
rative, albeit of a non-traditional kind. The decision in all such borderline
cases is really based on what the reader takes the ultimate purpose of the
text to be: narrative, or hortatory, or persuasive, or informative, or argu-
mentative, and so on.
1.4 Story – text – narration

In the next four chapters we shall look at some of the most influential lin-
guistically-informed discussions of narrative as verbal art: written or oral,
traditional and collective, or innovative and individual. These chapters
thus review what has been suggested about the poetics of narrative. By
this I mean the relatively abstract and theoretical commentary on the
more systematic and recurrent aspects of stories and storytelling. The
poetics of any type of verbal performance (revenge tragedies, sonnets,
diaries, whatever) is the study of the ground rules that shape all the pro-
ductions within that type. And most of these ground rules are not logical
requirements at all, but conventional norms in the production of revenge
tragedies, sonnets, and so on. Thereafter, Chapters 6 to 8 explore narra-
tives in various societal manifestations: in the social world of conversation,
by and for children, and in the public worlds of storytelling in the media,
the courts, and identity-construction.
Some warnings are in order concerning the way theorists of narrative
poetics have split the subject of study into two, and more recently three,
major domains or levels of inquiry. Thus the early-twentieth-century
Russian formalists (Propp, Tomashevsky, etc.) spoke of fabula and
sjuzhet, roughly equivalent to the more recent French (Benveniste,
Barthes) terms histoire and discours. These are roughly equivalent in
turn to Chatman’s English terms, story and discourse. By the first term of
each of these pairs is meant a basic description of the fundamental
events of a story, in their natural chronological order, with an accom-
panying and equally skeletal inventory of the roles of the characters in
that story.
A fabula [story] is a series of logically and chronologically related

events that are caused or experienced by actors.
(Bal, 1985: 5)
This is the level at which we may expect the possibility of ‘total transfer’
from one medium to another: everything at the level of story in, say, A
Christmas Carol, can and perhaps should appear as easily in a film or
cartoon version, or a ballet, as in the original written version. It may be
worth applying the linguistic terms ‘paradigmatic’ and ‘syntagmatic’ here.
A linguistic paradigm is a set or class of words (or other elements) that are
especially related to each other in that they amount to alternatives, in con-
trast with each other, usable at the same point in the verbal sequence,
without any regard for matters of sequence or progression. So the paradig-
matic axis of language is a ‘vertical’ or static column cutting through the
‘chain’ of speech or writing: every ‘link’ in the chain relates to a distinct
paradigm, and at each point whatever has been chosen to fill the link or
slot is tacitly contrasted with all the other members of the set or paradigm
that might have been used but were not. The syntagmatic axis of language
is the horizontal one, and concerns all our possible syntactic or usage
options in chaining words and phrases together, fully focussed on the
onward sequencing of items and paying no attention to the contrastive
identity of those items. In a story outline all the events and characters are
presented synoptically, with the minimum attention to, for example,
complexities of sequence, as if we’re getting the paradigmatic raw mater-
ials or ingredients; the syntagmatic dimension – the linear distribution of
event and character presentation, disclosure, elaboration, and so on – is
severely attenuated.
Above, I used the term ‘version’ to refer to actual cinematic or dance
realizations of the core story of A Christmas Carol. And ‘version’ is as
good a word as any to refer to the business of distinctive and creative
working on a story to produce the discourse we actually encounter. In
other words, sjuzhet or discours roughly denotes all the techniques that
authors bring to bear in their varying manner of presentation of the basic
story. As far as literary-minded people are concerned, discourse is much
the more interesting area of narrative poetics. Story seems to focus on the
pre-artistic, genre- and convention-bound basic event-and-character pat-
terns of narrative, with scarcely any room for evaluative contrasts or dis-
criminations – a level at which authorship seems an irrelevant concern.
Discourse looks at the artistic and individualized working with and around
the genres, the conventions, the basic story patterns, in the distinctive
styles, voices, or manners of different authors.
For good or ill – as will become clear, I have my doubts – the above
binary picture (of histoire versus discours, fabula versus sjuzhet, or story
versus discourse) has in recent discussions been complicated by the argu-
ment that we need to posit three levels, not two. As I understand it, this
rearrangement does not involve adjustment of both the binary categories
outlined above, but rather is simply a bifurcation of the second one, dis-
course. In the accounts of poetics of Genette (1980; 1988), and in several
of the books I will recommend as supporting textbooks in this area (e.g.
Bal, 1985; Rimmon-Kenan, 1983), the business of technical manipulation
and presentation of the basic story is said to involve two levels. That is to
say, if we think of histoire/story as level 1 of analysis, then within discourse
we have two further levels of organization, those of text and of narration.
At the level of text, the teller decides upon and creates a particular
sequencing of events, the time/space spent presenting them, the sense of
(changing) rhythm and pace in the discourse. Additionally, choices are
made as to just how (with what detail, and in what order) the particularity
of the various characters is to be presented, together also with choices as
to whose perspective or viewpoint will be adopted as the lens through
which particular events or descriptions or characters are seen and reported
(the business of focalization – to be discussed in Chapter 3). At the level of
narration, the relations between the posited narrator and the narrative she
tells are probed. An obvious contrast is that between a stretch of narrative
embedded within a novel and told by a character, on the one hand, and a
narrative told as if by a detached, external and omniscient onlooker, on
the other. This is also the level at which speech presentation (the mimetic
effects of pure dialogue, the deliberate ambiguities of free indirect dis-
course) can be analyzed.
This distinction between what we will call text and narration comes
principally from Bal. It amounts to an attempt to separate a layer at which
a narrative agent relates the text (a level of narration) from all the other
aspects of text manipulation (involving choices over how the story is pre-
sented). So text presents story in a certain manner, and in the narration an
agent relates that presentation. However, this latter separation is still a
source of controversy, and we may well want to question this confident
separation of narration from presentation. Two-level analysts, who find
the story/discourse bifurcation complicated enough, will always counter
with the claim that types of narration, and strategies of speech and
thought-presentation, are aspects of the manner of presentation, part of a
single domain of discourse.
These complex arguments will be returned to in passing but for now the
chief thing to keep in mind is the disparity in terminology used. Latterly,
in place of:
STORY – DISCOURSE
we have:
STORY – TEXT – NARRATION (as in Rimmon-Kenan, 1983)
with the added complication that these three terms are translated in Bal
(1985) as, respectively:
FABULA – STORY – TEXT
In general I shall use the same terms as Rimmon-Kenan wherever pos-

sible, even though reference will also be made to Bal’s study.
Further reading
Studies containing interesting general reflections on the nature of narrative and
its typical or defining features, from a range of literary and linguistic perspec-
tives, include: Bal (1985); Chatman (1978); Georgakopoulou and Goutsos
(1997); Hoey (2001); Longacre (1983); Schiffrin (1994); and Scholes and Kellogg
(1966).
Notes and exercises
1 There have been many proposals as to how to make fundamental distinctions
between types of discourse. One such is presented in Longacre (1983), who sug-
gests that discourses may be first distinguished from each other depending on
whether or not they involve ‘contingent temporal succession’ and orientation to
a particular agent. These yield four broad discourse categories: Narrative, Pro-
cedural (both ⫹succession, but only the former is ⫹agent), Behavioural, and
Expository (both ⫺succession, but only the former is ⫹agent). Examples of the
three non-narrative types include cookbooks and car manuals (Procedural),
opinion-editorial exhortations and self-help books (Behavioural), budget pro-
posals and scientific papers (Expository). Longacre’s scheme, which forms a
backcloth to further generic description of narrative elements, can be usefully
applied to ‘borderline’ narratives.
2 A really strong sense of narrativity is often derived from contemplating (coher-
ent) texts which are in sharp structural contrast with narratives. Chapter 5 of
Hoey (2001) offers an informative account of the structure of ‘discourse
colonies’: overlooked or ‘Cinderella’ written texts which Hoey suggests com-
prise elements which accumulate or combine rather like ants in an ant colony
(and not like the non-interchangeable organs in a human body). Consider the
items set out on a shopping list, or the entries in a phone book, or an encyc-
lopaedia; or the distribution of distinct news items on the Sports page of a
newspaper, or someone’s listing of ‘my favourite links’ on their personalized
web pages: each of these (the shopping list, the phone book) is a discourse
colony. The ‘texts’ that make up a discourse colony may be one word long (e.g.
the items in a shopping list), or many thousands of words long (e.g. essay-length
articles in an encyclopaedia). So the texts within a discourse colony may well
have their own kind of internal structure, which may be worth further study in
their own right (e.g. the internal ‘grammar’ of the short film notices in the
‘What’s On’ sections of many weeklies and weekend newspapers). Hoey’s dis-
cussion of the ‘discourse colony’ as a neglected type of written text may be
compared with Schiffrin’s commentary on ‘the list’, as a ‘nearly’ narrative
extended turn at talk, to be found often in everyday conversations (Schiffrin
1994: 291–315). See also Eggins and Slade (1997), on non-narrative genres in
everyday conversation.
3 The idea of change (of state) is so crucial to narrative, and to the status of narra-
tive as a definingly human activity, because without the recording of change that
narrative enables, the enactment of further changes would be considerably ham-
pered. In other words talking to each other about changes/developments in the past
assists us in coming up with plans in which we initiate (or respond to) present or
future changes – and this amounts to a profound resource with which to ‘master’
one’s environment. (Cf. the discussion of narrative control in this chapter: ‘master’
needs to be put in hedging quotation marks here to acknowledge that attempted
mastery by no means ensures success.) But notice that this reverses our everyday
way of thinking about narrative: in everyday thinking, first something happens,
and then you report it in a narrative. I am suggesting here that our enormous
propensity for narrative (or narrativization of experience) is cognitively enabling:
it helps us shape or present and plan futures to a degree far beyond the scope of
other animals. In effect, first we make a narrative, and then we are able to make or
adapt to real-world changes. An ‘after the fact’ activity, as soon as humans reflex-
ively recognized it as such, must have equally become a ‘before-the-fact’ activity,
of planning and projection.
4 None of this addresses death, which is perhaps the deepest human fact or
anxiety underlying our narrative drive. Like a life terminated by death, narra-
tives have endings but – perhaps more satisfyingly than we fear our lives and
deaths sometimes are – narratives also very typically make sense (that is to say,
usually they are coherent, they resolve or explain something). So while we live
with the anxiety that our lives may have one of these key attributes but not the
other (they come to an end, but they do not make sense), we comfort ourselves
with narratives, which have both these desiderata: they have beginnings, middles
and ends, and they make sense.
2 Basic story structure
2.1 Story/fabula/histoire
Narrative poeticians have long worked with a theoretical division of their
subject-matter into the domains of story (or fabula, or histoire) and dis-
course (or sjuzhet, or discours). Story is the basic unshaped story material,
and (with qualifications) comprises events, characters and settings. The
relations between these three are remarkably variable, but examples of all
three are nearly always present in narrative, although it is possible to dis-
pense with any explicit establishing of setting. Simply within the novel
canon, compare the relative emphasis on event in an adventure novel with
the relative emphasis on character in a Henry James novel, and the rela-
tive emphasis on setting in an historical novel. If events, characters and
settings are all-important elements of story, the first of these three has
nevertheless always been treated as pre-eminent and foundational by the-
orists of plot. For many theorists, the expressions ‘basic story structure’
and ‘event structure’ seem virtually synonymous. And a similar preoccupa-
tion with events and event structure, especially to the neglect of character,
will be apparent throughout this chapter. Character and setting will be
examined, however, in Chapter 4.
In order to describe story we have to adopt a medium of communica-
tion such as language. But the notion of unshaped, uncrafted, ‘unaestheti-
cized’ story, underlying every organized, shaped narrative we encounter, is
one that tends to treat the basic stuff of narrative as medium-independent.
Terminology introduced by Chomsky (1957; 1965) to explain syntactic
relations between basic sentences and more complex ones, and how the
latter are derived from the former, may be of some use here. Chomsky
argued that beneath or behind the ‘superficial’ differences between such
pairs of sentences as John fed the cat and The cat was fed by John, native
speakers know there is structural relatedness and an identical meaning:
the two surface structures differ, but these are minor rearrangements of a
single deep structure. More generally, ‘deep structure’ has come to mean
the underlying and core format of one or more texts (or other cultural
product), out of which, with enrichments and transformations that do not
displace that deep format, particular texts are produced. The idea of deep
structure as the encapsulation of the essential elements and operations of
a phenomenon, before or beneath all elaborations and refractions by
culture and interpretation, lives on in a range of disciplines. And in narra-
tive study, it is the unshaped or pre-shaped story that has been equated
with deep structure. Story has been thought of as a chronologically-
ordered deep structure representation of all the primary and essential
information concerning characters, events and settings, without which the
narrative would not be well formed. The important point here is that this
representation, or ‘bald version’, is abstract but structured. We may then
think of the teller of a narrative (the creative artist, the eye-witness,
or whoever) as generating a concrete ‘finished product’, the presented
discourse.
Here the Chomskyan analogy is weakest, since clearly most narrative
transformations are not so much transformations as elaborations and
enrichments, a fleshing-out of the basic story stuff. Reordering transforma-
tions, however, in which events which would happen in the real world in a
particular sequence ABCD are reordered so that they are encountered in
the discourse in the order BACD, are very widespread. An extreme form
of reordering transformation, in which crucial explanatory information is
withheld until the very end of a narrative, is common in crime and detec-
tive fiction, where significant information and clues, known to the author
and sometimes even the fictional detective whose enquiries are portrayed,
are withheld from the reader. Such withheld information is common in
‘high’ fiction, too: consider the withheld information as to who Pip’s bene-
factor is in Great Expectations, or as to who Esther’s parents are in Bleak
House.
The flourishing of abstraction in various humanistic academic disci-
plines is perhaps what underwrites the assumption that there is an abstract
level of story from which all concrete narratives, embellished by variations
of content, are derived. But this is untenable. All narratives involve the
report of some state and some change or changes to that state, and even as
we attempt to specify the allegedly core events and characters of stories
(the core ‘types’ of which events and characters in particular narratives are
‘tokens’) we find that content still remains. It has not and cannot be wholly
removed. If we look at what Vladimir Propp (the pioneer Russian analyst
of story structure) and others actually did, we find that, in search of basic
story structure, they started – inevitably – with the rich performed narra-
tive, and tried to ‘sift through’ that material, discarding all but the most
basic patterns. And yet even those patterns – as we shall see with Propp –
are quite clearly at best (as he conceded) genre-specific, at worst corpus-
specific. We need to see the implications of saying that certain identified
patterns in fact hold only for a particular genre, or, more limitedly, hold
only for the small collection of narratives actually analyzed.
Basic story structure 17
2.2 Propp’s morphology of the Russian fairytale
The starting point of Propp’s famous study (Propp, 1968; originally pub-
lished in Russian in 1928) would seem to be very much the sort of mini-
malist definition of narrative introduced in Chapter 1 – a text in which
there is recounted a change from one state to a modified state. As noted
earlier, we can label the actual change of state an ‘event’. Thus ‘event’, or
‘change of state’, is the key and fundamental of narrative. And Propp’s
morphology of the Russian fairytale is basically an inventory of all and
only the fundamental events (which he calls ‘functions’) that he identifies
in his corpus, which comprises 115 Russian fairytales.
In other words, Propp analyzed his collection of fairytales, looking
particularly for recurring elements or features (constants), and random or
unpredictable ones (variables). He concluded that, while the characters or
personages of the tales might superficially be quite variable, yet their
functions in the tales, the significance of their actions as viewed from the
point of view of the story’s development, were constant and predictable.
Both the number and sequence of the functions are asserted to be fixed:
there are just thirty-one functions, and they always appear in the same
sequence.
Functions of characters serve as stable, constant elements in a tale,

independent of how and by whom they are fulfilled. They constitute
the fundamental components of a tale.
(Propp, 1968: 21)
Here, for convenience in subsequent analytical tasks, is a full list of

Propp’s set of thirty-one key, fairytale-developing actions (functions),
which bring sequential changes to a specified initial situation:
1 One of the members of a family absents himself from home. (An

extreme exponent of this function of ‘absenting’ is where the parents
have died.)
2 An interdiction is addressed to the hero.
3 The interdiction is violated.
4 The villain makes an attempt at reconnaissance.
5 The villain receives information about his victim.
6 The villain attempts to deceive his victim in order to take possession
of him or of his belongings.
7 The victim submits to deception and thereby unwittingly helps his
enemy.
8 The villain causes harm or injury to a member of a family (defined as
‘villainy’).
8a One member of a family either lacks something or desires to have
something (defined as ‘lack’).
9 Misfortune or lack is made known; the hero is approached with a
request or command; he is allowed to go or he is despatched.
10 The seeker agrees to or decides upon counteraction.
11 The hero leaves home.
12 The hero is tested, interrogated, attacked, etc., which prepares the way
for his receiving either a magical agent or helper.
13 The hero reacts to the actions of the future donor.
14 The hero acquires the use of a magical agent.
15 The hero is transferred, delivered, or led to the whereabouts of an
object of search.
16 The hero and the villain join in direct combat.
17 The hero is branded.
18 The villain is defeated.
19 The initial misfortune or lack is liquidated.
20 The hero returns.
21 The hero is pursued.
22 The rescue of the hero from pursuit.
23 The hero, unrecognized, arrives home or in another country.
24 A false hero presents unfounded claims.
25 A difficult task is proposed to the hero.
26 The task is resolved.
27 The hero is recognized.
28 The false hero or villain is exposed.
29 The hero is given a new appearance.
30 The villain is punished.
31 The hero is married and ascends the throne.
Propp notes some internal patterning within this sequence. Certain func-
tions, for example, clearly go together as pairs, such as a prohibition and
its violation (2 and 3), struggle and victory (16 and 18), and pursuit and
deliverance (21 and 22). And clusters of functions are grouped under
general headings. Thus functions 1–7 are potential realizations of the
preparation, 8–10 are the complication, and later general groups include
transference, struggle, return and recognition.
In addition to the thirty-one functions, Propp identifies seven basic
character types or roles:
villain dispatcher
donor/provider helper
hero (seeker or victim) princess (⫹ father)
false hero
Note that an actual character may fill more than one character role (for
example, some individual in the tale may be both villain and false hero)
and of course one role might be filled by several individuals (there could
be several people functioning as helper or villain). Demonstrating the
application of this descriptive apparatus to his corpus of stories in meticu-
lous detail, Propp concludes:
Morphologically, a tale . . . may be termed any development proceeding
from villainy . . . or a lack . . . , through intermediary functions to mar-
riage . . . or to other functions employed as a denouement. Terminal
functions are at times a reward . . . a gain or in general the liquidation of
misfortune . . . an escape from pursuit . . . etc. Each new act of villainy,
each new lack creates a new move. One tale may have several moves,
and when analyzing a text, one must first of all determine the number of
moves of which it consists. One move may directly follow another, but
they may also interweave; a development which has begun pauses, and
a new move is inserted.
(Propp, 1968: 92)
I will not spend time summarizing just how Propp applies this morphology
to the particular tales in his corpus, since my main purpose here is to
outline what he means by ‘function’, ‘role’, and ‘move’, so that we can
identify similar elements in other stories. And the striking thing is that
certain fictions rather remote from the Russian fairytale do seem to lend
themselves to Proppian analysis without too much strain (see the ‘Notes
and exercises’ section at the end of this chapter). To take an example from
popular culture, consider the Star Wars film trilogy: without itemizing all
the Proppian functions and moves, the characters filling six of the seven
core roles are easy to list:
Villain: Darth Vader

Dispatcher: Luke’s uncle
Donor/provider: Obi Wan Kenobi (magical power provided is the
Force)
Helper: Yoda
Hero (seeker or victim): Luke Skywalker
Princess (⫹ father): Leia
You can apply the Proppian categories to any composed narrative, across
the whole range from The Iliad and the Bible to Hollywood action movies,
to TV series like Buffy the Vampire Slayer, and children’s stories.
The following story written by a seven-year-old child, with Proppian
functions appended on the left, may serve to demonstrate how easily and
appropriately Propp’s grammar can fit simple tales.
Initial situation 1 Once upon a time there was a bunny named

Benjie
⫹ magical agent 2 and she had magic powers.
冦
3 One day she was walking in the woods
Departure and a
4 bunny boy appeared
Translation 5 and they went together for a walk
Reconnaissance 6 and a man appeared with a big net
冦
7 and he got the two bunnies and went in a big
ship.
Villainy
8 Poor bunnies.
9 They were caught now.
Struggle 10 But right then the girl bunny tripped the man
Villainy nullified 11 and they got free once again.
Reward 12 So the boy bunny thanked the girl bunny for
saving him.
冦
13 The boy bunny asked the girl bunny to marry
him
? Equilibrium 14 and she said yes.
(Wedding) 15 So they had six bunny babies
16 and they lived happily ever after.
(text and analysis from King and Rentel, 1982.
See Christie et al., 1984)
If Propp’s schema fits the above story with eloquent ease, we might now
put it to work on a far more complex tale, that of ‘Eveline’, from Joyce’s
Dubliners collection. (I will be discussing this story in future chapters in
relation to a number of issues.) Propp’s very first function seems almost
uncannily relevant:
One of the members of a family absents [himself] from home.
Relevant, but not applicable mechanically. We might say, for instance, that
the ‘action’ of ‘Eveline’ is a dramatization (chiefly a mental dramatization)
of a stage within that first function:
One of the members of a family reconsiders a decision to absent

herself from home.
While Propp’s fairytales proceed through developmental actions, ‘Eveline’

is very largely a mental projection, both forward to possible future events
and backward to actual past ones: remembered before-events and imag-
ined after-events. Notice how her opening revery, up to sentence 24, is a
conspectual review of past circumstances as they impinge on Eveline’s
present situation. As Propp notes, a story has to begin with an initial situ-
ation, one into which an element of disequilibrium is introduced by the
function of absentation or another of the seven preparatory functions. But
in ‘Eveline’, it is the situation itself that is extensively dwelt upon, and
none of the first 23 sentences appears to constitute a narrative-propelling
function. Nor does this tendency lapse with sentence 24. It is just that with
sentence 24 comes confirmation that introspective indirect discourse – the
processes of thinking about living out a narrative of functional departure
from the current and continuing situation – is the chief narrative mode
adopted in this story (a mode I will discuss further in Chapter 4). But the
emphasis on elaboration of the multiple habitual circumstances that com-
prise the initial situation, mere prologue to a story, remains.
However, we can do some reconstruction of a simple developmental
story when Eveline’s thoughts turn to Frank. She meets him, is taken out
by him, at first finds this merely ‘an excitement’ but later begins to like
him. All of that, one imagines, would count as simply one function in
Proppian terms – ‘The heroine meets with a (benevolent?) stranger.’
Eveline’s liking for Frank seems related, not wholly ironically perhaps, to
the latter’s implied story-tellings, which themselves form a skeletal story:
He had started as a deck boy, had sailed through the Straits of Magellan,
had fallen on his feet in Buenos Aires, and had returned to the old country
for a holiday.
If the heroine’s association with Frank is a first function, the second
and third must be her father’s discovery of that association and interdic-
tion of its continuance: ‘[he] had forbidden her to have anything to say to
him’. But subsequently (function 4?) ‘she had to meet her lover secretly’,
so that now (function 5?) she was to go away with him to Buenos Aires
to be his wife.
At this point in the text it seems that the gap between the story outlined
in the two paragraphs above and the character’s current reflections closes,
for the next series of thoughts that are reported, incidents of family life
(centred on her mother) both happy and grim, are the direct trigger of the
‘sudden impulse of terror’ she feels – a psychological impulse to act
(‘Escape! She must escape!’) quite as real and compelling as an encounter
with any forest-dwelling villain. Of course in this psychological story
impulses are not pure and simple, but complex and clashing – a counter-
impulse is to stay, keep her promise to her mother, and submit to a ‘life of
commonplace sacrifices’. The final paragraphs are all about that clash of
impulses, the ‘maze of distress’ that renders her helpless and inert, unable
to respond in any way to Frank’s summons.
But while, in one light, Eveline’s refusal to leave at the story’s close can
be designated as helpless failure, as an abortive move or episode or, worse,
a succumbing to the villainy of oppression at home and at work; in another
light, with Frank as the villain, her rejection of him may be viewed as
‘manipulation resisted’, as a positive act of sisterhood uniting Eveline with
her mother. We shall find the pattern of these contrary readings neatly
highlighted by Greimas’ typology of character roles (which he calls
actants), to be discussed in Chapter 4.
But now, and not for the last time, we may want to raise the question of
reductivism. Is not Propp’s bold anatomy of fairytales a procedure which
hopelessly distorts, since it sets aside the important and necessary cultural
context in which these tales occur, and also ignores the varying details of
stories? Propp might retort to this that a structuralist/morphological
approach has to ‘reduce’, and quite explicitly sets out to shear off the
detail, the non-essential. ‘Essential (and non-essential) with regard to
what?’, we might ask. Essential to the meaning, or rather, to meaning,
would come the reply! For just as the phonologist, in positing all the core
and distinct sound units of a language, the phonemes, deliberately dis-
counts all the phonetic variation which does not constitute a meaning-
bearing, word-changing variation, so the story structuralist will argue that
he is identifying the basic narrative units of a mini-language, here the lan-
guage of Russian fairytales, and in so doing is identifying the essential con-
ditions within which story meanings can arise.
How tenable is the analogy? The strength of the phoneme case is that
language use is something it’s hard to stand outside of: we have strong intu-
itive judgments about what are and are not English sounds, English words,
and so on. It’s less easy to see, or insist upon, a sharp boundary to any set of
stories supposedly covered by a Proppian analysis – although the boundary
was sharp enough as far as Propp himself was concerned, since his corpus
was quite specific: just those 115 stories. And, relatedly, we might ask why
thirty-one functions? Why not thirty or thirty-two? Because, Propp implic-
itly answers, just thirty-one functions are needed – for the given corpus. But
what, we might still ask, are the grounds of this ‘need’?
The thirty-one functions identified (with neither duplication nor unjus-
tified merging of types) are claimed, largely intuitively, to be the only func-
tions necessary to specify the essential action structure of the stories in the
corpus. The question is whether such intuitionism is defensible, or whether
the whole descriptive apparatus is invalidated. But we need also to keep in
mind just what the goals and expectations of a Propp or Barthes-based
analysis are. The business of really getting at our intuitive judgments
(rather than our public and conditioned ways of talking about plots and
plot structures) will always present difficulties. But we do readily find
groups of readers (even whole communities) disclosing substantial agree-
ment over what is essential and non-essential in plot, characterization and
so on – disclosing, in short, a common grasp of structure. This generality of
agreement and commonality of grasp are the essential justification for the
inductive speculations of Propp, Barthes, and others.
2.3 Barthes on narrative

Entirely appropriately, Barthes’ famous ‘Introduction’ (1977; first pub-
lished in French in 1966) begins with an argument about inductive versus
deductive methods (in linguistics and narrative study), and defends the
latter, despite the inevitable provisionality entailed in moving from the
particular observations to the general hypothesis. Barthes writes:
Narrative analysis is condemned to a deductive procedure, obliged

first to devise a hypothetical model of description (what American lin-
guists call a ‘theory’) and then gradually to work down from this
model towards the different narrative species which at once conform
to and depart from the model.
(1977: 81)
He suggests that linguistics ‘seems reasonable’ as a founding model for
narrative analysis, but notes that discourse study will require a ‘second lin-
guistics’ going beyond the sentence. But he does posit a homological rela-
tion between sentence and discourse, at least as far as semiosis –
‘message-bearingness’ – is concerned:
A discourse is a long ‘sentence’ . . . just as a sentence . . . is a short ‘dis-

course’.
(Barthes, 1977: 83)
More particularly, Barthes emphasizes the need to separate different

levels of analysis, and the need for a hierarchical typology of units, in this
early essay. He proposes three major levels of narrative structure:
1 functions (as in Propp, Bremond);

2 actions (by which he refers to ‘characters’, rather as Greimas – dis-
cussed below – refers to them as actants);
3 narration (equivalent to what we have termed discourse, discours, or
sjuzhet).
What follows is almost entirely to do with the first level, that of function,
that by which narrative is ‘driven’. The essence of a function is ‘the seed
that it sows in the narrative, planting an element that will come to fruition
later – either on the same level or elsewhere, on another level’ (1977: 89).
Thus function is teleological, by which we mean it is concerned with the
long-term goals or purpose (intellectual and moral as well as actional) of a
narrative. Where we assume that a narrative is thus teleological, we tend
to look for (and find) material dispersed through the narrative that is
designed to support, sustain, and lead to that goal. Function is the means
of achieving this overarching coherence in a narrative, rather than any
merely local or adjacent cause-and-effect logic.
Barthes proceeds to distinguish two types of functions: (a) functions
proper (which we might call ‘Propp-type functions’); and (b) indices,
which are a unit referring
not to a complementary and consequential act but to a more or less

diffuse concept which is nevertheless necessary to the meaning of the
story.
(Barthes, 1977: 92)
They include indices to characters’ psychological states, notations of

‘atmosphere’, and so on. While functions proper are distributional,
sequential, ‘completed’ further on in the story – and so have a kind of syn-
tagmatic ratification, indices are said to be integrational, hierarchically-
oriented, realized by relating them to some higher, integrated level, a
paradigmatic ratification. On a broad continuum, Barthes suggests, there
are heavily functional narratives such as folktales, to be contrasted with
heavily indicial ones such as psychological novels.
A further cut is now introduced. Functions proper are of two types:
(a)1 Cardinal functions or nuclei or, to use Chatman’s useful term,

‘kernels’ (in Chatman, 1969): these are real hinge-points of narra-
tives, moments of risk (when things can go ‘either way’); they occur
consecutively and entail important consequences.
(a)2 Catalysers (not the best of terms): these fill in the narrative space
between nuclei, and are described as parasitic and unilateral by
Barthes, areas of safety and rest. For example, a ringing telephone
or a delivered letter may herald a real nucleus in a story – and a pre-
liminary ‘hinge’ would be whether the summons is answered or not,
the letter opened – but all sorts of ‘business’, prevarications and
accompaniments may surround that action as catalysers.
Indices, too, can be sub-classified as:
(b)1 Indices proper (charged with implicit relevance);
(b)2 Informants (depthless, transparent, identificatory data).
Indices involve an act of deciphering, the reader is to learn to know

a character or an atmosphere; informants bring ready-made
knowledge . . . their functionality is weak.
(Barthes, 1977: 96)
Finally Barthes notes that a unit can be a member of more than one class
at a time: one could be both a catalyser and an index, for example. And he
notes that in a sense nuclei (kernels) are the special group, with the other
three unit types being expansions of nuclei. Nuclei provide the necessary
framework, the other three fill it out.
Barthes goes on to appeal for descriptive study not merely of the ‘major
articulations of narrative’ but of the organization of the smallest segments,
which he sees as combining into coherent sequences:
A sequence is a logical succession of nuclei bound together by a rela-

tion of solidarity: the sequence opens when one of its terms has no
solidary antecedent and closes when another of its terms has no conse-
quent.
(1977: 101)
For example, ‘having a drink’ is suggested as a closed sequence with the

following nuclei: order a drink, obtain it, drink it, pay for it. (But is paying
as obligatory and integrated as the other three nuclei?) Now the business
of seeing a sequence in such a string of reported events, and labelling it as
‘having a drink’ (rather than, say, ‘quenching one’s thirst’ or ‘making
oneself socially available’) is, for Barthes, the kind of projective interpre-
tive activity readers are always doing in their narrative processing. It’s
what he later subsumes under his proaieretic code (Barthes, 1970).
Naming is a key act of mental processing, under the assumption that the
reader does not remember everything they read, but remembers selec-
tively, largely on the basis of events’ importance. The reader registers, and
may even verbalize, the broad scenario, the main threads, in a narrative –
a kind of incremental and revisable précis-making and paraphrasing. It
does not appear that Barthes drew on psycholinguistic evidence in his
assumptions and theorizing of ‘sequence naming’, but there are some
interesting parallels between his proposals here and more recent psy-
cholinguistic research on narrative (see the brief mention in the ‘Further
reading’ section of Chapter 7).
An invaluable critical demonstration of the Barthesian machinery
applied to a text is Chatman (1969), which sets it to work on the same
story, ‘Eveline’, that I discuss extensively in this chapter. Chatman identi-
fies just eight core narrative functions or kernels in the story; I list these
below, with Chatman’s interpretive labellings appended:
1 She sat at the window watching the evening invade the avenue.
(SITTING AND LOOKING)
2 One time there used to be a field there in which they used to play . . .
(REMINISCING)
3 Now she was going to go away . . . to leave her home. (REHEARS-
ING THE DECISION TO GO)
4 Was that wise? (QUESTIONING)
5 She stood among the swaying crowd in the station at the North Wall.
(PREPARING TO EMBARK)
6 Out of a maze of distress, she prayed to God to direct her, to show her
what was her duty. (INDECISION CHANGING TO ANXIETY)
7 She felt [Frank] seize her hand. (FRANK’S URGING HER TO GO)
8 No! No! No! (REFUSAL)
This skeletal structure certainly tells a story, with the required connected-
ness of stages or moves. And Chatman’s capitalized glosses of the kernels
or functions proper reflect the story’s attention to mental activity rather
than physical change: the story is structured around reflection, reminis-
cence, thinking about doing something, and getting ready to do something,
rather than on actions themselves. In addition, there is ample evidence
that indices and informants of the character Eveline, and catalysers
accompanying the narrative development that is principally ‘driven’ by the
functions proper, all contribute to an integrated presentation. Even the
smallest textual details, we might argue, play a role. Notice, for example,
the words swaying and maze that are used in the course of functions 5 and
6 above. Besides their specific application in these sentences, they are also
indexical, with reference to the entire story, of indeterminacy and vacilla-
tion, and encircling confusion, respectively. Many readers, casting
the interpretive net wider, will proceed to extract dominating indices,
informants, and so on, as they build up a sense of the basic structure of the
Dubliners narrative as a whole. Among the indices that many generalize
from ‘Eveline’ are qualities of dependence, submission to duty, and inef-
fectuality.
Despite the attractions of Barthes’ basic four-way categorization of
narrative material, problems remain concerning the replicability of
Barthes’ model, and in particular over how (by what criteria) we can
confidently judge what is and is not a nucleus, a catalyser, an index
proper and an informant. Some of these problems can be outlined by
examining Chatman’s explanation of the difference between kernels
(i.e. nuclei) and catalysers. Kernels are said to be hinges, alternative
path openings, and so on, while catalysers (better, ‘satellites’) are non-
essential actions (‘business’) accompanying the kernels, but of no larger
prospective consequence. In the extract below, Chatman italicizes the
alleged kernels:
One of the telephones rang in the dark room. Bond turned and moved
quickly to the central desk and the pool of light cast by the green
shaded reading lamp. He picked up the black telephone from the rank
of four.
What we might question here is the assumption that the phone-ringing and
answering are inherently nuclei, that Bond’s moving across the room is
inherently secondary. Such decisions can only be made retrospectively, it
seems, in the light of a fuller scanning and assessment of adjacent text. But
if the categorization is only retrospective, then it did not guide our reading
and is of lesser psychological validity. If structuralist analysis (in terms of
kernels or functions) is to be of value, we clearly need a robust explanation
of the kind of interest or question that a kernel provokes, an explanation
that sets out the bases of our stronger and weaker impressions of kernel-
hood. I will presently propose that we can initially work ‘longitudinally’,
like a reader, through a narrative text’s grammar (in a broad sense of that
term), in pursuit of more local and grammatically-cued marks of the core
narrative events. Ultimately, however, when we offer a determination of
what the kernel narrative utterances are, we have to operate holistically
and teleologically. And this assumption is necessary in order to apply any
criterion of well-formedness.
An analogous truism is that you can’t parse a sentence until you’ve read
it. Notice, however, that you can start to parse a sentence before you have
finished reading it – I think we typically do. But we know that the analysis
is provisional, may not ‘go through’ if we find a configuration of relations
that is out of the ordinary, the unmarked form. We know, then, not to put
too much trust in our parsing until the reading is complete, and we’ve seen
all the structure there is to see. Thus as we read:
It was John who

we expect that John is the ‘underlying subject’, the ‘doer’, made the focus
through a clefting device. But if we read on and find John is in fact the
underlying object, the ‘done to’ –
It was John who the boys attacked.
– we are not at all troubled by the need for revision. Amending the truism,
then:
You can begin to parse a sentence or text before you’ve finished

reading it, but you know you may need to revise your analysis.
The well-formedness test mentioned above is related to the grammatical

distinction between obligatory and deletable material. In narrative
‘grammar’, there is a similar assumption: while catalyser and indexical
material on character and setting are deletable, the functional kernels are
obligatory material which, on their own, constitute a coherent ‘bare’ narra-
tive (and recognizably related to the full version) – the discarded bits
would constitute no sort of coherent narrative at all.
The upshot of these qualifications and reservations should be that we
see more clearly that kernels and catalysers are not so much textual
‘givens’ as analytical constructions; and as Culler explains using a Saus-
surean phrase, they are ‘relational terms only’:
What is a kernel in one plot or at one level of description will be a

satellite at another. . . . In ‘Eveline’, for example, the actions of the
past which the heroine recalls could be organized into kernels and
satellites, but within the story they become satellites or expansions of
a kernel such as ‘weighing the evidence’. . . . We must accept that we
recognize kernels only when we identify the role of an action in the
plot or, to put it another way, promote an action to a constituent of
plot. . . . One cannot determine the role or function of an action
without considering its consequences and its place in the story as a
whole.
(Culler, 1975b: 135–6)
In this retrospective process of sense-making and plot-determination,

Barthes’ notion of ‘sequence-naming’, or something similar, must be cru-
cially involved. Thus as we read through ‘Eveline’ we must be construct-
ing a model or scheme into which the disparate propositions (sitting at
the window, leaning her head, listening to the man’s footsteps, thinking
about childhood days) can somehow fit. Much of this model-building will
be heavily guided by our cultural background, by what we think – and
what we think the writer thinks – is important in life, about what is
‘normal’ behaviour, and so on. When we do any text-interpreting, it is
not done in vacuo, a truly solitary and private interrogation of the text.
Both texts and readers are inescapably shaped or framed by prior (but
not fixed or eternal) cultural assumptions as to what is significant,
salient. It is because there is always this context of cultural significances
and saliences ‘around’ all our involvements with narratives that Barthes
says we have ‘a language of plot’ within us even before we approach any
particular story.
An outward sign of having made sense of the textual data is the pro-
duction of reasonable paraphrase, i.e. a paraphrase that neither we nor
other readers find incongruous or absurd. The paraphrase could be very
long, many times the length of the original (as most literary critical art-
icles on ‘Eveline’ are), or it could be single-proposition length, or just a
single word. The opening paragraphs of ‘Eveline’ could be paraphrased
as:
A young woman reflects on her past life
or as:
Reflection
But (and here the analyst’s dilemma begins to look remarkably like that of
those who attempt to identify the structure of spoken discourse) does it
make sense to work ‘from the bottom up’, as Barthes and Chatman claim
to be doing? Do we not have to work ‘from the top down’, i.e. first setting
up some broad hypothesis as to what happens in the story as a whole, thus
specifying the top-level constituents of story, then trying to move down the
hierarchy, analysing so as to separate out the bits that have been com-
plexly bound together? Or is this dramatic opposition of top-down and
bottom-up itself a misdirection? Culler’s objection to Barthes’ model is
quite simple but fundamental:
[It] remains strangely atomistic, through the lack of any specification

of what one is moving towards as one collects kernels and satellites
and groups them into sequences. . . . [And] Chatman is able to pick
out [the sentence One time there used to be a field there] as a kernel
only because he has some sense of an abstract structure towards
which he is moving. . . . . What the reader is looking for in a plot is a
passage from one state to another – a passage to which he can assign
thematic value.
(Culler, 1975b: 138–9)
Again, we are back to a linked before and after. The before and the after
can be labelled, if we like, oppositionally as problem and solution, or logi-
cally as cause and effect. Or more neutrally, we might simply label them as
situation 1 and situation 2, linked by one or multiple experiences.
2.4 Plot-summarizing: modelling intuitions
In his paper on ‘Defining narrative units’, which focusses on problems of
generalizability and adequacy in certain proposed models of plot analysis,
Jonathan Culler concludes:
Competing theories of plot structure can only be evaluated by their

success in serving as models of a particular aspect of literary compe-
tence: readers’ abilities to recognize and summarize plots, to group
together similar plots, etc. This intuitive knowledge constitutes the
facts to be explained.
(Culler, 1975b: 127)
To this characterization we should add a couple of caveats. The first is

that, even more immediately apparent than in relation to the notion of lin-
guistic competence on which it is loosely based, literary competence seems
to be very largely learned rather than innate, and markedly culture-
specific. Thus when we talk of ‘readers’ abilities’ we have to keep clearly in
mind that we are talking only of the acquired and developed ability of a
group of readers, rather than some universal mental ability, comparable to
the near-universal ability to walk or subtract.
The second related point is that what it means to call these culture-
specific abilities ‘intuitive knowledge’ remains unclear. People can be
better or worse at plot-summarizing, and can get better at it: does this
suggest an intuitive faculty? I would be inclined to take a far more beha-
vioural approach and argue that we get good at plot-summarizing largely
because, in our kind of world, constructing and communicating summar-
ized plots is a valued skill. By ‘our kind of world’ I mean a world in which
children and adults frequently face examinations; in which we often have
our recall of events put to various kinds of test; in which it is common to
share ‘at second hand’ a narrative whose performance one’s addressee has
missed; and in which there is particularly sharp awareness of the universal
constraint that influences the shape of so much of our behaviour, namely,
limitation of time. Would plot-summarizing be such a valued skill in a
settled, integrated oral community, where a set time and social space was
reserved for storytelling? There, to précis a plot might be regarded as either
incomprehensible or, worse, proof that one was a very poor storyteller.
Thus we need to qualify the idea of ‘intuitive competence’ in relation to
narratives. It is clear that people, with their various non-universal cultural
predispositions, can and do become adept at producing and understanding
plot summaries. This amounts to saying that they develop community-
validated skills in specifying the more important characters and events in
narratives. Most crucially of all, they get good at identifying what, relative
to their own frameworks of world knowledge and cultural assumptions, is
the ‘main point’ of a story. One of the questions we need to try to answer is
how on earth they do this. That they should do it is surely no great surprise:
it is merely an instance of the process of ranking or ordering things that we
do all the time in all sorts of activities, making rational decisions about
what things most need our attention – again, given our limited resources of
time, money, and energy.
How does all this apply to ‘Eveline’? Of this story, Culler suggests that
readers can construct and agree on hierarchies of appropriate plot sum-
maries, from very succinct to rather detailed ones. Indeed, if a culturally
homogeneous group of people is set the task of summarizing a short story,
in 100, or 40, or just 20 words, the degree of agreement over what to
mention and what to discard is often gratifyingly strong (sharp disagree-
ment may be due either to weak skills of summarizing or some covert sub-
cultural clash). In ‘Eveline’ Culler notes how various potentially important
plot incidents are eventually rejected by us as not central to plot, more an
indirect description of Eveline’s consciousness: for example, the man out
of the last house going home, Eveline watching from her window. But ‘She
had consented to go away. . . . Was that wise?’ is different:
[This] is immediately recognized as an important structuring element

which enables us [to re-interpret preceding material] . . . and to struc-
ture the material.
(Culler, 1975b: 130)
Culler concludes his discussion making reference to an implicit three-level

hierarchy of actions (characterizing, attributive, and plot-determining),
listed in reverse order of importance:
As we move through ‘Eveline’ we must decide which actions serve

only to characterize her and the situation in which she has placed
herself, which of these are crucial attributes involved in the change
compassed by the plot, and which actions are in fact crucial as actions.
(134)
Perhaps we must. But in addition most readers reading ‘Eveline’ know

that it was written by James Joyce, that it is therefore ‘Literature’, and
that failure to nominate or attend to a clear developmental plot, in such
fiction, has its own cultural warrant. In other words, perhaps ‘Eveline’ is
too category-marginal, too de-automatizing of plot logic, to be a suitable
exemplification of our standard ability to summarize. (A rather more
straightforward story, also about leaving home and the familiar, is sup-
plied in the first exercise appended to this chapter.) When asked what
happens in ‘Eveline’ it is not absurd to reply ‘I’m not entirely sure’,
although such an account of the plot of, say, a James Bond novel, would
be felt to be defective. In the case of literary narrative fiction, perhaps
more than elsewhere, the force of James’s famous observation seems
especially telling:
What is character but the determination of incident? What is incident
but the illustration of character?
(1963: 80)
We can follow Culler, then, in accepting Propp’s teleological conception of

plot structure – a conception in which evaluation of acts is held in
abeyance until their significance within the encompassing sequence is per-
ceived. Such retrospective definition of units relies very much on internal-
ized cultural models shared by readers (thus, Eveline’s sitting by a window
and recollecting the past needs to be recognized as a cultural index – a
cliché even – signalling introspection, serious reflection, and often also
entrapment and longing). And on another level, readers need to see
leaving home (as Eveline contemplates doing) as emotionally and cultur-
ally significant (with, perhaps, a special resonance for Joyce and readers of
Joyce).
Propp is the first of many to be structuralist in certain of his procedures
of analysis, outlined above, but quite intuitive in his grounds for asserting
particular similarities or differences. He is intuitive, for example, in assert-
ing that various characters in various stories are mere variant concrete
realizations of a single abstract role (no replication test using informants
was ever used by him to support his judgments); and he seeks our intuitive
assent to those claims. His study remains a pathbreaking exploration of
the narrative ‘competence’ that readers seem to share: we do have quite
definite ideas about the basic plots of the narratives we read or hear; we
do have an ability to summarize plots; we can often agree in identifying
what is not essential to plot; and we can identify ‘transformationally-
related’ plots.
2.5 In search of the grammaticization of plot structure

To restate the complication that ‘Eveline’ presents: as we read this story
we’re not entirely sure whether we should – as normally – be looking for
incidents, or, instead, attend to character. But let us proceed by trying to
see why, as many readers claim,
She had consented to go away, to leave her home. (sentence 24)
is one of the more important disclosures of plot structure in the story. In

doing so, we should always remember that ‘importance to plot’ is a co-
text-relative criterion: in some sorts of stories, the fact that an individual
had consented to leave home might be of slight importance given an array
of more dramatic and immediate events: cf.:
She had consented to go away, to leave her home, while the place was
being redecorated.
This example is a reminder of the importance, already emphasized, of the

teleological status of events in narratives. On the other hand it is fair to
say that ‘permanently leaving home’ seems a significant and ‘tellable’
event in most cultures.
In probing the high plot-structuring status of sentence 24, we should
first compare it to the 23 text sentences that precede it. When we come to
look at things more holistically or teleologically, we must relate sentence
24 to all that follows it as well as all that precedes it, but it should be worth
first approaching the sentence the way a first-time reader would, ‘from the
left’, seen in the textual light only of those sentences that precede it. And
what we find, in those sentences, are various grammatical cues that deflect
or argue against treatment of them as carrying crucial narrative events. We
may begin by looking at the finite verbs of the opening sentences (my
emphasis of the finite verbs):
She sat at the window watching the evening invade the avenue.(1) Her
head was leaned against the window curtains and in her nostrils was
the odour of dusty cretonne.(2) She was tired.(3)
All are distinguished by being either stative verbs, or at least verbs of static
description, with no intrinsic implication of change. Stative verbs – be,
know, realize, suppose – contrast with dynamic verbs – sit, leave, go –
grammatically, in that the former resist progressivization while the latter
progressivize freely: compare *She was being tired with She was leaving
home. Stative verbs describe states of affairs, and ‘passive’ processes of
cognition or perception (he knew Spanish); dynamic verbs depict events
and active processes, and ‘active’ mental processes (he was learning
Spanish). In the extract from ‘Eveline’ above, even the dynamic verbs (sit,
lean) are used in static ways.
One questionable assumption made here is that, in looking for the
reporting of narrative events, we should look particularly but not exclu-
sively at finite verbs. Almost any major part of the clause can express
change-of-state, including non-finite verbs (underlined):
(a) Chick watched the men confront McAndrew, shoot him, carry his
body on the back of a pack-mule out to the bottom piece, and bury
him.
And narrative events can also be expressed through nominalizations

(explained more fully in 8.5), here underlined:
(b) Chick saw a confrontation between the men and McAndrew, the
latter’s death by shooting, the transportation of the body by mule to
the bottom piece, and its burial.
But I think most of us would agree that, for most purposes, (a) works and
feels better than (b) as a narrative, and that a third alternative which
reported the confrontation and shooting and so on through finite main
verbs would probably be preferable to either. In making those judgments
we are justifying the argument for looking for narrative events particularly
in finite verbs. The finite verb, in short, is the unmarked, preferred and
unexceptional vehicle for expression or realization of plot events, while
the other forms I have mentioned are two of the marked and noticeable
alternatives. Accordingly, and while mindful that our assessment is proba-
bilistic and corrigible in the light of later text and other circumstances, we
calculate that, for example, the embedded clause ‘the evening invade the
avenue’ (embedded under sat, a verb of static condition) does not express
a crucial plot event.
The text continues:
Few people passed.(4) The man out of the last house passed on his
way home; she heard his footsteps clacking along the concrete pave-
ment and afterwards crunching on the cinder before the new red
houses.(5) One time there used to be a field there in which they used
to play every evening with other people’s children.(6)
What of sentence 4? Are there any intrinsic grounds for doubting this
clause’s plot-structuring importance? Perhaps only if we compare it with
the nearly synonymous
A few people passed.
The latter could be used to describe the passage of a particular group of

pedestrians (e.g. ‘A few nuns passed’), where we respond by wondering
‘What sort of people?’. This is not true of the textual alternative without
the indefinite article, which lacks the required sense of deictic (explained
at 3.4) or spatiotemporal specificity; it is, rather, a summative comment,
retrospective and hence descriptive, rather than narrative. And non-
specific to the point of carrying a null or negative implication: compare He
invited a few people over for dinner [⫽specific people on a specific occa-
sion] and He invited few people over for dinner [⫽almost none, in general].
And the term denoting the participants here, people, is one of a group of
noticeably general, all-purpose items (others include thing, person, stuff)
noted in Halliday and Hasan (1976). In sum, we are likely to demote
(rather than promote) sentence 4 as a candidate for kernel or function
status, in spite of its finite dynamic verb, because of the overtly non-
specific nature of its only participant.
By contrast the next sentence (sentence 5) does exhibit the kind of
identificatory specificity I am positing as a preferred characteristic of
narrative clauses, particularly in its definite description of a specified
individual:
The man out of the last house passed on his way home.
This incident could be crucial in the plot, even in the face of the following
two counter-signals:
1 In none of the following sentences is the man named, though he is

apparently known: characters crucial to plots are usually denoted (by
name, profession, distinguishing characteristics) rather more specifi-
cally than is done here.
2 Important characters and their important denoted actions are not
usually mentioned and immediately discarded as discourse topics.
There is usually a ‘follow-up principle’, such that one or more following
clauses maintain either the prominent character or his/her action as
some inner clausal element (subject, predicate, object), with the pre-
ferred option being to maintain the character as subject and theme
(theme denotes whatever is the first major constituent in a sentence,
hence the starting-point of the sentence’s message and what the writer
chooses the message to be about; various interesting effects can arise –
e.g. when theme and subject diverge). Here, overriding the typical
option, there is an immediate switch (to she heard his footsteps), itself
then discarded as an extendable topic in the following sentences, which
revert to the distant past:
One time there used to be a field there in which they used to play
every evening with other people’s children.(6) Then a man from
Belfast bought the field and built houses in it not like their little brown
houses but bright brick houses with shining roofs.(7) The children of
the avenue used to play together in that field – the Devines, the
Waters, the Dunns, little Keogh the cripple, she and her brothers and
sisters.(8)
There are basically two grounds for discounting most of this material as
crucial to story development: one is the frequent emphasis on the events
reported as habitual and recurrent, an emphasis that becomes excessive in
the following sentence:
Her father used often to hunt them in out of the field . . . but usually
little Keogh used to keep nix . . .(10)
The second ground suggesting plot non-salience is the pervasive use of dis-
tancing deictic or spatiotemporal markers, reinforcing the relative remote-
ness of the events and situation from the speaker’s (or, in this case, the
thinker’s) present. Distal deictics (one time, there, then, that, and so on) are
remarkably prominent:
Still they seemed to have been rather happy then.(11) Her father was
not so bad then; and besides, her mother was alive.(12)
The above sentences are of course inherently non-event-implicating
anyway since their main verbs are stative and intensive (verbs like be and
appear, which link a following description – which is not a grammatical
Object – with the preceding Subject: e.g. Bill appeared angry, Helen
became an accountant). But the observation that follows recapitulates the
‘extradiegetic’ nature of these reflections: ‘That was a long time ago’. (The
term ‘extradiegetic’ will be properly introduced in the next chapter.) The
text continues:
That was a long time ago; she and her brothers and sisters were all
grown up; her mother was dead.(13) Tizzie Dunn was dead, too, and
the Waters had gone back to England.(14) Everything changes.(15)
Now she was going to go away like the others, to leave her home.(16)
Home! She looked round the room . . .
Again, we have a series of stative relational descriptions in sentences 13

and 14, closing with a distant, past perfect verb, and a timeless (i.e. deicti-
cally unanchored) generic sentence in 15.
Only with sentence 16 do we encounter an utterance grammatically
quite distinct from all those preceding. Introduced by the proximal deictic
‘Now’, in striking contrast to all the previous instances of ‘then’, it is ori-
ented to the speaker/thinker’s narrative present, with an explicitly
dynamic verb in progressive aspect (progressivizability of a verb is the
very test of dynamic status). Not only is the utterance oriented to the
thinker’s present, it also has futuritive force, expressing an intended future
course of action. The links between sentence 16 and the surrounding text
are plentiful; some of these may be briefly listed, making reference to
phrases here shown in italic:
Now she was going to go away like the others, to leave her home.
The ‘Now’ links by contrast with the ‘then’ of times past, recounted in the
foregoing sentences; the ‘go away’ is lexically related to the ‘gone back’ of
sentence 14; the ‘others’ in ‘like the others’ is linked to the previously-
mentioned brothers and sisters, or the Waters, or possibly even her dead
mother and Tizzie Dunn, or any combination of these; and the mention of
‘home’ at the close is echoed in the exclamatory ‘Home!’ that follows. All
such cohesive ties do nothing to detract from the narrative distinctness of
sentence 16 in terms of its deixis, aspect, futuritive force, dynamic verb
expressing a clear change of state, and use of an already textually-
prominent participant (‘she’, Eveline) as sentential subject and theme.
What I have done, perhaps a little laboriously, for these first two para-
graphs could be done for the third paragraph, leading up to sentence 24 –
‘She had consented to go away, to leave her home’ – which analysts agree
is important to plot. Briefly, paragraph 3 displays many of the same
non-narrative characteristics as the previous two, particularly either stative
main verbs or at least ones implying no change of state:
She looked round the room . . . Perhaps she would never see again
those familiar objects . . . He had been a school friend . . . ‘He is in
Melbourne now’.
habitual or iterative processes
which she had dusted once a week for so many years. . . . Whenever he
showed the photograph . . . her father used to pass it.
or both
from which she had never dreamed of being divided . . . during all
those years she had never found out the name of the priest.
(underlining added)
Again, by sharp contrast with all these, and now with the additional impact
of being a near-repetition of sentence 16, comes sentence 24:
She had consented to go away, to leave her home.
But it is, note, an iteration with some differences which make it all the
more salient to plot. For while sentence 16 can be read as the expression
of the subject’s own purely personal decision to act, it now emerges that
another party is involved, and has proposed a specified course of action to
which Eveline has agreed. And if we compare sentence 24 with:
She had decided to go away.

She had agreed to go away.
the subtle semiotic overtones of the verb consent – so prominently used in

quasi-legal discussion of sexual matters – seem quite evident.
In this final section of the chapter I have offered some simple principles
which may guide a reader’s ‘real time’ processing of text in the search for
plot. In this way I have tried to uncover links between grammar and plot-
structure, in the conviction that such links quite typically do exist. But I
repeatedly argue in terms of preferences, expectations, and tendencies (we
expect main events to come in main verbs, we expect main characters to
be designated – recurrently – in individualizing ways, and so on). The pro-
cedure is necessarily fuzzy and provisional, but not haphazard. There is
little that is arbitrary or haphazard about the grammar of a language or
the grammar of stories. And as conceded earlier, this whole exercise of
provisional plot-assessment sometimes needs radical recasting after the
fact and act of reading, when a synoptic and teleological perspective is
adopted.
Further reading
On basic story structure the seminal works to which I have referred are: Propp
(1968) and Barthes (1977). And for a powerful critique of formalism, see Jameson
(1972). Later, Barthes turned away from structural analysis to a more content-
oriented approach, postulating five ‘codes’ that he saw as invoked and integrated
in the creation of story texture (hermeneutic, semic, proairetic or actional, referen-
tial or cultural, and symbolic); his S/Z (1970) shows him applying the codes in an
analysis of Balzac’s story Sarrasine (see Fowler, 1981: 96–129 for discussion).
Chatman (1969) is both a sympathetic overview of the early story-structuralist
work of Barthes and Todorov, and a source of numerous narratological insights
into the story ‘Eveline’, while Culler’s Structuralist Poetics (1975a) contains lucid
critique of structuralist theories of plot. Fowler (ed.) (1975), has several valuable
papers, including ones by Chatman and Culler. Prince (1982; 1991) are reliable
guides to the key elements and proposals in narratological theory, and can be used
in conjunction with the present and subsequent chapters; see also Onega and
Garcia (1996). Chapter 2 of Rimmon-Kenan (1983) covers some of the same
ground as this chapter, and students might like the ‘bivocal’ effect of having two
introductions to the same topic; Rimmon-Kenan also includes many apposite liter-
ary examples. I particularly admire and recommend Chapter 1 (‘Fabula: Ele-
ments’) of Bal (1985) – an advanced and authoritative introduction to both theory
and practice. This text (a translation) requires attentive reading, but is rigorous
and insightful. Also important are the opening chapters of Chatman (1978);
Chapter 4 of Simpson (1997) – see exercise 8, below – presents and applies an
exercise that is simple and extremely effective theoretically and pedagogically,
particularly on very short literary narratives. From within the British text-linguistic
tradition, see Chapters 4 and 6 of Hoey (2001); the latter chapter sets out a ‘matrix’
approach to sequencing in narrative and non-narrative texts. For reflection on nar-
rative structuring and competence from outside literary narratology, Bruner’s
work is particularly to be recommended, focussing on narrative competence as an
enabling resource in infancy (Bruner, 1986; 1990).
Notes and exercises

1 Below is part of the written text of a text-and-pictures story for children by Nigel
Snell entitled Julie Stays the Night (London: Hamish Hamilton, 1982):
One morning the telephone started to ring.

Mummy answered it. She said, ‘It’s Sally’s Mummy. She has invited you to
go and stay. Would you like to go?’
‘Oh yes,’ said Julie. ‘I’d love to.’ And she ran to tell Alexander, her pet
beetle.
The days went by and soon it was time to pack. Julie began to feel scared.
She was going away from her Mummy and Daddy and her toys, her bed and
her home.
Julie started to cry. ‘Don’t be sad,’ said Mummy. ‘You will love it when you
get there.’
Julie wasn’t so sure.
She decided to take Alexander along in a matchbox, just in case.
Mummy took her to Sally’s house by train. Julie felt very homesick.
But when they got to Sally’s, her mother was very kind.
They had a big tea and played with lots of toys.
That night Julie and Sally were both allowed to stay up an extra half hour.
When they went to bed Julie showed Sally her beetle, Alexander, but he
crawled out of the matchbox and went down a crack.
‘He seems quite at home here,’ thought Julie.
Attempt an analysis of this story, labelling events and states in terms of one, two,
or all three of the following: Propp’s functions; Barthes’ functions and indices;
my proposal about the grammaticization of core narrative clauses. Which
method or combination of methods seems to you to highlight basic story struc-
ture most effectively?
2 Some of the ways we can use Propp, or a Proppian approach, are noted in
Dundes’s introduction to the Morphology. If structural analysis is not to be an end
in itself, it seems vital to relate the forms Propp found to the culture (Russian,
Indo-European) out of which they emerged. Dundes speculates pertinently:
Does not the fact that Propp’s last function is a wedding indicate that
Russian fairy-tale structure has something to do with marriage? Is the fairy
tale a model, a model of fantasy to be sure, in which one begins with an old
nuclear family . . . and ends finally with the formation of a new family? . . .
Propp’s analysis should be useful in analyzing the structure of literary forms
(such as novels and plays), comic strips, motion-picture and television plots,
and the like. . . . Do children become familiar enough with the general
nature of fairy-tale morphology to object to or question a deviation from it
by a storyteller? . . . Finally, Propp’s scheme could also be used to generate
new tales.
(Dundes, 1968: xiv–xv)
Apply Dundes’s questions to the next episode of any television series that you
watch. How predictable is the villainy or lack that is grappled with, and what
form does the resolving ‘marriage’ typically take? Compare these predictabilities
with the functions (villainy, ‘marriage’, etc.) of a rather less formulaic series.
3 Taking Propp’s 31 functions and 7 roles, attempt to apply them to one or two
literary narratives you know well. Determine how much adjustment of the
scheme is needed for it to capture the basic story of one of the following: a
medieval morality play; one of Chaucer’s Canterbury Tales – e.g. ‘The Wife of
Bath’s Tale’; one of Shakespeare’s comedies or problem plays; any Dickens
novel, though Oliver Twist, A Tale of Two Cities and Bleak House seem
particularly manageable candidates; Pope’s ‘The Rape of the Lock’; Melville’s
Moby Dick.
In fact the narratives you choose to analyze could be of any type. It might be
your grandmother’s oft-rehearsed story of her pursuit and courtship by your
grandfather (perhaps there’s no villain here, but presumably it concludes with
marriage), or you could look at any simple children’s story you might have to
hand. But if even these narratives look at first glance unmanageably different
from Propp’s stories and morphology, you could certainly begin with a story
which, intuitively, you feel is not too unlike those examined by Propp, e.g. a
Grimm or Hans Andersen folktale.
How well does the Propp ‘grammar’ fit your chosen text? Where do the prob-
lems and difficulties lie? As you move away from traditional oral-based folk-
tales, you should expect that the kinds of functions and roles involved in the
basic story structure will be rather removed from those of Propp: removed, but
not unrelated. And that’s the point: in looking for the kinds of functions and
roles from which multifarious stories are generated, we are indeed committed to
identifying relatedness of structure.
4 Are the blurbs that accompany books of fiction – the brief fliers or advertise-
ments for books that appear somewhere on their jackets – plot summaries? If
not, what are they? Analyze the blurbs of three books of fiction, thinking about
this question in particular. Whatever differences of detail you find, it will
become newly apparent to you just how structured, carefully-worded and
important such blurbs are. In a bookshop, taking a quick glance at the merchan-
dise, the blurb may be the only part of the text that the critical potential pur-
chaser has time to read. You might move on to compare your fiction blurbs with
the blurbs that accompany popular non-fiction, academic literary criticism, lin-
guistics textbooks, science textbooks, biographies, etc.
5 Look analytically at two or three newspaper reviews of new novels. Do these
reviews contain a full summary of the plot, or only an incomplete one? Is there
any discernible pattern to this variation, with certain types of fiction getting
fairly full summaries, others almost none at all? Why might it be that certain
reviews start a story summary (often identifying the characters and setting but
only the first few events to occur) but do not complete it? Compare and contrast
the reviews, in a range of publications, of a current notable fiction. Where in the
review does the plot summary come, if it appears at all? Does it come all in one
chunk, or intermittently? How much agreement does there seem to be over what
to put into the summary, and where do there seem to be differences of empha-
sis? What possible grounds might there be for those differences? How might all
this relate to the idea raised in this chapter that plot perception may be cultur-
ally relative?
6 In her review of Fowler (ed.) (1975) Smith challenges its seeming assumption of
‘substantial agreement’ over plot summaries:
The extent to which we produce comparable plot summaries may have a

good deal to do with who ‘we’ are and how we learned both what ‘plots’ are
and also what it means to ‘summarize’ them. It is doubtful, for example, that
a South American tribesman would produce the same plot summary of one
of his myths as would Claude Levi-Strauss. Or to put it another way,
wouldn’t our explaining to the tribesman (or, of course, to a child) what we
wanted when we asked for a ‘plot summary’ be the same as telling him how
to ‘process’ the story the way ‘we’ have learned to?
(Smith, 1978: 183)
Take any narrative familiar to you and your friends (it doesn’t have to be highly
literary), and attempt a summary of the story which is not a plot summary but
still seems to you to be a fair summary in some respects (e.g. of the mood or
tone). Now present your summary to one or more of those friends and ask them
their opinion of it as a summary of the story. Your problems in producing a non-
plot summary, and their reactions to it, should be instructive.
7 There are other approaches to the ‘basic grammar’ of stories than those dis-
cussed in this chapter, perhaps most notably the work of Prince (1973; 1982);
and a vast body of psycholinguistic research into our mental modelling of story
format. This research hypothesizes that we construct schemas, or frameworks, of
archetypal stories, which we ‘internalize’ and use as a mental aid when we
attempt to comprehend or recall particular stories. Space limitations forbid
review of the theory and its assumptions, but some references to relevant liter-
ature are included in the ‘Further reading’ section of Chapter 7.
8 Chapter 4 of Simpson (1997) presents a very effective exercise that can be used
to stimulate detailed discussion and reflection on the precise choices an author
has made in composing narrative text – choices in the ordering of reported
events and states, and in the wording of those reports. Discussants are given all
the sentences of a short narrative text, separated and jumbled (typically, typed
on strips of paper); they are then asked, working individually or in groups, to set
out these sentences in the order in which they originally appeared. Thereafter,
solutions are compared and evaluated. Simpson’s article reports what happens
when this exercise is performed on a very short story – just 11 sentences – by
Ernest Hemingway, which he based on a newspaper report of the execution of
six Greek cabinet ministers in Athens in 1922. When compositors’ solutions are
subsequently compared with the Hemingway original, it is found that there are
particular descriptive sequences of sentence in the Hemingway version that all
groups have avoided. The Hemingway sequences ‘dwell’ on certain details (here,
of rain and sodden-ness) which informants seem to avoid as repetitive or obses-
sive (he does something similar elsewhere – e.g. at the opening of his story ‘Cat
in the Rain’). In these ways the exercise forces us to examine not merely event-
ordering in a narrative, but also the nature and ordering of accompanying
descriptions – i.e. the entire texture of the text –in terms both of coherence or
logic, and of effect. I have adopted and applied this exercise to ‘scrambled’ short
stories by writers such as Raymond Carver (‘Little Things’), Donald Barthelme
(‘The Baby’ and ‘The New Owner’), and Charles Reznikoff (‘The shoemaker sat
in the cellar’s dusk beside his bench’). The solutions and discussions are always
illuminating; the inventive unorthodoxy of writers’ text-sequencing decisions is
made newly apparent and, after reconsideration, newly understandable.
3 The articulation of narrative
text I
Time, focalization, narration
3.1 Narrative text: a single level of analysis

In this and the following chapter we turn from characterization of the ‘ele-
mentary particles’ of narratives, grouped under the label ‘story’, to the
various expansions and individualizations of those elements that go on in
the production of text. We can think of the move from the abstract level of
story components to the concrete level of textual realizations as both a
process and an articulation. The main processes or articulations involved
are listed by Bal as the following (I have replaced her terms ‘fabula’ and
‘story’ by ‘story’ and ‘text’ respectively, to avoid confusion):
1 The events are arranged in a sequence which can differ from the
chronological sequence.
2 The amount of time which is allotted in the text to the various ele-
ments of the story is determined with respect to the amount of time
which these elements take up in the story.
3 The actors are provided with distinct traits. In this manner, they are
individualized and transformed into characters.
4 The locations [settingsl where events occur are also given distinct
characteristics and are thus transformed into specific places.
5 In addition to the necessary relationships among actors, events, loca-
tions, and time, all of which were already describable in the layer of
the story, other relationships (symbolic, allusive, etc.) may exist
among the various elements.
6 A choice is made from among the various ‘points of view’ from which
the elements can be presented.
(Bal, 1985: 7)
This chapter concentrates on items 1, 2 and 6 on this list (i.e. time and
focalization), before concluding with a discussion of narration – the
complex different ways of narrating or being a narrator. The next chapter
will focus on items 3 and 4 (character and setting) together with suspense
and surprise, and the chapter after that (Chapter 5) on the subtle tech-
niques for disclosing characters’ words or thoughts, including free indirect
discourse. Item 5 on Bal’s list will not be addressed adequately in this
42 The articulation of narrative text I
book, due partly to space-limitations and partly to its less developed analysis
by narratologists and discourse linguists.
3.2 Text and time

Time itself, in the sense of the systematic measurement of what separates
particular past states from our present one, is itself a structuring and struc-
turalist notion. Structuring, because it asserts and articulates relations
between particular states or changes of state, and structuralist, in so far as
it relies on our recognition of particular similarities and particular differ-
ences between specified states. What we call ‘a year’ must have been first
perceived as a single full cycle of the warmer and colder seasons and of the
natural world, matching a full cycle of the sun’s movements towards and
away from one’s earthly location. Similarly, the perceived recurring suc-
cession of day and night gives rise to the introduction of temporal meas-
urement in just those terms (‘last night’, ‘in two days’ time’). So:
Time is perceived repetition within perceived irreversible change.
In addition we must remember that there is something unreal and heavily

convention-laden about so-called story time and text time. In neither case
are we referring to actual temporal progression, but the linear verbal
representation of temporality. A kind of artifice is at work, in which we
look for a match between the ‘real-world’ intervals and sub-intervals of
time that the narrative implies, and our sense of time passing during our
experience of reading that narrative. Whether we regard a particular
match as appropriate or not is a judgment largely guided by our prior
reading experience, our familiarity with the particular genre that the text
belongs to, and the seriousness of the events depicted.
Because text time is inescapably linear, there is an obvious and imme-
diate disruption of any neat correlation of real time to text time as soon as
the narrative involves more than one storyline: that is to say, as soon as
there is more than one set of developing circumstances affecting different
sets of characters. But even when we compare text time to what Rimmon-
Kenan calls ‘an ideal natural chronology’ (1983: 45), we rarely find that the
standard of steady correspondence is maintained. Even, say, when a story
is purely direct speech (monologue or dialogue), it is likely that we spend
considerably longer reading that written ‘transcript’ than would be spent
in any actual speaking.
The most influential modern theorist of text time has been Gerard
Genette, who isolates three major aspects of temporal manipulation or
articulation in the movement from story to text:
1 Order: this refers to the relations between the assumed sequence of

events in the story and their actual order of presentation in the text.
The articulation of narrative text I 43
2 Duration: for Genette this chiefly concerns the relations between the
extent of time that events are supposed to have actually taken up, and
the amount of text devoted to presenting those same events.
3 Frequency: how often something happens in story compared with how
often it is narrated in text.
As may be apparent from the foregoing discussion, the most potentially

problematic of these aspects is that of duration, and I shall devote rather
greater space to this aspect.
3.2.1 Order
Any departures in the order of presentation in the text from the order in
which events evidently occurred in the story are termed by Genette
anachronies. An anachrony is any chunk of text that is told at a point
which is earlier or later than its natural or logical position in the event
sequence (i.e. what we postulate by reconstruction to be the story
sequence). Strictly speaking, we can find anachrony even within a single
sentence. In The king died of grief because the queen had died, the subordi-
nate reason clause is an anachrony, presented after the report of the king’s
death even though it contains the report of an event which logically and
naturally – in the story – preceded his death.
But anachronies in extended narratives are more complex than that
example. They naturally divide into flashbacks and flashforwards, or what
Genette calls analepses and prolepses. An analepsis is an achronological
movement back in time, so that a chronologically earlier incident is related
later in the text. A prolepsis is an achronological movement forward in
time, so that a future event is related textually ‘before its time’, before the
presentation of chronologically intermediate events (which end up being
narrated later in the text). Any delayed disclosure is thus analeptic (the
reader expected, on the basis of chronological sequence, to be told of this
event or episode earlier), while any premature disclosure is proleptic (the
reader did not expect to be told this until later, if strict chronology had
been maintained). Incidentally, Bal’s useful terms for the two types of
anachrony are retroversions and anticipations.
An analepsis may be either homodiegetic or heterodiegetic, depending
on whether it carries information about the same character, event or story-
line as has been presented in the immediately-preceding text, or about
some different character or event. Examples of homodiegetic analepsis are
easy to find. We could cite the second paragraph of ‘Eveline’, with its
movement back in time to the games of Eveline’s childhood days; other
more complex analepses in ‘Eveline’ will be discussed later.
Or consider the very effective brief analepsis in ‘The Dead’, where
Gabriel falls into a revery while trying not to listen to Mary Jane’s
complex, unmusical piano-playing. A photograph of his mother brings
back to him how she always acted to ensure his advancement. But the text
continues:
A shadow passed over his face as he remembered her sullen opposi-

tion to his marriage. Some slighting phrases she had used still rankled
in his memory; she had once spoken of Gretta as country cute and that
was not true of Gretta at all. It was Gretta who had nursed her during
all her last long illness in their house at Monkstown.
As well she might (the flint-hearted reader using hindsight might add),
racked with guilt at having failed to nurse Michael Furey. Gabriel’s snob-
bish mother is surely wrong about Gretta, but the analeptic insertion of
her assessment, so much at odds with Gabriel’s, steers the reader towards
seeing (or being unsurprised later to see) Gretta as a rounded character,
not a flat or historyless one.
If we turn now to Nabokov’s Pnin (1957), we find many clear cases of
homodiegetic analepsis dotted throughout its opening pages. The follow-
ing is the novel’s second paragraph, describing the middle-aged Pnin’s
appearance as he travels by train to an engagement as a guest lecturer:
His sloppy socks were of scarlet wool with lilac lozenges; his
conservative black Oxfords had cost him about as much as all the rest
of his clothing (flamboyant goon tie included). Prior to the 1940s,
during the staid European era of his life, he had always worn long
underwear, its terminals tucked into the tops of neat silk socks, which
were clocked, soberly coloured, and held up on his cotton-clad calves
by garters. In those days, to reveal a glimpse of that white underwear
by pulling up a trouser leg too high would have seemed to Pnin as
indecent as showing himself to ladies minus collar and tie; for even
when decayed Mme Roux, the concierge of the squalid apartment
house in the Sixteenth Arrondissement of Paris where Pnin, after
escaping from Leninized Russia and completing his college education
in Prague, had spent fifteen years – happened to come up for the rent
while he was without his faux col, prim Pnin would cover his front stud
with a chaste hand. All this underwent a change in the heady atmo-
sphere of the New World. Nowadays, at fifty-two, he was crazy about
sunbathing, wore sport shirts and slacks, and when crossing his legs
would carefully, deliberately, brazenly display a tremendous stretch of
bare shin. Thus he might have appeared to a fellow passenger; but
except for a soldier asleep at one end and two women absorbed in a
baby at the other, Pnin had the carriage to himself.
Here is analepsis within analepsis. From the story’s present time, the text
jumps back to a description of Pnin’s ‘staid European era’ and his relations
with Mme Roux. But then during that description, there is a further jump
back to youthful Pnin’s escape from Russia. Thus we see a simple demon-
stration of how complex transformations of temporal order, in the articu-
lation of story as narrative text, make that text more entertaining, engross-
ing, and character-expressive. Nowhere in Pnin do we get a plodding,
blow-by-chronological-blow account of Pnin’s youth in revolutionary
Russia, his escape to Prague, his fifteen years there, and so on. That
important background only emerges ‘naturally’, as it were, when, as part
of Pnin’s ongoing narrative present, his past briefly ceases to be distant
background and becomes currently experienced foreground, as in Pnin’s
grieving recollection of a girl he had loved in pre-revolutionary Russia
who was subsequently murdered at Buchenwald (see p. 110ff.).
In fact the recollection is remarkably intense, more like a living-
through. Middle-aged Pnin is taking a vacation, along with assorted other
Russian emigrés and their children, in an upstate country retreat. But
when the name of Mira Belochkin crops up in conversation, Pnin with-
draws into an intense memory of his youth: of country house summers at a
Baltic resort, of his father and Mira’s father engrossed in their chess game
in a corner of the verandah.
Timofey Pnin was again the clumsy, shy, obstinate, eighteen-year-old

boy, waiting in the dark for Mira – and despite the fact that logical
thought put electric bulbs into the kerosene lamps and reshuffled the
people, turning them into ageing emigrés . . . my poor Pnin, with hallu-
cinatory sharpness, imagined Mira slipping out of there into the
garden and coming toward him among tall tobacco flowers whose dull
white mingled in the dark with that of her frock.
(111)
Later in the same revery it is disclosed that
In order to exist rationally, Pnin had taught himself, during the last ten
years, never to remember Mira Belochkin . . . One had to forget –
because one could not live with the thought that this graceful, fragile,
tender young woman with those eyes, that smile, those gardens and
snows in the background, had been brought in a cattle car to an exter-
mination camp and killed. . . . And since the exact form of her death
had not been recorded, Mira kept dying a great number of deaths in
one’s mind, and undergoing a great number of resurrections, only to
die again and again, led away by a trained nurse, inoculated with filth,
tetanus bacilli, broken glass, gassed in a sham shower-bath with
prussic acid, burned alive in a pit on a gasoline-soaked pile of beech-
wood.
(112–13)
As such passages show, grief and trauma are powerful triggers (perhaps
the most powerful) of character-based analepsis, while the willed ‘not-
remembering’ that is Pnin’s means of coping with them amounts to a
powerful suppression of analeptic tendencies in a character, reflected in
our clichéd injunctions to the grief-stricken to ‘live in the present’ and
‘look to the future’.
An example of heterodiegetic analepsis, offered by Genette, is the
focus on Swann as protagonist of certain events in section 2 of Proust’s À
la recherche du temps perdu: events that clearly occurred long before the
quite different focus of section 1, Marcel’s boyhood. To return to Pnin,
consider the extraordinary final chapter of that novel, in which the narra-
tor himself, whom we gradually realize is the ‘old friend’ and academic
rival whose appointment at Waindell College causes Pnin’s departure,
steps forward to give his own account of events. That account begins with
heterodiegetic analepsis: a reversion to the narrator’s own youth in Russia
in 1911.
All these examples are also external analepses, moving back to a time
prior to the opening of the text. Internal analepses are a textual moving-
back in the story, but not such a radical moving-back as to involve crossing
the text’s opening, and notably include any repetitions of incidents previ-
ously narrated in their proper chronological place. Of course an analepsis
could straddle or overlap the previously-established start of the narrative,
in which case Genette labels the analepsis ‘mixed’. As for the function of
analepses, a first observation can be that they seem to be designed to ‘fill
gaps’ in stories, though these gaps may themselves be the contrivance of
the writer, and may not be perceived as gaps until after the analepsis has
appeared.
But what of those analepses which cover events previously reported?
On such repetitive analepsis, Bal makes some valuable observations:
The repetition of a previously described event usually serves to

change, or add to, the emphasis on the meaning of that event. The
same event is presented as more, or less, pleasant, innocent, or import-
ant than we had previously believed it to be. It is thus both identical
and different: the facts are the same, but their meaning has changed.
In Proust, such internal retroversions form a part of the famous and
specifically Proustian interruption of the linearity in searching for, and
recovering of, the elusive past. But in much simpler literature too, fre-
quent use is made of possibilities such as these. Detective novels and
all kinds of texts which are constructed around mysteries, masquer-
ades, and puzzles adopt this technique as an important structural
device.
(Bal, 1985: 61)
Prolepses – which are much rarer than analepses – undercut or remove

suspense, since they reveal future circumstances to the reader long
before any chronological imperative dictates that they be told. They
foster a different kind of engaged puzzlement: the reader is made aware
of their own bafflement as to how characters and events might get from
the current situation to the future one prematurely revealed, and is all
the more intrigued to learn of the intervening happenings. Prolepsis is
common in first-person narratives, where it seems natural for the narra-
tor to jump forward occasionally to subsequent events which are closer
to that narrator’s own present. Like analepses, prolepses can be homo-
or heterodiegetic, depending on whether they entail a switch of focus to
a different character or storyline; and they are internal, external or
mixed, depending on their chronological relation to the endpoint of the
basic narrative.
One qualification may be made concerning certain examples of prolep-
sis, such as this one from Muriel Spark’s The Prime of Miss Jean Brodie:
‘Speech is silver but silence is golden. Mary, are you listening? What
was I saying?’
Mary Macgregor, lumpy, with merely two eyes, a nose and a mouth
like a snowman, who was later famous for being stupid and always to
blame and who, at the age of twenty-three, lost her life in a hotel fire,
ventured, ‘Golden’.
Examples such as this one are better thought of as proleptic ‘traces’

inscribed in a narrative, rather than the full-fledged shift of temporal ori-
entation where an extended stretch of text – a chapter, or a section –
reports distantly future events. Similarly, my first example of homodiegetic
analepsis in Pnin might be regarded as fleeting tracery rather than struc-
tural temporal reordering (in Barthesian terms, more an index than a func-
tion). Thus there seems to be a continuum of different kinds of anachrony,
performing different functions, with ‘fleeting tracery’ and ‘wholesale tem-
poral dislocation’ at opposite endpoints. The former is a contribution to
dense narrative texture, but is often more a local insight into a character,
or the narrator, than a manipulation of the event line; the latter is more
substantially a contribution to narrative structure, sometimes requiring the
reader to revise their assumptions as to just what the story is that is being
told. In characterizing anachronies Genette also suggests we can distin-
guish their chronological distance from the present moment in the story
(which I will call their ‘reach’, and Bal calls ‘distance’), and their chrono-
logical duration (which I will call their ‘extent’, and Bal calls ‘span’).
Before leaving the topic of ‘order’ it is worth reiterating that in discrim-
inating the types of anachronistic deviations that a narrative text makes
from the underlying story, all our observations are relational. That is, we
proceed not with the goal of simply unravelling the sequential jumble of a
Faulkner or Joyce novel, restoring the wholesome chronology that might
satisfy a historian or detective. Certainly unravelling is involved and is
important, but so too is a sense of the relations between the various
chunks of the text, where chunks told in one order (B–C–D–A–E) in fact
denote events that must be assumed to have occurred in a different order
(A–B–C–D–E), perhaps with considerable gaps or ellipses between A and
B and C, temporal contiguity of D and E, and so on: a potentially quite
elaborate network of temporal relations. (The picture is yet more complex
if we find that any two or more textual chunks report events that must
have occurred at the same point in time in the reconstructible story.)
Anachronies may also differ in scale. Thus an external analepsis that
reaches back twenty years with a story extent of one month, embedded in
a story whose extent otherwise is just twenty-four hours, should be distin-
guished from an analepsis identical in extent and reach but embedded in a
story of year-long extent.
3.2.2 Duration
What is text duration? Can it be reduced simply to the reading time of a
narrative? But readers read at different speeds, decide to break off from
reading at different places (or not at all), so that every reader will have a
different reading time for a narrative. On the other hand, granted absolute
differences of reading pace, we might still want to argue that there are rel-
ative similarities of reading time, for fluent native speakers, for particular
types of text. The application of such posited norms of reading duration,
against which one would then compare the likely temporal duration of the
events that the text relates, apply chiefly to scenic passages reporting
monologues, dialogues, sequences of physical actions which are punctual
or of short duration, and short journeys. So much genre fiction these days
seems to be aimed at the travelling reader – the person travelling alone by
train or plane, on a journey lasting several hours – that one imagines some
powerful effects might be achieved by contriving that both the reading
time and the text duration of such a book approximated, say, five hours.
Some such pan-textual norms of reading time are sometimes invoked in
stylistic commentaries on effects of (variation of) pace, but this is not the
approach adopted by Genette. Genette opts for an intra-textual strategy,
where textual pace at any particular place in a narrative is assessed relative
to pace elsewhere in that same narrative, and that pace is then expressed
as a ratio between the indicated duration of the story (in minutes, hours,
days) and the extent of text (in pages) devoted to its telling. This leads to
identifying a norm of pace for a particular narrative, against which acceler-
ations and decelerations can be perceived. The norm is thus text-bound,
and a constancy of pace would emerge if the ratio between story-duration
and extent of textual presentation were invariable – e.g. a page for every
month of a character’s life.
Genette’s ratio is very often far more mechanical than actual texts are.
For example, suppose there were a chapter for every year of a character’s
life, but those chapters were of rather different lengths: is this constancy of
pace or variation of pace? And how should we reformulate this ratio for
application to oral narratives? Presumably we would be back to hypothe-
sizing a norm for the ‘duration of delivery’ – although we could alterna-
tively take an intra-performance perspective, and make judgments about
pace relative to just a particular performance of a narrative. Rather sim-
plistic, too, is the assumption that event or story time is easy to deduce or
infer from the narrative, as if the heading of each page of the text carried a
digital read-out of the time elapsed. For example, just how long, in story
time, is Eveline’s revery? It could be anything from a few minutes to
several hours: it takes place between early and late evening (when the
mail-boat goes). Since we can’t be sure about the pace of the revery pre-
sentation, neither can we be sure as to whether the later scene at the quay-
side is a presentational acceleration or deceleration.
Maximum speed is said to constitute ellipsis, where no text space is spent
on a piece of story duration; the opposite situation is descriptive pause: text
without story duration (for example, the descriptive openings of A Passage
to India, or Ellison’s Invisible Man, or Hardy’s The Return of the Native).
An example of ellipsis cited by Chatman is that between the close of
Chapter 5 of Hemingway’s The Sun Also Rises, where Jake has finished
lunch and sets off for his office, and the opening of Chapter 6, when he is
waiting for Brett at five o’clock. Such ellipses, Chatman argues, are wide-
spread in modernist fiction, where a series of detailed scenic presentations
are linked by abrupt spatio-temporal jumps. However, it seems worth dis-
tinguishing this sort of ellipsis, which is simply an exploitation of the tem-
poral discontinuity we not only tolerate but probably prefer in our
narratives (so we do not have to read every dull thing a character does),
from the accelerations of presentation to the sharply abbreviated summary,
which are perceived as involving change of pace. For we surely do not feel
any change of pace, any acceleration, at the junction of Chapters 5 and 6 of
The Sun Also Rises. I would suggest, then, that ellipsis, in the form of a spa-
tiotemporal gap or aporia, is a narrational strategy of varying importance
(depending on just what gets left out) but is not really a type of narrative
pace, if we conceive of the latter as dependent on our judgments about the
rapidity of the telling of story events. In other words I am arguing for a view
of pace as the rapidity of the telling of what does get told.
3.2.3 Summary and scene

But more common and interesting than these are the relatively accelerated
and the relatively decelerated presentation, known since Lubbock’s inter-
pretive account of Jamesian poetics as summary and scene respectively
(Lubbock, 1973 [1921]). In summary, the pace is accelerated through a
textual compression of a given story period into a relatively short state-
ment of its main featutes. Rimmon-Kenan (1983: 53) cites an amusing
example of Nabokovian summary, a foregrounded and conscious play with
our conventional expectations both of novel duration and of what Barthes
calls catalysers (satellites), indices and informants:
Once upon a time there lived in Berlin, Germany, a man called

Albinus. He was rich, respectable, happy; one day he abandoned his
wife for the sake of a youthful mistress; he loved, was not loved; and
his life ended in disaster.
This is the whole story and we might have left it at that had there
not been profit and pleasure in the telling, and although there is plenty
of space on a gravestone to contain, bound in moss, the abridged
version of a man’s life, detail is always welcome.
(Nabokov, Laughter in the Dark)
In scene, story and text duration are conventionally considered identical

(e.g. passages of seemingly verbatim dialogue). But there is a pace of
telling that is more sharply counterposed to summary, in which things
move more slowly than in scene: the situation in which text is more
extended than story, which Chatman terms ‘stretch’. Again, in discussing
scenic pace, naturalistic assumptions and a covert comparison with the
pace of real interaction creep back in. Hitherto I have been stressing that
the pace of any portion of a text (the rapidity of presentation of time and
events passing) is relative to the pace in other portions of that text, or
relative to the pace of the text as a whole, or relative to the pace of other
texts. But when we start thinking about scene, and especially the
representation of direct speech exchanges between characters, some-
thing revealing happens. We tend not to assess the pace of a scene in
relation to co-textual standards, but rather in relation to real-world
standards.
This is a telling reminder of how our usual unquestioning view of fic-
tional dialogue is that it is the most complete and mimetic representation
of real dialogue. We tend to assume that reading direct speech dialogue
‘amounts to the same thing’ as being a witness to actual spoken inter-
action, so that to talk of the ‘pace’ of such written-up scenes – as if they
could go faster or slower – barely makes sense to us: obviously, scenes go
at just the pace of the actual interaction. But what has come to seem
obvious and commonsensical is not a necessary and unavoidable feature of
fictional dialogue, but a convention and an effect. The fictional direct
speech representation of any dialogue can go faster or slower, and we
neglect the artifice in fictional dialogue (e.g. all the ways in which it is not a
transcription), to the detriment of our appreciation of narrative poetics.
We will return to the natural and the conventional in speech representa-
tion when examining free indirect discourse in Chapter 5.
When the reader encounters relative acceleration or deceleration of
presentation, they will often interpret the shift in pace as an authorial or
narratorial indication of the marginality (on the one hand) or the central-
ity and importance (on the other) of what is being presented. More
important events and conversations are usually given in scenic detail, less
important or background ones in summary précis. But again, these are
norms from which writers often depart for good reasons. A writer may
play with our conventional expectations, a narrator-character may attempt
to suppress or retreat from certain important but distasteful events, and so
on. Shock and irony can be created by disclosing a central event briefly,
following detailed presentation of trivial events.
Chatman (1978: 76–8) also notes the contrast between narratorial
summary common in the traditional novel, and what goes on in many
modern novels, where characters provide summaries in their own recollec-
tions of past events (as in Woolf’s Mrs Dalloway). Hence we are really wit-
nessing scenes here, if the duration of character-derived remembered
events, and the extent of text devoted to their presentation, are commen-
surate with ratios for reflection-cum-presentation elsewhere in the narra-
tive. The puzzle hinges on what counts as a story event, the presentational
duration of which is to be compared with comparable events elsewhere in
the novel. Is the story event ‘Clarissa reminisces one day, from morning
shopping to evening party’; or is there here a series of story events, the
various incidents from her past that Clarissa reviews in memory? Perhaps
the best answer to give is ‘both’; we see the framing narrative, one day in
the life of Clarissa Dalloway (and some other Londoners), a narrative
within which the reminiscence is just a single event among many. We also
see, as the reminiscence unfolds, that an embedded story is being dis-
closed, spanning a much greater period of time than the framing one (‘a
life in one day of Clarissa Dalloway’).
In saying this we can maintain a dual characterization, just as, in sentence
grammar, we can say that a particular embedded nominal clause (she loved
life) counts simply as an object relative to the clause in which it is embedded
(Clarissa realized she loved life), but has its own internal structural logic
when viewed independently of the embedding clause. Thus we can now say
that as a unit within the framing narrative the reminiscence maintains
normal order, text-normal duration (allegedly), and singulative frequency.
But as a separate narrative embedded within the framing narrative, we
encounter the text of a story (of thirty years of Clarissa’s earlier life) which
involves various anachronies (many of the events are analeptically recol-
lected, and their normal order of occurrence is departed from), variations of
duration (some material is sharply summarized), and some instances of
repetitive frequency (see 3.2.4).
But even this dual characterization, arguably, leaves unremarked
certain effects that merit comment. A dual perspective seems all that is
strictly required, distinguishing the embedded story of incidents-remem-
bered-by-Clarissa-Dalloway from the superordinate story of incidents-on-
the-day-of-Mrs Dalloway’s-party. Thus, two stories at different levels: two
distinct configurations of order, duration and frequency. But while this
account of order can say things about that dimension in the two separate
stories, it fails to identify our experience of retroversion as we find
Clarissa’s thoughts jumping back to moments in her youth. As readers
processing the linear unfolding of the text, it seems we do not keep levels
and stories quite as separate as the analytical approach outlined above
does. In sophisticated ways that resist analytical unravelling, we unite the
story of things-happening with that of things-remembered-as-having-
happened, and understand Mrs Dalloway’s recollected experience to be
both timely and achronistic, utterly past in actuality, wholly present to
consciousness. Nor are these the only stories that we unite or hold in view
as we read this novel. There are the ongoing and remembered stories of
Lucrezia Warren Smith, of her husband, of Peter Walsh, and others. All
these stories we hold in view, unable to integrate them into a single
‘macro’ story, but also unable to separate them out as distinct: the textual
form denies us that solution. In the reading, our requirements of chronol-
ogy and causal relation in narrative are subordinated to attention to the
logic (in Clarissa Dalloway and all the rest) of human reaction, whether
physical or mental.
My conclusion, then, is that in a many-storied novel like Mrs Dalloway
our experience of the temporal order resists analysis – at least analysis in
terms of chronology. We seem to recognize and overlook anachronies at
the same time; we neither merge (in the sense of assimilate) nor unravel
these separable story sequences, but hold them in a complex enigmatic sus-
pension. These claims may be worth recalling when we come to examine
free indirect discourse in Chapter 5, for it seems that a very similar
unmergeable duality, resistant to analytical dissection, operates there too.
Hence the above paragraphs can be viewed as related both to the current
topic and, proleptically, to that of speech and thought reporting – the
embedding of voice within voice.
I have described Mrs Dalloway’s reminiscence as allegedly of normal
duration: but the strict test is to find other cases of reminiscing in the
novel, of apparently similar ‘real time’, and compare how much text is
used up in the different cases – a fairly rough and ready reckoning. And
perhaps an inappropriate one. There still seem to be several problems with
Genette’s notion of duration – it does not always seem to be addressing
what readers are responding to when they say one passage is pacey,
another passage is slow. And it seems that there is an underpinning of
text-internal vraisemblance (the means adopted in a text to convince the
reader that the text is a true and faithful representation of reality) or
verisimilitude in Genette’s approach. In other words, there is an unspoken
reliance on what we recognize as realistic and reasonable lengths of text to
spend on particular incidents, given some sense of the overall pace of a
narrative. In linguistic terminology, we could complain that his simple cal-
culation of duration leans too far towards assuming a natural fit, at least
within the narrative, and too far away from acknowledging that the norms
of text-extent in the presentation of various kinds of salient incidents may
be quite arbitrary and conventionalized.
In analyzing pace, supratextual comparisons are one interesting area
that calls for more study. For instance, we could usefully assess the pace of
Mrs Dalloway’s reminiscences relative to similar reveries in other Woolf
novels, in other modernist novels of the period, and in other British novels
of that time (James, Hardy, Galsworthy and Bennett). For it does seem we
become accustomed, as experienced readers, to certain kinds of duration
linked with certain kinds of culturally-salient narrative incident. We get
used to a certain kind of length of presentation of the details of a hero’s
early childhood in novels by Dickens and Thackeray, a certain degree of
extendedness in their presentations of love scenes and death scenes and
murders, so that our perceptions of duration must be not only intratextual
but partly intertextual (sensitive to author, period, and genre) as well. And
intratextually, as in the case of narrative discontinuity in The Sun Also
Rises cited earlier, we accept all sorts of ellipses related to cultural taboo
or tellability: no one has sex or goes to the bathroom in Austen or Dickens
– or more precisely, such actions are not narrated, being taboo or felt to be
uninteresting. This is no longer true when we get to Joyce and Lawrence.
But, as I have argued earlier, by duration we should mean something more
than the selectivity and discontinuity that can always be ascribed to a nar-
rative (we can always think of things that must have happened but haven’t
been reported).
3.2.4 Frequency
By the term ‘frequency’ we denote the business of repeated textual telling
of a single story incident. If the norm is ‘singulative’ frequency (telling n
times what happened n times), ‘repetitive’ frequency is exemplified in
Faulkner’s Absalom, Absalom!, in which the murder of Charles Bon by
Henry Sutpen is told thirty-nine times, by various tellers. Sometimes sin-
gulative presentation would strike us as strangely redundant and verbose:
where there are multiple occurrences of an incident of a single type. An
option in such cases is ‘iterative’ frequency – telling once what happened n
times. This is what happens in Faulkner’s ‘Barn Burning’, where one of the
wretched Snopes family’s hurried decampings from one farm to another is
narrated, but we are told that this is just the latest of a dozen such moves
that the young boy Sartoris Snopes (from whose viewpoint events are
told) remembers:
To-morrow they were there. In the early afternoon the wagon stopped
before a paintless two-room house identical almost with the dozen
others it had stopped before even in the boy’s ten years, and again, as
on the other dozen occasions, his mother and aunt got down and
began to unload the wagon, although his two sisters and his father and
brother had not moved.
(Faulkner, ‘Barn Burning’, 8–9)
The same story appears to embark upon an erroneous singulative telling

when the boy, Sartoris, finds that his father is in court a second time, before
a Justice of the Peace, and assumes that, like last time, it is to face and deny
(falsely) a charge of arson. The boy accordingly bursts out, ‘He ain’t done it!
He ain’t burnt . . .’ before his father silences him. Actually it would be truer
to say, ‘He ain’t done it yet’, for Snopes is indeed a compulsive barn-burner,
and before long he has torched the property of his present landlord, Major
de Spain. It is so as to break free from the awful singulative ‘again and
again’ of relocation, confrontation, barn-burning and flight, that Sartoris
alerts de Spain with the result that de Spain catches Snopes in the act, and
kills him.
3.3 Temporal refractions in text: Nabokov’s Pnin

So much for theorizing. We now need to return to some text, of manage-
able length, and see just how insightful the Genettian model can be. And
here it is important to remember that, although it seemed best to present
order, duration and frequency separately, in the practice of text-
articulation they are parameters that reinforce or interact with each other
in significant ways. Creativity or abnormality in one parameter may give
rise to exceptionality in the others: an event or episode told with repetitive
frequency will inevitably involve anachronisms in terms of order, and
more complex oddities in duration. But we should also keep in view the
possibility that other, non-Genettian methods – such as stylistic ones
focussing on a narrative’s patterns of tense and aspect in the verb, adver-
bials of temporal qualification – and an appeal to readers’ norms (rather
than the textual norm) may be more in tune with our judgments of textual
reorderings, pace and frequency. As attractive a text as any for these pur-
poses of demonstration is Nabokov’s Pnin. In what follows I will focus on
temporal processes in the first section of the first chapter of that novel.
Chapter 1, section 1 of Pnin recounts part of middle-aged Russian
emigré, US-naturalized, university professor Pnin’s mishap-ridden journey
to a women’s college where he is due to present a guest lecture. The first
paragraph is almost purely descriptive (I will return to this ‘almost’), with,
therefore, no scope for temporal classifications. It consists of two sen-
tences: the first locates and names Professor Timofey Pnin (and hence con-
cerns character and setting, not event), the second supplies a brief
evaluation of the prominent features of his appearance:
The elderly passenger sitting on the north-window side of that inex-

orably moving railway coach, next to an empty seat and facing two
empty ones, was none other than Professor Timofey Pnin. Ideally
bald, sun-tanned, and clean-shaven, he began rather impressively with
that great brown dome of his, tortoise-shell glasses (masking an infan-
tile absence of eyebrows), apish upper lip, thick neck, and strong-man
torso in a tightish tweed coat, but ended, somewhat disappointingly, in
a pair of spindly legs (now flannelled and crossed) and frail-looking,
almost feminine feet.
The second paragraph, which I have already quoted in full in 3.2.1, has an
early trivial anachrony:
His Oxfords had cost him as much as all the rest of his clothing . . . (my
emphasis)
But, as we have seen, more significant homodiegetic analepses follow. Are

these – the references to his European era, and its several stages (Russia,
Prague, Paris) – internal or external to the basic story? That, of course,
depends entirely on what we take to be the basic story; since the whole
Genettian exercise is an uncovering of relations between parts, it is quite
reasonable, in principle, for different analysts to adopt different – but con-
gruent – solutions.
Thus if we adopt the implied time point of paragraph 1 (October 1950)
as the opening of the story sequence as well as that of the text, then the
European references in paragraph 2 are clearly analepses (of varying
reach and extent), of sharply summarized duration and – at this stage at
least – of singulative frequency. Such a working hypothesis is, I think, sup-
ported by the content of the rest of the novel: I take the novel’s basic story
to be an account of Pnin’s last few months as a professor at Waindell
College, 1950–1 (an account interspersed with analepses providing retro-
spective glimpses of Pnin’s earlier life).
But one could, alternatively, see the basic story as about and including
Pnin at all the stages of his life. There are several reasons why such a solu-
tion, in this case at least, seems unattractive. Broadly, it casts the novel in
its entirety as a very oddly formed narrative: some fairly full presentations
of parts of Pnin’s childhood and adolescent experiences, often of brief
temporal extent and repeatedly interrupted by extremely lengthy ‘pro-
lepses’ concerning Pnin in 1950–1, with – and this is oddest of all – very
few references to developments during a huge chronological span from
c. 1920 to c. 1950. And more specifically, tense and temporal qualification
in the opening pages compel us to treat the Pnin of 1950, on the train, as
the Pnin of the narrative present –
Nowadays, at fifty-two, he was crazy about sunbathing, wore sports

shirts and slacks . . .
– while the Pnin of earlier days is the pre-story Pnin.

The third paragraph confirms our preferred hypothesis of the temporal
structure of the novel, with amusing retrospective explanations as to
Pnin’s current, unrecognized, error:
Now a secret must be imparted. Professor Pnin was on the wrong

train. He was unaware of it, and so was the conductor, already thread-
ing his way through the train to Pnin’s coach. As a matter of fact, Pnin
at the moment felt very well satisfied with himself. When inviting him
to deliver a Friday-evening lecture at Cremona – some two hundred
versts west of Waindell, Pnin’s academic perch since 1945 – the vice-
president of the Cremona Women’s Club, a Miss Judith Clyde, had
advised our friend that the most convenient train left Waindell at
1.52 p.m., reaching Cremona at 4.17.
But Pnin, we are told, consults his timetable and boards what he takes to
be a more convenient train:
Unfortunately for Pnin, his timetable was five years old and in part
obsolete.
Now, and only now, do we see the point, in the first sentence of the novel,
of the unexpected adverbial ‘inexorably’; and the occurrence of that word
is the single indicator of story temporal sequence in the first paragraph.
‘Inexorably’ implies something already begun, continuing, and impossible
to stop or cancel before its own predetermined and unwelcome conclu-
sion: a proleptic announcement of misfortune. En passant let us note the
word secret, used at the third paragraph’s opening (‘Now a secret must be
imparted’). Who is intentionally keeping the ‘secret’ of Pnin’s mistake
hidden from him, and are we readers, now in the know, complicit? Who, in
the circumstances, could possibly expose this ‘secret’? What kind of
teasing, darkly humorous, mildly sadistic narrator are we grappling with
here? Out of the accumulation of turns of phrase such as that one, we
readers build a picture of the kind of narrator we are being addressed by
in a novel or story, about which more is said towards the close of this
chapter.
There follow several pages of description of Waindell College, of Pnin’s
handful of students (each briefly characterized with the aid of deadly
analeptic commentary; e.g. ‘languid Eileen Lane, whom somebody had
told that by the time one had mastered the Russian alphabet one could
practically read ‘‘Anna Karamazov’’ in the original’), and of Pnin’s bizarre
teaching style. This latter is heavily dependent on digressions, personal
anecdotes, and the reading of comic passages from books.
But since to appreciate whatever fun those passages still retained one
had to have not only a sound knowledge of the vernacular but also a
good deal of literary insight, and since his poor little class had neither,
the performer would be alone in enjoying the associative subtleties of
his text. The heaving we have already noted in another connexion
would become here a veritable earthquake. Directing his memory,
with all the lights on and all the masks of the mind a-miming, toward
the days of his fervid and receptive youth . . . Pnin would get drunk on
his private wines as he produced sample after sample of what his lis-
teners politely surmised was Russian humour. Presently the fun would
become too much for him; pear-shaped tears would trickle down his
tanned cheeks. Not only his shocking teeth but also an astonishing
amount of pink upper-gum tissue would suddenly pop out, as if a jack-
in-the-box had been sprung, and his hand would fly to his mouth,
while his big shoulders shook and rolled. And although the speech he
smothered behind his dancing hand was now doubly unintelligible to
the class, his complete surrender to his own merriment would prove
irresistible. By the time he was helpless with it he would have his stu-
dents in stitches . . .
Here, emphatically, we see the significance of the two essential com-

ponents of narrative noted in Chapter 1: the teller and the tale. For while
Pnin himself is absorbed in the seemingly devastating humour of the texts,
his student audience’s amusement is entirely derived from the absorbing
spectacle of the teller. And framing these two is the reader’s amusement at
the artful telling of the students’ amusement at Pnin’s amusement: a
warming feeling of togetherness in humour, all the more poignant for our
awareness of the different kinds of estrangement that separate Pnin from
his students, and both of them from us. The practical irrelevance of what
the passage-content is that sets Pnin off is underlined by the fact that we
neither seek nor require any details about it whatsoever.
The other great attraction of this stylistic tour de force is its use of the
modal of iterated or habitual activity (would). The stages of physical col-
lapse in Pnin’s performance, culminating with the sudden emergence of his
false teeth, seem so detailed and specific as to make their recurrence all
the more extraordinary and laughable. And given the light-hearted nature
of what is told, such iterative-frequency narration – telling once what hap-
pened often – itself enhances the spirit of recognizability and empathy that
such incidents promote.
But if such a spirit were to render the reader too relaxed, the terse one-
sentence paragraph that follows ensures otherwise:
All of which does not alter the fact that Pnin was on the wrong train.
The point about Pnin’s habitual performance, described above, is that

while iterative in frequency, and of uncertain duration, it is, being habit-
ual, both analeptic and proleptic, a potential event at numerous points in
time. It cannot arise during this train journey of Pnin’s, to be sure, since
his students are not present with him: we may thus designate its textual
position, its being told here, as an achrony. (In a sense, then, ‘this is not
the time’, here on this train journey, for one of Pnin’s reading-comic-
passages laughing fits; but this is precisely where we readers are told of
them.) But being habitual events, these are not very firmly anchored to
any particular point or points in time: the only delimitation is the obvious
one that these hysterics can only arise during Pnin’s classes. (The habit is
rather less of a constant recurrence than, say, giving a little cough every
time one begins to speak.)
The reminder that in ‘our’ present-time Pnin is ‘still’ on the wrong train
is, in turn, followed by a long paragraph of characterization (again, a
descriptive pause in terms of duration, not tied to any point in terms of
order) concerning his fatal attraction to gadgetry. This, too, is followed by
a terse reminder of the situation in the ongoing basic story:
And he still did not know that he was on the wrong train.
The running joke continues, with a lengthy account of his eccentric

command of the English language followed by yet another resumption of
the story proper –
The conductor . . . had now only three coaches to deal with before
reaching the last one, where Pnin rode.
– before we are treated to the lovely ramifications of a ‘Pninian quandary’

as to where, on his person, he should store his lecture manuscript:
If he kept the Cremona manuscript . . . on his person, in the security of

his body warmth, the chances were, theoretically, that he would forget
to transfer it from the coat he was wearing to the one he would wear.
On the other hand, if he placed the lecture in the pocket of the suit in
the bag now, he would, he knew, be tortured by the possibility of his
luggage being stolen. On the third hand (these mental states sprout
additional fore-limbs all the time), he carried in the inside pocket of
his present coat a precious wallet . . .; and it was physically possible to
pull out the wallet, if needed, in such a way as fatally to dislodge the
folded lecture.
Pnin’s quandary is an acute form of the kind of anxiety we all feel over
unwelcome possible future consequences of our present actions, proleptic
imaginings all the more ironical since we know Pnin is already in a mess
(no need for him to imagine one). Incidentally, the detached conditional
‘if needed’, towards the close of the quoted extract, indicates how very
skilled is Nabokov’s use of the English language. On an unexceptional
reading it qualifies ‘the wallet’. But on a more bizarre – but appropriately
bizarre – reading it would qualify ‘pulling out the wallet so as fatally to dis-
lodge the lecture’. That is, if the wallet needed to be pulled out in such a
way as fatally to dislodge and lose the lecture, then Pnin could be relied on
to do the job!
The first section closes with further complex temporal reflections and
projections, as Pnin learns he is on the wrong train and is redirected by the
conductor. But I hope enough has been shown of just how widespread and
complex the manipulations of time-lines can be, even in quite short pas-
sages. To repeat, the point of such close analyses is not to unravel a text, to
return to some underlying singulative, steady-paced, linear chronology,
but rather to understand more fully how, in our narratives as in our lives,
we constantly demand and draw upon potential complexities of pace, iter-
ation, and reordering.
3.4 Focalization
Faulkner’s ‘That Evening Sun’ begins:
Monday is no different from any other weekday in Jefferson now.
Readers familiar with Faulkner will know that Jefferson is his fictionalized
version of Oxford, Mississippi, and that we are back in a fictionalized deep
South; but we do not yet know, even roughly, when this now is, or how
long ago the contrasted old-style Mondays were. Is the narrator contrast-
ing the 1950s with the 1930s, or the 1930s with the 1900s? We need to
know more about the viewpoint from which the story is being told. A
reader must – over the succeeding sentences of the story – come up with
answers to such questions, and in doing so will get an approximate ‘fix’ on
both the place and time from which the teller must have composed that
sentence, and the identity (insofar as this is disclosed) of the teller.
These are matters of orientation, and the linguistic term for all those
elements in a language that have a specifically orientational function is
deixis. The very presence in any discourse of features such as I and you, of
yesterday, today and tomorrow, of tense choices, and of contrastive
adverbs and adjectives such as here and there, this and that, now and then,
means that that discourse is consequently interpreted as grounded, or
anchored, as coming from a particular speaker at a particular place at a
particular time. Any text, then, that contains deictic information is thereby
understood as oriented from the spatiotemporal position that those deic-
tics imply. But the paradox is that the dedicatedly orientational words
(this, now, here, etc.) cannot be properly interpreted without information
about the speaker’s spatiotemporal location derived from elsewhere.
Compare It’s wonderfully hot here today (lots of deixis, but ambiguous
without co-text or context) and Friday 11 August 2000 was wonderfully hot
in Brittany (little or no deixis, largely unambiguous).
What applies to discourses in general applies particularly importantly to
narratives. In the process of telling a narrative, with its almost inevitable
and copious specifications of time and place, some perspective or another
has to be adopted as the vantage point from which the spatiotemporally
determinate events are related. Even
Once upon a time in a distant land there lived a beautiful princess
signals, through the emphasis on Once, distant and past tense lived, that
the perspective adopted in the telling assumes teller–listener proximity,
and a spatiotemporal remoteness, in the past, of the events to be narrated.
But, very importantly, the telling of a narrative does not have to maintain,
throughout, a single perspective or orientation. Simpler narratives will
tend to, and this orientation can be straightforwardly assumed to be the
narrator’s; but in the more complex literary and film narratives, the
viewpoint of the telling may move around, from narrator to one character
to another character.
‘Focalization’ is Gerard Genette’s term (the notion was further
developed by Bal) for this inescapable adoption of a (limited) perspective
in narrative, a viewpoint from which things are implicitly seen, felt, under-
stood, and assessed. By this is meant the angle from which things are seen
– where ‘seen’ is interpreted in a broad sense, not only (though often most
centrally) in terms of visual perception. As Rimmon-Kenan comments
(1983: 71), this term does not entirely shake off the optical–photographic
connotations that have made its Anglo-American critical equivalent, point
of view, problematic. I hesitate to offer another variant term to compete
with those we already have, but I do think orientation is a usefully wider,
less visual, term than ‘focalization’, and would help us to remember that
‘cognitive, emotive and ideological’ perspectives, in addition to the simply
spatiotemporal one, may be articulated by a narrative’s chosen focaliza-
tion. Accordingly, though I will mostly retain Genette’s term ‘focalization’
in what follows, the reader is welcome to substitute ‘orientation’ if this is
any help.
The great and continuing nuisance perpetuated by the term ‘point of
view’ is that it does nothing to discourage the conflation and confusion of
two distinct aspects of narrative practice:
1 The orientation we infer to be that from which what gets told is told.
2 The individual we judge to be the immediate source and authority for
whatever words are used in the telling.
These can be summarized in two distinct questions: ‘Who sees?’ and ‘Who
speaks?’
As noted, in many narratives, orientation and discourse-authorship are
sourced in a single individual. But speaking/thinking and seeing do not
have to come from the same agent. There are many cases where a narrator
‘undertake[s] to tell what another person sees or has seen’ (Rimmon-
Kenan, 1983: 72). In the early chapters of Great Expectations, for example,
the narrator is Pip the adult, with an adult’s extended vocabulary, a differ-
ent person from the focalizer, who is Pip the child. An inevitable corollary
of the notion of focalizer or subject-of-the-focalization is that there must
also be someone or something that is the object of the process, i.e. the
focalized. In the following section the different types of focalizer and
focalized identified by Bal are outlined.
3.4.1 Types of focalization

The basic contrast is between external and internal focalization. External
focalization occurs where the focalization is from an orientation outside
the story (the orientation is not associable with that of any character
within the text). In such cases the narrator/focalizer separation is neutral-
ized, so that the focalization is of no particular interest, independent of the
narration. Internal focalization occurs inside the setting of the events, and
almost always involves a character-focalizer, though some unpersonified
position or stance could be adopted. Thus in Faulkner’s ‘Barn Burning’,
the boy Sartoris is often the focalizer, but he is not the narrator (the
vocabulary in sentence 2 below is emphatically not Sarty’s). The story is
told largely from his orientation, but he is not directly responsible for the
words used, as this extract from the opening shows:
The store in which the Justice of the Peace’s court was sitting smelled
of cheese.(1) The boy, crouched on his nail keg at the back of the
crowded room, knew he smelled cheese, and more: from where he sat
he could see the ranked shelves close-packed with the solid, squat,
dynamic shapes of tin cans whose labels his stomach read, not from
the lettering which meant nothing to his mind but from the scarlet
devils and the silver curve of fish – this, the cheese which he knew he
smelled and the hermetic meat which his intestines believed he
smelled coming in intermittent gusts momentary and brief between
the other constant one, the smell and sense just a little of fear because
mostly of despair and grief, the old fierce pull of blood.(2) He could
not see the table where the Justice sat and before which his father and
his father’s enemy stood, but he could hear them . . .(3)
Here, sentence 1 is not exclusively the boy’s perspective, but that of

anyone inside the store: only someone who was present in the makeshift
courtroom at that time, or was later told what the room had been like by
someone else who had been present, can make this report. Much more
specifically, sentence 2 emphatically expresses the boy’s orientation and
sentence 3 mixes the boy’s focalization (‘he could hear’) with information
about the relative positions of the Justice and his father and the table,
which cannot be through the boy’s eyes since the text explicitly states he
could not see these things.
Like the two types of focalizers, there are also two types of focalized,
where the distinction is between viewing from outside or from within. In
the former, only the external, literally visible phenomena are reported; in
the latter, facts about the feelings, thoughts and reactions of a (or several)
character(s) are reported, so that a penetrating intrusive portrayal is
achieved. Molly Bloom in Ulysses is both internal focalizer and focalized
from within, while in Hemingway what is focalized is very commonly
viewed from without. Focalization may remain fixed, tied to a single focal-
izer, throughout a novel, as in What Maisie Knew. Of this novel Bal (1985:
104–5) writes: ‘The difference between the childish version of the events
[focalized by Maisie] and the interpretation that the adult reader gives
them determines the novel’s special effect.’ But the focalizing may vary
between two or more positions, as in Faulkner’s The Sound and the Fury.
The interest in focalization stems from the fact that it highlights the
‘bi-directionality’ of narrative: the fact that the focussing on a particular
object in a particular way reveals that object but also must reveal (or try
not to reveal) the perspective and ideology from which that subject is seen.
Furthermore, such revelatory focalization goes on regardless of whether
the object focalized really exists. In this respect it is useful to distinguish
between focalizeds that we as readers accept actually exist in the world of
the narrative from those we take to be dreams, fantasies or other figments
of one character-focalizer’s imagination. We might label this distinction
actual versus imagined focalizeds (a distinction slightly different from and
more difficult than Bal’s one between ‘perceptible’ and ‘non-perceptible’
focalizeds). Some narratives trade heavily on our uncertainty as to
whether what is focalized is actual and potentially ‘public experience’, or
imagined and perhaps an index of psychosis (a magnificent example is
James’s The Turn of the Screw).
3.4.2 Facets of focalization?

Bal (1985) seems to prefer not to attempt a detailed discrimination of
types of focalization, emphasizing rather the levels involved. But Rimmon-
Kenan (1983), evidently considerably influenced by Uspensky (1973), does
attempt a typology of what she calls the facets of focalization, the major
ones being perceptual, psychological and ideological, each of which
permits great variation in the power or breadth of the focalizing. For
example, with regard to the perceptual dimension of focalization, the
focalizer may enjoy (and relay to us from) a panoramic perspective which
allows holistic descriptions of large scenes, and even of several distinct but
simultaneous scenes; this obviously entails an external focalizer. On the
other hand, where the focalizer is a character within the narrative, the
limited view of that spatiotemporally limited observer is to be expected. A
similar broad contrast between the constrained and unlimited perspectives
(actually with many intermediate degrees of limitedness) applies also to
time focalization.
In addition to variation in spatiotemporal orientation, there is psycho-
logical variation, which Rimmon-Kenan separates into the cognitive (e.g.
the internal focalizer’s limited knowledge versus the external focalizer’s
theoretical omniscience) and the emotive (neutrality versus involvement
in presentation). And in involved emotive focalization, for example,
scenes are represented in a noticeably idiosyncratic way, such as seems
best attributed to the mood and personal evaluations of a character. The
focalized’s mind and emotions, too, are open to either external or pene-
trative/internal treatment (assuming the focalized is human or quasi-
human).
A final facet of focalization variation is the ideological. Often, it seems,
one ideology or world-view, of an external narrator–focalizer, is the domi-
nating norm, and any characters’ ideologies that deviate from this stan-
dard are at least implicitly (and sometimes explicitly) censured. On the
other hand, there may be a juxtaposition of different ideological orienta-
tions without any overt adjudication between them, so that the reader is
torn between different views of certain events in particular and (by exten-
sion) of the world in general. On this topic, Bakhtin (1981) is particularly
relevant. Insofar as both the psychological and the ideological facets of
focalization are a matter of how things are evaluated, there seems to be
plenty of room for overlap between these two. (Spatiotemporal perceptual
focalizing, by contrast, is not inherently evaluative, being merely a disclos-
ing by implication of the when and where of the witnessing.) In a typical
situation, it may well be that, as analysts, we talk of a deviant main charac-
ter as revealing weird psychology in their focalizations, this being high-
lighted and counteracted by the orthodox and ‘reasonable’ ideology of an
external narrator, focalizing elsewhere; the effect of this, for the reader,
is often that the character’s ideology appears to be ‘corrected’ by the
narrator’s.
3.5 Perceptual focalization as primary

While it may be accepted, in light of the foregoing examples, that narra-
tion always entails focalization, it is debateable whether we need to posit a
focalizer position distinct from the narratorial one in all texts, and whether
we should typically work on the assumption that we can identify a focal-
izer’s spatiotemporal, psychological and ideological orientations as distinct
from those of the narrator. Quite often theorists’ examples of psychologi-
cal or ideological focalization (e.g. those in Rimmon-Kenan) seem easily
accommodated within more orthodox characterizations of the particular
narrator as naïve or childlike or self-conscious or paranoid. Orientational
limitation attributed to a particular character’s perspective does seem to
make the best sense in relation to spatiotemporal matters; in the areas of
psychology and ideology it seems far less easy to resolve that a particular
emphasis is not that of the narrator – except, of course, where we are faced
with the speech or thought of a character, directly or indirectly rendered,
but that is another matter. In focalization the essential question remains:
‘Who sees?’ More fully, the question to ask in determining spatiotemporal
focalization is:
Who is the immediate seer here, and whose is the ‘zero-point’ for time
measurement here, to whom we attribute the spatiotemporal orienta-
tions we are given?
Similarly with psychological and ideological evidence, whether explicit or

only implicit:
To whom do we attribute these traces or revelations of psychological

or ideological orientation? Who is their immediate source?
3.6 Narrators and narration
In section 3.4 I introduced a distinction between
1 The orientation we infer to be that from which what gets told is told;
and
2 The individual or ‘position’ we judge to be the immediate source and
authority for whatever words are used in the telling.
The former we distinguish as focalization, the latter as narration, and it is

to this latter domain that I now turn.
The business of specifying just how many optional or obligatory roles
are involved in the process of narration can soon get remarkably complex.
The scheme most widely-adopted by narratologists is the following (see,
e.g. Chatman, 1978: 151):
real implied implied real

→ → (narrator) → (narratee) → →
author author reader reader
In Chatman’s discussion, the real author and reader are left out, on the
grounds that their ‘implied’ counterparts are the functioning substitutes in
the business of narrative transmission. Furthermore, Chatman claims that
the narrator–narratee pair are optional positions. I will propose an altern-
ative simplification which, I hope, will not ignore the potential complexities
involved; I will argue that, notwithstanding the potential relevance of the
‘implied author’, ‘narratee’ and ‘implied reader’ constructs, the three core
roles in literary narrative transmission are author, narrator and reader:
author → narrator → reader
It is no accident that of the six roles usually discussed in narratological

studies, these three are the ones that are robustly present on the surface of
narrative texts, and the most deeply recalled when we ordinarily think
about narratives. We (the reader) open Great Expectations and somebody
tells us how he stumbled over his own name, of Philip: whoever he is, this
Philip, the narrator, is distinct from the author, Charles Dickens. Or we (as
readers) read the opening paragraphs of ‘That Evening Sun’, and are told
‘about half the time we’d have to go down the lane to Nancy’s cabin and
tell her to come on and cook breakfast’ and, a page later, that ‘father’ tells
Quentin to go and see if Nancy has finished in the kitchen, whereupon ‘I
went to the kitchen’. So Faulkner, the author, has established Quentin as
narrator. And so on. By contrast, the implied author, the narratee, and the
implied reader are rather more notional: potentially analytically import-
ant, but in more oblique ways.
3.6.1 Implied author: a construct but not a primary role
The implied author is the mental picture of the author that a reader con-
structs on the basis of the text in its entirety. It is easiest to see that we do
(or at least can) perform such a (re)construction in the course of reading if
we imagine an anonymous text. Perhaps the text is a crime novel, with a
grizzled Metropolitan Police detective as protagonist, and the action is set
in the pulsing streets and edgy deals of contemporary Soho, involving
drugs, prostitution, protection-rackets and electronic fraud. However we
visualize and pigeonhole the implied author, we probably do not expect
that they are a retired nun who has never left Sri Lanka, or an overworked
junior doctor, or an eco-warrior, or a ten-year-old, or James Joyce. But
more important than the vague provenance a narrative might project as
belonging to the implied author (such as male, in his thirties, streetwise,
has knocked about a bit in London or somewhere very similar, knows
police-work ‘from the inside’, etc.) are the values that, rightly or wrongly,
the reader imputes to the implied author (perhaps, that he has a sordid
fascination with commercial sex, narcotics, transgression, violence, that he
seems to derive a certain pleasure from outsmarting the reader with his
inside knowledge of electronic financial dealings, that deep down he
mainly believes in private property, protecting your own, carrying a big
stick, and coming out ahead when you can . . .). If he didn’t tacitly sub-
scribe to those values – so the reader reasons, in explaining the picture of
the implied author they have drawn – why would he construct the kind of
narrator he has, and write about the kinds of things he does, in the ways
that he does?
An implied author can be retrospectively projected on any text, narra-
tive or otherwise. In the case of literary narratives, the account of an
implied author that a reader develops tends in practice to have much to do
with authorial intention and meaning. It amounts to an answer to the
reader’s own question: ‘What sort of person, with what sort of interests
and values, must the author be, to have produced this text with the preoc-
cupations and meanings that I take it to have?’. Thus one can imagine a
reader of ‘That Evening Sun’, perhaps one with no previous knowledge of
Faulkner or his other works, reading that story and being struck by the
tacit and not-so-tacit racism; the domesticity of the focus (little sense of a
larger world, beyond the semi-lawless town); the fractiousness and dys-
functionality of the Compson family; the cross-generational ‘gender
alliances’ that link Caddy with Nancy, and the narrator Quentin with his
father; the tangled dialogue on the page in which, repeatedly, A’s remark
to B is followed, before the latter’s reply is given, by C’s request to D;
Jason’s insufferable whininess; Nancy’s desperate, ruined circumstances;
and the way a story that includes cocaine-addiction, an attempted suicide,
and fear of murder unravels to a wretched close, with Nancy declaring
she’s ‘just a nigger’ and Caddy telling Jason he’s worse than one. In the
light of all that, and all that it means – the characterization, the way and
tone of narration, the topics and treatment – whatever picture we infer of
the mind of the author is our ‘implied author’.
The notion of the ‘implied author’ was first proposed by Wayne Booth:
As [the author] writes, he creates not simply an ideal, impersonal ‘man

in general’ but an implied version of ‘himself’ that is different from the
implied authors we meet in other men’s works . . . the picture the
reader gets of this presence is one of the author’s most important
effects. However impersonal he may try to be, his reader will
inevitably construct a picture of the official scribe who writes in this
manner.
(Booth, 1961: 70–1)
In subsequent discussions of the implied author, the emphasis has tended to

be on the word ‘implied’: in Booth the emphasis seems to me to be far more
on the word ‘author’. There the claim is that we project or reconstruct back,
from the text, some sort of version or picture of the author. Clearly that
version of the author is not the author herself; it may be a version that the
actual author vigorously rejects, or that is unduly flattering or derogatory of
her.
It is not clear, however, that the implied author is a distinct level of nar-
rative structuring. We can retain the term ‘implied author’ to refer to the
picture of Faulkner we conjure up from The Sound and the Fury (a very
different picture, surely, from that which we conjure up from The Reivers),
but we might as happily simply distinguish the Faulkner of The Sound and
the Fury and the Faulkner of The Reivers, or the Joyce of Dubliners and the
Joyce of Finnegan’s Wake. As far as these novels and narratives are con-
cerned, these are the only versions of the author we know, and there is no
‘real author’, unitary, unchanging, standing behind these narrative-derived
versions. If we’ve read Blotner’s biography of Faulkner, or Ellmann on
Joyce, we have read another narrative presenting (much more fully and
directly) another version of the author. Even if we also know an author
personally, we still perform the same process of forming a mental picture or
representation (itself a kind of narrative) of that author to ourselves, as an
integral part of the activity of knowing a person. In short, the pictures we
have of authors are always constructions, so that all authors are, if you like,
‘inferred authors’. But we can and should separate such pictures from the
actuality of authorial narrative production: those pictures may be import-
ant in narrative reception and critical theory, but they are irrelevant to
narrative production. The implied author is a real position in narrative pro-
cessing, a receptor’s construct, but it is not a core or necessary role in narra-
tive transmission. It is a projection back from the decoding side, not a real
projecting stage on the encoding side.
3.6.2 Narratees and implied readers
While the production side of narrative transmission can focus on author
and narrator, the reception side, similarly, should recognize the reader as
core participant, and narratee and implied reader as secondary. The narra-
tee is an individual, involved in or quite detached from the events of the
story, directly addressed by the narrator. Very occasionally a narrator
addresses her discourse to herself, but much more typically the specific
narrative addressee is a character–receiver within the story, if rather mar-
ginal to the action (the psychiatrist Dr Spielvogel who is told all the young
man’s hangups in Portnoy’s Complaint, the seafarers who are told
Marlow’s story in Heart of Darkness and so on). Sometimes, as in Sterne’s
Tristram Shandy, a narratee is playfully conjured up and has her ears
boxed for inattentiveness:
How could you, Madam, be so inattentive in reading the last chapter?

I told you in it, that my father was not a papist! You told me no such
thing, Sir. Madam, I beg leave to repeat it over again, That I told you
as plain, at least, as words, by direct inference, could tell you such a
thing . . .
(1967: 82, quoted in Rimmon-Kenan, 1983: 105)
We readers enjoy the joke in which a fictional entity, as if in our position,

gets ticked off in a way that we know we sometimes deserve. But we do
also see that this is an ‘as if’ relation, that this is a strategy or device (if the
narrator were really scolding us we would probably resent and resist that
impertinence). Any residual nervousness on the part of any Madam-
reader that she is being directly addressed is dispelled, of course, by the
direct speech attributed to this other, fictional Madam.
In all these cases and others besides, the narratee position is not a foun-
dational part of the framework of the telling, but an integral device in nar-
rational strategy. In all the cases cited above – and almost always, in fact –
the narratee is addressed by an intradiegetic narrator, i.e. not the narrator
of the narrative in our foundational sense (the source or agent for every-
thing that gets told). As Bal says of this ‘foundational’ narrator:
We . . . do not mean a story-teller, a visible, fictive ‘I’ who interferes in

his/her account as much as s/he likes, or even participates as a charac-
ter in the action. Such a ‘visible’ narrator is a specific version of the
narrator, one of the several different possibilities of manifestation. . . .
We shall rigorously stick to the definition of ‘that agent which utters
the linguistic signs that constitute the text’.
(Bal, 1985: 120)
Typically, a narratee is a visible fictional character whom we witness being

addressed by an even more visible second-order narrator, and behind their
fake dialogue is some storyteller (the first order or foundational narrator)
whom we take to be the agent of all their words (both the visible narra-
tor’s and the narratee’s), and any other material besides (e.g. the insertion
of chapter breaks in the telling, chapter numbers, headings, etc.). Narra-
tees, then, are real but secondary and infrequent textual entities.
By contrast the ‘implied reader’ is another reader-based construct, like
the implied author: it is a picture, based on the text in its totality, of the
kind of reader or archetypal reader that real readers assume that the text
has or had in mind as its audience. Inferencing and stereotyping is
involved in positing any text’s implied reader. We usually feel most secure
sketching the implied reader for narratives which are themselves rather
generic and which, though we may enjoy them, we do not hold in the
highest esteem. Thus it is easier to sketch the identity and values of the
assumed, targeted reader of novels by Barbara Bradford Taylor, Dick
Francis, Sue Grafton, and so on; it is harder to say who is, or is not, the
implied reader of Middlemarch or White Noise or Pnin. The notion of
implied reader, and its implication that certain readers are directly
addressed or interpellated while others are not, is always likely to provoke
contention and controversy. But the ascription of a particular kind of
implied reader to particular kinds of texts certainly happens, and to that
extent this secondary role or notion is valuable. Rightly or wrongly, in the
face of certain kinds of writing (which is, for example, noticeably sexist, or
class-conscious, or soft-pornographic, or religious, or technically-detailed
about some activity) we assume that the author had a particular kind of
reader in mind. But then again perhaps they did not, or perhaps – as
reader – I don’t care even if they did, and refuse to comply with that
oblique effort to include or exclude me. As Walter Ong (1975) pointed
out, the writer’s audience is always a fiction, a convenient provisional
target. Real readers, real audiences, can apprehend stories in quite unpre-
dicted ways, seeing a different point to them, and picturing quite dissimilar
authors of them.
3.7 Simpson’s typology of narratorial modes

Literary criticism operates with an array of well-established terms for dis-
tinguishing kinds of narrator and, by extension, kinds of narration. Many
of these terms constitute binary alternatives: third-person versus first-
person narration; within third-person narration, omniscient versus limited
narration; and, within omniscient narration, intrusive versus impersonal/
objective narration. Although invaluable, these terms are far from uncon-
troversial in application. But most importantly they immediately direct
attention to the narrator, whereas the schemes to be introduced below
direct our attention to the narration (and derive tentative characteriza-
tions of narrators therefrom). These latter schemes are thus more text-
linguistic and less psychological in orientation, and thus more suited to a
linguistic study of narrative.
As a preliminary step, we can follow Rimmon-Kenan (1983: 96–8),
based on Genette, in setting up different categories of narrator depending
on whether that narrator is extradiegetic or intradiegetic, and whether she
is a story participant or not (homodiegetic versus heterodiegetic). In addi-
tion to such a categorization, we can assess the visibility of a narrator by
looking for the following kinds of textual material, which are indicative –
in increasing order of intrusiveness – of narratorial presence:
1 descriptions of settings;
2 identification of characters;
3 temporal summaries;
4 definition of characters;
5 reports of what characters did not think or say;
6 commentary – interpretation, judgment, generalization.
The above six types of material reflect – in order – greater and greater nar-
ratorial knowledge and understanding of whatever story is articulated.
Narration embracing only types 1–3, for example, would not be unlike the
kind of official and minimally-interpretive account of an incident that
might be found in a police report or a description of an accident for insur-
ance purposes. In addition, within each of the six types of report there can
be greater or lesser degrees of specificity, insight, and understanding. (See
Rimmon-Kenan and Bal for demonstrations of incrementally intrusive
narration, and also the notes and exercises for this chapter.)
But there are other very useful typologies of narrational mode that are
worth consulting, including that of Uspensky (1973), and, a simplification
of this, Fowler (1986). More recently yet, Simpson (1993) has proposed a
revision and expansion of the Uspensky/Fowler scheme which is particu-
larly worthy of attention, not least since it is concerned with the kinds of
engagement that different narrators have with the material they narrate. It
is with a brief outline of and commentary upon Simpson’s scheme that I
will end this chapter.
To begin with, it is worth emphasizing that all these proposals are
attempting to look in an orderly and systematic way at the different kinds
of narration that a careful reader is almost inevitably going to notice, at
some level of consciousness. That is, the reader who reads this:
It is vain to say human beings ought to be satisfied with tranquility:

they must have action, and they will make it if they cannot find it. Mil-
lions are condemned to a stiller doom than mine, and millions are in
silent revolt against their lot. . . . Women are supposed to be very calm
generally: but women feel just as men feel. . . .
(Bronte, Jane Eyre, Penguin, p. 141; quoted in Simpson, 1993: 57)
knows that they are being addressed differently, or being addressed by a

different kind of narrator, than when they read this:
But suddenly a woman rose up before me, a big fat woman dressed in
black, or rather in mauve. I still wonder today if it wasn’t the social
worker. She was holding out to me, on an odd saucer, a mug full of
greyish concoction which must have been green tea with saccharine
and powdered milk.
(Beckett, Molloy, Picador, p. 23; quoted in Simpson, 1993: 52)
Both are first-person narrations, but neither fits more traditional categ-
orizations well (the former passage is both character-based and ‘omni-
scient’; the latter is both character-based and ‘estranged’), and clearly both
run through all six degrees of informativity on the Genette/Rimmon-
Kenan scale. What we need is a classification that will begin to recognize
and distinguish the fact that, whatever the narratorial intent, a reader may
find the former to be serious, even dogmatic, while the latter may appear
sometimes absurd or ludicrous. The former is full of sweeping generaliza-
tions and prone to the melodramatic (a stiller doom than mine); the latter
seems unsure even about the particular (dressed in black, or rather in
mauve) never mind moving to the general, or it combines vagueness with
uncanny detail, in ways which in this short extract feel closer to the low-
comedic than the sinister. Simpson’s model helps us to discuss these differ-
ences systematically. His approach entails looking at narration through
grammatical eyes, and thinking about the degree to which different narra-
torial stances are expressed through demonstrable aspects of the grammar
of their construction. So there is a continuity here with the interest in the
grammaticization of plot with which the previous chapter concluded (2.6):
the Simpson model probes the extent to which we can say that key narra-
torial modes are grammaticized. In what follows I rehearse Simpson’s
system, but add one or two glosses which may make the system easier to
memorize.
Simpson’s scheme distinguishes nine types or modes of narration. First
we can make a three-way distinction of narratives on the basis of relatively
explicit and well-recognized contrasts: is the narration first-person or
third-person, and if third-person, is the narration focalized through one
character’s consciousness (as happens in Henry James’s novels and many
since) or is the third-person narration ‘non-aligned’, emanating from some
detached point outside the consciousness of a particular I or a particular
she or he? In Simpson’s terms these three category alternatives are A, B
(R) and B (N), respectively, where A denotes first-person narration, B
denotes third-person, of which R denotes the Reflector or character-
mediated variety and N denotes the impersonal Narratorial option. This
tripartite distinction might also be annotated using personal pronouns,
labelling the three types I-, s/he-, and they-narration respectively. Calling
Simpson’s category A, or first-person, I-narration needs no explanation. In
a passage of she- or he-narration – Simpson’s B (R) – the third-person is
used, but the orientation is from the point of view of a particular she or he:
just what that he or she sees is what gets narrated. In a passage of they-
narration – Simpson’s B (N) – again the third-person is used, but there is
no adoption of the viewpoint of any particular she or he, rather we are
more detachedly told what they, one or more individuals viewed exter-
nally, do and say. The contrast between she/he- and they-narration is not
on the basis of singular versus plural, but on the basis of point of view
alignment (she/he-) or detachment (they-). They-narration need not report
the acts of two or more detachedly-viewed characters; it may report just
one person’s acts, but from a detached perspective, in which, I am suggest-
ing, that person is treated as a ‘they’.
The second axis upon which Simpson distinguishes types of narration,
again proposing three prominent types, is bound up with modality and
evaluation, a topic which is too complex for me to go into in any detail here
(the reader is directed to a number of useful summary accounts, starting
with Simpson’s own). In essence, in grammatical and textual studies,
modality refers to some of the crucial means by which a speaker qualifies
what would otherwise be absolute statements (like It’s wet and cloudy in
Lima; Tony borrowed my bike; and Tanya ate the pasta). So modality intro-
duces a kind of colouring of the discourse, investing utterances with some
of the commitments and reservations of its speaker or author (It seems that
it’s wet and cloudy in Lima; I didn’t mind Tony borrowing my bike; Tanya
must have eaten the pasta). Modality is a powerful indicator of point of
view, of the speaker’s or writer’s subjectivity; it is one of the means by
which an addressee feels they are hailed by a person with a voice and
human feelings, needs, burdens, and uncertainties. So it often supports or
fosters interactivity or connection between addresser and addressee. In dia-
logue and conversation, it gives rise to a sense of negotiation in which,
implicitly at least, the addressee feels that they have possibilities of taking
up the modalized claims of the speaker, rather than simply receiving and
submitting to them. In narrative – particularly completed written narrative
– the reader cannot practically ‘negotiate’ with the author, but the sense of
a modalized written telling as one that contrives to draw you into a
writer–reader conversation remains as a textual effect.
Simpson groups the various kinds of modality that one might identify
into two large tendencies: positive modality and negative modality. And
both of these must be contrasted with the third alternative, which is near-
total absence of modalization: flat, categorical, non-subjective statements
(of the Tanya ate the pasta variety). Within positive modality are all the
linguistic means of expressing the deontic (what must or ought to be done,
including duties and obligations and impositions in all their forms, from
strongest – I order you to – to weakest – I beg you to) or the boulomaic
(what is desired, wanted, wished for, or pleasing, again across the range
from strong to weak). To this ‘positive’ modality, we can add generalizing
and opinion-expressive sentences of the kind found in the Jane Eyre
passage above (there are many, the last of them being Women feel just as
men feel). And we should mention evaluative adjectives and adverbs
generally, along with verbs reporting a character’s thoughts, perceptions,
and reactions (she noticed . . . it annoyed her that . . .) collectively known as
verba sentiendi. A passage of narration which uses some of these resources
is, like the Jane Eyre passage, noticeably positive in modality, which here
means that the narrator appears to be engaged with and ‘upbeat’ about the
story they are telling, confident and in control of their material – also with a
sense that the material is ‘their’ material, that is, that they are in possession
of it. Relatedly, such a passage will tend to be marked by a backgrounding
or total absence of epistemic modality and words of estrangement (these
two will be explained below, under ‘negative modality’). The tone or tenor
is emphatic, confident, assured and addressee-reassuring.
By contrast, within negative modality fall the linguistic means of
expressing certainty and uncertainty about whatever is being reported –
again, using a range of means that include modal verbs such as might (have)
and must: She must have eaten the pasta (it was in the fridge this morning),
modal adverbs such as conceivably and undoubtedly, clausal constructions
like It is highly unlikely/probable that, and more idiomatic means: No ques-
tion, Tony was the one who took the bike. There may also be various verbs
of ‘speculative cognition’: she supposed that; I imagine that, etc. In addition
to these, narrative passages with negative modality include a number of
constructions that appear to refer to human perception (It’s obvious that . . .
Obviously . . . and similarly with other adjective/adverb partners: clear(ly),
evident(ly), seems/seemingly, looks/sounds like, and many more). Equally
predictably, the narrative will have few or no generic sentences, and verbs
reporting characters’ thoughts and reactions will be qualified: not she strode
along, intent on confronting the whole village but she strode along seemingly
intent on confronting the whole village. Or, most indicatively, using as if.
Many of these features can be grouped together and called ‘words of
estrangement’, in that they add to the reporting the impression that an out-
sider’s account is being given, rather than that of someone with ‘insider
knowledge’. A tone of doubt or guardedness, even of lostness or alienation,
may predominate: narration with negative-mode modality is often the
product of a narrator who is not in confident proprietorial control of the
story they are telling, but is tentative, confused, somewhat overwhelmed by
it or alienated from it (they are ‘self-questioning’, Simpson suggests). The
contrast between positive and negative narration, at the extreme, is all the
difference between:
1 the narrator who buttonholes you with their story because, themselves
knowing it completely, they wish to share it with you in the belief that
it is amusing, or instructive, or similar – at any rate, that it has a clear
point and effect.
and
2 the narrator who shares a story with you because its shape, point and
effect are, perhaps, unclear to them, as narrator, and they narrate as if
in the hope that sharing the story may help clarify the message, like
the messenger who conveys a message that they themselves do not
fully understand or associate with.
Clearly, type 2 narratives are much the more paradoxical and intriguing.
If positive narration implies, ‘Here’s a story and I, partly revealing
myself, know exactly what it means’ while negative narration implies,
‘Here’s a story but I, partly revealing myself, don’t know quite what it
means’, then neutral narration falls outside these two and implies simply,
‘Here’s a story (and I am neither revealing myself nor saying what it
means)’. In the neutral modes – which like positive and negative modes
can apply to first- or third-person (narratorial or reflectorial) narrations –
there is a nearly complete absence of narratorial modality. The teller tells
things categorically and ‘non-subjectively’, so that the tone is cool and
detached. Neutral modes are suited to physical description rather than
psychological development. Camus’s The Outsider is famed for its non-
modal reporting of topics (grief, desire, murderous hostility) where one
ordinarily expects modalization; other authors who draw heavily – but by
no means constantly – on the powerful effects of neutral narration include
Hemingway, Chandler, Carver, and a host of ‘hard-boiled’ crime-fiction
writers.
It is illuminating, in principle, to be able to create and compare versions
of a particular passage in several of the modes or styles listed above. Some
such transpositions are relatively easy to achieve, others are virtually
impossible, and those differences are themselves instructive. Thus, as
Simpson shows, transpositions between A, B(N) and B(R) versions of a
narrative passage, keeping the modality or lack of it the same, is straight-
forward. Thus transposing between what I have called I-narration, they-
narration, and s/he-narration, preserving the positiveness (for example) of
the original but modulating from an involved first-person narrator to an
external narrator to a reflector–narrator (or vice versa), is easily done and
can help to highlight the particular effects that an author’s chosen mode
permits. But ‘intracategory’ transpositions are significantly more problem-
atic. Transforming a neutral (unmodalized) account into a modalized one
(positive or negative) can be done, but not the reverse. For doing the
reverse, for example, converting a positive account to a neutral one, remov-
ing subjective and modalized language, usually renders the passage so full
of gaps as to render it incoherent. This is because many of the features we
have itemized as reflective of positive or negative narratorial shading are
not merely reflective of but constitutive of the passages in which they
appear. Consider, for example, the Jane Eyre passage cited earlier, which is
first person (note the mine in line 3) and positive:
It is vain to say human beings ought to be satisfied with tranquility:

they must have action, and they will make it if they cannot find it. Mil-
lions are condemned to a stiller doom than mine, and millions are in
silent revolt against their lot. . . . Women are supposed to be very calm
generally: but women feel just as men feel; they need exercise for their
faculties, and a field for their efforts as much as their brothers do;
Converting this to third person is straightforward, but dispensing with the

evaluative lexis, the generic sentences, the deontic and boulomaic modal-
ity, and replacing these with epistemic and perception modality, and
estranging comparatives (looked/seemed as if . . .) yields an entirely differ-
ent passage:
It seems that not all human beings are satisfied with tranquility: some
search out action, or they create their own. There may be many who
endure a more passive condition than that woman’s, and perhaps they
are in silent revolt against their lot. . . . Women appear very calm
generally: but some of them feel as some men do, and like them they
seem to need exercise for their faculties, and a field for their efforts.
Contriving such a ‘re-write’ is quite a dubious and controversial exercise,

so that what one is inclined to conclude is that the nine modes are like
horses for courses, and each has to be appreciated for its own qualities
rather than viewed in too much of a comparative perspective: there is no
way of writing A positive passages of Jane Eyre, in B(N) negative style,
that preserves the text as a Jane Eyre passage at all. The ‘conversion’
possibilities are summarized in the following diagram:
Positive Negative Neutral
(I ) A A A
(they) B(N) B(N) B(N)
(s/he) B(R) B(R) B(R)
Arrows with solid lines denote permissible transpositions (and their direction).
Arrows with dotted lines denote impermissible transpositions.
In analytical practice, it is often more appropriate to assess passages holis-
tically, for an overall impression of the positive, neutral or negative stance
of the narration. That is, it is unwise to approach a text hunting for epis-
temic or obligation modals, or generic sentences, or evaluative vocabulary,
or verbs of feeling and reaction and so on – very often one is bemused
to find few of these, or equally confusingly, a mixture which the
positive/negative contrast does not seem to predict. Rather it is best to
assess the passage as a whole, for its tone, before proceeding to look for
confirmation, in specific linguistic features, of its predominantly positive,
negative or neutral stance. So while this approach emphasizes the possibil-
ity of pinpointing tonal colouring and stance in the particularities of
textual choice (modals, verba sentiendi, generics, evaluatives, and so on), it
does not and cannot reduce to ‘item-spotting’. We have noted, for
example, that descriptions introduced by as if/as though, and like can
project estrangement and uncertainty (as if he were angry [was he angry?];
like a doctor [is she a doctor?]). But what of the following lines, from the
highly-descriptive opening of D. H. Lawrence’s story ‘Odour of Chrysan-
themums’, where both as if and like figure:
A large bony vine clutched at the house, as if to claw down the tiled
roof. . . . Beside the path hung dishevelled pink chrysanthemums, like
pink cloths hung on bushes. (emphases added)
Here we cannot simply note these forms and ‘read off’ narratorial
estrangement, removedness, or uncertainty. Here, both forms introduce
metaphors, and a narrator who adds metaphorizing elaboration (whether
of setting, character, or action) is usually asserting and enacting epistemic
engagement or confidence, rather than neutral or negative modality.
Indeed all the words of estrangement lack their negative-mode associ-
ations when they are used in figurative elaboration rather than literal or
factual report. This example is offered to show ways in which the nine-
mode model and its textual indices need to be applied alongside increas-
ingly detailed description.
With that said, it may be appropriate to close the chapter by looking at
a passage from one of this book’s ‘favoured’ literary narratives, and assess-
ing its choices among the narrational modes. What follows is the close of
‘Barn Burning’, where young Sartoris has run away from the entire patch-
work of misery and wretchedness engendered by his father:
The slow constellations wheeled on. It would be dawn and then sun-
up after a while and he would be hungry. But that would be tomorrow
and now he was only cold, and walking would cure that. His breathing
was easier now and he decided to get up and go on, and then he found
that he had been asleep because he knew it was almost dawn, the
night almost over. He could tell that from the whippoorwills. They
were everywhere now among the dark trees below him, constant and
inflectioned and ceaseless, so that, as the instant for giving over to the
day birds drew nearer and nearer, there was no interval at all between
them. He got up. He was a little stiff, but walking would cure that too
as it would the cold, and soon there would be the sun. He went on
down the hill, toward the dark woods within which the liquid silver
voices of the birds called unceasing – the rapid and urgent beating of
the urgent and quiring heart of the late spring night. He did not look
back.
This seems to fluctuate between B(R) and B(N) positive, depending on

the extent to which we feel that, despite being in language well beyond
that which we would associate with young Sartoris Snopes (but as we
know, ‘who speaks?’ is a different question from ‘who sees?’), nevertheless
much of this passage is told from Sarty’s perspective, judging that it is he
and not the narrator who reflects that walking will cure his cold and stiff-
ness. On balance it seems more appropriate to attribute these to Sartoris,
making him the focalizer of most of the passage – but they can also be
attributed to an external narrator, and there are no decisive indicators
either way. There are verba sentiendi here, abundant evaluative vocabu-
lary, and most strikingly a recurrent use of predictive would, postulating a
future that has not yet happened but which someone foresees will come to
pass: it would be dawn, he would be hungry, walking would cure his
feeling cold, and so on. Some of these latter are plausibly interpreted as
past tense reports of what Sartoris thought to himself, exploiting a style
known as Free Indirect Thought which will be fully explained in Chapter
5. Thus,
He got up. He was a little stiff, but walking would cure that too as it
would the cold, and soon there would be the sun.
can be annotated, to highlight Sartoris’s thoughts:
He got up. He was a little stiff [he realized], but walking would cure
that too [he thought to himself] as it would the cold, and soon there
would be the sun.
Too mechanical a search for modal markers might treat these would’s as
expressive of epistemic tentativeness, and the negative mode; but in
context there is clearly nothing tentative about them. In fact they are an
important element in a passage where, for the first time, the reader
glimpses young Sartoris not oppressed, tentative, or out of his depth: no
longer reacting in panic and despair to circumstances created by others.
Further reading
As in the previous chapter, the best places to begin further reading on the topics
discussed are Rimmon-Kenan (1983) Chapters 4, 6 and 7, and Bal (1985; revised
second edition: 1997). Rimmon-Kenan’s discussions of time and focalization, in
particular, include numerous useful examples from the literary canon, while Bal’s
long chapters on text and narration explain her distinctions painstakingly, with
much careful comparative analysis of simple constructed examples. Familiarity
with either of these should be adequate preparation for an analeptic move back to
their chief source, the work of Genette (1980; 1988). Genette (1980) includes a
lucid forward by Culler; it is essential to have some familiarity with Proust’s À la
recherche du temps perdu, on and around which the theory is elaborated, in order
to appreciate Genette’s analysis properly. Chatman’s books (1978 and 1990)
contain extensive stimulating discussion of these and related issues, with copious
exemplification from well known films and literary texts. Verdonk and Weber
(1995) contains many useful stylistic studies of narrative fiction. For the most direct
antecedent of the Simpson scheme of narrational modes, see Chapter 9 of Fowler
(1986), which concludes with a brief analysis of perspective switches – and their
functions – in a crucial scene in Mervyn Peake’s novel, Titus Groan. Among the
more advanced and technical discussions of narration and focalization, the follow-
ing are important contributions: Prince (1982); Lanser (1981); Jahn (1999); van
Peer and Chatman (2001); and Lodge (1977) and (1981) – Chapters 2 and 4 of the
latter are attractive applications by a critic and novelist interested in but not uncrit-
ical of narrative theorizing. Also to be recommended are, especially, Fludernik
(1993) and (1996), magisterial syntheses and proposals touching on many funda-
mental issues in narratology; Berendson (1981; 1984); Ryan (1981); Fowler (1981);
Uspensky (1973); Rifelj (1979); Nelles (1984); Warhol (1986); and Herman (1997).
Simpson’s nine-cell typology of narrators is set out thoroughly in Simpson (1993).
This chapter has entirely neglected the always intriguing business of second-person
narrative; a special issue of Style (28: 3), devoted to that topic, edited by Fludernik,
can be recommended.
Notes and exercises

1 Re-read William Faulkner’s story ‘That Evening Sun’, and then work through
the story’s opening paragraphs in sequence, making notes on what is going on in
each of them in terms of temporal order, duration and frequency; focalization;
and narration, and considering the following issues:
i Is paragraph 2 analeptic or is paragraph 1 proleptic?
ii What are the temporal boundaries of the basic story?
iii At what point(s) do we move from iterative-frequency narration to singula-
tive narration, and how is this shift achieved?
iv The description of Nancy in paragraph 3 invokes the image of her carrying a
balloon, and in paragraph 6 she refers to the children as ‘little devils’: in
relation to the story that unfolds, what unexpected significances may these
two allusions have, in terms of proleptic force, character elucidation, or
insight into the narrator–focalizer?
v What insights, however slight and inconclusive, into the psychological and
ideological partialities of the narrator–focalizer might be felt to be revealed
in paragraphs 1 and 2? Is the ‘we’ of paragraph 3 and onwards (‘we would
go a part of the way down the lane’) the same group as the ‘we’ of para-
graph 1 (‘we have a city laundry’)?
vi Most of this story seems focalized by the boy Quentin, aged nine. What is
the evidence to support this view? But paragraph 1 cannot have Quentin as
a child as focalizer: it is a view of a scene that only comes into existence
when that child is an adult of twenty-four. And who speaks, in the sense of
‘narrates’, in this story? Is it an adult narrator who writes (p. 292)
Dilsey was still sick in her cabin. Father told Jesus to stay off our place.
Dilsey was still sick. It was a long time. We were in the library after
supper.
vii Among the focalizeds, can we point to any differences (e.g. in detail of
description) between the way the other children, Caddy and Jason, are
focalized by Quentin, and the way Nancy is? Why might such differences
arise?
viii Which of the Simpson narrational modes does this story use?
ix In terms of the six degrees of narratorial intrusion and interpretation listed
early in 3.7, there seems scarcely any intrusive commentary here beyond
rank 4, i.e. few claims to special knowledge of characters’ motives and
impulses. What kinds of extra interpretive burdens does this place on the
reader? By contrast with the story’s ‘non-interpretive’ norm, consider and
discuss the following two cases of intrusive commentary. Are these depar-
tures from the general trend defensible, or are they blunders of ‘inconsis-
tency in telling’?:
A ‘What is it?’ I said. ‘What is it?’
‘I aint nothing but a nigger,’ Nancy said. ‘It aint none of my fault.’
She looked at me, sitting in the chair before the cold stove, the sailor hat
on her head. I went back to the library. It was the cold stove and all,
when you think of a kitchen being warm and busy and cheerful. And
with a cold stove and the dishes all put away, and nobody wanting to eat
at that hour.
B ‘Jason!’ mother said. She was speaking to father.
You could tell that by the way she said the name. Like she believed that
all day father had been trying to think of doing the thing she wouldn’t
like the most, and that she knew all the time that after a while he would
think of it. I stayed quiet, because father and I both knew that mother
would want him to make me stay with her if she just thought of it in time.
So father didn’t look at me.
2 Guided by the kind of questions raised above in the discussion of ‘That Evening
Sun’, attempt a similar analysis of time, focalization, and narration in one of the
following short stories: Joyce’s ‘Counterparts’, Flannery O’Connor’s ‘Revela-
tion’, Saul Bellow’s ‘Looking for Mr Green’, or John Updike’s ‘A & P’; or in the
opening paragraphs of Virginia Woolf’s Mrs Dalloway.
3 In the case of ‘Eveline’, it was argued, a Genettian relational assessment of dura-
tion might conclude that there was no clearly determinable change of pace, since
we cannot conclusively determine the real-time duration of events in either of
the story’s two sections. And yet many readers do feel a great shift in mood, an
extraordinary tension between Eveline’s frozen attitude and the urgency of
departure in the quayside scene. What are the grounds for that impression? The
answer may simply be that in the second and concluding section of the story, a
number of events are actually happening, and Eveline’s developing reflections
respond to those developing events. In the first section of the story, by contrast,
only one thing is happening at basic story level (she is reviewing her past and her
plans), even though that one ‘action’ involves recalling, as embedded particulars,
many events and incidents. In terms of perception of narrative pace, as in many
other respects, it seems that there is all the difference in the world between (a
report of) living a thing and simply recalling it. The second section may be short
in time, but is packed with sequenced events. The crucial thing seems not to be a
ratio of story time to textual extent but of story time to story events.
4 Much of the discussion in the later stages of this chapter leads towards a
consideration of what happens when a narrator wilfully or unwittingly distorts,
misleads, or suppresses. The topic of narrational unreliability is an extremely
rich theme, and has been probed by critics and theorists extensively for a very
long time (cf. Lawrence’s warning: ‘Never trust the teller, trust the tale’), espe-
cially vigorously in the last forty years, since Booth’s Rhetoric of Fiction
appeared. Of particular concern is the (un)reliability of intradiegetic narrators,
i.e. narrators visible – if only by way of the first-person pronoun – within the nar-
rative. That is to say, the detection of ‘corrupt’ narration is especially commonly
a challenge set for readers by intrusive/evaluative narrators (Simpson’s A and
B(R) positives). Narratorial unreliability in modern fiction is very widespread,
but often discussed examples include Bellow’s Dangling Man, Golding’s Free
Fall, Flannery O’Connor’s ‘Everything that Rises Must Converge’, J. M.
Coetzee’s Waiting for the Barbarians, Nadine Gordimer’s The Conservationist,
Ishiguro’s The Remains of the Day.
We attribute unreliability to any narrator the veracity of whose account we
come to suspect. Some narrators are liars, or consciously flatter themselves and
are clearly intended to be seen as attempting to deceive; other narrators mislead
for less culpable reasons: e.g. they may have the limited knowledge of a young
narrator, or be learning disabled like Benjy in The Sound and the Fury. Personal
involvement with events – especially when the narrator is a direct or indirect
victim of those events – may often give rise to narratorial suppression, distortion,
prevarication, and so on (as one example, consider Rosa’s account of events in
Absalom! Absalom!). In a more general way, abnormal values may give rise to a
type of unreliability that makes it difficult to decide whether we have a normal
narrator telling terrible things with much covert irony, or simply an awful narra-
tor. In assessing veracity and reliability, we have to act rather like a juror, weigh-
ing the evidence, looking for internal contradictions in what a narrator says
(especially when they serve that narrator’s purposes) or a clash between a narra-
tor’s representations of things and those of (other) characters whom we have
independent grounds for trusting and respecting. The great attraction and
danger of unreliable narration is, as Booth (1961) rather regrets, that no clear
moral or ideological stance is spelt out and held to, and we as readers are not
told what to think. But a fully articulated theory of what unreliability consists in,
and of the grounds for attributing it to one narrator but not another, remains
elusive and contentious. For important recent discussions of the topic, which
also usefully challenges us to rethink what we mean by reliable narration, see
Yacobi (1981), Wall (1994), Nünning (1999), and Fludernik (1999a).
5 In 3.2 I proposed that ‘time is perceived repetition within perceived irreversible
change’. What warrants that within? Why not say, on the contrary, ‘time is per-
ceived irreversible change within perceived repetition’? The former formulation
assumes change, variation, dynamism, chaos, out of which we pluck or impose
the orderly measuring that is the naming of times and dates; the latter formula-
tion assumes an underlying order, structure, and design, from within which erupt
changes, departures, and developments (whose location in time is, as it were,
already determined). These contrasting foundational assumptions about time,
repetition and change can be associated with two sharply counterposed world-
views, which might be called the Nietzschean and the Platonic (see discussion in
Lothe, 2000: 65–6, and references there to Deleuze and Hillis Miller).
text II
Character, setting, suspense, film
4.1 Character
Character, and everything it entails in the way of deep insight into the
minds of imagined others, their uniqueness of motive and difference of
worldview, is often what most powerfully attracts readers to novels and
stories. Yet it is the element in narratives that seems least amenable to sys-
tematic analysis. As a result it remains relatively neglected within narrato-
logical studies. To begin with, many narratologists were unconvinced that
here was a genuine topic to explore: what is called the ontological status of
character, individuals, and the self, was widely questioned (the ontological
status of a thing means its status as a part of existence: in what sense does
it ‘really’ exist, relative to other ‘really existing’ things?). This was prob-
ably part of a widespread reaction, in mid-twentieth-century literary
studies, to criticism and analyses which tended to assimilate characters in
literature to real people. Thus A. C. Bradley’s treatment of Shakespeare’s
tragedies as if they were the case-histories of real people, which formerly
had a profound influence on the study and performance of Shakespeare,
was increasingly seen as partial and distorting. There was also something
of a reaction against the heavy emphasis on character, the ‘bourgeois self-
determining Subject’, found in both nineteenth-century British novels and
traditional literary criticism about them. Nouveaux romanciers such as
Robbe-Grillet and Sarraute wrote their novels in a manner defiant of the
cult of the individual and any over-valuing of the allegedly unique
experience and response of particular personages whose psychologies
were to be dramatized. On the contrary, in such experimental fiction at
least, similarity of experience and personal behaviour, rather than dif-
ference, was asserted. And in tune with this, the structuralist preference
was to treat character ‘as a myth’, as Culler succinctly put it (Culler, 1975a:
230).
Character entails an illusion in which the reader is a creative accom-
plice. Out of words we make a person. A variety of descriptions of some
posited individual, together with descriptions – implicit or explicit – of that
individual’s actions and reactions, suffice to lead most readers to conceive
of a person of whom these references and insights are just glimpses. An
The articulation of narrative text II 81
iceberg principle is at work in the way most people read characters: we
operate on the assumption that the evidence we are shown is a necessarily
limited selection of material, that much more lies beneath the surface of
the novel, in the rest of that ‘person’s’ life.
But in the textualist–structuralist spirit alluded to above, many have
sought to revalue a new literalism which reminds us that novel characters
are really ‘just words’, are radically non-representational, and should not
be unthinkingly ‘recuperated’ by means of any direct and unguarded appli-
cation of amateur (or professional) psychological analysis. One of the
complaints about such responses – common enough in literature classes –
is that they ignore the art and textuality of novels, the degree to which
character, event, and everything else is a literary production, a construc-
tion. The mere verbal surface of novels having been ignored, there is no
clear limit to the arabesques of psychologizing: they can spiral on with pre-
cious little need for grounding in the text. Surface text can be dismissed as
censored testimony, heavily repressed, requiring startlingly unexpected
symbolic readings. By contrast, the ‘purist’ approach to character asserts
the non-referential or non-mimetic dimension in narrative art. In Wein-
sheimer’s memorable words (1979: 187):
Emma Woodhouse is not a woman nor need be described as if it were.
There are real differences of foundational assumptions separating the psy-

chologist student of character and the structuralist analyst of the actant.
But once we have acknowledged these contrasting viewpoints, it is often
sufficient to say that they are adopting quite different criteria for reading
narrative, and for identifying great narrative. A dual perspective can find
room for both:
In the text characters are nodes in the verbal design; in the story they
are – by definition – non (or pre-) verbal abstractions, constructs.
Although these constructs are by no means human beings in the literal
sense of the word, they are partly modelled on the reader’s conception
of people and in this they are person-like.
(Rimmon-Kenan, 1983: 33)
The latter part of this quotation tellingly emphasizes ‘the reader’s concep-
tion’ (strikingly rare in some of the earlier narrative theory), and recog-
nizes that characters are ‘partly modelled’ on real people. No matter what
some theories assert, readers continue to apprehend most novel characters
as individuals (whether seen dimly or sharply, whether recognizable,
comprehensible, lisible or impenetrable, alien, and unfathomable). And as
those apprehensions are built up, revised, and articulated, all sorts of
extra-textual knowledge, including our knowledge of characters in the real
world, is brought to bear. Thus while it is a mistake to assume that mimesis
or faithful ‘capture’ of the essential elements of real people is the truest
82 The articulation of narrative text II
and highest goal of fictional characterization, it would also be mistaken to
assume a more radical gulf between the ‘fiction’ of novels and the ‘fact’ of
the real world than there really is. It is not that literature in all its motile
fabricatedness cannot ‘map onto’ solid stable reality because of the
former’s instabilities and artifice; rather it is because our sense of reality
itself is a fluid construction, prone to shapechanging, that an enduring
mapping of it in literature or by other means is impossible. If, instead, we
think of the semiotic constructedness of people, things, and non-fictional
texts as of the same order as the semiotic constructedness of novels, then
we may come to see a middle way. From this perspective, art is not a sup-
plement to life, and hence cannot be simple-mindedly mimetic of it: both
art and life generally are understood to be representational.
4.2 Greimas’ actant model

One way in which character analysis has proceeded, both informally and
by more formal and systematic means, involves working with some notion
of ‘trait’, and that is the approach I will chiefly discuss in this chapter. It
amounts to a ‘bottom-up’ approach, a noticing of the accumulating evi-
dence to justify calling a particular character ‘athletic’ or ‘musical’ or
‘morbid’ or whatever the proposed trait might be. But here first, and
briefly, I will mention an explicitly ‘top-down’ approach, the actant model
of Greimas (1966). For all its reliance on intuitive schematization, it
remains workable and a stimulus to further reflection on character types.
Greimas worked along similar lines to Propp who, we have seen, had
earlier identified seven broad narrative roles filled by the main characters
in his Russian folktales, but he insisted that these roles were subordinate
to the 31 core functions or event-types, which he saw as the backbone or
main driving force of all the tales. Now Greimas proposed an inversion,
with events subordinated to character. As general categories underlying all
narratives, he suggested that there were just six roles, or actants as he
termed them, comprising three interrelated pairs:
Giver ⫹ Receiver
Subject ⫹ Object
Helper ⫹ Opponent
The six roles are usually diagrammatized as follows:
sender (superhelper) → object → receiver (beneficiary)

↑
helper → subject ← opponent
Such a model fits many traditional folk and fairytales remarkably well: the
subject or hero, perhaps a young man of lowly origin, seeks marriage to a
beautiful princess (object), in which case the man will also be the benefi-
ciary (possibly the princess and the country will too). In his quest he is
helped generously but with limited success by a friend or relative (helper),
but their combined efforts count for little in the struggle against specified
opponents (e.g. a wicked uncle of the princess, or some other eligible but
ignoble suitor), until a sender (often, in effect, a superhelper), such as the
king, or God, or some individual with magical powers for good, intervenes.
Here is the actant diagram again, but with possible exponent characters
filling the six core roles:
beneficent witch in → marriage to → young man, common

the woods princess weal, etc.
↑
friend of young man → young man ← wicked uncle of princess
Despite the model’s simplicity, and despite the need to annotate it vari-
ously so as to fit different genres better, the scheme can be usefully applied
to a range of texts. It is worth noting that two of the roles, sender and
object, are frequently not strictly characters at all in modern stories. The
sorts of things that count as object in our narratives, and the sorts of
special help we may get in our quest of those objects, are more often
abstract than concrete. Modern literary narratives are likely to have con-
ditions summarizable as fulfilment, liberation, happiness, self-knowledge,
or mental peace, as their object. At the same time there remain plenty of
spy stories, Westerns, romances and detective stories where a particular
tangible object clearly is the target: a particular criminal or lover, a secret
agent or document, or a coveted artefact of some sort. In many genres of
modern narrative, the role of giver/sender/superhelper has become quite
attenuated; but, again, there is an abundance of science fiction stories, TV
series, and Hollywood blockbusters where the intervention of a supporting
agent with higher powers remains de rigeur. One is tempted to think that
this persevering investment or faith in the idea of an intervening super-
power is part and parcel of a belief in a transcendent power, authority, and
standard. This is often much on display in American popular culture,
perhaps related to America’s being at present the sole superpower, but
versions of it can be found in every culture’s narratives. American popular
culture often restates a belief in America itself as a superpower, with a
special right and mission to intervene, rescue, and restore (the implicit
identification of the Rocky/Rambo personae with an imagined America, in
those films of the 1980s, particularly suggests this). As long as people
embrace myths of American can-do knowhow – or of plucky sangfroid
British resourcefulness, for that matter – or of a divine plan ensuring that
the virtuous and the oppressed will be finally rewarded – then the super-
helper role will surely survive in some of our narratives.
One among many applications of Greimas’ six-role schema is offered by
Vestergaard and Schroder (1985) in their book on the language of adver-
tising. They demonstrate how well the roles fit the dramatis personae in an
advertisement for Sanatogen multivitamin tablets. The addressee (‘you’) is
both subject and receiver, continued good health is the object, and your
pursuit of this goal is ordinarily assisted by vitamins and minerals from
meals (the helper), but made more difficult by some undesirable con-
sequences of your assumed busy life: snack lunches, dieting, reheated food
and skipped meals (the opponent). In steps Sanatogen as superhelper:
‘Sanatogen multivitamins give you essential vitamins and minerals.’ As
Vestergaard and Schroder observe:
Of particular interest are the facts that the role of object is not filled
by the product but by some quality or state associated with it, and that
the consumer is both subject and receiver. Advertising, in other words,
does not try to tell us that we need its products as such, but rather that
the products can help us obtain something else which we do feel we
need.
(1985: 29)
And the pattern revealed by the Sanatogen ad does seem to apply to a

variety of other utilitarian products too: face creams, tampons, beer, radial
tyres, breakfast cereals, washing powder, one-coat paint, toothpaste,
shampoo, after-shave lotion, and so on. In all advertisements in which,
then, there is a narrative-style emphasis on change of state, with a before
(‘How can I get rid of this dandruff?’/‘Why does my family need so many
fillings at the dentist’s?’) and an after (‘I’ve said goodbye to dandruff,
thanks to Glam’/‘Now we’ve switched to Gleaminfangs the dentist has
nothing to do!’), the product fulfils the role of important aid or accessory
in reaching or maintaining a certain quality of life whose inherent desir-
ability is – for some targeted audience – unquestionable. I add that caveat
about targeted audience because over the last few years there has been an
enormous growth of commercial advertising, in narrative format or other-
wise, which uses mass-audience media (TV, radio, billboards) while
implicitly singling out a particular group of addressees and calculatedly
opting to exclude and even alienate a significant proportion of the poten-
tial audience.
In some contrast to the advertising of everyday items like lotions and
breakfast cereals, it is noticeable that an explicitly narrative format tends
to fall away when it comes to promotion of truly luxury items such as per-
fumes, watches and jewellery, fur coats, very expensive cars, holidays, and
so on. In these, a more synecdochic or metonymic relation seems to
operate, where the product is presented as an intrinsic part (however
small) of the chic and elegant lifestyle that the advertisement typically por-
trays. (Synecdoche/metonymy is the rhetorical use of a phrase referring to
a part of or association with a larger whole, when reference to that larger
whole is to be understood; thus two common ways of synecdochically
referring to your car are by using the phrases ‘my wheels’ or ‘my motor’.)
Instead of representing the sender/superhelper, the product is part of the
object. As a consequence, the roles of superhelper, helper and opponent
(at least), and any marked sense of a before and an after, are usually
absent. Such luxury product advertisements are much more a description
than a narrative.
If we turn to application of the Greimas model to a literary narrative,
such as Joyce’s ‘Eveline’, we can use it to highlight the degree to which the
story is enriched by ambiguity and the paralysing effect of uncertainty, to
the point that these become part of the story’s theme. Eveline herself is
fairly clearly the subject and receiver, and her object would appear to be
freedom (especially freedom from verbal and physical abuse) and happi-
ness (at one point she reflects that ‘she wanted to live. . . . She had a right
to happiness’). In pursuit of this object it first appears that Frank is her
helper, while her potentially violent father is chief opponent. But at the
close of this very short story it is fairly clear that Eveline does not join
Frank on the steamer that would take her away to Buenos Aires. Why is
this? Is it due to the failure of a superhelper to appear? For the text
reports Eveline invoking such a superhelper:
Out of a maze of distress, she prayed to God to direct her, to show her
what was her duty.
But Eveline’s not going is more complex than this. Her prayer is not for
assistance with her escape, but with identifying ‘what was her duty’, i.e.
with deciding what should be her object, her first consideration. The neat,
five-role scenario sketched above is suddenly revealed as susceptible to a
rewriting, in which her current hard-working life (‘now that she was about
to leave it she did not find it a wholly undesirable life’), looking after her
father (‘Sometimes he could be very nice’), fulfilling her promise to her
mother ‘to keep the home together as long as she could’, collectively
undermine previous assumptions as to which (staying or going) is the
better object. In pursuit of this reinstated object of familial care and duty,
there seem to be two helpers: her father (on some occasions at least), and
her brother Harry, who ‘always sent up what he could’. The opponent is
now Frank, of course, and whether his intentions are honourable or not
becomes newly suspect.
But again, we seem to lack a powerful superhelper who might intervene
to arbitrate and resolve the choice between going and staying, choosing
Frank or the family. No doubt some will see Eveline’s ‘set[ting] her white
face to [Frank]’ as a clear choice, where the subject – as in so many
modern stories – has looked to resources within herself for the required
extra help. But the text does go on to describe her as ‘passive, like a help-
less animal’, which hardly fits an interpretation of Eveline as decisive
arbiter, come into her own powers. Ironically, a couple of textual clues
may prompt us to consider a physiological dysfunction, a weakness of the
heart in the face of acute anxiety, as a covert influence. In the course of
Eveline’s anguished uncertainty at the quayside, a distress which ‘awoke
nausea in her body’, we are told that ‘All the seas of the world tumbled
about her heart’. This may be sufficient textual warrant, perhaps, to recall
an earlier cryptic allusion to heart trouble:
Even now . . . she sometimes felt herself in danger of her father’s viol-
ence. She knew it was that that had given her the palpitations.
In their notes on revisions to ‘Eveline’ between its first magazine appear-

ance and its subsequent publication in the Dubliners collection, Scholes
and Litz remark on ‘the interesting addition of the palpitations’ (Scholes
and Litz, 1969: 239). As we read those earlier sentences we are entitled to
question this analeptic allusion: What palpitations? At the close of the
story, with Eveline paralysed physiologically as well as imaginatively, we
possibly see their effect.
4.3 Character traits and attributes

In many modern narratives of the more complex kind, the basic role or
function of a character as explored in the previous section – what the char-
acter does, in the plot – turns out to be far less interesting to the reader
than what the character is like. This is something of a paradox: what a
character does in a story may be essential for the text to count as a well-
formed narrative (‘He gets the girl in the end’/‘She detects and traps the
killer’, etc.), but what interests us is what kind of super-helper Sherlock
Holmes is shown to be, and just how he conducts himself in the course of
particular scenes and episodes. Details of characterization, the kind of
material that we have seen Barthes label as Indices or (mere) Informants,
often irrelevant to story, are equally often just what the reader finds
engrossing in a text. In deciding these judgments of type and manner of
character conduct we are inevitably very much guided – as Bal (1985: 80)
notes – by data from reality or extratextual situations. We carry to our
reading of a Sherlock Holmes tale plentiful knowledge gleaned from
various sources about doctors, detectives, crimes and human entangle-
ments. We may have ideas about more marginal aspects too, that are rele-
vant to a Holmes story, such as the nature of housekeepers, of pedestrian
traditional policework, and of Victorian beggars and urchins (and social
stratification in general). We may have even stood in London’s Baker
Street, looking – at least metaphorically – for 221b, Holmes’s fabled resi-
dence.
What this amounts to saying is that, in our making sense of any particu-
lar text, we have extensive resources of knowledge (sometimes called
extratextual knowledge, or knowledge of the world), which we can bring
to bear on our intepretation of the text under scrutiny. That ‘bringing to
bear’ will vary from reader to reader in at least two broad ways, to do with
the quality (depth, accuracy) of a reader’s knowledge, and the interpretive
evaluation the reader makes of that knowledge. We might summarize
these two components of adducible background knowledge as facts and
ideology.
To take a simple example, suppose I am reading a text which runs:
It is August 1880, and the prosperous streets of London’s West End

are busy with the carriages of ladies on leisurely shopping expeditions.
The afternoon is warm and sunny, but in the shaded doorway of a
house off Regent Street can be discerned the crouching form of a
beggar.
In the final words here an individual has been specified and, potentially at
least, we have encountered the first introduction of an important charac-
ter. But even without reading on, a certain fleshing out of that single
descriptive phrase, ‘a beggar’, takes place. I have some ideas about what ‘a
beggar’ is or means, but I additionally have ideas about what a London
beggar circa 1880 would be like, based on my own knowledge of late Vic-
torian social history, from whatever sources.
That knowledge is partial – in two senses, the factual and the ideo-
logical. The partial factual knowledge means that I cannot bring to mind a
depth of knowledge, and perhaps understanding, of the beggar that is
available to someone familiar with the details of the economic, social, spir-
itual, etc. life of Victorian beggars. The ideological partiality means that I
may have a certain view of Victorian beggars in general; I may regard
them as the victims of callous and inhuman neglect, a living index of a
moral degeneracy in the larger society. Ideologically of course that is just
one way of looking at Victorian beggars, and other readers may alterna-
tively assume that the dysfunction lies principally within the beggars them-
selves. Notice, too, that in these ideological glossings of the simple textual
denomination ‘a beggar’, I have been treating the beggar as some sort of
problem to be solved. That kind of emphasis seems triggered by the very
term ‘beggar’; it seems to be part of the interpretive semantic field the
term invokes, of a kind that is more extensive and subjective than anything
covered by a traditional dictionary entry. By comparison, consider what
may happen interpretively if the text instead ends in this way:
in the shaded doorway . . . can be discerned the crouching form of a

Buddhist monk.
Now a rather different body of extratextual knowledge is brought to bear

in a reader’s probabilistic characterization. Among other things, they
might note greater incongruity here, on the assumption that Buddhist
monks were rare in the London of 1880. But, despite the fact that Bud-
dhist monks rely on begged donations of food to live, the text has not
called this individual a beggar nor is the reader likely, ideologically, to
take the monk’s begging as the primary issue to have a view on.
Before proceeding it is perhaps worth stressing that limitations of
knowledge and partiality of view are not inherently disqualificatory. Some
readers will have fuller extratextual knowledge than others, or ideologies
more congruent with the narrator’s than other readers have, but we cannot
predetermine the relevance of those facts and views. The knowledgeable
reader may read too much into a marginal reference or character. Back-
ground facts and views are to do with actual Victorian beggars in general:
this particular beggar may be a very special one – Sherlock Holmes in dis-
guise, or one with mystical powers – so that the general type is of limited
help in our grasp of this particular character. The essential fact, despite the
need for revision and amendment of our probabilistic assessments as our
reading of a text proceeds, is that we do undertake this inference-based
fleshing-out of seemingly slight textual data, in character-comprehension
as in other matters. (The role of inference in relation to children’s narra-
tives will be examined in Chapter 7.)
4.4 Distinctive feature characterology

The interactive matching of textual facts and ideology with extratextual
facts and ideology that I have outlined above is, like the method of event
identification sketched at the close of Chapter 2, a ‘bottom-up’ type of
processing: it is, I would submit, an important feature of the experiential
real-time activity of reading, something we do as we read a text, not some-
thing postponed, or really postponable, until the entire text has been read,
when a holistic overview can be attempted. In principle, the ‘semantic
feature’ approach to character that I am about to discuss could be under-
taken either developmentally and incrementally or holistically, although it
tends to be presented in the latter way and may therefore appear overly
static. But if applied non-statically this resolutely structural approach can
be of value too.
A semantic feature analysis of the characters of a text involves specify-
ing a limited list of what the analyst takes to be the crucial features or
attributes which distinguish particular characters. Assuming that no two
characters are identical, a rather limited set of attributes needs to be
drawn up, attributes that (in a simple system) a character can either have
or not have, such that no two characters are assigned an identical set of
attributes. This method asks: What are the most important ways in which
any two characters differ? What are the narratively significant differences
between characters? In this way, it is intended, a simple but useable
schematic picture can be drawn up of all the crucial ways in which particu-
lar characters are the same as others and different from others. The
approach is based on distinctive feature phonology and componential
semantic analysis: influential methods of analyzing the fundamental
dimensions of sounds and meanings, respectively, in a language. For intro-
ductory discussion of these methods see, e.g. Finegan (1994), Fromkin and
Rodman (1978), or Bolinger (1981). An interesting demonstration of
distinctive feature (or semic) analysis of characters is to be found in
Fowler (1977: 36–8).
Although Bal suggests we might extend the modelling to note varia-
tions in degree (how weak or strong?) and modality (how probably?) of
attribute manifestation, this would be departing some way from the ana-
lytical simplicity of the basic technique. That technique, within its own
limits, is interesting in its highlighting of essential distinctions between
characters as the analyst sees them. There need be no requirement in dis-
tinctive-feature ‘characterology’ (as there would be in standard distinctive
feature analysis conducted by different analysts on the same dialect), that
all reader–analysts see the same set of features as significant, and agree
over their presence or absence, in a text’s characters. Such variant
emphases in analyses are in themselves, and in their reflection of analysts’
assumptions, an indication of interpretive and even ideological diver-
gences.
Among the fundamental elements in character analysis, a particular
delimitable topic that often attracts attention is simply the way a character
is designated in a text: in making reference to the character, in what ways
and in what proportions does the narrative make use of a proper name or
names, of definite descriptions, and of pronouns? The diverse and poten-
tially nuanced ways of naming a character, in the course of the narrative,
can create effects of irony or of sympathy, of narratorial approval or dis-
taste. And these can be quite complex, given that there is usually an
ongoing designating of several characters in any stretch of text by these
varied means. In the opening paragraph of Henry James’s story ‘The
Pupil’ (discussed in Leech and Short, 1981), for example, the young man
applying for the position of resident tutor is referred to, in order, in the
following varied ways:
The poor young man . . . him . . . he . . . (his) . . . he . . . he . . . he . . . (his)

. . . he . . . the candidate for the honour of taking his [the boy’s] educa-
tion in hand . . . this personage . . . he . . . (his)
None of the phrasal descriptions here (the poor young man, this
personage), we may surmise, would be adopted by the man to describe
himself or his position (we learn his surname, Pemberton, only in the
second paragraph). Embedded in the given text, the definite descriptions
in particular are a simple means of characterization which is also subtly
evaluative, covertly creating a tone of distanced sympathy for Pemberton,
distanced enough to permit ironical treatment of his situation. The picture
is further complicated in that, for example, in:
[the little boy] looked straight and hard at the candidate for the
honour of taking his education in hand
the description of Pemberton is as if it were the boy’s view of the situation.

And that example is a small indication of the complexity that emerges
once we consider the varied sources from which our information about
characters may come (from the character him- or herself, from another
character, from an external narrator, and so on).
Another heavily exploited means of characterization, often with the
assumption that outer surfaces reflect inner essence, is description of char-
acters’ outward appearance – especially their clothes and facial features.
Given the rarity (still) of surgical adjustments, we tend to think of the
latter as attributes a character is simply blessed with or stuck with, and we
may tend to reject any claimed causal link between appearance and
personality (such theories of physiognomy and phrenology, popular in the
nineteenth century, influenced writers such as Balzac and Dickens). On
the other hand powerful cultural and biological traditions associate
appearance with identity and character, the immediately ‘readable’ former
being taken as to some degree indicative, expressive and even constitutive
of the latter – as you can confirm if you frown at the next person you make
eye-contact with. Among countless literary examples we can mention
Cyrano de Bergerac’s protruding nose as one of his more significant attrib-
utes, thwarting his romantic quest, while Falstaff’s vast belly or ‘womb’ is
similarly non-incidental to character and story: his great waist reflects his
culpable great waste(fulness) as well as his verbal or situational fecundity,
as the play’s puns insist. And the smallpox that (mercifully, only temporar-
ily) disfigures Esther Summerson (in Bleak House) does threaten her like-
lihood of marrying and so threatens to alter the course of the story. Then
again, some features of physical appearance such as body shape, hair style,
facial expression (propensity to frown, with head lowered, as opposed to a
smiling disposition with an enquiring tilt of the head) are judged to be
partly under a character’s control, something for which they can be held
accountable.
The source of character apperception that I have neglected most
severely is the business of implicit characterization, based on how a char-
acter acts. We can always be judged by our deeds, of course, and in some
narratives oriented to events, descriptions and evaluations of characters
are so scarce that such indirect characterization is a reader’s only recourse.
Sometimes this points to the narrator’s rather limited interest in ‘telling
character’: instead, what happened is the overriding interest. But some-
times an indirectness of presentation is adopted for other reasons, where
an interest in characters and individuals remains strong. As Rimmon-
Kenan has noted (1983: 60–1), a good proportion of twentieth-century
fiction has tended to opt for indirectness of presentation, a ‘showing’ of
character; often this is done out of respect for the ability of the reader to
infer, evaluate, and draw conclusions on the basis of presented behaviour,
rather than of direct (and directive) presentation, an authoritative ‘telling’
of how a character is.
4.5 Setting
I have space only to touch on a few key aspects of the role of setting in
narrative text. Perhaps the first thing to say is that, although an elabora-
tion of setting is less essential to written narrative than event and character
are (we require a narrative to contain a sequence of events involving
change, and prefer those changes to involve or affect individuals with
whom we can sympathize or identify), the establishment of an identifiable
setting is a strong psychological preference in most readers. We like, in our
reading of narratives, to know where we are, and look for clear spatiotem-
poral indications of just where and when a thing happened. And it must be
emphasized that all of this discussion relates to setting in verbal narratives;
in film narratives, by contrast, setting is in a sense immediately and over-
whelmingly available, and plays a prominent role in the process of narra-
tion.
Just how specific a written narrative setting is expected to be seems rel-
ative to genre, type, and so on. Thus a story about a hen that outsmarts a
fox, a fable-like text whether for adults or children, does not need precise
anchorage to some particular farm of specified layout, with its fields,
machinery, and livestock described. Such particularity would even strike
us as bizarre, where the story’s power rests in its generic truth, its ‘pan-
situational’ universality. We can even cope quite happily with stories
about hens and foxes where the word ‘farm’ never appears: unless we
learn to the contrary, we simply assume a stereotypical rural background.
Relatedly, in Fielding and Austen novels that revolve around big houses
and their estates, parsonages, inns, and so on, the particularities of those
backcloths are rarely important. Settings in broad terms are of course
important here, for example, the fact that Fanny Price’s parents live in a
poky little house in Portsmouth while her uncle and aunt live in a grand
house at Mansfield Park. And, in the same novel, it is important that the
library in which the younger people present their entertainment is not
quite a suitable setting, just as the whole enterprise of the play is not quite
suitable. But the details of these settings are rarely crucial, since they are
rarely instrumental in plot development or a refraction of character.
In many modern novels, however, in which the oppressions of the built
environment (industrialized, technologized, homogenized) may be a
prominent theme, setting may be much more than backcloth. It may be
instrumental – like another character – in leading a character to act in a
certain way. This fuller role for setting – quasi-animate, menacing or
soothing, chorus-like or emblematic – goes back in the classical English
novel at least as far as Dickens, and continued to develop and modulate in
the work of George Eliot, Hardy, and Lawrence. But one genre of novels
where setting operates like a participant long antedates Dickens: the
Gothic novel tradition. Interestingly, for most modern readers, there is
‘too much setting’ in the Gothic novel, or it is cast in an implausibly promi-
nent role – with its dark, mysterious mansions bristling with medieval
towers, secret chapels, and concealed hallways and tunnels. In American
fiction setting seems to have been more prominent in the early literature
than more recently: in Cooper’s Leatherstocking novels, in Poe copiously
(most memorably, perhaps, in ‘The Fall of the House of Usher’), in the
American Gothic tradition that descended through Hawthorne to
Faulkner, and in Twain (what would Huckleberry Finn be without the raft
and the river?). Setting is probably to the fore in that literature for good
reason. Post-Independence American culture was sharply aware that it
constituted a new nation emerging on new terms in what was (for the set-
tlers) a new land: ‘received’ European settings would not do. But for all its
memorable depiction in more recent American fiction – in Ellison, Bellow,
Pynchon, Updike, Morrison, Silko and DeLillo – it is doubtful whether
setting routinely has quite the same instrumental role in postmodern
fiction that it perhaps had in modernist fiction. Increasingly – it may be
argued – our post- or late-modernist globalized condition makes particu-
larity of setting unimportant.
In simple terms, the relations between setting on the one hand, and
character and events on the other, may be causal or analogical: features of
the setting may be (in part at least) either cause or effect of how characters
are and behave; or, more by way of reinforcement and symbolic congru-
ence, a setting may be like a character or characters in some respects. As
an example of character:setting causal relations might be cited in Miss
Emily’s decaying house in Faulkner’s ‘A Rose for Miss Emily’, which in
turn is reminiscent of Miss Havisham’s wedding breakfast room in Great
Expectations. In most Dickens novels, it is easy to find causally-
determinative or -determined settings, from the fog-choked Chancery of
Bleak House to the debtor’s prison of Little Dorrit. In Faulkner, similarly,
setting counts: in ‘Barn Burning’, the wretched temporary hovel which the
Snopes family move into matches their miserable and oppressed tempera-
ments just as surely as the shining white mansion of their new landlord,
Major De Spain, suits the grandeur and refinement of the major and, espe-
cially, his Eastern-travelled wife. But contrary to Sartoris’s hope or assump-
tion, there will unavoidably be some dirty intercourse between these
impossibly different worlds – here in the form of the soiled, unrestorable rug.
Examples of analogical relation, where setting is like character in some
respect, are too much the norm to merit extensive comment. This is the
case particularly in relation to habitation where, in normal circumstances,
a character is likely to have some control over the kind of rooms or house
that they are shown to live in: in such cases, it should be unsurprising that
the house ‘fits’ or reflects its occupier. On the other hand in many narra-
tives a character does not have such control over their own living con-
ditions, so that an ‘unlikeness’ is equally possible. Toni Morrison’s Sethe
in Beloved and Dickens’s Oliver Twist are two examples of this, but we
might also mention Eveline, seemingly looking for the last time around the
dusty room which is not her room. But where it does obtain, analogical
relation of setting to character is more interesting when it involves a
matching of character with the larger natural environment, rather than
merely with residence. For this can hardly be by character design: Heath-
cliff and Catherine (in Wuthering Heights) have not made the wild, stormy,
elemental moorlands match their temperaments, so that if this is by design
it is that of a higher invisible power. And similarly Captain Ahab (in Moby
Dick) and Kurtz (in Heart of Darkness) are mesmerizingly congruent with
their extremity of setting. But perhaps more common than clear instances
of causality or analogy are texts which are indeterminate between these
two types of relation. Hardy’s ‘characterizations’ of Egdon Heath in The
Return of the Native, for example, are so distracting and compelling partly
because we sense, but cannot fathom, the influences that the heath has on
specific characters.
While many of these examples tend towards the broadly Gothic or per-
sonifactory, it is also true that far more conventional, ‘undramatized’, set-
tings play an important part in promoting verisimilitude and indirect
characterization. Senior civil servants have to live in a ‘civil servant’ style of
housing (unless there is good reason for the unexpected in domiciliary
setting): that is the unmarked option. Here the broad details of the house
(detached; with several bedrooms, a large kitchen and a detached brick
garage; in Surrey or Berkshire), and the garden (half-acre; well-kept; lawn,
roses, and fruit trees), in their emphases on the comfortable, rational and
unostentatious, will be assumed to be characterizing. So too will the specifics
of the internal furnishing (the living room has an original Victorian land-
scape over the genuine and occasionally-used fireplace, the comfortable
armchairs are from Heal’s, no television can be seen, but the hi-fi – with CD
and DVD player – is well-stocked with Deutsche Gramophon recordings,
etc.). All such details of setting articulate their owner’s intelligence,
conservative good taste, moderate wealth, cultured values – and utter
remoteness from, say, refugees or eco-warriors.
The above stereotyping need not be very close to the facts about real
senior civil servants’ homes and interests, some of whom may be so abnor-
mal as to vote for the Green Party and guzzle Diet Coke while playing
their Abba favourites. The point is that, in simplifying and standardizing
the world around us, we construct stereotyped portraits of civil servants –
and doctors, spies, politicians, travel agents, farmers, shopkeepers, garage
mechanics – and of how they live. In particular texts – again, unless there
is good reason for things to be to the contrary – we expect particular spies
and farmers broadly to match our stereotypes.
A false impression could easily be drawn from the foregoing concentra-
tion on stereotyping, extratextual, knowledge-based inferencing, and on
familiarity and predictability of characters and their behaviour. It could be
inferred that characters are mere assemblages of devices, artificial con-
structs whose seemingly natural properties are themselves convention-
bound, being the kinds of attributes particularly valued or ‘privileged’ by
the societies from which those narratives have emerged. But the view
of character as convention-based and convention-bound, common in
structuralist treatments, can easily be overstated. It will seem most plaus-
ible if the analyst focusses on similarities in characterizations within and
between texts to the neglect of those differences not amenable to struc-
turalist explanation. In the case of literary narratives at least, I would want
finally to emphasize their richness of narratorial and characterological
texture. This permits the construction or projection of characters who are
not ‘natural’ of course, but are still sufficiently distinctive and unique to
transcend dismissal as merely conventional; they are what Forster called
‘rounded’ characters. What would Dorothea Brooke have done if she had
met Ladislaw before Casaubon? How would Lambert Strether have lived
if he had stayed on in Paris? Did Cash spend the rest of his life with Anse
Bundren and his new wife? Will Sethe and Paul D marry? We may specu-
late (and it does not immediately seem absurd to do so), but we can hardly
project mechanically, for these characters are neither so static nor so pre-
dictable.
If these points are plausible, a revaluation of Bradley’s analyses of
Shakespeare’s characters seems in order. His approach cannot claim to
furnish adequate accounts of the tragedies as aesthetic wholes, but, as
Chatman shows in a spirited and persuasive defence (1978: 107–38), analy-
ses invoking relevant categories from the general vocabulary of psychol-
ogy can lead to original and plausible interpretative conclusions.
4.6 Character and setting in ‘The Dead’

In order to demonstrate in more detail the textual means by which charac-
ter and setting are articulated, I will concentrate on the final story in
Joyce’s Dubliners, ‘The Dead’. The story’s opening is as follows:
Lily, the caretaker’s daughter, was literally run off her feet. Hardly had
she brought one gentleman into the little pantry behind the office on the
ground floor and helped him off with his overcoat then the wheezy hall-
door bell clanged again and she had to scamper along the bare hallway
to let in another guest. It was well for her she had not to attend to the
ladies also. But Miss Kate and Miss Julia had thought of that and had
converted the bathroom upstairs into a ladies’ dressing-room. Miss Kate
and Miss Julia were there, gossiping and laughing and fussing, walking
after each other to the head of the stairs, peering down over the banis-
ters and calling down to Lily to ask her who had come.
It was always a great affair, the Misses Morkan’s annual dance.
Everybody who knew them came to it . . .
(175)
This opening is focalized from Lily’s orientation, and bears many traces of
her speech which we may remark on later in this chapter. But here the
chief things to note are that three named individuals are introduced, the
first an employee of the other two. We do not yet know whether any of
these will play a major role in the story, but in the given situation at least,
it being ‘the Misses Morkans’ annual dance’, the two sisters will be promi-
nent since they are the hosts. And the setting is evidently a modest private
house, one which is both a residence and a place of work, if the reference
to the ground-floor office is rightly interpreted. For this special occasion –
an annual dance – pro tem conversions of a ground-floor pantry and an
upstairs bathroom, into dressing-rooms, have been contrived.
Already the indices of genteel standards upheld despite straitened
financial circumstances are numerous. We note again, in the setting
description, the ‘wheezy hall-door bell’, the ‘bare hallway’, and the evi-
dently short distance separating Kate and Julia at the head of the stairs
from Lily down below. This is no grand house, nor one that is opulently
furnished. Modest means, and a sense of things worn but still functioning
(the bell is wheezy, but works) seem to be the tenor. Perhaps, analogously,
the Misses Morkan are ‘worn but still functioning’, but we cannot yet be
certain of their age. They’re old enough to throw – annually – a party for
ladies and gentlemen, and we may associate their ‘fussing’, in particular,
with stereotypical spinsterly behaviour. But equally important are the indi-
cations that they go together, behave alike, and have a shared life: they are
named together in a coordinate phrase, we are told (analeptically) that
they had resolved the dressing-room problem together, and they respond
identically to their sense of excitement, ‘walking after each other to the
head of the stairs’, and so on.
On the basis of the above evidence, we are already entertaining quite
elaborate ideas about the women introduced and their manner of living.
All this is under way even before the second paragraph, a mixture of
analepsis and descriptive pause, which reports that the residence is a dark
gaunt house on Usher’s Island, the upper part of which they had rented
from Mr Fulham, the corn-factor on the ground floor (176) and that the
Misses Morkan’s only niece, Mary Jane, lives with them and is now ‘the
main prop of the household’. Typical of the revisions we may often have to
make in our constructions of character and setting, we now – in the light of
the reference to the corn-factor on the ground floor – have to cancel any
supposition that the ground-floor office is the Morkans’ place of work.
Other inferences are confirmed, however: the sisters are old – Julia is quite
grey, Kate is too feeble to move from the house. Yet these women are
resilient, independent, committed to certain standards: from Lily’s focaliz-
ing perspective, they consume ‘the best of everything: diamond-bone sir-
loins, three-shilling tea and the best bottled stout’, and ‘would not stand
. . . back answers’.
To begin on a distinctive feature or trait analysis we need simply to
draw up a list of characters, set out across the page, and a list of attributes
set out down the page, and note the occurrence or otherwise of each
attribute in each character. But an ambiguity arises where the absence
of a trait is noted: absence may mean ‘no indicators of the presence
of’ or ‘evident lack of’. For example, if a character is marked as
‘⫹ strongminded’, does the negative counterpart denote average strength of
mind or downright pusillanimity? Here lies a problematic difference from
distinctive feature phonology; there, anything not [⫹labial] is inescapably
[⫺labial], but we cannot similarly say that anyone not strongminded is auto-
matically weakminded. I proceed below on the assumption that, in the case
of gradable attributes such as youthfulness (by contrast with absolute attrib-
utes, such as male/female, single/married), I am marking only presence or
absence of the specified trait, with no further assumption that absence of a
trait implies the presence of its opposite. Thus, in terms of the example
above, we will need to note positively weakminded characters on a separate
trait dimension from that of ⫹/⫺strongminded.
Table 4.1 is merely a first approximation. On rereading the story several
of these may be dropped from the reckoning as not being particularly
salient, while yet others may merit inclusion. A close look at the table will
show, for example, that ‘self-conscious’, ‘narcissistic’ and ‘superior’ have
identical profiles and thus do nothing to distinguish the characters: two of
these classifiers can be discarded, or the three categories can be merged.
And some of these attributes are greatly in need of annotation. Thus the
important adjective ‘generous’ I attribute to Kate and Julia (the party puts
them to real expense they can ill afford, their concern for their guests
seems utterly genuine), to Gretta in the generosity of her memory of
Michael Furey, but not to Gabriel, despite the numerous textual allusions
to his generosity. For all those allusions are suspect or qualified in some
respect, whether it is his self-restoring tipping of Lily; or his wife asserting,
Table 4.1 Character-trait inventory for ‘The Dead’
Kate/Julia Miss Ivors Gretta M. Furey Gabriel D’Arcy
Female ⫹ ⫹ ⫹ ⫺ ⫺ ⫺
Young ⫺ ⫹ ⫹ ⫹ ⫹ ⫹
Old ⫹ ⫺ ⫺ ⫺ ⫺ ⫺
Married ⫺ ⫺ ⫹ ⫺ ⫹ ⫺
Vulnerable ⫹ ? ⫹ ⫹ ⫺/⫹ ⫺
Vigorous ⫺ ⫹ ⫺ ⫺ ⫹ ?
Passionate ⫺ ? ? ⫹ ⫺ ⫺
Emotional ⫹ ⫺ ⫹ ⫹ ? ⫺
‘Generous’ ⫹ ? ⫹ ⫹ ? ⫺
‘Mortal’ ⫹ ⫹ ⫺ ⫹ ⫺/⫹ ⫺
Frank ⫹ ⫹ ⫹ ⫹ ⫺ ⫹
Covert ⫺ ⫺ ⫺ ⫺ ⫹ ⫺
Fussy ⫹ ⫺ ⫺ ⫺ ? ⫹
Narcissistic ⫺ ⫺ ⫺ ⫺ ⫹ ⫹
Superior ⫺ ⫺ ⫺ ⫺ ⫹ ⫹
Humble ⫹ ⫺ ⫺ ⫹ ⫺/⫹ ⫺
Self-conscious ⫺ ⫺ ⫺ ⫺ ⫹ ⫹
Erotic ⫺ ⫹ ⫹ ⫹ ? ⫺
Prosperous ⫺ ? ⫹ ⫺ ⫹ ⫹
‘You are a very generous person, Gabriel’ (217) even as Gabriel, fired by
covert physical desire for her, strives ‘to restrain himself from breaking out
into brutal language about the sottish Malins and his pound’; or the ‘gen-
erous tears’ that fill his eyes later (223), which are not so much tears for
Michael Furey but a self-pitying sorrow that he, Gabriel, has never felt so
selfless a passion for a woman. Similar qualificatory annotation should
accompany the trait ‘mortal’ – by which I mean whether or not a character
seems to have a sense of his or her own mortality. At the opening of the
story, mortal seems no more than a playful meiosis to Gabriel, complain-
ing of the ‘three mortal hours’ it takes Gretta to prepare for the party
(thereafter the story duration itself extends to little more than three
mortal hours); by the story’s close, all his thoughts are on death and fading
away. In this respect, and perhaps in others (from ⫺[sense of] vulnerabil-
ity to ⫹vulnerability?), it seems clear that Gabriel changes in the course of
the story. In the table I have separated earlier and later attributions with a
slash mark. Some such marking of characterological development – where
that occurs significantly – does require representation. So ⫺/⫹ in the table
does not mean ‘simultaneously having and not having the attribute’ but,
developmentally, ‘initially not having and later having the attribute’.
I shall leave the reader to explore in detail what the table implies about
individual characters and ‘overlapping’ characters. I will simply note that it
does draw our attention to the extent to which Gabriel is more like
D’Arcy than he is like Michael Furey, and the extent to which Gretta is
more like Michael Furey than she is like Gabriel. If we set aside the first
four attributes (none of which are conditions on the basis of which we
could reasonably pass a moral judgment on a character), it is quite striking
that, at least at the outset of the story, Gabriel and Michael Furey contrast
on practically every trait. We might also note that the table implies that
Gabriel is the only character that displays ‘trait-change’: might it be that
those characters showing trait changes, or question-marks against trait-
attributions, are those that readers find most interesting?
If we turn to setting in the story, we find subtle exploitations of the
ambience of setting, of change of ambience with change of setting, and of
noticeable inappropriateness of behaviour to setting. There is first the
public setting of the drawing room, where the music and dancing goes on,
and the more private back room, where the drinks are dispensed. In very
broad terms, there seem to be styles of interaction appropriate to those
domains, but it may be that the major determinant of these styles is the
fact that one room has the music and more women than men while the
other has the alcohol and more men than women. The public discourse of
the drawing-room setting can degenerate into the inconsequentiality of
Mrs Malins’s ramblings (rendered in free indirect discourse),
Her son-in-law was a splendid fisher. One day he caught a fish, a beau-
tiful big big fish, and the man in the hotel boiled it for their dinner.
(191)
while in the drinks room a discordant note may be struck when the pre-
vailing tenor, of masculine familiarity, is over-extended, as when Mr
Browne speaks ‘a little too confidentially’ to the young ladies. This
refreshments room undergoes a change of status, however, when it
becomes the supper-room dominated by a table laden with Christmas
delights, the stage for Gabriel’s speech of thanks extolling ‘genuine warm-
hearted courteous Irish hospitality’ (203), ‘good fellowship’ and cele-
brating ‘the Three Graces of the Dublin musical world’ (204). But
particularly telling are the arhythmic awkwardnesses of atmosphere in the
transitional scene that takes place in the hall as various guests prepare to
leave. This is the occasion both of D’Arcy’s needless but revealing rude-
ness and of his rendition (distantly) of ‘The Lass of Aughrim’, a song that
stirs such strong memories in Gretta (whose outward reactions stir such
strong but unrelated feelings in Gabriel).
Because the main dialogue takes place in the hall itself, but the charac-
ters and their words also relate both to the pantomime with the cabman
out on the street and to D’Arcy singing upstairs, complex spatial relations
between the near and the far develop. If the antics with the cabman are a
farce which the audience impatiently waits to depart, the song, expressing
tragedy, is a performance the audience yearns to draw near. Like Pnin’s
memories of Mira Belochkin, the song retells (brings near again) an old
grief, but only works properly when heard at a distance: D’Arcy’s voice is
‘made plaintive by distance’ (210). (Here, in fact, is a spatial articulation of
grief that we can set beside the temporal one sketched earlier in 3.2.1.)
And it is just the special configuration of perspectives and focalizations
that Gabriel experiences – ‘a woman standing on the stairs in the shadow,
listening to distant music’ – that makes him think of this as a picture, as
symbolic. The reader, relatedly, is prompted to interpret this ‘audio-
spatial’ scene symbolically – but with the enrichment that for us Gabriel,
too, is within the symbolic scene.
But nowhere in ‘The Dead’ is setting more powerfully used than in the
closing paragraphs, as Gabriel lies down beside his sleeping wife, and
reflects on the evening’s events, which have concluded in such an unfore-
seen way with his wife’s revelations. Critics continue to dispute whether
Gabriel here ‘transcends’ his earlier limitations and inversions of vision
and sympathy. But what seems beyond dispute is that setting here takes
on the role of companion and herald, then catalyst, and finally, in the
ubiquitously falling snow, of essence. Certainly, as he and Gretta
approach their hotel bedroom, and seeing things from Gabriel’s view-
point as we do, the setting they are approaching promises to determine
events: always usually surrounded by family and responsibilities, for this
one night Gabriel imagines they can be young lovers again, filling the
room with their passion and lust. But we see how his projection of setting
and events modulates into a different version as Gabriel comes to see the
room as a cold, dark, rented box to lie down in. There is a challenge here
to the stylistic analyst to chart the linguistic means by which, quite
rapidly, the categories of character and setting begin to dissolve, the text
announcing at one point, ‘His own identity was fading out into a grey
impalpable world’ (223).
4.7 Creating surprise and suspense in narratives

A crucial means of enriching character and event presentation, closely tied
to matters of plot progression, is the topic of this section: the ways in
which narratives are made more experientially engaging of the reader by
effects of either suspense or surprise. How exactly are these quite different
effects narratively created? My interest is specifically in plot-based sus-
pense and surprise and not, for example, the arguably less-powerful sus-
pense created by delayed identification of the true perpetrator in crime
and detection stories.
Consider Katherine Mansfield’s short story, ‘Bliss’. The story thema-
tizes Bertha’s ‘bliss’, her sense of charged gladness at being who she is and
where she is, and now, at the dinner party that is the story’s backcloth,
seemingly alive with sexual desire for her husband Harry for perhaps the
first time, blissfully in love with him and their life together. And then, to
our and Bertha’s devastating surprise, at the very close of the evening, she
finds that Harry is evidently embroiled in a passionate affair with Pearl
Fulton, the cool enigmatic beauty with whom she, Bertha, had thought –
when jointly admiring a pear tree in the garden – that she had achieved a
special communion. Turning towards the glittering pear tree, Bertha ends
the story on a note of painful suspense: ‘Oh what is going to happen
now?’. Two further examples will be given: from Jane Smiley’s novella,
The Age of Grief and Hemingway’s story, ‘Indian Camp’.
The Age of Grief tells the story of a marital break-up interspersed with
the story of a virulent influenza, which in turn hits every member of this
affluent professional-class Canadian family who, one might imagine, could
not be damaged by something so banal, so nineteenth century, as the ‘flu’.
Both parents, and each daughter in turn is laid low by the sickness, only to
gradually recover; and then Stephanie, the toughest child, succumbs, her
fever rising to a steady 105 degrees, and eventually falls unconscious. Smiley
brilliantly builds the suspense over whether Stephanie will actually die from
the ‘flu’ – all focalized from the viewpoint of the father, a successful dentist
like his estranged wife. And it is partly because the focalized narrative has
shown us, in detail, other family members falling sick but then recovering,
that we all the more fear that the outcome with Stephanie may be different,
and bad.
‘Indian Camp’ appears to be the story of how Nick Adams and his
doctor father and his uncle George, on a fishing trip, come to the aid of a
pregnant Indian woman who goes into labour prematurely. Her life and
that of the baby can only be saved – if they can be saved – if Dr Adams
performs a Caesarian section using whatever tools are to hand, and, of
course, without anaesthetic. Suspense is created as the reader hastens to
learn the fate of woman and baby. Despite the woman’s terrible pain (it
causes her to bite George, who is helping to hold her down), she and the
baby come through, and all appears to be ending well, with Dr Adams
particularly proud of his efforts: we are told he feels ‘exalted . . . as football
players are in the dressing room after a game’. But with the suspense past,
the reader is in for a powerful surprise. With his patients settled, the
doctor turns to the Indian woman’s husband.
‘Ought to have a look at the proud father. They’re usually the worst
sufferers in these little affairs,’ the doctor said. ‘I must say he took it
all pretty quietly.’
The reader is unlikely to have remembered the text’s earlier brief refer-
ences to the man, who is lying under a blanket on the bunk bed above his
wife’s bed: the wound in his foot, his smelly pipe smoking, or his sub-
sequent rolling over against the wall. Now the doctor draws back the
blanket to see the man, and the reporting is quite straightforward, matter-
of-fact, and unconcealing: ‘His throat had been cut from ear to ear’, and
this by his own hand – an open razor is found in the blankets – evidently
having found his wife’s suffering unbearable.
Stories such as The Age of Grief and ‘Indian Camp’ would suggest that
plot-based suspense is created where two broad conditions obtain:
1 the narrative ‘forks’ in a Barthesian sense of reaching a point of devel-

opment where very few (often just two) alternative continuations or
outcomes is highly predictable, so that one or two (just a few) narrra-
tive completions are clearly ‘foreseen’ by the reader.
2 at this point of narrative forking between broadly predictable comple-
tions, both or all such completions are ‘withheld’: the disclosure of just
which completion obtains in the present narrative is noticeably
delayed, beyond its earliest reasonable report.
A simple example of condition 1 would be where a character in a story

decides to smuggle an illegal item through customs. Several pages of nar-
rative might then report the character’s journey to their point of departure
from the first country. None of those pages would be suspenseful in the
sense specified here, though they could be ‘gripping’, interesting, anxiety-
inducing, etc. But once the character reaches customs, so that imminent
reporting can be expected as to whether or not they are stopped and
caught, then any delay or elaboration of those next steps creates and con-
stitutes suspense.
Where it occurs, suspenseful text is likely to be marked by quasi-
Labovian (see Chapter 6) evaluative text: descriptions of states, moods,
settings, ongoing circumstances without clear temporal limits (with intensi-
fying effects); reports of flashbacks, flashforwards, hypothetical and imag-
ined outcomes, and contemporary events other than those of the
presently-focussed-upon narrative line (comparator and correlative
effects); and reports of background causes, motivations (explication).
Sometimes, suspense can be achieved by ‘cutting’ to seemingly entirely
unrelated events, such as a wholly separate narrative line (as Dickens does
in cutting between two narratives in Bleak House).
Conditions 1 and 2 above do not include any requirement to the effect
that a resolution or outcome must be disclosed, so as to bring an end to the
prolonged uncertainty created during condition 2. In Raymond Carver’s
story ‘The Bath’, a little boy called Scotty is knocked down by a car on his
birthday, the day on which his mother was to have collected a birthday
cake ordered from a local baker. Scotty then lies in a coma for several
days at the hospital, attended by his desperately anxious parents. From
soon after Scotty’s removal to hospital, then, the story is suspenseful as
we, like the parents, wait for a completive resolution. But the story ends
with the beginning, only, of a phone call: ‘This has to do with Scotty,
yes . . .’ – so that no end to the suspense is reached. (In Carver’s later
re-working of the story, ‘A Small, Good Thing’, suspense is released since
we witness Scotty’s death; thereafter a minor secondary suspense is
created when the wife goes down to the baker’s premises intent on assault-
ing him for his harassing phone-calls.)
Narrative surprise of the kind that is of interest here seems to occur
where, somewhat in contrast to suspense, a different initial condition
obtains:
1 the narrative has approached and may be presumed to be passing an

unproblematic ‘milestone’, where there is little or no sense of poten-
tial forking into different sequels and where, rather, a stereotypical or
schematic next event or scene is strongly predicted. Hence, in ‘Indian
Camp’, the stereotypical story schema is that of the worried husband
attending his wife’s difficult labour, but since this turns out alright and
mother and child emerge safe and sound, one expects that if the narra-
tive turns back thereafter to the father, it will be to describe and
record his relief, outpouring of joy, humble gratitude, and so on.
Thereafter:
2 the expected ‘non-forking’ schematic or automatic progression does

not go through as predicted, and something relevant and related,
‘imaginable in the circumstances if we had operated with full foresight
or imagination’, happens. And this foreseeable but unforeseen devel-
opment pulls us up short, causes us to re-assess much of the narrative
whose shape we thought we already understood.
To summarize, the essence of narrative surprise is that a reader experi-

ences a new development as unforeseen but, upon reflection, foreseeable.
Thus not every kind of arguably ‘surprising event’ amounts to a narrative
surprise in this sense; when Scotty is knocked down by a car in both ‘The
Bath’ and ‘A Small, Good Thing’, this calamity is something of a surprise,
but it is not foreseeable on a long view, reaching back to the beginning of
the narrative. To say that an event is, on analytical reflection, foreseeable
is also to say that it ‘fits’ the larger structure of narrative conditions and
developments in the entire story: it is not a detached addendum, but a fact
or outcome that can be fully integrated (often belatedly) with everything
else, filling a gap we had not even noticed was there. Surprises, when they
come, are most effective when they are felt to be in no way absurd or
inexplicable, but reasonable and possible. To experience a really effective
narrative surprise is to be caught up in an activity of self-teaching, of
reflexive critique, and this is part of what makes them so valuable. They do
not entail simply a revision, by the reader, of their grasp of the narrative;
with a little jolt of correction, they also compel the reader to examine and
find lacking their own understanding: ‘I should have seen/realized that this
had happened or was going to happen or was the case. Why didn’t I?’ And
these surprises are finally about understanding, and not about informa-
tion-failure on the part of the narrator. Obviously, no shame or chastening
accrues to the reader who, before being so informed, fails to foresee that a
particular character has brown eyes, or was born in Buenos Aires, or
checked out of the Savoy hotel at 10 a.m. on the Monday morning; it is
entirely reasonable not to have foreseen such narrative facts and events
and not having foreseen them reflects no fallibilities of understanding. By
contrast, genuine narrative surprises present us not merely with unfore-
seen information; they display to us a moment or a space where our under-
standing (of facts and events integrated with motives, psychology, and
latent forces) has failed.
In Jane Smiley’s ‘The Age of Grief’, suspense is brilliantly created. The
influenza at first seems incidental to events, the ‘grief of the last weeks’
(187), in which the parents have teetered on the edge of the absolute
break-up of their relationship. Members of the family succumb to the flu
(and begin to recover) in series: the father, then Lizzie, the oldest, most
independent child; then Leah, the youngest and incredibly clingy one, then
the mother (Dana), and finally Stephanie (p. 195), at which point the nar-
rator-father comments: ‘I was nearly jovial. I thought I knew what I was
doing.’ He says to Stephanie: ‘but I have a feeling it will go away fast for
you and Mommy.’ Soon, however, the narrator is telling us ‘in myself I felt
panic, a little void, needle-thin but opening’ (p. 196), a typical proleptic
pointer to a negative outcome; as are the immediately following remarks
on ‘the permanent threat of death’. Even more ominous and suspense-
creating is the long ensuing reflection on commitment, especially ‘the
commitment of risk, . . . of heart’ that comes with being ‘an involved
father’. All such reflections, like the essence of Labovian evaluation (see
Chapter 6), serve to convey that the events the teller went through
are momentous experiences, and that, whatever the outcome not yet dis-
closed to us, the entire process was of great importance to the teller/
participant. What is conveyed to us, by all such extended event-suspending
observations, is that the events and their conclusion were not routine or
inconsequential. Similarly, even a relatively inconsequential event can be
invested with significance because of its placement – as here where the
steady charting of Stephanie’s temperature, in the 104.6 range, is interrupted
by a long description of the father carrying his sick wife back to her room,
and reflecting on his wife’s build, strength, dress preferences, and other sus-
pense-generating ‘distractors’ (p. 204). Similarly there is a dramatizing or
heightening effect achieved by the long paragraph situating Stephanie in her
room on pp. 205–6, culminating with the father’s attempt to wake her up.
Her unresponsiveness, floppiness, is, via presupposition – ‘I was glad she
seemed to be getting sounder sleep’ – coded as relaxed sleep; but the reader
is probably already inclined to interpret her behaviour as comatose uncon-
sciousness, which is confirmed by the thermometer reading of 105.2.
Again, rather than immediately reporting the panicked reaction we
assume is somewhere submerged within the father’s reaction, we are given
thoughts about his ‘vision’ of the stars as seen once before, before the text
circles back to focus on the child-as-tiny-star-in-the-universe. Given this
extensive build-up, culminating with the father’s tearful breakdown on the
phone to the hospital, followed by his preparations for taking Stephanie
in, the resolution of this mini-episode is powerfully swift, casual, and
embedded (208.3):
and I knew that the next morning, when Stephanie’s fever would have
broken, I would be extremely divorced from and a little ashamed of
my reactions, and it was true that I was. They sent us home from the
hospital about noon. Dana was making toast at the kitchen table, Leah
was running around in her pajama top without a diaper, and Lizzie
had escaped to school.
4.8 From prose to film: radical translation

After the following chapter’s study of characters’ discourse in fiction, this
book turns from narratological issues to approach narrative from a socio-
linguistic orientation. But before leaving the narratological phase, mention
– however necessarily brief – should be made of that contemporary narra-
tive genre with possibly the greatest cultural significance: film narrative.
The following observations are presented as postulates for discussion and
critique, and are chiefly confined to one particular issue among the very
many that might be addressed: What is involved – linguistically especially
– in making the film of a novel (or the book of the movie)? Comparing the
written and film narratives of The Dead or Sense and Sensibility, where are
the differences or changes, and what is added, lost, or cut?
a Consumption/processing time: with respect to the novel, the reader

has considerable control of this; with respect to film (especially in the
cinema), the director in large degree has control (a 110-minute
‘default’) and ‘one pace fits all’.
b The narrative film is multiple-authored, a joint telling, with director,
producer, camera and sound crews, location and set designers,
scriptwriters, all kinds of editors, and actors all having a direct ‘inten-
tional’ hand in how and what gets told. At the same time the film typi-
cally lacks a single ‘brooding’ or perceptible narrator. Novels by
contrast are almost invariably single-authored.
c Bakhtin did not write much about film; he would surely have
embraced its heterogeneity and hybridity (not merely of content – as
in the novel – but of production and authorship). Popular culture in
the last 100 years has seen the development of powerful new forms of
hybrid/heterogeneous narrative – now including pop songs with videos
attached, and hypertext fiction.
d Everything in the novel is achieved through written language (includ-
ing its representations of speech and thought, its showing or reporting
of emotion, etc.). But in film there is a blend of several modalities:
visual representation (depictions of settings, of characters, of actions
. . .); non-verbal aural representation (music, sound effects, indices of
setting); non-verbal human noises (of fingers typing on a keyboard, of
someone brushing their teeth, etc.); speech; and even writing (as a dis-
tinct subtype of non-iconic visual representation). The choices con-
cerning ‘how to tell it’ in the novel are between contrasting written
formulations. But the choices in telling, in film, are far more varied.
They are less constrained particularly concerning how, via which
modality, to tell the individual events of the story. For example, a
single event, such as ‘Jenny receiving bad news in a letter’ can be told
in (at least) any one, or any combination, of the following four modali-
ties:
1 and/or 2 and/or 3 and/or 4
visual aural speech writing
representation representation
And then within each of these broad modalities the director has a rich
variety of options. If they choose to tell of ‘Jenny receiving bad news
in a letter’ via visual representation, they may further choose to have
Jenny’s face and/or the letter in shot or not, close-up or at a distance,
darkly or brightly lit, centred in the frame or not, in one continuous
shot (with zooming or other movement or not) or several shots edited
to highlight discontinuity. Because the film is so much less formally
constrained than the novel, comparing the formal restriction or frame-
work of the novel with the open-ness of the film is analogous to com-
paring the delimitation of the sonnet form with the technical freedoms
of the novel.
e With respect to showing versus telling, in film narrative showing is
overwhelmingly the preferred option, and the chief dilemma is to
decide what to show (and in what detail, at what length, from what or
whose point of view).
f Language – and arguably written language in particular – is the best
modality for analysis (particularly of characters, their desires, aspira-
tions, fears, etc.). Such analytical presentation of character (‘inward’
portrayal) is largely denied to film, where we instead get outward or
external presentation, from which we must infer the significant details
of character. We have to judge characters by appearances, witnessed
reactions, displayed emotions, and interpretive visual representations
of mental processes – all on offer to us in abundance and (or but)
seemingly without teller-interpretation.
g It is useful to identify written narratives that would seem to be imposs-
ible to render in film, and films that would seem to be impossible to
conceive of as novels. And to explain why, in both cases. Presumably in
the former impossible ‘conversion’, crucial telling inheres in the texture
of the narration (which abstracts away from character, dialogue,
setting, and events), as in a Beckett novel or a John Barth story. In the
latter impossible ‘conversion’, crucial telling must inhere in those visual
and non-verbal aural modes that resist representation in words (e.g.
narrated experience of an overwhelming battle or disaster).
h The verbal record of setting and characters at the opening of a novel
must be replaced (or cut) in any film treatment. If replaced, replaced
by other, holistic, non-linear representations (e.g. a visual one).
i At the heart of film narration lies editing, the selection and combining
of one shot with the next, invariably paying attention to the kinds of
graphic, rhythmic, spatial and temporal relations that are thereby
created between the sequenced shots. There are no full equivalents of
shots and editing in written narrative, but there are arguably counter-
parts. Thus the sentence is a kind of shot, and the full stop is the point
of splicing or cutting. The author can contrive to make the reader,
proceeding from one sentence to the next, experience rhythmic,
spatial and temporal continuities or disjunctions just as the film direc-
tor can cause the viewer to, via editing choices. (But in the novel there
are arguably ‘higher level’ shots and cuts too: the coordination of one
paragraph with the next, and, optionally, the coordination of one
chapter or section with the next.)
j A major topic in novel versus film comparative study must concern the
sequentiality of the telling in one vis à vis the other. Written text draws
on all kinds of conjunctions and temporal and locative phrases: then,
right after that, six weeks later, but out at the back the house, and so on.
A judicious combination of these, together with reliance upon the
reader’s powers of inferencing in seeing the connections between
reported events or scenes, will be instrumental in creating narrative
coherence. In film, a great deal of reliance, typically, must be placed
upon visual processing to achieve similar ends. We are not literally
told, by verbal means, that ‘what happened next was this and then
right after that this happened’; typically, we are simply shown B hap-
pening after A, and C after B. We need to perform a ‘mental process-
ing’ in order to make sense of what we have seen, in order to store it
at least in short-term memory for the duration of the film, and in order
to be able to talk about the film afterwards, and form general interpre-
tive opinions. Just how we do these things is rather less clear, and less
stable, than in the case of the novel, where we typically have the text
in front of us to refer back to. The ‘reading’ of a written text is very
different from the ‘reading’ of a film, to a considerable degree
because, being in words and sentences, the written text is already
‘propositional’. A certain kind of interpretive work has thus already
been done for the novel-reader, by comparison with the film-viewer
(but in a sense ‘too much’ work has been done, creating a proposi-
tional ‘overload’). The comparative propositional scarcity of film is
what Chatman means by its verbal impoverishment: ‘Film gives us
plenitude without specificity. Its descriptive offerings are at once rich
and verbally impoverished’ (Chatman, 1990: 39).
k This last idea relates particularly to the intepreting of character. Take
the proposition ‘Emma Woodhouse is, for much of the story, selfish’.
In the novel, all sorts of verbal clues guide the reader to an observa-
tion of this kind (along with many other observations). How does a
viewer of the film Emma get to a similar point in interpretation? This
is a substantially different task because the attentive viewer will see a
range of depictions (not read descriptions) of Emma doing, saying,
perhaps thinking (e.g. via voice-over). And from those depictions the
viewer has to derive interpretive conclusions. But outside the film’s
representations of characters’ speech and thought, the film-viewer’s
evidence is to a considerable degree non-verbal. From that fusion of
characters’ speech and non-verbal depictions, the viewer must deduce
interpretive conclusions.
l To what extent must film and written renderings of e.g. Trainspotting,
or Like Water for Chocolate have different goals and in fact not be the
same narrative? Is the rendering of a book into a film or vice versa
comparable to literary translation? The French language has two
verbs for ‘changing’: changer (where mere alteration or adjustment is
involved) and échanger (where substitutionary displacement, is
involved, as when you change shoes or jobs). Perhaps the change from
a film to a written narrative, like interlingual literary translation, is an
échanger.
m Raymond Carver’s bleak short story, ‘Little Things’, about a couple
fighting over a baby begins with a naturalistic apartment setting, and it
getting dark outside. Then the text adds: ‘but it was getting dark on
the inside too’. How could you film this? (How would you film the
too?) The story ends without clear indication of who keeps the baby
(or whether it dies in the struggle). But the text reads: ‘in this manner,
the issue was decided’. Again, how would you film that?
n The contrast between novel and film is not simply words versus
images, but the processing of serially displaced words as opposed to
that of serially displaced images.
o If we conclude that between film and written narrative it’s a matter of
radical translation, of échanger and not merely changer, it may be
because finally the essence of a verbal narrative like Joyce’s ‘The
Dead’ or Carver’s ‘Little Things’ is in the (dispersed but accumulating
and mutually displacing) written words more than anywhere else. And
by the same token, the essence of a filmic narrative like Citizen Kane
is in the (dispersed but accumulating and mutually displacing) pro-
jected images – again, more than in anything else (spoken words,
music, sound).
p Because a prose narrative and subsequent film narrative (or vice versa)
are such incommensurably different discourses, brought together by
some commonality of underlying story, one relates to the other on the
basis of translation or transformation or échanger. Certainly ‘adapta-
tion’ seems inadequate to describe the process (and is only used to
describe prose-to-film conversions, admittedly much the commoner
progression). These prose-to-film radical translations are also perhaps
exceptional in their very frequency of occurrence: how rare, by compar-
ison, are musical ‘versions’ of paintings or poems, or sculptures of
ballets, or short-story versions of Beethoven string quartets.
4.9 The grammaticization of character and situation

In the final section of Chapter 2, I suggested a number of ways in which
the special status of certain clauses in a narrative, as main carriers of the
plot, may be typically reflected in the grammar of those clauses; and at the
close of Chapter 3 came a discussion of the role of modality in narrational
modes. Similar claims can be made about the grammar used in portraying
characters and settings: the grammar, examined closely, is found to match
the nature or conditions of characters. In making such a claim it is neces-
sary to stress that I am using the term ‘grammar’ in a relatively enlarged
sense, to denote not so much a formal syntactic description but a meaning-
oriented functional description of the language in question. By grammar,
then, I mean a systematic account of the principles governing choices of
words and sequences of words within a language, with the additional
requirement that the account is attentive to the meanings that speakers
associate with those choices. One of the most useful and usable
contemporary meaning-oriented grammars of English is that of Halliday
(1994). The newcomer to linguistics will find this to be a sometimes diffi-
cult read, but they are well-served by a number of invaluable introductions
to Halliday’s systemic grammar: Bloor and Bloor (1995); Butt, Fahey et al.
(1995); Eggins (1994); and Thompson (1996). Below I sketch that part of
Hallidayan grammar, namely choices in the transitivity of clauses, which
seems most relevant to the analysis of character and setting.
In Halliday’s systemic grammar, transitivity concerns rather more than
purely syntactic questions such as whether or not a particular verb takes a
direct object (cf. smile versus embrace). The theory assumes that the
semantic processes and participants expressed by particular noun phrases
and verb phrases in a clause are a representation of what we take to be
going on in the world. By means of choices from among a delimited set of
process types and participant roles, these being expressed in the grammar
of the clause and, in particular, its verb, we characterize our view of
reality. When we say John hates chocolate, we represent a particular situ-
ation as one in which an individual, John, interacts (negatively) with
chocolate by means of a mental process; and this is a standard English-
language way of talking about people not liking chocolate (cf. also the col-
loquial Chocolate doesn’t agree with me). In a different linguo-cultural
world, one might by contrast standardly say things equivalent to: Choco-
late fights John’s stomach, or John is square but chocolate is circular. But
we find these to be strange grammaticized representations by comparison
with our own, taken-for-granted, patterns.
A process consists, potentially, of three components:
1 the process itself (typically realized by a verb phrase); there are just
five core types of process;
2 the participants in the process (typically realized by noun phrases and,
in the case of attributes, adjectival phrases); this grammar recognizes
about twenty such participant types, four or five distinct ones associ-
ated with each process type;
3 circumstances associated with the process (realized by prepositional
phrases, adverbial phrases, and adverbial subordinate clauses); these
can be classified along fairly traditional adverbial distinctions (place,
time, manner, reason, and so on).
Transitivity (or process) analysis is a fundamental semantic parsing. The

analyst identifies the process or action that a clause expresses, and the
participants that are recognized. So in the sentence Bear drew Cat a big
plate of food with his magic pencil, the following parsing (and labelling)
would be applied:
process
material
Bear drew Cat a big plate of food with his magic pencil
agent beneficiary effected medium instrument
participant participant participant participant

In the given example, then, a particular occurrence (a mini-narrative, in
fact) is represented as one involving a material process of drawing and
four participants: an effected medium that is brought into existence, an
agent responsible for the process, and two further participants, the bene-
ficiary and instrument, which did not have to be mentioned grammati-
cally (cf., Bear drew a big plate of food) but were. Those, according to
this transitivity analysis, are the essential semantic components of the
given sentence, and by proceeding in the same way to analyse all the sen-
tences in a text, the essential semantic components of an entire text can
be identified. I will introduce and explain these labels (agent, material
process, etc.) very shortly. But straight away I should state that the
transitivity-parsing system is interesting in narrative analysis since it
systematically highlights, for any text, just what kinds of process tend to
go on, which individuals are (and are not) frequently agents, who or
what tends to serve as medium, and so on. The entire business of repre-
senting the processes and participants of reality is what Halliday has
termed the ideational function of language. In relation to this ideational
function, the clause is the basic vehicle for representing patterns of
experience. Below I offer a few more explanatory notes on the major
types of process and participant in the English clause (these notes are
heavily indebted to the lucid lecture hand-outs of my former colleague,
Anneliese Kramer-Dahl).
4.9.1 Material processes: processes of perceptible action

Material processes are those processes that might answer the questions,
‘What happened?’ or ‘What did (someone) do?’ By way of accompany-
ing participants, they always involve what is called a medium (the
affected entity, called the theme in some grammars) and this medium
usually pre-exists the process, but is further specified as the effected
medium if it only comes into existence by dint of the process (as when
someone bakes bread, paints a portrait, or builds a patio). Further
participants that may be involved are agent (a human or quasi-human
entity, acting intentionally), force (any inanimate agent), instrument
(always under implicit control of an agent), beneficiary and recipient.
Given limitations of space, I will do no more here than give a few
explained examples:
John lit the bonfire with paraffin

(where John is agent, the bonfire is medium, and paraffin is instru-
ment),
The lightning hit the tree
(the tree is medium; the lightning is not an agent, nor a controlled
instrument, but simply a force),
It is raining
(where there is really only a process, and no participant),
Emma painted a large watercolour
(where the watercolour is effected medium, Emma is agent),
Oliver collapsed
(where Oliver’s experience is hardly intentional: he is medium, not
agent, like the orchids below),
The orchids grew quickly
John wrecked his car
(where, assuming he acted unintentionally, John is both perpetrator
and victim, hence may be best labelled agent/medium).
And a final example where one embedded process serves to specify the
medium of another:
My grandmother bequeathed to me the jewellery that Grandpa once bought for her.
agent material recipient medium .................................................................

pro
[medium agent circ/ material beneficiary]

time process
4.9.2 Mental processes: processes of sensing

Mental processes entail verbs of feeling or mental reaction (like, fear),
perception (see, hear), and cognition (think, believe). These involve a
human or human-like participant, the senser, and a thing or fact that is
perceived, felt or thought, known as the phenomenon. Note that some
mental process verbs take an active, engaged agent rather than a mere
senser: the processor who listens, watches, learns, etc., is more agentive
than the one who merely hears, sees, or understands. The former
processes are more developmental, and can take a progressive aspect.
Grammatically, mental processes are distinguished by being able to
introduce, as the process’s phenomenon, an entire clause (which will
have its own transitivity analysis into process and participants). In this
they differ from all the other processes, particularly the behavioural ones
with which they seem to overlap. Mental processes are accordingly said
to ‘project’:
mental
senser process projected phenomenon
The boy knew he smelled cheese

4.9.3 Verbal processes
Verbal processes are processes of saying or telling in a broad sense, with
participants known as the sayer, the said, and the addressee:
He told me he was sorry.
4.9.4 Behavioural processes

Behavioural processes are processes of relatively spontaneous or uncon-
trolled human activity, internal or external. They lie at the physicalist or
materialist end of two other core processes (verbal or mental), so can be
thought of as instances of these represented in rather a materialist light:
looking, watching, staring, listening, breathing, coughing, grimacing,
frowning, smiling, dreaming, screaming, choking, blinking, and crying out.
They are semantically and grammatically halfway between material and
mental processes, with, typically, just one participant, a human behaver.
Like material processes they freely progressivize. Like typical verbal and
mental processes they involve a human do-er, but unlike both they do not
‘project’ (or they are such processes used in a non-projecting way: not I
worry that we’ll have no money left, but He worried about money).
4.9.5 Relational processes: processes of being

Relational processes are ones of characterization or identification, and
entail intensive verbs, or verbs expressing circumstantial or possessive
relations, the ‘default’ verb being be:
1 intensive: ‘x is y’
2 circumstantial: ‘x is at/about/like y’
3 possessive: ‘x has y’
George was victorious.

George was the winner.
It took place at 3 o’clock.
He has a great job.
To repeat, what can be gained from careful transitivity analysis of scenes

and characters is a preliminary picture of who is agentive and who is acted-
upon medium (for whatever reason: they may be passive, or powerless, or
just lazy), whether characters are doers or thinkers, whether instruments
and forces dominate in the world represented, and so on. Invaluable appli-
cations of the transitivity analysis apparatus in the description of characters
– their dispositions, ability to control things and infer causal connections, or
their powerlessness – include Halliday (1971), Burton (1982), and Kennedy
(1982). Hallidayan transitivity will be applied extensively in the course of
analyses in Chapter 8.
Further reading
Good places to begin further reading on character and setting are Rimmon-Kenan
(1983), and Bal (1985: 25–36, 79–99). Fowler (1977: 33–41), cited in 4.4, demon-
strates the application of a distinctive feature approach to character and setting in
The Great Gatsby. Chatman (1978: 96–145) is invaluable on, inter alia, character
‘traits’, A. C. Bradley, and the role of setting in ‘Eveline’. Interesting literary crit-
ical discussions of character include Price (1968), Harvey (1965), Bayley (1963),
and, with a more theoretical orientation, Culler (1975a), Frow (1986), Iser (1978),
Docherty (1983), Margolin (1989; 1996); Emmott (1997); and Culpeper (2000 and
forthcoming).
Lodge (1977: 73–124) contains an extended thesis to the effect that there are
two most fundamental and contrasting means of discourse-development: the
metaphoric (where inherently disparate things are somehow linked and some
illuminating context-specific similarity is posited), and the metonymic (where
inherent associations between things are drawn upon, and consequently a richer,
more textured picture of a single area of reality is achieved). A metaphoric
dynamic involves an adjoining of a second topic to a first on the basis of its local
and specific relevance (despite its logical and intrinsic irrelevance) – e.g. associat-
ing fog in the city of London with the Chancery law courts. The metonymic
dynamic advances the text by passing from one topic to another, logically or
traditionally associated topic. Lodge’s theoretical account, inspired by work by
Jakobson, leads to original and insightful analyses of just how the depiction of
setting proceeds in many literary texts, including the openings of Bleak House and
A Passage to India.
On thrillers, surprise and suspense, see, e.g. Rubin (1999), Palmer (1978),
Sauerberg (1984), and Luelsdorff (1995). On Hallidayan transitivity applied to
literature, Simpson (1993), Chapter 4, is very clear. Toolan (1998), Chapter 4,
could also be consulted, and also Short (1996). See also Halliday’s own influential
paper on the changing styles in Golding’s The Inheritors, Halliday (1971), together
with a recent reexamination of it, Hoover (1998).
Notes and exercises

1 Bedient (1969: 84) has declared this, of the characters in Middlemarch:
When we read the problems of Lydgate’s marriage, or about Casaubon’s
‘inward trouble’ or about Bulstrode’s public fall, it doesn’t occur to us that
these are imagined realities.
Do you agree?
2 One cluster of claims in Chatman (1978: 120) raises issues concerning real-world
representation and the interpretibility of character and setting in fiction by refer-
ence to real-world counterparts:
A trip to Dublin cannot but help us understand the special quality of paraly-
sis attributed to its denizens by Joyce, and meeting a Dublin working-class
girl, even in 1978, will give us deeper insight into Eveline’s predicament and
personality . . . I have never been to Ireland, but I know that the peculiar
sort of ‘strutting’ that Eveline’s father does would be clearer if I had.
On what grounds might one agree or disagree? The quotation appears to assume
that understanding Eveline, who represents a ‘Dublin working-class girl’ c. 1910,
can be facilitated by encountering an actual Dublin working-class girl c. 1980 (or,
now, c. 2000). Arguably all three descriptors (which might also be called traits),
that is, the meaning of ‘Dublin’ and ‘working-class’ and ‘girl’, have changed pro-
foundly in the years since ‘Eveline’ was written. But equally arguably, most
stable of these three, despite its material transformation, is ‘Dublin’, the sign
that denotes a setting. There was a Dublin in 1910, and there still is (while in fact
the story’s protagonist, Eveline, never was). Does all this reflect the fact that
setting, rather than character or event or discourse, is the narrative element that
most forcefully asserts a ‘referential connection’ with conditions in the real
world?
It is tempting to say that Dublin and working-classness and being a girl (young
woman) are so fundamentally changed today, by contrast with 1910, that inspec-
tion of contemporary exemplars will be of no help in interpreting the working
girl in ‘Eveline’. But in practice readers must get their background pictures,
which help them to interpret Eveline’s predicament and personality, from some-
where (these background pictures are the mental schemas and world knowledge
discussed in 4.3). And those background pictures, although they may be of spa-
tially or temporally remote categories (such as working-classness c. 1910) and
may be informed by a variety of historically and sociologically reliable sources,
will nevertheless carry an inescapable contemporaneity: they are present-day pic-
tures of what that projected world, back in 1910, was like. Are they, then,
fundamentally different from what is entailed in visiting present-day Dublin to
derive a picture of Dublin, 1910?
3 A common entry point for discussion of characters is the paying of particular
attention to their names. Dickens is commonly cited, with his roll-call of evoca-
tively- and characterizingly-named individuals: Podsnap, Pecksniff, Ebeneezer
Scrooge, Fagin, Gradgrind, Bounderby, Jupe, Wackford Squeers, and so on. But
one can take the names of characters in very many modern fictions and analyze
them for their expressivity, for their semiotic quality (see Doherty, 1985, for dis-
cussion). There seems almost a compulsion, in both authors and readers, to make
character names be message-bearing. And from the authorial point of view, since
a character’s name so often serves, when used, to trigger recall of the multi-
faceted notional substantiality of that character, it must be hard and unnatural for
an author to resist ordaining a name which is in some senses encapsulating. How
intentional that encapsulating or trait-evoking effect is will often remain uncer-
tain. At issue is a kind of enlarged onomatopoeia, encompassing all kinds of
semantic and evaluative association: did Faulkner’s prior selection of the name
Flem Snopes ‘guide’ the author to the qualities of curmudgeonly and calculated
meanness that the character has because these were already ‘in’ the name, or has
the character’s behaviour caused a kind of pejoration of a perfectly neutral and
arbitrarily-signifying word, in effect ‘giving the word Snopes a bad name’? And
what, to you, does the name Pnin signify? Conversely, try to find novels where
the names are ‘transparent’, semiotically null: are such novels from particular
authors only, or in one genre only? The naming/meaning conundrum is not con-
fined to fiction: most of us who become parents will find ourselves confronting it
when trying to decide (agree . . .) on a name for a new baby.
4 Perhaps the commonest and most important element in fictional settings is the
house. Everywhere in English fiction, from Mansfield Park to Bleak House, The
House of the Seven Gables, The Little House on the Prairie, Howards End,
Brideshead Revisited, The Mansion, The Hotel New Hampshire, and endlessly on,
there are novels whose titles themselves are names of houses that are an essential,
mood-determining, story-shaping, arena for characters and events. Take a novel
by any author of your choice (even at random), and examine the extent to which
allusions to the house(s) therein go beyond the basic required notation of place.
Can you find particular local narratological motivations – or a more general
generic one – for such ‘excesses’ of attention? Why are some fictions – e.g. Gothic
novels – apparently saturated with annotations of setting, while others – e.g.
Beckett’s later prose – have almost none?
‘The house’ is of course only one of many recurrent, emblematic or associ-
ation-rich ‘basic settings’, invoked and varied in countless fictions, and used as a
kind of hub for the entire narrative (often being mentioned in the work’s title).
Some others that immediately come to mind include the beach, the sea, the
college, and the hospital. But yet other commonly-experienced settings, I would
suggest, are rarely invoked in this way, as the hub or foundation of, say, an
entire novel. Thus we do not find novels ‘rooted’ in the park, in quite the way we
find them based on and around the beach. Mansfield Park is not an exception.
Without intending any disrespect to Mansfield (any town might be mentioned
here), it is fair to say that Mansfield Park is effective as Austen’s title so long as
we know – being able to decode the exclusivist metonymy – that Mansfield
denotes a big house and country estate and that the novel is not, heaven forbid,
a story set in a park in Mansfield. It may be interesting to attempt to explain why
house, beach and college earn the higher status while other settings do not. Why
is it that the park, the recreation ground, the airport, the shopping mall, to name
just a few, lack this ‘foundational setting’ status that seems accorded to houses
and beaches?
5 Compare and contrast the two basic settings in ‘That Evening Sun’ – Nancy’s
cabin and the Compsons’ house (actually there are clearly sub-domains within
the latter: the kitchen is a somewhat different setting from the library). Setting
plays a prominent role in this story. It could be argued that ‘change of setting’
appears to amount to a solution to Nancy’s problems: if she can stay with the
Compsons and sleep on their kitchen floor she will be all right, she says. But
this is a solution that the Compson parents are evidently quite reluctant to
accept.
6 Chatman notes the very effective use in ‘Eveline’ of climactic metaphors pre-
cisely at the moment of the heroine’s greatest anguish (Chatman, 1978: 142): the
bell clanging upon her heart, all the seas of the world tumbling about her heart,
Frank drowning her in those seas, and so on. These metaphors are especially apt
since, at the quayside, clanging bell and tumbling seas are part of the actual
setting. Eveline is subjected to ‘the onslaught of setting’. Could it be argued that
at the close of ‘The Dead’, similarly, Gabriel’s consciousness is subjected to the
onslaught of setting? We might look further in Dubliners to see whether this
gradual promotion of setting into a character, a shaping influence, applies to
other stories in the collection. Note that the title Dubliners itself is an act of defin-
ing characters by their setting.
As one means of probing the degree to which setting may seem to become
agent-like, the Hallidayan categorization of transitivity processes (4.9) could be
put to use. By contrast with earlier in the story, are there more occasions at the
close of ‘Eveline’ where some element of the setting is agentive or quasi-
agentive in the clause – typically, in subject position (in active sentences) with a
transitive verb and an affected object? My initial impressions are that the early
domestic setting – tame, unthreatening – comprises objects that are also syntac-
tic objects, things affected by a human agent (Eveline ‘had dusted’ the familiar
objects once a week for so many years). But at the quayside, setting includes ele-
ments that are quite agent-like, at least to the extent of taking up subject posi-
tion in either transitive or intransitive sentences, with no mention of a human
causer: ‘A bell clanged upon her heart’, ‘All the seas of the world tumbled about
her heart’. Does a more detailed analysis support those initial impressions of a
growing dynamism of setting?
7 Each of J. K. Rowling’s Harry Potter books is peppered with suspense and sur-
prises. Examine, for instance, the final pages of Chapter 16 and the opening of
Chapter 17 of Harry Potter and the Philosopher’s Stone (Bloomsbury): comment
in detail on how the suspense of Harry’s quest of the stone in its guarded hiding-
place is supplanted by the surprise, at the opening of Chapter 17, of finding that
Quirrell is not a p-poor s-stutterer.
8 Put any two or three of the alphabetized points in 4.8 to the test, using in evi-
dence any well-known fiction-to-film pairing.
text III
Representing character discourse
5.1 Achieving immediacy in the narration of thought

Although Bertha Young was thirty she still had moments like this
when she wanted to run instead of walk, to take dancing steps on and
off the pavement, to bowl a hoop, to throw something up in the air
and catch it again, or to stand still and laugh at – nothing – at nothing,
simply.
What can you do if you are thirty and, turning the corner of your
own street, you are overcome, suddenly, by a feeling of bliss –
absolute bliss! – as though you’d suddenly swallowed a bright piece of
that late afternoon sun and it burned in your bosom, sending out
a little shower of sparks into every particle, into every finger and
toe? . . .
Oh, is there no way you can express it without being ‘drunk and dis-
orderly’? How idiotic civilisation is! Why be given a body if you have
to keep it shut up in a case like a rare, rare fiddle?
‘No, that about the fiddle is not quite what I mean,’ she thought,
running up the steps and feeling in her bag for the key – she’d forgot-
ten it, as usual – and rattling the letter box. ‘It’s not what I mean,
because – Thank you, Mary’ – she went into the hall. ‘Is nurse back?’
(Katharine Mansfield, ‘Bliss’)
As we advance through these remarkable opening paragraphs of ‘Bliss’,

we can hardly fail to notice that a masterfully controlled storytelling is at
work. It is a storytelling in which external facts about a character on the
one hand, and a version of her intimate thoughts on the other, share the
same sentential envelope without strain. The routine information that
there is a woman called Bertha Young who is aged thirty gives way to dis-
closures of her sudden impulses – like this – to run, dance, laugh at
nothing, and so on. Such disclosures appear without either noisy signalling
or discoursal effort, and this is partly due to the avoidance of cumbersome
clauses of self-perception (She realized that, she felt that . . . etc.) or any
sudden appearance of confessional first-person pronouns (‘I want to . . .’, ‘I
feel . . .’). That much of this – from around line 3 onwards – is a subtle ren-
The articulation of narrative text III 117
dering of Bertha’s thoughts and emotions is supported by the belated
appearance of a ‘she thought’ in paragraph 4; but we do not need that ‘she
thought’ to know this.
The first intimation of a veiled immediacy of telling comes with the
unexpected phrase, ‘like this’ in the opening line. The expression swiftly
aligns us with the character Bertha’s own immediate orientation to her
world, ‘like this’ being not so much cohesive, tied to some co-textually
specified moment, but rather expressive and deictic, alluding to the current
moment, one we would have no difficulty identifying were we Bertha
herself. From that phrase onwards, for the entirety of the story, we can
argue that Bertha is the internal character-focalizer (and often the focal-
ized too).
In that first paragraph, too, the reader is cleverly led to assume that
Bertha is, at this point in the narrative, walking along a street, on a pave-
ment, in the open air – though none of this is directly reported. All that is
directly reported are her blissful impulses – and these might have arisen
even if in actuality she were lying in bed, or eating lunch, or kneeling in
church. Yet no such situation comes to mind as we read the paragraph: we
assume that reported impulses to frolic in the street indicate that the
woman is actually in the street, although perhaps only the exophoric defi-
nite article used in the reference to ‘the pavement’ is an explicit formal
cue. With an art that conceals art, the narrative has modulated from the
distanced description of the character, ‘Bertha Young was thirty’; to the
immediacy of the relaying of Bertha’s thoughts in, especially, paragraph 3;
and back to the externality of the direct speech in paragraph 4. Those
modulations have also involved switches in tense, from past to present to
past – but we are too distracted on a first reading by the larger textual
effect, its representation of bliss, to notice such mechanics. Abstracted
from its carefully crafted context, a sentence like this one in paragraph 4
might strike us as absurd, a clumsy mistake:
‘It’s not what I mean, because – Thank you, Mary’
But in context it is brilliantly effective.

Now consider another passage:
Isabel noted afresh that life was certainly hard for some people, and
she felt a delicate glow of shame as she thought how easy it now
promised to become for herself.(1) She was prepared to learn that
Ralph was not pleased with her engagement; but she was not prepared,
in spite of her affection for him, to let this fact spoil the situation.(2)
She was not even prepared, or so she thought, to resent his want of
sympathy; for it would be his privilege – it would be indeed his natural
line – to find fault with any step she might take towards marriage.(3)
One’s cousin always pretended to hate one’s husband; that was tradi-
tional, classical; it was a part of one’s cousin’s always pretending to
118 The articulation of narrative text III
adore one.(4) . . . You could criticize any marriage; it was the essence of
a marriage to be open to criticism.(5) How well she herself, should she
only give her mind to it, might criticize this union of her own!(6) She
had other employment, however, and Ralph was welcome to relieve
her of the care.(7) Isabel was prepared to be most patient and most
indulgent.(8) He must have seen that, and this made it the more odd he
should say nothing.(9) After three days had elapsed without his speak-
ing our young woman wearied of waiting; dislike it as he would, he
might at least go through the form.(10) We, who know more about
poor Ralph than his cousin, may easily believe that during the hours
that followed his arrival at Palazzo Crescentini he had privately gone
through many forms.(11) His mother had literally greeted him with the
great news, which had been even more sensibly chilling than Mrs
Touchett’s maternal kiss.(12) Ralph was shocked and humiliated; his
calculations had been false and the person in the world in whom he was
most interested was lost.(13) He drifted about the house like a rudder-
less vessel in a rocky stream, or sat in the garden of the palace on a
great cane chair, his long legs extended, his head thrown back and his
hat pulled over his eyes.(14) He felt cold about the heart; he had never
liked anything less.(15) What could he do, what could he say?(16) If
the girl were irreclaimable could he pretend to like it?(17) To attempt
to reclaim her was permissible only if the attempt should succeed.(18)
(Henry James, Portrait of a Lady: 338–9)
In this passage, as in that from ‘Bliss’, we see the subtle narrative presenta-
tion of characters’ thoughts, again with an attenuation of the sense that
any narratorial figure is relaying those thoughts. No matter how audacious
Isabel’s views become, no narratorial voice of ‘true judgment’ intrudes to
censure the character, dissociate itself from Isabel, and alert the reader to
the errors being broadcast. That is not James’s way. It is enough that we
see Isabel’s shallow petulance immediately juxtaposed to Ralph’s actual
views, again, relayed without noticeable narratorial editorializing.
And yet that this passage is still narrated seems clear. In particular,
Isabel continues to be designated by the third-person pronoun, she, as
someone different from the teller (who is always an I), and the tense con-
tinues to be the past tense of narration, rather than the present tense of
direct experience. Furthermore, the passage begins by telling us what
‘Isabel noted’ and by reporting as the text of an intrusive omniscient nar-
rator (Simpson’s type B(R) positive) what she thought and what she felt as
she thought. But by sentence 4 that narrator has almost completely with-
drawn (and remains so until sentence 10), their omniscience giving place
to Isabel’s own limited wisdom – a wisdom that confidently and blindly
assumes that Ralph ‘classically, traditionally’ always pretended to adore
her. Pretended, the reader who has reached page 338 asks? The mismatch
between Isabel’s distorted picture of things and the reader’s own fuller
one is too striking to ignore. The narrative mode adopted so as to narrate
Isabel’s thoughts in an uncensored form engages the reader in the comput-
ing or construing of many such ironies. The irony just noted may be chiefly
poignant. But when, in sentences 5 and 6, this newly-engaged young
woman announces to herself that the essence of a marriage is its openness
to criticism, and that she could – if she thought about it – criticize her own
impending union, but that she had other things to do (impliedly more
important), then a serious and extensive waywardness of judgment is laid
bare.
The narrative mode at work here, in sentences 3 to 7, 9 and the latter
half of 10, is Free Indirect Discourse, which will be explained and discussed
at length in the following sections. The phenomenon is often referred to by
other names; perhaps commonest, from a very long list, are: free indirect
style, style indirect libre, represented speech and thought, quasi-direct dis-
course, erlebte Rede, and combined discourse. Among the many narrative
tasks Free Indirect Discourse can help perform are ironies and mismatches
of the kind found in the commentary on Isabel above; equally, it can be
used in the service of empathetic disclosure as in the depiction of Ralph or,
in the earlier passage, Bertha. Wherever it is employed, it seems to involve
subtlety and complexity and demands a commensurate care from the
reader. Note, for example, how despite Isabel’s insistence on her own
patience towards Ralph, her indulgence, her refusal ‘to resent his want of
sympathy’, she does resent his unspoken reservations and wearies of
waiting. Or, somewhat differently, consider the narrator’s clever collusion
with the reader – ‘We, who know more about poor Ralph than his cousin’ –
when he picks up Isabel’s complaint that ‘he might at least go through the
form’ (i.e. wishing her joy and happiness, etc.), and uses the same phrase in
the next sentence in a tellingly different sense.
5.2 Modes of speech and thought presentation

We should begin with fundamental principles. When a narrator sets about
telling some individual’s story, there will be various actions and events
involving the character, which they will wish to narrate. Where these
actions are physically overt and observable by any careful witness – She
sat at the window watching the evening invade the avenue – I call these
Pure Narrative sentences. Very often, exercising their option to be as
revealing of a character’s inward feelings as they wish (the narratorial
option to disclose ‘up to omniscience’), information that goes beyond
external witnessing will be related: in her nostrils was the odour of dusty
cretonne. As long as those inward details remain matters of which the
character is not consciously aware, I would still call them part of the Pure
Narrative (PN). Similarly, reports of mental or verbal activity which do
not purport to be a character’s articulated speech or thought – His mind
went blank; he couldn’t get his thoughts straight – are here grouped within
PN, however intrusive into a character’s psyche they may be. These tend
to be metadiscoursal commentaries on how the character was thinking or
speaking, rather than reports on what they thought or said (but see Short
et al.’s proposal, as an alternative approach, discussed in Notes and exer-
cise 3, pp. 141–2). In simplest contrast to Pure Narrative sentences are
those that relay what are vouched to be a character’s actual speech –
direct speech (DS), here set in italics:
Lily called out: – Miss Kate, here’s Mrs Conroy.
And by extension, where a narrator reports what is offered as a reliable

version of the sense, but not the precise words, tense, and pronouns of a
character, we encounter (here in italics) Indirect Speech (IS) :
Both of them kissed Gabriel’s wife, said she must be perished alive . . .
To a greater and lesser degree, respectively, direct and indirect speech ‘give
voice’ to the character whose utterances are reported, making for a more
vivid and dramatized story. But very often characters may not outwardly say
things in response to their circumstances, they simply think them, and by a
further extension beyond Direct and Indirect Speech, narrators have come
to make extensive use of two standard options for reporting thoughts. The
first of these purports to convey the exact words a character has formulated
in thought, their precise mental utterance as it were (Direct Thought); the
second, parallel to Indirect Speech, offers the sense but not the precise
grammar or wording of a character’s thought (Indirect Thought). The two
direct forms show a character’s words – spoken or thought – without
internal amendment; and the narrator doing the showing in effect moves
away from the quoted utterance, which expresses purely the character’s
viewpoint. The two indirect forms tell a character’s words, suitably cast to
make sense from the point of view of the telling narrator, who remains invis-
ibly present and responsible for the wording.
At this point we should mention what I call the ‘framing clause’, also
known as the clause of communication or the inquit clause (inquit being
the Latin for ‘he said’ or ‘she said’). Normally when speech or thought is
included in narration, the particular character who is the source of that
speech or thought is also identified, in a framing clause, as in the following
examples, which are examples of DS, IS, DT, and IT, in italics, preceded
by framing clauses, from ‘The Dead’:
DS After a pause she [Aunt Julia] asked: – And what are galoshes,
Gabriel? (189)
IS As they [the young ladies] said they never took anything strong he
opened three bottles of lemonade for them. Then he asked one of the
young men to move aside . . . and filled out for himself a goodly
measure of whiskey. (191)
DT He thought: ‘I have never felt like that myself towards any woman.’
(based on 235)
IT He realized that the time had come for him to set out on his journey
westward. (based on 236)
By convention the standard version of all four modes mentioned so far,

DS, IS, DT and IT, is one in which a framing clause of the kind exempli-
fied occurs in the text, unambiguously anchoring or attributing it to a
particular character. But all four modes are equally possible without a
framing clause, hence without unambiguous anchorage:
‘And what are galoshes, Gabriel?’

They never took anything strong.
I have never felt like that myself towards any woman.
The time had come for him to set out on his journey westward.
When the four modes occur without a narrator’s framing clause, they are
said to be Free. Thus the four immediately preceding example sentences
are Free Direct Speech, Free Indirect Speech, Free Direct Thought, and
Free Indirect Thought, respectively. Of these, Free Indirect Speech and
Free Indirect Thought are immediately the more interesting, because in
actuality a range of formulations of a character’s speech or thought, indi-
rectly but freely rendered, is always available: an author can opt for a
more distant and narratorial variant, or a more character-proximate
variant one. Take, for example, the third utterance above (FDT), but
recast in Free Indirect Thought mode:
He had never felt like that himself towards any woman.
While that is a fairly soberly narratorial version, an author could alterna-

tively compose a more informal and character-expressive version:
He had never ever felt so . . . so lost in passion, he supposed it was, for

any woman.
Or even:
He had never – how did Swinburne put it? – ‘hold in the music of the
Almighty’ Gretta or anyone else, not Mother, not even the children,
the way that fellow Furey did . . .
Any of these, and innumerable others, which would count as Gabriel’s

Thought, but Indirectly relayed, and lacking a framing clause, would on
those grounds be classed as Free Indirect Thought. It is the most versatile
hybrid form which can carry all of the idiosyncrasy of phrasing, lexical
choice, dialect, epithets and exclamations of the character, as if this were a
direct rendering, with the sole proviso that the grammatical orientation
devices of deixis, the tenses and pronouns, are those of the narrator and
narrative.
Having elaborated matters to the point that we can distinguish a total of
eight distinct modes which a narrator can draw upon in the presentation of
characters’ speech or thought, mention should be made of the fact that
very often in analyses and commentaries, one subsuming category, of dis-
course, is referred to rather than the two subtypes, speech and thought, of
which it is composed. Thus eight modes are reduced to four: Direct Dis-
course (i.e. DS and DT), Free Direct Discourse, Indirect Discourse, and
Free Indirect Discourse. On the other hand, it is often quite useful to
discuss and evaluate the full set of eight modes, keeping speech and
thought distinct, for reasons that will become clear later.
In addition to the eight modes introduced so far, just three more need to
be noted at this point. Most important of these is Stream of Consciousness
(SOC), also known as Direct Interior Monologue. This can be thought of as
lying beyond Free Direct Thought on the thought–representation con-
tinuum: like FDT, it presents the source character’s thoughts in present
tense and first person, but it is usually more inward-looking, less orderly or
‘public in grammar’, more reliant on unstated inferential connections and
associations, less censored as it were, than FDT. Quite often it is implied
that in SOC passages we are scrutinizing the impulse-driven flow of a char-
acter’s mental activity, put into words that are only partially ordered and
marshalled: it is as if the character him- or herself has not really experi-
enced these words as an articulated sequence ‘in the head’. A final pair of
modes to mention are those where the narrator reports a character’s speak-
ing or thinking, but in a way that is sufficiently recast and summarized that
they need to be seen as different from the Indirect Forms, being even
closer to entirely narratorial text; following Leech and Short (1981) I will
call these Narrative Report of Speech Acts and Narrative Report of
Thought Acts respectively (NRSA and NRTA).
By way of practice at identifying some of the eleven categories that
have been introduced, two short extracts from ‘The Dead’ follow, with
annotations of portions of the text that appear here in italics (the remain-
der of the text is what I term Pure Narrative, which can be thought of as a
twelfth category):
Gabriel asked [Mrs Malins] whether she had had a good crossing. She
lived with her married daughter in Glasgow and came to Dublin on a
visit once a year. She answered placidly that she had had a beautiful
crossing and that the captain had been most attentive to her. She spoke
also of the beautiful house her daughter kept in Glasgow, and of all the
nice friends they had there. While her tongue rambled on Gabriel tried
to banish from his mind all memory of the unpleasant incident with
Miss Ivors. Of course the girl or woman, or whatever she was, was an
enthusiast but there was a time for all things. Perhaps he ought not to
have answered her like that. But she had no right to call him a West
Briton before people, even in joke. She had tried to make him ridicu-
lous before people, heckling him and staring at him with her rabbit’s
eyes.
Gabriel’s anodyne conversation with Mrs Malins is initially reported via

Indirect Speech (his whether she had had a good crossing is anwered by her
she had had a beautiful crossing and that the captain had been most attentive
to her. But the narrative rapidly modulates to one even more removed from
the character, that of NRSA, lightly summarizing her numbing banalities.
An elegantly crafted middle sentence narratorially reports that Mrs Malins
is still talking but, like Gabriel, it does not bother to note her words, turning
instead to Gabriel’s internal reflections. These begin in NRTA format (it is
unlikely that the words banish all memory of the unpleasant incident are pre-
cisely the words that run through Gabriel’s mind: these are a narrative
summary of his first thoughts here). But the narratorial presence drops
away, and Gabriel’s inner voice takes charge with Of course the girl or
woman, and so on, until the close, in full-dress FIT, eloquently disclosing
both Gabriel’s sexist pomposity and his hypersensitive vulnerability, which
seem to foster a mean and petty reaction.
Gabriel asked himself was he the cause of her abrupt departure. But
she did not seem to be in ill humour: she had gone away laughing. He
stared blankly down the staircase.
At that moment Aunt Kate came toddling out of the supper-room,
almost wringing her hands in despair.
– Where is Gabriel? she cried. Where on earth is Gabriel? There’s
everyone waiting in there, stage to let, and nobody to carve the goose!
– Here I am, Aunt Kate! cried Gabriel, with sudden animation,
ready to carve a flock of geese, if necessary.
For systematicity’s sake, I would classify was he the cause of her abrupt
departure as IT and not FIT, even though it has the question inversion
which is characteristic of the latter (canonical IT would be Gabriel asked
himself if he was the cause of her abrupt departure), because it remains
explicitly introduced by a framing clause. But as indicated, it is a quite
‘free’ variety of IT (alternatively, some analysts might say that here a
sentence which begins as IT modulates to FIT at some point as it
unfolds). The following sentence, although it could be purely narratorial,
in the context seems more likely to be a further report of Gabriel’s
impressions, now without disambiguating frame, hence FIT. But the
third sentence (He stared blankly . . .) cannot plausibly be Gabriel’s
thought: the narratorial camera has ‘pulled back’ to a medium-distance
shot, now outside Gabriel’s consciousness, to give us a quick observer’s
glimpse of what Gabriel might have looked like from the viewpoint of
another guest. Equally interesting is the final sentence quoted: Here I
am, Aunt Kate! is of course Gabriel’s direct speech, immediately framed
by cried Gabriel, with sudden animation, which is the narrator’s evalua-
tive report. But ready to carve a flock of geese, if necessary could be
either the narrator’s meiotic interpretation of Gabriel’s animation, or it
could be a resumption of Gabriel’s direct speech, a playful exaggerating
reply to Aunt Kate’s request that he carve the goose. On balance I
assume the final phrase is Gabriel’s direct speech, and not narrative
report, but the implied modulation from narratorial with sudden anima-
tion to the semantically congruent ready to carve a flock of geese from
Gabriel is so seamless as to be experientially imperceptible in the ordin-
ary process of reading. Such examples justify claiming that in some uses
of the discourse-representing modes (especially FDD and FID) we find
the subtlest local or temporary alignments of narrator and character, in
wording as well as viewpoint.
Just one of the attractive features of Free Indirect Discourse is that
most readers are not consciously aware of it being at work. It is a hybrid
and a marked or exceptional form, neither pure narrative nor pure charac-
ter-expression, and in many situations it manages to blend into the narra-
torial background. As we have seen, it comprises two sub-types, not
always easily distinguished, Free Indirect Speech and Free Indirect
Thought (henceforth, FIS and FIT respectively). The former of these is
often easier to perceive (and was particularly drawn upon in nineteenth-
century novels), but ultimately more attention needs to be directed to FIT,
which is more common and more important in modern fiction.
But to end this section with the modes with which it began, Direct
Speech and Indirect Speech, let us note how widespread these are in tradi-
tional novels, as well as in other discourses in which the speech of indi-
viduals is reported. Both hard news and human interest stories in
newspapers, for example, are heavily dependent on direct and indirect
speech reporting of the actual words of those involved. Direct speech pre-
tends to be a faithful verbatim report of a person’s actual words (although
we accept, in literature and in non-literary contexts, that all sorts of likely
pauses, reformulations, repairs, and dialectal features are partially if not
completely removed). If direct speech reporting of, e.g., a sports celebrity
in a newspaper story purports or pretends to be a faithful record of the
actual words spoken, then direct speech in fiction is a simulation of a pre-
tence; and sometimes the author may choose not to pretend or simulate
very hard – just as Shaw or Beckett did not feel bound in their plays to
give their characters dialogue that sounded undetectably similar to every-
day natural conversation.
In indirect speech, the narrator or reporter purports to provide an
accurate version of what the speaker said, but not by simply reproducing
that speaker’s own words: instead, the narrator’s words and deictic orien-
tation are retained. Accordingly, indirect speech versions can be fairly
remote from their hypothesized direct speech source. Thus Leech and
Short (1981: 323) note the use of quite summarized and remote versions of
indirect speech. And while the privileges of occurrence of all eleven
speech and thought presenting modes are notionally equal, the actual use
and distribution of the modes may vary in striking ways. Indirect Speech,
for example, is extremely sparingly used in the exemplary short stories
(‘Eveline’, ‘That Evening Sun’, etc.) of this study. On the other hand it is
extensively used in certain kinds of oral narratives and in newspaper
stories.
5.3 Differences between Direct and Indirect Discourse

We can list some of the formal ways in which Indirect Discourse (IS or IT)
normally differs from Direct Discourse (DS or DT). To make this com-
mentary more tangible, cumulative alterations (highlighted in italics on
first occurrence), in line with the list presented, are made to a simple sen-
tence which starts out as a narrative framing clause followed by Direct
Thought:
Gabriel thought: – The time has come for me to set out on my journey
westward.
1 Indirect Discourse will be in the same tense as that used in the encom-
passing pure narrative text (in traditional fiction, this is past tense; in
some contemporary literature, this may be present tense). Direct Dis-
course remains in the present tense which is the normal choice for our
expression of current actions and reactions.
Gabriel thought: – The time had come for me to set out on my

journey westward.
2 First and second person pronouns in the Direct version, denoting the
speaking character and his or her addressees, are matched by gender-
appropriate third person pronouns in the Indirect version, being the
narratorial designations of these characters (to themselves, every char-
acter is an I; but to an external heterodiegetic narrator [unlike, e.g. Pip
in Great Expectations] every character will be a he or a she). This
makes indirect speech more explicit than the direct with respect to the
gender of the addresser and addressee (an issue relevant to gender-
ambiguous or gender-concealing narratives).
Gabriel thought: – The time had come for him to set out on his
journey westward.
3 IS is not set off by inverted commas (speech marks) or other grapho-

logical flagging-up of directness, such as indentation and dashes and
capitalising the first direct word. While inverted commas used to be a
standard marker of direct speech, many twentieth century fiction
writers, including Joyce, dispense with speech marks for DS, instead
marking the onset of a character’s turn of talk by a new line, an
indent, and a long dash. (As a result the beginning of a character’s
direct speech at least – though not the end – is unmistakeable).
Gabriel thought the time had come for him to set out on his
journey westward
4 The reporting and reported clauses relate to each other differently: in

ID they are syntactically related hypotactically, rather than paratacti-
cally as in DD. That is to say, the clauses are linked together in a
superordinate:subordinate relation (the reporting clause being the
head or superordinate); by contrast the relation between reporting
and reported clauses in DD is one of coordination of ‘equals’, hence
paratactic. One consequence of this is that certain ‘robustly paratactic’
DD constructions, such as imperatives and exclamatives, resist incor-
poration into ID format. Thus when Mr Brown tries his whiskey:
– God help me, he said, smiling, it’s the doctor’s orders.
It is hard to devise a truly plausible ID counterpart of this: ?He smil-

ingly cried that God should help him but it was the doctor’s orders. The
main clause-subordinate clause relation in ID can be made explicit by,
typically, the presence of the subordinator that (but use of that to
introduce ID is optional, unlike obligatory que in comparable French
constructions). But other introducers besides that preface ID utter-
ances: a character’s polar (yes/no) question, indirectly reported, will
be introduced by if or whether (like that, impossible in Direct Dis-
course); but content questions will be introduced by the same Wh-
forms that would appear in the DD equivalent. Another consequence
of DD’s paratactic status is that the framing clause can freely occur
either before or after the DD:
He said: – It’s the doctor’s orders. ←→ – It’s the doctor’s orders,

he said.
But such reversibility of projecting clause and projected clause in ID is

much rarer, and sometimes impossible (e.g. with the that and if sub-
ordinators):
He said that it was the doctor’s orders. ←→ *That it was the

doctor’s orders, he said.
Gabriel thought that the time had come for him to set out on his
journey westward.
5 Points 1 and 2 above noted that in ID, tenses and pronouns are
aligned with the narrator’s perspective. This can be stated more gener-
ally: the grammar of ID entails a wholesale adoption of the deictic ori-
entation of the narrator. This means that all deictic indicators that
might appear in a character’s direct utterance, keyed to the character’s
assumed point in space and time (besides tense and personal pro-
nouns, words like here, today, this, and now), will be displaced by
deictic indicators which make sense from the narrator’s perspective
(typically, the ‘non-proximate’ deictic forms: there, on that (same) day,
that, then). Among the deictic indicators must be included the use in
DD of a few verbs of specific directional movement, implying move-
ment towards or away from the speaking character (Come here!; Go
away!); by contrast in ID the movement verbs will be directionally
non-specific, or, like other deictic elements, oriented to the narrator. It
is because come and go are as orientational as here and there that
some combinations of them sound unnatural or absurd: ?They will
come there tomorrow; ?Go here this minute!. In the example sentence,
in addition to now being displaced by then, there is arguably some
residual ‘Gabriel-orientedness’ in the verbs come and set out, which
can be expunged with a little effort:
Gabriel thought that the time had arrived for him to begin his
journey westward.
And if there had been a now in the original DD version – the time has
now come – then this would be substituted by then. But at this point
there is a danger of saturating the example with indirectness features,
to the point that it looks implausible.
6 Signally lacking so far from the list of differences between DD and ID
is the way that the ‘colourful’ idiosyncratic and dialectal qualities of
the speech or thought of a character who is reported directly fail to
transfer to the ID version. Being rooted in the specific subjectivity and
evaluations of that character, they cannot be retained in any standard
ID version, which purports to convey – and this usually quite sparingly
– the subjectivity of the narrator. There are difficulties, too, in repre-
senting DD commands, interjections and rhetorical questions, in ID.
But those are, like the items listed in the previous section, grammati-
cal or grammaticized features; here we are focussing on matters of
lexis, vocabulary, with their seemingly more direct expression of spe-
cific world-view and personality. It is the ‘loss’, as it seems, of the char-
acter’s vivid, colloquial, partisan and engaged lexis that is one of the
most striking characteristics of ID, by comparison with DD. Still, it is
worth emphasizing that what is certainly lost, in ID reporting, is not
‘colourful language’, necessarily, but ‘the character’s (possibly colour-
ful) language’. That is to say, typically and conventionally, especially
in literary fiction, narrators are more ‘neutral’ and formal than the
characters of whom they tell. But this is only a convention or
pronounced tendency, and will be overridden in many particular
cases. One is the novel The Adventures of Huckleberry Finn, where
Huck himself, as first-person narrator, reports nonstandardly the
presumably standard speech of some of the protagonists in his story.
Or consider a ‘lovable Cockney criminal’, about to go into court to
face a charge, who uses IS to report the advice of his lawyer to him:
‘The brief said it didn’t make no difference to the jury that I’d been
nicked twice before because they couldn’t be told that till after
they’d given their verdict’. The verb-choice nicked, and the nonstan-
dard multiple negation are assumed to be the speaker’s, not the
lawyer’s. Finally, returning to the Gabriel sentence we have been re-
working, out of an original DD mould and into an ID one, it has to
be conceded that the DD original shows no glaring lexical or phono-
logical idiosyncrasy of the kind that ID routinely removes. At the
same time, however, the DD original does use a construction which
is rather metaphorical and characteristic of Gabriel’s tendency to
think in tropes and poetic figures. It is quite possible that, if Joyce
had been compelled to render the sentence in ID, he would have
replaced Gabriel’s figurative formulation with something more
literal and narratorial:
Gabriel thought that the time had arrived when he should begin
confronting his own limited lifespan.
But we cannot be sure: in Joyce’s actual text, it is a FIT version

that appears:
The time had come for him to set out on his journey westward.
The preceding six-point listing of differences refers to particular features

being contrasted, in DD by comparison with ID, by their absence or pres-
ence. This was to avoid talking in terms of ‘removal’ (of, e.g. speech marks)
or ‘insertion’ (of, e.g. the subordinator that), which might have fostered the
false assumption that the Indirect Discourse was a transformed version of
the Direct Discourse, and that the latter ‘must have come first’, before the
former was derived from it. Although there is a conventional logical prior-
ity of real-world direct speech over real-world indirect speech, this does not
compel a treatment of all IS and IT as necessarily traceable to DD
antecedents. People are quite capable of ‘reporting’ things that their repor-
tees never said. Besides, as noted earlier, the notional priority of DD brings
few restrictions concerning the actual form that any indirect discourse
report may take.
In light of the six contrasts between DD and ID listed above, and in
particular in view of the last of these, we can begin to reflect on why and
where an author elects to use one and not the other. For example, where a
speaker’s use of highly interactive language features such as interjections,
and dialectalisms and exceptionalities of language is narratively significant,
then this may be sufficient justification for narratorial recourse to Direct
Speech. To opt for Direct Speech reporting is also to accept a scenic
slowing of pace, enhanced focus on the specificity and detail of an inter-
action, and a greater pressure on the author to make such text redeem-
ingly interesting (to offset the inevitable fact that the narrated action will
proceed – if it does at all during the DS dialogue – far more slowly).
This can be confirmed if we review some of the episodes of dialogue
in ‘The Dead’. Gabriel’s conversation with Mrs Malins is mercifully
briefly recorded, and mostly in indirect speech: any more extended airing
of her dullness would merely have laboured a joke and irritated the
reader. By contrast, Gabriel’s grilling from Miss Ivors is rendered in
Direct Speech, and at considerable length (the reader gets a far fuller
version of it than Gabriel is subsequently prepared to share with Gretta).
And it earns its place in the narrative: it is far better for Miss Ivors to be
seen putting Gabriel on the spot in her own words – ‘I have a crow to
pluck with you’ – than via some stilted, distancing phrasing, which would
simply fail to dramatize the scene: She abruptly announced that she had a
complaint to make to him. Equally vivid, and engaging (it makes Gabriel
colour) is her rhetorical question, ‘Who is G. C.?’. And when she then
declares:
O, innocent Amy! I have found out that you write for the Daily
Express! Now aren’t you ashamed of yourself?
it is more or less impossible to render this in Indirect Speech (or Narrative

Report of a Speech Act) in ways that are not grossly unsatisfactory:
She declared he was an innocent Amy, to have been found out writing
for the Daily Express, and asked him whether he wasn’t ashamed of
himself.
She told him she was shocked to find he was writing for the Daily
Express. She thought he should be ashamed of himself.
These are unsatisfactory because these Indirect versions make the report-
ing too prominent: it intervenes and distracts our attention from the
reported. The ‘she declared’ and the ‘she told him’ keep reminding us that
a narratorial third party is present, seemingly taking notes on this lively
spat. But we don’t need such reminders or such third parties (even if at the
back of our minds we know they must be on hand): we prefer the sensa-
tion of being as close to this quarrelling couple as if we were dancing next
to them in the Lancers, catching Miss Ivors’s every ego-chafing word
directly and unedited.
In broad terms, then, DD and ID may be said to differ in their effect on
the reader. Usually the reader feels a greater distance and detachment
from characters and their words when these are mediated via Indirect
Discourse. DD is an environment where characters appear to be in control
and speak for themselves, while in ID the narrator is more overtly still in
control, and reports on behalf of the characters. But the appearance or
illusion of character control should not be overstated: behind all the fic-
tional individuals, however reported, stands the controlling teller, as is
made newly vivid when, for example, stretches of direct speech are set in
sharply evaluative framing contexts (see Fludernik (1993) and Sternberg
(1982) for discussion).
But if character vividness and seeming autonomy are potential corollar-
ies of DD reporting, then equally ID becomes positively desirable when a
narrator judges that projecting such vividness is not appropriate. This
might be because the topic of speech or thought is mundane, or has
already been recorded earlier in the narrative. Or it may be that projecting
character depth, authenticity and autonomy is inappropriate because the
particular character is quite minor in the larger story, and it would be mis-
leading to endow them with so much individuality. On other occasions,
including a second reporting of thought or talk more fully presented previ-
ously, something even more abridged than ID is desirable: an entire phase
of a relatively unimportant conversation or meeting or telephone call or
mental assessment needs to be condensed into a sentence or two, on the
grounds that a fuller telling would be aesthetically ill-advised. Those are
the occasions on which NRTA or NRSA will be most suitable.
5.4 Different again: free Indirect Discourse

The narratorially most fascinating styles of discourse reporting, however,
are those that appear to lie between orthodox Direct and Indirect Dis-
course, revealing some of the formal features of each of those two formats.
If we list again the six major contrasts between DD and ID, in abbreviated
form:
1 DD has character’s tense; ID has narrator’s tense.

2 DD has character’s pronouns; ID has narrator’s pronouns.
3 DD is graphologically set apart; ID is not.
4 DD is paratactic and complementizer-free; ID is hypotactic and
complementizer-prone.
5 DD has character’s deixis; ID has narrator’s deixis.
6 DD has character’s lexis/colouring; ID has narrator’s lexis/colouring.
then FID is a remarkable selecting and blending of the ID or narratorial

option for 1, 2, and 3 (tense, pronouns, and graphological non-
removedness), but the DD or characterological option for 4, 5 and 6 (main
clause syntax – especially noticeable in interrogatives and imperatives and
exclamations; no complementizers; and character’s space/time deixis and
lexis/colouring). But FID is not simply a judicious combination of DD and
ID features, nor simply some middle way between these two. Both DD
and ID are accompanied by a framing or matrix clause: FID eschews one
altogether.
FID styles of thought and speech representation, then, are neither
direct nor indirect according to orthodox prescriptions, but mixings or
mergings of narratorial indirectness with characterological directness. As
noted, and despite first appearances, it really is not a blend, however
subtle, of DD and ID. This is because the mode it most crucially comple-
ments or contrasts with is neither DD nor ID, but PN, the narrator’s direct
interpellation or address of the reader. Accordingly I have elsewhere pro-
posed a schematic spatial distribution of DD, ID, FID and PN (where the
latter is taken to include NRTA and NRSA) along these lines:
ID
PN FID DD
This is intended to suggest, for example, that PN and DD are most sharply
distinct, with ID and FID as intermediate, but with the further condition
that FID is closer to both PN and DD than ID is to either of these (see
Toolan, 1998: 113 for fuller discussion). Despite formally appearing to be a
mixture of DD and ID, it is rarely confused with either of these, but rather
with PN. And this is rather more interesting precisely because between
DD, ID and FID we finally have only variations in the fullness or edited-
ness of rendering of the same individual’s ‘voice’; but where we cope with
the risk of confusing FID and PN we are in danger of misattributing or
misidentifying two entirely different voices and two different individuals,
the character and the narrator.
If there is one linguistic feature that seems noticeably more prominent
in FID than in alternative modes of discourse representation, it is modal-
ity. FID is marked by frequent use of modal verbs (must, should, had to,
could, might, would) and sentence adverbials (certainly, perhaps, maybe,
surely, of course, etc.) expressing judgments about the likelihood or neces-
sity or desirability of some action or state transpiring. All such modality
discloses the character’s needs and wants. All such modals are woven into
contestable judgments. They prompt an FID reading of the text they
accompany in just those cases where we find it implausible to imagine that
it is the teller who, perhaps rather abruptly, intrudes into the story to tell
us what some character ought to do, or what possibly had happened or
would happen.
Whenever a reader wishes to try to confirm for themselves that a
particular portion of text is FID essentially sourced in the character, and
not Narrative report sourced in that usually abstract narrator, a simple
framing or commutation test is sometimes useful. The test helps sharpen
vague or uncertain impressions, and uncertain judgments about the inton-
ation of a passage. All the reader needs to do is to assess the plausibility,
or conversely the jarring effect, when either of two frames is added to the
target stretch of text. Those frames are:
I, the narrator, tell you, the reader [insert text to be probed, unmodi-
fied]
and, alternatively,
[insert text to be probed, with any pronouns referring to the putatively

discoursing character converted to first person, and with tenses con-
verted to the present tense of thinking/speaking], the character
remarks, to themselves or other characters
The former probe explicitly casts the utterance as a narrator-to-reader

communicative act, or Pure Narrative; the latter casts the utterance as a
character-to-self-or-other-character communicative act. Any sentence, or
sentence-part, can be put to these framing tests to help clarify their FID or
PN status. Consider these snippets from an early point in Gabriel’s final
revery:
Perhaps she had not told him all the story.(1) His eyes moved to the
chair over which she had thrown some of her clothes. . . . (2) Poor
Aunt Julia! She, too, would soon be a shade . . .(3)
Taking these slightly out of order, the second of these sentences is clearly
narratorial: while it fits the narratorial frame quite unexceptionally, it is
clearly ludicrous if explicitly cast as characterological, via the second
frame:
?My eyes move to the chair over which she has thrown some of her
clothes, the character remarks (to himself).
People (characters) simply do not think things like that. Conversely, sen-
tence (3) is almost as incontestably characterological, FIT. This could be
an intrusive omniscient judging narrator’s disclosure to the reader, but this
would be an inexplicable and unmotivated departure from the norm estab-
lished throughout the book so far, and it would not sit well with adjacent
text. The adjacent text tells us that Gabriel is looking at things and think-
ing about things, and it is far more fitting to assume that this is one more
thing Gabriel thinks, and indeed that it is a thought he takes further in
subsequent sentences of this paragraph. And it fits the second probe test
perfectly:
Poor Aunt Julia! She, too, will soon be a shade, the character remarks
(to himself).
Finally, the first sentence Perhaps she had not told him all the story, taken
alone would seem to pass both the narratorial and characterological probe
test: good evidence that, in theory, the sentence could be either narrator’s
suggestion or character’s speculation. Again, context and co-text (and
content: despite his rather protracted interrogation of her, it would be in
Gabriel’s character to fear or suspect that not even Gretta had been
entirely honest with him) make it overwhelmingly more appropriate to
read this potentially ambiguous observation as Gabriel’s, in FIT.
As much as anything else, it is the modality in this last example –
Perhaps she had not told etc. – and the expressive evaluation in the previ-
ous example – Poor Aunt Julia! She, too, etc. – that encourage us to assign
FID interpretations. In seemingly narrative text, FID has the capacity to
‘put back’ all the immediacy and subjectivity and expressivity that other-
wise, in standard ID, gets edited out. And this ‘restored’ subjective expres-
sivity can extend to eccentricities of spelling and writing in general,
reflecting the non-standardness of the character: such features are comfort-
ably accommodated in some kinds of FID. All of this amounts to a restora-
tion of the colourful individuality of characters’ direct expression. We feel
FID to be more vivid and colourful than PN and ID report because both
the latter tend to be more detached, sober, restrained, and standard-
English-speaking than the words of the character in situ, undergoing the
experience, and talking and thinking their way through that experience in
frank and uninhibited ways. Of course the latter is an effect of direct experi-
ential reaction, but no less significant for that.
But these contrasts of PN sobriety and FID expressivity are reversible
patterns rather than rules. As suggested earlier, we could easily encounter
vivid and earthy narration of the life of a dull bureaucrat (himself stilted in
word and thought), a style that then gave way, where the character’s own
discourse was conveyed by either direct or free indirect means, to plodding
banality, arid and colourless. As elsewhere in narrative study, it is the per-
ception of difference (here between character speech and narrator
speech), which is the true criterion. And in some narratives, as in the
novels and stories of Henry James, where narrator and characters seem to
share a single lect so that contrasts scarcely arise at lexicogrammatical
levels at all, the differences can only be painstakingly derived from the
content of protagonists’ discourse (e.g. its varying degrees of reasonable-
ness), and the content of their character.
5.5 Who speaks, who thinks?

If the crucial question in focalization was ‘From whose spatiotemporal ori-
entation is this conveyed?’, here in FID it is ‘To whom do we attribute
these spoken words or articulated thoughts?’ The ‘Who is saying/thinking
this?’ question is useful to keep in mind, since FID is a very open category.
Despite the foregoing list of usual indicators of FID, there is no single
necessary feature of FID, which you have to have in order for a phrase or
clause to qualify as FID. Narratives in the present tense (where there will
be no tense-difference between direct and indirect discourses) or with
first-person narrators can just as easily contain FID as other types. As
Leech and Short (1981: 328–33) show in the following two examples, a
single character-attributed expletive, or even a single character-expressive
punctuation mark, can signal the presence of FID rather than narratorial
indirect discourse:
He said that the bloody train had been late.

He told her to leave him alone!
We assume that the bloody is the character’s (He), and not the narrator’s,
and similarly that the exclamation mark (and presumably at least the word
immediately preceding it) is the character’s. We thus have marks of what
must surely be FID, but we do not have clear indications of its extent – of
‘how much’, before and after the word bloody, is also essentially attribut-
able to the character. This amounts to another interestingly undecidable
dimension of FID. In the second example, do we decide that the entire
clause ‘leave him alone!’ is FID, or only ‘alone!’, or merely ‘!’?
FID is a long-established technique that can be usefully related to such
fundamental literary distinctions of narrative method as those between
showing and telling, or mimesis and diegesis. While the former places the
emphasis on a direct characterological representation or impersonation,
the latter implies a more indirect, detached teller-oriented conspectual
presentation. In the narration of events mimesis is associated with a scenic
presentation, while diegesis is linked to a condensed or ‘edited’ and
summarized account – with, consequently, a larger role played by the
teller who condenses or edits. Mimesis presents ‘everything that hap-
pened’ in one sense, but really only everything as it would be revealed to a
witness within the scene; in these latter respects it is quite partial and far
from comprehensive. And as Stanzel (1981) notes, it typically comes with
internal character-focalization. Diegesis presents ‘everything that hap-
pened’ in another sense, but only everything that a detached external
reporter decides is worth telling – a reporter who is able to reflect, reor-
ganize, and decide upon the point or teleology of the story prior to narrat-
ing it. We can accordingly predict that diegetic narration will have
more manipulations of temporal order, duration and frequency, and more
evident ranking or hierarchical ordering of event-presentation.
5.6 FID: functions and effects

In the neo-Platonist terms mentioned earlier (mimesis and diegesis), FID
amounts to a mimetic diegesis (a telling that shows or presents aspects of
the character’s ‘own’ words or thoughts). But there are numerous ways of
characterizing FID, in fact: as substitutionary narration; as combined dis-
course; as a contamination, tainting or colouring of the narrative; as a dual
voicing. My own preference is for viewing it as a strategy of (usually tem-
porary or discontinuous) alignment, in words, values and perspective, of
the narrator with a character. I favour the word ‘alignment’ because it
doesn’t prescribe whether that closeness of narrator to character is going
to be used for purposes of irony, empathy, as a vehicle for stream-of-
consciousness or the clashing of two voices, or whatever: the alignment is
perceived, then the function (or ‘naturalization’) is worked out by the
reader. The term ‘alignment’ also helps us keep in mind that, in terms of
lexicogrammatical markers and aesthetic or narrative effect, there is a con-
tinuum from pure narrative words to pure character words, with any
number of points on that continuum.
Like McHale and others (Ginsberg, 1982; Pascal, 1977), Jefferson rec-
ognizes that a key source of the impact of the FID sentence is its ambigu-
ous mixture of proper narrative and proper speech or thought:
The dual voice of FID which is responsible for the superficially realist
effect of immediacy is also an ambiguity which is highly unrealistic.
From a realist point of view, FID is a doubly disconcerting use of lan-
guage: its ambiguities cut it adrift from the two points at which we
commonly imagine language to be anchored to reality, the speaker
and the referent. It is neither fully expressive nor fully referential, and
this invraisemblance differentiates it most profoundly from other
forms of reported discourse.
(Jefferson, 1981: 42)
By invraisemblance Jefferson alludes in part to FID’s lack of vraisem-

blance: the effect or convention of faithful and ‘realistic’ representation of
the real world that many novels foster and many readers demand. The
invraisemblance of FID stems from the fact that it gives rise to sentences,
like Perhaps she had not told him all the story, and She too would soon be a
shade and Yes, the newspapers were right which, in just those forms and
with just those referential commitments, none of us in the discoursing of
our everyday lives, including our storytellings embedded in conversations,
could ever actually say. FID sentences are in this sense unsayable or
unspeakable, impossible to anchor to one speaker at one place and time,
since they are impossibly divided between two distinct speakers (narrator
and character) and anchorages. We can no more ordinarily speak (that is,
use) an FID sentence than we can ‘do both voices’ when we greet
someone else and are replied to by them: *How are fine, and you?
At this point mention should be made of Banfield’s challenging and
wide-ranging theory of narration and discourse (Banfield, 1982). Banfield’s
theory of unspeakable sentences argues that not merely FID but all narra-
tive sentences are marked off from ordinary communicative discourse by
their unspeakability. Narration is unspeakable because, unlike discourse
(which is both communicative and expressive), it is a text with neither a
genuine addressee (nor any textual traces of one) nor a genuine expressiv-
ity-disclosing speaker. FID, in partial contrast, carries abundant expressiv-
ity (that of the character), but still no genuine I–you communicativity, and
as a result remains unspeakable. The theory is fascinating, wide-ranging,
and presented in great technical and critical detail, and for those among
other reasons it merits scrutiny; but its declared adherents are few. In the
first edition of this book, I presented and critiqued Banfield’s argument at
length, and will not do so here; interested readers may wish to read that
earlier assessment, together with those of McHale (1983) and Simpson
(1993: 35–8), both of whom, like me, dissent from Banfield’s conclusions
(which, inter alia, include a rejection of the ‘dual voice’ approach to FID
outlined above).
As a final point of orientation we should note some of the influential
work of Short and his colleagues (e.g. Leech and Short, 1981; Semino,
Short and Wynne, 1999; Short, Semino and Culpeper, 1996) which surveys
and analyses trends in literary and non-literary discourse representation,
increasingly using corpus materials and methods. It is they who have
drawn attention to the fact that, in newspapers particularly, a newsworthy
item of written discourse (e.g. an open letter of resignation from the CEO
of some major public body) may be slightly – or even significantly – differ-
ently reported on separate occasions, even where both purport to be the
verbatim record. For example, the letter of resignation may be quoted, in
the front page story, as
‘I now feel, in the light of this shabby persecution by the press, in

which I have been subjected to cruel and unreasonable harassment,
that I must relinquish my post to defend my reputation.’
But on an inside page, one might find the following fuller version, equally
purporting to be the verbatim record:
‘I now feel, in the light of this shabby persecution by the press, in

which I and my family have been subjected to cruel and unreasonable
harassment, without the possibility of setting the record straight in a
timely fashion, that I must relinquish my post in order to defend my
reputation.’
In this example, an original letter has been reproduced in slightly different

versions, on the inside and front pages of the newspaper. To situate this
practice alongside other kinds of discourse representation, Short et al.
propose that both versions be called Direct Writing (DW). But of course
where you can get Direct Writing report you can also get report via Indi-
rect Writing:
In his letter Sir Nigel said that his whole family had been subjected to
harassment by the press.
Indeed, the entire menu of Writing report options, parallel to those for
Speech and Thought, is available: FDW (Free Direct Writing), DW (Direct
Writing), FIW (Free Indirect Writing), IW (Indirect Writing), NRWA
(Narrator’s Representation of Writing Act), and NW (Narrator’s Report of
Writing). One application of these terms could be to plagiarism: Free
Direct Writing passed off as (genuine) writing. Another is to literature:
near the opening of Jane Austen’s Mansfield Park, for example, the narra-
tive reports a letter written to Lady Bertram by her impoverished sister
Frances. The letter is particularly concerned with the future prospects of
Frances’ many children:
Her eldest son was a boy of ten years old, a fine spirited fellow who
longed to be out in the world; but what could she do? Was there any
chance of his being hereafter useful to Sir Thomas in the concerns of
his West Indian property?
This is Free Indirect Writing, a deictically indirect or narratorially-framed

version of the written discourse which, we understand, Frances Price has
sent to Lady Bertram.
But perhaps the most influential of the observations made by Short and
his colleagues are some that appeared in Leech and Short (1981) concern-
ing the contrasting ‘norms’ of speech-reporting and thought-reporting,
and, as a consequence of these different norms, the customary and differ-
ent reception that readers accord to FIS and FIT. Here, then, the two
modes are not collapsed into a superordinate FID category, and rather the
distinctive effect of each type – FIS and FIT – is assessed comparatively,
but separately. Leech and Short’s starting-point has the force of common
sense. Under normal conditions, we expect that any ordinary close witness
(W) to an incident, a series of events involving speech and action, should
be able to report both the actions of the observed individuals and, in
Direct Speech, their actual words. (A court of law would prefer it: ‘Can
you tell the court the exact words that the defendant used?’, the barrister
will ask.) This remains the case even though numerous studies show that
verbatim recall is rarely achieved.
By contrast we do not accept that an ordinary witness (W) to an inci-
dent can reliably report the sententially-composed thoughts – or even the
chaotically-jumbled ones – that ran through A’s head, or B’s, at any point
during the incident. At best, if W were a close friend of A or B, and inter-
viewed them afterwards, they might glean from them an account of what
they had been thinking – but in W’s subsequent narrative report of the
incident, any indications of what A or B had been thinking, however pre-
sented (across the range of options from FDT to NRTA), would in fact be
only a report of a report. W has no direct access to A’s thoughts, common
sense tells us; only A has. In light of this, and again other things being
equal, a witness/narrator does surprise us, and is regarded as in some
respects imposing on us, the addressees, and on the thinking character
reported, when they offer a heterodiegetic report of direct thought. (And
the rules of evidence in criminal law would completely exclude a witness
from anything so outrageous as ‘telling the court what thoughts ran
through the defendant’s mind’.)
The upshot of these common sense predispositions in readers and
addressees, Leech and Short suggest, is that while Direct Speech is
‘normal’ as the means for reporting speech – a character’s speech is
expectably reported directly unless there are contingent reasons for opting
for some other mode – Indirect Thought is ‘normal’ as the means for
reporting anyone else’s thoughts besides your own, including a character’s
in a heterodiegetic narrative. And again, if IT is the norm, departures from
that norm become ‘accountable’: we will tend to expect there must be a
(good) reason why the default option of IT has not been adopted. If one
visualizes the Speech-mode continuum on a horizontal line, with FDS at
the far left, then most departures from this putative DS norm will be ‘to
the right’, which is also to say, by recourse to modes which adjust the char-
acter–narrator weighting increasingly in favour of the narrator: as one
moves towards NRSA, one adopts modes that carry less and less of the
character’s expressivity and subjectivity in their own words, and more and
more of the narrator’s controlling and editorialized version of those words.
Even FIS, most crucially, gives the narrator ‘more of a say’, and more
opportunity to insinuate narratorial intonation and covert comment, than
the DS norm does. Accordingly, Leech and Short argue, FIS is often read
or heard as a narratorial ‘stepping back’ from allowing the character full
and uncritiqued rein, as DS would tend to do; FIS often discloses charac-
ter’s words laced with narratorial irony.
On the other hand, again if the Thought-mode continuum runs horizon-
tally, with SOC and FDT leftmost and NRTA rightmost, then a move
away from the IT norm ‘to the left’, to FIT, is a contrasting move to that
entailed in adopting FIS. FIT, by comparison with IT, gives the character’s
expressivity and subjectivity fuller disclosure, pushing down the narrator
into a less prominent role, carrying less scope for implicit editorial cri-
tique. And the assumption is that this option is taken typically where the
narrator has no inclination to critique or ironize the represented thinking
character; rather, there is implicit narratorial empathy with the character.
If these trends are as robust as is here suggested, they would be reminis-
cent of what are called scalar implicatures in linguistic pragmatics (Levin-
son, 1983: 132ff.). Consider the set of quantifying terms:
all, most, many, some, and few
We ordinarily expect a speaker to use the ‘leftmost’ appropriate term, so

as to be cooperatively informative, in any situation, even though any terms
to the right of a particular item must be true if that particular item is true.
Thus if all my ten siblings live in Belgium, I should normally say, ‘They all
live in Belgium’ and not that many or some of them live there, even
though these are true: they are true but would be insufficient and mislead-
ing. Anyone who knows or discovers that all my siblings live in Belgium
will assume I am signifying or implicating something if I use a lesser quan-
tifier like most. In a roughly comparable way, possibly, recourse to less
direct speech-presentation than DS may prompt the reader to infer some
narratorial implicature involving comment or criticism; while recourse to
more direct thought-presentation than IT may similarly invite the infer-
ence of narratorial respect, empathy, an ‘entering’ of the character’s inti-
mate mental space that is licensed on the grounds that the narrative does
so without any intent of controlling or managing that thought-world, but
simply of being its platform.
All this is schematically summarized in the following figure, based on
Leech and Short (1981: 344):
Characterological.....................................................................................Narratorial
FDS DS FIS IS NRSA PN
norm implied reduced character alignment/empathy
FDT DT FIT IT NRTA PN
implied increased character alignment/empathy norm
Finally, whether the theorist dwells on the undecidability of FID, or its

‘unspeakability’, or its dialogism, all commentators recognize that no
ordinary speaker, speaking to us face-to-face (rather than, for example,
narrating) would produce the kind of discourse that FID inhabits. That
is because in such FID-coloured narrative two radically distinct situ-
ations of utterance are simultaneously implied: situations with quite dif-
ferent participants (character-to-character versus narrator-to-reader)
and quite different spatiotemporal settings (the ‘then’ of the story
versus the ‘now’ of the telling/reading), and with convergence only in
the area of subject matter (a character’s represented thoughts or speech
are experiential subject matter for that character, but also the subject
matter of the narrative for its reader). The oddity of encountering two
amalgamated situations of utterance should be evident if we reflect
upon Lyons’s much-quoted characterization of ‘the canonical situation
of utterance’:
This involves one-one, or one-many, signalling in the phonic medium

along the vocal-auditory channel, with all the participants present in
the same actual situation able to see one another and to perceive the
associated non-vocal paralinguistic features of their utterances, and
each assuming the roles of sender and receiver in turn. . . . There is
much in the structure of languages that can only be explained on the
assumption that they have developed for communication in face-to-
face interaction. This is clearly so as far as deixis is concerned.
(Lyons, 1977: 637–8)
FID, we can see, is at a fascinating distance from this ‘canonical situation

of utterance’. By contrast the main subject-matter of the next chapter, oral
narratives of personal experience, are typically to be found wholly embed-
ded in precisely that canonical situation.
Further reading
The literature on both the theory of free indirect discourse and literary exploita-
tions of it is voluminous. A good starting place is Chapter 10 of Leech and Short
(1981). Pascal (1977) is a clear and detailed account of FID’s role in, especially,
nineteenth-century European fiction; McHale (1978) is an authoritative survey,
making connections to fundamental issues in the theory of poetics. Also to be rec-
ommended in this vein are Hernadi (1972) and Ginsberg (1982), while longer
innovative studies include Cohn (1978), Stanzel (1984) and Fludernik (1993).
Sternberg (1982), is excellent on ‘the indirections’ of direct speech. More linguistic
discussions of FID begin with Bally (1912), and include Voloshinov (1973), Bicker-
ton (1967), Jones (1968), Bronzwaer (1970), Fillmore (1974), Dillon and Kirchhoff
(1976), McKay (1978; 1982), Banfield (1982) and, for various languages other than
English: Hagenaar (1996), Redeker (1996), Kullmann (1995), Vuillaume (1998),
Hummel (1999), and Hirose (2000). Discussions of Banfield’s work in McHale
(1983) and Simpson (1993) are also very useful. Other work to be consulted
includes Duchan et al. (1995), Ehrlich (1990), Dry (1995), Oltean (1993), Mezei
(1996), Baron (1998) and Ferguson (2000). On reporting speech in non-literary
contexts, see, e.g. Slembrouck (1992), Waugh (1995), Vincent and Perrin (1999)
and Myers (1999). On the thorny issue of ‘faithfulness’ in discourse reporting, a
recent paper is Short, Semino and Wynne (2001). Advanced studies of FID and
related topics continue to appear in such journals as Poetics Today; Journal of Nar-
rative Technique; Narrative; Language and Literature; Journal of Literary Seman-
tics; and Style.
Notes and exercises

1 Much of the narrative of ‘Eveline’ is not plausibly FID, and close attention to
the text reveals a highly fluid, even volatile, style of narration that moves from
intrusive narration, summarizing her thoughts, in sentence 1 below, to the
almost entirely impersonal report – except for the word ‘mournful’ – in sentence
2, to the probable FID of sentence 3:
She felt her cheek pale and cold and, out of a maze of distress, she prayed
to God to direct her, to show her what was her duty.(1) The boat blew a
long mournful whistle into the mist.(2) If she went, tomorrow she would be
on the sea with Frank, steaming towards Buenos Ayres.(3)
(40)
Similarly, it is implausible to suggest that such sentences as the following, the
work of an intrusive omniscient narrator, represent Eveline in her own thoughts
and words:
As she mused the pitiful vision of her mother’s life laid its spell on the very
quick of her being – that life of commonplace sacrifice closing in final
craziness.
(40)
This sentence provoked an interesting difference of interpretation between two

noted Joycean scholars, Chatman and Hart. Of it Hart wrote that its ‘strong note
of falsity’ reflects Eveline’s inability to love, or think of her situation except in
‘tawdry clichés’ (Hart, 1969: 51). In rebuttal, Chatman protests that Hart has
confused character viewpoint and narrator’s voice, misreading the scene as a
result:
Surely the objectionable words are not Eveline’s but the narrator’s. It is he
who is parodying pulp-literature sentimentality in tawdry clichés . . . [using
language that is] not in her vocabulary.
(Chatman, 1978: 153, fn. 9)
Chatman may be right, but the issue is not at all easy to decide; and if he is right,
it leaves in need of explanation why it is that Joyce creates, for a story so much
told from Eveline’s point of view, a narrator who adopts parodying clichéd lan-
guage which inevitably tends to mock and belittle Eveline and ‘fictionalize’ her
situation. A careful charting of the strategic changes in narrative orientation in
‘Eveline’ will uncover many such complexities – and authorial risks.
2 It has already been implied, in references in the chapter to ‘coloured’ ID, that
some types of ID are a good deal closer to FID, having dual-voice qualities,
than other types of ID. Indeed some of the former would simply be called FID
by some analysts, despite the presence of a framing clause. It may be quite
instructive, using a corpus of literary examples or plausible composed
examples, to attempt to specify what character-oriented or FID-like grammati-
cal features do seem allowable in ID, and which such features seem positively
disallowed. That is, among the inventory of FID features, which are possible in
ID and which are impossible, and why? Among FID features worth testing for
are:
Subject–auxiliary verb inversion in reported questions

Topicalization (stating the clause Subject or Object initially, and then a
second time pronominally: He said these pistachio nuts, I can’t stop eating
them)
Language-mixing (narrator’s dialect with character’s dialect)
Attitudinal disjuncts attributable to the character not the narrator (i.e. not
He turned to her and said bluntly that was her problem, but He turned to
her and said frankly he couldn’t give a fig.)
Incomplete sentences
Subjectless Imperatives
Direct address (She went to the bottom of the stairs and yelled that Claud,
Charlotte and Lau-raaaah!, DINNERtiiime!).
Good commentaries on this topic appear in Banfield (1982) and Fludernik

(1993).
3 As mentioned in 5.2, over the years since Leech and Short (1981), Short and his
co-researchers have proposed a slight revision to the categorization of the kinds
of indirect reporting of a character’s speech or mental processing, to capture
kinds of representation that are somewhat distinct from all the extant types (see
Short et al. 1996). These are Narrator’s report of Voice (NV) and Narration of
Internal States (NI), the latter being similar to what Cohn (1978) has called ‘psy-
chonarration’. These comprise allusions to or annotations of characters’ speak-
ing and thinking which differ from all the other modes for reporting character’s
discourse in that they carry no mention of the content of the speaking or think-
ing, but chiefly report its manner or style. They arguably therefore fall outside
the domain of discourse-report altogether, and into that of Narration. Unlike all
the other discourse-representing modes, for example, it is hard to see how one
might convincingly re-cast an instance of NV as IS or DS (and similarly with
respect to NI), which is always to a degree possible, across all the other modes,
even if one begins with something as abridged as an NRSA occurrence. On the
other hand it is true that sometimes only a fine line distinguishes NV from
NRSA, or NI from NRTA. Thus one of Short et al.’s examples of NV, taken
from J. G. Ballard’s The Empire of the Sun, is
Voices fretted along the murmuring wire.
Presumably if this had run Their voices quarrelled back and forth along the tele-
phone line, this might be classified instead as NRSA.
6 Narrative as socially situated
The sociolinguistic approach
6.1 Labov and narrative structure

A great deal of important work on naturally-occurring narrative – particu-
larly the narratives of ordinary people in their extraordinary everyday
lives – stems from just two seminal essays by the American sociolinguist
William Labov. The first, written jointly with Joshua Waletzky, appeared
in Helms (1967), under the title ‘Narrative analysis: oral versions of per-
sonal experience’. The second appeared as Chapter 9 of Labov’s Language
in the Inner City (1972), and is titled ‘The transformation of experience in
narrative syntax’. Much more recently, Labov has published ‘Some
Further Steps in Narrative Analysis’ (1997), in a special issue of the
Journal of Narrative and Life History devoted to Labov-influenced com-
mentaries from a host of scholars celebrating the thirtieth anniversary of
the 1967 paper. That recent paper makes only slight adjustments to the
widely-adopted Labovian account.
Labov and Waletzky’s hypothesis is that fundamental narrative struc-
tures are to be found in oral versions of personal experience – the ordinary
narratives of ordinary speakers. They wish, by looking at many narratives,
to identify and relate formal linguistic properties of narrative to their func-
tions. And like all structuralists, their analysis is based on the perception
of a delimited set of recurrent patterns and the setting aside of what they
take to be local differences in the pursuit of the deeper structural sim-
ilarities: ‘We will be relying upon the basic techniques of linguistic analy-
sis, isolating the invariant structural units which are represented by a
variety of superficial forms.’ (Labov and Waletzky, 1967: 12).
They relate those identified linguistic-structural properties to functions,
nominating two broad functions in particular as core. The first of these is
the referential function: narrative’s functioning as a means of recapitulat-
ing experience in an ordered set of clauses that matches the temporal
sequence of the original experience. Versions of this ‘sequential recapitu-
lation’ function have arisen before in this book, in the definitions of story
or bare narrative. The second function has been less universally emph-
asized, partly because narrative has not always been properly related to its
contexts of occurrence, its role as an instrument or resource for its human
‘users’. The second function then, which Labov and Waletzky term ‘evalu-
ative’, attends to the users of narratives, and the requirement that a narra-
tive has a point, is worth telling, as far as the teller (and preferably the
addressee also) is concerned.
Unlike Propp, whom they suggest focussed on ‘large semantic units’
(13), Labov and Waletzky concentrate first on
the smallest unit of linguistic expression which defines the functions of

narrative [i.e. the smallest unit which can ordinarily realize those func-
tions] – primarily the clause . . .
(13)
And their first task becomes that of attempting to relate
the sequence of clauses in the narrative to the sequence of events

inferred from the narrative.
(20)
This task is more manageable than it might appear (if we think back to
Genette and Nabokov’s Pnin, discussed in earlier chapters), because, as
Labov has reiterated in his 1997 paper, among the kinds of ‘crisis’ narra-
tives of personal experience that were studied (usually someone’s account
of a situation where they genuinely feared for their life), without exception
clauses reporting events were told in the same order as the putative original
events would have happened. So Labov is prepared to assert that oral nar-
ratives of life-threatening personal experience are invariably recounted
chronologically.
This, in turn, points to the quite sharp difference between the kind of
oral and ‘real time’ production that sociolinguists such as Labov tend to
study, and the kind of crafted written literary production that narratolo-
gists like Genette study. A review of the enduring connections and com-
monalities between these seemingly different narratives and analyses will
be included in this chapter. Techniques of temporal reordering and Prous-
tian complexity, however, have no place in the Labovian studies:
The basic narrative units that we wish to isolate are defined by the fact
that they recapitulate experience in the same order as the original
events.
(1967: 20–1)
Accordingly, and in just these terms, the following sample constitutes a

narrative:
Well, this person had a little too much to drink and he attacked me
and the friend came in and she stopped it.
while the following version, where there is presentational reordering

Narrative as socially situated 145
through subordination, is simply not a narrative in Labov and Waletzky’s
terms, even though it is an acceptable recapitulation of experience:
A friend of mine came in just in time to stop this person who had had
a little too much to drink from attacking me.
It displays the kind of temporal/clausal manipulation that they simply did

not find actually produced by their informants, as an exemplar of the
particular text-type that interested them. This preference might be
summed up by saying that, in telling narratives of personal life-threatening
experience, a ‘blow by blow’ account is obligatory. In passing, Labov and
Waletzky acknowledge that narratives in their restricted sense are just one
distinct kind of storytelling.
The foregoing remarks should serve as a warning that their assumptions
and procedures differ in some respects from those we have previously
looked at. The attraction of their principles lies in their clarity, replicabil-
ity, and their search for a basic pattern from which more complex narra-
tives might be derived. And their differences from the narration analysts
should not be overstated: after all, Genette no less than they develops his
taxonomy of diegetic variations on a foundational ‘natural’ order of event
recapitulation.
6.2 Fixed narrative clauses, free evaluative clauses

The Labovian thesis, then, is that true narrative clauses, the backbone of
narrative as they narrowly define it, are temporally ordered independent
clauses (along with their dependent subordinate clauses) that must occur
in a fixed presentational sequence. There is a ‘shiftability’ of subordinate
clauses, around the main clause on which they depend, that excludes them
from consideration as fully narrative clauses; in Labov (1997) the
independent clause that is the ‘head’ of a narrative clause is termed a
sequential clause. The fixity of sequence of properly narrative clauses is
quite crucial: any reordering of narrative clauses would create a different
story. Compare
1 John fell in the river, got very cold, and had two large whiskies.
with
2 John had two large whiskies, fell in the river, and got very cold.
Clearly, despite many similarities of form, the differently-ordered clauses

here represent two radically different stories, with different cause-and-
effect relations between events, and different probable evaluations by
teller and hearer (the listener might comment ‘serves him right’ about the
second John but could hardly do so about the first).
Labov and Waletzky present a rather more formal account of the non-
shiftability of narrative clauses (relative to each other) by specifying the
displacement potential of clauses in narratives. Setting a narrative out on
the page with just one main or independent clause per line of text, and
labelling the clauses alphabetically, they annotate the letter-label of each
clause by noting, either side of it, the number of previous and subsequent
clauses that any particular clause could displace. Accordingly narrative 2
above might be annotated as follows:
0a0 John had two large whiskies,

0b0 fell in the river,
0c0 and got very cold.
which postulates that none of the three clauses can be moved earlier or
later without ‘changing the inferred sequence of events’ (Labov and
Waletzky, 1967: 21) in the original experience. But if the following obser-
vation is added at the story’s close:
This happened when he was still at school.
it seems that this is freely shiftable to any position earlier in the story
without interference with the inferred sequence of events:
John had two large whiskies – this happened when he was still at
school – fell in the river, and got very cold.
If this additional comment were to come at the end of the sequence, the
annotation would be:

0c0 and got very cold.
3d0 This happened when he was still at school.
The annotation 3d0 states that clause ‘d’ may occur as many as three
clauses earlier, but no later, in this narrative, without alteration to the
inferred sequence of actual events, the essence of the narrative. The previ-
ous three clauses, on the other hand, continue to be locked into their given
interrelated positions. A clause like ‘d’, describing the circumstances
surrounding the fixed sequence of events of a narrative and with the
potential of being moved anywhere in the text, is known in Labovian ter-
minology as a free clause.
Narrative clauses (reporting the ordered experience of the interrelated
events) and free clauses (reporting the context of the events, and partici-
pants’ perspectives) are the most sharply contrasted pair of clause types
that Labov identifies. Each is the basic means for enacting the two func-
tions of narrative cited earlier: the ‘referentially’ ordered recall of tempor-
ally ordered experience, and the ‘evaluative’ staging of the story so as to
convey its point and tellability. But there are two further types of clause,
intermediate between narrative and free ones; they are types of clause
which are movable within limits, and with intermediate functions. The first
intermediate type are those linked narrative clauses, usually only pairs or
triads, that can be mutually reordered ‘freely’. These are called coordinate
clauses – the coordination here being not so much in terms of grammar as
in terms of narrative sense. Many grammatically coordinated clauses are
not coordinated in this special sense of permissible narrative reorderability
(She took arsenic and died, for example, is grammatically but not narra-
tionally coordinate; the clauses in She drank coffee and alcohol. She
smoked cigars. She ate red meat all the time, are narrationally but not
grammatically coordinate). Revising the ‘drunken drenching’ story,
Labovian coordination can be demonstrated:

0c1 got very cold
1d0 and ruined his suit.
4e0 This happened when he was still at school.
Clauses c and d are coordinate: the order in which John’s getting cold and
ruining his suit are reported is freely reversible without alteration to the
basic narrative.
Conversely, there are some free clauses that are not entirely freely
shiftable to just any other place in the sequence. They have a limited
domain of occurrence or reach, and are termed restricted clauses. In the
following version:
0a2 John got this urge to be the star of the party.

0b0 He had two large whiskies,
0c0 performed a standing somersault on the embankment wall,
0d0 fell in the river,
0el got very cold
1f0 and ruined his suit.
6g0 This happened when he was still at school.
clause ‘a’ here is a kind of comment in parallel to the main action, and –
arguably, at least – could come a little later in the sequence. It could occur
after ‘b’ or ‘c’, but no later than this, if we interpret the contents of ‘d’ as
an unintentional lapse from ‘star of the party’ behaviour (he fell, he did
not simply jump). The event that ‘d’ reports, we imagine, must have con-
siderably dampened the urge that ‘a’ discloses. But a greater degree of
arguability of the analysis enters with the notion of restricted clauses:
some readers may contest the analysis offered above, and might want to
reclassify ‘a’ as a wholly free clause. We may expect, then, that in more
complex narratives these restricted clauses will be a source of much satis-
fying analytical-interpretive trouble as they generate uncertainties and
larger debates about the ‘what happened when, and why’ of particular
texts.
What then, for Labov, is a narrative? Minimally, and similar to my own
definition in Chapter 1,
a minimal narrative [is] a sequence of two clauses which are tempor-

ally ordered . . . a minimal narrative is defined as one containing a
single temporal juncture.
(Labov, 1972: 361)
And by temporal juncture, as we have seen, is meant the non-reversibility

of two narrative clauses without change of the original semantic interpre-
tation of the story. Having characterized narrative clauses, Labov and
Waletzky go on to specify the heads of those clauses: these are, as we
might expect, the main clause finite verbs, and usually occur in the simple
past or present. Progressive aspect (he was performing somersaults) and
perfective aspect (he had performed two somersaults) are rare though not
impossible in narrative clauses: such aspectual elaborations are much com-
moner in restricted and, especially, free clauses. Procedures are also out-
lined for isolating the primary sequence of any narrative, a postulated
simplest unmarked order of clauses, but such refinements need not occupy
us here.
We should instead consider now Labov’s essay of 1972, which, as noted,
expands the earlier one, applying the model to actually-occurring narra-
tives from New York Black English vernacular culture. In this later paper
a six-part structure is posited of a fully-formed oral narrative:
1 Abstract: What, in a nutshell, is this story about?

2 Orientation: Who, when, where?
3 Complicating action: What happened and then what happened?
4 Evaluation: So what? How or why is this interesting?
5 Result or resolution: What finally happened?
6 Coda: That’s it, I’ve finished and am ‘bridging’ back to our present
situation.
These are related in Labov’s famous ‘diamond’ picture of the progression

of an oral narrative, which I reproduce below. Notice how evaluation,
while associable with a ‘most expected’ place in the progression of a narra-
tive, namely at the ‘high point’ before the crisis is resolved, is shown – by
the spreading waves – to be something that can permeate the telling.
Evaluation can and does occur anywhere in narratives; in linguistic terms
it is often better to think of it as a prosody rather than a section.
What follows are more detailed commentaries on the six structural
Evaluation
Complicating
action Resolution
Orientation Coda
Abstract
parts or phases of this model. Besides reproducing some of Labov’s own

examples, I will refer extensively to a literary ‘oral narrative of personal
experience’, Mr Casey’s story of the ‘famous spit’, which he tells at the
Christmas dinner table, early in Joyce’s Portrait of the Artist as a Young
Man. Besides being a magnificently crafted story, set with a richness of
highlighted contextualization that seems particularly available in literary
narratives, the story exemplifies the similarities between ordinary narra-
tives and literary ones. Since it is presented as spoken by Mr Casey and yet
known by us to have been written by Joyce, we could argue that as a type
the story lies somewhere between the purely oral narrative and the purely
written one.
6.3 Abstracts and orientations

Together with the coda, the abstract can be thought of as one of the
optional margins of a narrative. They fall not so much at the beginning and
the end of a narrative, as before the beginning and after the end, respec-
tively. In sharpest contrast with them, the Complicating Action is the
phase in which most of a story’s positionally fixed narrative (=sequential)
clauses occur; it is, as it were, the obligatory nucleus. The term ‘nucleus’ is
here taken over from its use in describing the structure of the English syl-
lable: the syllable is standardly said to contain a vowel as obligatory
nucleus, this being optionally flanked by an onset and coda, each of which
comprises one or more consonants. Interestingly, the alternative British
terms for the margins of a syllable, arrest and release, seem applicable to
narratives too: an oral narrative arrests the flow of multi-party talk, but-
tonholing the addressees; with the story’s completion there is a freeing-up
of the suspended turn-taking possibilities and a release, too, from the state
of being absorbed and ‘caught up’, as teller or listener, in an experience of
the narrative.
A typical abstract outlines or advertises the story that a teller intends
will follow. (I choose that form of words, rather than ‘intends to tell’,
because, as will be discussed later, one person can produce an abstract
intending that someone else goes on to tell the fuller story.) An abstract
may then sketch a narrative in a severely abridged form, but it is never a
proper telling of it, as the following examples should indicate:
1 An’ then, three weeks ago I had a fight with this other dude outside.
He got mad ’cause I wouldn’t give him a cigarette. Aint that a bitch?
(Labov, 1972: 356)
2 – Tell me, did I tell you that story about a very famous spit?
– You did not, John, said Mr Dedalus.
– Why then, said Mr Casey, it is a most instructive story.
(Joyce, Portrait of the Artist, 34)
But abstracts also do something else (at least in the view of many socio-
linguists), as example 2 indicates. They often contain requests for the
extended turn at talk necessary to tell a story. The teller’s addressees are
thus politely given the opportunity to indicate their preference not to hear
the story, for whatever reason (they may be pressed for time, or they may
have something very important they want to say first, or they may have
heard the story before). But where an abstract is a bid for the floor among
friends, interacting casually or competitively, it may have few marks of a
polite request and rather more marks of a friendly battle for attention.
This is reflected in a third common characteristic of abstracts, besides
those of story summarizing and requesting telling rights: abstracts are
often advertisements or trailers for stories, making exaggerated claims for
what will follow, promising more than gets delivered. This third character-
istic is noticeably absent both from Labov’s data and from his description
of the element. But that abstracts performing requesting, advertising or
floor-wresting functions are absent from Labov’s corpus is hardly surpris-
ing: typically his informants are supplying stories after they have been
invited to do so, so that for any of them to then request-to-narrate would
be rather odd, indicative of some misunderstanding, or an extreme of
timidity or politeness. In sum, then, abstracts in one respect mark an
exiting of direct interaction in their summary heralding of the monologue
text that is intended to follow, but in another respect they remain interac-
tive in their functions of checking or announcing – or insisting upon – the
tellability in principle of a narrative.
A variant of the normal pattern in which a single teller provides both
abstract and following story is that where a co-conversationalist supplies an
abstract of a story considered worth the telling, so as to prompt or invite
another participant to tell it. This technique may appear in any setting
where at least two of the participants know each other well, e.g. from a
couple at a social gathering. One interesting consequence of the technique is
that worries about tellability are no longer, as usually, directed at one’s
addressees: the prompter has effectively gone on record as asserting the
tellability of a story, and the prompted teller has the luxury of proceeding or
not, depending on whether he or she thinks the story is tellable.
The collaborative story introduction is, however, fraught with uncer-
tainties when compared with the solo story introduction. We might
compare it to the difference, in soccer, between shooting at goal yourself
or passing to a team-mate for them to shoot at goal. Some prompted
tellers dislike being ‘forced’ into telling a story, but are also aware – as is
everyone else – of the awkward ‘misfire’ feel if they decline to proceed.
Such an action puts either the prompter or the prompted in a bad light,
and may be adjudged uncooperative if it is viewed as a refusal to make a
contribution that has been explicitly requested (see Grice, 1975, on the
cooperative principles that may be said to underwrite our construction and
reception of utterances in conversations). Even with cooperative goodwill,
however, a prompted teller may fail to tell a story, or fail to tell it ade-
quately, due to poor recall of the detail of the story events. In such cases
the initiating prompter has made a mistaken judgment as to just how well
their partner remembers the story – a clear case of poor teamwork.
The orientation specifies the participants and circumstances, especially
of place and time, of the narrative, and is equivalent to what was called
setting (not to focalization) in the narratological chapters. Labov notes
that while orientation material can be embedded within opening narrative
clauses, it more commonly comes in a block of free clauses prior to the
development of the narrative action. Or at least we might think of the
position between abstract and complicating action as the unmarked and
logical position for orientation. In the orientation we can expect verb
forms other than extensive verbs (i.e. transitive or intransitive verbs
involving action) in the simple past tense. Thus here we may find past per-
fectives and past progressive verb phrases, and intensive verbs (be,
become, seem, etc.) in relational process clauses specifying attributes and
identities (as described in 4.9).
But the most interesting use of orientation is where components of it
are strategically delayed, and we are told salient facts about the setting
late in a narrative. This, like analepsis, may create effects of considerable
surprise, even of shock. It may be particularly used where the teller
himself, as a participant in a sequence of events, only belatedly learned
some salient facts of a situation and wants to put his audience in a similar
experiential position.
6.4 Evaluation
Evaluation consists of all the means used to establish and sustain the
point, the contextual significance and tellability, or reportability, of a story.
It may take many forms and appear at almost any point in the telling,
although it is often particularly clustered around the ‘hinge’ or climactic
point of the action, just before – and in effect delaying – the resolving
action or event. It is the pre-eminent constituent by means of which the
narrator’s personal involvement in a story is conveyed. In Labov’s words,
it is
the means used by the narrator to indicate the point of the narra-
tive, its raison d’etre: why it was told, and what the narrator is
getting at.
(Labov, 1972: 366)
The Labovian sub-types of evaluation are first distinguished according to

whether they appear inside or outside the fixed-position clauses of narra-
tive. Evaluations appearing outside the narrative clause are of five sub-
types.
1 Wholly external evaluations – as external as you can get; here the nar-
rator breaks the frame of the story-telling itself to address the listener
directly, interrupting the narrative to express a speaker’s current or
still valid general evaluation of the distant events:
It is a most instructive story . . . the story is very short and sweet

... (Joyce: 35, 36)
It was the strangest feeling.
It was quite an experience.
2 Evaluation embedded as a comment reported as made by the teller-

as-participant at the time of the events themselves:
And I said to myself: ‘This is it!’
3 Or embedded as a comment made by the teller-as-participant to

another participant:
And I said to Mary: ‘This is it!’
4 Or embedded as an evaluation coming from another participant:
And Mary said to us: ‘This is it!’
5 Evaluative action: how some participant responded in physical rather

than merely verbal terms to the ongoing events:
I never prayed to God so fast.

I was shakin’ like a leaf.
The first of these last two examples, it may be noted in passing, reflects
one set of cultural assumptions, in which prayer is implicitly classified as a
kind of evaluating rather than a kind of doing. Prayer here is not con-
ceived of as narrative action, occurring at a fixed point in a sequence of
causally-linked events. Devout believers might see things differently, inter-
preting any happy outcome that followed the prayers as caused by (the
deity supplicated in) those prayers.
All these modes of evaluation involve a temporary suspension of the
action, a brief ‘time out’ from the telling of the story proper. When well
placed, such manoeuvres do indeed create suspense, set apart whatever
narrative follows (often the concluding resolution), and heighten the lis-
tener’s interest. But all such external suspensions can be contrasted with
the narrative-clause-internal modes of evaluation to which I will turn in
Section 6.6. Before doing so, a few observations on a central assumption of
Labovian method are necessary.
6.5 Doing and saying

Labov and Waletzky studied ‘danger of death’ narratives. There they
found, and thereafter adopted as a working assumption, that what is said
(by teller or others) will not be the core of the story; rather, the narrative
core is occupied by what is done. The ‘what is done’ then becomes the core
narrative text of clauses – actions – while the ‘what is said’ becomes evalu-
ative commentary on those actions. Now this pattern is common enough to
be a reasonable assumption in many cases, and one can immediately see
why ‘danger of death’ stories of the sort Labov elicited would promote
identification of a sharp division between salient actions such as physical
assaults and accidents on the one hand, and verbal reactions on the other.
But more complicated integrations of words and actions are also possible,
where the sayings are the most important doings – are the ‘action’ of the
narrative – revealing a fixity of sequence of those sayings, temporal junc-
ture, and so on. From the work of the philosopher J. L. Austin has
developed a renewed recognition that our use of words in interaction is
typically a performing of actions and not merely an asserting of true or
false (hence evaluative) statements. Labov’s own informants demonstrate
this vividly in other interactions that he analyzes (1972): as their enhanced
skills of duelling, rapping and sounding make very clear, there can be
verbal contests (where participants are in ‘danger of loss of social pres-
tige’) as well as physical ones.
But no narrative of sayings, you might complain, can break your bones
or put you in danger of death. Again, however, we might question this
assumption. There are too many recorded cases of individuals placed in a
position of great danger by their own or others’ verbal acts, which some
violent party finds provocative or threatening. Or someone comes to the
brink of suicide or murder, incited by acts of speech, and then is carefully
‘talked down’ by individuals who are themselves at risk. Sayings may
indeed put someone in danger of death. As they can, also, in a murder trial
where the oral testimony of a key witness might be decisive. And in our
increasingly textual worlds, where growing numbers of us make our living
through our verbal work rather than our non-verbal actions, we may predict
that ‘what is said/written/promised/denied/argued, etc.’ will more commonly
constitute the essential complicating material in narratives.
Not least in literature: in the ‘famous spit’ story of Mr Casey, for
example, the ‘doing’ that makes up the complicating action to be evalu-
ated is a verbal ‘doing’: the old woman’s series of increasingly offensive
insults. And the terminate or resolutory action of spitting itself is
reported less as a doing (Casey does not announce ‘I spat at her’) than as
a saying:
– And what did you do, John?

– Do! said Mr Casey. She stuck her ugly old face up at me when she
said it and I had my mouth full of tobacco juice. I bent down to her
and Phth! says I to her like that.
He turned aside and made the act of spitting.
– Phth! says I to her like that, right into her eye.
Casey’s Phth! has some of the marks of external-evaluation of type 3, but we

know it is also a crucial sequentially-fixed action: an inextricable merging of
evaluation and narrative action, of saying and doing. Similarly we can
ponder the old woman’s reaction to this verbal-cum-non-verbal action:
He clapped his hand to his eye and gave a hoarse scream of pain.
– O Jesus, Mary and Joseph! says she. I’m blinded! I’m blinded and
drownded!
We infer, from Mr Casey’s accompanying gesture of clapping his hand to

his eye, that the old woman reacted in other ways than simply saying these
things. But again, the woman’s verbal response is itself very important to
the story: the unpleasant physical consequences of the spit are far less
to the point, but in any event do not include literal blinding or drowning.
The tellable reaction is the verbal one, directly mimicked, with its non-
standardism, its ludicrous hyperbole, and the irony of its probably profane
invocation coming from the lips of this self-declared defender of the faith
and guardian of morality.
The above unravelling of the narrative/evaluative dichotomy is broaching
a very important topic, for in Labov as in other fabula/sjuzhet theorists the
separability of the plot from surrounding or interpolated discoursal elabora-
tions and recastings is an operational necessity. And in working with that
dichotomy, all our language reflects an assumption that plot is core, that the
clauses of narrative action are the heart of the matter, the inner narrative,
and that evaluation is to a degree external, and always intrusive. Now that
assumption is one I am prepared to accept, with the caveat voiced above
(that core narrative clauses may be ones of speech rather than action, and
look very like evaluations) and a further caveat I will now introduce. This is
that adopting the categories of fixed narrative clause and freer evaluative
clause is not in itself a claim about the ‘status in reality’ of the actions and
sayings these report. In particular, it is not an assumption that material pre-
sented as main-clause narration ‘really did’ happen the way the narrator
claims. Put thus the warning seems obvious, that narrators ‘aestheticize’
their experiences, assert cause-and-effect chains where no chains are there,
and so on; but the point is well made by Culler (1981: 184–5); see also the
related point in Hoey (2000: 105), that happenings do not exist as structured
forms, prior to their telling. If we imagine that first we have the sequence of
actions (narrative) and then we work on the reporting of them to enhance
their point or tellability (evaluation), we are ignoring the possibility of a
reverse order of impulses, namely that, guided by the prior awareness of the
tellability-requirement, our evaluations shape our plots.
6.6 Internal evaluation

Internal evaluations, as noted, are those evaluations woven into the struc-
ture and composition of the core Narrative clauses – those clauses, with
their dependents, that carry the core sequential ‘spine’ of the narrative,
filling the Complicating Action and Resolution sections. Such evaluations
are grouped by Labov into four sub-types:
1 Intensifiers: here are included:

gestures (often accompanied by deictic that),
He turned aside and made the act of spitting. . . . He clapped his

hand to his eye.
expressive phonology, onomatopoeic sounds, and intonational

emphases:
Phth!; ‘I’m drownded!’
exaggerating quantifiers such as all,
such booing and baaing; paid all her attention to me; right into her
eye
repetitions,
I let her bawl away [reported twice]; Phth! says I to her like that
[reported twice];
and ritual utterances.

Unlike the following three types, Intensifiers do not significantly
complicate the syntax of the narrative clause. They are clearly a kind
of embellishment, as their name would suggest, akin to the use of
underlining and bold face in writing.
2 Comparators. Why, in the telling of what happened, would you
mention various things which did not happen? That is the question
that Comparators invite. For while Intensifiers evaluate directly,
telling what happened in an exaggerated or elaborated way, Compara-
tors evaluate indirectly, by drawing attention away from what actually
happened by alluding to what might have but didn’t happen. The main
types of comparators include:
expressions of negation,
You never heard such booing and baaing; I couldn’t say a word;
modality and modulation,
We had to make our way to the railway station;
and futurity:
I won’t sully this Christmas board.
All of these are concentrated in, but not confined to, the auxiliary
verbal elements; also included are questions, hypothetical claims,
imperatives, and, most overtly, comparative or superlative construc-
tions. They involve an indirect evaluative departure, in the lexi-
cogrammar, from the simple direct telling of the narrative actions. A
more complex departure is the use of simile or metaphor, when these
occur in narrative clauses (we may surmise that they would be more
common in free clauses). And they all have the effect of introducing a
branch or fork, into an otherwise simple or un-branched point in the
telling; every comparator evokes an alternative narrative development
which becomes background or relief or illuminating contrast to the
actual narrative development that proceeds.
3 Correlatives: these bring together events in a single independent
clause, and require complex syntax (and are hence often beyond the
control of young narrators):
progressives (be ⫹ V-ing), and appended participles (adjacent verbs in
non-finite V-ing form), both of these emphasizing simultaneity of dis-
tinct actions,
She kept dancing along beside me in the mud bawling and scream-
ing.
double appositives,
there was one old lady, and a drunken old harridan she was surely
double attributes,
a drunken old harridan
and ‘left-hand’ participles or ‘deverbal’ adjectives,
an unsavoury-looking character.
4 Explicatives: These are appended subordinate clauses which qualify,
or give reasons for, the main events reported: for example, clauses
introduced by while, although, since, because, and so on:
sure I couldn’t say a word in any case because my mouth was full
of tobacco juice.
Such clauses enhance tellability by more fully specifying the extent or

motivation for a particular action. In a structural sense they are strictly
inessential to the telling of a narrative’s fundamental sequenced
events; but in the reality of interaction, they are virtually indispens-
able: rarely are they completely absent from oral narratives of per-
sonal experience, in western cultures at least.
6.7 Coda
As indicated earlier, codas and abstracts can be seen as related. A coda
signals the sealing off of a narrative, just as an abstract announces the
opening up of one. There seem to be two most common devices within
codas. One is the explicit declaring that the narrative proper is over, so
that for an addressee now to ask, ‘And then what happened?’ should be
absurd (or demoralizing: it indicates they did not ‘get the point’). The
element is often realized by a near-redundant narrative-external comment,
using pro-forms that are both textually anaphoric (pointing backwards to
earlier co-text for interpretation) and distancing in their deixis:
1 And that is the end of the story.

2 And that was that.
3 And that – that was it, you know.
In versions 1 and 3 the item that, it can be assumed, is cohesively tied to

some previous text in which a state or resolution has been reported; but it
is also deictically pointing to that resolution, and now locating it at a dis-
tance from the speaker and her current position (that, not this or here), no
matter how vivid and immediate parts of the telling of the narrative might
have been.
This brings us to the second device common in codas. As the above
indicates, codas are commonly the site of a deictic shift, especially in the
more involving narratives of personal experience. In telling such narra-
tives, the teller who is also a principal participant often switches the deictic
anchorage to the spatiotemporal orientation of himself-as-participant,
selecting items such as this, here and now relative to that individual, rather
than relative to the currently present one telling the story to an addressee.
But it would make no sense, once the story was finished and the inter-
action was in the process of returning to conversational mode, in ‘real’
present time, to persist in using present-time deictics with reference to the
past story. Putting this another way, the teller seems best advised to signal,
before the close of his long narrative turn, that he has exited from the
marked past narrative to present deictics mode, and that all can resume
normal use of present deictics to designate relatedness to the present
context of situation. Accordingly, tellers often do signal such a switch in
their codas:
4 And ever since then I haven’t seen the guy ’cause I quit, I quit, you
know. No more problems.
5 And you know that man . . . is a detective in Union City and I see him
every now and again.
These use distal deixis (the or that) now to denote elements of the story,
and present tense (with or without perfective, ‘current relevance’, aspect)
to denote the shared current time of the speaker-as-conversation-partici-
pant, no longer speaker-as-narrative-participant. Labov talks of codas as
having
the property of bridging the gap between the moment of time at the
end of the narrative proper and the present. They bring the narrator
and the listener back to the point at which they entered the narrative.
(Labov, 1972: 365)
Codas bridge a constructed gap, mutually agreed upon in the abstract,

which creates a distinctive space for the narrative to occupy. Relatedly,
we may look for a counterpart in reverse, near the outset of narratives,
of the coda’s deictic switch or bridging. This does seem to occur fairly
often, although it is centred more in the orientation than the abstract.
Thus abstract 1 in Section 6.3 alerts the addressee to orientational switch
by alluding to ‘this other dude’ (cf. ‘another dude’ or ‘some other dude’),
and by marking the temporal gap from the present (‘three weeks ago’).
While abstract 2 does not perform an orientational switch, Mr Casey’s
very next turn, which is both orientational and externally evaluative,
does:
– Why then, said Mr Casey, it is a most instructive story. It happened
not long ago in the county Wicklow where we are now.
(Joyce, Portrait of the Artist: 34)
The seeming redundancies of the second sentence here should give us

pause. Why, we might ask, doesn’t Casey simply say:
It happened here not long ago.
or at the most,
It happened not long ago here in the county Wicklow?
Is the loosely-appended adverbial, where we are now, tellable material at

all? Ordinarily not, but here perhaps the sentence as a whole is a nicely-
judged establishment of both the story’s detachment from the present and
its heightened current relevance (the time was ‘not long ago’, the place
was hereabouts). Casey is a cunning storyteller, and this story’s underlying
purpose is to suggest that the foul-mouthed old woman who denigrated
Parnell within the story, and Dante, who is one of its present addressees
and has similarly condemned Mr Casey’s political hero, are analogues,
meriting analogous treatment (to be spat upon). The seeming redundan-
cies of Casey’s introductory turn highlight the situational contiguities just
as, later, the old woman’s heaping of abuse will be reported as if parallel to
Dante’s censure.
But having dwelt at length on the form and function of codas, it should
now be stressed that in an oral personal narrative a coda may not appear.
This is especially so in narratives told in the course of a single conversation
where sequences of stories may get told ‘back to back’ by one teller or
several. In such sequences it may happen that some aspect of the resolu-
tion of one story serves, in some direct or remote way, as the trigger of the
next. The associative links that particular tellers construct, so as to
enhance a story’s sense of being ‘locally occasioned’ and ‘sequentially
implicative’ (Jefferson, 1978) are immensely varied.
6.8 Stories in societies

Another approach to evaluation in stories, less directly attentive to linguis-
tic form than Labov and more interested in community-wide motivations,
comes in the work of Polanyi (1978; 1981). She has argued that the kinds
of things that we seem to agree make stories tellable reflect and disclose
our cultural presuppositions and values. Similarly, Tannen (1979) demon-
strates in some detail how our ‘structures of expectation’ based on past
experience influence the particular ways we construct our stories, and
interpret those of others. This in turn would predict that the kinds of
stories that get told, and are valued, in one cultural milieu may differ quite
considerably from those that get told in another. And more recent studies
by Tannen, on both ethnic-based and gender-based differences in ‘horizon
of expectation’ – concerning narrative construction and other verbal
behaviours besides – confirm this (Tannen, 1991; 1994).
One very influential study of this kind was reported in Heath (1983), a
comparative study of language use in two small-town working-class
communities in the rural south-east of the United States, one black
(Trackton), the other white (Roadville). These two communities reveal
quite different views of stories and storytelling, as Romaine summarizes:
In Roadville stories stick to the truth and are factual. They maintain
strict chronicity, end with a summary statement or moral, and serve
the function of maintaining values and reaffirming group membership.
Any fictionalized account is a lie. Trackton stories, on the other hand,
are hardly ever serious. The best stories are ‘junk’, and the best story-
tellers are those who can ‘talk the best junk’, i.e. make the most wildly
exaggerated comparisons and tell outlandish fictional narratives.
(Romaine, 1985: 102–3)
And as Heath (1983: 189) concluded, of these radically diverging norms of

behaviour:
For Roadville, Trackton’s stories would be lies; for Trackton, Road-

ville’s stories would not even count as stories.
In the face of Heath’s findings, which are supported by extensive data, it

should be evident that reading off cultural values or ideology from the
points that a community’s stories have will not be easy, precisely because a
community’s cultural presuppositions and values are often, being presup-
positions, not explicitly stated or rehearsed in everyday cultural activity.
The ‘said’ of a culture is deeply embedded in its ‘unsaid’ (Tyler, 1978). As
the Trackton–Roadville study confirmed, it is not the case that everything
else besides a story’s main evaluated point is kept constant across
communities. It may well be that communities have different kinds of
story points because they have different perspectives on the proper func-
tions and nature of storytelling. If that is so, we are driven back from the
easier task of correlating delimited story points with possible cultural
values, to the harder task of holistically assessing all of a community’s ten-
dencies in narrative use in relation to inferred cultural values.
6.9 Narrative performance

Over the past few years a number of sociolinguistic studies have appeared
using the folkloristic term performance to describe a certain type of
particularly involved and dramatized oral narrative. The term has been
adopted and espoused despite resistance from some traditional folklorists,
for whom the term standardly denotes the retelling of stories that are part
of a tradition, are collectively known, and are non-innovative.
The idea that narratives are often performances seems to have first
emerged in the work of Hymes and Goffman, and received fuller sociolin-
guistic exposition in Wolfson (1982). Wolfson’s work is most directly relat-
able to Labov’s; but she makes reference to a larger corpus of stories, told
in a greater variety of contexts (of situation, formality, and of familiarity,
age, sex and status of the tellers and addressees, and so on). Working with
this set of possible variables, Wolfson probes the varying formal reflexes of
individuals’ experiential involvement in their storytellings as dramatized
re-enactments:
When a speaker acts out a story, as if to give his audience the

opportunity to experience the event and his evaluation of it, he may
be said to be giving a performance.
(Wolfson, 1982: 24)
To perform a story is to furnish one’s addressees with a more vivid and

involving experience of that story, while exploiting special performance
features as resources for highlighting the story’s main point. The perfor-
mance features Wolfson particularly singled out for their instrumentality
were the following:
1 direct speech
2 asides
3 repetition
4 expressive sounds
5 sound effects
6 motions and gestures
7 conversational historic present (CHP), alternating with narrative past
tense.
Roughly, the more your story has of all of these, the more ‘performed’ it
is. Most of these items are already implicit in some part of Labov’s typol-
ogy of evaluation devices; only items 2 and 7 are noticeably new. I will
comment briefly on both. In an aside a teller exits briefly from the time
reference of the story dialogue he is recounting in order to add some
comment about the content which has continued relevance in the present
time of the teller and listeners. Such comments highlight, for the listeners,
the teller’s own viewpoint and sympathies with regard to the interrupted
dialogue. Thus asides are an interruption of direct speech, and are there-
fore somewhat related to that mode of evaluation; they are also somewhat
like external evaluation, with time reference to the present time of the
teller-as-conversationalist; while in their explanatory function they are
rather like Labov’s explicatives.
The most innovative section of Wolfson (1982) is the account of the use
of historic present tense in storytelling. Wolfson contests the traditional
grammarian’s view that historic present, in conversation-embedded narra-
tives, is there simply to make a story more vivid and immediate. Rather,
the historic present when used in conversation-embedded narration has a
distinct function – hence the emphasis on conversational historic present –
and that function relates not directly to the occurrence of CHP itself, but
to the patterns of switching between historic past and present tense that it
permits. The switches or alternations are said to facilitate three main
effects: (a) an intervention by the teller; (b) a focussing of attention on
certain portions of the narrative; and (c) a dividing-up of the flow of action
into distinct events. Relatedly, Schiffrin (1981), found historic present
tense almost wholly confined to the complicating action section of per-
sonal narratives (significant because this is precisely the section of stories
where tense does not have to do any special task of temporal orientation,
since temporal and presentational order can be assumed to be congruent).
So CHP is emphatically an internal evaluation device (Schiffrin, 1981: 58).
Perhaps unsurprisingly, Wolfson (1982) also found that stories were
more likely to be performed (not merely told) when conversationalists
were of the same sex, of similar age and status, similar in attitudes and
background, and so on. In principle, at least, what Wolfson did for per-
formedness – probing what kind of conversationalists in what kind of rela-
tion opted to perform their narratives – could be done for the whole range
of Labovian evaluation devices. But a major difference would then
emerge, for evaluation undertakes a larger, more crucial task, than that of
performance: the task of articulating the point of a story and persuading
the audience of its tellability. Thus while performed stories give the
impression that they contain a number of extra evaluative features added
to stories told between intimates but not between people remote from
each other in various ways, a reverse trend may apply when we compare
the amount of evaluation in the stories of mature tellers. In other words, it
may be the case that the fewer the dimensions of similarity between con-
versationalists, the greater the role that evaluation devices may have to
play, in spelling out the significance of a tale for a teller to a listener who
does not share unspoken understanding.
6.10 Dispersed, embedded, and group oral narratives

There are other types of personal narrative, more subtly interleaved with
the ongoing conversation, than those Labov extracted. Polanyi, for
example, draws attention to the ‘diffuse story’ (1978: 109ff.), in which a
chunk of story is followed by a chunk of multi-party conversation glossing,
clarifying and amplifying aspects of the story chunk just told. In such a
format, story evaluation can become a collaborative exercise. Diffuse story
format shades into practices of embedding where, for example, several
short stories are spread out and interrupted by conversational interludes,
but can also be treated as the several sections of a single, overarching
story.
Uppal (1984) has just such an example, a story comprising four conver-
sationally embedded stories (I am most grateful to Ms Uppal for permis-
sion to reproduce data from her study). The stories emerged in the casual
conversation of a mixed-sex group of Singaporean university students,
good friends, reminiscing about their schooldays, sitting around a table in
a college canteen. The numbers (1) to (4) in the lefthand margin mark, in
my analysis, the onset of each of the four embedded stories within the
ongoing conversation. Text in square brackets is my glossing of utterances
or explanation of the ongoing interaction; I have provided only minimal
‘translation’ of the many features of colloquial Singaporean English in the
teller’s speech. At the opening of the transcript, S is referring to his time in
Sixth Form, the final two years of high or secondary school (here the
classes or grades are called 6A and 6B), where less academically successful
students are sometimes required to repeat a year of schooling once or
even, here, twice.
(1) S: My one ah- my one ah- I was saved by my intelligence y’know.

Because ah- I was in 6A, y’know. Eight of us were caught gam-
bling, y’know. Seven 6-repeat-2. That time they got 6-repeat-once,
repeat-2 y’know. Repeating the second time repeating the first
time. Seven 6-repeat-2. Then 6A. All public caning. Because the
teacher caught us gambling y’know. So all public caning. Some
more [= what is more] I prefect y’know. Kena [got (⫹ Verb)en,
caused to happen] lost badge, all. Everything gone nah, koyak
[spoilt, broken]. So kena, er, public caning. Some more I gabrah
[make/made a mess of things], boy. Some more public is like, prin-
cipal office no sweat, y’know, but public is like they call your
parents round.
V: Huh?
S: Ya, so parents will stand by the side, know.
G: Ayooo!
[some dialogue ellipted]
(2) S: Because, because it seems that you know is so- the son something
goes wrong, parents have to be notified y’see. So they give public
caning. Then- some more standard y’know. Okay next, come ah.
Fellow’ll squeeze your bums all, know. Because beginning stages I
think someone put book all. /katabah/ Got sound, y’know.
G: [uninterpretable]
S: So the fellow he squeeze squeeze your bum all. He make sure soft,
know # Cushion # You see funny funny things ah. The fellow is,
ah, the senior assistant is the one who canes nah, not the principal.
Principal was a woman nah, so senior assistant. He feel feel. Then
he smile at the school, y’know. Then he take out wallet. Beautiful
wallet # Three handkerchiefs [i.e. wallet and handkerchiefs were
intended to protect the boy’s backside]. Take out all. One by – ((S
makes lashing sound)) Wahh Lan [Damn!]. I tell you cry, dah. 6-
repeat boys crying, dah. 6-repeat that time you respect. Wah, that
time the terrors of the school.
B: This one your grandfather stories, ah?
S: Public caning terok [severe], y’know. I tell you.
V: Is it really that painful or because of the ( ).
H: Pain boy.
S: Damn pain na, I tell you. Damn pain. I never kena but I know
damn pain. See boys cry, I know damn pain. The canes also got
different ones, know. You know Kim Seng.
(3) I escaped because I was in 6A. The principal came. He saw- he saw
all 6-repeat. Then he called down my teacher. ‘Explain why your
6A boy is doing with the 6-repeat student.’ The teacher got no
explanation. So he said, ‘I want you to account for his behaviour.’
N: Like real, ah?
S: Then I excused y’know, I excuse. Go back. My teacher to account
for my behaviour for the rest of the year lah. So I kena [got to, had
to] sit sit in front. Smile smile everyday.
N: So unfair, know. How come you never kena [sc. got caned]?
S: Exactly. I kena- after the six boys moody already.
N: Ya boy.
H: No members [= friends] na, alamak.
N: Some more, never mind. They are the terrors of the school,
y’know you go home nah.
(4) S: But they- they’re okay nah. The reason I got to know them was
badminton. You see, I played badminton for school, y’know. Then
the whole school team formed by 6-repeat students, except I. Only
6A feller playing badminton for school. So slowly lah. Slowly got
into the ideas of life lah. Then ah you know school, afternoon
school. Twelve o’clock come. What do you do?=
G: =True=
S: All first started with the Saturday, know, play Saturday. Carry on.
Then slowly switched to five cents nah.
G: ((/)) [backchannel acknowledgement noises]
S: Then slowly slowly switched to ten cents, twenty cents. Then after
that ting-tong already.
G: Mm.
S: Move lah, advance. Ting-tong, kena sway sway [had some bad
luck, a catastrophe]. Cards only kena caught. Moody. Advance
through what.
G: ((/)) [backchannel acknowledgement noises]
S: Actually we tell them, know. We say, ‘Sir, this one second time
only.’ Tell them, ‘Gambling playing for fun.’
G: Mm.
S: Actually six months ago we started already.
Story 1: How the teller, with others, once contemplated a public caning
for gambling.
Story 2: How public caning at school was an ordeal, involving removal of
padding/protection, and painful.
Story 3: How the teller escaped caning on the condition that his teacher
could thereafter account for his behaviour to the principal.
Story 4: How the teller, a 6A boy, got in with the 6-repeat boys, and
their gambling.
There are ways in which these are four largely free-standing narratives, as
the summaries immediately above suggest. But there are respects too in
which these are four episodes or components within a single elaborated
narrative: story 1 supplies an abstract, orientation, and some complicating
action, story 2 adds further orientation and evaluative description, story 3
contains the resolution, while story 4 is a flashback explaining how the
situation at the beginning of the complicating action was arrived at. But
each of these stories is itself adequately formed, with temporally-ordered
action and resolution, and all four come from a single speaker.
To strengthen the claim that they are not merely a ribbon of loosely-
related stories but rather also constitute a single ‘macro-story’, we should
at least require some degree of temporal juncture between these stories,
and further evidence that these four parts all cohere around a specific
topic and resolution. Here the main story we might call a ‘danger of
caning’ one: the teller recalls how, when he, a model student, was caught
at school gambling for money with boys from a ‘repeat’ class, he alone
escaped the pain and humiliation of a public caning. The main story reso-
lution is as follows:
Then I excused y’know, I excuse. Go back. My teacher to account for

my behaviour for the rest of the year lah. So I kena sit sit in front.
Smile smile everyday.
This appears towards the close of subordinate story 3. Prior to it, story 1
perfunctorily reports the fact that the boys were caught gambling, and the
minor actual resolution (the teller is stripped of his prefect’s badge) plus
the possibility of a more major resolution (public caning). But most promi-
nent in story 1 is emphasis on the crucial orientational facts that the other
boys are in a repeat stream, distrusted hard cases, while the teller was in
the prestige fast stream. Story 2 tells the embedded story of public canings
in general, and the caning of his co-gamblers in particular, all the more
vivid and hilarious in colloquial Singaporean English.
It is clear that the public caning reported in 2 must have been preceded
by the conditional letting-off of the teller-as-participant reported in 3, and
that orderliness of event recapitulation has not been maintained (the order
of actual occurrence of events reported in the stories being 4^1^3^2; that
is, Gambling^Detection^Reprieve^Caning). On the other hand we may
still maintain that here is a well-formed macro-story if we classify story 2
as evaluative within the larger context, an extended comparator exploring
the path of punishment that, to the teller’s relief, was not taken in his case.
As noted earlier, the main point of this macro-story is then not what hap-
pened to the other boys, but how this boy got away lightly – which indeed
is how S announces his narrative at the outset. Story 2 effectively height-
ens our interest in just what treatment the teller will receive, artfully sus-
pending the action before that resolution.
We may notice, too, even from the sparse transcript above, the different
roles taken up by different addressees of S’s extended telling. G is a sup-
portive backchanneller, as less prominently are H and N. B on the other
hand is briefly sceptical, but doesn’t pursue the doubt about the
veracity/credibility of the story; but not insignificantly, B’s challenge is
immediately followed by V’s, a serious challenge to the story’s tellability
(is public caning truly painful?). That challenge is comprehensively
crushed, with S telling four times over that (though he himself never got
caned) the pain is ‘damn pain’.
If the above is accepted as a type of dispersed story, the more common
phenomenon is for stories to relate more directly to surrounding conver-
sational discourse and prior stories in noticeably relevant ways. Usually
the onus is on conversationalists as individuals to insert their stories
appropriately, showing attentiveness to what others have said, and so
on. But sometimes group stories are told, by all or several conversational-
ist–participants working together: a narrating that does not involve sus-
pension of the turn-taking mechanisms. At first glance the group story
appears to be an admirably egalitarian and collaborative storytelling
mode, built on a heightened degree of interdependence. The division
between tellers and listeners falls away as does anxiety over tellability,
and a spirit of benevolent mutual indulgence may prevail: all are con-
tributing to a story that each already knows. But, in some ways, and with
more scope than usual, a competitive rivalry can also shape the several
contributions to the telling: there may not be full agreement on the point,
on whom among the teller-participants comes out of the story well, and so
on. And since there will be general familiarity with the broad outline of
the story in question, the exercise may become more oral-literary, with
great attention paid to the most effective and entertaining methods of
verbal expression.
As Polanyi notes (1978: 143) the group story is an area where greater
fictionality may be tolerated; credibility or vraisemblance is not a prime
concern, since focus is less on any particular teller than on the tale for the
tale’s own sake. Or, as Bennett (1983) notes of her related story type, the
group saga, the retelling is for purposes of reminiscence and social
binding, with the emphasis less on the incidents themselves than on paint-
ing a rich picture of the situation. Alternatively, as Uppal (1984) notes, a
group story may become more a series of hypotheses or conjectures about
what might have been the case (the true resolution, the proper evaluation)
concerning some set of events, particularly where all the group-tellers
were detached witnesses of those events, with limited inside knowledge.
And the more that a group’s talk becomes an unordered set of overlapping
conjectures, the more the talk returns from narrative mode to that of
ordinary conversation or gossip. Putting things the other way around, it
may be that a collective style of telling is more frequently adopted where
there is individual and collective uncertainty about just what happened
and just how things ended up: each in the group helps to piece the story
together.
Along these lines, we may see crime and detective fiction as a literary
genre partly built upon the group story principle. When P. D. James’s
Superintendent Dalglish or Sue Grafton’s Kinsey Milhone piece together
the testimony of a range of individuals connected in some way to a
crime, sleuth and reader must process a radically dispersed and potentially
defective group story. The dispersal stems not merely from the fact that
there is usually no Aristotelian unity of time or place as far as the charac-
ters’ contributions (within the story) are concerned, but more directly
from the fact that in crime stories the guilty party, often along with other
investigated individuals who have information they wish to remain hidden,
will contribute misleadingly rather than sincerely cooperatively: any
implicit declarations, by all parties, that they are collaboratively intent on
helping the investigator to ‘get the true and full story’ told have to be
treated as unreliable. Defectiveness and contradiction are probabilities
since one or more of the group may construct their contributions so as to
conceal their own guilt: fictionality with a vengeance, for self-protective
and not merely aesthetic purposes.
6.11 From Labov to literature

The extent to which the Labovian six-part formalist analysis of the oral nar-
rative of personal experience applies or is relevant to literary narratives has
also become a matter of some contention. This section will review some of
those matters of contention, and conclude with the suggestion that rele-
vance, rather than direct application, is the more helpful final emphasis.
In Pratt (1977), a substantial chapter is devoted to exploring some of
the ways in which Labovian analysis could be extended and applied to lit-
erary fiction – ‘written narratives of invented others’ experiences’. Numer-
ous examples are given of novels and short stories revealing various forms
of abstracts, orientations, and so on. Of course the sheer scale of a novel
makes for huge differences of magnitude of these narrative components by
comparison with a brief oral narrative. Thus Pratt notes that Hardy’s The
Return of the Native opens with an entire chapter of orientation, and really
only a small part of orientation at that: a description of the Egdon Heath
landscape on a particular November afternoon. Pratt adds:
The human characters and complicating action do not arrive until the
second chapter, aptly titled ‘Humanity Appears Upon the Scene,
Hand in Hand with Trouble.’
(Pratt, 1977: 53)
Furthermore, as Pratt notes, all sorts of departures from the ‘underlying’

unmarked or canonical narrative format, with adequate abstract, orienta-
tion, and so on, are to be found in literary narratives. Nor, in the face of
such creative departures from the standard format, are we inclined imme-
diately to reject the text as a defective narrative: we are used to paying
more attention, working harder, being more challenged by literary texts
than non-literary ones. Thus many novel- and story-openings project us in
medias res, into the middle of ongoing events, with deliberately insufficient
orientational briefing, as if we were already familiar with the world and
actors depicted. Cast into this position, we sometimes feel we indeed
become familiar with the world presented very rapidly; on other occasions,
however, we remain somewhat disoriented throughout, and effects of
mystery, disorder, obscurity, and lostness, are achieved (intentionally or
otherwise). Texts by Kafka, Beckett and Pinter are among those that
immediately come to mind.
The broad point Pratt makes, by further copious exemplification of
forms of resolution, abstract and coda in literary narratives, is that no
radical gulf separates literary from ‘everyday’ narrative: the same devices,
used for the same purposes, emerge in both.
What is important about the fact that literary narratives can be ana-
lyzed in the same way as the short anecdotes scattered throughout our
conversation? To begin with, it casts grave doubt on the Formalist and
structuralist claims that the language of literature is formally and func-
tionally distinctive. . . . Unless we are foolish enough to claim that
people organize their oral anecdotes around patterns they learn from
reading literature, we are obliged to draw the more obvious conclu-
sion that the formal similarities between natural narrative and literary
narrative derive from the fact that at some level of analysis they are
utterances of the same type.
(Pratt, 1977: 67, 69)
At this distance those last claims seem overstated, a hinting at some ‘deep
structural’ human rules for storytelling (applied in superficially different
ways in conversation and in novel-writing) that subsequent studies have
not endorsed. The implied picture is vertically or hierarchically-oriented:
deep structural formats ‘surface’ in oral or literary narratives. Currently
theorists tend more to subscribe to a horizontally-oriented picture, in
which nineteenth-century novels, modern short stories and contemporary
personal oral narratives are generically separated, like languages or cul-
tures across a continent, with no ‘governing’ or derivational relations
between them, and plenty of scope for borrowing, appropriation, and
other hybridities. Within this more Bakhtinian picture of intertextual and
cross-genre clashings and mergings, it is not at all implausible to suggest
that people sometimes introduce into other parts of their everyday lives,
including their oral storytelling, forms and patterns they have learnt from
reading literature. As we have seen, Hart has suggested that this is how
Eveline comes to use pulp-romance-fiction language in the Dubliners story
(see Notes and exercises 1, in the previous chapter).
Since Pratt’s study, narratologists have differed over the usefulness to
narrative poetics of the Labovian approach. Chatman (1990), for example,
is sceptical about the ‘contextualism’ that he sees Labov’s and similar
studies harbouring. He suggests for instance that evaluation in literary
narrative is rather different from that in personal storytelling:
The insistence that every narrative include an evaluation of the

experience narrated seems to fly in the face of the efforts by novelists
since James and Flaubert to eliminate judgement and other commen-
tary from the narrator’s pronouncements.
(1990: 317–18, fn. 9)
And as noted in this chapter, everyone agrees that oral personal narra-
tive’s iconic chronology is radically different from the repertoire of
anachronies to be found in contemporary artistic fiction and film where,
arguably, they are so common a feature as to have become the unmarked
norm. And what of Free Indirect Discourse: where the teller is also pro-
tagonist, narrating in praesentia, as in Labovian narratives, what need or
scope is there for this complex device? Character and setting, too, are
radically minimized in oral personal narratives when compared with
typical fictional stories, although here we should not overlook the way
some writers on occasion supply even less in the way of setting or charac-
terization than an oral teller. Nor should we overlook the extent to which
all these differences may be artefacts of the profoundly different contexts
of production and consumption in which the novel and the conversation-
ally-embedded personal narrative are situated.
In view of these manifold differences, many students of narrative have
questioned whether any meaningful rapprochement between Labovian
analyses and narratology remains possible. But on the other hand a
number of relations and continuities remain. Fleischman, for example,
argues that literary-narrative anachrony remains grounded in cultural and
common sense chronology:
If our cultural and literary competence did not include a narrative

norm, one component of which is iconic sequence, then the
anachronies of literary and cinematic narrative could not produce the
effects they do on readers and spectators.
(1997: 164)
For Fleischman, too, the palpable differences between robust Labovian
Resolution and the kind of unresolved or open-ended terminations of
modern novels, which may even introduce further complicating action and
orientation, is not beyond explanation. It is not so much that modern nar-
rative art has ‘abandoned narrativity’ as that, under various cultural pres-
sures, it has developed or changed quite substantially. Besides, it is a
mistake to neglect the extent to which tellers and addressees continue to
crave or orient to The Sense of an Ending, to quote the title of an import-
ant book on this topic (Kermode, 1967). Of most modern artistic films and
novels it is still appropriate to ask someone who has seen or read them,
‘So what finally happens?’. More generally, sequentiality remains a defin-
ing property of narrative fictions, ancient and modern, even when these
choose not to affirm that sequentiality in the simplest chronology-iconic
narration.
Equally comparatively complex, and thus different from Labovian nar-
ratives in degree rather than in kind, are the Evaluation resources of liter-
ary narrative. Let us begin with Chatman’s observation, quoted above, to
the effect that Jamesian fiction with its impersonal non-arbitrating narra-
tion, is designed to produce evaluation-free narrative. In a sense this is
surely true, but a number of factors might be added to the account:
1 Almost no-one comes away from a James (or Flaubert, etc.) novel or
story complaining that the narrative had no point, or that it wasn’t
worth telling; some readers may react with a, ‘so complex!’, but few if
any ask, ‘So what?’. Jamesian fiction fulfils the Evaluation function or
rationale somehow, even if not in ways immediately akin to those
identified in oral personal narratives.
2 In terms of Simpson’s nine-mode schema, while James and Flaubert
both adopt B(R) narration, only Flaubert is characterized as typically
neutral, while James is said to standardly adopt B(R) Positive mode. It
seems reasonable to suppose that all six of Simpson’s modalized
modes (all the positive and negative ones) are by definition Evalu-
ative. If this presumption is correct, we should note the overlaps and
differences between Simpson’s formal markers and Labov’s: the way
Labov emphasizes negation while Simpson alludes to it only indirectly
(although with further analysis and specification it might be possible
to relate negation to negative mode narration), or the way generic
sentences figure in Simpson’s profile of positive narration but are not
directly invoked by Labov (although, again with careful qualification,
generic sentences might be recognized as ‘absolutely’ free clauses in
narratives – so narratively unbounded or free from punctual particu-
larity as to be a kind of ‘anti-narrative’ device – see Toolan, 1998: 64).
Among overlaps one would have to note the ways that both Labov
and Simpson highlight modal constructions as evaluative. Finally one
might reflect further on repetition, recognized as prominently evalu-
ative in Labov but not mentioned by Simpson: while repetition may be
found in all kinds of narration, it is arguably most noticeable in the
third-person narratorial styles which Simpson has called neutral B(N)
and B(R) – Hemingway and Flaubert are his best exemplars, respec-
tively. Like hard news reporting with which they are sometimes com-
pared, these deploy extensive repetition, as a brief glance at the
opening of almost any Hemingway story will confirm. Such repetitions
may be neither positive nor negative in Simpson’s terms, but they are
surely evaluative in Labov’s.
3 Even the Jamesian or Flaubertian narrator who may eschew Labovian
external evaluation, and many forms of internal evaluation (most
Intensifiers, Comparators, and Explicatives), can hardly be held to
shun Labov’s fourth category of internal evaluation, Correlatives. On
the contrary, these are used in abundance.
4 Finally there is a sense in which all the above factors miss the main
point, or at least, have to be re-situated once the main point is made.
And this main point is that a Jamesian novel is not merely more
complex than an oral personal narrative in the sense that it contains a
multitude of inter-connected narratives, each of which is locally
tellable and evaluated by its teller who is potentially different from all
other tellers within the novel; it is also more complex in that entire
narratives may constitute part or all of the Evaluation of a different or
larger narrative with which it connects – and connects in ways which
do not reduce to the grammatical choice between chaining (parataxis)
and embedding. Consider Gretta’s narrative of her loving and losing
Michael Furey; this is abundantly evaluated by Gretta of course, its
teller, and indirectly by Gabriel, its reactive recipient. But how does it
‘fit’ into the encompassing narrative called ‘The Dead’? In a sense it is
Complicating Action (‘what happens next is that Gretta explains . . .);
but it is also Orientation (it tells us more about the ‘who’ that Gretta
is than we – and even Gabriel – had imagined); and it contributes to
the Evaluation of the encompassing story in a number of ways: just one
of the ways is the way that Michael’s reported selfless unconditional
love for Gretta becomes a comparator in the evaluation of Gabriel’s
love for Gretta and other women: it is an order of love that, Gabriel
recognizes, he himself has never felt. So Gretta’s story contributes to
the evaluation of Gabriel and the narrative about him. But you could
hardly specify these relations in advance, on the basis of form or
expression held up for inspection, segregated and detached from its
embedded use. Nor could one assert with any confidence that all
readers, at all times and in all places, will agree with the attributions of
value and significance that I have made of these situated elements.
And were any theorist to propose, as one of the ‘rules’ of evaluative
narration, that a late-disclosed love affair and tragic death, revealed in
just these circumstances, will always be Orientational and Evaluative,
this would be clearly a mistake: it would be an attempt to subject phe-
nomena that can only be assessed in their integrated situatedness as if
they were decontextualizable and segregatable elements (on integra-
tionism, see Harris, 1998). This is why Chatman is ultimately wrong to
call the Labovian approach ‘contextualist’: by the standards of thor-
oughgoing integrationist commentary, Labov is not nearly contextual-
ist enough. In short, literary narratives are not merely more complex
than oral personal ones, they are exponentially more complex, exploit-
ing resources for evaluation (narrative ellipsis and pause are two more
that come to mind) which are virtually non-existent in the simpler
form.
Further reading
Labov and Waletzky (1967) and Labov (1972) are obviously the place to begin
further reading, and then Labov (1981) on speech actions and physical reactions
(relevant to 6.5). Thereafter, Chapter 2 of Pratt (1977) applies the Labovian model
to literary examples; see also Watts (1984), and Toolan (1988), while Culler (1981:
169–87) contains useful critique. Among the sociolinguistically-oriented studies
that have applied Labov’s model are: Carter and Simpson (1982), Bennett (1983)
and Polanyi (1982). Narrative performance and CHP are discussed in Wolfson
(1978; 1979; 1981; 1982), Schiffrin (1981) and Silva-Corvalan (1983). On cultural
presuppositions and their shaping of the stories we tell, see Polanyi (1978), Tannen
(1979) and Linde (1993).
If there is one theme that predominates in the sociolinguistic study of narra-
tives, stories, and performed tellings, it is that diversity of situation generates
diversity of stories, structures, features, and expectations. From one situation to
the next, it seems, ‘the same person’ can tell the same story so differently that it is
as if they were not the same person. See, among others, Johnstone (1990; 1996),
Johnstone and Bean (1997), Harré (1998), McConnell-Ginet and Eckert (1995),
Neisser and Fivush (1994), Linde (1999) and Norrick (2000). And on narrative
analysis applied to healthcare and psychology, see e.g. Crawford, Brown, and
Nolan (1998), Crossley (2000) and Payne (2000).
As a result narrative as a genre or kind has been problematized anew, and
viewed through the lens of social theory (using Bourdieu and Foucault among
others). Recently, the work of Gee (particularly on the diversity of narrative styles
that children and adults disclose, defying easy class- or ethnicity-categorization),
who has developed something of an alternative to the Labovian model, a verse-
stanzaic approach, has been very influential: see Gee (1991; 1996) and Hymes
(1982).
A great deal of attention has been directed to the question of the ways in which
spoken and written discourse differ. For a revision of the prevailing assumptions
about the greater syntactic complexity of written narratives, see Beaman (1984).
Tannen (1982) sees written narrative as combining the syntactic complexity of
writing along with spoken discourse features used to enhance interpersonal
involvement – hence a merging of oral and literate strategies; see also Tannen
(1988). Ultimately we may need to pay more conscious attention not simply to the
sociolinguistic environment but also to the wider cultural setting from which a
story emerges. For recent anthropologically-minded study of narratives, see
Bruner, who has been widely admired (1984; 1986; 1990; 1992) and Bauman
(1986).
But one publication stands out as a ‘snapshot’ of the range of applications (and
partial rebuttals or reformulations, it should be noted) of the Labovian structural-
ist approach to narratives, a snapshot that captures the range of ways, towards the
end of the 1990s, in which those seminal papers inspired researchers and practi-
tioners in education, first and second language acquisition, psychology, psychother-
apy, law, social constructionism, cultural and media studies, anthropology, and so
on – an array of disciplines ranging far beyond sociolinguistics. This is the 1997
special issue of Journal of Narrative and Life History, edited by Bamberg (1997a),
an absolutely essential resource. It opens with a reprint of the 1967 Labov and
Waletzky paper and closes, as noted earlier, with Labov’s more recent thoughts on
narrative analysis. In between, some fifty distinguished scholars contribute to this
compendium of concise and focussed papers, showing how their work has adopted,
modified or otherwise profited from the Labovian proposals, whether in harmony
with them or by means of the more clearly-articulated dissent or critique that they
helped make possible.
Notes and exercises

1 There is no better place to start, in seeing how effective the Labovian scheme of
narrative analysis is, than with data that you yourself have elicited. Record a
story from someone you know well. The story could be prompted by, for
example, asking them if they have been in a car crash, if they’ve ever had an
operation, or whatever, or the choice of story topic could be left up to the infor-
mant. The recording could be open or surreptitious. Just whom you record, obvi-
ously, will make a very great difference to the type and structure of story told:
five-year-olds produce stories somewhat unlike those of teenagers, who produce
ones somewhat unlike those of octogenarians. And the teller’s closeness to you,
the recorder, will also cause considerable differences. That ‘closeness/distance’
can be measured in all the ways listed in Section 6.9, but also in countless other
variable ways as well. With all the 6.9 factors held constant: is the teller ‘in a
good mood’ at the time of the recording? Have they always openly or secretly
felt competitive with you? Is there a lot of distracting background noise, or activ-
ity? What time of day is it (how tired are you both)? And so on, potentially end-
lessly. But at this stage the chief thing is to ‘get’ a story.
Once that is done, you will need to transcribe it. Or so runs conventional
sociolinguistic wisdom and practice. It is important however to reflect on how in
so doing one level of abstractive artificiality – setting about recording oral stories
by overt or covert mechanical means – is being compounded by another – the
reinterpretation of embedded speech-involving activity as displayed in
inspectable written text. Proceeding from taped record to written transcript is
full of paradoxes: most analysts would agree that an audiovisual record of a con-
versation was fuller and in some sense better than a merely audio record, and
yet most might also agree that, as a readable transcript, one of an audio-only
recording is far more manageable than one of an audiovisual one. The latter are
often regarded as unworkable. But the reality is that good-quality transmissable
audiovisual recording is becoming increasingly easy to achieve (as easy as good
quality sound-recording in the mid-1960s), and that this is often preferable to
pure audio in involving less exclusion of potentially relevant signifying material.
The real challenge is to develop annotative systems that can be applied selec-
tively in commentary and analysis of played and displayed sequences: for inter-
esting proposed descriptive categories, see Kress and van Leeuwen (1996) and
Baldry et al. (2000). But already, with present capabilities for computer handling
of digitized sound and image records, there should arguably be a good deal less
reliance on transcription of such records which, as noted, tends to displace ana-
lytical attention from the recorded activity to the transcript.
Notwithstanding those remarks, it is reasonable to begin the gathering and
analysis of oral narratives by working with audio recordings initially, together
with written transcripts of these. There is no great mystery about transcriptions,
because there is no possibility of getting them perfectly ‘correct’. It is best to
start by trusting your own judgment, simply setting out as full a written record of
the taped material as you can. You will soon encounter difficult questions of spe-
cific practice and general theory. Should one measure pauses: are they some-
times an evaluative device? Are there different kinds of pauses? Should one
annotate stress and intonation, and changes of pace of delivery? What about eye
contact – or lack of it – between teller and listeners; and both the teller’s and lis-
teners’ body movements? Can we legitimately ignore these dimensions of the
speech event? Where the answer is no, this highlights those respects in which an
audio-tape may mislead us more than it assists. But all empirical study must
resolve such questions: for useful discussions of sociolinguistic methodology see
Labov (1972) who famously drew attention to the Observer’s Paradox, as it has
been called: that the sociolinguist needs to observe how interactants behave
when they are unobserved. See also Wolfson (1976), Stubbs (1983: 218–46),
Ochs (1979) and more recently Wolf (1996), who proposes an integrational lin-
guistic dissolving of the Labovian paradox, at least in his study of New Orleans
residents with French-origin names, and those name-bearers’ reactions to
variant pronunciations of those names: “the facts are constituted in and by the
process of observation, and by virtue of the interaction between interviewer and
[name-] bearer, as between bearer and interlocutor generally.”
Now at last you can test out the Labovian model outlined in this chapter on
your sample narrative. Identify the orientation, complicating action, resolution,
and abstract and coda (if any) of the narrative. Is there any significantly delayed
orientation, and if so, are such delays motivated or accidental? Tellers may well
display something other than ‘perfect competence’: they may forget to include
important background, and may even neglect to relate important narrative
events. What store can be set by such errors or infelicities? Focussing now on
evaluation, is there a distinct section of this, as Labov’s diamond diagram sug-
gests, prior to the resolution, and/or is evaluation dispersed throughout the text?
Are certain types of evaluation device more prominent than others (e.g. many
intensifiers, few explicatives or comparators); is this a performed narrative, in
Wolfson’s sense? (Many of these questions only become really interesting once a
comparative analysis of a number of narratives is attempted.) Finally, in what
ways do you feel – or suspect – that the Labovian model is inadequate relative to
your target narrative? Is there temporal reordering of the kind disallowed by
Labov? Are there dimensions of evaluation that his categories seem not to
cover?
2 Attempt a six-part (abstract, etc.) analysis of Joyce’s ‘The Dead’. How satisfac-
tory does this seem to be? Does it help if we treat the text as two narratives – the
story of the Christmas party and the story of Gabriel and Gretta’s night at the
hotel – each with its own trajectory from orientation to resolution? Argue the
case for or against this proposal, taking into consideration the kinds of sequen-
tial cause-and-effect relatedness that each approach might highlight. Think also
of the possible evaluative function of the text’s recurring references to, for
example, warmth, generosity, mortality, darkness, and coverings. These may
operate at a level of narrative texture, mentioned at the opening of Chapter 4,
which I have entirely neglected. Thus, in the copious references to darkness and
covering and gauntness, from the ‘three mortal hours’ it allegedly takes Gretta
to dress, to the lugubrious conversation about the monks who sleep in their
coffins – a macabre topic which is ‘buried in the silence of the table’ (201) – there
are numerous allusions to the fact of dying. These, too, the reader incorporates
into their sense of why the story is being told.
To say so much is not, however, to claim that the ‘point’ of ‘The Dead’, or any
moral or message, can be easily extracted and summarized in a sentence or two.
That this is so first begins to be apparent when we try to decide where (if at all)
resolution leaves off and coda (if any) begins. Is Gabriel’s dreamlike sense of
dissolving, at the close of the story, a coda appended to the already-past events
of the party and of Michael Furey’s life and death? Or, more integrated into the
narrative, are these concluding thoughts a kind of resolution to all the preceding
events: selfconscious as codas often are, but inescapably a result as a resolution
is? Is our choice between seeing this as coda or resolution affected by whether
we feel the revery is genuinely revelatory insight, or maudlin self-indulgence?
3 But at least as interesting as applying Labov to such large narrative units as
whole novels or stories is applying the model to the stories within stories. Here
in ‘The Dead’, analyze Gretta’s direct speech account of her involvement with
Michael Furey, and comment on any oddities in the structuring of her story that
emerge. How important to the telling is its addressee, Gabriel?
4 If the question is asked, ‘Where or what are the evaluation elements in
Faulkner’s ‘‘That Evening Sun’’?’ – to say nothing of a more extended narrative
– how might one go about developing an answer? Ultimately I think it has to be
concluded that in anything as notional as evaluation, answers and modes of pro-
ceeding analytically will be contentious. But two potentially useful procedures
may be starting-points.
One of these begins by assuming that it is worth attending to what a reader
takes away, after the event, from a narrative, and combines this with a focus on
point and tellability. With both these in mind, it seems reasonable to ask one or
more informants (individually or jointly), who have recently read ‘That Evening
Sun’, what they recall being the point of the story, and what they recall being
some of the most striking means by which Faulkner or the narrator conveyed
that point.
A second procedure (bottom-up rather than, as in the previous suggestion,
top-down) would entail asking first-time readers of the story, as they proceed
through their first reading of one or more passages, to mark all those stretches of
text which they sense are not strictly telling what happened (next), nor supplying
orientation. By either means, it may be that readers agree that some parts of the
story are saturated with evaluation, while other passages seem to be largely
sequences of action (although not necessarily actions with a clearly reached
resolution, hence liable to attract interpretive evaluation, by the reader, as
scenes of ineffectuality or routineness or even disorder – depending on whether
the reader assumes that the protagonists would like to reach particular goals).
In the course of ‘That Evening Sun’ Nancy attempts to tell a story to the chil-
dren, who are overtly resistant addressees (one challenges her: ‘You don’t know
any stories.’). She of course, throughout the larger story, is convinced she is in
‘danger of death’, and that Jesus (hiding in the ditch) will come in the night and
kill her: it is to detain the children, to keep them with her that bit longer, that
she attempts to tell the story. Comment in detail on whether Nancy’s story –
only fragments of which are directly related to us – is one of vicarious or per-
sonal experience; comment, too, on how young Quentin, the narrator, who is
here one of the addressees of Nancy’s story, conveys evaluations of Nancy’s
efforts at evaluated narrating; and also comment on Caddy’s and Jason’s
reported remarks, during and just after Nancy’s storytelling, and how they get to
the nub of the evaluation and credibility issues that sociolinguists emphasize are
crucial to stories told in the course of talk, as this one is.
5 In this chapter I have emphasized that in order to justify the telling of a story,
what gets told must be to some degree special, and that non-trivialness must be
adequately communicated. On the other hand, from the ethnomethodologist
Harvey Sacks has come an emphasis on another aspect of people’s stories, an
aspect potentially at variance with the non-trivialness condition mentioned
above. This is that, to protect our stories from dismissal as outrageous fabrica-
tion, immediately suspect, we render our stories as ordinary as possible. Sacks
suggests that the stories we typically encounter in everday interaction are ‘over-
whelmingly banal’, and that people strive to establish the ‘nothing happened’
sense of truly catastrophic events. Thus ordinary people have to work their way
between the Scylla of the ‘so what?’ question (by making their stories not unin-
teresting) and the Charybdis of the ‘I don’t believe you!’ reaction (by making
their stories not incredible).
It would be interesting to explore the repertoire of linguistic means that we
might identify as, in part at least, performing this task of ‘ordinariness construc-
tion’. Again, we need recordings or transcripts of narratives of incidents that we
know (from personal involvement or further questioning of the teller) to be
extraordinarily ghastly, fortuitous or whatever. Some of the means we might
predict will contribute to ‘ordinariness construction’ might include downtoning
modifiers (rather, a bit, somewhat, etc.); hedges (sort of, kind of, etc.); vague lan-
guage; epistemic modals and qualifying verbs (I suppose, I guess, seemingly); less
emotive predicates (hit instead of smash); plentiful description of the ambient
routine events, settings and background; a flattening off of pitch and volume
variations; and restricted changes in pace of delivery. Sacks’s ordinariness
hypothesis would be to some extent confirmed if we could find a teller telling the
same story twice over, once to close friends, once to more distant acquaintances,
and found more downtoning in the latter case than in the former. It may also be
the case that the expression of ordinariness varies with different tellers: more
prominent in the stories of introverts, relatively neglected in the stories of extro-
verts.
6 Compare and contrast a spoken and written telling of the same story. If you are
lucky, the informant you used for exercise 1 might be persuaded to write down the
story she has told you. It’s probably best not to mention you’d like a written
version of the story until you have extracted the oral version, even if that infor-
mant declines to continue. And perhaps the most crucial thing to do is to keep the
data manageably short: many informants just won’t do a good job of writing up a
story that, in the oral telling, took ten minutes or more. The best way to grasp
what, as a data-gatherer, you are up against is to do the exercise yourself: produce
oral and written versions of some brief narrative from your own life before inflict-
ing this chore on others!
Now compare the two versions of the ‘same’ story in any ways you think
appropriate, but focussing especially on the formal differences. Are abstracts,
codas, evaluations, etc., ‘done’ the same way in the different media? Despite dif-
ferences of form, are there pairings or counterparts: for example, are there ele-
ments of the written form that do the job done by stress and intonation and pace
of delivery in the spoken version?
7 The single most delimiting aspect of Labovian description seems to be the
requirement that only independent clauses can carry the fundamental, fixed-
order clauses of narrative. In practice, Labov found sufficient apparent obser-
vance of a main-clause/main-event chronological pairing to suggest that this was
a requirement – a requirement, at least, in the kind of narratives, from the kind
of speakers, in the kind of settings, and so on, that he studied. But, as suggested,
in other kinds of narratives different clause/event pairings may arise.
Look for instances of main-event reordering in any spoken or written narra-
tives of personal experience that you have access to. Do these appear in particu-
lar types of subordinate clauses (e.g. of time and reason)? Do they relate to
particular types of happening (e.g. hidden motivations)? Are they more
common in the written narratives than the spoken? Does the temporal distance
of the narrated events from the teller’s present time seem to affect any tendency
to reorder? And – widening the net now – are they more frequent in narratives
of vicarious experience than of personal experience?
Finally, a deeper difficulty, which links this exercise up with some of the
remarks appended to exercise 1 above. This exercise (as at various points in the
chapter) has talked about the independent clauses and subordinate clauses in
the oral narrative a speaker delivers. But just how conclusively instances of these
categories are in a teller’s narrative, as distinct from being found by an analyst to
be in the narrative, is a contentious matter. Bear in mind, first of all, that
analysts almost invariably ‘find’ their independent and subordinate clauses in the
transcript (often their own transcript) of the discourse rather than the discourse
itself or its oral/audio recording. Analysts almost never (I suggest), in situ, as the
story is being told, make a mental note to themselves that ‘Ah, there’s an
independent clause, and there’s a subordinate clause which goes with the pre-
ceding clause, no, the following clause . . . ’ and so on. Nor do they do this with
the audio tape. Instead they put the thing in writing, at which point a host of
resources or categories are almost unthinkingly brought to bear on the trans-
formed discourse, in ways which make the analyst’s task considerably more man-
ageable. To begin with, a helpful and consistent spacing appears between what
are now usually and recognizably words (and where an analyst has found words,
they can – and do – the more easily find sentences (or ‘sentoids’), full stops and
initial capitals, and clauses). This despite the well-attested fact that there is no
regular audible gap between words in speech. No analyst in practice would
record the Singaporean story presented earlier and transcribe the opening as
follows:
myoneahmyoneahIwassavedbymyintelligenceyknowbecauseahIwasinsix-
ayyknoweightofuswerecaughtgamblingyknowsevensixrepeattwothat time
The only people whom I recall presenting ‘transcripts’ of this kind are poets. To
summarize, when the analyst makes a transcript of oral interaction this is
nothing like a mere ‘copying out’, but rather entails a range of spoken and
unspoken interpretive assumptions and decisions, a kind of ‘marking up’ of the
text in ways which will ensure that such basic elements of written language as
words, sentences, and clauses will be easily found. And that marking up into
writing is guided by resources which fall entirely outside the system of writing
itself, most notably intonation. It is the speaker’s heard intonation, in situ and
recorded on the audio tape, that is most determinative of an analyst’s placing a
period and sentence boundary at the end of the following chunk of talk:
MyoneahmyoneahIwassavedbymyintelligenceyknow
So categories of writing (the sentence, the clause) are being found on the basis
of criteria (intonation, etc.) which are no integral part of writing at all.
7 Children’s narratives
7.1 Stories for, by, and with children

Even within stories for children (as within those by and with) we can
observe and contrast many sub-types: spoken or written; face-to-face or
‘removed’ (e.g. radio or cassette); one-to-one or one-to-many; sponta-
neously constructed or text-dependent or scripted; and many more, in
numerous distinct combinations. Even with everything else held constant
except the prompt or priming, significant differences can emerge: Stein
and Albro (1997) have found that children ‘primed’ with a story opening
about a big grey fox produced considerably more complex narratives than
when given a story ‘stem’ to elaborate about a little boy or girl. So the
potential dimensions of situational variation are numberless. But a very
useful preliminary categorization of situations can be achieved by consid-
ering Halliday’s three macro-dimensions of discourse, which he calls the
field, mode, and tenor of the discourse.
Field designates, broadly, the nature of the social activity that is going
on and that the participants are engaged in (e.g. children sharing a per-
sonal-experience story with their teacher and peers). The discourse topic
will be important, but field concerns larger behavioural questions: under-
standing what type of social activity is in progress, what the activity is or is
about, and not merely what the discourse is about. In a segment of court-
room interaction, the discourse topic might be a particular incident involv-
ing particular stolen goods, but the field remains the administration of
criminal justice. Like other Hallidayan categorizations, field assumes some
degree of motivated ‘fit’ between language practice and cultural practice,
where a society’s language patterns articulate or ‘construe’ its view of the
world.
Tenor concerns participant relations as shaped by the statuses and roles
that are being observed (e.g. friend to friend, father to son, doctor to
patient, sales assistant to customer); this is a fascinatingly rich area since
most of the roles we occupy, and consequent relations with others, are
more temporary and contingent than we imagine. Focussing on tenor
chimes with new approaches, broadly social-constructionist (for a variety
of appproaches, see among others, Muhlhausler and Harré, 1990;
Children’s narratives 179
Bamberg, 1997b; Edwards, 1997) which emphasize and explore how we
‘take on’ particular roles or statuses in particular circumstances – the kind
of thing that conversation analysts vividly characterize as ‘doing being X’.
Thus no-one is spontaneously, a-socially and at birth, Catholic or female
or a professor or a general practitioner or a single parent or British or
Maori; rather all these categorizations (cf. the character attributes dis-
cussed at 4.4) must or will to a large extent be performed in the course of
communicative interaction, including verbal discourse. As indeed we know
from the possibilities of ‘passing’ and impersonation. Nor are these cat-
egories stable or permanent, with a fixed value: the categories ‘housewife’
or ‘divorced person’ or ‘atheist’ have nothing like the social force (and
consequences for tenor, for the ways in which people so categorized tend
to be discoursally addressed, and talked or written about) that they once
had. And new categories arise, such as ‘single parent’.
Mode encompasses variations in the medium of the text and the expec-
tations that the addresser and addressee have of the language they are
using (e.g. whether the text is written or spoken, planned or unplanned,
integral to some ongoing activity or a detached reconstruction; and
whether the language aims to instruct, persuade, emote, and so on). For
much fuller discussion of field, tenor and mode see Halliday (1978), Halli-
day and Hasan (1985), Martin (1992), and Matthiessen (1995).
The above reminders of situation-dependent variation are also relevant
to two caveats that should be entered concerning analysts’ assessments of
tellers and tales. The first concerns the use of data from children’s perfor-
mance with stories (skills at constructing, comprehending, summarizing,
and so on) as evidence of their cognitive skills or linguistic competence
more broadly defined. If situation-dependency makes major differences to
the stories that tellers produce, we must be circumspect about moving
from an assessment of a particular performance in a particular (perhaps
highly problematic) situation, to any general assessment of narrative (let
alone cognitive) ability. The second concerns the actual analysis of a
child’s tales: if our judgments are dominated by unexamined preconcep-
tions as to what a ‘good’ story should be like, and neglect the likelihood
that what is good in one set of circumstances may be quite inappropriate
in a different set, then any conclusions drawn may be valueless, if not posi-
tively harmful. We should never end up with the young wag dogged by
his tale.
On the other hand, while ackowledging differences, we can and should
also look for patterns of similarity and relationship across a range of cases.
Otherwise we fail to identify connections within what we can intuitively
accept as a coherent area of study. Ultimately analyses of children’s narra-
tives lead to some fairly sharp issues of relation and causation. Can specific
linguistic evidence be identified that supports the frequent claim that chil-
dren’s oral storytelling is a preparation for written narrative composition?
Similarly, if written narrative is a distinct kind of writing and a competence-
base from which older primary-school children build their developing
control of other varieties of language such as argumentation and descrip-
tion, which particular linguistic features of the known variety are a help,
and which are a hindrance, in progress towards those ‘target’ varieties?
Are children’s narratives evaluatively and structurally simplified versions
of adult ones, modelled on the narratives they hear from adults, or are
they constructed according to other principles; what linguistic evidence is
there on either side? Can we nurture production of more complex stories
(written or spoken) in children – and if so, how? What role do recall exer-
cises, probe questions, and shared retellings play in the development of
narrative skills? How does reading narrative relate to the writing of narra-
tive, should the former be graded for optimal effectiveness, and how do we
determine grades of narrative complexity or difficulty?
7.2 Storytelling and emergent literacy

An increasingly well-explored area of children’s language development is
that of the child’s nurtured receptiveness to writing and reading, and
familiarity with some of the disciplines and constraints of writing and
reading, somewhat before they are formally taught to read or write. The
fundamental recognition here is that, as the introduction to one recent col-
lection of studies of children’s emergent literacy puts it,
Practically all children in a highly literate society such as ours do bring

knowledge about literacy to school with them.
(Farr, 1985: vii)
A number of related theses, which I have space only to list in summary

fashion here, accompany this basic insight. Literacy begins at home, these
studies show, where that home is characterized by valued relationships
with parents, siblings and friends. Recurringly, close ethnographic studies
such as those reported in Farr (1985) and Bissex (1981) show that where
the interactant is highly attuned to the child’s current stage of writing com-
petence, is supportive and enthusiastic and not overly corrective of
‘errors’, then that interactant can be instrumental in developing the child’s
literacy skills in a greater variety of ways and more rapidly than if the child
were left to their own resources. It also seems that the effectiveness of the
support an interactant offers cannot be separated from the value the child
puts on that interactant as a person. This is a sharp reminder of the non-
autonomous nature of narrative productions and receptions: what you give
and what you get may vary sharply depending on the relations – of respect,
affection, or otherwise – between addresser and addressee. Halliday’s
notion of tenor, introduced earlier, attempts to address at least the lexical-
ized and grammaticized aspects of these complex and changing ties.
The idea of the interactant as an essential ally in the assault on ‘forward
positions’ of language development, otherwise beyond the individual
child’s current powers, is based solidly on the psychologist Vygotsky’s
notion of the zone of proximal development. Indeed Vygotsky’s theories
of cognitive development in the individual are the dominant theoretical
underpinning of much current research in the field. While Piaget’s theories
have the attraction of indicating quite explicit stages in cognitive develop-
ment which, at differing pace, all children must go through in the specified
order, following a ‘biological timetable’, those theories have been felt to
concentrate on development driven by interaction between the learner
and the object-of-knowledge (e.g. a language) to the neglect of develop-
ment driven by growing and changing cultural and personal interaction. In
a sense Vygotsky is more comprehensive than Piaget, allowing biology a
role but arguing that cultural development, fostered by interaction with
significant others in one’s immediate environment, dominates. And from
that interaction, as Vygotsky famously asserted, internalization proceeds:
‘An interpersonal process is transformed into an intrapersonal one’
(Vygotsky, 1978: 57). That which was learned through social interaction
becomes knowledge in the individual. And Vygotsky’s emphasis on the
social environment makes it all the easier for researchers to expect and
allow for variation in different children’s orders and manner of narrative
development, given different social environments and different interests
and purposes in the child. These brief remarks on Vygotsky may serve to
give some indication of the theoretical background of current orientations
in child development research.
The specific strategies adopted by kindergarten teachers in the nurtur-
ing of children’s skills of writing and storytelling are of many kinds. But
most commonly discussed are the activities of ‘sharing time’ (reviewed in
the next section), dictation and dramatization. In dictation, a child dictates
a new story, or retells a familiar one to a teacher-scribe. The teacher-scribe
is thus in a position to monitor the telling as it is in progress: suggesting
improvements, pointing to information gaps, etc., while the child retains
responsibility for the production. Dictation is in effect ‘doing writing
without doing the writing’, and gives the child a particularly immediate
sense of how marvellous a knack it is to be able to record one’s own story
productions with fixed graphic symbols (reproducible, increasingly unmys-
terious). Dictation entails a degree of detachment from the events of a
story that we noted earlier was essential for consistency of orientation in
any telling, and it also nurtures reflexive self-awareness. The dictated story
is usually read back to the child, with the teacher checking with the child
that she is happy with the story as it stands. The larger world of ‘second
thoughts’, self-correction and assisted improvement, of verbal production
as externalizable and separable from its originator but something that the
originator may retain responsibility for and ownership of, begins to take
shape. All this goes a step further when such dictated stories are then dra-
matized, with the teacher declaiming the story, in chunks, as its original
teller and other children act out its events: the teller learns the need for a
story to be coherent and tellable to its addressees in the most direct way
possible. In acting out a story the children reconstruct the world of events
which the teller’s words represent, and in the process, awkwardnesses of
match, between words and world, can become newly apparent.
7.3 Differing styles, differing orientations

In 6.8 I mentioned Shirley Brice Heath’s findings that what counted as
narratives in two small-town North Carolina working-class communities
(one black, one white) were drastically different. How easy is it, one
might wonder, for anyone from Trackton to tell a good, appreciated story
in Roadville, or vice versa? And not just in Trackton and Roadville. Much
recent research from psycholinguistic, sociolinguistic and social psycholog-
ical perspectives has uncovered just how much cultural specificity and rela-
tivism attends narrative practice. Let a hundred ways of telling flourish,
the observer might suggest, but the consequences may be undesirable
where contrasting ‘ways’ meet. Sometimes the exponents of the preferred
style get more rewarded, even without the rewarders being aware that it is
chiefly to differences of style that they are responding. Heath (1982)
vividly presents the moral and educational dilemmas that emerge when
orientation to narrative and literacy is recognized as plural.
Another persuasive elucidation of the problems broached here is
charted in the work of Michaels (1981; 1985; Michaels and Collins, 1984),
who shows that
[First-grade] children from different backgrounds come to school with

different narrative strategies and prosodic conventions for giving nar-
rative accounts.
(1981: 423)
This matters because ‘sharing time’ activities in which, prompted and sup-
ported by the teacher, individual children describe things or narrate past
experiences to the whole class, are an important and widely-used educa-
tional resource; they are a potentially rich practice ground for using (and
learning) literate discourse strategies. In an empirical study Michaels found
that children whose already-established discourse style matched the
teacher’s own literate style and expectations did well at ‘sharing time’
activities and benefited from this ‘oral preparation for literacy’ (1981: 423).
But where the child’s style and the teacher’s expectations diverged,
teacher–child collaboration was often unsuccessful and longer-term
adverse effects on school performance and evaluation were a danger.
Among the discourse expectations that Michaels found a particular
teacher disclosing, in her regulatory reactions to her children’s stories,
were the following:
1 explicit spatiotemporal grounding of the talk;

2 full description and naming of objects involved in a story, even of
those in plain sight;
3 minimal assumptions of shared background knowledge between the
teller and his or her addressees; and
4 lexicalization of topic shifts, marking any thematic relatedness that
persisted despite change of topic.
In sum, the teacher – ‘Mrs Jones’ – ‘seemed to expect a literate-style,

decontextualized account centering on a single topic’ (Michaels and
Collins, 1984: 223).
The problem that Michaels identifies related to the sharing style of the
African–American children in Mrs Jones’s class, a style which turned out
to have its own ‘systematicness’ but was at variance with the teacher’s
preferred style. While the white children brought to sharing time a dis-
course style that was already at least embryonically ‘topic centered’ in the
ways outlined above, the black children brought to the task an established
‘topic associating’ style. In topic associating discourse, the teller presents
‘a series of implicitly associated personal anecdotes’ (1981: 429), which
are actually built around an unstated thematic focus. Information is
broken into chunks prosodically, by means of a high sustained pitch on
and, followed by a pause; no other lexical connectives besides and were
used to link anecdotes.
This kind of rhythmically chunked, topic associating discourse is evi-

dently difficult to follow for those who, like the teacher and student
teacher, expect the discourse to focus on a single topic and to be
prosodically marked with sharp rising contours (signalling ‘more to
come’) or falling contours (signalling full closure).
(Michaels, 1981: 430)
Children who used this ‘topic associating’ style had acute problems with
meeting the teacher’s requirement, for their sharing time stories, that they
tell of just one thing of importance: the style itself made it seem (to the
teacher) as if there was no particular topic whatsoever. Telling about just
one thing made sense ‘only if one had a topic centered schema to begin
with’ (Michaels, 1981: 434).
In order to get a fuller lexicogrammatical description of the differences
in the discourse styles (topic-centred versus topic-associating) of such chil-
dren, Michaels and Collins analyzed the oral retellings, by four first-
graders (two white, two black), of a short film. To enlarge the comparative
exercise, they also looked at two fourth-graders’ narratives, both oral and
written, of the same film. They found that
Of the four first-grade narratives, two use a wide variety of lexical and
syntactic devices to signal agent focus, causal connections, old v. new
distinctions, and coreference relations. We call this a literate discourse
style. . . . The other first grade narratives rely more on prosodic cues
such as duration and special contouring to signal agent focus, causal
connections, and so forth. We call this an oral discourse style because
prosodic cues such as duration and contouring, although essential for
oral communication, are precisely what is not available in written lan-
guage.
(Michaels and Collins, 1984: 232)
Repeatedly, necessary textual connections are marked lexically in the lit-

erate discourse style, but prosodically in the oral discourse style. For
example one of the literate-style speakers first introduces one of the main
protagonists thus:
there was a man/. . . that was . . . picking some pears/
Twenty-four lines later she can recycle the relative clause to confirm that
here is second mention of the same man:
they walked by the man who gave/. . . wh-who was picking the pears//
By contrast one of the oral-style speakers introduces the same character

thus:
it was about/ . . . this man/ he was um/. . . um . . . takes some . . . peach-

/. . . some . . . pea:rs off the tree/
using two independent clauses. Twenty-five lines later, on second mention

of this character, the speaker relies almost entirely on prosody (vowel
elongation and high rise-fall contour on man) to signal old information,
subsequent mention:
and when that . . . when he ‘pa:ssed/ by that ma:n/ the man . . . the ma:n
came out of the tree/
In fact there is some lexical signalling of coreference here: the speaker’s

use of the deictic word that in the phrase that ma:n. But it is reasonable to
argue that this is an insufficient lexical marker of the coreference intended.
Another lexis/prosody contrast concerns the marking of an important
resultative connection between two narrative events in the story. This is
marked in the literate-style speaker’s account by the standard written con-
nective, so, while in the oral-style speaker’s account it is signalled prosodi-
cally by a stressed high fall on then. In other words when then is
prosodically marked in this particular way it is intended to convey causal
relation and not merely temporal relation; whether addressees derive that
meaning-difference or not is precisely the problem at issue.
Similar oral style versus literate style disparities, with attendant dif-
ficulties, were found in the oral narratives of the fourth-graders (both of
whom were fluent readers and writers ‘at the top of their class’). And the
potential problems in the oral-discourse speaker’s spoken narrative
become actual defects when that person constructs his written narrative.
For Paul, the literate-discourse speaker, ‘learning to write involves enrich-
ing a system he already knows and uses effectively in oral discourse’.
Geoffrey, on the other hand, tends to rely heavily on prosodic cues in

speaking; these cues are often the sole indicators of highlighted
information, coreference relations, and perspective shifts in his oral
narrative. His written narrative is characterized by weakly signalled
transitions and ambiguous identity relations. For him, with prosodic
options lost, learning to write means learning a new system for sig-
nalling thematic cohesion.
(Michaels and Collins, 1984: 241)
7.4 Children’s narrative development

The previous two sections have chiefly considered some of the ways that
emergent narrativity in the social setting of the school seems to be best
supported – together with some of the ways in which, despite the best
intentions of child and teacher, frustrations may arise. But we also need to
know more about just what development does occur typically in children’s
narrative ability. For these purposes several large-scale studies of chil-
dren’s stories can be referred to, and I will turn to these shortly.
One classic study of children’s linguistic development is Piaget’s pio-
neering work (1926). Focussing on story-recall ability in 6- to 8-year-olds,
Piaget identified a number of failings in the recall productions of his
younger subjects, failings which are far less frequent in the older children.
Those major failings concerned order, causality and orientation: the actual
order of events in a story was not reproduced in recall, cause-and-effect
relations were not properly marked, and egocentric use of pronouns
(tending to orient narrative events to the child herself) led to misdirec-
tions, in the recall texts, over who was doing what to whom in the narra-
tive. Despite many decades of critical review, Piaget’s basic findings are
still widely accepted (with the qualification that children can produce non-
egocentric text at a rather earlier age – four years – than he claimed). But
just what conclusions we should draw from such findings remains a matter
of debate.
More recently, children’s developing grasp of temporal order and
causality in stories, as well as of complex qualities such as that narratives
should appear adequately ‘completed’ and evaluated have been probed in
invaluable studies by Peterson and McCabe (1983) and others (McCabe
and Peterson (eds), 1991; McCabe and Peterson, 1996). The 1983 study
was important because, despite the authors’ broadly psycholinguistic inter-
ests, they eschewed highly artificial recall and comprehension tests in
favour of analysis of the ordinary stories that particular children chose to
tell them – not quite spontaneously, but certainly in relaxed and informal
situations. Peterson and McCabe conducted an analysis of 288 stories,
elicited during casual conversation, from a controlled sample of ninety-six
children ranging in age from 4 to 9 years (eight children for each of the
twelve age/sex combinations). Each story was subjected to analysis accord-
ing to three methodologies: the Labovian, the psycholinguistic ‘story-
grammar’ approach (see the ‘Further reading’ of the present chapter), and
a third, more syntax-focussed method; only the first two of these are dis-
cussed here.
First applying Labovian methods (or what they call ‘High Point analy-
sis’, referring to the top of the diamond in Labov’s famous diagram),
they found some interesting shifts in the type of narrative that children
tend to produce at different ages. ‘Leap-frog’ narratives were the com-
monest single type of story produced by the 4-year-olds, narratives
‘where the child jumps from event to event unsystematically, leaving out
important events’ (1983: 48), which put a heavy processing burden on the
listener. At age 4, the classic pattern (i.e. the full Labovian diamond,
with a clear climax and resolution) was still relatively rare. Within a year,
however, the leap-frog pattern itself seemed to become rare and then
disappeared completely, while the classic pattern became increasingly
common, dominating in the productions of the 7-year-olds and above.
But in addition they found that ‘chronological’ stories persist, albeit with
some decrease in frequency, at all ages. Stories classified as chronologi-
cal are those where there is temporal sequence of events without ade-
quate point or integratedness to the material, i.e. stories that fulfil the
referential function of narrative but not the evaluative one. Narratives
that Peterson and McCabe labelled impoverished (too few events, too
often repeated) or disoriented (confused, contradictory) also become
rare.
When Peterson and McCabe subjected the same corpus of stories to a
story-grammatical analysis they uncovered an equally developmental
picture. There was a steady increase with age in both the length of narra-
tives produced, and their structural complexity. A broad and cumulative
shift emerges, from production of mere sequences (such as clusters of
statements that report descriptions or actions without any protagonist’s
motivations linking them up) to production of episodes (where some sort
of purposeful planned behaviour is explicitly asserted, or can be inferred,
from the actions of a protagonist). Sequences outnumber episodes in the
4-year-olds’ narrative structures by more than two to one; episodes out-
number sequences in those of 9-year-olds by four to one.
But one type of sequence persists (often alongside episodes) in narra-

tives at every age examined: this is the reactive sequence, in which a
set of changes . . . automatically cause other changes with no planning
involved. . . . Something happens that causes something else to
happen, although there is no evidence of goals.
(Peterson and McCabe, 1983: 71)
An example is the following, from a 6-year-old boy:
E: What happened in the accident that you saw?

L: Car got burned up.
E: A car got burned up? Tell me about what happened when the car got
burned up.
L: There was three kids in there. Everybody got out in, just in time, and,
and, and then, my Dad didn’t keep his eyes on the road and we were
almost wrecked.
E: You were almost wrecked?
L: Yeahhhhh. I wouldn’t want that to happen. I’d be out of school about
a week.
(Peterson and McCabe, 1983: 73)
Peterson and McCabe noted both the increasing tendency for older chil-
dren to focus on planning and the unexpected persistence of reactive
sequences:
Several investigators . . . have suggested that the goal-directed,

problem-solving episode may have psychological reality. It may be an
important underlying pattern that is used by people in processing and
producing both fictional stories and narratives about actual life experi-
ences. However, the present research suggests that it may not be the
only one, since reactive sequences are common at all ages and cannot
be described as goal-directed behaviour. It is inevitable that important
things occur in the lives of children that are not planned, although
they elicit reactions. Reactive sequences seem to be the best way of
capturing the sense of these important externally imposed events.
(1983: 99)
If we acknowledge that ‘plan-less’ reactive sequences are sometimes a per-

fectly appropriate narrative to share – there being plenty of tellability, for
example, in a car that burns up, nearly killing three kids and nearly
causing Dad to crash the car – we can equally recognize the common
ground of such sequences with many ‘hard news’ stories (briefly discussed
in Chapter 8). Another broad trend is the increased elaboration and
embedding (roughly akin to coordination and subordination) within older
children’s narratives: multi-structure narratives, and structures within
structures, yielding ‘a more fine-grained analysis of their experience’
(1983: 100). In specific response to Piaget’s claim that young children have
a deficient grasp of causality (a claim which itself is based on findings of
children’s deficient mastery of causal connectives), Peterson and McCabe
argue the need to probe for children’s incipient understanding of causality
by means other than how they use particular linguistic expressions (such as
causal connectives).
If Peterson and McCabe supply abundant evidence of the structural
development in the narratives of children from age 4 to 9, Umiker-Sebeok
(1979) gives us a vivid insight into the younger child’s coming-into-
awareness of just what narratives are and can do as a distinct type of social
activity, a distinct field of discourse. She reports a study of the intra-
conversational narratives of infants (3- to 5-year-olds) in their nursery
schools. One striking finding was the tendency of the youngest children to
produce narratives about situationally proximate matters, i.e. about events
that had just occurred in their immediate environment (not just within the
school but even within the same play period). The 3-year-olds did this in 89
per cent of their intraconversational narratives. The 5-year-olds, however,
described events that occurred outside the preschool environment in 53 per
cent of the total narratives, i.e. narratives about ‘remote’ matters. This
looks like a striking example of learning to walk before attempting to run.
Umiker-Sebeok also found that older children’s narratives were far more
interactive with their listeners, eliciting relevant responses from their child
listeners (e.g. questions about orientation, evaluation or result).
It is interesting to compare all the above findings with those of Kernan
(1977), who undertook a Labovian analysis of a much smaller corpus of
narratives from somewhat older children – girls in three age bands: 7- to
8-year-olds, 10- to 11-year-olds, and 13- to 14-year-olds. Kernan’s findings
are interestingly complex or mixed. Thus the older girls tended to supply
an abstract or introducer for their stories (such as ‘Well see, this what
happen’) far more regularly than the youngest group; but why they did so
is open to speculation. In line with others’ findings, the older children also
did more elaborate identification of their stories’ main characters and loca-
tions (i.e. supplying more than a bald name). But Kernan also adds that
the teens’ provision of richer background information was done less for
reasons of structural or factual completeness, but rather to help ensure
proper ‘uptake’ of the story.
Also of interest in Kernan’s study is evidence of a change in relative fre-
quency of use of two simple sets of clausal connectives: and, then, and and
then, on the one hand (where the clauses so linked are said to be either quite
independent of each other, or have only a temporal-sequential interrela-
tion); and so, so then, and and so then, on the other hand (where the follow-
ing clause is said to depend on the preceding one in order to be understood
properly). Frequency of use of connectives from the two sets, combined,
declines steadily from the youngest to the oldest girls; and the distribution
between the two sets, heavily skewed towards the ‘and’ set in the youngest
girls’ narratives, is much more balanced in the oldest girls’ narratives.
But perhaps most significantly of all, Kernan finds a steady growth with
age of occurrence of clauses that serve what he calls an expressive function:
clauses that indicate in some way the feelings of the narrator toward
the events he is relating and that are used to attempt to convey that
feeling to the audience.
(Kernan, 1977: 101)
In Labovian terms, these are the means of evaluation that are external to
the narrative clause: expressions of the teller’s attitude to the events she is
recounting. It can be argued that they involve the self-as-teller maintaining
a kind of distance from the self-as-protagonist – sufficient distance for that
teller to be able to ‘take up a position’ on the events and protagonist.
Viewing these expressivities in this light, it is highly significant that the
most orientationally detached of these external devices, framed direct
speech – e.g. I said, I’m getting out of here – is never used by Kernan’s
youngest age group, but becomes frequent later.
Confirmation of this development of perspectival control, this ability to
frame one’s own narration, is found in Hickmann’s study (1985) of chil-
dren’s ability to transform dialogues into cohesive texts. Hickmann’s
youngest children (4-year-olds) never used direct or indirect quotation
frames to separate the act of reporting from the speech that was reported,
but used pitch and intonation signals to distinguish different speakers.
Older children (7- to 10-year-olds), however, did use direct quotation
frames. (It is a curious mirror symmetry that has children learning to dis-
ambiguate attribution by adopting framing clauses in their spoken narra-
tives, while sophisticated writers in turn have learned to dispense with
disambiguating framing clauses, in free discourse – direct or indirect.)
Thus the progression is similar to that noted by Kernan, though Hick-
mann’s tellers are using framed quotations earlier. In other situations,
reported speech in narratives may in effect combine the Labovian refer-
ential and evaluative functions. In a number of studies (e.g. Ely et al.,
1996), Ely has found a positive correlation between parents’ focus on talk
in their interactions with their children and the children’s subsequent
resort to quotation in their narratives. Ely also found a consistent gender
difference, with females using reported speech more frequently than
males.
7.5 Children’s narratives and the development of registers

and genres: the systemic-linguistic approach
If narrative is only one use of language, how does it compare and contrast
with others, and what are those others? And with regard to the develop-
ment of language skills in children, is there perhaps too much attention to
narrative style, and insufficient attention to other styles (argumentative,
expository, etc.)? The work of Martin and Rothery, two linguists who
adopt Halliday’s systemic-linguistic approach, tackled just these questions.
They sought ‘to identify some of the distinctive features of different genres
and some features of development in [children’s] writing’ (Rothery and
Martin, 1980: preface); their work is presented in this section. And in the
following section, a summary is given of recent systemic-linguistic pro-
posals for revisions to our categorizations of (adults’) conversational
stories: something of an enlargement and a corrective to the Labovian
model that was presented in the previous chapter. (This summary of sys-
temicists’ work could have appeared in that chapter, and is not exclusively
relevant to children’s storytelling; on the other hand it is only in the
present chapter that the systemic approach to register and genre has been
introduced, so on balance this seems the best place to report these devel-
opments.)
On the basis of their analysis of the kinds of writing children actually
produce from kindergarten to high school, Rothery and Martin argue that
there is a standard order in which registers and genres are mastered. The
first genre to appear is what they call observation/comment, as in the
following production from a 6-year-old:
My surprise
Oun day my mum bought me
o some books. and I falte
glad.
(from Christie et al., 1984: 69)
This most basic type of written text, in which some personal experience is
recorded together with an evaluative comment, is the foundation from
which, it is claimed, two major styles of writing develop: a narrative style
and an expository style. The narrative style entails temporality and an
affective trajectory; the expository style is a-temporal, has no affective tra-
jectory, and is basically a process of describing by means of increasing
depth and detail of analytical observation.
Accordingly, two genres appear a year or two after observation/
comment in children’s writing. Recount represents the first narrative-style
genre to be grasped, while report is the first expository-style genre to be
mastered. Recounts are akin to a chronological sequence in Peterson and
McCabe’s terms: a series of events temporally sequenced, often with ‘and
then’ as conjunctive link, optionally framed by an orientation and, at the
close, a reorientation. Reports are a more factual and objective description
than observation/comments, and are relatively depersonalized. Below I
reproduce the developmental genre typology that Rothery and Martin
propose. The two strands of the typology develop from left to right across
the page, with the year of school in which a typical child begins to control
a particular genre indicated underneath.
As the diagram shows, in fourth grade children begin to write personal
narratives, a genre distinct from recounts: in a narrative, awkward or
unforeseen events arise (complications requiring resolution). Fleshing out
the picture of genre-development we can again draw on the notions of
field, mode and tenor. As far as story subject-matter is concerned (one
aspect of field), children progress from writing stories about personal
experiences to writing ones about vicarious experience. In mode we see a
shift to relative context-independence (far less reliance on accompanying
pictures, less assumption of addressee’s shared knowledge, and a growing
recount personal vicarious thematic

narrative narrative narrative
observation/comment
report exposition literary criticism
Year of K 1 2 3 4 5 6 7 8 9 10 11 12
school
Age in 5 6 7 8 9 10 11 12 13 14 15 16 17
years
grasp of the fact that texts may address a remote hypothetical audience,
and not just the teacher and classmates). In tenor there are transitions
from stories disclosing an inflexible spirit of solidarity with, for example,
friends of the teller, to stories where there is far greater neutrality of
tenor, where evaluations are no longer the stock counters such as ‘yuck’,
‘great’ and ‘lovely’. Growing flexibility in control of tenor is, of course,
part of a developing non-egocentric sense of point of view, of the possibil-
ity of writing a story from another’s viewpoint.
Martin (1983) provides an insight into the enlarging linguistic repertoire
that indicates developing mastery of the register of narrative. He reports
the performance of ninety British schoolchildren (in three age groups: 6 to
7, 8 to 9, and 10 to 11 years old), of both sexes, of average ability and pre-
dominantly working-class, on several storytelling and re-telling tasks. In
particular, he tested their skills at suitably introducing and referring to
story participants. But first it is worth outlining what a teller’s linguistic
options for referencing are. In the chart of reference options below (taken
from Martin, 1983: 11), the most basic choice between introducing an
entity as new or alluding to an entity as given, is labelled as a choice
between presenting and presuming.
TYPES OF REFERENCE
Presenting Well there was a frog
Context of culture sitting in the sun

Reference (homophora)
Direct
He began to croak
(anaphora)
Presuming Textual
(endophora) Indirect
Context of because the pond was so cool.
(bridging)
situation
Extra-textual then that one over there arrived.

(exophora)
Within the types of referring that presume the listener can retrieve the
information, each terminal label indicates just where the listener has to go
to identify the entity referred to. Thus homophoric reference obtains
when, even on first mention of the entity, definite reference is used, as if
there were only one entity of this kind in the ambient culture. Hence refer-
ences to the moon and the sun will normally be homophoric; and by the
same token, indefinite reference would be distinctly odd for the targets of
homophoric reference: ?Make sure you keep out of a sun in the middle of
the day. Presuming reference that involves a disambiguating co-textual
verbal tie may be direct (e.g. by way of pronouns) or indirect, via the
bridging links that are discussed in 7.7 (where there are frogs we can
expect ponds). Most significant in analysis of children’s text production is
the reliance, on both first and subsequent mention, on exophoric reference
to immediate context, standardly by means of a definite noun phrase (the
cat) where more mature storytellers might be expected to follow the con-
vention of presenting the entity as new on first mention, and drawing on
anaphoric means of reference in subsequent mentions. Compare these two
retellings of the same previously-presented story, from a 6- to 7-year-old
boy and a 10- to 11-year-old girl respectively (exophoric nominal groups
are underlined, noun phrases involving bridging inferences are in italics):
1 Well there was a frog in the jar. The boy was looking it in the jar with
the frog. then the boy was in bed and the frog jumped out then it was
morning. the frog then the boy looked in the jar and he saw the frog
was gone. then he got dressed no then he looked in the boots his
boots. Then it wasn’t there. Then he looked out the window. The dog
got the jar on and then the dog fell out. Then the jar broke. then he
called for the frog. Then the then a deer came. He got then he climbed
on the uh rock and then a deer came and got the boy on the antlers.
Then he ran and then the deer stopped. Then the boy fell in the river.
Then they sat down and then the dog was on the boy’s head. Then
they looked over the log. Then they saw the family. Then the little
babies came out. Then the boy walked in the river.
2 Tommy was laying at the bottom of his bed looking at his pet
frog with his dog. While he was asleep with the dog on his bed the frog
tried to get out of the glass jar. The next morning they saw the
frog had gone. So they looked out the window and the dog had a glass
on his head. The dog fell out the window and smashed the glass and
Tommy came out and held the dog and the dog licked his face and
they called for him. They went down to the woods and there was a
swarm of bees coming out of a bee-hive. He went to the top of a rock
and the next minute he was on top of a reindeer. The dog went behind
the rock and the next thing they knew the reindeer was running after
the dog with Tommy on his head. They fell into the water off a cliff
top and Tommy saw a hollow tree so they both went over the tree and
Tommy told his dog to be quiet and they saw his pet frog and another
frog and lots of baby frogs. Tommy took one of the baby frogs home
with him.
(Martin, 1983: 19–20)
By contrast with the seven exophoric nominals introducing participants in

text 1, Martin finds only two in text 2 (arguably the first mention of Tommy
is a third). The generally appropriate management of entity-introduction
(appropriate in relation to mainstream storytelling conventions) in text 2
clearly contrasts with inappropriate means in text 1. (And the ‘definite ref-
erence’ introduction of ‘Tommy’ in text 2 can be defended as conforming
to story introduction conventions: we often dispense with the near-
redundant ‘There was a boy called Tommy’). Again, when we consider the
bridging references in the two texts, it is the mainstream norms of appropri-
ateness that highlight ‘the little babies’ in text 1 as rather awkward. The
three noun phrases classified in text 2 as bridging references can be inter-
preted differently: two simply as temporal connectives, not entities in the
story, and the third – the woods – as exophoric, but easy to interpret given
the preceding text with its dog and frog. These alternative interpretations
highlight the way that whenever a lexico-grammatical descriptive system
attempts to capture differences to do with meaning, then different analysts
are unlikely to produce absolutely identical analyses.
What is more interesting to realize is that, for example, frequency of
noun phrases involving bridging is not in itself an index of maturity: rather
it is increasingly appropriate recourse to bridging that we should look for.
Nor do the above remarks even broach the question of intention, though
that issue is surely important. For example, did the speaker of text 1 intend
us to bridge between ‘the family’ and ‘the little babies’? One final point to
consider would be the ways in which entity introductions that are identical
in terms of the system of choices set out above are yet substantially differ-
ent as regards effective storytelling. Thus first mention of the frog in both
texts is by means of presenting, but notice how much more appropriate is
the fuller description in text 2, his pet frog, than the bald a frog in text 1.
The special relationship between boy and frog conveyed in text 2 fails to
be explicitly recorded in text 1. Similarly contrast the narrative adequacy
of his pet frog and another frog and lots of baby frogs in text 2 with the
multitudinous possible referents of the family, the description that is its
counterpart in text 1.
7.6 The systemic-linguistic account of story genres

A little more will now be said about recent developments in systemic-
linguistic studies of narrative structure. If there is one tendency that most
palpably unites most of this recent work it is a focus on the situatedness
and cultural embeddednes of language patterns, seeking to specify what is
typical in the lexicogrammatical, register, and genre choices in, say, a
medical case conference as distinct from a disciplinary meeting, as distinct
from a child’s bathtime interaction as distinct from local council guidelines
on keeping a dog; and so on. As well as describing these diverse cultural
typicalities, systemicists wish also to understand and explain them, without
neglecting the ideological issues of power and positioning, affiliation and
exclusion, that may be motivating factors.
The more that systemic linguists (and other socially-oriented linguists)
have focussed on how context shapes the typical grammar of types of
interaction, the more the notion of genre has been invoked and elabor-
ated. And because of this ‘generic’ turn in socially-oriented linguistics,
Bakhtin’s long essay on ‘The problem of speech genres’ has been accorded
a place of theoretical priority and prominence – not as ‘the final word’ on
these matters but rather since it contains the complex reflections of this
most innovative and inspiring of proto-sociolinguists (other influences on
recent systemic studies include: Mitchell, 1957; Sinclair and Coulthard,
1975; Hasan, 1984). The core of Bakhtin’s argument is summarized by
Jaworski and Coupland:
Bakhtin presents a highly dynamic view of speech genres, inter-

penetrating and ‘re-accentuating’ each other and being continually
renewed as they are used. These claims are the basis for an intertex-
tual perspective on discourse, seeing discourse as the recontextualizing
of already existing forms and meanings, one text echoing and partially
replaying the forms, meanings and values of another.
(Jaworski and Coupland, 1999: 53)
And, it may be added, in partially replaying the forms and meanings of

other texts, each new text partially does not replay prior forms and mean-
ings. In short, nothing semiotic is fixed, or determined in advance, even if
everywhere there are would-be fixers. If these things are true of language
activity generally, they are true also of narrative and narratives.
By ‘genres’ systemic linguists mean, following Bakhtin in this, ‘relat-
ively stable types’ of interaction, extending beyond the traditional literary
genres to all the interactional types:
Thus, a transactional encounter such as buying meat at the butcher’s is

a genre, as is a recipe in a magazine or a staff meeting in the work-
place. [And] systemic linguists define genres functionally in terms of
their social purpose. Thus, different genres are different ways of using
language to achieve different culturally established tasks, and texts of
different genres are texts which are achieving different purposes in the
culture.
(Eggins and Martin, 1997: 236)
In systemic-linguistic accounts, a community’s lexicogrammar and the

arrays of options this fosters are tightly woven into the community’s social
and cultural worlds: the language reflects and construes the social order
that sponsors it. There are three crucial strata or phases of organization:
the lexicogrammar itself; the level of register (where combinations of
options from among the field, tenor and mode possibilities alluded to at
the opening of this chapter arise); and the most culturally-embedded level
of genre (where ‘staged, goal-oriented social processes’ [Eggins and
Martin, 1997: 243] are modelled).
At the same time it is probably as well to guard against assuming that,
armed with grammar, register and genres, the analyst can show that every-
thing we do with language ‘fits’ without gaps or overlaps. The basic idea of
a genre, if as indicated it can include such things as buying meat at the
butcher’s, is perhaps not so far away from Barthes’ idea of the narrative
sequence alluded to in 2.3: a ‘marking off’ of a normative and conventional
goal-directed sequence of moves, treated as a unified instance. But just as
we would be hard put to state just how many events had happened to us in
the last week (or in the last hour) – that is, event-hood is either culturally
or subjectively relative and arbitrary – so we would be hard put to show
that our interactional lives involve a steady progression through one
genre-instantiation to the next, without residues, blends, or awkward
hybrids. Perhaps the well-recognized genres are like mountain peaks, the
easiest parts of the landscape to see and to orient by; but there seems to be
plenty of ‘travelling’ between these established activity types, where any
sense or evidence of being genre-embedded is attenuated.
All of this is relevant to narrative, since systemicists argue that several
genres of story need to be recognized as occurring in the course of every-
day conversation. This account stems particularly from the work of Plum
(1988), but is advanced also by Martin, Eggins and Slade, and others. They
contend that the Labovian crisis or ‘high point’ narrative (often involving a
‘triumph over adversity’), is only one type in a set of alternatives, and
should not be promoted as the normal or unmarked format. For a range of
reasons, people often tell ‘well-formed’ stories in conversation that are
other than the six-part Narrative with all the trimmings (just as the English
breakfast is no less a meal for lacking a dessert or cheeses). Four chief story
genres, in partial contrast with each other, are identified: Narrative,
Recount, Anecdote and Exemplum. All may have stages that approximate
an Abstract, Orientation, and Coda, but then the differences emerge (see
table below). The Narrative, as in Labov’s model, has all three ‘internal’
constituents: Complication, Evaluation and Resolution. The Recount is
structurally nearest to the Narrative, perhaps, but it signally lacks Evalu-
ation and Resolution (cf. the reactive sequences noted by McCabe and
Peterson, discussed in 7.4): without a high point or crisis, what is there to
resolve? The kinds of story contribution that is easiest to recognize as
Recounts, I would suggest, are those produced at someone else’s initiation,
often shaped by politeness factors. Take yesterday evening: in a sense, very
probably, nothing much happened to you, nothing that is easily fashioned
into a Narrative that you currently spontaneously feel like contributing to
the conversation. But if your friend says to you ‘What did you do last
night?’, you might well reply with a brief sequence of unexceptional, non-
problematic events or activities: ‘Oh I got home, made pasta, watched ER,
phoned my mum, and then went to bed . . .’. No high-point, no Resolution.
The Anecdote has just one remarkable or problematic event rather
than several actions, this being followed by a quasi-evaluative Reaction
and, certainly, no Resolution: telling of being disturbed by a cockroach
running over your foot when you were taking an exam, or of suffering a
‘whiteout’ as a novice skier on a steep slope (both examples from Eggins
and Slade, 1997: 245–51). So the major emphasis is on plentiful ongoing
Evaluation, including listener-contributed Evaluation. And Eggins and
Slade suggest that there may be some socially-driven, gender-based prefer-
ences at work here, with men often preferring to tell their lives as Narra-
tives and women preferring to tell Anecdotes.
The Exemplum is arguably more minimalist than the Anecdote: it also
has just one event, but not a particularly remarkable or exceptional one,
reported in order to be interpreted. So the core structure is Incident ⫹
Interpretation (it is debatable whether Exemplums really have a distinct
coda section in addition, or whether the Interpretation section is a required
coda). Plum describes the Exemplum as ‘an anecodote designed to point a
moral’ (1988: 233) and there is certainly a message of ‘which just goes to
show that x (where x encapsulates a moral)’ made reasonably explicit at the
story’s close, having been at least implicit throughout. It is this characteris-
tic, and the sense that the teller has planned the telling of the incident so as
to further some ulterior purpose, to instruct the addressee rather than to
entertain, which distinguishes the Exemplum from the Anecdote. Where
the incident in the Exemplum is a means to an end, a warrant for a claim,
namely the conversationalist’s ‘offloading’ of a generic judgment, the entire
structure can be said to be marked by attenuated narrativity: sometimes the
Exemplum administers the pill of strong opinion sweetened by a rich narra-
tive instance. To summarize the structures of these conversational story
genres (with optional elements in parentheses):
Genre Generic structure
Narrative (Coda) (Orientation) Complication Evaluation (Coda)

Resolution
Recount (Coda) (Orientation) Record of events (Coda)
Anecdote (Coda) (Orientation) Remarkable event Reaction (Coda)
Exemplum (Coda) (Orientation) Incident Interpretation (Coda)
(adapted from Eggins and Slade, 1997: 268)
Having identified these four relatively distinct types, the implications of

those different types can be explored: why they develop and are sustained
culturally and socially, who uses which type in which situation, what types
are ‘barred’ from what contexts, and so on. Generic analyses of texts, in
this Hallidayan tradition, are always also gravitating towards producing
social and even ideological analyses.
7.7 Stories for and with children

Much of the research reported above, on the kinds of stories children
produce, would suggest that even at a very early age, even under 2, children
have a well-developed sense of what normally happens in stories and what
can be expected, in structure and content. But when we turn to popular,
commercially-published stories for children, it is clear that children want
and appreciate some creative departures from and exploitations of the
mainstream norms. This blend of the standard and the unexpected, the
given and the new, very probably answers to a child’s need for both the
reassurance provided by the familiar and the mental stimulation provoked
by the unexpected.
Now it might be argued that the requirements of reassurance and
mental disturbance apply to narratives for all ages, but there seems not to
be as strong an emphasis on the instructional function of stories for older
readers. Accordingly, both traditional and original stories for the young
are often marked by their structural similarity to fables, and their ‘happy
outcome’ codas often carry an implicit moral that could be cast as a
generic sentence or proverb. Relatedly, fiction for adults permits all sorts
of unreliability of narration, and immoralities of thought and action that
go uncondemned and unpunished within the narrative. In stories for
young children, by contrast, narration is almost invariably – and marvel-
lously – reliable, reliably reporting good conquering evil.
In what follows I want to discuss the text of one or two stories for young
children, noting both those elements that confirm and conform to the
canonical format and those aspects that are absorbing and entertaining
breaches of the norm. I will begin with the text of a very simple story that
combines – as do all those I will discuss – text and pictures. What follows is
the entire text of one episode about Mooty and Grandma; this comprises
six captions, as listed here, and each in turn accompanies a distinct,
double-page picture:
Mooty and Grandma
1 Mooty was a lovely little mouse. He had twinkling eyes and a long
curly tail.
2 One day, Mooty fell asleep under Grandma’s favourite chair. His long
curly tail lay in a coil on the floor. He dreamt of cheese and bread
crumbs. He dreamt of sweets and juicy plums.
3 His tail twitched as he dreamt. It curled and it straightened. It straight-
ened and it curled till his dreams stopped. Then it lay quietly in a little
gentle curve on the floor.
4 Grandma came into the room. ‘Oh, how careless I am,’ she thought.
‘I’ve dropped my sewing thread again.’ She bent to pick the long
strand of thread from the floor.
5 Poor Mooty! Poor Grandma! ‘Help!’ screamed Mooty as he tried to
wiggle free. ‘Help!’ screamed Grandma as she dropped him on the
floor.
6 Mooty ran into his hole as fast as his legs could carry him. Grandma
ran into the kitchen as fast as her legs could carry her. They both quiv-
ered and shook. Poor Mooty! Poor Grandma!
(Jessie Wee and Kwan Shan Mei, Mooty and Grandma, Singapore:
Federal Publications, 1980)
The role played by the illustrations accompanying this text is of great

interest. The pictures perform many functions: tautologically, we can say
they depict, but we should clarify what that in turn involves. First, to
depict is to supply visual validation or confirmation of what the text
asserts. Thus the first frame depicts a green Mooty, in lovable pose, with
huge ears, large bright eyes, and long curling tail, dressed in a red lungi
and red slippers. The text’s asserted verbal descriptions (He had twinkling
eyes and a long curly tail, etc.) are thus confirmed, and young listeners in
need of reminders of what ‘mouse’, ‘twinkling’ and ‘curly tail’ mean can
profit from the kind of ostensive definition the picture provides. In addi-
tion, for viewers alert to such things, the lungi and slippers, together with
his name, identify Mooty as an Asian mouse.
Pictures thus regularly confirm, clarify, explain and elaborate; they
carry on the work of narration by other than verbal means. For example,
the picture of Mooty’s dream shows more types of edible goodies than the
text lists. At a more abstract level, pictures stimulate the child to grasp
that words can represent pictorial scenes even when no scenes are pro-
vided in advance. The business of experiencing and understanding the
implications of text/scene matching, which all illustrated stories nurture, is
a crucial step to the more decontextualized children’s story, the one with
text alone, where the child is required to produce in her own mind, using
her imaginative resources, satisfying mental pictures of what is going on.
This important progression is surely one of the clearest cases of Vygot-
skyan interpersonal capability becoming transformed or enlarged so as to
become an intrapersonal one, particularly where the storytelling is shared
between adult and child. The adult can engage the child with reflexive
tasks and questions, such as ‘Show me Mooty’s red slippers’ and ‘Do you
see what Mooty’s dreaming of?’. At the later, fully intrapersonal stage, the
child in effect sees and shows themselves the things represented in the nar-
rative text, and those representations will be in words alone.
Grammatically, we have here several of the characteristic means of
supplying an orientation and of reporting complicating action. The main
character is introduced by means of stative verbs of relation, in which his
name, his species and his most crucial plot-determining attribute (his long
tail) are recorded. From these permanent states, caption 2 focusses on a
singular occasion of Mooty’s sleeping and dreaming, with imaginative
depiction of what Warren et al. (1979) would call the internal psychologi-
cal cause for his tail’s enthusiastic twitching. Accordingly, the text’s elab-
orate focussing of the reader’s attention to the moving tail in caption 3
seems well motivated, even though the tail really needs to be still when
Grandma sees it. In reporting the preparatory steps to that encounter, the
tail’s actions are reported in the expected format: dynamic verbs in past
tense and active voice, with the tail as subject, and repetition of the
salient feature of setting – ‘on the floor’ in captions 2 and 3. The most
important narrative development prior to Grandma’s intervention is
reported at the close of caption 3, where the verb lay, open to both stative
and dynamic interpretations, is used. We readers–listeners know that here
the stative description is also terminative, a point of rest after all the
excitement (of the dream) is spent. For Grandma, however, there is not
even a residual impression that what she sees as inert could ever be
mobile.
We should also take account of the significant role played by inference
and implicit messages, even in such a simple text as this Mooty episode.
Notice, for example, that there is no explicit verbal confirmation, in caption
4, that what Grandma interprets to be a thread is in fact Mooty’s tail. (The
accompanying picture manages to avoid disambiguation, too: it shows
Grandma’s reaching hand, and the thread/tail, but not the rest of Mooty’s
body: see p. 198).
Similarly, although caption 4 ends by reporting that Grandma ‘bent to
pick the long strand of thread from the floor’, there is no direct reporting
of Grandma’s act of picking it up, nor of her immediate shock of realiza-
tion. Both of these have to be inferred in the light of the caption and
picture that follow. Even in simple stories for the very young, then, it
seems that important plot-developing judgments and events may some-
times be left implicit, to be inferred, rather than spelt out. Whether this is
a good thing deserves more detailed review. My own impressions of my
children’s reactions to the Mooty story are that they do make the neces-
sary inferences, and are aware of the logic of cause and effect involved.
But we should not dismiss out of hand a very different view, which might
argue that in stories for very young children who may not yet have prop-
erly grasped the orthodoxies of sequence and consequence we require in
narratives, it is ill-judged to have stories that fail to reinforce those ortho-
doxies as explicitly as they might.
The text above comprises the first of the two episodes that make up the
story. In the second episode a flour-blanched Mooty, and Grandma in her
white nightclothes again manage to terrify each other, with the same paral-
lel reporting of their ludicrous and reciprocal misapprehensions:
‘I’ve seen a ghost,’ they both cried as they pulled their blankets over
their heads.
The outcome of these episodes, then, is an interesting variant of the

‘happy outcome’ norm, heavily mitigated, however, since the child knows
that the characters’ misery is unjustified, laughable. And behind the laugh-
ter lies a clear fable-like message, of how unfortunate inter-species mis-
apprehensions may stem from fear, ignorance and prejudiced first
impressions. The clear analogy here with inter-racial relations arguably
should not be spelt out to the child (cf. the systemic linguistic category of
Exemplum). It is more important that they get the general point that
appearances and impressions can be deceptive.
Another story that can be seen creatively blending departures from and
conformities to mainstream cultural and textual norms for children’s
stories is Burglar Bill by Janet and Allan Ahlberg. Bill is a decent, hard-
working thief who in the course of his work acquires a ‘nice big brown box
with little holes in it’. Upon returning home he discovers that it contains a
baby, and is introduced to the responsibilities of care-giving (as distinct
from those of property-taking). That same night Bill himself is burgled,
but he catches the intruder – Burglar Betty – redhanded. When these two
professionals get into conversation it transpires, in the fine traditions of
the eighteenth-century novel, that Betty is the widowed mother of the
baby, Bill has acquired. This well-matched pair repent their wrongdoings,
make literal restitution by putting back everything they’ve stolen, and
begin a new life together (a church wedding, in white). Burglar Bill
becomes Baker Bill.
That story summary should indicate fairly clearly how the unorthodoxy
of having a burglar as hero becomes only transitional to the final state of
rampant nuclear family orthodoxy. But there are textual unorthodoxies
here too: there is a jaunty sense of the unusual in the way Bill’s life as a
burglar is described, very much at odds with the high moral assumptions of
the orthodox children’s story. Much of the rhythmic poise of the opening
lies simply in the brazen recurrence of the adjective stolen, even in cases
where we suspect the narrator of dishonest exaggeration:
Burglar Bill lives by himself in a tall house full of stolen property.

Every night he has stolen fish and chips and a cup of stolen tea for
supper. Then he swings a big stolen sack over his shoulder and goes
off to work, stealing things. Every morning Burglar Bill comes home
from work and has stolen toast and marmalade and a cup of stolen
coffee for breakfast. Then he goes upstairs and sleeps all day in a com-
fortable stolen bed.
But the non-standardness extends to speech as well as behaviour: Bill and

Betty speak vivid working-class Cockney English. Together they produce
dialogue such as the following, a long way from the ‘correct’ mainstream
speech that dominates children’s stories:
‘Who are you?’ says Burglar Bill.

‘I’m Burglar Betty,’ says the lady. ‘Who are you?’
Burglar Bill puts on his own mask.
‘Oh,’ says Burglar Betty, ‘I know you – it’s Burglar Bill! I seen your
picture in the Police Gazette.’ Then she says, ‘Look here, I’m ever so
sorry – breaking in like this. If I’d have known . . . ’
‘Don’t mention it,’ says Burglar Bill. He holds out his hand.
‘Pleased to meet you.’
‘Likewise, I’m sure,’ Burglar Betty says.
And with this tenor of discourse established between the pair, there is
more than a hint of tolerant mockery of our standard expressions when
Bill, repenting his thieving, declares:
I can see the error of my ways . . . I’ve been a bad man.
The final children’s story I will consider is the magnificent Bear Goes to
Town by Anthony Browne. Bear has a magic pencil, so that whatever he
draws comes into existence – itself a subtle exploitation of the requirement
of text-to-pictures match discussed earlier. Bear thus has the power to
‘narrate’ everything but events. Although his good friend Cat is seized and
imprisoned by menacing guards, Bear uses his pencil to free Cat and other
animals, and remove them to a created pastoral haven, far from the threats
of the town. The core of the plot, then, is slight. What is absorbing,
however, is the extent to which the details over and above the basic plot
create a richness of thematic texture. This texturing is achieved more by
the messages in the pictures than the words, and it is quite striking just
how dependent the story is on the pictures to articulate developments and
connections that the text does not spell out. Consequently a rather high
degree of inference-making is required in the reader/listener. And it is
only when higher levels of theme and analogy are considered that several
scenes can be understood as properly motivated parts of the narrative
rather than somewhat incidental occurrences in a mere report or
sequence. Thus the second frame of the story book depicts Bear, in town,
being knocked over by the human traffic that dwarfs him, and the accom-
panying text reads, in part,
It was rush hour. Bear was small and people could not see him. They
knocked him down.
Strictly irrelevant to the main plot, and evidently an accidental mishap, this
incident can nevertheless be treated as thematic prolepsis, presaging the
threat to the animals that humans can be and, soon enough in this story, will
be. The next frame has a cat’s face looking down into Bear’s, the latter still
lying flat on his back from the fall. The caption simply reads ‘Bear saw big
yellow eyes looking down at him.’ Without any further introduction, Cat
becomes a full participant in the story – the text continues:
‘What is that?’ asked Cat, looking at Bear’s pencil.

‘It’s my magic pencil,’ said Bear.
‘Then draw me something to eat,’ said Cat.
And on the spot Bear draws, and thus creates, a rich variety of foods for
Cat to eat. Somehow Cat has immediately understood just what Bear’s
magic pencil does – and the child reader has to accept that Cat under-
stands. This seems to have less to do with standard inference-making and
more to do with this particular story’s chosen dynamics of informativeness.
In other words, the interpretive leap assumed in Cat, expected in the
reader, is just one of many sense-making bridges that the child has to
supply when reading this story. And, what is equally important to report, if
the experience of my own child and her peers is representative, it seems
that these sense-making demands are neither baffling nor beyond them,
but rather make Bear Goes to Town a favourite, preferred to stories with
plodding explicitness of informational links between actions and reactions.
In the course of this chapter I have attempted to show that, behind their
seeming simplicity and playfulness, stories for and by children are, on
closer consideration, remarkably complex, and an aid to and index of the
interactional and cognitive development of the child. Relative to respec-
tive ages and quantities of experience, the fact that many adults do cope
with Joyce’s ‘Eveline’ (and Ulysses) seems a far less impressive achieve-
ment than that 3-year-olds manage to cope with Bear Goes to Town.
When one makes allowances for the age of the child and her limited
acquaintance with story processing and story production, one is almost
drawn to end on a gloomy note, and remark on how scant is the enlarge-
ment of reading and writing skills from childhood into adulthood. Be that
as it may, what should be clear is that all the key elements of sophisticated
narratives, as discussed in the foregoing chapters of this book, are present
in simpler form in children’s narratives: causal connectivity of events, tem-
poral order and reordering, scenic and summarized presentation, focaliza-
tion, inferred linkage, and so on.
One curious effect in Bear Goes to Town seems striking confirmation
that 3- and 4-year-olds are already aware of the way narrative voice can
shift. In the middle of the story, there is a scene where Bear is absorbed by
a shop’s window display and Cat walks on along the street, unaware that a
menacing black-coated figure is lurking in a doorway. In a sudden depar-
ture from the narrative’s neutral reportage stance, the text suddenly cries
out (without the quotation marks used elsewhere to frame the words or
thoughts of participants):
Look out, Cat!
It is fascinating to see just how regularly children comment on this as they

hear the story – especially if they are hearing it for the first time. ‘Who
says that?’, they ask, or ‘Where did that come from?’. Who indeed?
Whether this modulation of voice is best described as one from the imper-
sonal to the personal, or from the extradiegetic to the intradiegetic, the
important thing is to have recognized – as these infant story-recipients,
astonishingly, clearly do – that a shift in narratorial voice has occurred.
Further reading
On children’s narrative development: Peterson and McCabe (1983), Cazden
(1981), Applebee (1978), Piaget (1926), Vygotsky (1962); and, from a systemic-
linguistic perspective, Rothery and Martin (1980), Christie et al. (1984), Hasan
(1984), Ventola (1987), Eggins and Slade (1997), White (1997) and Christie and
Martin (1997). On emergent literacy and children’s storytelling, see Heath (1983),
Romaine (1985), Bennett-Kastor (1983; 1986; 1999), Farr (1985), Cook-Gumperz
and Green (1984), Eaton, Collis and Lewis (1999). On children’s narratives in
health and clinical contexts, see Perez and Tager-Flusberg (1998) and Reilly, Bates
and Marchman (1998). Relevant journals include Journal of Education, Journal of
Child Language, Applied Psycholinguistics, and Language and Education.
As indicated at 7.4, a large body of psycholinguistic research has been under-
taken into the question of whether humans develop mental models of story struc-
ture, which they use as an aid in the comprehension, storage and recall of actual
stories. These mental frameworks are termed story schemas, and their probable
configuration can be represented on the page as a grammar; hence researchers in
this area are often described as story-grammarians. The psychological status and
the implications of such modellings remain controversial, but have been exten-
sively explored. Among key texts adopting a broadly psycholinguistic approach,
see: Rumelhart (1975; 1977), Mandler and Johnson (1977), Stein and Glenn (1979),
Warren, Nicholas and Trabasso (1979), Clark (1977), Brown and Yule (1983);
Schank and Abelson (1977) and Johnson-Laird (1981).
In recent years, seemingly part of the same movement away or beyond the
delimited formalism of Labovian analysis (as noted in the ‘Further reading’ of the
previous chapter), child language theorists and researchers have urged a renewed
attention to the symbolic uses to which children’s narratives are put. In a broader
sociocultural or social-constructionist framework, it is argued, the functions of nar-
ratives as vehicles of children’s meaning-creation and identity-construction, and of
self-projection, can and must be addressed (see, e.g. Bruner, 1986; Bamberg,
1997b; Muhlhausler and Harré, 1990; Nicolopoulou, 1996; 1997; Ely, Bleason,
and McCabe, 1996). One simple but powerful point to note is that in much
formalist-structuralist analysis of children’s stories, there has been a remarkable
disregarding of content, of what children actually choose to talk about. Only in
very recent years, for example, have researchers begun systematically to study the
ways children talk in their stories about being sad, or angry, or feeling they have
failed – all, unquestionably, developmentally important affective responses. In a
very useful survey Nicolopoulou (1997) speaks for many in arguing that a sociocul-
tural approach that may reveal children’s narrativized development of counter-
posed realities (differences of gender, ethnicity, class, and so on), must perforce be
a comparative and contrastive analysis.
Notes and exercises

1 The story-grammars mentioned in the ‘Further reading’ section above often
adjudge that the simplest stories are about a situation in which some want arises
and prompts an action which helps children’s narratives fulfil the want, or
prompts proceeding to some different action. An ‘if . . . then . . .’ logic is particu-
larly visible in such schemes.
It may be that such simple schemes are themselves most apparent in stories
for young children, who are only just beginning to encounter stories in books.
Consider the following text, for example, which is taken from a words-and-pic-
tures storybook entitled Dear Zoo, by Rod Campbell.
I wrote to the zoo to send me a pet. They sent me an . . . [flapped picture

of an elephant]. He was too big! I sent him back.
So they sent me a . . . [flapped picture of a giraffe]. He was too tall! I sent
him back.
So they sent me a . . . [flapped picture of a lion]. He was too fierce! I sent
him back.
So they sent me a . . . [flapped picture of a camel]. He was too grumpy! I
sent him back.
So they sent me a . . . [flapped picture of a snake]. He was too scary! I sent
him back.
So they sent me a . . . [flapped picture of a monkey]. He was too naughty!
I sent him back.
So they sent me a . . . [flapped picture of a frog]. He was too jumpy! I sent
him back.
So they thought very hard, and sent me a . . . [flapped picture of a puppy].
He was perfect! I kept him.
(comments in square brackets are my additions)
The storybook is aimed at children still some way from actual reading, but old
enough, and familiar enough with books, to participate while an older person
reads the text. At every point in the text where there is ellipsis indicated above,
the child has to lift a flap representing one side of a crate or container. And
behind the flap is a picture of the animal to be identified and named by the child.
Comment on the sorts of skills and understanding that such a book may
nurture in the child. Think especially about the idea of storytelling collaboration
mentioned both in 5.10 and 6.2; about appropriate timing of the flap-lifting;
about the skills of recognition, recall and naming that are involved; of how the
text performs the two fundamental Labovian functions, of reference and evalu-
ation, in a particularly overt way; and of the association of attributes with indi-
viduals in the course of characterization, which was discussed in relation to
literature in 4.3 and 4.4.
2 There is a fundamental assumption within Michaels’ and others’ treatment of
the dichotomy between oral-style and literate-style storytelling by infants that
needs fuller review. This is the assumption that the two styles, the two ways of
narrativizing experience, are ‘equal but different’. Similarly, Wolf (1984: 71)
argues that ‘children from different language backgrounds all have systematic
repertoires of narrative genres and registers’ and that the challenge is for edu-
cators to recognize those skills, and use them as a link to literacy. This seems
incontestable. Whatever the repertoires a child is identified using (and however
‘literacy-removed’ these are deemed to be), somehow these systems have to be
built on, rather than simply ignored or overriden, in the fostering of literate lan-
guage use, i.e. written language with its reduced integration with immediate situ-
ational context but its complex cotextual articulation.
But the question remains: are the oral and literate styles equally effective ways
of making sense of human experience? Can we demonstrate that both styles can
satisfactorily present logical and cognitive grasp of the events or situations? The
secondary issue, of whether those two styles are equal in the real world, is one
we can address subsequently; but we should first consider whether these two
styles are equal as effectively sense-making, suitably interpretive, and success-
fully message-conveying formats. If they are, then we might wonder whether
promoting a rather sharp change of style, as when oral-style users are encour-
aged to develop a literate ‘code’, is justified. Why shouldn’t such oral-style users
continue with their ‘equal but different’ style in their written language? The self-
evident answer to this, that oral-style prosody-based signalling just won’t work
in written language, should prompt us to ask further why the oral-style will not
work.
3
Once there was a dog named Whiskers. He got run over, because he ran in
front of a car. He was very sick after. He had to be rushed to hospital by
Ambulance and fast. At the end he ended up dieing isn’t that ‘Sad’.
(quoted in Christie et al., 1984: 78)
Analyze the text above, or others gathered from children, in relation to the lin-
guistic features of narrative genre that systemic linguists have drawn attention
to:
a the types of processes the verbs express (on process-types, see 4.9);
b the types and variety of lexical items;
c the consistency or variation of sentential theme (see 2.6);
d the manner of introducing protagonists and making subsequent reference to
them;
e the types of conjunctive relations that there are between clauses, and whether
these are given explicitly or only implicitly. Under the umbrella term ‘conjunc-
tion’ consider here two independent bases of subclassification:
ii semantic or logical type – the four basic logical types of connective are
additive, temporal, consequential (including purpose, cause, condition and
concession), and comparative (contrast and similarity);
ii implicit or explicit – is the conjunctive relation explicitly expressed or
simply to be logically inferred?
8 Narrative as political action
8.1 The contexts of narratives

In the previous chapter we explored children’s growing control of narra-
tive texts as decontextualized texts. ‘Decontextualized’ here does not
mean ‘entirely removed from context’ (an impossibility and theoretical
incoherence) but ‘freed from immediate reliance upon and disambiguation
by a co-present teller and addressee’. As the previous chapter suggested
this involves learning to use language and genres and conventions. But
while narrative text can be decontextualized as far as immediate situation
is concerned, it can hardly be so as far as the broader cultural framework
is concerned. A narrative is never without contexts which both shape and
come to be shaped by the story that is told and heard. Contexts may be put
in the plural for the too-often-neglected reason that the teller and
addressees of a narrative may assume quite different grounds for a
particular tale being told, and may separately deduce rather different
morals or consequences. It is because any narrative inevitably has some
effect on its addressees and consequences in the real world that we have to
recognize that narratives are, among other things, a kind of political
action. This is true of all narratives, even the most escapist, or those turned
to ‘just for a bit of fun’. Narratives, in short, invariably carry political and
ideological freight.
In the following sections there is space to present only brief glimpses of
the political contexts and contents of narratives. But by looking particu-
larly at the language of stories in the press, in the legal courtroom, and at
the narrative aspect of certain types of cultural commentary, I hope to
show that the worlds our narratives represent and make sense of may be
politically distinct worlds, with differing and even clashing assumptions.
8.2 Hard news stories in the newspaper

A newspaper is a hybrid compendium of discourse colonies (see Notes and
exercises 2, in Chapter 1), with each colony somewhat generically distinct:
hard news, soft (‘fancy that/who’d have thought it?’) news, features,
lifestyle discourse, sports reports, business news, opinion-editorials and so
Narrative as political action 207
on. These vary in narrativity. But even hard news stories have a structure
profoundly unlike oral narratives of personal experience. Most typically,
the modern English-medium hard news story is oriented around the
opening sentence, which will include the most tellable and critical world-
disrupting event of the story that it introduces. Thereafter, orientational
and complicating action reports will follow, in an order that often has
more to do with salience relative to the lead and headline than with
chronology. Or as Allan Bell puts matters:
Perceived news value overturns temporal sequence and imposes an

order completely at odds with the linear narrative point. It moves back-
wards and forwards in time, picking out different actions on each cycle.
(Bell, 1991: 153)
In fact, a news story’s reordered telling exploits and relies upon the
reader’s ability to reassemble events in their experienced order insofar as
the reader feels the need to do this. But the relayed version clearly privi-
leges news value, or ‘salience of latest outcome’, over chronology – pre-
sumably judging that modern readers do, too. The headline is often
composed last, a simplified and highly-evaluated version of part of the lead
sentence, and usually by a different, editorial, hand. While it would be
hard to maintain that headlines are neutral or impartial, this has often
been claimed of the body-copy of hard news items, particularly in the
English-medium press. Journalists pride themselves in their objective
reporting, and eschewing of opinion and subjective evaluation. As White
(1997) has shown, the typical news story of English-medium western jour-
nalism has a lead-dominated and orbitally-organized generic structure,
these contributing to the effect of establishing an impersonal authorial
voice, that purports to convey impartial and objective common sense. By
orbital structure, White means the way in which successive segments of
reported action, or orientation, or comment from story-participants are
not so much linked to each other in series, but separately linked (by
longer, weaker loops) back to the lead sentence which, each in its own
way, they explicate. He identifies five broad kinds of satellite, each directly
attached to the lead and supporting of its content: Elaboration; Cause and
Effect; Justification; Contextualization, and Appraisal-Attribution. The
lead sentence is thus the nucleus of the news story, carrying the event that
is represented as the most acute departure from and greatest threat to
ordinary life as we readers know and cherish it: White calls this nucleus a
constructed ‘peak of social-order disruption’ (1997: 112).
In reflecting upon these news structure choices it is important to bear in
mind that, for all their ‘naturalness’ for us today, these are cultural choices
that could be otherwise: hard news items could be told chronologically,
with headlines that only introduced the story without divulging the
outcome, as they were in the past. What, then, is the motivation for the
lead-dominated and orbiting-satellite structure? It has often been
explained in terms of repetitive amplification of the previously-selected
lead: but more recently it has been suggested that it also helps ‘naturalise
and portray as commonsensical the ideology which informs’ the selection
of just this story, handled in just this way (White, 1997: 111). More gener-
ally proponents of Critical Discourse Analysis (Fairclough, 1989; 1992; van
Dijk, 1988; 1993) emphasize the power that dominant social groups wield
through their discourses. Those controlling discourses construe and
project what counts as common sense, ordinary and reasonable, often in
subtly partial ways, effecting degrees of exclusion or marginalization of
subordinated social groups. The construal and projection may be calcu-
lated and considered; but quite often it is unconscious and seemingly invis-
ible to those beneficiaries who enact or sanction it (cf. presupposition, in
8.5): it is usually all the more effective for being so.
In summary, newspaper hard news stories have developed a distinct nar-
rative structuring, quite at variance with the Labovian high point oral Narra-
tive or indeed Recount, Anecdote or Exemplum. In being one of the most
powerful articulators and representers of its proprietor’s and readers’ world
(especially of that world’s ‘new’ or emergent culture), functioning as mirror,
lamp and lens, the newspaper is a powerful ideological instrument. It is not
that the newspapers carry ‘more ideology’ than most other processes and
participants in our world; but rather that the ideology is exceptionally widely
disseminated, influential and inspectable. With the lead sentence of hard
news items established as the ‘nucleus’ of the story, and the headline as
pithy paraphrase of the lead, there is a powerful warrant for directing crit-
ical narrative analysis particularly at those elements.
8.3 Political narratives in the news

A particularly influential linguistic study of the covert and even uncon-
scious ideological pressures on the hard-news story was conducted by
Tony Trew (1979), who applied a kind of linguistic criticism of text that
was in some respects a precursor of Critical Discourse Analysis, using the
Hallidayan system of linguistic description (sketched here at 4.9; see also
3.7 and 7.5). Trew’s thesis was that mainstream newspapers typically
espouse and legitimate some version of the dominant ideology that oper-
ates in a society, and must do so even in the course of reporting news that
is ‘awkward’ for that ideology. Trew writes:
Often one can see over a period of days a sequence in which some-
thing happens which is awkward from the point of view of the news-
paper reporting it, and this is followed by a series of reports and
comment over the succeeding days, perhaps culminating in an editor-
ial comment. By the time the process is finished, the original story has
been quite transformed and the event appears as something very dif-
ferent from how it started.
(Trew, 1979: 98)
Trew supports these claims in a comparison of two British newspapers’
divergent reports (in June 1975) of a single incident in pre-independence
Rhodesia (now Zimbabwe). Below are the openings of those stories, from
The Times and the Guardian respectively:
From The Times:
RIOTING BLACKS SHOT DEAD BY

POLICE AS ANC LEADERS MEET
Eleven Africans were shot dead and about 2,000 in the African Highfield
15 wounded when Rhodesian police township of Salisbury this afternoon.
opened fire on a rioting crowd of
and from the Guardian:
POLICE SHOOT 11 DEAD

IN SALISBURY RIOT
Riot police shot and killed 11 African here today in the Highfield African town-
demonstrators and wounded 15 others ship on the outskirts of Salisbury.
Despite similarities of content, The Times report is in the passive voice,

while the Guardian is active; and while The Times’s headline mentions the
agent of the killings (by police), this agent is deleted, as passivization com-
monly allows, in the lead sentence, and has to be retrieved by inference
from the subordinate temporal clause (when Rhodesian police . . .). Trew
notes a number of differences, of grammar and vocabulary, between the
two reports, many of these contributing to a picture (in The Times) of the
rioters and even the ANC (African National Congress) as chiefly respons-
ible for the deaths, or instead (in the Guardian) of the police as chiefly
responsible. But he more particularly focusses on the daily revisions to the
story of the Salisbury killings in subsequent issues of The Times, and how
the killings are incorporated into allusions to Salisbury’s riots and these, in
turn, become mere backcloth or precursor to ANC rifts. By means of
subtle and quite possibly unconscious retroactive re-casting, the larger nar-
rative that emerges is one of tragedies and disorders ultimately caused by
violent ‘tribal’ factionalism among black Africans, ‘with the whites con-
cerned merely to promote progress, law and order’ (Trew, 1979: 105–6).
At some point, press ‘reformulation’ no doubt becomes misrepresentation
and even defamation (where individuals are named), with attendant legal
liabilities. But even where these are not involved, a more general process
operates: over time, and in the hands of intermediaries, one story turns
into a different story.
The specific incident reported occurred twenty-five years ago, but there
are a number of respects in which the topic and sub-genre is anything but
dated. Even as I write, within and beyond Harare (formerly Salisbury)
there are demonstrations and acts of public disorder, sometimes ruthlessly
put down by government forces and sometimes noticeably not, as a power-
struggle involving the ruling party (ZANU PF), the opposition party
(MDC), landholding white Zimbabweans and landless black Zimbab-
weans, unfolds. Stories and headlines in several respects cognate with
those discussed by Trew are quite likely to appear.
Other contentious incidents, as distant in time, in which demonstrators
were shot and killed by security forces, continue to reverberate in our
media and in the public conscience – most notably the ‘Bloody Sunday’
shooting and killing, on 30 January 1972, in Londonderry, Northern
Ireland of 13 unarmed civil rights marchers by British Army paratroopers
who claimed to be returning fire on terrorists. A new commission of
inquiry chaired by Lord Saville, is even now (2001) in the process of gath-
ering evidence; it is charged with getting nearer the whole truth than the
report prepared by Lord Widgery in 1972 (a report increasingly regarded
as, at best, unreliable).
The public march, demonstration, or protest, because it involves some
effort and discomfort on the part of protesters and invariably attracts
media attention, is in a democracy or quasi-democracy one of the most
powerful resources available to the marginalized and one of the threats
most feared by the powerful (significantly, in contemporary Britain, polic-
ing of public protests is increasingly guided by statutes, such as the Public
Order Act of 1986). In a democracy the powerful are assumed to rule with
the consent of the governed; but a sizeable public march makes exception-
ally clear that the consent of some of the governed is withheld. So more
recently in Britain we have had the ‘Countryside Alliance’ demonstration
in London, and the failed mass demonstration against high fuel prices; and
internationally we have seen the range of articulate and organized dissent
during the World Trade Organization summit in Seattle, in December
1999.
Compare the treatment of those Seattle ‘manifestations’ in the
Independent and The Times, both of 1 December. In the Independent, the
headline, subhead, accompanying graphic and opening body-text are
reproduced opposite.
Spirit of the Sixties returns with a

vengeance to harass capitalism
WORLD TRADE TALKS Riot police fire tear gas as chanting protesters form a
human barricade to stop delegates attending the summit in Seattle
[large photograph of stationary protesters linking arms, no police or

opposed group apparent, one (female) prominently centred, wearing a
white shirt emblazoned with the message ‘This is a Peaceful Protest’]
Tens of thousands of protesters, cheering and whooping, thronged into down-

town Seattle yesterday to form a human barricade blocking delegates from
the World Trade Organisation, turning the opening of the much-contested
ministerial meeting into a colourful and passionate “carnival against capital-
ism”.
In The Times, headline, subhead, graphic and opening are as follows:
WTO protesters wreck start

to talks
Police fire gas as carnival turns sour, Bronwen Maddox reports from Seattle.
[large photograph, long shot, of crowded city intersection, with baton-wield-

ing police in riot gear centred and in focus, accompanied by an armoured
vehicle and with smoke/gas in the further distance, between the police and a
large throng of protesters; upper bodies of other protesters also in the near dis-
tance, at bottom of photograph; one or two placards – ‘Profit at what cost? No
to WTO’ – in focus.]
Demonstrators forced the World Trade Organisation to postpone and then

cancel its opening ceremony yesterday. Police firing pepper spray and rubber
bullets waded into ranks of steelworkers and costumed environmentalists as
the demonstration’s carnival atmosphere turned to menace, threatening severe
embarrassment for President Clinton.
There are a number of ways in which these two versions of events diverge:
to focus solely on the processes in the headlines, where the Independent
talks of a returning and harassing, The Times unequivocally reports that
the protesters wrecked the start to the talks. And both papers, intriguingly,
invoke the process of one state turning (in)to another state as what cru-
cially happened; but while the Independent reports that the opening of the
talks turned into a carnival, The Times asserts that an initial carnival atmo-
sphere turned to menace (or turned sour, as the subhead runs).
Why is The Times so evidently decided that the protestors’ disruptions
of the WTO talks lead to a ‘turning sour’ or ‘menace’? The reason must be
that The Times underlyingly presupposes the legitimacy of those WTO
talks: the newspaper has (and expresses) no doubts that in a normal and
ordered world the WTO talks would commence smoothly, and that this in
itself (as distinct from anything the delegates decided) would be no more
tellable than if a parliamentary session scheduled to begin at 2 p.m. on a
certain day did indeed begin at that time. By contrast the Independent
reports the actual events with no such entrenched commitment to the
assumption that the talks commencing on time is simply normal, reason-
able and ordinary. It treats the carnivalesque disruptions as legitimate and
as reasonable as the WTO gathering (possibly more so). So the Independ-
ent tells the story of the talks-disruption in a way that implies that, as far as
the newspaper is concerned, it would not be a disaster or a great threat to
the social order if the WTO talks collapsed entirely (and had to be re-
opened on some quite different basis, perhaps much more responsive to
the various interests of those protesting here). The Independent is not at
all sure that the protesters are endowed with less legitimacy, and are not
expressive of a better emerging social order, than the WTO delegates.
Perhaps they have more legitimacy. Similar uncertainties of construal and
legitimacy-assigment shadow some of the media coverage of other con-
tentious issues where two partly-reasonable but counterposed world-views
vie for support: e.g. divergent responses to the BSE/CJD crisis, or to
experimental planting of genetically-modified crops.
8.4 The unfolding news story: a contemporary example

For the purposes of a more extended analysis of the political and ideo-
logical determinants of a hard-news story (and of the narrativization of
politics), I will turn now to a very recent British news story, and five news-
papers’ treatment of it. The news item concerned comments made by the
leader of the opposition Conservative Party, Mr Hague, about crime and
policing, soon after the brutal killing in London of a black child by still-
unidentified individuals. Offering a synopsis of ‘the background’ to an
ongoing hard news story is inevitably contentious, but it is also essential. I
shall try to be neutral in recording the following facts:
1 On 28 November, 2000, a ten-year-old boy called Damilola Taylor bled

to death from a stab wound in his leg, after reportedly being attacked by
three teenagers, possibly black. Damilola was Nigerian-born and black,
and had lived in England for less than a year. The attack and death
occurred in an acutely run-down housing complex in London, report-
edly haunted by drug addicts, drug pushers, and petty criminals.
2 In the weeks that followed this tragic death, a good deal of public
anguish and soul-searching ensued, particularly since the police were
unable rapidly to identify, arrest and charge any suspects. Why and
how had this happened? Where had society gone wrong? Was this a
failure of the schools, of parenting, of social provision, of policing; did
such outrages (social workers on the estate said that stabbings and
violence were routine, not exceptional) follow from a profit-driven
‘selling’ of violence by the media and the larger culture? While the
chattering classes obsessed about the new Human Rights Act, how
had Damilola Taylor’s rights been protected?
3 The ‘frontline’ agency in crime prevention is the police. Since the
early 1990s, during both Conservative and Labour governments,
recruitment to the police has failed to keep up with numbers departing
(retirement, resignations). The police forces, particularly the London
(Metropolitan) one, have been been widely described as ‘demoralized’
in the aftermath of the MacPherson report. The MacPherson report
was a government-commissioned enquiry, chaired by a senior judge,
into the mishandling, by the Metropolitan police, of the investigation
of the murder of Stephen Lawrence, a black teenager, allegedly by
one or more of a gang of known racist white youths. The eventual,
belated trial of those youths failed to secure any convictions due to
lack of corroborating evidence. Among many other observations, the
MacPherson committee concluded that there was ‘institutional racism’
(unwitting or unconscious race-based differential treatment) in the
Metropolitan police force and that this needed to be addressed imme-
diately. This conclusion was praised in some quarters as a facing up to
reality, denounced in others as a slur on all decent police officers and
bound to damage further the public trust.
4 Partly as a consequence of the MacPherson sentiments, London police
officers on the streets began using their powers to ‘stop and search’
individuals acting suspiciously much more infrequently (statistics
showed that those subjected to ‘stop and search’ treatment – other
nouns used here include indignity, harassment, and provocation – were
disproportionately members of racial minorities). Therefore, in a
sense, post-MacPherson race-sensitivity may have caused more defen-
sive policing, and less interventionist vigilance over street crime. This
is summarized in certain quarters as ‘race-sensitivity (political correct-
ness) impeding the fight against crime’. At the same time it is also
widely recognized that, for example, male black youths are dispropor-
tionately the victims as well as the perpetrators of street crime.
5 It was suggested by some that the combination of lower police
numbers, demoralization and a ‘retreat’ from the pro-active stance of
routine stopping and searching, made it more easy or more likely for
opportunist street crime, against property or the person, to go unde-
terred and undetected.
6 Police numbers in many regional forces have begun to rise in recent
months, and the decline in numbers in London has ceased. All agree
that pay as much as conditions is a factor hampering recruitment to
public service professions in London (e.g. the high cost of living).
All the above are only part of the background, but they are a necessary
contextualization for the newsworthy items, duly reported by most national
newspapers, below. In essence these were a series of widely-publicized
statements made in the days following Damilola Taylor’s death, by Mr
William Hague, leader of the main party opposed to the Labour govern-
ment, the Conservative Party. In answer to some of the questions listed
under 2 above, he declared that the Labour government were in part to
blame, for presiding over conditions which made tragic deaths like
Damilola’s more likely to happen.
Now at last to the news items, all taken from reports of Mr Hague’s
intervention published on one day (Tuesday, 19 December, 2000). Con-
sider first just their headlines:
Guardian 1. Hague race jibe angers ministers
Independent 2. Race and policing: Hague’s defiance inflames the
anger
Telegraph 3. Tory leader ‘won’t be gagged on crime’
Mirror 4. CRAZY HAGUE DEFIES DAMI DAD’S
PLEA
The Times 5. Hague links Labour with murder rise
The differences between these headlines suggest that these papers are not
even telling entirely the same story. Judging by references to him in all five
headlines (four times by family name, once via his political position), it is
agreed that the story is ‘about’ Mr Hague. But thereafter divergences
appear: only the two centre-left broadsheets (Guardian and Independent)
mention race in their headlines. By contrast the centre-right broadsheets
(The Times and Telegraph) mention crime or murder (but not race), while
the Independent headline alone invokes a distinct issue which, it implies,
lies at the tellable core of the story: policing. The tabloid Mirror is differ-
ent from all the above in its acute personalization of the story, as a clash of
a crazy, defiant Hague and ‘Dami’s pleading Dad’.
Simply by lining up these five headlines, it is easy to see which account
is relatively sympathetic to Mr Hague or, conversely, unsympathetic to
him, and which reports tend to be at least not unsympathetic to Labour or
‘ministers’ or those feeling ‘anger’. Without further analysis it is possible
to categorize the Guardian headline as unambiguously critical of Hague,
respectful of ministers, the Independent as more neutral but mildly critical
of Hague, the Telegraph as approbatory of Hague, the Mirror – the only
tabloid here – as sensationalist in its anti-Hague sentiment, sympathetic to
Damilola’s dad, and The Times as uncritical of Hague, and undeclared or
guarded in its criticism of Labour. The crucial openings to these five con-
trasting news treatments will now be briefly discussed in series, but con-
trastively. Due to space limitations, nearly all commentary is directed to
the headlines, but occasional brief mention is also made of the following
bodycopy, which certainly warrants further analysis. Hard-news story
headlines (as distinct from those – often oblique or trivial or merely witty
– that accompany ‘soft news’ or human interest stories) are invariably of
particular importance, by virtue of their prominence and careful composi-
tion. The headline is the narrative’s Abstract, with the complication that it
has been composed after the preparation of the body-copy. They are a
kind of ‘text-bite’ that emerged long before sound-bites were recognized:
pithy, crucial, viewpoint-shaping encapsulations.
From the Guardian
Hague race jibe angers

ministers
Nicholas Watt and Nick Hopkins
Tuesday December 19, 2000
William Hague yesterday raised the Hours after the Taylors had accused Mr
stakes in the row over law and order Hague of using their son’s death as a
when he ignored a plea for restraint “political football”, the Tory leader
from the parents of Damilola Taylor to went on the offensive to declare that
issue a blunt warning that a Labour politicians could not bury their heads in
election victory will lead to more tragic the sand over crime levels.
murders in Britain’s inner cities.
Only the Guardian and the Mirror, the most clearly anti-Hague and pro-
Labour of the five papers canvassed here, specify in their headlines a
particular affected human as recipient of the verbal process (a jibe in the
Guardian, defiance in the Mirror) initiated by Mr Hague. That human recip-
ient is ministers in the Guardian, and Dami’s dad in the Mirror. And there is
nothing to suggest that either the Guardian view of ministers, or the Mirror’s
of Dami’s dad, is at all negative. By contrast the Independent is unspecific
about who exactly is made angry by Mr Hague’s defiance, and in the Tele-
graph and The Times, in the implicit matrix clause in each case, there is
really only one stated participant: Mr Hague, who won’t be gagged (in the
Telegraph) or who projects an entire claim, that Labour is linked with
murder increases (in The Times). So only in headlines 1 and 4 above is there
the most direct representation of a human-to-human confrontation. In the
Guardian something Mr Hague has said has ‘angered ministers’. And it is
striking, to say the least, that what he has crucially, tellably said according to
the Guardian is so different from what, according to The Times, he has said.
For in the latter he has linked Labour with a rise in incidence of murder; but
in the Guardian he has uttered a ‘jibe’ about race.
Here the key and powerfully evaluative term is jibe: nothing remotely
similar occurs in the headlines or copy of the other newspapers surveyed.
Standardly a jibe is a wounding and negative, even destructive remark,
typically about something of relatively minor importance. Often the
source of the jibe, in being so represented, is also regarded as frivolous,
playful, or non-serious, while the target or addressee of the jibe suffers
embarrassment or humiliation. But a jibe is not a critique, it is not bal-
anced, fair or reasonable, it is a ‘scoring points off’ someone – ‘a rude or
insulting remark about someone that is intended to make them look
foolish’ (according to the corpus-based Cobuild dictionary). On several
scores, then, jibe is a marked and therefore informative choice here. If we
do not ordinarily make jibes about really serious matters – religion, death,
cancer – then a ‘race jibe’ implies a kind of deepseated inappropriateness
in Mr Hague’s way of broaching the topic; besides, as the dictionary indi-
cates, we standardly jibe at a person, not an abstract category such as race.
The newspaper is perhaps suggesting – but again only tacitly – that Mr
Hague is underlyingly intent on embarrassing particular individuals (the
Home Secretary and other Labour ministers), under cover of protesting
that he is concerned only with ‘the issues’ (crime and policing).
From the Independent:
Race and policing: Hague’s

defiance inflames the anger
By Paul Waugh and Andrew Grice
19 December 2000
William Hague rejected a plea yesterday But Mr Hague was unrepentant and
from the parents of Damilola Taylor to stop warned that he would not be “bullied” into
using their son’s death as a “political foot- dropping the issue of low police morale and
ball” and even suggested that voting Labour numbers. He triggered further condemna-
would lead to similar tragedies in future. tion by suggesting a second Labour term
Richard and Gloria Taylor accused the would result in more crime and “more
Tory leader of seeking to use the killing of tragedies” similar to that of Damilola.
the 10-year-old boy for political advantage
when he said that a fall in police numbers
was partly to blame.
The Independent’s headline somewhat clumsily reports that Mr Hague has

‘inflamed the anger’; on the assumption that ‘inflaming’ or causing
increase in anger is invariably undesirable and ill-advised, it is implicit in
this representation that Mr Hague has acted in regrettable ways.
From the Daily Telegraph, Tuesday, 19 December 2000:
Tory leader ‘won’t be gagged on crime’

By George Jones Political Editor
WILLIAM HAGUE claimed last night that ogise for talking about the implications of
another Labour government would mean the murder. No areas of public debate on
more tragedies such as the murder of crime should be off-limits, he said.
Damilola Taylor. Labour described the Tory leader as “a
Despite criticism from the boy’s parents desperate man who has made a terrible
that he was using their son as a “political mistake”. Jack Straw, the Home Secretary,
football”, he said he would not be bullied refused to apologise for suggesting that
into silence on inner city crime, falling he was a racist. “I believe the Leader of
police numbers and morale. Mr Hague the Opposition is playing the race card,”
said he was “very sorry” for Damilola’s he told MPs.
parents but refused to back down or apol-
The Telegraph headline is one of the more ingenious reworkings of events,

using passive voice which enables the teller to dispense with prominent
naming of who it might be who is intent on gagging Mr Hague. But notice
the direct speech representation of Mr Hague’s words; actually this is
pseudo-direct speech, common in newspaper headlines: it is not a faithful
record of Mr Hague’s words since, judging by the body-copy that follows,
he never used the word gagged; it seems that he did, however, say that he
would not be ‘bullied’ into silence (see also bodycopy in the Independent)
about low police morale and rising crime. Using gagged in a passive con-
struction with agent-deletion is probably judicious for the Telegraph, in
view of the fact that Mr Hague was urged not to speak further by two
quite different parties: Damilola’s grieving parents, and the Home Secre-
tary and other Labour ministers. It would be narratorially incoherent for
Mr Hague, or a newspaper, to represent the Taylors as in any position to
gag or bully a leading political figure.
From the Mirror:
CRAZY HAGUE DEFIES DAMI DAD’S PLEA

WILLIAM Hague last night defied a tragedies” from another term of
plea by the parents of murdered Labour government.
Damilola Taylor to stop using his Mr Hague was rounded on by all
death as a “political football”. sides as he insisted police numbers had
The Tory leader astonishingly contributed to the stabbing of
ignored the request by dad Richard Damilola, 10, three weeks ago.
and mum Gloria and predicted “more
The Mirror report uncompromisingly casts Mr Hague in a negative light,

from its headline evaluative description of him as Crazy Hague – the crazy
being highly evaluative and ‘in excess of’ anything in the co-text to warrant
this verdict. While Hague is crazy, what he has most tellably done is defy a
plea – a significantly unidiomatic collocation (compare ignore a plea or defy
critics: in the second sentence of the report he ignores the request, but again
how the reader is to take this is not left to chance, since we are told that he
astonishingly ignored the request). Standardly, those who plead are weak
and vulnerable, so that Hague’s contraversion of the plea earns him no
credit (compare torturers ignoring the pleas of victims), despite the use of
the potentially approbatory verb defy. Indeed someone who defies the pleas
of others is being represented enacting rather an odd procedure, since pleas
are far too weak and subordinate (compare defying a threat) to warrant
being defied: hence, perhaps, the attribution of craziness, bolstered by the
following characterization of this behaviour as astonishing. In addition, Mr
Hague is represented as in conflict, in the first instance, not with the Labour
government but with the father of the murdered boy, Damilola, or Dami’s
Dad (notice the gender bias here, which ‘promotes’ just the father as oppo-
nent, from the couple cited in the bodytext: dad Richard and mum Gloria).
The intimacy or solidarity reflected in the informal naming of the Taylors is
clearly in contrast with the distanced (The Tory leader) or even hostile
(crazy Hague) namings of Mr Hague. Consistently, across headline and all
the quoted copy, Mr Hague is Subject and Theme; in the Mirror the text is
unequivocally about Mr Hague and his defying, ignoring, predicting and,
finally, being rounded on. Consistently he is presented as acting in ways that
are mildly absurd, if not crazy. Thus he is said to have insisted that ‘police
numbers contributed to the stabbing of Damilola’: here the paper has used a
nominalization – police numbers – of a potentially quite misleading kind.
Anyone reading this who was unversed in the ongoing debate could be for-
given for thinking that Mr Hague had absurdly claimed that by their numer-
ousness the police had facilitated the crime; whereas police numbers is a
kind of contracted nominalization, denoting the relative decline in police
numbers in the last few years.
From The Times:
Hague links Labour with murder rise

BY PHILIP WEBSTER AND TOM BALDWIN
WILLIAM HAGUE inflamed the growing In a series of interviews organised after
political tensions over crime and race last the parents’ criticism, Mr Hague voiced his
night by suggesting that the re-election of a personal sympathy for them. But he
Labour government could lead to more pitched straight into more controversy
tragedies such as the murder of the school- when asked in a BBC interview if Labour’s
boy Damilola Taylor. crime policies would lead to more cases
Hours after being accused by the parents such as Damilola’s. “There would be more
of Damilola of using their son’s death as a crime if police numbers and police morale
political football, a defiant Conservative continue to fall,” he said. “It will mean,
leader said he would not be “bullied” into yes, that there are more tragedies we will
letting the issue of police numbers and end up discussing.”
morale drop.
The Times’s headline, although it makes Mr Hague Theme and Subject,
also makes him the source of a projected proposition, which has been
complexly grammatically transformed; the ‘congruent’ and spelled-out
sense is something like:
Mr Hague has said that government by the Labour party has indirectly
caused a rise in the incidence of murders.
But, in newspaper discourse, to link carries associations and insinuations

entirely missing from more neutral projecting verbs like say, allege and
claim. The difference is partly reflected by the lexicogrammatical fact that
you cannot write a coherent counterpart headline that runs:
*Hague alleges/claims Labour with murder rise
This is because link, unlike allege, has incorporated into it the sense that
the following nominal phrases ‘go together’: so to link is not merely to
name two or more entities, but also to assert their connection (possibly
causal). Furthermore, where the verb link is used with a human agent and
two inherently human objects, it seems to have a negative ‘semantic
prosody’. By a negative semantic prosody I mean that despite its seeming
neutrality, in particular contexts of use the word usually carries and
spreads to the co-text a negative connotation – but this usage tendency is
sufficiently subtle that it has not become codified in dictionary definitions
as ‘part of the meaning’ of the word. Link can be used neutrally, or even
positively, particularly when the two entities linked are inanimate; but
with human agent and one or more human objects, as corpus research con-
firms, the linking is of suspects with crimes or, in situations of actual or
simulated scandal, adulterous sexual partners or shady business dealings:
to detect a link is to uncover a suspectly-concealed connection.
Even quite preliminary scrutiny of the Cobuild corpus confirms these
observations. On the assumption that Hague links is ‘headlinespeak’ for
Hague has linked (Labour rule with rise in violent crime), I selected and list
below the first instances of linked I found in the corpus. (This sort of selec-
tivity is necessary so as not to be distracted by all the occurrences of links
as a noun, references to the organisation Link, and so on). When link is
used as a verb, it is overwhelmingly past tense, and here are the first five
such random and unrelated examples that the corpus offered me:
Fat . . . particularly saturated fat, has been linked to coronary heart

disease and other diseases . . .
discovering the mechanism that linked India’s monsoon with Peru’s El
Niño – . . .
Your addiction to laxatives may be linked to fears about growing up.
They . . . [pre-cognition and telepathy] . . . They are strongly linked in
some meaningful way . . . .
his Christian Democrat power base is linked to the Sicilian Mafia. . . .
Three of these five are unequivocally ‘negative’ linkages – saturated fat

with heart disease, laxative–addiction fears, and a power base and the
mafia, respectively. The other two here are not noticeably negative – but
they are also different structurally from our target construction, namely
where link is used with at least one human-reference object. Again,
outside of newspaper headlines, past or perfective constructions rather
than present tense are much commoner in this usage. The following (non-
random) example from the corpus is absolutely typical, and again confirms
the normally negative semantic prosody of the construction:
A study in Canada has linked warm winds – called chinooks – to

migraines in some sufferers.
Notice too that in the corpus the following preposition of choice is to and
not with: it is as if The Times is a little uncomfortable with the explicit
causality implied by the standard ‘directional’ preposition (which would
also require a change in order of entities: Hague links murder rise to
Labour) and elects to ‘temper’ the connection by a semantically weaker
and vaguer choice, with (weaker yet, and not a preposition at all, would
have been and).
We have already seen that the small differences between these news
treatments are arguably as important as larger ones. To mention another,
consider contrasting use here of the phrase political football. While all five
papers report the phrase ‘political football’, in four papers the Taylor
parents are said to have accused Mr Hague of using their son’s death (a
nominalization) as a political football; but in the Telegraph, the only fully
Hague-aligned paper in this group, the parents are said to have accused
him of using their son (a mere nominal entity) as a political football. If that
characterization makes the parents’ criticism marginally less coherent and
retrievable, it thereby makes Mr Hague’s position more justified.
Fascinatingly, in News International’s other British national newspaper
besides The Times, the tabloid Sun, the Hague/police numbers story does
not figure at all on the 19 December: it is as if the furore simply has not
happened. And this extraordinary fact serves as a reminder of how these
news stories, and the further narratives they give rise to, multiply and
spread in complex ways. Thus these stories are not just about crime and
policing and race and Mr Hague; they are also ‘about’ entirely undeclared
characters and events, only apparent to the insider–reader, such as the
approaching general election and Mr Michael Portillo. Mr Portillo is Mr
Hague’s chief rival for the leadership of the Conservative Party and,
assuming that party’s success at some future general election, for the
prime ministership. But the pro-Tory press is said to be divided over the
competing claims of Hague and Portillo, with the Telegraph remaining
pro-Hague while the News International papers, The Times and the Sun,
have become more receptive to the idea of a Portillo takeover. This back-
ground may explain the way that The Times treatment of Mr Hague’s
intervention is comparatively lukewarm, while as noted, the Sun chooses
not to feature it at all.
One of the things to conclude is that incremental sociopolitically
embedded national stories such as the Hague/policing one are a good deal
more nuanced and heterogeneous, incorporating parts of other important
ongoing politicized narratives, than news reports of outrages and disasters
in distant settings. The average British reader’s connection to a Zimbabwe
news item, even where this reports actual deaths in a blatantly ideologi-
cally-motivated way, remains less direct than in relation to British or
European news, in terms of political, social, military, or economic involve-
ment.
A further general conclusion that may be advanced is that detailed
linguistic analysis shows the news media struggling to narrate the news
they relate – particularly where that news concerns political initiatives
and responses (rather than, say, accidents and disasters). What exactly
did Mr Hague say or do, that needed reporting on 11 December? Each
newspaper’s editorial team has to fashion an answer to this question, in
setting about telling the story. Despite or because of their position as
privileged witness and narrator, the newspapers themselves grapple with
a kind of narratorial and discursive ‘turbulence’ as they strive both to
report with accuracy and at the same time articulate a reasonable evalua-
tive construal of events. The turbulence is reflected in the incongruous
usages and idiomaticity tensions of the headlines and lead sentences,
where expressions heavy with clashing presuppositions or discordant
semantic prosodies are yoked in awkward collocation. There, despite the
great care and effort expended on word- and phrase-selection, we have
found a powerful public figure reported as ‘defying a plea’, ‘inflaming
the anger’, making a ‘race jibe’, ‘inflaming the tensions’, and so on. And,
it should be emphasized, these collocational incongruities in headlines
and lead sentences are not instances of the intentional or unintentional
humour that are often noted. Rather they are Bakhtinian clashings or
wrenchings of verbal patterns which it seems reasonable to trace to
inconsistency of narratorial stance, or turbulent conjunction of voices
and evaluations. They arise as journalists contrive to ‘get the story
straight’ but remain aware that there are always at least two sides to any
story.
8.5 The linguistic apparatus of political construal: notes on

key resources
Here is a brief review of some of the key linguistic phenomena it seems
sensible to examine when analyzing the political orientation of a narrative.
Most are lexicogrammatical systems where choice of formulation, or
‘slant’, is possible; they can all contribute to the discoursal variations,
transformations, or alternations to be found where different treatments of
a news item are compared.
1. Transitivity. Following Halliday’s account of clause transitivity as the

representation of reality (4.9), it is reasonable to look at just which entities
are presented as participants in a text’s representation of events. Which
individuals or groups tend to be cast as agent (sayer, thinker), and which
tend to be cast as affected medium, in the text (see also Passivization,
below)? What kinds of process are particular protagonists reported as ini-
tiating? And which relevant parties are scarcely mentioned at all?
Consider the protest against the Seattle WTO talks: was this chiefly a
physical battle (as the abundant television coverage of police–protestor
confrontations and the now-standard designation as The Battle of Seattle
would suggest) or a conflict of wills, or more one than the other? Similarly,
are protests against foxhunting, or against genetically-modified experimen-
tal crops, as essentially physical (as distinct from mental and ethical) as
they are often represented? The media, in their pursuit of exciting copy
and a more dramatic portrait of the day’s happenings, are surely likely to
represent matters in a more material and less mental way than might be
the case. The success or failure of various campaigns (political, military,
commercial) may have even more to do with winning hearts and minds
than we tend to think. And it may be that the dominant party seeks to
reinforce whatever physical/material advantage they have, by their empha-
sis on that advantage in their statements (= stories) to the press, so that
mental (and verbal) opposition is worn down. These are highly speculative
comments, but they seem to be important issues to do with textual presen-
tation – and they are issues that a Hallidayan anatomy of transitivity
enables us to explore.
Besides the processes and participants of transitivity analysis, it is often
crucial to examine that content that a text casts as the third and back-
grounded component, circumstantial elements. Circumstances can extenu-
ate or incriminate, and they often have more to do with underlying causes
than is first apparent. In the case of the controversial sinking of the
General Belgrano during the Falklands War (2 May 1982), a number of
surrounding circumstances that might have influenced the public’s evalu-
ation of the action were absent or variably presented in initial reports. For
example, just where was the Belgrano in relation to the British 200-mile
exclusion zone; in which direction was it steaming; how close to British
ships was it at the time of the attack; how near to it were other Argentine
naval vessels (which might have rescued the crew); and how long did it
take to sink?
Deciding whether surrounding circumstances are ever indirect causes is
often difficult, and a circumstance that is actually foregrounded in a head-
line is definitionally no longer neutral background: at the very least, to
highlight a circumstance is to suggest that it is rather less tenuously
involved in whatever incident has occurred. All of these implications are
strengthened by the use in headlines of such conjunctions as as. Recall The
Times headline discussed by Trew:
RIOTING BLACKS SHOT DEAD BY POLICE

AS ANC LEADERS MEET
As rarely explicitly expresses a causal connection in headlines, usually

serving as the most succinct means of reporting temporal connection or
simultaneity; but since it does elsewhere express causal relation, readers
may attribute a residually causal emphasis even while assuming that the
‘declared’ interpretation is temporal.
2. Passivization, especially with agent-deletion. Passive voice sentences

are a significant representational variant, which need to be seen alongside
their closest counterpart, the active mood sentence. For example, an initial
representation of a key event may use a transitive material process clause,
that is to say, a clause containing a physical process done by one partici-
pant to another. Active voice, the normal and simpler ordering, places the
agent as subject, followed by the process undergone, followed by the
affected entity as object, e.g.
Police shoot Africans

Youths stab boy
A common variant of this is the complex transitive clause where an

attribute or condition of the affected, perhaps arising as a result of the
process stated, is also mentioned:
Police shoot Africans dead

Youths stab boy to death
But a further alternative, often with a distinctly different effect, is the

passive construction. By passivization the affected participant is brought to
the focal subject position in the sequence, and the semantic agent can
optionally be deleted:
Africans (are) shot dead (by police)

Boy stabbed to death (by youths)
With reference to informativeness, the reformulation says less: we no

longer have an indication of the cause or agent of the process of shooting.
But it does thematize (bring to the front) the most affected or changed
participants.
3. Suppletion of agentless passives by intransitive clauses. An (agentless)

passivized or complex-transitive clause can be supplanted by an
intransitive clause relatively smoothly. Typically, both clauses will have
the structure S–P–(A), and while there may have to be a change of lexical
verb, the new verb choice can be close in meaning to the original:
Africans shot dead (in Salisbury riot)

→ Africans die (in Salisbury riot)
Boy stabbed to death on crime-ridden estate
→ Boy dies on crime-ridden estate
The affected participant formerly in object position is now the sole stated
participant, occupying subject position, and the former description of a
causal relation, what x did to y, is now simply a report of what happened
to y, or even, of what y ‘does’.
4. Nominalization. A nominalization is a conversion and encapsulation of

what is intrinsically a clausal process in the syntactic form of a noun
phrase, hence treating the entire process as an established ‘thing’ – which
can then serve as a participant in some other more directly reported and
inspectable process. Nominalization ‘de-narrativizes’ a process, making
the process mere background to a product or thing. This formulation:
Damilola Taylor died of stab wounds
is a process and a narrative. But the following equivalent
The death of Damilola Taylor from stab wounds
is a nominalization, assuming a narrative but not telling it. Similarly, in the

Salisbury riot story –
The deaths of 13 Africans
– nominalization can attenuate the sense of ‘shooting dead’ or ‘dying’ as

experienced processes. The reformulations deflect the reader’s attention
further from questioning whether these deaths were killings or not, and if
so who the killers were. Nominalization transformations such as this recast
an implicit process into the form of a static condition or thing. This
nominal condition or thing can then be used as the agent or affected par-
ticipant or carrier of some other process, now become the focus of our
attention:
The Taylors had accused Mr Hague of using their son’s death as a

‘political football’.
The deaths of 13 Africans triggered a further wave of violence in Salis-

bury townships today.
Nominalization is one of the crucial linguistic resources deployed in news
reports. But as the examples above suggest, it can also be exploited and
abused: it enables the user to refer without narrating, without clear and
explicit report. The teller can be economical with the facts (as they see
them). Nominalizations are exploited, used as sword and as shield, by every
political and ideological faction or persuasion. In the explicitly political
arena, they often serve to contrive implicit or explicit transfers of respons-
ibility – unsurprisingly, in view of the fact that varying answers to the ques-
tion, ‘Who is to blame?’ lie at the heart of much political discourse. It is not
that nominalization is inevitably ‘wrong’ or undesirable – it is an invaluable
means of textual condensation, e.g. in academic and scientific writing. But it
clearly can be used, in barely perceptible ways, to background what
arguably should be in the foreground. In some ways nominalization is more
threatening than the ultimate kind of backgrounding, declining to report a
story at all: nominalizing reformulations are still a kind of reporting of what
happened, usually with most of the main participants noted, and the ‘hard’
facts (numbers, times, places, etc., involved) recorded correctly. Given that
degree of accuracy, busy readers may not have the resources to reflect on
the lexicogrammatical slantedness of the interpretation.
On the other hand, readers busy or idle are not dupes or stooges: they
are often adept at sniffing out the linguistic and communicative manipula-
tions that newspapers perform. Readers of every political persuasion can,
given time and motivation, reinterpret the incidents that their newspapers
interpret for them. But such reading against the grain is effortful and, in
the long run, rather a perverse exercise in indirectness. Even the most
active of readers is likely, in the longer run, to be habituated and ‘recon-
ciled’ to the categorizations and evaluations constantly relayed to them by
the news outlets they patronize. Cumulatively, through a lengthy text, and
across a series of stories filed over days and weeks on the same issue, and
given the likely congruence of a newspaper’s treatment of the particular
issue with their treatment of a broader grouping of issues, the effect of
enforcing and reinforcing a particular view of the world can be almost irre-
sistible. George Orwell’s novel 1984 dramatizes these efforts of world-view
construction and reconstruction at a tyrannical extreme.
5. Modality and evaluation (see 3.7, 6.4 and 6.6). Within systemic linguis-
tics more recently, the systems of lexical resource available for conveying
evaluation have been brought together under the cover-term of Appraisal.
Appraisal theory postulates that, by means of complex networks of vocab-
ulary expressing various kinds of appreciation, judgment, and affect,
speakers can ‘encode’ their interpersonal evaluations of the subject-
matter. For detailed accounts, see Martin (2000), White (2000) and Eggins
and Slade (1997).
6. Namings and descriptions (see also 4.4). Consider again Trew’s news
stories about riots and killings (we can note, relevantly, that the Salisbury,
Rhodesia of the 1970s has become Harare, Zimbabwe). Were those who
acted rioters, demonstrators, or troublemakers; were they a unified group or
was there a ruthless, violent minority amongst the majority; were those
actually shot representative of the entire group, or innocent bystanders, or
ringleaders, or what? Notice, in passing, the contrasting evaluations carried
by the words leader and ringleader. Were the police ordinary police or
special police, black or white or mixed or both-but-stratified. If trained to
deal with riots, trained by whom, with what objectives? How many police
actually opened fire, were they young, nervous recruits, or hardened old
stagers? All the foregoing questions are probing both the facts and, by
implication, their proper reported description; and all descriptions carry
some interpretative and political charge. Parallel scrutiny of facts, namings
and descriptions can be directed at the characterizations of the protesters,
delegates and police in the ‘Battle of Seattle’ story, or indeed of Mr Hague,
the police, and Damilola’s attackers, in the Hague/policing story. How can
or should Mr Hague have been named, for example? In one local South
London newspaper’s version of the Hague/policing story (SouthLondon
Online), Mr Hague is referred to as the Yorkshireman, thereby presumably
intending to intimate that he is ‘an outsider’ and perhaps ill-informed. In
the most salient areas of political contention, journalists – like the rest of us
– have rich inventories of overlapping descriptors to choose from as they
sort out which characterization fits their (ideologically contextualized)
account of things best. Some sets of variant description are particularly
familiar, and often contain noticeably positive or negative evaluation: cf.
terrorist versus freedom fighter versus gunman versus men [sic] of violence;
question versus allegation; reply versus rebuttal; answer versus refute;
opinion versus allegation; opinion versus fact; answer versus justification;
national security versus government cover-up; policy versus expediency.
At the same time it is quite possible for there to be lexical ‘gaps’ in the
vocabulary standardly available for describing events and participants.
Take the case of workers going on a one-day strike: how might the strike’s
proponents best describe their action? One common description, ‘day of
industrial action’, seems particularly ill-advised, a kind of ready-supplied
ammunition for critics who, quite reasonably, point out that such occasions
are days of industrial inaction. Alternatively, nominalizations like ‘one-day
stoppage’ (which gives no kind of purpose to the action) and ‘one-day
strike’ draw some attention to those doing the striking, but fail to impli-
cate any other party or circumstances as underlying causes of the stop-
page. Some unionists like the emphasis on workers as powerful, as having
some control over production, that phrases like ‘industrial action’ and ‘day
of protest’ suggest. But that in itself may be bad strategy if your narrative
is to be taken up by a hostile press. There really does seem to be some-
thing of a lexical-descriptive gap here, with no standard pithy expression
that would represent the one-day stoppage as a demonstration, by suspen-
sion of work (and remuneration), of a collective sense of deep grievance:
perhaps the phrase ‘day of hardship’ would come close.
Over the past fifteen years, public discourse has become even more self-
aware, under the pressures of monitoring by one interested group or
another, who wish things to be named (represented) a certain way and are
prepared to contest namings in some other objectionable way, and who
are furthermore able to wield some kind of ‘capital’ (publishing house-
style; or sociocultural support from liberal-progressive opinion; even legal
rulings) to uphold or exclude particular namings. Some of this has (itself)
been dismissively labelled as political correctness; but it is a broader and
more important process than that tainted phrase suggests: in a thorough
study, Cameron (1995) calls it verbal hygiene. Whatever the label, it is a
discoursal self-consciousness that is thoroughly characteristic of
contemporary Western culture, where ‘spin’, branding, image, surface, the
‘look’ (of the model, the Web page, the gesture) is deemed so important
in, as before, winning hearts and minds (if not also lawsuits).
Besides considering single lexical points in a text, it is important also to
look at the lexis of stories more holistically, and consider what kinds
predominate. Is the lexis abstract or concrete, and in either case, which
kind of domain is it standardly used to refer to? Matches and mismatches
(intentional or otherwise) between the basic message to be conveyed and
the language used to convey it can rapidly be uncovered in this way (e.g.
advocating peace and reconciliation but using the language of force and
conflict – smashing through ignorance by mobilizing the forces for change).
7. Collocational incongruity. Under this label may be gathered all the

demonstrably atypical or infelicitous constructions – clearly having most
impact in headlines and leads – of the kind discussed towards the end of
8.4 and thus not elaborated here. By contrast with some of the other dis-
course-transformative resources discussed in this list, these clashes of
idiom or usage may well not be deliberate, even though, ironically, they
seem to arise after much careful deliberation over individual word selec-
tion. Mr Hague is defiant so that the verb defies is selected; the Taylors’
request for their tragedy not to be used in a policy debate is nominalized
as a plea. But the chaining of these in text gives rise to the incongruity of
a powerful person defying a plea, possibly reflective of ideological and
narrative turbulence, amenable to Bakhtinian explanation as a clash of
voices.
8. Presupposition. This is the term used to describe a speaker’s back-

grounding, in their utterance, of various kinds of assumptions that are nev-
ertheless retrievable from that utterance. Thus if I say the car battery is flat
again I presuppose that the battery has been flat before. Even if the sen-
tence is cast in the negative – Thankfully the battery isn’t flat again – the
presupposition remains true (the battery was flat before). See Chapter 8 of
Toolan (1998) for a simple literary-minded exposition, Simpson (1993) and
Levinson (1983) for a detailed linguistic account. The most strategic use of
presupposition in public or political discourse is so as to ‘insert’ into the
record (but in the background, as if they were uncontroversial facts,
common ground) propositions that are contestable or slanted. To take two
examples from the Daily Telegraph story discussed above, when this
reports
[Mr Hague] said he would not be bullied into silence on inner city
crime, falling police numbers and morale.
it relays without objection Mr Hague’s presupposition that somebody has

been intent on bullying him into silence on inner city crime, etc. Notice
by contrast that the Independent questions Mr Hague’s presupposition by
using scare quotes:
[Mr Hague] warned that he would not be ‘bullied’ into dropping the
issue.
On the other political side, the Telegraph also reports the Home Secretary
as saying ‘the Leader of the Opposition is playing the race card’ – a
formulation which presupposes that there exists such a thing as ‘the race
card’.
But presuppositions can shape controversial narratives in much less
explicitly political contexts, too. Suppose, for example, that someone has
tragically bled to death after having been stabbed by unknown assailants;
news bulletins might refer to the brutal murder. This presupposes that the
term murder validly applies – but in the circumstances this is prematurely
to categorize and in a sense to pre-judge the violent attack: there may have
been neither intent to cause life-threatening harm, nor recklessness.
The above are examples of relatively direct, linguistically-encoded pre-
supposition. Often as interesting to reflect upon are the cultural or societal
presuppositions that can be inferred from what members of a community
talk or write about, and from how they do so. The community or culture
involved need not be coterminous with nationality, but could be based on
a sport, music, a religion, a hobby – every imaginable grouping that might
sustain a monthly magazine or an Internet discussion group. Clearly, very
different kinds of things will be narratively ‘tellable’ (being significant, or
calamitous, or amusing, or a triumph, etc.) for each of these communities,
and strikingly different cultural presuppositions will form a backcloth to
the foregrounded tellable material. Here is the headline to an inside-page
news item in the latest issue of my university’s Library Bulletin:
Periodicals Information Point Moves
But in the more ‘encompassing’ media, such as national newspapers, radio

and TV, where single compendious texts contrive to address a relatively
heterogeneous audience, a greater degree of guessing about audience
interests, and of creation or shaping of audience interests, will be involved.
It is quite possible for a newspaper to attempt to ‘induce’ and nurture
certain presuppositions in its readers, which they would not spontaneously
subscribe to. Indeed to some extent competing commercial media have to
‘force’ reader presuppositions in this way: with each succeeding publica-
tion or broadcast, they have to project the meta-message that the stories
they are relaying are highly relevant and tellable.
This section should not end without a reiteration that narratives are
enabling props as well as insidious reconstructions. Accordingly, and
bearing in mind that ideological perspectivism in language-use is
inescapable, we should recognize the good side of news stories in the
press. Readers are afforded some interpretation, some halfway coherent
construction of what is past, present and to come in the larger world
around them. And obviously these constructions are supplied far more
rapidly and presented far more effectively than any individual reader
could manage on their own.
Hard-news narratives, we have seen, construe and reconstrue newswor-
thy facts and events (both awkward and convenient) so that these are con-
gruent with the ‘macro-story’ of a newspaper’s broader view of the world.
In the case of long-running news items, the stories form a sequence of
tellings in which changes of emphasis, over time, are likely. We cannot
always safely predict that a particular paper’s angle on an issue or story will
stay constant – as Trew’s example showed. But the world has moved on a
long way since the Salisbury killings of June 1975 that Trew discusses, and
it would be interesting to chart the shifts in The Times’s own broader
account of ‘the Zimbabwe story’. One might take shifts and revisions in the
presentation of just one participant, for instance, by way of illumination.
What sort of character, for example, has The Times construed Robert
Mugabe to be, over the quarter-century from 1975 to 2000? My impression
is that, along with many other conservative journals, The Times first cast
Mugabe as a Marxist ideologue presiding over senseless terrorist violence,
bent on the annihilation of the white factions in Africa, and impervious to
conciliatory dialogue. After the Whitehall conference, however, in which
ZANU (PF) participated – and especially after the substantially free and
fair democratic elections (1980) in which, to the surprise of many Western
commentators, Mugabe emerged a clear victor over his rivals – it was,
quite literally, a different story. Now Mugabe appeared reasonable after
all, educated and religious: his two Western degrees were emphasized, as
was his devout Catholicism. In the decades thereafter, there was a steady
cooling towards Mugabe and the new order in Zimbabwe as more stories
emerged of inter-tribal persecution and sharp control of political opposi-
tion. (Notice the use of nominalizations here, for events that undoubtedly
have agents and affecteds.) Present-day Times narratives depict an ageing
and unstable, quasi-dictatorial figure, presiding over economic chaos and
failure, fomenting racism in his attacks on political opponents, intent on
maintaining a one-party state within a democratic shell. All these shifts in
characterization suggest that either Mugabe has changed rather radically,
or The Times has, or – and this is the most likely – both have, in ways that
complexly interact. But insofar as it is true that The Times (the teller) has
changed its mind about Mugabe over the years, notice how different this is
from literary narration where, for example, we might expect Joyce’s
Gabriel or Nabokov’s Pnin to change and develop (drop some attributes
and gain others), but do not expect the author, in the course of the telling,
to change their mind. If Mugabe were a character in a novel, we would
regard such drastic shifts in presentation as reflecting an incompetent or
negligent narrator, who didn’t seem to know their own story (incompe-
tence different in kind from the ‘uncertainty’ of Simpson’s negative-
modality narrators). But this is one major way in which press narratives
differ from literary ones.
President Mugabe is still in power as I write; he is one of those extra-
ordinary political survivor–endurers, like Presidents Castro of Cuba, Moi
of Kenya, and Mandela of South Africa, Senior Prime Minister Lee Kuan
Yew of Singapore and former Prime Minister Thatcher of Great Britain,
General Pinochet of Chile and – a Mandela still awaiting release from her
Robben Island – Aung San Suu Kyi of Myanmar. A careful tracking of the
narrative construction and re-construction of any of these figures over the
years, in and out of office, is always likely to be revealing.
Unlike a literary narrative, press narratives are never ‘finished’: there is
always tomorrow’s edition, which may have to assimilate new and
awkward events, even to the extent of revising the newspaper’s back-
ground construction of events. So caveat lector: let the reader beware.
Newspapers – like governments – often do not know what they are talking
about. They may know no more than Bertha at the end of ‘Bliss’ wonder-
ing what is going to happen now; but unlike her they must at once provide
a satisfactory account. Their macro-narratives are interpretations of the
past and ‘authoritative’ predictions about the future, but the future is a
foreign country, ill-understood, so that reading newspapers for the truth is
a bit like religious belief – an act of faith. At the present time in Britain,
perhaps no narratives so confirm this as those fumblingly related, by politi-
cians and journalists alike, concerning the gradual but much-resisted
development, by federation, of these United States of Europe.
8.6 News stories online

Over the last ten years or so, a profoundly important development with
respect to print or text news media has emerged which must be factored
into any analysis of news narratives: the Internet, which, besides many
other things, is taking the paper out of newspapers. Increasingly, people
are accessing their newspapers – especially the broadsheets – via their
computer screens rather than a paper copy. The online version of any
newspaper is inevitably different in content (including fewer and different
accompanying advertisements) and format and reading experience than
the print version. It is likely, too, that the sequence in which a reader
‘samples’ the linked screens of text, of the online version, is different, and
less linear or even ‘orderly’, than in the case of the newsprint copy.
Equally important are all the hypertext possibilities of online media. I
will mention here only their simplest forms, involving hotlinked text, but
increasingly hypermedia allows the possibility of, for example, a text story
about the Prime Minister’s latest pronouncement to be accompanied by an
audiovisual file carrying a forty-second cache of the relevant speech; simi-
larly, reviews of music recordings, concerts, plays and new books can, in
principle, be accompanied by brief samples (audiovisual where applicable)
of the material being discussed. Sites such as BBC Online have been doing
this for some time, and were traditional newspapers to move in that direc-
tion in their online sites, it would be no surprise in this postmodern culture
of consumerist infotainment, and discoursal plenitude or excess (the infor-
mational plenitude that Chatman noted in the traditional narrative film –
4.8(j) – can be seen as quite modest and manageable by comparison).
At its simplest, then, hypertext possibilities mean that where, for
example, the Guardian runs an article on harassment of the MDC opposi-
tion party by the Government and ZANU (PF) in Zimbabwe, the editors
can choose to embed, within the article, hotlinks to a range of directly-
related, or more background items, which the reader can surf to. Equally
commonly, a menu of such links can be displayed in a sidebar running
alongside the news item. And those linked items themselves may contain
links to numerous further reports, profiles, opinion pieces, government
documents and so on. In view of which, several related questions arise:
what is the sequential structure of the online news report: should one
‘jump’ to linked pages – particularly if these are linked pages within the
same day’s edition of the newspaper – at precisely the point where that
link is offered in the body of the report, subsequently jumping back to the
original text? And if one does pursue some of the advertised links, at what
point do you know that you have reached the end of the news item? For
our experience of any of the larger news services’ online materials is that
in practical terms these are, as the Guardian site explicitly declares,
‘Unlimited’. For such reasons it is quite possible to conclude that the
Internet, like a multi-purpose library and leisure complex, is in many
respects an anti-narrative device, exploiting the technology that banishes
several desiderata of narrative listed in Chapter 1: teleology, a critical
high-point, an individualized agonist, a palpable teller, and closure.
There are, of course, ‘more narrative’ (in the adjectival sense) places on
the Internet, such as the individual’s Web pages, particularly where these
have remained stolidly unmodified for months: here we do sense individu-
alization, a depicted teller, and closure, but no teleology or genuine crisis
point. And every ‘complete’ online edition of a newspaper would appear
to provide closure for that day at least, just as the purchased newsprint
version does: but even that ‘temporary’ closure is undercut by the online
newspapers’ increasing provision of ‘breaking news’ items, posted on the
site as soon as possible: thus, checking the Guardian at 10:50 a.m., I find a
new item of breaking news has been added as recently as 10:46: the ants in
the discourse colony that is the online news site are always coming and
going (but are not the same ants).
The emergence of a huge array of competing TV channels in Britain
and elsewhere is representative of a greater diversification and, import-
antly, confusion of the media that shape public opinion and sentiment.
The diversification and confusion take many forms. One example concerns
newspaper-allegiance: in the past someone might subscribe to the Daily
Telegraph and have it delivered to their door every day, as a result deriv-
ing their picture of, say, Zimbabwe or the leader of the Opposition’s state-
ments solely from this consistently conservative source. But the
contemporary online reader of the Telegraph is in a very different posi-
tion: by means of a few clicks and without moving from their seat, their
computer can jump to the Guardian news pages, or the Mirror’s, or CNN
or indeed Zimbabwe’s Daily News (http://www.dailynews.co.zw).
As noted, online newspapers today are far from offering merely the
‘bare text’ of the news item: even where it does not offer audio or video
links, the online page will bristle with panels and sidebars, carrying com-
mercial advertisements (usually more colourful) of all kinds – sometimes
grotesquely at odds with the content and mood of the news story – as well
as the previously-mentioned links to other parts of the current online
issue, or to items in past online issues of the newspaper, or to external
websites. In the Telegraph’s version of the Hague/police numbers story, a
sidebar on the left offers links to a number of external sites (listed below),
and arguably these contribute to the evaluation or ‘uptake’ of the story
they frame. The editorial teller intimates that these pages are part of the
background to the narrative, tellable if the addressee believes they need to
be told about them: in a sense, each offered hotlink of this kind approxi-
mates an Abstract, a point at which, by clicking or not clicking, the
addressee can be told the supporting narrative, or not. But from the point
of view of the ‘centered’ story, items in these sidebar menus are, even
when not clicked on and expanded upon, highly evaluative. Consider the
selection, and presentational order, of the organizations listed as links in
the online Telegraph’s story, including the preponderance of police organi-
zations:
Conservative Party
Labour Party
The murder of Damilola Taylor – Metropolitan Police
Racial and Violent Crimes Taskforce – Metropolitan Police
Police – Home Office
Metropolitan Police Federation
Police Services of the UK
Police Federation of England and Wales
8.7 Stories of class and gender
Besides race, the other two most universal axes along which division,
exclusion and ideologically-motivated discrimination proceed are gender
and class. To put this starkly, in Western culture and elsewhere, narratives
of all kinds have tended to ‘tell’ a story about the inferiority or subordin-
ation of women to men, and of the working class to the middle class. In a
sense that outcome is an inevitable corollary of a culture embracing and
putting store by these gender and class categories: part of the reason for
subscribing to the categorization of people into working class and middle
class is so that you can tell stories that assert differences and asymmetries
between these two groups. So classist and gendered narratives perform-
atively sustain the class- and gender-oppositions which their tellers and
addressees wittingly or unwittingly endorse and live by.
If class- and gender-oppositions were genuinely ones of identities that
were merely ‘different but equal’, like, perhaps, a person’s preferring tea to
coffee, the consequences might be slight. But in sociocultural practice,
gender and class differences are robustly hierarchical, contriving to empower
one group by dint of disempowering or marginalizing others. As countless
stories imply. Again using a Hallidayan analysis to specify the choices of
process and participant that have been made in relevant narratives, we can
ask with whom lies the power to do, to cause, and to affect? Asymmetry will
soon be apparent, whether the texts studied are nineteenth-century British
fiction or Hollywood blockbuster action movies of the last twenty years or
contemporary television commercials. These trends do not hold true for all
narratives, of course, and they have probably weakened as tendencies over
the last few decades; but they remain powerfully present.
To the extent that the above generalizations hold, it seems that such nar-
ratives simply replicate mainstream reality and main-stream ideology. But
there are numerous departures – and ways of departing – from the narra-
tive-ideological norms, whether we are looking at published novels or per-
sonal stories embedded in conversation over lunch. Such ideologically
non-standard narratives are like ‘minority reports’ on the way things are
(or were, or might be). They express a local sub-group’s preferences as
being at some degree of variance from the larger society’s assumed domi-
nant norms. And although such local contestings of the sociopolitical ortho-
doxies are not the norm, still they may occur anywhere; but they may fail to
be sharply visible because those contestings are themselves heterogeneous
syntheses of normative positions on some issues and radical positions on
others. For example, a story might contest orthodoxies of class and race dis-
crimination and yet remain covertly sex-discriminatory (or patriarchal),
and so on: using the list of linguistic features in 8.5 as a guide, it is possible
to identify the aspects of a discourse which represent one position, and
those which articulate a perhaps conflicting one. Limitations of space
prevent a fuller exploration and demonstration of these issues here.
Such limitations also forestall discussion of the various other respects in
which we can talk about gendered narratives, and the desirability of a
feminist narratology (Lanser, 1986; 1995; 1999; Warhol, 1989; Mills 1998a,
b; Fludernik, 1999b). Is the Labovian ‘high point’ narrative structure, with
a mounting action rooted in conflict and leading to an abrupt climax a
‘masculinist’ format? If so, can it be contrasted with a preferred feminist
story structure, equally ‘valid’: an unfolding of events, in waves of progres-
sion, that more clearly brings the story and its audience back to its point of
departure, in a more circular fashion? From time to time analysts have
speculated along these lines, and various individual narratives can be
adduced in support (one might cite Michelle Roberts’ Une Glossaire/A
Glossary, briefly discussed at 1.3; or stories by contemporary writers such
as Margaret Atwood, Grace Paley, Jeanette Winterson, Ann Beattie,
Carol Shields, Jane Smiley, and Alice Munro – see discussion of her
‘Circle of Prayer’ in the Notes and exercises below). In the modern period
of English literature, much ‘structurally female’ writing can be traced to
Virginia Woolf; Cather, Chopin, Richardson and Mansfield may also be
mentioned.
Different again are questions of gender ‘in’ the written narrative. For
example, where the gender of a story’s author or third-person narrator is
unknown, to what extent do readers like to determine this, or simply make
assumptions, and if the latter, on what textual bases? Or where a first-
person narrator figures within the text, with gender references excluded or
suppressed, again how important is this to the addressee and how do they
go about ‘resolving’ the ambiguity by reference to textual content? As
indicated earlier, race and gender and class are some of the most ingrained
categories that we live by, but all these and others are now widely recog-
nized to be culturally-created and contingent, rather than biological
essences inherited at birth. Accordingly, in literary narrative especially,
contemporary authors often challenge or deny any expectations readers
may have that a character will be unambiguously presented as female,
middle-class, heterosexual, and black (or, in each separate respect, not).
8.8 Prejudice in ethnic narratives

The sociopolitical partiality of certain types of everyday discourse emerges
very vividly in van Dijk’s (1984) studies of ethnic prejudice in cognition and
conversation. Again, limitations of space prevent a proper treatment, but
this work is widely reported elsewhere. Suffice it to say that, in the dis-
courses of mainstream European cultures, minority ethnic groups (like
homosexuals, immigrants, refugees, asylum-seekers, and possibly religious
fundamentalists) are construed in the way that the male middle-class con-
strues women, only much more so. In Holland, van Dijk found that the con-
versational stories involving minority ethnic groups told by majority-group
members characterized these ‘foreigners’ as culturally and even physically
threatening, alien and strange, incomprehensible, irrational, untrustworthy,
irreligious and prone to criminality. Such crude and consistent stereotyping
means that a Proppian folktale analysis, even without much amendment,
fits the negative stories told about minorities:
The Dutch people are the heroes or the victims, and the outgroup
members are the villains. Even the events and the actions involved
may be drawn from a limited repertory. Minorities cause us trouble of
various kinds: they make noise, cause dirt in the street, take our jobs
or houses, or are engaged in crime. Minority stories are becoming a
specific genre of the folklore of ingroup prejudice.
(van Dijk, 1984: 81)
Amidst this dreary predictability of content, one unpredictability of struc-
ture was interesting: fully half of van Dijk’s corpus of stories lacked a solu-
tion or resolution to the problems raised in the complication.
And that is precisely how the ingroup storyteller sees the social con-
flict: minorities cause problems, but we cannot do anything against
that. Instead of a heroic success story, we then have a complaint or
accusation story.
(90)
If such narratives are, indeed, geared to generalized complaint, then they
have some of the properties of what systemicists distinguish from Labovian
narrative as exempla (see 7.6). As exempla they facilitate negative evalu-
ations that can, if needed, be cast in generic sentence form – ‘they can’t be
trusted’; ‘they send all their money back home’ – and these in turn can be
grounds for deontic prescriptions about what should be done (with them).
The power of the prejudice in ethnic narratives persists today in Western
Europe and beyond, in diverse forms. Important studies continue to appear
on the racialization of political discourses, narrative or otherwise (e.g. van
Dijk, 1991; 1992; 1993; Blommaert and Verschueren, 1998). In recent years,
Wodak and her co-researchers have turned their attention to a critical dis-
course analysis of European discourses on immigration, and the virulently
anti-immigrant far-right political discourse of Jorg Haider and his Freedom
Party, currently part of the government coalition in Austria (see e.g. Wodak
et al., 1999; van Leeuwen and Wodak, 1999). In Britain, arguably, the accept-
able (at least, legal) face of racism today is the ‘europhobia’ to be found in
some of the tabloid newspapers’ treatment of European Union stories.
8.9 Stories in court

A final arena in which I want to draw attention to the political dimension
to assessments of tellers and their stories is that of the law. That, contrary
to official assumptions, not all tellers are equal under the law is highlighted
in the work of O’Barr and his co-researchers. Lind and O’Barr (1979), for
example, report the findings of a detailed analysis (using actual US court
transcripts) of the different kinds of language used by witnesses. Their first
finding was that
Many of the most common and important dimensions of variation in
court speech seemed to carry information about the power, status, and
control of the speakers.
(Lind and O’Barr, 1979: 70)
In particular, speakers seemed to divide rather noticeably into two camps:

those who used a way of speaking with most or all of the features of what
has been called ‘women’s language’ but is better termed ‘powerless lan-
guage’, and those who adopted a way of speaking that is socially recog-
nized as ‘powerful language’. Powerless language style, first identified in
Lakoff (1975), is marked by the following features:
a high frequency of intensifiers like so and very; of ‘empty adjectives

like cute and charming and ghastly; hyper-correct grammar; polite
forms (would you like to . . .?); higher frequency of accompanying ges-
tures; more hedges (well, I guess); rising (question-type) intonation in
declarative sentences; greater pitch range and more rapid pitch
changes (heard as more dramatic than the powerful male’s drone).
(Powerful language is defined negatively, as speech that is largely devoid

of the above features). Correlation of social position with use of powerful
versus powerless language was unmistakeable: witnesses who were pro-
fessionals, people in authority, of either sex, used few or no powerless
language features; but low social-prestige witnesses, including the
unemployed, did. The researchers now re-recorded actual powerless-
speech testimonies in four versions: versions with and without the power-
less-speech features, and with male and female actors as speakers. When
these versions were presented to judges under controlled conditions, they
found that for both the male and female witnesses, the power speech pro-
duced perceptions that the witness was markedly ‘more competent,
attractive, trustworthy, dynamic and convincing than did the powerless
speech testimony’ (Lind and O’Barr, 1979: 72).
Lind and O’Barr went on to observe that witnesses’ testimony tended
to vary in another, immediately evident, way: in the length of a witness’s
response to a lawyer’s question. Some witnesses supplied (and were
allowed to supply) lengthy narrative answers to relatively few lawyer ques-
tions, while others produced fragmentary testimony – brief answers to
numerous lawyer questions. Both intuition and courtroom manuals (which
advise lawyers to have their own witnesses give them narrative answers
but to restrict opposition witnesses to fragmentary ones) suggest that nar-
rative answers imply trust in a witness. That trust and evaluation is of
course the lawyer’s, but there seems to be plentiful evidence that subjects
in the role of jurors will, to some extent, accept and adopt a lawyer’s assumed
evaluation of a witness. And, of course, this encapsulates a major goal of a
lawyer’s examination of witnesses, in the adversarial system: that this witness
be accepted as truth-telling, that that witness for the opposing side be
discounted as dishonest or unreliable, and so on. In short, as barristers have
long known informally, the manner in which witnesses and attorneys speak in
the presentation of testimony – the manner and format of narratives in court
– can affect social and even judicial evaluation of them by jurors.
It is possible to treat criminal cases, in particular, as situations in which
two pre-eminent tellers, the counsels for the prosecution and the defence,
are the architects of two extended and partially conflicting narratives.
Each of those narratives is a collaborative effort, with attempts at valida-
tion and corroboration from perhaps a number of witnesses, and each is
also a fragile and threatened narrative, in that witnesses’ assertions may be
challenged and probed under cross-examination. But the two leading
counsels are undoubtedly chief narrators – particularly during cross-
examination, where both sides are so intent on ‘putting their case’ as a
coherent narrative that witnesses’ denials are largely disregarded by the
lawyer involved: a semblance of question-and-answer dialogue frames
what is essentially narrative monologue (see Heffer, in preparation). See
also the work of Atkinson and Drew (1979), which shows how subtle can
be the negotiations of description, evaluation and inferrable intention
during cross-examination of tribunal witnesses.
The upshot of hearing both the case for the prosecution and that for the
defence is that judge and jury are presented with two clashing narratives,
possibly with partially divergent sequences of events, and contrasting
claimed motivations, resolutions and evaluations. Judge and jury have to
attempt a sifting of these two narratives, a merging which may involve
rejecting very many of one party’s assertions regarding the events, charac-
ters, setting and motivations. The particular goal that will delimit these
activities of sifting and merging is that of upholding or rejecting a specific
charge that certain individuals are guilty of criminal wrongdoing. Matters
are not quite this straightforward, however, since, as Heffer (adopting
Bruner) shows, somewhat orthogonally to the sequential, discursive, every-
day narrativizing of the facts of a case lie the legal rules and principles,
which form an equation or paradigm. Thus theft is the (a) dishonest
(b) appropriation of (c) property (d) belonging to another (e) with intention
permanently to deprive them of it, and if those five elements fail to be estab-
lished, then the narrative is not a ‘theft’ narrative and the criminal charge
will fail (even if some other narrative, such as ‘transfer of property using
undue influence’, seems to fit). Broadly, the law applies paradigmatically;
but lay jurors, it is postulated, proceed syntagmatically or narratively. In the
course of their ‘narrativizing’ of the case, barristers and the judge (who in
English courts provides a summing-up), must negotiate a difficult integra-
tion of the paradigmatic and the narrative (see Heffer, in preparation).
One interesting feature to notice is the way witnesses are required to
state only those things they know to be the case, and to avoid airing opin-
ions or judgments (except when specifically asked to do so, as in, for
example, ‘Did the defendant appear troubled or anxious at that time?’). In
other words, judicial procedure attempts a separation of what Labov has
distinguished as the referential and the evaluative: witnesses are to
provide unevaluated referential reports; lawyers incorporate these in their
submission of evaluated, point-laden narratives; and judge and/or jury
decide whether to accept those narratives, in part or in whole. The
persuasive but inherently suspect narratives (suspect given that, as every-
one is aware, they cannot both be true) that opposing lawyers weave are
always constructed with the end in view of compelling either assent to or
dissent from the conclusion that the jury ‘are sure that the defendant is
guilty’ of the crime. Permeating these narratives are complex attributions
of blame, responsibility, and negligence. What applies to courtroom narra-
tives applies to a large degree to coroner’s inquiries and legal tribunals as
well (one may predict that a rich array of kinds of narrative inequalities
and disvaluations are likely to be found in the discourses of employment,
immigration and mental health tribunals).
Like the analyses of O’Barr and his co-researchers, those of Ruth
Wodak (1985) address the question of the glaringly different consequences
– for example, in terms of sentence – that different strategies of self- and
testimony-presentation may have in, for example, trials of motorists for
dangerous driving – where ‘different modes of class- and sex-specific
socialization’ (1985: 184) and narrative ‘effectiveness’ militate against
female working-class defendants of limited education.
And what local or extended metaphors are invoked to give shape and
vividness to narrative representations in the criminal court? I will cite just
one that span the wrong way. The prosecution lawyers in the O. J.
Simpson criminal trial, in their final summing up to the jury, drew heavily
on the ‘jigsaw’ metaphor, suggesting that over the preceding months
various witnesses had brought ‘pieces’ of the puzzle to court, and that the
complete picture could now be assembled (Cotterill, 2000). The defence
team very effectively ‘turned’ this metaphor by reminding jurors that in a
typical jigsaw puzzle, the player is guided by the picture on the front of the
box, available in advance of all efforts to put the pieces together. And that,
they contended, was precisely how the prosecution had preceded: first they
decided on the picture, one in which Simpson was pre-judged as the mur-
derer, and then they sought the pieces to ‘fit’ that picture. This rejoinder
may have had only minor impact in this most atypical of trials, but it could
only strengthen the defence case.
Finally, linguistic analysis of narratives in the (chiefly criminal) justice
system has most prominently established itself as part of ‘forensic
linguistics’, one of the fastest-growing new fields of linguistics (Coulthard,
1994, 1996, and in preparation). When in 1952 the state convicted and later
hanged Derek Bentley, a learning-disabled young man, on the grounds
that he was the knowing accomplice of Chris Craig, a minor, who shot and
killed a police officer, were there irregularities and reasonable doubts
about the written narrative statement of events that the police took from
Bentley? Was Bentley’s narrative all his own words, without police
additions or ‘scaffolding’ via leading questions? Making proper allowance
for the stressful circumstances in which such narrative statements are
made and taken, and with due consideration for whatever may be distinc-
tive about a particular detained person’s narrative idiolect, are there reli-
able indicators in structure, word-choice, and so on, that can distinguish
genuine and uncoerced statements from those where there may have been
interference and fabrication? In the Bentley case, on posthumous appeal
in 1997, the 1953 conviction was ruled unsafe (largely on the grounds of
the original judge’s biassed summing-up, but the forensic linguistic evi-
dence concerning Bentley’s statement was not discounted). Linguistic
analysis of the narrative portions of suspect documents – statements in
custody, witness statements, suicide notes, anonymous blackmail or threat-
ening letters, etc. – is becoming increasingly robust and reliable, and the
courts’ growing attention to it reflects this. The stakes could hardly be
higher. Nowhere is it more palpably evident that, individually and collec-
tively, our narratives are our life stories.
Further reading
An excellent place to begin with is Fowler et al. (1979) which includes the work by Trew
that I have discussed; Bolinger (1980) is a similarly invaluable general treatment. There-
after, specifically on Critical Discourse Analysis, see Fairclough (1989; 1992; 1999);
Caldas-Coulthard and Coulthard (1996); and special issues of the journal Discourse &
Society (e.g. 1992). On the language of the news media, especially of hard news narra-
tives, see Fowler (1991), Bell (1991), White (1997), van Dijk (1988) and Bell and
Garrett (1998). Theoretical commentary on the Internet, hypermedia and their implica-
tions is voluminous; see Aarseth (1997) and Murray (1998) for a taster. On gender,
feminism and narrative, a beginning can be made with Lanser (1986), Warhol (1989),
Mezei (1996), Mills (1998a), Fludernik (1999) and Bucholtz (1999). On the political
aspects of ethnic stories, see Sykes (1985), van Dijk (1984; 1991; 1992; 1993), Wodak et
al. (1999), Blommaert and Verschueren (1998). And on the political consequences of
courtroom discourse, see Lind and O’Barr (1979), O’Barr (1982), Atkinson and Drew
(1979), Bennett and Feldman (1981) and Wodak (1981; 1985). On stories in US
Supreme Court proceedings: Amsterdam and Bruner (2000). Kress and van Leeuwen
(1996) is a rich and multi-stranded account of the semiotic import and logic of multi-
modal texts, particularly those that include graphics. This study is full of contributions
towards the development of a systematic understanding of what is involved in visual lit-
eracy. On all these overlapping areas there is now a host of journals to consult, includ-
ing Discourse & Society, Forensic Linguistics, Journal of Pragmatics, Text and
International Journal of the Sociology of Language.
Notes and exercises

1 Hodge (1979) offers an analysis of the processes and participants that appear in the
headlines of one British broadsheet’s foreign news coverage on just one day. The
broad trends are as follows:
1 Public persons say and tell, they do not otherwise act and are not acted on.
2 Private persons only exist if they are the subject of violent action, but what they
say or feel does not exist as news.
3 The world outside Britain is unrelievedly a world of conflict, usually between
states, or within states, or between governments and subversive forces. The conflict
is occasionally resolved, temporarily at least, but is more usually the motive for the
reported action.
4 Since they are mediated through public persons, conflicts exist mainly through
statements and attitudes.
5 Since public figures predominate over private, words and feelings predominate
over actions. (Hodge, 1979: 161–2, slightly abridged)
Extract the headlines of the foreign news stories of any newspaper you have to hand.
Analyze these headlines, in terms of their transitivity (4.9) and draw up lists of the
agents and mediums in material processes, the sayers and sensers in clauses of verbal
or mental process, and so on. To what extent do you find Hodge’s claims still valid,
nearly twenty-five years later? Attempt a similar analytical exercise on the same
newspaper’s headlines for (a sample of) its domestic news stories. Which – if any – of
Hodge’s generalizations need amending if they are to apply to this rather different
sample? Do any amendments have to do with domestic journalism’s greater attention
to ‘human interest’ stories? In fact, can we use a Hallidayan transitivity analysis to
uncover some of the basic ingredients of what newspapers count as ‘human interest’
or soft-news stories?
2 Using the critical discourse-analytic ‘toolkit’ set out in this chapter, analyze and
comment upon any current newspaper account of President Mugabe, and events in
Zimbabwe.
3 If one wants to make the case for a female or feminist story structure, the short story
by Canadian writer Alice Munro called ‘Circle of Prayer’ may be as good a test-case
as any (a quite different alternative would be Angela Carter’s ‘The Company of
Wolves’). Only a few preliminary observations can be offered here, possible pointers
to rather than defining criteria of feminist structuration. To begin with temporal
sequencing, in my analysis, ‘Circle of Prayer’ comprises eighteen distinct temporal
segments (which I label ‘a’ to ‘r’), which are presented in an order quite at variance
with their reconstructible chronology. Thus the chronologically earliest segment,
about how Trudy and her now ex-husband Dan met, is eighth to be told. The discour-
sal telling begins with the account of how Trudy throws a jug across the room at or at
least towards her teenage daughter, Robin (the jug, made some time ago by Dan in a
pottery class, formerly held a necklace, a precious keepsake to Trudy, which she has
just discovered that Robin in a grand gesture has dropped into the coffin of a tragi-
cally-killed school acquaintance). What is told first in the discourse is chronologically
the eighth episode. But even this episode, which falls into three subparts, is told
achronologically: first we are told that Trudy threw the jug, then we are told the
history of the jug and where it sits and what is kept in it, and then we are told how
Trudy came home from work, got the jug down, found as she expected that the neck-
lace was missing, confronted Robin in her room, and threw the jug. But it is not
simply the anachrony of the telling that sets this structuring at a remove from typical
‘high-point’, crisis-and-resolution narration. It is more broadly the sense that no
single particular crisis is available, as it were, to be resolved: Trudy’s complex of con-
cerns is not that simple, that it can be cast as a linear narrative with trajectory and
endpoint. In these respects, a couple of brief textual segments are particularly strik-
ing. They are brief, remarkably similar and symmetrical segments, which I reproduce
here:
b You threw a jug at me that time. You could have killed me.
Not at you. I didn’t throw it at you.
You could have killed me.
q You threw the jug. You could have killed me.
Yes.
As far as chronological sequence is concerned, these, in the order presented here, are
unquestionably the ‘latest’ moments in the story. In fact they record two scenes or
episodes that may have occurred indefinitely long after the entirety of the story
proper – even ten or twenty years later. And it seems reasonable to treat them as
temporally distinct from each other since by the time that (q) arises, but only then,
Trudy is prepared to admit she might have killed Robin. But as far as narrational
sequence is concerned, as my labelling indicates, episode (b) is the very second scene
to be told, right after the jug-throwing episode discussed above; and episode (q) –
and here is the symmetry – is the penultimate scene to be told. In a long and complex
story, these belated moments are told prematurely but separately, in counterposed
positions.
The story is also feminist in its eliding of a straightforward answer to the ‘about-
ness’ question. It is hard to say what the story is chiefly about, and even harder to say
‘what finally happens’: these feel like the wrong questions to dwell on. ‘Circle of
Prayer’ is substantially about Trudy’s strained relations with her infuriating
independent-minded daughter (or housemate?). At least they are talking again by
the end of the story proper, when Robin – a running champion – phones her mother
at the end of Trudy’s workshift and asks, ‘Can I run over and ride back with you?’: a
simple form of constructive circularity or recurrence. But the story is also about
Trudy’s feelings for and resentments of her ex-husband Dan and his infidelity; and it
is about how she now lives, her work, her friends. These include her friend Janet,
who has joined a Circle of Prayer, and explains to Trudy:
‘What the Circle is really about is, you phone up somebody that is in it and tell
them what it is you’re worried about, or upset about, and ask them to pray for
you. And they do. And they phone one other person that’s in the Circle, and
they phone another and it goes all around, and we pray for one person, all
together.’
With due allowances for differences, this is not so far removed from being a fair
description of the entire story, which seems at core to dwell, therapeutically, on what
Trudy is worried or upset about. As a therapeutic narrating cure, there is no mas-
culinist ‘triumph’ at the close of the story, although it does end with a kind of male
triumph. The final narrated episode, (r), involves learning-disabled Kelvin, a young
male resident in the Home that Trudy works at; Kelvin has worries of his own, about
other residents and how he relates to them. But when Trudy asks him if he ever prays
and if so what he prays for, Kelvin has an answer:
‘If I was smart enough to know what to pray for,’ he says, ‘then I wouldn’t have to.’
In context, Kelvin’s observation is remarkably powerful, and shows Kelvin has an

insight which his self-description (not smart enough) denies. Not knowing quite what
to pray for applies not just to Kelvin but to Trudy too, and is perhaps a liberation as
much as a problem. While Bertha in ‘Bliss’ wonders what will happen now, Trudy –
like Kelvin – seems not to know what she would want to happen now or next, and the
cyclical trajectory-less structure of the story in a sense validates this, even making the
narrative with beginning, middle and end seem simplistic. The Trudy-focalized narra-
tive reflects on what it calls Kelvin’s ‘halfway joke’:
It’s not meant as comfort, particularly. Yet it radiates – what he said, the way he
said it, just the fact that he’s there again, radiates, expands . . .
If the case is to be made out for an alternative, feminist structural format or tendency
in some literary (and non-literary) narratives, then evidence from a variety of facets
of text-construction, including those exemplified above, are likely to be relevant.
Other facets likely to be of relevance will include focalization, narrational mode and
speech- and thought-representation. Calibrating a story’s deployment of all these ele-
ments, in combination, it may be possible to identify tendencies which might warrant
classifying the story as ‘more feminist’ or ‘more masculinist’, without needing – or
wanting – to specify two well-defined categories, ‘feminist’ and ‘masculinist’ struc-
tures. But the starting-point remains literary competence and cultural judgment, and
mine is that ‘Circle of Prayer’ contrasts with ‘Barn Burning’, the jug-throwing so dif-
ferent from the rug-throwing, in ways which might conceivably be related to femin-
ism and masculinism. The challenge is to spell out those ways.
Bibliography
Aarseth, E. (1997) Cybertext: Perspectives on Ergodic Literature, Baltimore: Johns

Hopkins University Press.
Ahlberg, J. and A. Ahlberg (1979) Burglar Bill, London: Collins.
Amsterdam, A. and J. Bruner (2000) Minding the Law, Harvard: Harvard Univer-
sity Press.
Atkinson, M. and P. Drew, (1979) Order in Court, Atlantic Highfields, NJ: Human-
ities Press.
Austin, J. L. (1962) How To Do Things With Words, Oxford: Clarendon Press.
Bakhtin, M. M. (1981) The Dialogic Imagination: Four Essays, M. Holquist (ed.)
Austin: University of Texas Press.
Bakhtin, M. M. (1986) Speech Genres and Other Late Essays, Austin: University of
Texas Press.
Bal, M. (1985 [1997: 2nd edition]) Narratology: Introduction to the Theory of Nar-
rative, Toronto: University of Toronto Press.
Baldry, A., P. Thibault, et al. (2000) Multimodality and Multimediality in the Dis-
tance Learning Age, Campobasso: Palladino Editore.
Bally, C. (1912) ‘Le style indirecte libre en francais modeme’, Germanisch-
Romnanisch Monatsschrift, 4, 549–56, 597–606.
Bamberg, M. (ed.) (1997a) Oral Versions of Personal Experience: Three Decades of
Narrative Analysis [special issue of the Journal of Narrative and Life History],
Mahwah, NJ: Lawrence Erlbaum.
Bamberg, M. (ed.) (1997b) Narrative Development: Six Approaches. Mahwah, NJ:
Lawrence Erlbaum.
Banfield, A. (1982) Unspeakable Sentences, New York: Routledge & Kegan Paul.
Baron, H. (1998) ‘Disseminated consciousness’, Essays in Criticism, 48, 4, 357–78.
Barthes, R. (1970) S/Z, Paris: Seuil.
Barthes, R. (1977) ‘Introduction to the structural analysis of narratives’, in
Image–Music–Text, London: Fontana.
Bauman, R. (1986) Story, Performance, and Event: Contextual Studies of Oral Nar-
rative, Cambridge: Cambridge University Press.
Bayley, J. (1963) The Character of Love: A Study in the Literature of Personality,
New York: Collier Books.
Beaman, K. (1984) ‘Coordination and subordination revisited’, in D. Tannen (ed.)
Coherence in Spoken and Written Discourse, Norwood, NJ: Ablex.
Bedient, C. (1969) ‘Middlemarch: touching down’, Hudson Review, 22.
Bell, A. (1991) The Language of the News Media, Oxford: Blackwell.
Bell, A. and P. Garrett (eds) (1998) Approaches to Media Discourse, Oxford: Blackwell.
Bibliography 243
Bennett, G. (1983) ‘ “Rocky the police dog” and other tales’, Lore and Language
3, 8, 1–19.
Bennett, W. and S. Feldman (1981) Reconstructing Reality in the Courtroom: Justice
and Judgement in American Culture, New Brunswick, NJ: Rutgers University
Press.
Bennett-Kastor, T. (1983) ‘Noun phrases and coherence in child narratives’,
Journal of Child Language, 10, 135–49.
Bennett-Kastor, T. (1986) ‘Cohesion and predication in child narratives’, Journal
of Child Language, 13, 353–70.
Bennett-Kastor, T. (1999) Predications and nonreferential cohesion in Irish-speak-
ing children’s narratives’, Functions of Language, 6, 1, 195–241.
Benveniste, E. (1966) Problemes de linguistique generale, Paris: Gallimard.
Berendson, M. (1981) ‘Formal criteria of narrative embedding’, Journal of Literary
Semantics, 10, 79–94.
Berendson, M. (1984) ‘The teller and the observer: narration and focalization in
narrative texts’, Style, 18, 140–58.
Bhaya, R., R. Carter and M. Toolan (1988) Clines of metaphoricity, and creative
metaphors as situated risk-taking. Journal of Literary Semantics, 17, 1: 20–40.
Bickerton, D. (1967) ‘Modes of interior monologue: a formal definition’, Modern
Language Quarterly, 28, 2, 229–39.
Bissex, G. (1981) Gnys at Work: a Child Learns to Write and Read, Cambridge,
MA., Harvard University Press.
Blommaert, J. and J. Verschueren (1998) Debating Diversity: Analysing the Dis-
course of Tolerance. London, Routledge.
Bloor, T. and M. Bloor (1995) The Functional Analysis of English: a Hallidayan
approach, London: Arnold.
Bolinger, D. (1980) Language, the Loaded Weapon, London: Longman.
Bolinger, D. (1981) Aspects of Language, New York: Harcourt Brace Jovanovich.
Booth, W. (1961) The Rhetoric of Fiction, Chicago: University of Chicago Press.
Bordwell, D. (1986) Narration in the Fiction Film. London and New York: Routledge.
Bronzwaer, W. J. M. (1970) Tense in the Novel, Groningen: Wolters Noordhoff.
Brown, G. and G. Yule (1983) Discourse Analysis, Cambridge: Cambridge Univer-
sity Press.
Brown, P. and S. Levinson (1987) Politeness, Cambridge: Cambridge University
Press.
Browne, A. (1983) Bear Goes to Town, London: Arrow Books.
Bruner, E. (ed.) (1984) Text, Play and Story: The Construction and Reconstruction
of Self and Society, Washington DC American Ethnological Society.
Bruner, J. (1986) Actual Minds, Possible Worlds. Cambridge, MA: Harvard Uni-
versity Press.
Bruner, J. (1990) Acts of Meaning, Cambridge, MA: Harvard University Press.
Bruner, J. (1992) The narrative construction of reality, in H. Beilin and P. Pufall
(eds) Piaget’s Theory: Projects and Possibilities. Hillsdale, NJ: Lawrence
Erlbaum, 229–48.
Bucholtz, M. (1999) ‘You da man: Narrating the racial other in the production of
white masculinity’, Journal of Sociolinguistics, 3: 4, 443–60.
Burton, D. (1982) ‘Through glass darkly: through dark glasses’, in R. Carter (ed.)
Language and Literature, London: Allen & Unwin, 195–214.
Butt, D. et al. (1995). Using Functional Grammar: an Explorer’s Guide, Sydney:
National Centre for English Language Teaching and Research.
244 Bibliography
Caldas-Coulthard, C. and M. Coulthard (eds) (1996) Texts and Practices: Readings
in Critical Discourse Analysis, London: Routledge.
Cameron, D. (1995) Verbal Hygiene, London: Routledge.
Carter, R. and P. Simpson (1982) ‘The sociolinguistic analysis of narrative’, Belfast
Working Papers in Linguistics, 6: 123–52.
Carver, R. (1989) ‘A Small, Good Thing’, in Cathedral and other Stories, London:
Vintage Books.
Carver, R. (1994) ‘Little Things’, in Fires: Essays, Poems, Stories, London: Harvill.
Cazden, C. (1981) Language in Early Childhood Education, Washington DC:
National Association for the Education of Young Children.
Chatman, S. (1969) ‘New ways of analysing narrative structures’, Language and
Style, 2: 3–36.
Chatman, S. (1978) Story and Discourse, Ithaca: Cornell University Press.
Chatman, S. (1990) Coming to Terms: the Rhetoric of Narrative in Fiction and Film,
Ithaca: Cornell University Press.
Chomsky, N. (1957) Syntactic Structures, The Hague: Mouton.
Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge: MIT Press.
Christie, F. et al. (1984) Children’s Writing: Reader, Victoria: Deakin University
Press.
Christie, F. and J. Martin (eds), (1997) Genres and Institutions: Social Processes in
the Workplace and School, London: Cassell.
Clark, H. (1977) ‘Inferences in comprehension’, in D. Laberge and S. J. Samuels
(eds) Basic Processes in Reading, Hillsdale, NJ: Erlbaum.
Cohn, D. (1978) Transparent Minds: Narrative Modes for Presenting Consciousness
in Fiction, Princeton, NJ: Princeton University Press.
Colby, B. (1970) ‘The description of narrative structures’, In P. Garvin (ed.) Cogni-
tion: A Multiple View, New York: Spartan Books.
Cook-Gumperz, J. and J. Green (1984) ‘A sense of story: influences on children’s
storytelling ability’, in D. Tannen (ed.) Coherence in Spoken and Written Dis-
course, Norwood, NJ: Ablex, 201–18.
Cotterill, J. (2000) Representing Reality in Court: Power and Persuasion in Trial
Discourse, unpublished PhD dissertation, University of Birmingham.
Coulthard, M. (1994) ‘Powerful evidence for the defence: an exercise in forensic
discourse analysis’, in J. Gibbons (ed.) Language and the Law, London:
Longman, 414–42.
Coulthard, M. (1996) ‘The official version: audience manipulation in police records
of interviews with suspects,’ in C. Caldas-Coulthard and M. Coulthard (eds)
Texts and Practices: Readings in Critical Discourse Analysis, London: Routledge,
pp. 166–78.
Coulthard, M., in preparation, Language as Evidence, London: Routledge.
Crawford, P., B. Brown and P. Nolan (1998) Communicating Care: the Language of
Nursing, Nelson Thornes, Cheltenham.
Crossley, M. (2000) Introducing Narrative Psychology: Self, Trauma, and the Con-
struction of Meaning. Buckingham: Open University Press.
Culler, J. (1975a) Structuralist Poetics, London: Routledge & Kegan Paul.
Culler, J. (1975b) ‘Defining narrative units’, in R. Fowler (ed.) Style and Structure
in Literature, Oxford: Blackwell.
Culler, J. (1981) The Pursuit of Signs, London, Routledge & Kegan Paul.
Culpeper, J. (2000) ‘A cognitive approach to characterization: Katherina in Shake-
speare’s The Taming of the Shrew’, Language and Literature, 9:4, 291–316.
Bibliography 245
Culpeper, J. (forthc.) Language and Characterization: People in Plays and Other
Texts, London: Longman.
Dijk, T. van (1984) Prejudice in Discourse: An Analysis of Ethnic Prejudice in Cog-
nition, Amsterdam: John Benjamins.
Dijk, T. van (1988) News as Discourse, Hillsdale, NJ: Lawrence Erlbaum.
Dijk, T. van (1991) Racism and the Press, London: Routledge.
Dijk, T. van (1992) ‘Discourse and the denial of racism’, Discourse & Society, 3:1,
87–118.
Dijk, T. van (1993) Discourse and Elite Racism, London: Sage.
Dillon, G. and F. Kirchnoff (1976) ‘On the form and function of free indirect style’,
Poetics and Theory of Literature, 1, 3, 431–40.
Docherty, T. (1983) Reading (Absent) Character: Towards a Theory of Characteri-
zation in Fiction, Oxford: Clarendon Press.
Dry, H. (1995) ‘Free indirect discourse in Doris Lessing’s “One off the short list”,’
in P. Verdonk and J. J. Weber (eds) Twentieth Century Fiction: from Text to
Context, London: Routledge, 96–112.
Duchan, J., G. Bruder and L. Hewitt (eds) (1995) Deixis in Narrative: A Cognitive
Science Perspective, Hillsdale NJ: Lawrence Erlbaum.
Dundes, A. (1968) Introduction to Propp, The Morphology of the Folktale, Austin:
University of Texas Press.
Eaton, J., G. Collis and V. Lewis (1999) ‘Evaluative explanations in children’s nar-
ratives of a video sequence without dialogue’, Journal of Child Language, 26(3),
699–720.
Edwards, D. (1997) Discourse and Cognition, London: Sage.
Eggins, S. (1994) An Introduction to Systemic Functional Linguistics, London: Pinter.
Eggins, S. and J. Martin (1997) Genres and registers of discourse. In T. Van Dijk
(ed.) Discourse as Structure and Process, London: Sage, 230–56.
Eggins, S. and D. Slade (1997) Analysing Casual Conversation, London: Cassell.
Ehrlich, S. (1990) Point of View: A Linguistic Analysis of Literary Style, London:
Routledge.
Ely, R., J. Bleason and A. McCabe (1996) ‘“Why didn’t you talk to your mommy,
honey?”: gender differences in talk about past talk’, Research on Language and
Social Interaction, 20, 7–25.
Emmott, C. (1997) Narrative Comprehension: A Discourse Perspective, Oxford:
Oxford University Press.
Fairclough, N. (1989) Language and Power, London: Longman.
Fairclough, N. (1992) Discourse and Social Change, Oxford: Polity Press.
Farr, M. (ed.) (1985) Children’s Early Writing Development, Norwood, NJ: Ablex.
Faulkner, W. (1942) Go Down, Moses, New York: Random House.
Faulkner, W. (1950) Collected Stories of William Faulkner, New York: Random
House.
Ferguson, F. (2000) ‘Jane Austen, Emma, and the Impact of Form’, Modern Lan-
guage Quarterly, 61:1, 157–80.
Fillmore, C. (1974) ‘Pragmatics and the description of discourse’, Berkeley Studies
in Syntax and Semantics, 1, Chapter 5.
Finegan, E. (1994) Language: its Structure and Use, New York: Harcourt Brace.
Fish, S. (1980) Is There a Text in this Class?, Harvard: Harvard University Press.
Fleischman, S. (1997) ‘The “Labovian model” revisited with special consideration of
literary narrative’, in M. Bamberg, (ed.) Oral Versions of Personal Experience:
Three Decades of Narrative Analysis, Mahwah, NJ: Lawrence Erlbaum, pp. 159–68.
246 Bibliography
Fludernik, M. (1993) The Fictions of Language and the Languages of Fiction,
London: Routledge.
Fludernik, M. (1996) Towards a ‘Natural’ Narratology, London: Routledge.
Fludernik, M. (1999a) ‘Defining (In)Sanity: The Narrator of The Yellow Wallpaper
and the Question of Unreliability’, in W. Grünzweig and A. Solbach (eds) Tran-
scending Boundaries: Narratology in Context, Tübingen: Gunter Narr Verlag,
75–95.
Fludernik, M. (1999b) ‘The genderization of narrative’, GRAAT, 21, 153–75.
Fowler, R., (ed.) (1975) Style and Structure in Literature, Oxford: Blackwell.
Fowler, R. (1977) Linguistics and the Novel, London: Methuen.
Fowler, R. (1981) Literature as Social Discourse, London: Batsford.
Fowler, R. (1986) Linguistic Criticism, Oxford: Oxford University Press.
Fowler, R. (1991) Language in the News, London: Routledge.
Fowler, R. et al. (1979) Language and Control, London: Routledge & Kegan Paul.
Fromkin, V. and R. Rodman (1978) An Introduction to Language, New York:
Holt, Rinehart & Winston.
Frow, J. (1986) ‘Spectacle binding: on character’, Poetics Today, 7, 2, 227–50.
Gee, J. P. (1991) A linguistic approach to narrative, Journal of Narrative and Life
History, 1, 15–39.
Gee, J. P. (1996) Social Linguistics and Literacies: Ideology in Discourses, London:
Taylor & Francis.
Genette, G. (1980) Narrative Discourse, Ithaca: Cornell University Press.
Genette, G. (1988) Narrative Discourse Revisited, Ithaca: Cornell University Press.
Georgakopoulou, A. and D. Goutsos (1997) Discourse Analysis: An Introduction,
Edinburgh: Edinburgh University Press.
Georgesen, J. and C. Solano (1999) ‘The effects of motivation on narrative
content and structure’, Journal of Language and Social Psychology, 18(2), 175–95.
Ginsberg, M. P., (ed.) (1982) ‘Free indirect discourse: a reconsideration’, Language
and Style, 15, 2, 133–49.
Greimas, A. (1966) Sémantique Structurale, Paris: Larousse.
Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. Morgan (eds) Syntax
and Semantics 3: Speech Acts, New York: Academic Press, 41–58.
Hagenaar, E. (1996) Free Indirect Speech in Chinese. In T. Janssen and W. van der
Wurff (eds) Reported Speech: Forms and Functions of the Verb, Amsterdam:
Benjamins, 289–98.
Halliday, M. A. K. (1971) ‘Linguistic function and literary style’, in S. Chatman
(ed.) Literary Style: A Symposium, London: Oxford University Press, 330–65.
Reprinted in J. Weber (ed.), The Stylistics Reader, London: Arnold, 56–86.
Halliday, M. A. K. (1978) Language as Social Semiotic, London: Edward Arnold.
Halliday, M. A. K. (1994) An Introduction to Functional Grammar, 2nd edition,
London: Edward Arnold.
Halliday, M. A. K. and R. Hasan (1976) Cohesion in English, London, Longman.
Halliday, M. A. K. and R. Hasan (1985) Spoken and Written Language, Victoria:
Deakin University Press.
Harré, R. (1998) The Singular Self: an Introduction to the Psychology of Person-
hood, London: Sage.
Harris, R. (1981) The Language Myth, London: Duckworth.
Harris, R. (1998) Introduction to Integrational Linguistics, Oxford: Pergamon.
Hart, C. (1969) ‘Eveline’, in C. Hart (ed.) James Joyce’s Dubliners, London:
University of California Press.
Bibliography 247
Harvey, W. J. (1965) Character and the Novel, Ithaca: Cornell University Press.
Hasan, R. (1984) ‘The nursery tale as a genre’, Nottingham Linguistic Circular, 13.
Hawthorn, J. (ed.) (1985) Narrative: From Malory to Motion Pictures, London:
Edward Arnold.
Heath, S. B. (1982) ‘What no bedtime story means: narrative skills at home and
school’, Language in Society, 11, 49–76.
Heath, S. B. (1983) Ways with Words, Cambridge: Cambridge University Press.
Heffer, C. (in preparation) Making a Case: Narrativization and Categorization in the
Courtroom Language of Legal Professionals. Unpublished PhD dissertation, Uni-
versity of Birmingham.
Hemingway, E. (1964) ‘Indian Camp’, in The Essential Hemingway, Har-
mondsworth: Penguin, pp. 279–83.
Herman, D. (1997) Scripts, Sequences, and Stories: Elements of a Postclassical
Narratology, PMLA, 112:5, 1046–59.
Hernadi, P. (1972) ‘Free indirect discourse and related techniques’, appendix to
Beyond Genre, Ithaca: Cornell University Press, 187–205.
Hickmann, M. (1985) ‘Metapragmatics in child language’, in E. Mertz and R. J.
Parmentier (eds) Semiotic Mediation: Sociocultural and Psychological Perspec-
tives, Orlando, FL: Academic, 177–201.
Hickmann M. (ed.) (1987) Social and Functional Approaches to Language and
Thought, New York: Academic Press.
Hirose, Y. (2000) Public and private self as two aspects of the speaker: a con-
trastive study of Japanese and English, Journal of Pragmatics 32, 1623–56.
Hodge, R. (1979) ‘Newspapers and communities’, in R. Fowler et al., Language
and Control, London: Routledge & Kegan Paul, 157–74.
Hoey, M. (2001) Textual Interaction: An Introduction to Discourse Analysis,
London: Routledge.
Hoover, D. (1998) Language and Style in The Inheritors, New York: University
Press of America.
Hummel, M. (1999) ‘Polyphonie, Appell und Perspektive bei erlebter Rede: Fran-
zosisch, Portugiesisch, Spanisch’, in S. Grosse and A. Schonberger (eds) Dulce et
Decorum est Philologiam Colere, I–II, Berlin: Domus Editoria Europaea,
1633–49.
Hunston, S. and G. Thompson (eds) (2000) Evaluation in Text: Authorial Stance
and the Construction of Discourse, Oxford: Oxford University Press.
Hymes, D. (1972) ‘Models of the interaction of language and social life’, in J.
Gumperz and D. Hymes (eds) Directions in Sociolinguistics, New York: Holt,
Rinehart & Winston.
Hymes, D. (1982) ‘Narrative form as a “grammar” of experience’, Journal of Edu-
cation 2, 121–42.
Iser, W. (1978) The Act of Reading: a Theory of Aesthetic Response, London: Rout-
ledge and Kegan Paul.
Jahn, M. (1999) ‘More aspects of focalisation: refinements and applications’,
GRAAT 21, 85–110.
James, H. (1963 [1884]) ‘The art of fiction’, in M. Shapira (ed.) Henry James:
Selected Literary Criticism, Harmondsworth: Penguin.
James, H. (1966 [1888]) The Portrait of a Lady, Harmondsworth: Penguin.
Jameson, F. (1972) The Prison-House of Language, New York: Princeton Univer-
sity Press.
Jaworski, A. and N. Coupland (1999) The Discourse Reader, London: Routledge.
248 Bibliography
Jefferson, A. (1981) ‘The place of free indirect discourse in the poetics of fiction:
with examples from Joyce’s ‘Eveline’, Essays in Poetics, 5:1, 36–47.
Jefferson, G. (1978) ‘Sequential aspects of story-telling in conversation’, in
J. Schenkein (ed.) Studies in the Organization of Conversational Interaction, New
York: Academic Press.
Johnson-Laird, P. (1981) Mental Models, Cambridge: Cambridge University Press.
Johnstone, B. (1990) Stories, Community, and Place: Narratives from Middle
America, Bloomington: Indiana University Press.
Johnstone, B. (1996) The Linguistic Individual: Self-expression in Language and
Linguistics, New York: Oxford University Press.
Johnstone, B. and J. Beam (1997) Self-expression and linguistic variation, Lan-
guage in Society, 26, 221–46.
Jones, C. (1968) ‘Varieties of speech presentation in Conrad’s The Secret Agent’,
Lingua, 20, 162–76.
Joyce, J. (1956 [1914]) Dubliners, Harmondsworth: Penguin.
Joyce, J. (1960 [1916]) A Portrait of the Artist as a Young Man, Harmondsworth:
Penguin.
Kennedy, C. (1982) ‘Systemic grammar and its use in literary analysis’, in R. Carter
(ed.) Language and Literature, London: Allen & Unwin, 83–99.
Kermode, F. (1967) The Sense of an Ending : Studies in the Theory of Fiction, New
York: Oxford University Press.
Kernan, K. (1977) ‘Semantic and expressive elaboration in children’s narratives’, in
S. Ervin-Tripp and C. Mitchell-Kernan (eds) Child Discourse, New York: Acad-
emic Press.
Kress, G. and T. van Leeuwen (1996) Reading Images, London: Routledge.
Kullmann D. (ed.) (1995) Erlebte Rede und impressionistischer Stil: Europaische
Erzahlprosa im Vergleich mit ihren deutschen Ubersetzungen, Gottingen: Wall-
stein, 528 pp.
Labov, W. (1972) Language in the Inner City, Philadelphia: University of Pennsyl-
vania Press.
Labov, W. (1981) ‘Speech actions and reactions in personal narrative’, Georgetown
University Round Table on Languages and Linguistics, 219–47.
Labov, W. (1997) ‘Some further steps in narrative analysis’, in M. Bamberg (ed.)
Oral Versions of Personal Experience: Three Decades of Narrative Analysis
[special issue of the Journal of Narrative and Life History], Mahwah, NJ:
Lawrence Erlbaum.
Labov, W. and J. Waletzky (1967) ‘Narrative analysis: oral versions of personal
experience’, in J. Helms (ed.) Essays on the Verbal and Visual Arts, Seattle: Uni-
versity of Washington Press.
Lakoff, R. (1975) Language and Woman’s Place, New York: Harper Colophon.
Lanser, S. (1981) The Narrative Act: Point of View in Prose Fiction, Princeton:
Princeton University Press.
Lanser, S. (1986) ‘Toward a Feminist Narratology,’ Style, 20, 341–63. Reprinted in
R. Warhol and D. Herndl (eds) Feminisms: An Anthology of Literary Theory
and Criticism, New Brunswick, NJ: Rutgers University Press, 1997, 674–93.
Lanser, S. (1995) ‘Sexing the narrative: propriety, desire, and the engendering of
narrative’, Narrative, 3, 85–94.
Lanser, S. (1999) ‘Sexing narratology: toward a gendered poetics of narrative
voice’, in W. Grunzweig and A. Solbach (eds) Transcending Boundaries: Narra-
tology in Context, Tubingen: Gunter Narr, 167–83.
Bibliography 249
Lawrence, D. H. (1955) ‘Odour of chrysanthemums’, The Complete Short Stories,
London: Heinemann.
Leech, G. and M. Short (1981) Style in Fiction, London: Longman.
Leeuwen, T. van and R. Wodak (1999) ‘Legitimizing immigration control: a
discourse–historical Analysis’, Discourse Studies, 1:1, 83–118.
Levinson, S. (1983) Pragmatics, Cambridge: CUP.
Lind, E. and W. O’Barr (1979) ‘The social significance of speech in the courtroom’,
in H. Giles and R. St Clair (eds) Language and Social Psychology, Oxford:
Blackwell, 66–87.
Linde, C. (1993) Life Stories: the Creation of Coherence, New York: OUP.
Linde, C. (1999) The transformation of narrative syntax into institutional memory,
Narrative Inquiry, 9:1, 139–74.
Lodge, D. (1977) The Modes of Modern Writing, London: Edward Arnold.
Lodge, D. (1981) Working with Structuralism, London: Routledge & Kegan Paul.
Longacre, R (1983) The Grammar of Discourse, New York: Plenum Press.
Lothe, J. (2000) Narrative in Fiction and Film, Oxford: Oxford University Press.
Lubbock, P. (1973) The Craft of Fiction [1921], New York: Viking Press.
Luelsdorff, P. A. (1995) ‘A grammar of suspense’, Journal of Literary Semantics,
24:1, 1–20.
Lyons, J. (1977) Semantics, 2 vols, Cambridge: Cambridge University Press.
McCabe, A. and C. Peterson (eds) (1991) Developing Narrative Structure, Hills-
dale, NJ: Lawrence Erlbaum.
McCabe, A. and C. Peterson (1996) Meaningful mistakes: the systematicity of chil-
dren’s connectives in narrative discourse and the social origins of this usage
about the past. In J. Costermans and M. Fayol (eds) Processing Interclausal
Relationships in the Production and Comprehension of Text, Hillsdale, NJ:
Lawrence Erlbaum, 139–54.
McConnell-Ginet, S. and P. Eckert (1995) Constructing meaning, constructing
selves: snapshots of language, gender, and class from Belten High. In K. Hall
and M. Bucholtz (eds) Gender Articulated; Language and the Socially Con-
structed Self, London and New York: Routledge.
McHale, B. (1978) ‘Free indirect discourse: a survey of recent accounts’, Poetics
and Theory of Literature, 3, 249–87.
McHale, B. (1983) ‘Linguistics and poetics revisited’, Poetics Today, 4, 1, 17–45.
McHale, B. (1985) ‘Speaking as a child in U.S.A.’, Language and Style, 17, 351–70.
McKay, J. H. (1978) ‘Some problems in the analysis of point of view in reported
discourse’, Centrum, 6, 1, 5–26.
McKay, J. H. (1982) Narration and Discourse in American Realistic Fiction,
Philadelphia: University of Pennsylvania Press.
Mandler, J. and N. Johnson (1977) ‘Remembrance of things parsed: story structure
and recall’, Cognitive Psychology, 9, 111–51.
Mansfield, K. (1981) The Collected Short Stories, Harmondsworth: Penguin.
Margolin, U. (1989) ‘Structuralist approaches to character in narrative: the state of
the art’, Semiotica, 75, 1–24.
Margolin, U. (1996) ‘Characters and their versions’, in C. Mihailescu and W.
Hamarneh (eds) Fiction Updated: Theories of Fictionality, Narratology, and
Poetics, Toronto, University of Toronto Press, 113–32.
Martin, J. (1983) ‘The development of register’, J. Fine and R. Freedle (eds)
Developmental Issues in Discourse, Norwood, NJ: Albex, 1–40.
Martin, J. (1992) English Text: System and Structure, Amsterdam: Benjamins.
250 Bibliography
Martin, J. (2000) Inter-feeling: gender, class and appraisal in Educating Rita. In S.
Hunston and G. Thompson (eds) Evaluation in Text, Oxford: Oxford University
Press.
Matthiessen, C. (1995) Lexicogrammatical Cartography, Tokyo: International Lan-
guage Sciences Publishers.
Mezei, K. (1996) ‘Free indirect discourse, gender, and authority in Emma,
Howards End, and Mrs. Dalloway’, in K. Mezei (ed.) Ambiguous Discourse:
Feminist Narratology & British Women Writers, Chapel Hill: University of North
Carolina Press, 66–92.
Mezei, K. (ed.) (1996) Ambiguous Discourse: Feminist Narratology & British
Women Writers, Chapel Hill: University of North Carolina Press.
Michaels, S. (1981) ‘ “Sharing time”: children’s narrative styles and differential
access to literacy’, Language in Society, 10, 423–42.
Michaels, S. (1985) ‘Hearing the connections in children’s oral and written dis-
course’, Journal of Education, 167:1, 36–56.
Michaels, S. and J. Collins (1984) ‘Oral discourse styles: classroom interaction and
the acquisition of literacy’, in D. Tannen (ed.) Coherence in Spoken and Written
Discourse, Norwood, NJ: Ablex, 219–42.
Mihailescu, C. and W. Hamarneh (eds) (1996) Fiction Updated: Theories of Fic-
tionality, Narratology, and Poetics, Toronto: University of Toronto Press.
Mills, S. (1998a), ‘The gendered sentence’. in D. Cameron (ed.) The Feminist Cri-
tique of Language: A Reader, London: Routledge, 65–77.
Mills, S. (1998b) ‘Post-feminist text analysis’, Language and Literature, 7:3,
235–53.
Mitchell, T. (1975 [1957]) The language of buying and selling in Cyrenaica. In T. F.
Mitchell, Principles of Neo-Firthian Logic, London: Longman.
Muhlhausler, P. and R. Harré (1990) Pronouns and People, Oxford: Blackwell.
Murray, J. (1998) Hamlet on the Holodeck, Cambridge, MA: MIT Press.
Myers, G. (1999) ‘Functions of reported speech in group discussions’, Applied Lin-
guistics, 20, 3, 376–401.
Nabokov, V. (1957) Pnin, Harmondsworth: Penguin.
Neisser, U. and R. Fivush (eds) (1994) The Remembering Self: Construction and
Accuracy in the Self-narrative, New York: Cambridge University Press.
Nelles, W. (1984) ‘Problems for narrative theory: The French Lieutenant’s
Woman’, Style, 18, 2, 207–17.
Nicolopoulou, A. (1996) Narrative development in social context. In D. Slobin, J.
Gerhardt, A. Kyratzis and J. Guo (eds) Social Interaction, Social Context, and
Language: Essays in Honor of Susan Ervin-Tripp, Mahwah, NJ: Lawrence
Erlbaum, 369–90.
Nicolopoulou, A. (1997) Children and Narratives: toward an interpretive and
sociocultural approach. In M. Bamberg (ed.) Narrative Development: Six
Approaches, Mahwah, NJ: Laurence Erlbaum, 179–215.
Norrick, N. (2000) Conversational Narrative. Storytelling in Everyday Talk,
Amsterdam: John Benjamins.
Nünning, A. (1999) Unreliable, compared to what? Towards a Cognitive Theory of
Unreliable Narration: Prolegomena and Hypotheses. In W. Grünzweig and A.
Solbach (eds) Transcending Boundaries: Narratology in Context, Tübingen:
Gunter Narr Verlag, 53–73.
O’Barr, W. (1982) Linguistic Evidence: Language, Power, and Strategy in the
Courtroom, New York: Academic Books.
Bibliography 251
Ochs, E. (1979) ‘Transcription as theory’, in E. Ochs and B. Schieffelin (eds)
Developmental Pragmatics, New York: Academic Press, 43–72.
Oltean, S. (1993) ‘A survey of the pragmatic and referential functions of Free Indi-
rect Discourse’, Poetics Today, 14:4, 691–714.
Onega, S. and J. Garcia (eds) (1996), Narratology, London: Longman.
Ong, W. J. (1975) Orality and Literacy, London: Methuen.
Palmer, J. (1978) Thrillers: Genesis and Structure of a Popular Genre, London:
Arnold.
Pascal, R. (1977) The Dual Voice, Manchester: Manchester University Press.
Payne, M. (2000) Narrative Therapy: an Introduction for Counsellors. London: Sage.
Peer, W. van and J. Chatman (eds) (2001) New Perspectives on Narrative Perspective,
Albany: State University of New York Press.
Perez, C. and H. Tager-Flusberg (1998) ‘Clinicians’ perceptions of children’s oral
personal narratives’, Narrative Inquiry, 8(1), 181–202.
Peterson, C. and A. McCabe (1983) Developmental Psycholinguistics: Three Ways
of Looking at Child’s Narrative, New York: Plenum Press.
Peterson, C., B. Jesso and A. McCabe (1999) ‘Encouraging narratives in
preschoolers: an intervention study,’ Journal of Child Language, 26(1), 49–67.
Piaget, J. (1959 [1926]) The Language and Thought of the Child, London: Rout-
ledge & Kegan Paul.
Plum, G. (1988) ‘Textual and Contextual Conditioning in Spoken English: A
Genre-Based Approach’, unpublished PhD thesis, Department of Linguistics,
University of Sydney.
Polanyi, L. (1978) The American story: cultural constraints on the structure and
meaning of stories in conversation, unpublished doctoral dissertation, University
of Michigan.
Polanyi, L. (1981) ‘Telling the same story twice’, Text, 1, 315–36.
Polanyi, L. (1982) ‘The nature of meaning of stories in conversation’, Studies in
Twentieth-Century Literature, 6, 1–2, 51–65.
Polanyi, L. (1985) Telling the American Story: From the Structure of Linguistic
Texts to the Grammar of a Culture, Norwood, NJ: Ablex.
Pratt, M. L. (1977) Toward a Speech-Act Theory of Literary Discourse, Blooming-
ton: Indiana University Press.
Price, M. (1968) ‘The other self: thoughts about character in the novel’, in M. Mack
and I. Gregor (eds) Imagined Worlds, London: Methuen, 279–99.
Prince, G. (1973) A Grammar of Stories, The Hague: Mouton.
Prince, G. (1982) Narratology: The Form and Function of Narrative, The Hague:
Mouton.
Prince, G. (1991) A Dictionary of Narratology, Aldershot: Scholar Press.
Propp, V. (1968 (1928]) The Morphology of the Folktale, Austin: University of
Texas Press.
Redeker, G. (1996) ‘Free indirect discourse in newspaper reports’, in C. Cremers and
M. den Dikken (eds) Linguistics in the Netherlands, Amsterdam: Benjamins,
221–32.
Reilly, J., E. Bates and V. Marchman (1998) ‘Narrative discourse in children with
early focal brain injury’, Brain and Language, 61(3), 335–75.
Rifelj, C. (1979) ‘Time in Agatha Christie’s novels’, Language and Style, 11, 4,
213–27.
Riffaterre, M. (1973) ‘Interpretation and descriptive poetry’, New Literary History,
reprinted in Young (ed.) (1981).
252 Bibliography
Rimmon-Kenan, S. (1983) Narrative Fiction: Contemporary Poetics, London:
Methuen.
Roberts, M. (1994) Une Glossaire/A Glossary, in During Mother’s Absence,
London: Virago, pp. 131–81.
Romaine, S. (1985) ‘Children’s narratives’, Linguistics, 23, 83–104.
Rothery, J. and J. Martin (1980) Writing Project, Papers I (Narrative: Vicarious
Experience) and 2 (Exposition: Literary Criticism), Sydney, Department of Lin-
guistics, University of Sydney.
Rubin, M. (1999) Thrillers, Cambridge: CUP.
Rumelhart, D. (1975) ‘Notes on a schema for stories’, in D. Bobrow and A. Collins
(eds) Representation and Understanding, New York: Academic Press.
Rumelhart, D. (1977) ‘Understanding and summarizing brief stories’, in D. Laberge
and S. J. Samuels (eds) Basic Processes in Reading, Hillsdale, NJ: Erlbaum.
Ryan, M.-L. (1981) ‘The pragmatics of personal and impersonal fiction’, Poetics,
10, 517–39.
Said, E. (1982) The World, the Text, and the Critic, London: Faber.
Sauerberg, L. O. (1984) Secret Agents in Fiction, London: Macmillan.
Saussure, F. de (1983) Course in General Linguistics, trans. R. Harris, London:
Duckworth.
Schank, R. and R. P. Abelson (1977) Scripts, Plans, Goals, and Understanding,
Hillsdale, NJ: Erlbaum.
Schiffrin, D. (1981) ‘Tense variation in narrative’, Language, 57, 1, 45–62.
Schiffrin, D. (1994) Approaches to Discourse, Oxford: Blackwell.
Schlesinger, P. (1987) Putting ‘Reality’ Together: BBC News, London: Methuen.
Scholes, R. and R. Kellogg (1966) The Nature of Narrative, New York: Oxford
University Press.
Scholes, R. and A. W. Litz (1969) Joyce’s Dubliners: A Critical Edition, New York:
Viking.
Semino, E., M. Short and M. Wynne (1999) ‘Hypothetical words and thoughts in
contemporary British narratives’, Narrative, Fall.
Short, M. (1988) ‘Speech presentation, the novel and the press’, in Willie van Peer
(ed.) (1988) The Taming of The Text, London: Routledge, pp. 61–81.
Short, M. (1996) Exploring the Language of Poems, Plays and Prose, London:
Longman.
Short, M., E. Semino and J. Culpeper (1996) ‘Using a corpus for stylistics research:
speech and thought presentation’, in J. Thomas and M. Short (eds) (1996) Using
Corpora in Language Research, London: Longman, pp. 110–31.
Short, M., E. Semino and M. Wynne (2001) ‘Revisiting the notion of faithfulness in
discourse report/(re)presentation using a corpus approach’, Language and Liter-
ature, 10.
Silva-Corvalan, C. (1983) ‘Tense and aspect in oral Spanish narrative: context and
meaning’, Language, 59, 4, 760–80.
Simpson, P. (1993) Language, Ideology and Point of View, London: Routledge.
Simpson, P. (1997) Language Through Literature: an Introduction, London: Routledge.
Sinclair, J. McH. and M. Coulthard (1975) Towards An Analysis of Discourse,
London: Oxford University Press.
Slembrouck, S. (1992) ‘The parliamentary Hansard “verbatim” report: the written
construction of spoken discourse’, Language and Literature, 1:2, 101–19.
Smiley, J. (1988) The Age of Grief, London: Collins.
Smith, B. H. (1978) On the Margins of Discourse, Chicago: Chicago University Press.
Bibliography 253
Snell, N. (1982) Julie Stays the Night, London: Hamish Hamilton.
SouthLondon Online (icSouthLondon.co.uk)
Spark, M. (1961) The Prime of Miss Jean Brodie, London: Macmillan.
Stanzel, F. (1971) Narrative Situations in the Novel, Bloomington, IN: Indiana Uni-
versity Press.
Stanzel, F. (1981) ‘Teller-characters and reflector-characters in narrative theory’,
Poetics Today, 2, 2, 5–15.
Stanzel, F. (1984) Theory of Narrative, trans. C. Goedsche, Cambridge: Cambridge
University Press.
Stein, N. and E. Albro (1997) ‘Building complexity and coherence: children’s use
of goal-structured knowledge in telling stories’, in M. Bamberg (ed.) Narrative
Development: Six Approaches, Mahwah, NJ: Lawrence Erlbaum, 5–44.
Stein, N. and C. Glenn (1979) ‘An analysis of story comprehension in elementary
school children’, in R. Freedle (ed.) New Directions in Discourse Processing,
Hillsdale, NJ: Ablex.
Sternberg, M. (1982) ‘Point of view and the indirections of direct speech’, Lan-
guage and Style, 15, 2, 67–117.
Sterne, L. (1967 [1760]) Tristram Shandy, Harmondsworth: Penguin.
Stubbs, M. (1983) Discourse Analysis, Oxford: Blackwell.
Sykes, M. (1985) ‘Discrimination in discourse’, in T. van Dijk (ed.) Handbook in
Discourse Analysis: volume 4, London: Academic Press, 83–101.
Tannen, D. (1979) ‘What’s in a frame?’, in R. Freedle (ed.) New Directions in Dis-
course Processing, Norwood, NJ: Ablex, 137–81.
Tannen, D. (ed) (1982) Spoken and Written Language: Exploring Orality and Lit-
eracy, Norwood, NJ: Ablex.
Tannen, D. (1988) Talking Voices, Cambridge: Cambridge University Press.
Tannen, D. (1991) You Just Don’t Understand: Women and Men in Conversation,
London: Virago.
Tannen, D. (1994) Talking from 9 to 5, London: Virago.
Taylor, T. J. (1981) Linguistic Theory and Structural Stylistics, Oxford: Pergamon.
Thompson, G. (1996) Introducing Functional Grammar, London: Arnold.
Todorov, T. (1977) The Poetics of Prose, Oxford: Blackwell.
Toolan, M. (1988) ‘Analyzing conversation in fiction’, Poetics Today, 8, 2, 393–416.
Toolan, M. (1998) Language in Literature, London: Hodder.
Traugoft, E. and M. L. Pratt (1980) Linguistics for Students of Literature, New
York: Harcourt Brace Jovanovich.
Trew, T. (1979) ‘Theory and ideology at work’, in R. Fowler et al., Language and
Control, London: Routledge & Kegan Paul, 94–116.
Tyler, S. (1978) The Said and the Unsaid: Mind, Meaning and Culture, London:
Academic Press, 1978.
Umiker-Sebeok, J. D. (1979) ‘Preschool children’s intraconversational narratives’,
Journal of Child Language, 6, 91–110.
Uppal, G. (1984) ‘Narratives in Conversation’, unpublished MA thesis, National
University of Singapore.
Uspensky, B. (1973) A Poetics of Composition, Berkeley: University of California
Press.
Verdonk, P. and J. J. Weber (1995) Twentieth-Century Fiction: From Text to
Context, London: Routledge.
Vestergaard, T. and N. Schroder (1985) The Language of Advertising, Oxford:
Blackwell.
254 Bibliography
Vincent, D. and L. Perrin (1999) ‘On the narrative vs. non-narrative functions of
reported speech: a socio-pragmatic study’, Journal of Sociolinguistics, 3(3),
291–313.
Voloshinov, V. N. (1973) Marxism and the Philosophy of Language, trans. L. Matejka
and I. R. Titunik, New York: Seminar Press.
Vuillaume, M. (1998) ‘Le discours indirect libre et le passe simple’, in S. Vogeleer, A.
Borillo, C. Vetters and M. Vuillaume (eds) Temps et Discours, Louvain: Peeters,
pp. 190–201.
Vygotsky, L. A. (1962 [1978]) Thought and Language, Cambridge, MA: MIT Press.
Wall, K. (1994) ‘The Remains of the Day and its challenges to theories of unreliable
narration’, Journal of Narrative Technique, 24, 18–42.
Warhol, R. (1986) ‘Toward a theory of the engaging narrator’, PMLA, 101, 811–18.
Warhol, R. (1989) Gendered Interventions: Narrative Discourse in the Victorian Novel,
New Brunswick, NJ: Rutgers University Press.
Warren, W. H., D. W. Nicholas and T. Trabasso (1979) ‘Event chains and inferences
in understanding narratives’, in R. Freedle (ed) New Directions in Discourse Pro-
cessing, Hillsdale, NJ: Ablex, 23–52.
Watts, R. J. (1984) ‘Narration as role-complimentary interaction’, Studia Anglia Pos-
naniensia, 17, 157–64.
Waugh, L. (1995) ‘Reported speech in journalistic discourse: The relation of function
and text’, Text, 15, 1, 129–73.
Weinsheimer, J. (1979) ‘Theory of character: Emma’, Poetics Today, 1, 185–211.
White, P. (1997) ‘Death, disruption and the moral order: the narrative impulse in
mass-media hard news reporting’, in F. Christie and J. R. Martin (eds) Genres and
Institutions: Social Processes in the Workplace and School, London, Cassell, 101–33.
White, P. (2000) ‘Media objectivity and the rhetoric of news story structure’, in E.
Ventola (ed.) Discourse and Community: Doing Functional Linguistics, Tübingen,
Gunter Narr Verlag, 378–97.
Wodak, R. (1981) ‘Discourse analysis and courtroom interaction’, Discourse
Processes, 3, 369–80.
Wodak, R. (1985) ‘The interaction between judge and defendant’, in T. van Dijk,
(ed.) Handbook of Discourse Analysis: volume 4, London: Academic Press, 181–91.
Wodak, R., R. de Cillia, M. Reisigl and K. Liebhart (1999) The Discursive Construc-
tion of National Identity, Edinburgh: Edinburgh University Press.
Wolf, D. (1984) ‘Ways of telling: text repertoires in elementary school children’,
Journal of Education, 167, 1, 71–87.
Wolf, G. et al. (1996) ‘Pronouncing French names in New Orleans’, Language in
Society, 25, 407–26; reprinted in R. Harris and G. Wolf (eds) Integrational Linguis-
tics: A First Reader, Oxford: Pergamon, 324–41.
Wolfson, N. (1976) ‘Speech events and natural speech’, Language in Society, 5,
189–209.
Wolfson, N. (1978) ‘A feature of performed narrative: the conversational historical
present’, Language in Society, 7, 215–37.
Wolfson, N. (1979) ‘The conversational historical present alternation’, Language, 55,
1, 168–82.
Wolfson, N. (1981) ‘Tense-switching in narrative’, Language and Style, 14, 226–30.
Wolfson, N. (1982) CHP, The Conversational Historical Present in American English
Narrative, Cinnarminson, NJ, Foris Publications.
Yacobi, T. (1981) ‘Fictional reliability as a communicative problem,’ Poetics Today,
2:2, 113–26.
Index
Aarseth 239 Bell 207, 239

Absalom, Absalom! 53 Bell and Garrett 239
abstract (Labovian) 148, 149–51 Bellow 141
active vs. passive voice 223 Bennett 166, 172
addressee 2–4 Bennett and Feldman 239
advertisements, narratives in 83–5 Bennett-Kastor 203
Age of Grief, The 99–100, 102–3 Berendson 77
Ahlberg and Ahlberg 200 Bickerton 140
À la recherche du temps perdu 46 Bissex 180
Amsterdam and Bruner 239 blame-attribution in political discourse 225
anachrony 43, 54–8 ‘Bliss’ 99, 116
analepsis 43–8, 55–7 Blommaert and Verschueren 235, 239
Anecdote, genre of in systemic account Bloor and Bloor 107
195–6 Bolinger 88, 239
annotation of clauses, Labovian 146–7 book blurbs 38
anti-narrative potential of the Internet 231–2 book reviews 38
Applebee 203 Booth 66, 79
Appraisal, systemic-linguistic theory of 225 Bradley 80, 112
Aristotle 7 Bronzwaer 140
Atkinson and Drew 237, 239 Brown and Yule 203
Atwood 234 Browne 200
Austen 91 Bruner 37, 172, 203
Austin 153 Bucholtz 239
Burglar Bill 200–1
Bakhtin 63, 104, 169, 194, 221, 227 Burton 111
Bal 11, 12, 33, 39, 43, 46, 47, 60, 61, 62, 67, 77, Butt et al. 107
86, 89, 111
Baldry et al. 173 Caldas-Coulthard and Coulthard 239
Bally 140 Cameron 227
Bamberg 173, 179, 203 Campbell 204
Banfield 135–6, 140, 141 canonical situation of utterance 139–40
‘Barn Burning’ 9, 53–4, 61, 75–6, 241 cardinal functions 22
Baron 140 Carter and Simpson 172
Barthelme 38 Carver 38, 101, 106
Barthes 10, 22–28, 37, 38, 47, 86, 194 catalysers 22, 57
‘Bath, The’ 101 Cather 234
Bauman 172 causality, children’s grasp of 185, 187, 188
Bayley 112 Cazden 203
Beaman 172 character(s) 41, 80–90, 93–9; traits/attributes
Bear Goes to Town 201–3 of 86–90 95–8; dialects of 127–8;
Beattie 234 representing discourse of 116–40;
Beckett 70 ontological status of 80–1
Bedient 112 Chatman 10, 12, 24–6, 28, 37, 50, 51, 64, 94,
behavioural processes 111 112, 114, 141, 169, 170, 172
256 Index
children’s narratives 178–205 direct and indirect writing 136–7
Chomsky 15–16 direct discourse 120–30; differences between
Chopin 234 direct and indirect discourse 125–30
Christie and Martin 203 discourse colony 13, 206
Christie et al. 20, 190, 203, 205 displacement potential 149
‘Circle of Prayer’ 240–1 Docherty 112
circumstantial elements in transitivity Dry 140
analysis 222 Dubliners 114
Citizen Kane 107 Duchan et al. 140
Clark 203 Dundes 38
class, stories of 233–4; impact of on duration 43, 48–53
narrative-reception 238 dynamic vs. stative verbs 32
Cobuild corpus 219
coda (Labovian) 148, 157–9 Eaton, Collis and Lewis 203
Coetzee 241 Edwards 179
Cohn 140 Eggins 107
Colby 7 Eggins and Martin 194–5
collective storytelling 162–7 Eggins and Slade 13, 195–6, 203, 225
collocation, unidiomatic or strained in hard ellipsis 56
news heads and leads 214–21 Ely 189
collocational incongruity 227 Ely, Bleason and McCabe 203
comparators, Labovian evaluative 156 embedded oral narratives 169–74
complicating action (Labovian) 148 Emmott 112
conjunctions in story-tellings 205 endophoric reference 206
construal of political events in news reports Erlich 140
214–30 estrangement, words of 72
contexts of narratives 178–9 ethnic prejudice in narratives 247–9
contextualism 169, 172, 206 evaluation: external (Labovian) 148, 151–3;
conversational historic present 161–2 narrative-clause-internal (Labovian)
Cook-Gumperz and Green 203 155–7; in literary narratives 170–2
coordinated and restricted clauses evaluative clause (Labovian) 146
(Labovian) 147–8 ‘Eveline’ 20–21, 25–28, 30–32, 37, 43, 49, 78,
correlatives, Labovian evaluative 156, 171 85, 112–3, 114, 140–1, 169
Cotterill 238 ‘Evening Sun, That’ 59, 64, 65, 77–8, 114, 175
Coulthard 238 events 6, 31–37
courtroom narratives 235–9 Exemplum, genre of in systemic account
Crawford, Brown and Nolan 172 195–6, 200, 235
critical discourse analysis 208 explicatives, Labovian evaluative 157
Crossley 172
Culler 26–28, 29–31, 37, 77, 112, 154, 172 fabula/sjuzhet 10–12, 15, 154
Culpeper 112 Fairclough 208, 239
cultural expectations, differences of in Farr 180, 203
storytelling 159–60 Faulkner 47, 59, 61, 66, 77
feminist narratology and narrative structure
Damilola/policing story, analysis of 234, 240–1
reporting of 212–21, 226–8, 232 Ferguson 140
‘Dead, The’ 43–4, 94–9, 107, 120–33, 171–2, field of discourse 178
174–5 Fillmore 140
Dean’s December, The 141 film narrative 103–7; editing of 105; multi-
Dear Zoo 204 modality of 104; inference in the reception
decontextualization of narratives 206 of 105–6; plenitude in 106
deep structure 15–16 Finegan 88
deictic, deixis 33–5, 59, 117, 126–7, 158 first vs. third-person narration 69–76
demonstrations, news reporting of 208–12 Fleischman 169–70
descriptive pause 56 Fludernik 77, 79, 130, 140, 141, 234, 239
Dickens 113 focalization 11, 39, 59–63; facets of 62–3;
dictation in children’s story-making 181 perceptual 63; vs. narration 64
diegesis 134 forensic linguistics 238–9
diffuse oral stories 169–70, 173 foreseeability 101–2
Dijk, van 208, 234–5, 239 Forster 94
Dillon and Kirchhoff 140 Fowler 37, 38, 69, 77, 89, 112, 239
Index 257
free clause (Labovian) 145–9 Hirose 140
free direct speech and thought 121 Histoire/discours 10–12, 15
free indirect discourse 76, 119, 130–40, 141, Hodge 239–40
169 Hoey 12, 13, 37, 154
free indirect speech and thought 76, 123–4 homodiegesis and heterodiegesis 43, 46
frequency 43, 53–4, 57 Hoover 112
Fromkin and Rodman 88 Hummel 140
Frow 112 Hymes 161, 172
functions, in Labov’s analysis 146–7 hypotaxis 126
functions, in Propp’s analysis 17–18
functions proper, in Barthes’ analysis 23–6 ideology 63, 207–8, 212–21
illustrations in children’s stories 197–203
Galda and Pellegrini 216 implied author 65–6
Gee 172 implied reader 67–8
gender, stories of 233–4 ‘Indian Camp’ 99–100
General Belgrano sinking 222 indices, in Barthes’ analysis 23–8
Genette 11, 12, 42–9, 52, 54–5, 60, 69–70, 77, indirect discourse 120–30
144, 145 inference, in character-interpretation 86–8
genres, systemic-linguistic approach to 189–97 in stories for children 197–203
Georgakopoulou and Goutsos 12 informants 22, 57, 97
Ginsberg 140 integrational linguistics 172, 174
Glasgow Media Group 256 intensifiers, Labovian evaluative 155
Go Down, Moses 123 intensive verbs 37
Goffman 161 interior monologue 122
gothic 91 Internet, and news narratives 230–2
grammatical metaphor 256–8 intransitive clauses, in headlines 223–4
grammaticization of character and situation intuitions about plot 20, 29–31
107–11 Invisible Man 49
grammaticization of plot structure 31–7 Iser 112
grammaticization of political emphasis
238–45 Jahn 77
Great Expectations 60, 64 James 31, 49, 89, 118, 170
Great Gatsby, The 100–1 Jameson 37
Greimas 21, 23, 81–5 Jane Eyre 69, 73–4
Grice 151 Jaworski and Coupland 194
grief and narration 45–6, 98 Jefferson A. 135
group narratives 162–7 Jefferson, G. 159
Johnson-Laird 203
habitual or iterative activity, and narrative Johnstone 172
events 34, 36, 57 Johnstone and Bean 172
Hagenaar 140 Jones 140
Hall 256 Joyce 47, 66, 94, 202
Halliday 107–111, 112, 178–9, 180, 208, 222 Julie Stays the Night 37–8
Halliday and Hasan 33, 179
hard news story 207 Kennedy 111
Hardy 175 Kermode 170
Harré 172 Kernan 188–9
Harris 5, 41, 172 kernels 24–5
Harry Potter and the Philosopher’s Stone ‘Killers, The’ 73
114–15 knowledge (background or extratextual)
Hart 141, 169 97–9
Harvey 112 Kramer-Dahl 109
Hasan 194, 203 Kress and van Leeuwen 173, 239
Hawkes 41, 256 Kullmann 140
Hawthorn 2–3
headlines, in newspapers 258 Labov 102, 143–59, 167–72, 173, 174, 176,
Heath 160, 182–5, 203 186, 188–9, 194, 203, 208, 234
Heffer 237 Labov and Waletzky 143–8, 173
Hemingway 73, 99 Labovian approach and literary narrative
Hernadi 140 167–72
Hickmann 189 Lakoff 236
258 Index
Lanser 77, 234, 239 naming and description 225–7; of characters
Laughter in the Dark 50 101, 141, 205; and processes 240–1, 261
Lawrence, D. H. 75, 79 narratee 67–8
‘leap-frog’ narratives, children’s 186 narration 10–11, 41–76
Leech and Short 89, 124, 134, 136, 137–9, 140 Narration of Internal States 142
van Leeuwen and Wodak 235 narrative, definition of 6–10; Labov’s
Leitner 256 definition of 148
Levinson 138, 227 Narrative, genre of in systemic account
Levi-Strauss 41, 42 195–6
lexical choice 225–6 narrative, typical characteristics of 4–6
lexical ‘gaps’ 226 narrative clause (Labovian) 145–9
lexical vs. prosodic marking of textual narrative constructedness 4
connections 204–5 narrative development in children 193–202
Life and Times of Michael K., The 241–2 narrative displacement 5
Lind and O’Barr 235–7, 239 narratives in court 235–9; and the law’s
Linde 172 paradigmatic counter-pressures 237
literacy, emergence of in children 180–93 narrative report of speech or thought acts
‘Little Things’ 106–7 122–3
Lodge 77, 112 narrative trajectory 4–5
Longacre 12, 13 narrative vs. conversation 3
Lothe 79 narratorial omniscience, degrees of 69
Lubbock 49 narratorial modes, Simpson’s typology of
Luelsdorff 112 68–76
Lyons 139–40 narrator(s) 2, 67–76; trust in 3
Narrator’s Report of Voice 142
MacPherson 213 natural narrative and literary narrative,
McCabe and Peterson 185–7, 195 unified approach to 168–9, 171–2
McConnell-Ginet and Eckert 172 Neisser and Fivush 172
McHale 135, 136, 140 Nelles 77
McKay 140 neutral modes of narration 73–5
Mandler and Johnson 203 newspaper narratives 206–21
Mansfield 99, 116, 234 Nicolopoulou 203, 204
Mansfield Park 114, 137 nominalization 32–3, 224–5
Margolin 112 ‘Non-narratives’ 8–10
Martin 179, 191–3, 225 norms, function of in stories for children
material processes 109–10 211–6
Matthiessen 179 norms of speech and thought reporting
mental processes 110 137–9; and scalar implicature 138–9
metaphors in courtroom discourse 238 Norrick 172
Mezei 140, 239 Nünning 79
Michaels 182, 183, 204
Michaels and Collins 182–5 observation-comment 190
Middlemarch 112 observer’s paradox 174
Mills 234, 239 O’Barr 238, 239
mimesis 134 Ochs 174
Mitchell 194 ‘Odour of Chrysanthemums’ 75
modality in narration 71–6; and evaluation Oltean 140
225; positive and negative modality 71–2; Onega and Garcia 37
in free indirect discourse 133 Ong 68
mode of discourse 179 online news and newspapers 230–2
Molloy 70 ontological status 90
Mooty and Grandma 197–200 oral narratives of personal experience 143–72
morphology of the Russian fairytale 14–20 oral vs. literate discourse strategies 182–5,
Mrs Dalloway 51–2 204–5
Mugabe, President 229–30 orbital structure of hard news story 207
Muhlhausler and Harré 178, 203 order, temporal 42–8
Munro 234, 240–1 ordinariness-construction 176
Murray 239 orientation 60–3; Labovian orientation 148,
Myers 140 149–51
Nabokov 44, 49 pace of narrative 48–53

Index 259
Palmer 112 Rimmon-Kenan 11–12, 31, 37, 42, 49, 60, 62,
paradigmatic 10, 22, 41 63, 67, 69–70, 77, 81, 90, 111
parataxis 126 Roberts,8, 234
Pascal 135, 140 roles, in Propp’s analysis 18–19
passivization 223 Romaine 203
Payne 172 Rothery and Martin 189–90, 203
van Peer and Chatman 77 Rowling 114
perceptual focalization 63–4 Rubin 112
Perez and Tager-Flusberg 203 Rumelhart 203
performed narrative 160–2 Ryan 77
Peterson and McCabe 185–7, 190, 203
phonology and phonetics, analogy with Sacks 175–6
13–14 Sauerberg 112
Piaget 185, 187, 203 Saussure
pictures and depiction in stories for children Saville 210
197–203, 204 scene (and summary) 49–51
plans, and narrative 13 Schank and Abelson 203
Plum 194, 196 schema-theory 39
Pnin 44–5, 47, 54–8, 98 Schiffrin 12, 13, 162, 172
poetics of narrative 9 Scholes and Kellogg 12
point of view 68 Scholes and Litz 86
Polanyi 159, 162, 166, 172 science, narratives in xi
political narratives 206–39 Semino, Short and Wynne 136
popular culture, narratives in 83 sense-making, through narratives 13–14
Portrait of a Lady, The 117–19 sequences, in Barthes’ analysis 23
Portrait of the Artist as a Young Man, A 69, setting 39, 91–9; causal vs. analogical 92–3;
149–50, 154–7, 158–9 and character-stereotyping 93–4; houses as
powerful and powerless language 236 113–4
Pratt 167–8, 172 sharing time, and children’s storytelling
prefabrication of narratives 4 development 182–5
presuming vs. presenting reference 191–2 Shields 234
presupposition 227–9 Short 112, 120
Price 112 Short, Semino and Culpeper 136, 141
Prime of Miss Jean Brodie, The 47 Short, Semino and Wynne 140
Prince 37, 38, 77 Silva-Corvalan 172
processes and participants, grammatical Simpson 37, 38, 68–76, 77, 112, 118, 136, 140,
108–11, 205 170–1, 227
projecting and projected clauses 126 Sinclair and Coulthard 194
prolepsis 43–8, 58; proleptic traces 47 Singaporean oral narrative 163–6
Propp 7, 16–22, 37, 38, 144, 235 Slembrouck 140
Proust 46 ‘Small, Good Thing, A’ 101
psycholinguistic approach to narrative and Smiley 99, 234
story 203 Smith 38
‘Pupil, The’ 89 social constructionism 178–9
Sound and the Fury, The 61, 66, 79
racial or ethnic prejudice in narratives 234–5 speech and thought in narratives,
reactive sequences 186–7 representation of 116–40
Recount, genre of in systemic account 195–6 Stanzel 140
Redeker 140 Star Wars 19
reference, development of in children’s stative vs. dynamic verbs 32
stories 191–3, 205 Stein and Albro 178
reflector narration 70–6 Stein and Glenn 203
register and genre 189–93 stereotyping 105
register-development 190–3 Sternberg 130, 140
Reilly, Bates and Marchman 203 Sterne 79
relational processes 111 stories for children 178–80, 197–203
resolution (Labovian) 146 story and discourse 10–12, 15
Return of the Native, The 49, 167–8 story-dictation, by children 187
Reznikoff 38 story genres 193–7
Richardson 234 story-grammatical analysis 186–7, 203
Rifelj 77 stream of consciousness 122
260 Index
stretch 50 Ulysses 61
structuralist theory of plot, evaluation of Umiker-Sebeok 188
25–31 Une Glossaire/A Glossary 8–10
Stubbs 174 unreliability of narrator 79
styles of storytelling 182–5 unspeakable sentences 135–6
summarizing plot 29–31 Uspensky 62, 69, 77
summary (and scene) 49–53 Uppal 163, 166
Sun Also Rises, The 49, 53
surprise and suspense 99–103 Ventola 203
Sykes 239 verbal hygiene 227
syntagmatic, 10 verbal processes 111
systemic-linguistic analysis of narrative, verba sentiendi 72, 76
189–203, 205 Vestergaard and Schroder 83–4
Vincent and Perrin 140
tale 1–4, 65 Voloshinov 140
Tannen 159, 160, 172 vraisemblance 60
teleology of narratives 26–7, 231–2 Vuillaume 140
teller 1–4, 5, 65 Vygotsky 180–1, 199, 203
temporal juncture 144–6
tenor of discourse 178 Wall 79
testimony, styles of in court 235–7 Warhol 77, 234, 239
text 10–11 Warren, Nicholas and Trabasso 199, 203
theme 34, 205 Watts 172
Thompson 107 Waugh 140
time 39, 42–58; and repetition 79 Wee and Kwan 198
Todorov 6–7, 37 Weinsheimer 81
Toolan 112, 170, 172, 227 What Maisie Knew 61
topic-associating vs. topic-centred White 203, 207–8, 225, 239
storytelling style 183–5 Widgery 210
traits of characters 86–90, 95–8 Wodak 235, 238, 239
transformation 6–7, 15–16 Wolf, D. 205
transcription of oral narratives 173–4, Wolf, G. 174
176–7 Wolfson 161–2, 172, 174
transitivity, systemic-linguistic account of Woolf 234
107–111, 114, 222–3 WTO–Seattle demonstrations, reporting of
Trew 208–9, 210, 229, 239 210–12, 222
Tristram Shandy 67
turbulence, in news report construals of Yacobi 79
complex events 221
Turn of the Screw, The 62 Zimbabwe (Rhodesia), in news reporting
Tyler 160 208–10, 229–32

Untitled

Uploaded by

Copyright:

Available Formats

Untitled

Uploaded by

Document Information

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

Untitled

Uploaded by

Copyright:

Available Formats

Narrative

‘An excellent critical survey of recent research in narratology and linguis-

Michael Toolan is Professor of Applied English Linguistics at the Univer-

Language, Literature and Critical Feminist Stylistics

The Series Editor

London and New York

2 Basic story structure 15

3 The articulation of narrative text I: time, focalization,

5 The articulation of narrative text III: representing

6 Narrative as socially situated: the sociolinguistic

8 Narrative as political action 206

Narratives are everywhere, performing countless different functions in

1.1 Teller, tale, addressee

However, since tellers can become intensely absorbed in their self-

In short, narratives always involve a Tale, a Teller, and an Addressee, and

Narrative focusses our attention on to a story, a sequence of events,

One of the distinctive characteristics of narrative concerns its necessary

1.2 Typical characteristics of narratives

1 A degree of artiﬁcial fabrication or constructedness not usually appar-

and notice the ﬁnality and permanence conveyed by the ever/never

And that is the end of the story.

which has near-identical counterparts in the closing sequences of radio

Bees do not regale one another with reminiscences of the nectar

This is a lovely image (or narrative), partly because in fact it is some-

A ﬁrst attempt at a minimalist deﬁnition of narrative might be:

a perceived sequence of non-randomly connected events

This deﬁnition recognizes that a narrative is a sequence of events. But

The simple relation of successive facts does not constitute a narrative:

The one-line deﬁnition of narrative also suggests that consequence is not so

Folk narrative in its simplest form is the verbal description of one or

1.3 Narratives and non-narratives

A narrative is a perceived sequence of non-randomly connected

This deﬁnition introduces (1) one or more foregrounded individuals as

• sequenced and interrelated events;

Now consider a literary text such as Michelle Roberts’ novella – if that is

1.4 Story – text – narration

A fabula [story] is a series of logically and chronologically related

STORY – TEXT – NARRATION (as in Rimmon-Kenan, 1983)

FABULA – STORY – TEXT

In general I shall use the same terms as Rimmon-Kenan wherever pos-

Functions of characters serve as stable, constant elements in a tale,

Here, for convenience in subsequent analytical tasks, is a full list of

1 One of the members of a family absents himself from home. (An

Villain: Darth Vader

Initial situation 1 Once upon a time there was a bunny named

One of the members of a family absents [himself] from home.

One of the members of a family reconsiders a decision to absent

While Propp’s fairytales proceed through developmental actions, ‘Eveline’

2.3 Barthes on narrative

Narrative analysis is condemned to a deductive procedure, obliged

A discourse is a long ‘sentence’ . . . just as a sentence . . . is a short ‘dis-

More particularly, Barthes emphasizes the need to separate different

1 functions (as in Propp, Bremond);

not to a complementary and consequential act but to a more or less

They include indices to characters’ psychological states, notations of

(a)1 Cardinal functions or nuclei or, to use Chatman’s useful term,

Indices involve an act of deciphering, the reader is to learn to know

A sequence is a logical succession of nuclei bound together by a rela-

For example, ‘having a drink’ is suggested as a closed sequence with the

It was John who