Carol Myers-Scotton - Codes and Consequences PDF

Codes and Consequences:
Choosing Linguistic Varieties
Carol Myers-Scotton,
Editor
OXFORD UNIVERSITY PRESS

Codes and Consequences
This page intentionally left blank
Codes and Consequences
Choosing Linguistic Varieties
Edited by
Carol Myers-Scotton
New York Oxford

OXFORD UNIVERSITY PRESS
1998
Oxford New York
Athens Auckland Bangkok Bogota Bombay
Buenos Aires Calcutta Cape Town Dar es Salaam
Delhi Florence Hong Kong Istanbul Karachi
Kuala Lumpur Madras Madrid Melbourne
Mexico City Nairobi Paris Singapore
Taipei Tokyo Toronto Warsaw
and associated companies in
Berlin Ibadan
Copyright © 1998 by Carol Myers-Scotton

Published by Oxford University Press, Inc.
198 Madison Avenue, New York, New York 10016
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data

Codes and conseuqences: choosing
linguistic varieties / edited by Carol Myers-Scotton
p. cm.
Includes bibliographical references and index.
ISBN 0-19-511522-8; 0-19-511523-6 (pbk.)
1. English language—Discourse analysis.
2. English language—Social aspects.
3. English language—Variation.
I. Myers-Scotton, Carol.
PE1422.C63 1998
401'.41—dc21 98-7061
1 3 5 7 9 8 6 4 2
Printed in the United States of America
on acid-free paper
ACKNOWLEDGMENTS
The editor, authors, and publisher gratefully acknowledge permission to re-

produce extracts from the following:
All the Pretty Horses by Cormac McCarthy, © 1992, Alfred A. Knopf. Re-
print, 1993 Vintage International. By permission of Alfred A. Knopf.
"Stylistic variation and cognitive constraint in Cormac McCarthy's All the

Pretty Horses," © 1995 in Wade Hall and Rick Wallach (eds.) Sacred Vio-
lence: A Reader's Companion to Cormac McCarthy. El Paso: Texas Western
Press. By permission of Texas Western Press.
The Great Gatsby by F. Scott Fitzgerald. © 1925, Charles Scribner's Sons.

Copyright renewed 1953 by Frances Scott Fitzgerald Lanahan. © 1991,1992
by Eleanor Lanahan, Matthew J. Bruccoli, and Samuel J. Lanahan as Trust-
ees under agreement dated 7/3/75 created by Frances Scott Fizgerald Smith.
Excerpts from The Great Gatsby ("authorized text") by F. Scott Fitzgerald.
Edited by Matthew J. Bruccoli. By permission of Scribner, a division of Simon
& Schuster.
As I Lay Dying by William Faulkner, from © 1930, renewed 1957. From

1985, Faulkner, Novels 1930-1935. New York: Library of America (Liter-
ary Classics of the United States, Inc. By permission of Random House, Inc.
The Sirens of Titan by Kurt Vonnegut, Jr. © 1959 Dell. By permisson of Kurt
Vonnegut, Jr., and Dell Publishing Co.
"The Snow Man" from Collected Poems by Wallace Stevens. © 1923, re-
newed 1951 by Wallace Stevens. By permission of Alfred A. Knopf, Inc.
"The Tyger," "A Poison Tree" and an excerpt from "London" by William
Blake from Songs of Innocence and of Experience. © 1977. New York: Oxford
University Press. By permission of Oxford University Press.
I am very grateful to Travis Gordon for his fine work in formatting this
volume. I wish to thank Janice L. Jake and Agnes Bolonyai for comments on
my own contributions to the volume. I also appreciate the help of Cathy
Bridgeman, Noreen Doughty, and CeCe Mikell as editorial assistants.
CONTENTS
Contributors ix
I. OVERVIEW
1 Introduction 3
Carol Myers-Scotton
2 A Theoretical Introduction to the Markedness Model 18
Carol Myers-Scotton
II. STYLISTIC CHOICES IN LITERATURE

3 Implicatures of Styleswitching in the Narrative Voice
of Cormac McCarthy's All the Pretty Horses 41
Nancy Kreml
4 Marked Grammatical Structures: Communicating Inten-
tionality in The Great Gatsby and As I Lay Dying 62
Carol Myers-Scotton
5 Markedness and References to Characters in Biblical
Hebrew Narratives 89
Timothy Wilt
6 Literariness, Markedness, and Surprise in Poetry 101
Amittai F. Aviram
7 Villainous Boys: On Some Marked Exchanges in Romeo
and Juliet 124
Trevor Howard-Hill
III. STYLISTIC CHOICES IN SPOKEN ENGLISH

8 Markedness and Styleswitching in Performances
by African American Drag Queens 139
Rusty Barrett
9 Styleswitching in Southern English 162
Margaret Mishoe
10 Marked Versus Unmarked Choices on the Auto
Factory Floor 178
Janice Bernsten
IV. STYLISTIC CHOICES AND SECOND-LANGUAGE ACQUISITION

11 "Not Quite Right": Second-Language Acquisition
and Markedness 195
Mary Sue Sroda
Index 215
CONTRIBUTORS
Amittai F. Aviram is associate professor of English and comparative litera-

ture and a consulting faculty member in the linguistics program at the Univer-
sity of South Carolina, where he teaches literary theory and the reading of
poetry. He has published a research monograph, Telling Rhythm: Body and
Meaning in Poetry (1994) and a chapbook of his own poems, Tender Phras-
es, Brassy Moans (1994). His poems have also appeared in magazines and
anthologies.
Rusty Barrett is a doctoral candidate in the linguistics department at the

University of Texas-Austin. In addition to his work on language and gender/
sexuality, he has conducted a great deal of research in Mayan linguistics.
Based on fieldwork in Guatemala, his dissertation is a grammar of Sipaka-
pense Maya.
Janice Bernsten is an associate professor of linguistics at the University of

Michigan's Flint campus. She is a sociolinguist continuing her work on the
auto factory floor. In addition, she pursues research on African linguistic
issues, especially in East and Southern Africa. She is the coauthor of a gram-
mar of Shona designed for U.S. university students.
Trevor Howard-Hill is the C. Wallace Martin Professor of English at the

University of South Carolina. He is also the editor of the Oxford Shakespeare
Concordances (Oxford: Clarendon Press, 1969-1973). He has written exten-
sively on bibliography, textual criticism, and editing and is editor of Papers
of the Bibliographical Society of America. His most recent book is Middle-
ton's "Vulga Pasquin": Essays on "A Game of Chess" (1995).
Nancy Kreml is a member of the English faculty at Midlands Technical Col-

lege in Columbia, South Carolina. She received her Ph.D. in linguistics from
the University of South Carolina in 1992. She is the author of several articles
on Cormac McCarthy and on stylistics in writing about science.
Margaret Mishoe is a member of the English faculty at the University of

North Carolina-Charlotte, where she teaches English and linguistics cours-
es. She received her Ph.D. in linguistics from the University of South Caro-
lina in 1995. Her principal areas of research are the structure of dialects
among lower socioeconomic class speakers in the South and their use of
styleswitching.
X CONTRIBUTORS
Carol Myers-Scotton is a Carolina Distinguished Professor at the University

of South Carolina where she is a professor of linguistics and English and has
been director of the linguistics program. She has done extensive fieldwork in
Africa on bilingual situations and has published widely on codeswitching,
including two books, Social Motivations for Codeswitching: Evidence from
Africa (1993) and Duelling Languages: Grammatical Structure in Codeswitch-
ing (1993), both with Oxford.
Mary Sue Sroda is director of the M.A. TESOL program at Murray State
University (Kentucky) and a doctoral candidate in the linguistics program at
the University of South Carolina. Her research interests are second-language
acquisition and pragmatics and she has coauthored an article on perceptions
of reference in young children. Her dissertation is entitled "Relevance Theory
and the Markedness Model: A Cognitive Approach to Second-Language
Acquisition."
Timothy Wilt is a translation consultant for the United Bible Societies. He

has worked with translation teams in Africa for several years, especially in
Zaire, Cameroon, and Chad. He holds a Ph.D. in linguistics from Michigan
State University.
1
Overview
1
Introduction
CAROL MYERS-SCOTTON
A s THE TITLE of this volume indicates, the subject is the selection of one
linguistic code (or variety) rather than one or more others and its conse-
quences. Code and variety are cover terms for linguistic systems at any level,
from separate languages to dialects of a single language to styles or substyles
within a single dialect. As all linguists know, they are useful—even necessary—
terms because the divisions between linguistic systems are not always discrete. In
the intended meaning here, consequences are easier to label, although this does
not mean they are necessarily separable, either. The consequences considered here
have to do with either effects on ease of cognitive processing or social/psycholog-
ical effects. However, all effects are cognitive in the sense that they depend on the
addressee's making mental calculations about the speaker's intentions. (Address-
ee will be used as a cover term for addressee and reader, and speaker as a cover
term for speaker and writer.)
Intentionality refers to the messages conveyed by utterances in addition to
those which the utterances literally denote (i.e., more than the meaning of
words and their combinations). Grice (1975) gave the name implicatures to
such consequences or outcomes, and one of the ways in which Sperber and
Wilson (1995) extend Grice's ideas is by referring to implicatures as both strong
and weak. In my own work, I generally refer to intentionality rather than to
implicatures. I see the relation between the two terms as this: Intentionality
becomes apparent in implicatures.
Sperber and Wilson argue that any utterance can introduce a number of
implicatures. They also reject the notion that implicatures are necessarily
determinate, that speakers always have in mind a specific set of implicatures
that they expect addressees to recover. They say, "Some implicatures are
made so strongly manifest that the hearer can scarcely avoid recovering them.
Others are made less stongly manifest" (Sperber and Wilson 1995:197). While
the analyses in this volume make specific claims about the intentional
messages that choices of linguistic varieties convey, they also support the point
3
4 OVERVIEW
of view of Sperber and Wilson that we are dealing with open-ended sets of
intentions or implicatures.
Messages of intentionality can be conveyed in various ways. The examples
in Sperber and Wilson 1995 rely on the way an utterance is phrased to convey
implicatures but without the type of variation in phrasing involved in a change
in linguistic variety. In contrast, the data sets considered in this volume were
selected to show how intentionality can arise through changes in linguistic
variety, from one style to another.
Messages of intentionality are additional messages but not necessarily
secondary; in fact, messages of intentionality may be the primary message
in a discourse. For example, in Barrett's analysis of performances by
African American drag queens (chapter 8), the message of intentionality
conveyed by styleswitching is, in many ways, the main message. This
message is that social categories related to gender, ethnicity, class, and
sexuality are not stable. Also, based on my own discussion of Faulkner's
As I Lay Dying (chapter 4), I would argue that the psychological effects
of the style Faulkner imparts to Darl give the reader a better key to under-
standing Darl's character than the referential import of his monologues. In
other chapters, the consequences of using one style rather than another may
be more than an overlay on referentially based messages but less than the
primary messages.
MARKEDNESS AND CHOICES OF VARIETIES

A major theme in this volume is how the construct of markedness can be
related to the choice of one linguistic variety over other possible varieties. The
argument developed in each of the chapters is that writers or speakers choose
what can be considered marked choices to convey certain messages of inten-
tionality. To varying degrees, authors discuss their data in terms of the dis-
tinction between unmarked and marked choices of linguistic varieties, a
distinction formalized in the Markedness Model (Myers-Scotton chapter 2, in
this volume; but also 1993).
Markedness in Structural Linguistics

Markedness, of course, is a well-known concept in structural linguistics where
it has been used at least since the first half of the twentieth century to explain
how linguistic systems are structured (Trubetzkoy [1939] 1969; Jakobson
1971; Battistella 1990,1996). Specifically, markedness figures in four ways
in structural analyses. First, it is an organizational construct used to describe
various polarities within the different systems of language. For example,
within phonology, two sounds may be related in terms of their articulatory
properties, but one member is referred to as unmarked because it requires
fewer distinctive features in its description than does the marked member.
Second, markedness also is used in reference to frequency. The more un-
Introduction 5
marked member of a set is that which occurs more frequently. For example,
a particular word order in dependent clauses in a given language is called
unmarked, with other orders called marked. Third, and related to frequency,
one member of a pair may occur as a cover term for both members. For
example, in Spanish muchachos 'boys' functions as a cover term for both
muchachas 'girls' and muchachos. Fourth, markedness is used in reference to
language change. A general argument (but with many dissenting views) is that a
more marked structure is susceptible to change (i.e., becoming more unmarked)
whether through internally or externally motivated change.
All these uses of markedness are largely descriptive although they could
also be incorporated into a theoretical (i.e., an explanatory and/or predictive)
framework. For example, within the theory of universal grammar (UG), marked-
ness relates structural features to the grammar as core versus periphery. This
notion is extended in research in language acquisition. However, within UG
markedness is not absolute. For example, consider head position. The un-
marked version of the head parameter is that all languages have their
heads on the same side of their complements. In a specific language, the
central core may be considered unmarked with respect to head direction
(e.g., in English, nouns as heads precede their complements, as in movie about
dinosaurs). However, parts of the structure may be marked; for example, in
English, determiners and adjectives precede the noun head, as in the new
movie. (See Cook and Newson 1996 for a discussion of Chomsky and uni-
versal grammar.)
Markedness and Social Consequences

In my various discussions of the markedness model (MM), I use the concept
of markedness in ways related to those discussed previously. However, I am
not interested in linguistic forms as members of a set based on structural
characteristics but rather of a set based on social or psychological import. My
use of the term markedness is meant to imply that choices within such a set
also can be characterized within a system of oppositions, with markedness as
a property of oppositions. The premise underlying this approach is that all
linguistic codes or varieties come to have social and psychological associ-
ations in the speech communities in which they are used. Given these
associations, the use of a particular code is viewed in terms of the
unmarked versus marked opposition in reference to the extent its use
"matches" community expectations for the interaction type or genre where it
is used: What community norms would predict is unmarked; what is not
predicted is marked. However, this approach does not assume that these
oppositions are necessarily categorical in reference to markedness. Rather,
codes fall along a continuum as more or less unmarked. Further, there is
not necessarily a single unmarked or marked choice, although there is often
a dominant unmarked choice (in terms of frequency and in reference to the
6 OVERVIEW
norms of the dominant group), especially when the situation or genre is

relatively conventionalized.
The MM employs the unmarked versus marked distinction as a theoretical
construct to explain the social and psychological motivations for making
one code choice rather than another. The MM itself is based on the premise
that, as part of their innate language faculty, all language users have a
predisposition to view linguistic choices as more or less marked, given the
social and intellectual context. That is, all persons come equipped with the
competence to assess linguistic choices. To develop the ability to put this
competence to use, everyone must be exposed to language use in specific
interaction types in their speech community. Thus, all language users develop
a sense of markedness regarding varieties in the community's linguistic
repertoire for given interaction types or for given genres in writing in their
community.
In brief, choices are labeled unmarked when they constitute predicted
behavior, given the context; that is, they negotiate the unmarked rights and
obligations set for that context. In some sense, they maintain the community's
status quo. Choices are marked when they can be seen as negotiations to
invoke a rights and obligations set other than the unmarked one for the
context. (See chapter 2 in this volume for more details on the MM.)
Applying the Markedness Model to New Types of Data

Until now, I mainly employed the MM to explain choices in codeswitching in
conversation. Further, even though codeswitching can occur at any structural
level of language, I dealt largely with switching between two or more languag-
es (e.g., Myers-Scotton 1988a, 1993; but see Myers-Scotton 1985,1988b, on
switching between styles and Myers-Scotton and Zhu 1983 on choice of terms
of address). Because I often wrote about switching between languages in
conversation whenever I presented the MM, many assumed that I intended
such switching as the model's primary purview. Also, most other researchers
who apply the MM in their work use it to discuss switching between languages
(e.g., Gibbons 1987, Swigart 1992, Adendorff 1993, Wei 1994; but see East-
man 1985 and DeBose 1992 on style or dialect switching). Because it was
never my intention to limit the MM to switching between languages, a major
motivation for editing this volume is to demonstrate that the MM can be used
profitably in other ways.
The goal of this volume, therefore, is to show how theoretical frame-
works that include the MM—or at least the concept of markedness—can be
applied profitably to other levels of language. Therefore, there are no
chapters on codeswitching between languages. Instead, all the chapters
deal with choices among available styles or subsets of styles (e.g., direc-
tives, in Bernsten chapter 10). The choices are among styles in English in
all cases but one—and even that one treats English translations of styleswitching
(Wilt chapter 5).
Introduction 7
EXPLAINING STYLISTIC CHOICES IN DIVERSE AREAS
The table of contents gives at least a clue to the diversity of subject matter.
Chapter 2 offers a general theoretical overview of the MM as I conceptualize
it today. The next nine chapters offer applications of how the concept of
markedness contributes to the social or psychological message in different
choices of linguistic form. Although some studies are clearly within the scope
of sociolinguistics, because many messages of intentionality in conversation
are about a speaker's view of his or her own persona or social group member-
ships and relationships to other participants, the way that the authors view
choices among linguistic alternatives also places this volume within cogni-
tively based pragmatic studies as well. This is clear in the two ways that the
notion of speakers as rational actors is developed.
First, viewing language users as rational actors means supporting the premise
that speakers make choices from an opportunity set (i.e., their linguistic reper-
toires) to achieve certain goals important to them as individuals. In so doing, these
speakers must calculate the costs and rewards—social or otherwise (e.g., aesthet-
ic)—of one choice over another. This is what distinguishes these studies from
sociolinguistic investigations that concentrate on language use as largely identi-
fying the user with social group memberships or social networks.
Second, viewing language users as rational actors includes the notion that
their choices demonstrate that these users not only recognize, but also exploit, the
potential for utterances to convey intentional as well as referential meaning. As
noted earlier, intentional messages enlarge the referential or social meaning.
This comment about style conveys some of the flavor of the questions which
this volume addresses:
Whenever a speaker communicates he must make a decision as to what he choos-
es to make explicit and what he chooses to leave implicit. It is not as if [I promise
that Tom is coming] communicates something different from [Tom's coming].
Rather, they differ in the amount of help the reader is given in recovering whatever
is communicated. In other words, they differ in style. (Blakemore 1992: 7-8)
THE INTENDED AUDIENCE
The volume is designed to interest faculty and students in the disciplines of

linguistics, literature, and foreign language. Because it deals with the expli-
cation of literature as well as spontaneous speech, it is a useful text for
courses in either literary stylistics or rhetoric and composition in English
departments. It also addresses issues covered in discourse analysis or
(socio)pragmatics in linguistics departments. Moreover, the volume ought to
be accessible to the general educated reader. Technical discussions are
minimal and when they occur, technical terms are fully explained, with the
nonspecialist in mind. None of the chapters has been published before,
except for portions of chapter 3.
8 OVERVIEW
ORGANIZATION OF THE VOLUME
Following this introduction, Part I includes a chapter giving a theoretical

overview of the markedness model. The five chapters in Part II deal with
marked choices in literary works, ranging from Shakespeare to poetry to
modern novels to Hebrew biblical narratives. Part III includes three chapters
that analyze spontaneous speech. Their subject matter, however, is very di-
verse, ranging from directives at an auto plant to ingroup conversation among
smalltown southerners to drag queen performances. Part IV consists of only one
chapter, which considers ways in which to use markedness in instructing
second-language learners of English on how to speak and write as communi-
catively competent users of the language. What differentiates almost all the
chapters from discussions of stylistic effects by literary stylists is that most of
the analyses are supported with quantitative evidence to corroborate the claims
that are made.
The chapters vary in the extent to which the MM figures in the explanations
of linguistic choices. Several of the chapters specifically use the MM as their main
theoretical framework; for example, Mishoe (chapter 9) employs the MM exclu-
sively in explaining her data set. Others use several theoretical approaches. For
example, Kreml (chapter 3) utilizes the MM along with relevance theory
(Sperber and Wilson). Others refer only to the general notion of markedness in
regard to text types or linguistic varieties (Wilt chapter 5; Aviram chapter 6).
THE MARKEDNESS MODEL AS A RATIONAL ACTOR MODEL
In chapter 2, "A Theoretical Introduction to the Markedness Model," I present

an overview of the MM as it stands today. Those familiar with earlier versions
of the model will certainly recognize it; however, the current version goes
beyond previous explications in emphasizing how the MM is a rational actor
model. Rational actor models are concerned with explaining why actors (speak-
ers) make the choices they do; such models view rationally based choices as the
means by which actors try to achieve their goals as well as possible, thereby
optimizing an outcome.
With the MM explicated as a rational actor model, linguistic choices are
now cast in terms from Elster's (1989) conception of how social choices are
made. I also introduce the somatic marker hypothesis of Damasio (1996).
Features of the social context, including the speaker's social identity features,
are considered structural constraints that determine an opportunity set; in the
case of linguistic choices, the opportunity set is a speaker's linguistic repertoire.
A speaker's opportunity set (and the structural constraints that "produce" it),
thus, are external constraints on the speaker's choices. At this point, I introduce
a second set of constraints that are internal to the speaker. I have in mind the
speaker's innately available "markedness evaluator" and also somatic mark-
ers, as characterized by Damasio. Finally, as a third filter, rationality comes
Introduction 9
into the picture. To act rationally means that the speaker selects the "best"
choice, given the degree of internal consistency of his or her desires, values, and
beliefs and the available evidence.
In concluding the chapter, I emphasize that the attempt to optimize on some
goal limits choices in ways that features of neither the social nor structural
context can do: The important term regarding choices is not possible but
feasible. By feasible I mean choices that are not simply accessible but deemed
advantageous based on the prescriptions of rationality just stated. Such choices
are necessarily the result of generally unconscious cognitive calculations.
Few of the following chapters in which specific data sets are analyzed in
reference to markedness mention the MM explicitly as a rational actor model
and none discusses the mechanisms of choice as outlined previously. Most are
much more concerned with identifying a particular style as marked or not and
with explaining the social or cognitive effects of the use of a marked style when
it occurs. However, in reading my theoretical chapter together with the anal-
yses of specific data sets, I hope the reader can see that what the authors are
doing is applying my new emphasis on the mechanisms of linguistic choice
(i.e., the how and why of choice).
LITERARY STUDIES
Part II consists of five chapters on literary works. It opens with chapter 3,

"Implications of Styleswitching and the Narrative Voice of Cormac McCar-
thy's All the Pretty Horses," in which Nancy Kreml is concerned with how two
styles play against each other. Her analysis shows that almost all of the novel
is written in one style (the unmarked style) that shows the action of the novel,
but action for which McCarthy does not explicitly assign cause and effect.
A second style (the marked style) occurs very infrequently, but it suggests
how the reader is to interpret the actions. She provides a quantitative
analysis of linguistic features (e.g., length of sentences, use of metaphors, and
types of lexical choices) to substantiate the existence of the two different styles.
Kreml employs both relevance theory (Sperber and Wilson 1995) and the
MM, and her analysis can be seen as arguing that the two styles contrast at
the level of implicature as well as in terms of their surface linguistic
features. Although both styles introduce many implicatures, those in the
unmarked style are left very open: By describing an event, McCarthy indi-
cates it is important, but he leaves indeterminate how is it important. In
contrast, implicatures arising from the more marked style are highly con-
strained in various ways. For example, not only is the marked style's often
convoluted syntax at odds with the simple syntax of the world of action,
but such structures give rise to implicatures that introduce the notions that
actions are complex and have consequences, that they do not just happen.
Similarly, the metaphors (e.g., stars swarmed), many modifications (e.g.,
young thieves in a glowing orchard), and uncommon lexical choices (e.g.,
10 OVERVIEW
serried) of the marked style require the reader to spend more effort in interpret-
ing the text; such stylistic devices suggest complication—even disorder—in
the type of implicatures that arise. In this way, McCarthy introduces the
notion of responsibility and, as Kreml says, ultimately the moral signifi-
cance of actions.
Chapter 4, "Marked Grammatical Structures: Communicating Intentional-
ity in The Great Gatsby and As I Lay Dying," is my own contribution to
analyses of data sets. The chapter considers different syntactic features in
each of the novels, concluding that the marked syntactic structures in each
of them require a different type of cognitive processing from the unmarked
structures in each novel. The particular type of processing contributes sub-
stantially to the specific cognitive effects produced by the structures in question.
In The Great Gatsby I demonstrate that five narrative passages which carry
"crucial" messages in the novel have more "elaborating" (versus "core")
Projections of Complementizer (CPs) than do five matched passages with less
than crucial messages. I argue that the difference in the number of elaborat-
ing CPs means that the crucial passages require more cognitive processing.
Elaborating CPs include adverbial adjuncts and sentential complements as
well as relative clauses.
In As I Lay Dying, 1 consider differences in two monologues each from eight
different characters. My overall argument is that certain syntactic features
distinguish the monologues of Darl, the apparent hero of the novel. Statistical
analysis demonstrates that Darl's monologues show significantly more de-
tached participial phrases than monologues of the other characters, as well as
certain types of prepositional phrases. I argue that because these phrases con-
tribute little to the propositional weight of the sentence (they do not contain
tensed verbs and their arguments), Darl's monologues have a dream-like qual-
ity. My claim is that such cognitive effects are a main means by which Faulkner
builds Darl's characterization.
In regard to the structures studied, the sources of stylistic effect contrast with
each other. In The Great Gatsby, the author slows down the action by adding
cognitive weight to sentences in the crucial passages; he does this by adding
additional propositions to the sentences by adding elaborating CPs. I explain
how this is different from the effect of merely adding more words. In As I Lay
Dying, the author builds up the reader's view of Darl through monologues that
are not loaded with extra propositional weight. Overall, such an analysis of
both the novels emphasizes the role of the more abstract level of predicate-
argument structure over surface-level considerations of style.
In chapter 5, "Markedness and References to Characters in Biblical Hebrew
Narrative," Timothy Wilt argues that translators of texts need to pay attention
to the markedness of certain aspects of a text to arrive at translations that are
faithful to the tenor of the original. Wilt, a Bible translator himself, discusses
the use of terms of address in biblical Hebrew narratives and how they are
translated into English. What piqued Wilt's interest in terms of address was his
Introduction 11
noticing in the book of Judges that some historical figures, especially leaders
of Israel's enemies, even after their first reference, are still referred to by their
full names and titles rather than by only their proper names or even a pronoun.
In a quantitative analysis, Wilt demonstrates that such full references for
Israel's enemies are marked compared with the references to Israel's heroes.
Elsewhere in the Bible, Wilt finds other instances of marked terms of address.
He develops an argument as to why the references to enemies, as well as to other
persons in other biblical texts, should include a relational phrase, suggesting
that the sense of their names contributes to the development of the narrative.
That is, the relational phrases foreshadow what is going to happen. For exam-
ple, when an enemy is referred to as X the crow, the implication is that such
an enemy is no more of a threat than a crow. When an enemy is referred to as
X the Sacrifice Whose Protection Was Refused, the import is that such an
outcome will befall those who stand opposed to God.
Bible translators have tended either to reproduce literally the marked terms
of address or eschew them for the sake of a fluent style. Instead, Wilt suggests
that the lesson to be learned is that (1) certain terms of address are marked and
(2) they therefore warrant special attention. He argues that translators would
be more faithful to the original if they would attempt to reproduce the inten-
tionality conveyed by these usages with equivalently marked expressions in the
target language of translation, especially when irony or humor is intended in the
original. He says, "The equivalence will be not only with regard to the seman-
tic content of expressions but also with regard to the pragmatic inferences that
the original speakers expected their audiences to make."
In chapter 6, "Literariness, Markedness, and Surprise in Poetry," Amittai
F. Aviram offers a characterization of poetry—and any literary text—as a sort
of virtual communication and therefore as marked in relation to ordinary verbal
communication. Aviram, a poet himself who also writes about literary theory and
poetry, argues that although the primary function of ordinary (or unmarked)
communication is to convey various types of meaning, the function of the literary
text is aesthetic. As Aviram puts it, "the proper response to an aesthetic object
... is contemplation, interpretation, the pleasure of discovering how the parts
come together to make a whole, and a sense of wonder at its wholeness, complete-
ness, perfection, ingenuity, and forms of surprise."
In relation to this view, Aviram disagrees with analyses of poetry which
emphasize sequential aspects of a text, such as certain specific structural ele-
ments. In this sense, he is something of a contrarian in this volume in which
most chapters emphasize the role that stylistic choices play in conveying—and
not just adding to—the overall intentionality imparted in a literary work or
conversation. Aviram claims that "stylisticians assume that what makes poetry
poetry is the use of various marked linguistic devices in order to convey
meaning...." He goes on to say, "Stylisticians tend to assume that the literary
text is a message like any other, only more punchy or expressive, using more
elaborate devices to get the message across." Instead, he argues that the way
12 OVERVIEW
poems work as aesthetic structures is to "deliver surprise when given the right
context and when read as totalities rather than sequences of elements."
To reinforce his claims as to how literariness conditions the proper ap-
proach to reading a poem, he provides explications of a number of poems,
including Emily Dickinson's "I cannot live with you," William Blake's "The
Tyger" and "A Poison Tree," and Wallace Stevens's "The Snow Man."
The final chapter (chapter 7) in this part is "Villainous Boys: On Some Marked
Exchanges in Romeo and Juliet" by Trevor Howard-Hill, a well-known editor of
Shakespeare and writer about Elizabethan drama in general. In his contribution,
he considers the extent to which the MM can be applied beyond conversation. His
specific subject is the occurrence of boy as a term of address in Romeo and Juliet;
he argues that the marked usage of boy is the basis for the revenge plot on which
the play depends. Specifically, Howard-Hill discusses three crucial dramatic
moments involving the marked use of boy together with repetition of villain.
Early in the play, Capulet, the head of one of the two feuding houses, censures
Tybalt, a hot-headed young man in the Capulet clan, by referring to him as boy.
Later, Tybalt applies this term to Romeo (from the other feuding house, that of
Montague) in challenging him to fight. Howard-Hill argues that the distribution
of boy and villain in the play is by design, not accident. Significantly, villain is
coupled with boy in the last scene of the play.
In his concluding remarks, Howard-Hill brings up the interesting question
whether dramatic characters can be seen as making their own linguistic choic-
es. He raises a number of reasons as to why dramatic speakers are not rational
actors in the usual sense. For example, he points out that their words are
governed by "such matters as the initial choice of genre, the principle of
decorum within the general theory of genre, and the specific purposes of the
playwright." He also questions what type of expectations readers of drama can
have in the absence of the effects of performance on a stage. In the end, however,
he leaves the answer open to the question about dramatic characters and their
"ability" to make linguistic choices. It is a subject deserving more thought; yet,
surely with a "suspension of disbelief" both readers and a theater audience can
view characters as rational actors and, therefore, they can form expectations
about choices that a specific character will make, given the way the character
has already been developed by the playwright up to the point of a questioned
choice. Both types of audiences also have the acquired ability to approach the
entire drama with a sense that certain types of utterances (e.g., calling a young
man boy) will be marked.
SWITCHING IN NATURALLY OCCURRING TALK

Part III of the volume consists of three chapters that analyze data from spon-
taneous discourse. In chapter 8, "Markedness and Styleswitching in Perfor-
mances by African American Drag Queens," Rusty Barrett considers how
African American drag queens switch among three styles in their performances.
Introduction 13
He labels the styles as white woman style, African American style, and gay
style. Although Barrett details defining features for each style, he also stresses
that there is no clear line dividing them. In switching among styles in their
performances, drag queens exploit the social and psychological associations of
each style by mocking the associations, in a sense. For example, the unmarked
association of the white woman style is with white, middle- or upper-class fem-
ininity; yet, here is a drag queen who is "none of the above," claiming this style
as her own. Similarly, in switching to an African American style—especially
when it is a "street" style, as is the case in many of Barrett's examples—the
performer is using a spoken style at odds with her clothes and jewelry, which
effect a stylish image of glamor. Finally, in switching to a gay style, again the
performer projects an image at variance with her dress.
Barrett argues that African American drag queens engage in switching to
make a number of social statements. First, they undermine audience assump-
tions that gender, ethnicity, class, and sexuality are categorical social catego-
ries. Second, drag queens demonstrate—by switching styles—that although
they are capable of producing language that would allow them to fit into
mainstream society (white woman style), they consciously choose to maintain
their identities as African American gay men. Third, switching also emphasizes
the social inequalities associated with white society. Fourth, these drag queens
demonstrate that they refuse to be categorized according to a single set of social
attributes.
Barrett also considers another topic in this chapter, how the types of switch-
ing which the MM delineates can be used to discuss certain asymmetries
between dialect, register, and genre as types of linguistic varieties. Going beyond
earlier treatments that largely simply define the three types of varieties, Barrett
shows how the MM might provide a principled basis for discussing the ways
in which dialects, registers, and genres relate to each other in actual use.
Chapter 9, "Styleswitching in Southern English," by Margaret Mishoe is
one of the two chapters in the volume that deal with naturally occurring speech
in ordinary conversation. Mishoe describes and explains styleswitches by
lower-socioeconomic-class white speakers living in a small community in the
North Carolina foothills.
The switches are between the two main styles in their repertoires, what
Mishoe calls home style and local standard. While local standard approxi-
mates Standard American English except for its southern phonological fea-
tures, home style has a number of distinctive morphological and syntactic
features that separate it from both the national standard dialect and local
standard. Further, a defining feature of local standard that distinguishes it from
home style is that local standard is spoken at a clearly slower rate.
Mishoe defines home style as the unmarked choice in the conversations she
studies, based on the fact that she could claim it was used in 95% of the
conversational turns. In the twelve sequences of ingroup conversation that
Mishoe studies, switches generally are marked switches from home style to
14 OVERVIEW
local standard. For example, when Christmas presents are being unwrapped,
it turns out one couple has been given a new camcorder but does not know how
to use it. In asserting his expertise and right to assume control of the
camcorder, one participant switches to local standard. In another example
the switch is from local standard to home style. In this case, the family
matriarch greets her guests for a party in local standard to establish her
identity as someone who understands her duties as "hostess." However, she
switches to home style as her guests are seating themselves in order to re-
affirm her identity as a close friend and community member. Because the
speakers are Mishoe's own relatives and their friends, the type of ingroup
conversations she was able to record resonate with a homespun humor and
verisimilitude not often found in the codeswitching literature. Yet, the
speakers' motivations for switching all fall under the "classic" motivations
that the MM attributes to speakers who make marked choices. In the main,
Mishoe's speakers are negotiating a change in the personae that they wish
to project and therefore a change in their interpersonal relationships in the
current exchange.
The final chapter in this part, chapter 10, "Marked versus Unmarked
Choices on the Auto Factory Floor," by Janice Bernsten, considers "the rules
of engagement" between assembly-line workers and supervisors. Bernsten,
who teaches in the auto factory town of Flint, Michigan, studies talk at work
in an auto factory. In her analysis, she concentrates on how both unintentional
and intentional marked choices have negative effects in an environment in
which not only tasks on the assembly line are very conventionalized.
Although Bernsten was unable to gain entrance to the factory floor herself,
she did collect quantitative data from sixty-four workers via a questionnaire
and interviews. She found that factory floor language is direct and unmitigat-
ed. In fact, she reports that workers expect direct language from supervisors
to the extent that mitigated language often receives a negative evaluation.
Several of her examples vividly illustrate this. Bernsten suggests that the lack
of politeness markers and the pervasiveness of bald imperatives used by both
management and workers may index an attempt by each group to assert and
maintain equal power.
Bernsten also comments at length about the workers' underlying knowledge
of marked and unmarked behavior in the factory. This is apparent, she claims,
in the way in which workers responded to hypothetical situations on the question-
naire. She points out that they showed much insight into the major differences in
interpretation that minor differences in syntactic and lexical choices could make.
Also, many workers wrote about possible interpretations, based on possible
differences in the supervisor's perceived assumptions and motivations. As
Bernsten writes, "It is evident that given social identity features (e.g., age and
gender) and given linguistic choices do not mechanistically determine listeners'
interpretations." Instead, perceived attitudes of the speakers play a large part
in the listeners' evaluations of possible meanings.
Introduction 15
In conclusion, Bernsten states that a practical consequence of such a study
is that the findings can be used for training employeees in the most effective
discourse patterns, particularly new managers coming from outside the fac-
tory culture.
TEACHING WHAT'S MARKED, WHAT'S NOT

The final study in the volume (chapter 11), "'Not quite right': Second-Lan-
guage Acquisition and Markedness," clearly considers consequences in mak-
ing a marked rather than an unmarked choice, even if the choices are not
always between styles. The author, Mary Sue Sroda, is an experienced com-
position instructor, who deals with the problems of the second-language learner
in mastering "appropriate" language use. As Sroda puts it, "the fundamental
issue of language use for teachers of second language (L2) learners is, 'why
do we say what when?'" The problem for these teachers is that instead of there
being one "correct" answer, there is often a range of "right" answers, some
answers that are "not quite right," and other answers that would not be accept-
able at all.
Sroda claims that the MM provides an explanatory framework for the range
of linguistic choices that second language learners can make that are "not quite
right." Sroda proposes that the examples of "odd" language can be redefined
as choices that are marked for a given rights and obligations set and, therefore,
interfere with or confound communication.
Sroda develops her argument by showing in a case study how both native
and nonnative speakers may make marked choices, but that there are differ-
ences in the types made across the two groups. (In the case study she presented
both native and non-native speakers enrolled in freshman English with a
perplexing situation: "You've locked your keys in the car. You have to call
someone to bring your spare keys. You have to ask a stranger for the money
to call—what do yousabsay?")On the basis of her results and the frameworks she
employs (both the MM and Sperber and Wilson's relevance theory), Sroda
proposes that nonnative speakers can be taught to include certain crucial
information features for given interaction types. In the keys-in-the-car scenario,
information features that must be included to satisfy the requirements of the
unmarked rights and obligations set between speaker and stranger would be
"ask for money," "no keys," and "use for money." Responses which omitted
"no keys" are marked because hearers cannot pragmatically recover the sit-
uation that has caused the speakers to ask for the quarter. The unmarked rights
and obligations set between speaker and stranger can be maintained in spite
of the imposition of the request if the speaker provides a good reason, as defined
by the community.
Sroda sees the MM as useful in her emphasis on teaching students about
crucial information features, as opposed to teaching them set phrases. She uses
two notions from the MM in this regard. First, she states that the claim of the
16 OVERVIEW
MM is not that native speakers always try to make unmarked code choices
but that they have the ability to make both marked and unmarked choices in
interactions and through these choices do "social" work. This leads to one of
Sroda's main points: Second-language learners also want to be able to make
both marked and unmarked choices. She writes, "Without the ability to have
the option of marked choices in their linguistic repertoires, L2 learners will
never acquire a native-like pragmatic ability." Second, Sroda argues that
"the ability of the MM to describe normative choices while allowing for
surface variation can be of use in explaining to L2 learners how their choices
are marked."
In a final section, Sroda discusses the range of marked choices that both
native and nonnative speakers make in their academic writing. She argues
that although there are countless discussions in the literature of composition
and rhetoric about the influence that cultural knowledge has on language use,
the MM operationalizes the interaction between language and culture in the
concept of L1-defined rights and obligation sets; that is, what nonnative
speakers need to learn is not set phrases but the way that the unmarked rights
and obligations set underlies the selection of unmarked information units and
lexical choices.
CONCLUSION
The diversity of subject matter in the chapters in this volume demonstrates—

with vigor—how the general theoretical construct of markedness can be in-
voked to explain the social and psychological effects of choosing one style of
a language rather than another. The specific usefulness of the MM in such
analyses also seems clear, even when it has not been explicitly employed by
particular authors.
In conclusion, several findings of general importance emerge. First, as is
especially evident in chapter 10 on auto factory floor talk , distinguishing the
unmarked choice from a marked choice (or an unmarked style from a marked
style) is not just a matter of "common sense." Factory discourse differs from
the discourse of the other worlds the workers inhabit. This suggests that it is not
a speaker's linguistic repertoire or global norms about politeness that deter-
mine choices but, rather, the interaction-specific cognitive calculations that
distinguish rational choice.
Second, what is implicit in all the chapters and made explicit in Kreml's
chapter on the Cormac McCarthy novel (chapter 3) is that the interaction of
styles is what gives rise to the richest implications, thereby setting up the most
implicatures. That is, although each style has import on its own, the fact
that each style is necessarily perceived in terms of a markedness relation-
ship offers us insights into that import that would not be available without
the opposition. Contrast is not everything, but it is a major constraining force
on interpretation.
Introduction 17
REFERENCES
Adendorff, Ralph D. 1993. Ethnographic evidence of the social meaning of Fanakalo
in South Africa. Journal of Pidgin and Creole Languages 8: 1-28.
Battistella, Edwin L. 1990. Markedness: the evaluative superstructure of language.
Albany: State University of New York Press.
. 1996. The logic of markedness. New York: Oxford University Press.
Blakemore, Diane. 1992. Understanding utterances. Oxford: Blackwell.
Cook, Vivian, and Mark Newson. 1996. 2nd ed. Chomsky's universal grammar, an
introduction. Oxford: Blackwell.
Damasio, Antonio. 1996. The somatic marker hypothesis and the possible functions
of the prefrontal cortex. Philosophical Transactions of the Royal Society of Lon-
don B 351: 1313-1420.
DeBose, Charles. 1992. Codeswitching: black English and standard English in the
African-American linguistic repertoire. Journal of Multilingual and Multicultural
Development 13: 157-167.
Eastman, Carol M. 1985. Establishing social identity through language use. Journal
of Language and Social Psychology 4: 1-20.
Elster, Jon. 1989. The cement of society. Cambridge: Cambridge University Press.
Gibbons, John. 1987. Code-mixing and code choice: a Hong Kong case study. Clev-
edon, Avon: Multilingual Matters.
Grice, H. P. 1975. Logic and conversation. In Syntax and semantics, vol. 3, ed. Peter
Cole and Jerry L. Morgan, 198-211. New York: Academic Press.
Jakobson, Roman. 1971. Selected writings, Vol. 2: Word and language. The Hague:
Mouton.
Li, Wei. 1994. Three Generations, two languages, one family. Clevedon, England:
Multilingual Matters Ltd.
Myers-Scotton, Carol. 1985. "What the heck, sir?" Style shifting and lexical colour-
ing as features of powerful language. In Sequence and pattern in communicative
behaviour, ed. Richard L. Street and J. N. Cappella, 103-119. London: Arnold.
. 1988a. Codeswitching as indexical of social negotiation. In Codeswitching:
anthropological and sociolinguistic perspectives, ed. Monica Heller, 151-186.
Berlin: Mouton de Gruyter.
. 1988b. Self-enhancing Codeswitching as interactional power. Language and
Communication 8: 199-211.
. 1993. Social motivations for Codeswitching: evidence from Africa. Oxford:
Oxford University Press.
Myers-Scotton, Carol [Scotton, Carol Myers], and Zhu Wanjin. 1983. Tongzhi in
Chinese: conversational consequences of language change. Language in Society
12: 477-494.
Sperber, Dan and Deirdre Wilson. 1995. 2nd ed. Relevance, communication and
cognition, Oxford: Blackwell.
Swigart, Leigh. 1992. Two codes or one? The insider's view and the description of
Codeswitching in Dakar. Journal of Multilingual and Multicultural Development
13: 83-102.
Trubetzkoy, Nikolai. 1939. Principles of phonology. Trans. Christiane A. M. Bal-
taxe. Reprint, Berkeley: University of California Press, 1969.
2
A Theoretical Introduction to
the Markedness Model
CAROL M Y E R S - S C O T T O N
I N EVERY SPEECH community, more than one way of speaking exists. That
is, no community is without at least two different speech styles, and in
many communities, more than one dialect or more than one language is
spoken, even if some people speak only one dialect of one language. In the
case of dialects and languages, these varieties typically can be associated with
different social groups; that is, not everyone in the community has command
of all the varieties in the community's linguistic repertoire, or not everyone
uses the varieties he or she knows with the same frequencies. To a lesser
extent, the same can be said for different styles or registers. Although the
same styles may be more or less in everyone's repertoire, clear degrees of
difference in an individual's access to different styles exists and certainly
differences in when and where they are used exist. (Varieties is a cover term
for selections at all linguistic levels so that choices between varieties include,
for example, choices of one language rather than another, one dialect over
another, one style or register over another, and one form of a directive or
refusal over another.) Transforming these assertions about varieties and their
associations with social groups and social contexts into quantitatively based
facts is the bread-and-butter activity of much sociolinguistic research. One
basic claim of the markedness model (MM) follows on these facts. This claim
is simply that in their own language use, individuals exploit the relationships
that become established in a community between a linguistic variety and who
uses the variety, and where and how it is used. That is, individuals take
advantage of the associations that their addressees/readers make between a
variety and its typical users or uses. (From now on, I use speaker as a cover
term for both speaker and writer and addressee as a cover term for both
addressee and reader.} Of course, at least some other researchers who
study social discourse have made the same observation. For example, Bell
(1984) argues not only that individual speakers design their conversational
18
The Markedness Model 19
contribution with their audience in mind but that the particular design chosen
is based on the speech associated with a particular social group. Also, in claim-
ing that speakers use certain politeness strategies Brown and Levinson (1987)
surely imply that speakers expect that addresees will interpret these strategies
as "polite" because the strategies are customarily associated with "being
polite" in various ways in specific speech communities (see also Tracy 1990
for other uses of such strategies). Further, speech accommodation theory,
revitalized as communication accommodation theory (CAT), relies on speak-
ers and addressees recognizing that a speaker's choice of a particular variety
is accommodation or divergence (Giles and Coupland 1991; Giles, Coup-
land, and Coupland 1991). And, of course, to recognize accommodation/
divergence, language users have to associate varieties with particular users
and "scripts" for situations.
The way in which the MM is related to the approaches cited previously is
that they all accord to language users the ability to make choices regarding
the varieties they employ, choices that necessarily involve cognitive calcula-
tions about their potential effect. In brief, these approaches are all cognitive-
ly based approaches. All these approaches would argue that choices are in-
tentional in the sense that they are made to achieve certain social ends. Fur-
ther, not only are speakers' intentions behind choices, but speakers make
choices with the expectation that addressees will recognize a choice as carry-
ing a particular intention. On this view, speakers' choices are not determined
by their social group memberships—although their individual linguistic rep-
ertoires (what varieties are available to each individual) are very much so
determined. Unfortunately, some researchers mistakenly equate limitations
on choices with limitations on repertoire (e.g., Meeuwis and Blommaert 1994
in their criticism of the MM). I discuss the relationship between repertoire
and choices further later.
The MM, as well as the other approaches cited previously, all assume that
speakers are rational in the sense that, at some level of consciousness, they are
making choices that do not simply reflect their social group memberships or
the type of speech event in which they are participating or the structure of the
event. In fact, later I argue that the MM belongs to the class of rational actor
models, and the same could be said of these other approaches. However,
although all these approaches exploit the sociopsychological associations of
linguistic varieties in explaining the choices speakers make, a second claim of
the MM that sets it apart in subtle but important ways is the issue of goals.
To put it explicitly, the question to be answered by any of these approaches
is, To what end are speakers exploiting the social (or psychological) relation-
ships of linguistic varieties in a given community? In what general sense can
their goals be characterized? Under the MM, the goal of speakers is to en-
hance rewards and minimize costs; in two words, the goal is to optimize. What
this means is that speakers choose one variety over another because of the
benefits they expect from that choice, relative to its costs. Although the other
20 OVERVIEW
approaches just mentioned can also be considered rational actor models to an

extent, they explain linguistic choices more as instances of blanket strategies
than does the MM. For example, such a strategy as accommodation or polite-
ness has a certain uniformity in the source of its explanation of choices.
Under the MM, speakers may well accommodate to the addressee's style or
may use politenesss strategies, but they also may not. Choices depend on the
strategy that would optimize for self. Thus, making choices is necessarily seen
under the MM as very customized. Often it will mean putting together com-
binations of choices; it will always mean taking account of all available evi-
dence regarding the best strategies for the specific exchange at hand and
considering the internal consistency of a set of choices. This difference is
discussed further later.
While the claim of optimization as the speaker's goal has been a major
ingredient in the MM since its original formulation (Myers-Scotton 1983), I
made it increasingly prominent in more recent presentations (e.g., Myers-
Scotton 1988a, 1988b). In the most recent publications on the MM, I began
to refer specifically to the MM as a rational actor model (Myers-Scotton
1993,1997). I develop this line here. As Elster (e.g., 1979) pointed out, such
a model does not claim that people are always rational in their choices;
however, it does claim that the premise that a choice can be explained as
rationally based does cover the phenomenon. Next, I develop further the
ways in which the MM is a rational actor model.
First I offer an up-to-date outline of the basic tenets of the MM that have
appeared in the previous discussions of the model mentioned earlier. The
most recent interpretations of the MM were in reference to codeswitching
between languages (Myers-Scotton 1988,1993); they may have misled some
readers into thinking the model applies only to such codeswitching. This
limitation is most definitely not the case; to reiterate what was said in earlier
explications, the MM applies to code choices at all levels of language.
INTERPRETATION AS MORE THAN DECODING

THE LINGUISTIC SIGNAL
Like the cooperative principle and its related maxims (Grice 1975) and rele-
vance theory (Sperber and Wilson 1995), the MM starts with the premise
that comprehension of an utterance involves much more than decoding the
linguistic signal. All three models include as a second premise that the gap
between this decoding and what is actually meant to be communicated can be
filled in by inference, a process driven by the certainty that the message
carries intentionality in addition to referentiality. Further, each of these three
models has its own "superpremise" that is a crucial part of the interpretation
of all choices. In this way, they are more "global" than an approach that
seeks to explain why particular strategies are used (e.g., politeness or accom-
modation strategies). For Grice (1975), the superpremise is that all partici-
pants proceed with the expectation that utterances are so framed that they
are "cooperative" and that this premise will aid in recovering intended mes-
sages. For Sperber and Wilson (1995), "relevance" is the key premise. Their
principle of relevance is this: "Every act of ostensive communication commu-
nicates a presumption of its own optimal relevance" (Sperber and Wilson
1995: 158). In the case of the MM, the global premise is "negotiation,"
encapsulated in my negotiation principle. The principle is:
Choose the form of your conversational contribution such that it indexes the
set of rights and obligations which you wish to be in force between speaker
and addressee for the current exchange.
Clearly, the negotiation principle is modeled in its phrasing after Grice's
cooperative principle. As such, it pays tribute to Grice's (1975) insight that
participants interpret verbal messages with preconceived notions about forc-
es that color how interpretation is to proceed. Like Grice's principle, the
negotiation principle is intended to clue in addressees that, in addition to
conveying information, the speaker has an interactional goal. However, at
this point, the resemblance between the two principles ends. As noted earli-
er, the clue that Grice's principle gives to addressees is that they should be
able to assume that speakers are being cooperative. In contrast, the negoti-
ation principle says nothing about interactants as cooperating with each
other. The only cooperation the negotiation principle necessarily implies is
that both speaker and addressee approach the interaction with the same
preconception about the purpose of the form of the speaker's contribution:
that this form can be interpreted as indexing a negotiation. This negotiation
can have one of two related concerns: Either it refers more directly to the
speaker's persona itself or more to this persona in relation to that of other
participants in the interaction. (In many ways, of course, the speaker's per-
sona can never stand on its own but only in a contrastive/complementary
relationship with that of others.)
An important aspect of all three of these models in which there is an
intentional message implied is that successful communication depends on
having all participants "involved"—at least to the extent that they recognize
that either "cooperation" or "relevance" or "negotiation" is a crucial basis
for interpretation. In this way, all instances of language use must be joint
enterprises in some sense. This view of language is not novel today, of course,
at least in many circles. For example, the thesis of Using Language is that
"Language use is really a form of joint action" (Clark 1996: 3). Further, the
view that language use involves some sort of "co-construction" is a major
premise in the many studies under the rubric of conversation analysis.
However, just because these various views share the idea of joint enter-
prise does not mean all their premises are complementary. Conversation
analysis, for example, emphasizes the contribution of the surface structural
features of a conversation to the ultimate meaning that participants derive
22 OVERVIEW
from an interaction. Sequential organization in particular is a major feature

studied. In contrast, the three models under discussion in this section (Grice's
cooperative principle model, Sperber and Wilson's relevance theory, and my
MM) would say that in addition to the referential messages in words and
their configurations, the critical messages in discourse are in mental calcula-
tions of what is being inferred, not in how surface structures are aligned.
Thus, the "joint enterprise" for these models is at the cognitive level. That is,
such models are based on the idea that speakers "know" that they can employ
utterances for more than referential messages and that they know that ad-
dressees know this also and therefore will be looking for organization at this
level. Note that although in all three of these models an intentional message
is implied, only in the MM is this "extra" message sociopsychological in
nature, not referential. This is not to say that the other two models could not
be so modified as to refer to sociopsychological inferences; however, as con-
ceputalized, they do not.
OUTLINING THE MARKEDNESS MODEL
I turn now to the components of the MM. First, the MM presupposes that as
part of their linguistic capacity—or their general cognitive capacity—all speak-
ers have what I have called a markedness metric and which I now prefer to call
a markedness evaluator. (Metric implies discrete and linear measuring; this is
not how I now realize markedness should be conceived.) The existence of a
markedness evaluator implies that the capacity to conceptualize markedness is
part of any innate competence. Conceptualizing markedness means possessing
the potential to develop two abilities: (1) the ability to recognize that linguistic
choices fall along a multidimensional continuum from more unmarked to
more marked and that their ordering will vary, depending on the specific dis-
course type, and (2) the ability to comprehend that marked choices will receive
different receptions from unmarked choices. Actually, to develop either of these
abilities requires exposure to the use of unmarked and marked choices in
actual community discourse, in the same way that a speaker requires expo-
sure to a language in use to acquire its grammatical structures. (Innateness
requires a data base to be instantiated.) From this exposure, language users
develop the general understanding of markedness in relation to alternative
linguistic varieties for a given interaction type and "know" that making
marked choices has different consequences from making unmarked choices.
I argue later how this exposure actually sets up "readings of markedness"
in the markedness evaluator in a way very similar to how Damasio (1994,
1996) claims that somatic markers are set up in the body and in the brain's
representation of the body that have to do with "survival" at a number of
levels. In referring to his somatic marker hypothesis, he states, "The hypoth-
esis rejects attempts to limit human reasoning and decision making to mech-
anisms relying, in an exclusive and unrelated manner, on either conditioning
alone or cognition alone" (Damasio 1996: 1413). As should become clearer

as this chapter proceeds, the MM's argument is in a similar vein. Speakers can
make links between the use of a linguistic variety and its effect in a certain
situation; this linking is a form of associative learning. These linkages are
only set up through experience with language in use. While the MM definite-
ly recognizes the crucial role of the speaker's actual experience with social
factors, the model argues that speakers do not make choices of linguistic
varieties on the basis of such direct linkages in a cause-and-effect fashion.
Instead, the MM builds in a device that will prune multiple options and
multiple future outcomes in some way. This device is a cognitive device, the
markedness evaluator. It takes these individual linkages as input and apprais-
es them with reference to the theoretical construct of markedness. Marked-
ness, then, is best considered as a means to compare the effects associated
with one particular linkage between the use of a linguistic variety with link-
ages associated with other varieties. Choices then are biased by the readings
of markedness for linkages contained in the markedness evaluator.
To repeat, even though the markedness evaluator has an innately based
presence (whether as a task-specific module or—more likely—as the output
of specific weightings in general neural networks), it comes with no innate
evaluations. Rather, it only comes programmed with the ability (and bias) to
make evaluations. When the speaker is exposed to language in use, this expe-
rience provides the basis for the assessments the evaluator will make. It does
this by taking the linkages that experience has shown hold between a certain
choice for a given discourse and its outcome in relation to an individual's
position in a given speech community. These markedness readings are always
relative in three ways. First, they are interaction-specific. Second, they are
dynamic, meaning they are subject to change with changing circumstances,
even within a single interaction; that is, what was a marked choice—in regard
to the relationship between participants at the outset of an interaction—can
become an unmarked choice in the process of negotiation. Finally, they al-
ways involve multidimensional ordering, meaning choices are more or less
marked, not categorically marked.
Note that choices are seen as more or less unmarked for a particular
rights and obligations set (RO set). "Rights and obligations" is a theoretical
construct for referring to what participants can expect in any given interac-
tion type in their community. For example, a person who calls a business
establishment on the phone and asks to speak to a particular employee can
expect that the receptionist will either connect her to that employee or take
a message if the employee is not available. The term rights and obligations
does not generally refer to actual "rights" in any legal or moral sense; for
example, a customer standing at the cash register in a drugstore with an item
he wishes to buy in plain view cannot a priori sue the idle salesperson at the
cash register who ignores him for more than a minute or so. (However, of
course, the salesperson's boss can fire him for neglect of duty.) In reference
24 OVERVIEW
to some types of behavior, rights and obligations is another term for norms,
codes of behavior that are established and maintained by the social group.
Norms are discussed further later.
THE PROBLEM OF RO SETS
The unmarked RO set for a given interaction type (or genre, if literary works
are the subject) is derived from whatever situational features are salient for
the community for that interaction type. As I noted in Social Motivations
(1993), I do not attempt to specify what those situational features are. I avoid
doing so for three reasons: First, the same situational features are not relevant
in all interaction types (or genres) or across all communities. For example, in
one type, socioeconomic status or occupation may be relevant, but in another
it may not. When I made an initial visit to a law firm recently, I was careful
to "dress as a professional"; the attorney I was to see already knew my
occupation, but I thought I might get better treatment from the receptionist
if she could infer my status. However, men who are emptying my garbage
container in front of my house while I write this do not care whether I am a
professional person or a blue-collar worker; it will not change how "well"
they empty my garbage cart. Second, the content of a situational feature may
change. For example, what qualifies a person to be considered a member of
ethnic group x may change because the defining features of that ethnic group
may change. Third, the hierarchical relation of one feature to another fea-
ture in an interaction type may change. This change can happen within the
interaction itself when the topic changes. For example, the relative expertise
of group members regarding how a particular technological device, such as
a camcorder, works may be relatively unimportant until using the device
becomes a goal in a group (Mishoe chapter 9, this volume). Hierarchial
relations of features can change over time as well. For example, participants
may revise their views of how much certain technological skills are to be
valued, or gender or ethnic group may take on increased salience over other
features as national political debate centers on these features. At the same
time, it is true that one can predict that some factors will figure in the
establishment of the unmarked RO set in many interaction types. These
factors include the main social identity features of participants (e.g., age,
sex, ethnic group, socioeconomic status, and occupation) as well as topic/
goal and setting.
Another reason I avoided attempting to come up with a taxonomy of the
features necessary to characterize the relations of participants in an un-
marked RO set is that such attempts very quickly run into problems because
a binary opposition along the lines of relevant versus irrelevant is hard to
apply. What should count as "relevant"? Some features seemingly must be
specified, but what about others which are desirable? And what about "ir-
relevant"? Is there a difference between calling a feature overspecified and
calling it out of place? In addition, even to classify two participants in rela-

tion to a "must specify" criterion is difficult. For example, consider the
relative statuses of the speaker versus addressee in Sroda's keys-in-the-car
situation (chapter 11, this volume) in which the speaker who has locked his
or her keys in the car asks a stranger for a quarter to call a friend. From one
point of view, the stranger has higher status in this interaction and deserves
deference, because the speaker is asking a favor. However, from another
point of view, to think that he or she has the "right" to ask a stranger for a
quarter, the speaker almost necessarily must consider that some "egalitarian
norm" prevails, making his or her status as far as the interaction type is
concerned (if not this particular interaction) equal to that of the stranger.
Thus, it is difficult to specify exhaustively and discretely the situational
features that go into determining the unmarked RO set for a given interac-
tion type. Further, to assume that such a specification, such a taxonomy, has
any explanatory value on its own is a mistake. I return to this point later in
a discussion of how truly explanatory comprehensive models of language
use are valuable, even in their "infancy."
In sum, utilizing the input of their experiences in their community and the
cognitive equipment in their markedness evaluator, speakers (and address-
ees) are able to arrive at readings of markedness of two types. First, they can
establish the outlines of the unmarked RO set for a particular interaction type.
To do this, they take into account whatever situational features are salient,
given a specific community and interaction type. These features may well
include participants, topic, and setting. Second, they calculate the relative marked-
ness of code choices (i.e., linguistic varieties) to index the unmarked RO set.
THE MARKEDNESS MODEL EXPLAINED
In the following sections, I present the MM as a rational actor (RA) model.

As such, the MM is integrated into a more comprehensive view of how social
behaviors arise. First, I give the maxims of the MM from the Social Motiva-
tions version and examples of unmarked and marked choices that show stylistic
variation. Next, I outline how RA models work in general and clarify how the
MM can be considered a type of such a model. In relation to this, I explore
research suggesting how situational and cognitive features of the MM can be
integrated into a coherent explanatory and predictive model of language use as
an instance of general participant behaviors. I also outline the steps involved in
linguistic choices according to the MM as an RA model. Finally, I offer some
reasons to attempt to formulate such a model to explain language use.
Maxims under the Markedness Model

Given specific readings of the markedness evaluator, speakers make code
choices with the following five maxims in mind. Although these maxims
were originally formulated to refer to naturally occurring conversations,
26 OVERVIEW
with a little modification, writers can apply at least the unmarked-choice

and the marked-choice maxims to literary texts, and the characters in liter-
ature can apply them all.
The Unmarked Choice Maxim: Make your code choice the unmarked
index of the unmarked rights and obligations set in talk exchanges
when you wish to establish or affirm that rights and obligations set.
The Marked Choice Maxim: Make a marked choice which is not the
unmarked index of the unmarked rights and obligations set in an inter-
action when you wish to establish a new rights and obligations set as
unmarked for the current exchange.
The Exploratory Choice Maxim: When an unmarked choice is not
clear, use switching between speech varieties to make alternate ex-
ploratory choices as (alternate) candidates for the unmarked choice
and thereby as an index of a rights and obligations set which you
favor.
Deference Maxim: Switch to a code which expresses deference to others
when special respect is called for by the circumstances.
Virtuosity Maxim: Switch to whatever code is necessary in order to
carry on the conversation/accommodate the participation of all speak-
ers present.
How Speakers Apply the Maxims

What motivates speakers to follow any one of these maxims? If speakers
follow either the unmarked-choice maxim or the marked-choice maxim, it is
clear that they are negotiating the RO set they see as beneficial to them in
some way. To follow either the deference maxim or the virtuosity maxim has
fewer direct advantages for the speaker; however, there certainly are indirect
benefits. Presumably, speakers follow the deference maxim when they "want
something" from the addressee; that is, speakers choose deference because
they expect it will have a payoff for them, even if it is only avoided costs.
Following the virtuosity maxim allows speakers to present themselves as
enablers in that they make it possible for a conversation to take place; in this
way, they put themselves in a good light. Also, by utilizing their ability to
switch from one code to another they show off their linguistic repertoire and
therefore show themselves as multidimensional individuals.
Note that the model assumes that different individuals have similar no-
tions of what choices are unmarked for given situations. How can this be so,
since any community includes people from different social backgrounds? The
key to understanding that a particular linguistic choice is unmarked is not to
say that it is the one that gives the most members of the community the best
outcome; rather it is only the one most expected, and what makes it expected
is that it meets the beliefs and desires of those persons in the community who
have sufficient power to set norms. The fact is simply that opportunities to
designate markedness are not equally distributed across a community. This
inequity is also recognized in one of the propositions of an extended version
of communication accommodation theory: "In formal and status-stressing
situations, many speakers are likely to converge to the sociolinguistic mark-
ers and behavior of the dominant group" (Gallois, Giles, Jones, Cargile, and
Ota 1995: 142). Note that the model does not operationalize markedness in
its maxims. Markedness is a theoretical construct; such constructs are de-
fined by their properties, but their existence is not subject to direct falsifiabil-
ity. For example, as it is used in the MM, the major property of markedness
is that different linguistic choices link different social meanings to specific
RO sets in specific interaction types or genres. Theoretical constructs differ
from the hypotheses derived from them because these prediction-making
hypotheses must be so stated so that it is clear what type of evidence would
falsify them. The relation of the hypotheses to the construct is that if the
hypotheses are supported, this attests to the construct's value as a heuristic
that abstracts information from data sets. Information so framed has an
economy; it is an organized system, not just a description. The main argu-
ment of the MM is that the unmarked versus marked opposition is just such
a heuristic.
A hypothesis identifying the unmarked choice can be derived from the
construct of markedness as applied to code choice. This is the frequency
hypothesis:
One linguistic variety, structural type, or discourse pattern occurs with more
frequency than other possible varieties or structures as the most unmarked
index of a specific RO set in a specific interaction type.
How are the terms of this hypothesis operationalized? Languages, dialects,
and styles operationalize "linguistic variety." Structural types consist of lex-
ical, morphosyntactic and phonological patterns (e.g., what information units
comprise an "ask a stranger for a quarter" scenario, as discussed in chapter
11). Note that except for linguistic variety, the unmarked choice is not a
single entity, but rather a set of entities with like features (e.g., bald impera-
tives, some with tags, some not). And, of course, linguistic varieties them-
selves are not really discrete entities, either.
Speakers concur on the unmarked choice because, thanks to the marked-
ness evaluator, they "know" (consciously, but more often unconsciously)
that the unmarked choice has occurred with more frequency than other choices
in like circumstances in their community. They also have tacit knowledge
that other community members share this recognition.
That unmarked choices emerge when data sets are examined quantitative-
ly is supported empirically (e.g., Scotton and Bernsten 1988; Blum-Kulka,
House, and Kasper 1989). Admittedly, though, I would like to see more
studies that test the hypothesis that frequency in outcome types positively
28 OVERVIEW
correlates with the unmarked choice. Other hypotheses also can be derived
from the construct of markedness, specifically ones about the types of speak-
ers and situations (including social and intellectual climates) that favor un-
marked versus marked choices. However, such a discussion goes beyond the
goals of this chapter.
Earlier, I implied that the unmarked choice does not "treat" everyone
"equally." Why, then, should most speakers make the unmarked choice?
This point is addressed in some detail later when I discuss what it means to
behave rationally. Here, suffice it to discuss why people "build" norms and
behave in terms of them. One of the major proponents of RA models, Elster
(1989) characterizes social norms as the "cement of society." He writes that
"social norms coordinate expectations" (Elster 1989: 97). Because norms,
by their very nature, depend on an often unspoken consensus, norms neces-
sarily favor unmarked choices, which themselves depend on a consensus
(i.e., the fact they occur more frequently than marked choices implies a
consensus). Since speakers know that most group members (not all!) per-
ceive social norms as legitimate, they tend to see unmarked choices in the
same light. Even if they do not find norms to be legitimate, speakers gener-
ally find them to be compelling because of the potential costs resulting from
violating them. Again, this expectation favors unmarked choices. Persons
can use the "compelling" component of norms to their own advantage through
their own choices. And, whether they do this by making an unmarked or a
marked choice, their choice is a negotiation to make it a norm and thereby
to bind other participants.
Thus, although each person is free to make choices on his or her own,
what unites individual choices is that they are made against the same back-
drop of what is unmarked or marked. This tacit knowledge not only pro-
motes unmarked choices, as I have just argued, but also accounts for the
ability of participants (who have been socialized in the community in ques-
tion) to interpret marked choices; that is, at the very least, they know that a
marked choice is not an unmarked choice.
Examples of unmarked choices should be easy to find, and they are. In a
very clear example, Ervin-Tripp (1976: 123) gives what are contrasting un-
marked choices for asking a physician to hand an associate a chart. From a
nurse, the unmarked choice would be an embedded imperative (modal + you
interrogative). In Ervin-Tripp's version, the nurse includes a mitigating open-
er and closer: Ob, by the way, doctor, could you leave me that chart when
you're through? However, for one physician to another, the unmarked choice
would be a bald imperative. In Ervin-Tripp's example, an informal attention-
getter precedes the bald imperative and a tag question follows it: Hey, Len,
shoot the chart to me, willya?
To illustrate the dynamic nature of choices and the use of a marked choice
to index a new RO set, consider the following. At a university where I used
to teach, a male professor whom I knew socially was named provost. One day
soon after his appointment I called him on departmental business. As a con-

versational opener, I identified myself and said, How are you? Rather than
respond with the unmarked choice in such an interaction between acquain-
tances, if not close friends (Fine, how are you?), he said, What can I do for
you? This choice, marked for conversational openers between friends and
even only social acquaintances, of course, was his negotiation for a new RO set
in which we were not to interact as friends but as provost and faculty member.
The Markedness Model as a Rational Actor Model

In this section, I show the ways in which the MM can be considered an RA
model. In doing so, I develop the argument that linguistic code choices, far
from being best explained as chance behavior, or linear responses to sequenc-
ing rules in conversation, are best explained as instances of the speaker acting
rationally. To act rationally simply means that a choice reflects a goal to
enhance rewards and minimize costs—as I said at the outset, to optimize.
Indirectly, this section is an argument to support the claims of the authors of
the various chapters about the goals of the various speakers and authors who
are their subjects. In a final section, I discuss the merits of explanation and
prediction inherent in this approach.
RA theorists come from a variety of disciplines, including economics and
sociology. My treatment of RA models relies heavily on discussions in a
number of places by Jon Elster, a philosopher. I adapt many of his ideas and
those of other RA researchers as a means of explaining the choice of one
linguistic variety over others. The general claim of such models can be stated
very simply: Choices in specific interactions are best explained as cognitively
based calculations that depend on the actor's (speaker's) estimation of what
actions offer him or her the greatest utility. Note that RA models are often in
opposition to equity theory which Clark (1996) promotes as a major influ-
encing factor in linguistic choices. Clark's discussion results in defining equi-
ty theory in a number of different, not necessarily complementary, ways. I
can certainly agree with one view he presents and do not see it contradicting
RA models: "Equity theory rests on the assumption that people try to avoid
the distress they feel in inequitable situations" (Clark 1996: 292). He goes on
to relate ideas about "protecting face" from Goffman (1967) and Brown and
Levinson (1987) to equity theory. Again, I can see how reciprocity in protect-
ing face does not contradict the premises of RA models that I am supporting
here. However, I cannot say the same for "the equity principle" he proposes:
In proposing a joint project, speakers are expected to pesuppose a method for
maintaining equity with their addressees. (Clark 1996: 295)
The possibility that speakers make marked choices which are negotiations
against the existing status quo is a major point in the MM. Whether this
possibility is compatible with the equity principle all depends on what is meant
30 OVERVIEW
by "maintaining equity." Clark never makes that clear. If "maintaining

equity" means "achieving a mutually acceptable outcome," then the MM
(and presumably other RA models) can accept "equity." Certainly, a point
that needs to be stressed about the MM is that speakers' choices-—and most
especially marked choices—are only negotiations. For them to "go through,"
they must be accepted as an acceptable medium to all participating speakers.
(Parenthetically, let me say that Clark's citation of David Lewis's insight
[1969] "that language use is really people solving coordination problems"
[Clark 1996: 62] is something I can endorse.) However, just because all
participants accept a linguistic variety as an acceptable medium (and there-
fore accept the RO set for which it is unmarked) does not mean that they
necessarily "avoid the distress they feel in inequitable situations" or that they
"maintain equity with their addressees." Many times, some participants must
endure inequity; it is the nature of human interactions, like it or not. Of
course, Clark is correct in pointing out that there are a number of linguistic
choices that can be bids to restore or maintain equity, such as deference
strategies and avoidance rituals. However, the speaker in the more powerful
position is not obliged to use them.
Above all, RA models are mainly concerned with explaining why actors
(speakers) make the choices they do. And they explain choices as the
means by which actors try to achieve their ends as well as possible. As
such, RA models focus less on choices in relation to the effect of their
choices on others than some other approaches that I have indicated also are
cognitively based. I have in mind Bell's (1984) model of audience design, as
well as Brown and Levinson's (1987) politeness model or the CAT model
of Giles and his associates (1991). CAT also presents speakers as, sometimes
at least, making choices based more on "social comparisons," the desirabil-
ity of presenting oneself as a member of one group versus another (e.g.,
Giles, Coupland, and Coupland 1991: 27; Giles and Wiemann 1987). This
difference in focus need not be taken as a shortcoming of any of the ap-
proaches, but it is an important difference. While speakers as rational actors
include an evaluation of the effects of choices on others, the actor is central.
Speakers seek to optimize their own outcomes, not those of their addressees.
A second point favors RA models over other models of language choice:
RA models provide a mechanism that activates behavior. Rationality is such
a mechanism, Elster argues. This is so because RA models provide a means of
telling us how things happen on two levels. First, in its emphasis on cost-
benefit analysis, an RA model necessarily explains choices as goal directed. In
this way it tells us why choices are made. Second, rationality is also the
mechanism that answers the question of how choices are made. According to
Elster, to behave rationally includes three optimizing operations: (1) consid-
ering given desires and values and prior beliefs, to find the best action; (2)
making sure that these desires, values, and beliefs are internally consistent;
and (3) making sure these desires, values, and beliefs are based on available
evidence. Note that to be considered rational, to simply hold intentions based

on desires, values, and beliefs is not enough; the tests of internal consistency
and available evidence are crucial.
SOCIAL NORMS AND RATIONAL CHOICES
Previously, I introduced the role of social norms in code choice. I implied that
one reason unmarked choices develop is that they themselves are norms.
However, I now argue against the view that social norms are the major
mechanism in choosing among possible varieties, that choices are made merely
to support norms. It is true that norms designate marked and unmarked
choices, but speakers make choices.
Yet, norms certainly do influence speakers. Recognizing this tells us why
unmarked choices are the most frequent. First, to be norms, courses of action
or views have to be shared by others and they are partly sustained by the
approval/disapproval of others. Even when violating norms, nonsociopathic
individuals also "preserve" norms by their own feelings of embarrassment,
guilt, or shame. Thus, a major reason to make the unmarked choice (i.e.,
within the terms of the MM, to follow a norm in one's choice of linguistic
variety) is to avoid group disapproval and personal distress, as noted in the
discussion of Clark (1996) earlier.
Risk aversion is a second reason for making the unmarked choice, even if
making the unmarked choice (i.e., following norms) does not give the speaker
what he or she might consider to be the "best" outcome. That is, a second
reason for not making a marked choice which might optimize the outcome,
which would negotiate a desired RO set, is that the speaker's subjective and
objective evaluations of the probability of success tell the speaker that chanc-
es of success are low. That is, making a marked choice might not be following
the stricture of rationality to base decisions on available evidence. Further, it
is often simply safer to make the unmarked choice. Many economists (e.g.,
Kagel 1995) demonstrate that people are simply afraid of taking risks (i.e.,
risk aversion is stronger than the profit motive).
Third, another reason to opt for the unmarked choice is that making and
negotiating the desired consequences of marked choices may require more
effort. Selection in general has a moving target. To make marked choices that
actually achieve their goals, one must recognize the "strategic" nature of the
environment and assess successfully the intricacies of the interrelation of
salient factors in a specific interaction and the possibility of change in their
relative saliency. That is, the environment itself can change because it in-
cludes other actors who are similar to the speaker and therefore who will be
trying to maximize their gains, too. (In chapter 4,1 argue that processing a
kind of marked choice in a novel requires more effort from the reader [ad-
dressee] than processing unmarked choices.)
Yet, at times, speakers do make marked choices. These are choices that do
32 OVERVIEW
not index the unmarked RO set; therefore, in making such choices, the speaker's
aim is to establish a new RO set as unmarked for the current exchange.
Marked choices derive their intentional meaning from two sources. First,
because a marked choice is not the unmarked choice for the unmarked RO set
for the current exchange, its choice is a negotiation against that unmarked
RO set. Second, it is a call for another RO set in its place. The choice that is
marked under the current RO set is, in fact, unmarked in the new RO set
which the speaker is negotiating.
THE MARKEDNESS MODEL AND THE

EVOLUTION OF SURVIVAL MECHANISMS
Note that both Elster's ideas about RA theory and mine about the MM rest
on the premise that speakers are innately disposed to make decisions that
they judge will result in the best outcome. In effect, such a premise assumes
that actors are innately disposed toward optimizing survival. This premise
receives support from empirical research which, at first glance, is very distant
from a cost-benefit analysis of social behaviors and specifically of linguistic
choices. This area of research is the neurobiological study of reactions to a
variety of stimuli, with the organism's survival as a goal. The research of
Damasio and his associates introduced earlier deals with possible functions
of the prefrontal cortices. Damasio argues that "marker signs" influence the
types of responses which an organism will have to stimuli. The connection
with human responses to social situations, which Damasio himself suggests,
is made in the following way. First, Damasio lays the groundwork by giving
evidence that organisms have "preorganized mechanisms" to handle basic
biological activities that contribute to survival. Organisms then develop pref-
erences for assessing new events/things as "good" or "bad." He calls these
preferences somatic markers:
The marker signals arise in bioregulatory processes, including those which ex-
press themselves in emotions and feeling, but are not necessarily confined to those
alone. This is the reason why the markers are termed somatic: they relate to body-
state structure and regulation even when they do not arise in the body proper but
rather in the brain's representation of the body. (Damasio 1996: 1413)
Damasio goes on to argue that somatic markers help limit "the space" nec-
essary for decision making and also allow the organism to call on previous
experience and make decisions quickly. Although such a system would have
evolved to maximize basic survival,
a very large range of other problems including those which pertain to the social
realm, are indirectly linked to precisely the same framework of survival versus
danger, of advantage versus disadvantage, of gain and balance versus loss and
disequilibrium. It is plausible that a system geared to produce markers and
signposts to guide basic survival, would have been pre-adapted to assist with
"intellectual" decision making. (Damasio 1996: 1417)
In the same article, Damasio presents empirical evidence that patients with
damage to the ventrodemial sector of the prefrontal cortex do not do as well
as normal control subjects in a card game in which they must adopt a
certain strategy to end up winning and not losing. Further, the experi-
ment demonstrated that superior memory and IQ in control (i.e., normal)
subjects did not seem to be the source of the ability to adopt the winning
strategy; rather, subjects relied on their ability "to sense," overtly or not, the
best strategy.
As I have indicated earlier, Damasio's somatic markers have certain sim-
ilarities with the MM's markedness evaluator. Based on input from experi-
ence, the evaluator enables speakers "to sense" the degree to which alterna-
tive linguistic choices are unmarked or marked for a given interaction type.
In this way, the markedness evaluator is like somatic markers in that it
enables speakers to take a shortcut in the decision-making process. Howev-
er, the two types of preferences the "devices" offer are somewhat different.
Although somatic markers actually direct an actor toward a specific judg-
ment for a scenario, in contrast, the markedness evaluator offers only infor-
mation as to degree of markedness of a potential choice for that scenario vis-
a-vis others—it does not offer the type of "good" or "bad" judgment of the
somatic marker. However, it is true that in giving a reading of "more un-
marked" versus "more marked" the evaluator, in effect, gives the individual
the equivalent of the green light for "go" (for unmarked choices) or a yellow
light for "caution" (for marked choices).
MODIFYING ELSTER'S RA MODEL FOR THE

MARKEDNESS MODEL
In this sketch of the outcome process, I redraft Elster's two-filter RA model.
In one of his books, Elster describes any given piece of human behavior as the
end product of two successive filtering devices; I add a filter and also add
other constraints to the first filter.
The first filter is the set of what Elster refers to as structural constraints.
These are defined as "all the physical, economic, legal and psychological
constraints that an individual faces" (Elster 1989: 14). In turn, these con-
straints produce what Elster calls an opportunity set. In the case of linguistic
choices, I say that structural constraints are those factors generally sub-
sumed under the term social context. Thus, the most important structur-
al constraints for this analysis are the participants' social identity features
(age, sex, socioeconomic status, etc.) as well as the characteristics of the
discourse situation (setting, topic, etc.). In linguistic terms, an actor's oppor-
tunity set, which is a product of relevant structural constraints, is his or her
linguistic repertoire.
It is true, of course, that larger societal factors affect the content and
salience of an actor's individual structural constraints. For example, in some
34 OVERVIEW
societies, if one is a woman, the chances of higher education are less great
than those for men. Similarly, family connections or membership in a certain
ethnic group may largely influence access to certain socioeconomic statuses.
However, neither the MM specifically nor RA models in general have as their
goal to explain how large-scale societal forces limit a person's access to par-
ticipation in certain interaction types by limiting their linguistic repertoire.
Rather, their focus is on how individuals make choices for themselves, given
their available repertoires.
An important distinction exists between this model and "social context
models" in the Labovian tradition. In those models, the social context factors
(i.e., structural constraints) are correlated with linguistic choices (with the
implication that they determine choices in some sense, at least). Under RA
models, including the MM, what social context factors do is determine not
choices but only the speaker's opportunity set, that is, those varieties the
speaker is able to use. The distinction is very important.
In addition, I add to the first filter of structural constraints those structur-
al features that organize discourse, particularly those having to do with
sequential organization. As I noted already, many practitioners of conversa-
tion analysis see such facts as whether an utterance is a first-part or second-
part member of a specific adjacency pair as limiting the possibilities—whether
in content or form—as to what can fill that slot. I accept that such structural
constraints are features affecting the opportunity set for a specific interac-
tion type; however, they do not determine the actual choice from that set
which the speaker makes.
All the constraints that this first filter imposes—whether the speaker's
social identity features or situational variables or surface linguistic structural
features of the interaction type or the specific conversational turn—are exter-
nal. That is, speakers do not control them, at least directly.
At this point, I add, as a second filter, those features that are internal to the
speaker but still not under conscious control. These are Damasio's somatic
markers and the innately "available" markedness evaluator. Like the con-
straints of the first filter, speakers have no direct control over how these
elements constrain the choice process. Within the terms of this filter model,
what these elements do is further bias the selection of alternatives from the
initial, structurally determined opportunity set, this time in terms of "suc-
cesses" or "failures" based on the actor's previous factual experience,
facts previously categorized in an unconscious cost-benefit analysis.
The third filter includes the mechanisms responsible for which alternative
from the opportunity set is actually chosen. The main element in this filter, of
course, is rationality. Earlier, I outlined how rationality operates as a mech-
anism: From the speaker's set of possible choices, rationality indicates the
"best" choice, given the degree of internal consistency of his or her desires,
values, and beliefs and the available evidence. Whereas desires, values, and
beliefs are reasons for a choice, actually to act rationally means that the
speaker does not just choose the most favored choice but rather the most
feasible choice. Note that this does not mean that the actor always has self-
interest in mind. Acting rationally simply means making the optimal choice,
given any value system the actor has.
In acting rationally, the speaker does take account of social norms as a
component of "the available evidence." In contrast with rationality, which is
forward-looking ("I do X because I hope to achieve Y"), norms are back-
ward-looking, simply having the form, "consensus has favored doing X; there-
fore do X." Earlier, I outlined reasons why following norms, and therefore
making the unmarked choice, is an attractive alternative for many actors.
CONCLUSION
In this final section, I characterize RA models (and, necessarily, the MM) in
a theoretic sense, contrasting them in brief with other approaches to linguis-
tic code choices. Elster repeatedly refers to rationality as a normative theory
in the sense that it tells people, "if you want X, then do Y." As such, RA
models are not necessarily predictive, although they are explanatory: They
explain what people would do if they were rational and explain the choices
that people actually do make, based on the assumption that they are acting
rationally. It is in this sense that rationality is a mechanism to explain social
behavior. RA models are not necessarily explanatory theories that make pre-
dictions about what people really do, and the MM does not predict in any
global sense when speakers will make marked versus unmarked choices (but
see Myers-Scotton, chapter 4, this volume, on Fitzgerald and marked choices
for "crucial passages"). However, the MM is generally predictive in the sense
that it predicts how choices will be interpreted.
In contrast to RA models, models of linguistic choices that hold that
choices depend primarily on the social context do not provide a mechanism for
evaluating and making choices. Such models typically state that choices are
determined by the social identities of speakers and situational factors such as
topic, setting, genre, and the like. Such models ignore or, at the very least,
downgrade the existence of rationally based choice. The problem is that nei-
ther the situational features nor the participants' identity features of such mod-
els can be considered mechanisms of choice. It is true that in such models pre-
dictions can be correlated with observable outcomes; in fact, in this sense they
can even identify the unmarked choice. A general example of such a "predic-
tion" is this: "If situational features X are present, then choice Y is unmarked."
Another model of linguistic choices is contained in the various approaches
to discourse under the rubric, conversation analysis (CA). Like social context
models, CA does account for the data descriptively, although it takes quite a
different approach to context. In fact, a main goal for CA is "to develop an
empirical analysis of the nature of context" (Drew and Heritage 1992: 17).
They go on to say that " 'context' and identity have to be treated as inherently
36 OVERVIEW
locally produced, incrementally developed, and, by extension, transformable

at any moment" (Drew and Heritage 1992: 19).
In a sense, structural organization in CA is similar to social situational
factors in the social context models because in most CA approaches, the
emerging conversational structure, especially sequences, "organizes" choic-
es. However, in the case of both social context and CA models, formulating
a prediction that identifies a choice (i.e., social context models) or describes
what choices are made in a specific context (i.e., CA models) is not the same
as explaining why a choice obtains or, even more interestingly, why a choice
is not always observed. That is, such "predictions" are too restrictive. Most
importantly, they do not allow for—or account for—marked choices. Nor
do such models provide a way for interpreting marked choices. As will be-
come apparent, the "meat" of many of the chapters in this volume are, in
fact, marked choices.
In sum, RA models, including the MM, offer a far-reaching advantage
over other current models of linguistic choice. From the outset, "being ratio-
nal" constrains choices in an important way: Every choice in a speaker's
repertoire (opportunity set) does not have an equal chance of occurring.
Instead, the goal to enhance rewards and minimize costs limits choices in a
way that neither situational factors nor structural organization can do: The
operative word regarding choices is not possible but feasible. This narrowing
of choices to those that are feasible or advantageous assumes that such choic-
es are the result of conscious or unconscious cognitive calculations.
Finally, and perhaps most important, as noted early in this chapter, al-
though RA models such as the MM do not claim that people are always
rational, they do claim that the assumption of rationality accounts for the
data. That is, not only do RA models provide an explanation of why every
potential choice does not occur with the same frequency, but they also pro-
vide a principled means for interpreting the choices that do occur.
REFERENCES
Bell, Alan. 1984. Language style as audience design. Language in Society 13: 145-204.
Blum-Kulka, Shoshana, Juliane House, and Gabrielle Kasper (eds.). 1989. Cross-
cultural pragmatics: requests and apologies. Norwood, N.J.: Albex.
Brown, Penelope, and Stephen Levinson. 1987. Politeness, some universals of lan-
guage usage. Cambridge: Cambridge University Press.
Clark, Herbert H. 1996. Using language. Cambridge: Cambridge University Press.
Damasio, Antonio. 1994. Descartes's error: emotion, reason, and the human brain.
New York: Grosset/Putnam.
. 1996. The somatic marker hypothesis and the possible functions of the
prefrontal cortex. Philosophical Transactions of the Royal Society. London B
351. 1313-1420.
Drew, Paul, and John Heritage. 1992. Analyzing talk at work: an introduction. In
Talk at work, ed. Paul Drew and John Heritage, 3-65. Cambridge: Cambridge
University Press.
Elster, Jon. 1979. Ulysses and the sirens. Cambridge: Cambridge University Press.
. 1989. The cement of society. Cambridge: Cambridge University Press.
. 1991. Rationality and social norms. Archives of European Sociology 32:
109-29.
Ervin-Tripp, Susan. 1976. Speech acts and social learning. In Meaning in anthropol-
ogy, ed. Keith Basso and Henry Selby, 123-153. Albuquerque: University of New
Mexico Press.
Gallois, Cynthia, Howard Giles, Elizabeth Jones, Aaron C. Cargile, and Hiroshi Ota.
1995. Accommodating intercultural encounters, elaborations and extensions. In
Intercultural communication theory, ed. Richard L. Wiseman, 115-147. Thou-
sand Oaks, Calif.: Sage Publications.
Giles, Howard, and Nikolas Coupland. 1991. Language: contexts and consequences.
Pacific Grove, Calif.: Brooks Cole.
Giles, Howard, Nikolas Coupland, and Justine Coupland. 1991. Accommodation
theory: communication, context, and consequences. In Contexts of accommoda-
tion: developments in applied linguistics, ed. Howard Giles, Nikolas Coupland,
and Justine Coupland, 1-68. Cambridge: Cambridge University Press.
Giles, Howard, and John M. Wiemann. 1987. Language, social comparison and pow-
er. In Handbook of communication science, ed. Charles R. Berger and Steven H.
Chaffee, 350-386. Newbury Park, Calif.: Sage.
Goffman, Erving. 1967. The presentation of self in everyday life. Harmondsworth:
Penguin.
Grice, H. Paul. 1975. Logic and conversation. In Syntax and semantics, vol. 3, ed.
Peter Cole and Jerry L. Morgan, 41-58. New York: Academic Press.
Kagel, John H. 1995. Auctions: a survey of experimental research. The handbook of
experimental economics, ed. John H. Kagel and Alvin Roth, 501-559. Princeton,
N.J.: Princeton University Press.
Lewis, David K. 1969. Convention: a philosophical study. Cambridge, Mass.: Har-
vard University Press.
Meeuwis, Michael, and Jan Blommaert. 1994. The "Markdness Model" and the
absence of society: remarks on codeswitching. Multilingua 13: 387-423.
Myers-Scotton, Carol M. 1983. The negotiation of identities in conversation: a theory
of markedness and code choice. International Journal of the Sociology of Lan-
guage 44: 115-136.
. 1988a. Codeswitching as indexical of social negotiation. In Codeswitching:
anthropological and sociolinguistic perspectives, ed. Monica Heller, 151-186. Berlin:
Mouton de Gruyter.
. 1988b. Self-enhancing codeswitching as interactional power. Language
Communication 8: 199-211.
. 1993. Social motivations for codeswitching: evidence from Africa. Oxford:
. 1995. What do speakers want? Codeswitching as evidence of intentionality in
linguistic choices. SALSA II: proceedings of the second annual symposium on
language and society-Austin, ed. Pamela Silberman and Jonathan Loftin, 1-17.
Austin: University of Texas Department of Linguistics.
. 1997. Rational actor models and social discourse analysis. In Discourse anal-
38 OVERVIEW
ysis: proceedings of the first international conference on discourse analysis, ed.

Emilia Ribeiro Pedro, 177-199. Lisbon: Edicoes Colibri/Associacao Portuguesa de
Linguistica.
Myers-Scotton, Carol M. [Scotton, Carol Myers], and Janice Bernsten. 1988. Natural
conversations as a model for textbook dialogue. Applied Linguistics 9: 372-384.
Sperber, Dan, and Deirdre Wilson. 1995. Relevance, communication and cognition.
2nd ed. Oxford: Blackwell.
Tracy, Karen. 1990. The many faces of facework. In Handbook of language and social
psychology, ed. Howard Giles and W. Peter Robinson, 209-226. Chichester, En-
gland: John Wiley.
II
Stylistic Choices in Literature
3
Implicatures of Styleswitching in the Narrative

Voice of Cormac McCarthy's
All the Pretty Horses
NANCY KREML
C ORMAC MCCARTHY'S NOVELS pose many problems for the reader: Issues
of history, philosophy, religion, abnormal psychology, and mythology
are intertwined with problems of style and structure. Adding especially to the
difficulty of reading All the Pretty Horses (1992) is the problematic narrative
voice. The narrative voice is the reader's primary access to the events of the
novel and their significance, yet the experience of reading the novel is often
like that of observing a play or film: We hear dialogue (often in page-long
interchanges without even repeated indications of speaker identity) and see
details of setting and action clearly, but without overt interpretation, without
description of inner states or even of possible ambiguous tones of voice or
facial expression, without clearly stated evaluations, suggestions, or judg-
ments of characters and their actions. Nevertheless, McCarthy provides the
reader access not only to the inner workings of the characters' minds but also
to interpretations of the events of the novel. To understand the means by
which such effects are achieved, we may turn to theories of communication
of intentionality, especially relevance theory (Sperber and Wilson 1995) and
the markedness model (Myers-Scotton chapter 2, this volume).
If these theories can explain the significance of the stylistic choices in All
the Pretty Horses, they may enable a reader to choose among the widely
varying interpretations of the plot and characters of the novel. To help us
understand the problems of interpretation of the novel, a brief summary
provides context. The protagonist, John Grady Cole, a sixteen-year-old Tex-
an bereft of family and heritage, travels to Mexico with a friend (Rawlins)
and befriends a younger man (Blevins) who rides a beautiful, but possibly
stolen, horse. When Mexicans steal the horse from the boys, they regain it,
but Blevins kills a man in the process and leaves the other boys. John Grady
41
42 S T Y L I S T I C C H O I C E S IN LITERATURE
and Rawlins find a hacienda where John Grady demonstrates his remarkable
ability to break and train horses. He is accepted until he falls in love with the
young heiress, Alejandra, and her family apparently betrays him to the law to
separate the girl from this lover, unsuitable both culturally and financially.
After Blevins is killed and Rawlins wounded, John Grady defends his life in
prison by killing a probably hired killer. Released from prison by some un-
known external agent, John Grady makes a futile attempt to reunite with
Alejandra and finally locates the beautiful horse and returns with it to Texas.
At the novel's end, he is seeking its rightful owner.
Even in this precis, we see that many of the most important actions and
relationships in the novel must be qualified as "apparent" or "possible,"
communicated only through implicature; McCarthy does not explicitly as-
sign cause and effect for many crucial elements in the plot. John Grady's
responsibility for his fate is thus a central problem in reading the novel: Is it
his own self-deception and misguided idealism that lead to his losses, or is he
the undeserving target of evil (especially as personified by the Mexicans and
women who wrong him)? This question has been debated: Bell (1992) and
Morrison (1993) see him as victimized by "the autocratic rule of families, at
best, and at worst, of brute power instead of law" (Bell 1992: 926), but Luce
(1995) argues that much that happens to John Grady results from his own
lapses in judgment and ignorance of cultural differences.
THEORETICAL FRAMEWORKS
Because the narrative voice offers little or no overt interpretation, only recov-
ery of the implicatures of the novel, especially those created by stylistic choic-
es, may help the reader to resolve this question. By manipulating the interplay
between a marked and an unmarked style in the narrative voice, McCarthy
constrains with some precision the reader's interpretation of the implicatures
of the narrative voice. An examination of these implicatures can in fact lend
weight to one aspect of the argument that it is the protagonist John Grady
Cole's own self-deception and misguided idealism, rather than foreign and
feminine villainy, that lead to his losses. Beyond that, offering an answer to
the problems posed by the ambiguities of the plot, the implicatures suggest
that the world does have meaning, shaped by responsibility, cause and effect,
and knowledge of consequences.
Relevance theory and other theories of the communication of intentional-
ity enable us to see how the reader is led to recover the appropriate implica-
ture by McCarthy's use of two sharply distinct styles in the narrative voice.
In analyzing literary language, Sperber and Wilson (1995: 222) have been
especially concerned with authors' use of weak implicatures to signal to read-
ers the necessity to recognize or create multiple meanings of texts. In the
cooperation between writer and reader, the reader must determine which
contexts to use for the recovery of meaning from the text, so that multilay-
Cormac McCarthy 43
ered and complex interpretations result when a writer leaves open the possi-
bility of various contexts. However, multiplicity of meaning can dissolve
into meaninglessness without some guidance from the writer as to which
inferences can most appropriately be drawn, so writers must supply what Sper-
ber and Wilson call "acts of ostension" (1995: 49) or noticeable signals that
give hints or set limits for selection of appropriate, or most relevant, contexts.
Relevance theory's analysis of communication of intentionality helps ex-
plain McCarthy's stylistic effects: Although he makes communication rich by
opening many possible implicatures, he makes it precise by using stylistic
choices to constrain the reader's selection of contexts for "optimal Rele-
vance" (Blakemore 1992: 177). This constraint is made manifest to the read-
er through acts of ostension.
McCarthy creates the most significant act of ostension in this novel by
using two distinct styles. He sets up an unmarked style for the uninterpreted
transmission of observations. He then introduces a marked style that leads
the reader to question the simplicity of the unmarked; indeed, it constrains or
forces us to interpret. The markedness model explains the choices of a speak-
er or writer as resulting from the making of rational (though not necessarily
overtly conscious) choices to optimize outcome— that is, to influence the
behavior of the hearer or reader (Myers-Scotton chapter 2, in this volume). In
the case of the reader of literature, that behavior is primarily internal: the
selection of one interpretation against another (or the maintenance of the
tensions of coexistent multiple readings). The choices that produce this out-
come, according to the markedness model, are the code choices that are
rationally believed to satisfy the various constraints of the situation at hand
(Myers-Scotton chapter 2, in this volume). In literature, these constraints are
partly those in social life but partly also those created by the novelist. Each
novel establishes its own linguistic community, with its own unmarked style,
and "sets of rights and obligations" (Myers-Scotton chapter 2, in this vol-
ume) attendant on it, its own "cognitive environment" (Sperber and Wilson
1995: 38-46), and thus develops its own constraints.
THE NOVEL AND COGNITIVE CONSTRAINTS
As an example, a writer who intends to be read by the majority of the Amer-

ican public is constrained to write in English, a choice made by McCarthy in
the main part of All the Pretty Horses; however, he may choose to represent
a world that includes characters who speak Spanish and may, as McCarthy
does, include untranslated Spanish dialogue. The reader must then either
know Spanish, find some way of translating the Spanish passages, or remain
ignorant of their content. McCarthy thus adds his own constraint (the ability
to interpret Spanish) to the larger social constraint (the ability to read En-
glish). Thus, untranslated Spanish in an English text becomes for this novel
the unmarked norm, though in many social contexts it would be marked.
44 STYLISTIC CHOICES IN L I T E R A T U R E
Thus, the unmarked style of this novel assumes a world of permeable borders
between languages.
This is only one example of an author's ability to create a style that is
unmarked for that particular text, and thus a style that is marked for that text
as well. Other obvious examples of contrast between the unmarked norm of
this novel and that outside it occur in punctuation: McCarthy never uses
quotation marks or indeed any punctuation to set off dialogue from narra-
tion and routinely drops apostrophes in contractions, though he maintains
them for the possessive—again lessening the distinction between the narrator
and the characters, between various levels of usage.
McCarthy creates a basic style for the novel that becomes the norm of
discourse, a simple, straightforward, undecorated style with the same un-
marked force as the linguistic and social unmarked choices of the natural
world. The majority of the novel is in this style, and the simple interpretation
of John Grady as wounded hero derives from it. But relevance theory ex-
plains how the styles of the narrative voice take us beyond that initial simplic-
ity. In the natural world and in the novel, the violation of constraints is a form
of communication. The writer's intention is communicated by the use of a
marked choice, one that clearly violates the norms of the unmarked choices.
Such a marked choice is vividly apparent in All the Pretty Horses: In only
a few passages McCarthy employs a style that is complex and metaphoric.
The style is marked both by the salience of the features—the unusual vocab-
ulary and occasional archaic structures—and by the quantitative differences
seen in Tables 3-1, 3-2, and 3-3.1 The use of this marked style is the act of
ostension that signals to the reader the need to recover complex implicatures,
to look beyond surface actions for motivations, assumptions, and mispercep-
tions. In the use of this style we find some of the clues that lead to the more
complex evaluation of John Grady's own complicity in his fate.
Finally, an even more complex message is communicated by the conver-
gence of the two styles, signaling the possibility of the convergence of the
1. The tables are based on six passages in the marked style and six in the unmarked, equaling
a total of sixty-five lines of type each (this edition has about twelve words of type per line). The
frequencies are determined by counting the total number of occurrences of each feature and
dividing by the number of sentences or lines. Sentences are considered to be groups of words
ending with periods, regardless of syntactic features. McCarthy's punctuation is very deliberate
(he often writes two words together to form an unhyphenated compound not found in dictionar-
ies, he uses apostrophes for contractions rarely and quotation marks for dialogue not at all, and
he does not italicize non-English words), and thus his periods must be taken as intentional
division of segments of the text.
For each passage in the marked style, a passage in the unmarked is selected from as closely
as possible the same segment of the book and from within ten pages when possible. The pairs of
passages appear throughout the novel. Because only twenty-two passages of more than five lines
in the marked style are found in the novel, and few of those occur during the second hundred
pages (approximately pp. 90-230), most pairs are taken from either the earlier or the later parts
of the novel. None contain dialogue. The pages and numbers of lines given for each passage refer
to the printed text (McCarthy 1992). (See Appendix)
Cormac McCarthy 45
simple world of action, lived out by the "heroic" protagonist and described
in the unmarked, simple style, and the more complex world of meaning
suggested and reflected by the marked, elaborated style. The gradual devel-
opment of this fluidity between the borders of the unmarked and marked
styles, like the increasing fluidity of the border between Spanish and English
in the novel, also signals the emergence of a new set of relations between
writer and reader: The reader of the mixed style must see the unity of action
and meaning.
NARRATIVE VOICE IN TWO EXAMPLES
To show in detail the developments of these constraints, I examine the stylis-

tic qualities of the narrative voice in two passages, using those passages to
demonstrate how McCarthy teaches us to recognize the constraints he im-
poses. As the dominant style of the narrative voice he establishes a transpar-
ent style, with little limitation of interpretation, against which he plays a
secondary, more highly constrained foregrounded style. Further, he uses the
elements of the foregrounded style to signal a thematic shift, an intrusion of
another level of meaning. Early in the novel, the foregrounded style occurs in
fairly extended passages. As we learn to recognize its characteristics, he uses
it more briefly and less often, so that a sentence or even a phrase evokes the
significance of the style. This association of style and meaning becomes one
of the constraints on interpretation that guide us in recovering the implica-
tions of the text.
COMPARING THE STYLES QUANTITATIVELY
McCarthy's styles are so clearly distinct that we can quantify the differences
in passages easily recognizable throughout the novel. Tables 3-1, 3-2, and 3-
3 detail a number of contrasts between the two styles. The tables show how
different features occur in the two styles both by frequency per sentence and
frequency per line of type. We must know the frequency of features per line
to understand how long the sentences are in each of the two styles. If we know
only the frequency per sentence, the longer sentences of the marked style will
artificially inflate the number of features. Thus, the frequency per line gives
a more accurate representation of the density of the reading experience.
However, the frequency with which each feature occurs in a sentence is also
important because co-occurrence of features in a sentence requires the reader
to sustain complex syntactic relationships.
One of the more obvious differences between the two styles is in the length
of sentence: The unmarked sentence has only 23.44 words per sentence,
whereas the marked has 42.11. The difference is less when we contrast the
words per line: 11.54 in the unmarked versus 12.32 in the marked. The
difference between these frequencies results from the much higher number
46 STYLISTIC C H O I C E S IN LITERATURE
Table 3-1
Syntactic Features
Per Sentence Per Line

Features Unmarked Marked Unmarked Marked
words 42.11 23.44 11.54 12.31

NP+VP 1.90 3.84 .94 1.12
VP 4.06 6.52 2.00 1.91
VP+VP .72 .42 .35 .12
copula .21 .58 .11 .17
non-finite VP .94 1.53 .46 .45
PP 3.13 5.6.3 1.54 1.65
ADJP 1.31 3.95 .65 1.15
ADVP .38 .68 .18 .20
NPVP+NPVP .40 1.63 .20 .48
overt COMP .31 1.63 .15 .48
ADJP post NP — .55 — .15
levels of mod. 1.53 2.89 .75 .85
of prepositional phrases in the marked style (5.63 in the marked, compared

to only 3.13 in the unmarked) and other constructions requiring short connec-
tive words.
Modifiers are more common in the marked (e.g., adjectives 3.95/sentence
marked to 1.31/sentence unmarked, nonfinite verbs 1.53/sentence marked to
0.94/sentence unmarked), but even more significantly, the marked has a much
higher level of modification (modifiers that in turn modify other modifiers—
2.89/sentence marked to 1.53/sentence unmarked). In the marked style, too,
adjectives often follow rather than precede the nouns they modify or follow
intransitive verbs of motion (0 in the unmarked, 0.15 per line in the marked).
The marked style also has more extremes—one-syllable words occur at the
rate of 10 per line in the marked to 9.37 in the unmarked; three-syllable
words at 0.49 per line, compared to 0.34 per line in the unmarked.
The syntax of the marked style is thus more complex than that of the
unmarked, and the word length much more varied. The effect of these longer
sentences is most obviously to foreground the language simply by the fact of
the marked difference. But an additional effect is the nature of the marked
style: The longer, more complex sentences and words require more effort and
slow down the reader's processing of thought, thus making these passages
doubly noticeable and more difficult to skim past without stopping to absorb
meaning. The length and complexity,of the sentences literally, almost phys-
ically, constrain the reader to find meaning.
Literary features are also much more common in the marked style: 0.20
metaphors occur per line in the marked, while 0 are in the unmarked, and
such prosodic features as alliteration are also heavily used in the marked
Cormac McCarthy 47
Table 3-2
Lexicon

words 23.43 42.11 11.54 12.30

Spanish words .22 .37 .11 .11
English words 23.29 41.74 11.43 12.20
syllables/word:
one 19.03 34.21 9.37 10.00
two 3.63 12.44 1.78 1.72
three .69 1.68 .34 .49
more than three .09 .32 .05 .09
(0.88/line) and rare in the unmarked (0.32/line). Other features are more
difficult to quantify: The rhythms—the variations of stress—in the unmarked
style are more nearly those of common speech, whereas those of the marked
are literary or biblical (resulting in part from the higher frequency of words
of more than two syllables). Again, these features not only mark the style but
involve the reader more intensely in processing the language.
Perhaps the most famous of McCarthy's stylistic devices is his use of rare
and arcane vocabulary, a feature that in All the Pretty Horses occurs almost
exclusively in the marked style. Few of these words are old or rare enough to
be marked as archaic in common dictionaries, yet the reader recognizes them
as rarely occurring in common speech—spume (161), isinglass (242), serried
(242), espaliered (73). This vocabulary in the marked style is itself a very
Table 3-3
Literary Features

metaphor — .68 — .20

simile .03 42 .02 .12
appositive .06 .32 .03 .09
parallelism .34 .95 .17 .28
formulaic structure — .70 — .23
alliteration .66 3.00 .32 .88
assonance .40 2.74 .23 .80
consonance .34 1.79 .17 .52
rhyme .13 .53 .06 .15
repetition .32 4.22 .15 .58
48 STYLISTIC CHOICES IN LITERATURE
specific act of ostension, requiring the reader to draw heavily on memory,

inference, or dictionary. Like the use of Spanish in the novel, the use of these
rarer words presupposes a reader who shares this knowledge of language or
is willing to acquire it for the sake of comprehending the passage. The words
themselves, like the syntax, constrain the reader to spend more time and
effort on interpreting the text and thereby to feel the weight of meaning
more heavily.
Though these quantified comparisons substantiate the existence of two
different styles and suggest their general nature, we can understand the work-
ings of the novel most clearly by examining several specific passages in detail.
To clarify the developments of these marked and unmarked styles, I examine
the stylistic qualities of the narrative voice in passages first of unmarked, then
of marked, and finally of the two styles mixed to reveal how McCarthy
teaches us to recognize the constraints he imposes.
THE U N M A R K E D STYLE
As the unmarked style of the narrative voice, McCarthy establishes a trans-

parent style used for 97% of the passages in narrative voice (see Table 3-1),
especially the narratives of action, as illustrated here in account of John
Grady and Rawlins as they wait for Blevins to retrieve his stolen horse:
Passage A
The boy slid from the horse and picked his way gingerly with his bare feet
across the road to the house and looked in. Then he climbed through the
window.
What the hell's he doin? asked Rawlins.
You got me.
They waited. He didnt come back.
Yonder comes somebody.
Some dogs started up. John Grady mounted up and turned the horse and
went back up the road and sat the horse in the dark. Rawlins followed. Dogs
were beginning to bark all back through the town. A light came on.
This is by God it, aint it? said Rawlins.
John Grady looked at him. He was sitting with the carbine upright on his
thigh. From beyond the buildings and the din of dogs there came a shout.
You know what these sons of bitches'll do to us? said Rawlins. You thought
about that?
John leaned forward and spoke to the horse and put his hand on the horse's
shoulder. The horse had begun to step nervously and it was not a nervous horse.
He looked toward the houses where they'd seen the light. A horse whinnied in
the dark.
That crazy son of a bitch, said Rawlins. That crazy son of a bitch. (82)
This style seems to be the simplest possible in English. Most sentences are
SV(subject-verb): the boy slid, Rawlins pulled, they rode, the leather creaked,
Cormac McCarthy 49
and so on. Verbs and their arguments stand in clear and strong relation to
each other; rarely do introductory or intervening phrases interfere with the
clear statement of action (only 3.13 PP (prepositional phrases) per sentence,
0.94 nonfinite verbs, including one-word modifiers—Table 3-2). The sen-
tences are short (23.43 words per sentence—Table 3-1) and often mono-
clausal (1.9 clauses per sentence—Table 3-1); clauses are combined more
often with and than with a complementizer of any kind. The directness and
immediacy of the syntax embody that of the scene. We find few adjectives
(1.25 per sentence—Table 3-3), either as preceders of nouns or as comple-
ments, and few adverbs (0.38 per sentence)—only 5% of all words are mod-
ifiers, whereas 14% are verbs (Table 3-2). Most of the adverbial information
concerns time and place, rarely manner: across the road, in the dark. Both
nouns and verbs are usually morphologically simple (Table 3-2 shows that
only 4% of the words in the unmarked passages have more than two sylla-
bles, including Mexican), and of Anglo-Saxon derivation (see Table 3-2):
slid, rode, shout, high, and so on.
This simplicity makes the correspondence between sentence form and
content even more noticeable, as we can see in the beginning of the next
passage (the departure of John Grady and Rawlins for Mexico):
Passage B-l
They rode out along the fenceline and across the open pastureline. The leather
creaked in the morning cold. They pushed the horses into a lope. The lights fell
away behind them. They rode out on the high prairie where they slowed the
horses to a walk. (30)
Here, most sentences consist of subject and intransitive verb, with few
direct objects in these passages—the actions happen without connections in
the intransitive sentences, without an actor who acts on something else—in
short, without cause and effect. The characters, especially John Grady, un-
dertake a venture without recognizing the extent of its possible consequenc-
es, an ignorance that ultimately leads to loss, imprisonment, and death. The
simplicity of the syntax also makes more prominent the semantics of the
passage. Case grammar analysis of this part of the passage shows how the
thematic roles—the semantic as opposed to syntactic cases of nouns and
verbs according to Hurst's (1990) schema—reveal the underlying nature of
the characters. Most notable is the absence of nouns acting not only as gram-
matical subjects but also as thematic agents, that is, as actors: The agents are
they. When we find a noun in subject position, it is thematically the recipient
of the action: leather creaked, bell that tolled, and, most vividly, lights fell—
although it is actually the riders who leave the lights, these riders do not take
here even the responsibility of moving through the landscape.
This is the standard style, the unmarked norm, of the novel. Remarkably,
it is used for the most crucial scenes: We find it not only in scenes of getting
dressed and feeding horses, buying food and closing gates but also in scenes
central to the plot, such as the execution of the younger boy, Blevins (177-
178), John Grady's killing of the man in prison (199-202), and the last
meeting of John Grady and the Mexican heiress Alejandra (247-248), among
others. In these passages the language appears limited, precise, but powerful;
the style mimics the action.
However, the unmarked style is in some ways also the least constrained.
By describing any action (even the buying of baloney) and therefore drawing
our attention to it, McCarthy suggests that it is significant; by leaving it
uninterpreted, he allows the reader to supply many contexts for its possible
interpretation. As our precis of the novel shows, many essential aspects of the
narrative are unknown; the lack of constraint in the telling of the story is so
great that the story is not completely told.
The term constraint refers to the limits set on an otherwise unlimited
cognitive process. According to Sperber and Wilson's (1995) formulation of
relevance theory, there is no limit on the interpretations, the weak implica-
tures, a reader may find in a text, nor, significantly, is there any need for a
reader to look beyond the most obvious strong implicatures. In Passage A, we
know that Blevins goes in the window, but we do not know what happens
once he is inside; we see only the horses' reactions. There came a shout, but
we are not told who shouts nor what is said. Even the coherence of the text
itself is almost completely implied; and is so ubiquitous, so constant, that its
meaning must often be interpreted from the context (it may imply sequence,
as in he leaned . . . and spoke . . ., contrast, as in and it was not a nervous
horse, etc.). Similar ambiguities arise from the use of pronouns which often
occur with no antecedent (the novel's initial appearance of the character
Luisa, for example, on page 3), so that their reference can be recovered only
from subsequent action. Throughout the novel, dialogue is not set off by
quotation marks, nor are speakers designated, so that the reader must assign
speech to character, possibly erroneously. Even in those scenes of crucial ac-
tion, therefore, much of the interpretation is left to the reader. If the scenes of
action were indeed the entire book, told only in this unmarked style, we might
well read John Grady as a simple man of action. If we are not forced to look
beyond his own stated motivations for supporting Blevins, for his silence at
his first appearance at the ranch, for courting Alejandra, we have the right to
impute to him only noble motivations, and we are not constrained to see more.
THE MARKED STYLE
Such undirected interpretation is not found in another style of the narrative

voice, the marked style, which appears all the more opaque by its contrast
with the transparent unmarked style. Because the unmarked style is so trans-
parent, so unforegrounded, the appearance of the marked style suddenly
draws the reader's attention to the language; in turn we are led to the signif-
icance of McCarthy's stylistic choices for understanding John Grady's char-
Cormac McCarthy 51
acter. We first recognize the style's characteristics in extended passages like

the remainder of the paragraph describing the departure for Mexico:
Passage B-2
They rode out on the high prairie where they slowed the horses to a walk and
the stars swarmed around them out of the blackness. They heard somewhere in
that tenantless night a bell that tolled and ceased where no bell was and they
rode out on the round dais of the earth which alone was dark and no light to it
and which carried their figures and bore them up into the swarming stars so that
they rode not under but among them and they rode at once jaunty and circum-
spect, like thieves newly loosed in that dark electric, like young thieves in a
glowing orchard, loosely jacketed against the cold and ten thousand worlds for
the choosing. (30)
The introduction of metaphor—stars swarmed—marks the transition of the
riders from the town to the high prairie (the beginning of their journey) and
stylistically marks the introduction of the marked style. Here we find subor-
dination: where no bell was, so that they rode, like thieves newly loosed; and
modification: that tenantless night, the swarming stars, young thieves in a
glowing orchard. Indeed, in this passage and others like it, there are 3.84
clauses per sentence, and 1.63 complementizers, as well as 1.53 nonfinite
verbs (Table 3-2).
Now we are in a world of myriad syntactic connections, though all are not
what they appear to be on the surface. The simple SVO sentence pattern is
contorted and obscured by intervening modifiers, inversions, ellipses. In this
maze of true and false syntactic leads, the use of the marked style argues the
possible falsity of the romantic heroism underlying John Grady's journey to
Mexico. The complementizer that in a bell that tolled dislocates the subject from
its action. Next, we find an apparent copula, was, actually followed by no
complement (where no bell was) and so joins nothing. We are also misled by the
elaborate repetitions, appositives, and parallel structures: they heard and they
rode; which . . . and which . . . ; like thieves newly loosed and like young
thieves . . ., loosely jacketed—all lead us to expect that the sentence has a
tightly controlled syntax in which every phrase fits. But our expectations are
confounded in the remainder of the sentence: they rode... like young thieves
..., loosely jacketed against the cold and ten thousand worlds for the choos-
ing. Parallelism of syntactic form—in this case the two sides of the coordinating
conjunction and—suggests a corresponding parallelism of ideas, but no way
does ten thousand worlds for the choosing parallel syntactically any phrase
related to it in content, certainly not loosely jacketed. The false parallelism
suggests a fundamentally erroneous assumption in the riders' view of their place
in the world. Just as the reader erroneously assumes that the first part of the
sentence will provide a pattern for the last, so the boys wrongly assume that the
Mexican world they enter will match the American world they leave. Thus, the
syntax of the marked style in this passage offers clear implicatures of meaning
underlying the simple world of action described in the unmarked passages.
Also an index of the marked style of All the Pretty Horses is the lexicon.
In earlier works, McCarthy used an enormous vocabulary, drawn from a
surprising range of social levels, historical periods, mental disciplines, profes-
sions, and contexts. Witek (1991: 51-53) speaks of the disparity between the
language of the characters and that of the narrative voice in his other novels,
Suttree (1973) and Blood Meridian (1985). In All the Pretty Horses, how-
ever, even in the marked style most of the vocabulary is more nearly that of
common speech but turned at an odd and jolting metaphoric angle by context
(as with stars swarmed, above). Only occasionally does a word from another
time or place or level of usage appear, as in the unexpectedly archaic in that
dark electric; again, the rarity of the occurrence emphasizes its effect and its
importance. The use of the word dais is one of the means by which the simple
cowboys become the mythological figures of the final lines. In one of the later
phrases, they rode at once jaunty and circumspect, the emphasis arises not
only from the unexpected register and combination but also from the syntax,
with the adjective complement (jaunty and circumspect) following a verb
(rode) that is not a copula. This structure is in fact formulaic in the novel; this
and other such structures (in that N, of some AdjN, etc.) appear only in the
marked style (see Table 3-3).
Finally, the marked style foregrounds sound through repetition. Witek
(1991: 56-63) and Morrison (1993:183-186) demonstrated the function of
repetition in the structure of several McCarthy novels, but we can see that
repetition of sound here also marks a shift of style. Repetitions of structure
and sound, especially of assonance and consonance, rhyme and alliteration,
occur throughout, even in the unmarked style, but they become much stron-
ger in the marked. In this passage, the unifying phrase is they rode: It recurs
four times and in each case is the core of the sentence. Other repetitions also
figure: stars, swarmed, bell, dark, loose, thieves. We hear assonance in the
repeated o sounds—no, bore, rode, glowing, and cold—and assonance as
well when the o gives way to oo—loosed, loosely, and choosing—to or and
ar horses, swarmed, where, earth, dark, orchard, and worlds—and rhyme
in the series fell and bell. Alliteration, too, is strong: stars swarmed, tenantless
. . . tolled, rode ... round; thieves . . . thousand. The passage is thus unified by
the interwoven recurrence of sound patterns. Rhythm here, too, is more varied:
A complex interplay of secondary and primary stresses can give way to a sudden
powerful burst of strong syllables (ten thousand worlds). This foregrounding of
prosody marks the style more insistently as a different kind of speech.
The marked style differs cognitively from the unmarked, also. As we saw
earlier, the unmarked style is less constrained; the reader is left to supply the
connections between the unadorned facts presented, and thus the interpreta-
tions of them. Although the marked style often suggests more possible inter-
pretations, it also more sharply limits those that might be most appropriate.
This appears on the most obvious level in the more frequent use of modifiers
and in the overt statements of precise sentence connectives (such as relative
Cormac McCarthy 53
pronouns) and in the preference for overtly stated (and thus more limited)
similes (like thieves) to more open metaphors.
The constraining function of this style is especially strengthened in Pas-
sage B-2 by the foregrounding of function words—adverbs (0.68 per sen-
tence; Table 3-3), complementizers (1.6 per sentence; Table 3-2), and espe-
cially prepositions (3.13 per sentence; Table 3-1), the means whereby Mc-
Carthy creates the image of the riders as raised into the heavens. Dominant
in reading the passage are the changing impressions of the riders first as a part
of the landscape (Cheuse 1992), then as on the surface of the planet (and
thereby raised in size), and finally as riding among the stars, heroic in size (an
apotheosis whose falsity is underscored by the young thieves simile). Al-
though the verbs account for some of the visual impression of upward move-
ment (rode, carried and bore), the image grows and is sustained throughout
the passage by a class of words that usually escape the attention of a reader:
prepositions. At least fifteen prepositions express motion, from their riding
along and across to the stars' swarming out of the blackness, the earth's
bearing them up into and their being loosed in the starry darkness. The
prepositions are further foregrounded by repetition—out on (11. 3-4 and 7)—
and negation—not under but among (1. 10). With such foregrounding, the
style itself becomes further associated with constraint, in the sense of the
limiting of infinite possible implicatures to the appropriate choice.
The shift to the marked style in such a sustained passage clearly functions
to focus our attention on the language itself rather than on the actions the
language describes. Such passages occur at several points where the reader
must not so much envision characters and actions as understand those actions
in their context—the context of history, the physical landscape, and the myth-
ical, moral, even ontological universe in which those characters act. The novel
opens with the most sustained of such passages—almost all of the first five pages
are in this style, including John Grady's sight of his grandfather in the casket (3),
a howling train (3-4), and a vision of Indian ghosts (5-6). Later, such passages
color the travels of the boys before the loss and recovery of Blevins's horse
(73) and the arrival at the Mexican hacienda, La Purisima (93). They tell of
John Grady's love for the horses (105, 128) and for Alejandra (141) and give
us his dreams of horses while in prison (161), the death of a doe (282), and the
final passage where he rides away, still leading Blevins's horse (302). The use
of the marked style in these passages teaches us to associate that style with
such subject matter: the sudden realization of the greater dimensions of ac-
tions. In these passages, McCarthy constrains us to see that events have mean-
ing; we are not at liberty to take them as simple actions with no significance.
MIXING MARKED AND UNMARKED STYLES
Substantial passages of five or more lines in the marked style are rare: Only
6.1 pages (about 2,600 words), or 3% of the 95,000-word novel, are in this
style. The passages are not more than a paragraph each (often only part of a
longer paragraph) and vary in length from 60 to 220 words. But the passages
teach us their language; we learn to associate it with this kind of meaning so
that we do not need a sustained passage to recognize the shift to another level
of the narrative, the level of meaning. A sentence, a phrase, or even a word
can be so clearly marked with this style that we recognize the sign, just as even
one word of Spanish summons up Mexico.
We can see also that when the marked style is used with greater economy
it may also have even more force, as in this passage narrating Don Hector's
first long interview with John Rawlins:
Passage C-l
They sat at a long table of english walnut. The walls of the room were covered
with blue damask and hung with portraits of men and horses. At the end of the
room was a walnut sideboard with some chafing dishes and decanters set out
upon it and along the windowsill outside taking the sun were four cats. Don
Hector reached behind him and took a china ashtray from the side board and
placed it before them and took from his shirtpocket a small tin box of english
cigarettes and opened them and offered them to John Grady and John Grady
took one.
Gracias, he said.
The hacendado placed the tin on the table between them and took a silver
lighter from his pocket and lit the boy's cigarette and then his own.
The hacendado nodded again. He sipped his coffee. He was seated sideways to
the table with his legs crossed. He flexed his foot in the chocolatecolored veal
boot and turned and looked at John Grady and smiled.
Why are you here? he said.
John Grady looked at him. He looked down the table where the shadows of the
sunning cats sat in a row like cutout cats all leaning slightly aslant. He looked
at the hacendado again.
I just wanted to see the country, I reckon. Or we did. (112-113; emphasis
added)
Here we can see how McCarthy's use of an image to suggest John Grady's
lapse in honesty is attended by stylistic signals with which it comes to be
associated in the passage. In C-l, the unmarked style dominates the narrative
voice. In most passages, the characteristically direct and straightforward sen-
tence pattern appears. Words are morphologically simple, of one or two
syllables, and concrete (english walnut, ashtray, sipped, looked); modifica-
tion is moderate and also clearly concrete (blue damask, silver lighter, thin
stream). The language is drawn from a lexicon which is standard for the time,
place, occasion, and level of society. The style suddenly shifts from unmarked
to marked in two places only: the references to the shadows of the cats. Both
cats and shadows appeared as images elsewhere (indeed, the book opens and
closes with the image of reflection and shadow). During this apparently inno-
cent conversation, ostensibly directed toward confirming John Grady's place
Cormac McCarthy 55
in the community of horsemen, the image of the cat's shadow lurks in the
background, pouncing only when John chooses falsely to deny affiliation
with Blevins, a denial of reality that eventually plays a part in his dismissal from
the hacienda and eventual imprisonment. The significance of the image is ech-
oed and underscored by the abrupt change to the marked style precisely at the
point in the conversation at which John Grady makes his ill-fated choice.
The first appearance of the cats actually occurs at the beginning of the
conversation, when they are catalogued among the items in the room. Here
in the unmarked style they are but one more item in the room, but later, when
the conversation turns to the potentially threatening question of the purpose
of John Grady's coming to the hacienda, the image of the cats recurs, cast in
the marked style: the shadows of the sunning cats sat in a row like cutout cats
all leaning slightly aslant (113). Stylistically this sentence differs considerably
from the previous narrations of the conversation, showing the marked fea-
tures we saw in B-2: the unexpected simile (like cutout cats), nonfinite verbs
(sunning, leaning, and perhaps aslant), inverted modification (all leaning
slightly aslant follows cutout cats), levels of modification (slightly modifies
aslant which modifies leaning, which in turn modifies cats, etc.), and fore-
grounded prosody (alliteration of s: shadows, sunning, sat, slightly; asso-
nance of a: shadows, cats, aslant; and repetition and rhyme or near-rhyme:
cats, sat, cutout, cats). This intrusion of the marked style thus constrains our
interpretation of the scene. By this time, we have learned to associate such
language with the linking of actions to their larger context; it forces us to
recognize John Grady's choice, the choice to lie about his reason for coming
there: I just wanted to see the country, I reckon.
Then the conversation leaves this dangerous subject, and the cats are
unmentioned during the conversation and the unmarked style narration of
three pages. At the end of the interview, however, the conversation again
turns to a question that should lead John Grady to discuss his connection
with Blevins, and the cat image recurs:
Passage C-2
The hacendado leaned back in his chair. One of the cats rose and stretched.
You rode here from Texas.
Yessir.
You and your friend.
Yessir.
Just the two of you?
John Grady looked at the table. The paper cat stepped thin and slant among the
shapes of cats thereon. He looked up again. Yessir, he said. Just me and him.
The hacendado nodded and stubbed out his cigarette and pushed back his chair.
Come, he said. I will show you some horses. (116; emphasis added)
The danger lies in John Grady's choice to lie, not only because of his dishon-
esty to the hacendado but also because of his unwillingness or inability to
recognize the evil he has already chosen. By denying his connection with
Blevins to Don Hector, he denies it to himself: John Grady looked at the

table. The paper cat stepped thin and slant among the shapes of cats thereon.
He looked up again. Yessir, he said. Just me and him [Rawlins] (116). The
danger of this false innocence is embodied in the image of the cat, now
narrowed to one of the cats, suggesting that the danger, too, is no longer
general but specific: The image appears while John Grady considers yet
another denial of his connection with Blevins.
The danger implicit in John Grady's refusal to accept responsibility for
choice is embodied also in the startling shift of stylistic choice, which again
operates to limit our interpretation of the scene—we must recognize that the
conversation is connected with a greater context because we have learned to
recognize the style as carrying that significance, and we know more precisely
what context, because of the repetition of stylistic features from the earlier
scene (C-l). The term paper cat is a leap into the metaphoric without warn-
ing, and the reference is extremely distant, linked to the metaphor of the cats'
shadows as cutouts mentioned only once, three pages earlier, just as the word
slant is a repetition also carried from that distant previous passage. This
distant link mirrors the distant connections with Blevins deliberately ignored
by John Grady. The sibilant s of the earlier passages recurs as well: stepped,
slant, shapes, cats; the hissing is reinforced by the fricatives 9 and 9: the, thin,
thereon and the nasal n: thin, slant, among, thereon. There is rhyme—paper
and shape—and repetition—cat. Again the syntax is marked, with adjective
complements (thin and slant) following a verb of action rather than a copula.
The word thereon, drawn from a more formal lexicon, is emphasized by its
sentence-final position.
The marked style is thus so clearly marked, and in this extended passage
so clearly associated with the intrusion of another element, that McCarthy
can achieve a powerful effect with only a single sentence. We cannot miss the
importance of John Grady's failure to speak; our interpretation of this is
constrained by the use of the style we have come to recognize as representing
a particular kind of significance. Throughout the novel, such small but pow-
erful occurrences accumulate even more significance. Often they occur be-
fore or after some event crucial to the action, to prepare us (constrain us) to
interpret that action in a certain way: a thunderstorm that will eventually
lead to the loss of Blevins's horse, for example, glowed mutely like welding
seen through foundry smoke. As if repairs were under way at some flawed
place in the iron dark of the world (67); riding across the countryside, after
parting from Alejandra, he thinks of universal pain as some formless parasitic
being seeking out the warmth of human souls wherein to incubate (256).
Even these brief passages occur rarely—once for every seven or eight pages of
text, often clustered before and after crucial scenes—but they serve to control
our understanding of much of the action narrated in the more open un-
marked style. Most especially in the final sections of the book do we see the
Cormac McCarthy 57
marked style as commentary on the events narrated by the unmarked. It is

also in these final scenes that John Grady, by becoming more aware of the
consequences of his actions, comes closer to seeing those actions as happen-
ing in that larger context (indeed, even a brief appearance of the marked style
reminds us of that larger context). The mixing of the styles to some extent
embodies his painfully acquired awareness of the web of cause and effect
within which he acts. The reader is also constrained to recognize that the
larger context of the novel includes notions of responsibility, cause and ef-
fect, and ultimately therefore a moral significance of actions.
INTERACTION OF STYLES: ACTIONS AND MEANINGS
Thus, it is the interaction of styles in this novel that allows us access to the
inner workings of characters' minds and, even more, to the workings of the
narrative. The unmarked style shows us the action of the novel, but action
whose causes and consequences are unclear and so may mean anything or
nothing. The marked style suggests the framework for recognizing the mean-
ings of these actions and requires or constrains the reader to interpret them.
Even the occasional use of the marked style offers the reader the implicatures
that suggest a more precise and complex interpretation of the characters and
the world of the novel. By so doing, the style limits the infinite options of
interpretation and especially denies the possibility of random action, action
without meaning. The convergence of styles at the end of the book thus
comes to signify the necessary impingement of meaning upon action.
ACKNOWLEDGMENT: Portions of this chapter were originally printed as Kreml, Nancy (1995).
REFERENCES
Bell, Vereen. 1992. Between the wish and the thing the world lies waiting. Southern
Review 1992: 920-927.
Blakemore, Dianne. 1992. Understanding utterances. Cambridge, Mass.: Blackwell.
Cheuse, Alan. 1992. A note on landscape in All the pretty horses. Southern Quarterly
30: 146-148.
Hurst, Mary Jane. 1990. The voice of the child in American literature: linguistic
approaches to fictional child language. Lexington: University of Kentucky Press.
Kreml, Nancy (1995). Stylistic variation and cognitive constraint in Cormac McCar-
thy's All the pretty horses. In Sacred violence: a reader's companion to Cormac
McCarthy, ed. Wade Hall and Rich Wallach, 137-148. El Paso: University of
Texas at El Paso.
Luce, Dianne. 1995. When you wake: John Grady Cole's heroism in All the pretty
horses. In Sacred violence: a reader's companion to Cormac McCarthy, ed.
Wade Hall and Rich Wallach, 155-168. El Paso: University of Texas at El Paso.
McCarthy, Cormac. 1973. Suttree. New York: Random House.
. 1985. Blood Meridian or the evening redness in the west. New York:
Random House.
1992. All the pretty horses. New York: Knopf.
Morrison, Gail. 1993. John Grady Cole's expulsion from paradise. In Perspectives on
Cormac McCarthy, ed. Edwin T. Arnold and Dianne C. Luce, 173-192. Jackson:
University of Mississippi Press.
Sperber, Daniel, and Deidre Wilson. 1995. Relevance, cognition and communication.
Witek,Terri. 1991. "He's hell when he's well": Cormac McCarthy's rhyming diction.
Shenandoah 41: 51-66.
APPENDIX
Unmarked Passages (32 sentences, 750 words)
p. 6 (9 lines):
The house was built in eighteen seventy-two. Seventy-seven years later his grandfather
was still the first man to die in it. What others had lain in state in that hallway had been
carried there on a gate or wrapped in a wagonsheet or delivered crated up in a raw
pineboard box with a teamster standing at the door with a bill of lading. The ones that
came at all. For the most part they were dead by rumor. A yellowed scrap of news-
print. A letter. A telegram. The original ranch was twenty-three hundred acres out of
the old Meusebach survey of the Fisher-Miller grant, the original house a oneroom
hovel of sticks and wattle.
pp. 22-23 (8 lines):

They rode together a last time on a day in early March when the weather had already
warmed and the yellow mexicanhat bloomed by the roadside. They unloaded the
horses at McCullough's and rode up through the middle pasture along Grape Creek
and into the low hills. The creek was clear and green with trailing moss braided over
the gravel bars. They rode slowly up through the open country among scrub mesquite
and nopal. They crossed from Tom Green County into Coke County.
p. 74 (13 lines):
The riders got their plates and utensils out of the saddlebags and John Grady got the
little enameled pot out of the blackened cookbag and handed it to Blevins together
with his old woodenhandled kitchen fork. They went to the fire and filled their
plates with beans and chile and took each a couple of blackened corn tortillas
from a piece of sheetiron laid over the fire and walked over and sat under the
willows a little apart form the workers. Blevins sat with his bare legs stretched
out before him but they looked so white and exposed lying there on the ground
that he seemed ashamed and he tried to tuck them up under him and to cover his
knees with the tails of the borrowed shirt he wore. They ate. The workers had for
the most part finished their meal and they were leaning back smoking cigarettes
and belching quietly.
Cormac McCarthy 59
p. 154 (7 lines):
They sat side by side on a bench of iron slats in the little alameda. A pair of the
guards stood a little ways off with their rifles and a dozen children of different
ages stood in the dust of the street watching them. Two of the children were girls
about twelve years of age and when the prisoners looked at them they turned shyly
and twisted at their skirts. John Grady called to them to ask if they could get
them cigarettes.
p. 246 (11 lines):

He offered to pay in advance but the proprietor dismissed him with a small wave of
the hand. He walked out into the sun and down the street where he caught the bus
back to town.
He bought a small awol bag in a store and he bought two new shirts and a new pair
of boots and he walked down to the train station and bought his ticket and went to a
cafe and ate. He walked around to break in the boots and then went back to the hotel.
He rolled the pistol and knife and his old clothes up in the bedroll and had the clerk
put the bedroll in the storage room and he told the clerk to wake him at six in the
morning and then went up to bed. It was hardly even dark.
p. 279 (17 lines):

Ascending into low hills they passed a small estancia and they dismounted and went
afoot through the ruins of a cornfield and found some melons and sat in the stony
washedout furrows and ate them. He hobbled down the rows and gathered melons
and carried them out through the field to where the horses stood and broke them
open on the ground at their feet for them to feed on and he stood leaning on the rifle
and looked toward the house. Some turkeys stepped about in the yard and there was
a pole corral beyond the house in which stood several horses. He went back and got
the captain and they mounted up and rode on. When he looked back from the ridge
above the estancia he could see that it was more extensive. There was a cluster of
buildings above the house and he could see the quadrangles laid out by the fences
and the adobe walls and irrigation ditches. A number of rangy and slatribbed cattle
stood about in the scrub. He heard a rooster crow in the noonday heat. He heard a
steady distant hammering of metal as of someone at a forge.
Marked Passages (19 sentences, 800 words)

p. 6 (9 lines):
The last of the day's light fanned slowly upon the plain behind him and withdrew
again down the edges of the world in a cooling blue of shadow and dusk and chill and
a few last chitterings of birds sequestered in the dark and wiry brush. He crossed the
old trace again and he must turn the pony up onto the plain and homeward but the
warriors would ride on in that darkness they'd become, rattling past with their stone-
age tools of war in default of all substance and singing softly in blood and longing
south across the plains to Mexico.
p. 30 (10 lines):
They rode out on the high prairie where they slowed the horses to a walk and the stars
swarmed around them out of the blackness. They heard somewhere in that tenant-
less night a bell that tolled and ceased where no bell was and they rode out on the
round dais of the earth which alone was dark and no light to it and which carried
their figures and bore them up into the swarming stars so that they rode not under but
among them and they rode at once jaunty and circumspect, like thieves newly loosed
in that dark electric, like young thieves in a glowing orchard, loosely jacketed against
the cold and ten thousand worlds for the choosing.
p. 73 (9 lines):
Bye and bye they passed a stand of roadside cholla against which small birds had been
driven by the storm and there impaled. Gray nameless birds espaliered in attitudes of
stillborn flight or hanging loosely in their feathers. Some of them were still alive and
they twisted on their spines as the horses passed and raised their heads and cried
out but the horsemen rode on. The sun rose up in the sky and the country took
on new color, green fire in the acacia and paloverde and green in the roadside run-
off grass and fire in the ocotillo. As if the rain were electric, had grounded circuits that
the electric might be.
p. 161-162 (17 lines):

That night he dreamt of horses in a field on a high plain where the spring rains had
brought up the grass and the wildflowers out of the ground and the flowers ran
all blue and yellow far as the eye could see and in the dream he was among the
horses running and in the dream he himself could run with the horses and they
coursed the young mares and fillies over the plain where their rich bay and their
rich chestnut colors shone in the sun and the young colts ran with their dams and
trampled down the flowers in a haze of pollen that hung in the sun like powdered
gold and they ran he and the horses out along the high mesas where the ground
resounded under their running hooves and they flowed and changed and ran and
their manes and tails blew off of them like spume and there was nothing else at
all in that high world and they moved all of them in a resonance that was like a
music among them and they were none of them afraid horse nor colt nor mare
and they ran in that resonance which is the world itself and which cannot be
spoken but only praised.
p. 242 (7 lines):
He crossed a dry gypsum playa where the salt crust stove under the horse's hooves like
trodden isinglass and he rode up through white gypsum hills grown with stunted datil
and through a pale bajada crowded with flowers of gypsum like a cave floor uncov-
ered to the light. In the shimmering distance trees and jacales stood fugitive in the clear
morning air.
p. 282 (14 lines):

The sky was dark and a cold wind ran through the bajada and in the dying light a cold
blue cast had turned the doe's eyes to but one thing more of things she lay among
in that darkening landscape. Grass and blood. Blood and stone. Stone and the
dark medallions that the first flat drops of rain caused upon them. He remem-
bered Alejandra and the sadness he'd first seen in the slope of her shoulders
which he'd presumed to understand and of which he knew nothing and he felt a
Cormac McCarthy 61
loneliness he'd not known since he was a child and he felt wholly alien to the
world although he loved it still. He thought that in the beauty of the world were
hid a secret. He thought the world's heart beat at some terrible cost and that the
world's pain and its beauty moved in a relationship of diverging equity and that
in this headlong deficit the blood of multitudes might ultimately be exacted for
the vision of a single flower.
4
Marked Grammatical Structures:

Communicating Intentionality in The Great
Gatsby and As I Lay Dying
CAROL MYERS-SCOTTON
Gatsby believed in the green light, the orgastic future that year by year recedes
before us. It eluded us then, but that's no matter—tomorrow we will run faster,
stretch out our arms farther. . . And one fine morning—So we beat on, boats
against the current, borne back ceaselessly into the past.
- F. Scott Fitzgerald, [1925] 1991: 141
M ANY SCHOLARS of American literature can recite these final lines of F.

Scott Fitzgerald's The Great Gatsby. Many think the work is the great
American novel. Fitzgerald's style—his evocative metaphors and similes—of
course, are one reason Fitzgerald's prose is valued. For example, many read-
ers remember such lines as Gatsby believed in the green light or that Dai-
sy's voice sounded like money. Another twentieth-century American nov-
elist whose style is cited as an essential part of his greatness is William
Faulkner. The Faulkner work studied here is As 7 Lay Dying, the novel
Faulkner himself said is the result of setting out to write a tour de force. In
that novel, the style of each chapter represents the point of view of a partic-
ular character.
However, this chapter provides a stylistic analysis of the less obvious
aspects of the style of Fitzgerald and Faulkner than either the crafting of met-
aphors and similes or the perfection of the point-of-view technique. Here, we
consider the interplay of syntax and what I call cognitive weight. Kintsch (1974)
reported that reading times increase as a function of the number of propositions
in the text. Even controlling for other factors, such as the number of words in the
text, does not alter the result. Different types of structures are the subject in the
two different novels, but both have one feature in common: They are marked
in the sense that they depart from the prevailing, or unmarked, syntax of
62
Fitzgerald and Faulkner 63
each of the novels. Whereas the syntactic structures to be discussed are less
obvious aspects of style than the stylistic features commonly associated
with either Fitzgerald or Faulkner, I argue that the relation between struc-
ture and cognitive weight is perhaps an even more important aspect of their
styles as the carrier of intentional messages. Specifically regarding Fitzgerald, I
claim that through the syntactic structures he employed, whether consciously
or unconsciously, he made sure that his "crucial passages" would stand out
as "the figure in the carpet." I demonstrate that Fitzgerald adds to the import of
certain passages "crucial" to the novel's central message through using struc-
tures (Longacre 1985: 86, 87) referred to as "packing the event line" or "slow-
ing the camera down." In regard to Faulkner, I show how he amplifies especially
one key character in As I Lay Dying by giving him a distinctive "syntactic
signature" that sets his monologues off from those of other characters.
HYPOTHESES UNDER CONSIDERATION
My general hypothesis is in line with the theme of this volume: Linguistic

virtuosos will exploit marked linguistic choices to convey intentional mean-
ings in their discourse. Writers who are linguistic virtuosos either "mark" the
crucial passages in their works or mark key characters to set them off from
the rest of the work or the rest of the characters.
Marked choices, of course, are shifts from the expected or the norm, given
the social situation or the genre involved. Although all humans have tacit
knowledge that there is a multidimensional markedness continuum for every
interaction type in their repertoire, what choices are more unmarked or marked
is both community specific and interaction specific. Further, within a genre,
such as the novel, each specific novel has its own unmarked style. Humans
are innately predisposed to view linguistic choices in terms of markedness;
however, specific readings of markedness depend on experience in the real
world (Myers-Scotton chapter 2, this volume).
Intentional meaning refers to implicatures, to use the term of Grice (1975)
and Sperber and Wilson (1995). Implicatures convey meaning in addition to
that which is contained in the semantics of individual words and which is
transmitted in their semantic and syntactic relationships with other words in a
discourse. Although messages of intentionality may simply augment referen-
tial messages, they generally go beyond them in various ways. A major source
of inference, of course, is what is generally called style—not the words them-
selves but, rather, the cognitive associations they call up or which are called up
by one way of "saying something" out of the larger set of possible ways. In this
vein, departures from expected style, for example, the fact that a marked syntac-
tic structure is employed, are sources of implicatures or intentional meaning.
The more specific hypotheses studied here are as follows:
1. Fitzgerald marks the crucial narrative passages in The Great Gatsby
by using syntactic structures that are demonstrably different from the
64 STYLISTIC C H O I C E S IN L I T E R A T U R E
typical (or unmarked) narrative passages in the novel as a whole. That

is, a sample of five key crucial passages will differ from five superficially
similar passages in regard to a set of syntactic structures. "Crucial passag-
es" mean those passages literary critics often identify as "carrying the
author's message" (Matthew Bruccoli, personal communication). All the
syntactic structures to be studied are discussed here, but the main investi-
gation concerns the incidence of projections of complementizer (CP).
2. Faulkner distinguishes the "more heroic" of the characters in As I Lay
Dying—that is, Darl and to a lesser extent Vardaman—by their exten-
sive use of a syntactic structure, which I call the trailing constituent.
The frequency of this structure is marked in comparison with its use by
other characters. Trailing constituent is a cover term for both what
Thompson (1983) called detached participial phrases and for certain
types of prepositional phrases—those of condition (referring to time
and location) and those of manner, process, and direction.
I N D E P E N D E N T MOTIVATION FOR HYPOTHESES
The general hypothesis is motivated by previous theories in literary stylistics,

as well as by several linguistic theories of discourse structure. First, the mo-
tivation for the general hypothesis comes from the markedness model (My-
ers-Scotton chapter 2, in this volume). A claim of the model is that a major
component of the indexicality of any single linguistic form or variety is its
markedness reading. A markedness reading refers to the extent to which the
form or variety in question fits the pattern of what is unmarked or expected
in a given interaction type or genre. Of course, all interpretations of linguistic
choices call up schema—composed of motives, desires, ideas, and peopled
events—with which a linguistic choice has become associated through expe-
rience. What is of special interest here are choices which are marked. The
argument of the markedness model is that writers and speakers are motivated
to employ marked choices at times just because they have the special value of
evoking an "otherness." As such, they send a special message to the reader/
addressee that is over and above the referential meanings they include.
Speaking of marked structures is reminiscent of the notion of foreground-
ing, that has its roots in the work of the Russian formalists. As one of the
members of this school, Viktor Sklovskij, argued, the function of art is to
make people aware of the world in a fresh way. A device for achieving this is
defamiliarization or ostranenie (Russian) "makingstrange." One of Sklovskij's
elaborations on this view is particularly relevant to this chapter, which claims
that authors add cognitive weight to their words by making marked choices:
[Art] exists to make one feel things, to make the stone stony. The purpose of art
is to impart a sensation of things as they are perceived and not as they are
known. The technique of art is to make objects "unfamiliar," to make forms
difficult, to increase the difficulty and length of perception because the process
of perception is an aesthetic end in itself and must be prolonged, (quoted in Van
Peer 1986: 1; emphasis added)
The formalists, as their name indicates, studied literature mainly in its for-
mal aspects, eventually arriving at the concept of foregrounding. This term
was introduced to the West by Garvin (1964) as a translation of the Czech
aktualisace, employed in the work of several Prague scholars. One of the
problems with the Russian formalists and the Prague school of (linguistic)
structuralists, however, was a failure to indicate how one could identify
what would count as defamiliarization or foregrounding. Another problem
is with the way in which the Prague school looked at structure: They saw
grammar— or language—as largely a surface phenomenon of items arranged
in slots. That is, they operated as if linguistic forms can be separated from
their conceptual structure, and they did not consider that more abstract
structural hierarchies might underlie surface strings.
Recently, a model relying on the idea that nonfigurative language is un-
marked whereas figurative language (including literary language) is made
strange or foregrounded was criticized by Freeman (1996). Freeman seized
on two notions. First, even though he does not explicitly state it, he clearly
objects to interpretations that rely primarily on decomposing surface fea-
tures. Second, he more explicitly opposes the notion that has always been
associated with foregrounding, that literary language is "deviant" and
that its effect depends on this feature. With a specific interest in charac-
terizations of metaphor, Freeman takes this notion a step further. He
attributes to the foregrounding model (he refers to it as the objectivist
position) the idea that "a structure is parasitic upon ordinary language struc-
ture and that it can be explained in terms of its deviance from semantic
interpretations of ordinary, non-metaphorical language interpretable by a
compositional semantics ..." (Freeman 1996: 280). According to Freeman,
this model does not take account of cognitive processing but, rather, ex-
plains metaphors in terms of their deviance in reference to structural
features. He writes "The 'deviance' is with respect to a language-univer-
sal set of semantic features and cornbinatory rules external to individual
cognition." (1996: 80) As an alternative explanation to how metaphor works,
Freeman offers Lakoff's (1987: xv) experientialist position. On this view,
metaphor is "embodied human understanding" (Freeman 1996: 281); that
is, based on their experience, humans extract schema that they use to inter-
pret metaphors.
Contrary to Freeman, I would argue that the foregrounding model can be
"rescued," especially if it is combined with the markedness model. First,
there is no reason to assume that deviant means parasitic. That is, marked
choices do not depend for their intentionality on unmarked choices. Rather,
what markedness accomplishes in this regard is simply to highlight the fact
that what is said or written is not unmarked and therefore deserves special
attention. After all, the primary meaning of foregrounding is to draw atten-

tion. Second, under the markedness model, all uses of language require cog-
nitive processing for the addressee/reader to arrive at an interpretation. Al-
though it is possible that a compositional view of semantics and grammar
may suffice for the addressee to arrive at referential meaning (but this is a
controversy beyond the scope of my consideration here), it seems clear there
is no way that an addressee can achieve intentional meaning without engag-
ing in cognitive calculations. (This matter is discussed more extensively in
chapter 2 in this volume; see also Sperber and Wilson 1995). Further, in the
formulation of the message (the work of the writer/speaker) and in interpreting
the message (the work of the reader/addressee), both parties must necessarily
bring to bear their own experience, as members of the salient speech communi-
ty, to achieve the desired communicative effect. That is, the process neces-
sarily is not very different from that which defines the position for which
Freeman argues: "The Experientalist position claims that we create met-
aphor by projecting onto an abstract target domain the entities and struc-
ture of a concrete source domain, a schematized real or vicarious bodily
experience" (1996: 280-281). Thus, I see no need to equate the foreground-
ing model with objective, mechanistic decomposition, especially as this model
is reinterpreted within the markedness model, for two reasons. First, the defin-
ing feature of a foregrounded or marked structure is not its deviancy but rather
its "otherness." One must consider not only that it departs from the unmarked
style but also what its markedness in the context where it occurs accomplishes,
in terms of both the effects of its structure and its associations. Second, to
interpret the implicatures a marked structure evokes, the reader/addressee clear-
ly must carry out cognitive calculations, both to process the structural effects
and to call up the real-world associations of the marked structure.
The idea that there should be a relation between "crucial" passages or
character development and marked syntactic structures is specifically moti-
vated by Longacre's discussion of "discourse peaks" (Longacre 1985).1 An
overall feature of discourse peaks is that the "event line" fades at a peak. The
reason to orchestrate this fading is that "priority is given to marking the peak
of the discourse. In effect, then, the very disappearance of the routine event-
line markers is itself a marking of peak" (Longacre 1985: 85). In many differ-
ent places, Longacre discusses the "conspiracy of features" found in and
around peaks or climatic points in discourse. He lists ten structural features,
all showing a "level of excitation peculiar to the peak as contrasted with the
rest of the discourse" (Longacre 1985: 85). These include rhetorical underlin-
ing by means of repetition and paraphrase, heightened vividness by a tense
shift or a person shift, or dramatic shifts toward action with many more verbs
1. Longacre and his students, often missionary linguists connected with the Summer Institute
of Linguistics, regularly translated texts in what might be called relatively exotic languages. They
championed the need for basing grammatical analysis on discourse samples long before discourse
analysis as it is known today had any stature in the discipline of linguistics.
than found in other passages. I argue that Fitzgerald highlights his crucial
passages by "packing the event line"; Longacre says that one way this can
be accomplished is by increasing the ratio of verbs to nonverbs. Howev-
er, I argue that Faulkner distinguishes his character, Darl, by doing what
might be considered the opposite—by avoiding the weight that finite
verbs and their arguments involve. Longacre states that the pace can be
slowed by treating structures that are not usually on the event line as if
they were; how this characterization applies to Darl's monologues is dis-
cussed later.
ANALYSIS OF STRUCTURE IN "THE GREAT GATSBY"
In the case of The Great Gatsby, the specific hypothesis predicts that a com-
parison of five narrative paragraphs containing crucial messages with five
paragraphs whose contribution is more only to develop the story or event line
will show different syntactic patterns in the two sets. To identify crucial
passages, I consulted a leading Fitzgerald scholar, Matthew Bruccoli, for
advice. Five passages he chose were paired with five nearby narrative
passages. These matched the crucial passages as closely as possible in regard
to both subject matter and length. The longest crucial passage has 101 words
and the shortest one has 80 words, for a total of 736 words and 23 sentences
versus 751 words and 25 sentences in the unmarked passages.
By examining randomly chosen narrative paragraphs throughout the nov-
el, I confirm that the "matching" passages are representative of most narra-
tive passages in the novel. These matching passages are defined as unmarked
passages, whereas—if the hypothesis is supported—the crucial passages are
defined as marked passages.
Based on Longacre's findings and claims, the specific hypothesis predicts
that more structures would load the event line in the crucial passages.
Four main categories were studied: (1) embedded or subordinating CPs;
(2) Inflection (INFL) phrases (projections of INFL, or IPs ); (3) preposi-
tional phrases (PPs); and (4) types of verbs. I also considered other features
such as (1) configurations of noun phrases (NPs); (2) conjunctions as links for
IPs versus links for NPs and other XPs (X being some variable), such as verb
phrases (VPs), adjective phrases (APs), and PPs. In addition, I examined parti-
cipials and their functions. I used Eyeball (a computer parsing program) to
count lexical categories.
To test the hypothesis, I drew detailed syntactic trees of the sentences in
each set of passages. The difference in the trees of the two sets was immedi-
ately striking, supporting the hypothesis. Most obvious is that the trees for
the crucial passages simply had more "foliage"; that is, the sentences includ-
ed more hierarchical branching. This is so even though the average number
of words per sentence (thirty-two in the crucial passages compared to thirty
in the unmarked ones) was very similar.
INTERPRETATION OF RESULTS FOR "THE GREAT GATSBY"
Because it was immediately clear from the tree diagrams that the two types of
passages differed so that, at least in some way, the passages containing the
crucial messages were distinct from the unmarked passages, from now on the
two types are referred to as the unmarked and marked passages. Figures 4-1
through 4-4 present tree diagrams that are discussed next.
Interestingly, the differences between the unmarked and marked passages
have little to do with incidence of lexical categories per se. The parsing anal-
ysis program showed that there was little difference in the two types of pas-
sages in regard to the type of features often counted in discourse analysis
studies (e.g., ratio of nouns to adjectives to the total number of words, or
auxiliaries to the total number of verbs).
Results for Elaborating CPs

The story is quite different, however, when one considers the incidence of
elaborating CPs. CPs, of course, can be either independent or dependent
clauses. Only elaborating CPs were counted. This category included such CPs
as the following from a marked passage: (i) adverbial adjuncts such as when
I came back from the East last autumn; (ii) sentence complements such as
that I wanted the world to be in uniform and at a sort of moral attention
forever in the sentence [ I felt [that I wanted the world to be in uniform and
at a sort of moral attention forever]]; and also (iii) relative clauses such as
which is dignified under the name of creative temperament (Fitzgerald
[1925]1991: 5-6). There are many more elaborating CPs in the marked pas-
sages than the unmarked ones. The overall count is thirty-one (in the marked
passages) to fifteen. This difference is statistically significant, according to a
X2 test. (See table 4-1.)
Table 4-1
Comparisons: Modifying CPs, IPs, NP & NP (& NP)
The Great Gatsby
Crucial/Marked passages Unmarked Passages

(words: 736) (words: 751)
Category N % N % X2
CPs 31 4.2 15 1.9 *5.564 (p <.02S)

IPs 7 0.9 54 0.53 0.362yc
NP&NP(8cNP) 4 0.54 13 1.7 *3.764YC ( <.10)
Note: N = number of tokens; % = proportion of category tokens expressed as a

percentage of word total; 1 df; .10 and .025 = the probability level; YC = Yates'
Correction applied; * = significant.
Incidence of IPs
The marked passages use almost twice as many IP-and-IP constructions as the
unmarked ones; however, at seven versus four the frequencies are not large
enough to give me confidence that this ratio would stand across a larger
corpus and are not large enough to perform a test of statistical significance.
(See table 4-1.) Still, when they occur, the IP-and-IP constructions are distinc-
tive. For example, the first sentence in the marked passage from the Bruccoli
edition (Fitzgerald [1925] 1991: 34) shows a three-part IP-and-IP construc-
tion: The lights grow bright (as the earth lurches away from the sun); now the
orchestra is playing yellow cocktail music; and the opera of voices pitches a
key higher. These three IPs are joined by the conjunction and. In contrast, in
the unmarked passages, this conjunction tends to link not IPs but other XPs,
especially NPs, but also VPs, APs, and PPs. For example, in figure 4-4, which
presents a tree diagram of a sentence from an unmarked passage, there are six
NPs linked by and.
Incidence of Prepositional Phrases

Although the unmarked passages show more PPs, eighty-one to seventy-three,
the marked passages use more PPs to show "process," "direction," or "man-
ner" of activity or action. In the unmarked passages, PPs tend to indicate
"state or condition," "time," or "place/location." For example, consider the
PPs (in bold) in the sentence from one marked passage, that is illustrated as
a tree diagram in figure 4-3:
The lawn started at the beach and ran toward the front door for a quarter of a
mile jumping over sun-dials and brick walls and burning gardens—finally it
reaches the house drifting up the side in bright vines as though from the momen-
tum of its run. (Fitzgerald [1925]1991: 9)
All but the first PP clearly show process and manner. Compare these with the
PPs in a sentence from a nearby unmarked passage:
I never saw this great-uncle, but I am supposed to look like him—with special
reference to the rather hard-boiled painting [condition] that hangs in my fa-
ther's office [location].
Table 4-2 presents the statistics on the incidence of PPs. If the overall inci-
dence of PPs is considered (seventy-three in the marked passages compared to
eighty-one in the unmarked passages), the difference is not statistically signif-
icant. However, when we compare the passages in regard to specific types of
PPs, we see a clearly statistically significant difference. Many more PPs refer
to process, manner, and direction in the marked passages (N = 31) than in the
unmarked passages (N = 6). In contrast, the unmarked passages have many
more PPs referring to condition, time, and location (N = 50) than the marked
passages (N = 20).
Table 4-2
Comparisons: PPs, Class I PPs, Class II PPs
The Great Gatsby
Crucial/Marked Passages Unmarked Passages

Category N % N % X2
PPs 73 9.9 81 10.7 0.414

Class I PPs 31 4.2 6 0.79 * 16.89 (p<.001)
Class II PPs 20 2.7 50 6.6 *12.856 (p < .001)

percentage of word total; 1 df; .001 = the probability level; * = significant.
Verbal Categories
There are more verbs in the marked passages, seventy-three to fifty-eight,
although the difference is not statistically significant, as shown in Table 4-3.
What is more interesting, however, is that the marked passages show more
successive verbs or verbs in sequence. For example, consider the sentence
from a marked passage that begins with a sequence of verbs—The groups
change more swiftly, swell with new arrivals, dissolve and form ...'and goes
on to add more— confident girls who weave here and there . . . become . . .
the center of a group and then excited with triumph glide on . . .' (Fitzgerald
[1925]1991: 34).
Also, the marked passages show many more verbs which I call phrasal
verbs. True, it is difficult to define a phrasal verb absolutely. Rather than try
to do this, Hopper (1993), for example, simply refers to all potential candi-
dates as multiply articulated verbal expressions. For an example of the prob-
lem, consider the verb in this clause: which is dignified under the name of the
Table 4-3
Comparisons: Vs, Phrasal Verbs
The Great Gatsby
Crucial/Marked Passages Unmarked Passages

Category N % N % X2
Vs 73 9.9 58 7.7 1.716
Phrasal Verbs 26 3.5 7 0.9 "10.938 (p<.001)
Ratio (PVs to Vs) 1:2.8 1:8.28

"creative temperament" (Fitzgerald [1925]1991: 5-6). Is the verb simply is

dignified, or is it is dignified under? I define phrasal verbs in two ways. First,
the category includes those verbs followed by particles which cannot easily be
separated from the main verb by a parenthetical intrusion, (for example,
stretch out in this fragment from a marked passage: When we pulled out
into the winter night and the real snow, our snow, began to stretch out
beside MS ... (Fitzgerald [1925]1991: 137). Second, the category also in-
cludes those combinations of verb + prepositions in which the preposition does
not assign a thematic role on its own; that is, the thematic role of the comple-
ment is assigned only by the verb in concert with the preposition as satellite. For
example, by this definition, think of is a phrasal verb and not a verb + a
preposition heading a PP. Note that think of is equivalent to consider. This
phrasal verb occurs in this sentence in a marked passage: As I sat there,
brooding on the old unknown world, I thought of Gatsby's wonder . . .
(Fitzgerald [1925]1991: 141). However, dignified under in dignified under
the name is not a phrasal verb. Under these two criteria, of the seventy-three
verbs in the marked passages, thirty-six are phrasal verbs. This means that
one out of every two verbs is a phrasal verb. In contrast, in the unmarked
passages, out of fifty-eight verbs, only fourteen are phrasal verbs, or one
in four. This difference is statistically significant according to a X 2 test.
(See table 4-3.) Some examples of phrasal verbs from the marked passag-
es include turned out in Gatsby turned out all right at the end (Fitzgerald
[1925]1991: 6); and tipped out in Laughter is easier . . . spilled with
prodigality, tipped out at a cheerful word (Fitzgerald [1925]1991: 34).
What special effects do phrasal verbs have? Hopper (1993) remarked that
in contrast to simple transitive verbs, which suggest backgrounded or unwit-
nessed events, "multiply articulated verbal expressions" suggest an authorial
perspective on the action—that is, experienced or witnessed details. I would
add that because phrasal verbs may be considered as icons of an "unfold-
ing,"2 their presence is consistent with a view of narrative as at times con-
structive rather than just reportorial. If phrasal verbs carry such intentional
meanings, it would be predicted that more of them should occur when the
author is giving his or her perspective on events. This is just what is found in
the marked (crucial) passages, of course.
STRUCTURAL TREE DIAGRAMS
The four figures included in this chapter provide structural tree diagrams of
example sentences. Figures 4-1, 4-2 and 4-3 provide examples of structural
2. Janice L. Jake points out one potential effect of phrasal verbs is that the verb portion of
verb + satellite contributes its own referential meaning and also contributes to the meaning of the
entire phrasal verb. For example, in the phrasal verb turn out, turn on its own means "to change"
as in change a direction. In combination with out it implies not only a change but a result in which
the final outcome was not predictable from the outset.
If personality is an unbroken series of successful gestures, then there was something gorgeous
about him, some heightened sensitivity to the promises of life, as if be were related to one of
those intricate machines that register earthquakes ten thousand miles away.
Figure 4-1
Phrase Structure Tree for a Sentence from a Crucial/Marked Passage
The Great Gatsby (1991: 6)
The lights grow brighter as the earth lurches away from the sun and now the orchestra is playing
yellow cocktail music and the opera of voices pitches a key higher.
Figure 4-2
The Great Gatsby (1991:34)
trees of one sentence from a marked passage. Figure 4-4 gives an example of
the structural tree of a sentence from an unmarked passage. As noted earlier,
the marked passages have many elaborating CPs. These can be seen in figure
4-1, for example. Note these elaborating CPs: (i) if personality is an unbro-
ken series of successful gestures, (ii) as if he were related to one of those
intricate machines, (Hi) that register earthquakes ten thousand miles away.
(The main clause or root CP is there was something gorgeous about him.)
They also show many phrasal verbs. For example, in the sentence diagramed
in figure 4-2, lurch away occurs: The lights grow brighter as the earth lurch-
es away from the sun. . . .
DISCUSSION OF ELABORATING CPS AND CONJOINED IPS
Whereas differences in the types of PPs and the incidence of phrasal verbs
across the two types of passages warrant comment, what is most striking in
the comparison is the incidence of elaborating CP chaining and conjoined IP
constructions in the marked passages. When there are a number of elaborat-
ing CPs and IPs within a single sentence, the sentence is more "thoughtful."
But how is this accomplished? As background to an answer, consider that
sentences in general satisfy thematic (theta) role projections—they contain
The lawn started at the beach and ran toward the front door for a quarter of a mile, jumping over
sun-dials and brick walks and burning gardens—finally when it reached the house drifting up the
side in bright vines as though from the momentum of its run.
Figure 4-3
The Great Gatsby (1991:9)
74
i remember the fur coats of the girls returning from Miss This or That's and the chatter of frozen
breath and the hands waving overhead as we caught sight of old acquaintances and the matchings
of invitations: "Are you going to the Ordways'? the Herseys'? the Schultzes'?" and the long green
tickets clasped tight in our gloved hands.
Figure 4-4
Phrase Structure Tree for a Sentence from an Unmarked Passage
The Great Gatsby (1991: 136)
75
thematic role assigners (verbs or prepositions) and thematic role receivers

(NPs and some pronouns). Thematic roles include agent, patient, beneficia-
ry, locative, and others. Thus, any sentence includes a verb that will specify
internal thematic role requirements, and the necessary complement NPs
and, in some cases, PPs to satisfy the possible subcategorized thematic roles
projected by the verb and VP. Note that APs and many PPs do not satisfy
thematic role requirements. Rather, as adjuncts they elaborate on already
projected thematic roles. Therefore, they differ from NPs, which are proto-
typical complements.
The reason thematic roles concern us here is that processing the relation
between thematic role assigners (e.g., a verb) and thematic role receivers
(e.g., a noun) requires "cognitive energy." Thus, whenever there are nu-
merous thematic roles presented in the same sentence, this "slows down
the camera"—to rephrase Longacre's (1985) words about narrative peaks.
Because CPs and IPs express thematic role requirements by containing
both finite verbs and NPs, it should now be clear why the number of CPs
and IPs in the marked passages is important: The more CPs and IPs there
are, the more thematic role relations there are to process, even for elab-
orating CPs.
Having more than one CP or IP in a sentence increases cognitive weight
in three ways. First, some CPs and IPs are like NPs in that they themselves
can receive thematic roles. For example, as a sentential complement, a CP
may function as a patient (e.g., that the rain falls mainly on the plain in the
sentence [I think [that the rain falls mainly on the plain]]. Second, other
additional adjunct CPs that are not part of the argument projection of the
main verb (as the complement CP is in the sentence just discussed), "create"
more thematic roles. Thus, the CP when I came back from the East last
autumn includes thematic roles in addition to those in the main CP of the
sentence. That is, the elaborating CP includes the thematic roles of actor (/)
and direction/source [the East). Third, and perhaps most important, all
additional CPs and IPs necessarily create additional propositions, all with
their own cognitive weight.
Now, it is true that any multiplication of phrases, whether they are CPs
or IPs or NPs or PPs or even APs, creates new information and therefore
"slows down the camera," thus making the passage more "memorable."
The difference, however, is that the information created by CPs or IPs is of
a different order from that created by other phrases. This difference occurs
because, as noted above, new propositions are created. In the case of other
types of multiple phrases, the only thing added is new information. For
example, multiple adjunct PPs or adjectives simply offer an elaboration on
a proposition. In the case of multiple NPs, the elaboration is slightly dif-
ferent, but still it is only an elaboration. That is, in an NP-and-NP con-
struction, still only one thematic role projection is satisfied, even though
there are two nouns.
ANALYSIS OF "AS I LAY DYING"
I turn now to Faulkner's macabre novel, As I Lay Dying. This novel is the
story of a poor, white rural Mississippi family and their grisly odyssey while
bringing the dead body of the wife and mother of the family to town to be
buried next to her kin. As it progresses, the story of the journey is told
entirely through the monologues of the characters, most of whom are mem-
bers of the family.
Two monologues for each of eight characters were studied, seven family
members and the doctor, Peabody. The first monologue to appear in the
novel for each of these persons was analyzed. The second set of monologues
studied include the next one for each of the same characters that consisted
largely of narration. In some cases, only part of a monologue was studied to
keep all texts to approximately the same length and to treat narrative mate-
rial only.3 For the sake of brevity, I refer to all samples as monologues even
if they are only portions of monologues. Again, the hypothesis is that there
will be differences in syntactic structures across the passages studied (in this
case, across the monologues as presentations of the characters).
To establish a baseline to serve as Faulkner's unmarked style at the time,
passages from Light in August, written about the same time, were also ana-
lyzed. However, results show that the use of CPs, IPs, and VPs by most of
the characters in As I Lay Dying is fairly similar to usage in narrative pas-
sages in Light in August. Thus, there is nothing "marked" about As I Lay
Dying in regard to these features. Further, because there is no clear dif-
ference across characters in As I Lay Dying in regard to these features,
any differences in characterization to be attributed to syntactic features
would have to involve structures other than the relative incidence of CPs,
IPs, and VPs.
Trailing Constitutents
Even in a "holistic" reading of the first monologues of the eight characters,
the reader is struck by a feeling that Darl and, to a lesser extent, Vardaman,
are given styles that set them apart from the other characters. Darl is the
second oldest son and the hero of the novel—if there is one—whereas Vard-
aman is the youngest child and considered mentally "odd," even at the begin-
ning of the novel. The difference in their style seems to have something to do
with the extent to which their monologues (especially Darl's) achieved a
similar length to the other monologues only by including many phrases as
optional additions to the main verb and its arguments.
3. The two sets of passages studied from As / Lay Dying (Library of America edition 1995)
are the following: Sample I: Addie (p. 114), Anse (p. 24), Cash (pp. 157-158), Darl (p. 1), Dewey
Dell (p. 18), Jewel (p. 11), Vardaman (p. 36), and Peabody (p. 28). Sample II: Addie (pp. 117-
118), Anse (p. 71), Cash (pp. 177-178), Darl (pp. 149-150), Dewey Dell (pp. 78-79), Vardaman
(pp. 170-171), and Peabody (pp. 29-30).
Thus, my more specific hypothesis is the following:

Darl's monologues would show many more of what I call trailing constituents
than the monologues of the other characters. Also, the monologue of Varda-
man would pattern more like that of Darl than the other characters in regard to
these constituents.
As noted previously, trailing constituents is used as a cover term for what
some other writers call detached participial phrases (Thompson 1983) and
also for certain types of prepositional phrases. Thompson defines detached
participial phrases in conversation in reference to intonation patterns. She
refers to them as "being set off by pauses, of exhibiting a clause-final falling
intonation contour characteristic of independent clauses, or of being preced-
ed by a clause ending with a clause-final falling intonation." She goes on to
say that "these intonational signals of detachment are virtually without ex-
ception marked by commas in writing." (Thompson 1983: 43) Although
Thompson's definition is useful, I used the following ways to define them.
First, detached participial phrases are distinct from PPs; thus, they cannot
begin with a preposition. They are also distinct from reduced relative clauses.
By these two criteria, with the water bubbling up in It would be quiet there
then, with the water bubbling up . . . . (Faulkner [1930]1985: 114) is not a
detached participial phrase. Second, they can begin with a participial verb,
which, of course, is tenseless. For example, consider this participial phrase
from Darl's second monologue: This time Jewel is riding upon it, clinging to
it...' (Faulkner [1930]1985:150) Third, they may begin with a NP as subject
of the participle, although these are very rare in my sample. Sometimes this
subject is coreferential with a main-clause nonsubject NP, but more often it
is in a part-whole relationship with the main-clause subject. This example
occurs in Darl's first monologue, with the participial clause beginning with a
single broad window (in caps): Square, with a broken roof set at a single
pitch, it leans in empty and shimmering dilapidation in the sunlight, A SIN-
GLE BROAD WINDOW IN TWO OPPOSITE WALLS GIVING ON TO
THE APPROACHES OF THE PATH (Faulkner [1930]1985: 1).
If the trailing constituents are highlighted in the text, it is easy to see how
many of these types of constituents occur in Darl's monologue . (Detached
participial phrases are in caps and the prepositional phrases that show pro-
cess, manner, or direction are in bold):
Jewel and I come up from the field, FOLLOWING THE PATH IN SINGLE
FILE. Although I am fifteen feet ahead of him, anyone watching us from
the cottonhouse can see Jewel's frayed and broken straw hat a full head
above my own.
The path runs straight as a plum-line, WORN SMOOTH BY FEET AND
BAKED BRICK-HARD BY JULY, between the green rows of laid-by cotton, to
the cottonhouse in the center of the field, where it turns and circles the cotton-
house at four soft right angles and goes on across the field again, WORN SO BY
FEET in fading precision.
The cottonhouse is of rough logs, from between which the chinking has
long fallen. Square, with a broken roof set at a single pitch, it leans in empty and
shimmering dilapidation in the sunlight, A SINGLE BROAD WINDOW IN
TWO OPPOSITE WALLS GIVING ON TO THE APPROACHES OF THE
PATH. Jewel, fifteen feet behind me, LOOKING STRAIGHT AHEAD, HIS
PALE EYES LIKE WOOD SET INTO HIS WOODEN FACE, he crosses the
floor in four strides with the rigid gravity of a cigar-store Indian dressed in patch
overalls and endued with life from the hips down, and steps in a single stride
through the opposite window and into the path again just as I come around the
corner. In single file and five feet apart and Jewel now in front, we go on up the
path toward the foot of the bluff. Tull's wagon stands beside the spring,
HITCHED TO THE RAIL, THE REINS WRAPPED ABOUT THE SEAT STAN-
CHION. In the wagon-bed are two chairs. Jewel stops at the spring and takes
the gourd from the willow branch and drinks. I pass him and mount the path,
BEGINNING TO HEAR CASH'S SAW. When I reach the top he has quit
sawing. STANDING IN A LITTER OF CHIPS, he is fitting two of the boards
together. (Faulkner [1930]1985: 1).
Compare Darl's monologue with the first five sentences of Vardaman's
monologue and similar samples from other characters, Cash and Dewey Dell:
If I jump I can go through it like the pink lady in the circus, into the warm
smelling, without having to wait. My hands grab at the bushes; beneath my feet
the rocks and dirt go rumbling down. Then I can breathe again, in the warm
smelling. I enter the stall, TRYING TO TOUCH HIM, and then I can dry then
I vomit the crying. As soon as he gets through kicking I can and then I can cry,
the crying can. (Faulkner [1930]1985: 36)
Cash:
So Jewel got the team and come for me and they fixed me a pallet in the wagon
and we drove across the square to the corner where Pa said, and we was waiting
there in the wagon, with Dewey Dell and Vardaman eating bananas, when we
see them coming up the street. Pa was coming along with that kind of daresome
and hangdog look all at once like when he has been up to something he knows
ma ain't going to like, CARRYING A GRIP IN HIS HAND . . . (Faulkner
[1930]1985: 177)
Dewey Dell:
We picked on down the row, the woods getting closer and closer and the secret
shade, PICKING ON INTO THE SECRET SHADE with my sack and Lafe's
sack. Because I said will I or wont I when the sack was half-full because I said
if the sack is full when we get to the woods it wont be me. I said if it don't mean
for me to do it the sack will not be full and I will turn up the next row but if the
sack is full, I cannot help it. (Faulkner [1930]1985: 18)
Darl's Monologues as Distinctive

A comparison across all monologues (ranging from a total of 562 to 796
words when both samples are considered) shows that Darl is very distinctive
Table 4-4
Comparisons: Detached Participial Phrases (Samples I & II)
As I Lay Dying
Character Words N %
Addie 796 3 0.4

Anse 562 3 0.5
Cash 716 5 0.7
Darl 743 22 3.0
Dewey Dell 724 4 0.6
Vardaman 747 10 1.3
Peabody 730 4 0.5
2 YC
X = *34.907 (p <,.001)

percentage of word total; 6 df; .001 = the probability level; YC= Yates' Correction
applied; * - significant.
in his use of detached participial phrases. He has twenty-two, eleven in each

of his monologues. Vardaman has the next largest number of these constitu-
ents, a total of ten. There are four detached participial phrases in the mono-
logue from which the extract of Vardaman, cited previously, comes; one
occurs in the extract. Cash has one detached participial phrase in the portion
of the monologue quoted previously; altogether he has five of these constit-
uents in both of his samples. In the extract from Dewey Dell's monologue,
there is one detached participial phrase; in both her monologues she has a
total of four. In this regard, the monologues of Cash and Dewey Dell are
Table 4-5
Comparisons: All PPs (Samples I & II)
As I Lay Dying
Class I PPs (process Class II PPs (condition,

manner, direction) time, location)
Character Words N % N %
Addie 796 17 2.1 44 5.5

Anse 562 9 1.6 26 4.6
Cash 716 18 2.5 18 2.5
Darl 743 51 6.9 41 5.5
Dewey Dell 724 28 3.9 16 2.2
Vardaman 747 32 4.3 25 3.3
Peabody 730 7 1.0 34 4.7
x2 = *60.617 (p <.001) *24.32 (p < .001)

typical of those of the other characters and in stark contrast with those of
Darl and, to a lesser extent, those of Vardaman.
Overall statistics for all characters showing their use of detached par-
ticipial phrases are found in table 4-4. With his twenty-two examples,
Darl shows 3.0 detached participial phrases for every 100 words. For
Vardaman, with his ten examples, the ratio is 1.3 per 100 words. Cash has
five examples, but all other characters show only three or four, or fewer than
0.5 or 0.6 per 100 words. Note that these differences are significant accord-
ing to a x2 analysis.
In addition to showing a high usage of detached participial phrases, Darl
also shows a high frequency and ratio of both class I (manner, process, and
direction) and class II (condition, time, and location) prepositional phrases
compared to the other characters. These are in boldface in Darl's monologue
included above. Darl has fifty-one PPs which encode process, manner, or
direction. The high incidence of such PPs definitely sets him apart from any
other character, giving him a ratio of 6.9 PPs of this type per 100 words.
Vardaman follows with thirty-two PPs of this class and a ratio of 4.3 PPs
of this type per 100 words. Darl is less distinctive in his use of class II-
type PPs (i.e., PPs which show condition, time, or location). He has forty-
one PPs in class II, but Addie has forty-four and Peabody has thirty-four.
Still, overall, the distribution of findings across all eight characters is signif-
icant for both types of PPs. (See table 4-5.) Again, the results are statistically
significant.
In addition to showing a high usage of both of the types of structures for
which I use the cover term trailing constituents, an analysis of only sample
I shows that Darl has many more dynamic verbs, more nouns, and more
descriptive adjectives than other characters. He has fewer elaborating CPs
that other characters and fewer coordinated phrases (N+N or V+V, etc.).
However, reporting the details on these categories is beyond the scope of
this chapter.
INTERPRETATION OF RESULTS FOR "AS I LAY DYING"
So how does this grammatical profile build Darl's character? Impression tells
us that Darl is dream-like—fanciful at the same time he is pensive. In con-
trast, the other characters are more commonplace, as would fit our stereo-
types of the rural, white, southern lower class. In her quantitative study of
detached participles Thompson (1983) commented on the relative scarcity of
this construction in conversation and stated that it is more typical of planned
texts and therefore more often found in writing than speech. She argues that
the purpose of the detached participle is to describe, not state logical or
temporal relations. When the detached participle occurs, Thompson charac-
terizes such a text as "discourse that attempts to describe by creating an
image" (1983:46). She goes on to say: "Writing that tends to spark the visual
imagination, then, tends to abound in detached participles" (1983: 51). Else-

where, she states that the detached participle "serves as a device that allows
the speaker/writer to present certain material as background against which
other material can be put forth as 'figure' in the Gestalt sense" (Thompson
1983:44).
In many ways, the typical monologue in this novel is similar to conversa-
tion, and the contrast between conversation and written text is a clue to the
evocative effect of Darl's syntax. Resembling conversation, the monologues
of most of the other characters seem to be directed at an imaginary listener.
These monologues impart facts in the relatively straightforward way of hav-
ing sentences largely composed of IPs and CPs. The main task of these types
of constituents, as indicated in the earlier discussion of Gatsby, is to intro-
duce new propositions. They do this, of course, by virtue of the fact that the
heart of such a constituent is a finite verb and NPs in the primary argument
positions of subject and object. These NPs fulfill the thematic roles of
agents and patients or goals and beneficiaries. However, Darl's mono-
logue does not show this emphasis on argument structure, perhaps
because his character seems less concerned with presenting informa-
tion than with painting a scene. Although Darl's first monologue has more
concrete nouns than that of any other character, most of these nouns are not
part of the argument structure of the main clause or of elaborating CPs. They
occur as thematic roles of either participials or PPs. Because these trailing
constituents are not part of the main argument structure (both participials
and PPs are adjuncts, not complements of the main VP in a CP), they do not
modify the propositional content of the sentence but simply provide back-
ground for what is already conveyed. In this way they contrast with comple-
ments in the main VP, that is, with NPs as objects.
This lack of contribution to the main propositional weight of the sentence,
this lack of attachment to finite verbs, means that some subject NPs in Darl's
monologue (i.e., those in detached participial phrases) are not really an-
chored in any sense. They are left floating, if you will. Thus, the argument
here is that it is this lack of relationship between the nouns and verbs (i.e.,
tenseless participles) in detached participial phrases and the main predicate-
argument structure of the sentence that gives Darl's monologues their dream-
like and ethereal quality. To a lesser extent, the monologues of Vardaman
also have this quality.
Of course other characters also have a few detached participials, but the
point is that they are just that: a few. When they do occur, the effect is similar
to that found in Darl's monologues. For example, the first monologue of the
father, Anse, has several detached participials in the first few lines:
Durn that road, and it fixing to rain, too. I can stand here and same as see it with
second-sight, A-SHUTTING DOWN BEHIND THEM LIKE A WALL, SHUT-
TING DOWN BETWIXT THEM AND MY GIVEN PROMISE. I do the best
I can, much as I can get my mind on anything, but durn them boys. (Faulkner
[1930]1985:24)
Although in many ways Anse is not a sympathetic character, these detached
participials give him—just for a second—the same dream-like quality Darl has
in general. For once, he evokes images rather than confining his contributions
to the pedestrian. At the least, they add a reflective dimension to his portrayal.
CONCLUSION
This chapter provides evidence that in Fitzgerald's The Great Gatsby five
passages identified as crucial in carrying the "authorial message" differ in
certain syntactic ways from five matched passages which largely function
only to carry the story line forward. This chapter also provides evidence that
the syntax of Darl in Faulkner's As I Lay Dying is marked in relation to the
syntax of other characters in showing many more trailing constituents. Most
especially, Darl's monologues have many detached participial phrases.
Interestingly enough, the sources of stylistic effect in the two novels in
reference to the structures studied are exactly opposite each other. In Gatsby
the author slows down the action by adding "cognitive weight" to sentences
in the crucial passages; he does this by adding additional propositions to the
main proposition of the sentence. This is done by adding elaborating CPs
containing finite verbs and their arguments (NPs). These types of structure
require additional cognitive processing because the reader needs to work out
the relationship between the verbs and the thematic roles contained in their
arguments, roles such as agent or patient. In As / Lay Dying the author builds
up the reader's picture of Darl by giving him monologues that are not packed
with "extra" propositional modification of the central proposition (i.e., core
CP) to process. Instead, Darl's monologues contain constituents that do not
have finite verbs (detached participial phrases) or have no verbs at all (prep-
ositional phrases). The abundance of these constituents in Darl's monologues
contrasts with what is found in the monologues of the other characters.
Reading such constituents provides the reader with images or impressions
rather than information. The result is that the cognitive processing in terms
of thematic content and the mapping of propositional content required of the
reader is reduced. The net stylistic effect is that Darl himself comes across as
detached, not so much a part of the everyday world of the cognitive process-
ing of actor and event as the other characters but a reflective observer.
These conclusions result from a frequency count of the major components
of phrase structure in paired discourse samples from the two novels, elabo-
rating CPs in the case of Gatsby and detached participials in the case of As
I Lay Dying. To reach the most interesting claim of the analysis in both
novels—that stylistic effects are achieved through varying the propositional
weight—one must go beyond surface-level considerations of style. This
analysis emphasizes the role of the more abstract level of predicate-argument

structure. Nouns and verbs and their relations, including thematic role as-
signment, are what affect propositional weight. Thus, this chapter argues
for quantitative analysis in stylistic research, but quantification that goes
beyond counting lexical categories and phrase structures only as surface
configurations.
Finally, because a goal of this chapter was to show that linguistic virtuosos
achieve certain effects by using marked choices, some discussion of the rele-
vant structures as marked choices is in order. It has been argued that many
linguistic choices are motivated by the wish of the speaker/author to convey
messages which, in addition to conveying the referential message of words
and their syntactic configurations, suggest messages that are intentional in
nature. Elsewhere I commented on the fact that messages of intentionality are
social or psychological in nature (e.g., Myers-Scotton 1993). With the marked-
ness model, marked choices function as negotiations to change either the
speaker's previously established persona or the interpersonal relationship
with other participants. Is it a stretch of the imagination to claim that marked
choices novelists make are socially or psychologically motivated? To answer
this question fully is beyond the scope of this chapter. Here, suffice it to say
that for novelists, marked choices, at the very least, draw attention to them-
selves. In so doing, they clearly signal a higher degree of authorial involve-
ment at the point at which they occur. Can one argue that when marked
choices are made, the story line becomes secondary to "something else"? I
would answer "yes." I suggest that marked stylistic choices in novels are
negotiations to change the writer's relationship with the reader and perhaps
also the reader's relationship with a character. In making marked choices,
writers are inviting the reader to look for "something else" beyond the refer-
ential message—in the parlance of Grice (1975) and Sperber and Wilson
(1995), to look for implicatures. Marked choices invite the reader to partic-
ipate in a new way, to engage in the cognitive processing necessary for "working
out" additional meaning. It is in this sense that one can argue that marked
choices in a novel are also socially or psychologically motivated.
ACKNOWLEDGMENT: I am extremely grateful to Longxing Wei for a good deal of help in

quantifying features for this analysis.
REFERENCES
Faulkner, William. 1930. As I lay dying. New York: J. Cape and H. Smith. Reprint,
New York: Literary Classics of the United States, 1985.
. 1932. Light in August. New York: Modern Library. Reprint, New York:
Literary Classics of the United States, 1985.
Fitzgerald, F. Scott. 1925. The Great Gatsby. New York: Charles Scribner's Sons.
Reprint, Cambridge: Cambridge University Press, 1991, ed. Matthew H. Bruccoli.
Freeman, Donald C. 1996. "According to my bond": King Lear and re-cognition. In
The stylistics reader: from Roman Jakobson to the present, ed. Jean Jacques We-
ber, 280-297. New York: St. Martin's Press.
Garvin, Paul. 1964. A Prague reader on esthetics, literary structure and style. Wash-
ington, D.C.: Georgetown University Press.
Grice, H. Paul. 1975. Logic and conversation. In Syntax and semantics, vol. 3, ed.
Peter Cole and Jerry L. Morgan, 41-58. New York: Academic Press.
Hopper, Paul. 1993. The English verbal expression in written discourse. Paper
presented at Rice University symposium on alternative linguistics in Houston,
Texas.
Kintsch, Walter. 1974. The representation of meaning in memory. Hillsdale, N. J.:
Erlbaum.
Lakoff, George. 1987. Women, fire, and dangerous things. Chicago: University of
Chicago Press.
Longacre, Robert E. 1985. Discourse peak as zone of turbulence. In Beyond the
sentence: discourse and sentential form, ed. Jessica Wirth, 83-98. Ann Arbor,
Mich.: Karoma.
Myers-Scotton, Carol. 1993. Social motivations for codeswitching: evidence from
Africa. Oxford: Oxford University Press.
Sperber, Dan, and Deirdre Wilson. 1995. Relevance: cognition and communication.
Thompson, Sandra A. 1983. Grammar and discourse: the English detached participial
clause. In Discourse perspectives on syntax, ed. Flora Klein-Andreu. 43-65. New
York: Academic Press.
Van Peer, Willie. 1986. Stylistics and psychology, investigations in foregrounding.
Wolfeboro N.H.: Croom Helm.
APPENDIX
PASSAGES STUDIED FROM "THE GREAT GATSBY"
marked ([1925]1991: 5-6)

When I came back from the East last autumn I felt that I wanted the world to be
in uniform and at a sort of moral attention forever; I wanted no more riotous excur-
sions with privileged glimpses into the human heart. Only Gatsby, the man who gives
his name to this book, was exempt from my reaction—Gatsby who represented every-
thing for which I have an unaffected scorn. If personality is an unbroken series of
successful gestures, then there was something gorgeous about him, some heightened
sensitivity to the promises of life, as if he were related to one of those intricate ma-
chines that register earthquakes ten thousand miles away. This responsiveness had
nothing to do with that flabby impressionability which is dignified under the name of
the "creative temperament"—it was an extraordinary gift for hope, a romantic readi-
ness such as I have never found in any other person and which it is not likely I shall ever
find again. No—Gatsby turned out all right at the end; it is what preyed on Gatsby,
what foul dust floated in the wake of his dreams that temporarily closed out my
interest in the abortive sorrows and short-winded elations of men.
marked ([1925]1991: 9)
And so it happened that on a warm windy evening I drove over to East Egg to see
two old friends whom I scarcely knew at all. Their house was even more elaborate
than I expected, a cheerful red and white Georgian Colonial mansion overlooking the
bay. The lawn started at the beach and ran toward the front door for a quarter of a
mile, jumping over sun-dials and brick walks and burning gardens—finally when it
reached the house drifting up the side in bright vines as though from the momentum
of its run. The front was broken by a line of French windows, glowing now with
reflected gold, and wide open to the warm windy afternoon, and Tom Buchanan in
riding clothes was standing with his legs apart on the front porch.
marked ([1925]1991: 34)

The lights grow brighter as the earth lurches away from the sun and now the
orchestra is playing yellow cocktail music and the opera of voices pitches a key higher.
Laughter is easier, minute by minute, spilled with prodigality, tipped out at a cheerful
word. The groups change more swiftly, swell with new arrivals, dissolve and form in
the same breath—already there are wanderers, confident girls who weave here and
there among the stouter and more stable, become for a sharp, joyous moment the
center of a group and then excited with triumph glide on through the sea-change of
faces and voices and color under the constantly changing light.
Suddenly one of these gypsies in trembling opal seizes a cocktail out of the air,
dumps it down for courage and moving her hands like Frisco dances out alone on the
canvas platform. A momentary hush; the orchestra leader varies his rhythm obligingly
for her and there is a burst of chatter as the erroneous news goes around that she is
Gilda Gray's understudy from the "Follies." The party has begun.
marked ([1925]1991: 137)

When we pulled out into the winter night and the real snow, our snow, began to
stretch out beside us and twinkle against the windows, and the dim lights of small
Wisconsin stations moved by, a sharp wild brace came suddenly into the air. We drew
in deep breaths of it as we walked back from dinner through the cold vestibules,
unutterably aware of our identity with this country for one strange hour before we
melted indistinguishably into it again.
marked ([1925]1991: 141)

And as I sat there, brooding on the old unknown world, I thought of Gatsby's wonder
when he first picked out the green light at the end of Daisy's dock. He had come a long
way to this blue lawn and his dream must have seemed so close that he could hardly fail
to grasp it. He did not know that it was already behind him, somewhere back in that vast
obscurity beyond the city, where the dark fields of the republic rolled on under the night.
Gatsby believed in the green light, the orgastic future that year by year recedes
before us. It eluded us then, but that's no matter—tomorrow we will run faster, stretch
out our arms farther . . . . And one fine morning—
So we beat on, boats against the current, borne back ceaselessly into the past.
unmarked ([1925] 1991: 6)

I never saw this great-uncle but I'm supposed to look like him—with special ref-
erence to the rather hard-boiled painting that hangs in Father's office. I graduated
from New Haven in 1915, just a quarter of a century after my father, and a little later
I participated in that delayed Teutonic migration known as the Great War. I enjoined
the counter-raid so thoroughly that I came back restless. Instead of being the warm
center of the world the middle-west now seems like the ragged edge of the uni-
verse—so I decided to go east and learn the bond business. Everybody I knew was
in the bond business so I supposed it could support one more single man. All my
aunts and uncles talked it over as if they were choosing a prep-school for me and
finally said "Why—ye-es" with very grave, hesitant faces. Father agreed to finance
me for a year and after various delays I came east, permanently, I thought, in the spring
of twenty-two.
unmarked ([1925]1991: 8)
I lived at West Egg, the—well, the less fashionable of the two, though this is a
most superficial tag to express the bizarre and not a little sinister contrast between
them. My house was at the very tip of the egg, only fifty yards from the Sound, and
squeezed between two huge places that rented for twelve or fifteen thousand a sea-
son. The one on my right was a colossal affair by any standard—it was a factual
imitation of some Hotel de Ville in Normandy, with a tower on one side, spanking
new under a thin beard of raw ivy, and a marble swimming pool and more than forty
acres of lawn and garden. It was Gatsby's mansion. Or rather, as I didn't know Mr.
Gatsby it was a mansion inhabited by a gentleman of that name. My own house was
an eye-sore, but it was a small eye-sore and it had been overlooked, so I had a view
of the water, a partial view of my neighbor's lawn and the consoling proximity of
millionaires—all for eighty dollars a month.
unmarked ([1925]1991: 33)

Every Friday five crates of oranges and lemons arrived from a fruiterer in New
York—every Monday these same oranges and lemons left his back door in a pyramid
of pulpless halves. There was a machine in the kitchen which could extract the juice
of two hundred oranges in half an hour, if a little button was pressed two hundred
times by a butler's thumb.
At least once a fortnight a corps of caterers came down with several hundred feet
of canvas and enough colored lights to make a Christmas tree of Gatsby's enormous
garden. On buffer tables, garnished with glistening hors d'oeuvre, spiced baked hams
crowded against salads of harlequin designs and pastry pigs and turkeys bewitched to
a dark gold. In the main hall a bar with a real brass rail was set up, and stocked with
gins and liquors and with cordials so long forgotten that most of his female guests
were too young to know one from another.
unmarked ([1925]1991: 136)

One of my most vivid memories is of coming back west from prep school and later
from college at Christmas time. Those who went farther than Chicago would gather
in the old dim Union Station at six o'clock of a December evening with a few Chicago
friends already caught up into their own holiday gaieties to bid them a hasty goodbye.
I remember the fur coats of the girls returning from Miss This or That's and the chatter
of frozen breath and the hands waving overhead as we caught sight of old acquaintan-
ces and the matchings of invitations: "Are you going to the Ordways'? the Herseys'?
the Schultzes'?" and the long green tickets clasped tight in our gloved hands. And last
the murky yellow cars of the Chicago, Milwaukee & St. Paul Railroad looking cheer-
ful as Christmas itself on the tracks beside the gate.
unmarked ([1925]1991: 140)

I spent my Saturday nights in New York because those gleaming, dazzling parties
of his were with me so vividly that I could still hear the music and the laughter faint
and incessant from his garden and the cars going up and down his drive. One night I
did hear a material car there and saw its lights stop at his front steps. But I didn't
investigate. Probably it was some final guest who had been away at the ends of the
earth and didn't know that the party was over.
5
Markedness and References to Characters

in Biblical Hebrew Narratives
T I M O T H Y WILT
S EVERAL YEARS A G O , a translation team and I considered how to deal

with passages such as the following from the prophet Haggai, in the
Hebrew Scriptures:
I will fill this house with splendor, says the LORD of hosts. The silver is mine,
and the gold is mine, says the LORD of hosts. The latter splendor of this house
shall be greater than the former, says the LORD of hosts; and in this place I will
give prosperity, says the LORD of hosts.1
In accordance with a common approach to translation, 2 I suggested that the
team might want to reduce the number of says the LORD of hosts if such
repetition would not occur in their language, as it would not, so I said, in
contemporary English. T w o days later, I came across the following in a Kurt
Vonnegut novel:
"As far as I'm concerned," said Constant, "the universe is a junk yard, with
everything in it overpriced. I am through poking around in the junk heaps,
looking for bargains. Every so-called bargain," said Constant, "has been con-
nected by fine wires to a dynamite bouquet." He spat again.
"I resign," said Constant.
"I withdraw," said Constant.
"I quit," said Constant. (Vonnegut 1950: 21)'
1. Haggai 2:7b-9. This and all other quotes from the Hebrew Scriptures are, unless otherwise
indicated, cited according to the translation of the New Revised Standard Version (1993).
2. The approach to translation as developed by Eugene Nida and his followers (Nida and Tabor
1976:197). Lawrence Venuti, although opposed to Nida's approach, refers to Nida as "prolific and
influential" and "typical of other theorists in the Anglo-American tradition" (1995: 21).
3. A similar repetition in the speech formula occurs on page 297, where Rumfoord is the
speaker.
89
I recalled the same pattern when later my eight-year-old daughter said, "Sup-
per's ready, Dad. Come to the table, Dad. You sit here, Dad."4
As well as proving the inaccuracy of my statement concerning what "would
not occur in contemporary English," the texts from Vonnegut and my daugh-
ter also point to a serious flaw in the approach to translation that I was
following in my discussion with the translation team. Emphasis on the im-
portance of using "natural" style in translation is untempered by sufficient
attention to the fact that sometimes natural style involves the use of marked
choices. It also overlooks the function of these marked choices and how
marked choices may be represented in translation.
In this chapter, I consider some other examples of repetition in the He-
brew scriptures that represent marked choices by the narrator. Translators
tend to treat them as ancient Hebrew "style," failing to recognize that in
terms of comparable ancient Hebrew texts they are marked. Consequently,
they have paid insufficient attention to their communicative function. In the
cases to be discussed, the marked structures function to enhance humor,
irony, and/or thematic development. Suggestions are made for using marked
structures with equivalent communicative functions in translations.
MARKED REFERENCES TO CHARACTERS
A problem that leaps out at the reader of traditional Bible translations and
troubles the translator coming at it anew is the extensive repetition of names
and titles such as that found in the references to leaders of Israel's enemies in
Judges 3:8. The reference to the first of these leaders, Cushan-risthathaim, is
typical of this section:
The Lord sold them into the hand of Cushan-rishathaim King of Aram-naha-
raim; and the Israelites served Cushan-rishathaim eight years. But... The Lord
raised up a deliverer. . . Othniel son of Kenaz, Caleb's younger brother. The
spirit of The Lord came upon him, and he judged Israel; he went out to war, and
The Lord gave Cushan-rishathaim King of Aram into his hand; and his hand
prevailed over Cushan-rishathaim. (Judges 3:8-10)
The enemy king is referred to four times. Each time his proper name is used,
twice with his title and origin. A pronoun could have been used at least in the
second occurrence and in such a short extent of text the second King of Aram
is mnemonically unnecessary. The references to the enemy king contrast
with those to Othniel where the proper name is followed by a series of
unambiguous pronouns.
Similar repetition is found in Genesis 29:10:
4. She was smiling as she employed this highly marked repetition. I was dining with the family
for the first time after a work trip, so the repetition may have been to underline my familial role.
It also comes at a time when the preferred papa is being replaced by dad due to the influence of
schoolmates in a new cultural setting.
Hebrew Biblical Narratives 91
Table 5-1
References to Characters in Judges 3:8
More Highly Marked

Unmarked Marked Proper Name +
Pronoun Proper Name Relational Phrase
Israel's heroes 58 (68%) 20 (24%) 7 (8%)

Enemy leaders 15 (42%) 10 (28%) 11 (31%)
When Jacob saw Rachel, the daughter of Laban the brother of his mother and
the flock of Laban the brother of his mother, Jacob went to the well, took the
stone off of it, and watered the flock of Laban the brother of his mother.
The narrator employs marked structures to refer to Cushan-rishathaim
and to Laban. Assuming referential unambiguity, the unmarked structure, in
any language, employed to refer to characters after their initial introduction
is that of the pronoun, 5 as used in the references to Othniel, cited earlier. In
contexts in which the reference of a pronoun could be ambiguous, the un-
marked structure would be the use of the proper name of a character or a
relational phrase such as King of Aram or brother of his mother. In both
instances, the use of the proper name of a character plus a relational phrase
is marked: It occurs less frequently and in fewer linguistic environments and
it is structurally more complex. The following sections elaborate on this
observation and attempt to explain why the marked structures are used.
MARKED REFERENCES TO THE "ENEMY X KING OF Y"
Taking frequency as a rough indication of markedness, we compare the ref-

erences to the leaders of Israel's enemies with the references to the heroes who
delivered Israel, in Judges 3:8.6 (See table 5-1.)
In reference to Israel's heroes, the unmarked, pronominal structure occurs
about twice as frequently as the combined total of marked structures. But in
reference to the enemy leaders, the unmarked, pronominal structure is used less
often than marked structures. Particularly striking is the difference between
the use of the most highly marked structure (proper name + relational phrase):
It is relatively infrequent in reference to the heroes (8% or 7/85) but almost as
frequent as pronouns (31% or 11/36) in reference to the enemy leaders.
If the references via proper name + relational phrase to the enemy
leaders were spread throughout the accounts and used for the purpose of
5. Givon (1991: 257) and Dressier (1992: 12).

6. Pronouns are only counted once if occurring within the same clause (as proper names will
occur only once per clause). In a structure such as "And X answered and [pronoun] said," the
pronoun (verbal prefix) is not counted because a noun phrase rarely occurs in this context.
References in direct speech are not counted.
disambiguating, these frequency counts would not be so interesting. But this

is not the case: There is a clustering of their proper names which clearly
stands out in contrast to the references to the heroes. Although there are more
than twice as many references to the heroes than to the enemy leaders, a
hero's name is repeated within a clause or in immediately adjoining clauses
only once (Judges 4:9-10), but each of the enemy leaders' names is repeated
at least once in this environment:
The Lord sold them into the hand of Cushan-rishathaim King of Aram-naha-
raim; and the Israelites served Cushan-rishathaim eight years.... The Lord gave
Cushan-rishathaim King of Aram into his hand; and his hand prevailed over
Cushan-rishathaim. (Judges 3:8, 10)
Then he [Ehud] presented the tribute to King Eglon of Moab. Now Eglon was
a very fat man. (Judges 3:17)
God subdued, on that day, Jabin King of Canaan before the Israelites. The hand
of the Israelites bore harder and harder on Jabin King of Canaan, until they
destroyed Jabin King of Canaan. (Judges 4:23-24)
They [the men of Ephraim] captured the two captains of Midian, Oreb and
Zeeb; they killed Oreb at the rock of Oreb, and Zeeb they killed at the wine
press of Zeeb, as they pursued the Midianites. They brought the heads of Oreb
and Zeeb to Gideon beyond the Jordan. (Judges 7:25)
[Gideon] said, "Here are Zebah and Zalmunna, about whom you taunted me,
saying 'Do you already have in your possession the hands of Zebah and Zal-
munna'. . . ?" (Judges 8:15)7
Whereas this dense repetition may occur in varying parts of the battle
episodes, each account, apart from the one concerning Eglon, contains the
dense repetition in the closure as well. This further marks the references; it is
unparalleled in all other accounts of Israel's conquest of Canaan. Of all the
battle accounts, from the exodus to the appointment of Israel's own king,
there are none but these, which repeatedly refer to the enemy leaders' names
in the closure. Repetition of a king's name in adjoining clauses occurs only in
one other narrative (Numbers 21:23). Repetition of the structure [Personal
Name + king of + Territorial Name], as in Jabin King of Canaan, occurs in no
other accounts, let alone in adjoining clauses.8
7. This quote refers to Judges 8:5-6 where the repetition of the couplet of proper names is
separated from the preceding one by only the narrative-tensed formula introducing the direct
speech. In 8:12 and 8:21, Zebah and Zalmunna are also repeated in very closely adjoined clauses;
in each case, they could have been used unambiguously in the second occurrence.
8. References to enemy leaders in battle accounts are made in the following passages, aster-
isks indicating that the name of the leader is given: Exodus 14* (Pharaoh is treated like a proper
name); Numbers 21:1-3, 21-31*, 33-35*; 31:1-12*; Joshua 2:2-3; 8:1-29; 10:1-27*, 28, 29-
30, 33*, 36-37, 38-39; 11:1-14*, 16-20; Judges 1:5-7*; 3:7-11*, 12-25*, 4:2-24*, 7:25*,
8:5-21*; 11:12-33.
Comparison of reference strategies can also be made with accounts of encounters of enemies
from within such as Korah (Numbers 16) and Achan (Joshua 7).
Having observed these highly marked references to the characters, we

must now ask why they are so marked. In general terms, marking can be done
primarily for structural or for thematic purposes. The observations in the
preceding paragraphs indicate that structural marking (such as that used to
introduce a new setting, a new character, the peak, or closure of an account)
is not involved here,9 so we look for thematic significance.
One key aspect of these prominently marked names is that they all argu-
ably have a meaning related to the thematic development of the section in
Judges that follows the report of Joshua's death. The basic perspective is
given in the exposition of Judges 2:11-3:5: "abandonment of The Lord re-
sulted in affliction by the enemies 'left to test all those in Israel' but crying out
to The Lord resulted in deliverance." This perspective is then illustrated in the
narratives concerning the enemy leaders. The marked use of their names sug-
gests that the sense of their names contributes to the thematic development of
the narrative10: Cushan the Doubly Wicked,11 Fat Calf,12He'll Understand,
Captains Crow and Coyote,13 and Sacrifice and Protection Refused.14
A number of the marked extensions on biblical names refer to conse-
quences which befell the name-holder. One of the salient consequences of
Israel's "doing what was evil in the sight of The Lord" (Judges 2:11; 3:7,12,
12; 4:1; 6:1) was their falling victim to such disgusting rulers as these. But
when Israel "cried out to The Lord" (Judges 3:9, 15; 4:3; 6:6, 7), The Lord
would raise up leaders who could deliver them from the Doubly Wicked
9. Examples of marking for structural purposes would be the use of formulas such as "In the
nth year of King A son of B, X son of Y began to reign" to introduce portrayal of a succeeding
king's reign, "and it was after these things" to introduce a new episode, and a particle to mark
story-line events (markers such as these are discussed in detail in Longacre (1989).
10. The meaning of a character's name frequently has thematic significance in biblical nar-
ratives. One example is in Hosea 1:6-9: "Gomer had a second child . . . the LORD said to Hosea,
'Name her Unloved, because I will no longer show love to the people of Israel. . . .' After Gomer
had weaned her daughter, she became pregnant again and had another son. The LORD said to
Hosea, 'Name him Not-My-People, because the people of Israel are not my people. . . . ' " (Amer-
ican Bible Society 1978).
11. Or better: King Cushan, the Scoundrel of Aram-Naharaim.
12. The Hebrew root can mean "calf" or "round," depending on the vowel pointing. A better
translation might be Big Bull. Its weakness is that it could suggest an image of power. But the
alliteration and monosyllables would work with the context to indicate that the suggestion of
power is facetious and there would be a link in English between the name, "Bull's" excretion
(Verses 22, 24), and a common expression for pretense which would serendipitously harmonize
with the Hebrew's bawdy satire.
13. Usually translated Raven and Wolf, but my translation uses terms that are more pejora-
tive, in American English at least, while referring to closely related animals. (Anchor Bible
Dictionary [1992] lists both raven and crow as translations for the Hebrew reference to the bird.)
14. This is probably the most frequently suggested translation of the name (e.g., Boling in
Judges [1975: 155] and HarperCollins Study Bible [1993: 383]; Soggins (1987: 131). Others
such as wandering shade (New Jerusalem Study Bible note) and Salm protects (Mendenhall in
Anchor Bible Dictionary [1992(6): 1055]) have also been suggested and would all harmonize
with the irony of this section.
kings. The Fat Calf demanding "tributes" (elsewhere, the Hebrew word can
refer to temple offerings) is as easily slaughtered as calves at the altar. The
enemies of the people of The Lord should be no more of a threat than a Crow
or a Coyote. Such enemies are destined to become like the Sacrifice Whose
Protection Was Refused. These accounts illustrate the nature and the out-
come of those who stand opposed to God. The readers/hearers of these bibli-
cal texts are called to understand this, with the hope that they will understand
more quickly than the one at the center of the accounts understood, the one who
was referred to by the most highly marked structures: King He'll Understand.
MARKED REFERENCES TO "X THE RELATIVE OF Y"
X the relative of Y is a marked reference to a character: It is structurally more

complex than the use of the character's proper name or pronoun and occurs
less frequently and in fewer environments.
In introducing his discussion of the structural and thematic uses of X ben
Y, Clines stated, "The long form . . . is not always employed when one of the
[environments he identifies] arises; one can only hope to show why X ben Y is
used when it is used, and one cannot usually speculate about why it is not used"
(1972:267). But the notion of markedness provides a simple explanation in this
respect: Unmarked structures, the default choices, occur more frequently
than marked structures. The degree to which the narrator wishes to mark a
reference will determine how frequently he or she uses marked structures.
As Berlin pointed out, the use of relationship terms in addition to the
proper name occurs thematically "to [call] attention to [to mark] specific
relationships between certain characters" (1983: 18). But there is consider-
able variation in the degree to which the narrator marks these relationships.
TERMS OF ADDRESS IN THE STORY OF JACOB AND LABAN

In the Genesis Jacob-steals-the-blessing episode, each of the four main char-
acters (the two parents and the two children) are referred to twice as fre-
quently by their proper names than by the more marked structure X the
relative of Y.15 However, when Isaac and his parents are in the same scene
(Genesis 21-25:11), just the opposite ratio occurs: The more highly marked
form is used twice as frequently as the proper name.16 When Isaac and his
15. Totals of use of proper name (e.g., Isaac) versus use of X the relative of Y (e.g., his father
Isaac): Isaac—nine versus five; Rebekah—five versus one; Esau—seven versus {our; Jacob—nine
versus four).
16. When only the proper name occurs, the immediate environment would make use of the
structure X the relative of Y awkward if not ungrammatical: "Abraham gave the name Isaac to
his son" (Genesis 21:3), "on the day that Isaac was weaned" (21:8), "It is through Isaac that
offspring shall be named for you" (21:12), "Isaac said to his father Abraham" (22:7); "Abra-
ham gave all he had to Isaac" (25:5 Masoretic Text of the Hebrew); his son occurs in other
manuscripts).
parents are not in the same scene, the structure Isaac ben Abraham/Sara is
never used while the less marked structure Isaac is used about seventy times.17
The frequency with which the marked structure X the relative of Y is used
hits an all-time high in the passage concerning Laban, which was quoted earlier:
three times in four consecutive clauses. With little text-structural motivation for
these repetitions, the prominent marking points to thematic considerations.
The intense highlighting of the familial relationship between Jacob and
Laban, along with other references in the immediately following verses to the
joy of family,18 heighten the irony of the situation: Family, the crucial source of
joy and support, will soon turn into a locus of bitterness and rivalry. Marking
this relationship in conjunction with the flock—the wealth of Laban—foreshad-
ows one of the key elements in the rivalry. In the verse immediately following the
happy reunion scene, Laban mentions both "kinsman" and "wages." The
kinsman relationship between Laban and Jacob will never again be mentioned.
Their wage relationship will come to the fore.19 The end of joy, the disruption
of the extended-family unit is evident in the following references, which so
sharply contrast with the exuberant discovery of the brother of his mother:
Jacob to Laban: . . . my own home and country . . . my wives . . . my children
for whom I have served you . . . (Genesis 30:25-26)
The sons of Laban: Jacob has taken all that was our father's. . . . (31:1)
The immediately preceding verse: [Jacob] grew exceedingly rich, and had large
flocks....)
Narrator: Jacob saw that Laban did not regard him favorably. . . . (31:2)
Narrator: Jacob deceived Laban the Aramean (31:20)
Narrator: [Laban] took his kinsman (31:23)
Narrator: God appeared to Laban the Aramean
Jacob to Laban: my kinsfolk and your kinsfolk (literally, "my brothers and your
brothers") (31:37)
17. The references to Lot are similar to those of Isaac. In three of the first four references
to Lot (Genesis 11:27, 31; 12:4, 5), he is referred to in terms of his relationship to Abram. He
is a passive character counted among Abram's other dependents and possessions. The use of
the X the relative of Y structure accords with the references to Isaac and Jacob, mentioned
earlier, used in referring to dependency relationships. It is a marked structure but, in this
context, not highly marked.
In chapter 13, Lot is a wealthy adult with whom Abram must resolve territorial disputes;
he is referred to simply by his unmodified proper name: Lot (eight times). But in chapter 14,
telling of his capture by enemies, Lot's name is used only in the X the relative of Y structure
(14:12, 16): Y equals Abram on whom Lot depends as much for rescue as he did for nurturing
in his younger days.
18. The following four verses contain several references to familial relationships (her father's
kinsman; Rebekah's son; his sister's son; you are my bone and my flesh) and to the joy of this
relationship (kissed [twice], wept aloud, ran, embraced).
19. Eleven of the twenty-three occurrences in Genesis of the verb work/labor/slave occur in
the Jacob-Laban narrative; nine out of the ten references in Genesis to wages occur here.
Laban devolves from Jacob's mother's brother to one involved in my family

versus your family disputes with Jacob, to the Aramean: from a person famil-
ially attached to Jacob to one familially then geographically separated.
THE TRANSLATION OF MARKED REFERENCES

We now briefly consider how various translations treat marked structures by
looking at how they render the Laban and the enemy king passages referred
to in the preceding section.
There are three basic options for translating structures that are marked in
the original text: (1) reproducing the marking patterns of the original, (2)
reducing or omitting the marking patterns, or (3) replacing the marking pat-
terns of the original with marking patterns in the target language that will
have the same, or at least a similar, communicative effect.
The first option is commonly chosen by what are often referred to as
literal translations (from Genesis 29:10):
Now when Jacob saw Rachel, the daughter of his mother's brother Laban, and
the sheep of his mother's brother Laban, Jacob went up and rolled the stone
from the well's mouth, and watered the flock of his mother's brother Laban.
(New Revised Standard Version 1993)
And when Jacob saw Rachel, the daughter of his uncle Laban, and the flock of
his uncle Laban, Jacob went up and rolled the stone off the mouth of the well,
and watered the flock of his uncle Laban. (New Jewish Publication Society 1988)
The problem is obvious. The Hebrew underlines the thematic importance of
the relationship and the flock. The English translations might possibly under-
line this aspect, but they also signal to the reader that the narrative was
clumsily produced. The translation of the New Revised Standard Version
sounds foreign; that of the New Jewish Publication Society sounds silly.
Today's English Version translates according to the second option. It trans-
lates the Genesis text as if it were completely unmarked, eliminating two of
the references to the familial relationship:
When Jacob saw Rachel with his uncle Laban's flock, he went to the well, rolled
the stone back, and watered the sheep.
The English style is much better than in New Revised Standard Version and
New Jewish Publication Society Version but the thematic underlining is lost.
The following suggests how the passage, in the context of the immediately
preceding and following verses, could be translated by using English, rather
than Hebrew, marking techniques:
While he was speaking with them, Rachel had come near with her father's
sheep, under her care. Jacob again looked at her: the daughter of his uncle, his
mother's brother! He then looked at the sheep: his uncle's! He took the stone off
the well so that the sheep could drink. Then, crying for joy, he kissed Rachel.
SUGGESTIONS FOR REFERENCES TO THE ENEMY LEADERS
Most, if not all, translations choose one of the first two options listed in the
introduction to this section to render the references to the enemy leaders in
Judges 3-8. However, they would be better translated in a way that would
allow the reader to appreciate the thematic relevance of their names to the
narrative's presentation. Rather than being relegated to footnotes, the mean-
ings of these names should be represented in the text of the translation, and
the humor and irony of the original kept. This could be done in various ways;
thus I offer the following suggestions as examples:
So The Lord became angry with Israel and let them be conquered by Aram-
naharaim's King Cushan the Scoundrel. For eight years, they were enslaved by
The Scoundrel. (Judges 3:8)
The Lord empowered Moab's King Calf. . . . The Israelites were King Calf's
slaves for eighteen years. . .. The Israelites sent a gift to Moab's Calf. . . . Then
[Ehud] presented the gift to Moab's King Calf—a really fat Calf, he was. (3:12,
14, 16, 17)
On that day, God made Canaan's King He'll Understand give way to the Isra-
elites. Their power over him steadily increased until they destroyed He'll Under-
stand, King of Canaan. (4:23-24)
They killed Raven at Raven Rock and Wolf at Wolf's Winery. (7:25)
Then Sacrifice and Protection Withheld said, "You come and kill us. . . ." So
Gideon slaughtered like sacrifices those from whom protection had been with-
held. . . . (Judges 8:21)
CONCLUSION
This chapter focused on the translation of marked references that are in-
creasingly marked by repetition of the structures in narrow text space. The
interpretation of this literary device is viewed as proceeding in the same way
as the interpretation of the marked use of language in everyday conversation,
as explained by the markedness model (Myers-Scotton chapter 2, in this
volume).
Just as interlocutors choose and evaluate marked codes in terms of social
values and communication goals, so, too, the producer of a literary text
makes marked choices to signal communicative goals. In turn, the audience
evaluates these choices in terms of literary norms shared with the speaker.
The translator of literary texts may attempt to use equivalently marked ex-
pressions and structures in the target language to represent the original speech
event. The equivalence will be not only with regard to the semantic content
of expressions but also with regard to the pragmatic inferences that the orig-
inal speakers expected their audiences to make. Similar to the switch from one
language to another, when semantically unnecessary but pragmatically signifi-
cant, the occurrence of an expression like says The Lord of Hosts three times in
one verse is semantically equivalent to says The Lord of Hosts occurring once
in the verse, or not at all if it has already been established that Yahweh is the
origin of the utterance. However, the repetition is pragmatically significant in
the text we considered, in that it rhetorically underlines the importance of the
source of the message (divine, not human) and the validity of the prophet's
identity as messenger (not inventor) in keeping with the religious traditions
recognized by the hearers as valid.
I have indicated how translators of biblical texts tend either to reproduce
fairly literally the marking patterns of the original text or to diminish or
ignore them for the sake of fluent, contemporary style.
The former practice can lead the translation's audience to make inferences
that were unintended by the original speaker, because the target audience
tends to interpret the marked structures according to the norms of their own
communicative context rather than those of the original text's communica-
tive context. This misinterpretation is similar to that which can arise in inter-
cultural communication when the second-language speaker uses such stylis-
tic features as loudness, verbosity, or directness, which are not intended to be
marked, or are intended to be marked but with the intention that different
inferences be made than those made by the first-language speaker of the
language. Miscommunication often results, especially when the hearer does
not share a sympathetic bond with the struggling speaker.
On the other hand, when translators diminish or ignore the marking pat-
terns in the original text, a more readable style may result, but the text can be
significantly misrepresented as the reader will not be able, or will be less likely,
to make the inferences that the original speaker desired the audience to make.
In most Bible translations attempting to use contemporary language, an ex-
treme reduction in the use of marked structures contributes to a literarily unsat-
isfactory leveling of style. In the case of the Judges narratives discussed previous-
ly, the stylistic leveling resulted in changing stories that were originally filled
with bawdy and taunting humor to unemotional, pious narratives in the
target language of translation. A more accurate representation of the origi-
nal results when the translator uses marked structures in the target language
that enable communicative effects similar to those intended by the original
speaker's use of marked structures. There may be an overlap in the two lan-
guages' use of marked structures for similar communicative effects, but the two
languages will often differ in how they mark structures to obtain similar results.
The repeated use of the proper names of Israel's enemies in the book of
Judges was evaluated in terms of norms suggested by the statistical study of
references to characters in other biblical narratives, especially in the subgenre
of the battle accounts. It was concluded that the repetition of names was for
thematic purposes, encouraging the audience to attribute significance to the
meaning of the character's name as well as enabling them to identify the
character. Suggestions were given for how equivalent marking could be
achieved in translation.
The thematic use of repetition may be understood as a means of minimiz-

ing communicative costs. While signaling secondary themes, repetition does
not foreground these themes at the expense of distracting the reader from
appreciation of either the primary themes or the basic story line. In the He-
brew Scriptures, primary themes are often made explicit through speeches of
key characters, including Israel's god, or through the narrator's introductory
or concluding comments. The primary theme of both the Genesis and Judges
narratives discussed previously is, broadly stated: The Lord takes care of his
chosen people. This theme is explicitly stated by the Lord in, for example, the
episode immediately preceding the Jacob-meets-Rachel episode and by the
narrator in the introduction to the book of Judges.
Marked repetition is one of the narrative means used to invite the reader
to consider what other themes lie behind the most transparent ones. As one
literary critic says:
If the biblical truth is explicit, then the whole truth is implicit; and the more you
bring to this art of implication, the more secrets and prizes it yields. No one goes
away empty-handed. But the challenge to the ... reader is as omnipresent as the
solicitude for all. (Sternberg 1987: 52)
REFERENCES
American Bible Society. 1978. Good News Bible: the Bible in today's English version.
New York: American Bible Society.
Anchor Bible dictionary. 1992. New York: Doubleday.
Berlin, Adele. 1983. Poetics and interpretation. Sheffield, England: Almond.
Biblia Hebraica Stuttgartensia. 1966/77. Stuttgart: Deutsche Bibelstiftung.
Boling, Robert G. 1975. Anchor Bible commentary on Judges. New York: Anchor.
Clines, David J. A. 1972. X, X ben Y : Personal names in Hebrew narrative. Vetus
Testamentum 22: 266-287.
Dressier, Wolfgang U. 1992. Marked and unmarked text strategies within semiotical-
ly based textlinguistics. Language in context, ed. Shin Ja J. Hwang and William
Merrifield, 5-18. Dallas: Summer Institute of Linguistics and University of Texas-
Arlington.
Givon, Talmy. 1991. Markedness in grammar: distributional, communicative and
cognitive correlates of syntactic structure. Studies in Language 15: 335-370.
HarperCollins Study Bible: new revised standard version. 1993. New York: Harper-
Collins.
Longacre, Robert E. 1989. Joseph: a story of divine providence. Winona Lake, Ind.:
Eisenbrauns.
New Jewish Publication Society. 1988. Tanakh: the New Jewish Publications Society
translation according to the traditional Hebrew text. Philadelphia: Jewish Publi-
cation Society.
The New Jerusalem Bible. 1985. New York: Doubleday.
Nida, Eugene, and Charles Tabor. 1976. The theory and practice of translation.
Leiden: J. Brill.
Soggins, J. Alberto. 1987. he livre des Juges. Geneva: Labor et Fides.
Sternberg, Meir. 1987. The poetics of biblical narrative: ideological literature and the
drama of reading. Bloomington: Indiana University Press.
Venuti, Lawrence. 1995. The translator's invisibility: a history of translation. New
York: Routledge.
Vonnegut, Kurt, Jr. 1959. The sirens of Titan. New York: Dell.
6
Literariness, Markedness, and

Surprise in Poetry
AMITTAI F. AVIRAM
POETRY VERSUS COMMUNICATION
T HAT POETRY is a marked instance of language and that poems make

expressive use of marked stylistic devices in both vocabulary and syntax
are known facts about poetry for decades (Jakobson 1960, 1968, 1973;
Mukarovsky [1948]1989). Somewhat accordingly, many scholars bring con-
cepts from linguistics to bear upon literary texts, specifically by analyzing the
stylistic devices that seem to make the literary text effective (Austin 1981;
Epstein 1981; Fairley 1981; Freeman 1981b; Keyser 1981; Kiparsky 1981;
Tarlinskaya 1987). Nevertheless, many such stylisticians commenting on
poetry tend to focus on these stylistic expressive devices while often seeming
to miss the point of the texts they are reading. These often inadequate, super-
ficial readings result not from mere accident but from unexamined and incor-
rect assumptions about the nature of poetry and the literary text, assump-
tions that neglect the text's literariness in favor of what is believed to be its
communicative function as a message.
Normal communication requires not only utterances but exchanges of
utterances—a communicative context—to establish reference. People must
be able to ask each other questions and have answers; the communicative
exchange is a constant dialectic aiming toward mutual understanding. But
the literary text, as Socrates points out about the "written" text in general,
cannot answer any questions—it always says the same thing, no matter how
things change around it (Plato 1982, 564-567 [275d-e]). By contrast, a poem,
and, indeed, any literary text, is a special case of the use of language, in which
the referential—and even the apparently pragmatic—sense of the utterance is
permanently removed from its communicative context and therefore cannot
truly be said to communicate anything at all. The literariness of the literary
text is, first of all, not a matter of its inherent makeup but of how it is
101
received. Literariness is, in the first instance, a cognitive, not a physical,

category. Indeed, the person who, in a lover's quarrel, responds to a partner's
yelling in anger at him or her by observing, "You're so cute when you're
mad," is, in a sense, transforming a communicative message into a literary
text. This person is inappropriately regarding the utterance that expressed
the anger as an object of contemplation and aesthetic pleasure, attending to
the attractiveness of the form of expression rather than the anger it is intend-
ed to convey. A response that shows appropriate communicative engagement
in this case might be fear, righteous indignation, or defensiveness rather than
admiration at the display of anger as if it were intended as a work of art. But
the literary text is intended as a work of art, even if it represents an utterance
expressing anger.
As in the foregoing example, literariness is not so much a quality of an
utterance itself but rather a quality of how an utterance is perceived, a frame
of mind on the part of the hearer or reader that treats the utterance as some-
thing worthy of contemplation, discovery, and admiration rather than the
conveyance of an intended message. Nevertheless, literary texts usually have
formal features that prompt readers to receive them as literary rather than
communicative. When the reader perceives the text as literary, the apparent
communicative content of the utterance is a sort of virtual rather than actual
communication, one that would only take place within an imaginary world
that would provide it with a communicative context. Because that imaginary
world is created only out of inferences derived from the text's signs, it has
limits that distinguish it from the infinitude of the real world. Insofar as the
literary text communicates an intention, it does so at a level higher than that
represented directly by its own language—outside, as it were, the text's fictive
frame. The fact that the literary text is bracketed off from the exchanges of
normal communication also causes the text to be apprehended as a totality
rather than strictly as a sequence. Whereas normal communication is a po-
tentially endless sequence of signs and responses through time, the literary
text is a completed whole unto itself, acquiring a timeless spatiality in the
reader's mind (on completion of reading) at the same time that it also produc-
es the appearance of a normal sequence.
The markedness of a literary text such as a poem then serves simulta-
neously to express intention and to provide the aesthetic pleasure of sur-
prise—the two, intention and surprise, become inseparable—but only insofar
as such markedness is placed within its properly literary context and the
reader assumes the correct default expectations for what would be unmarked.
In short, the language of a poem does not mean but seems to mean. The poem
means only insofar as it uses this seeming within a larger context, which is not
communicative in the normal sense but literary and aesthetic.
By aesthetic, I mean appealing to a sense of pleasure in the contemplation
of the object, in its totality and as a pattern. This is the purpose, or primary
function, of an aesthetic object. By contrast, the primary function of commu-
Poetry 103
nication is to convey meaning. If one communicates effectively, the aesthetic
pleasure given by the form of communication should be negligible or trivial
compared to the content being communicated. Effective communication usu-
ally means clear, relatively transparent, easily understood communication,
with little distraction of the listener's attention to the formal features of
expression. But if an utterance is considered an aesthetic object, it does not
matter so much what meaning it conveys as how the parts of the utterance
work together to make a satisfactory or successful whole. Two opposite
mind-sets are required for appropriate responses, respectively, to what is
understood as a communicative utterance and to what is understood as an
aesthetic object—a literary text. The proper response to a communicative
utterance is directly relevant to the content of the utterance—a reply, an
action, or an appropriate affect (e.g., fear or defensiveness in the case of angry
speech). By contrast, the proper response to an aesthetic object, such as a
literary text, is contemplation, interpretation, the pleasure of discovering
how the parts come together to make a whole, and a sense of wonder at its
wholeness, completeness, perfection, ingenuity, and forms of surprise.
Communicative messages are not subject to the same interpretive attitude
as aesthetic objects because communicative messages are embedded within
ongoing communicative contexts that can restrict and control their mean-
ings. By contrast, literary texts, as aesthetic objects, are appropriate objects
pondering their meanings, as they are bracketed off from any communicative
contexts other than the limited, imaginary ones they imply. Much of the
aesthetic pleasure of the literary text is precisely in the pondering over possi-
ble meanings and the construction of plausible patterns of meaning. By con-
trast, communicative messages should aim at clarity, and ambiguity is not
generally a source of pleasure but of temporary frustration, remedied by-
further exchanges of communication intended to clarify and disambiguate.
EXAMPLE: EMILY DICKINSON'S "I CANNOT LIVE WITH YOU"
"I Cannot Live with You," a poem by Emily Dickinson, (1960: 317-318, no.
640), illustrates some of these points. It begins with the statement (or, to be
more accurate in speaking about poetry, the pseudostatement), "I cannot live
with you— / It would be Life—." We can infer that the speaker must just have
been asked by someone to live with him—that someone has just proposed
marriage to the speaker (and that the speaker is a woman, the person charged
with either rejecting or accepting a marriage proposal). Later lines in the text
clearly corroborate this inference. Thus we imagine a communicative con-
text, in which an interlocutor, the suitor, asks a question, "Will you marry
me?" and what we "hear" is the response. But we cannot infer how the
interlocutor would respond to this response, nor is it a relevant consideration
for the experience of the poem. The focus of the poem is, rather, on the
abstruse reasons the speaker gives for declining the proposal, all of which
involve, ironically, a way of living up only too well to the (Congregationalist

Protestant Christian) ideal of marriage, given the inherent sex inequalities
within that ideology. The poem thus allows an ideological system to play
itself out against itself. The neatness with which the speaker, in all innocence,
can give voice to a standard ideology in such a way as to make the fulfillment
of that ideology—marriage— impossible is, itself, a source of great aesthetic
pleasure: We wonder at the brilliant wit of this argument that turns so seam-
lessly and yet ironically against its own assumptions. At the same time, the
neatness of the poem's metrical and syntactic arrangements coincides with
the feeling of inevitability about its argumentative structure, again to contrib-
ute to our delight in its ironic perfection. The meter and other poetic elements
do not, then, contribute to the "message" of rejecting a marriage proposal;
they contribute to the surprising aesthetic effect of the text for us, its readers.
The rejection is not aimed at us. We get to delight in the ingenuity, brilliance, and
perfection of the utterance. Within the imaginary world implied by the text, the
imaginary recipient of the message would not be able to enjoy these aesthetic
qualities. For him, it would be a message, not a literary text. By contrast, in
our real world, outside the imaginary world of the text, we, the readers, can
and should take pleasure in the wit and ingenuity with which the utterance
is crafted—that is, its aesthetic qualities. The reader experiences much of
this aesthetic pleasure in the process of solving the riddle of the poem—
figuring out that it could be imagined as the rejection of a marriage proposal
and then marveling at the surprising form of logic such a rejection takes.
This chapter attempts to clarify the problems of stylistics when applied to
poetry, and, in so doing, sketches a concept of the literary text—the poem—
in a way that is more rationally defensible than the concept apparently held
by many stylisticians. I hope that this concept of the literary text would also
be more useful for teaching people how to read and appreciate poetry and
other literature. I discuss the concept of literariness first, defining poetry
accordingly. Then, I examine some errors that I think typify the potential
pitfalls of stylistics-oriented reading. With these criticisms, I urge readings of
the same texts more faithful to the concept of literariness and show how
markedness functions in these readings in the realm of the literary rather than
in the realm of normal communicative utterances. Finally, drawing on these
examples, I review the phenomenon of markedness in relation to communi-
cation in the special case of the literary utterance, suggesting how what is
"communicated" is an aesthetic experience rather than a message.
THE APPROPRIATE ATTITUDE TOWARD THE LITERARY TEXT
A literary text is not communication in a direct and simple sense. If the

utterance takes a propositional form ("Roses are red"), it is subject to the
truth test in normal conversation ("Are they really?") but not in a literary text
(i.e., it matters not whether they are or not; the poem says they are, for its
Poetry 105
purposes). If the utterance takes the form of a question, it would be liable to

an answer or, if a rhetorical question, a response in real life; in literature, it
is only open to literary interpretation. The same with imperatives: In real life,
one responds to "Take out the garbage!" by doing so or by refusing to do so,
whereas, in reading poetry, one responds by figuring out the relevance of the
seeming imperative to the rest of the seeming utterances that make up the
poem as a whole, to discover the whole poem's point, its overall aesthetic
effect. In poetry, even illocutionary acts have no force in the real world but
only contribute elements to an aesthetic whole. If a character says "I cannot
live with you," this does not mean that a real marriage proposal was declined
but that the formula of rejection is presented to us in a surprising way that
gives us aesthetic pleasure.
The communicative context ennabling a communicative utterance has
two elements: (1) a social context of exchange, and (2) a referential context,
giving the utterance meaning and relevance. The social context includes with-
in it the possibility that the recipient of the utterance will be able to ask
questions or answer the utterance in some way, either directly or indirectly
(as, for instance, by writing a letter after hearing a lecture in a large lecture
hall). Even the refusal to respond, the failure to respond, or the accidental
impossibility of responding contributes to the social context. The referential
context must be clear enough to all parties to the communication for commu-
nication to occur at all. Reference is not exclusive to propositional utteranc-
es. If I say "How are you today," it must be clear who the you is and which
day today is for the utterance to qualify as normal communication. By con-
trast to all of this, the literary text requires (1) a reader and (2) inferrable
fragments of a communicative context (with both social and referential ele-
ments) sufficient to make the utterance make imaginary sense in an imagi-
nary world. The elements of this "world" would be insufficient to provide the
infinite possible preconditions, environments, and consequences of a real
utterance. The fragments of a communicative context that can be inferred
from the clues in the text are just enough to make the literary text "work,"
that is, deliver its aesthetic effect of surprise and satisfaction. The limitation
of the communicative context to these few inferrable fragments of an
imaginary world gives every literary text an air of ambiguity and invites
the reader to ponder its interpretive possibilities. But that same limitation,
the result of the text's being bracketed from the constant give-and-take of
normal communicative exchanges, is what gives the text its finitude and
completeness as an aesthetic whole, so that the entirety of its pattern can be
appreciated and enjoyed.
Because the communicative message and the literary text are received
cognitively in completely different ways, the appropriate responses to each
are different. When the speaker of Wallace Stevens's "Blanche McCarthy"
exhorts, "Look in the terrible mirror of the sky," our first task is not to look
up (and thus cease reading the poem) but to consider what the "terrible
mirror of the sky" might, figuratively, mean, and thus how it will fit into that
text's system of figurative signs as an aesthetic experience. When the speaker
of John Keats's "Ode on Melancholy" admonishes, "No, no, go not to Lethe,"
we are to imagine that the utterance would be properly directed to an imag-
inary person—a virtual literary character—who is about to commit suicide.
It is unlikely that we would happen to receive the message appropriately
when we, ourselves, are on the verge of committing suicide and would pause
long enough to read the poem through to take the rest of its admonishments.
Here, as elsewhere in poetry, even the seeming message—the content of the
imaginary communicative utterance in an imaginary context—is far too com-
plex and ambiguous to serve as a practical message in a real-life situation.
This richness is not a necessary feature of a literary text, but it follows natu-
rally, on both the poet's and the reader's parts from the assumption of the
appropriate frame of mind that the reader would hold in receiving a literary
text rather than a communicative message.
That the appropriate response to the imperative beginning "Ode on Mel-
ancholy" is the imaginary reconstruction of a possible scene in which the
utterance would be a communicative message introduces a crucial paradox
about the literary text. This paradox is that the literary text's existence de-
pends on the hypothetical possibility that the very same utterance, the same
signs, could be employed in an imaginary communicative exchange within an
imaginary referential context.
LITERARINESS AND THE AESTHETIC PLEASURE OF SURPRISE
Literary meaning, then, is a special case of meaning. What a literary text

"communicates" to its reader is its aesthetic effect, which may be understood
as some form of surprise. Literary texts use the material ordinarily used for
communication to deliver surprise, which pleases the reader, edifies him or
her, sharpens readerly skills, and performs other functions associated with
aesthetic experience rather than with communication per se. Literary texts
achieve surprise by means of various kinds of markedness. The very removal
of the utterance from any conceivable actual communicative exchange and
referential context already makes the utterance as a whole, by definition,
marked as literary, and it is in part the task of the literary writer to produce
sufficient linguistic evidence to make sure that the reader does not mistake
the text for a communicative message. Verse in poetry, among its functions,
sets the language of poetry off from that of normal communication, as, in
verse, attention is drawn toward the mere surface of language—its pure sounds
or visual appearance in writing—rather than to its meaning. This shifting of
attention would be inefficient and counterproductive if the utterance were
intended for communicative purposes, because the aim of a communicative
utterance is, first and foremost, transparency. When a parent tells a child to
wash the dishes, the parent does not want the child to dwell on the sound of
Poetry 107
the parent's consonants or the repetition of sh in wash and dishes. This
abnormal focus on the surface of language rather than on its meaning is
specific to poetry. Nonpoetic literary texts use a variety of other devices to
ensure the reader's response to the text, as a whole and in the last instance,
as aesthetic rather than communicative.
But the mere existence of verse in poetry is not generally sufficient to
deliver surprise, except in a very vague way. Poetry produces surprise by
means of marked choices against default assumptions on many different lev-
els, all depending either on literary contexts alone or on the peculiar and
paradoxical interaction between literary contexts and the nonliterary, real
contexts that they imitate or to which they allude, in actual communication.
For example, "Because I Could Not Stop for Death" (Dickinson, 1960: 712,
no. 350; see Aviram 1994: 263-279) surprises the reader in many ways,
including the following.
1. The speaker apparently has already died and is thus speaking from be-
yond the grave. This doesn't generally happen in real life. In literature,
such speeches are usually attended by signs of the supernatural, and the
speeches are often cryptic and oracular. Here, the speech is a matter-of-
fact past-tense narrative of the events leading up to the speaker's present
(dead) state.
2. The speaker has died, but she does not sound as if she is in heaven or hell
but, rather, in a state of complete emptiness, where nothing happens.
(This also makes the idea of her speaking from such a state even more
unfathomable.)
3. The poem is in tetrameter quatrains (four beats per line, four lines), with
the second and fourth lines truncated so that the fourth beat is observed
as silence rather than being realized with a syllable of a word. This is the
conventional hymn form, which would be conventionally associated, in
the Christian context to which other signs in the poem allude, with praises
for the happy afterlife promised to the saved Christian soul. Yet this poem
contains no such praise, and thus the poem forms a sort of anti-hymn
when read against the cultural context of the Christian beliefs to which it
alludes or that can assumed to be known (and possibly held) by its imme-
diate readership in nineteenth-century America.
4. The allegorial figures of the poem describe death as if it were marriage—
to the figure Death. This itself is a small surprise—the bridegroom being
Death. But death has long been figured in Christian literature as like a
marriage—to Christ, not to death. So having Death appear in the place of
Christ is another surprise.
The surprises all work together to create what appears to be the speech of a
naive speaker who now knows that death leads not to heaven or hell but
simply to an immense vacancy. In other words, the poem dramatizes a rever-
sal of fundamental Christian beliefs about the afterlife. The effect of the poem
does not depend on this apparent revelation being either true or false; it
depends on its being imaginatively successful and giving us the pleasure of
surprisingly reversing ordinary, normative expectations. All this is accom-
plished by an intricate network of markedness on thematic, formal, and
figurative levels.
The capacity of a text to deliver pleasing surprise depends, therefore, on
the specific knowledge of its readership, knowledge drawn both from life
and—especially—from literature. What is delightfully surprising to one read-
er may be simply chaotic to another and too predictable to a third. This
difference is what accounts largely for differences in literary tastes, critical
receptions of various poets, and the constant changes in academic and cul-
turewide canons of good poetry. This fact, too, creates a paradox, this time
about the teaching of literature. Students appreciate a literary text insofar as
they have sufficient prior knowledge to view the marked features of the new text
with the delight of surprise. Yet it is that very pleasure that would motivate a
student to wish to acquire sufficient knowledge—sufficient literacy—to have
it in the first place. In teaching literature, we constantly find ourselves in the
position of trying to get students to read things in the hope that, eventually,
they will realize—always retrospectively, and often too late for us—that the
texts we assigned did, in fact, give them pleasure. But this is precisely why
academic canons cannot be based on prevalent popular tastes and are, in some
ways, necessarily conservative. The task of teaching literacy includes that of
continually reintroducing sufficient literary context from the past to make the
pleasurable surprises of texts available to succeeding generations of students.
Learning to read poetry is not the same thing as learning to read messages,
although the former recruits the skills of the latter to gain access to surprise
and literary pleasure. Yet the approach that some linguists take, in comment-
ing on the style of particular poems, presumes an identity between these two
skills. As a result, stylisticians of high professional repute and considerable
expertise in their field have, in their commentaries on poems, completely
missed the surprise constituting the point of those poems and belabored
issues that are somewhat more obvious. Teachers and students, in turn, read
these commentaries and are likely to repeat the stylisticians' mistakes: to
confuse the literary text with a communicative message, to relegate literary
devices to the role of "boosters" for the message's meaning, to read parts in
sequence at the expense of the whole, and to oversimplify the sense of the
literary text.
MISTAKING POEMS FOR MESSAGES, STYLISTICS INVITES

SUPERFICIAL READINGS
Because stylisticians assume that what makes poetry poetry is the use of
various marked linguistic devices in order to convey meaning (i.e., to commu-
nicate) stylisticians tend to assume a given meaning for a poetic text and then
Poetry 109
elaborate all the various devices that "boost" the expression of this meaning.
In many cases, the meaning they infer as their starting point is oversimplified
and sometimes illogical when applied to the totality of the poem, the whole
text rather than parts of it in sequential order. This superficiality, flatness, or
even wrongheadedness is not simply an accident. It proceeds from the project
of "revealing" the expressive devices of a text rather than, first, understand-
ing the way literary texts such as poetry work as aesthetic structures whose
function is to deliver surprise when given the right context and when read as
totalities rather than sequences of elements. Stylisticians tend to assume that
the literary text is a message like any other, only more punchy or expressive,
using more elaborate devices to get the message across. Accordingly, the
message is a sequence of signs, bearing the normal relation of beginning to
end rather than a sequential elaboration of a complex structure whose parts
must be remembered and held together in the reader's mind to appreciate the
totality. These assumptions are incorrect and lead, ultimately, to disappoint-
ing misreadings of rich and powerful poems.
EXAMPLE: BLAKE'S "THE TYGER"
My first example of this problem is E. L. Epstein's (1981) otherwise interest-

ing essay, "The Self-Reflexive Artefact: The Function of Mimesis in an Ap-
proach to a Theory of Value for Literature." Epstein discusses many great
poems in a variety of languages, including, briefly, William Blake's "The
Tyger." His general thesis is that literary value, especially in poetry, derives
from the extent to which linguistic devices are used in a mimetic fashion—to
imitate the content of the poem. Thus, Epstein's sense of the term self-reflex-
ivity differs sharply from its meaning in literary theory. For Epstein, self-
reflexivity means the "reflection" or imitation (or merely expression, repre-
sentation) of the meaning of the message in its form—as if the form were a
dressing or clothing for the meaning. By contrast, in modern literary theory,
the term refers to the way that the literary text, bracketed off from normal
communication, represents in its thematic content its own nature as a literary
text. That is, the poem is "about itself." (See Aviram 1994: 17-28, 43-58,
223-246.)
In the specific case of "The Tyger," Epstein spells out quite neatly the
content he finds imitated (or expressed) in the form:
What Blake achieves in Tyger [sic] is a subjective mimesis, based on syntax, for
a very high degree of value. It seems obvious from the poem that what Blake is
conveying is his own awe at perceiving the energy that drives the universe and
the poet, a force here symbolized as a tiger, a power beyond good and evil.
(Epstein 1981: 187)
Obvious, indeed. He also notes, beforehand, that, after the reader does "his
'first complete reading,' as it could be called, all further readings contribute
nothing new and become increasingly automatized scannings of a familiar

linear pattern" (Epstein 1981: 186).
Is rereading "The Tyger" that useless? I doubt it, especially if the results
of the first complete reading are as Epstein summarizes them, quoted here.
Let us consider the poem as would a good literary reader rather than a stylis-
tician. The good reader's goal will be, first, to read the whole of the poem,
and to understand how the parts relate to the whole that comprises them.
It is striking that "The Tyger" is made up exclusively of questions.1 From
an interpretive point of view, these questions beg a question: Are the ques-
tions to be taken as straightforward, wherein the speaker seeks answers, or
are they rhetorical questions, not intended to be answered or having self-
evident answers? Rhetorical questions are marked uses of the normal ques-
tion structure, which is normally assumed to seek an answer. One of the ways
that rhetorical questions are marked as such is by containing within them the
information that would otherwise be left for the answerer—that is, the rhe-
torical question contains its own answer, and thus prompts a response that
accords with its implied answer.
1. What are you doing today?
2. Are you going swimming today?
3. Surely, you're not going swimming today, are you?
The third question would likely be understood, in conversation, as a rhetor-
ical question, wherein the statement, "You are not going swimming," is
contained, as it were, within the question. Whereas item 1 leaves open any
possible answer suggesting a plan or activity, and item 2 leaves open either a
yes or no answer, question 3 requires either assent—"Yes, you're right, it's
too cloudy"—-or remonstrance—"Actually, I was planning to swim; why
not?" Question 3, in other words, is a version of the statement "You
shouldn't go swimming today," or even the command, "Don't go swim-
ming today."
In "The Tyger," the questions are not clearly rhetorical, but not clearly
straightforward, either.
Tyger tyger, burning bright
In the forest of the night;
What immortal hand or eye,
Could frame thy fearful symmetry? (Blake [1789-1797]1967: plate 42)
As a rhetorical question, this sentence would be equivalent to the statement
"Surely, no immortal hand or eye could frame thy fearful symmetry." Yet,
such a statement makes no sense, semantically, within the context of social
and cultural beliefs that would seem applicable to the poem. If, that is, the
tiger were "created" at all, it would have to have been created by some sort
1. My reading of Blake's "The Tyger" follows, and develops, the one given by Bloom
(1971: 35-39).
Poetry 111
of immortal being—a being with powers sufficiently divine to create mortal

beings, and thus, itself, immortal—whether God or a classical figure such as
Prometheus or some other imaginary immortal being. In this case, then, the
question is not rhetorical but straightforward, asking for the identity of the
divine being who created the tiger. And yet, like a rhetorical question, this
question contains within it sufficient information to make the answer neces-
sarily one from a limited number of choices—some being that has an immor-
tal hand or eye.
In the second quatrain, the speaker further informs our concept of the
being about which the questions are posed:
In what distant deeps or skies,
Burnt the fire of thine eyes?
On what wings dare he aspire?
What the hand, dare seize the fire? (Blake [1789-1797]1967: plate 42)
The maker of the tiger must have had wings and must have dared to aspire.
It would be inappropriate to speak of God as daring to do anything. The
image in the questioner's mind seems to derive from that of Lucifer, leader of
the rebel angels, in Milton's Paradise Lost. (The image of Daedalus aspiring
to fly may be relevant but probably not as clearly so as the image of Satan.)
Following this thought, the last line presumes that the tiger's maker had
seized the fire used to make the tiger's eyes, in a way reminiscent of
Prometheus's theft of fire to give to man. The transition from Lucifer to
Prometheus is possible because (1) both of them are types of rebel divinity,
and (2) in some sense, neither of them could really have made the tiger—
neither the story of the Fall of the Angels nor that of Prometheus's creation
of man (after which he gives man fire, against the will of the other gods) quite
fits in an imaginary story of making the tiger—and, in any case, there is no
actual story of the tiger's creation in canonical literature or myth. Milton's
Satan specifically is incapable of creating life—a notion following that of
Spenser's arch-demon Archimago, who can only make illusions, not real
living beings. And Prometheus only creates man and steals fire to give it to
man; Prometheus would not make an animal that man finds frightening. So,
in both cases, the questions contain within them hypotheses that make the
questions unanswerable. These are special, marked questions, because
they contain enough information to act as rhetorical questions, but the
answers that they would prompt if they were rhetorical questions would
be patently incorrect.
One of the most problematic aspects of Epstein's reading of "The Tyger"
is his assumption that the speaker of "The Tyger" is the same as the actual
author of the poem, William Blake. As we have seen thus far, the speaker of
the imaginary speech that the poem represents is somewhat befuddled by his
own questions. He asks questions that are impossible to answer because they
at once contain too much information to be open to an answer as if they were
straightforward questions, and, as rhetorical questions, presume incoherent

or impossible answers. The befuddlement of the speaker becomes even more
dramatically obvious in the next quatrain:
And what shoulder, & what art,
Could twist the sinews of they heart?
And when thy heart began to beat,
What dread hand? & what dread feet? (Blake [1789-1797] 1967: plate 42)
The first question here continues in the line of the previous ones, suggesting
that whoever made the tiger must have (or have had) some great physical
powers to make something of such power to terrify. But the second question
seems to confuse the object of creation with the creator. The "dread feet"
couldn't really be attributes of the creator. Whereas the hand would be used
to make the tiger and would follow in the series of metonymies, shoulder, art,
hand—as well as recalling "What immortal hand" from the opening and "What
the hand dare seize the fire" in the second quatrain, feet are not used to create.
This is not to mention that it would be hard to imagine in what way feet can be
dread, except insofar as a tiger has claws on both fore and hind paws and springs
to the attack on its hind legs, and thus has dread feet. The very confusion or
misattribution of qualities from the tiger onto the creator, or perhaps vice
versa, illustrates something about neither the tiger nor the creator but, rather,
the speaker himself. The speaker seems to be caught up in a sort of hysteria,
compatible with his inability to accept the real answer to his original question.
The mood of hysteria rises even higher in the next quatrain:
What the hammer? what the chain,

In what furnace was thy brain?
What the anvil? what dread grasp,
Dare its deadly terrors clasp? (Blake [1789-1797]1967: plate 42)
One can imagine a hammer and an anvil involved if the making of the tiger
is likened to a product of the Greek god Hephaestos (Roman Vulcan), yet
another potential but impossible creator, because Hephaestos also never ac-
tually makes living beings, but rather artefacts such as Achilles's shield. It is
hard even to imagine how a chain would come into play here. And, although it
makes some sense to think of the tiger's brain as being full of "deadly terrors,"
it is hard either to imagine the brain, even a tiger's brain, as the product of a
hammer, an anvil, and a chain or to imagine why a "dread grasp" would be
necessary to take hold of it. Most of all, the questions at this point seem to fall
thick and fast, contributing to the mounting emotional pitch of the speaker.
The next quatrain both alludes again to Paradise Lost and makes clear to
us, if it wasn't before, why the speaker is having so much trouble:
When the stars threw down their spears
And water'd heaven with their tears:
Did he smile his work to see?
Did he who made the Lamb make thee? (Blake 1789-1797]1967: plate 42)
Poetry 113
The first three lines, again, seem to presume that it was Lucifer—or Satan, as
he is known after the fall from heaven—who created the tiger, again, a doc-
trinal impossibility. Here, the presumption leads to a fearful imagining of
Satan taking pleasure in making something terrifying to humanity. But the
closing line asks the very opposite question: How could the very same God
who made a soft, friendly, and woolly thing like the lamb also make the
ferocious tiger? How could the God who made Good also make Evil? This
last gloss of the speaker's question is further assisted by the proverbial
identification of the Lamb of God with the person of Jesus, and also by the
poem, "The Lamb" (Blake [1789-1794]1967: plate 8), in Songs of Inno-
cence, composed a few years before Songs of Experience, in which "The
Tyger" appears. The title, "The Tyger," may in some sense allude to this
earlier poem, as both titles are generic names of animals. In "The Lamb," the
speaker, who identifies himself gleefully as a child, says that the lamb was
created by God, and both the lamb and the speaker (a child) typify the inno-
cence that God has created. The contexts thus reinforce the idea that the
problem the speaker has in "The Tyger," the problem pushing him into a
state of hysteria, is the traditional Problem of Evil: How could God, who is
good, have made evil things in this world?
The final quatrain, repeating the first but for one word, once again seems
to settle—and yet not to settle—upon the alternative answer, that it was not
God, but Satan, who created the tiger:
Tyger Tyger burning bright,
In the forests of the night:
What immortal hand or eye,
Dare frame thy fearful symmetry? (Blake 1789-1797]1967: plate 42)
The word dare recalls "One what wings dare he aspire," alluding to Lucifer's
rebellion against God, as if to suggest that the creation of the tiger was, itself,
a gesture of rebellion and usurpation of God's creative powers.
To review, two important points come out of our close reading of Blake's
"The Tyger." First, the speaker's questions have the form of rhetorical ques-
tions insofar as they contain enough information to imply their own answers,
but the answers that such rhetorical questions would prompt would be both
inconsistent with each other and generally impossible according to the doc-
trinal and mythological contexts to which they allude. Second, the speaker's
insistent questions seem to boil down to a refusal to believe the most obvious
answer to the question of who made the tiger, God, because it seems impos-
sible to the speaker that a good God could make evil things.
But the very same logic that enables us to question the assumptions that
the speaker's apparently rhetorical questions assume should also enable us to
question the assumptions underlying the central question, the Problem of
Evil. For this problem is only relevant insofar as it is assumed, to begin with,
that the tiger is evil, an assumption manifest with all the terms suggesting
terror, fear, dread, and so on, throughout the poem. Just as the speaker has
created for himself his own incapacity to answer his own questions by insis-
tently posing questions that are too closed to allow the correct answer (God),
so, too, he has created his own problem to begin with by assuming that the
tiger is evil. In a sense quite consistent with Blake's other work, the speaker
himself has "created" the tiger insofar as he has created in his own imagina-
tion the object of which he is terrified and projected that object onto the
actual being, the tiger.
This idea that evil is created in the eye of the beholder is consistent with
other poems in Songs of Experience, in particular, "A Poison Tree":
I was angry with my friend:
I told my wrath, my wrath did end.
I was angry with my foe:
I told it not, my wrath did grow.
And I water'd it in tears,

Night and morning with my fears,
And I sunned it with smiles,
And with soft deceitful wiles;
And it grew both day and night,
Till it bore an apple bright,
And my foe beheld it shine,
And he knew that it was mine,
And into the garden stole

When the night had veil'd the pole;
In the morning glad I see
My foe outstretched beneath the tree. (Blake 1789-1797]1967: plate 49)
Blake's retelling of the fall in the Garden of Eden—a story originally about
the acquisition of knowledge of the difference between good and evil—draws
our attention to the minimal difference between friend and foe—supported
by the minimal phonological difference in the first and third lines: 1 was angry
with my f-. We have no reason to know, or to think, that the friend and the
foe are inherently different, until the speaker chooses to treat them differ-
ently. The speaker feels wrath equally toward his friend and his foe.
What defines the foe as a foe is not that the foe incurs the speaker's wrath
but that the speaker does not tell his wrath to his foe, leading eventually to
disastrous consequences in which the foe lives up to his role as foe and
the enmity between speaker and foe reaches the extremes of deception
and death.
This idea also corresponds to the meaning of the phrase in "London," also
in Songs of Experience, "the mind-forg'd manacles":
I wander thro' each charter'd street
Near where the charter'd Thames does flow
Poetry 115
And mark in every face I meet

Marks of weakness, marks of woe.
In every cry of every man,

In every infant's cry of fear,
In every voice, in every ban,
The mind-forg'd manacles I hear. (Blake 1789-1797]1967: plate 46)
Everywhere the speaker goes, he witnesses a suffering connected to restric-
tion and confinement—even the Thames river is "chartered" and thus not
free to all. But the manacles of confinement are forged in the mind. They are,
implicitly, the products of bad ideas and lack of imagination.
Finally, "The Tyger" has its own most immediate visual context to give us
a strong clue to corroborate the above reading. Blake originally published his
Songs of Innocence and of Experience himself, etching each poem by hand
into a copper plate, as if to imitate medieval illuminated manuscripts, along
with illustrations to the poems, some of them exquisite, which he colored in
by hand painting them on the newly printed books. "The Tyger" has one such
illustration: a picture of a tiger, below the words of the poem and in the
bottom quarter of the page. (A tree along the right side of the page may
suggest an Edenic setting and pehaps reinforce the connection with "A Poison
Tree.") The tiger hardly looks frightening or terrible. If anything, with eyes
wide open and shoulders low, he looks a bit frightened himself, which would
make perfect sense if he is the target of the speaker's frightened— and thus
frightening—projections. Even if the other literary contexts mentioned here
as corroborating evidence should be considered inaccessible to the beginning
reader—and they should not—still, the picture itself is evidently as much a
part of the "text" of "The Tyger" as the words are.
We have come a long way from Epstein's reading of "The Tyger" as an
expression of Blake's "own awe at perceiving the energy that drives the uni-
verse and the poet" (1981:187). The shallowness of Epstein's reading comes
despite his meticulous commentary on specific stylistic choices within the
poem, some of which may be correct, but all of which are trivial if, to begin
with, they are assumed to support an unsatisfactory reading. But the reduc-
tive, superficial reading is not an accident. Epstein's failure to recognize the
various kinds of irony at play in " The Tyger" arises from the theoretical
assumption that the poem is a communicative message-in this case, express-
ing Blake's awe-and that the task of a good reader is to notice how expres-
sive this message is. "The Tyger" is not a communicative utterance. Rather,
it is a dramatic exercise, involving an imaginary speech. Our task as good
readers is to notice where the surprises lie in the content and form of this
speech, and what sort of meaningful pattern the surprises form. This pattern
gives us, simultaneously, the poem's meaning and its aesthetic effect.
The poem's surprises depend on markedness on a variety of levels, many
of which require knowledge of appropriate contexts. For instance, the clash
of implied answers between Satan, Prometheus, and Hephaestos depends on

our recognition of literary and mythological allusions. The summary of the
questions as the Problem of Evil depends somewhat on familiarity with this
problem in Western thought, especially in religion. In the former case, the
unmarked condition would be a single, coherent field of allusive reference,
making the questions neatly rhetorical and their answers easy to produce. In
the latter case, the unmarked condition would be the simple answer, yes, God
made all animals, both the lamb and the tiger. Without these contexts, one
cannot fully appreciate the way that the speaker seems to be answering his
own questions with answers that he cannot accept and yet cannot accept the
answer (God) that would seem more obviously appropriate, given his con-
cerns, because of the Problem of Evil. And without these contexts, the reader
cannot arrive at the point of noticing that the speaker is in some sense inventing
his own difficulties by categorizing the tiger as "evil" from the start. The un-
marked condition here would be either a celebration of the tiger as a natural
creature—an attitude somewhat typical of the Age of Sensibility—or a com-
plaint about the appearance of evil in an otherwise good world—a sentiment
we might find in Renaissance literature, in which things of nature might be
employed allegorically to represent moral categories such as good and evil.
Blake creates a marked utterance on this level by bringing together these
two, incommensurate viewpoints—an allegorical and a sentimental view of
Nature-—to reveal the factitiousness of both views as products of the perceiv-
er's projections onto the world. This last point, however, is not Blake's mes-
sage, as such, but rather the effect of his drama. Poetry does not tell, it makes
things happen with words.
In "The Tyger," Blake causes us to recognize the difficulties someone can
get into by making moral assumptions about things in nature that have no a
priori moral essence. In this sense, we could say that the poem "communi-
cates" to us, but what it communicates is not a message but an experience.
This experience—what happens if we have this attitude—is indistinguishable
from the poem's aesthetic effect, which is the effect of surprise. Here, the
surprise is, ultimately, a series of surprises: (1) the speaker is saying nothing
but questions; (2) the questions can't be answered, even though they sound
like rhetorical questions; (3) the questions reveal great consternation; (4) the
speaker has created the cause of his own consternation; and (5) the answer to
the question turns out to be "God" only in the first instance but "the speak-
er" or "the perceiver" in the last. It is this succession of intellectual surprises
that makes the work simultaneously deep and delightful.
EXAMPLE: WALLACE STEVENS'S "THE SNOW MAN"
My second (and last) example of an unsatisfactory reading produced by a

stylistician is Samuel J. Keyser's (1981) reading of "The Snow Man." Let us
first recall the poem.
Poetry 117
The Snow Man

One must have a mind of winter
To regard the frost and the boughs
Of the pine-trees crusted with snow;
And have been cold a long time

To behold the junipers shagged with ice,
The spruces rough in the distant glitter
Of the January sun; and not to think

Of any misery in the sound of the wind,
In the sound of a few leaves;
Which is the sound of the land

Full of the same wind
That is blowing in the same bare place
For the listener, who listens in the snow,

And, nothing himself, beholds
Nothing that is not there and the nothing that is. (Stevens 1954: 9)
Keyser (1981:112-120) comes closer to a convincing reading of this poem
than Epstein does with "The Tyger," but not close enough to keep him from
missing the point entirely. To decide the semantic meaning of the poem,
Keyser first quotes Frank Kermode and then Stevens himself, clearly accord-
ing more authority to the latter than the former. Here is Kermode first:
Out of "The Snow Man" grows the recurring metaphor of winter as a pure
abstracted reality, a bare icy outline purged clean of all the accretions brought
by the human mind to make it possible for us to conceive of reality and live our
lives. So purged, reality has no human meaning, nor has a man; he is
. . . the listener, who listens in the snow,
And, nothing himself, beholds
Nothing that is not there and the nothing that is.
(quoted in Keyser 1981: 119)
He continues:
In contrast to Kermode's interpretation of The Snow Man we have Wallace
Stevens's own explanation which he gives in a letter to Hi Simons, dated 18
April 1944: "I shall explain the Snow Man as an example of the necessity of
identifying oneself with reality in order to understand it and enjoy it." (Keyser
1981: 119)
The New Critics of the early to middle twentieth century generally urged that
readers view a poet's own comments about his poems, especially brief and
cryptic comments in personal correspondence, with the utmost skepticism:
We cannot really know how much thought or knowledge was assumed as the
context for the letter, and we cannot know whether Stevens willfully, or
perhaps unknowingly, reduced the complexity of his own poem for his cor-
respondent's sake. Nevertheless, not heeding this principle, Keyser takes the
beginning of the poem, the words One must, to indicate that the entire poem
is an admonishment to the reader, a behest telling the reader that he or she
must be or do as follows. Granted, the bulk of Keyser's comments focus—
correctly—on the remarkable structure of the poem's single sentence, which
is built so that the reader will complete a given portion and think that that is the
whole of it, only to keep reading and find out that what he or she had thought
to be a whole turns out to be only a part of a much more complex sentence.
Despite this fine insight, ironically, Keyser still allows the apparent imper-
ative form of the opening, One must, to dominate his sense of the whole. He
treats the complex structure of the sentence as an expressive ornament to
dramatize the sense of the original imperative, in effect, "One must be like a
snow man." Keyser therefore misses the point that the sentence turns out to
have an altogether different, surprising meaning when read as a whole. The
sentence begins as an apparent imperative but ends up actually being more
like a complex conditional-resultative sentence.
It helps if we reduce the poem's sentence to something of a skeletal para-
phrase. We can take advantage of the poem's many parallelisms—such as the
parallel of regard in the first tercet and behold in the second, and the frost and
the boughs and pine trees crusted with snow within the first tercet, as well as
both of these with the junipers shagged with ice and the spruces rough in the
distant glitter I Of the January sun.
One must be a creature of winter to perceive wintry things and not perceive
emotion (misery) in any of them; only a snow man, who is nothing himself (not
a perceiving subject), can (as it were) perceive nothing that is not there and the
nothing that is.
The sentence thus has an idiomatic structure something like that of the sen-
tence, "You have to be heartless to know her and not love her." This latter
does not urge someone to be heartless; it explains what it would have to take
to avoid loving a lovable person.
Stevens's sentence is not a recommendation to become like a snow man,
which would be impossible. If it were possible for the reader to become like a
snow man, that would mean that the reader would bring nothing to the reality
he or she perceives—would impose no conceptions such as "misery in the
sound of the wind." But that would also have to mean that he or she would not
bring any knowledge to bear on perception, even the perception of the marks on
the page that are the letters and words of the poem. In short, a snow man cannot
read poetry. So, if the reader were (hypothetically) successful at becoming a
snow man, he or she would do so only at the expense of his or her ability to read
the poem or even to know that he or she has succeeded. The snow man is
nothing himself. Even before this, the poem gives us a strong clue in the phrase
equivalent to have a mind of winter, that is, have been cold a long time. The
Poetry 119
latter phrase sounds like a reference to the dead. This does in fact make sense
later on, once the reader figures out that the snow man is not a human being
and that one would have to be dead—be nothing—to "see" the world the way
"he" does—that is, not see at all. The poem then works, as it happens, as a
nice allegory for one of the main points in this chapter: the importance of the
extrinsic literary and cultural contexts brought to bear on the literary text.
The surprising shifting of the sentence in Stevens's "The Snow Man" is
not a demonstration of the openness that being a snow man would entail: We
cannot possibly think like a snow man, because a snow man does not, by
definition—even the poem's definition—think. Rather, it is precisely the turn,
from apparent injunction to actual observation about the nature of human-
ity, that surprises and shocks the reader out of the complacency of his post-
Romantic expectations. A commonplace of Romantic thought was and con-
tinues to be the idea that humanity is alienated from Nature, in permanent
exile, because of its self-consciousness. The very things that make us hu-
man—language, reason, morality, civilization—also make us somehow sep-
arated from Nature and fill us with anxiety and a sense that we do not belong
where we are. With this idea comes a yearning for the blissful state of oneness
with Nature. To put the thought simply (at the risk of caricature), squirrels
are lucky because they never have to worry about whether they are doing the
right thing and never have to wonder about the meaning of life. But whereas,
in the Romantics, the object of such yearning might be a bird (such as Keats's
nightingale), which may or may not in fact have some awareness of its own
mortality (for all we know), Stevens's snow man is most certainly without
any awareness of anything, because it is not a living being at all. Indeed, even
the name for it, snow man, is the product of human projections and hu-
man creation: It is only figuratively a "man" made of snow. The un-
marked condition in this case is, precisely, the expectation that one must
have a mind of winter and be like the snow man. The surprise comes when
we realize that, far from having to be a snow man, one has no possible
choice of being a snow man, because being one would be neither being nor
choice at all.
There is, indeed, a deeper paradox in the poem, however. The common
human attributes that include the hearing of misery in the sound of the winter
wind also include our ability to contemplate what it would be like if we were not
ourselves. Imagining other subjectivities is part of the human condition
because it is one of the capabilities of imaginative projection, and this in-
cludes at least trying to imagine impossible subjectivities, beings that cannot
possibly be subjects at all, such as the snow man, "nothing himself." The
poem does not so much tell us that this is an aspect of the human condition as
allow us to experience this paradoxical fact as another result of the surprise that
the poem delivers. Again, the poem's "meaning"—that is, its point-—is insep-
arable from its aesthetic effect, its surprise, and its paradox.
Like Epstein, Keyser makes his mistake precisely because he expects
the poem to express the author's ideas or feelings as a communicative

utterance. This renders him insensible to the possibilities of surprise on the
level of ideas, even though Keyser is quite sensitive to surprises in syntax. He
assimilates the surprises of syntax to an incorrect reading of the content
because he assumes a simplistic reading of the content to begin with and then
assumes that the syntactic patterning will enhance or boost the message. This
logic is very similar to that of Renaissance theorists who viewed such
things as verse or poetic imagery as a kind of rhetorical "garb" to "clothe"
the truth carried within the text as its message. Then as now, this confu-
sion results in a tendency to read parts at the expense of the whole.
Perhaps most fascinating here is how the poem itself seems to prepare the
way for the mistaken reading—not only by setting up the surprise but also by
then addressing the very questions about reading that the mistake—and
the surprising turn to correct it—would naturally pose. The poem pre-
sents to us the paradox of how, like reading, human perception always
includes projection, but our ability to project can also include our imag-
ining impossible nonsubjects as test cases, nonsubjects who, if they could
exist, would not project when they perceived. The reader who most "truly"
understands the text is the one who projects nothing at all, but this would
mean that he or she actually understands nothing at all and is not in fact
reading. Reading always involves distortion, even though its aim is the ab-
sence of distortion. It is this very paradox that keeps the poem from being
either a simplistic recommendation to a Zen-like nonprojection or an equal-
ly simplistic celebration of human projections. Rather, it dwells on the paradox
that makes (perfect) reading simultaneously possible and impossible. It takes a
good enough reader to recognize, ultimately, that the poem is also saying that
the perfect reading is impossible—and yet that is the perfect reading, because no
other reading will do. This self-reflexivity—the poem allegorically representing
how poems work—rather than the mimesis of imitative form that Epstein
calls self-reflexivity, accords best with literariness. For when the utterance is
removed from normal communicative and referential contexts, it can only
ultimately refer to itself and illustrate the paradox of language that only
appears to communicate—having no communicative exchange to make it do
so; and yet, in a sense, that very paradox is what the text "communicates."
Every literary text approaches this paradox of self-reflexivity in a different
way, but the ultimate pleasure of surprise and fascination with the paradox
of a language that has the substance but not the social circumstances of
communication remains always at the end of the road of interpretation.
CONCLUSION
As we have seen, literary texts, including poems, have the function of deliv-
ering surprise, provided the reader can recognize the appropriate contexts as
the unmarked conditions against which the texts' features are marked. This
Poetry 121
means that literary pleasure presupposes by its very nature literacy—or what
Culler (1981) aptly calls literary competence. Literary competence is not a
stable state, however, but an infinitely progressive accumulation. Reading is
a lifelong apprenticeship in the art of literary competence, as an increasing
number of texts becomes increasingly accessible to the studious reader. But
what about readers who aren't there yet? How does a reader get started? Are
we not faced with a closed loop, where the pleasure of reading requires prior
reading, which requires the capacity for the pleasure of reading?
The pleasure of literary texts, fortunately, is available to various degrees
on various levels simultaneously. At the most rudimentary, the necessary
context of the appropriate "unmarked" default expectations comes from
very broadly shared cultural traditions, ideas, beliefs, concepts, or images.
For instance, almost everyone who has learned to read in a Western country
will recognize the collocation of a tree with a dangerous fruit and a garden
into which the foe steals as some sort of allusion to the Fall in the Garden of
Eden, because this web of myths is so widely known in some form or other
throughout Western culture, even among people who never read the Bible or
Milton's Paradise Lost. In the case of "The Tyger," the incoherency among
the specific divine figures to which the questions allude would require a fair
bit of prior reading. But the fact that the speaker utters nothing but questions,
and that the questions, while containing sufficient information to be rhetor-
ical, do not seem to permit the very answers that they seem to prompt—these
facts should be evident to a reader attentive to mere syntactic and rhetorical
conventions and only the vaguest and broadest cultural beliefs or traditions
about creation and divinity. In the case of "The Snow Man," knowledge of
the Romantic tradition of yearning for an unalienated unity with Nature only
provides extra elaboration beyond the more basic point, which is that the
sentence winds up not urging the reader to be like a snow man but rather says
what being a snow man would be like, if it were possible (and it turns out to
be, paradoxically, quite impossible). The reader who gets this point only need
read the sentence as a whole rather than stopping too soon, and need think
about the logical implications of what it would mean to urge someone to be
"nothing himself" and therefore unable to perceive anything at all, including
the poem making this supposed behest. Literacy in both of these instances
requires, above all, that the reader read the entire poem through, all of its
sentences, trying to see a pattern in the whole and only then consider how
each part contributes to that whole rather than moving sequentially and
making controlling inferences at every point from limited information. The
procedure of accepting parts of the whole as dominant—without sufficient
attention to possible irony—characterizes normal communication rather than
the reading of literary texts. A literary text is first and foremost a whole, an
aesthetic pattern of effects rather than a sequence delivering a message.
The reader must attend to the literary contexts that enable a literary text
to deliver surprise by means of marked effects, and the reader must view the
entire poem as a completed whole, a pattern whose function is to deliver that

surprise. The surprises of literary texts form aesthetically satisfying patterns,
and the reader's discovery of a satisfying pattern is, itself, the major pleasur-
able surprise that reading and interpreting a literary text provide. These two
principles, context and pattern, can guide the new reader of literature from
the outset. But it is the very nature of literary texts, because they draw on
literary contexts to condition markedness and form patterns of surprise, that
reading leads to more reading, and more reading deepens and enriches the
experience of rereading familiar literary texts, making them forever new. In
a sense, reading has no beginning and no end. It has no end because the
addition of more literary contexts will always enrich the experiences of marked-
ness, pattern formation, and surprise. And it has no beginning because one
must always have some sort of prior knowledge, even if only rudimentary, to
find pleasure in reading a literary text. Into this beginningless and endless
labyrinth of texts, the teacher leads new readers, for literacy cannot be sep-
arated from pedagogy. The first step in this journey into the world of literary
texts is to understand the difference between texts and normal communica-
tion and how literariness conditions the proper approach to literature.
REFERENCES
Austin, Timothy R. 1981. Constraints on syntactic rules and the style of Shelley's
"Adonais": an exercise in stylistic criticism. In Essays in modern stylistics, ed.
Donald C. Freeman, 138-165. London: Methuen.
Aviram, Amittai F. 1994. Telling rhythm: body and meaning in poetry. Ann Arbor:
University of Michigan Press.
Blake, William. 1789-1797. Songs of innocence and of experience. Facsimile reprint,
New York: Oxford University Press, 1967, with introduction and notes by Geof-
frey Keynes.
Bloom, Harold. 1971. The visionary company. Ithaca, N.Y.: Cornell University Press.
Culler, Jonathan. 1981. Literary competence. In Essays in modern stylistics, ed. Donald
C. Freeman, 24-41. London: Metheun.
Dickinson, Emily. 1960. The complete poems of Emily Dickinson, ed. Thomas H.
Johnson. Boston: Little, Brown.
Epstein, E. L. 1981. The self-reflexive artefact: the function of mimesis in an approach
to a theory of value for literature. In Essays in modern stylistics, ed. Donald C.
Freeman, 166-199. London: Methuen.
Fairley, Irene R. 1981. Syntactic deviation and cohesion. In Essays in modern stylis-
tics, ed. Donald C. Freeman, 123-137. London: Methuen.
Freeman, Donald C. (ed.). 1981a. Essays in modern stylistics. London: Methuen.
. 1981b. Keats's 'To Autumn': poetry as process and pattern. In Essays in
modern stylistics, ed. Donald C. Freeman, 83-99. London: Methuen.
Jakobson, Roman. 1960. Concluding statement: linguistics and poetics. Style in lan-
guage, ed. Thomas Sebeok, 350-377. Cambridge, Mass.: MIT Press.
Poetry 123
. 1968. Poetry of grammar and grammar of poetry. Lingua 21: 597-609.
. 1973. Principes de versification. In Questions depoetique, ed. Tzvetan Todorov,
40-55. Paris: Seuil.
Keyser, Samuel Jay. 1981. Wallace Stevens: form and meaning in four poems. In
Essays in modern stylistics, ed. Donald C. Freeman, 100-122. London: Methuen.
Kiparsky, Paul. 1981. The role of linguistics in a theory of poetry. In Essays in modern
stylistics, ed. Donald C. Freeman, 9-23. London: Methuen.
Mukarovsky, Jan. 1948. Standard language and poetic language. In The critical tra-
dition, ed. David H. Richter, 860-868. Reprint, New York: St. Martin's Press,
1989.
Plato. 1982. Phaedrus. Plato in twelve volumes, Vol. I, ed. and trans. Harold North
Fowler, 405-579. Cambridge, Mass.: Harvard University Press. (Numbers and
letters in brackets are the canonical form of citations to standard editions of Plato.)
Stevens, Wallace. 1954. The collected poems of Wallace Stevens. New York: Random
House-Vintage.
Tarlinskaya,Marina. 1987. Rhythm and meaning: "rhythmical figures" in English
iambic pentameter, their grammar, and their links with semantics. Style
21(l):l-35.
— 7—
Villainous Boys: On Some Marked

Exchanges in Romeo and Juliet
TREVOR HOWARD-HILL
T HE FUNDAMENTAL PRINCIPLE of the markedness model, the negotiation

principle, can apply to literary works of the early modern period as well
as to conversations:
Choose the form of your conversation contribution such that it indexes the set
of rights and obligations which you wish to be in force between speaker and
addressee for the current exchange. (Myers-Scotton 1993: 113)
In this formulation the "conversation contribution" is the literary work that
an author wishes to contribute to the literary tradition, the "form" is the
mode (fiction/nonfiction, poetry, or prose) and genre in which the work is
written, the "set of rights and obligations" are the conventions of the genre
that establish the reader's expectations, and the speaker and addressee are the
author and reader, respectively. On the level of the literary work the signif-
icant distinction between the principle's application to literature and conver-
sation seems to be that there is only a single speaker (collaborative works
aside) and that therefore there is no exchange: A literary composition is not
a conversation.1 However powerful they may be in determining the success or
failure of the author's work, or assigning meanings to it, readers have no
power to alter the form of words before them or to affect the progress and
outcome of the literary composition.
Nevertheless, most literary works contain conversations in which the
negotiation principle and markedness may be observed in operation. Indeed,
plays in particular may be considered simply a sequence of speeches and
1. The associated maxims on codeswitching (CS) as an unmarked choice (Myers-Scotton

(1993: 114), CS as a marked choice (131), CS as an exploratory choice (142), CS as a deferential
strategy (147), and the virtuosity maxim (148) are less readily applied to literature. The first two,
however, seem relevant to the initial choice of genre and a change of genre (or dominant style within
a work), and the exploratory maxim may relate to "experimental" literature.
124
Romeo and Juliet 125
conversations, providing intrinsically fertile ground for the consideration of
the markedness model. Without enlarging the theoretical context further, in
order not to repeat remarks made elsewhere (Howard-Hill 1996), I selected
Shakespeare's Romeo and Juliet as a play likely to be widely familiar to
illustrate the application of the model to drama.
SHAKESPEARE'S BOYS
Shakespeare's plays are full of boys, natural parts for the youths who were
apprenticed to master actors and usually played the female leads.2 Among
them are great lubberly boys, wimpled, whining, purblind, wayward boys
(Cupid), little scrubbed boys, peevish and prating boys, changeling boys,
very dishonest paltry boys, and even sweet and lovely boys. As well, there
are more than a few unembellished usages of the noun. Of the 455 occur-
rences of boy or boys in Shakespeare's plays, two- thirds (297) are vocatives,
apparently used to refer to servants, though some of them could be marked
in the particular situation. About a fifth of these have qualifying adjectives
but some twenty-one of them (my boy, sir boy) can be set aside, leaving only
a small number (37) that are more likely to display markedness in conversa-
tional situations.3 These marked instances are fairly evenly divided between
cases with derogatory and cases with approbatory adjectives, an observa-
tion that suggests that Shakespeare himself had no particular attitude to
boydom: The dramatic situation of the moment seems more likely than the
playwright's predisposition to determine how speakers characterize boys.
And many of the nonvocative references to boys are objective, occurring in
apparently neutral contexts.4
On the other hand, outside Romeo and Juliet, in two cases young male
protagonists are designated as boys in conspicuously contemptuous terms.
In All's Well That Ends Well, Bertram, the young Count Rossillion (the most
unsympathetic character ever rewarded with a Shakespearian heroine), is
characterized as a proud scornful boy by the King of France (2.3.151). Even
his mother describes him as a rash and unbridled boy (3.2.28), and to the
2. Spevack (1969) records 352 occurrences of boy, 1 hyphenated compound, 9 genitives,
and 93 occurrences of boys in the plays. Romeo and Juliet contributes 11 occurrences of boy,
2 of boys. For convenience, the text is quoted from the Riverside Shakespeare (1974) used by
Spevack.
3. Even though dramatic speeches usually have on-stage auditors, substantial parts of them
may offer no opportunities for negotiations, in which markedness is present, to occur. There-
fore, instances of a form of address, boy (like man, woman, child), when obviously a speaker
is interacting with an auditor or auditors, identify the situations in which markedness is most
likely to come into play.
4. This observation, like others to this point, is made from study of the concordance rather
than analysis of their broader contexts in the individual plays. Nonvocative uses of boy are not
ignored here; indeed, often the noun with negative attributes occurs to reinforce the marked
vocative usage.
parasite Parolles he is a foolish idle boy (4.3.215) and a lascivious boy (4.3.220,
300). The general feeling, both within the play and in modern criticism, is
that the epithets are just.
More significant for consideration of the earlier Romeo and Juliet is the
case of Coriolanus. Coriolanus is presented as a heroic but intemperate char-
acter who turns against Rome even to the extent of allying himself with his
rival, Aufidius, Rome's enemy. When at last the Romans send out Coriola-
nus's mother, wife, and son to plead for the city to be spared, he yields.
Aufidius, envious of the honor Coriolanus garnered among the Volscians, is
already plotting Coriolanus's assassination when he reports that he has made
a treaty for peace with Rome and accuses Coriolanus of betraying his allies:
Auf. . . . He has betray'd your business, and given up,
For certain drops of salt, your city Rome,
I say "your city," to his wife and mother,
Breaking his oath and resolution . . .
. . . at his nurses's tears
He whin'd and roar'd away your victory . . .
Cor. Hear'st thou, Mars?
Auf. Name not the god, thou boy of tears!
Cor. Ha?
Auf. No more.
Cor. Measureless liar, thou hast made my heart
Too great for what constrains it. "Boy"? O slave!
[8 verses omitted]
. . . "Boy," false hound!
If you have writ your annals true, 'tis there
That, like an eagle in a dove-cote, I
Flutter'd your Volscians in Corioles.
Alone I did it. "Boy"! (5.6.91-115)
Indignation spurs Coriolanus to repeat the offensive boy incredulously three
times among a scattering of equally derogatory terms directed at his rival:
liar, slave, cur, hound. Clearly even from these few examples, in Shakes-
peare's world the term boy was polyvalent, attracting derogatory and ap-
probatory qualifiers according to the social situation of its use but also, even
without immediate qualification, possessing either a neutral or offensive force;
that is, a simple (unqualified) use of the term boy may be marked or unmarked.
THREE CRUCIAL DRAMATIC MOMENTS IN "ROMEO AND JULIET"
In Romeo and Juliet the marked usage of boy supplies the affective basis of the
mechanism of the revenge plot on which the play depends.5 Three crucial dra-
matic moments in Romeo and Juliet embody an interesting pattern of the marked
5. It is imprudent to claim any novelty of observation in Shakespearian criticism. If a reader
knows of a discussion of this topic that has not come to my attention, I should very much like to
learn of it.
sense of boy together with significant repetition of villain. The main purpose of
this chapter is to consider boy and associated expressions in relation to the
dramatic situations in which they occur, with the aid of the markedness model.
Certainly, Tybalt's speech threatening a disastrous outcome (1.5.89-92,
quoted later) marks the boundary between the protasis and the epitasis, the
crucial point at which the introductory development translates to dramatic
action. One might even assert against the common view that it is not the
families' feud that kills the young lovers (and three other young men in the
play) but rather, the simple monosyllable boy spoken in anger, piercing the
heart of the young man to whom it was addressed, Tybalt. Except for Benvo-
lio, reserved to sum up at the end of the action, all the young men and women
in the play are killed by themselves or each other because of a word. 6
Shakespeare does not reveal why the Capulets and Montagues are at
odds, but it seems that the feud is conducted on the fringes, as it were, of
the principal families of Verona, by servants and by such zealots as the
fiery Tybalt (1.1.109), not the son but apparently the leading young
gentleman of Capulet's clan. Shakespeare assigns each of the principal
young men a somewhat limited function in the play, to embody a major
attribute of the conventional character of the young man as it developed from
classical drama through the sixteenth century. 7 Benvolio ("well-wisher")
appears as the good friend; Mercutio as the sociable, volatile humorist; Romeo
the young man as romantic lover; and Tybalt, the young man, as the upholder
of honor. In the first scene Shakespeare sketches Tybalt's character econom-
ically according to conventional Renaissance delineations of character. Ty-
balt reveals his essential quality and foreshadows his function in the play in
his first speeches at his entrance into the fray between the servants of the
Montagues and Capulets:
Tyb. What, art thou drawn among these heartless hinds?
Turn thee, Benvolio, look upon thy death.
Ben. I do but keep the peace. Put up thy sword,
Or manage it to part these men with me.
Tyb. What, drawn and talk of peace? I hate the word
As I hate hell, all Montagues, and thee.
Have at thee, coward. (1.1.66-72)
Besides his bellicosity, Tybalt's use of epithets to characterize his opponents
6. This is not to neglect the influence of the circumstances in which the word is spoken, as I
mention later.
7. See Herrick ([1950]1964: 150-154) Young man. Among other qualities, the young man is
amorous, does everything to excess, is ambitious for honor, and is sociable and fond of laughter.
8. Besides villain, mentioned below and in note 14, see Tybalt's heartless hinds which can be
rendered as cowardly servants; for hind OED records a transferred meaning of "rustic, a
boor." Folk etymology might have associated coward and cow-herd in Shakespeare's time
when (unusually) they shared similar spellings. Later in the play, "slave" (1.5.55) meaning "a
servant completely divested of freedom and personal rights" is also "used as a term of contempt"
(OED I.1.b.).
as social inferiors is obvious.8 In the three scenes of the play in which he

appears (I discuss the other scenes later), Tybalt is depicted as being aggres-
sively conscious of his social standing.
More remarkably, within a few verses, Shakespeare begins a series of
uses of the term villain that can be shown to be literally marked. At line 79,
the head of the opposing house enters. The first words Montague (Romeo's
father) utters, in the inflamed Tybalt's presence to the head of his house,
are "Thou villain Capulet!" Villein (OED sb.1. "One of the class of serfs
in the feudal system") clearly contributes to villain (OED sb. l "Origi-
nally, a low-born base-minded rustic; a man of ignoble ideas or instincts.
. . . a. Used as a term of opprobrious address."). The next four occurrenc-
es of the word (1.5.62, 64, 75, 3.1.61) come significantly from Tybalt's
mouth to characterize Romeo, the heir of the house of Montague.
Later, at Capulet's "old accustomed feast" (1.2.20), Tybalt is present
as, apparently, the youngest adult male of his family.9 Although uninvit-
ed, Romeo accompanies Mercutio in order to see Rosaline but immedi-
ately falls in love with Juliet. Tybalt overhears his infatuated couplets:
Tyb. This, by his voice, should be a Montague.
Fetch me my rapier, boy. What dares the slave
Come hither . . .
Now, by the stock and honor of my kin
To strike him dead I hold it not a sin. (1.5.54-59)
Tybalt appoints himself as protector of his family's honor and appealing to
the head of his house, old Capulet, expects to be answered with similar
devotion to the family's interests. Capulet, however, knows nothing bad
about Romeo, "a virtuous and well-governed youth" (68). ("Youth" here is
obviously not derogatory).
It is my will, the which if thou respect,
Show a fair presence and put off these frowns,
An ill-beseeming semblance for a feast. (72-75).
To his chagrin, Tybalt finds that his well-intentioned notice of Montague's
presence is rejected and, further, that his demeanor is criticized. He has
assumed that his uncle shared his concern for the family's honor and that he
himself had a special responsibility to redress the injury of Romeo's pres-
ence. Injured self-esteem is a potent cause of anger and motive for revenge,
9. Tybalt is the son of Lady Capulet's brother (3.1.146) and Capulet is his uncle by marriage;
he has a younger brother, Valentine, who has no part in the play. The Capulets themselves have
no other child than Juliet (3.5.164-165). The point needs to be established because (1) with the
death of Tybalt and later of Juliet, Capulet loses his heirs as the Montagues lose theirs on
Romeo's death, ensuring the extinction of both households; and (2) Shakespeare may have
intended Tybalt to regard himself as having a special responsibility to protect the family's
honor. (Shakespeare establishes the general agedness of the guests—except for Juliet and the
mute Rosaline—by Capulet's conversation at the beginning of 1.5.)
as Campbell ([1930] 1970) pointed out for King Lear.10 With rising anger,
Tybalt sets his judgment against his uncle's and reacts forcefully:
Tyb. It fits when such a villain is a guest.
I'll not endure him. (75-76)
Capulet's response is immediate and predictable: Old men, as the conven-
tional theories of character have it and experience confirms, are quickly moved
to anger when the young go against their will.
Cap. He shall be endured.
What, goodman boy? I say he shall, go to!
Am I the master here, or you? go to!
You'll not endure him! God shall mend my soul,
You'll make a mutiny among my guests!
You will set cock-a-hoop! you'll be the man!
Tyb. Why, uncle, 'tis a shame.
Cap. Go to, go to,
You are saucy boy. (76-83)
Capulet's speeches here are marked rhetorically by the short periods and the
profusion of exclamation and interrogation marks (mostly supplied by the
modern editor). The ten verses of Capulet's previous speech (1.5.65-74)
advising Tybalt to tolerate Romeo's presence contain five periods (here de-
fined as sequences of words terminated by a semicolon, a period, an excla-
mation mark, or a question mark) of an average length of sixteen words. The
first ten verses of Capulet's following speech (1.5.76-85) contain fourteen
periods, averaging 5.42 words in length. An actor does not have to supply
passionate intensity to these lines: Shakespeare's language virtually directs
him to deliver the speeches in the appropriate manner.
Rendered economically by Shakespeare's mastery, the psychological mech-
anism that drives Tybalt's speeches here is readily recognized. Tybalt as-
sumes that he and Capulet share a distaste for Romeo's presence at the feast
and expects that Capulet would respond to the notice of Montague's atten-
dance with pejorative speeches about Romeo, language like his own. But
Capulet responds in a different register. The breakdown in the negotiation
of a common style of speaking occurs when Tybalt refuses to take Capulet's
lead, and it is represented dramatically by Capulet's resort to abusive speech,
particularly the marked use of the usually denotative boy.
Capulet calls Tybalt goodman boy and saucy boy and mocks his pretensions
10. Describing an Elizabethan philosophy of anger descended from Aristotle {Rhetorica, 1378a,
31-34), Campbell uses words that apply well to Tybalt:
We see [him] demonstrating what we know to have been an accepted principle, that a man is
angered by an injury to his self-esteem, that he is soonest angered when that respect in which he
has thought himself most worthy seems to be disregarded, that he is soonest angry . .. with those
who have previously treated him becomingly and now change, and with those who do not
appreciate his kindness. (Campbell [1930]1970: 185-186)
to manhood: You'll be the man! (81), thus explicitly marking out the domain
of his insults.11 In Capulet's eyes, and in his speeches, Tybalt has fallen from
the prized state of manhood to unprivileged boyhood. Shakespeare elsewhere
remarks of Cupid, his disgrace is to be called boy, but his glory is to subdue
men [Love's Labor's Lost, 1.2.179-182). Humiliated publicly, in the midst
of his kinsmen, but unable to challenge the patriarch of his family, Tybalt
cannot subdue men. He can, however, challenge his coeval, Romeo, who
supplied the occasion of his disgrace.
Tyb. Patience perforce with willful choler meeting
Makes my flesh tremble in their different greeting.
I will withdraw, but this intrusion shall,
Now seeming sweet, convert to bitt'rest gall. (89-92)
The quatrain form and the alliteration and assonance that contribute to the
intensity of this speech mark its significance for the prosecution of the action.
Within a short time, Benvolio and Mercutio know of Tybalt's letter to Romeo
challenging him to a duel (2.4.6-7). Interspersed with scenes of romantic
lyricism, the tragedy now surges toward its catastrophe.
On his third and final appearance in the play (in 3.1), Tybalt addresses
Benvolio and Mercutio correctly (and neutrally) as gentlemen (38). Although
the play is set in Verona, Shakespeare observed his usual practice of establish-
ing the ranks and familial relationship of the characters with the terms famil-
iar to his London audiences: He did not attempt to supply an Italian veneer
with such terms as signior. Consequently, the speeches are sprinkled with the
common terms of address: sir, sirrah, my lord, lady, madam, nurse (for the
Nurse) and father (for Friar Lawrence), as well as nephew, niece, uncle,
cousin, mother, and daughter often used vocatively between relatives. Such
expressions, like Tybalt's gentlemen are the tokens of common civility and
relationship and therefore are usually not marked. However, on Romeo's
entrance Tybalt applies to him in person the derogatory epithet by which he
earlier described Romeo three times, the characterization that Capulet reject-
ed: Romeo . . . thou art a villain (60-61). Moreover, Mercutio passes on to
Romeo the even more strongly marked term by which Capulet so humiliat-
ingly addressed him, challenging him to fight with boy (66). The markedness
of this word derives first from the dramatic circumstances of its application
to Tybalt in 1.5 as well as the contrast with the socially correct term of
address used to the other young men in this scene. From the beginning of their
exchange here Romeo and Tybalt were talking at cross purposes: Tybalt
wanted Romeo to fight him and Romeo wanted Tybalt to regard him as a
member of his family (on account of his marriage to Juliet in the previous
scene). The use of boy here then probably marks the boundary of Tybalt's
willingness to negotiate with Romeo, but we cannot be sure. Romeo's conse-
11. G. B. Evans remarks that "goodman boy is a double-barreled insult to Tybalt, goodman
being a title inferior to that of gentleman." (Riverside 1974: 77 note)
quent soft reply infuriates Mercutio who draws on Tybalt, bringing about his
own death. At last Romeo, a reluctant revenger, fights Tybalt:
[Row.] Now, Tybalt, take the "villain" back again
That late thou gavest me. . . . (3.1.125-126).
The modern editor makes the markedness of the expression transparent in his
text.12 Initially linked with boy in scene 1.5, its force continues through the play:
Tyb. Thou wretched boy, that didst consort him here,
Shalt with him hence.
Rom. This shall determine that.
They fight; Tybalt falls. (130-131)
There remains only one more significant use of boy to notice, a coda to
the plot line of Tybalt's revenge. The boy Romeo received from Tybalt at
3.1.129 before the duel, he passes on to Paris as he seeks entrance to the
Capulet family tomb.
[Par.] Stop thy unhallowed toil, vile Montague!
Can vengeance be pursued further than death?
Condemned villain, I do apprehend thee.
Obey and go with me, for thou must die.
Rom. I must indeed, and therefore came I thither.
Good gentle youth, tempt not a desp'rate man. (5.3.54-59)
Once more addressed as villain (56), Romeo returns good gentle youth (59,
youth 61) to the challenger, whom he does not recognize.13 But when Romeo
is brought to the point of fighting, in his speech the gentle youth is demoted
to a mere boy: Then have at thee, boy! (70).
THE USE OF "BOY" AND "VILLAIN"
The distribution of boy and villain in the play, particularly their collocation
in crucial episodes, demonstrates that their usage is marked by design rather
than by statistical accident. Villain in particular occurs seventeen times in the
play. It is first used of Capulet by Montague in Tybalt's presence, applied
four times to Romeo by Tybalt, and it is twice rejected or returned to Tybalt
by Romeo (in 3.1, where also Mercutio uses it twice, once neutrally and once
of Tybalt). On learning of Tybalt's death at Romeo's hands, Juliet debates
with herself the respective shares of Tybalt's and Romeo's villany in 3.2,
using the term three times, and again with Lady Capulet in 3.5, where the
term occurs four times in three verses:
12. The same editor notes of villain that it is "A very serious insult demanding reprisal,
carrying not only the sense of 'depraved scoundrel' but undertones of 'low-born fellow.'" (Evans
1984: 123)
13. Romeo's reference to himself as "man" (59) is generic and neutral in the context.
[La. Cap.] . . . the villain lives which slaughter'd him.

Jul. What villain, madam?
La. Cap. That same villain Romeo.
Jul. [Aside.] Villain and he be many miles asunder. — (3.5.79-81)
Like others of her clan, Juliet knows how important names are: It was she
who first raised the question of nominal value.14 Finally, as I just mentioned,
villain is coupled with boy in the last scene of the play.15 In this play, then,
villain is not a general term of abuse but, rather, a marked one.
CONSIDERATION OF "CONTEXT" IN DRAMA
Whether or not the foregoing discussion correctly illustrates the application

of the markedness model to a literary work, it does raise larger issues that
deserve consideration. The social context in which the early modern play-
wright conducts his "conversation" with the reader or spectator is rhetorical.
(By "social" I mean the whole complex of factors that govern the relationship
of an author and his public in a particular time.) More narrowly, his linguistic
choices are constrained by the initial choice of genre, as was foreshadowed
earlier by the initial mention of the negotiation principle. By undertaking to
write The Most Excellent Lamentable Tragedie, of Romeo and Juliet (to give
the title of the first, 1599, edition), that is, by selecting the genre of tragedy,
Shakespeare undertook to make certain linguistic choices. He knew also that
those among his audience who cared would have certain stylistic expecta-
tions. As well as the decorum of character that I already mentioned, there is
decorum of language: speech or style—register in a sense. The style of speech
is expected to be appropriate both to the social status of the character and
also to the specific context, emotional, dramatic, and so on. A play by one of
Shakespeare's early contemporaries affords a striking illustration of stylistic
variation. In Thomas Kyd's The Spanish Tragedy (1587?) the elderly Lord
Marshall (chief judge) Hieronimo laments the murder of his son in highly
wrought formal verses:
O eyes, no eyes, but fountains fraught with tears;
O life, no life, but lively form of death;
O world, no world, but mass of public wrongs,
Confus'd and fill'd with murder and misdeeds;
[17 verses omitted]
14. In fact, an important question in the play for the Capulets is precisely what name Romeo
should bear: hence Juliet says, O, be some other name! I What's in a name? That which we call
a rose / By any other word would smell as sweet;/So Romeo would, were he not Romeo call'd,
/ Retain that dear perfection which he owes / Without that title. Romeo, doff thy name, / And for
thy name, which is no part of thee, / Take all myself. (1.2.42—49).
15. To summarize, villain occurs at 1.1.79, 1.5.63, 1.5.65, 1.5.75, 3.1.61, 3.1.64, 3.1.94,
3.1.102, 3.1.125, 3.2.79, 3.2.100, 3.2.101, 3.5.79, 3.5.80, 3.5.81, and 5.3.56.
Eyes, life, world, heavens, hell, night, and day,
See, search, shew, send, some man, some mean, that
may— A letter falleth.{Edwards [1959] 1969:3.2.1-4,22-23).
But when a letter falls in his path, surprised and suspicious, his style of speech
alters remarkably:
What's here? A letter? tush! it is not so:
A letter written to Hieronimo. (24-25).
Similarly (returning to Romeo and Juliet), though the abrupt change in
Capulet's speech in 1.5 with its short, angry phrases is stylistically significant,
that does not mark it for auditors or readers: It is the appropriate style of
speech for an angry man and therefore should be considered the unmarked
choice.16 Stylistic variation, therefore, may not invariably produce marked-
ness although, as in this instance, it may be remarkable. That is, variation is
only marked when it brings in a style that is unexpected or inappropriate for
the relationship depicted.
DO DRAMATIC CHARACTERS MAKE CHOICES?
Not all the speeches in plays are conversations, but the conversations are the
obvious site for application of the markedness model. Nevertheless, any-
one who analyzes conversations in plays must be uncomfortably aware
of the influence of the author. The exchanges between speakers that make
ordinary conversations worth overhearing become, in light of the negotia-
tion principle, curiously uninteresting. The exchanges are not dynamic for
the speakers except within the scope predetermined by their creator. In fact,
they are not conversations at all but, rather, the representations of conversa-
tions, in plays selectively scripted with greater or lesser skill by the play-
wright. Generally, they lack the mumblings, false starts, hesitations, ir-
relevancies, repetitions, syntactical errors, or confusions of conversa-
tions in ordinary life, and rarely do characters on a stage speak at the
same time as other speakers or interrupt the speeches of other characters.
The dramatic speakers are not "rational actors" in the usual sense, possessing
the possibility to make linguistic choices as conversations continue, but, rath-
er, they are puppets whose linguistic constructions are governed by such
matters as the initial choice of genre, the principle of decorum of lan-
guage within the general theory of genre, and the specific purposes of the
playwright.
Consideration of the relevance of such matters may also bring to question,
in their relation to early modern literature, principles as fundamental as Grice's
(1975) maxim of quantity. They also identify fruitful areas of investigation
16. What would mark Capulet's speech, difficult though it might be to conceive it, would be
an orotund, expansive manner of speech employing long periods and ornamental figures.
beneficial to students of linguistics and literature alike.17 For instance, rhetor-

ical embellishment and ornamented speech were thought desirable for a large
part of European history. The plain style is a style in early modern practice.
Consequently, there may be a paradoxical reversal of modern readers' expec-
tations, so speeches where Grice's maxim pertains may be strongly marked
for a modern reader. For instance, when Hamlet declares that he had never
loved Ophelia (Hamlet, 3.1.118), Ophelia sheds her previously artificial style
of speaking: I was the more deceiv'd(\\9), one of Shakespeare's telling "qui-
et" speeches. More pertinently, in Romeo and Juliet, on learning the false
news of Juliet's death, her usually prolix lover speaks only the plain but
memorable Is it e'en so? Then I defy you, stars! (5.1.24). On the other hand,
even though readers or spectators might form some expectation—from their
understanding of the characters in the plays—of the ways in which Ophelia
or Romeo would react to those speeches whose function is largely informa-
tional, it is difficult to conceive that spectators or readers could have expec-
tations of the style of expression of their speeches. Readers may form expec-
tations of the actions of speakers in a play and judge them more or less
appropriate accordingly, but they cannot form expectations about the specif-
ic form (language) of the speeches unless the play was written in language so
thoroughly conventional or trite as to lack any tinge of originality, individu-
ality, or unexpectedness. Humiliation or grief can be expressed in different
ways even within the convention of decorum of speech. In fact, the idea of
reader's expectations, though acceptable on a broad level from a shared knowl-
edge of generic or rhetorical conventions, may not readily be invoked in less-
expansive dramatic situations which are, as established, controlled by the
playwright without mediation (save in his own mind during the process of
composition) with spectators or readers. Whether expectation functions on
the microlevel of individual speeches and variations of style in early modern
plays is questionable, particularly if expectations are formed from readers'
understanding of intentions, whether the playwright's or the individual
speaker's, as revealed by the prevailing register and such of the dramatic
speeches alone.
Moreover, the embeddedness of speeches within the play complicates the
analysis of markedness because, more often than not, considerations broader
than the local contexts of speeches determine the shapes they take and there-
fore the identification of markedness. In ordinary conversations, as linguists
recognize, facial expression, physical posture and gesture, and verbal deliv-
ery contribute to communication and, therefore, to a degree, to verbal nego-
tiations and markedness. In drama, markedness is supplied by such perfor-
mance means (i.e., nonverbal means) as physical action (movement, stage
position, gesture), as well as by lighting, sound effects and accompanying
17. To mention existing studies of these matters would divert the thrust of this chapter, which
is essentially literary rather than linguistic.
music, and the use of properties and costume, and a theatrically experienced
reader cannot escape their influence. The extent to which such considerations
affect the analysis of a printed text of speeches written for performance on a
stage is uncertain: The relationship of markedness and the semiotics of per-
formance might deserve some examination. I make these remarks with the
tentativeness that befits one who has limited knowledge of recent linguistic
studies but, nevertheless, great interest in their outcome.
REFERENCES
Campbell, Lily. 1930. Shakespeare's tragic heroes: slaves of passion. Cambridge:

Cambridge University Press. Reprint, London: Methuen, 1970.
Edwards, Philip (ed.). 1959. In Revels plays. The Spanish tragedy: Thomas Kyd.
London: Methuen.
Evans, G. B. (ed.) 1984. Romeo and Juliet. New Cambridge ed. Cambridge: Cam-
bridge University Press.
Grice, H. P. 1975. Logic and conversation. In Syntax and semantics, vol. 3, ed. Peter
Cole and Jerry L. Morgan, 41-58. New York: Academic Press.
Herrick, Marvin T. 1950. Comic theory in the sixteenth century. Urbana: University
of Illinois Press. Reprint, 1964.
Howard-Hill, T. H. 1996. U and non-U: class and discourse level in Othello. In
Shakespeare's universe: Renaissance ideas and conventions: essays in honor of W.
R. Elton, ed. John M. Mucciolo [and others], 175-186. Aldershot, Hertsford-
shire: Scolar Press.
Africa. Oxford: Clarendon Press.
Riverside Shakespeare. 1974. Textual ed. G. Blakemore Evans. Boston: Houghton
Mifflin.
Spevack, Marvin. 1969. A complete and systematic concordance to the works of
Shakespeare. Vol. 4: A concordance to the complete works: "A"-Hilding.
Hildesheim, Germany: G. Olms.
III
Stylistic Choices in
Spoken English
8
Markedness and Styleswitching in

Performances by African American
Drag Queens
RUSTY BARRETT
T HIS CHAPTER CONSIDERS the ways in which African American drag

queens (AADQs) use language style in their performances, particularly
the ways in which they use choices that would be considered "marked" under
the markedness model (Myers-Scotton [Scotton] 1988; 1993; chapter 2, this
volume). After discussing drag queen as a social category, I consider the con-
cept of style, and some of the confusion surrounding the terminology used for
linguistic varieties. I argue that the Markedness Model provides a useful tool
for examining differences related to the notions of dialect, register, and genre.
Finally, I use examples from AADQ speech to demonstrate how AADQs use
both unmarked and marked choices as rhetorical devices in their performances
to highlight the instability of social categories related to gender, ethnicity,
class, and sexuality.
DISTINGUISHING DRAG QUEENS
First, it is important to clarify the meaning of drag queen. Basically, the social
category of drag queens is gay men who dress as women, especially those who
perform in gay bars. As a social group, drag queens are often confused with
other groups: transsexuals, transvestites, cross-dressers, and female imper-
sonators.1 Transsexuals are individuals who feel that the sex they were as-
signed at birth does not correspond with their true gender identity. The term
transvestite is used to refer to those who wear the clothing of the opposite sex
(i.e., opposite from the sex they were assigned at birth). Unlike transsexuals,
1. The terms transgender person or transgenderist are often used as umbrella terms for various
social groups. ( For discussion see Bullough and Bullough 1993; Devor 1989; MacKenzie 1994;
and Feinberg 1996.)
139
140 STYLISTIC CHOICES IN SPOKEN ENGLISH
transvestites (or cross-dressers) categorize themselves as members of the gen-

der corresponding to their assigned sex. Cross-dressing refers to anyone who
wears the clothing associated with the opposite sex/gender, regardless of their
sexual orientation. Neither transsexualism nor transvestitism/cross-dressing
is a specifically gay/lesbian phenomenon. In contrast, drag is specifically part
of lesbian and gay culture. Although it is possible to say that heterosexuals
such as Milton Berle or Dustin Hoffman (in Tootsie) are "in drag," these men
would not be considered drag queens.
The term female impersonator is closer to drag queen than any of the other
terms considered here. Many professional drag queens refer to themselves as
female impersonators or illusionists as a means of distancing themselves from
nonprofessional drag queens. For example, one participant in Esther New-
ton's ethnography of gay female impersonators said that the term drag queen
sounded "sort of like a street fairy puttin' a dress on" (1972: 17). However,
female impersonators generally attempt to produce the illusion of being a
"real" woman, usually performing as a specific celebrity (e.g., Marilyn Mon-
roe, Liza Minnelli, or Cher).
Professional drag queens typically have their own unique persona (as
opposed to only reproducing the persona of a single celebrity). In addition,
most drag queens generally make no pretense about not being male but, rath-
er, use their performances as a means of playing on the irony of crossing
genders. Finally, although there are self-described female impersonators who
claim to be heterosexual, drag queens are openly (and proudly) gay.
AADQ performances generally include lip synching to records and emcee-
ing a variety of shows, including lip synching by other drag queens, male
strippers, or talent shows. The emcee usually presents comedic monologues
that involve interaction with the audience. Performances often encode direct
information about the relationship between language (and performance as a
whole) and social and cultural issues. As Turner noted, the relationship be-
tween performance genres and society is "reciprocal and reflexive—in the
sense that the performance is often a critique, direct or veiled, of the social life
it grows out of" (1986:22). This is true of AADQ performances, which often
present critiques of social structure and are often highly political. As Briggs
argues, "performers are not passive, unreflecting creatures who simply re-
spond to the dictates of tradition or the physical and social environment. They
interpret both traditions and social settings, actively transforming both in the
course of their performances" (1988: 7). Building on the work of Bauman
(1975) and Hymes (1981), Briggs (1988: 8-9) notes that one of the primary
components of performance is the responsibility of the performer to his or
her audience. The audience evaluates a performer not only on the content of
the speech but also on the performer's communicative competence in accor-
dance with the criteria for a specific performance genre. A primary goal of
drag performance is to highlight mismatches between the performer's "per-
ceived" identity (as a woman) and her "biographical" identity (as a man). A
African American Drag Queens 141
successful drag performance is dependent on the ability to use language in a
way that demonstrates that categories based on gender, ethnicity, class, and
sexuality are indeed performances (cf. Butler 1993) and cannot be taken as
obvious or constant reflections of some "authentic" identity.
DRAG QUEENS IN THIS STUDY
The drag queens in this study all belong to the class of glam(our) queens, that
is, drag queens who attempt a highly stylish image of glamour, dressing up
like movie stars at the Academy Awards or contestants in the evening gown
competition at a beauty pageant. "Glam" queens typically wear evening gowns
with lots of beads or fringe, exaggerated jewelry (such as big earrings or wide
flashy bracelets), and big-haired black wigs.2 Their clothing is often fairly
revealing, with very short skirts or high slit dresses and bare arms. Although
they dress in a way to accentuate their ability to look "feminine," they often
undermine that image of femininity in their performances. Thus, instead of
producing humor through their appearance, they play off the irony that a
man could create the believable image of a woman.
Most of the language examined in this chapter is taken from public per-
formances by AADQs in gay bars in Texas. From January to May 1993, I
observed between six and ten AADQ performances per month. In addition to
dealing with drag queens in Texas, some of the data in this study come from
television appearances by RuPaul, a drag queen originally from Atlanta whose
dance song "Supermodel (You'd better work!)" became very popular in early
1993. RuPaul recorded two albums, appeared in several films, and was on
numerous television shows and even wrote her3 autobiography (RuPaul 1995).
She currently hosts her own talk show on the VH-1 cable channel. In this
chapter, data involving RuPaul are taken from two talk show appearances
and a speech given at the 1993 March on Washington.
LANGUAGE STYLE AND THE MARKEDNESS MODEL
I take style to be an overarching term for all linguistic varieties below the level
of language, namely the types of variation that Ferguson (1994) categorizes
as dialect, register, and genre.4 Ferguson states that "identifying markers of
language structure and language use" (1994: 18-21) set apart one dialect
from another, one register from another, and one genre from another. Among
2. Unlike all of the other drag queens in this study, RuPaul almost always wears a blond wig.
3.I generally use feminine pronouns to refer to drag queens, especially in drag. The use of he
to refer to a drag queen may be insulting as it insinuates that her performance is somehow flawed.
(For a study of uses of she among gay men in general, see Rudes and Healy 1979.)
4. Here I depart from Ferguson's use of the term conversational variation in his definitions of
variation. For me, conversation is a particular mode of discourse (or genre) included under the
cover term style.
themselves, the three varieties differ in this way: (1) dialects distinguish one
social group from another; (2) registers distinguish one communication situ-
ation from another; and (3) genres distinguish one message type from anoth-
er. I follow Ferguson in distinguishing the three types of varieties in this way.
Whether a given set of linguistic variables (on any of the levels of phonol-
ogy, morphosyntax, or lexicon) is classified as dialect, register, or genre is
determined by the discourse (i.e., through usage). For example, for speakers
with multiple group identities (i.e., with more than one dialect), a given dia-
lect may be reserved for speech in a particular communication situation (e.g.,
with particular interlocutors). However, because that dialect is also a feature
of the situation, that dialect functions as a register for those speakers. This
means that in different interactions the terms dialect, register, and genre may
apply to the same set of linguistic variables. For this reason, I use the term
style to represent all the sets of linguistic variables included in any of the three
categories. Thus, I employ the following definition of style:
Style: The set of linguistic variables that are characteristic of a given dialect,
register, or genre.
This definition is similar to Ferguson's use of the term variety; however, for
many researchers, variety includes language as well as dialect, register, and
genre. For this reason, I prefer style for my purposes. Recognizing style as a
cover term for dialect, register, and genre enables us to make distinctions in
speaker variation more explicit. Such an interpretation contrasts with La-
bov's (1972) use of the term style. He used it to refer to a continuum between
"formal" and "informal" speech and operationalized the continuum in terms
of specific tasks (or genres), such as talking in an interview situation, or read-
ing word lists or narrative materials. He also brought in the notion of social
setting (or communication setting in Ferguson's definition of register). Thus,
for Labov, register and genre were collapsed into a single dimension of "for-
mality," which was then used to compare speech across social groups (dia-
lects). The lack of explicit distinctions between the categories of dialect, reg-
ister, and genre and the vagueness of distinctions such as formal/informal
reduces the usefulness of style in this sense as an analytical concept. Similar
problems arise in research that examines register in terms of "oral" versus
"literature" genres (e.g. Finnegan and Biber 1994).
STRUCTURAL CHARACTERISTICS OF DIALECTS, REGISTERS, GENRES
As Ferguson notes, dialects, registers, and genres differ "in the degree of
cohesiveness they show as systems and the sharpness of the boundaries be-
tween them" (1994: 23). Thus, these three systems, which I consider together
under the rubric style, consist of linguistic variables that form "fuzzy" sets,
and the degree to which variables belong to a given set must be considered
before meaningful discussions of style can proceed. However, as already noted,
overlap is inevitable such that one linguistic variable is found across more
than one type of style. For example, Sherzer (1989) points out that some verb
suffixes in Kuna occur exclusively in specific genres, whereas others occur in
everyday speech in general but with greater frequency in a given genre.
Certainly, the more a variable is restricted to a particular style, the higher
its degree of membership in that style will be. For example, compare two
variables from African American Vernacular English (AAVE) in the Southern
United States: the copula verb be with continuative or durative aspect (e.g.,
I be working there for five years) and double modals (e.g., might could as in
She might could do that). Both forms have been attributed to AAVE. Yet,
compared to double modals, this useofbe would be the more salient (and less
"fuzzy") marker of the AAVE style set, at least in the Southern United States,
as it has less overlap with the speech of whites.
Because they are more salient, variables with the more exclusive member-
ship in the set of variables defining a particular style are typically more useful
in studying the ways in which speakers use styles, although combinations of
low membership variables might also produce a salient index of identity.
Dialects that are more distinct from one another have fewer salient, high
membership variables in common. There is a continuum from very similar
dialects (e.g., "men's language" versus "women's language" in English), which
will have many variables in common, to distinct, unrelated languages, which
may have no variables at all in common.
MOTIVATIONS FOR SWITCHING BETWEEN STYLES
Myers-Scotton's markedness model (Scotton 1988; 1993; chapter 2 of this

volume) analyzes switches between different codes (languages or styles) of a
given language based on the relationship between the code and the rights and
obligations sets (RO sets) indexed by the code. Sets of linguistic variables may
be seen as indexing the particular RO set associated with a given group iden-
tity, situational context, or speech event. Thus, individual dialects, registers,
and genres index particular RO sets. Within the markedness model, each type
of interaction has an unmarked RO set. According to the model, "the un-
marked RO set is derived from whatever situational features are salient for
the community for that interaction type" (Myers-Scotton 1993: 84). Para-
mount among those features are the identities of participants. Speakers use
codes to index the RO sets that are unmarked for a given interaction or to
attempt to change or explore the nature of the RO set present in an interac-
tion. Thus, the RO set is dynamic: At the outset of an interaction an unmarked
RO set is in place, but it is open to change as well.
In regard to motivations for codeswitching, Myers-Scotton (1993) claims
that there are four major types of switches: sequential unmarked code switch-
ing, codeswitching itself as the unmarked choice, codeswitching as the marked
choice, and codeswitching as an exploratory choice. These are described in
the discussion that follows. By analyzing all switching as one of these four
types, the markedness model enables us to explain the choices of stylistic
variables speakers make from their linguistic repertoires to achieve specific
goals in the course of particular interactions.
In addition, however, the markedness model could provide a means for
exploring the relationships and distinctions among the fuzzy sets comprising
style: dialect, register, and genre. For example, consider sequential unmarked
codeswitching, which refers to switches from one code to another as unmarked
when the unmarked RO set changes within an interaction. That is, with a
change in situational factors, the unmarked RO set changes, and with this
change comes a change in unmarked code to index the new RO set. This type
of switching would be predicted for both dialects and registers. For dialect
switching, sequential switches from one dialect to another as unmarked would
be predicted when there are changes in the group identities that are salient;
for register switching, sequential unmarked switching would be predicted when
there are changes in the situational context.
The second type of switching, codeswitching as the unmarked choice, would
be predicted to occur with dialects. This type of switching consists of a pat-
tern of switching back and forth between two styles in the same conversation,
with the overall pattern carrying the message of multiple identities for the
speakers. However, such switching is not predicted when styles function as
registers. By definition, a register occurs in a particular situational context
and therefore we cannot speak of a change of register without a change in
situation. Switching as the unmarked choice takes place without a change of
situation and therefore cannot involve registers. However, we would predict
that the third and fourth types of switching, marked switching and explor-
atory switching, would occur with register rather than dialect in the major-
ity of cases. Codeswitching as a marked choice is a switch to a code that is
not the unmarked choice for the current RO set and therefore is a call for
another RO set in its place. Codeswitching as an exploratory choice occurs
when an unmarked RO set (and therefore an unmarked choice) is not clear.
In such cases, speakers use switching to propose alternate choices as candi-
dates for an unmarked choice and thereby as an index for the RO set which
they favor.
I suspect that all types of switching can occur between genres. For exam-
ple, the inclusion of a personal narrative style in the academic writing of some
feminist theorists would have originally been a marked change from one genre
to another (within a given work); such an inclusion might be seen as an at-
tempt to alter the exclusion of women from the RO set indexed by academic
writing. Over time, however, this type of switching has become convention-
alized, making switches between personal narrative and more academic styles
an unmarked norm in the writing of some scholars. Unmarked sequences of
genre switching should be quite common, such as moving from jokes to sto-
ries within the course of a conversation. When the appropriateness of a given
genre is unclear, or the responsiveness of an audience is uncertain, speakers

might use exploratory genre switching as well.
The asymmetry of dialect and register switching, within the terms of the
markedness model, reflects the fact that expressions of a particular identity
are often reserved for specific situational contexts. The fact that the marked-
ness model has (unintentionally) encoded the division between dialects and
registers provides external support for the model, suggesting that the model
might provide a principled basis for studying the ways in which dialects, reg-
isters, and genres relate to one another in actual use.
STYLISTIC CHOICES OF AFRICAN AMERICAN DRAG QUEENS
Although AADQs usually employ a wide variety of styles, three basic styles
recur in their speech, reflecting membership in three different social groups.
These are an AAVE style, a gay male style, and a style based on stereotypes
of white women's speech. The AAVE and gay male styles index the identities
of AADQs as African Americans and as gay men, respectively. The white
woman style indexes their identity as drag queens while also indexing the RO
set associated with actual white women and implying a variety of social at-
tributes associated with white women, including stereotypical notions of fem-
ininity and glamour (cf. Barrett 1994).
When used by AADQs, each of these styles reflects a stereotype of the
linguistic behavior for the prototypical member of the social groups indexed
by the style. As Le Page and Tabouret-Keller note, when speakers construct
their identities through language "what they recognize and imitate are stereo-
types they have created for themselves" (1985: 142). As linguists, we often
pride ourselves on the fact that we study language as it is actually used, often
attempting to demonstrate the falseness of stereotyped attitudes. When dis-
cussing choices of linguistic style, however, it is important to remember
that speakers often base their linguistic behavior on stereotypes (as they
do not have our privileged access to detailed sociolinguistic studies). Thus,
each of the styles I am presenting does not reflect the natural speech of any
actual white woman, gay man, or African American. Rather, the styles
reflect prevailing stereotypes concerning the speech of members of these
three social groups and are used by AADQs to index the RO sets of those
social groups.
Because each of these styles indexes a social group, each constitutes dia-
lects in Ferguson's (1994) framework.5 Nevertheless, because each can serve
to index a particular situation, each also can be considered as a register as
5. Traditionally, style and dialects are differentiated as follows. Dialect is used for specific
social groups (differentiated along ethnic, regional, and sometimes class divisions). Dialect also
often connotes an opposition to the standard dialect. Typically two styles differ from each other
by fewer linguistic variables than two dialects do.
well as a dialect (depending on usage). This is why I use the broader term style
as a cover term. As is often the case with styles, there is a certain amount of
overlap in the sets of linguistic variables making up these three styles. Al-
though I will not arrange the variables in a given style in a hierarchy of "sa-
lience," it is important to keep in mind the fact that the degree of indexicality
is not constant across variables. In addition, it should be noted that not all
variables will be present in any given style at any given moment (and the
presence/absence of variables will not be constant across speakers). Speakers
may even choose a subset of linguistic variables from a style to index specific
attributes associated with speakers of that style. For example, Sunaoshi (1995)
demonstrated that Japanese women in managerial positions choose particu-
lar features (but not the complete set of variables) from the "Motherese" style
in Japanese to index the authoritative status of mothers. By not using the full
set of variables associated with Motherese, these women are able to assert
their authority without fully indexing the powerful differential between moth-
ers and children (which might be seen as condescending to their employees).
In the data that follow, I generally classify a particular utterance as be-
longing to a given style according to the following (rather arbitrary) criteria:
(1) a linguistic variable represents a particular style if the variable's distribu-
tion does not overlap with it in some other style and (2) a particular style is
present if two linguistic variables from that style are present that do not both
overlap with another style (although they may each individually overlap with
two different styles). The basic features of each of these styles are outlined in
the following sections. Those features characterizing the white woman style
come from Lakoff (1975), because Lakoff's description reflects a stereotyped
image of how white women should speak (cf. Barrett 1994; Bucholtz and
Hall 1995). The features associated with stereotypes of gay men's speech are
those I have discussed elsewhere (Barrett, 1997); they are compiled from a
variety of sources (Lakoff 1975; Rodgers [1972]1979; Hayes 1981; Walters
1981; Goodwin 1989; Moran 1991; Gaudio 1994). The features of AAVE
have been widely studied by linguists; the list presented here is primarily that
of Walters (1992). Wyatt provides a similar list (1994).
In the remainder of this chapter, I present examples that demonstrate the
ways in which AADQs choose particular stylistic variables from this reper-
toire to enhance their performances.
White Woman Style (Lakoff 1975: 53-56)

Specific lexical items related to their specific interests, generally relegated
to them as "woman's work": dart (in sewing) and specific color terms (ecru,
magenta)
"Empty" adjectives like divine, charming, cute . . .
Question features for declaratives: tag questions [It's so hot, isn't it?), ris-
ing intonation in statement contexts (What's your name, dear? Mary Smith?)
The use of hedges of various kinds (e.g., well, y'know, kinda)

The intensive use of so
Hypercorrect grammar, superpolite forms, and euphemisms: women are
not supposed to talk rough.
Women don't tell jokes.
Gay Male Speech Style (Barrett 1997)

The use of lexical items and structures included as part of Lakoff's wom-
en's language, e.g., specific color terms and the "empty" adjectives (mar-
velous, adorable, etc.), as well as hedges and boosters.
The use of a wider pitch range for intonational contours than in speech of
straight men.
Hypercorrect pronunciation; the presence of phonologically nonreduced
forms and the use of hyperextended vowels (the probable source of the
"lisping" stereotype).
The use of lexical items specific to gay language (Rodgers [1972]1979 gives
a somewhat dated lexicon of many such expressions).
The use of a H*L intonational contour, often co-occurring with extended
vowels (as in FAAABulous).
African American Vernacular English (AAVE) Style (Walters 1992)

AAVE Realizations of Standard English Phonological Features :
Interdental fricatives as labiodental fricatives: "teeth" [thijf], "other" [Ave]
Final consonant reduction: /r/ deletion, "sister" [sIstha]; /l/ palatalization
or deletion, "all" [a];/l/cluster vocalization or reduction, "film" [fIm] or
[fI.lm]; unreleased stops as glottal stops, "cat" [khas?]; cluster simplifica-
tion, "best" [bEs]
Intervocalic weakening: /r/ deletion, [v] [B] "every" [EBij]
Strident cluster metathesis or modification: "ask" [aeks], "stray" [ trej]
Prenasal raising of [E] to [I], "pen" [phIn]; or lowering to [ae]: "thing" [6ae
Monophthongization of [ai]: "my" [ma]
Stress fronting: "police" [ph6w:lijs]
AAVE Morphological and Lexical Innovations and Realizations of SE Features:
Absence of nonsyllabic inflectional endings: third person singular {-s},
noun plural {-s}, possessive {-s}
Reduction of unstressed syllabic inflections {ing}: [In], or [.n], and {to}:
[e] "going to" [gAne]
Lexical verbs as aspect markers: perfective "done," I done finish my work;
inceptive "come" don't come coming in here...; future perfect be done;

intensive continued progressive be steady; habitual copula: she be ner-
vous vs. she nervous (non-habitual) (Green 1994)
Negation: multiple negation, didn't do nothing; ain't as negative of forms
of to be and to have
"You" plural distinctions: y'all and y'all's
Special intonation patterns, e.g., H*HL in rhetorical speech
CONTEXT AND UNMARKED CHOICES
For in-group communication outside performances, AADQs typically ex-

hibit styleswitching as the unmarked norm. This switching consists of all
three of the styles discussed here, mirroring the fact that AADQs are all Af-
rican American, gay men, and drag queens. Thus, by unmarked switching
between styles, they index these three aspects of their group identities.
This continuous switching between the three styles is typical of AADQ
speech, so much so that the styles may even co-occur, with phonological
variables from one style overlapping with syntactic variables from another
style. In the utterance in [1] (from RuPaul on the Arsenio Hall Show), the
phonology is that of the white woman style (i.e., it is typical of Standard
English, but not the hypercorrect pronunciation sometimes associated with
the speech of African American women or gay men). The syntax, however, is
typical of the AAVE style.
[1] You know, in my mind's eye I always been a superstar, you know. And
nobody couldn't tell me no differen[t].
The white woman phonology occurs up until the final word different, which
is pronounced [dIfrIn] (i.e., without the final [t]). After the word different, the
speech continues in AAVE style, so that the switch actually occurs before the
end of the sentence. This meshing of styles is possible because of the fuzzy
boundaries between styles and points to a problem that was not addressed by
purely quantitative studies of language style. A quantitative analysis would
most likely misanalyze the utterance in [1] as belonging to only one style,
depending on whether the coding was for syntactic or phonological charac-
teristics. This suggests that possible cases of styleswitching as the unmarked
norm may have gone unnoticed because the styles were studied as closed sets
that can be easily isolated from one another.
Consider the following example, a speech given by RuPaul at the 1993
March on Washington for Lesbian, Gay and Bisexual Rights:
[2] Speech given by RuPaul
1 Hello America
2 My name's RuPaul
3 Supermodel of the worl[d]

4 [begins chant, audience joins in] Hey . . . Ho . . . Hey . . . Ho . . .
Hey . . . Ho
5 You know people ask me all the time
6 Where I see myself in ten years
7 And I say I see myself in the White House, baby!
8 Miss Thing goes to Washington
9 Can you see it? Wha- we gonna paint the mother pink, OK?
10 We put one president in the White House, I figure you can do it again.
11 Everybody say love! [audience responds, "love!"]
12 Everybody say love! [audience responds again, "love!"]
13 Now drive that down Pennsylvania Avenue!
14 Peace, love and hairgrease!
15 I love you!
In this example, RuPaul moves back and forth between each of the three styles,
often allowing the styles to overlap. For this reason, and the fact that the
situational context does not change, this type of switching would be referred
to as dialect, as a subset of style, switching as the unmarked choice.
RuPaul begins her speech with typical standard, albeit feminized, English
pronunciation reflecting the white woman style. The phrase supermodel of
the world, however, is produced without a final [d] on the word world, a
form typical of the AAVE style. The chant in line 4 reflects the call-response
tradition, a genre often associated with African American preachers (and
frequently used by drag queens). After the chant, RuPaul returns to a white
woman style up until line 7 and the phrase in the White House, baby. This
phrase is spoken with an increasingly widening pitch range, typical of the gay
male speech style ending with the word baby spoken with an extended vowel
and long falling (H*L) intonation (typical of the gay male style). In line 8,
RuPaul refers to herself as Miss Thang, a term that indexes gay male identity
and/or African American identity. She pronounces the word thing with stan-
dard English [I], however, rather than with the [ei] or [ae] vowels typical of
both the gay and AAVE pronunciations. Thus, a term related to both gay and
African American identity occurs with a "white woman" phonology, letting
the various styles overlap. In line 9 the question Can you seeit?is spoken with
a white woman phonology. The phonology then switches to AAVE with the
word mother, which is pronounced without the final [r] and ends with the
exaggerated H*L gay male style intonation on the word OX. RuPaul's pho-
nology returns to a white woman style in lines 10 and 11. Although line 11
has white woman phonology, it begins a call-response routine, thus again
indexing African American identity. In lines 12 and 13, the intonation becomes
more like that found in African American preaching (cf. Queen 1992). Through
the use of codeswitching as the unmarked norm, RuPaul uses the speech to
simultaneously index her membership in several different social groups.
Even though codeswitching as the unmarked norm is typically used for
in-group communication and some performance contexts, a change in the
situational context may mean a change in style. In this case, we can say that
the performer is switching registers from one context to the next. For exam-
ple, the table below compares the speech of RuPaul during her first appear-
ances on the Joan Rivers Show and on the Arsenio Hall Show. These two
televised interviews occur about one year apart from each other. The Arsenio
Hall episode aired in March 1993; the Joan Rivers episode aired in May 1994.
The Arsenio Hall interview was RuPaul's first appearance on national televi-
sion. Although both were recorded before live audiences, the two programs
constitute quite different contexts. The Arsenio Hall Show was hosted by an
African American male, aired late at night, and was generally targeted to
young adults. The Joan Rivers Show was hosted by a white woman, aired
during the morning, and was generally targeted to women (especially women
who do not work outside the home). The differences in context on the two
programs can be seen in the stylistic choices made by RuPaul on the two
programs. Table 8-1 compares RuPaul's use of the three styles discussed above
on each of the two programs. Because the styles overlap a great deal and may
co-occur (as already noted above), the table compares utterances (rather than
particular phonological or syntactic linguistic variables). The utterances are
sorted according to those that occur exclusively in a given style, those in which
a switch between two styles occurs or in which two styles overlap, and those
in which all three styles occur. The tokens were collected from the first six
minutes of the interviews.6
On the Joan Rivers Show, which has a predominantly white studio audi-
ence, RuPaul does not use the AAVE style at all. More than 80% of the time,
she uses the white woman style exclusively. This can be seen as a form of
accommodation (cf. Giles, Coupland, and Coupland 1991) to the speech of
Joan Rivers, an actual white woman. Nevertheless, RuPaul does continue to
reference gay identity, demonstrating that drag queen identity must, at the
very least, combine some elements of gay male speech with stereotyped wom-
en's speech.7 In contrast, on the Arsenio Hall Show, RuPaul used the white
woman style exclusively only half of the time and used the AAVE style exclu-
sively 13.3% of the time. Thus, although the white woman style continues to
6. This is the time up until the first commercial break on the Joan Rivers Show (in which the
entire interview lasts an hour minus commercials). On the Arsenio Hall Show the interview was
only about ten minutes long in its entirety.
7. For example, as noted by Queen (1997), the actors in the film To Wong Foo ... Thanks for
Everything, Julie Newmar do not sound like actual drag queens, largely because they use white
women's speech (and in the case of Wesley Snipes, AAVE), but they do not employ gay male
speech as would an actual drag queen.
Table 8-1
Stylistic Variation in the Speech of RuPaul
on Two Television Talk Shows
Joan Rivers Arsenio Hall

style N % N %
WW 70 81.4 38 50.6
GM 8 9.3 7 9.3
AAVE 0 — 10 13.3
WW/GM 8 9.3 6 8
GM/AAVE 0 4 5.3
WW/AAVE 0 — 2 2.6
AA/AAVE/GM 0 — 2 2.6
Totals 86 100 75 99.8
WW = white woman style, GM = gay male style, AAVE = African Amer-

ican Vernacular English style, N = number of utterances in a give style or
combination of styles.
predominate, it co-occurs with the AAVE style, reflecting the context and
shared African American identity between RuPaul and Arsenio Hall. Although
some of the stylistic choices on the Arsenio Hall Show reflect marked choic-
es, it is clear that the context of the program allows for a much wider range
of stylistic choices. Thus, context in a broad sense may influence what is the
unmarked choice as well as the general range of stylistic choices possible.
Moreover, although switching as the unmarked norm is quite common among
AADQs, the styles may also occur in isolation, acting as registers indexing
the RO set of a given situation. This is the case with the absence of the AAVE
style on the Joan Rivers Show, in which RuPaul does not index her African
American identity, reflecting the RO set of the interaction with a white woman.
Because each style may work individually, switches between any of the
three styles may be a marked switch or a sequence of unmarked styles. Recall
that unmarked sequences are predicted with changes in audience, topic, or
context. A number of unmarked sequences occur in a specific genre of come-
dic monologues, that which offers critiques of the political and economic
situations of African Americans. In this genre, certain comments may be di-
rected to certain members of society, whether or not they are present in the
audience. The white woman style is used to index the power held by whites,
whereas the AAVE style is used to index the social situation of African
Americans. The following example is from a drag queen performing in a
gay bar. The style shifts correspond to a change in topic; as such they are
changes in register. The AADQ is posing as a salesperson trying to sell rat
traps to the audience. She offers three different types of rat traps for use in
three different neighborhoods.
[3] Segment of performance by African American drag queen in Texas

1 OK! What we're gonna talk about is, um, rat traps, um . . .
2 [holds up a mouse trap] This is a rat trap from <name of upper-class
white neighborhood>
3 It's made by BMW. It's real compact.
4 It's, thank you . . . <obscured> . . .
5 It's really good, it's very convenient, and there's insurance on it.
6 And this is from <name of same white neighborhood>
7 OK, now for the <name of housing project> . . . [holds up a large rat
trap]
8 This rat trap is made by Cadillac, it's a big mother fucker.
9 [holds up a gun] Now for the <name of inner city area>
10 You just don't need no rat trap.
11 Cause those mother fuckers look like dogs out there.
12 Shit!
13 I put in a piece of cheese, the mother fucker told me,
14 "'Next time put in some dog food."
This example begins in the white woman style, switching to the AAVE style
beginning in line 7. The sentences in 1-7 all end with a final rising tone, typ-
ical of the white woman style (Lakoff's "question intonation"). The phonol-
ogy and intonation switch to AAVE in line 7, although the first syntactic fea-
tures of AAVE do not occur until line 10, when the speaker begins to talk about
the third location (which has the strongest reputation as a primarily African
American low-income area). Thus, the number of linguistic variables used to
index African American identity increases when discussing the neighborhood
that is the strongest representative of the social conditions stereotypically as-
sociated with African Americans. This increase supports the claim of the marked-
ness model that linguistic choices are indices of particular RO sets. In addition,
it demonstrates that the density of variables from a particular style may increase
with an increased desire to index the RO set corresponding to that style.
In situations of performances or public appearances, the unmarked
choice should generally be clear, as the expected audience (and correspond-
ing RO set) is probably known to a speaker before the performance begins.
Thus, cases of exploratory switching are quite rare in public performances
by AADQs. One example of exploratory switching, however, occurred dur-
ing the opening moments of RuPaul's appearance on the Arsenio Hall Show.
This appearance was RuPaul's first on national television and, according to
her autobiography (RuPaul 1995: 176-7), she was extremely nervous. At
the very beginning of the interview, RuPaul greeted the audience by saying
how much she loved their enthusiastic response to her introduction:
[4]
H*L
1 I love it.
H* HL
2 I feel aaaall the love . . .
H*L
3 I love that.
L H
4 I feel all the love you're sending me.
Here, the first utterance of I love it carries a H*L intonational contour that
reflects stereotypically gay speech (as in FAABulous). In line 2, the word all
is extended with a H*HL intonation that is particular to certain types of
rhetorical speaking in AAVE. This intonational pattern, which Michaels (1986)
described in African American children's show-and-tell narratives, con-
tains a high pitch on an extended vowel followed by a shortfall. In addi-
tion to indexing African American identity, H*HL indexes the genre public
speaking, typically in formal settings, such as sermons (cf. Queen 1992).
In line 3, RuPaul returns to the "gay" style of line 1. Finally, in line 4, the
final H of the white woman style is used. In this example, the same referential
information is conveyed four times, with alternations between each of the
three styles. These alternations can be seen as exploratory switching that
indexes a variety of possible identities as a means of addressing the vari-
ety of potential audience members in the studio and watching on televi-
sion. Thus, exploratory switching may occur when the unmarked choice
is not clear. This is not surprising, considering that these data represent
RuPaul's first utterances during her first appearance on national television, a
context in which she may not have been certain as to which stylistic choice
would be most appropriate.
MARKED CHOICES IN AADQ PERFORMANCES
Although off stage AADQs generally use styleswitching as the unmarked

norm for ingroup communication, in their performances they make more
use of marked stylistic choices to draw attention to their speech. The per-
formers use language style to play on the audience's assumptions sur-
rounding issues of gender, ethnicity, and sexuality. A major part of the
communicative skill displayed in AADQ performances relies on the speak-
er's ability to anticipate audience members' interpretations of the situa-
tion and then to use marked stylistic choices to disrupt the audience's
assumptions.
Often, these marked choices reflect instances of signifyin(g), an African
American speech event in which the full meaning of an utterance cannot be
understood from referential meaning alone (cf. Gates 1988; Mitchell-Ker-

nan 1972). s In signifyin(g), an utterance takes on special value through
indexing a particular rhetorical figure or a speaker's skill at verbal art. In
particular, the examples of signifyin(g) in these marked stylistic choices are
cases of what Morgan (1994) discusses as "reading dialect," in which the
linguistic style itself may be crucial in conveying the full meaning of a given
utterance.
AADQs often directly address particular audience members, using marked
stylistic choices to put the addressee "on the spot." This technique is quite
common in drag performances. For example, this device is used by The Lady
Chablis, a Savannah, Georgia, drag queen discussed in the best-selling Mid-
night in the Garden of Good and Evil (Berendt 1994). In her autobiography,
The Lady Chablis describes how she approaches audience members during
performances:
Whenever I performed my monologue, I made sure to comb the audience for a
"victim." I'm always gonna pick on someone—that's part of my act. Still is. I'll
usually find somebody in the front row who's got a certain afraid look on their
face, like "Oh my God, please don't say nothing to me." 'That's when I move in
for the kill. If I see a woman, and she's draped in diamonds, I might ask, "Girl,
what didja do to get those jewels? Didja suck dick that good? Share y'secrets
with The Doll!" (The Lady Chablis 1996: 103)
Here, The Lady Chablis uses a highly marked stylistic choice for comic effect.
The highly informal style and coarse language directed at a woman "draped
in diamonds" is unexpected, indexing a (false) closeness and familiarity be-
tween the drag queen and the audience member. As such, it represents both
a dialect switch (imaginary change in participant identity) and a register switch
(imaginary change in type of interaction). For example, discussing sexual
exploits and using the term girl would normally be reserved for conversa-
tions with a close friend (most likely another gay man or another drag queen).
The drag queen suggests that the audience member is not a wealthy upper-
class woman but someone who simply received her jewels as a reward for her
sexual abilities. As Myers-Scotton argues, such a marked choice is often a
powerful feature of language "because it deviates from the expected and
because its motivations may not be clear, therefore leaving the addressee off
balance and unable either to predict the RO set in effect for the rest of the
exchange or to explain the speaker's precise motivations" ([Myers-
Scotton] Scotton 1985: 112). By using a marked choice, AADQs disrupt the
expected RO set between audience and performer. Such marked choices are
most often used to address white audience members, especially if they are
clearly heterosexual or upper middle class. In these cases the marked choice
produces comic effect because the AADQ used stylistic language choices to
assert power over individuals belonging to more dominant social groups.
8. For a fuller discussion of signifyin(g) in AADQ performances, see Barrett (1994).

In a similar example from a performance in Texas, the drag queen approach-

es a white heterosexual couple sitting near the stage. The setting is a gay bar with
a predominantly African American clientele. When speaking to the white cou-
ple, the drag queen switches to the white woman style and then asserts that she
herself is also white. Note that the switch does not occur within example [5]
but rather the example itself represents a marked dialect, given that a dialect
associated with African Americans would be unmarked in this setting.
[5] African American drag queen approaching a white couple during a per-
formance in a gay bar:
1 Oh, hi, how are you doing?
2 White people. Love it.
3 I . . . I'm not being racial cause I'm white.
4 I just have a <obscured> . . . I can afford more sun tan.
In line 1, how are you doing is spoken with an exaggerated L*H intonation
pattern typically associated with wealthy suburban young white women such
as sorority members (cf. McLemore 1991). This marked choice implies that,
as whites, the audience members must be spoken to in white English. The
stylistic change indexes the issues of power created by the presence of a white
heterosexual couple in a bar for gay African Americans, a dialect switch. The
heterosexual couple assumes a certain privilege in invading what is otherwise
a "safe space" for gays and lesbians. By producing humor at the expense of
the white audience members, the AADQ makes it clear that their decision to
come and "observe" (and appropriate) gay African American culture is, at
the very least, problematic.
In the next example a Texas drag queen uses a marked choice to under-
mine the image of femininity indexed by the white woman style. The switch
here is from one dialect to another, as if the drag queen belongs first to one
social group and then to another.
[6] Drag queen in Texas introducing a male stripper in a gay bar
1 Are you ready to see some muscles?
2 (audience yells) Some dick?
3 Excuse me I'm not supposed to say that, words like that in the micro-
phone,
4 Like shit, fuck, and all that, you know?
5 I am a Christian woman.
6 I go to church.
7 I'm always on my knees.
In this example, the AADQ uses the white woman style in line 1. After
using the word dick in line 2, she apologizes for using words that aren't "lady-
like" (cf. Lakoff's "women don't talk rough"). Here the apology emphasizes
the fact that the use of obscenities is a marked stylistic choice (given that it
occurs within the white woman style). The performer further undermines the
white woman style by explaining the apology in line 4 by using even more
obscenities. These marked choices disrupt the RO set indexed by the white
woman style, emphasizing the fact that the drag queen does not actually hold
claim to the image of genteel femininity that the style indexes. The remainder
of this example further exploits this irony with a joke about the drag queen
being a Christian woman who is always on her knees. Here, the reference to
being on her knees carries the dual meaning of "always praying" and "always
performing fellatio," thus playing off of the stereotypes of white women and
drag queens, respectively. The stylistic choice highlights the fact that the
audience cannot assume that the "performed" identity of an upper-class,
sophisticated woman is "authentic" because the chosen style produces infer-
ences (about the character of the drag queen) that violate the expected behav-
ior associated with the white woman style.
In example [7], RuPaul uses a marked switch into the AAVE style to
emphasize her identity as a biologically male African American. Hall is using
the standard dialect of American English, but RuPaul switches to AAVE.
[7] RuPaul during her first appearance on the Arsenic Hall Show
Arsenio Hall: I'm sure there're some people who would like the question
answered do you . . . would you have rather been born a
woman?
RuPaul: No. No no. I'm very happy with being (.) a big o' black
man.
Up to this point, RuPaul uses primarily the white woman style, indexing an
RO set associated with a high level of sophistication and femininity. Follow-
ing the pause after being, however, she switches into the AAVE style to say
a big o' black man. This marked switch enhances the referential content of
the utterance by using a linguistic style associated with being black. In addi-
tion, the choice undermines the audience's assumptions concerning RuPaul's
performance of a sophisticated feminine identity by reminding them (both refer-
entially and stylistically) that she is indeed both African American and male.
In the next example, an AADQ in Texas also uses AAVE as a marked
choice to point out that she is biologically male. This example comes from a
bar with a predominantly white clientele. As the drag queen is introducing a
male stripper, a young African American man enters the bar and walks past
the front of the stage. In line 7, the performer makes a marked switch to
another dialect to address the newcomer.
[8] Another introduction of male stripper in a bar in Texas
1 Please welcome to the stage, our next dancer,
2 He is a butt-fucking tea,9 honey.
9. Butt-fucking tea refers to anything that is exceptionally good.

3 He is hot.
4 Masculine, muscled, and ready to put it to ya, baby.
5 Anybody in here (.) hot (.) as (.) fish (.) grease?
6 That's pretty hot, idn't it?
7 (Switches to exaggerated low pitch) Hey what's up, home boy?
8 (Switches back) I'm sorry that fucking creole always come around when
I don't need it.
9 Please, welcome,
10 hot, gorgeous, sexy, very romantic,
11 and he'd like to bend you over and turn you every which way but loose.
Given that lines 1-6 are primarily in the white woman style, the switch to
AAVE (accentuated by an accompanying drop in pitch) makes the switch
highly emphasized. The switch creates momentary solidarity with the Afri-
can American who entered the bar, thus disrupting the prevailing RO set in
a bar filled primarily with whites. In line 8, the AADQ apologizes for that
fucking creole. In this line, however, the word creole is pronounced with a
vocalized [1], and the verb come occurs without the Standard English /+s/
inflection. Although the drag queen apologizes for using AAVE, the apology
itself occurs in the AAVE style. This stylistic choice continues to index soli-
darity with the African American patrons in the bar, undermining the sincer-
ity of the apology to the predominantly white audience. In lines 9-11, the
performer returns to the white woman style, returning to the RO set that was
in place before the marked switch in line 7. This return to the white woman
style serves to further emphasize the markedness of the switch to AAVE and
the change in RO sets relating to the show of solidarity with the African
American patron.
As these examples demonstrate, the use of marked choices disrupts the
status quo during a given performance. By using such marked stylistic choices
to call the prevailing RO sets into question the performers can draw attention
to themselves and their communicative skills. In addition, these marked choices
highlight the various questions of social difference brought out in the perfor-
mances by indexing disunities between perceived/performed and "actual"
(biographical) identity.
CONCLUSION
One of the goals of glam drag is to present an image that is as "real" as pos-
sible. Through the use of clothing, jewelry, hair, and cosmetics glam queens
create an external image of exaggerated femininity, often taking great pride
in how much they appear to be "real women." The Lady Chablis, for exam-
ple, often announces that she is not only a real woman, but "a pregnant uptown
white woman" (Berendt 1996). Thus, glam queens often make a specific ef-
fort to create the illusion of a femininity associated with upper-class white

heterosexual women. Language style plays a crucial role in the construction
of this image of femininity. AADQs demonstrate their ability to draw on
the power and prestige of white society through the creation of an exter-
nal image of a wealthy woman (i.e., clothes, jewelry, and makeup) and
the use of the white woman style of speaking. The white woman style is
thus crucial in creating a real, believable presentation of uptown white
womanhood.
While the use of the white woman style adds to the external image of fem-
ininity in drag performances, styleswitching into AAVE or stereotyped gay
male speech undermines the rich white feminine persona indexed by the white
woman style. By switching into these other language styles, AADQs demon-
strate that although they are capable of creating the visual and linguistic sym-
bols of white upper-class society, they are not actually attempting to create a
personal identity associated with white society. In AADQ performances,
demonstrating pride in one's African American identity is usually as im-
portant as the ability to produce white women's language. AADQs who
wear blond wigs or who prefer to lip-synch to songs by white artists are
often criticized as "trying to be white." Styleswitching thus demonstrates that
although an AADQ is capable of producing language that fits into white
society (and thus, take on the prestige associated with white society), she
consciously chooses to maintain her identity as an African American gay
man (see Barrett 1994, 1995b). The white woman style of speaking is
thus used both to index the power and prestige of white upper-class so-
ciety and to produce a critique of the social inequalities associated with
white society.
AADQs use styleswitching to undermine audience assumptions concern-
ing the personal identity of the performer. Marked switches are often used to
renegotiate assumptions concerning AADQ identity, demonstrating that the
performer refuses to be categorized according to a single, given set of social
attributes. Styleswitching thus plays a crucial role in AADQ performances,
keeping the audience aware that the performer cannot be tied down to a sin-
gle identity. In turn, interest in the performance is heightened by the aware-
ness that the current persona of the performer could change at any moment
through a change in language style.
The examples presented here demonstrate the ways in which AADQs choose
from sets of linguistic variables to attune their language to particular social
settings. Using their knowledge of what the unmarked or marked code choice
will be in a given interaction, they choose linguistic forms to produce specific
effects. The markedness model provides a means of understanding the ways
in which these stylistic choices convey the performers' attitudes about the
issues brought up in the course of a performance. In some cases, they may
choose the unmarked choice, whether that be a particular style or codeswitch-
ing as the unmarked norm. It is the marked choices, however, that are used to
add emphasis and rhetorical force to AADQ performances. By using marked
choices, AADQs tune their performances to undermine audience assumptions
concerning issues of social difference such as ethnicity, sexuality, class, or gender.
ACKNOWLEDGMENT: Thanks to Grainger Saunders, Gregory Clay, and Kathryn Semolic for help
and support in conducting this research and to Carol Myers-Scotton for comments on earlier
drafts. Finally, special thanks to Keith Walters for helpful suggestions and advice. (Barrett 1995a
is an earlier version of this chapter. See also Barrett 1994, 1995b.)
REFERENCES
Barrett, Rusty. 1994. " She is not white woman!": the appropriation of white women's
language by African American drag queens. In Cultural performances: proceed-
ings of the third Berkeley women and language conference, ed. Mary Bucholtz,
A. C. Liang, Laurel Sutton, and Caitlin Hines, 1-14. Berkeley, Calif.: Berkeley
Women and Language Group.
. 1995a. The Markedness Model and style switching: evidence from African
American drag queens. In SALSA II: Proceedings of the second annual sympo-
sium on language and society-Austin (Texas Linguistics Forum 34), ed. Pam-
ela Silberman and Jonathan Loftin, 40-52. Austin: University of Texas Depart-
ment of Linguistics.
. 1995b. Supermodels of the world, unite!: political economy and the language
of performance among African American drag queens. In Beyond the lavender
lexicon: authenticity, imagination and appropriation in lesbian and gay languages,
ed. William L. Leap, 203-223. Newark, N. J.: Gordon and Breach.
. 1997. The "homo-genius" speech community. In Queerly phrased: language,
gender, and sexuality, ed. Anna Livia and Kira Hall. New York: Oxford Univer-
sity Press.
Bauman, Richard. 1975. Verbal art as performance. American Anthropologist 77.
290-311. Revised and expanded as Verbal art as performance. Prospect Heights,
111.: Waveland Press, 1977.
Berendt, John. 1994. Midnight in the garden of good and evil. New York: Random
House.
. 1996. Chablis and me. In Hiding my candy: the autobiography of the grand
empress of Savannah, by The Lady Chablis with Theodore Bouloukos, 12-18.
New York: Pocket Books.
Briggs, Charles. 1988. Competence in performance: the creativity of tradition in
Mexicano verbal art. Philadelphia: University of Pennsylvania Press.
Bucholtz, Mary, and Kira Hall. 1995. Introduction: twenty years after Language and
woman's place. In Gender articulated: language and the socially constructed self,
ed. Kira Hall and Mary Bucholtz, 1-22. New York: Routledge.
Bullough, Vern L., and Bonnie Bullough. 1993. Cross dressing, sex, and gender. Phil-
adelphia: University of Pennslyvania Press.
Butler, Judith. 1993. Bodies that matter: on the discursive limits of "sex." New York:
Routledge.
Devor, Holly. 1989. Gender blending: confronting the limits of duality. Bloomington:
Indiana University Press.
Feinberg, Leslie. 1996. Transgender warriors: making history from Joan of Arc to
RuPaul. Boston, Mass.: Beacon Press.
Ferguson, Charles A. 1994. Dialect, register, and genre: working assumptions about
conventionalization. Sociolinguistic perspectives on register, ed. Edward Finegan
and Douglas Biber, 15-30. New York: Oxford University Press.
Finegan, Edward, and Douglas Biber. 1994. Register and social dialect variation: an
integrated approach. In Sociolinguistic perspectives on register, ed. Edward Fine-
gan and Douglas Biber, 315-347. New York: Oxford University Press.
Gates, Henry Louis, Jr. 1988. The signifying monkey: a theory of African-American
literary criticism. New York: Oxford University Press.
Gaudio, Rudolf P. 1994. Sounding gay: pitch properties in the speech of gay and straight
men. American Speech 69: 30-37.
Giles, Howard, Nikolas Coupland, and Justine Coupland. 1991. Accommodation
theory: communication, context, and consequence. In Contexts of accomodation:
developments in applied sociolinguistics, ed. Howard Giles, Justine Coupland, and
Nikolas Coupland, 1-69. Cambridge, England: Cambridge University Press.
Green, Lisa. 1994. Study of verb classes in African American English. Linguistics and
Education 7: 65-81.
Goodwin, Joseph P. 1989. More man than you'll ever be: gay folklore and accultur-
ation in middle America. Bloomington: Indiana University Press.
Hayes, Joseph J. 1981. Gayspeak. In Gayspeak: gay male and lesbian communication,
ed. James W. Chesebro, 45-57. New York: Pilgrim's Press.
Hymes, Dell. 1981. "In vain I tried to tell you": essays in Native American ethnopo-
etics. Philadelphia: University of Pennsylvania Press.
Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsyl-
vania Press.
[The] Lady Chablis, with Theodore Bouloukos. 1996. Hiding my candy: the autobi-
ography of the grand empress of Savannah. New York: Pocket Books.
Lakoff, Robin. 1975. Language and woman's place. New York: Harper and Row.
Le Page, R. B., and Andree Tabouret-Keller. 1985. Acts of identity: Creole-based
approaches to language and ethnicity. Cambridge, England: Cambridge Universi-
ty Press.
MacKenzie, Gordene Olga. 1994. Transgender nation. Bowling Green, Oh.: Bowling
Green State University Popular Press.
McLemore, Cynthia Ann. 1991. The pragmatic interpretation of English intonation:
sorority speech. Unpublished PhD dissertation. Austin: University of Texas De-
partment of Linguistics.
Michaels, Sarah. 1986. Narrative presentations: an oral preparation for literacy with
first graders. In The social construction of literacy, ed. Jenny Cook-Gumperz.
Cambridge, England: Cambridge University Press.
Mitchell-Kernan, Claudia. 1972. Signifying and marking: two Afro-American speech
acts. In Directions in sociolinguistics, ed. John J. Gumperz and Dell Hymes, 161-
179. New York: Holt, Rinehart & Winston.
Moran, John. 1991. Language use and social function in the gay community. Paper
presented at NWAVE 20 (New Ways of Analyzing Variation). Philadelphia.
Morgan, Marcyliena. 1994. No woman no cry. the linguistic representation of Afri-
can American women. In Cultural performances: proceedings of the third Berke-
ley women and language conference, ed. Mary Bucholtz, A. C. Liang, Laurel Sut-
ton, and Caitlin Hines, 525-541. Berkeley, Calif.: Berkeley Women and Language
Group.
[Myers-Scotton, Carol]Scotton, Carol Myers. 1985. "What the heck, sir": style shift-
ing and lexical colouring as features of powerful language. In Sequence and pat-
tern in communicative behaviour, ed. R.L. Street, Jr., and J. N. Cappella, 103-19.
London: Arnold.
. 1988. Codeswitching and types of multilingual communities. In Language
spread and language policy, ed. P. Lowenberg, 61-82. Washington, D.C.: Geor-
getown University Press.
. 1993. Social motivations for codeswitching: evidence from Africa. Oxford:
Newton, Esther. 1972. Mother camp: female impersonation in America. Chicago:
University of Chicago Press.
Queen, Robin M. 1992. Prosodic organization in the speeches of Martin Luther King.
Proceedings of the IRCS workshop on prosody in natural speech, 151-60.
Philadelphia: Institute for Research in Cognitive Science, University of Pennsyl-
vania.
. 1997. "I don't speak spritch": locating lesbian language. In Queerly phrased:
language, gender, and sexuality, ed. Anna Livia and Kira Hall. New York: Oxford
University Press.
Rodgers, Bruce. 1972. The queen's vernacular. Reprint as Gay talk: a (sometimes
outrageous) dictionary of gay slang, New York: Paragon Books, 1979.
Rudes, Blair A., and Bernard Healy. 1979. Is she for real?: the concepts of femaleness
and maleness in the gay world. In Ethnolinguistics: Boas, Sapir and Whorf revis-
ited, ed. Madaleine Mathiot, 49-61. The Hague: Mouton.
RuPaul. 1995. Lettin' it all hang out: an autobiography. New York: Hyperion.
Sherzer, Joel. 1989. The Kuna verb: a study in the interplay of grammar, discourse,
and style. In General and Amerindian ethnolinguistics, ed. Mary Ritchie Key and
Henry M. Hoenigswald, 261-72. Berlin: Mouton de Gruyter.
Sunaoshi, Yukako. 1995. Your boss is your "mother": Japanese women's construc-
tion of an authoritative position in the workplace. In SALSA II.- Proceedings of the
second annual symposium about language and society-Austin (Texas Linguistics
Forum 34), ed. Pamela Silberman, and Jonathan Loftin, 175-88. Austin: Univer-
sity of Texas Department of Linguistics.
Turner, Victor. 1986. The anthropology of performance. New York: PAJ Publica-
tions.
Walters, Keith. 1981. A proposal for studying the language of homosexual males.
Master's thesis, University of Texas at Austin.
. 1992. Supplementary materials for AFR 320/LIN 325: Black English. Mas-
ter's thesis, University of Texas at Austin.
Wyatt, Toya A. 1994. Language development in African American English child speech.
Linguistics and Education 7: 7—22.
9
Styleswitching in Southern English

MARGARET MISHOE
W HAT MOTIVATES the same speaker to sometimes remark on heavy

rain by saying / ain't never seen nothing to beat all this rain and, at
other times, to say something more like I've never seen a rain like this'i In this
chapter, I argue that speakers switch between the styles in their linguistic rep-
ertoire to project different dimensions of themselves. The subjects studied
here are a close-knit group of rural southern whites who have relatively little
education and little exposure to how others negotiate the fast track of mod-
ern life. Yet, their language use patterns demonstrate that their tacit knowl-
edge of the intentional messages in contrasting style choices serves them as
well as it does any urban high-flyer. The subject here is strategic switching
between the two varieties in the community's linguistic repertoire to index
how they wish to present themselves in relation to new topics or other partic-
ipants' contributions in an ongoing conversation.
When speakers have access to two different styles in the dialect of their
language, factors in a large-scale social pattern help to account for individual
choices in styleswitching. These social patterns can be discussed in terms of
marked versus unmarked style choices ([Myers-Scotton]Scotton 1983). Most
speakers have at least two forms of their language, switching from one to the
other in predictable settings. For example, they may use one form in the home
and another in the market. Blom and Gumperz (1972) refer to this as situ-
ational switching. Distinct varieties of the language are employed for dis-
tinct situations (church, school, home) or activities (public speaking, cere-
monial speech).
However, switches also occur among family members and close
friends in the same conversation, often with no change in setting or
even participants. This occurrence indicates that the phenomena of
styleswitching is not confined to interactions in selected places but, rather, is
a natural part of conversation that can occur for a variety of reasons. Myers-
Scotton (1993) discusses non-situationally motivated switching under three
rubrics: switching itself as the unmarked choice, switching as the marked
162
Southern English 163
choice, and switching as an exploratory choice. Most of the switching dis-

cussed in this chapter can be characterized as switching as the marked choice.
Data for this study were collected from lower socioeconomic white speak-
ers who live in a small community located in the foothills of North Carolina.
I call the community Cedar Falls. Over a two-year period, the subjects were
videotaped taking part in various social activities, including two Christmas
Eve parties, a small church service, casual interactions among friends and
family members, and a Thanksgiving dinner. All the participants in this study
are family members or close friends living in the same speech community.
GOALS AND METHODS
Using examples from the data gathered for this study, I offer support for the
existence of strategic switching between the two linguistic varieties spoken in
Cedar Falls. Whether these varieties are best characterized as different dia-
lects or styles of the same dialect is open to discussion. Sociolinguists form no
consensus on this issue, and, in fact, most largely sidestep it altogether. This
sidestepping is what the term variety accomplishes, after all. Generally speak-
ing, if two varieties are labeled styles of the same dialect (rather than dialects
of the same language), their similarities outweigh their differences. Clearly,
the two varieties in question here have many structural features in common.
Yet, identifiable differences separate them at all three levels of linguistic struc-
ture (i.e., phonology, morphosyntax, and lexicon); the examples given here
make that separation apparent. However, assessing these differences via
quantification is beyond the scope of this chapter and not relevant to its
purpose. Therefore, I simply refer to one dialect, the Cedar Falls dialect, but
recognize two styles. The style that is the common medium for in-group con-
versations in this community is called home style, and the style more reserved
for outgroup interactions there is called local standard. Because the speakers
studied used home style for 95% of their conversational exchanges, home
style is considered the unmarked choice for conducting conversation between
family and close friends. Local standard represents the participants' approx-
imation of Standard American English. However, local standard only ap-
proaches the standard dialect, as exemplified on national network television
news programs. Speakers in this study may styleswitch into a different variety
that they believe is Standard American English but still use words or patterns
that are part of the rural southern vernacular—what I call home style here.
Sentence [1] exemplifies some of the most distinctive features of home
style:
[1] What I need me is some time and some money, but I ain't got neither one right
now.
The use of the personal dative me, the negative form of ain't, and the double
negatives, ain't and neither mark this sentence as home style. When speakers
164 STYLISTIC C H O I C E S IN SPOKEN E N G L I S H
use other forms that approach Standard American English, as in sentence [2],
I claim that they are using local standard, the local version of Standard
American English.
[2] I don't go to saloons to eat—even if they have good food, I don't want that kind
of environment.
To point out distinguishing features of home style is not to suggest, however,
that every sentence spoken in home style contains these features. In fact, if a
speaker were to use such features in every sentence, it would be very unusual
indeed. A strong indicator of a switch to local standard is that speakers speak
more slowly than usual. Thus, sentences that may appear to be in local stan-
dard in the following examples may be considered as part of a home style
discourse because they lack the feature of slower speech rate that is a defining
feature of the local standard. The question for this study is: If home style is
the unmarked choice for these speakers, and if they are surrounded by family
and close friends, why would these speakers styleswitch at all? If the situation
is unchanged and the participants remain constant, what is the motivation
for styleswitching? The following examples involve styleswitches from local
standard, from home style into local standard, or are examples wherein we
might anticipate a switch that is not forthcoming. Examining these examples
gives us clues as to why speakers choose the marked choice of styleswitching
during certain speech events.
The participants in this study are my family members and friends and
acquaintances of those family members. The participants were chosen on the
basis of the following sociocultural factors: their geographic identity (they
were all born and presently live in rural North Carolina), their educational
background (one female member of the group went to nursing school at the
local technical college; none of the other participants graduated from high
school), and their occupations and income (if employed, they have low-pay-
ing blue-collar jobs). Their community was once a thriving cotton mill town,
but the mills closed in the early 1960s, and the people who were not closely
tied to the community left for the cities. The people who remained behind
usually owned a piece of land and a small home. They now work in the near-
est small town, about 12 miles from the community. These jobs are generally
in fast-food restaurants and retail stores (Wal-Mart or K-Mart). There is also
seasonal work in the furniture industry. The members of this community are
generally very close and depend on each other emotionally and spiritually.
The older members attend church together regularly; the younger members
attend sporadically.
Instead of attempting to present all the features that constitute
home style, I give fairly long segments of the conversation showing
styleswitching, and for each conversation, I refer to the features of
the switch to make it clear to the reader that a switch took place. The
switch into a different variety (most commonly from home style to local stan-
dard) is in caps. The data were gathered by me, or by a person in the commu-
nity who was a member of my family. A videorecorder was used.
ANALYSIS OF DATA
The first example centers on Pearl, the matriarch of the family, greeting two
of her guests for a small Christmas Eve party at her door. She greets them in
local standard, and then immediately switches to home style. Note that (.) =
pause, // // indicates overlap. Switches are in caps.
[3] Pearl: 1 it's so nice to see y'all
2 I'm so glad you could make it
3 please come in and find a seat
[The guests seat themselves and Pearl continues]
[4] Pearl: 4 AIN'T THIS RAIN AWFUL?
5 I'M AWISHIN I COULD GO SOMEWHERE AND SEE ME
SOME SUN
6 I AIN'T NEVER SEEN NOTHING TO BEAT ALL THIS
RAIN
The only nonstandard variant that we see in sequence [3] is the use of the
plural form Y'all, but this is not a stigmatized form in southern English, be-
cause it is used across class boundaries in the South and thus is considered a
component of local standard. In sequence [4], Pearl sits down with the same
guests that she just greeted at the door. Now we see five separate examples of
the more distinctively local variety, home style. First, Pearl uses the negative
ain't to begin her conversation about the weather: AIN'T THIS RAIN AW-
FUL. Then a is prefixed before the verb wishing: I'm AWISHIN I could go
somewhere. In this same sentence Pearl uses the personal dative me, as in
AND SEE ME SOME SUN. Also, standard auxiliary verb selection is not
followed in I AIN'T NEVER SEEN. Finally, there is another occurrence of
the negative form ain't in a negative sentence to end the exchange: I AIN'T
NEVER SEEN NOTHING.
In sequence [3], Pearl greets her guests in local standard. When she goes
to the door, even though she is greeting close friends, she is aware that the
situation requires a relatively formal greeting. She wants to be seen as some-
one who is a "hostess," someone who knows and understands her duties as
hostess. Therefore, she uses local standard to establish this identity. Howev-
er, in the same exchange, as her guests are seating themselves, it is also impor-
tant for Pearl to reaffirm her identity as close friend and community member.
Her code choice in sequence [3] is a choice that permits her to present
herself with the status of hostess. Notice that Pearl uses three formulaic utter-
ances which are all echoic of what even an upper-class hostess at a formal
party might say to guests whom she does not know well. Pearl's choice in [4]
is a move toward declaring solidarity with her guests; she accomplishes this
by switching to home style, the style the guests recognize as their unmarked
medium of in-group social exchanges.
In sequence [5], we see Pearl facing a similar set of options. The partici-
pants in this sequence are Pearl and Dan; Tessa; Tessa's husband, Ed; and
Pearl's daughter Sheila. Pearl brings Tessa a chicken breast covered in sour
cream and bacon. Everyone is speaking in home style. The phone rings and
Pearl answers it, switching to local standard.
[5] Pearl: 1 [to Tessa] How ya like that?
Tessa: 2 It's good, [looking at her husband] You wouldn't like it (.} so
don't bother.
Ed: 3 I noticed you hadn't offered me none.
Sheila: 4 Mitch [her brother] don't eat em good neither.
5 His girlfriend keeps bringing em to him.
Dan: 6 What's that?
Sheila: 7 Them chicken breast with sour cream.
Dan: 8 [to Ed] You wouldn't like em so don't insult your taste buds.
//[laughter]//
Ed: 9 I don't plan on it (.) I can't eat no sour cream.
Sheila: 10 Speakin' of food (.) has anybody tried sushi?
Tessa: 11 Idn't that raw fish?
Sheila: 12 Oooh yes.
Tessa: 13 D'you like em?
Sheila: 14 Oooh no:o.
//[laughter]//
Sheila: 15 I kept thinkin' iffen they'd abattered em up an dropped em in a
deep fryer we'd all be eatin' good.
16 I mean might jus as well swallow up a goldfish.
17 I don't want me no raw fish.
[phone rings]
Pearl: 18 LOFTIN RESIDENCE (.} HAPPY THANKSGIVING.
19 HOW IS SHE?
20 UHHUH
21 DO THEY THINK THAT'S BASED ON HER MEDICATION?
22 I SEE (.) YOU KNOW I'M PRAYING FOR YOU BOTH
23 MY HEART IS WITH YOU
24 WHATEVER HAPPENS HERE WE'LL ALL MEET IN THE
RAPTURE,WON'T WE?
25 TELL HER TO REST ...
Ed: 26 //[to Sheila]//1 member when you throwed eggs on us at Easter.
Sheila: 27 I never did!
Pearl: 28 [still on the phone] GOODBYE DEAR.

Sheila: 29 I ain't never chunked the first egg at nobody (.) never.
//[laughter]//
[Pearl hangs up the phone as Sheila says line 29. She turns back to
the people in the room]
Pearl: 30 Well good Lord ( . ) what in tha(. )
31 I'm athinkin' y'all are all crazy.
In this sequence, Pearl is again faced with a limited number of options when
she answers the phone. The subset of ways to answer a telephone is relatively
small. Pearl, who is conversing with her family in home style, can choose to
answer the phone in home style or she can switch to local standard. Because
Pearl works part time at a shelter for battered women, there is the possibility
that she is being called about a problem at work. Pearl's work puts her in
contact with people of varying social status, including workers from the local
hospital and law enforcement agents. Features of home style, such as multi-
ple negation and nonstandard subject-verb agreement, are stigmatized by
the educated people in the town in which Pearl works, and Pearl is aware of
this; therefore, the cost of staying in home style when she answers the phone
could be a loss of status among her colleagues.
That explains why Pearl switches into local standard when she answers
the phone. The question is, Why does she choose to stay in local standard when
she finds out that the person calling is a close friend who was invited for dinner?
Faced with a set of choices similar to that in sequence [3], Pearl this time chooses
to continue to use local standard in the exchange. The answer to this question
lies in the topic of the conversation. Pearl is a retired practical nurse and the
topic that immediately comes up concerns a medical problem. Local standard
would be used by most of the community members if they went to the hospi-
tal or a doctor's office; it is the variety used for more formal occasions. Here,
Pearl can claim expert status (particularly among people in her speech commu-
nity, most of whom did not graduate from high school). So in line 21, when
Pearl asks about her friend's medical condition, she does so in local standard.
21 DO THEY THINK THAT'S BASED ON HER MEDICATION?
In lines 22 through 24 the topic shifts away from medicine, yet it still con-
cerns the physical well being of the caller's wife. Pearl shows her concern by
saying in line 22 that she is praying for you both and in line 23 that her heart
is with you. Line 24 is a serious religious reference that comes directly from
the Bible and therefore Pearl treats it with formality.
24 WHATEVER HAPPENS HERE WE'LL MEET IN THE RAPTURE WON'T WE ?
We don't know what the caller's response to this utterance is, but we do know
that even though the conversation going on behind Pearl is taking place in
home style and is very lighthearted, Pearl is not influenced by it. We can see
this is line 28 when Pearl closes the phone conversation with goodbye dear.
In this community, this is a marked way to end a conversation with a friend.
Phone calls in home style usually end with phrases like bye bye, love ya, or I'll
call ya soon. Pearl's choice of a marked closure indicates the seriousness of
the topics that were discussed in the phone call. When she turns back to her
family and rejoins their conversation, she immediately switches back into home
style.
As we can see, even though the situation limits the subset of responses, the
speaker still has choices. Before each switch in sequences [3], and [4], Pearl
is presented with these choices, and she makes them based on her knowledge
of her community and the social message she wants to convey. The situation
limits, but does not control, her choices.
In sequence [6], we see another example of styleswitching to establish iden-
tity. This time the speaker wants to move from his identity as a close friend
of the family into the identity of someone who is an expert or who has excep-
tional knowledge in a particular situation. Again, the situation propels the
styleswitch. The participants in this situation are family members: Dan (fa-
ther), Pearl (mother), Sheila (daughter), Mitch (son), and friends of the fam-
ily, Frank and Myra (husband and wife). Frank uses his styleswitch to com-
mandeer the new camcorder that Dan and Pearl received for Christmas. The
conversation begins in home style.
[6] Frank: 1 What in the world have y'all got now?
Pearl: 2 We got us a prize (.) the young'uns tol us ta open it up early.
Frank: 3 Well lookit that (.) ain't that sompin?
Dan: 4 January one (.) This says January one.
Pearl: 5 Well (.) that ain't us.
Dan: 6 This here things a year behind.
Pearl: 7 Hell (.) maybe it is us.
//[laughter]//
Sheila: 8 It says ta press initial button.
Mitch: 9 //no (.) that ain't it//
Myra: 10 Honey, [speaking to her husband Frank] don't you know sompin'
bout these things (.) ain't there sompin on tha side?
11 There's spose ta be sompin on tha side.
Frank: 12 WE NEED SOME EXTRA LIGHT (.) DO YOU HAVE AN OVER-
HEAD LIGHT?
Pearl: 13 We usta (.) but Ralph knocked it out with his numchucks.
Dan: 14 Is the thing on?
Pearl: 15 I think it's been steady on.
Frank: 16 JUST LET ME SEE WHAT I CAN DO.
Pearl: 17 Honey do your thing.
//[laughter]//
Frank: 18 I DON'T BELIEVE IT WILL GIVE YOU A PICTURE WITHOUT
MORE LIGHT.
Sheila: 19 Idn't it spose ta work in any light (.) don't it say something bout
candle light?
Frank: 20 No (.) that's another model (.) this'un NEEDS MORE LIGHT.
21 PUT A LAMP ON EITHER SIDE OF THE ROOM AN IT'LL
WORK FINE.
22 (.) I THINK IT'LL WORK FINE.
In line 1, Frank enters the conversation by inquiring about the present that
Dan is opening:
1 what in the world have y'all got now?
Pearl responds in home style, and in line 3, Frank assesses the gift using home
style:
3 Well lookit that (.) ain't that sompin?
Lines 4 through 10 demonstrate that Dan has no knowledge about the cam-
corder, in line 12 Frank styleswitches into local standard and takes control of
the situation.
12 WE NEED SOME EXTRA LIGHT (.) DO YOU HAVE AN OVER
HEAD LIGHT?
The other participants remain in home style throughout the rest of the ex-
change while Frank continues using local standard through line 22:
22 I THINK IT'LL WORK FINE.
Speakers have multiple identities; one way to change identities in an ex-
change is to change the linguistic variety that is indexing one identity to a
linguistic variety that will index another. Frank wants to move from someone
who is a guest and close friend to someone who knows more than the other
speakers about the camcorder. None of the other speakers challenge his right
as expert; they continue to use the unmarked home style. The host, Dan, turns
over the operation of the camcorder to Frank, and the other speakers start to
move lamps and set the room up as Frank requests. In other words, Frank's
styleswitch successfully establishes his identity as the expert in charge. This
example illustrates the point that different speakers in the same exchange may
have different marked and unmarked choices. When Frank uses local standard
to negotiate a new identity, that variety becomes his unmarked choice. His bid
for a new rights and obligations set (RO set) is successful, even if it seems that
he is using marked choices for the situation without response. His status as expert
is accepted by the other speakers, as evidenced by their behavior. To negotiate
a new RO set for an exchange does not necessarily mean to negotiate a new
variety for all participants in the exchange. In fact, in these particular
circumstances, if other speakers were to switch as well it would be interpreted
as a challenge to Frank's authority, as though the speakers who switched

were also establishing themselves as experts, and an argument would almost
certainly ensue.
In sequence [7], someone is making the marked choice to enter a conver-
sation using local standard in response to a face-threatening act. This se-
quence starts with a conversation between Pearl and her niece Tessa. Dan
(Pearl's husband) is listening to this conversation but does not enter until the
last exchange. He joins the conversation using local standard.
[7] Pearl: 1 We usta would go ridin in tha mountains.
2 But I hate the ole interstate (.)
3 There ain't a blamed thing ta see on an ole interstate.
4 But Dan (.) he won't never drive off the interstate.
5 He's always afraid he's agonna get lost.
6 I tol him (.) honey there ain't no way—
7 I mean you jus can't get lost in America.
//[laughter]//
8 but we jus ride ride that ole interstate jus ride that ole interstate
Tessa: 9 Is he worried bout gettin lost do ya think?
10 Maybe he's worried bout havin ta ask some mountain man for di-
rections.
//[laughter]//
Pearl: 11 That's more likely.
12 He says you get back up in there where nobody don't know ya and
y'll get shot,
//[laughter]//
Tessa: 13 Are ya jus a little paranoid Dan?
//[laughter]//
Dan: 14 NO (2) I'M A REALIST.
15 I KNOW PEOPLE UP IN THERE DO THINGS THEY DON'T
WANT FOLKS TA KNOW ABOUT AN YOU RIDE UP ON A
STILL UP THERE Y'LL BE MISTOOKEN FOR SOMEBODY ELSE.
16 YOU CAN GET YOURSELF SHOT
In this sequence, Pearl and Tessa laugh about Pearl's husband. They have
insinuated that he might be unnecessarily afraid of both driving in the moun-
tains and asking for directions. Although Dan cannot dispute the facts of the
story because he did indeed tell his wife he would not drive off the interstate
for fear of being attacked, he can still establish himself as a reasonable person
with reasonable fears. He does this in a number of ways. First, he starts the
exchange with a simple denial in 14, no. He then pauses for a full two seconds
before he continues. When he says, I'M A REALIST, he is using a lexical
item, realist, which is not normally in his code; this is the first indication he
has switched to local standard from home style. (Note that Dan has only two
switches in eight hours of data.) The fact that he switches here indicates that
he feels a strong need to take control and assert power; therefore, he assumes
the prerogative of initiating a change in the code that is being used. He tells
us in lines 15 and 16 that he knows something about the hills of North Caro-
lina, and that there is a real danger in venturing off the main roads and stum-
bling on a moonshiner's still. When Dan says, I'M A REALIST, he implies
that Pearl and Tessa are not being realistic when they talk about driving
through the mountains. Pearl and Tessa have to choose whether they will
continue the exchange, possibly counterattacking Dan, or whether they will
yield to Dan's bid for power. They choose to yield, possibly because Dan's
use of local standard alerts them to the fact that they have insulted him. After
he speaks, Pearl and Tessa nod seriously at each other and at Dan. They then
take up another topic.
The styleswitches seen in sequences [3] through [7] are all switches where-
in the speakers are negotiating power in the interaction. With this power comes
a change in the speaker's position in the RO set in effect for the conversation-
al exchange, an acknowledgment of authority and expertise. Therefore, to
negotiate such a change they make a marked choice, in this case the use of
local standard, so that they might gain in some way. They "rock the social
boat" (Myers-Scotton 1988: 202). In the first sequences, Pearl's styleswitch-
es gain her a strengthened social role. When Frank switches to local standard
in sequence [6], he gains status as an expert. And when Dan uses local stan-
dard in sequence [7], he establishes himself as a thinking person who knows
how to act responsibly. By virtue of initiating a change in the relationship,
the speakers indicate a desire to communicate superiority or authority. The
success of the choices they make can be seen in the responses of the other
participants.
However, other reasons to initiate a codeswitch in a conversation exist.
In sequences [8] and [9], the speakers styleswitch from home style into local
standard when they become contemplative, remembering a poignant past
event. In sequence [8], the participants are Pearl, her daughter Sheila, her
niece Tessa, and her friend Myra. It is Thanksgiving Day, and they are re-
membering a Christmas Eve party they attended the year before. While this
conversation goes on with all the speakers using home style, another family
friend, Louise is seen turning into the driveway of Pearl's residence. At first
the arrival of this friend triggers sarcastic remarks because Louise's husband
(Richard) makes a point of not joining these gatherings. He always has an
excuse; usually he begs off by claiming illness. Pearl suddenly urges her other
guests not to expect too much of Louise during the holidays because this
is her first Thanksgiving and Christmas since the death of her father. Pearl
reminds everyone how hard the holidays are for someone who is grieving
for a loved one and this reminder triggers a memory about her youngest
daughter who was killed twelve years earlier. Pearl styleswitches into local
standard as she remembers her daughter, the funeral, and the first Christ-
mas she spent without her. In this example, the conversation goes from being
very lighthearted to sarcastic to very serious, and we can see that the topic
change triggers the styleswitch. The discussion of death and the memory of
the funeral of Pearl's daughter motivates the switch.
[8] Pearl: 1 Frank was crazy over there with Sheila.
2 Tryin ta open her presents.
3 And she opened one and it was panties and he says what size are
they? an she says my size!
//[laughter]//
Sheila: 4 He din't need ta know.
5 They was my size.
Pearl: 6 An he was opening up uh uh
7 Eve'rything she'd open he was into it.
8 An he says "coulda been mine."
Sheila: 9 //I think it's Louise//
Pearl: 10 "If it hadn't been for the name tag."
Pearl: 11 An he wadn't gettin his present fast enough.
12 Mitch was passin um out an he was makin all manner of fun at it
ya know.
13 Mitch said (.) "finally it's yours" an held it over to look at it.
14 An he said "I don't believe it (.) I wanna see tha tag"
15 Louise is all by herself
Tessa: 16 She are? [sarcastic]
Pearl: 17 [very sarcastic] Ri:chard jus coul:dn't fa:ce it.
Sheila: 18 [High pitch] Wimpy wimpy wimpy.
Tessa: 19 Do we do we do we hear a hint of?
//[laughter]//
Pearl: 20 She is real (.) she's kinda down in the dumps.
21 IT'S HER FIRST HOLIDAY WITHOUT HER DADDY
22 IF SHE DON'T ACT PLUMB NORMAL IT'S CAUSE SHE CAN'T
HARDLY (.) IT'S AWFUL. IF YOU'VE NEVER BEEN
THROUGH IT
23 THERE'S HOLIDAYS STILL AIN'T THE SAME.
24 I BROKE DOWN THE OTHER DAY AN ASKED SHEILA (.)
WHEN DOES THE GRIEF EVER END?
25 IT'LL SOON BE 12 YEARS SINCE DEBRA(.)
26 I AIN'T (2) GOT OVER IT YET(.) I REALLY MISS HER HOLI-
DAYS.
Myra: 27 DIDN'T DADDY DIE THA SAME YEAR? 81?
Pearl: 28 YEA(.) YOUR DADDY DIED WHEN?
Myra: 29 June.
Pearl: 30 SHE DIED JULY 8TH.
Tessa: 31 Lord it don't seem like its been that long.
Pearl: 32 81 IN JULY AND IT'LL BE 12 YEARS COMING UP (.) TO ME
IT SEEMS LIKE A MILLION YEARS (.) AN OTHER TIMES IT
SEEMS LIKE YESTERDAY.
33 I can still smell them horrible flowers at the funeral home. (2)
34 NOT THAT I WADN'T GRATEFUL FOR THEM I WAS (1) BUT
THAT'S A HARD SMELL TO FORGET.
35 THAT'S A HARD CHRISTMAS THAT COMES NEXT
The recollection of a formal and sad occurrence triggers a shift to formal and
serious language (local standard). This type of switch is brought on by the
internal factors of contemplation and memory. Pearl makes the marked choice
to switch to a formal code to signal that she is talking about a formal and
solemn occasion.
It is important to recognize that there is a difference between conversation
about a formal and solemn topic and conversation that is serious. Partici-
pants do not switch to local standard every time the topic shifts to a very
serious subject. For example, sequence [9] is a conversation about two friends
of Pearl whose son was arrested for murder. The friends, Lynn and Pam,
cannot attend the Thanksgiving day dinner because Pam is too upset about
her son (Cody) being in jail (for murdering his friend, Sid) to socialize. In this
sequence, Sheila explains to Tessa why Lynn and Pam will not be at the gath-
ering and what happened to Cody. Even though she is talking about a very
serious chain of events, she does not shift to local standard.
[9] Sheila: 1 They're saying he [Cody] killed that guy [Sid].
2 But Sid weren't no account.
3 He was meaner than a snake.
4 If Cody did kill him it weren't no more than he deserved (.) got
what he deserved.
5 Did you know sompin?
6 He [Sid] rode him around with a head in tha trunk of his car for a
week once.
7 Everybody knowed he was bad (.) awful.
8 But folks was scared of him.
9 Now tha law is actin all crazy cuz they found him shot dead in
his car.
10 I don't know if Cody did it or what.
11 But his poor momma is going wild.
12 An for what (.) for some trash that's dead an nobody cares.
13 He [Sid] was a sorry son of a bitch.
14 Whoever shot him oughta get a medal.
Sheila recounts this narrative to Tessa without any styleswitches to local

standard, even though the topic itself could not be more serious. A topic shift
to a serious subject is not enough in itself to trigger a shift. Sheila becomes
more vehement as she talks, but she does not styleshift.
However, in sequence [10] we have another narrative with a serious topic
that does cause the speaker to shift from home style to local standard. Se-
quence [10] is taken from data gathered at a small rural church. The church
has a very small membership and shares a pastor with two other small churches.
When the minister is unavailable, the members gather in the sanctuary and
share personal testimony about the power of God in their lives. This sharing
is referred to as testifying. The following transcript comes from a male in his
early sixties who rises and tells the congregation about the time that he knew
he was saved by the Lord. Testifying in church calls for a certain discourse
structure. If the testimony is about the moment in time when the speaker was
saved by God, the speaker starts by recounting his life before he was saved.
He then discusses the moment or time of salvation and recounts how his life
changed from that moment to the present. In the following sequence, the
speaker starts his testimony using home style. When he remembers the time
that he found God, he styleswitches to local standard. The switch seems to be
triggered by the recollection of this solemn event in his life.
[10] 1 Hello friends (.) I come ta talk to you today bout tha power of the Lord
2 I member tha time as a youngun (.)
3 My momma and daddy'd pray for me ever night.
4 I wuz wilder than a hare.
5 I wuz headed for that no good place (2) an they knowed it (.) an they prayed.
6 Didn't nothing they said make no differance ta me.
7 I made ma poor momma and daddy cry many's tha night.
8 I got married (.) but it didn't change me none.
9 I kept on with ma ways (.) kept on.
10 THEN THERE CAME A DAY
11 IT WAS SEPTEMBER 9TH 1963
12 I WAS RIDING IN MY CAR COMING BACK FROM RALEIGH
13 AND I SUDDENLY FELT THE POWER OF THE LORD IN THA CAR
WITH ME
14 AND GOD SPOKE AS PLAIN (2).
15 AND HE SAID IT'S TIME TO COME HOME.
16 AN I FELT THE POWER OF SALVATION RIGHT THEN AND THERE
IN MA HEART AND I KNEW NOTHING WOULD EVER BE THE
SAME.
17 FOLKS (.) YOU HAVE TO FEEL THA POWER OF THE LORD FOR
YOURSELF.
18 AS MUCH AS THEY WANTED TO (.) MY DEAR MOTHER AND
FATHER COULDN'T PRAY ME INTO HEAVEN.
19 I HAD TA FIND THAT ROAD.
20 I HAD TA HEAR THA VOICE.
21 AN I THANK GOD EVER DAY.
22 AN I PRAISE THA LORD EVER DAY.
The speaker in this sequence starts his narrative using home style. This use
allows him to signal his solidarity with the rest of the congregation. By keep-
ing the details of his past vague and by using the more casual form of the
dialect (home style), he allows them to share in the first nine lines of his story.
In this small rural Baptist church all the listeners believe that you must first
fall and then accept Christ in your life. By starting his story with his failure
to lead a good life and the worry that caused his parents, he is telling the
hearers something that they can understand and with which they can iden-
tify. In lines 3 and 7, the speaker refers to his parents as momma and daddy.
He seems to be using home style to identify that time when he was a child,
not young in years, but too young in his mind to accept God, so that his
parents had to try to pray for his salvation because he would not do it him-
self. In line 11, the speaker recounts the exact day and year that he found
God. The memory of this powerful event marks the styleswitch into local
standard, and the speaker stays in local standard throughout the rest of his
narrative. This switch is powerful and shows the speaker is manipulating his
language to make a point. On the day he found God, he became an adult; his
life changed completely. To further demonstrate this change, in line 18 the
speaker refers to his parents as my dear MOTHER AND FATHER, termi-
nology used by an adult.
In sequences [8] and [10], styleswitching is triggered by contemplation
and memory. It is used by the speakers to signal that they are discussing a time
or an event that is both important and serious. The topics in these shifts (death
and salvation) are solemn subjects, and the speakers styleswitch to local stan-
dard when they remember and want to convey how serious the events were.
In sequence [9], Sheila recounts a serious narrative to Tessa, but even though
she is discussing the problems faced by friends of the family, the message is
neither personal nor formal for her: Therefore, Sheila makes the unmarked
choice and stays in home style. We might predict that if Cody's mother was
telling the story related by Sheila, she might styleswitch into local standard.
For her, these events are personal, solemn, and tragic, and her code choice
would probably indicate this for the listener.
The final two examples come from the patriarch of the family, Dan. In
sequence [11], Dan is telling about a retirement dinner given for an executive
in the company for which he worked. The dinner was being held at the local
country club, a place to which Dan had never been invited before. The menu
was pre-selected, with prime rib as the main course. Dan tells the family
about the evening.
[11] Dan: 1 So we's all at this party at the Country Club an they's serving
prime rib.
2 An I hadn't never had none of that though I'd heard of it
3 So they come abringin' it out an the boy sets mine down in front
of me
4 An let me tell you the thing was aswimin' in pure ole red blood an
there weren't no way I was gonna eat it
5 So I looked at the boy an I says to him
6 SAYS I (.) YOUNG MAN (.) I BELIEVE I'LL HAVE MINE
COOKED.
//[laughter]//
Dan is attempting to achieve at least two goals with his shift. First, he is
making fun of the posh surroundings. At this event, it probably feels to him
that everybody is putting on airs, and he shows that he can put them on, too.
But also, he shows his family and friends that even though he thought the
behavior of those around him was funny, he is a man who knows how to use
what would be considered proper speech in this type of situation. He knows
how to send unsatisfactory food back and how to behave properly at the
country club.
Sequence [12] is an example of someone refusing to styleshift when to
styleswitch suddenly becomes the unmarked choice. The reason we can say
that styleshifting into local standard here is the unmarked choice is that the
speakers are all talking about travel overseas. Some of the members of the
community have traveled overseas, but only if they were in the military and
stationed at an overseas base. Usually this means Germany, which is one of
the few destinations to which servicemen and -women can take their families.
The people in the group start discussing their travels to Germany, and they
switch into local standard. This may be because a variety approximating
Standard American English is the unmarked choice when they are overseas,
and thus local standard is the unmarked choice when they are reminiscing.
The participants who were abroad begin to talk about daytrips to France,
and they discuss how beautiful Switzerland is. The listeners respond in short
sentences in local standard, such as it sounds wonderful and I would love to go
someday. At one point in the conversation, Pearl turns to her husband and says,
"Oh Dan wouldn't you just love to go there." Dan responds in sequence [12].
[12] Dan: 1 I done been ta the ocean on the one side and the mountains on
t'other
2 An that's all the fartherest I intend ta go
Dan's refusal to switch to the style that is prevailing in the group marks a
powerful manipulation of his language. In home style, the term done been
marks finality, that is, something that was done and probably will not be
repeated in the near future. The adjective all in line 2 indicates the full amount,
or rather the total limit of what Dan intends to do. And finally the invention
of a new superlative, fartherest, indicates that Dan will absolutely not travel
any farther from home than the mountains or the beach. He could have said
he has gone "as far as he wants to go," or that he wants "to go no farther."
But Dan combines the terms done been, all, and fartherest to let his listeners
know that he is not interested in travel to Europe or anywhere else too far
from home.
When we examine Dan's word choice, along with his refusal to accommo-
date his style (once we know that styleswitching is a possible choice in his
repertoire), we see a powerful manipulation of language. Dan's refusal to
styleswitch puts not only himself but also his world in a positive position. He
has the mountains on one side and the ocean on the other. This puts him and
his home in the center of the universe, and his staunch refusal to move far,
either in language or in location, lets the hearer know that this is exactly
where he wants to be.
CONCLUSION
The data from this study show us that people alternate between styles to make
very subtle changes in day-to-day relationships. The ability to shift styles, to
change both mood and tone, to switch identities to what appears to be the
most favorable incarnation for an interaction, is at the heart of styleshifting.
The ultimate reason for styleshifting is to promote ones self in the most pos-
itive light, and here we see ordinary people negotiating language in their home,
showing themselves to be linguistic virtuosos.
REFERENCES
Blom,J. P., and J.J. Gumperz. 1972. Social meaning in linguistic structures: codeswitch-
ing in Norway. In Directions in sociolinguistics, ed. John J. Gumperz and Dell
Hymes, 407-435. New York: Holt, Rinehart and Winston.
[Myers-Scotton, Carol] Scotton, Carol Myers. 1983. The negotiation of identities in
conversation: a theory of markedness and code choice. International Journal of
the Sociology of Language 44: 115-136.
. 1988. Codeswitching as indexical of social negotiation. In Codeswitching: an-
thropological and sociolinguistic perspectives, ed. Monica Heller, 151-186. Ber-
lin: Mouton de Gruyter.
. 1993. Social motivations for Codeswitching: evidence from Africa. Oxford:
10
Marked versus Unmarked Choices on the

Auto Factory Floor
JANICE BERNSTEN
D ISCOURSE RESEARCH on talk at work largely focuses on the analysis of

naturally occurring language in settings such as law and medicine (Drew
and Heritage 1992). It is easy to see how special conditions apply in a venue
such as a courtroom, with all the rules constraining the talk of judges, lawyers
and witnesses. What is less evident is that the rules of engagement are just as
stringent in such work settings as an automobile plant. There, it is not the law
books that provide the rules but the contracts between union and manage-
ment. Equally important from the point of view of the workers are the un-
written expectations for appropriate talk between management and labor.
The purpose of this chapter is to show how talk in the automobile factory is
constrained by participants' tacit knowledge of expected or unmarked lan-
guage choices and how participants interpret marked choices. I demonstrate
that communication in the factory was hampered in the last few years by the
entrance of supervisors lacking assembly-line experience and, therefore, lacking
the mental representations for making appropriate choices under the marked-
ness matrix characteristic of the factory setting.
Drew and Heritage (1992) point out the special nature of institutional
talk, which makes it of interest to researchers. Institutional goals lead to
constraints and expectations such that participants may be expected to fol-
low a different set of rules than those that apply to conversations in a social
setting. As Miller writes:
Institutional discourses are shared and standardized frameworks for anticipat-
ing, acting in, and reflecting on social settings and interactions. They allow and
constrain setting members to organize their interactions as instances of stan-
dardized types of social relationships and produce conditions for responding to
issues in predictable ways. (1994: 282)
In the case of a unionized automobile plant, the standardized social relation-
ships on the factory floor are those between worker, supervisor, and union
178
Auto Factory Workers 179
representative. Each of these players contributes to institutional talk based

on scripts developed over years of negotiation between union and manage-
ment. The data that follow show that a principal goal of the workers is to
fight against the power asymmetry that Drew and Heritage note as a custom-
ary characteristic of institutional settings. The resulting scripts show unmit-
igated language to be the code choice for all the players.
I became interested in investigating shop talk because of linguistic exam-
ples provided in my sociolinguistics classes in Flint, Michigan, by students
who had strong family connections in the local auto industry. Most of these
students are the first generation in their family to attend college, and most
come from the families of auto workers or are auto workers themselves, coming
to college to retool for the changing demands of today's economy. These stu-
dents reported that linguistic principles were very valuable in explaining some
of the communication problems they were experiencing on the job. At the
end of a lecture on Myers-Scotton's markedness model (1993), a student came
up and said, "It was a marked choice!" He had a new manager who was
perceived as totally ineffective on the factory floor. The man would approach
two workers talking on the line and wait until one was finished before ad-
dressing them. Then he would say something like "I hate to bother you, but
could you possibly do X." My student explained that on the factory floor, the
unmarked directive form is Do X. The marked language choices of the man-
ager were clear evidence to his workers that he lacked the skill to manage. I
became very interested in finding out more about what did constitute effec-
tive management language in the eyes of local factory workers. How could
the markedness model be applied in this setting to account for the workers'
evaluation of appropriate and inappropriate linguistic choices by supervisors?
MYERS-SCOTTON'S MARKEDNESS MODEL

Myers-Scotton (1993) argues that speakers have, as part of their linguistic
capacity, a markedness metric which allows them to judge linguistic choices
as more or less expected for given interaction types. Each setting has norms
specifying the expected or unmarked set of rights and obligations for each
participant. Both marked and unmarked choices emanate from Myers-Scot-
ton's negotiation principle:
Choose the form of your conversation contribution such that it indexes the set
of rights and obligations which you wish to be in force between speaker and
addressee for the current exchange. (1993: 11)
Speakers sometimes make the unmarked choice, which has the effect of main-
taining the status quo in terms of rights and obligations between speakers.
On the other hand, they may make an unexpected or marked choice, which
has the effect of changing the set of rights and obligations between speaker
and addressee.
Of course, each speech community has its own set of norms and its own
set of choices which will be viewed as marked or unmarked within given
contexts. Markedness can only be interpreted within specific interactions in
specific communities. As noted, talk at work has a discourse structure that
differs substantially from ordinary conversation. An outsider in the work
setting may err, as did the supervisor in my student's story, by making a code
choice that in a general setting could be viewed as unmarked. However, in the
auto factory setting, his directive choice was viewed as a marked choice, lead-
ing workers to view him as abdicating the power a supervisor should display
by using bald imperatives in the work setting. Although it was not his inten-
tion to alter the rights and obligations between himself and the workers, his
marked choice led to that result.
DATA COLLECTION
Whereas the students provided brief anecdotal examples of encounters in the

factory, I needed more evidence to determine the whats, hows and whys of
discourse in the automobile factory. To discover the marked and unmarked
language in the factory setting, it was essential to collect more examples of
naturally occurring talk and also the workers' evaluation of this speech. Al-
though there is a fair amount of literature on talk in the courtroom and medical
office (Atkinson and Drew 1979; Fisher and Todd 1983; Mishler 1984; West
1984; Levi and Walker 1990; Conley and O'Barr 1990), relatively little work
has been done on shop talk. This lack of data is due in large part to the fact
that factory managers consistently prohibit outsiders from doing such
studies within their facilities. Auto production companies fear industrial
espionage, litigation, and negative publicity which, according to Form
(1976: 140), kept investigators out of plants for many years. I ran into
similar resistance when I applied for permission to collect data in local
auto production facilities.
When it became clear that doing research within the factory setting was
not possible, I proceeded to collect data by interviewing factory workers
outside the workplace. Fifteen male and fifteen female workers were inter-
viewed, resulting in thirty-five hours of transcripts. The workers initially
talked with me individually at a local coffee shop and then started keeping
notebooks with linguistic examples, particularly those in which there was a
communication breakdown between supervisor and worker which could
illuminate marked and unmarked linguistic choices. This sort of data collec-
tion was advocated by Cameron et al. (1992) who criticized traditional so-
cial science research methodology, particularly as it is applied in studying
relatively powerless groups, such as factory workers. They encourage mov-
ing away from research on subjects to research on, for, and with subjects.
This sort of research implies "the use of interactive or dialogic research
methods, as opposed to the distancing or objectifying strategies" (Cameron
et al. 1992: 22) traditionally used in social science research. Indeed, the
interviewees and data collectors for this study were enthusiastic partici-
pants. One man said, "I am so glad someone is interested in our opinions
and sees us as skillful workers and worthwhile individuals—not as stupid,
lazy factory workers."
Workers were equally enthusiastic when it came to participating in a more
traditional data collection procedure—filling out questionnaires designed to
follow up on insights gleaned in the original interviews. A mid-Michigan local
of the United Auto Workers (UAW) allowed me to take up the first twenty
minutes of its monthly meeting administering a brief questionnaire which
combined free response and multiple-choice items.
Sixty-four workers completed the questionnaire (forty-one males and
twenty-three females). These workers are representative of the current
UAW membership in mid-Michigan in being part of a shrinking work
force. General Motors has done a great deal of downsizing in the last twen-
ty years and very little hiring. The average age of the workers is forty-four;
only one worker below age thirty participated in the survey. The workers
have an average of twenty-two years on the job and are thus intensely steeped
in factory culture.
According to my informants, institutional talk on the factory floor is highly
constrained by the rules in national and local union-management contracts.
These codes have the effect of depersonalizing communication and creating
"scripts" for many encounters between workers, supervisors, and union rep-
resentatives. Before taking a look at the actual talk, it is useful to take a step
back into the history of the auto industry to find the sources of current fac-
tory discourse.
HISTORICAL DEVELOPMENT OF SHOP TALK
In the early days of the auto industry, the production process was not yet
standardized. Making cars was a craft in which the work was varied and
workers controlled production. Management was paternalistic. Gartman notes
that in those early days, Henry Ford "relied on personal intimacy and asso-
ciation to control his workers" (1986: 180). The result was talk between
workers and managers which focused on solidarity rather than power. Nev-
ins and Hill write of Henry Ford in 1904:
Everybody used to call him Hank or Henry... and he used to know everybody
by name. He seldom gave a direct command. Instead, he would say, "I wonder
if we could get this done right away," or, "It would be fine if you could do so-
and-so." These hints were effective. The men would just break their necks to see
if they could do it. They knew what he wanted. (1954: 270-271)
As the factories grew, the work also became more mechanized and workers
had far less control over their labor. Foremen took over supervision and in
182 STYLISTIC CHOICES IN SPOKEN E N G L I S H
those days, the power of the foreman was arbitrary and unchallengeable. The
speed of the line, one's chances for promotion, indeed, keeping the job
itself—all were under the control of the foreman (Gartman 1986: 182).
One woman told me about a foreman who would pay a different piece work
rate for exactly the same job, depending on how well he liked the worker. When
the women complained, he told them, "There are plenty of people waiting out-
side the plant for this job if you don't want it." Another man told me of a
foreman who gave higher wages to men willing to mow his lawn and bring
him cases of beer.
Pressure to standardize arbitrary work practices came from both above
and below. Management established employment departments to centralize
worker discipline and base sanctions on written rules and records instead of
the manager's personal opinions. The UAW, established in 1937, worked
for job classifications so that wages were associated with work done rather
than the individual doing it. The UAW also put in place the seniority system
and overtime pay policies. These efforts stripped the foremen of power and
distanced their personal involvement in conflict. As Form writes: "Griev-
ances are not resolved by on-the-spot consultation with the foreman because
grievances are often categoric and not individual matters. An elaborate griev-
ance machinery involving increasingly high echelons of the union and man-
agement handles thousands of complaints annually" (1976: 140).
SHOP TALK SCRIPTS
Three different publications outline the "grievance machinery": the national

agreement between UAW and the auto production company, the local agree-
ment between a plant and a specific union local, and the employee hand-
book for the plant (UAW 1993; North American Truck Platforms/Flint
Assembly 1994; Employe Handbook n.d.). When a conflict arises, the
first step is for worker and supervisor to attempt to solve the problem
themselves. If the conflict is task related, supervisors must make expec-
tations clear; this comes in the form of the statement I'm giving you a
direct order. If the worker doesn't comply with the order, he or she is in
danger of being written up and eventually suspended from the job. In
that case, the supervisor must say I'm writing you up. Workers can be
written up not only for disobedience but also for any of thirty-six other
code violations in the Employe [sic] Handbook (n.d.) of the Flint Truck
and Bus Group. These range from being late to fighting to throwing refuse
on the floor.
I'm giving you a direct order and I'm writing you up have the force of
performatives as described by Austin (1962) because their utterance re-
sults in a change in the world. In the factory setting, when the supervisor
utters these statements, the result may be the loss of the worker's job. The
seriousness with which workers take these utterances is illustrated in an ex-
ample from Kearns, quoting a worker who cut her hands doing her job on
the assembly line.
On both my hands, I'm bleeding all over the place, through my gloves and I said,
"I've got to go to first aid now. Now I have to go."
He said, "No, you finish your job, then you can go."
But what am I going to do? I don't want to get fired for not following a direct
order. Okay? And that is an automatic fire. (1990: 67)
The woman kept working, the parts she was assembling continuing down the
line covered with blood, much to the dismay of the workers after her on the line.
If a worker disagrees with the supervisor's view of a situation, or if the
worker wants to initiate a grievance, the script is Get me the committee man.
The committee man, as the union representative is called in the factory, comes
to the floor to negotiate the conflict between worker and supervisor. Con-
flicts not solved by discussion among worker, supervisor, and committee
man may be appealed through several levels—first, to the shop committee,
then to the corporation and international union, and, finally, to an impartial
umpire at the highest level (UAW 1993).
One source which details several conflicts on the shop floor is Rivethead,
Ben Hamper's account of his twenty years on the auto assembly line. The
following example is a conflict script featuring the talk described previously
plus another feature of shop talk described by all the interviewees in the study:
the pervasiveness of profanity on the factory floor.
Supervisor: Hamper, that was the General Foreman on the line. According to
him, you've run three wrong axles in the past hour. Goddamnit, I'm writin'
your ass up NOW!
Worker: I don't give a shit how many times you try to write me up. The whole
problem is that I am too fucking short to see what the hell I'm doin' on that
job. How can I hit the carrier arms when I can't even SEE the bastards?
Supervisor: Don't try to peddle me that line of bullshit. I've had short people run
that job before.
Worker: I want my committee man down here—NOW! (Hamper 1986: 107)
This shop talk differs quite dramatically from Henry Ford's hint direc-
tives of days gone by. In fact, these linguistic forms are consistent with insti-
tutional talk on the factory floor in general: Language is direct and unmit-
igated. There are several origins of this direct language. First, as noted ear-
lier, clear orders and directives are mandated by union contracts. Second,
the factory environment is often so loud that it is difficult for workers to hear
heavily embedded directives. Third, according to Ervin-Tripp (1976), we
can expect unmitigated directives when persons are asked to do something
generally expected to be their responsibility. Children say to mom I need a
cookie, customers say to a McDonald's clerk Give me a hamburger, and
supervisors in the factory say to the worker Go and sweep. Workers expect
direct language from supervisors and as shown later, mitigated language often
receives a negative evaluation.
What may be less expected is the unmitigated language, which workers
use to address supervisors. We find that workers address everyone—worker,
supervisor, plant manager—up to the top management by first names. One
woman showed me a piece she wrote in the plant newsletter called Why
can't we compete, Pete? criticizing the policies of the general manager of
the plant whose first name is Peter. When asked on a questionnaire how they
would request time off from a supervisor, more than half of the union work-
ers chose not a question (e.g., Could I have some time off tomorrow to go
to the doctor's? Would it be possible to have some time off tomorrow?)
but the statement / won't be here tomorrow afternoon, I'm going to a
doctor's appointment. Cavendish, in describing eight months working in
a British auto firm in the late 1970s, relates an incident with her super-
visor which illustrates the unmitigated nature of talk between workers
and supervisors:
He [the supervisor] acted as if he was doing us a great favour whenever he brought
components or took away boxes, and stood waiting for us to say "please" and
"thank you." We bawled him out because it was his job to fetch and carry and
no one said please and thank you to us every time a tray came down the line.
(1982: 104)
It is clear that talk on the factory floor is full of unmitigated directives and
vulgarity and lacks politeness markers such as please and thank you. Again,
looking at history can put this language into perspective. Literature from the
UAW paints the traditional relationship between workers and management
as a long-term class struggle dating back to late sixteenth-century Europe
(UAW n.d.: 3). According to Nash, "the auto worker, despite high pay and
other benefits, remains among the most militant and angry members of the
American labor movement" (1976:61). Nash points out that the major source
of job dissatisfaction among auto workers is powerlessness in the work sit-
uation. The UAW was formed in the first place to offer resistance to the
power asymmetry in the factory. Workers can fight against this asymmetry
of power by using the language of equals to their supervisors, knowing the
union will back them when conflict arises.
The desire to be treated with respect as individuals and workers was a
common theme in Collinson's (1992) work in a British factory and was echoed
in mid-Michigan interviews and questionnaire comments. When asked to de-
scribe their least effective supervisors, 44% of the UAW members included
comments involving respect. A thirty-six-year-old female assembler wrote of
her supervisor, "She was disrespectful and treated employees like children that
needed to be babysitted." A forty-year-old male wrote, "He acts like I'm the
boss and it doesn't matter what you think." Workers respond to perceived
inequalities with straight talk. A forty-eight-year-old male assembler wrote:
My supervisor was acting like a "boss" not as a supervisor. I'm not asking him
to "kiss my ass" but I'm certainly not going to kiss his. I finally told him, "I can
be a cooperative employee or an uncooperative employee. Which do you want?"
We at present time get along much better.
The unmitigated nature of shop talk marks the discourse as very different
from that expected in ordinary conversational interaction. In their polite-
ness model, Brown and Levinson (1987) describe in detail the efforts speak-
ers usually make to defuse or mitigate face threatening acts through positive
and negative politeness strategies. Coulthard notes that bald on-record acts
"are not frequent" in conversation (1985:51). Brown and Levinson observe:
"The majority of natural conversations do not proceed in this brusque fash-
ion at all" (1987: 95).
If the conversational rules between supervisor and worker in the auto
factory differ from those in other discourse settings, we can assume that there
is a potential for misunderstanding when a person who lacks experience in
the factory becomes one of the players. Traditionally, supervisors came from
the ranks of workers on the line who were looking for advancement. Re-
cently, however, due to increased educational requirements and reluc-
tance of workers to leave the security of the union, supervisors coming to
the floor lack experience on the line (Chinoy 1992). Some of the misunder-
standings described by the workers can be traced to a lack of knowledge of
shop talk. On the other hand, other conflicts arise between supervisors
and workers who know the factory discourse "rules" perfectly well but
choose not to follow them. The interviews and questionnaires provide
examples of both intentional and unintentional marked choices on the part
of supervisors.
MARKED CHOICES ON THE FACTORY FLOOR
An Intentional Marked Choice

The following exchange really puts into focus the unmitigated directive as the
unmarked and expected code choice of a supervisor in an auto factory.
[1] Conversational sequence between experienced supervisor and experienced worker
who is serving as an all-purpose replacement (Bernsten 1995).
Supervisor: "Go do Bill's job." [Worker complies]
Supervisor: "Go do Tom's job."
Worker: "I can't do Tom's job."
Supervisor: "I'm giving you a direct order." [Worker tries to comply—returns
to supervisor]
Worker: "I can't do Tom's job."
Supervisor: "Go sweep." [Worker complies]
Supervisor: "Do you think you can do Helen's job?"
Worker: "Get me the committee man."

[Worker complies; supervisor writes request for union rep]
In this case, we have an experienced manager supervising a woman who
came to his department as a replacement for the day. This highly skilled
woman knows all the jobs on the line and can do almost all of them. The
supervisor in the first encounter with this worker makes the unmarked choice
in giving orders in the form of a bald imperative. The manager started the
day by telling her to replace Bill so he could go on a break; the request took
the form Go do Bill's job, which she did. Then the manager asked her to
cover one of the few jobs she cannot do on the line because of the heavy
lifting involved. According to the worker, the supervisor knew she couldn't
do it. In fact, he wanted to keep the man who was doing it on the line.
When he gave her the direct order, she had to make the attempt. She
returned, and he sent her to sweep, saying Go sweep. At this point, she
was mildly irritated with him but continued to comply with his orders. His
next request took the form of a question directive, the type of mitigated request
form which on the surface looks far more polite than his previous directives.
He said Do you think you can do Helen's job? The worker took this request
as a highly marked form and a direct insult to her capabilities as a replace-
ment worker. She told me, "Of course I could do Helen's job. How could
he say such a thing to me?" She complied with the request, but on her
way out, she said Get me the committee man. When the union representa-
tive arrived to negotiate, the foreman agreed he had been out of line and gave
her a formal apology.
An Unintentional Marked Choice

Example [1] shows a supervisor making a deliberately marked choice, which
the worker took as unnecessary sarcasm and throwing his power into the
exchange. His remark certainly had the effect of challenging the rights and
obligations set between the conversational participants, with a visit from the
union representative as the result. In contrast, we find some supervisors
making linguistic choices which from their standpoint would be unmarked;
in fact, these choices would be unmarked in the general speech community.
However, in the context of the factory, the choices are perceived by workers
to be marked.
[2] Conversation between new supervisor and experienced worker.
Supervisor: "Are you going to be here tomorrow?"
Worker: "I hadn't planned on it."
[Supervisor leaves abruptly and returns with committee man]
Supervisor: "You have disobeyed a direct order."
Worker: "What was it?"
Supervisor: "To come back to this department and work here tomorrow."
Example [2] is a conversational sequence between a new supervisor and

an experienced worker who was told he would be serving as a replacement
in a department for three days. At the end of the third day, the supervisor
came up to the worker and asked, Are you going to be here tomorrow? When
the worker replied I hadn't planned on it, the supervisor left angrily to return
with the committee man. She said to the worker, You have disobeyed a direct
order. The worker said, What was it? The supervisor believed she had told
the worker to be back in the department for a fourth day; her question direc-
tive was, of course, ambiguous and the worker took it as an information
question rather than a directive. The unmarked choice on the factory floor
would have been Be back here tomorrow.
When asked to give examples of ineffectual management language, workers
talked about supervisors who used polite or ambiguous directives which ac-
tually sound a lot like the ones attributed to Henry Ford in the early days of
the auto industry. These supervisors were given nicknames like "Mr. Rog-
ers" and/or "The Lamp Post" and were generally regarded with contempt by
the workers.
Exploiting Ambiguity
Although the misunderstanding in example [2] was due to a true commu-
nication breakdown between supervisor and worker, interviewees gave other
examples in which workers exploited the ambiguity of mitigated directives to
ignore ones they felt were too indirect. Example [3] features such an exchange.
[3] Conversational sequence between supervisor and repair person.
Supervisor: "Aren't you going to fix that?"
Worker: [Ignores supervisor]
Supervisor: "I'm telling you to do that."
Worker: [Makes repair]
In this case, the worker chose to interpret "Aren't you going to fix that?"
as an information question rather than a directive. He waited until the super-
visor produced the unmarked directive for the factory setting, the direct order
I'm telling you to do that. Workers told me the indirect requests in the form
of questions or hints were sometimes ignored.
In an effort to learn more about workers' perceptions of marked direc-
tive choices, I included the following problem on the union workers' ques-
tionnaire:
Suppose you had a supervisor who said, "I wonder if you could possibly do this
job." How would you evaluate his or her effectiveness? Do you think workers
would comply with this type of directive?
Twenty percent of the workers commented on the ambiguity of the request
and stated that their effectiveness rating and their compliance would depend
on their relationship with the supervisor. A forty-two-year-old male wrote,
"A good supervisor who has a rapport with workers can use this approach.
If a supervisor has to rely on positional power, this approach will not work."
A male zone committee man wrote, "It depends on if it is a genuine question
or if after you answer 'no' you know a direct order will follow."
Forty-two percent of the respondents reacted negatively to the if you could
request. They labeled a supervisor who used such a directive form as ineffec-
tual and believed workers would not comply with the request. A forty-one-
year-old female noted two possible negative interpretations. She wrote, "There
is the negative of if you could possibly which could be taken as an insult,
but the question also sounds like the supervisor is scared to death to ask
the question." Indeed, some workers responded as did the worker in exam-
ple [1], taking such a request as a personal insult. A forty-six-year-old male,
wrote "I might be defensive in thinking he was questioning my ability to
complete the job." Another group viewed such a request as a sign of weak-
ness. A male electrician wrote, "No leadership, workers would not com-
pletely respect them." A final group assumed that such a request would
only be for work out of one's job classification—something definitely to
be avoided. A thirty-seven-year-old woman wrote, "Being that this would
be a rare request, my first response would be 'sorry, I don't do favors for
engineers.' This practice could come back to haunt you in added duties."
Interestingly, 37% of the workers reported that they would be very pleased
to be addressed in such a way by a supervisor. A female worker with thirty
years of experience on the assembly line wrote, "For me, it would be very
effective. I would do more for him or her when asked in that manner." A
male wrote that the supervisor would be effective because "it's not an order,
it's a request to do your best effort." Another said, "I feel more workers
would respond to this type of language because it is working with the work-
ers and our jobs are important to us."
DISCUSSION
Workers describing actual encounters with supervisors and those respond-
ing to hypothetical situations on the questionnaire showed a great deal of
insight in discussing the alternative implications of language choices. Work-
ers described minor differences in syntactic and lexical choices which could
result in major differences in interpretation. Also, many workers wrote at
length about possible interpretations based on the differences in the super-
visor's perceived assumptions and motivations. It is evident that given social
identity features (e.g., age and gender) and given linguistic choices do not
mechanistically determine listeners' interpretations. Myers-Scotton writes:
"While norms largely determine the interpretations of choices, speakers, not
norms, make choices" (1993: 102). It is clear from this data that perceived
attitudes of the speakers play a large part in the listeners' evaluations of pos-
sible meanings.
The markedness model is characterized as a universal cognitive structure

in which the "actual assignment of readings of markedness of codes is only
developed in reference to a specific community through social experiences in
interactions there" (Myers-Scotton 1993: 79). The social competence to com-
municate effectively in the factory requires workers to activate their knowl-
edge of the way in which factory discourse differs from the other discourse
worlds which they inhabit. The workers' intricate responses reveal cognitive
calculations based on their underlying knowledge of marked and unmarked
behavior in the factory.
Although all the conversational maxims in the markedness model are af-
fected by the discourse structure of the factory, one which is particularly
impacted is the deference maxim: "the switch to a code which expresses def-
erence to others when special respect is called for by circumstances" (Myers-
Scotton 1993: 147). There appear to be very few settings in the factory in
which the deference maxim is activated. The factory discourse patterns re-
vealed in this study are the result of the historical development of talk be-
tween managers and workers where management and union have maintained
a basically adversarial relationship for many years. I would argue that the
lack of politeness markers and the bald imperatives used by both manage-
ment and workers can be indexed with the attempt of each group to assert
and maintain equal power. Unmitigated talk is the unmarked choice for both
management and worker in an effort to maintain the status quo in rights and
obligations established in rigorous union-management negotiation and cod-
ified in the omnipresent union-management contract (UAW 1993). Any de-
viations from the unmitigated code choices are perceived as marked and are
interpreted as an abnegation of the power stance expected from both the
supervisor and the workers.
These data offer implications for training employees in the most effective
discourse patterns, particularly new managers coming from outside the fac-
tory culture. It is clear that managers need to be prepared for unmitigated
language from workers as the unmarked choice. They also need to know the
expectations of workers regarding the unmarked choices for directives from
them. On the other hand, mastering scripts does not automatically make a
respected supervisor. Workers described the best managers as those who used
good judgment in managing workers by giving clear orders on the one hand
and treating workers with respect on the other. The best managers were those
who avoided conflict talk altogether. One woman described her favorite su-
pervisor as a man who "let's [sic] his line run itself. He sees potential prob-
lems and deals with them in advance."
It would be beneficial to see how talk on the shop floor differs in two
other types of auto factories. What is shop talk like in factories where work-
ers lack the protection of the union, particularly in these days of uncertainty
about job security? Can we expect more politeness and mitigated talk from
workers in such factories? Will the unmarked choices for management and
worker show a more asymmetrical relationship? An alternative venue is the

factory run under Japanese management. How does the cooperative/team
approach brought in to Japanese-owned plants in the United States play out
on the factory floor? Does it result in both workers and supervisors being
less direct and more polite? Do the unmarked choices in these factories re-
veal mitigation on the part of both management and worker? Contrasting
shop talk in these plants with that in unionized Michigan auto production
facilities will constitute the next phase of this research.
ACKNOWLEDGEMENT: Thanks to Kevin Beard, Barbara Eastman, Loren Shipprett, Gary Ber-
nath, the members of UAW Local 605, Lansing, Michigan, and ail the other workers who pro-
vided the insights that made this chapter possible.
REFERENCES
Atkinson, John Maxwell, and Paul Drew. 1979. Order in the court: the organisation
of verbal interaction in judicial settings. London: Macmillan.
Austin, John L. 1962. How to do things with words. New York: Oxford Univer-
sity Press.
Bernsten, Janice. 1995. "You know you're in trouble when she swears": female and
male managers on the automobile factory floor. Paper presented at the American
Association of Applied Linguists Annual Meeting, Los Angeles.
Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: some universals in lan-
guage usage. Cambridge: Cambridge University Press.
Cameron, Deborah, Elizabeth Frazer, Penelope Harvey, M. B. H. Rampton, and Kay
Richardson. 1992. Researching language: issues of power and method. London:
Routledge.
Cavendish, Ruth. 1982. Women on the line. London: Routledge and Kegan Paul.
Chinoy, Eli. 1992. Automobile workers and the American dream. 2nd ed. Urbana:
University of Illinois Press.
Collinson, David L. 1992. Managing the shopfloor: subjectivity, masculinity and
workplace culture. Berlin: Walter de Gruyter.
Conley, John M., and William O'Barr. 1990. Rules versus relations: the ethnography
of legal discourse. Chicago: University of Chicago Press.
Coulthard, Malcolm. 1985. An introduction to discourse analysis. London: Long-
man.
Drew, Paul, and John Heritage. 1992. Analyzing talk at work: an introduction. In
Talk at work: interaction in institutional settings, ed. P. Drew and J. Heritage, 3-
65. Cambridge: Cambridge University Press.
Employe [sic] Handbook, n.d. Truck and Bus Group, Flint Assembly Plant, G-3100
Van Slyke Road, Flint, Mich.
Ervin-Tripp, Susan. 1976. "Is Sybil there?" The structure of some American English
directives. Language in Society 5: 25-66.
Fisher, Sue, and Alexandra Dundas Todd (eds.). 1983. The social organization of
doctor-patient communication. Washington, D.C.: Center for Applied Linguistics.
Form, William H. 1976. Blue-collar stratification: autoworkers in four countries.
Princeton, N.J.: Princeton University Press.
Gartman, David. 1986. Auto slavery: the labor process in the American automobile
industry, 1897-1950. New Brunswick, N.J.: Rutgers University Press.
Hamper, Ben. 1986. Rivethead: tales from the assembly line. New York: Warner Books.
Kearns, Josie. 1990. Life after the line. Detroit: Wayne State University Press.
Levi, Judith N., and Anne Graffam Walker (eds.). 1990. Language in the judicial process.
New York: Plenum.
Miller, Gale. 1994. Toward ethnographies of institutional discourse. Journal of Con-
temporary Ethnography 23: 280-306.
Mishler, Elliot G. 1984. The discourse of medicine: dialectics of medical interviews.
Norwood, N.J.: Ablex.
Africa. Oxford: Oxford University Press.
. 1995. What do speakers want? Codeswitching as evidence of intentionality in
linguistic choices. SALSA II: Proceedings of the second annual symposium on lan-
guage and society-Austin, ed. Pamela Silberman and Jonothan Loftin, 1-17. Aus-
tin: University of Texas Linguistics Department.
Nash, Al. 1976. Job satisfaction: a critique. In Auto work and its discontents, ed. B.
J. Widick, 61-88. Baltimore: Johns Hopkins University Press.
Nevins, Allan, and Frank Ernest Hill. 1954. Ford: the times, the man, the company.
New York: Scribner's Sons.
North American Truck Platforms/Flint Assembly. 1994. Local Agreement between
NATP Flint Assembly and Local 598 United Autoworkers.
United Auto Workers. 1993. Agreement between UAW and the General Motors
Corporation.
. n.d. Heritage of struggle: the historical setting for the Flint sitdown strikes.
Detroit: UAW Education Department.
West, Candace. 1984. Routine complications: troubles in talk between doctors and
patients. Bloomington: Indiana University Press.
IV
Stylistic Choices and

Second-Language Acquisition
11
"Not quite right": Second-Language
Acquisition and Markedness
MARY SUE SRODA
N o ONE CONTESTS the assertion that acquiring a second language as an

adult is an arduous and extremely complex process. This chapter con-
siders one part of that complex process—the acquisition of native-like use of
a second language, also known as pragmatic ability or pragmatic competence.1
I propose that reanalysis through the markedness model (Myers-Scotton
1993b) yields an explanatory account of the complexity of second language
use through a simple cognitive mechanism underlying complex socioprag-
matic effects.
The nature of language use is so flexible and unpredictable that it is almost
impossible to exhaustively describe what is "appropriate" language use in
any situation. That is, the fundamental issue of language use for teachers of
second-language (L2) learners is, "why do we say what when?" Teachers of
a second language often get questions from students in the following form:
"What should I say in situation X?" At first, the answer for highly formulaic
situations seems easy. In the case of purchasing stamps at the post office, for
example, a teacher might give an answer such as / would like a book of 20
stamps please. However, there are many other lexical and morphosyntactic
variations that are just as acceptable as in examples [l]-[4]:
[1] Do you have books of 20 stamps?
[2] Could I have a book of 20 stamps?
[3] A book of 20 stamps, please.
[4] 20 stamps, please.
The set of allowable, appropriate responses is even larger in situations that
are less formulaic. In fact, in everyday communication, situations in which
1. This ability contrasts with other linguistic abilities, the ability to produce well-formed ut-
terances, for example.
195
196 STYLISTIC CHOICES AND SECOND-LANGUAGE ACQUISITION
there is only one appropriate lexical or morphosyntactic code choice are the
exception rather than the rule.
That is, instead of there constantly being one "correct" answer, there is
often a range of "right" answers, some answers that are "not quite right,"
and other answers that would not be considered acceptable at all. Following
the post-office situation, examples [5] and [6] could be considered less "right"
than [l]-[4].
[5] Give me 20 stamps.
[6] Pardon me, would you be so kind as to allow me to purchase a book of 20
stamps?
Example [7], on the other hand, would not be considered acceptable at all for
this type of service transaction:
[7] 20 stamps, now!
One point from which to start examining this phenomenon is to examine
what is meant by right in light of the post-office example; a code choice that
is right would mean that it is appropriate for the status of the speakers and
appropriate for what is communicated. Even though it violates no structural
rules of English, example [7] is not right in two ways: First, bald imperatives
are not normative (unmarked) for this type of service transaction; second,
inclusion of the lexeme now stands in contrast to the typically unvoiced as-
sumption that the transaction would normally take place immediately after
the request is made.
Identifying code choices that are absolutely not right or appropriate for
a given situation such as the post-office transaction, is quite easy. The type
of language that is harder to discuss involves examples that fall into the "not
quite so right" category as seen in examples [5] and [6]. Unfortunately, lan-
guage use that is not quite so right is exactly the type of language use that
teachers of nonnative speakers of English often encounter. Students say or
write things in their L2 which strike their teachers, as one colleague put it,
as "odd English." That is, many times, especially in the case of intermediate
or advanced nonnative speakers, what the students say or write is not
grammatically wrong in the sense of violating the rules of well-formed En-
glish but, rather, odd in the sense of not being the way a native speaker
would say it. In such cases, the effect of the odd language choices L2 learn-
ers make can vary in degree from simply emphasizing the fact that the learn-
er is not a native speaker of English to completely obscuring the learner's
intended meaning.
The issue of differences in the sociopragmatic effects of various morpho-
syntactic or lexical choices in L2 acquisition and the difference in the inter-
pretation and/or perception of these structures by native and nonnative speak-
ers of English is well studied in the field of L2 acquisition (see Blum-Kulka
et.al. 1989; Kasper and Blum-Kulka 1993, among others). But Turner (1995,
Second-Language Acquisition 197
1996) points out that these studies have limited implications for pedagogy as
applied to natural language use.
The goal of this chapter is to look at the acquisition of pragmatic compe-
tence by L2 learners within a theoretical framework that provides an expla-
nation of the variations in the morphosyntactic/lexical choices in natural
language use such that teachers of English as a Second Language may have
a new way to talk to L2 learners about how the choices they make are (and
are not) like the choices native speakers make. That is, this chapter shows
how Myers-Scotton's markedness model provides an explanatory framework
for the range of linguistic choices that L2 learners can make which are "not
quite right."
THE MARKEDNESS MODEL
In the last ten years much research was done concerning the development of
theories and models of communication that involve the intentionality of speak-
ers when communicating and their cognitive awareness of linguistic choices.
Such a model is Myers-Scotton's markedness model (MM; 1993b) propos-
ing sociolinguistic motivations for codeswitching among multilingual speak-
ers. To limit the MM to simply motivating choices made in codeswitching
would be a mistake, however, as it is applicable to all linguistic choices that
speakers make.
A detailed description of the MM is included in chapter 2 (Myers-Scotton
in this volume). The discussion here is intended to emphasize the parts of the
model that specifically concern this analysis. The idea behind the MM is that
speakers make linguistic choices to negotiate social distances in linguistic
interactions. For each interaction, an expected, normative, unmarked rights
and obligations set (RO set) exists for each speaker. However, RO sets are
community specific.
During an interaction, speakers either endorse or reject the expected RO
set for the interaction by making linguistic choices that are or are not compat-
ible with that RO set. A choice that is compatible with an RO set is unmarked
for that particular set, and a choice not compatible with the RO set is marked
for that RO set. If a speaker makes a marked choice, he or she is making a bid
for another RO set. For example, in an employer-employee discussion, an
employer may use a more casual style of speech and lexical choice to indicate
a willingness to switch from a very formal employer-employee RO set to a
more informal RO set that colleagues would use, thus making a bid for a
decrease in social distance.
THE MARKEDNESS EVALUATOR
The markedness evaluator (ME) is the mental faculty involved with assign-
ing assessments of markedness to linguistic choices. On this faculty, Myers-
Scotton writes:
198 STYLISTIC CHOICES AND SECOND-LANGUAGE A C Q U I S I T I O N
This metric [evaluator] is part of the innate cognitive faculty of all humans. It
enables speakers to assess all code choices as more or less unmarked or marked
for the exchange type in which they occur. A critical distinction is that, while the
metric [evaluator] is a cognitive structure and therefore a universal, it underlies an
ability which is particular. The ability, consisting of the actual assignment of
readings of markedness to codes, is only developed in reference to a specif-
ic community through social experience in interactions there. Thus, while it is a
universal feature of language use that all choices are interpreted in terms of their
markedness, one can speak of the markedness of a particular code only in refer-
ence to a specific speech event in a specific community. (1993b: 80)
Speaking of choices as marked or not assumes that they take place in a norma-
tive framework . . . while speakers are innately equipped with a markedness
metric [evaluator], they only make actual readings through experience with
language use in that community. They then view the codes in their community's
repertoire as more unmarked or marked, according to community norms, as the
index of the unmarked RO set between certain participants in a given talk ex-
change. This argument implies that conversations are more or less convention-
alized, that speakers have some sense of the unmarked "script" or "schema" for
them. (1993b: 109)
The analysis of L2 morphosyntactic lexical choices is concerned with two

key ideas from the previous passages: "the actual assignment of readings of
markedness to codes" and the "normative framework" within which the
readings of markedness make sense. I propose that both concepts are part of
the markedness evaluator and provide explanations for why nonnative speak-
ers (NNSs) have difficulty using an L2 the way that native speakers (NSs) use
the same language. The innate non-language-specific cognitive mechanisms of
the MM, which are in place before a speaker acquires even one language, under-
lie sociopragmatic competence in any language. As a basic cognitive faculty
of speakers of any language, the markedness evaluator (ME) does not change
when a speaker begins to acquire and use an L2. Instead, during L2 acqui-
sition, all the things that serve as input to these abilities and mechanisms that
allow NNSs to make choices reflecting their communicative goals are what
change. In this sense, then, part of what an L2 learner acquires in becoming
a native-like L2 communicator is not only information about the RO sets
specific to that L2 community but also community-specific "readings," as it
were, about which lexical and morphosyntactic structures index those sets.
In short, I propose that the examples of "odd," awkward, or ineffective
language can be redefined as choices marked for a given RO set and, there-
fore, interfere with or confound communication. Yet they represent a speak-
er's "best attempt" at a code choice that reflects his or her communicative
intention. As markedness is delineated in the MM, the ways in which an
utterance can be marked are complex but not infinite, and this markedness
is always defined relative to some sort of unmarked starting point.
My hypothesis is that many of the seemingly infinitely complex problems
of awkwardness and ineffectiveness that occur in student writing of Standard
American Academic English can be accounted for within the structure of the
MM. I predict that both NS and NNS writing may show these errors but that
there will be differences in the type of errors that NSs and NNSs make. I
further contend that the way the MM accounts for these differences provides
clues for teachers of NNSs to help their students become more native-like in
their language use.
METHODOLOGY
The data from composition writing for this chapter were gathered from NSs
and NSSs of English from a one semester, first-year composition class (En-
glish 101) at the University of South Carolina. Normally, there are special
sections of English 101, called English 101B, which are solely for interna-
tional students enrolled at the university. Yet, through a computer error at the
registrar's office, four native speakers of English were allowed to enroll in a
section of 101B in the fall of 1994. The corpus consists of drafts and in-class
writing collected throughout the semester from thirteen NNSs and four NSs.
The example from the corpus that brought the possible applications of
the MM for the L2 acquisition of pragmatic competence to my attention was
an essay from a female student, Miko, a native Japanese speaker. Miko was
beginning her freshman year at the university after having spent one year as
a student in the university's intensive English program for international stu-
dents. From her writing on the first day of class, she expressed a love of
writing but also recognized the differences between communicating in Jap-
anese and English. She expressed frustration common to many students who
have an intermediate-level command of English in writing: "I tried to get
used to American writing style, but it was very hard for me."2 However, one
particular paper of Miko's, an assignment asking students to write a narra-
tive essay, brought into relief the way the MM may help explain the devel-
oping sociopragmatic ability of L2 learners.
In the first draft of her narrative essay, Miko told the story of losing her
wallet one afternoon while at the university. After a frenzied search of places
she had been shortly before she noticed the wallet was missing, she remarked,
"I wanted to call my boyfriend, but I didn't have even 25 cents."
The possible applications of the MM can be seen in her use of internal
dialogue to decide what to do then while standing in front of the campus
student activity center:
"If I ask somebody, will he or she give me money?" There was a lot of people there.
"Who can I ask?" I've never asked somebody for money like a homeless person
in my life. I was scared. I was thinking of a polite English sentence in my Brain.
"Could I have your 25 ?" "Is that correct?" I had no idea. I was totally lost.
2. All examples from student writing are presented exactly as they occurred, including all
nonstandard punctuation, spelling, and grammar. Except when indicated otherwise, examples
come from writing that students did not edit or revise.
Another English sentences came to my mind. " Excuse me, today I lost my wallet.
I'd like to call my boyfriend, but I don't have money. Can I borrow money?"
This time, kind of long sentences came to my mind. It sounded ok for me.
However, if the person say No, what should I do?
Because she could not decide about the best way to ask a stranger for money
for the pay phone, Miko never asked anyone. In her essay, Miko came up
with two possibilities for asking for a quarter: Could I have your 25 ? and
Excuse me, today I lost my wallet. I'd like to call my boyfriend, but I don't
have money. Can I borrow money? Either choice may have resulted in Miko
achieving her communicative goal, getting a quarter from a stranger, but Miko
didn't just want to get a quarter, she wanted to ask the right way. Stated in
the terminology of the MM, Miko could not make a code choice that she
thought was unmarked for the RO set she considered to be in place when
asking a stranger in English for money to make a phone call.
Ironically, the second option Miko proposed, Excuse me, today I lost my
wallet. I'd (ike to call my boyfriend, but I don't have money. Can I borrow
money?, was relatively unmarked and probably would have resulted in some-
one understanding what she needed and giving her a quarter to call.
MEASURING UNMARKEDNESS
I felt that augmenting my intuitions about what constituted an unmarked choice

in Miko's situation should be possible by designing a task to elicit the intuitions
of other speakers of English. My prediction was that comparing the responses
should yield quantifiable characteristics of unmarked choices consistent with
the MM. Once the characteristics of the unmarked choice were described, the
unmarked choice was compared to NNS responses to the same task.
The eventual goal of this chapter is to apply the operationalized descrip-
tion of unmarked choices from the elicitation task to the problems in marked-
ness that L2 learners have in their writing to help them understand precisely
how their linguistic choices are not native-like.
ELICITATION TASK AND RESPONSES
Every speaker has intuitions about markedness, but often explicitly discuss-
ing those intuitions can be very difficult. To help average speakers avoid this
difficulty, I developed the task that described a situation similar to the one
Miko described in her narrative essay. Subjects for this task were NSs of English
enrolled in two sections of English 101 (n - 30). For purposes of comparison,
the same task was given to students enrolled in one section of 101B contain-
ing only NNSs of English (n = 21).
All students were given a blank index card, told to put their native lan-
guage at the top, and asked to write their responses to the following situation:
One day you drive your car to school and lock your keys inside. You can call
home to have someone bring your spare set of keys, but there is only a pay
phone and you have no money. You have to ask a stranger for the money to call
. . . what would you say?
Two NS responses and one NNS response were excluded from this analysis
because the subjects wrote down summaries of their communication strate-
gies as opposed to writing the actual words they would use; that is, 7 would
try to politely ask to borrow a quarter versus Excuse me, I locked my keys in
my car. Examples [8]-[10] and [11]-[13] illustrate responses which were
included in the analysis from the NS and NNS groups, respectively.
NS Responses
[8] Excuse me sir, but I locked my keys in my car and I don't have any change.
Would you happen to have some change I could have.
[9] Excuse me do you have a quarter so I can call home? I have locked my keys in
my car. I have some change in my car and will give it back to you when my car
is opened.
[10] Excuse me. I've locked my keys in my car and need to call someone. May I please
borrow a quarter for a phone call?
NNS Responses
[11] Would you please give me a quarter to make a phone-call, 'cause I've forgotten
my car's key and I want my family to pick me up.
[12] Sorry to bother you, but I really need to borrow a quarter. See I locked my keys
in my car and I just need to phone a friend to come pick me up.
[13] Excuse me sir. I have a slight problem. I wonder if you could help me. I seem to
have locked my keys in my car and I don't have any money. May I please borrow
a quarter so I can call home.
RESULTS OF "LOCKED CAR" TASK
As seen in Table 11-1, the responses made by the NSs and NNSs of English
do not on the surface seem very different. Responses in both groups varied
from the brief to the verbose with responses from the NNS group spreading
Table 11-1
Analysis of Responses
Native Speakers Nonnative Speakers

(N = 30) (N = 21)
Briefest response 16 words 11 words

Longest response 41 words 63 words
Mean length of response 25.9 words 27.8 words
Median length of response 29 words 37 words
out toward the two extremes more than the NS group. The striking differ-
ence between the two groups in this type of analysis is a difference in the
median of each group, but this difference is mostly due to one extremely long
response from one of the NNS subjects. If that one very long response is not
counted, the next longest response is forty-two words and the median for the
NNS group drops to twenty-seven words. Even though NNSs tended to use
slightly more words in their responses, the variation of length of response in
both the groups indicates that the number of words is not particularly re-
vealing in terms of distinguishing features of unmarked responses.
CODING THE "LOCKED CAR" DATA
The problem still remains of how to describe these responses such that an
unmarked NS response for this situation can be characterized and compared
to the NNS responses. To do this, the responses were analyzed in terms of
the information they contained. Following the model of communication pro-
posed in Sperber and Wilson (1995), one can define the information content
of the responses in terms of the effect that a piece of information has on the
assumptions of a hearer. Sperber and Wilson describe assumptions of a hear-
er as thoughts treated by the individual as representations of the world.
When linguistic input is processed, it has the effect of causing the hearer
to construct a new assumption, strengthen or weaken existing assump-
tions, or abandon existing assumptions. I propose that linguistic input
can be divided into "information features." For example, all responses
contained an information feature that I call ASK FOR MONEY. When the hear-
ers process the information feature of ASK FOR MONEY, they construct a new
assumption: This person is asking me for money. Describing the informa-
tion in an utterance by the effect that it has on the assumptions of the
hearer provides a way to compare utterances that are, in terms of mean-
ing, roughly equivalent. In fact, I propose that this method is the only way
to compare such utterances because different combinations of lexical items
can have the same social effect. For example, no two NSs chose the same
way to ask for money, as the following nine randomly selected NS examples
help illustrate.
[14] May I please have a quarter
[15] ... and need some change to call home, do you have some I could have?
[16 Could you please spare me a quarter.
[17] I was wondering if you had a quarter I could use for the pay phone.
[18] ... but I have no money to use the pay phone. Could you please help me out?
[19] Would you happen to have some change I could have.
[20] do [you] happen to have a quarter
[21] do you think you can spare a quarter
[22] could you lend me a quarter
All the information features in the request introduce or modify assumptions.

For instance, if the overall request contains a feature that mentions specific
use for the money in addition to the ASK FOR MONEY feature, the assumption
introduced by the ASK FOR MONEY feature is modified to something like this
person is asking me for money to make a phone call. The assumptions intro-
duced or modified by information features serve as input to the ME that
produces readings about the relative markedness of the information features.
The information features that appeared in the "locked car" data are de-
scribed next, illustrated with examples from the NS data:
POLITE OPENING. This category includes instances of opening the request
with a polite attention-getter such as excuse me or excuse me ma'am or
sir, although other alternatives such as bare honorifics like Sir? occurred.
Other examples of Polite Opening in the data were could you help me?,
I'm sorry to bother you, and could I have a second of your time?
ASK FOR MONEY. This represented the actual request for the money to
make the call. Responses varied widely, as examples [14]-[22] discussed
earlier show.
NO KEYS. This category included any mention that the subject has locked
the keys in the car such as I've just locked my keys in the car, I seem to have
locked my keys in my car, or I acidently [sic] locked my keys in my car.
USE FOR MONEY. This category represents all mention of what the subject
will do with the money such as so that I can call home, for a phone call,
and / need to call home for an extra set of keys.
NO MONEY. Items in this category included all mention of why the sub-
ject has no money and thus provides the stranger with a reason for asking
for money for the phone. Examples include / locked my keys and my
money in my car, I don't have any money for the pay phone, and / don't
have any change.
PROMISE TO GIVE MONEY BACK. This category includes assurances that
the subject will get the money back to the stranger—I'll pay you back and
/'// give it back to you once my car is opened.
THANK FOR TIME. One NNS subject thanked the stranger for taking the
time to listen to the request even if the request were to be to be refused.
Table 11-2 shows the breakdown of information categories by speaker
group. As seen previously, not all requests included all information fea-
tures.3 The only information feature that all subjects in both groups included
3. For the purposes of this analysis, I leave further theoretical explication as to how informa-
tion features may be cognitively represented and stored for future study. However, I do see strong
similarities between what I consider to be the content of an information feature and the semantic-
pragmatic feature bundles of the matrix language frame model described in Myers-Scotton
([1993a]1997) and Myers-Scotton and Jake (1995).
Table 11-2
Analysis of Responses by Mention of Information
Native Speakers Nonnative Speakers

N = 30 N = 21
Type of Response Number % of N Number % of N
Polite opening 24 80.0 14 67.0

Ask for money 30 100.0 21 100.0
Mention keys 25 83.3 16 76.1
Use for money 25 83.3 21 100.0
No money 7 23.3 2 9.0
Promise . . . money back 2 6.7 0 0
Thank f o r time 0 0 1 4 . 8
in their responses was ASK FOR MONEY, which the final sentence of the in-
structions explicitly told them they were to do—You have to ask a stranger
for the money to call. . . what would you say?
ANALYSIS OF "LOCKED CAR" DATA —NS GROUP
I propose that the frequency of occurrence of certain types of information

categories delineates an unmarked choice for the RO set in the locked car
scenario. Thus, the unmarked choice for this scenario can be described as ASK
FOR MONEY, USE FOR MONEY, and MENTION KEYS. I did not include the cat-
egory of POLITE OPENING in the unmarked description here because it seemed
more a general characteristic of attention getting when starting an interac-
tion, and several of the NS responses seemed (to me) perfectly fine without it
as seen in examples [23] and [24].
[23] Could I please borrow a quarter to use the phone? I've locked my keys in my car.
[24] I locked my keys in the car. Do you happen to have a quarter so that I can call
home and have my roommate bring the spare set?
A somewhat unexpected outcome of this analysis was that the features of
the unmarked choice for the car key scenario did not come in any special
order. Thirteen NSs started their response with an explanation about locking
the keys in the car, but seventeen began with a request for the money to call.
Five NS subjects gave responses that omitted one of the above character-
istics describing the unmarked choice for the locked car scenario. Two speak-
ers did not include USE FOR MONEY, as seen in example [25].
[25] Do you have a quarter I can borrow please, because I locked my keys and my
money in my car.
Even though example [25] explicitly omits one information feature judged to
be required for the entire utterance to be unmarked, I claim that it is still an
unmarked response because the feature USE FOR MONEY is pragmatically re-
coverable by the hearer. That is, the hearer will use assumptions of shared
cultural knowledge, namely that quarters are used in pay phones, to under-
stand that the speaker wishes to make a phone call.
Three NS subjects omitted the feature MENTION KEYS, as seen in examples
[26] and [27].
[26] do you think you can spare a quarter so I can make a phone call? I haven't got
any money.
[27] Could you help me? I need a quarter to call home, can I get it from you?
Responses that omitted MENTION KEYS are marked because the hearers of
examples [26] and [27] cannot pragmatically recover the situation that caused
the speakers to ask for the quarter. Thus, the reason for the request was not
communicated by the speaker contrary to the element of the RO set that re-
quires unmarked requests to be justified to maintain social distance. The social
distance between strangers can be maintained in spite of the imposition of the
request if the speaker provides a "good" reason, as defined by the communi-
ty. In this case, the emergency of locking one's keys in the car is sufficient, but
the omission of this information prevents the hearer from knowing this.
In contrast to responses that were marked because of information that
was omitted, two NS subjects gave marked responses by providing extra in-
formation by promising to pay the money back, as seen in example [28].
[28] Could you please lend me a quarter? I've locked my keys in the car and I need
to call home so that my mother can bring me my spare keys. I'll pay you back.
Even though half of both subject groups (fifteen NSs and twelve NNSs) ac-
tually phrased the request as borrowing a quarter, this request is not usually
seen as solicitation of a loan.4 Therefore, the specific offer to pay back the
quarter is marked.
ANALYSIS OF "LOCKED CAR" DATA —NNS GROUP
So far, this analysis focuses on the responses provided by the NS subjects to

determine what characterizes the unmarked choice in the locked car situa-
tion. How do NNS responses compare with respect to the unmarked choice?
Examples [11]-[13] and Table 11-2 show some differences between the two
groups, but few of the differences seem consequential. A great deal of simi-
larity between responses of the two groups is not surprising, given the sim-
plicity of the task and the intermediate level of the L2 learners. The NNS also
showed the same varied order of information features as the NS group—ten
NNSs mentioned the keys first, eleven NNSs mentioned the need of a quarter
first. However, out of the group of twenty-one NNSs, five gave marked re-
sponses and the analysis of the NNS marked responses is revealing.
4. I am indebted to Amy Albert for her observations and comments about this.
In addition to NNS responses as a group being relatively similar to the

unmarked choice, four NNS subjects gave marked responses like those made
by the NSs when omitting one of the information features (see earlier discus-
sion). Example [29]-[31] illustrate the NNS responses omitting the informa-
tion feature MENTION KEYS:
[29] Excuse me, but could you give me some quarters? I need to call my home.
[30] Can you give me some quarters to make a phone call?
[31] Excuse me, would you mind giving a quarter to make a call? Because I don't
have any coin.
An interesting aspect of example [29] is that it is the only example in the
corpus from a Japanese speaker. Example [29] is also the only response in
which the subject wrote out a response in her native language and then trans-
lated it into English. This example is somewhat similar to Miko's first formu-
lation of a request, Could I have your 25 ? It is possible that part of the
difference between Miko's two formulated responses, Could I have your 25 ?
and Excuse me, today I lost my wallet. I'd like to call my boyfriend, but I
don't have money. Can I borrow money? may have been that for the first
response Miko may have been relying on features of unmarked choices from
her native language and culture.
When comparing NS and NNS responses that are marked due to the
omission of an information feature, a question of why any of the native speakers
omitted an information feature remains. One possible explanation is that a
few native speakers in the group did not have a complete understanding of
the unmarked features indexed by the locked car RO set.
The goal of soliciting so many responses in this task is to demonstrate
that, in the face of so much lexical and structural variation, features can be
extracted which describe a normative response and identify its associated
RO set. This claim is not to imply, however, that all individuals will make
that normative response.
Example [32] is the only marked response from the NNS group that is
marked because of additional information features.
[32] Excuse me sirl May I take of your precious time a minute. I'm affraid [sic] I
locked my keys inside my car if you could possibly give me a quarter so that I
can make a call to my wife who would come and bring me the other keys from
home, if you cannot I understand and / thank you for listening to me.
It was not only the longest response but the most excessive response in terms
of including eight politeness markers (indicated in italics). What makes
example [32] marked is that whereas politeness markers are usually un-
marked in indexing social distance or an imposition, the politeness should
be proportional. That is, in a culture that looks favorably on citizens helping
one another with problems and perceives a quarter as a very small amount
of money, there is not a great degree of social distance or imposition in
the locked car situation. Therefore, not only may incorporating a great
degree of politeness be interpreted as marked, it may be interpreted as a sign
of sarcasm.
The claim of the MM is not that NSs always try to make code choices that
they consider to be unmarked, but that they have the ability to make marked
and unmarked choices in interactions and through these choices do "social"
work. Recall that some NS choices were marked in various ways. For exam-
ple, speakers who offered to pay the quarter back may have been making a
bid for a slightly different RO set in which they would be on a more equal
footing with the stranger because they were explicitly not asking for the stranger
to give them the money outright. Or, perhaps these subjects supposed that
their offer to pay back the money would not be taken seriously but would
emphasize the paucity of the sum they had asked for and thus increase their
chances of being successful. In either case, it is possible to intuit a purpose
behind these marked choices regardless of whether the marked choice helps
the subjects achieve their communicative goals.
Second-language learners want to do the same: make marked and unmarked
choices when they wish to. However, the instances of odd English, which I
now define as marked choices that confound communication or seem to com-
municate information that does not seem intentional, indicate that L2 learn-
ers do not have the same sort of ability that native speakers do over their
linguistic choices.
In this section I tried to show that it is possible to come up with intuitive
and empirical descriptions of unmarked choices while still allowing for the
lexical and structural variations that occur in natural speech. The ability of
the MM to describe normative choices while allowing for surface variation
can be of use in explaining to L2 learners how their choices are marked. That
is, the problems L2 learners have making unmarked choices is of primary
importance. Their goal is not only to make unmarked choices when they
communicate but to have native-like marked choices as part of their linguistic
repertoires. Without the ability to have the option of marked choices in their
linguistic repertoires, L2 learners will never acquire a native-like pragmatic
ability. The next question to consider, then, is the following: Can the MM
be applied to natural language use and help nonnative learners be more
native-like?
MARKEDNESS FOR NATIVE AND NONNATIVE

SPEAKERS OF ENGLISH
Because, as discussed earlier, the cognitive faculties underlying the MM are

the same in a person's L1 as in her L2, the elements that affect an L2 learner's
ability to produce marked and unmarked utterances have to do with the in-
puts to the process that produce readings about marked and unmarked forms.
Furthermore, because of the allowable variation in unmarked forms, these
inputs can rarely be a closed set of specific lexical items or morphosyntactic

structures. I propose these readings about relative markedness should be
described in terms of information features. Specific lexical items or combina-
tions of lexical items will be associated with information features in an L2
learner's encyclopedic memory or mental lexicon.
In applying the conclusions of the locked car task to spontaneous lan-
guage use, an examination of an L2 learner's ability, or lack thereof, in mak-
ing unmarked choices in natural language use starts with the odd English
examples that are marked. The rest of this chapter examines instances in written
discourse when marked choices made by L2 writers did not work. In such
examples two explanations are possible: Writers make what they think is an
unmarked choice but is not, or, writers are intentionally making a marked
choice that is not serving the purpose they intended. In either case, both choices
are dependent on the L2 writers' having the same understanding as their
intended NS audience of what constitutes an unmarked choice in that par-
ticular situation.
On this view, the elements of the MM as discussed previously (LI commu-
nity-defined RO sets, information features, and specific lexical/morphosyn-
tactic choices) become descriptors around which a principled discussion of
non-native-like markedness is centered. The next section uses examples from
the corpus of writing from NNSs described earlier to illustrate each of the
levels at which marked choices can be unsuccessful—that is, at the level of the
L1 community-defined RO set that indexes certain information features as
unmarked, at the level of the unmarked information features associated with
that particular RO set, or at the level of the lexical/morphosyntactic choices
chosen to reflect those information features.
RO SETS
Essays from L2 students enrolled in an American university provide insight
into the writers' perceptions of the RO sets they believe to be in effect. RO
sets are not linguistic choices—they index information features and provide
input to the ME for markedness readings on all possible linguistic choices,
marked or not. As opposed to other types of RO sets, the RO set of writing
in American academic discourse has the advantage of being well established
and very normative, and bids for other RO sets are generally not acceptable
in academic discourse. Therefore, student academic writing reflects an effort
to generally stay within the linguistic choices indexed by the RO set. For
example, one of the elements that characterizes the academic RO is the estab-
lishment of authority. That is, one of the expectations of the academic com-
munity is to present information in a way in which it is credible. If the writers
are not authoritative sources themselves, the RO set indicates that they should
present evidence or cite authoritative sources to give credibility to the infor-
mation they are presenting. In addition, this expectation of authority indexes
formality in information features and the linguistic choices they underlie.
Example [33] shows the NNS student trying to work within what she per-
ceives to be the required formality of the RO set in place.
[33] Just when I feel about something deeply, the good words seem to come to my
mind one after another, and I pretend to be a very good poet, but actually what
I am writing is nothing but a piece of feces.
The student's use of the clinical term feces, instead of a more colloquial word,
reflects her attempt to stay within the formality required by the RO set in
place. However, her use of a more formal term was not compatible with the
RO set, in spite of its formality, because references to bodily waste in nonlit-
eral ways are almost never appropriate in formal academic writing. Note that
other formal terms, such as excrement, would have been just as unacceptable.
The student seems to realize that different tasks require different code choic-
es, but is not able to make a code choice that works.
Linguistic choices can be marked with respect to the academic RO set in
different ways, as in examples [34] and [35].
[34] For example, I paid $1000 to take psychology 101 and I hate that class. It
doesn't help me with my major; HRTA. Even though I study hard, it is difficult
to get a good grade for me and if I don't pass it, I have wasted the money. I can
buy a car with that money. I feel like I neglect my important classes just to study
for a meaningless Psychology class.
Example [34] does not work because of the hierarchical nature of the RO set
for academic discourse. That is, in addition to the requirement for presenting
information in an authoritative way, there is some social distance between
the writer and the addressee, and the addressee has more status. I claim this
is true even when students are told that the audience for their essays is com-
posed of other college students (as opposed to the professor or a general aca-
demic audience). In the RO set of academic writing, writers behave as if their
readers have higher stature even if this is not true in the real world. Unmarked
linguistic choices must show deference and reflect the degree of social distance
in the author/reader relationship as understood by the academic community.
Therefore, linguistic choices that index solidarity, as in example [32], are
marked. In that example, the student was trying to express her frustration at
certain university regulations. Yet, instead of presenting a compelling argu-
ment for her position, she simply seems to be venting her frustrations to her
audience and appealing to the audience's sympathy—choices that are consis-
tent with an equal degree of social status between author and reader. There-
fore, appeals to sympathy are marked in the academic RO set because audi-
ence has more social status than the author.
Example [35] is marked with respect to the social relationships indexed by
RO sets in a different way.
[35] We would like to involve ourselves in the Travel and Tourism segment of this
booming industry and the title of our research paper gives an indirect hint to
what we want to pursue as a career choice. [The title of the paper was The
Friendly Skies]
In example [35], the oblique reference to the intended title comes across as
coy, raising the question, "Why didn't the authors come right out and say
they wanted to work in the airline industry?" The authors' use of indirectness
makes a bid for a decrease in social distance by requiring readers to work
harder to understand what they mean. The RO set of academic discourse has
requirements of directness consistent with the higher status of the addressee.
For their code choices to be unmarked, authors should be as direct as possible
and, thus, not waste the time and effort of their readers.
This section is not intended to describe exhaustively all the characteristics
of the academic discourse RO set. I am not sure that doing so is useful ped-
agogically or even possible because I have only been able to describe charac-
teristics of RO sets in the context of mistakes in natural language use. Because
they only index linguistic choices, RO sets work more as boundaries than as
predicators of language use—linguistic choices "within" the boundary are
unmarked, and those outside the boundary are marked. The way to get L2
learners to understand where an L1 community sets its boundaries depends
on showing them when their L2 language use crosses those boundaries while
they are trying make their linguistic choices stay inside.
INFORMATION FEATURES
Information features that make up unmarked choices indexed by RO sets

represent another type of markedness problem that L2 learners encounter.
For example, L2 learners could have a native-like understanding of the ele-
ments of the RO set in place but have a non-native-like knowledge of how
certain information features reflect the unmarked choices they are trying to
make, as examples [36] and [37] illustrate.
[36] Some people do this gradually while some do it overnight by getting involved
in an unfamiliar field or by getting involved into something negative.
[37] This question poses a problem to many students and therefore requires ade-
quate thinking.
The issues that [36] and [37] raise have to do with leaving out information
features that are required for the choices to be unmarked. Something negative
in [36] and adequate thinking in [37] do not follow the requirement to be
specific. An information feature of the unmarked choice should be a specific
referent. Note that no requirement for exactly what that referent should be
exists; specific lexical choices depend on what the writers actually meant by
something negative and adequate thinking.
[38] If it [the Writing Studio] is for our benefit, why do most people avoid this free
help?
Example [38] reflects an error in the information feature requirement I call

claim-proof. That is, for a linguistic choice to be unmarked, claim-proof re-
quires the authority of the writer's claims to be supported by some sort of
information about how he or she knows this to be so.
Some information features affecting markedness can appear in writing on
a more global level—across clauses and paragraphs and sometime through-
out the essay, as example [39] shows.
[39] Finally, it's true because it changed not only his life but also affirmed his writ-
ing style.
Out of context, the sentence in example [39] looks perfectly fine except that
it was part of an autobiographical essay about the writer's writing experi-
ences. Referring to oneself in the third person throughout the entire essay
when it is clear that the author and the subject are the same person was
certainly marked.
LEXICAL/MORPHOSYNTACTIC CHOICES
The third type of markedness problem that L2 learners encounter is seen in

the way that information features are reflected in specific lexical and mor-
phosyntactic choices as examples [40] and [41] demonstrate:
[40] Like many other college students I find higher education expensive and like
many others who don't live in SC—twice as expensive
[41] The other day I heard a mad professor saying that we are not in high school.
In example [40], I remarked to the student that she did live in South Caroli-
na—she just wasn't from here. She responded that she had picked live as a
synonym for reside because the school policy allows state residents to pay
lower tuition. In this case the student chose a lexeme that did not communi-
cate the information feature she intended. The same is true in example [41];
the student knew of the semantic relationship between mad and angry, but
the word order he chose unintentionally communicated something else.
NATIVE SPEAKER CHOICES
Although the focus of this chapter is the developing pragmatic abilities of

NNSs, the data collected from the NSs of English were not free from prob-
lems in markedness. Although the NSs seemed to have fewer problems over-
all, I found a least one example of each type in the part of the corpus made up
of writing from the four NSs in the mixed 101B class. The most salient feature
of the markedness problems in the NS writing was that all the choices ap-
peared to be made with clear intentions of the choice being marked; however,
the markedness simply did not work from an academic point of view as seen
in examples [42] and [43]:
[42] [From a persuasive essay that argues against hunting] He is an older buck, this
may be the last season that his hormones will turn him into the aggressive love
machine of his younger days.
[43] I interviewed Mr. Owen in his office in the main building of the high school
which was permeated with trophies, plaques, and certificates exhibiting a life-
time of success.
The tone of the language in example [42] stood out in such sharp contrast to
the overall tone of a persuasive essay against hunting that it came across as
flippant. The student explained that he was trying to be expressive and got
carried away. In example [43] the student explained that she chose the word
permeated because she wanted a really impressive word.
So although it seems that NSs and NNSs overall do make the same types
of mistakes concerning markedness, they make them for different reasons.
Native speakers do not make marked choices because they think those choic-
es are unmarked. Rather, in errors concerning markedness, the NSs inten-
tionally make marked choices to achieve their communicative goals, but these
marked choices are sometimes unsuccessful. Handbooks of academic writ-
ing, which students often use in introductory writing classes, such as the class
from which these data were collected, encourage students to use expressive,
effective language while avoiding "tired expressions and meaningless lan-
guage" (Carter and Skates 1996: 217). The mistakes NSs make often reflect
their unsuccessful attempts to do this without straying too far outside the
norms of academic writing.
IMPLICATIONS FOR PEDAGOGY
The markedness model provides a principled explanation as to why employ-

ing student-generated language is successful as a basis for illustrating and
teaching native-like language use. This intersection of theory and application
is crucial to teachers who need simple explanations for the commonalties that
seem to run through seemingly endless variations in possible linguistic choic-
es. For example, there are countless discussions in the literature of composi-
tion and rhetoric about the influence that cultural knowledge has on lan-
guage use, but the MM operationalizes the interaction between language and
culture in the concept of LI-defined RO sets, the effect that RO sets have on
the selection of information features, and the way that information features
underlie specific lexical items.
The MM is not a predictive framework in the following way: It does not
claim to say when, how often, or under what conditions L2 learners will have
problems with markedness in their language use. However, the MM is pre-
dictive in the sense that it does state that if RO sets index unmarked choices
and an unmarked choice is selected, that unmarked choice will not produce
marked effects. Because they have tacit knowledge regarding how marked-
ness "works" in their native language, native speakers know they have a range
of marked and unmarked options for any communicative task. Nonnative

speakers, even ones who know "a lot of words," are not always aware of
what their options are. From the MM perspective, developing a native-speak-
er-like pragmatic competence means having the same range of choice as a
native speaker.
The corpus of writing from L2 learners and native speakers used in this
analysis contained many examples of marked and unmarked choices that
were successful. Using language that students generate themselves, teachers
can talk about the features of lexical choices, which of these features are
considered unmarked, and how marked choices might be interpreted by the
hearer/reader.
It is impossible to list all the possible "correct" lexical choices that stu-
dents can make in a given situation, be it asking a stranger for a quarter to
make a phone call or writing the opening of a paper for a sociology class. By
the same token, it is impossible to list all the choices that students should not
make in a certain situation. As example [28] shows, a teacher would have to
be psychic to tell students before the fact not to mention bodily wastes in their
academic essays.
One way to discuss the intuitions that NSs have about RO sets, informa-
tion features, and lexical choices may be to do a detailed analysis of a partic-
ular communicative situation and the possible unmarked choices associated
with it, as discussed earlier. However, the very detailed analysis of situations,
such as the locked car scenario, is not necessary (and would be impossible )
for every foreseeable type of communicative situation. Instead, teachers can
use their own intuitions about markedness to help students understand the
sociopragmatic effects of the linguistic choices they make. Students use this
information as input to their own productive, innate cognitive mechanisms
that underlie the complex choices about markedness available to them.
Note that the projected outcome of this type of pedagogical approach is
not to have learners make only unmarked choices; the marked/unmarked
distinction is vital to give students the linguistic flexibility to reach all their
communicative goals. As competent speakers, L2 learners intentionally make
marked choices when it reflects the best way to reach their communicative
goals. Instead, this approach should help students build and use their linguis-
tic repertoires as native speakers do.
REFERENCES
Blum-Kulka, Shoshana, Juliane House, and Gabrielle Kasper (eds.). 1989. Cross-
cultural pragmatics: requests and apologies. Norwood, N.J.: Alex.
Carter, Bonnie, and Craig Skates. 1996. The Rinebart guide to grammar and style.
Orlando, Fla.: Harcourt Brace College Publishers.
Kasper, Gabrielle, andShoshana Blum-Kulka (eds.). 1993. Interlanguagepragmatics.
Oxford: Oxford University Press.
Myers-Scotton, Carol. 1993B. Social motivations for codeswitching: evidence from

Africa. Oxford: Oxford University Press. Reprint, 1995.
. 1993a. Duelling languages: grammatical structure in codeswitching. Oxford:
Myers-Scotton, Carol, and Janice Jake. 1995. Matching lemmas in a bilingual lan-
guage competence and production model: evidence from intrasentential
codeswitching. Linguistics 33: 981-1024.
Sperber, Dan, and Deirdre Wilson. 1995. Relevance, communication and cognition.
Turner, Ken. 1995. The principal principles of pragmatic inferencing: co-operation.
Language Teaching 28: 67-76.
. 1996. The principal principles of pragmatic inferencing: politeness. Language
Teaching 29: 1-13.
INDEX
Adendorff, Ralph D., 6 Brown, Penelope, 19, 29, 30, 185

aesthetics Bucholtz, Mary, 146
definition of aesthetic pleasure, 102 Bullough, Bonnie, 139
effect, 104,106,115, 119 Bullough, Vern L., 139
experience, 106 Butler, Judith, 141
as function of literary text, 11
pattern, 121 Cameron, Deborah, 180, 181
pleasure, 112-3 Campbell, Lily, 129
poems as aesthetic structures, 11-12, Cargile, Aaron C, 27
109 Carter, Bonnie, 212
All the Pretty Horses, 9, 41-61 Cavendish, Ruth, 184
American Bible Society, 93 Cheuse, Alan, 53
Anchor Bible Dictionary, 93 Chinoy, Eli, 185
As I Lay Dying, 10, 64, 77-83 choice, of linguistic variety, see also style
Atkinson, John M., 180 for aesthetic pleasure, 101-22
audience design model, 30 as appropriate, 195-6
Austin, Timothy R., 101 directives and power relations, 178-
Austin, John L., 182 90
auto factory directives, 14, 178-91 by dramatic characters, 12, 124-35
Aviram, Amittai, 11-12, 101-123 explaining stylistic choices, 7, 44-
45, 57, 63, 83-84, 157-58
Barrett, Rusty, 12-13,139-161 exploiting marked choice, 63
Battistella, Edwin L., 4 giving rise to implicatures, 3—4,41-
Bauman, Richard, 140 43,63
Bell, Vereen, 42 marked lexical choice, 47-48
Bell, Alan, 18, 30 marked syntax, 45-46, 62-88
Berendt, John, 154, 157 marked terms of address, 89-100,
Berlin, Adele, 94 124-33
Bernsten, Janice, 14-15, 27, 178-191 and markedness, 4, 5, 9-15, 22-23,
Biber, Douglas, 142 25-28, 35, 183, 185-7, 189, see
Blake, William, 12, 109-16 also markedness
"The Lamb," 113 as message of intentionality, 3-4
"London," 114-5 as negotiation, 21, 30
"The Poison Tree," 114 as projecting self-image, 177
"TheTyger," 109-14, 115 styles in performances, 139-58
Blakemore, Diane, 7, 43 styles in southern English, 162-77
Blom, Jan-Petter, 162 of variety and rationality, 7-9, 18-
Blommaert, Jan, 19 38, 43
Bloom, Harold, 110 Clark, Herbert H., 21, 29, 30, 31
Blum-Kulka, Shoshana, 27, 196 Clines, David J. A., 94
Boling, Robert, 93 code, see also choice, dialects, styles
"boy" as derogatory term of address, 12, as cover term, 3
124-35 social and psychological functions, 5
Briggs, Charles, 140 cognitive calculations, 9, 16, 19, 66, 189
215
216 INDEX
cognitive constraints, 43-44 social statement, 13, 157-8

cognitive processing, 10, 22, 83 styleswitching, 145-6
cognitive weight, 10, 62-64, 76, 83 Dressier, Wolfgang U., 91
Collinson, David L., 184 Drew, Paul, 35, 36, 178, 180
communication accommodation theory
(CAT), 19, 30 Eastman, Carol M., 6
competence Edwards, Philip, 133
communicative competence, 8, 15, 140 Elster,Jon, 8, 20, 28, 33
pragmatic competence, 195-7, 199, Employe [sic] Handbook, 182
207, 213 Epstein, E. L., 101, 109, 110, 115
Conley, John M., 180 equity principle, 29-30
consequences, see choice, markedness Ervin-Tripp, Susan, 28, 183
constraint Evans, G.B., 131
on cognitive process, 50 exploiting social and psychological asso-
of context in drama, 132 ciations, 13, 145-6, 157-8
of genre, 133 exploratory choice maxim, 26
on interpreting implicatures, 42, 45
in relevance theory, 50 Fairley, Irene R., 101
structural constraints, 33—4, see also Faulkner, William, 10, 77-84
rational actor model Feinberg, Leslie, 139
conversation Ferguson, Charles A., 141, 142, 145
conversation analysis, 21, 35-6 Finegan, Edward, 142
information features, 202-13 Fisher, Sue, 180
structural organization, 36 Fitzgerald, F. Scott, 10, 62-76, 83-84
Cook, Vivian, 5, 185 foregrounding model, criticism, 65-66
cooperative maxims derived from, 133-4 foregrounding, 45-46, 64-65
cooperative principle, 20-22 Form, William H., 180, 182
costs and rewards, 7, 19, 29, 36, see also formulaic speech, 52, 165
optimizing outcomes Frazer, Elizabeth, 180, 181
Coulthard, Malcolm, 185 Freeman, Donald C., 65, 66, 101
Coupland, Justine, 19, 30, 150
Coupland, Nikolas, 19, 30, 150 Gallois, Cynthia, 27
Culler, Jonathan, 121 Gartman, David, 181, 182
Garvin, Paul, 65
Damasio, Antonio, 8, 22, 23, 32 Gates, Henry Louis, Jr., 153
DeBose, Charles, 6 Gaudio, Rudolf P., 146
deference maxim, 26,189 genre, 12
detached participial phrases, 10, 78-83 as constraining linguistic choices, 132
Devor, Holly, 139 conventions of, 124
dialects, see also choice of variety, styles Gibbons, John, 6
African-American Vernacular En- Giles, Howard, 19, 27, 30, 150
glish, 145-8 Givon, Talmy, 91
definition of dialect, register, genre, Goffman, Erving, 29
139, 142-45 Goodwin, Joseph P., 146
Southern English, 162-77 Green, Lisa, 147
Dickinson, Emily, poems, 12, 103-04, Grice, H. Paul, 3, 20, 21, 63, 84, 133
107-08 Gumperz, John J., 162
discourse peaks, 66
drag queens, 12, 139-61 Hall, Kira, 146
INDEX 217
Hamper, Ben, 183 knowledge of consequences, 42-43,

Harper Collins Study Bible, 93 56-57
Harvey, Penelope, 180, 181 Kreml, Nancy, 9-10, 41-61
Hayes, Joseph J., 146
Healy, Bernard, 141 Labov, William, 142
Hebrew [The] Lady Chablis, 154
language, 10, 89, 93, 96 Lakoff, Robin, 146
marked terms of address, 90 Lakoff, George, 65
scripture, 89-100 LePage, Robert B., 145
translation, 10-11, 89-90, 96-97 Levi, Judith N., 180
Heritage, John, 35, 36, 178 Levinson, Stephen, 19, 29, 30, 185
Herrick, Marvin, 127 Lewis, David K., 30
Hill, Frank Ernest, 181 lexical choice
Hopper, Paul, 70 nonnative speaker problems, 211-2
House, Julian, 27, 196 sociopragmatic effects, 196, 198
Howard-Hill, Trevor H., 12, 124-35 terms of address, 10-12, 89-99,
Hurst, Mary Jane, 49 124-35
Hymes,Dell, 140 Li, Wei, 6
literariness, 101-03
implicatures, 3, 41-57, 63 and surprise, 106-08
complex, 44 Longacre, Robert, 63, 66, 76, 93
constraint on interpretation, 9, 43 Luce, Dianne, 42
in marked/unmarked style, 9
recovery of, 42 MacKenzie, Gordene Olga, 139
strong, 3, 50 marked choice maxim, 26
weak, 3, 42, 44 markedness
information features reflecting unmarked competence to assess markedness, 6,
choices, 15, 210 22, 33
institutional talk, 178-9 evaluator, 8, 22-23, 33-34, 197-8,
intentionality, 3-4, 7, 20-22, 63, 102 208
attributed to dramatic characters, 133- implications for pedagogy, 212
34 markedness model, 5-6, 18-36, 43,
how literary texts communicate, 101 124, 141-5, 162, 179-80,197-98
in marked choices, 11, 44, 185-6, markedness model as rational actor
211-13 model, 7-9, 19-20, 29-31
of speaker/writer, 44, 63-64, 84, 98, for native and nonnative English
102, 197 speakers, 207-8
theory of, 3-4, 20-22, 42-43, 63 opposition as meaningful, 16
social consequences, 5-6, 13, 15
Jake, Janice, 71, 203 for styles in novel, 50-53, 63
Jakobson, Roman, 4, 101 as term in structural linguistics, 4-5
Jones, Elizabeth, 27 in terms of address for Israel's ene-
mies, 91-94
Kagel, John H., 31 translating marked biblical referenc-
Kasper, Gabrielle, 27, 196 es, 96-99
Kearns, Josie, 182 in Universal Grammar, 5
Keyser, Samuel Jay, 101, 116, 117 McCarthy, Cormac, 41, 44, 52
Kintsch, Walter, 62 McLemore, Cynthia Ann, 155
Kiparsky, Paul, 101 Meeuwis, Michael, 19
218 INDEX
Michaels, Sarah, 153 pragmatic comparison native and non-

Miller, Gale, 178 native speakers responses, 200-7
Mishler, Elliot G., 180 projection
Mishoe, Margaret, 13-14, 162-77 of complementizer (CP), 10
Mitchell-Kernan, Claudia, 153 elaborating CPs, 10, 68, 72-76, 83
Moran, John, 146 of inflection (IP), 69, 72-76
Morgan, Marcyliena, 153
Morrison, Gail, 42, 52 quantitative comparison of styles, 45-
Muka ovsky, Jan, 101 48, 68-71, 91, 125, 151, 202-04
Myers-Scotton, Carol, 3-17, 18-38, 62- Queen, Robin M., 149, 150, 153
88, 124, 139, 143, 154, 162, 171, 179,
188, 189, 195, 197, 203 Rampton, M. B. H., 180,181
rational actors, 7, 12, 19-20, 34-36, 133
narrative peaks, 76 rational actor model, 8-9, 29-35, see
Nash, Al, 184 also markedness model
negotiating relationships, 84, 132-3, filters in an RA model, 33-35
145-6, 169, 171, 188-90, 197 as normative theory, 35
negotiation principle, 21-22, 124, 179 structural constraints in RA model,
Nevins, Allan, 181 33-34
New Critics, 117 rationality's role in code choice, 8-9,19,
New Revised Standard Version, 89 34-35
New Jewish Publication Society, 96 readers' expectations in drama, 12, 134
Newson, Mark, 5, 185 register, 139, 141-45
Newton, Esther, 140 relevance theory, 8-9, 15, 20-22, 41-43
Nida, Eugene, 89 Richardson, Kay, 180, 181
norms and markedness, 28, 31, 35, 98, rights and obligations
188-90 in composition writing, 208
North American Truck Platforms/Flint effects on information features, 208-
Assembly, 182 13
ROsets, 6-7, 15-16, 23-25, 43, 124,
O'Barr, William, 180 143, 154, 169, 171, 179-80, 197
opportunity as linguistic repertoire, 33 Riverside Shakespeare, 125, 130
opportunity set, 7, 8, 34 Rodgers, Bruce, 146, 147
optimizing outcomes, 9, 19-20, 29-31, Rudes, Blair A., 141
36, 43 RuPaul, 141, 148-51, 152
Ota, Hiroshi, 27
Scotton, Carol see Myers-Scotton
phrasal verbs, 70-71 scripts, 181, 198
Plato, 101 shop talk scripts, 182-5
poetry, 101-23 second language agmatic competence,
aesthetic pattern, 121 195-213
aesthetic pleasure, 106-8 Shakespeare, William
as other than "messages," 108-9 All's Well That Ends Well, 125
versus communication, 101-3 Coriolanus, 126
politeness Hamlet, 134
face threatening acts, 170, 185 King Lear, 129
politeness model, 19-20, 30, 185 Romeo and Juliet, 12, 125-134
strategies, 184-5 Sherzer, Joel, 143
pragmatic competence, 195, 199 Skates, Craig, 212
INDEX 219
Skolvskij, Viktor, 64-65 survival mechanisms, 32-33
social context, role in code choice, 14, Swigart, Leigh, 6
19, 33
Soggins, J. Alberto, 93 Tabor, Charles, 89
somatic marker hypothesis, 8, 22, 32-34 Tabouret-Keller, Andree, 145
southern American English styles Tarlinskaya, Marina, 101
local standard, 13, 163-77 terms of address, 10-11, 12, 89-99,
home style, 13, 163-77 124-35
Spanish, 5, 43, 45, 48, 54 The Great Gatsby, 10, 62-64, 68-76,
Sperber, Dan, 3, 4, 9, 20, 21, 41, 42, 43, 83-84
50, 63, 66, 84, 202 Thompson, Sandra A., 64, 78, 81, 82
Spevack, Marvin, 125 Todd, Alexandra Dundas, 180
Sroda, Mary Sue, 15-16, 195-214 Tracy, Karen, 19
Sternberg, Meir, 99 trailing constituents, 10, 64, 77-83
Stevens, Wallace, 12 Trubetzkoy, Nikolai, 4
"The Snow Man," 116-20 Turner, Victor, 140
styles, 7, 141—42 see also choice, code, Turner, Ken, 196
dialects, markedness
African-American style, 13, 145, 147, UAW, 182, 183, 184, 189
148-53, 157-8 unmarked choice maxim, 26
cognitive constraints, 42—45
cognitive processing, 62, 83-84 Van Peer, Willie, 65
Darl's monologues, 10, 78-83 variety, see choice, dialects, styles
family gatherings, 13, 162-77 Venuti, Lawrence, 89
gaymale style, 13, 145, 147, 149-53, villain, 130-32
157-8 virtual communication, 11, 102
interaction of styles in novel, 53-57 virtuosity maxim, 26
literariness, 101-22 Vonnegut, Jr., Kurt, 89
marked style in novel, 50-53, 63-64
unmarked style in novel, 48-50, Walker, Anne Graffam, 180
63-64 Walters, Keith, 146,147
white woman style, 13,146-8, 155, West, Candace, 180
157-8 Wiemann, John M., 30
styleswitching, 6, see also styles Wilson, Deirdre, 3, 4, 9, 20, 21, 41, 42,
motivations for switching, 44-45, 43, 50, 63, 66, 84, 202
57, 63-64, 83-84, 92-94, 98- Wilt, Timothy, 10-11, 89-100
99, 131-32, 143-5, 157-8, 177, Witek, Terri, 52
189, 212 Wyatt, Toya A., 146
Sunaoshi, Yukako, 146
surprise in poetry, 106-08, 122 Zhu, Wanjin, 6

Carol Myers-Scotton - Codes and Consequences PDF

Uploaded by

Copyright:

Available Formats

Carol Myers-Scotton - Codes and Consequences PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Carol Myers-Scotton - Codes and Consequences PDF

Uploaded by

Copyright:

Available Formats

Codes and Consequences:

Choosing Linguistic Varieties

OXFORD UNIVERSITY PRESS

Choosing Linguistic Varieties

New York Oxford

Copyright © 1998 by Carol Myers-Scotton

Library of Congress Cataloging-in-Publication Data

The editor, authors, and publisher gratefully acknowledge permission to re-

"Stylistic variation and cognitive constraint in Cormac McCarthy's All the

The Great Gatsby by F. Scott Fitzgerald. © 1925, Charles Scribner's Sons.

As I Lay Dying by William Faulkner, from © 1930, renewed 1957. From

II. STYLISTIC CHOICES IN LITERATURE

III. STYLISTIC CHOICES IN SPOKEN ENGLISH

IV. STYLISTIC CHOICES AND SECOND-LANGUAGE ACQUISITION

Amittai F. Aviram is associate professor of English and comparative litera-

Rusty Barrett is a doctoral candidate in the linguistics department at the

Janice Bernsten is an associate professor of linguistics at the University of

Trevor Howard-Hill is the C. Wallace Martin Professor of English at the

Nancy Kreml is a member of the English faculty at Midlands Technical Col-

Margaret Mishoe is a member of the English faculty at the University of

Carol Myers-Scotton is a Carolina Distinguished Professor at the University

Timothy Wilt is a translation consultant for the United Bible Societies. He

MARKEDNESS AND CHOICES OF VARIETIES

Markedness in Structural Linguistics

Markedness and Social Consequences

norms of the dominant group), especially when the situation or genre is

Applying the Markedness Model to New Types of Data

THE INTENDED AUDIENCE

The volume is designed to interest faculty and students in the disciplines of

ORGANIZATION OF THE VOLUME

Following this introduction, Part I includes a chapter giving a theoretical

THE MARKEDNESS MODEL AS A RATIONAL ACTOR MODEL

In chapter 2, "A Theoretical Introduction to the Markedness Model," I present

Part II consists of five chapters on literary works. It opens with chapter 3,

SWITCHING IN NATURALLY OCCURRING TALK

TEACHING WHAT'S MARKED, WHAT'S NOT

The diversity of subject matter in the chapters in this volume demonstrates—

approaches just mentioned can also be considered rational actor models to an

INTERPRETATION AS MORE THAN DECODING

from an interaction. Sequential organization in particular is a major feature

OUTLINING THE MARKEDNESS MODEL

alone or cognition alone" (Damasio 1996: 1413). As should become clearer

THE PROBLEM OF RO SETS

calling it out of place? In addition, even to classify two participants in rela-

THE MARKEDNESS MODEL EXPLAINED

In the following sections, I present the MM as a rational actor (RA) model.

Maxims under the Markedness Model

with a little modification, writers can apply at least the unmarked-choice

How Speakers Apply the Maxims

soon after his appointment I called him on departmental business. As a con-

The Markedness Model as a Rational Actor Model

by "maintaining equity." Clark never makes that clear. If "maintaining

evidence. Note that to be considered rational, to simply hold intentions based

SOCIAL NORMS AND RATIONAL CHOICES

THE MARKEDNESS MODEL AND THE

MODIFYING ELSTER'S RA MODEL FOR THE

locally produced, incrementally developed, and, by extension, transformable

ysis: proceedings of the first international conference on discourse analysis, ed.

Implicatures of Styleswitching in the Narrative

x2 = 60.617 (p <.001) 24.32 (p < .001)