INTERLANG PRAGMATICS KASPER.pdf - STIBA Malang
INTERLANG PRAGMATICS KASPER.pdf - STIBA Malang
INTERLANG PRAGMATICS KASPER.pdf - STIBA Malang
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Interlanguage Pragmatics
This page intentionally left blank
Interlanguage<br />
Pragmatics<br />
EDITED BY<br />
Gabriele Kasper<br />
and Shoshana Blum-Kulka<br />
New York Oxford<br />
OXFORD UNIVERSITY PRESS<br />
1993
Oxford University Press<br />
Oxford New York Toronto<br />
Delhi Bombay Calcutta Madras Karachi<br />
Kuala Lumpur Singapore Hong Kong Tokyo<br />
Nairobi Dar es Salaam Cape Town<br />
Melbourne Auckland Madrid<br />
and associated companies in<br />
Berlin Ibadan<br />
Copyright © 1993 by Oxford University Press, Inc.<br />
Published by Oxford University Press, Inc.<br />
200 Madison Avenue, New York, New York 10016<br />
Oxford is a registered trademark of Oxford University Press, Inc.<br />
All rights reserved. No part of this publication may be reproduced,<br />
stored in a retrieval system, or transmitted, in any form or by any means,<br />
electronic, mechanical, photocopying, recording or otherwise,<br />
without the prior permission of Oxford University Press.<br />
Library of Congress Cataloging-in-Publication Data<br />
Interlanguage pragmatics / edited by Gabriele Kasper<br />
and Shoshana Blum-Kulka.<br />
p. cm. ISBN 0-19-506602-2<br />
1. Interlanguage (Language learning) 2. Pragmatics. 3. Speech<br />
acts (Linguistics) I. Blum-Kulka, Shoshana.<br />
P53.15193 1993<br />
410.4'4—dc20 92-34791<br />
1 35798642<br />
Printed in the United States of America<br />
on acid-free paper
CONTENTS<br />
Interlanguage Pragmatics: An Introduction, 3<br />
Gabriele Kasper and Shoshana Blum-Kulka<br />
I<br />
Cognitive Approaches to Interlanguage<br />
Pragmatic Development, 19<br />
1 Consciousness, Learning, and Interlanguage Pragmatics, 21<br />
Richard Schmidt<br />
2 Symbolic Representation and Attentional Control<br />
in Pragmatic Competence, 43<br />
Ellen Bialystok<br />
II Speech Act Realization, 59<br />
3 Expressing Gratitude in American English, 64<br />
Miriam Eisenstein and Jean Bodman<br />
4 Perception and Performance in Native and Nonnative Apology, 82<br />
Marc L. Bergman and Gabriele Kasper<br />
5 Interlanguage Features of the Speech Act of Complaining, 108<br />
Elite Olshtain and Liora Weinbach<br />
6 Interlanguage Requestive Hints, 123<br />
Elda Weizman<br />
1 Cross-Linguistic Influence in the Speech Act of Correction, 138<br />
Tomoko Takahashi and Leslie M. Beebe<br />
III Discourse Perspectives, 159<br />
8 Toward a Model for the Analysis of Inappropriate Responses<br />
in Native/Nonnative Interactions, 161<br />
Juliane House<br />
9 Explaining NNS Interactional Behavior:<br />
The Effect of Conversational Topic, 184<br />
Jane Zuengler
vi<br />
Contents<br />
10 The Metapragmatic Discourse of American-Israeli Families<br />
at Dinner, 196<br />
Shoshana Blum-Kulka and Hadass Sheffer<br />
11 Notes on the Interlanguage of Comity, 224<br />
Guy Aston<br />
ABOUT THE AUTHORS, 251
Interlanguage Pragmatics
This page intentionally left blank
Interlanguage Pragmatics:<br />
An Introduction<br />
GABRIELE <strong>KASPER</strong> and SHOSHANA BLUM-KULKA<br />
The Scope of Interlanguage Pragmatics<br />
Interlanguage pragmatics (ILP) is a second-generation hybrid. As its name betrays,<br />
ILP belongs to two different disciplines, both of which are interdisciplinary. As a<br />
branch of Second Language Acquisition Research, ILP is one of several specializations<br />
in interlanguage studies, contrasting with interlanguage phonology, morphology,<br />
syntax, and semantics. As a subset of pragmatics, ILP figures as a sociolinguistic,<br />
psycholinguistic, or simply linguistic enterprise, depending on how one<br />
defines the scope of "pragmatics." For thorough discussion of definitional issues,<br />
see Leech (1983) and Levinson (1983). The perspective on pragmatics we adopt is<br />
an action-theoretical one, viewing pragmatics as the study of people's comprehension<br />
and production of linguistic action in context. Interlanguage pragmatics has<br />
consequently been defined as the study of nonnative speakers' use and acquisition of<br />
linguistic action patterns in a second language (L2) (e.g., Kasper, 1989b). Yet tying<br />
interlanguage pragmatics to nonnative speakers, or language learners, may narrow<br />
its scope too restrictively. As Blum-Kulka (1991; Blum-Kulka & Sheffer, Chapter<br />
10) demonstrates through the case of American immigrants to Israel, speakers fully<br />
competent in two languages may create an intercultural style of speaking that is both<br />
related to and distinct from the styles prevalent in the two substrata, a style on which<br />
they rely regardless of the language being used. The intercultural style hypothesis is<br />
supported by many studies of cross-cultural communication, notably interactional<br />
sociolinguistics (e.g., Gumperz, 1982; Tannen, 1985) and research into the pragmatic<br />
behavior of immigrant populations across generations (e.g., Clyne, 1979;<br />
Clyne, Ball, & Neil, 1991). It also receives strong anecdotal support, worthy of<br />
systematic investigation, by highly proficient nonnative speakers whose L2 conversational<br />
behavior carries interlanguage-specific traits, and who claim at the same<br />
time that they do not abide by native norms any more when conversing in their<br />
native language. For instance, one of us was told by several of her Chinese students<br />
that in response to invitations and offers they wish to accept, they no longer engage<br />
3
4 Interlanguage Pragmatics: An Introduction<br />
in ritual refusal, as required by traditional Chinese culture. Some of her Japanese<br />
students claim that they are much more direct in their interaction in Japanese than they<br />
used to be before extended exposure to Western ways. Emerging intercultural styles,<br />
so prevalent in the international academic community, deserve interlanguage pragmaticists'<br />
close attention. Hence, it appears useful to include under ILP the study of<br />
intercultural styles brought about through language contact, the conditions for their<br />
emergence and change, the relationship to their substrata, and their communicative<br />
effectiveness. A look at the literature on ILP (cf. the overview in Kasper & Dahl,<br />
1991), however, suggests that the populations studied have invariably been nonnative<br />
speakers, reflecting the status of ILP as a branch of second language research. While<br />
the present collection largely follows this line, Blum-Kulka and Sheffer (Chapter 10)<br />
extend the perspective to include native speakers' intercultural styles.<br />
Furthermore, according to researchers' labeling of their objects of study, ILP<br />
predominantly refers to the comprehension and production of linguistic action,<br />
including discourse regulation. An area of study that most investigators would<br />
clearly recognize as "pragmatic" yet that is not usually included under ILP is<br />
communication strategies. The de facto separation of pragmatics and communication<br />
strategies in second language studies reflects different alignments chosen by<br />
researchers in each area. The study of communication strategies has predominantly<br />
been grounded in psycholinguistic models of cognitive processing (e.g., Poulisse,<br />
1990; Bialystok, 1990), whereas ILP has derived its theoretical and empirical foundation<br />
from general and especially cross-cultural pragmatics (e.g., Blum-Kulka,<br />
House, & Kasper, 1989). For the most part, research on communication strategies<br />
has examined learners' solutions to referential problems; ILP has focused on the<br />
illocutionary and politeness dimensions of speech act performance. While this<br />
division of labor reflects different research traditions, it has little theoretical support.<br />
In Bachman's (1990, 84ff.) model of communicative competence, for example,<br />
pragmatic competence, a component of language competence, subdivides into<br />
sociolinguistic and illocutionary competence, which in turn entails the ability to<br />
express a variety of communicative functions, such as making reference. Strategic<br />
competence is seen as processing ability, operating on the language competence in<br />
its entirety and including "strategic" solutions to comprehension or production<br />
problems. In this volume, strategic aspects of speech act performance and discourse<br />
participation are examined by Weizman (Chapter 6), Aston (Chapter 11), and<br />
House (Chapter 8); processing considerations for pragmatic development are proposed<br />
by Schmidt (Chapter 1) and Bialystok (Chapter 2).<br />
Pragmatic Comprehension<br />
Domains of ILP<br />
Early studies focused on learners' attribution of illocutionary force and perception<br />
of politeness. Research on the attribution of illocutionary force has centered on the<br />
comprehension of indirect speech acts, factors contributing to ease or difficulty of<br />
pragmatic comprehension, the role of linguistic form and context information, and
Interlanguage Pragmatics: An Introduction 5<br />
learner variables influencing force attribution. Carrell (1979) demonstrated that<br />
advanced L2 learners have complete access to conversational implicature, and make<br />
full use of their inferential ability in the comprehension of indirect speech acts. The<br />
only stumbling block for these learners was the "pope answer," a highly culturespecific<br />
violation of the maxim of relation. Bouton (1988) was interested in the<br />
impact of cultural background on the comprehension of indirect answers. He found<br />
a significant difference between six groups of learners from different cultural backgrounds<br />
and native speakers of American English. Comparison of the learner<br />
groups showed similar perceptions for German, Spanish-Portugese, and Taiwanese<br />
learners, differing from those of Korean, Japanese, and Chinese learners from the<br />
People's Republic of China. In addition to influence from learners' cultural background,<br />
Bouton also established an effect for type of implicature. Comprehension<br />
was easiest when the relevance maxim had been violated, whereas understated<br />
negative evaluation, a violation of the maxim of quantity, was more difficult to<br />
understand, for nonnative speakers as well as native speakers. In examining the<br />
relative effect of linguistic form and context information on learners' perception of<br />
indirect requests, Carrell (198 la, b) found that learners primarily relied on linguistic<br />
utterance features, regardless of their linguistic and cultural background, age, and<br />
proficiency. Her findings are at odds with those established by Ervin-Tripp, Strage,<br />
Lampert, and Bell (1987), who found that children acquirifig their first (LI) and<br />
second language strongly base their pragmatic comprehension on situational information,<br />
disregarding form. Other studies in developmental pragmatics (e.g., Reeder<br />
& Wakefield, 1987) support Ervin-Tripp et al.'s results (cf. Bialystok, Chapter 2).<br />
The apparent conflict in outcomes of Carrell's and Ervin-Tripp et al.'s studies can be<br />
reconciled by invoking Gibbs's (e.g., 1984) Conventional Meaning Model. In a<br />
series of experiments, Gibbs demonstrated that listeners directly access nonliteral<br />
meanings if linguistic forms and situational contexts are conventionalized. Absence<br />
of familiar and appropriate contexts and novel, nonconventionalized utterances<br />
requires (sequential) processing of literal and nonliteral meaning. Carrell's studies<br />
precisely illustrate the second condition, since the highly abstract task her subjects<br />
were faced with could only be solved via bottom-up processing. Ervin-Tripp et al.'s<br />
task conditions, on the other hand, fullfil the criteria that would bias listeners<br />
toward immediate processing of indirect meaning: the indirect requests were embedded<br />
in everyday situations familiar to the children, allowing them to apply<br />
situational schemata to their understanding of directive utterances. In partial replication<br />
of Gibbs's research on the processing of conventionally indirect requests (e.g.,<br />
1983), Takahashi and Roitblat (1992) examined through latency measurement<br />
whether Japanese learners of English reconstruct implied requestive force by processing<br />
both literal and implied meanings or immediately recover the nonliteral<br />
meaning (see also Takahashi, 1990, for a review of the literature on nonliteral<br />
utterance comprehension). Models of pragmatic comprehension, representing a variety<br />
of theoretical orientations, are reviewed by House (Chapter 8), who also offers<br />
an analysis of learners' pragmatic comprehension in ongoing interaction.<br />
In all of the studies cited above, English figures as L2. The only study on force<br />
attribution by learners of a different target language is a small investigation by<br />
Koike (1989), who examined how beginning classroom learners of Spanish with
6 Interlanguage Pragmatics: An Introduction<br />
English as LI understood a Spanish request, apology, and command. Learners<br />
correctly identified the illocutions in 95% of the cases or better. In seeking cues to<br />
decide on force, learners were shown to rely most frequently on formulaic illocutionary<br />
force indicators such as "por favor" and "lo siento," and on words with key<br />
prepositional meanings.<br />
A related line of inquiry has examined how learners assess the politeness value<br />
of different speech act realization strategies. While most of the research on force<br />
attribution has studied on-line utterance comprehension, politeness perceptions have<br />
been investigated through off-line metapragmatic judgment tasks such as card sorting<br />
(Carrell & Konneker, 1981; Tanaka & Kawade, 1982), paired comparison<br />
(Walters, 1979), multiple choice (Tanaka & Kawade, 1982), and rating scales<br />
(Eraser, Rintell, & Walters, 1980; Rintell, 1981; Olshtain & Blum-Kulka, 1985;<br />
Kitao, 1990). Results confirm learners' ability to distinguish different degrees of<br />
politeness in conventions of means and forms, although their perceptions do not<br />
always agree with those of native speakers. Japanese learners of English largely<br />
agreed with American native speakers in their relative politeness judgments of<br />
syntactic modes (imperative, declarative, interrogative), the politeness marker<br />
"please," and of deferential address terms, but they differed in the politeness values<br />
attributed to request modification by tense and modals. Whereas American informants<br />
perceived positively worded requests as more polite than negatively worded<br />
ones, this assessment was not shared by the Japanese raters. Negative politeness<br />
strategies (Brown & Levinson, 1987) were rated as more polite by Americans and<br />
Japanese in the United States than by Japanese in Japan (Kitao, 1990).<br />
Learners' differential politeness perceptions have been attributed to a variety of<br />
factors. Learners were found to differ in the extent to which they base their politeness<br />
perceptions in L2 on those in LI. Spanish learners of English did not transfer<br />
their LI perceptions of formally equivalent requestive strategies to L2, rating their<br />
Spanish requests as more deferential than their English counterparts (Eraser et al.,<br />
1980). Learners of Hebrew, who appeared to base their politeness perceptions of L2<br />
requests and apologies initially on LI, increased their tolerance for directness and<br />
positive politeness (Brown & Levinson, 1987) with length of residence in the target<br />
community (Olshtain & Blum-Kulka, 1985). Length of residence rather than L2<br />
proficiency accounted for differential politeness perceptions in learners and L2<br />
native speakers. The rec.eiver's age and sex influenced politeness assessment in<br />
Spanish learners' perception of English requests, whereas no such effects were<br />
noticeable in suggestions (Rintell, 1981). Comparing the perceptions of politeness<br />
in request strategies by Japanese EFL and ESL learners, Kitao (1990) found that<br />
exposure to English accounted for different ratings between those groups.<br />
In addition to learners' assessments of pragmalinguistic information, a few<br />
studies have also examined nonnative speakers' sociopragmatic perceptions. Probing<br />
into learners' "universal" and culture-specific assumptions about apology frequency<br />
and realization, Olshtain (1983) found that for Russian learners of Hebrew,<br />
the event rather than culture and language were the decisive variables. Englishspeaking<br />
learners felt less need to apologize in Hebrew than in English, thus suggesting<br />
a culture-specific approach. Asked to assess the weight and values of<br />
contextual factors in apologizing, German learners of English were found largely to
Interlanguage Pragmatics: An Introduction 7<br />
agree with native speakers of British English, except for degree of imposition<br />
involved in the apology, which the German raters found to be consistently higher<br />
than the English judges (House, 1988). Ratings of contextual factors in apologies<br />
were also provided by Thai learners of English and American native speakers<br />
(Bergman & Kasper, Chapter 4). The factor on which these informants differed<br />
most was obligation to apologize. Japanese learners of English reported that in<br />
American society, refusal was a more socially acceptable act than in Japan, and<br />
could therefore appropriately be carried out more directly (Robinson, 1992). There<br />
are very few studies that examine learners' sociopragmatic perceptions by direct<br />
probing, such as rating tasks (e.g., House, 1988; Bergman & Kasper, Chapter 4;<br />
Olshtain & Weinbach, Chapter 5) and self-report in interviews (Cohen & Olshtain<br />
1991; Robinson, 1992). Many more studies infer learners' L2 sociopragmatic<br />
knowledge through their production. Thus Beebe and collaborators (e.g., Beebe,<br />
Takahashi, & Uliss-Weltz, 1990; Beebe & Takahashi, 1989a, b), focusing on the<br />
effect of status on the performance of face-threatening acts by Japanese learners of<br />
English, found that the Japanese informants style-shifted more according to interlocutor<br />
status than speakers of American English (but see Takahashi & Beebe,<br />
Chapter 7, for a counterexample). What research still needs to demonstrate is how<br />
learners' sociopragmatic perceptions change over time, and how such change is<br />
reflected in their linguistic action patterns.<br />
Production of Linguistic Action<br />
The available evidence suggests that regardless of a particular L1 and L2, and of the<br />
type of learning context (naturalistic vs. instructed), learners have access to the<br />
same range of realization strategies for linguistic action as native speakers, and<br />
demonstrate sensitivity to contextual constraints in their strategy choice (Blum-<br />
Kulka, 1982; Kasper, 1989b; Rintell & Mitchell, 1989). The main obstacle to<br />
learners' exploiting their "general pragmatic knowledge base" (Blum-Kulka, 1991)<br />
appears to be their restricted L2 linguistic knowledge, or difficulty in accessing it<br />
smoothly (e.g., Blum-Kulka, 1982; Koike, 1989; Edmondson & House, 1991). But<br />
other factors intervene: a lack of L2 pragmalinguistic sophistication, combined with<br />
negative transfer of sociopragmatic norms from LI or nonnative perceptions of L2<br />
sociopragmatic norms, or even purposeful loyalty to LI cultural patterns, may yield<br />
deviations from native use at high proficiency levels as well (Blum-Kulka, 1991).<br />
While ILP research has by now covered a wide variety of typologically different<br />
Lls, there is still only a handful of languages studied as L2: in addition to different<br />
national varieties of English, a few studies have examined Hebrew (Blum-Kulka,<br />
1982, 1991; Olshtain, 1983; Olshtain & Cohen, 1989; Olshtain & Weinbach, Chapter<br />
5), German (Faerch & Kasper, 1989), Norwegian (Svanes, 1989), Spanish<br />
(Koike, 1989), and Japanese (Sawyer, 1992) as targets. Until the scope of target<br />
languages has been considerably broadened, universality claims need to be voiced<br />
with caution.<br />
Learners' distribution patterns of strategies and forms have been shown to vary<br />
from those of native speakers. Some studies report that learners prefer more direct<br />
modes of conveying pragmatic intent than native speakers; others suggest the oppo-
8 Interlanguage Pragmatics: An Introduction<br />
site. Preference for higher directness in IL than in both or either LI or L2 has been<br />
documented in learners' requests (Tanaka 1988; Koike, 1989; Fukushima, 1990),<br />
making and rejecting of suggestions (Bardovi-Harlig & Hartford, 1990), refusals<br />
(Robinson, 1992), and a variety of conflictive acts (Kasper, 1981). By contrast,<br />
nonnative speakers of Hebrew preferred less direct requests (Blum-Kulka, 1982,<br />
1991) and complaints (Olshtain & Weinbach, Chapter 5) than native speakers.<br />
Explanations of these findings have been sought in the scope of learners' linguistic<br />
IL knowledge, transfer from LI, and perceptions of what is sociopragmatically<br />
appropriate in the target community. Some studies noted differences in politeness<br />
approach between native speakers and nonnative speakers. Venezuelan Spanishspeaking<br />
learners of English systematically used positive politeness strategies when<br />
apologizing to a host for not having attended her party, whereas American English<br />
native speakers preferred a negative politeness approach (Garcia, 1989). Conversely,<br />
in pre-trial interviews, Athabaskan defendants employed negative politeness<br />
strategies, whereas white defendants as well as the interviewers deployed more<br />
positive politeness (Scollon & Scollon, 1983). In order to repair uncomfortable<br />
moments in academic advising sessions at an American university, students who<br />
were native speakers of English predominantly relied on positive politeness strategies,<br />
while Taiwanese students adopted a negative politeness approach (Fiksdal<br />
1990). Less reliance by nonnative speakers on positive politeness has also been<br />
observed in the performance of corrections by Japanese speakers of English (Takahashi<br />
& Beebe, Chapter 7). American informants in status-higher positions tended<br />
to preface a correction by a positive remark, whereas Japanese nonnative speakers<br />
used this solidarity strategy very infrequently.<br />
Nonnative speakers' strategy choice is sometimes less responsive to contextual<br />
factors than native speakers'. Japanese learners of English used the same (direct,<br />
barely mitigated) requestive strategies in conversation with status-unequal and socially<br />
distant interlocutors as with status-equal and familiar coparticipants, whereas<br />
American English controls varied their request behavior in the two conditions (Tanaka,<br />
1988). Japanese learners of English also underdifferentiated their realizations<br />
of offers and requests in three conditions of social distance (Fukushima, 1990). In<br />
expressing gratitude, native speakers of American English varied the length of their<br />
speech activity according to degree of indebtedness; no such effect was found for<br />
nonnative speakers of different linguistic backgrounds (Eisenstein & Bodman 1986,<br />
Chapter 3; Bodman & Eisenstein, 1988).<br />
The quality and range of linguistic forms by which linguistic action can be<br />
implemented and modified has consistently been shown to differ between native<br />
speakers and nonnative speakers, the nonnative speakers' repertoire typically being<br />
more restricted and less complex than native speakers' (Scarcella, 1979; Kasper,<br />
1981; Blum-Kulka, 1982, 1991; Schmidt, 1983; House & Kasper, 1987; Trosborg,<br />
1987, Faerch & Kasper, 1989; Eisenstein & Bodman, Chapter 3). While some of<br />
these findings may simply reflect the state of learners' lexical and syntactic knowledge,<br />
the issue becomes a clearly pragmalinguistic one when learners demonstrably<br />
"know" a particular lexical item or syntactic structure yet use it in a way that does<br />
not convey the intended illocutionary force or politeness value. In the interlanguage
Interlanguage Pragmatics: An Introduction 9<br />
of nonnative speakers of Hebrew, requests are lended an unintended whining emphasis<br />
by the use of "bevakasha" (please) in sentence-initial (rather than intra- or<br />
postsentential) position, and such pragmalinguistic deviations are singled out by<br />
native judges as "nonnative" (Blum-Kulka, 1991). One area where insufficient<br />
control of pragmalinguistic knowledge is particularly obvious is that of pragmatic<br />
routines. Coulmas's (1981) contention that routine formulas are a serious stumbling<br />
block for nonnative speakers has been supported by nearly every ILP study (e.g.,<br />
Scarcella, 1979; House, 1988; Eisenstein & Bodman, 1986; Kasper, 1989a;<br />
Wildner-Bassett, 1984; Fukushima, 1990), yet there has been little systematic investigation<br />
of this phenomenon. There is evidence of learners supplying near-literal<br />
translations of LI routines: for example, the German, "entschuldigen Sie bitte"<br />
(English, "excuse me, please") instead of "I'm sorry" (House, 1988); of using a<br />
translation equivalent for an LI routine where none is used in L2, such as prefacing<br />
a high-imposition request with "I'm sorry" (from Japanese "sumimasen" or<br />
"gomennasai" [Fukushima, 1990]); and of failing to use any kind of routine where<br />
one would be required, such as failing to offer an expression of gratitude (Kasper,<br />
1981). Bodman and Eisenstein (1988) observed that in the attempt to express gratitude<br />
in English, learners would use literal translations of LI proverbial expressions<br />
in written production questionnaires but not in role-plays, which displayed considerable<br />
disfluencies but no overt use of LI proverbs (also Eisenstein & Bodman,<br />
Chapter 3).<br />
At the nonroutinized end of speech act production, learners have been found to<br />
engage in more speech activity than native speakers. The "waffle phenomenon"<br />
(Edmonson & House, 1991) has been noted in requests (Blum-Kulka & Olshtain,<br />
1986; House & Kasper, 1987; Faerch & Kasper, 1989; Weizman, Chapter 6),<br />
apologies (House, 1988, Bergman & Kasper, Chapter 4), and complaints (Olshtain<br />
& Weinbach, Chapter 5), as well as in referential communication (Bongaerts, Kellerman,<br />
& Bentlage, 1987; Tarone & Yule, 1987; Yule & Tarone, 1990). According<br />
to Blum-Kulka and Olshtain (1986), waffling is proficiency-dependent, being<br />
strongest at an intermediate stage when learners possess the linguistic means to say<br />
as much as they wish, yet at the same time feel more of a need to be explicit about<br />
their communicative goals and the reasoning behind them than more acculturated<br />
nonnative speakers do. Edmondson and House (1991) point out, however, that the<br />
waffling effect in speech act realization is observable only in learners' written<br />
responses to production questionnaires, not in role-plays. This observation corroborates<br />
Bodman and Eisenstein's findings about the differential use of proverbial<br />
material in written and oral-interactive production, and the much shorter contributions<br />
made in role-plays by nonnative speakers in comparison with native speakers<br />
(Eisenstein & Bodmann, Chapter 3). In Edmondson and House's analysis (see<br />
Bergman & Kasper, Chapter 4, for more discussion), learners' extensive use of<br />
supportive strategies in the absence of formulaic routines suggests that nonroutinized<br />
material functions to compensate for the lack of automatized discourse<br />
routines. It will be worthwhile for IL pragmaticists to examine whether this hypothesis<br />
bears out across languages and tasks, and whether it interacts with factors such<br />
as proficiency and length of residence.
10 Interlanguage Pragmatics: An Introduction<br />
Development of Pragmatic Competence<br />
The bulk of ILP research focuses on nonnative speakers' use of pragmatic knowledge<br />
in comprehension and production, rather than on development. This focus is<br />
also adopted by the data-based studies in this book. Of the available developmental<br />
investigations, some cross-sectional studies did not find proficiency effects in<br />
learners' strategy selection (Takahashi & Beebe, 1987; Trosborg, 1987; Svanes,<br />
1991) whereas others did (Takahashi & DuFon, 1989; Maeshiba, Yoshinaga, Kasper<br />
& Ross, in press). However, developmental effects were observable in learners'<br />
repertoires of pragmatic routines and modality markers (Scarcella, 1979; Trosborg,<br />
1987). Possibly, the inconsistent results reflect instrument effects and the difficulty<br />
of determining precisely learners' proficiency levels across studies. The few longitudinal<br />
studies to date indicate distinct developmental patterns in learners' request<br />
realization (Schmidt, 1983; Ellis, 1992) and use of a sentence-final particle in<br />
Japanese (Sawyer, 1992). They strongly suggest the need for more longitudinal<br />
studies in naturalistic settings, observing learners from the very beginning of their<br />
language acquisition process. Equally important as a reliable and valid empirical<br />
data base is a theoretical framework to account for pragmatic learning. In this book<br />
two such frameworks, grounded in different models of cognitive processing, are<br />
proposed by Bialystok (Chapter 1) and Schmidt (Chapter 2; see also Kasper &<br />
Schmidt, 1992).<br />
Pragmatic Transfer<br />
Influence from learners' native language and culture on their IL pragmatic knowledge<br />
and performance has been amply documented. Because of its potential risk to<br />
communicative success, the focus has been on negative transfer; that is, the influence<br />
of LI pragmatic competence on IL pragmatic knowledge that differs from the<br />
L2 target. Positive transfer, that is, pragmatic behaviors or other knowledge displays<br />
consistent across LI, IL, and L2, have received less attention. We think this is<br />
so because positive transfer usually results in communicative success and therefore<br />
is less exciting to study. Furthermore, it is methodologically difficult to disentangle<br />
positive transfer from universal pragmatic knowledge and generalization on the<br />
basis of available IL pragmatic knowledge.<br />
Negative transfer has been attested to at the sociopragmatic level, influencing<br />
learners' perception of status relationships (Beebe, Takahashi, & Uliss-Weltz, 1990;<br />
Takahashi & Beebe, Chapter 7), of the appropriateness of carrying out refusals<br />
(Robinson, 1992), of the need to apologize (Olshtain, 1983) and to express gratitude<br />
(Eisenstein & Bodman, Chapter 3); their lack of accommodation to target norms for<br />
complimenting, responding to compliments, and negotiating invitations (Wolfson,<br />
1989); their choice of politeness style (Garcia, 1989; Olshtain & Cohen, 1989,<br />
Blum-Kulka, 1982) and of particular strategic options (House, 1988; Beebe, Takahashi,<br />
& Uliss-Weltz, 1990). Evidence for pragmalinguistic transfer at the level of<br />
form-force mapping has been less documented in the literature than anecdotal<br />
accounts would suggest (Blum-Kulka, 1982; Faerch & Kasper, 1989 in requesting;<br />
House, 1988; Olshtain & Cohen, 1989 in apologizing; Bodman & Eisenstein, 1988,
Interlanguage Pragmatics: An Introduction 11<br />
in thanking); possibly learners' proficiency levels in most ILP studies were too high<br />
for negative form-force transfer to occur. Most of the reported pragmalinguistic<br />
transfer affects the strategic options and forms that modify the politeness value of a<br />
linguistic act. Learners' choices of semantic formulas and linguistic tokens for<br />
apologizing were influenced by LI patterns (Olshtain & Cohen, 1989; Trosborg,<br />
1987; House, 1988; Bergman & Kasper, Chapter 4; Maeshiba, Yoshinaga, Kasper<br />
& Ross, in press). LI lexical and syntactic material used to mitigate requestive force<br />
was transferred to learners' IL request performance (Blum-Kulka, 1982; Blum-<br />
Kulka & Levenston, 1987; Faerch & Kasper, 1989). Questioning patterns employed<br />
in Japanese to express disagreement were used as disagreement strategies in Japanese<br />
learners' English (Beebe & Takahashi, 1989b).<br />
Although the literature abounds in evidence for pragmatic transfer, little has yet<br />
been done to investigate more closely the conditions under which pragmatic transfer<br />
is or is not operative. Whereas there is some indication from performance data<br />
(Kasper, 1981; House & Kasper, 1987; Faerch & Kasper, 1989) and retrospective<br />
reports (Olshtain, 1983; Cohen & Olshtain, 1991; Robinson 1992) that learners<br />
invoke criteria such as assessment of linguistic and cultural distance, and of the<br />
specificity or sameness of pragmalinguistic or sociopragmatic patterns in L1/L2,<br />
there is only one study to date that systematically examines pragmatic transferability.<br />
Takahashi (1992) found that the transferability of conventionally indirect<br />
requests from Japanese to English was highly context-dependent, and varied with<br />
learners' proficiency and familiarity with the situation.<br />
At a conceptual level, it is important to note that "negative" pragmatic transfer<br />
does not necessarily reflect lack of competence in the pragmatics of the target<br />
community. When nonnative speakers communicate in a style different from native<br />
ways of speaking, it is a matter of attribution if this style is seen as lacking in some<br />
way, or just different, and if its maintenance over time is considered negatively, as<br />
fossilization, or positively, as a marker of cultural identity. The degree of sociocultural<br />
accommodation to the L2 culture may be as well a matter of choice as of<br />
ability. A foreign accent, for example, can well shield a nonnative speaker, identifying<br />
her as nonnative, and thereby flexing the norms by which she is judged and<br />
lending her a certain latitude in choosing her ways of speaking. The desirable goal<br />
for the high-proficiency second language speaker, be it in contexts of immigration<br />
or in the use of L2 in cross-cultural communication, may well be that of disidentification,<br />
rather than absolute convergence. There is thus much room for future<br />
research, which will be necessary in order to advance beyond merely ascertaining<br />
pragmatic transfer to understanding its differential modes of operation and symbolic<br />
functions.<br />
Communicative<br />
Effect<br />
Grammatical or phonological IL deviations from target language norms have the<br />
"advantage" of being easily recognizable by native speakers. A "nonnative" identification<br />
also serves to protect such speakers from the risk of the peculiarities of their<br />
speech being attributed to flaws in their personality, or ethnocultural origins. Higher<br />
levels of L2 proficiency allow for ease of communication, but still leave open the
12 Interlanguage Pragmatics: An Introduction<br />
possibility of pragmatic failure at the pragmalinguistic and sociopragmatic level<br />
(Thomas, 1983). Differing from grammatical errors, pragmatic failure is neither<br />
easily recognizable by interlocutors without training in pragmatics, nor explained<br />
away by recognizing the speaker as nonnative. Blum-Kulka and Olshtain (1986)<br />
explain pragmatic failure as linked to cultural variability in the implementation of<br />
Grice's (1975) conversational model: certain types of pragmatic deviations from<br />
target norms, such as the learner's tendency to verbosity (discussed above) are seen<br />
as violations of a cultural norm for the balance required between the maxims of<br />
clarity and quantity. Others (e.g., Riley, 1989) are concerned with defining the<br />
scope of communicative behavior amenable to pragmatic failure, insisting that both<br />
verbal and nonverbal phenomena be included. Interactional sociolinguistics has<br />
provided rich evidence for miscommunication resulting from interlocutors' differences<br />
in the use of contextualization conventions (Gumperz, 1982) and different<br />
conversational styles (Tannen, 1981; 1985). While disturbing in everyday communication,<br />
incompatible styles, resulting in mutual misreadings of the other person's<br />
intention and attitude, can have serious consequences in gate-keeping encounters,<br />
leading to unfavorable outcomes for the client (Erickson & Shultz, 1982; Scollon &<br />
Scollon, 1983).<br />
There have been three major approaches to the study of pragmatic failure.<br />
Miscommunication research as referred to above provides micro-sociolinguistic<br />
analyses of naturalistic encounters, minutely identifying problematic features at the<br />
levels of prosody, pragmatics, syntax, lexis, discourse organization, conversational<br />
management, and nonverbal behavior. Qualitative analysis of performance data is<br />
sometimes combined with quantitative measures of particular features and with<br />
retrospective interviews of the participants. Micro-sociolinguistic analysis ascertains<br />
conversational style differences and identifies instances where such differences<br />
become problematic, but does not usually inquire into the origin of different conversational<br />
styles. A second approach is contrastive pragmatics, involving the crosscultural<br />
and cross-linguistic comparison of speech act realization patterns through<br />
identifying similarities and differences between the pairs or groups of languages<br />
studied. Such research is purely descriptive, having no predictive power for the<br />
study of IL pragmatics and actual communicative practices in cross-cultural encounters<br />
but serving an important hypothesis-generating and explanatory role in studies<br />
of interlanguage pragmatic performance and knowledge. In order to study the<br />
relationship between learners' prior knowledge and pragmatic performance, a third<br />
line of investigation needs to be called upon: ILP. In its canonical form, ILP<br />
research, following received methodology in interlanguage studies (Selinker, 1972)<br />
by comparing learners' IL production and comprehension with parallel LI and L2<br />
data, provides the methodological tool to determine where and how learners' pragmatic<br />
performance differs from L2, and to establish where IL specific behaviors<br />
appear to be influenced by learners' LI knowledge. (As argued above, determining<br />
with any certainty whether or not transfer has actually been operative requires<br />
additional measures, such as retrospective reports and transferability studies.) Furthermore,<br />
ILP can establish which IL-specific pragmatic behaviors reside in sources<br />
other than transfer (e.g., see Kasper, 1981; Blum-Kulka, 1982; Schmidt, 1983;<br />
Olshtain & Cohen, 1989).
Interlanguage Pragmatics: An Introduction 13<br />
However, just as contrastive pragmatic study is unable to identify pragmatic<br />
transfer, learner-focused ILP, unless supplemented by other measures, such as<br />
ratings of learners' performance by native speakers (Fraser et al., 1980; Eisenstein<br />
& Bodman, 1986, Chapter 3), cannot make claims about communicative effect.<br />
"Negative" pragmatic transfer may or may not result in pragmatic failure. Erickson<br />
and Shultz (1982) demonstrate that successful co-membershiping can neutralize<br />
communicative style differences. Tannen (1985) cites examples of "negative" transfer<br />
that lead to positive attributions (reminiscent of "charming" foreign accents),<br />
and styles that, though different, are complementary rather than conflicting, thus<br />
leading to successful outcomes. Clyne (1979) concludes that communication conflict<br />
arises not so much from local differences in linguistic action patterns but from<br />
features that impinge on interlocutors' perceptions of power, trust and solidarity.<br />
While there is a legitimate ecological interest in the identification of miscommunication<br />
and its causal relationship to LI communicative practices, nonnative<br />
communicative styles, whether transfer-induced or not, do not necessarily result in<br />
pragmatic failure. The time-honored contrastive equation "difference = negative<br />
transfer = error" has proved to be just as little true for pragmatics as for other<br />
domains of nonnative language learning and use, though it has an indisputable<br />
heuristic value.<br />
To summarize, the strength of interactional sociolinguistics is identifying pragmatic<br />
failure; of contrastive pragmatics, identifying cross-cultural and crosslinguistic<br />
pragmatic differences and similarities; and of ILP, identifying learnerspecific<br />
pragmatic behaviors and their relationship to learners' LI and L2. A fullfledged<br />
research program that sheds light on the relationship between cross-cultural<br />
differences, IL-specific pragmatic features, including transfer, and communicative<br />
effects will usefully combine methods from all three areas of investigation.<br />
Reflecting the various approaches employed in studying interlanguage pragmatics,<br />
this book is organized into three parts: (I) Cognitive approaches to interlanguage<br />
pragmatic development, (II) Speech act realization, and (III) Discourse perspectives.<br />
For an overview of the chapters, the reader is referred to the introduction to<br />
each part.<br />
References<br />
Bachman L. F. (1990). Fundamental considerations in language testing. Oxford: Oxford<br />
University Press.<br />
Bardovi-Harlig, K., & Hartford, B. S. (1990). Congruence in native and nonnative conversations:<br />
Status balance in the academic advising session. Language Learning, 40, 467-<br />
501.<br />
Beebe, L. M., & Takahashi, T. (1989a). Do you have a bag Social status and patterned<br />
variation in second language acquisition. In S. Gass, C. Madden, D. Preston, &<br />
L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics<br />
(103-25). Clevedon and Philadelphia: Multilingual Matters.<br />
Becbe, L. M., & Takahashi, T. (19895). Sociolinguistic variation in face-threatening speech
14 Interlanguage Pragmatics: An Introduction<br />
acts. In M. Eisenstein (Ed.), The dynamic interlanguage (199-218). New York:<br />
Plenum.<br />
Beebe, L. M., Takahashi, T., & Uliss-Weltz, R. (1990). Pragmatic transfer in ESL refusals.<br />
In R. C. Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative<br />
competence in a second language (55-73). New York: Newbury House.<br />
Bialystok, E. (1990). Communication strategies. Oxford: Basil Blackwell.<br />
Blum-Kulka, S. (1982). Learning how to say what you mean in a second language. Applied<br />
Linguistics, 3, 29-59.<br />
Blum-Kulka, S. (1991). Interlanguage pragmatics: The case of requests. In R. Phillipson,<br />
E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds.), Foreign/second<br />
language pedagogy research (255-72). Clevedon and Philadelphia: Multilingual<br />
Matters.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests<br />
and apologies. Norwood, NJ: Ablex.<br />
Blum-Kulka, S., & Levenston, E. (1987). Lexical-grammatical pragmatic indicators. Studies<br />
in Second Language Acquisition, 9, 155—70.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Studies in Second Language Acquisition, 8, 47-61.<br />
Bodman, J., & Eisenstein, M. (1988). May God increase your bounty: The expression of<br />
gratitude in English by native and nonnative speakers. Cross Currents, 15, 1-21.<br />
Bongaerts, T., Kellerman, E., & Bentlage, A. (1987). Perspective and proficiency in L2<br />
referential communication. Studies in Second Language Acquisition, 9, 171-99.<br />
Bouton, L. (1988). A cross-cultural study of ability to interpret implicatures in English.<br />
World Englishes, 17, 183-96.<br />
Brown, P., & Levinson, S. C. (1987). Politeness. Cambridge: Cambridge University Press.<br />
Carrell, P. L. (1979). Indirect speech acts in ESL: Indirect answers. In C. A. Yorio,<br />
K. Perkins & J. Schachter (Eds.), On TESOL '79 (297-307). Washington, DC:<br />
TESOL.<br />
Carrell, P. L. (1981a). Children's understanding of indirect requests: Comparing child and<br />
adult comprehension. Journal of Child Language, 8, 329-45.<br />
Carrell, P. L. (1981b). Relative difficulty of request forms in L1/L2 comprehension. In<br />
M. Hines & W. Rutherford (Eds.), On TESOL '81 (141-52). Washington, DC:<br />
TESOL.<br />
Carrell, P. L., & Konneker, B. H. (1981). Politeness: Comparing native and nonnative<br />
judgments. Language Learning, 31, 17-31.<br />
Clyne, M. (1979). Communicative competences in contact. ITL, 43, 17-37.<br />
Clyne, M., Ball, M., & Neil, D. (1991). Intercultural communication at work in Australia:<br />
Complaints and apologies in turns. Multilingua, JO, 251-73.<br />
Cohen, A. D., & Olshtain, E. (1991, October). The production of speech acts by nonnatives.<br />
Paper presented at the Theory construction and research methodology in second<br />
language acquisition conference, Lansing, MI.<br />
Coulmas, F. (1981). "Poison to your soul." Thanks and apologies contrastively viewed. In<br />
F. Coulmas (Ed.), Conversational routine (273—88). The Hague: Mouton.<br />
Edmondson, W., & House, J. (1991). Do learners talk too much The waffle phenomenon in<br />
interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood<br />
Smith, & M. Swain (Eds.), Foreign!second language pedagogy research (273—86).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by<br />
native and non-native speakers of American English. Applied Linguistics, 7, 167-85.
Inlerlanguage Pragmatics: An Introduction 15<br />
Ellis, R. (1992). Learning to communicate in the classroom. Studies in Second Language<br />
Acquisition, 14, 1-23.<br />
Erickson, F., & Shultz, J. (1982). The counselor as gatekeeper: Social interaction in interviews.<br />
New York: Academic Press.<br />
Ervin-Tripp, S., Stragc, A., Lampert, M, & Bell, N. (1987). Understanding requests.<br />
Linguistics, 25, 107-43.<br />
Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request<br />
realization. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics<br />
(221 -47). Norwood, NJ: Ablex.<br />
Fiksdal, S. (1990). The right time and pace: A microanalysis of cross-cultural Datekeeping<br />
interviews. Norwood, NJ: Ablex.<br />
Fraser, B., Rintell, E., & Walters, J. (1980). An approach to conducting research on the<br />
acquisition of pragmatic competence in a second language. In D. Larsen-Freeman<br />
(Ed.), Discourse analysis in second language research (75-91). Rowley, MA: Newbury<br />
House.<br />
Fukushima, S. (1990). Offers and requests: Performance by Japanese learners of English.<br />
World Englishes, 9, 317-25.<br />
Garcia, C. (1989). Apologizing in English. Politeness strategies used by native and nonnative<br />
speakers. Multilingua, 8, 3-20.<br />
Gibbs, R. W. (1983). Do people always process the literal meaning of indirect requests<br />
Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 524-33.<br />
Gibbs, R. W. (1984). Literal meaning and psychological theory. Cognitive Science, 8, 275-<br />
304.<br />
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and<br />
semantics. (Vol. 3): Speech acts (41-58). New York: Academic Press.<br />
Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press.<br />
House, J. (1988). "Oh excuse me please . . .": Apologizing in a foreign language. In<br />
B. Kettemann, P. Bierbaumer, A. Fill, & A. Karpf (Eds.), Englisch alsZweitsprache<br />
(303-27). Tubingen: Narr.<br />
House, J., & Kasper, G. (1987). Interlanguage pragmatics: Requesting in a foreign language.<br />
In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance.<br />
Festschrift for Werner Mullen (1250-88). Tubingen: Narr.<br />
Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr.<br />
Kasper, G. (1989a). Interactive procedures in interlanguage discourse. In W. Oleksy (Ed.),<br />
Contrastive pragmatics (189-229). Amsterdam: Benjamins.<br />
Kasper, G. (1989b). Variation in interlanguage speech act realization. In S. Gass, C. Madden,<br />
D. Preston, & L. Selinker (Eds.), Variation in second language acquisition:<br />
Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.<br />
Kasper, G., & Dahl, M. (1991). Research methods in interlanguage pragmatics. Studies in<br />
Second Language Acquisition, 13, 215-47.<br />
Kasper, G., & Schmidt, R. (1992, March). Interlanguage pragmatics and language learning.<br />
Paper presented at the Sociolinguistics and TESOL Colloquium. TESOL Congress,<br />
Vancouver.<br />
Kitao, K. (1990). A study of Japanese and American perceptions of politeness in requests.<br />
Doshida Studies in English, 50, 178-210.<br />
Koike, D. A. (1989). Pragmatic competence and adult L2 acquisition: Speech acts in interlanguage.<br />
Modern Language Journal, 73, 79-89.<br />
Leech, G. N. (1983). Principles oj pragmatics. London: Longman.<br />
Lcvinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press.
16 Interlanguage Pragmatics: An Introduction<br />
Maeshiba, N., Yoshinaga, N., Kaspcr, G., & Ross, S. (in press). Transfer and proficiency in<br />
interlanguage apologizing. In J. Neu & S. Gass (Eds.), Speech acts across cultures.<br />
Berlin: Mouton de Gruyter.<br />
Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In<br />
S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49).<br />
Rowley, MA: Newbury House.<br />
Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to<br />
native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language<br />
acquisition (303-25). Rowley, MA: Newbury House.<br />
Olshtain, E., & Cohen, A. (1989). Speech act behavior across languages. In H. W. Dechert<br />
& M. Raupach (Eds.), Transfer in language production (53-67). Norwood, NJ:<br />
Ablex.<br />
Poulisse, N. (1990). Compensatory strategies in L2 production. Amsterdam: Foris.<br />
Reeder, K., & Wakefield, J. (1987). The development of young children's speech act comprehension:<br />
How much language is necessary Applied Psycholinguistics, 8, 1-18.<br />
Riley, P. (1989). "Well, don't blame me": On the interpretation of pragmatic errors. In<br />
W. Olesky (Ed.), Contrastive pragmatics (231-51). Amsterdam: Benjamins.<br />
Rintell E. (1981). Sociolinguistic variation and pragmatic ability: A look at learners. International<br />
Journal of the Sociology of Language, 27, 11-34.<br />
Rintell, E., & Mitchell, C. (1989). Studying requests and apologies: An inquiry into method.<br />
In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics (248-<br />
72). Norwood, NJ: Ablex.<br />
Robinson, M. A. (1992). Introspective methodology in interlanguage pragmatics research. In<br />
G. Kasper (Ed.), Pragmatics of Japanese as a native and foreign language. Technical<br />
Report No. 3 (27-82). Second Language Teaching & Curriculum Center, University<br />
of Hawaii at Manoa.<br />
Sawyer, M. (1992). The development of pragmatics in Japanese as a second language: The<br />
sentence-final particle ne. In G. Kasper (Ed.), Pragmatics of Japanese as a native and<br />
foreign language. Technical Report No. 3 (83-125), Second Language Teaching &<br />
Curriculum Center, University of Hawaii at Manoa.<br />
Scarcella, R. (1979). On speaking politely in a second language. In C. A. Yorio, K. Perkins,<br />
& J. Schachter (Eds.), On TESOL '79 (275-87). Washington, DC: TESOL.<br />
Schmidt, R. (1983). Interaction, acculturation and the acquisition of communicative competence.<br />
In N. Wolfson & E. Judd (Eds.), Sociolinguistics and second language acquisition<br />
(137-74). Rowley, MA: Newbury House.<br />
Scollon, R., & Scollon, S.B.K. (1983). Face in interethnic communication. In J. C. Richards<br />
& R. Schmidt (Eds.), Language and communication (156-88). London: Longman.<br />
Selinker, L. (1972). Interlanguage. 1RAL 10, 209-30.<br />
Svanes, B. (1989). En undersoekelse av realisasjonsmoensteret for spraakhandlingen "aa be<br />
noen om aa gjoere noe." Maal og Minne, 1-2, 89-107.<br />
Takahashi, S. (1990). Exploring the comprehension process of nonliteral utterances and some<br />
implications for automaticity. University of Hawaii Working Papers in ESL, 9, 67-97.<br />
Takahashi, S. (1992). Transferability of indirect request strategies. University of Hawaii<br />
Working Papers in ESL, 11, 69-124.<br />
Takahashi, S., & Roitblat, H. L. (1992, July). Comprehension of'L2 indirect requests. Paper<br />
presented at Pacific Second Language Research Forum, Sydney.<br />
Takahashi, S., & DuFon, P. (1989). Cross-linguistic influence in indirectness: The case of<br />
English directives performed by native Japanese speakers. Unpublished manuscript,<br />
University of Hawaii at Manoa.
Interlanguage Pragmatics: An Introduction 17<br />
Takahashi, T., & Beebe, L. M. (1987). The development of pragmatic competence by<br />
Japanese learners of English. JALT Journal, 8, 131-55.<br />
Tanaka, N. (1988). Politeness: Some problems for Japanese speakers of English. JALT<br />
Journal, 9, 81-102.<br />
Tanaka, S., & Kawade, S. (1982). Politeness strategies and second language acquisition.<br />
Studies in Second Language Acquisition, 5, 18—33.<br />
Tannen, D. (1981). Indirectness in discourse: Ethnicity as conversational style. Discourse<br />
Processes, 4, 221-38.<br />
Tannen, D. (1985). Cross-cultural communication. In T. A. Van Dijk (Ed.), Handbook of<br />
discourse analysis (Vol. 4, 203-15). London: Academic Press.<br />
Tarone, E., & Yule, G. (1987). Communication strategies in East-West interactions. In<br />
L. Smith (Ed.), Discourse across cultures (49-65). New York: Prentice-Hall.<br />
Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.<br />
Trosborg, A. (1987). Apology strategies in natives/nonnatives. Journal of Pragmatics, 11,<br />
147-67.<br />
Walters, J. (1979). The perception of deference in English and Spanish. In C. A. Yorio, K.<br />
Perkins, & J. Schachter (Eds.), On TESOL '79 (288-96). Washington, DC: TESOL.<br />
Wildner-Bassett, M. (1984). Improving pragmatic aspects of learners' interlanguage.<br />
Tubingen: Narr.<br />
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House.<br />
Yule, G., & Tarone, E. (1990). Eliciting the performance of strategic competence. In R. C.<br />
Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative competence<br />
(179-94). New York: Newbury House.
This page intentionally left blank
I<br />
COGNITIVE APPROACHES<br />
TO <strong>INTERLANG</strong>UAGE<br />
PRAGMATIC DEVELOPMENT<br />
Although there is a comprehensive literature on children's development of pragmatic<br />
competence, little work has been done on the acquisition of pragmatic ability by<br />
adult second language learners. Schmidt (Chapter 1) and Bialystok (Chapter 2)<br />
discuss two central issues in adult pragmatic learning, each proposing a different<br />
theoretical framework to address the problem. Schmidt explores the role of conscious<br />
awareness in the acquisition of pragmatic competence. Based on a critical<br />
examination of recent work in experimental psychology and support from diary<br />
reports, Schmidt concludes that the necessary condition for pragmatic learning to<br />
take place is attention to the pragmalinguistic or sociopragmatic information to be<br />
acquired. While implicit and incidental learning seems possible, noticing and generalizing<br />
about relevant features in the input is highly facilitative. Schmidt's analysis<br />
has important implications for research methodology in pragmatics and language<br />
teaching. Theory and cited illustrations encourage the use of self-report as a method<br />
of data collection in interlanguage pragmatics. For second language instruction,<br />
Schmidt's analysis indicates activities that alert the learner to pragmalinguistic and<br />
sociopragmatic features as well as a consciousness-raising approach to the teaching<br />
of L2 pragmatics.<br />
Bialystok examines the learning task in the pragmatic domain of adult second<br />
language learners as opposed to child first language acquirers. Applying her twodimensional<br />
model of language learning and use to pragmatics, she concludes that<br />
the learning problem for children acquiring their first language and adults learning a<br />
second language is quite different. Children primarily need to develop an analyzed<br />
knowledge of form-function mappings. For them, acquiring pragmalinguistic resources<br />
and their contextual distribution patterns takes precedence over the need to<br />
develop the control strategies required for efficient use of pragmatic knowledge.<br />
With adults, the order of learning tasks is reversed. While they also have to acquire<br />
new pragmalinguistic knowledge, their primary problem appears to be one of control,<br />
increasing their ability to process pragmatic information smoothly in contexts<br />
and making socially and contextually appropriate selections of linguistic forms.<br />
19
20 Cognitive Approaches to Interlanguage Pragmatic Development<br />
Schmidt's and Bialystok's proposals offer different but compatible approaches to<br />
pragmatic interlanguage development: Schmidt is concerned with the conditions for<br />
initial intake; Bialystok addresses the cognitive dimensions on which pragmatic<br />
interlanguage competence evolves. Both approaches, we hope, will inspire databased<br />
studies of adult pragmatic learning.
1<br />
Consciousness, Learning<br />
and Interlanguage Pragmatics<br />
RICHARD SCHMIDT<br />
During the past decade, the study of interlanguage pragmatics has produced important<br />
empirical findings, primarily through the identification and comparison of<br />
speech act realization patterns in various languages based on data from both native<br />
and nonnative speakers. In addition to this focus on product, some attention has<br />
been paid to the processes of comprehension and production in second language<br />
pragmatics (Faerch & Kasper, 1984, 1989; Kasper, 1984). In contrast to these<br />
concerns, there has been little discussion of how pragmatic abilities are acquired in<br />
a second language.<br />
This chapter is concerned with the ways in which consciousness may be involved<br />
in learning the principles of discourse and pragmatics in a second language.'<br />
The role of conscious and nonconscious processes in the acquisition of morphosyntax<br />
has been hotly debated within the field of second language acquisition (Krashen,<br />
1981, 1983; Munsell & Carr, 1981; Rutherford & Sharwood Smith, 1985; Seliger,<br />
1983; Sharwood Smith, 1981), but these debates have ignored pragmatic and discoursal<br />
abilities. My discussion will of necessity be speculative, drawing on current<br />
theories of the role of consciousness in human learning in general, drawn primarily<br />
from cognitive science and experimental psychology, with some suggestions for the<br />
extension of general principles to the learning of pragmatics. This is an issue with<br />
important pedagogical implications. In second language teaching, as Richards<br />
(1990) points out, there are currently two major approaches to the teaching of<br />
conversation in second language programs. The first is an indirect approach, in<br />
which conversational competence is seen as the product of engaging learners in<br />
conversational interaction; the underlying assumption is that the ability to carry on<br />
conversation (which includes pragmatic ability and other factors as well) is something<br />
that is acquired simply in the course of doing it. In practice, this leads to the<br />
use of group work activities or other tasks that require interaction. The second, a<br />
more direct approach, focuses explicitly on the strategies involved in conversation<br />
and emphasizes consciousness-raising concerning these strategies.<br />
21
22 Cognitive Approaches to Interlanguage Pragmatic Development<br />
Is Pragmatic Knowledge Conscious or Unconscious<br />
Wolfson has argued that native speaker knowledge of what she calls rules of speaking<br />
(which include both pragmatic and discoursal rules) is mostly unconscious:<br />
Rules of speaking and, more generally, norms of interaction are ... largely<br />
unconscious. What this means is that native speakers, although perfectly competent<br />
in the uses and interpretation of the patterns of speech behavior which prevail<br />
in their own communities are, with the exception of a few explicitly taught<br />
formulas, not even aware of the patterned nature of their speech behavior. [Native<br />
speakers] . . . are not able . . . to describe their own rules of speaking. (Wolfson,<br />
1989, 37)<br />
Wolfson cites several types of evidence in support of her claim that speakers do<br />
not have reliable information concerning the ways in which they use language:<br />
people who are bilingual or bidialectal may switch from one language or variety to<br />
another without being aware of it and cannot accurately report their use of these<br />
languages or varieties (Blom & Gumperz, 1972); native speakers often report that<br />
they typically use or do not use specific forms, but their descriptions do not match<br />
reality (Wolfson, D'Amico-Reisner, & Huber, 1983); even highly trained linguists<br />
who rely on intuition to describe such phenomena as the differences between men's<br />
and women's speech (e.g., Lakoff, 1973) may find their intuitions proven incorrect;<br />
textbook writers, who almost always rely on intuition rather than empirical data,<br />
provide information regarding language use that is frequently wrong (Cathcart,<br />
1989; Holmes, 1988; Williams, 1988).<br />
There are several reasons why we should expect native speakers' intuitions<br />
about these matters to be fallible. First, there is the obvious problem of the intrusion<br />
of prescriptive norms, stereotypes, and folk-linguistic beliefs; when asked what they<br />
do, informants are likely to report what they think they should do. Second, this kind<br />
of introspection violates basic principles distinguishing between potentially accurate<br />
and inaccurate verbal reports (Ericsson & Simon, 1984; Nisbett & Wilson, 1977),<br />
because such intuitions are general rather than specific, retrospective rather than<br />
concurrent, and sometimes call for information that could not be reported even if the<br />
other conditions were met. Ericsson and Simon (1984) propose that the only information<br />
that is potentially available for accurate self-report is information that is<br />
attended to in short-term memory in the performance of a task. In other words, in<br />
order to give an accurate report of your own performance, you must have been<br />
paying attention and aware of what you were doing at the time. Speech act realizations<br />
and other aspects of rules of speaking are often produced by fluent speakers<br />
with little conscious reflection or deliberation during their performance, and are<br />
therefore not accurately reportable. If accurate self-reports are limited to reporting<br />
information that has been stored as a result of one's own conscious thought processes,<br />
intuitions about the linguistic behavior of groups are particularly suspect<br />
(Cameron, 1985).<br />
The evidence cited by Wolfson (1989) shows that native speakers do not necessarily<br />
have access to their own rules of speaking, but it fails to show that speakers<br />
never have any access to such rules. Blum-Kulka (Chapter 10) and Olshtain and
Consciousness, Learning, and Interlanguage Pragmatics 23<br />
Blum-Kulka (1989) have argued that Hebrew-English bilinguals in Israel exhibit<br />
heightened metapragmatic awareness and are aware of their code-switching behavior.<br />
Odlin suggests that linguistic forms that are important for communicative competence<br />
are, in general, highly salient and accessible to awareness, which may be<br />
why the metalanguage observed in anthropological linguistics tends to describe<br />
linguistic functions more accurately than linguistic form (Odlin, 1986). The fact that<br />
communicative behavior is sometimes accurately reportable is also compatible with<br />
the principle that accurate self-report depends on information that is attended to<br />
during performance. Pragmatic and discoursal knowledge is not always used automatically<br />
and unreflectively. Conversations vary a great deal in terms of spontaneity<br />
and planning (Ochs, 1979). Some people preplan telephone conversations, and<br />
writing involves a great deal of conscious deliberation and choices in discourse<br />
organization. There are many occasions on which particular care is given to producing<br />
appropriately polite language. Students may worry about how to address professors,<br />
and many aspects of the use of personal address are not unreflecting<br />
responses to a determining context but represent strategic and sometimes manipulative<br />
choices (Kendall, 1981).<br />
Pragmatic knowledge therefore seems to be partly conscious and partly accessible<br />
to consciousness, although it cannot be the case that all pragmatic knowledge is<br />
accessible to consciousness. Just as linguists seek to discover general principles of<br />
language that are reflected in the effortless control of grammar by native speakers<br />
but of which they have no conscious awareness, research in pragmatics seeks to<br />
identify patterns and general principles that native speakers are equally unable to<br />
articulate based on introspection. However, even if a great deal of pragmatic knowledge<br />
is held implicitly and cannot be articulated, this does not tell us how such<br />
knowledge was established. Skillful performance that currently relies on automatic<br />
processing and makes little demand on either attention or consciousness may have<br />
originated from conscious declarative knowledge (Lewis & Anderson, 1985). General<br />
principles, patterns, and rules of pragmatics may be beyond the reach of<br />
introspection, but this does not inform us of the possible role that awareness of<br />
crucial features of language rules, however incomplete and transitory, may play in<br />
the establishment of such knowledge (Munsell & Carr, 1981).<br />
Consciousness and Principles of Language Learning<br />
Our ordinary language use of words like conscious, consciousness and consciously<br />
is ambiguous. This is one reason why theorists in psychology and applied linguistics<br />
have preferred to use related technical terms such as explicit versus implicit knowledge<br />
(Bialystok, 1979, 1981;Krashen, 1981; Odlin, 1986; Sharwood Smith, 1981),<br />
controlled versus automatic processing (Bialystok, Chapter 2; Bialystok &<br />
Bouchard-Ryan, 1985; Carroll, 1981; McLaughlin, Rossman & McLeod, 1983;<br />
Posner & Klein, 1973; Shiffrin & Schneider, 1977); declarative versus procedural<br />
knowledge (Anderson, 1982; Ellis, 1989b; Faerch & Kasper, 1984; O'Malley,<br />
Chamot & Walker, 1987), serial versus parallel processing (McClelland,<br />
Rumelhart, & the POP Research Group, 1986), and so on. Unfortunately, the use of
24 Cognitive Approaches to Interlanguage Pragmatic Development<br />
technical terms does not by itself eliminate the ambiguities. Odlin (1986) has<br />
discussed the various ways in which the contrast between explicit and implicit<br />
knowledge has been understood, and Norman and Shallice (1986) have identified<br />
ambiguities inherent in the concept of automatic processing, some of which are<br />
exact parallels to the ambiguities of consciousness. Since a great deal of debate<br />
about conscious and unconscious processes has been fueled by conceptual and<br />
definitional disagreements (Bowers, 1984; White, 1980), it is preferable to grapple<br />
with these issues directly, rather than masking them with alternative terms.<br />
It seems to me that when we speak of having been conscious of something, we<br />
most often mean that we were aware of it, that we subjectively experienced it as part<br />
of the "stream" of consciousness (Battista, 1978; James, 1890; Natsoulis, 1987).<br />
However, when we speak of having done something consciously, we may mean<br />
either that we did it with awareness of what we were doing or that we did it<br />
deliberately. This is one of the main ambiguities involved in most discussions of<br />
consciousness: consciousness as awareness versus consciousness as intent (Ceci &<br />
Howe, 1982). When we speak of consciousness as awareness, there is also a<br />
question of the degree or level of our awareness. We may mean that we simply<br />
noticed the occurrence of something or that we had a more abstract understanding of<br />
it (Bowers, 1984). Therefore, when we speak of language learning as being conscious<br />
or unconscious, we might be thinking of several distinct aspects of the<br />
problem of consciousness in learning, including at least the following: whether a<br />
learner is trying to learn something; whether the learner is aware that he or she is<br />
learning; whether the target language forms that are learned are consciously noticed<br />
or picked up through some kind of subliminal perception; whether learners acquire<br />
general rules or principles on the basis of conscious understanding and insight or<br />
more intuitively; or whether learners are able to give an accurate account of the rules<br />
and principles that seem to underlie the construction of utterances.<br />
There is experimentally based literature from psychology that bears on all of<br />
these issues, along with a small amount of evidence from second language acquisition<br />
studies. It is useful to summarize the relevant research in terms of three<br />
principal distinctions.<br />
Conscious Perception versus Subliminal Influences in Learning<br />
My personal choice of a label for the key concept here is noticing, although there<br />
are a variety of technical terms for this, including focal awareness (Atkinson &<br />
Shiffrin, 1968), episodic awareness (Allport, 1979), conscious perception (Dixon,<br />
1971) and apperceived input (Gass, 1988). Each of these constructs presupposes the<br />
allocation of attentional resources to some stimulus and identifies the level at which<br />
perceived events are subjectively experienced and are reportable by the person who<br />
experiences them. 2<br />
Events may remain unnoticed for several reasons—because attention is directed<br />
elsewhere, because the information is too complex to be processed, or because it is<br />
presented too quickly or too softly to be consciously seen or heard. While it is<br />
virtually impossible when observing naturalistic language learning to know exactly
Consciousness, Learning, and Intel-language Pragmatics 25<br />
what the learner has or has not noticed, the existence of unnoticed information can<br />
be established under experimental conditions by the failure of subjects to report their<br />
awareness of a stimulus if asked immediately following its presentation. This criterion<br />
of subjective awareness can be contrasted with an objective measure of perception,<br />
which various experimenters have argued is best established by a subject's<br />
ability to discriminate among two or more alternative stimuli in a forced choice task<br />
(Cheesman & Merikle, 1986; Eriksen, 1960; Moore, 1988).<br />
Although many theorists believe that unconscious learning (in some sense)<br />
predominates in second language learning, it is very unlikely that what language<br />
learners consciously perceive or notice in input is unimportant for learning. A more<br />
difficult question is whether it is necessary to notice what is said in a language in<br />
order for that information to be stored in memory and to play a role in language<br />
learning, or whether it is also possible for some learning to be based on unnoticed<br />
information, information that is perceived at some level and perhaps processed<br />
subliminally without being consciously registered.<br />
There is a widespread belief (at least in North America) that the existence of<br />
subliminal learning of some kind has been established for decades. In the 1950s,<br />
Packard objected to the covert manipulation of consumers through the use of subliminal<br />
messages in advertising (Packard 1957), a theme expanded upon by Key<br />
(1973). Beginning in the 1980s, subliminal audiocassettes were aggressively marketed<br />
that promised everything from cures for obesity and drug addiction to enhanced<br />
visual acuity, improvement in examination performance, and more effective<br />
language learning. However, there seems to be virtually no scientific support for<br />
claims of behavior modification through subliminal messages. Moore has reviewed<br />
the research on subliminal techniques in advertising, concluding that the advertising<br />
stories everyone has heard about (such as the stimulation of movie theater patrons to<br />
buy popcorn or softdrinks through subliminal messages) are apocryphal. Such techniques<br />
probably never were used, and even if they were, "there is no evidence that<br />
subliminal messages can influence motivation or complex behavior" (Moore, 1988,<br />
293). Merikle has examined commercially distributed "subliminal" audiotapes and<br />
subjected them to both psychophysical experimentation and spectographic analysis,<br />
reporting that the cassettes analyzed contained no embedded subliminal messages<br />
whatsoever that could conceivably influence behavior (Merikle, 1988, 355).<br />
There is a well-attested phenomenon of subliminal perception. Stimuli that are<br />
presented too rapidly for conscious detection or in competition with tasks that are<br />
assumed to consume all attentional resources may activate existing memory structures<br />
and associations (Dixon, 1971, 1981; Marcel, 1983). Eich (1984) has reported<br />
experiments in which pairs of words were both presented to the unattended channel<br />
in a shadowing task, one of which was ambiguous (e.g., fair or fare), while the<br />
other word biased its less common interpretation (e.g., taxi). Recognition of both<br />
members of such pairs was poor but in a spelling test subjects were biased in the<br />
direction of the disambiguated meaning. These and other similar demonstrations<br />
show that words that are not consciously perceived or noticed can be processed to<br />
the level of word meaning. However, all demonstrations of subliminal perception so<br />
far have involved subtle effects resulting from the unconscious detection and pro-
26 Cognitive Approaches to Interlanguage Pragmatic Development<br />
cessing of very familiar stimuli. Such effects do not imply the creation of new<br />
memory structures, the establishment of new associations, or the learning of new<br />
concepts (Ericsson & Simon, 1984; Underwood, 1976, 1982), and certainly nothing<br />
remotely analogous to learning a second language.<br />
At the present time, the available evidence is compatible with the strong assertion<br />
that there is no such thing as subliminal language learning or any other kind of<br />
subliminal learning. Second language forms that are not noticed do not affect<br />
learning. This allows the concept of intake in second language learning to be<br />
defined in terms of what the learner attends to and notices (Schmidt, 1990).<br />
Explicit versus Implicit Learning<br />
The contrast between subliminal learning and implicit learning, or learning without<br />
understanding, has to do with the level of awareness involved. I use noticing to<br />
mean registering the simple occurrence of some event, whereas understanding<br />
implies recognition of a general principle, rule, or pattern. For example, a second<br />
language learner might simply notice that a native speaker used a particular form of<br />
address on a particular occasion, or at a deeper level the learner might understand<br />
the significance of such a form, realizing that the form used was appropriate because<br />
of status differences between speaker and hearer. Noticing is crucially related<br />
to the question of what linguistic material is stored in memory (Atkinson & Shiffrin<br />
1968; Kihlstrom, 1984); understanding relates to questions concerning how that<br />
material is organized into a linguistic system.<br />
Implicit learning refers to nonconscious generalization from examples. The<br />
general phenomenon of implicit learning has been well established in the psychological<br />
literature and is viewed as a natural product of attending to structured input<br />
(Hartman, Knopman, & Nissen, 1989; Reber, 1989). There is a gathering consensus<br />
within psychology that the mechanisms of implicit learning probably involve the<br />
strengthening and weakening of connections between nodes in complex networks as<br />
the result of experience, rather than through the unconscious induction of rules<br />
abstracted from data. An example of this recent shift in perspective can be seen in the<br />
work of Reber, who has carried out numerous experiments involving exposing<br />
subjects to strings of letters generated by an artificial grammar. After training,<br />
subjects were able to make accurate judgments about the well-formedness of novel<br />
strings, without being able to articulate the rules of well-formedness (Reber, 1976;<br />
Reber, Allen & Regan, 1985; Reber, Kassin, Lewis & Cantor, 1980). Until recently,<br />
Reber (1976) argued that knowledge resulting from implicit learning was encoded in<br />
the form of unconscious abstract representations. In a more recent publication,<br />
Abrams and Reber (1988) have suggested that implicit learning as demonstrated in<br />
these experiments probably rests upon some kind of covariation counter, a system that<br />
logs both event frequencies and event co-occurrences. One model that simulates the<br />
mechanisms currently believed to underlie implicit learning is Parallel Distributed<br />
Processing (PDP). PDP has been used to model the acquisition of the German definite<br />
article (MacWhinney et al., 1989), the past tense in English (Rumelhart & McClelland,<br />
1986), the development of visual word recognition skills (Seidenberg & Mc-<br />
Clelland, 1989), and the acquisition of gender in French (Sokolik & Smith, 1989).
Consciousness, Learning, and Interlanguage Pragmatics 27<br />
Explicit learning, that is, conscious problem solving, relies on different mechanisms,<br />
including attempts to form mental representations, searching memory for<br />
related knowledge and forming and testing hypotheses (Mathews, Buss, Stanley,<br />
Blanchard-Field, Cho & Druhan, 1989; Johnson-Laird, 1983). Both implicit learning<br />
and explicit learning have particular strengths. Implicit learning appears to be<br />
superior for the learning of fuzzy patterns based on perceptual similarities and the<br />
detection of nonsalient covariance between variables, while explicit learning is<br />
superior when a domain contains rules that are based on logical relationships rather<br />
than perceptual similarities (Mathews et al., 1989).<br />
Intentional versus Incidental Learning<br />
Whereas the concepts of subliminal and implicit learning are both related to the<br />
consciousness as awareness, incidental learning refers to consciousness as intent. If,<br />
as I have claimed, it is necessary to notice the occurrence of linguistic forms in order<br />
for them to serve as intake for learning, is it also necessary to deliberately pay<br />
attention to such features in order to notice them More generally, is it necessary to<br />
want to learn in order to learn<br />
This is not so difficult a question as the others I have raised. In many cases, it<br />
does not matter if a language learner intends to pay attention or not. A language<br />
learner's limited processing abilities may make it impossible to notice something<br />
regardless of an intent to do so. There are other cases in which some task to be<br />
performed forces the learner's attention to be focused on some pieces of information<br />
rather than others, and in such cases, what is stored in memory is the information<br />
that must be attended to in order to complete the task (Ericsson & Simon, 1984); the<br />
learner's intention to learn is irrelevant (Anderson, 1985). On the other hand, there<br />
are many situations in which a language learner is free to opt in and out of learning<br />
contexts and to pay attention or not, depending on one's personal hierarchy of deep<br />
goals and momentary dispositions (Baars, 1988; Kahneman, 1973; Kihlstrom,<br />
1984); in such cases paying attention is crucial.<br />
Extensions to the Learning of Pragmatics and Discoursal Rules<br />
I have argued that linguistic forms can serve as intake for language learning only if<br />
they are noticed by learners; that paying attention to such forms is certainly helpful,<br />
but not necessary if other factors in the learning context focus attention on them so<br />
that they are noticed; and that general principles of the organization of language may<br />
be discovered through the use of either explicit or implicit learning mechanisms. I<br />
have also suggested that even in cases where what native speakers "know" about the<br />
pragmatic principles of their language is inaccessible to consciousness, such knowledge<br />
may nevertheless be based on insights and understanding at the time of learning.<br />
What evidence is there that these claims are relevant for the learning of pragmatics
28 Cognitive Approaches to Interlanguage Pragmatic Development<br />
First Language Learning of Pragmatics<br />
Research on the acquisition of first language pragmatics suggests that both noticing<br />
and some level of understanding are important in such learning. Clark (1978) has<br />
observed that the types of metalinguistic abilities shown by preschool children are<br />
primarily related to communicative interaction rather than grammatical form. The<br />
ethnographic literature on language socialization shows that an important childrearing<br />
goal is to develop the child's communicative competence. Demuth (1986)<br />
has reported on the prompting routines for appropriate verbal behavior that play an<br />
active role in the social development of Basotho children. Ochs (1986) has described<br />
the ways in which Samoan caregivers use prosodic strategies for teaching<br />
children how to encode affect-laden utterances. Clancy (1986) has shown how<br />
Japanese mothers interweave questions and declarative hints to socialize children in<br />
the use of indirectness. Watson-Gegeo and Gegeo (1986) describes how Kawara'ae<br />
caregivers use repeating routines to teach children what to say and when to say it.<br />
These and many similar reports suggest that while parents and other caregivers use<br />
different socialization strategies in different cultures, there is probably universal<br />
validity to the observation of Gleason and Perlmann:<br />
Unlike the acquisition of syntax, semantics, and even some sociolinguistic rules,<br />
when it conies to speaking politely adults do not leave it to the child to construct<br />
the rules on his or her own. Here, they take an active, even energetic part in<br />
directly instructing their children in the use of the various politeness devices.<br />
(Gleason & Perlmann, 1985, 102)<br />
Snow, Perlmann, Gleason and Hooshyar (1990) have examined parent-child<br />
interactions in 110 families in order to see what kinds of information concerning<br />
politeness strategies are made available to children from their interaction with<br />
parents. Assuming that the basic dimensions of power, social distance, and degree<br />
of imposition underlie the general rule system for politeness, Snow et al. looked for<br />
evidence for three types of information that might be made available to children:<br />
direct teaching of general rules of politeness, manipulation of the dimensions of<br />
politeness so that the relevant covariations were made more salient, and information<br />
about the use of specific forms. Snow et al. found that the first type of information<br />
was rare, but that there was plentiful evidence in their data that children were<br />
explicitly told what forms to use in particular situations, and that correlations<br />
between forms and the dimensions of politeness were made salient in interaction.<br />
These findings suggest (though they do not prove this point) that children are not<br />
only exposed to but also notice surface forms. Children are also presented with<br />
information that could be used to induce more general principles (through either<br />
implicit or explicit learning mechanisms), but are not taught the underlying principles<br />
directly.<br />
Second Language Learning of Pragmatics<br />
Since adults can report their understandings much more readily than children, it<br />
ought to be possible to examine the role of noticing and understanding in the<br />
development of pragmatic ability by adult second language learners directly, by
Consciousness, Learning, and Interlanguage Pragmatics 29<br />
asking learners to report their experiences. Even so, the relevant data are difficult to<br />
obtain, requiring both a sound methodology for eliciting self-reports (Faerch &<br />
Kasper, 1987) and opportunities to catch learners in the actual process of learning,<br />
rather than simply performing their current competence. Unfortunately, there have<br />
been few studies of any aspect of the phenomenology of second language learning<br />
and no studies at all which have attempted systematically to ascertain what learners<br />
have been conscious of as pragmatic principles were learned.<br />
Anecdotally, there is evidence for a relationship between what learners notice<br />
and understand about pragmatics and discourse and what is learned. The following<br />
six examples are from my own experience, either as a language learner or from<br />
interacting in English with second language speakers with different linguistic and<br />
cultural backgrounds; they concern interactions about which I wrote brief notes to<br />
myself shortly after they happened. The first four examples represent the coincidence<br />
of recognition and insight with rapid learning; the last two represent instances<br />
of less successful learning.<br />
(1) In the course of a 22-week stay in Brazil, during which I progressed from no<br />
proficiency at all in Portuguese to the S-2 level on the FSI scale (see Schmidt &<br />
Frota, 1986, for details), I kept a language learner's diary. Several entries illustrate<br />
the phenomenon of being told about some aspect of the pragmatics of Brazilian<br />
Portuguese in class and then almost immediately noticing it in input, such as the<br />
following:<br />
Journal entry, Week 6. This week we were introduced to and drilled on the<br />
imperfect. . . . The basic contrast seems straightforward enough: ontem eufui ao<br />
clube ["yesterday 1 went to the club"] vs. antigamente eu ia ao clube ["formerly I<br />
used to go to the club"]. L gave us a third model: ontem eu ia ao clube ["yesterday<br />
I was going to the club"], which L says is a common way of making excuses. . . .<br />
Wednesday night Amos came over to play cards, and the first thing he said was eu<br />
ia telefonar para voce ["I was going to call you"], exactly the kind of excuse L<br />
had said we could expect.<br />
(2) I noted in my diary several times the difficulties I had with telephone<br />
conversations, especially in knowing when and how to end a conversation (Schmidt<br />
& Frota, 1986, 276). I knew that with friends the closing move would be for both<br />
parties to say ciao, but I could never identify the point at which I could say it, so I<br />
would often stand holding the phone waiting patiently for the other person to say it<br />
first. Finally, during the last week of my stay, a friend came to my apartment and<br />
used my telephone to make several calls. I listened carefully, and noticed that in two<br />
successive calls, shortly before saying ciao, my friend said the phrase entdo td,<br />
which means no more than "so, then." Suspecting that this might be a preclosing<br />
formula, I immediately called another friend and after a few minutes of talk, said<br />
entdo td, paused briefly and plunged ahead with ciao in the same turn. It worked,<br />
and after that I had no trouble at all getting off the phone efficiently. I subsequently<br />
asked several native speakers how to close a telephone conversation. None could<br />
tell me, but when 1 suggested the use of entdo td, they agreed that was right.<br />
(3) Midway through my stay in Brazil, I took a trip to another city for several<br />
days, and later wanted to send postcards to people I had met there. I wrote a few<br />
cards, and then asked a native speaker to rewrite one for me. I noticed that he began
30 Cognitive Approaches to Interlanguage Pragmatic Development<br />
rephrasing my message with the expression E ai, como estdo ("So, how are<br />
you"), so I did the same with each subsequent card. A week after sending the<br />
cards, I got a call from one of the recipients (a native speaker of English who was a<br />
long-term resident in Brazil) who began the conversation by commenting that my<br />
Portuguese must be improving rapidly, given the colloquially appropriate style of<br />
my postcard.<br />
(4) On the first day of a 2-week trip to Thailand, I presented a paper at the end of<br />
the day at a national conference. After the lecture, several Thais with whom I would<br />
be working for the following week approached me and made some brief remarks in<br />
English (I know no Thai) and then slipped away. I found myself standing by myself<br />
much quicker than I expected, and had the unsettling feeling that my talk must have<br />
been very poorly received. I returned to my hotel feeling quite depressed about this.<br />
That evening, I looked over some materials that I had collected during the day,<br />
including an article by Sukwiwat and Fieg (1987) on greeting and leave-taking in<br />
Thai. Sukwiwat and Fieg pointed out that conversations are closed quickly in Thai<br />
but tend to be drawn gradually to a close in English, so that Americans are often<br />
taken aback by what appear to be abrupt, brusque, and sometimes rude departures.<br />
Thais, on the other hand, think that American leave-takings drag on excessively and<br />
involve unnecessary verbiage. I immediately realized that I might have misinterpreted<br />
the significance of what had happened earlier. For the remainder of my<br />
stay, I tried my best to beat the Thais at their own game by closing conversations<br />
faster than they could, for example, by suddenly announcing, "well, I'm leaving<br />
now." I never succeeded in getting away faster than they did, but my disquiet at this<br />
aspect of Thai behavior evaporated and I suffered no discomfort from behaving in a<br />
way that would be rude by my own cultural norms.<br />
(5) Between the early 1960s and mid-1970s, I lived mostly in Arabic-speaking<br />
countries and became fairly proficient in both Egyptian and Lebanese Arabic. In<br />
some varieties of Arabic, parents and other relatives may address children with what<br />
Ayoub (1964) has called bipolar kin terms, ego addressing alter with the term that in<br />
its literal sense would be appropriate for alter addressing ego; e.g., a grandfather<br />
may address his granddaughter with a term equivalent to "grandpa." This occurs<br />
when the senior wants the junior to do something, but chooses a conciliatory request<br />
form, metaphorically reversing the power relationship between the two—what<br />
Brown and Levinson (1987) would call a point-of-view operation. I knew of this<br />
phenomenon only from Ayoub's article, however. I never noticed parents using it<br />
with their children in either Egypt or Lebanon, although I often observed parents<br />
and other family members interacting with children. Years later, I noticed the use of<br />
such a form when visiting friends from Lebanon in California. Playing in the<br />
swimming pool, the mother said to her son, in English, "Okay, Baron, swim down<br />
the other end of the pool now, Mommy." I have since been assured by speakers of<br />
both Egyptian and Lebanese Arabic that they do use such forms, but second language<br />
speakers of Arabic whom I have asked have reported that, like me, they have<br />
never noticed it being used.<br />
(6) In several publications (Schmidt, 1983, 1984), I reported on a case study of a<br />
Japanese learner of English whose overall level of communicative competence was<br />
superior to his rather rudimentary control of English grammar. In looking at the<br />
development of pragmatic ability by my subject, Wes, I found that he often used
Consciousness, Learning, and Interlanguage Pragmatics 31<br />
hints that native speakers of English, including myself, did not realize were intended<br />
as directives. For example, once in a theater, Wes turned to me and asked me<br />
if I liked my seat. I responded that my seat was fine, not realizing at all that he was<br />
indirectly requesting that we change places. After many years of interacting with<br />
Japanese speakers of English, I think that I now recognize such hints on most<br />
occasions, but this learning process has been slow.<br />
All of these anecdotes indicate an apparently very close connection between<br />
noticing what was present in input and learning. Each case of successful learning<br />
also involved more than just noticing the forms used, but also an appreciation of<br />
their functional meaning: that an imperfective signaled an excuse, that entao td was<br />
a preclosing device, that e af was useful for greetings, that an abrupt departure did<br />
not necessarily imply a problem. Two of my Portuguese examples also illustrate<br />
intentional rather than incidental learning—I was deliberately seeking speech routines<br />
for openings and closings, and discovered them. By contrast, in the case of<br />
abrupt Thai departures, the learning was incidental; I had no prior awareness of a<br />
learning problem or intent to learn anything. 1 noticed the behavior and may or may<br />
not have carried on some conscious inferencing in arriving at my conclusion that my<br />
poor lecture lay behind it (those thought processes are not recoverable), but my<br />
corrected understanding of the significance of such behavior was fortuitous and the<br />
information was externally provided. In the case of the Portuguese imperfect used<br />
for excuses, explicit information about pragmatic function seems to have made the<br />
input more salient, though it is virtually certain that such forms were in input all<br />
along.<br />
In contrast to the Thai example, externally provided information about Arabic<br />
bipolar kin terms had no effect on my learning. I never noticed their occurrence in<br />
the dialects to which I was exposed, and they never became part of my competence<br />
in Arabic. This example suggests some of the difficulty in accounting for what<br />
becomes conscious and what does not. This is a complex issue beyond the scope of<br />
this paper, but part of such an account would include Baars's observation that events<br />
remain unnoticed if they are either uninterpretable in context or so stable as to be<br />
part of the context (Baars, 1983). The Arabic use of bipolar kin terms seems to be<br />
especially opaque to native speakers of English, who find them nearly uninterpretable.<br />
3 In the case of both Thai departures and Wes's hints in English, the problem lay<br />
not in noticing what was said but in understanding what was intended. The interpretation<br />
of hints is problematic for native speakers as well as learners (Ervin-Tripp,<br />
1972), but it is not clear to me why externally provided information was sufficient to<br />
block future inferences from abrupt departures to perception of a problem in the<br />
case of my Thai example, whereas knowledge about Japanese speech behavior at a<br />
similar level of generality did not lead quickly to the establishment of the appropriate<br />
inferencing behavior.<br />
Explicit and Implicit Learning of General Principles<br />
While all of my examples involve understanding in the sense of matching surface<br />
forms with meaning, none of them are good examples of generalization from specific<br />
examples to more general principles. However, there are cases in which the
32 Cognitive Approaches to Interlanguage Pragmatic Development<br />
learning of pragmatics and discourse must involve such generalization, for example,<br />
not just the recognition and use of frozen routines such as entao td and e ai, como<br />
estdo, but learning less frozen formulas, as well as fully productive structures for<br />
speech act realizations.<br />
A good example of the involvement of consciousness in generalizing a formula<br />
has been provided by Ferguson (1976) in recalling his learning of Arabic root-echo<br />
responses. There are numerous adjacency pairs in Arabic in which a greeting,<br />
compliment, or other initiating utterance requires a formulaic response that contains<br />
a lexical item (usually a verb) derived from the triconsonantal root of the most<br />
important lexical item in the initiating utterance. On one occasion, Ferguson bought<br />
an article of clothing in a market, and when the purchase was complete, the seller<br />
said to him "mabruuk" (congratulations). He did not know the response formula for<br />
this, but did know that an appropriate response form would be one which contained<br />
the root BRK from the first part of the adjacency pair. By analogy with several other<br />
response formulas that he did know, he guessed what the root-echo response form<br />
might be. Ferguson comments:<br />
Probably 'alia ybaarikfiik was the root-echo response to BRK. I tried it, and the<br />
smile showed I had given the right reply. The whole analysis took only a split<br />
second, and was just like getting an instance of grammatical concord or case<br />
government right. (141)<br />
This is an example of conscious problem solving or explicit learning, but I have<br />
indicated that implicit language learning is also possible. It may be useful, therefore,<br />
to spell out in some detail how more general principles of pragmatics might be<br />
acquired without being conscious of them.<br />
Following Fox (1987) and Spolsky (1989), I suggest that some pragmatic and<br />
discoursal principles are better represented as associative networks rather than as<br />
prepositional rules, and that connectionist models are promising in accounting for<br />
those aspects of pragmatic knowledge that do appear to be unconscious. This may<br />
involve less of a paradigm shift in the areas of pragmatics and discourse than in<br />
syntax, since researchers in pragmatics have had a less fixed notion of what is meant<br />
by a "rule" of pragmatics than have syntacticians, and relatively little attention has<br />
been given to consideration of how such rules might be represented psychologically.<br />
Some ethnomethodologists have rejected the concept of rules as conceived in formal<br />
logic as a model of social action (Mehan & Wood, 1975). Probabilistic network<br />
approaches have been suggested for the analysis of code-switching (Dearholt &<br />
Valdes-Falles, 1978); and Pomerantz (1978) has described compliment responses as<br />
the result of the cooperation of multiple constraints. Each of these approaches is<br />
compatible with a connectionist interpretation.<br />
There is one type of representation of pragmatic rules that I think is psychologically<br />
implausible, but that can also be recast in network form. The distribution<br />
of address forms has been represented by both Geoghegan (1971) and Ervin-Tripp<br />
(1972) in the form of a flowchart, illustrating decision points in the form of serially<br />
ordered binary selectors. For American English, Ervin-Tripp indicates that the first<br />
question to be asked is whether or not the addressee is a child or adult; then, if an<br />
adult, whether the interaction takes place within a status-marked setting; if the<br />
setting is not status-marked, whether the addressee's name is known; if the name is
Consciousness, Learning, and Interlanguage Pragmatics 33<br />
known, whether the addressee is kin; if not kin, whether the addressee is a friend or<br />
colleague; if so, whether the addressee is of higher rank; and so on, finally exiting<br />
the system with an appropriate address form. Ervin-Tripp (1972) and Geoghegan<br />
(1971) state explicitly that while paths through the flowchart represent rules, such<br />
flowcharts are like a formal grammar in representing a logical model and are not<br />
intended as psychological models of decision making.<br />
There are various problems with these models. Kendall (1981) has pointed out<br />
that they are too deterministic, and that a factor called "dispensation," meaning<br />
essentially to disregard all other factors, is introduced to get around the problem of<br />
variability. Positing serially ordered selectors also implies complete scalability<br />
(each selector must be listed only once, and selectors encountered first in the<br />
flowchart must outweigh all subsequent selectors in their influence), which cannot<br />
be empirically supported. For the present discussion, the most important drawback<br />
to such models is that they are unlikely to have any psychological reality. While<br />
conscious choices of which address forms in unclear situations might indeed involve<br />
sequential consideration of the types of selectors contained within flowcharts, there<br />
is little reason to suggest that automatic choices are made on the basis of speeded-up<br />
serial processing; most psychological accounts of automatic processing assume that<br />
parallel processes dominate. However, flowcharts such as those suggested by Ervin-<br />
Tripp (1972) and Geoghegan (1971) can easily be restructured into connectionist<br />
architecture, and the choice of address forms can be reconceptualized as a network<br />
of unordered connections between features of social context (addressee age, rank,<br />
marital status, etc.) and linguistic outputs. Some connections between social context<br />
features and address forms may be so heavily weighted that the connection is almost<br />
categorical, while others may be very weak, leading to fuzzier, less determinate<br />
outcomes.<br />
Other kinds of pragmatic knowledge that may be similarly represented include<br />
the complex patterns of covariation among features of social context and the linguistic<br />
realizations of speech acts that have been empirically documented by analysis of<br />
CCSARP data (Blum-Kulka, House & Kasper, 1989). The theoretical framework<br />
proposed by Brown and Levinson (1987), which attempts to relate a very wide range<br />
of pragmatic realizations to variation in three basic contextual features—social<br />
distance, power, and culture-specific evaluations of threat to face—is a similar case<br />
since these cooperating (or conflicting) constraints exert probabilistic influences.<br />
If we assume that associative network models have some face validity as a<br />
model of implicit pragmatic knowledge, we may then ask how such knowledge may<br />
be acquired, and specifically the role that consciousness is likely to play in the<br />
establishment of a network. There is some evidence from experimental psychology<br />
that bears on several aspects of this question.<br />
(1) Do learners have to keep track (consciously, by counting) of the frequency<br />
with which contextual or pragmatic features occur The answer to this is almost<br />
certainly no. Learners may be able to make reasonably accurate estimates of the<br />
relative frequency of such things, but they do not do so by counting, and it is widely<br />
accepted that attention to a stimulus event is sufficient to trigger the automatic<br />
(effortless and unintended) encoding of its frequency of occurrence (Hasher &<br />
Zacks, 1984).<br />
(2) Do learners need to notice the specific relevant pragmalinguistic or contex-
34 Cognitive Approaches to Interlanguage Pragmatic Development<br />
tual features of an event in order to trigger such encoding of frequency This<br />
question is somewhat controversial, but the answer is probably yes. Hanson and<br />
Hirst (1988) point out that an event may be thought of as a cluster of attributes. They<br />
report experiments supporting the hypothesis that attention to specific stimulus<br />
attributes is necessary in order to encode frequency information for those attributes.<br />
(3) Do learners need to understand the significance of co-occurring linguistic<br />
and social context features in order to acquire a network of complex covariations<br />
This is perhaps the most interesting question, and strikes to the heart of what is<br />
meant by implicit learning. Experiments in implicit learning suggest that implicit<br />
learning may be self-organizing, and that it is not necessary to realize the significance<br />
of one event for another in order to establish connections.<br />
Lewicki (1986) has reported a series of experiments designed to demonstrate the<br />
nonconscious detection of covariations involving social stimulus material. Subjects<br />
were presented with a series of descriptions of persons which mentioned a number<br />
of psychological and social characteristics. Some of these traits were manipulated<br />
by the experimenter, either to confirm or disconfirm preexisting stereotypes. After a<br />
learning phase, subjects rated new stimulus material. The experiments showed that<br />
correlations built into the personality descriptions during the learning phase influenced<br />
judgments in the testing phase. By running different versions of the same<br />
basic experiment, Lewicki was able to assess subject awareness at various points in<br />
learning. Subjects did notice the manipulated traits (as intended by the experimenter)<br />
and were momentarily aware of their co-occurrence in single stimulus descriptions.<br />
That is, they were able to recall both of them when questioned immediately<br />
after exposure. 4 However, the subjects were unaware of any systematic relationship<br />
between the manipulated traits. When told that some traits (out of a large number<br />
used as descriptors) had been systematically manipulated by the experimenter,<br />
subjects were unable to identify which ones had been manipulated.<br />
Analogously, we can specify the minimum requirements of learning an address<br />
system (or any other system of complex covariations) in a second language. If the<br />
task is to acquire an address system in which the ingroup/outgroup distinction is<br />
relevant or in which address forms systematically vary by sex of addressee, learners<br />
must attend to and notice in input both the linguistic forms and the relevant contextual<br />
features. This may mean attending to features of context that either are not<br />
relevant or are defined differently in the native language, so that learning a new<br />
pragmatic system often entails learning how to make new interpretative assessments<br />
of the world. However, it does not seem to be necessary for learners to make any<br />
conscious connection between the address forms encountered and the contextual<br />
factors that are correlated with such forms. For example, in learning the address<br />
system of Japanese, when you hear someone you know as Mr. Morita addressed as<br />
Morita-kun, where kun is an address form, you must notice both the form and the<br />
relevant contextual factors (these include sex, age, and rank of both speaker and<br />
addressee, intimacy, tone, and setting) if this is to be intake for learning, but need<br />
not draw the conclusion that Morita was addressed that way because of any of these<br />
factors.<br />
Nevertheless, it would certainly be extremely helpful to be consciously aware of<br />
such connections. It is sometimes argued that implicit learning is superior to conscious<br />
problem solving (Krashen, 1981), but this seems to be true for only some
Consciousness, Learning, and Intel-language Pragmatics 35<br />
types of learning tasks. Reber has reported several times that subjects learning<br />
artificial grammars under an implicit learning condition (subjects were told to<br />
memorize examples, which presumably interfered with any attempt to analyze the<br />
input) were better able to recognize valid new strings generated by the grammar than<br />
those subjects who were told to try to figure out the rules of the underlying grammar.<br />
However it cannot be assumed that subjects who attempted to discover the<br />
rules succeeded in doing so. In a recent publication, Reber (1989) makes exactly<br />
this point. Arguing that the particular artificial grammar to be learned was constructed<br />
in such a way that subjects were unlikely to be able to find the rules they<br />
were searching for, Reber now argues that "looking for rules will not work if you<br />
cannot find them," but "looking for rules will work if you can find them" (Reber,<br />
1989, 223). McLeod and McLaughlin (1986) report anecdotal evidence for the<br />
frequent occurrence of rapid restructuring following "clicks of comprehension."<br />
Conclusions<br />
The data from experimental psychology clearly support a conservative hypothesis<br />
that whatever learning might result from unattended processing is insignificant<br />
compared to the results of attended processing. The data seem to me to be also<br />
compatible with two much stronger hypotheses, that attention to input is a necessary<br />
condition for any learning at all, and that what must be attended to is not input in<br />
general, but whatever features of the input play a role in the system to be learned.<br />
For the learning of pragmatics in a second language, attention to linguistic forms,<br />
functional meanings, and the relevant contextual features is required. I also claim<br />
that learners experience their learning, that attention is subjectively experienced as<br />
noticing, and that the attentional threshold for noticing is the same as the threshold<br />
for learning. Finally, I argue that, while incidental and implicit learning are both<br />
possible, consciously paying attention to the relevant features of input and attempting<br />
to analyze their significance in terms of deeper generalizations are both highly<br />
facilitative.<br />
I do not claim that the anecdotal examples from my own language learning<br />
experiences prove these points, because the most that language learner diary reports<br />
can establish is that learners have noticed crucial facts about language use. What is<br />
needed is much more systematically gathered data on what learners notice (and are<br />
able to report) and what they do not notice (are unable to report) as they are<br />
learning. 5 Suspicions have been voiced that "it is doubtful that [introspection] can<br />
shed light on how the learner moves from one state to another, i.e., how input<br />
becomes intake" (Ellis, 1989a). I think that investigation of the learner's thoughts at<br />
such points of change is just what needs to be investigated. Even the harshest critics<br />
of reliance upon introspective methods agree that individuals do know the focus of<br />
their attention at any given point in time, as well as the content of their current<br />
thoughts, emotions, evaluations, and plans (Nisbett & Wilson, 1977) and that these<br />
conscious thought processes can be reported. A priori conclusions that there will be<br />
no relationships between such phenomena and language development are unwarranted.<br />
No strong prescriptions for the teaching of second language pragmatics can be
36 Cognitive Approaches to Interlanguage Pragmatic Development<br />
drawn from this discussion, but some general observations seem in order. Simple<br />
exposure to sociolinguistically appropriate input is unlikely to be sufficient for<br />
second language acquisition of pragmatic and discoursal knowledge because the<br />
linguistic realizations of pragmatic functions are sometimes opaque to language<br />
learners and because the relevant contextual factors to be noticed are likely to be<br />
defined differently or may be nonsalient for the learner. Second language learners<br />
may fail to experience the crucial noticings for years. The fact that this does not<br />
seem to happen in first language learning is attributable not to any sort of pragmatics<br />
acquisition device, but to the efforts that parents and other caregivers make in order<br />
to teach communicative competence to children, using a variety of strategies.<br />
Motivation is an important determinant of the allocation of attentional resources<br />
(Crookes & Schmidt, 1991). Because of the close connections among pragmatic<br />
realization strategies, assessments of role and status relationships between speaker<br />
and hearer, and the expression of personality, it is likely that there is a stronger<br />
relationship between motivation, acculturation and other affective factors in the<br />
development of pragmatic and discoursal ability than in other aspects of language<br />
learning, such as syntax (Schmidt, 1983). Those who are concerned with establishing<br />
relationships with target language speakers are more likely to pay close attention<br />
to the pragmatic aspects of input and to struggle to understand than those who are<br />
not so motivated. But since intentional learning is unnecessary when some task<br />
causes attention to be focused on what is to be learned, one way to develop pragmatic<br />
competence in classroom contexts could be through task-based language teaching<br />
(Long, in press). Tasks can be selected that focus the learner's attention on pragmatic<br />
forms, functions and co-occurring features of social context.<br />
Explicit teacher-provided information about the pragmatics of the second language<br />
can also play a role in learning, provided that it is accurate and not based<br />
solely on fallible native speaker intuitions. Explicit teaching is often more efficient<br />
than attention to input for identifying the pragmalinguistic forms of the target<br />
language. The understanding of general rules and patterns may be unnecessary for<br />
learning, but Grossberg (1988) has argued that the learning mechanisms modeled in<br />
connectionist networks are slow because they result only from gradual changes in<br />
the bottom-up adaptive weights of the network, whereas top-down processes such<br />
as focused attention and expectations greatly speed up and actively reorganize the<br />
way in which input is processed. This is not to claim that explicit knowledge<br />
somehow "becomes" implicit knowledge, but to recognize a synergistic relationship<br />
between the mechanisms of implicit and explicit learning (Mathews et al. 1989),<br />
which justifies a consciousness-raising approach to the teaching of pragmatics.<br />
Notes<br />
1. I am grateful to Michael Long, Paul Munsell, and Danny Steinberg for very helpful<br />
comments on an earlier draft of this paper.<br />
2. I owe the distinction made here between information that is perceived and information<br />
that is noticed to Bowers (1984), who argues that information becomes conscious when it is<br />
processed to the level of short-term memory and selectively attended to. Bowers also distin-
Consciousness, Learning, and Interlanguage Pragmatics 37<br />
guishes between two senses of unconscious, referring to information that is unnoticed and<br />
information that is unappreciated or uncomprehended, and I have drawn upon his model in<br />
my description of subliminal versus implicit learning.<br />
3. When I have presented this example to native speakers of English, they have often<br />
assumed that the mother must have meant for her son to "swim to Mommy," but this is not a<br />
correct interpretation; that is, this is not an example of a missing preposition.<br />
4. In another series of experiments, Lewicki (1986) attempted to demonstrate that information<br />
which is presented subliminally or which is not attended to may also lead to learning.<br />
These experiments did not successfully demonstrate learning, but some interesting subtle<br />
effects were found. Subjects responded more slowly to questions mentioning those stimulus<br />
traits that had been presented subliminally. Lewicki argues that this demonstrates the internalization<br />
of weak processing algorithms, which could eventually result in more demonstrable<br />
learning effects. Such an experimental demonstration would disprove my claim that there is<br />
no subliminal learning whatsoever. Baars (1988) has claimed that this zero-point question is<br />
essentially unanswerable and has obscured the more important and answerable question of<br />
whether more conscious involvement is needed to learn more information, the answer to<br />
which is clearly affirmative.<br />
5. Michael Long (personal communication) has pointed out to me that language learners<br />
may sometimes produce a vocabulary item in a second language that they did not know they<br />
knew until that moment, not being sure that it is right and certainly not knowing how it ever<br />
got into the mental lexicon. However, the issue of whether the learner noticed such a lexical<br />
item in input (which must have occurred, if my account is correct) is quite separate from the<br />
question of whether the learner will be able to say much later when it was encountered. We<br />
know all sorts of things without being able to recall the circumstances under which we<br />
acquired that knowledge.<br />
References<br />
Abrams, M., & Reber, A. (1988). Implicit learning: Robustness in the face of psychiatric<br />
disorders. Journal of Psycholinguistic Research, 17(5), 425-39.<br />
Allport, D. (1979). Conscious and unconscious cognition: A computational metaphor for the<br />
mechanism of attention and integration. In L. Nilsson (Ed.), Perspectives on memory<br />
research. Hillsdale, NJ: Lawrence Erlbaum, pp. 61-89.<br />
Anderson, J. (1982). Acquisition of cognitive skill. Psychological Review, 89, 369-406.<br />
Anderson, J. (1985). Cognitive psychology and its implications (2nd ed.). New York: W. H.<br />
Freeman.<br />
Atkinson, R., & Shiffrin, R. (1968). Human memory: A proposed system and its control<br />
processes. In K. Spense (Ed.), The psychology of learning and motivation: Advances<br />
in research and theory (Vol. 2). New York: Academic Press, pp. 89-195.<br />
Ayoub, M. (1964). Bi-polarity in Arabic kinship terms. In H. Lunt (Ed.), Proceedings of the<br />
ninth international congress of linguistics. The Hague: Mouton, pp. 1100-1106.<br />
Baars, B. (1983). Conscious contents provide the nervous system with coherent, global<br />
information. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.), Consciousness and<br />
self-regulation: Advances in research and theory (Vol. 3). New York: Plenum Press,<br />
pp. 41-79.<br />
Baars, B. (1988). A cognitive theory of consciousness. New York: Cambridge University<br />
Press.<br />
Battista, J. (1978). The science of consciousness. In K. Pope & .1. Singer (Eds.), The stream
38 Cognitive Approaches to Interlanguage Pragmatic Development<br />
of consciousness: Scientific investigations into the flow of human experience. New<br />
York: Plenum Press, pp. 57-87.<br />
Bialystok, E. (1979). Explicit and implicit judgements of L2 grammaticality. Language<br />
Learning, 29, 81-103.<br />
Bialystok, E. (1981). The role of linguistic knowledge in second language use. Studies in<br />
Second Language Acquisition, 4, 31-45.<br />
Bialystok, E., & Bouchard-Ryan, E. (1985). A metacognitive framework for the development<br />
of first and second language skills. In D. Forest-Prcssley, G. MacKinnon, &<br />
T. Waller (Eds.), Metacognition, cognition, and human performance. Orlando, FL:<br />
Academic Press, pp. 207-52.<br />
Blom, J., & Gumperz, J. (1972). Social meaning in linguistic structures: Code-switching in<br />
Norway. In J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics (407-34).<br />
New York: Holt, Rinehart & Winston, pp. 407-34.<br />
Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-cultural pragmatics: Requests and<br />
apologies. Norwood, NJ: Ablex.<br />
Bowers, K. (1984). On being uncousciously influenced and informed. In K. Bowers &<br />
D. Meichenbaum (Eds.), The unconscious reconsidered. New York: Wiley, pp.<br />
227-72.<br />
Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge:<br />
Cambridge University Press.<br />
Cameron, D. (1985). Feminism and linguistic theory. London: Macmillan.<br />
Carroll, J. (1981). Conscious and automatic processes in language learning. Canadian Modern<br />
Language Review, 37, 462-74.<br />
Cathcart, R. L. (1989). Authentic discourse and the survival English curriculum. TESOL<br />
Quarterly, 23(1), 105-26.<br />
Ceci, S., & Howe, M. (1982). Metamemory and the effects of intending, attending and<br />
intending to attend. In G. Underwood (Ed.), Aspects of consciousness (Vol. 3):<br />
Awareness and self-awareness. London: Academic Press, pp. 147-64.<br />
Cheesman, J., & Merikle, P. M. (1986). Distinguishing conscious from unconscious perception.<br />
Canadian Journal of Psychology, 40, 343-67.<br />
Clancy, P. (1986). The acquisition of communicative style in Japanese. In B. Scheiffelin &<br />
E. Ochs (Eds.), Language socialization across cultures. Cambridge: Cambridge University<br />
Press, pp. 213-49.<br />
Clark, E. (1978). Awareness of language: Some evidence from what children say and do. In<br />
A. Sinclair, R. Jarvella, & W. Levelt (Eds.), The child's conception of language.<br />
Berlin: Springer-Verlag, pp. 17-43.<br />
Crookes, G., & Schmidt, R. (1991). Motivation: Reopening the research agenda. Language<br />
Learning, 41(4), 469-512.<br />
Dearholt, D., & Valdes-Fallis, G. (1978). Toward a probabilistic automata model of some<br />
aspects of code-switching. Language in Society, 7(3), 411-19.<br />
Demuth, K. (1986). Prompting routines in the language socialization of Basotho children. In<br />
B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures. Cambridge:<br />
Cambridge University Press, pp. 51-79.<br />
Dixon, N. (1971). Subliminal perception: The nature of a controversy. London: McGraw-<br />
Hill.<br />
Dixon, N. (1981). Preconscious processing. New York: Wiley.<br />
Eich, E. (1984). Memory for unattended events: Remembering with and without awareness.<br />
Memory and Cognition, 12, 105-11.<br />
Ellis, R. (1989a). Review of C. Facrch&G. Kasper (Eds.), Introspection in second language<br />
research. Applied Linguistics, 10(3), 360-64.
Consciousness, Learning, and Interlanguage Pragmatics 39<br />
Ellis, R. (1989b). The role of practice in classroom learning. Revue de L'AILAIAILA Review,<br />
5, 20-39.<br />
Ericsson, K., & Simon, H. (1984). Protocol analysis: Verbal reports as data. Cambridge,<br />
MA: MIT Press.<br />
Eriksen, C. W. (1960). Discrimination and learning without awareness: A methodological<br />
survey and evaluation. Psychological Review, 67, 279-300.<br />
Ervin-Tripp, S. (1972). On sociolinguistic rules: Alternation and co-occurrence. In J. Gumperz<br />
& D. Hymes (Eds.), Directions in sociolinguistics. New York: Holt, Rinehart &<br />
Winston (pp. 213-50).<br />
Faerch, C., & Kasper, G. (1984). Pragmatic knowledge: Rules and procedures. Applied<br />
Linguistics, 5(3), 214-25.<br />
Faerch, C., & Kasper, G. (Eds.). (1987). Introspection in second language research. Clevedon<br />
and Philadelphia: Multilingual Matters.<br />
Faerch, C., & Kasper, G. (1989). Interlanguage request modification. In S. Blum-Kulka, J.<br />
House & G. Kasper (Eds.), Cross-cultural pragmatics. Norwood, NJ: Ablex, pp.<br />
221-47.<br />
Ferguson, C. (1976). The structure and use of politeness formulas. Language in Society, 5,<br />
137-51.<br />
Fox, B. (1987). Variation and patterning in discourse and grammar. In K. M. Denning, S.<br />
Inkeles, F. C. McNair-Knox, & J. R. Rickford (Eds.), Variation in language:<br />
NWAVE-XV at Stanford. Stanford: Stanford University, Department of Linguistics.<br />
Gass, S. (1988). Integrating research areas: A framework for second language studies.<br />
Applied Linguistics, 9, 198-217.<br />
Geoghegan, W. (1971). Information processing systems in culture. In P. Kay (Ed.), Explorations<br />
in mathematical anthropology. Cambridge, MA: MIT Press, pp. 3-35.<br />
Gleason, J., & Perlmann, R. (1985). Acquiring social variation in speech. In H. Giles &<br />
R. St. Clair (Eds.), Recent advances in language, communication, and social psychology.<br />
London: Lawrence Erlbaum.<br />
Grossberg, S. (1988). Nonlinear neural networks: Principles, mechanisms, and architectures.<br />
Neural Networks, I, 17-61.<br />
Hanson, C., & Hirst, W. (1988). Frequency encoding of token and type information. Journal<br />
of Experimental Psychology: Learning, Memory, and Cognition, 14(2), 289-97.<br />
Hartman, M., Knopman, D. S., & Nissen, M. J. (1989). Implicit learning of new verbal<br />
associations. Journal of Experimental Psychology: Learning, Memory, and Cognition,<br />
15(6), 1070-82.<br />
Hasher, L., &Zacks, R. (1984). Automatic processing of fundamental information: The case<br />
of frequency of occurrence. American Psychologist, 39, 1372-88.<br />
Holmes, J. (1988). Doubt and uncertainty in ESL textbooks. Applied Linguistics, 9(1), 21-<br />
44.<br />
James, W. (1890). The principles of psychology. New York: Holt.<br />
Johnson-Laird, P. (1983). Mental models. Cambridge, MA: Harvard University Press.<br />
Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.<br />
Kasper, G. (1984). Pragmatic comprehension in learner-native speaker discourse. Language<br />
Learning, 34(4), 1-20.<br />
Kendall, M. (1981). Toward a semantic approach to terms of address: A critique of deterministic<br />
models in sociolinguistics. Language & Communication 1(2/3), 237—54.<br />
Key, W. B. (1973). Subliminal seduction. Englewood Cliffs, NJ: Signet.<br />
Kihlstrom, J. (1984). Conscious, subconscious, unconscious: A cognitive perspective. In<br />
K. Bowers & D. Meichcnbaum (Eds.), The unconscious reconsidered. New York:<br />
Wiley, pp. 149-211.
40 Cognitive Approaches to Intel-language Pragmatic Development<br />
Krashcn, S. (1981). Second language acquisition and second language learning. Oxford:<br />
Pergamon.<br />
Krashen, S. (1983). Newmark's "ignorance hypothesis" and current second language acquisition<br />
theory. In S. Gass & L. Selinker (Eds.), Language transfer in language learning.<br />
Rowley MA: Newbury House, pp. 135-53.<br />
Lakoff, R. (1973). Language and woman's place. Language in Society, 2(1), 45-80.<br />
Lewicki, P. (1986). Nonconscious social information processing. New York: Academic<br />
Press.<br />
Lewis, M., & Anderson, J. (1985). Discrimination of operator schemata in problem solving:<br />
Learning from examples. Cognitive Psychology, 17, 26-65.<br />
Long, M. H. (in press). Task-based language teaching. Oxford: Basil Blackwell.<br />
MacWhinncy, B., Leinbach, J., Taraban, R., & MacDonald, J. (1989). Language learning:<br />
Cues or rules Journal of Memory and Language, 28, 255-77.<br />
McClelland, J., Rumelhart, D., & the POP Research Group. (1986). Parallel distributed<br />
processing: Explorations in the microstructure of cognition (Vol. 2): Psychological<br />
and biological models. Cambridge: MA: MIT Press.<br />
McLaughlin, B., Rossman, T., & McLeod, B. (1983). Second language learning: An<br />
information-processing perspective. Language Learning, 33, 135-57.<br />
McLeod, B., & McLaughlin, B. (1986). Restructuring or automaticity Reading in a second<br />
language. Language Learning, 36(2), 109-23.<br />
Marcel, A. (1983). Conscious and unconscious perception: An approach to the relations<br />
between phenomenal experience and cognitive processes. Cognitive Psychology, 15,<br />
238-300.<br />
Mathews, R. C., Buss, R. R., Stanley, W. B., Blanchard-Fields, F., Cho, J. R., &<br />
Druhan, B. (1989). Role of implicit and explicit processes in learning from examples:<br />
A synergistic effect. Journal of Experimental Psychology: Learning, Memory, and<br />
Cognition, 15(6), 1083-1100.<br />
Mehan, H., & Wood, H. (1975). The reality of ethnomethodology. New York: Wiley.<br />
Merikle, P. M. (1988). Subliminal auditory messages: An evaluation. Psychology and Marketing,<br />
5(4), 355-72.<br />
Moore, T. E. (Ed.). (1988). Subliminal influences in marketing (Special issue]. Psychology<br />
and Marketing, 5(4), 291-372.<br />
Munsell, P., & Carr, T. (1981). Monitoring the monitor: Review of Second language acquisition<br />
and second language learning. Language Learning, 31, 493-502.<br />
Natsoulis, T. (1987). The six basic concepts of consciousness and William James' stream of<br />
thought. Imagination, Cognition and Personality, 6, 289-319.<br />
Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Verbal reports on mental<br />
processes. Psychological Review, 84, 231-59.<br />
Norman, D., & Shallice, T. (1986). Attention to action: Willed and automatic control of<br />
behavior. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.), Consciousness and selfregulation:<br />
advances in research and theory (Vol. 4). New York: Plenum Press, pp.<br />
1-18.<br />
Ochs, E. (1979). Planned and unplanned discourse. In T. Givon (Ed.), Syntax and semantics,<br />
(Vol. 12): Discourse and syntax. New York: Academic Press, pp. 51-80.<br />
Ochs, E. (1986). From feelings to grammar: A Samoan case study. In B. Schieffelin &<br />
E. Ochs (Eds.), Language socialization across cultures (251-272). Cambridge: Cambridge<br />
University Press.<br />
Odlin, T. (1986). On the nature and use of explicit knowledge. IRAL, 24(2), 123-44.<br />
Olshtain, E., & Blum-Kulka, S. (1989). Happy Hebrish: Mixing and switching in American-<br />
Israeli family interactions. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.),
Consciousness, Learning, and Interlanguage Pragmatics 41<br />
Variation in second language acquisition: Discourse and pragmatics. Clevedon &<br />
Philadelphia: Multilingual Matters, pp. 59-83.<br />
O'Malley, J. M., Chamot, A. U., & Walker, C. (1987). Some applications of cognitive<br />
theory to second language acquisition. Studies in Second Language Acquisition, 9,<br />
287-306.<br />
Packard, V. (1957). The growing power of admen. Atlantic Monthly (September), 55-59.<br />
Pomerantz, A. (1978). Compliment responses: Notes on the co-operation of multiple constraints,<br />
(n J. Schenkein (Ed.), Studies in the organization of conversational interaction.<br />
New York: Academic Press, pp. 79-110.<br />
Posner, M., & Klein, R. (1973). On the functions of consciousness. In S. Kornblum (Ed.),<br />
Attention and performance IV. London: Academic Press, pp. 21-35.<br />
Reber, A. (1976). Implicit learning of synthetic languages: The role of instructional set.<br />
Journal of Experimental Psychology: Human Learning and Memory, 2, 88-94.<br />
Reber, A. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology:<br />
General, 118(3), 219-35.<br />
Reber, A., Allen, R., & Regan, S. (1985). Syntactical learning and judgment, still unconscious<br />
and still abstract: Comment on Dulany, Carlson, and Dcwey. Journal of Experimental<br />
Psychology: General, 114, 17-24.<br />
Reber, A., Kassin, S., Lewis, S., & Cantor, G. (1980). On the relationship between implicit<br />
and explicit modes in the learning of a complex rule structure. Journal of Experimental<br />
Psychology: Human Learning and Memory, 6, 492-502.<br />
Richards, J. C. (1990). The language teaching matrix: Curriculum, methodology, and materials.<br />
New York: Cambridge University Press.<br />
Rumelhart, D., & McClelland, J. (1986). On learning the past tense of English verbs. In<br />
McClelland, Rumelhart & the POP Research Group, 216-71.<br />
Rutherford, W., & Sharwood Smith, M. (1985). Consciousness raising and universal grammar.<br />
Applied Linguistics, 6, 274-82.<br />
Schmidt. R. (1983). Interaction, acculturation, and the acquisition of communicative competence:<br />
A case study of an adult. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and<br />
language acquisition. Rowley, MA: Newbury House, pp. 237-74.<br />
Schmidt, R. (1984). The strengths and limitations of acquisition: A case study of an untutored<br />
language learner. Language Learning and Communication, 3(1), 1-16.<br />
Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics,<br />
11(2), 17-46.<br />
Schmidt, R., & Frota, S. (1986). Developing basic conversational ability in a second language:<br />
A case study of an adult learner of Portuguese. In R. Day (Ed.), Talking to<br />
learn: Conversation in second language acquisition. Rowley, MA: Newbury House,<br />
pp. 137-326.<br />
Seidenberg, M., & McClelland, J. (1989). A distributed, developmental model of word<br />
recognition and naming. Psychological Review, 96(4), 523-68.<br />
Seliger, H. (1983). The language learner as linguist: Of metaphors and realities. Applied<br />
Linguistics, 4, 179-91.<br />
Sharwood Smith, M. (1981). Consciousness-raising and the second language learner. Applied<br />
Linguistics, 2, 159-68.<br />
Shiffrin, R., & Schneider, W. (1977). Controlled and automatic human information processing.<br />
II. Perceptual learning, automatic attending, and a general theory. Psychological<br />
Review, 84, 127-90.<br />
Snow, C. E., Perlman, R. Y., Gleason, J. B., & Hooshyar, N. (1990). Developmental<br />
perspectives on politeness: Sources of children's knowledge. Journal of Pragmatics,<br />
14(2), 289-305.
42 Cognitive Approaches to Interlanguage Pragmatic Development<br />
Sokolik, M., & Smith, M. (1989). French gender recognition: A network model and implications<br />
for second language acquisition. Paper presented at the Second Language Research<br />
Forum, Los Angeles.<br />
Spolsky, B. (1989). Communicative competence, language proficiency, and beyond. Applied<br />
Linguistics, 10(2), 138-56.<br />
Sukwiwat, M., & Fieg, J. (1987). Greeting and leave-taking. PASAA, 17(2), 1-12.<br />
Underwood, G. (1976). Attention and memory. Oxford: Pergamon Press.<br />
Underwood, G. (1982). Attention and awareness in cognitive and motor skills. In G. Underwood<br />
(Ed.), Aspects of consciousness (Vol. 3): Awareness and self-awareness. London:<br />
Academic Press, pp. 111-45.<br />
Watson-Gegeo, K., & Gegeo, D. (1986). Calling-out and repeating routines among Kwara'ae<br />
children. In B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures<br />
(17-50). Cambridge: Cambridge University Press.<br />
White, P. (1980). Limitations on verbal reports of internal events: A refutation of Nisbett and<br />
Wilson and of Bern. Psychological Review, 57(1), 105-12.<br />
Williams, M. (1988). Language taught for meetings and language used in meetings: Is there<br />
anything in common Applied Linguistics, 9(1), 45-58.<br />
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House.<br />
Wolfson, N., D'Amico-Reisner, L., & Huber, L. (1983). How to arrange for social commitments<br />
in American English: The invitation. In N. Wolfson & E. Judd (Eds.), Sociolinguistics<br />
and language acquisition. Rowley, MA: Newbury House, pp. 116-28.
2<br />
Symbolic Representation<br />
and Attentional Control<br />
in Pragmatic Competence<br />
ELLEN BIALYSTOK<br />
Pragmatic competence entails a variety of abilities concerned with the use and<br />
interpretation of language in contexts. It includes speakers' ability to use language<br />
for different purposes—to request, to instruct, to effect change. It includes listeners'<br />
ability to get past the language and understand the speaker's real intentions, especially<br />
when these intentions are not directly conveyed in the forms—indirect requests,<br />
irony and sarcasm are some examples. It includes command of the rules by<br />
which utterances are strung together to create discourse. This apparently simple<br />
achievement to produce coherent speech itself has several components—turntaking,<br />
cooperation, cohesion.<br />
These three aspects of pragmatic competence converge most prominently on the<br />
ability to use and interpret nonliteral forms, such as metaphorical uses of language<br />
and indirect requests. Native speakers are expected to see through the forms to<br />
retrieve the speaker's intentions. When listeners fail to do this, conversation (and<br />
sometimes more than just conversation) breaks down.<br />
All three aspects of pragmatic competence are implicated in some respect for all<br />
conversation, but the challenge is greater when the conversation includes nonliteral<br />
forms to signal indirect requests. It is at these times that the ability to perform<br />
according to pragmatic rules is more rigorously tested. To participate successfully in<br />
such conversations, speakers need to have mastered the three aspects of language<br />
listed above. First, speakers must be capable of using language for different purposes<br />
so that the speech act of requesting is properly distinguished from other<br />
intended effects of language use. Second, speakers must be capable of modifying<br />
the form of the request to reflect social aspects of the context, embedding the<br />
request in the forms necessary to indicate the socially expected degree of politeness<br />
or deference. Finally, speakers must be capable of participating in an interaction<br />
following the conventions of conversation in order to make the request.<br />
Descriptions of how speakers come to use and understand indirect and meta-<br />
43
44 Cognitive Approaches to Interlanguage Pragmatic Development<br />
phoric speech provide an important component of the explanation of the development<br />
of language proficiency both for children learning a first language and adults<br />
learning a second language. For children, this skill is reflected in their ability to<br />
understand the intention of forms used in the following types of conversations<br />
(examples taken from Shatz, 1983, 843):<br />
(1) MOTHER: He's three and three-quarters. (Child looks puzzled).<br />
MOTHER: How old are you<br />
CHILD: Four and two dollars.<br />
(2) EXPERIMENTER: (inquiring about a 3-year-old's ability): Can you tie your shoes<br />
CHILD: They are tied.<br />
This chapter will examine the nature and development of the linguistic representations<br />
that underlie pragmatic competence, specifically the competence involved in<br />
the use and interpretation of indirect and metaphoric speech. Knowledge for rules of<br />
use must be learned, represented, and transformed in the same way as the knowledge<br />
that controls other, more formal, aspects of the linguistic system. How is this<br />
knowledge represented, how do these representations change for speakers of a<br />
second language, and how do these representations support the use and interpretation<br />
of nonliteral speech This chapter will begin with a brief summary of the<br />
descriptions offered for children's development of pragmatic abilities. I shall organize<br />
this summary around children's developing competence with the three aspects<br />
of pragmatic ability identified above: increase in the variety of speech acts, greater<br />
modification of speech according to the context, and mastery of conversation and<br />
discourse. In the next section, I outline a model of language processing. Finally, I<br />
shall apply this model to the problem of adult second language learners' development<br />
of pragmatic competence.<br />
Children's Pragmatic Competence<br />
The most frequently studied aspect of pragmatic competence is the ability to make<br />
use of a variety of language functions. These functions have been identified and<br />
classified in various taxonomies. Searle's (1979) five-part classification system<br />
provides a moderate number of distinctions and is widely used. He includes the<br />
following as a finite and exhaustive list of possible speech acts: representatives,<br />
directives, commissives, expressives, and declarations. Other systems differ in<br />
number and detail. Halliday (1975), for example, considers the relevant distinctions<br />
to be among the ideational, interpersonal, and textual functions of language. The<br />
point, however, remains essentially the same, namely, that these categories constitute<br />
different linguistic functions that native speakers would normally be expected to<br />
master.<br />
How do children master the variety of language functions that native speakers of<br />
a language routinely express According to Clark and Clark (1977), the speech of<br />
young children learning their first language at the earliest stages is mostly confined<br />
to assertions (representatives) and requests (directives). Early assertions generally<br />
indicate the existence of objects, describe their location, or comment on their
Symbolic Representation and Attentional Control 45<br />
properties. Requests are typically motivated by needs or wants, asking for food,<br />
drink, toys, and the like.<br />
Within this limited number of categories and primitive syntactical resources for<br />
language structure, children manage nonetheless to express a range of semantic<br />
relations. These semantic relations are perhaps a better estimate of the children's<br />
ability to manipulate language to express a variety of intentions. Bloom (1973)<br />
astutely pointed out that when her daughter Kathryn said "mommy sock" on two<br />
different occasions, her intention in each was quite different. In one case, Kathryn<br />
was describing the spatial relation between an object and a location, and in the<br />
other, the interactive relation between an actor and an object. Similarly, a young<br />
child's utterance of "apple" can serve as a request, a declaration, or a question in<br />
different circumstances. In the preschool years, therefore, the range of pragmatic<br />
functions that children typically engage in is more limited than the range of semantic<br />
relations they express through these functions. Nonetheless, these semantic<br />
relations still constitute a finite set, including such intentions as presence of object,<br />
location of object, request for action, request for information (Bloom, 1973; Bowerman,<br />
1973; Schlesinger, 1971).<br />
In spite of these limitations, Shatz and McCloskey (1984) argue that young<br />
preschool children know a great deal about different speech acts. They point out that<br />
by age 2 children are able to distinguish between such speech acts as questions and<br />
nonquestions regardless of how they are formed. There were even some children<br />
under 2 years who were found to discriminate between types of questions—yes/no<br />
questions from wh- questions. Shatz and McCloskey claim that these achievements<br />
indicate that "young children's responses are not governed by a literal interpretation<br />
strategy" and that they have an "early understanding of the constraints of discourse"<br />
(25). These two insights correspond to the second and third aspects of pragmatic<br />
competence identified here.<br />
As children enter the school years, there is an expansion in the range of language<br />
functions available to them. This increase is achieved in two ways: by adding new<br />
speech acts to the repertoire, such as commissives and expressives, and by learning<br />
to use a wider range of formal structures to express the speech acts they have already<br />
been using. This addition of new and more flexible resources to express old intentions<br />
enhances the child's pragmatic competence, even without the addition of new<br />
speech acts. By the early school years, children's control over a variety of speech<br />
acts is fairly impressive, although the use of declarations is still rare.<br />
The second aspect of pragmatic competence is the ability to adjust the form of<br />
speech to comply with contextual factors. This manipulation is most clearly signaled<br />
by children's ability to use and interpret such forms as indirect requests. As<br />
this aspect of competence is developing, young children often make conspicuous<br />
errors, such as the frequently cited example of the following telephone conversation:<br />
(3) CALLER: Is your mother home<br />
CHILD: Yes.<br />
In this case the child has failed to identify the question as an indirect request for<br />
action.
46 Cognitive Approaches to Interlanguage Pragmatic Development<br />
The development of children's use of requests and the variety of direct and<br />
indirect means for expressing those requests is an impressive achievement in the<br />
preschool years. Becker (1982) documents the changes that occur in children's<br />
formulation of requests from the earliest attempts based only on gestures, sounds,<br />
and primitive utterances such as "more," to requests based on well-formed syntactic<br />
structures, and by about age 3 to the inclusion of some indirect requests. By age 5<br />
children's command of requests is quite sophisticated, including a variety of indirect<br />
forms, hints and utterances that involve sequencing several steps to achieve the<br />
requested goal. The proportion of indirect to direct requests increases between the<br />
ages of about 3 and 5 years, but then remains relatively constant until about 8 years.<br />
Ervin-Tripp (1977) documents as well the development of children's ability to<br />
both produce and interpret indirect speech. She points out that there is considerable<br />
variability among children in this development and identifies some of the social and<br />
contextual factors that account for that variability. Perhaps more surprising is variability<br />
in the interpretation of indirect speech that she attributes to linguistic differences,<br />
even by the same children (Ervin-Tripp, 1989). She reports that French-<br />
English bilingual children in Switzerland interpreted indirect speech correctly in<br />
French, accounting for such devices as irony and sarcasm, but incorrectly in English,<br />
taking the speech for its literal rather than its intended meaning.<br />
The third aspect of pragmatic competence is understanding the structure of<br />
conversation: What are the rules that govern how to structure discourse and participate<br />
in conversations Knowing how to carry on a conversation includes knowing<br />
the rules for mechanical aspects, such as turn-taking, as well as linguistic aspects,<br />
such as coherence and knowledge of adjacency pairs, or what kinds of utterances<br />
are expected to follow others. McTear (1984) summarizes these achievements by<br />
claiming that the important feature of children's developing conversation is their<br />
increased awareness of and use of the contingency relations that characterize normal<br />
conversational interactions.<br />
Bruner (1983) claims that a great deal of the knowledge about conversations is<br />
learned in infancy. Adults and infants engage in conversation-type activities that<br />
include turn-taking, focusing on common objects, and vocalization (simply cooing<br />
or gurgling for the infant). From this basis, he argues, the essential dynamics of<br />
conversation are learned and easily transferred to linguistic interactions.<br />
Even with the prelinguistic experiences in which Bruner invests so much explanatory<br />
power, children must still learn the rules that govern linguistic conversations.<br />
Bloom, Rocissano, and Hood (1976) studied the development of discourse coherence<br />
by children to determine the extent to which utterances were appropriately<br />
related to preceding utterances. They found that children were contributing to conversations<br />
with considerable accuracy by the age of two. Shatz and McCloskey<br />
(1984) point to the need for specific knowledge of conversational rules; such knowledge<br />
is learned from early on but is certainly not complete in the preschool years:<br />
Speaking appropriately, then, is not just a matter of speaking grammatically or<br />
matching forms in initiation with forms in response. Appropriate language behaviour<br />
also involves knowledge of the conversational rules, conventions, and social<br />
circumstances governing linguistic interactions. (19)
Symbolic Representation and Attentional Control 47<br />
There is some dispute about when these rules are learned. McTear (1984) attributes<br />
knowledge of them to preschool children but Ackerman (1981) defers their emergence<br />
until the school years.<br />
In summary, children continue to advance in their mastery of pragmatic uses of<br />
language through the preschool and early school years. They learn to use a wider<br />
range of speech acts, they learn to modify the forms of their utterances to accommodate<br />
social aspects of language use and to interpret indirect forms, and they learn the<br />
conventions for participating in conversations. Although the rudiments of all these<br />
abilities are evident from a very early age, possibly from the child's first words, real<br />
competence with these pragmatic functions of language cannot be claimed to be<br />
complete until the school years. Adult second language learners must also gain<br />
control over the pragmatic uses of language in contexts, but they certainly do not<br />
begin with a childlike naivete about the social uses of language. What sort of<br />
explanation, then, could account for development of pragmatic competence both by<br />
children who are learning their first language and by adults who already have<br />
mastered another linguistic system<br />
Learning and Using Language<br />
The analysis of pragmatic competence presented here is based on a model of<br />
language processing that we have been using as the framework for research into a<br />
variety of aspects of language acquisition and use. The model has two sides to it and<br />
requires two kinds of descriptions. One of these is the description of learner competence.<br />
For this purpose, the goal of the model is to describe the processing ability of<br />
language learners in terms of the cognitive mechanisms responsible for learning and<br />
using language. The second is the description of task demands. For this purpose,<br />
language functions are analyzed for the cognitive demands they place upon learners.<br />
Different uses of language involve different processing abilities of language<br />
learners; the assumption is that these demands can be systematically determined and<br />
analyzed for their difficulty. Language proficiency, then, is considered in terms of<br />
the fit between the processing abilities of the learner and the task demands imposed<br />
by a specific language use situation. Where the two are congruent, learners will<br />
perform well; where the task demands are excessive relative to the learner's ability,<br />
learners will struggle.<br />
The descriptions that are applied to these two sides of the problem (namely,<br />
speakers learning and using a language and the tasks they are mastering) are based<br />
on two cognitive components of language processing. These processing components<br />
are the basis for both assessing learners' abilities and determining task difficulty.<br />
The processing components are called analysis of knowledge and control of<br />
processing. Each component is specialized for a different aspect of the complex<br />
process of learning and using a language. Moreover, each processing component<br />
develops with experience and maturity on its own course. The development of these<br />
two processing components is normally correlated, as children generally move in a<br />
roughly synchronous manner toward greater mastery of each. Some situations,
48 Cognitive Approaches to Interlanguage Pragmatic Development<br />
however, accelerate the development of one of them and create an imbalance in the<br />
learner's ability to process language. These imbalances would either enable children<br />
to succeed or prevent them from performing in language use situations that would<br />
normally be expected for children whose development followed the more usual<br />
course.<br />
Analysis of knowledge is the process of making explicit, or analyzing, a<br />
learner's implicit knowledge of a domain. As a result of the process of analysis,<br />
mental representations of a domain of knowledge become both more explicit and<br />
more organized around formal (abstract) categories. For language, knowledge of the<br />
linguistic system becomes explicit in terms of the constituents and categories that<br />
make up the language (words, letters, phonemes) and in terms of the rules that<br />
govern the structure and combination of those categories (subject-verb-object—<br />
SVO, negation). None of this explicitness is necessary for the uses of language<br />
undertaken by young children. One does not need explicit access to these concepts<br />
and rules in order to produce grammatical sentences or understand conversation.<br />
Once a domain has been analyzed, that knowledge can be used for functions not<br />
supported by implicit representations. The behavioral outcome of high levels of<br />
analysis is the ability to articulate structural principles of organization for the<br />
domain. At the same time, new operations become possible only with greater levels<br />
of analysis. Advanced or specialized uses of language, such as literacy skills, are<br />
based on more analyzed representations of language than are necessary for children's<br />
earlier uses of the system. The construct of analysis is compatible with<br />
theoretical notions of linguistic structuring by children posited by Berman (1986),<br />
Bowerman (1982), Karmiloff-Smith (1986), and others as well as the explication of<br />
implicit representations in other cognitive domains, such as classification, as described<br />
by Gelman and Baillargeon (1983).<br />
Control of processing is the process of controlling attention to relevant and<br />
appropriate information and integrating those forms in real time. Language presents<br />
multiple sources of information, both linguistic and nonlinguistic, and part of effective<br />
language processing is being able to attend to the required information without<br />
being distracted by irrelevant or misleading cues. Oral language, for example,<br />
provides potential information about meanings as well as formal information about<br />
the words and structures chosen, the style of speech used, and features of the<br />
pronunciation. Carrying on a conversation, however, requires that attention be<br />
directed primarily to the meanings.<br />
Donaldson (1978) contends that children need to develop control over attentional<br />
procedures to succeed in school and to progress in many of the tasks usually<br />
associated with Piagetian operational thought. These include such tasks as classification<br />
and conservation. The construct is similar to the executive control schemes,<br />
such as those posited by Case (1985), which regulate the selection and coordination<br />
of lower level schemes responsible for problem solving. The difference between<br />
executive schemes and control in the present system is that control of processing is<br />
not hierarchically related to other schemes. Rather, it exists at the same level as a<br />
different but equivalently powerful aspect of processing (cf. Jackendoff, 1987). In<br />
this sense, control of processing is the selective attention to the relevant aspects of a<br />
representation for the purpose of carrying out a specific task.
Symbolic Representation and Attentional Control 49<br />
In my research into children's language development, I have investigated the<br />
role of the two processing components in the development of children's language<br />
proficiency. In this way, I have shown that children's progress with such activities as<br />
learning to read and developing metalinguistic awareness can be traced directly to<br />
their progress in mastering each of these component processes (e.g., Bialystok<br />
1986, 1988). When tasks (such as understanding written texts or performing specific<br />
metalinguistic judgments) are analyzed for the demands they place upon each of the<br />
processing components of analysis and control, the result is generally that children<br />
who have been shown by means of independent measures to possess sufficient levels<br />
of the relevant processing component are the children who are likely to succeed on<br />
the language task.<br />
Consider, then, the development of each of the processing components in somewhat<br />
more detail. Regarding the process of analysis, the changes in representation<br />
of language brought about through analysis of the linguistic system can be divided<br />
into three levels of representation (Bialystok, 1991), I have called these conceptual<br />
representation, formal representation, and symbolic representation. In conceptual<br />
representation, language is organized only around the meanings it represents. Language<br />
is transparent and representations of meanings are based on semantic and<br />
conceptual information carried through the language. This kind of representation is<br />
children's first access to language. They understand language in context almost<br />
regardless of the specific forms being used. Similarly, children are not aware of the<br />
words they use to express their intentions, and the categories they form are built out<br />
of the semantic properties of those intentions, not out of the formal means for<br />
expressing them. In this view, oral uses of language can proceed from conceptual<br />
representations.<br />
Formal representations are coded in terms of the structure of language. The<br />
categories that organize the representations are the formal categories of language<br />
itself: language is made of words and sounds, rules of order must be observed when<br />
forming sentences, and the like. Children have formal knowledge of language<br />
structure when they can identify letters and tell you what sound a letter makes, count<br />
the number of words in a sentence, and detect and possibly correct grammatical<br />
violations. This level, then, refers to explicit knowledge of language structure and<br />
corresponds most closely to what most investigators mean by metalinguistic knowledge.<br />
In symbolic representation, the coding is a relation between a form and a<br />
referent. Symbolic representation is an explicit accounting of the way in which<br />
language refers. The knowledge of letters coded at this level is a symbolic relation<br />
between a form (the letter) and the sound it signifies (the referent). The representation<br />
is in the form of the relation stands for. Words are symbols that stand for<br />
meanings, so a symbolic coding of a word explicitly includes this relational property.<br />
Symbolic representation involves understanding that letters and written words do<br />
not have meaning in themselves but rather signify words and concepts that do have<br />
meaning. It is an advance for children to move from understanding letters and<br />
sounds as objects (the letter B makes the sound "b") to understanding letters as<br />
symbols (the letter B stands for the sound "b"). Symbolic representation deals<br />
directly with the problem of how language refers. In these terms, children may
50 Cognitive Approaches to Interlanguage Pragmatic Development<br />
become literate only when their representations of linguistic knowledge have become<br />
analyzed to the point that the written forms for the language are understood as<br />
symbols and not merely as formal objects.<br />
Control of processing develops as well throughout childhood. As children move<br />
from using language as a strictly oral system to becoming acquainted with its<br />
written and formal properties, new attentional strategies are required. Reading, for<br />
example, demands a delicate balance between attention to the semantic and formal<br />
properties of the written text. Children's competence in early reading can be traced<br />
in part to their level of control of processing (Bialystok & Mitterer, 1987).<br />
If this model of language processing is applied to the development of pragmatic<br />
competence, then the abilities subsumed under descriptions of pragmatic competence<br />
must be stated in terms of their basis in levels of analysis and control. If these<br />
two processing components are responsible for language performance, then they<br />
must also be responsible for the pragmatic abilities of language users. Moreover,<br />
both components must be integrated into the account of pragmatic competence,<br />
even if it turns out that there is an imbalance in the levels of each that speakers must<br />
attain. Accordingly, two issues must be addressed. First, it is necessary to determine<br />
the ways in which representations of language must change to support the pragmatic<br />
functions of language and the level of analysis necessary for these uses. Second, it is<br />
necessary to determine the attentional strategies that must be developed to use<br />
language appropriately in contexts and the way in which these strategies relate to the<br />
ones normally used during conversation. These analyses will allow one to determine<br />
the level of proficiency within these two processing components necessary to demonstrate<br />
pragmatic competence and also permit a more detailed description of the<br />
processes that adult second language learners must master in order to achieve this<br />
competence.<br />
Processes in Pragmatic Competence<br />
Throughout the descriptions of children's mastery of the three aspects of pragmatic<br />
competence described in the first section, two factors were consistently cited in one<br />
way or another. First is the need for children to expand their linguistic resources.<br />
Shatz and McCloskey (1984), for example, attribute the improved ability of older<br />
children to interpret speech acts to their increased vocabulary. These children have<br />
developed a range of possible acceptable responses to various forms, and these<br />
forms are selected more appropriately on the basis of contextual cues.<br />
Second is the ability to select between two possible interpretations of an utterance.<br />
In the case of the indirect request given in (3), for example, the choice is<br />
between the literal meaning (Is your mother home) and the intended meaning (Call<br />
her to the telephone). Shatz and McCloskey (1984) make the following claim for<br />
development: "Children's earliest appreciation of speech acts, then, depends on<br />
their ability to attend to either contextual or surface features of an utterance. Their<br />
responses grow more sophisticated as their capacity to process information selectively<br />
from additional sources develops" (30).
Symbolic Representation and Attentional Control 51<br />
These two aspects of development correspond more or less to advances in<br />
analysis and control respectively. Shatz (1983) similarly points to this kind of dual<br />
achievement for an explanation of the development of pragmatic competence: "A<br />
theory of the acquisition of communicative competence, then, must make reference<br />
to children's representational and processing capabilities as well as to the content of<br />
their social and linguistic knowledge" (844). Consider, then, some of the processing<br />
demands made upon analysis and control by pragmatic competence.<br />
For pragmatic knowledge, conceptual representation is the stage in which utterances<br />
can be formed to satisfy specific speech acts but the speaker is focusing on<br />
the intended meaning and not on the forms being selected to express that intention.<br />
Formal representations are organized around categories. At this stage, speakers are<br />
able to classify a group of utterances as being requests or assertions and are aware of<br />
the relation between direct and indirect speech acts. The representation of utterances,<br />
that is, can be in terms of the illocutionary force of those utterances. For<br />
symbolic representation there is additionally an explicit representation for the relation<br />
between the pragmatic intention and the forms used to achieve that intention.<br />
Representations at this level could include the relation between explicit devices such<br />
as politeness markers and their modifying effect on the interpretation of utterances<br />
in specific contexts. At this explicit relational level of representation, the learner<br />
would also have direct access to a repertoire of alternative forms of expression that<br />
indicate more or less the same intentional content. The critical aspect of the representation<br />
at this level is not the explicit forms or speech acts known by the learner as<br />
was the case for the formal level, but rather the relation between these forms and the<br />
meaning they convey in contexts. Such representations would make explicit the way<br />
in which different forms, such as politeness markers, modify the social interpretation<br />
of the utterance.<br />
Pragmatic competence is an achievement that depends minimally on formal<br />
representations, but ideally on symbolic representations. The interpretation of<br />
meaning in contexts is inherently a relational problem. Specifically, it involves the<br />
relation between a set of linguistic forms and the meanings intended by those forms<br />
in specific contexts. The representation that underlies this performance consists of a<br />
relation between a given meaning and a range of possible forms that give rise to that<br />
meaning. Selecting the appropriate form requires an assessment of contextual and<br />
social factors. Thus the mapping is not between form and meaning—the usual<br />
problem in semantics—but between form and social context, with meaning held<br />
constant across intentions within a socially defined situation. Put this way, the<br />
problem for pragmatics is to develop the resource of equivalents from which selections<br />
can occur. The richer the repertoire, the greater would be the pragmatic<br />
competence.<br />
Children need to learn all of this. Children acquire pragmatic competence by<br />
building up their representations of the linguistic system through all three stages.<br />
First they simply understand meaning in contexts. Macnamara (1972) cited this<br />
ability to understand utterances in context regardless of forms as the very basis of<br />
language acquisition. Next they learn about different speech acts and different forms<br />
and realize that these signal different meanings. But the relation between these
52 Cognitive Approaches to Interlanguage Pragmatic Development<br />
forms and contexts is not worked out until this knowledge is represented symbolically.<br />
At this stage, selection from the formal alternatives can be deliberate and<br />
effective, and comprehension can be more accurate.<br />
Becker (1982) offers a description of pragmatic development based primarily on<br />
children's development of requests that is largely compatible with the view being<br />
described here. She points out that children's first requests are "unanalysed wholes"<br />
that eventually become decomposed to create new forms. Children build up more<br />
complete representations that allow them to manipulate the components of these<br />
wholes: "Toddlers can vary their requests systematically in several ways, which is<br />
only possible once they have acquired a more abstract system" (30). This "more<br />
abstract system," I would argue, is the formal or symbolic representation that is the<br />
result of the process of linguistic analysis.<br />
Children younger than 5 years old, according to Gordon and Ervin-Tripp (1984)<br />
do not, theoretically, have the processing ability to manage indirect requests, such<br />
as "It's cold in here." Yet, children under 5 years of age routinely produce and<br />
understand these forms correctly. How is this possible The solution offered by<br />
Gordon and Ervin-Tripp is that these children have learned direct mappings between<br />
forms and functions and can use and understand them in familiar activities and<br />
interactions. They argue that for these children, the form "it's cold in here" is not an<br />
indirect request at all, but a perfectly direct means of achieving a goal.<br />
Shatz (1978) addresses the same question and suggests that 2-year-olds' (appropriate)<br />
responses to such forms as indirect requests may be based more on primitive<br />
response strategies than on inferential analysis and propositional content. Children<br />
may respond on the basis of some prototypical meanings (Shatz & McCloskey,<br />
1984) that are suggested by the situation and not respond to the forms of language<br />
being used. Children may have heard sentences like "Can you talk on the telephone"<br />
in contexts in which action rather than a verbal answer was expected, creating a bias<br />
toward that response. This bias overestimates children's ability to interpret the<br />
illocutionary force of such indirect requests.<br />
On both these views, children's language competence is consistent with the<br />
argument that their knowledge of the pragmatic functions of language is based on<br />
conceptual representation, not formal or symbolic representation. Therefore, while<br />
children may appear to function appropriately, they cannot be described as having<br />
pragmatic competence. With conceptual representations, children are not even<br />
aware that the form is functioning as a request. When children come to understand<br />
that this form is one way of achieving a request, then their representation of that<br />
relation can be said to be formal. Symbolic representation would be signaled by the<br />
speaker's ability to deliberately select unusual or indirect forms, to manipulate<br />
aspects of the request such as the politeness or deference implied, or to understand<br />
the relation between the formal features of the utterance and its illocutionary function.<br />
Adult second language learners generally begin the task of mastering the pragmatic<br />
structure of the new language at the second (formal) level of representation<br />
and attempt to develop a symbolic representation from that system. Children begin<br />
with the problem of "how to mean," how to interpret indirect requests, how to string<br />
together discourse into conversation. Adults learning a second language have al-
Symbolic Representation and Attentional Control 53<br />
ready sorted out the nature of meaning and already have explicit formal categories<br />
for concepts corresponding to speech acts. They may even have explicit categories<br />
for the formal pragmatic markers, such as politeness terms, used in the second<br />
language. The problem for adults is to learn the symbolic relation between forms<br />
and contexts appropriate to the second language. This may also entail expanding<br />
formal linguistic resources by learning new forms of expression or new sets of<br />
equivalents for given expressions in much the same way that vocabulary is learned.<br />
Some forms can even be learned in terms of a social context: when in a restaurant<br />
we say it this way.<br />
The claim here is strong regarding the amount of learning that is a precondition<br />
to pragmatic competence for children. In this view, children's knowledge of language<br />
must be drastically restructured and organized around a new set of formal<br />
explicit categories before children can even develop the symbolic representations<br />
that underlie real pragmatic competence.<br />
Adult learners of a second language also need to develop more analyzed representations<br />
of the language. Culturally specific forms and rules for pragmatic language<br />
use may require some specific analysis by adult learners both in the sense of<br />
organizing implicit knowledge into new explicit categories and in the sense of<br />
increasing the repertoire of language structure by learning new forms. New categories<br />
may be necessary because the second language may make social distinctions not<br />
followed in the speakers' first language, distinctions relating, for example, to social<br />
status, age, or sex of the listener. More forms may be necessary because the second<br />
language may include a richer variety of alternatives for expressing the same intentions.<br />
Conversational features such as rules for turn-taking, interrupting, and opening<br />
and closing conversations may also require explicit learning and analysis by<br />
adult second language learners. At the same time, some languages may be more<br />
indirect, and speech acts (such as requests) may be more buried. Second language<br />
learners would have to learn all these pragmatic conventions and the forms for<br />
conveying them.<br />
The control issue for pragmatic competence is the problem of attention. In<br />
addition to the usual problem of a formal (surface structure) and semantic (meaning)<br />
representation of an utterance competing for attention, the difficulty of deciding<br />
between literal and metaphorical (or indirect) semantic interpretations must also be<br />
resolved. How do children learn to ignore the literal meaning and respond by<br />
attending to the intention of utterances<br />
In Clark's (1979) model for the interpretation of indirect meanings, indirect<br />
utterances give rise to two representations in listeners corresponding to the literal<br />
and intended meanings of the utterance. Adults have both interpretations simultaneously<br />
available and solve the problem of meaning by attending to one of them as<br />
a function of contextual factors. This selective attention may be especially difficult<br />
to execute in a second language where the conventions are less familiar. Cultural<br />
differences such as the use of indirect requests for action, the use of sarcasm and<br />
irony, and the forms for signaling politeness and deference may make the decision<br />
regarding the intended meaning difficult. Thus for adults the problem of correct<br />
interpretation is the responsibility of control of processing.<br />
Children, according to Shatz and McCloskey (1984), do not have two represen-
54 Cognitive Approaches to Interlanguage Pragmatic Development<br />
tations available for the interpretation of indirect speech. They rely exclusively on<br />
contextual conditions and do not, in general, form a representation for the literal<br />
meaning. This direct mapping between the utterance and a unique meaning in<br />
context is a consequence of the language being represented at the conceptual level.<br />
De Villiers (1984) puts the case somewhat more tentatively but makes essentially<br />
the same point. She argues that children do not have multiple mappings between<br />
forms and meanings and that their interpretation of meanings is based on their<br />
knowledge of familiar contexts and familiar utterances. Thus the control problem is<br />
reduced, as there are not two interpretations competing for attention. Intended<br />
meaning is resolved more on the basis of social-contextual circumstances than on<br />
the basis of linguistic forms.<br />
In summation, the problem presented to children learning their first language<br />
and adults learning their second language regarding pragmatic competence is different.<br />
For children, the primary task is the process of analysis. Children need to<br />
explicate and expand their linguistic resources to cope with the demands of using<br />
language for different purposes, in different contexts, and to different effects. Adults<br />
need to learn some social conventions for the new language and to solve some<br />
mapping problems between forms and social conditions, but the task of organizing a<br />
linguistic system along the lines of speech functions so that the system is operable in<br />
contexts has been solved for a first language. Adjustments in the organizing structures<br />
in which the first language is already represented as a formal or symbolic<br />
system carrying pragmatic functions are a relatively small problem.<br />
For adults, the problem to be solved for pragmatic competence is essentially to<br />
develop the control strategies to attend to the intended interpretations in contexts<br />
and to select the forms from the range of possibilities that satisfy the social and<br />
contextual needs of the communicative situation. Adults make pragmatic errors, not<br />
only because they do not understand forms and structures, or because they do not<br />
have sufficient vocabulary to express their intentions, but because they choose<br />
incorrectly. They attend to the wrong possibility for meaning, they fail to attend to a<br />
social distinction that needs to be marked linguistically, or they select the incorrect<br />
politeness marker for the situation or the listener. Many examples of the kinds of<br />
new selection strategies adults must learn when speaking a second language and the<br />
kinds of social and contextual errors they make while learning them are described in<br />
the research reported in the collection by Blum-Kulka, House, and Kasper (1989a).<br />
Pragmatic competence is achieved when control of processing is mastered for a<br />
richly analyzed representation of the language.<br />
Adults do, of course, still need to worry about the analysis problem. They need<br />
to continue to build up their repertoire of formal linguistic resources and to verify<br />
that their organization of the system has followed the correct categories. In my view,<br />
this problem is relatively minor; other views, however, lead to different conclusions.<br />
Blum-Kulka, House, and Kasper (1989b) describe two opposing views on<br />
this issue. Eraser (1985) argues that speech act markers are universal but have<br />
different distributions across languages and cultures, while Wierzbicka (1985) argues<br />
speech act markers are unique realizations created by languages and cultures.<br />
The prevailing opinion in the literature appears to be that of Fraser, and it is from<br />
that assumption that my argument, which attributes little analysis difficulty to adult
Symbolic Representation and Attentional Control 55<br />
second language learners, follows. If Wierzbicka is correct, the analysis problem<br />
for adults learning a second language is considerably exacerbated.<br />
Similarly, while children's greatest obstacle for mastering pragmatic competence<br />
is in developing analyzed representations of language, they still need to pay<br />
attention to control. Errors such as the one in (3) do occur, and children have to<br />
learn the processes for attending to the appropriate meaning possibility. Children<br />
use and understand indirect speech from an early age, and so the role of control in<br />
early language processing for pragmatic proficiency is tangible. The argument<br />
presented here, however, is that the greatest challenge for each group, children<br />
learning their first language and adults learning a second language, lies in a different<br />
component.<br />
This division, in which children face a more serious barrier from analysis and<br />
adults from control, is not accidental. The two processing components are intricately<br />
connected to each other, but are nonetheless ordered in their priority. They are<br />
connected because each entails the other. Analyzed representations carry with them<br />
the prospect of selective attention to aspects of a representation or deciding between<br />
competing representations. Yet attention is a vacuous concept in the absence of<br />
content, so control of processing presupposes representations. This is not a hopeless<br />
tangle; in my empirical work I have been able to measure learners' levels of mastery<br />
of these two processing components independently and show their separate involvement<br />
in language tasks. Thus, although proficient reading, for example, depends on<br />
both analysis and control, analysis is prior (Bialystok, 1988). Further research with<br />
the processing abilities of children and adults as they master the pragmatic functions<br />
of language will address these claims directly.<br />
References<br />
Ackerman, B. P. (1981). The understanding of young children and adults of the deictic<br />
adequacy of communication. Journal of Experimental Child Psychology, 31, 256-70.<br />
Becker, J. A. (1982). Children's strategic use of requests to mark and manipulate social<br />
status. In S. A. Kuczaj (Ed.), Language development (Vol. 2): Language, thought<br />
and culture (1-35). Hillsdale, NJ: Erlbaum.<br />
Berman, R. A. (1986). A step-by-step model of language learning. In I. Levin (Ed.), Stage<br />
and structure: Reopening the debate (191-219). Norwood, NJ: Ablex.<br />
Bialystok, E. (1986). Factors in the growth of linguistic awareness. Child Development, 57,<br />
498-510.<br />
Bialystok, E. (1988). Aspects of linguistic awareness in reading comprehension. Applied<br />
Psycholinguistics, 9, 123-139.<br />
Bialystok, E. (1991). Letters, sounds, and symbols: Changes in children's understanding of<br />
written language. Applied Psycholinguistics, 12, 75—89.<br />
Bialystok, E., & Mitterer, J. (1987). Metalinguistic differences among three kinds of readers.<br />
Journal of Educational Psychology, 79, 147-53.<br />
Bloom, L. (1973). One word at a time: The use of single word utterances before syntax. The<br />
Hague: Mouton.<br />
Bloom, L., Rocissano, L., & Hood, L. (1976). Adult-child discourse: Developmental interaction<br />
between information processing and linguistic knowledge. Cognitive Psychology,<br />
8, 521-52.
56 Cognitive Approaches to Interlanguage Pragmatic Development<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989a). Cross-cultural pragmatics: Requests<br />
and apologies. Norwood, NJ: Ablex.<br />
Blum-Kulka, S., House, J., & Kasper, G. (1989b). Investigating cross-cultural pragmatics:<br />
An introductory overview. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural<br />
pragmatics: Requests and apologies (1-34). Norwood, NJ: Ablex.<br />
Bowerman, M. (1973). Early syntactic development: A cross-linguistic study with special<br />
reference to Finnish. Cambridge: Cambridge University Press.<br />
Bowerman, M. (1982). Reorganizational processes in lexical and syntactic development. In<br />
E. Wanner & L. R. Gleitman (Eds.), Language acquisition: The state of the art (319-<br />
346). Cambridge: Cambridge University Press.<br />
Bruner, J. S. (1983). Child's talk: Learning to use language. New York: Norton.<br />
Case, R. (1985). Intellectual development: Birth to adulthood. New York: Academic Press.<br />
Clark, H. H. (1979). Responding to indirect speech acts. Cognitive Psychology, 11, 430-77.<br />
Clark, H. H., & Clark, E. V. (1977). Psychology and language. New York: Harcourt Brace<br />
Jovanovich.<br />
de Villiers, J. (1984). Form and force interactions: The development of negatives and questions.<br />
In R. L. Schiefelbusch & J. Pickar (Eds.), The acquisition of communicative<br />
competence (193-236). Baltimore, MD: University Park Press.<br />
Donaldson, M. (1978). Children's minds. Glasgow: Fontana.<br />
Ervin-Tripp, S. (1977). Wait for me, roller skate! In S. Ervin-Tripp & C. Mitchell-Kernan<br />
(Eds.), Child discourse (165-188). New York: Academic Press.<br />
Ervin-Tripp, S. (1989). Children's social concepts in language development. Invited address<br />
to the Society for Research in Child Development, Kansas City, MO.<br />
Fraser, B. (1985). On the universality of speech act strategies. In S. George (Ed.), From the<br />
linguistic to the social context (43-69). Bologna, Italy: CLUEB.<br />
Gelman, R., & Baillargeon, R. (1983). A review of some Piagetian concepts. In J. H. Flavell<br />
& E. M. Markman (Eds.), Handbook of child psychology, (Vol. 3): Cognitive development<br />
(167-230). New York: Wiley.<br />
Gordon, D., & Ervin-Tripp, S. (1984). The structure of children's requests. In R. L. Schiefelbusch<br />
& J. Pickar (Eds.), The acquisition of communicative competence (83-106).<br />
Baltimore, MD: University Park Press.<br />
Halliday, M.A.K. (1975). Learning how to mean. London: Arnold.<br />
Jackendoff, R. (1987). Consciousness and the computational mind. Cambridge, MA: MIT<br />
Press.<br />
Karmiloff-Smith, A. (1986). From metaprocess to conscious access: Evidence from children's<br />
metalinguistic and repair data. Cognition, 28, 95-147.<br />
Macnamara, J. (1972). The cognitive basis of language learning in infants. Psychological<br />
Review, 79, 1-13.<br />
McTear, M. F. (1984). Structure and process in children's conversational development. In<br />
S. A. Kuczaj (Ed.), Discourse development: Progress in cognitive development research<br />
(37-76). New York: Springer-Verlag.<br />
Schlesinger, I. M. (1971). Production of utterances and language acquisition. In D. I. Slobin<br />
(Ed.), The ontogenesis of grammar: A theoretical symposium (63-101). New York:<br />
Academic Press.<br />
Searle, J. R. (1979). A taxonomy of illocutionary acts. In J. R. Searle (Ed.), Expression and<br />
meaning (1-27). Cambridge: Cambridge University Press.<br />
Shatz, M. (1978). On the development of communicative understandings: An early strategy<br />
for interpreting and responding to messages. Cognitive Psychology, JO, 271-301.<br />
Shatz, M. (1983). Communication. In J. H. Flavell & E. M. Markman (Eds.), Handbook of<br />
Child Psychology, (Vol. 3): Cognitive Development. New York: Wiley, pp. 84-89.
Symbolic Representation and Attentional Control 57<br />
Shatz, M., & McCloskey, L. (1984). Answering appropriately: A developmental perspective<br />
on conversational knowledge. In S. A. Kuczaj (Ed.), Discourse development: Progress<br />
in cognitive development research. New York: Springer-Verlag, pp. 19-36.<br />
Wierzbicka, A. (1985). Different cultures, different languages, different speech acts. Journal<br />
of Pragmatics, 9, 145-78.
This page intentionally left blank
II<br />
SPEECH ACT REALIZATION<br />
This section includes five studies of interlanguage speech act performance:<br />
thanking, apologizing, complaining, requesting, and correcting. While there is a<br />
large literature on requests and only somewhat less on apologies, the other three<br />
speech acts have not yet been studied much. Consequently, for the investigation of<br />
requests and apologies, there is a much richer empirical and theoretical foundation<br />
for interlanguage pragmaticists to work from.<br />
According to various semantic criteria, the five speech acts form intersecting<br />
groups. In Searle's (1976) typology, thanking, apologizing, and complaining represent<br />
expressive speech acts; requesting is a directive act; and correcting, a representative<br />
act. Following Leech's (1983) functional classification, thanking and apologizing<br />
are convivial, while complaining, requesting, and correcting are competitive<br />
acts. While these groupings may not seem to be of much interest beyond speech act<br />
theory, they have behavioral correlates in speech act realization patterns, suggesting<br />
that there is indeed a relationship between speech act theorists' analyses and language<br />
users' communicative practice. Because illocutionary and social goals are<br />
compatible in the case of convivial acts, thanking and apologizing tend to be<br />
aggravated in force. Complaining, requesting, and correcting tend to be mitigated,<br />
reflecting the tension between illocutionary and social purpose. However, these are<br />
rough generalizations: as has frequently been pointed out in the literature, not only<br />
would it depend on the context of a request whether and how much it needs to be<br />
mitigated, it might even be perceived as quite consistent with the hearer's facewants;<br />
for instance, if the very fact of being asked to do something, even if involving<br />
some effort on the part of the speaker, conveys appreciation of the hearer's<br />
competence or attributes.<br />
Indirectness and conventionalized means of implementation are common features<br />
of most linguistic action, yet speech acts differ in the extent to which fixed<br />
conventionalized expressions are routinely used. Since thanking and apologizing are<br />
regularly implemented by means of standardized routines (though, obviously, not<br />
only by them), learners not only have to find out when those acts are appropriate or<br />
even required in the target community but also what formulas need to be used, and<br />
in conjunction with what other means of expressing gratitude or regret. In case of<br />
acts that are contextually perceived by the speaker as posing a threat to the hearer's<br />
face, such as the competitive acts under study in the following chapters, the main<br />
decisions for learners (as well as native speakers) to make are the ones formulated<br />
59
60 Speech Act Realization<br />
by Brown and Levinson (1987): whether to carry out the face-threatening act at all,<br />
and if so, whether to do it in an on-record or some kind of off-record fashion. The<br />
literature documents considerable cross-cultural variability and learner-specific performance<br />
patterns on these issues (see the Introduction).<br />
Eisenstein and Bodman (Chapter 3) examine the expression of gratitude by<br />
native and nonnative speakers of English, representing a variety of linguistic and<br />
cultural backgrounds. Their study combines data from a number of sources: observation<br />
of naturalistically occurring events, role-plays, oral and written production<br />
questionnaires, and retrospective interviews. LI data were obtained using written<br />
questionnaires. Learners' responses were rated for appropriateness, using native<br />
speakers' thanking performance as baseline data. Differences between learners' and<br />
native speakers' expressions of gratitude were noticeable at the sociopragmatic and<br />
pragmalinguistic level. Sociopragmatic differences involved the kinds of events and<br />
interlocutor relationships that require thanking, as well as the learners' familiarity<br />
with the situational context. Thanking behavior was further influenced by the<br />
culture-specific prominence of values such as modesty, gratefulness, and indebtedness.<br />
While the learners' performance was pragmalinguistically successful in contexts<br />
that required simple, ritualized responses, they did much less well when more<br />
complex, creative expressions were called for. Even the advanced learners often did<br />
not fluently command the range of conventions of means and forms by which the<br />
native speakers expressed gratitude.<br />
Because Eisenstein and Bodman used a variety of data types, the impact of these<br />
instruments on the subjects' responses could be assessed. Questionnaires, roleplays,<br />
and naturalistic data exhibited the same types and wording of semantic<br />
formulas. However, not only did the interactive conditions produce more linguistic<br />
activity in terms of restatements and reformulations, they also emphasized the role<br />
of the giver in the speech event: it was through her conversational contributions that<br />
the receiver's choice of specific semantic formulas was prompted. Thanking thus<br />
appeared as a jointly developed event, involving mutually adjusted contributions by<br />
both participants.<br />
One of the issues demonstrated by Eisenstein and Bodman's chapter is the crosscultural<br />
variability of contextual factors in the event of thanking. Bergman and<br />
Kasper (Chapter 4) focus on the assessment of contextual factors in situations where<br />
the speaker has committed some offense, and the way in which contextual assessment<br />
influences informants' selection of apology strategies. Informants were Thai<br />
learners of English and native speakers of American English and Thai. Data were<br />
collected by means of assessment and production questionnaires. Informants largely<br />
agreed in their perception of the relationship between context-external factors (social<br />
distance and dominance) and context-internal factors (severity of the offense,<br />
offender's obligation to apologize, the likelihood for the offended party to accept the<br />
apology, and offender's loss of face). However, when it came to assessing individual<br />
offense contexts, cross-cultural differences prevailed, Thai and American respondents<br />
differing most in their perception of the offender's obligation to apologize. Of<br />
the apology strategies, aggravating the apology was most sensitive to contextinternal<br />
factors. Respondents assumed more explicit responsibility for the offense<br />
the closer they perceived the relationship between the interlocutors. Social distance
Speech Act Realization 61<br />
thus was important in apology performance. Status, on the other hand, was unrelated<br />
to learners' and native speakers' strategy choice. While learners and English<br />
native speakers displayed the same frequency patterns in expressing explicit apology<br />
and responsibility as well as in their upgrading of apologetic force, the learners<br />
made more use of context-dependent strategies such as downgrading responsibility<br />
and severity, offering repair, and different forms of verbal redress. In more than half<br />
of the cases, the difference between learners' and American English native speakers'<br />
strategy suppliance could be traced to sociopragmatic transfer from Thai.<br />
Bergman and Kasper discuss the learners' oversupplying, relative to the target<br />
norm, of context-dependent strategies as yet another instance of the verbosity phenomenon<br />
noted in some of the interlanguage pragmatics literature. Evidence is<br />
accumulating that verbosity is very likely to be instrument-induced. While learners'<br />
responses to (written) production questionnaires, such as in Bergman and Kasper's<br />
study, are typically characterized by verbosity, their contributions to interactive<br />
tasks such as role-plays are not. In fact, Eisenstein and Bodman's (Chapter 3) study<br />
shows that learners' expressions of gratitude in role-plays are 50% shorter than those<br />
of native speakers of English. These output differences reflect differential cognitive<br />
demands placed on learners in the two conditions, and highlight the research purposes<br />
to which the two types of data collection instruments might best be put.<br />
Production questionnaires serve to elicit knowledge displays without making demands<br />
on learners' fluency or interactional skills. Role-plays, on the other hand,<br />
require the ability to compute contextual factors and assemble relevant linguistic<br />
material in a highly automatized fashion. Because they make simultaneous demands<br />
on learners' comprehension and production systems, they are useful tools for probing<br />
learners' ability to instantiate sociopragmatic and pragmalinguistic knowledge<br />
in interaction.<br />
Apologies and complaints may occur in response to the same event: apologies as<br />
remedial activity produced by the offender, and complaints as expression of disapproval<br />
voiced by the offended party. Olshtain and Weinbach (Chapter 5) examine<br />
cross-cultural and situational variability in the choice of complaining strategies, and<br />
the interaction of contextual assessment with strategy choice. Data were collected<br />
by means of the same types of instrument as in Bergman and Kasper's apology<br />
study: production questionnaires to elicit responses to offensive events, and assessment<br />
questionnaires to probe for informants' perceptions of contextual factors.<br />
Olshtain and Weinbach's chapter reports on three studies: a study of complaint<br />
performance by native speakers of Hebrew, a cross-cultural comparison of complaining<br />
by native speakers of Hebrew and of British and American English, and an<br />
interlanguage study, comparing complaint realization by nonnative speakers of<br />
Hebrew at intermediate and advanced proficiency levels with that of native speakers.<br />
The responses by the native speakers of Hebrew were found to cluster around<br />
three strategies: expressing disapproval, voicing an explicit complaint (such as "you<br />
should have done x"), and giving a warning. Strategy choice was determined by<br />
interlocutors' status relationship, low-high constellations prompting less severe<br />
complaints; equal and high-low relationships, more severe ones. The cross-cultural<br />
comparison yielded highly consistent response patterns across the three native<br />
speaker groups. An important feature to note is that although informants were given
62 Speech Act Realization<br />
the choice to opt out, two-thirds of the respondents did choose to complain. It will<br />
be a matter for future studies to examine whether members of more<br />
interdependence-oriented cultures choose to abstain from complaining more frequently.<br />
Learners' complaints differed from those of the target group on all selected<br />
measures. The nonnative speakers produced longer complaint utterances, chose<br />
more severe complaining strategies, and used both more softeners and intensifiers.<br />
Strategy choice was influenced by interlocutors' relative status, social distance, and<br />
the hearer's obligation to have avoided the offensive act. Both groups produced the<br />
most talk when no explicit hearer obligation existed, and the least when there was a<br />
clear contractual violation. Increasing obligation resulted in more severe strategies<br />
in both groups. However, the learners produced longer utterances when the hearer's<br />
obligation was implicit, and they opted for more severe strategies than the native<br />
speakers when an explicit obligation had been violated. Just as in Eisenstein and<br />
Bodman's findings, the learner responses in this study, too, displayed more variability,<br />
suggesting that these nonnative speakers were not quite accustomed to target<br />
conventions of means and forms.<br />
Conventionality of means and forms has predominantly been studied in the<br />
context of requests. In a variety of request contexts, native and nonnative speakers<br />
of different languages were shown to prefer conventionalized indirect request strategies.<br />
In interlanguage pragmatics in particular, little attention has been given to<br />
indirectness in requests that is not conventionalized in terms of particular illocutionary<br />
force indicating frames (though conventionality of means is extant in indirect<br />
requests, too). Weizman (Chapter 6), therefore, in her study on requestive hints,<br />
makes an important contribution to understanding interlanguage requesting. Since<br />
previous research has shown hints to be less polite than conventionally indirect<br />
requests, their prime function is not one of politeness. Rather, their inherent opacity<br />
(opacity of illocution, proposition, or both) enables speaker and hearer to exploit<br />
their deniability potential. Based on this functional analysis, Weizman hypothesized<br />
that learners make more use of hints than native speakers, employing them as a<br />
communication strategy in compensation for the lack of conventionalized forms.<br />
Comparison of the hinting patterns in native and nonnative speakers' responses to<br />
eight request contexts in a discourse completion questionnaire did not reveal greater<br />
reliance on hinting by learners than by native speakers, hence no support for the<br />
communication strategy hypothesis. Length of residence in the target community<br />
had no effect on the amount of hints used, suggesting that learners possess the<br />
ability to employ nonconventional indirectness from early on. The nonnative speakers<br />
matched native speaker options in the contextual distribution of hints and preference<br />
for substrategies. Both groups preferred hints with the greatest illocutionary<br />
opacity, confirming the deniability potential hypothesis. The nonnative speakers,<br />
however, employed more redundant combinations, suggesting that learners invest<br />
more linguistic activity in reaching their communicative goals even when hinting.<br />
This part concludes with a study by Takahashi and Beebe (Chapter 7) on corrections<br />
performed by Japanese learners of English in comparison with native speakers<br />
of Japanese and American English. In their previous studies of face-threatening acts<br />
carried out by the same groups of native and nonnative speakers, the authors had<br />
found distinctive patterns of style shifting according to interlocutor status. Focusing
Speech Act Realization 63<br />
on the modification of corrections by means of positive remarks and softeners, the<br />
Japanese learners' style-shifting patterns were clearly influenced by transfer from<br />
Japanese. While Japanese learners, reflecting native sociopragmatic norms, styleshifted<br />
more than American respondents when refusing, contradicting and disagreeing,<br />
this study indicated dramatic style-shifting in the American speakers' use of<br />
positive remarks. Their prevalent use of positive remarks in the high-low condition,<br />
which was not matched by the Japanese learners or native speakers, provides more<br />
evidence of a positive politeness orientation in American interaction, and greater<br />
emphasis on status congruence in Japanese conversational behavior. The study also<br />
confirms Beebe and Takahashi's earlier claim that pragmatic transfer prevails in<br />
higher proficiency learners.<br />
The five chapters on interlanguage speech act realization highlight a number of<br />
important issues: (1) even quite proficient learners tend to have less control over the<br />
conventions of forms and means used by native speakers in the performance of<br />
linguistic action; (2) differences between learners' and native speakers' sociopragmatic<br />
perceptions of comparable speech events are systematically related to differences<br />
in their speech act performance; (3) transfer at the pragmalinguistic and<br />
sociopragmatic level persists at higher levels of proficiency; (4) learners produce<br />
more speech than native speakers when the task is less demanding on their control<br />
skills; and (5) researchers must pay close attention to the constraints of different data<br />
collection instruments on learners' performance.<br />
References<br />
Brown, P., & Levinson, S. C. (1987). Politeness. Cambridge: Cambridge University Press.<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
Searle, J. (1976). A classification of illocutionary acts. Language in Society, 5, 1-23.
3<br />
Expressing Gratitude<br />
in American English<br />
MIRIAM EISENSTEIN and JEAN BODMAN<br />
Expressing gratitude has important social value in American English. One indication<br />
of its importance is that it is one of the few functions that most speakers can<br />
remember being explicitly taught as children. Used frequently in a wide range of<br />
interpersonal relationships, this function, when appropriately expressed, can engender<br />
feelings of warmth and solidarity among interlocutors. Conversely, the<br />
failure to express gratitude adequately can have negative consequences for the<br />
relationship of speaker and listener.<br />
Expressions of gratitude can range from simple, phatic utterances to lengthy<br />
communicative events mutually developed by both the giver and the recipient of a<br />
gift, favor, reward, or service. Most native speakers of English on a conscious level<br />
associate the expression of gratitude with the words "thank you"; however, they are<br />
unaware of the underlying complex rules and the mutuality needed for expressing<br />
gratitude in a manner satisfying to both the giver and recipient. Similarly, second<br />
and foreign language learners are unaware of the underlying rules for expressing<br />
gratitude in English; in fact, they usually assume that the expression of gratitude is<br />
universal and remain unaware of significant differences in its cross-cultural realization<br />
(Eisenstein & Bodman, 1986). Because of this, the function of expressing<br />
gratitude is particularly difficult for learners to perform successfully.<br />
Background<br />
The importance of social interaction rituals, such as expressing gratitude, was<br />
addressed by Goffman (1967). He explained that every individual has a potential<br />
emotional response to others which is related to his or her "face," defined as "an<br />
image of self, delineated in terms of approved social attributes" (5). Violation of<br />
social norms may result in loss of face with accompanying negative feelings on the<br />
part of one or both participants. Brown and Levinson (1987) identify the desire for<br />
approval as "positive face" and define "negative face" as "the desire to be unim-<br />
64
Expressing Gratitude in American English 65<br />
peded in one's actions" (13). Some acts, verbal and nonverbal, may be counter to<br />
the "face wants" of speaker or hearer and, therefore, are "face-threatening acts"<br />
(65). Brown and Levinson categorize expressing thanks as a face-threatening act in<br />
which the speaker acknowledges a debt to the hearer—thus threatening the speaker's<br />
negative face. Searle (1969) stresses the positive aspects of thanking, which he<br />
defines as an illocutionary act performed by a speaker based on a past act performed<br />
by the hearer that was beneficial. Leech (1983) describes thanking as a convivial<br />
function whose goal of stating appreciation helps maintain a polite and friendly<br />
social atmosphere. Thus, it is apparent that expressing gratitude is a complex act<br />
potentially involving both positive as well as negative feelings on the part of giver<br />
and receiver.<br />
Thomas (1983) identifies the difficulty encountered by nonnative speakers in the<br />
cross-cultural realization of speech acts. She notes that misunderstandings can arise<br />
not only from language limitations (pragmalinguistic failure) but also from inadequate<br />
utilization of social conventions and values in the target culture (sociopragmatic<br />
failure). Coulmas (1981) views thanking in cross-cultural perspective and underscores<br />
the challenge for speakers of European and Asian languages to express<br />
thanking adequately to each other. He posits a useful distinction between thanks that<br />
entail indebtedness to the addressee and thanks that imply no indebtedness.<br />
It is useful to keep in mind that not all expressions using the words "thank you"<br />
refer to gratitude. Rubin (1983) collected natural data on uses of the words "thank<br />
you" that referred not only to gratitude but also to other language functions, such as<br />
compliments and closings. In fact, Hymes (1971) states that "thank you" as it is<br />
used in British English is often more of a formal marker than an expression of<br />
gratitude, which is its more common function in American English.<br />
Experiment 1<br />
We began investigating how gratitude is expressed by observing its use in natural<br />
contexts by native speakers of American English. Data were either audiotaped and<br />
transcribed at a later time or written down as field notes. We focused on only those<br />
utterances whose illocutionary force was that of gratitude in response to receiving a<br />
gift, favor, reward or service. 1 Fifty situations were identified in which expressions<br />
of gratitude occurred. These ranged from short, formulaic expressions that appeared<br />
to be highly ritualized to intricate and lengthy interchanges conveying deeply felt<br />
emotions. As a result of a series of pilot studies, we constructed a questionnaire<br />
containing fourteen of these situations designed to elicit expressions of gratitude.<br />
The pilot studies revealed the importance of describing the roles and relationships of<br />
the interlocutors in the questionnaire in detail as well as carefully describing specific<br />
contexts for the situations; we found that small changes in these variables significantly<br />
affected the nature of the responses. Fifty-six native speakers of American<br />
English, representing both males and females, a range of ages, diverse social<br />
backgrounds, and natives of a variety of regions with the United States, were asked<br />
to write responses to each of the fourteen situations.<br />
The written data elicited by the questionnaire were analyzed by coding each
66 Speech Act Realization<br />
utterance in terms of its underlying speech act. For example, on leaving a dinner<br />
party, a guest said:<br />
Thank you for inviting me. I had a great time.<br />
We coded this utterance as "Thanking + Expressing Pleasure." Wherever possible,<br />
we used functional categories described in the literature (van Ek, 1976, Searle,<br />
1969); however, in some instances we had to create our own tentative terminology<br />
where appropriate descriptors had not been previously identified. For example, after<br />
opening a gift, a subject responded:<br />
Oh, how beautiful. How did you know It's just what I wanted!<br />
The italicized utterances were clearly not intended to express the functions of<br />
Asking for Information or Expressing Need. We tentatively coded them as "Expressing<br />
Intimacy: Mind Reading," an indirect compliment acknowledging the accuracy<br />
of the giver's understanding of the receiver's unexpressed desires.<br />
The fourteen questionnaire items are briefly summarized in Appendix 1. Some<br />
of these items produced phatic, ritualized responses, such as brief comments to a<br />
bus driver (Item 2), a cashier (Item 5), a garage attendant (Item 13), and a friend<br />
handing over a newspaper (Item 12). In post hoc interviews, native American<br />
English speakers characterized utterances, such as<br />
Thanks.<br />
Thank you. Have a nice day.<br />
as virtually automatic. The expression of gratitude in situations such as these appears<br />
to be a social amenity rather than a genuine expression of gratitude and has<br />
been identified by Rubin (1983) as the "bald thank-you."<br />
Some other items on the questionnaire produced relatively short, but more<br />
creative responses than those described above. Swift thanking followed by a single<br />
brief comment was typical for Item 6 (a friend bringing attention to a bit of food on<br />
a diner's face) and Item 8 (thanking a spouse for spontaneously helping around the<br />
house). Typically, our respondants felt (although a great more could have been said)<br />
that comments like<br />
Thanks. You're a sweetheart.<br />
(coded as Thanking + Expressing Affection) and<br />
Thanks. That was really nice of you.<br />
(coded as Thanking + Complimenting the Giver) were sufficient in recognizing a<br />
spouse's thoughtfulness.<br />
The remaining items elicited and seemed to require much more complex and<br />
lengthy expressions of gratitude. These were most successfully analyzed as speech<br />
act sets (Cohen & Olshtain, 1981), groups of semantic formulae that together<br />
achieve the appropriate language for a particular situation. In appreciation for a<br />
generously offered $500 loan, a characteristic response was:<br />
You're a lifesaver. Thanks. I'll never forget it. You really can't imagine what this<br />
means to me.
Expressing Gratitude in American English 67<br />
(Coded as Complimenting the Person/Action + Thanking + Expressing Indebtedness<br />
+ Expressing an Inability to Articulate Deep Feelings). Our respondents, when<br />
interviewed later, indicated that in these situations gratitude was much more challenging<br />
to express. It was not uncommon for the respondents to state with humility<br />
that their linguistic skills were inadequate for the task. American native speakers,<br />
rating the appropriateness of the utterances, found humbling admissions such as this<br />
not only expressed the depth of the feelings in a satisfactory manner, but also<br />
adequately expressed gratitude.<br />
It is useful to consider two characteristics that Goffman (1967) describes that are<br />
operant in social interactions—"demeanor" and "deference"—in order to understand<br />
the complex linguistic task speakers face. "Demeanor" refers to the social<br />
desirability of an individual reflected through his or her appearance and behavior.<br />
"Deference" is the appreciation an individual exhibits to another through his or her<br />
words and actions. Expressing gratitude requires that the recipient of a gift, favor, or<br />
service exhibit both proper demeanor and proper deference in situations in which he<br />
or she is feeling especially vulnerable. The recipient must show humility and<br />
gratitude without losing dignity and control. The giver must remain sensitive to the<br />
needs of the receiver and also behave with adequate deference and demeanor. The<br />
struggle to find the words and exhibit acceptable demeanor and deference that are<br />
mutually satisfactory is highly challenging. In addition, the difficulty of the task is<br />
further compounded by the fact that both demeanor and deference in expressing<br />
gratitude are culturally bound and, hence, difficult to translate from one sociolinguistic<br />
context to another.<br />
The lengthiest speech act sets were produced by situations that caused the<br />
recipient to feel unusually grateful or indebted to the giver. None of the individual<br />
semantic formulas constituting a set could be identified as more salient or of a<br />
higher order than the others. Instead, the members of each set interacted synergistically<br />
to express gratitude appropriately. Furthermore, the functions within the<br />
speech act set did not appear in a fixed order. The direct expression of thanking, for<br />
example, could be stated at the beginning, the middle, or the end of the set.<br />
While greater emotion sometimes provoked longer speech act sets, this did not<br />
occur when there was considerable social distance between interlocutors. Item 4 (a<br />
vice president of personnel offers a relatively new employee a raise) elicited surprisingly<br />
brief expressions of gratitude. This confirms one aspect of Wolfson's<br />
Bulge Theory (1989), identifying brevity in communications between socially distant<br />
interlocutors. However, we did not find in all cases the terseness that Wolfson<br />
(1989) identified in communications among intimates. While the situation in which<br />
a spouse is helpful around the house did produce a brief response, the situations<br />
involving gift giving and loaning a large amount of money evinced quite lengthy<br />
responses despite the intimacy of the interlocutor relationship.<br />
Our data revealed that, although speakers are free to say anything they wish in<br />
expressing gratitude (and there were occasional examples of highly creative speech<br />
act sets), most speakers seemed to draw from a finite pool of conventionalized<br />
expressions and ideas. Within the speech act set of expressing gratitude for a gift,<br />
many native respondants referred to the lack of necessity for such generosity by<br />
using expressions such as "Oh, you shouldn't have" and "You didn't have to."
68 Speech Act Realization<br />
In accepting the loan, native speakers stated their inability to express their<br />
appreciation sufficiently:<br />
God, I don't know how to thank you.<br />
I can't tell you how much I appreciate this.<br />
1 can't thank you enough.<br />
I can't tell you what this means to me.<br />
Experiment 2<br />
The same questionnaire that was administered to the native speakers of American<br />
English was given to 67 nonnative college students in advanced-level ESL classes.<br />
They were middle-class and represented fifteen language backgrounds—the largest<br />
groups were Chinese, Korean, Japanese, Russian, and Spanish. The majority of<br />
these learners had been in the United States for approximately two years, although<br />
the range was as low as three months and as high as five years. The nonnative<br />
speakers were asked first to respond to the items in the questionnaire in English.<br />
Subsequently, they were given the opportunity to respond to the same situations in<br />
their native languages if they wished. Twenty-five of the subjects provided us with<br />
LI responses to the situations. (Bilingual speakers were asked to provide literal and<br />
figurative translations of the LI responses.) This allowed us to consider instances of<br />
cross-linguistic influence.<br />
Using native speaker data as a baseline, we rated nonnative written responses.<br />
Rating descriptors can be found in Table 3.1. The nonnatives were successful on the<br />
whole in responding in a nativelike manner to situations requiring simple, phatic,<br />
ritualized expressions of gratitude. However, items requiring complex speech act<br />
sets (Items 1, 3, 4, 7, 9, 10, 14) were problematic for many of our nonnative<br />
subjects. They were not able to approximate acceptable native speech acts in a<br />
significant number of cases. In fact, their difficulties in adequately expressing<br />
gratitude in a nativelike manner were extensive and severe. (See Table 3.2 for a<br />
summary of ratings on nonnative responses.)<br />
Table 3.1. Acceptability Scale for Nonnative Responses<br />
Not acceptable<br />
Problematic<br />
Acceptable<br />
Nativelike perfect<br />
Not comprehensible<br />
Resistant<br />
A violation of a social norm, a faux pas, a likely instance of sociopragmatic<br />
failure.<br />
An error that might cause misunderstanding, but of a less serious nature.<br />
Language so strange, unexpected, or garbled that interpretation is difficult.<br />
Instances of pragmalinguistic and/or sociopragmatic failure.<br />
Clear and appropriate language, but containing small errors that do not interfere<br />
seriously with native speakers' understanding.<br />
Close to native responses in content, syntax, and lexicon.<br />
An utterance that is extremely hard, if not impossible, to comprehend.<br />
Often an instance of pragmalinguistic failure.<br />
Nonnative participants, although finding it possible to answer some items,<br />
refused to answer others, or gave reasons why they could not or would<br />
not answer particular items.
Expressing Gratitude in American English<br />
69<br />
Table 3.2. Summary of Results of All Students on Individual Questions<br />
Question<br />
Question<br />
topic<br />
No<br />
response<br />
Not<br />
acceptable<br />
Problematical<br />
Acceptable<br />
Perfect<br />
NC*<br />
Resistant<br />
1<br />
3<br />
$5.00<br />
Sweater<br />
5<br />
7.5%<br />
2<br />
3.0%<br />
3<br />
4.5%<br />
3<br />
4.5%<br />
9<br />
13.4%<br />
19<br />
28.4%<br />
9<br />
13.4%<br />
14<br />
20.9%<br />
39<br />
58.2%<br />
29<br />
43.3%<br />
1<br />
1.5%<br />
0<br />
0%<br />
1<br />
1.5%<br />
0<br />
0%<br />
4<br />
7<br />
Raise<br />
$500.00<br />
5<br />
7.5%<br />
9<br />
13.4%<br />
10<br />
14.9%<br />
3<br />
4.5%<br />
16<br />
23.9%<br />
14<br />
20.9%<br />
14<br />
20.9%<br />
15<br />
22.4%<br />
18<br />
26.9%<br />
21<br />
31.3%<br />
1<br />
1.5%<br />
1<br />
1.5%<br />
3<br />
4.5%<br />
4<br />
6.0%<br />
9<br />
Lunch<br />
5<br />
3.0%<br />
11<br />
16.4%<br />
22<br />
32.8%<br />
9<br />
13.4%<br />
14<br />
20.9%<br />
1<br />
1.5%<br />
5<br />
7.5%<br />
10<br />
Farewell<br />
14<br />
20.9%<br />
7<br />
10.4%<br />
21<br />
31.3%<br />
11<br />
16.4%<br />
13<br />
19.4%<br />
0<br />
0%<br />
0<br />
0%<br />
14<br />
Dinner<br />
7<br />
10.4%<br />
2<br />
3.0%<br />
24<br />
35.8%<br />
23<br />
34.3%.<br />
11<br />
16.4%<br />
0<br />
0%<br />
0<br />
0%<br />
*Not comprehensible.<br />
In post hoc interviews, some nonnative speakers expressed a belief that they did<br />
not do well in responding to some situations. They attributed their shortcomings to a<br />
lack of real-life experience with some of the situations either because of their limited<br />
social roles (e.g., homemakers who had no experience in the business world) or<br />
because our situations lacked social reality in their native cultures (e.g., individuals<br />
who felt very uncomfortable with the situation involving the large loan). Although,<br />
in post hoc interviews, some native speakers expressed the same reservations, all<br />
were, nevertheless, willing and able to find the appropriate language to respond to<br />
the situations.<br />
Analysis showed that ease or difficulty in responding appropriately to particular<br />
items stemmed from linguistic factors as well as cultural ones. An analysis of poor<br />
responses indicated that nonnatives lacked the words and the syntax to work their<br />
way through culturally familiar and unfamiliar situations. Our data contained many<br />
examples of syntactic errors (e.g., "I consider change a job") and lexical problems<br />
including difficulties with intensifiers (e.g., "I very appreciate"), word order (e.g.,<br />
"I'll pay back you"), idioms (e.g., "Thank you. Sound is good."), prepositions<br />
(e.g., "I hope to see you by us"), and choice of words (e.g., "It is so glad to me that<br />
I have such kind of good friend").<br />
Table 3.3 shows the percentages of students whose responses were scored "Acceptable"<br />
and "Nativelike/Perfect." While Japanese speakers of English are generally<br />
thought to be very polite, they performed poorly on paper compared to the other<br />
groups in our study. Possible explanations for this include the lack of cultural<br />
congruity between Japanese and American cultures and the fact that written data do<br />
not reflect nonverbal cues and prosodic features which might serve to soften responses<br />
in real-life situations. The difficulty in cross-cultural communication between<br />
Japanese and English speakers is well documented in the research of Beebe<br />
and Takahashi (1989a,b). A revealing study by Coulmas (1981) shows that Japanese<br />
speakers have particular difficulty in English with those expressions of gratitude that
70 Speech Act Realization<br />
Table 3.3. "Acceptable" and "Native-like/Perfect"<br />
Ratings by Language Group<br />
Language<br />
background<br />
Japanese<br />
Spanish<br />
Korean<br />
Chinese<br />
Russian<br />
Acceptable and<br />
nativelike/perfect (%)<br />
30.2<br />
34.0<br />
36.0<br />
48.5<br />
66.6<br />
Nativelike/perfect (%)<br />
14.8<br />
20.0<br />
23.0<br />
28.1<br />
58.3<br />
imply indebtedness. In Japanese, Coulmas explains, an appropriate expression of<br />
gratitude acknowledges this debt by using the same expressions as would be required<br />
for an apology (e.g., Sore-wa kyoshuku desu, which literally means "I feel<br />
ashamed," can function as "thank you" or "I'm sorry" [89]). Brown and Levinson<br />
(1987) explain that, unlike England and the United States, Japan is a "debt-sensitive<br />
culture" in which thanks can be expressed, in effect, by saying, "I am humiliated, so<br />
awful is my debt" (247).<br />
By contrast, the relative success of the Russian participants in the study does not<br />
reflect the general American perception that Russians often express themselves<br />
inappropriately. Perhaps expressing gratitude is not one of the problem areas for<br />
them, or it is possible that judgments in this study reflect the limitations in the<br />
written channel, as might be the case with the Japanese participants. Interestingly,<br />
the Russian participants were most resistant in expressing themselves when they felt<br />
insecure or when they felt that the situations were not to be taken seriously. It may<br />
be that maintaining adequate demeanor played an important part in the behavior of<br />
these respondants. (See Eisenstein & Bodman, 1986, for more information on these<br />
data.)<br />
Experiment 3<br />
In our earlier work, discussed above, we noted that naturally occurring expressions<br />
of gratitude were different in some respects from our written data. 2 In an attempt to<br />
include the influence of intonation and prosodic features in our data, we administered<br />
the questionnaire orally by reading each situation to ten native speakers and<br />
taping their responses. Transcripts were almost identical to the language of the<br />
written questionnaires. Some prosodic features seemed to add to the perceived<br />
sincerity and effectiveness of the gratitude expressed. Emotion was conveyed by<br />
tone as well as words. The range of emotion was more evident in the audiotapes.<br />
Gratitude expressed to socially distant individuals was clearly reserved in tone in<br />
comparison with effusive warmth and enthusiasm evidenced between friends and<br />
intimates.<br />
In post hoc interviews, the native speakers stated that they had found the process<br />
awkward and had ambivalent feelings about the naturalness of their responses. They<br />
felt that they would say much more in a natural situation. In fact, this was corroborated<br />
by the frequent uneasy pauses in their responses. It was as though the subjects
Expressing Gratitude in American English 71<br />
were waiting for someone to reply to their initial expression of gratitude before they<br />
could continue. Dubin (personal communication) has noted similar difficulties with<br />
people leaving messages on telephone answering machines. This led us to postulate<br />
that expressing gratitude might not be merely a language function that existed as a<br />
response to a beneficial event, but might, instead, be a negotiated and interactive<br />
event that had greater social significance.<br />
Experiment 4<br />
In light of the importance to our subjects of interacting with the giver in order to<br />
express themselves fully and naturally in expressing gratitude, we used the same<br />
situations to set up role-plays that we hoped would reveal the interactive nature of<br />
how gratitude is expressed. In addition, during this period, we continued to collect<br />
naturally occurring examples. Some samples of role-plays and natural dialogues<br />
appear in Appendix 2. Our corpus consisted of 98 role-plays. Thirty-four were<br />
performed by native pairs, 40 by nonnative pairs and 24 by natives paired with<br />
nonnatives. The participants were literate, middle-class adults from a variety of<br />
social and ethnic backgrounds. Role-plays were recorded and transcribed for subsequent<br />
analysis. The procedures described in the earlier experiments were then<br />
followed.<br />
We found that the role-plays and natural situations incorporated the same words<br />
and semantic formulas that appeared in the written data. In this sense, our written<br />
data were representative of certain aspects of natural language use. For example, the<br />
following expressions occurred under all conditions in response to receiving a gift:<br />
It's beautiful.<br />
It's just what I needed.<br />
How did you know<br />
You shouldn't have.<br />
Yet the role-plays went beyond the written data by providing us with additional<br />
insights into the functions of thanking and responding to thanks as they exist in a<br />
conversational interaction. Similar findings comparing oral and written data have<br />
been reported by Beebe and Cummings (1985) and DeCapua and Dunham (1989).<br />
Just as role-play data gave us additional information through the inclusion of conversational<br />
turns, natural data was even richer and more revealing because it evidenced<br />
restatements of the same information using slightly different language and<br />
showed that the same function recurred over time.<br />
In analyzing the role-plays, we found that the language expressed by the giver<br />
(of the gift, favor, reward, or service) is crucial to enabling the receiver to convey<br />
gratitude successfully. The giver prompts and comments throughout the development<br />
of the speech act set. Prompts appear to function as linguistic enabling devices,<br />
allowing the receiver to reassure the giver of his or her gratitude. Examples of<br />
prompts are:<br />
1 hope it fits.<br />
1 hope you like it.<br />
How was your lunch . . . did you like it<br />
Are you sure it's enough
72 Speech Act Realization<br />
Comments seem to keep the conversation moving by filling a turn-taking opportunity<br />
and giving the thanker additional cues on how to thank the giver.<br />
That'll keep you warm.<br />
It should be washable. Machine washable.<br />
We're going to miss you.<br />
It was nice having your.<br />
There is evidence in our data that the act of giving may be perceived as disrupting<br />
the social equilibrium. Giving places the receivers in a position of obligation. The<br />
givers can mitigate the disequilibrium by downplaying the importance of the gift,<br />
favor, reward, or service (e.g., "Here's a little something I picked up for you").<br />
Some givers also attempt to downplay the seriousness of the event by introducing<br />
humor. They can also reassure receivers by stating that the act of giving brings<br />
them pleasure and by claiming that the gesture has caused them no inconvenience.<br />
The giver can have an important role in reassuring the receiver that it is all right to<br />
take. Note the following interchange from natural data regarding the $5.00 loan:<br />
1: Oh, no, I don't want to bother you.<br />
2: Really, I have plenty. Believe me, you can have it.<br />
Likewise, to restore the balance of the social relationship, the receiver needs to<br />
express the appropriate amount of appreciation to the giver. In our natural data, we<br />
noted that the receiver continued to thank the giver using a variety of strategies until<br />
the giver indicated that enough had been said. The giver often indicated his or her<br />
intent to end the thanking episode by abruptly nominating a change of topic.<br />
The nonnative speakers had difficulty replicating several aspects of the native<br />
model. Nonnative role-plays were often as much as fifty percent shorter than those<br />
of natives. In some instances, even the advanced nonnatives in our study lacked the<br />
language to express themselves more fully. (See Appendix 2, Dialogues 6 and 7.)<br />
This is in contrast to results reported by Blum-Kulka and Olshtain (1986). In<br />
attempting to find the right words, the advanced learners they studied said too much<br />
and were perceived as verbose. While our data showed that some individual turns of<br />
nonnatives were longer because of hesitations, cirumlocutions, and searches for the<br />
right words, the conversations taken as a whole were not long enough for natives to<br />
feel that gratitude had been adequately conveyed.<br />
Another aspect that nonnatives failed to replicate was the warm and sincere tone<br />
that natives conveyed when they expressed gratitude. In our first study, we questioned<br />
whether the nonnatives would be able to compensate for their lexical and<br />
syntactic difficulties by sounding and acting very appreciative. We found, unfortunately,<br />
that those students who lacked fluency and had to search continuously for<br />
words also spoke with an unconvincing tone. When native speakers were asked to<br />
judge these nonnative samples, they stated that the nonnatives sounded insincere.<br />
Those students who were more fluent, by contrast, had a more nativelike tone and<br />
sounded truly grateful.<br />
In our native data, there was a definite rhythm in the interactions of the interlocutors.<br />
In order for the gifts, services, rewards, or favors to be accepted successfully,<br />
a mutually agreed-upon "script" needed to be followed. As a result, there
Expressing Gratitude in American English 73<br />
was more frequent communication breakdown when two nonnatives conversed than<br />
when one member of the pair was a native. The nonnatives' language flowed better<br />
when they had the support of a native speaker. When a favor or a gift needed to be<br />
repaid, native speakers often explicitly stated their intention to do so as a part of<br />
their thanks. For example, in expressing gratitude for the small loan, a considerable<br />
number of natives were very specific about repayment: "Thanks a million. I'll pay<br />
you back next Monday." In some instances the intention to repay was stated in more<br />
general terms: "We hope you'll come over to our place sometime soon."<br />
Nonnatives sometimes did not express reciprocity or did so in an inappropriate<br />
manner; e.g., "Thank you very much for dinner. You will come to our house next<br />
week" In terms of the time mentioned, natives characterized this offer of reciprocity<br />
as being offered too specifically. They felt that the nonnatives were in too much of<br />
a hurry to pay them back. The overall offer, they felt, sounded too abrupt and<br />
demanding. Natives felt that they were being "put on the spot" to obligate themselves.<br />
In studying the groups and subgroups in our sample, it became apparent that<br />
cross-cultural differences in language values and customs all contributed to potential<br />
misunderstanding. These issues are addressed by Thomas (1983) in her discussion<br />
of pragmalinguistic and sociopragmatic failure. Furthermore, Blum-Kulka and<br />
Olshtain (1986) note that pragmatic principles are subject to both intercultural and<br />
intracultural variation. Some Anglo-American subjects said that they could not<br />
conceive of revealing to their friends that they were in need of money. Russian,<br />
Ukranian, Chinese, and German subjects stated that they, too, would find lack of<br />
money a source of shame and would have serious difficulty borrowing it from a<br />
friend. In receiving a gift, several groups indicated that they would probably put the<br />
gift aside and open it at a later time to avoid social embarrassment. Chinese participants<br />
explained that their cultural values demanded that an emphasis be placed on<br />
the presence of the guest rather than on a material object presented by the guest.<br />
Other Asian, Greek, and Anglo-American subjects indicated that they would probably<br />
say little about the gift and open it later when others were not present. However,<br />
some Americans and Hispanics informed us that it was customary for them to open<br />
a gift in front of the giver to share the enjoyment.<br />
Expressions of gratitude occur at the time of giving and, in our natural data, are<br />
found to reoccur later. For Anglo-American subjects, this reentry of thanks is as<br />
important as the original thanks, which would have to be performed out of politeness.<br />
A written note or phone call at a later date is not only highly valued but often<br />
expected. However, in traditional Jewish-American culture, important interactions<br />
tend to occur face to face; writing or calling are considered less meaningful. Our<br />
Polish and Puerto Rican subjects also preferred thanks to be conveyed in person.<br />
The cross-cultural complexity of this language function was also revealed by<br />
subjects who stated that expressions of gratitude offered in a context they considered<br />
inappropriate could be perceived as distancing, insulting, or rude. In many cultures,<br />
the words "thank you" are not commonly used to express appreciation to family<br />
members for acts of kindness considered part of their social roles. A Puerto Rican<br />
informant who had lived for many years in the United States described how hurt and<br />
angry her father became when she thanked him for helping her take care of her son,<br />
his grandchild. Her mother called berating her:
74 Speech Act Realization<br />
How could you have been so thoughtless You thanked your father. He was happy<br />
to take care of Johnnie. Have you forgotten how to behave He's your father and<br />
he loves you. How could you be so cold—to thank him<br />
Compliments relating to the father's kindness would have been better received.<br />
The opposite occurred in the case of an American family who welcomed a distant<br />
relative from Argentina into their home. After a month had passed during which the<br />
family had extended themselves personally and financially for their guest, they<br />
became increasingly annoyed that the relative rarely expressed appreciation for their<br />
efforts with the exception of an occasional terse "thank you." In Argentina, favors<br />
for family members are expected as part of the warmth and bonding of the family<br />
unit. While this is also true in American culture, the explicit statement of gratitude<br />
is required after each event. In many cultures it is felt that gratitude and the intent to<br />
reciprocate need not be expressed overtly. In fact, it might be considered insulting to<br />
do so. Among friends and relatives, reciprocity is taken for granted. In the Republic<br />
of Croatia in Yugoslavia, for example, thanking behavior must be judicious. If<br />
people are thanked excessively (more times than necessary or with inappropriate<br />
length or enthusiasm), this behavior is interpreted not as expressing gratitude but as<br />
an indirect solicitation for more gifts or favors.<br />
Another area of cross-cultural difference arose from the high value placed on<br />
modesty and humility by several cultures in our sample. A Chinese informant<br />
indicated that modesty is required in responding to the offer of a raise:<br />
Thank you very much. But I think I have not done so well to get a raise. Anyway,<br />
I'd try to do better.<br />
Also, as mentioned above, apologizing appears to be associated with expressions of<br />
gratitude in some cultures. A Japanese participant responded to the offer of the raise:<br />
I'm sorry. 1 will try harder in the future.<br />
Another Japanese, in response to the $500 loan, said:<br />
I'm sorry. I'll always remember the debt of gratitude.<br />
Americans find these kinds of utterances difficult to interpret and find them uncomfortable<br />
and confusing.<br />
Conclusion<br />
Our data show that thanking is a speech act that is mutually developed. It can<br />
involve a complex series of interactions and encodes cultural values and customs.<br />
Both the giver and the thanker collaborate in the development of a successful<br />
thanking episode. Because expressing gratitude involves "face wants" in the sense<br />
that the speaker expresses both approval of and indebtedness to the receiver of<br />
thanks, the feelings of the interlocutors are closely tied to this interaction. Our study<br />
and those of others reveal that, while thanking appears to be a universal function, it<br />
is accomplished differently in contrasting cultures, whose values may focus differentially<br />
on the various components that constitute expressions of gratitude.
Expressing Gratitude in American English 75<br />
In methodological terms, our studies reveal that, while written and oral questionnaire<br />
data miiTor the words and expressions used in conveying gratitude, roleplays<br />
reveal the interactive aspects of the function more fully. Although natural data<br />
are most indicative of the whole process, role-plays have the advantage of providing<br />
enough control to allow meaningful comparisons to be made. An integration of all<br />
four kinds of data was effective in helping us understand how native and nonnatives<br />
express gratitude in English.<br />
Finally, it was evident that even advanced learners of English have considerable<br />
difficulty adequately expressing gratitude. They need information on the nature of<br />
what to say, the language used to express it, and the context in which it is needed,<br />
"rom a pedagogical perspective, role-plays can serve not only as a diagnostic tool<br />
for teachers, but can also provide much needed practice for learners. Ideally, with<br />
their heightened awareness, learners can also observe outside the classroom the use<br />
of pragmatic functions in social interaction and vehicles that depict such interaction<br />
(e.g., dramas, soap operas). It may also be helpful for students to compare English<br />
models to their own speech in order to enhance their awareness of the pragmalinguistic<br />
and sociopragmatic rules of English.<br />
Appendix 1: The Questionnaire<br />
Please read the following short descriptions of situations in which you might find yourself.<br />
Think of what you might say in response to this situation. Write your response (if any) in the<br />
space provided. Say as much or as little as you wish—you may choose to say nothing in<br />
several circumstances.<br />
1. It's Friday. You look in your wallet and notice that you only have $2.00. Your good<br />
friend at work notices this and hears you say, 'Darn, I'll have to go to the bank.' Your<br />
friend asks if you need money, and you say that you forgot to go to the bank. Your<br />
friend says, 'I have plenty. How much do you need' You say, 'Could you lend me<br />
$5.00 I'll pay you back on Monday.' Your friend says, 'Sure. Are you sure you<br />
don't need more than that' You say you don't. Your friend gives you the $5.00.<br />
2. You board the bus, pay your money and take a seat near the front of the bus. Just<br />
before your stop, you signal the driver to stop. You move to the front, the bus comes<br />
to a stop, and the doors open.<br />
3. It's your birthday, and you're having a few people over for dinner. A friend brings<br />
you a present. You unwrap it and find a blue sweater.<br />
4. You work for a large company. The Vice-President of Personnel calls you into his<br />
office. He tells you to sit down. You feel a little nervous, because you have only been<br />
working there for six months. The Vice-President says, 'You're doing a good job. In<br />
fact, we are so pleased with you that I'm going to give you a $20.00 a week raise.'<br />
5. In the supermarket, the cashier puts your groceries in bags and turns to begin<br />
checking out the next customer. You pick up your bags to leave.<br />
6. At the table in a restaurant a friend says, 'You have something on your face.' You ask<br />
where. Your friend tells you. You rub your face and ask, 'Is it off" Your friend says<br />
that it is.<br />
7. You find yourself in sudden need of money—$500.00. You mention this to a friend.
76 Speech Act Realization<br />
Your friend immediately offers to lend it to you. You are surprised and very grateful.<br />
Your friend writes out a check for $500.00 and gives it to you. At first you say, 'Oh<br />
no, I didn't mean for you to lend it to me. I couldn't take it.' Your friend says,<br />
'Really, it's all right. What are friends for' After your friend insists again, you take<br />
the check.<br />
8. You are married. Both you and your spouse work. You come home late from work<br />
and find that your spouse has done some work around the house that you had<br />
promised to do, but had not had a chance to do.<br />
9. Your friend suggests going out to lunch. You say that you'd like to go, but you only<br />
have $2.00. Your friend says, 'Ah, don't worry. I'll take you today.' Your friend<br />
takes you to a very nice restaurant—a much more expensive one than the ones you<br />
usually go to. You have a wonderful meal. Your friend pays, and you get up to leave.<br />
10. You have just gotten a new and better job. A friend at the office tells you she has<br />
organized a farewell party for you.<br />
11. You have just gotten your hair cut in a new style, and you like it better than the old<br />
way. Your friend sees you and says, 'Hey, you've got a new haircut. It looks nice.'<br />
12. You are sharing an apartment with a friend. You're both sitting and relaxing in the<br />
living room. You ask your friend to hand you the newspaper which is nearby. Your<br />
friend gives you the newspaper.<br />
13. You pick up your car in a parking garage. As the attendant who drove up your car<br />
walks past you to get the next person's car, you hand him a tip.<br />
14. You have been invited to the home of a rather new friend. You have dinner with him<br />
and his wife and a few other friends of theirs. The food was great, and you really<br />
enjoyed the evening. As you leave, your hosts accompany you to the door.<br />
Appendix 2<br />
Dialogue 1<br />
(A role-play between two native speakers)<br />
WOMAN 1: I've got something for you.<br />
WOMAN 2: You do (takes package and unwraps) Oh, my goodness. Why did you do<br />
this ... a blue sweater, (giggles) How did you know this is just what I<br />
needed It's wonderful. Thank you. I'm so surprised!<br />
WOMAN 1: (laughing) Price tag in it<br />
WOMAN 2: I can really use this . . .<br />
WOMAN 1: Good. I hope it fits.<br />
WOMAN 2: Oh, I'm sure it will. Thank you. (goes to try it on) It's medium, I'm sure it<br />
will. TA-DA!<br />
WOMAN 1: (laughing) Yes, it does fit nicely.<br />
WOMAN 2: Yes, it's just what I needed. That's really nice.<br />
WOMAN 1: Wear it in good health.<br />
WOMAN 2: Thank you.<br />
WOMAN 1: So, how's everything (topic changes.)<br />
Dialogue 2<br />
(A role-play between two native speakers)<br />
WOMAN 1: I've only got $2.00. What am I going to do now I guess I'll have to go to the<br />
bank.
Expressing Gratitude in American English 77<br />
WOMAN 2: Oh, you don't have to go to the bank. I have money. I can lend you some.<br />
WOMAN 1: Oh no, I don't want to bother you. Thanks.<br />
WOMAN 2: Really I have plenty. Believe me, you can have it.<br />
WOMAN 1: Can you lend me $5.00, do you think<br />
WOMAN 2: Yeah. Sure. Here.<br />
WOMAN 1: Oh, I really appreciate this. Thanks.<br />
WOMAN 2: You're welcome.<br />
Dialogue 3<br />
(Natural data—a conversation between two native speakers)<br />
MOTHER: Another box, huh Oooo . . . what did you buy Oh, I shouldn't . . .<br />
I thought when she went out I should tell her not to buy one thing.<br />
DAUGHTER: Us. This is for you.<br />
MOTHER: What<br />
DAUGHTER: And this one is for me.<br />
MOTHER: This is for me Ohhhhh.<br />
DAUGHTER: That's for you.<br />
MOTHER: Ohhh, Jeanie, what'd you do that for now<br />
DAUGHTER: (giggles)<br />
MOTHER: That's too much, (opening the gift box) Oooo! Oh, my, that's lovely. Lovely.<br />
Oh my, Jeanie, that's pretty.<br />
DAUGHTER: Open it up.<br />
MOTHER: Ooo. Beautiful. Oh, yes.<br />
DAUGHTER: See . . . That'll keep you warm.<br />
MOTHER: Yes, I should say so. That's beautiful.<br />
DAUGHTER: Good. I'm glad you like it. It should be washable. Machine washable.<br />
MOTHER: Is it<br />
DAUGHTER: I think so. Just toss it in the machine.<br />
MOTHER: Ohh. Oh, my, that's lovely. Thank you.<br />
DAUGHTER: Oh. you're welcome, (they kiss and laugh.) What are we having for dinner<br />
Dialogue 4<br />
(Natural data—a conversation between two native speakers)<br />
MOTHER: (looking in a gift box of assorted pastries) Oooo. Oo. Oo. Oo.<br />
DAUGHTER: You've got a choice.<br />
MOTHER: Ohhh. You spend too much money.<br />
DAUGHTER: No.<br />
MOTHER: Oh, dear.<br />
DAUGHTER: No.<br />
MOTHER: Yes.<br />
DAUGHTER: No, it was fine.<br />
MOTHER: You shouldn't. I'll have to pay you.<br />
DAUGHTER: No, no, no. (switches topic.)<br />
Dialogue 5<br />
(Natural data—A conversation among native speakers, a 12-year-old boy,<br />
two men, and two women)<br />
AUNT: Why don't you give Erik his<br />
UNCLE: Erik Here's . . .<br />
ERIK: Oh, thank you.<br />
UNCLE: Your Christmas present.<br />
ERIK: Thank you. (opening present) Oh! Excellent!
78 Speech Act Realization<br />
AUNT: Do you like that<br />
ERIK: I ... I've always wanted a good sweater. We were going to get one. Now I<br />
don't need one. Excellent, thanks.<br />
AUNT: Good.<br />
UNCLE: Good.<br />
AUNT: I don't know. It may be a little big.<br />
DAD: Ooo. That looks great.<br />
MOM: Oo-la-la!<br />
DAD: Hey, Joe College.<br />
MOM: Ooo. And that's a real wool one. Aww. That looks like an Eddie Bauer.<br />
ERIK: Thank you.<br />
MOM: Is it wool<br />
AUNT: Yeah. It may be a little big. But I figure that . . . Erik . . . any minute now, it'll<br />
fit.<br />
MOM: Yeah. Oh ... handsome!<br />
AUNT: It's a little big.<br />
MOM: He was just talking about how he wanted some good-looking sweaters to wear to<br />
school. That's what the kids are wearing. Oh, boy.<br />
AUNT: (looking at the television) Hey, look, the Superbowl's coming on.<br />
Dialogue 6<br />
(A role-play between a Turkish man and a Brazilian woman)<br />
MAN: This is very kind of you inviting ... eh ... to you birthday ... eh ...<br />
party ... eh ... may I introduce you ... eh ... angora sweater if you<br />
mind ... eh ... if you like a sweater.<br />
WOMAN: Ay! Are you give me a sweater ... a blue sweater<br />
MAN: A blue sweater . . . yes.<br />
WOMAN: Ah ... Oh, that's wonderful. I really do like it.<br />
MAN: I'm happy you like it.<br />
WOMAN: I will. And ... eh ...<br />
MAN: That's fine. And ... eh ... hoping to celebrate you next birthday all together<br />
again.<br />
WOMAN: And so you can give me another sweater! (laughing) This time red, okay<br />
MAN: Yes, okay.<br />
Dialogue 7<br />
(A role-play between a French man and a Malaysian Chinese woman)<br />
WOMAN: I know today's a special day. I brought something for you.<br />
MAN: Why is this a special day Is a surprise<br />
WOMAN: It's your birthday.<br />
MAN: Oh, is birthday. Are you sure How did you know that<br />
WOMAN: (long pause)<br />
MAN: I'm really surprised, but ... uh ... I'm really at ease, but thank you very<br />
much because that's very nice and I didn't know you learned it was my<br />
birthday. I appreciate it.<br />
WOMAN: Don't mention it. (gives him the present)<br />
MAN: Wow! My God! But I like! Ah ... that's ... oh ... I like that. Thank you<br />
very much.<br />
WOMAN: Don't mention it.
Expressing Gratitude in American English 79<br />
Dialogue 8<br />
(A role-play between a Chinese woman and a Russian woman)<br />
WOMAN 1: Since you have been work here ... uh ... very good and very diligent. I<br />
like you very much. From next week I will raise your salary $15.00 every<br />
week. Would you like this<br />
WOMAN 2: Yes. Thanks a lot. Thanks a lot.<br />
WOMAN 1: Don't mention it. . . . And ... uh ... that's all. Would you please go back<br />
to your work<br />
Dialogue 9<br />
(A role-play between native English speakers)<br />
1: How about going out to lunch this afternoon Are you free<br />
2: I'm free, but I ... I don't know. 1 just didn't take enough with me. . .<br />
1: Well, don't worry about it, I'll treat you.<br />
2: No, I don't want you to do that.<br />
1: It's okay. It's my pleasure.<br />
2: (silence)<br />
1: Next time you can treat me. Okay<br />
2: Well, all right.<br />
Dialogue 10<br />
(A role-play between a Spanish man and a Hispanic woman)<br />
WOMAN: Joe, 1 have a little problem. 1 just realize that I don't have any money with me<br />
and the banks are closed now. Do you think is possible for you to lend me at<br />
least ten dollars for the weekend<br />
MAN: Ten dollars Do you need that money for what To eat or . . . <br />
WOMAN: No . . . just to have some money with me for the weekend in case any emergency<br />
comes up ...<br />
MAN: Yes, I can. I can borrow . . .<br />
WOMAN: I can lend you . . .<br />
MAN: I can lend you, I can lend you ten dollars, but don't forget next Thursday . . .<br />
when you . . . Yes, when you get paid, give me back ten dollars and if you<br />
want to give me some interests . . . well . . .<br />
WOMAN: (laughing) Oh ... thank you very much. I really appreciate that.<br />
Appendix 3<br />
The Role of the Giver<br />
Offer<br />
Here.<br />
I'll take you.<br />
How much do you need<br />
Downplay the offer<br />
Here's a little something.<br />
Really, I have plenty.<br />
I hate to eat alone.<br />
Comment<br />
It's washable.<br />
Role of the Thanker<br />
Hesitate<br />
Oh, I couldn't.<br />
Are you sure it's all right<br />
Oh, come on. No!<br />
Express surprise<br />
For me<br />
You're kidding!<br />
Compliment<br />
It's lovely.<br />
This is nice.
80 Speech Act Realization<br />
Prompt<br />
I hope it fits.<br />
Do you like it<br />
Did you have a good time<br />
Reassure<br />
Really, it's all right.<br />
What are friends for<br />
React<br />
Oh, good.<br />
I'm glad you like it.<br />
It was nice to see you.<br />
Close<br />
Well, good.<br />
Don't worry about it.<br />
It is/was my pleasure.<br />
Don't mention it.<br />
(Change topic)<br />
You're so nice.<br />
You're the best husband/wife in<br />
the whole world.<br />
Offer or suggestion of repayment<br />
I'll pay you back on ...<br />
Next time, I'll take you.<br />
I hope we can do this again<br />
sometime (soon).<br />
If there is anything I can ever<br />
do for you . . .<br />
Re-thank<br />
I love the sweater you gave me.<br />
That was such a nice evening.<br />
Thanks again for helping me<br />
out. I don't know what I<br />
would have done without you.<br />
Notes<br />
1. We noted that utterances containing words such as "thank you" or "thanks a lot" did<br />
not always express gratitude. For example, in the exchange:<br />
1: That haircut doesn't do a thing for you.<br />
2: Thanks a lot! You really know how to make me feel good!<br />
the illocutionary force of "thanks" is irritation rather than gratitude.<br />
2. Further information on our findings regarding cross-linguistic and cross-cultural elements<br />
in the written data will be integrated with our discussion on role-plays below since<br />
many of the same elements appeared in the written data as well as the oral data.<br />
References<br />
Beebe, L. M., & Cummings, M. C. (1985). Speech act performance: A function of the data<br />
collection procedure Paper presented at the Sixth Annual Sociolinguistics Colloquium,<br />
TESOL Convention, New York.<br />
Beebe, L. M., & Takahashi, T. (1989a). Do you have a bag Social status and patterned<br />
variation in second language acquisition. In S. Gass, C. Madden, D. Preston &<br />
L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics<br />
(103-125). Clevedon and Philadelphia: Multilingual Matters.<br />
Beebe, L. M., & Takahashi, T. (1989b). Sociolinguistic variation in face-threatening speech<br />
acts: Chastisement and disagreement. In M. Eisenstein (Ed.), The dynamic interlanguage:<br />
Empirical studies in SL variation (199-218). New York: Plenum.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Studies in Second Language Acquisition, 8, 165-79.<br />
Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge:<br />
Cambridge University Press.<br />
Cohen, A., & Olshtain, E. (1981). Developing a measure of sociocultural competence: The<br />
case of apology. Language Learning, 31, 113-34.
Expressing Gratitude in American English 81<br />
Coulmas, F. (1981). Poison to your soul: Thanks and apologies contrastively viewed. In<br />
F. Coulmas (Ed.), Conversational routine (66-91). The Hague: Mouton.<br />
DeCapua, A., & Dunham, J. F. (1989). Interaction strategies in the requesting and giving of<br />
advice. Paper presented at the Second Language Acquisition Circle, October 13,<br />
1989, New York University, New York.<br />
Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by<br />
native and non-native speakers of American English. Applied Linguistics, 7, 167-85.<br />
Goffman, E. (1967). Interaction rituals. New York: Pantheon Books.<br />
Hymes, D. H. (1971). Sociolinguistics and the ethnography of speaking. In E. Ardener (Ed.),<br />
Social anthropology and language (47-93). London: Tavistock.<br />
Leech, G. N. (1983). Principles of pragmatics. London: Longman.<br />
Rubin, J. (1983). The use of "Thank you." Paper presented at the Sociolinguistics Colloquium,<br />
TESOL Convention, Toronto, Canada.<br />
Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge:<br />
Cambridge University Press.<br />
Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.<br />
van Ek, J. A. (1976). The threshold level for modern language learning in schools. London:<br />
Longman.<br />
Wolfson, N. (1989). The social dynamics of native and nonnative variation in complimenting<br />
behavior. In M. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in SL<br />
variation (219-36). New York: Plenum.
4<br />
Perception and Performance<br />
in Native and Nonnative Apology<br />
MARC L. BERGMAN and GABRIELE <strong>KASPER</strong><br />
"I'm sorry, so sorry,<br />
Please accept my apology ..."<br />
BRENDA LEE, "I'm Sorry" (1958)<br />
In terms of their actional characteristics, apologies relate in important ways to other<br />
frequently studied expressive speech acts. Like the speech acts of thanking, complimenting,<br />
and complaining, apologies occur post-event. Complimenting and thanking<br />
involve events deemed praiseworthy by prevailing social norms; complaints and<br />
apologies refer back to events that constitute norm infringements. By complimenting<br />
and thanking, an event is made into a praiseworthy occasion; by complaining<br />
about or apologizing for its occurrence, an event is made into a transgression. In<br />
terms of value attribution, the relationship between event and speech act is reflexive<br />
rather than unilateral. Complaints, thanks, and compliments commonly hold the<br />
addressee to be causally involved in the preceding event; the agent assuming responsibility<br />
for the event necessitating the apology is the speaker. Compliments differ<br />
from thanks in that thanks require that the addressee's preceding action be beneficial<br />
to the speaker; compliments refer to addressee-related events, which do not need to<br />
be beneficial to anybody in particular. The actional descriptors referred to above<br />
allow us to distinguish compliments, thanks, complaints, and apologies schematically<br />
(Table 4.1).<br />
Apologies can be defined as compensatory action to an offense in the doing of<br />
which S was causally involved and which is costly to H. This conceptualization is<br />
supported by Goffman's (1971) view of apologies as remedial interchanges, remedial<br />
work serving to reestablish social harmony after a real or virtual offense. Following<br />
Goffman's distinction of ritual and substantive compensation, apologies can be<br />
classified into (1) those redressing virtual offenses, which are remedied by the sole<br />
offering of an apologetic formula, and (2) those redressing actual damage inflicted<br />
on the addressee, sometimes including an offer of material compensation. Both<br />
kinds of apology have been demonstrated to vary cross-culturally.<br />
82
Apology, Perception, and Performance 83<br />
Table 4.1. Actional Features of Four Expressives<br />
Event<br />
e-Evaluation*<br />
e-Actor<br />
e-Recipient<br />
Compliment<br />
Thanking<br />
Complaint<br />
Apology<br />
post<br />
post<br />
post<br />
post<br />
good<br />
good<br />
bad<br />
bad<br />
H<br />
H<br />
H<br />
S<br />
H<br />
S<br />
S<br />
H<br />
*e = event.<br />
Ritualistic apologies are sometimes distinguished from substantive ones by different<br />
formulae. Thus in (American) English, ritualistic apologies with the formula<br />
"excuse me" are offered as territory invasion signals when addressing strangers<br />
(e.g., prior to asking direction), as announcements of temporary absence from<br />
ongoing interaction (e.g., in order to answer the phone), or upon virtual or real<br />
intrusion of another person's physical space (e.g., passing somebody in a narrow<br />
hallway). Borkin and Reinhart (1978) define the function of "excuse me" as "a<br />
formula to remedy a past or immediately forthcoming breach of etiquette or other<br />
light infraction of a social rule" (61). "I'm sorry," in their analysis, is used in a<br />
wider range of contexts, especially "in remedial interchanges when a speaker's main<br />
concern is about a violation of another person's right or damage to another person's<br />
feelings" (61). Borkin and Reinhart's analysis is given strong empirical support by<br />
House (1988). She found that in seven apology contexts involving substantive<br />
offenses, native speakers of British English (n = 100) used "I'm sorry" up to 80%<br />
of the time. In six of these contexts, "excuse me" did not figure at all, in one of them<br />
it was used with negligible frequency (3%).<br />
The ritualistic function of much apologetic behavior, and its cross-cultural variability,<br />
has also been noted by Coulmas (1981) in his analysis of expressions of<br />
gratitude and indebtedness in a number of Western languages as opposed to Japanese.<br />
Coulmas notes that in many contexts requiring expressions of gratitude in<br />
Western cultures, such as upon receiving a gift, Japanese requires an apologetic<br />
formula such as "sumimasen." The function he ascribes to ritualistic apologies in<br />
Japanese concurs with the functional properties of "excuse me" described by Borkin<br />
and Reinhart (1978); namely, indicating "the speaker's willingness to conform to<br />
conventional rules and social expectations. . . . Verbal apology occurs even if there<br />
was no serious or real offence as a precaution against inadvert misconduct or<br />
unanticipated negative interpretation of one's performance" (Coulmas, 1981,84).<br />
While Borkin and Reinhart's (1978) analysis suggests that acquiring appropriate<br />
formulas for ritualistic apology is problematic for nonnative speakers (NNS), substantive<br />
apologies confront the learner with a more complex learning task. First (as<br />
with ritualistic apologies), she has to identify the occurrence of an event that requires<br />
apology. This may require restructuring her cultural knowledge as such<br />
events have been shown to vary cross-culturally (Olshtain, 1983). Second, the<br />
severity of the offense and the weights of contextual variables such as power and<br />
distance need to be assessed—another potential trap as perceptions of these social<br />
variables are also subject to cross-cultural variation (House, 1988; Vollmer &<br />
Olshtain, 1989). Finally, appropriate output strategies have to be selected. While all
84 Speech Act Realization<br />
available evidence points to a universally valid apology speech act set (Olshtain,<br />
1989), preferences for strategy choice are contextually and cross-culturally at variance<br />
(Olshtain, 1989). As Garcia (1989) has demonstrated, cultural differences do<br />
not only obtain in preferences for local strategic choices but for global approaches to<br />
the speech event, such as opting for a deference versus a solidarity style.<br />
While we do not believe that language users actually go through the motions of<br />
planning linguistic action in the serial fashion outlined above (Schmidt, Chapter 1),<br />
appropriate apologizing requires the specified knowledge components, and the language<br />
user's ability to access this knowledge fast and flexibly. In interlanguage (IL)<br />
pragmatics, only a few studies have addressed the issue of how NNS apologize in<br />
ongoing interaction with a native (NS) or NNS interlocutor, and the available<br />
studies (Rasper, 1981; Trosborg, 1987; Garcia, 1989) have analyzed NNS's performance<br />
in terms of their pragmatic knowledge (what semantic formulas do they use)<br />
rather than examining how pragmatic knowledge is accessed under immediate processing<br />
constraints.<br />
Studies of IL apologizing have essentially addressed the same research<br />
question—the accessibility of apology strategies to NNS—yet they have examined<br />
this issue by means of different data-gathering procedures. Rintell and Mitchell<br />
(1989) compared NS and NNS use of apology strategies in their responses to written<br />
and oral Discourse Completion questionnaires and found only slight differences<br />
between the two conditions. In order to review previous substantive findings and<br />
assess further instrument effects, the results of five studies are summarized in<br />
Table 4.2.<br />
Holmes (1989) investigated the realization of apologies in an ethnographically<br />
collected corpus of 183 remedial interchanges produced by adult NS of New Zealand<br />
English (NS NZE). Hers is not an interlanguage study but has been included<br />
because it allows us to compare elicited data with naturally occurring speech.<br />
Olshtain (1983) examined apology performance in role-plays, enacted by NS of<br />
American English (NS AE), Russian (NS Rus), and Hebrew (NS Heb), and by<br />
American and Russian learners of Hebrew (IL E-Heb, IL Rus-Heb). Role-play was<br />
also employed by Trosborg (1987) in a study involving NS of British English (NS<br />
BE), NS of Danish (NS Dan), and three groups of Danish learners of English at<br />
different levels of proficiency. Table 4.2 includes the figures for the lower advanced<br />
level (IL Dan-En). House (1988) used a Discourse Completion Task (DCT) to study<br />
apology realization by NS of British English (BE NS), NS of German (NS Ger), and<br />
German learners of English (IL Ger-En). By means of the same DCT questionnaire,<br />
Kasper (1989) looked at the apology responses provided by NS of Danish (NS Dan)<br />
in comparison with Danish learners of English (IL Dan-En) and Danish learners of<br />
German (IL Dan-Ger). Throughout the five studies, apology realizations were<br />
coded according to the semantic formulae identified as constituting the apology<br />
speech act set (Olshtain & Cohen, 1983; Blum-Kulka, House & Kasper, 1989).<br />
Figures reported in the studies have been summarized according to these categories.<br />
Most subjects apologized explicitly by means of an Elocutionary Force Indicating<br />
Device (IFID) such as "I'm sorry" and stated whether they assumed responsibility<br />
for the offense (responsibility): e.g., by blaming a third party ("the kitchen<br />
must have mixed up the orders"), admitting the offensive act ("I forgot to bring your
Apology, Perception, and Performance 85<br />
Table 4.2. Selection of Apology Strategies (%) by Fifteen Groups of Informants<br />
Verbal<br />
IFID Intensifier Responsibility Account Minimization Repair redress<br />
ethno. (Holmes)<br />
NS NZE<br />
RP (Olshtain)<br />
NS AE<br />
NS Rus<br />
NS Heb<br />
IL E-Heb<br />
IL Rus-Heb<br />
RP (Trosborg)<br />
NS BE<br />
NS Dan<br />
IL Dan-E<br />
DCT (House)<br />
NS BE<br />
NS Ger<br />
IL Ger-E<br />
DCT (Kasper)<br />
NS Dan<br />
IL Dan-E<br />
IL Dan-Ger<br />
61<br />
92<br />
66<br />
66<br />
69<br />
64<br />
7<br />
6<br />
8<br />
80<br />
69<br />
73<br />
72<br />
75<br />
68<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
40<br />
31<br />
48<br />
22<br />
27<br />
15<br />
9<br />
100<br />
75<br />
58<br />
92<br />
71<br />
24<br />
33<br />
41<br />
70<br />
66<br />
80<br />
49<br />
82<br />
75<br />
23<br />
42<br />
33<br />
33<br />
23<br />
21<br />
22<br />
18<br />
13<br />
3<br />
5<br />
6<br />
18<br />
5<br />
10<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
NR<br />
21<br />
25<br />
14<br />
13<br />
11<br />
16<br />
9<br />
16<br />
12<br />
5<br />
0<br />
0<br />
8<br />
0<br />
0<br />
22<br />
12<br />
21<br />
15<br />
13<br />
14<br />
13<br />
16<br />
23<br />
2<br />
0<br />
0<br />
0<br />
0<br />
7<br />
5<br />
6<br />
2<br />
9<br />
4<br />
8<br />
8<br />
5<br />
1<br />
book"), or by self-blame ("it's my fault"). The extremely low frequencies of IFID<br />
provided by the British-English and Danish speakers in Trosborg's role-plays do not<br />
seem to be the result of a language or instrument effect since high frequencies of<br />
IFID were supplied by speakers of the same languages in Kasper's (1989) DCT<br />
study and by Olshtain's (1983) role-players. This suggests differential contextual<br />
effects in Trosborg's data as opposed to the other studies. Intensification of apology<br />
("Im very/terribly sorry") was only reported in the DCT studies, where the English<br />
NS and German-English IL users upgraded their apologies more than the Danes and<br />
the German NS. Except for the Danish NS, who provided the same amount of<br />
explanations or accounts for the offense in the role-play and DCT, the interactive<br />
conditions elicited considerably more accounts than the noninteractive DCT. This<br />
instrument effect may be indicative of different psychological and discoursestructural<br />
conditions. Subjects may feel more psychologically pressed to account for<br />
purported misdeeds in a direct encounter with the offended party, and in ongoing<br />
discourse where they can distribute various apologetic acts over several turns.<br />
Accounts may even be requested by the interlocutor, a possibility precluded by the<br />
one-turn response in a DCT. The British-English and Danish NS tended to minimize<br />
the severity of the offense (e.g., "it's only a tiny scratch (on your car)") more in the<br />
role-plays than in the DCTs, which suggests a similar effect as in the case of<br />
accounts. The greater effort of these speakers to minimize the offense is consistent<br />
with their reluctance to explicitly apologize and take on responsibility. Repair, or<br />
compensation for the incurred damage ("I'll bring it in tomorrow"), was offered<br />
with about the same frequencies as minimization. Olshtain's (1983) role-plays and
86 Speech Act Realization<br />
Holmes' (1989) ethnographic data provided very few instances of repair. Expressions<br />
of concern for the hearer ("I hope you didn't wait long") or promises of<br />
forebearance ("it won't happen again"), conflated as "verbal redress," were used<br />
extremely rarely by all groups across conditions.<br />
For the majority of NS and IL users, expressing an explicit apology and making<br />
a responsibility statement were the essential components of apology, whereas providing<br />
explanations, minimizing the offense, and offering repair and verbal redress<br />
were optional and, as demonstrated by Olshtain (1989), context-dependent strategies.<br />
This summary thus extends Olshtain's findings (1983, 1989) to a wider range<br />
of languages.<br />
The variation in supplied apology strategies raises a number of questions. For<br />
instance, why do Holmes's (1989) informants make considerably fewer responsibility<br />
statements than any other group Our supposition is that this is due to the<br />
type of apology included in Holmes's corpus. Even though Holmes does not comment<br />
on this, her offense categories "space" and "talk" offense in particular suggest<br />
that she included ritualistic apologies in her corpus. The apologetic formulas by<br />
which ritualistic apology is performed very often do not co-occur with other redressive<br />
activity. As the role-plays and DCT questionnaires exclusively provide contexts<br />
for substantive apologies, the naturalistic data set does not provide an adequate<br />
baseline for the two types of elicited data. Furthermore, the conspicuous difference<br />
between Trosborg's (1987) role-plays and the remaining elicited data suggests that<br />
close attention be given to instrument effects in the study of apology.<br />
Research to date has examined apology behavior in a variety of Western cultures<br />
and languages. While these studies have been important in providing preliminary<br />
evidence for a universally valid apology speech act set, and the differential<br />
selections from this set according to contextual factors, it is requisite to<br />
extend the scope of study to non-Western languages and cultures to advance the<br />
fundamental issue in cross-cultural pragmatics; namely, the universality and specificity<br />
of linguistic action. One of the methodological problems in cross-cultural and<br />
cross-linguistic comparison is to determine whether the contextual conditions in<br />
which the speech act behavior under study occurs are perceived as the same or<br />
different by the groups to be compared (Arndt & Janney, in press; Blum-Kulka &<br />
House, 1989). With specific reference to apologies, Wolfson, Marmor, and Jones<br />
(1989) comment that "a cross-linguistic study of apologies may well reveal that the<br />
notions of offense and obligation are culture-specific and must, therefore, become<br />
an object of study in themselves" (180). Brown and Levinson's (1987) politeness<br />
theory predicts that the weightiness of face-threatening acts (FTAs), computed by<br />
adding the values of social distance, dominance, and degree of imposition as perceived<br />
by actors in a given context, determines the kind and amount of redress<br />
afforded in the performance of FTAs (76). According to theory and empirical<br />
evidence (Kasper, 1990 for overview), both the weights and values of the contextual<br />
factors vary cross-culturally. In the cross-cultural study of apology, it is therefore<br />
essential to establish what constitutes an offense, how members of different cultures<br />
perceive offense contexts, and how these perceptions are reflected in output strategies.
Apology, Perception, and Performance 87<br />
The Present Study<br />
The central research questions we wish to examine in this study are: (1) How are<br />
contextual factors in a variety of offense contexts perceived by Thai and American<br />
informants (2) How is the selection of apology strategies determined by contextual<br />
factors and (3) What patterns of intracultural and intercultural variability are observable<br />
in the selection of apology strategies by Thai NNS of English as compared to<br />
NS of Thai and American English<br />
Informants in this study were 423 Thai graduate students at Chulalongkorn<br />
University, Bangkok, at an intermediate level of proficiency in English, and 30 NS<br />
of American English who were students at the University of Hawaii at Manoa. Data<br />
were collected by means of two questionnaires. The Assessment questionnaire<br />
included 20 items, each of which specified a different offense context (see Appendix<br />
1 for sample item). Informants were asked to rate these contexts on a 5-point rating<br />
scale for 4 context-internal factors (Severity of offense, offender's Obligation to<br />
apologize, Likelihood for the apology to be accepted, offender's Face-loss) and 2<br />
context-external factors (social Distance and Dominance) (for the distinction between<br />
context-internal and context-external factors, see Brown & Fraser, 1979;<br />
Blum-Kulka & House, 1989). The selection of the context-external factors Distance<br />
and Dominance (power) follows Brown and Levinson's (1987) weightiness formula<br />
(76) and their argument that Distance and Dominance constitute "very general<br />
pancultural dimensions which nevertheless probably have 'emic' correlates" (76).<br />
The context-internal factors are assumed to function as components of Brown and<br />
Levinson's dimension "degree of imposition," specified for the speech event of<br />
apologizing. Severity of offense and Obligation to apologize were found to distinguish<br />
different types of offense by Olshtain (1989). Likelihood for the apology to be<br />
accepted was added in view of an analysis of apology responses; however, it is<br />
included in the present study to examine possible interdependence between it and<br />
other contextual factors, and effects on the selection of apology strategies. Offender's<br />
Face-loss was included since, as Vollmer and Olshtain (1989) note, "the<br />
expression of responsibility seems to be linked directly to the S's cost and loss of<br />
face which results from performing the speech act of apologizing" (198), and which<br />
consequently can be expected to have an impact on the redress afforded in carrying<br />
out an apology.<br />
The Dialog Construction (DC) questionnaire included the same offense contexts<br />
as the Assessment questionnaire. Informants were asked to supply the offender's<br />
and the offended person's turn (see Appendix 2 for sample item). Since the apology<br />
response was elicited from the informants rather than being provided as part of the<br />
questionnaire items, the instrument differed from the standard Discourse Completion<br />
questionnaire (Blum-Kulka, House & Kasper, 1989). For this study, only the<br />
first pair parts were analyzed.<br />
The two questionnaires included the following contexts:<br />
1. A and B are friends. A has had an accident with a car borrowed from B.<br />
(Damaged Car)
88 Speech Act Realization<br />
2. A and B are friends. A borrowed a magazine from B and poured coffee over<br />
it. (Ruined Magazine)<br />
3. At a staff meeting, teacher A contradicts teacher B. (Contradiction)<br />
4. At a staff meeting, teacher A accuses teacher B of being a poor teacher.<br />
(Poor Teacher)<br />
5. At an office, a junior colleague forgets to pass on a private message to a<br />
senior colleague. (Private Message Low-High)<br />
6. At an office, a senior colleague forgets to pass on a private message to a<br />
junior colleague. (Private Message High-Low)<br />
7. At an office, a junior colleague forgets to pass on an important business<br />
message to a senior colleague. (Business Message Low-High)<br />
8. At a office, a senior colleague forgets to pass on an important business<br />
message to a junior colleague. (Business Message High-Low)<br />
9. At a restaurant, a customer changes her mind after the order has already<br />
been taken. (Order Change)<br />
10. At a restaurant, a waiter spills food on a customer's clothes. (Food on<br />
Customer)<br />
11. At a restaurant, a waiter brings the wrong order. (Wrong Order)<br />
12. At a restaurant, a customer spills food on a waiter. (Food on Waiter)<br />
13. At Bangkok Airport, a customs official messes up a traveler's suitcase.<br />
(Messed-up Bag)<br />
14. At Bangkok Airport, a traveler is caught trying to smuggle a Buddha out of<br />
the country. (Smuggled Buddha)<br />
15. At Bangkok Airport, a customs official breaks a legally purchased statue<br />
when searching a traveler's suitcase. (Broken Status)<br />
16. At Bangkok Airport, a traveler is unable to produce a customs form. (Customs<br />
Form)<br />
17. A professor has not yet graded a term paper that a student was supposed to<br />
pick up. (Ungraded Paper)<br />
18. A student forgets a book she was supposed to return to her professor.<br />
(Borrowed Book)<br />
19. A professor misplaces a student's term paper and fails the student. (Failed<br />
Student)<br />
20. A student plagiarizes from a published book and is found out by a professor.<br />
(Cheating Student)<br />
Items in both questionnaires were randomized. The Assessment questionnaire<br />
was filled out by 30 American NS of English and by 30 Thai NNS of English. Both<br />
versions were in English; however, for the Thai study, conceivably difficult vocabulary<br />
items were glossed in Thai. The DC questionnaire was prepared in an English<br />
and in a Thai version. The English version was filled out by 30 NS of English and<br />
by 288 Thai NNS of English. The Thai version was filled out by 136 Thai NS.<br />
The Assessment Study<br />
The mean ratings and standard deviations of contextual factors in each offense<br />
situation are included in Appendix 3. Calculation of Pearson Product Moment
Apology, Perception, and Performance 89<br />
Coefficients revealed no statistically significant correlation between the contextexternal<br />
factors Distance and Dominance, nor between any of the context-external<br />
and the context-internal factors. This finding lends support to Brown and Levinson's<br />
(1987) thesis of the mutual independence of contextual factors (80f.). However, we<br />
hesitate to conclude that such independence would continue to obtain in ongoing<br />
interaction. While a static, snapshotlike view of an interactional configuration, as<br />
represented by the questionnaire items, suggests the integrity of Distance and Dominance<br />
relative to each other and to the interactional content, conversational interchanges<br />
where apologies are jointly negotiated may well evidence a dynamic and<br />
interdependent fluctuation of participants' ascriptions of Distance and Dominance<br />
(e.g., Aronsson & Satterlund-Larsson, 1987), and of these context-external with<br />
context-internal factors.<br />
The ratings of context-internal factors by both Thai and American informants<br />
yielded high correlations between Severity and Obligation (American r = .7210,<br />
p < .000, Thai r = .6815, p < .001), Severity and Likelihood (American r =<br />
- .6382, p < .002, Thai r = - .7505, p < .000), Severity and Face-loss (American<br />
r = .7564, p < .000, Thai r = .5848, p < .007), and Obligation and Face-loss<br />
(American r = .8528, p < .000, Thai r = .7401, p < .000). Overall, then, it can be<br />
concluded that Severity of offense is systematically related to the offender's Obligation<br />
to apologize (corroborating Olshtain's (1989) finding about the covariance of<br />
Severity and Obligation [160]), the Likelihood for the apology to be accepted—the<br />
more serious the infringement, the less the likelihood for the apology to be<br />
successful—and the extent to which the offender's (positive) Face is adversely<br />
affected. Furthermore, Face-loss and Obligation can be viewed as codeterminant:<br />
more Face-loss requires more Obligation to apologize, and more Obligation entails<br />
more Face-loss.<br />
The demonstrated interrelatedness of context-internal factors can be understood<br />
as an explication of Brown and Levinson's (1987) contextual dimension "degree of<br />
imposition" (76) for the speech act of apology. What is noteworthy is the strong<br />
congruence by which Thai and American informants perceived the relationships<br />
between the context-internal factors, and the lack of interrelation expressed in the<br />
ratings of the context-external factors. However, rather than drawing hasty conclusions<br />
about congruent social perceptions by members of widely different cultures,<br />
we feel that two caveats are in order. First, the Thai questionnaire was administered<br />
in an English version. The assessments given by the Thai informants may therefore<br />
represent their perceptions of offense contexts in English-speaking cultures rather<br />
than in their native culture. Their approximation of the American ratings might be<br />
taken as an expression of intercultural competence, a construct parallel to interlanguage<br />
competence. To what extent the Thais' intercultural perceptions are reflective<br />
of, or different from, their social perceptions in their native culture can only be<br />
examined by comparative Thai material, which we unfortunately were unable to<br />
collect.<br />
Second, the noted correlations obtained across the 20 contexts represented in the<br />
questionnaire. They do not allow predictions of the values assigned to contextual<br />
factors in individual contexts. In the remainder of this section, we shall take a closer<br />
look at the contextual assessments of different offense situations. Statistically signif-
90 Speech Act Realization<br />
icant differences between the Thai and American ratings were determined by multivariate<br />
analysis (MANOVA). In order to facilitate the following presentation, we<br />
shall divide the continuous ratings into three categories: low (1-2.3), medium (2.4-<br />
3.7), and high (3.8-5).<br />
SEVERITY OF OFFENSE<br />
Only Order Change and (according to the American raters) Borrowed Book were<br />
perceived as light offenses. Medium Severity was assigned to offenses involving<br />
low material costs (Ruined Magazine), inconveniences rather than infringements<br />
of legal rights or entitlements (Ungraded Paper, Contradiction, Messed-up<br />
Bag, Private Messages), or mishaps occurring as part of a job (Food on Waiter,<br />
Wrong Order). These offenses represent minor impositions on somebody's time,<br />
money, physical space, energy, or face-wants; they warrant no legal claims to<br />
redress.<br />
High Severity offenses have major real-life consequences rather than involving<br />
easily repairable inconveniences. They may constitute illegal action (Cheating Student),<br />
high material costs (Broken Statue), potential or real obstruction of regular<br />
procedures and negligence of professional obligations (Business Messages, Failed<br />
Student), or violation of a person's physical integrity without this being an accepted<br />
job or task risk (Food on Customer).<br />
Offenses that were rated as less severe by American than by Thai informants<br />
were Borrowed Book, a routine fact of American academic life, and Smuggled<br />
Buddha. For cultural outsiders, taking a Buddha statue out of Thailand constitutes<br />
no more than a petty offense, in legal terms. For Thais, the act is not just illegal but<br />
sacrilegious, expressing contempt of religious beliefs and of the most significant<br />
symbol of Thai culture. Americans gave higher Severity ratings than the Thais to<br />
Wrong Order, Private Message High-Low, Poor Teacher, and Cheating Student.<br />
Whereas the Thais assigned contradicting a colleague and questioning his or her<br />
professional competence the same medium Severity rating, the Americans perceived<br />
the "Poor Teacher" allegation as much more serious.<br />
OBLIGATION TO APOLOGIZE<br />
Except for Order Change, which was rated low on Obligation by the Thais, both<br />
groups of informants perceived the offender's Obligation to apologize as medium or<br />
high in all contexts. However, the American raters gave as many as 12 offenses<br />
higher scores on Obligation than the Thais. A consistent relationship was discernable<br />
between Obligation and Severity in the ratings by both groups. First, Offense<br />
contexts received the same ratings on Obligation as on Severity. Thais and Americans<br />
perceived 13 contexts as similar on these dimensions. This finding is consistent<br />
with the assessment of seven offense contexts by Israeli raters, reported by Olshtain<br />
(1989). Second, Offenses received a categorically higher rating on Obligation than<br />
on Severity. In the American ratings, the low Severity offenses Order Change and<br />
Borrowed Book scored medium on Obligation, and five medium Severity offenses<br />
(Ruined Magazine, Wrong Order, Food on Waiter, and the Private Messages) registered<br />
high on Obligation. The Thai raters assessed six medium Severity items as<br />
high on Obligation (Private Message Low-High, Wrong Order, Food on Waiter,<br />
Borrowed Book, Damaged Car, Messed-up Bag). This relationship between Severi-
Apology, Perception, and Performance 91<br />
ty and Obligation has not been reported in previous apology assessment studies<br />
(House, 1988; Olshtain, 1989; Vollmer & Olshtain, 1989).<br />
The ratings of individual offense items on Obligation thus permit us to elaborate<br />
upon the correlational finding, "the more severe an offense, the more it warrants<br />
apology": Either the obligation to apologize is directly proportionate to the severity<br />
of the offense, or the perceived need to apologize categorically exceeds the degree<br />
of the norm infringement. This latter finding strongly suggests the precedence of<br />
relational over transactional concerns in much human interaction: in the interest of<br />
restoring social harmony, remedial action in the form of apology is seen as required<br />
even if the offense is a minor matter. Building on Goffman's (1971) discussion of<br />
ritual and substantive apology (also Owen, 1983), the following analogy comes to<br />
mind: ritual apology is offered as redress to virtual offenses, that is, to events that<br />
might, but did not, cause offense. By the same token, factual minor offenses can be<br />
magnified into virtual major infringements through disproportionate offer of redress.<br />
The benefit accruing from a disproportionate (and acted-upon) need to apologize<br />
is to reestablish the implied parties—offender and offended person—as fully<br />
competent and responsible participants in the business of social interaction.<br />
There was, however, one exception to the patterned relationship between Severity<br />
and Obligation. In the Thai ratings, Smuggled Buddha was assessed as the<br />
highest offense of all (4.7) yet assigned to medium Obligation (3.5). There were no<br />
indications from the ratings of the other context-internal factors to explain this<br />
irregularity; although Smuggled Buddha registered low on Likelihood and high on<br />
Face-loss, so did other items that were assessed as strongly requiring apology. We<br />
think that perhaps the informants reasoned strategically on this item: since an<br />
apology is concomitant with admitting to an offense that is a major cultural and legal<br />
infringement, an escape route for the offender might be to pretend ignorance by not<br />
apologizing. We shall see whether this speculation bears out in the Thais' choice of<br />
apology strategies.<br />
LIKELIHOOD OF APOLOGY ACCEPTANCE<br />
The Thai raters did not perceive Likelihood for the apology to be accepted as low in<br />
any context, whereas the American raters felt that acceptance was unlikely in<br />
Broken Statue, Cheating Student, and Smuggled Buddha. These three contexts,<br />
which (except for Smuggled Buddha in the American ratings) registered highest on<br />
Severity in Thai and American assessments, were also given the lowest ratings in<br />
the medium category by the Thais. Likewise, Thai and American raters perceived<br />
the same seven offenses as having a high Likelihood for apology acceptance. All of<br />
these offenses had been given low ratings on Severity.<br />
OFFENDER'S FACE-LOSS<br />
All offense contexts were perceived as involving medium or high damage to the<br />
offender's face by both Thai and American raters. However, the Thais considered<br />
only four contexts—Contradiction, Poor Teacher, Order Change, and Private Message<br />
High-Low—as less than highly face-threatening. The Americans, on the other<br />
hand, felt in as many as nine contexts that the offender's face was not severely<br />
threatened. In addition to the four contexts also rated as medium for Face-loss by the<br />
Thais, the Americans attached only minor Face-loss to low and medium Severity
92 Speech Act Realization<br />
offenses that were due to an oversight rather than to an intentional norm infringement,<br />
and that did not involve high costs to the offended party (Ruined Magazine,<br />
Messed-up Bag, Customs Form, Ungraded Paper, Borrowed Book). For the Thais,<br />
minor Severity of an offense did not entail minor Face-loss. Out of twelve medium<br />
Severity offenses, the Thais rated nine high on Face-loss. The highest Severity<br />
offenses were also given the highest ratings on Face-loss by Thais and Americans.<br />
The emerging pattern is thus that for the Americans, offender's Face-loss and<br />
Severity of offense were codeterminant. In the Thai informants' perception, by<br />
contrast, high Severity and high Face-loss were interrelated, but lower Severity<br />
might be outweighed by other factors that could make an event highly threatening to<br />
the offender's face even in the absence of much damage in objective terms. Such<br />
factors seem to include demonstration of undesirable personal attributes such as<br />
clumsiness, carelessness, or forgetfulness. We are reluctant, however, to draw<br />
further inferences from these differential ratings, since they may well be indicative<br />
of conceptual differences between the notion of "face" in Thai and American culture.<br />
Comparative ethnosemantic study of the concept efface" in both cultures and<br />
languages is needed, exploring its meaning and function within each cultural context<br />
(cf. Hu 1944 for an examination of the concept of face in Chinese culture).<br />
Of the context-external factors Distance and Dominance, we shall only examine<br />
Distance, as no effects were found for Dominance on any of the other contextual<br />
factors, nor on the selection of apology strategies.<br />
DISTANCE<br />
With few exceptions, the Thai and American ratings of Distance reflected the social<br />
role relationship between offender and offended party. Thai and American informants<br />
agreed in perceiving the closest relationships between friends, and the most distant<br />
relationships between strangers in service and administrative encounters. Differences<br />
are noticeable in the two groups' perceptions of student-professor and work relationships.<br />
Both groups assessed the relationships between student and professor as<br />
medium Distance, corresponding to "acquaintance" as opposed to familiars or strangers.<br />
Yet for the Thais, student-professor tended more towards the low end of the<br />
medium Distant category (> 3), whereas for the Americans, it approximated closer to<br />
the high end of medium Distance (< 3). Possibly, Thais perceive the prototypical<br />
relationship of students and professors more similar to that of distant family members,<br />
while Americans regard it more as a work relationship involving participants on<br />
different levels of positional hierarchies. The latter view is supported by the ratings of<br />
relationships between colleagues, which the Thais perceived as more distant than the<br />
Americans in four out of six cases. Previous studies with different populations are<br />
consistent with either the Thai or the American assessments of social Distance in the<br />
two types of role relationships. German and British raters assessed relationships<br />
between colleagues as less distant than student-professor relationships (House, 1988,<br />
305), whereas Israeli informants perceived more Distance between colleagues than<br />
between student and professor (Olshtain, 1989,160). Neither the Thai nor the American<br />
ratings indicated any evidence of differential perceptions of Distance between the<br />
relationships of teachers and office colleagues on the one hand, and between coworkers<br />
at different ranks on the other hand.
Apology, Perception, and Performance 93<br />
Our findings on the Assessment study can be summed up as follows:<br />
1. Thai and American raters perceive context-external and context-internal factors<br />
as unrelated.<br />
2. Context-internal factors in offense contexts are highly interrelated. Severity<br />
covaries with Obligation to apologize, Likelihood for the apology to be<br />
accepted, and offender's Face-loss, lending support to Olshtain's (1989)<br />
hypothesis that "severity of offense is the representative contextual factor in<br />
the sociopragmatic set of the apology" (160).<br />
3. Thai and American raters consistently perceive Obligation, Likelihood, and<br />
Face-loss as higher than Severity of offense, suggesting the primacy of<br />
interpersonal concerns over transactional goals in much remedial exchange.<br />
4. Despite the overall consistent relationship of contextual factors in the American<br />
and Thai assessments, individual offense contexts may be characterized<br />
by constellations of contextual factors that are not predictable from the<br />
general pattern, and that vary cross-culturally. In each of the 20 situations,<br />
Thais and Americans differed in their perception of at least one contextual<br />
variable, most on Obligation, least on Likelihood.<br />
By way of summary, Table 4.3 presents the two language groups' differential<br />
assessments.<br />
Table 4.3. Differences in American and Thai Assessments of Contextual Factors<br />
Obligation Face-loss Distance Dominance Severity Likelihood<br />
Borrowed Book<br />
Messed-up Bag<br />
Poor Teacher<br />
Private Mess. H-L<br />
Wrong Order<br />
Smuggled Buddha<br />
Ruined Magazine<br />
Cheating Student<br />
Private Mess. L-H<br />
Business Mess. L-H<br />
Business Mess. H-L<br />
Food on Customer<br />
Food on Waiter<br />
Customs Form<br />
Contradiction<br />
Damaged Car<br />
Order Change<br />
Failed Student<br />
Ungraded Paper<br />
Broken Statue
94 Speech Act Realization<br />
The Dialog Construction Study<br />
The DC data were coded into the following major categories:<br />
IFID—Illocutionary Force Indicating Device, specifying the force of apology<br />
(I'm sorry, I'm afraid)<br />
Upgrader—Element increasing apologetic force (I'm terribly sorry, I really<br />
didn't mean to hurt you) Taking on Responsibility—speaker admitting the<br />
offense, including self-blame (How stupid of me), lack of intent (I didn't<br />
mean to do this), and admission of fact (I haven't graded it yet)<br />
Downgrading Responsibility or Severity of offense—(a) utterance reducing<br />
speaker's accountability for the offense, including excuse (My watch had<br />
stopped), justification (I was suddenly called to a meeting), claiming ignorance<br />
(1 didn't know you were expecting me), problematizing a precondition<br />
(we weren't supposed to meet before 12:00), or denial (I didn't do it); (b)<br />
utterance reducing severity of offense (I'm only 10 minutes late)<br />
Offer of Repair—speaker offering to remedy damage inflicted on offended party<br />
by an action to restitute H's entitlements (I'll pay for the damage, I'll have it<br />
marked tomorrow)<br />
Verbal Redress—speaker showing concern for offended party (I hope you<br />
weren't offended), efforts to appease (let me buy you a drink) or promise of<br />
forebearance (it won't happen again)<br />
Interrater reliability was established through consensus coding by three raters<br />
(Thai and English NS data) and two raters (Thai-English interlanguage data).<br />
The distribution of apology strategies across offense contexts and language<br />
groups is included in Appendix 4. The overall strategy distribution by language<br />
groups is displayed in Figure 4.1. Consistent with previous studies, IFID and Taking<br />
on Responsibility were used in more than half of the possible cases. Upgrading<br />
Fig. 4.1. Overall use of apology strategies (%) by NS of English (English, N = 30), NS of<br />
Thai (Thai, N = 136), and Thai NNS of English (IL, N = 288). IFID = Illocutionary Force<br />
Indicating Device, UG = Upgrader, TR = Taking on Responsibility, DG = Downgrading,<br />
RE = Offer of Repair, VR = Verbal Redress.
Apology, Perception, and Performance 95<br />
apologetic force, Downgrading Responsibility, and offer of Repair were supplied in<br />
one third of all contexts, and Verbal Redress was expressed on about every tenth<br />
occasion or less. The contextual distribution of strategies was examined by comparing<br />
strategy use in each context between pairs of language groups (NS English-NS<br />
Thai, NS English-IL users, NS Thai-IL users).<br />
IFID<br />
Overall, correlational analysis revealed no effects from the contextual factors on the<br />
selection of IFIDs, the exception being Obligation, which was found to covary with<br />
the use of IFIDs by the IL group (r = .6588, p < .002).<br />
In 16 out of the 20 situations, the three language groups agreed in their frequency<br />
of IFID suppliance. All groups used few IFIDs (between 14% and 50%) in low<br />
Severity contexts such as Order Change and Customs Form, as well as in high<br />
Severity situations such as Cheating Student and Smuggled Buddha. Thus the Thais<br />
matched their medium rating of Smuggled Buddha on Obligation by providing few<br />
apologetic formulas. Conversely, IFIDs were supplied with high frequency (80%-<br />
96%) in response to high Severity offenses such as Broken Statue and Food on<br />
Customer, which received identical IFID frequencies as the medium Severity offense<br />
Food on Waiter. The inverse relationship, for some offenses, between Severity<br />
and IFID was especially pronounced in the English data: in the lowest Severity<br />
context, Borrowed Book, 83% IFIDs were supplied, as in the medium Severity<br />
contexts Ungraded Paper and Wrong Order (86%), while the high Severity item<br />
Damaged Car only elicited 47% IFIDs. Contradiction and Poor Teacher were assigned<br />
to different severity categories (medium and high) in the American Assessments,<br />
yet no significant difference obtained in the frequencies of IFIDs in these<br />
contexts. These two items were both assessed as medium offenses by the Thai<br />
raters, which is reflected in the frequencies of IFIDs provided by Thai NS and IL<br />
users. All three language groups used the same frequencies of IFIDs in response to<br />
the Messages offenses (60%-81%). The difference in Severity between Private and<br />
Business messages was not matched by IFID suppliance.<br />
The theoretically interesting point emerging from the highly variable use of IFIDs<br />
is its intricate relationship with Severity of offense. By responding to low Severity<br />
offenses with an explicit apologetic formula, the offender symbolically emphasizes<br />
her eagerness to repair whatever minor norm infringement has occurred. Not using<br />
IFIDs to remedy high Severity offenses may have either of two functions. First, an<br />
offender may avoid admitting responsibility for the offense committed, which would<br />
be concomitant with an explicit apology. Second, and conversely, where the offender<br />
is prepared to assume responsibility, an all-purpose apologetic formula, which is also<br />
used for ritualistic apology, might not be felt to convey adequately a substantive<br />
apology for a major offense. An expression of apology propositionally related to the<br />
specific offense might be more apt to convey the sincerity of the speaker's regret.<br />
UPGRADING THE APOLOGY<br />
Upgrading of apology correlated highly with Obligation (r = .7628, p < .000) and<br />
Face-loss (r = .7248, p < .000) in the English data, and with Obligation in the<br />
Thai-English IL data (r = .6887, p < .001). Except for Obligation and IFID, which
96 Speech Act Realization<br />
covaried in the IL data, Upgrading, remarkably, was the only strategy that correlated<br />
with context-internal factors; it reveals that for the English NS, Obligation to<br />
apologize and Face-loss have their behavioral correlates not in the use of apologetic<br />
formulas, nor in expressing responsibility, but in the intensification of either or both<br />
of these strategies. This intensification makes good sense, considering the routinized<br />
nature of apologetic formulas: in order for them to count as sincere apology,<br />
they warrant Upgrading, which serves to emphasize their substantive nature. The IL<br />
users, while sharing with the English NS the preference to express Obligation<br />
through Upgrading, also increased their use of IFIDs with greater Obligation. If this<br />
increase can be taken to indicate that the NNS regard IFIDs just as apt a means to<br />
express Obligation as Upgraders, it might reveal a tendency for the NNS to underdifferentiate<br />
the apologetic function of IFIDs, compared to English NS use.<br />
In all three groups, Upgrading correlated highly with IFID (English: r = .6906,<br />
p < .001, Thai: r = .7636, p < .000, IL: r = .8014, p < .000). To the extent<br />
Upgrading was used to intensify an IFID, it could be claimed that the covariance<br />
expresses the relationship between an independent (IFID) and a dependent variable<br />
(Upgrader). Yet it must be remembered that Upgrading was coded when it operated<br />
on other apology strategies as well, for instance, on Lack of Intent as a subcategory<br />
of Responsibility (e.g., I really didn't mean to hurt you). However, no covariance<br />
could be established between Upgrading and any other apology strategy.<br />
Least Upgrading (0%-20%) was afforded across language groups to the low<br />
Obligation and low Face-loss context Order Change, the medium Obligation offenses<br />
Customs Form, Contradiction, and Smuggled Buddha, and to Poor Teacher,<br />
which had been rated high on Obligation by the American and medium by the Thai<br />
informants. Consistent Upgrading was furthermore provided in Business Message<br />
Low-High. In half of the offense contexts, Upgrading was supplied differentially by<br />
the three groups. In four situations, the Thai NS provided fewer Upgrading than the<br />
English NS, which is reflective of their lower ratings of these offenses for Obligation<br />
(Failed Student and the restaurant situations Wrong Order, Food on Waiter, and<br />
Food on Customer). The reverse—the Thais providing more Upgrading than the<br />
English NS—was true in Messed-up Bag, where the English informants did not<br />
upgrade at all, and Ungraded Paper. The IL users closely approximated target norms<br />
in all but one context (Messed-up Bag).<br />
TAKING ON RESPONSIBILITY<br />
Taking on Responsibility was the only apology strategy that covaried with a contextexternal<br />
factor (Distance) in all three groups (English: r = — .6514,p < .001, Thai<br />
r = -.6418, p < .001, IL: r = -.7165, p < .000). English and Thai NS<br />
and Thai-English NNS thus agreed in expressing more responsibility for the offensive<br />
act, the closer they were to the offended person; conversely, the more distant<br />
the relationship to the offended party, the less they were likely to admit accountability.<br />
Consistent with the ratings for Distance, in Damaged Car and Ruined Magazine,<br />
involving two friends, all groups of informants expressed Responsibility in 89%-<br />
98% of the possible cases. In the four Message offenses, taking place between coworkers,<br />
offenders assumed Responsibility with a frequency of 66%-93%. The
Apology, Perception, and Performance 97<br />
Thai NS and IL users afforded the same amount of Responsibility in the four<br />
Message contexts, regardless of status relationship (senior to junior or vice versa) or<br />
Severity of offense (private versus business message). The Thai NS consistently<br />
provided very high Responsibility frequencies (91%-93%), the IL users offering<br />
less (76%-84%). Curiously, the English NS expressed more Responsibility for<br />
failing to pass on a private message than for a business message. Two of the studentprofessor<br />
contexts also registered high for Responsibility: Borrowed Book and<br />
Failed Student. These items clearly illustrate that Severity of offense is not operative<br />
in determining whether or not an offender explicitly assumed Responsibility: Borrowed<br />
Book, which ranged lowest on Severity in the American ratings, elicited the<br />
highest amount of Responsibility (I forgot to take it along); Failed Student, one of<br />
the highest Severity offenses, was afforded almost as much Responsibility. High<br />
Distance relationships such as all waiter-customer interactions registered low on<br />
Responsibility (3%—47%). The failure of Severity to predict whether or not offenders<br />
take on Responsibility was also manifest at the lower end of the Responsibility<br />
scale: the lowest frequencies of Responsibility were observed in the lowest<br />
Severity contexts Order Change as well as in the high Severity context Smuggled<br />
Buddha. However, the Thais' reluctance to assume Responsibility for the Smuggled<br />
Buddha reflected its medium rating on Obligation. In seven contexts, more Thai NS<br />
than English NS assumed Responsibility, but only in two of these situations (Food<br />
on Customer and Broken Status) did the IL users follow their native pattern.<br />
DOWNGRADING RESPONSIBILITY OR SEVERITY<br />
All three groups used considerably fewer strategies to downgrade responsibility for<br />
the offense than to explicitly assume responsibility. The high Severity offense<br />
Smuggled Buddha was downgraded by far the most frequently by all groups. This<br />
supports further our earlier speculation about the low rating of Smuggled Buddha on<br />
Obligation to apologize by the Thai informants. Cheating Student, another high<br />
Severity offense, was heavily downgraded by the Thai informants, whereas the<br />
English NS downgraded this offense in less than half of the possible cases. Contexts<br />
in which high Responsibility corresponded to low Downgrading were Damaged<br />
Car, Ruined Magazine, Borrowed Book, Failed Student, and Customs Form. Moreover,<br />
informants downgraded responsibility less in some contexts with high occurrence<br />
of IFIDs (Food on Customer, Wrong Order, Food on Waiter, Broken Statue).<br />
Yet some offenses, notably the undelivered Messages, were redressed both by<br />
highly frequent use of Responsibility and fairly frequent Downgrading strategies<br />
(all groups between 25% and 52%).<br />
With only 6 out of 20 situations in which the three language groups did not differ<br />
in downgrading responsibility, this category displayed the most intergroup variability.<br />
In most of these contexts, the Thai NS downgraded more than either or both<br />
English NS and IL users. The IL users downgraded responsibility more than did the<br />
English NS in three of the restaurant situations and in the two teachers' meetings.<br />
OFFER OF REPAIR<br />
In the American data, Repair correlated negatively with Downgrading (r =<br />
— .6824, p < .001): offenders will be less prone to offer Repair the more they<br />
downgrade the offense. The same logic did not obtain for the Thai data. Neverthe-
98 Speech Act Realization<br />
less, for more than half of the offense contexts, English NS, Thai NS, and IL users<br />
agreed on whether or not they offered Repair.<br />
Least Repair (0%-10%) was offered in compensation for verbal or verbally<br />
conveyed offenses (Contradiction and Poor Teacher, the Private Message contexts,<br />
Order Change). Offenses that elicited Repair offers in more than half of the possible<br />
cases were those in which material compensation could be provided (Broken Statue,<br />
Ruined Magazine, Damaged Car, Food on Customer) or a neglected obligation<br />
honored (Ungraded Paper, Borrowed Book, Wrong Order, Failed Student). The<br />
language groups offered Repair differentially in three of the Message situations.<br />
Very few Repair offers were made by the Thai NS in all of the Message contexts,<br />
whereas the English NS and IL users offered Repair more frequently to compensate<br />
for the undelivered Business Messages. The IL users provided more Repair than the<br />
English NS in one type of high Distance context: customs official-traveler (Messedup<br />
Bag, Customs Form, Smuggled Buddha).<br />
VERBAL REDRESS<br />
This strategy ranged lowest in frequency across language groups and offense contexts.<br />
By far the most frequently chosen subcategory of Verbal Redress was Concern<br />
for Hearer. Groups agreed in affording very little Verbal Redress in the four<br />
custom official-traveler contexts, three of the waiter-customer contexts and of the<br />
student-professor interactions, and in all Message contexts (0%-19%). Only one<br />
offense, Contradiction, elicited more than 20% Verbal Redress from all groups,<br />
which is consistent with previous findings for a comparable offense in English and<br />
German (House, 1988,310) and Hebrew (Olshtain, 1989,161). In three contexts,<br />
only one of the groups supplied over 20% Verbal Redress: the Thai NS in Failed<br />
Student, and the IL users in Poor Teacher and Food on Waiter. In 11 contexts, either<br />
or both Thai NS and IL users provided more Verbal Redress than the English NS.<br />
Expressing concern for the offended party was thus a more common strategy for the<br />
Thais, regardless of the language they used.<br />
Table 4.4 summarizes the contexts in which the IL users differed from the<br />
English NS in their use of apology strategies. No differences were obtained for<br />
Ruined Magazine, Private Message High-Low, and Failed Student. Of the six<br />
apology strategies, learners and English NS differed least in Upgrading apologetic<br />
force and in the canonical strategies IFID and taking on Responsibility, whereas<br />
they differed most in the context-dependent strategies Repair, Verbal Redress, and<br />
Downgrading responsibility or severity of offense. If we apply Selinker's (1969)<br />
operational definition for determining effects of language transfer to these data, the<br />
white dots indicate effects of pragmatic transfer in the Thai-English IL users' apology<br />
performance. Fifty-five percent of the differences in the use of apology<br />
strategies can thus be attributed to pragmatic transfer. We are not at all convinced<br />
that this is an appropriate way to determine pragmatic transfer; however, since we<br />
are not aware of a comparably rigorous method that might be more in accordance<br />
with current thinking on cross-linguistic influence, we shall accept this procedure,<br />
and the obtained results, until further notice.<br />
To summarize our findings on the contextual distribution of apology strategies<br />
across the three languages:
Apology, Perception, and Performance 99<br />
Table 4.4. Differences in Apology Strategies Used between NS of English<br />
and Thai-English IL Users<br />
Messed-up Bag (13)<br />
Business Mess. L-H (7)<br />
Cheating Student (20)<br />
Ungraded Paper (17)<br />
Broken Statue (15)<br />
Borrowed Book (18)<br />
Food on Customer (10)<br />
Food on Waiter (12)<br />
Order Change (9)<br />
Poor Teacher (4)<br />
Customs Form (16)<br />
Smuggled Buddha (14)<br />
Wrong Order (11)<br />
Private Mess. H-L (6)<br />
Business Mess. H-L (8)<br />
Damaged Car (1)<br />
Contradiction<br />
UG TR IFID RE VR DG<br />
IFID = Illocutionary Force Indicating Device, UG = Upgrader, TR = Taking on Responsibility, DG — Downgrading,<br />
RE = Offer of Repair, VR = Verbal Redress.<br />
= IL users differ from English NS atp < 0.05.<br />
= IL users differ from English NS at/ < 0.05; English NS differ from Thai NS at/) < 0.05; no difference between<br />
Thai NS and IL users at p < 0.05.<br />
1. Contextual factors operated differentially in the strategy selection. While<br />
Obligation had been found to determine the choice of IFID by NS of German<br />
and British English (House, 1988), in the present study it was only the IL<br />
users who made their choice of IFID contingent on perceived Obligation to<br />
apologize.<br />
2. The strategy most sensitive to contextual factors was Upgrading: the more<br />
Obligation and Face-loss involved in an offense, the more Upgrading of<br />
apology would be provided. Previous studies had shown NS of Hebrew to<br />
increase apologetic force in direct proportion to Severity of offense and in<br />
indirect proportion to the offender's status vis-a-vis the offended party (Olshtain,<br />
1989); the latter finding was also obtained for NS of British English and<br />
German (House, 1988). These concurring results provide consistent evidence<br />
for the underspecification of IFID as an apology strategy: since routine<br />
formulas are used both in ritual and substantive function, sincere expression<br />
of regret warrants specific marking through intensification (House, 1988;<br />
Vollmer & Olshtain, 1989).<br />
3. Informants were more prone to explicitly express Responsibility for the<br />
offense the closer the relationship between the offender and the offended
100 Speech Act Realization<br />
person. This finding appears to contradict Wolfson's (1989) bulge hypothesis,<br />
according to which less redress is offered in interaction between very low<br />
and very high distance participants, whereas the most redress is provided in<br />
medium distance relationships. The contradiction is resolved if we reconsider<br />
the relationships comprised by our questionnaire contexts. Whereas they<br />
do include total strangers and acquaintances such as colleagues and friends,<br />
they do not extend to intimate relationships. Because the relationships spectrum<br />
covered in this study does not include the lowest end of the social<br />
distance continuum, our finding does not present counter-evidence to the<br />
bulge hypothesis.<br />
4. No effect of contextual factors was found on Downgrading Responsibility,<br />
Repair, and Verbal Redress. All three of these context-dependent strategies<br />
tended to be used more frequently by the IL users than by the English NS.<br />
5. The Thai-English IL users differed least from the English NS in their suppliance<br />
of Upgrading and the canonical strategies IFID and taking on Responsibility.<br />
Most differences occurred in the context-dependent strategies.<br />
6. More than half of the differences in apology suppliance can tentatively be<br />
attributed to pragmatic transfer from Thai apology patterns.<br />
By examining Thai learners' contextual perceptions and choices of apology<br />
strategies, we have shed light on their sociopragmatic knowledge of apology. We<br />
have not analyzed their actual wordings of apology strategies, that is, their pragmalinguistic<br />
knowledge of apologizing in English, which would require another<br />
study.<br />
How can we account for the relative oversuppliance of context-dependent strategies<br />
by the Thai learners Our finding is consistent with House's (1988) observation<br />
that NNS tend to do "too much of a good thing," a phenomenon she labelled<br />
"gushing," or, less benevolently, "waffling" (Edmondson & House, 1991). Waffling<br />
had been reported to characterize request realization by learners of different native<br />
and target languages (e.g., Blum-Kulka & Olshtain, 1986; Kasper, 1989). House<br />
(1988) was the first to draw our attention to a similar phenomenon in NNS apologizing,<br />
demonstrating that German learners of English provided more apology upgrading<br />
and expressions of responsibility than NS of British English. Edmondson and<br />
House (1991) point out that all the studies in which waffling is evident are written<br />
DCTs, and remind us that waffling was not present in the interactional, negotiated<br />
discourse elicited by open-ended role-plays (Kasper, 1981; see also Eisenstein &<br />
Bodman, Chapter 3). Therefore, waffling may well be an instrument effect of<br />
DCT—both of the Discourse Completion and Dialog Construction variety. But<br />
Edmondson and House (1991) go further in their account of waffling. In both roleplay<br />
data (Kasper, 1981; Edmondson, House, Kasper, & Stemmer, 1984) and DCT<br />
data (House, 1988, 1989), German learners of English were consistently shown to<br />
use fewer conventionalized requestive and apologetic forms than British NS. At the<br />
same time, waffling was extant in the DCT though not in the role-play data. In order<br />
to reconcile these seemingly contradictory findings, Edmondson and House refer to<br />
the differential cognitive demands of discourse production in face-to-face interaction<br />
(role-play) and written questionnaires. They argue that paucity of adequate
Apology, Perception, and Performance 101<br />
requestive and apologetic routines characterizes learners' speech act realization in<br />
both production tasks because such routines, though available in IL, are not "integrated<br />
into learners' discourse production systems" (285). Yet when learners' planning<br />
and execution of linguistic action does not operate under the pressures of<br />
conversational turn-taking, they will compensate for the lack of pragmatic routines<br />
by oversupplying nonconventionalized speech act realization strategies.<br />
Our study provides partial support for Edmondson and House's (1991) hypothesis.<br />
The Thai learners did indeed waffle, but they did not do so in compensation for<br />
lacking apology routines. Rather, our findings suggest that DCT, in whatever version,<br />
provides learners with an opportunity for knowledge display that is precluded<br />
for many NNS by the cognitive demands of face-to-face interaction. Comparison of<br />
learners' pragmatic performance in interactive discourse and DCT may thus throw<br />
light on the state of learners' discourse production systems, or the extent to which<br />
available pragmatic knowledge is readily accessible in conversation.<br />
Acknowledgments<br />
The research reported in this paper was supported in part by the University of<br />
Hawaii Project Development Fund, grant R-90-866-F-034-B-101. We gratefully<br />
acknowledge the assistance we received from colleagues and students in carrying<br />
out the study. Staff and students at the Language Institute, Chulalongkorn University,<br />
Bangkok, helped us collect the Thai data. Naoko Maeshiba and Naoko<br />
Yoshinaga made their English NS data available to us. Peggy DuFon and Satomi<br />
Takahashi have a major share in the data analysis. Special thanks are more than due<br />
J. D. Brown, Thorn Hudson, and Steven Ross for their assistance with the statistical<br />
processing of the data during different phases of the project.<br />
Sample item from Assessment questionnaire<br />
At a friend's home<br />
Appendix 1<br />
John and Paul are good friends. John borrowed Paul's car for the weekend. Unfortunately,<br />
when he was backing up to park, he didn't see a lamppost. He hit it and damaged the rear of<br />
the car. He is now returning the car to Paul.<br />
1. How close are John and Paul in this situation<br />
1 2 3 4 5<br />
very close<br />
very distant<br />
2. What is the status relationship between John and Paul<br />
1 2 3 4 5<br />
John higher John = John lower<br />
than Paul Paul than Paul
102 Speech Act Realization<br />
3. How serious is John's offense<br />
1 2 3 4 5<br />
not serious at all<br />
very serious<br />
4. Does John need to apologize<br />
1 2 3 4 5<br />
not at all<br />
absolutely<br />
5. How likely is Paul to accept John's apology<br />
1 2 3 4 5<br />
very likely<br />
very unlikely<br />
6. Does John gain or lose face in this situation<br />
1 2 3 4 5<br />
gains face<br />
loses face<br />
Appendix 2<br />
Sample item from Discourse Completion questionnaire<br />
At a friend's home<br />
John and Paul are good friends. John borrowed Paul's car for the weekend. Unfortunately,<br />
when he was backing up to park, he didn't see a lamppost. He hit it and damaged the rear of<br />
the car badly. He is now returning the car to Paul.<br />
PAUL: Is everything okay<br />
JOHN:<br />
PAUL:
Appendix 3: Mean Ratings (X) and Standard Deviations of Contextual<br />
Factors by American (n = 30) and Thai Informants (n = 30)<br />
Severity<br />
Obligation<br />
Likelihood<br />
Face-loss<br />
Distance<br />
Dominance<br />
Item<br />
Am<br />
Th<br />
Am<br />
Th<br />
Am<br />
Th<br />
Am<br />
Th<br />
Am<br />
Th<br />
Am<br />
Th<br />
1 X<br />
SD<br />
2 X<br />
SD<br />
3 X<br />
SD<br />
4X<br />
SD<br />
5 X<br />
SD<br />
6 X<br />
SD<br />
IX<br />
SD<br />
8 X<br />
SD<br />
9 X<br />
SD<br />
10 X<br />
SD<br />
11 X<br />
SD<br />
12 X<br />
SD<br />
13 X<br />
SD<br />
14 X<br />
SD<br />
15 X<br />
SD<br />
16 X<br />
SD<br />
17 X<br />
SD<br />
18 X<br />
SD<br />
19 X<br />
SD<br />
20 X<br />
SD<br />
4.10<br />
0.99<br />
2.60<br />
1.05<br />
2.90<br />
1.12<br />
4.50*<br />
0.56<br />
3.60<br />
0.96<br />
3.50*<br />
1.02<br />
4.60<br />
0.48<br />
4.40<br />
0.80<br />
2.10<br />
1.14<br />
4.60<br />
0.56<br />
3.30*<br />
0.89<br />
3.50<br />
1.28<br />
3.20<br />
1.19<br />
3.60*<br />
1.30<br />
4.60<br />
0.55<br />
2.90<br />
1.33<br />
2.80<br />
1.10<br />
1.00*<br />
0.91<br />
4.70<br />
0.53<br />
4.80*<br />
0.37<br />
3.70<br />
0.97<br />
2.50<br />
0.89<br />
3.40<br />
0.89<br />
3.60*<br />
1.05<br />
3.20<br />
1.02<br />
2.90*<br />
1.03<br />
4.50<br />
0.72<br />
4.10<br />
0.76<br />
1.80<br />
1.05<br />
4.30<br />
0.59<br />
2.60*<br />
1.05<br />
3.00<br />
1.05<br />
3.40<br />
1.09<br />
4.70*<br />
0.65<br />
4.50<br />
0.85<br />
2.80<br />
1.24<br />
3.20<br />
1.07<br />
3.10*<br />
1.20<br />
4.50<br />
0.96<br />
4.50*<br />
0.67<br />
4.90*<br />
0.30<br />
4.70*<br />
0.47<br />
3.10<br />
1.31<br />
4.20*<br />
0.87<br />
4.50*<br />
0.56<br />
4.30*<br />
0.97<br />
4.90*<br />
0.30<br />
4.90*<br />
0.40<br />
3.40*<br />
1.26<br />
5.00*<br />
0.18<br />
4.80*<br />
0.42<br />
4.80*<br />
0.50<br />
3.50*<br />
1.09<br />
3.40<br />
1.43<br />
4.90<br />
0.25<br />
3.30<br />
1.24<br />
3.70<br />
1.21<br />
3.80<br />
0.86<br />
4.90*<br />
0.25<br />
4.60<br />
0.80<br />
4.30*<br />
0.95<br />
3.50*<br />
0.97<br />
3.00<br />
1.25<br />
3.10*<br />
1.15<br />
3.90*<br />
1.00<br />
3.50*<br />
0.92<br />
4.50*<br />
0.62<br />
4.10*<br />
0.76<br />
2.30*<br />
0.96<br />
4.60*<br />
0.71<br />
4.00*<br />
1.03<br />
4.10*<br />
0.75<br />
4.10*<br />
0.73<br />
3.50<br />
1.63<br />
4.70<br />
0.59<br />
3.10<br />
1.16<br />
3.20<br />
1.26<br />
4.10<br />
0.93<br />
4.40*<br />
0.72<br />
4.60<br />
0.61<br />
3.60<br />
1.08<br />
4.50<br />
0.88<br />
3.30<br />
0.82<br />
2.70<br />
0.99<br />
3.80<br />
0.76<br />
3.90<br />
0.96<br />
3.00<br />
1.06<br />
3.60<br />
0.99<br />
4.10<br />
1.09<br />
2.90<br />
1.06<br />
2.90*<br />
1.45<br />
4.00<br />
1.13<br />
3.10*<br />
1.11<br />
2.30<br />
1.23<br />
2.20<br />
1.08<br />
3.10<br />
1.14<br />
4.10<br />
0.81<br />
4.50*<br />
0.56<br />
3.60<br />
0.98<br />
2.20<br />
0.99<br />
3.60<br />
1.02<br />
4.40<br />
0.67<br />
3.00<br />
0.96<br />
2.90<br />
1.19<br />
4.00<br />
0.87<br />
4.20<br />
0.82<br />
3.40<br />
1.05<br />
3.70<br />
0.85<br />
4.10<br />
1.15<br />
3.10<br />
1.12<br />
4.00*<br />
0.94<br />
4.00<br />
0.98<br />
3.70*<br />
0.91<br />
2.50<br />
1.26<br />
2.70<br />
1.30<br />
3.50<br />
1.12<br />
4.40<br />
1.03<br />
4.00*<br />
0.91<br />
3.50<br />
1.20<br />
2.40<br />
1.02<br />
4.30<br />
0.83<br />
3.50<br />
1.15<br />
2.60<br />
0.96<br />
3.70*<br />
1.06<br />
4.00<br />
1.02<br />
3.60<br />
0.87<br />
4.50<br />
0.76<br />
3.90<br />
0.94<br />
2.90<br />
1.20<br />
4.80*<br />
0.40<br />
4.20<br />
0.84<br />
4.50*<br />
0.88<br />
2.50*<br />
1.23<br />
3.80*<br />
1.39<br />
4.30<br />
0.86<br />
3.60*<br />
0.95<br />
3.30*<br />
0.97<br />
3.20*<br />
1.12<br />
4.60<br />
0.84<br />
4.70<br />
0.53<br />
4.30<br />
0.80<br />
3.90<br />
0.80<br />
2.60<br />
1.07<br />
2.60*<br />
1.14<br />
3.80<br />
0.86<br />
3.70<br />
0.83<br />
4.10<br />
0.88<br />
4.10<br />
0.91<br />
3.00<br />
0.41<br />
4.40*<br />
0.80<br />
3.80<br />
0.79<br />
3.80*<br />
0.82<br />
3.80*<br />
0.82<br />
4.60*<br />
0.66<br />
4.70<br />
0.59<br />
4.10*<br />
0.81<br />
3.90*<br />
0.77<br />
4.10*<br />
0.60<br />
4.50<br />
0.67<br />
4.70<br />
0.57<br />
1.50<br />
0.62<br />
1.80*<br />
0.90<br />
2.80*<br />
0.95<br />
3.10<br />
1.24<br />
3.10*<br />
0.69<br />
3.20*<br />
0.73<br />
3.10<br />
0.75<br />
3.00*<br />
0.91<br />
4.50<br />
0.76<br />
4.70<br />
0.87<br />
4.40<br />
0.84<br />
4.50<br />
0.96<br />
4.80*<br />
0.54<br />
3.90<br />
1.36<br />
4.80<br />
0.56<br />
4.60<br />
0.72<br />
3.20<br />
0.78<br />
3.10*<br />
0.75<br />
3.10<br />
0.79<br />
3.50*<br />
0.85<br />
1.40<br />
0.66<br />
2.80*<br />
0.30<br />
3.90*<br />
0.86<br />
3.50<br />
1.33<br />
3.60*<br />
0.76<br />
3.70*<br />
0.86<br />
2.90<br />
0.87<br />
3.70*<br />
0.94<br />
4.60<br />
0.72<br />
4.60<br />
0.60<br />
4.60<br />
0.66<br />
4.60<br />
0.72<br />
4.30*<br />
0.83<br />
4.40<br />
0.80<br />
4.60<br />
0.80<br />
4.20<br />
0.83<br />
2.90<br />
1.00<br />
2.50*<br />
1.12<br />
2.80<br />
1.01<br />
2.90*<br />
1.09<br />
2.80<br />
0.56<br />
3.00*<br />
0.55<br />
2.90*<br />
0.36<br />
3.10<br />
0.63<br />
1.70<br />
0.94<br />
4.40<br />
0.91<br />
1.30*<br />
0.54<br />
4.40<br />
0.88<br />
4.20<br />
1.01<br />
1.60<br />
1.14<br />
1.70<br />
0.88<br />
4.00<br />
1.28<br />
3.50<br />
1.41<br />
2.30*<br />
1.42<br />
3.40*<br />
1.52<br />
2.00*<br />
1.10<br />
4.60<br />
0.76<br />
1.90<br />
1.11<br />
4.30<br />
1.04<br />
1.40<br />
0.92<br />
2.70<br />
0.85<br />
4.00*<br />
1.17<br />
2.20*<br />
1.31<br />
3.10<br />
0.50<br />
1.40<br />
1.05<br />
4.60<br />
0.84<br />
1.90*<br />
0.98<br />
4.50<br />
0.96<br />
3.90<br />
1.37<br />
1.50<br />
0.81<br />
2.00<br />
0.87<br />
3.60<br />
1.17<br />
2.90<br />
0.50<br />
3.30*<br />
0.94<br />
2.30*<br />
0.94<br />
2.90*<br />
0.44<br />
4.40<br />
0.99<br />
1.60<br />
0.87<br />
4.40<br />
0.84<br />
1.60<br />
0.80<br />
Item numbers in the leftmost column refer to the following offense contexts: 1 — Damaged car, 2 = Ruined magazine, 3 —<br />
Contradiction, 4 = Poor teacher, 5 = Private message low-high, 6 = Private message high-low, 7 = Business message low-high, 8<br />
= Business message high-low, 9 = Order change, 30 = Food on customer, 11 = Wrong order, 12 = Food on waiter, 13 = Messedup<br />
bag, 14 = Smuggled Buddha, 15 = Broken statue, 16 = Customs form, 17 = Ungraded paper, 18 - Borrowed book, 19 =<br />
Failed student, 20 ~ Cheating student.<br />
* = p < 0.05.
Appendix 4: Use of Apology Strategies (%) by NS of English (EN, n = 30),<br />
NS of Thai (TN, n = 136), and Thai NNS of English (IL, n = 228)<br />
IFID<br />
UG<br />
TR<br />
DG<br />
RE<br />
VR<br />
Item<br />
EN<br />
TN<br />
IL<br />
EN<br />
TN<br />
IL<br />
EN<br />
TN<br />
IL<br />
EN<br />
TN<br />
IL<br />
EN<br />
TN<br />
IL<br />
EN<br />
TN<br />
IL<br />
1 n<br />
%<br />
2 n<br />
%<br />
3 n<br />
%<br />
4 n<br />
%<br />
5 n<br />
%<br />
6 n<br />
%<br />
7 n<br />
%<br />
8 n<br />
%<br />
9 n<br />
%<br />
10 n<br />
14<br />
47<br />
20<br />
67<br />
18<br />
60<br />
20<br />
67<br />
23<br />
77<br />
19<br />
63<br />
22<br />
76<br />
18<br />
60<br />
8<br />
27<br />
28<br />
82<br />
60<br />
85<br />
66<br />
79<br />
59<br />
79<br />
63<br />
104<br />
78<br />
98<br />
73<br />
107<br />
81<br />
92<br />
70<br />
31<br />
23<br />
115<br />
185<br />
64<br />
211<br />
75<br />
185<br />
65<br />
180<br />
64<br />
219<br />
77<br />
217<br />
76<br />
219<br />
77<br />
225<br />
78<br />
41<br />
14<br />
273<br />
11<br />
37<br />
8<br />
27<br />
4<br />
14<br />
5<br />
17<br />
14<br />
47<br />
10<br />
33<br />
16<br />
55<br />
16<br />
53<br />
2<br />
7<br />
23<br />
35<br />
27<br />
21<br />
16<br />
17<br />
13<br />
22<br />
18<br />
73<br />
55 C<br />
63<br />
47<br />
73<br />
55<br />
58<br />
44<br />
4<br />
3<br />
64<br />
83<br />
29<br />
66<br />
23<br />
41<br />
14<br />
49<br />
18<br />
99<br />
35 C<br />
102<br />
36<br />
153<br />
54<br />
117<br />
41<br />
4<br />
1<br />
206<br />
27<br />
90<br />
27<br />
90<br />
11<br />
37"<br />
19<br />
63<br />
27<br />
90<br />
24<br />
80<br />
19<br />
66»<br />
22<br />
73<br />
1<br />
3»<br />
2<br />
117<br />
89<br />
117<br />
91<br />
94<br />
71"<br />
81<br />
65<br />
121<br />
91<br />
125<br />
93<br />
122<br />
92"<br />
121<br />
92<br />
21<br />
16 a<br />
62<br />
262<br />
91<br />
276<br />
98<br />
139<br />
49<br />
136<br />
49<br />
217<br />
76<br />
236<br />
81<br />
238<br />
84<br />
238<br />
81<br />
25<br />
9<br />
108<br />
1<br />
3.b<br />
1<br />
3 a<br />
13<br />
43 b<br />
15<br />
50 b<br />
12<br />
40<br />
13<br />
43<br />
12<br />
41"<br />
8<br />
27<br />
12<br />
40ab<br />
3<br />
29<br />
22 a<br />
28<br />
22"=<br />
72<br />
54=<br />
62<br />
50 C<br />
63<br />
47<br />
70<br />
52=<br />
42<br />
32<br />
44<br />
34<br />
9<br />
7ac<br />
22<br />
44<br />
15"<br />
7<br />
2'<br />
49<br />
[7bc<br />
66<br />
24 bc<br />
86<br />
30<br />
85<br />
30»<br />
68<br />
24b<br />
68<br />
26<br />
0<br />
Qbc<br />
7<br />
20<br />
67<br />
18<br />
60<br />
0<br />
0<br />
1<br />
3<br />
1<br />
3<br />
0<br />
0<br />
10<br />
35*<br />
8<br />
21'<br />
1<br />
3<br />
25<br />
94<br />
72<br />
88<br />
68<br />
10<br />
8<br />
9<br />
7<br />
7<br />
5<br />
11<br />
8<br />
3<br />
2ac<br />
9<br />
7»c<br />
14<br />
10'<br />
84<br />
217<br />
75<br />
196<br />
70<br />
16<br />
6<br />
11<br />
4<br />
14<br />
5<br />
13<br />
5<br />
37<br />
I3bc<br />
51<br />
18''<br />
3<br />
l c<br />
188<br />
2<br />
7<br />
1<br />
3<br />
7<br />
23<br />
2<br />
8"<br />
4<br />
13<br />
2<br />
7"<br />
1<br />
5 ab<br />
1<br />
3ab<br />
3<br />
[Jab<br />
1<br />
18<br />
14<br />
12<br />
9<br />
41<br />
31<br />
20<br />
16<br />
15<br />
11<br />
14<br />
10<br />
18<br />
[4a<br />
22<br />
17"<br />
2<br />
2 a<br />
10<br />
31<br />
11<br />
13<br />
5<br />
74<br />
26<br />
58<br />
21 b<br />
50<br />
18<br />
53<br />
19 b<br />
38<br />
13"<br />
34<br />
12"<br />
3<br />
l b<br />
10
%<br />
11 n<br />
%<br />
12 n<br />
%<br />
13 n<br />
%<br />
14 n<br />
%<br />
15 n<br />
%<br />
16 n<br />
%<br />
17 n<br />
%<br />
18 n<br />
%<br />
19 n<br />
%<br />
20 n<br />
%<br />
93 87<br />
24 85<br />
80 64<br />
28 115<br />
93 87<br />
10 76<br />
33» b 60=<br />
11 39<br />
39 30<br />
23 108<br />
79 82<br />
7 49<br />
24 37<br />
25 75<br />
83 ab<br />
57=<br />
25 93<br />
83 b 70<br />
19 98<br />
63 73<br />
7 39<br />
24 b 31 =<br />
96<br />
236<br />
84<br />
263<br />
92<br />
186<br />
65 b<br />
105<br />
37<br />
236<br />
83<br />
108<br />
38<br />
152<br />
53"<br />
166<br />
59 b<br />
221<br />
78<br />
135<br />
48 be<br />
77=<br />
13<br />
43*<br />
22<br />
73=<br />
0<br />
Qab<br />
7<br />
25<br />
11<br />
38<br />
5<br />
17<br />
4<br />
13=<br />
10<br />
33<br />
13<br />
43=<br />
4<br />
14<br />
48 »=<br />
26<br />
20»<br />
57<br />
43=<br />
41<br />
32'<br />
21<br />
16<br />
61<br />
46<br />
24<br />
18<br />
60<br />
46 ac<br />
50<br />
38<br />
29<br />
22"<br />
17<br />
13<br />
73=<br />
84<br />
30<br />
177<br />
63<br />
75<br />
26"<br />
44<br />
15<br />
177<br />
62<br />
60<br />
21<br />
71<br />
25=<br />
64<br />
23<br />
83<br />
29<br />
52<br />
18<br />
7ab<br />
9<br />
30<br />
6<br />
20=<br />
14<br />
47<br />
3<br />
11 =<br />
7<br />
24 = b<br />
26<br />
87<br />
28<br />
93<br />
29<br />
97<br />
27<br />
90<br />
10<br />
35<br />
47 a<br />
41<br />
30<br />
57<br />
43=<br />
66<br />
52<br />
32<br />
25==<br />
91<br />
69=<br />
106<br />
80<br />
94<br />
71<br />
124<br />
93<br />
120<br />
89<br />
59<br />
46<br />
38 b<br />
78<br />
28<br />
99<br />
35<br />
133<br />
47<br />
28<br />
10=<br />
140<br />
49 b<br />
231<br />
81<br />
212<br />
74<br />
212<br />
81<br />
246<br />
86<br />
122<br />
43<br />
10 b<br />
4<br />
13<br />
9<br />
30"<br />
4<br />
13= b<br />
22<br />
79<br />
2<br />
7<br />
7<br />
24<br />
14<br />
47=<br />
2<br />
7ab<br />
2<br />
7 a<br />
14<br />
48 ab<br />
16 C<br />
24<br />
18<br />
28<br />
21<br />
42<br />
33=<br />
105<br />
80<br />
15<br />
11<br />
28<br />
21<br />
106<br />
80==<br />
40<br />
30=<br />
36<br />
27ac<br />
92<br />
72*<br />
2bc<br />
31<br />
11<br />
36<br />
13 b<br />
79<br />
28 b<br />
260<br />
91<br />
13<br />
5<br />
48<br />
17<br />
114<br />
40=<br />
114<br />
ll b<br />
7<br />
2=<br />
201<br />
71"<br />
83<br />
19<br />
63<br />
10<br />
33<br />
1<br />
3 b<br />
4<br />
14 b<br />
13<br />
45<br />
2<br />
7ab<br />
12<br />
40ab<br />
17<br />
57<br />
21<br />
70<br />
3<br />
10<br />
63<br />
79<br />
60<br />
50<br />
38<br />
12<br />
9<br />
21<br />
16=<br />
75<br />
57<br />
23<br />
17»c<br />
84<br />
64=<br />
78<br />
59<br />
85<br />
63<br />
22<br />
17<br />
66<br />
230<br />
82<br />
124<br />
44<br />
54<br />
19 b<br />
146<br />
5]bc<br />
181<br />
64<br />
123<br />
44 be<br />
216<br />
76 h<br />
216<br />
65<br />
197<br />
69<br />
45<br />
16<br />
2<br />
0<br />
l b<br />
1<br />
4»b<br />
0<br />
0<br />
0<br />
0<br />
0<br />
l=<br />
b<br />
0<br />
0<br />
0<br />
1 =<br />
1<br />
3"<br />
2<br />
7=<br />
0<br />
]ab<br />
8<br />
11<br />
23<br />
17=<br />
5<br />
4<br />
4<br />
3<br />
8<br />
6"<br />
0<br />
0<br />
7<br />
5=<br />
18<br />
14==<br />
32<br />
24 a<br />
6<br />
5=<br />
4<br />
16<br />
6"<br />
82<br />
29 b<br />
10<br />
4<br />
3<br />
1<br />
12<br />
4"<br />
5<br />
1<br />
7<br />
2<br />
7<br />
6=<br />
36<br />
13<br />
21<br />
7 b<br />
The leftmost column specifies offense contexts, cf. Appendix 3.<br />
IFID = Illocutionary Force Indicating Device, UG = Upgrader, TR ~ Taking on Responsibility, DC = Downgrading, RE = Offer of Repair, VR = Verbal Redress.<br />
a = EN and TN different at p < 0.05.<br />
>> = EN and 1L different at p < 0.05.<br />
= = TN and IL different at p < 0.05.
106 Speech Act Realization<br />
References<br />
Arndt, H., & Janney, R. W. (in press). Universality and relativity in cross-cultural politeness<br />
research: A historical perspective. Multilingua.<br />
Aronsson, K., & Satterlund-Larsson, U. (1987). Politeness strategies and doctor-patient<br />
communication: On the social choreography of collaborative thinking. Journal of<br />
Language and Social Psychology, 6, 1-27.<br />
Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requestive<br />
behaviour in five languages. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural<br />
pragmatics: Requests and apologies (123-54). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Appendix: The CCSARP coding<br />
manual. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (273-94). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Studies in Second Language Acquisition, 8, 65-119.<br />
Borkin, A., & Reinhart, S. M. (1978). "Excuse me" and "I'm sorry." TESOL Quarterly, 12,<br />
57-79.<br />
Brown, P., & Eraser, C. (1979). Speech as a marker of situation. In K. Scherer & H. Giles<br />
(Eds.), Social markers in speech (33-63). Cambridge: Cambridge University Press.<br />
Brown, P., & Levinson, S. D. (1987). Politeness: Some universals in language usage.<br />
Cambridge: Cambridge University Press.<br />
Coulmas, F. (1981). "Poison to your soul." Thanks and apologies contrastively viewed. In<br />
F. Coulmas (Ed.), Conversational routine (69-91). The Hague: Mouton.<br />
Edmondson, W., & House, J. (1991). Do learners talk too much The waffle phenomenon in<br />
interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood<br />
Smith, & M. Swain (Eds.), Foreign/second language pedagogy research (273-86).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Edmondson, W., House, J., Kasper, G., & Stemmer, B. (1984). Learning the pragmatics of<br />
discourse. Applied Linguistics, 5, 113-27.<br />
Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by<br />
native and non-native speakers of American English. Applied Linguistics, 7, 167-85.<br />
Garcia, C. (1989). Apologizing in English: Politeness strategies used by native and nonnative<br />
speakers. Multilingua, 8, 3-20.<br />
Goffman, E. (1971). Relations in public. Harmondsworth: Penguin.<br />
Holmes, J. (1989). Sex differences and apologies: One aspect of communicative competence.<br />
Applied Linguistics, 10, 194-213.<br />
House, J. (1988). "Oh excuse me please . . .": Apologizing in a foreign language. In B.<br />
Kettemann, P. Bierbaumer, A. Fill & A. Karpf (Eds.), Englisch ah Zweitsprache<br />
(303-27). Tubingen: Narr.<br />
House, J. (1989). Politeness in English and German: The functions of "please" and "bitte." In<br />
S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests<br />
and apologies (96-119). Norwood, NJ: Ablex.<br />
House, J., & Kasper, G. (1981). Politeness markers in English and German. In F. Coulmas<br />
(Ed.), Conversational routine (157-85). The Hague: Mouton.<br />
Hu, H. C. (1944). The Chinese concepts of "face." American Anthropologist, 46, 45-64.<br />
Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr.<br />
Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden,<br />
D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse<br />
and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.
Apology, Perception, and Performance 107<br />
Kasper, G. (1990). Linguistic politeness: Current research issues. Journal of Pragmatics, 14,<br />
193-218.<br />
Kasper, G. (1989). [Native and nonnative apologies]. Unpublished manuscript.<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In<br />
S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49).<br />
Rowley, MA: Newbury House.<br />
Olshtain, E. (1989). Apologies across languages. In S. Blum-Kulka, J. House, & G. Kasper<br />
(Eds.), Cross-cultural pragmatics: Requests and apologies (155-73). Norwood, NJ:<br />
Ablex.<br />
Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd<br />
(Eds.), Sociolinguistics and second language acquisition (18-35). Rowley, MA:<br />
Newbury House.<br />
Olshtain, E., & Weinbach, L. (1987). Complaints: A study of speech act behavior among<br />
native and non-native speakers of Hebrew. In J. Verschueren & M. Bertucelli-Papi,<br />
(Eds.). The pragmatic perspective (195-208). Amsterdam: Benjamins.<br />
Owen, M. (1983). Apologies and remedial interchanges. Berlin: Mouton.<br />
Rintell, E., & Mitchell, C. L. (1989). Studying requests and apologies: An inquiry into<br />
method. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (248-72). Norwood, NJ: Ablex.<br />
Selinker, L. (1969). Language transfer. General Linguistics, 9, 67-92.<br />
Trosborg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11,<br />
147-67.<br />
Vollmer, H. J., & Olshtain, E. (1989). The language of apologies in German. In S. Blum-<br />
Kulka, J. House & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies<br />
(197-218). Norwood, NJ: Ablex.<br />
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House.<br />
Wolfson, N., Marmor, T., & Jones, S. (1989). Problems in the comparison of speech acts<br />
across cultures. In S. Blum-Kulka, J. House & G. Kasper (Eds.), Cross-cultural<br />
pragmatics: Requests and apologies (174-96) Norwood, NJ: Ablex.
5<br />
Interlanguage Features<br />
of the Speech Act of Complaining<br />
ELITE OLSHTAIN and LIORA WEINBACH<br />
In the speech act of complaining, the speaker (S) expresses displeasure or<br />
annoyance—censure—as a reaction to a past or ongoing action, the consequences<br />
of which are perceived by S as affecting her unfavorably. This complaint is usually<br />
addressed to the hearer (H) whom the S holds, at least partially, responsible for the<br />
offensive action. For the purpose of this study, censure will be assumed to have been<br />
expressed whenever S chooses to verbalize her disapproval of the violation.<br />
From the speaker's point of view, the following preconditions need to be fulfilled<br />
in order for the speech act of complaining to take place:<br />
1. H performs a socially unacceptable act (SUA) that is contrary to a social<br />
code of behavioral norms shared by S and H.<br />
2. S perceives the SUA as having unfavorable consequences of herself, and/or<br />
for the general public.<br />
3. The verbal expression of S relates post facto directly or indirectly to the<br />
SUA, thus having the illocutionary force of censure.<br />
4. S perceives the SUA as: (a) freeing S (at least partially) from the implicit<br />
understanding of a social cooperative relationship with H; S therefore<br />
chooses to express her frustration or annoyance, although the result will be a<br />
"conflictive" type of illocution in Leech's terms (Leech, 1983, 104); and<br />
(b) giving S the legitimate right to ask for repair in order to undo the SUA,<br />
either for her benefit or for the public benefit. It is the latter perception that<br />
leads to instrumental complaints aimed at "changing things" that do not meet<br />
with our standards or expectations. The main goal of such instrumental<br />
complaints is to ensure that H performs some action of repair as a result of<br />
the complaint.<br />
Since complaining or expressing censure is a face-threatening act directed toward<br />
H, and since its conflictive nature might result in a breach of the social goal of<br />
maintaining comity and harmony between S and H (Leech, 1983), S is faced with a<br />
series of "payoff" considerations before the actual realization of censure can take<br />
108
The Speech Act of Complaining 109<br />
place. Obviously, the annoyed and frustrated S needs to air her feelings in order to<br />
regain emotional balance or in order to get "things done properly." But how does<br />
this compare to the risk of losing a friend or causing another person considerable<br />
embarrassment and possibly even anger which might affect his willingness to carry<br />
out repair In other words, to what extent might S be willing to risk certain social<br />
advantages in order to rid herself of the frustration caused by the SUA or in order to<br />
improve matters and correct the situation (in case of bad service, bureaucracy, etc.)<br />
In cases where the damage done could be repaired by H, S needs to consider the<br />
most effective route for such a remedy to occur; a straightforward reproach may not<br />
ensure the repair process, while the avoidance of complaining might.<br />
Payoff Considerations<br />
According to Brown and Levinson (1987, 60) the realization of a face-threatening<br />
act (FTA), such as the speech act of censure, follows a certain path of decision<br />
making. We contend that with respect to the speech act of complaining this series of<br />
decision-making steps is really a sequence of payoff considerations. There are at<br />
least three main junctures of decision making:<br />
1. At the first juncture of decision making, the speaker has the possibility of<br />
completely opting out from performing the act; in other words, refraining from<br />
expressing censure altogether. Such opting out may have certain payoffs from S's<br />
point of view: she can get credit for being tactful and polite; she can avoid the<br />
consequences of potentially damaging H's face; and she can give H the feeling that<br />
S is considerate and cares for H (who is responsible for the violation). On the other<br />
hand, S might remain frustrated since she has not expressed her feelings of annoyance<br />
and disappointment and has not reprimanded H for the SUA. Furthermore, in<br />
cases of instrumental complaints (leading to expected repair) mentioned earlier,<br />
performing the act may risk the potential for repair and at times might endanger S's<br />
position with respect to H. Such considerations of risk will affect S's decision to<br />
carry out the act or opt out.<br />
2. At the second node, S has the possibility of choosing to carry out the act "on"<br />
or "off" record. Off record in the case of censure would be some hint related to the<br />
inconvenience that resulted from the SUA, without explicitly mentioning either the<br />
SUA or H. Such an evasion of the clear and obvious expression of censure would<br />
present similar payoffs for S as in step 1 above, but if H chooses to pick up the hint,<br />
then some reaction from H, either an apology or an offer of repair, might set things<br />
straight and S would no longer feel annoyed. For example, if in the situation where<br />
there is a line of people waiting to get to a counter a person suddenly tries to push in<br />
front of S, the latter might say something like, "I'm the last in line." No real facedamage<br />
has been committed since there was no direct mention of the offender or the<br />
offense, yet the offender will, most probably, go and stand behind S and thus resolve<br />
the problem. On the other hand, if S decides to express censure "on" record, she<br />
moves on to the next juncture of decision making and payoff considerations.<br />
3. At the third step of Brown and Levinson's path, S has a choice to realize the<br />
speech act with or without redress. If she chooses not to use redress, then censure is
110 Speech Act Realization<br />
expressed unmitigatedly as a statement or request that explicitly mentions the SUA<br />
and/or H as violator. In this case, payoffs have been considered irrelevant or less<br />
important and S chooses to express her frustration without any mitigation. In fact,<br />
there is a possibility that S will not even refrain from a full realization of the<br />
complaint and will indeed use a personal threat or insult. The risk that S is willing to<br />
take is open conflict with H. In turn, H might react to such a complaint with an open<br />
attack on S, serving as retaliation for the face-damage suffered by H.<br />
On the other hand, if S chooses to use redress as part of the realization of the<br />
speech act of complaining, then two options are open: the positive politeness orientation<br />
or the negative politeness orientation. In the case of censure, a speaker opting<br />
for a positive politeness orientation would probably still prefer the explicit mention<br />
of both the SUA and H but with some expression of mutual concern and understanding,<br />
which creates mitigation. The payoff for such a choice would be the fact that S<br />
would fully air her frustration but at the same time would express personal interest<br />
or understanding concerning H. In an in-group context, this orientation will not<br />
necessarily lead to real conflict and the risk will be relatively small. Thus, if a<br />
mother tells her son, "You have left some dirty dishes in the sink again," and there is<br />
an understanding between them that he should place the dishes in the dishwasher,<br />
such a complaint will not greatly damage the relations between the two.<br />
If the decision is negative politeness oriented, the strategy to perform the act of<br />
censure might take the form of a mitigated expression. In the case of censure, such<br />
mitigated strategies might be realized as a conventional request for repair, where<br />
applicable, or as a statement relating to the SUA but not directly to H. Thus in the<br />
case of pushing into the line, S might say "Excuse me, sir. There's a line here," or in<br />
the case of a latecomer to a meeting, "I was really hoping to start our meeting on<br />
time." It seems that in the negative politeness orientation, even when the complaint<br />
is explicitly expressed, mitigation might lower the risk and thus lessen the effect on<br />
payoff.<br />
A Cross-Cultural Comparison of Native Speakers' Norms<br />
In order to investigate features of interlanguage with respect to the speech act of<br />
complaining, it was necessary to first establish baseline intra-cultural and crosscultural<br />
norms. We started with a study of the realization of complaints in Hebrew<br />
(Olshtain & Weinbach, 1987; Weinbach, 1988). Data were collected via responses<br />
to a discourse completion questionnaire, consisting of 20 situations. Each situation<br />
presented respondents with a detailed description of the context, the social status<br />
and the social distance between the interlocutors, and details on whether there exists<br />
mutual commitment or an explicit/implicit contract (social obligation) that requires<br />
H to be responsible for the SUA (see Appendix). The data collected from 35 Israeli<br />
university students led to the development of a scale for the perception of the<br />
severity of the complaint. This scale enabled us to categorize complaint realizations<br />
on five points along a continuum. The categories were defined in terms of the<br />
degree of face-threat that S was willing to undertake when expressing censure or, in<br />
other words, in terms of payoff considerations as described above. The following<br />
five realization patterns were established.
The Speech Act of Complaining 111<br />
1. Below the level of reproach are various realizations that enable S to avoid<br />
explicit mention of the offensive event or direct focus on S. In fact, this category<br />
deliberately leaves room for H's interpretation as to whether or not a complaint was<br />
actually voiced. A typical reaction in this category would be a general remark: for<br />
example, if H spilled something and damaged the tablecloth or some papers, S<br />
might say: "Such things happen," or "Don't worry about it, there's no real damage."<br />
In terms of payoff considerations, S makes sure (in this realization) that there is no<br />
real breach in social harmony as a result of the SUA and that H does not feel<br />
reprimanded too strongly.<br />
2. Expression of annoyance or disapproval encompasses various realizations<br />
that are vague and indirect and do not explicitly mention either the SUA or H, but do<br />
express general annoyance at the violation. S still tries to avoid open confrontation<br />
with H, but makes it clear that there was a violation of some sort. Thus, a realization<br />
of a complaint in a statement such as: "Such lack of consideration!" or "This is<br />
really unacceptable behavior" does not specify exactly what was wrong and who<br />
was responsible. H can choose to interpret it as a complaint but he can also ignore<br />
the illocutionary intent.<br />
3. Explicit complaint refers to realizations where S has made the decision to use<br />
an open face-threatening act toward H, but to instigate no sanctions. In this category<br />
S explicitly refers to either the SUA, or H, or both. Thus, any of the following<br />
realizations of complaints in a case where a doctor postponed an operation without<br />
informing the patient would be considered an "explicit complaint": "You're inconsiderate!"<br />
"One should not postpone this type of operation," or "You should not<br />
have postponed such an operation."<br />
4. Accusation and warning is expressed as a complaint when S chooses to<br />
perform an open face-threatening act and further implies potential sanctions against<br />
H. An example of this type of complaint addressed toward someone who was very<br />
late for a meeting is, "Next time I'll let you wait for hours."<br />
5. Immediate threat is expressed when S chooses to openly attack H. A realization<br />
of this strategy is, "You'd better pay the money right now"; "I'm not moving<br />
one inch before you change my appointment." This strategy can also consist of<br />
curses and direct insults, such as "You're an idiot."<br />
Olshtain and Weinbach (1987) suggested that these five categories of realization<br />
patterns of complaining make up the speech act set for this speech act (Olshtain &<br />
Cohen, 1983). Table 5.1 indicates that speakers of Hebrew tend to cluster around the<br />
three central strategies: disapproval, complaint, and warning. When the speaker is<br />
of lower status than the hearer, the tendency is to opt for less severe complaints<br />
(disapproval and complaint); when the interlocutors are equals or the speaker has<br />
higher social status, the tendency is stronger for the two more severe realizations<br />
(complaint and warning). Few respondents choose the softest strategy, "below level<br />
of reproach"—mostly in cases where the speaker is of lower status than the hearer;<br />
even fewer make use of "immediate threat," which seerns to increase with social<br />
power.<br />
The second study to be reported here was designed to compare different cultural<br />
groups with respect to a set of situations in terms of their preference for the optingout<br />
strategy, the direct strategy, the unmitigated, and the mitigated strategies (Katz,<br />
1987). In this study the realizations of complaining clustered for all three groups in
112 Speech Act Realization<br />
Table 5.1. Distribution of Strategies by Social Status<br />
among Native Speakers of Hebrew (% of actual responses<br />
of each strategy in the particular social status context)<br />
Strategy<br />
Below level of reproach<br />
Disapproval<br />
Complaint<br />
Warning<br />
Threat<br />
S(-P)<br />
H( + P)<br />
(%)<br />
18<br />
28<br />
38<br />
14<br />
2<br />
100<br />
S — speaker, H = hearer, P = (social) power.<br />
Equals<br />
(%)<br />
8<br />
15<br />
48<br />
25<br />
4<br />
100<br />
S(+P)<br />
H(-P)<br />
(%)<br />
6<br />
13<br />
43<br />
32<br />
6<br />
100<br />
three major categories: the unmitigated form, comparable to what we called earlier<br />
the "warning"; the mitigated strategy, in fact, the conventional complaint; and the<br />
indirect strategy, which resembles implicit disapproval. Unique to this study is that<br />
it also allowed for the strategy of opting out in cases where the respondents chose to<br />
indicate that they would prefer not to offer a verbal reaction to the situation given.<br />
Subjects in this study consisted of two English-language groups—one British<br />
(n = 27) and one American (n = 23)—and a Hebrew-speaking group (n = 25). The<br />
English-speaking informants were students and staff from Lancaster and Birmingham<br />
Universities in England, and from Columbia and New York Universities in<br />
New York City; the Hebrew speakers were from Tel Aviv University and Beit Berl<br />
Teachers' Training College. Their age ranged from 19 to 55.<br />
The instrument used to collect the data was a Reaction Elicitation Questionnaire,<br />
which is in fact a modified version of Blum-Kulka's (1982) Discourse Completion<br />
Questionnaire. The Modified Elicitation Questionnaire was designed to<br />
elicit expressions of censure from the three groups of respondents in their country of<br />
domicile. The purpose of modification was to obviate, as much as possible, constraints<br />
made on responses with respect to tone or content, and furthermore, to<br />
allow the option of zero response leading to the opting-out strategy.<br />
The questionnaire consisted of two main parts (beyond the section dealing with<br />
personal information): (1) 25 situations to which respondents were asked to react in<br />
writing; and (2) an evaluation sheet on which respondents were to indicate the<br />
degree of perceived severity of the SUA provided in the questionnaire. The 25<br />
situations provided in the questionnaire consisted of 20 descriptions related to acts<br />
expected to elicit expressions of censure and 5 distractor situations placed at randem<br />
among the 20 other situations.<br />
The acts described in the 20 censure situations refer to manifestations of socially<br />
unacceptable behavior in a Western culture. These situations were divided into 5<br />
main content categories:<br />
1. littering<br />
2. noise making<br />
3. unpunctuality
The Speech Act of Complaining 113<br />
Table 5.2. The Option to Express or Refrain from<br />
from Expressing Censure by Cultural Group<br />
British<br />
533 tokens<br />
(%)<br />
American<br />
431 tokens<br />
(%)<br />
Israeli<br />
491 tokens<br />
(%)<br />
Censure<br />
Avoidance<br />
65<br />
35<br />
100<br />
61<br />
39<br />
100<br />
64<br />
36<br />
100<br />
4. queue jumping<br />
5. petty stealing<br />
The first important finding of this study is that respondents from all three cultures<br />
opted out less than they chose to express censure, for the given situations. Table 5.2<br />
presents these data. About two-thirds of the respondents in each group chose to<br />
realize the speech act of censure, while about one-third opted out by indicating on<br />
the questionnaire that they would prefer "to say nothing" in the given situation.<br />
Table 5.3 presents strategy preferences among those who decided to carry out<br />
the act of censure. From considering the data in both tables 5.2 and 5.3, it becomes<br />
apparent that the differences among the three groups were negligible. This was<br />
contrary to our expectations: from other studies we had come to perceive Israeli<br />
society as much more direct and positive politeness oriented and the British culture<br />
as much more indirect and negative politeness oriented, with the American culture<br />
falling somewhere in between. It seems, therefore, that the chosen situations had a<br />
much stronger impact on strategy choice, and this impact was similar in all three<br />
cultures. It is conceivable that the seemingly "universal" situations selected for the<br />
instrument contributed to similar strategy choice in the three cultural groups.<br />
Figure 5.1 shows the variation of responses by the three groups of respondents<br />
according to the five situations used in this study. The distribution of responses<br />
further reinforces that the three culture groups differ in their speech act behavior of<br />
complaining only slightly with respect to three situations (noise, unpunctuality,<br />
line), while showing almost identical realization patterns for the other two. The<br />
overall picture emphasizes the similarity of performance among the three culture<br />
groups.<br />
Table 5.3. Strategies to Express Censure by Cultural Group<br />
British<br />
347 tokens<br />
(%)<br />
American<br />
264 tokens<br />
(%)<br />
Israeli<br />
313 tokens<br />
(%)<br />
Indirect<br />
Mitigated<br />
Unmitigated<br />
20<br />
15<br />
65<br />
100<br />
14<br />
17<br />
69<br />
100<br />
13<br />
16<br />
71<br />
100
114 Speech Act Realization<br />
Fig. 5.1. Variation of responses by three groups of respondents.<br />
The Intel-language of Complaints<br />
In this section we are concerned with the pragmalinguistic and sociopragmatic<br />
features of the interlanguage of complaining as exhibited by intermediate and advanced<br />
learners of Hebrew as a second language. We have tried to focus specifically<br />
on deviations from native norms that might lead to sociopragmatic failure (Thomas,<br />
1983; Blum-Kulka & Olshtain, 1986).<br />
In previous studies (Kasper, 1981; Olshtain, 1983; Olshtain & Cohen, 1983;<br />
Olshtain & Blum-Kulka, 1985; Blum-Kulka & Olshtain, 1986) it has been established<br />
that nonnative speech act behavior can deviate from native behavior: in<br />
strategy selection, in utterance length, in the consideration of social and pragmatic<br />
features, in carrying out or opting out from performing a speech act, and in varying<br />
the degree of external and internal modification. It is our objective to describe some<br />
of the potential deviations that might arise particularly during the acquisition of the<br />
speech act of complaining.<br />
The research presented in this section will focus on one large-scale study comparing<br />
native and nonnative realizations of complaints with respect to five measures:<br />
utterance length as expressed in number of words, utterance length as expressed<br />
in number of moves, position on the severity scale, use of softeners, and use<br />
of intensifiers. Additionally, two shorter studies dealing with cultural and ethnic<br />
preferences for strategy selection, within the nonnative group, will be discussed as<br />
an expansion of our understanding of the various factors that might affect interlanguage<br />
behavior.<br />
Native and Nonnative Realizations of Complaints in Hebrew<br />
The main goal of this investigation was to identify factors that distinguish native<br />
from nonnative realizations of complaints. Two main hypotheses were formulated to<br />
underlie the various studies reported here.
The Speech Act of Complaining 115<br />
1. According to previous work on interlanguage pragmatics (Blum-Kulka &<br />
Olshtain, 1986), we would expect nonnative speakers to produce longer utterances<br />
(as measured by number of words or by number of moves) than native speakers,<br />
given the same descriptions of situations on the discourse completion test (see<br />
Appendix).<br />
2. Although the cross-cultural comparison above demonstrated striking similarities<br />
in the performance of the three "central" strategies on the severity scale<br />
(Tables 5.2 and 5.3), learners will tend to move more toward the less severe<br />
strategies than do native speakers; new immigrants or newcomers to any culture<br />
might try to avoid straightforward face-threatening interactions at all costs. Strategy<br />
selection for complaining can be ranked on a 5-point scale, where 5 is the strongest,<br />
most severe complaint, and 1 is slightly below the level of reproach. Native speakers<br />
can be expected to have a higher average along this scale (tendency toward<br />
stronger complaint) than nonnative speakers. Furthermore, in their attempt to sound<br />
less aggressive and more polite, learners might tend to use more softeners, which<br />
lower the face-threatening nature of complaints. Accordingly, they will use fewer<br />
intensifiers, since such intensifiers aggravate complaining force and create even<br />
stronger conflict.<br />
Two groups of subjects—35 native speakers of Hebrew and 35 learners<br />
(nonnatives)—were compared on five separate measures in order to test the above<br />
two hypotheses. Length of utterance was tested on two measures—number of<br />
words and number of moves within the realization; severity of complaint was<br />
measured by the average position along the severity scale (from 1 to 5), and<br />
softeners and intensifiers were compared across all situations. A multivariate analysis<br />
of variance (MANOVA) was conducted and the differences between the two<br />
groups were found significant (F[5,64j = 25,7; p < .001). The results of a univariate<br />
analysis for the differences between the two groups on the five measures are<br />
presented on Table 5.4. The results indicate strong group differences for the first<br />
three measures: total length of utterance, number of moves, and the severity scale<br />
(strategy selection). The learners, as expected in hypothesis 1, use considerably<br />
more words in order to express the speech act of complaining. The number of moves<br />
may be related to the overall length of utterance since more moves would have to be<br />
realized in more words, but it is obvious that independent of the number of moves,<br />
the overall length of utterance indicates a strong difference between the groups. This<br />
result further reinforces the findings presented in Blum-Kulka and Olshtain (1986)<br />
and strengthens our belief that advanced learners tend to make use of more words in<br />
order to negotiate their intentions expressed via speech act realizations.<br />
The results presented in Table 5.4 also confirm hypothesis 2, at least partially.<br />
According to hypothesis 2 we expected native speakers to prefer more severe<br />
strategies in comparison to the learners. The average on the severity scale is 2.79 for<br />
the Hebrew speakers and only 2.61 for learners, which fully confirms this part of<br />
hypothesis 2. Furthermore, Table 5.4 also indicates that, as expected, learners tend<br />
to use more softeners (.68)' in comparison to natives (.54). However, the last part of<br />
hypothesis 2 was not confirmed, as it seems that learners use more intensifiers (.77)<br />
than natives (.68). It may well be that in their attempt to negotiate their intentions,<br />
the use of more words overcomes any other considerations, intensifiers being used<br />
mainly for message clarification rather than for intensification.
116 Speech Act Realization<br />
Table 5.4. Univariate F-Tests for Differences between Native Speakers and Learners<br />
on Five Measures (averaged over 20 situations)<br />
Native Ss<br />
Learners<br />
Measure<br />
X<br />
SD<br />
X<br />
SD<br />
F<br />
Total no. of words<br />
7.05<br />
1.03<br />
10.14<br />
1.84<br />
74.86***<br />
No. of moves<br />
1.67<br />
.18<br />
2.09<br />
.30<br />
50.03***<br />
Severity scale<br />
Total softeners<br />
2.79<br />
.54<br />
.21<br />
.16<br />
2.61<br />
.68<br />
.18<br />
.24<br />
15.24***<br />
8.05**<br />
Total intensifiers<br />
.68<br />
.16<br />
.77<br />
.18<br />
4.67*<br />
*p < .05.<br />
**p < .01.<br />
***p < .001.<br />
Some situations were particularly sensitive to the differences between the two<br />
groups. Consider situation 11: "Your car needs some urgent repair work and you<br />
take it to the garage close to closing time. There are two cars in line before yours.<br />
You ask nicely to have your car fixed in spite of the late hour. The owner of the<br />
garage apologizes for not being able to fix it and rejects any possibility of even<br />
looking at it at this late hour." The native speakers made relatively short complaints<br />
in this situation, such as: "I won't come here again"; "Okay, I'll do without it";<br />
"You just lost a client." The learners, on the other hand, had much longer utterances:<br />
"I was hoping for better treatment since I have been your client for quite<br />
some time. What can I do now"; "Sir, I have to have my car. You know Mr. X, he<br />
recommended that I come here and I don't want to complain to him about you—it<br />
might give your garage a bad name. So, why don't you fix it for me" These<br />
examples show how learners feel the need to "say more" in the hope that they might<br />
get the car fixed, after all.<br />
Situation 9 was particularly sensitive to the differences in strategy selection on<br />
the severity scale: "It is not the first time that loud rock music is heard from your<br />
neighbor's apartment quite late at night. You pick up the phone and you say . . ." In<br />
this situation the native speakers tended to opt for the most severe strategies, many<br />
using threats and ultimatums such as: "If you don't turn off your stereo soon I'll<br />
break your door down," "If you don't stop this loud music immediately I'll call the<br />
police." The learners were much more hesitant in their selection of strategies,<br />
opting for complaints rather than threats, such as: "I'm trying very hard to sleep but<br />
it is impossible with all this noise," "Some people would like a little peace and quiet<br />
but you are completely inconsiderate."<br />
Following the analysis above, we raised a number of additional questions.<br />
Given the high significance of length of utterance and selection of strategy on the<br />
severity scale as measures distinguishing the two groups, might particular aspects of<br />
the situation interact with these differences We were concerned, primarily, with<br />
features of social status, social distance, and the existence of some kind of a contract<br />
or social obligation on H's part to have avoided the reason for complaining. These<br />
three independent variables: social status, social distance, and social obligation
The Speech Act of Complaining 111<br />
were embedded in the description of the situations on the discourse completion<br />
questionnaire. Table 5.5 presents the means and standard deviations of utterance<br />
length and severity scale produced by the two groups, by social distance, social<br />
status and social obligation, while Table 5.6 gives the F values for the multivariate<br />
analysis of variance for these data.<br />
From the F values presented in Table 5.6 we see that social distance is a<br />
significant factor for both groups in terms of both length of utterance and severity of<br />
complaint (p < .001). It does not, however, interact with group and is nonsignificant<br />
as far as group distinctions are concerned. It is interesting to look at the means<br />
presented in Table 5.5 for the two groups: although length of utterance is always<br />
higher for learners, as was to be expected based on the main effect established<br />
earlier, it is surprisingly similar in its distribution according to different types of<br />
social distance. Thus, both groups use more words with acquaintances, as if there<br />
seems to be a need to negotiate more with interlocutors of this type. This finding<br />
goes along with Wolfson's (1989) description of "the bulge" phenomenon, which<br />
Table 5.5. Means and Standard Deviations (in parentheses) for Two Dependent Variables<br />
According to Groups by Social Distance, Social Status, and Social Obligation<br />
Number of words<br />
Severity scale<br />
Social factor<br />
Natives<br />
Learners<br />
Natives<br />
Learners<br />
Social Distance<br />
Strangers<br />
6.74<br />
(1.61)<br />
9.57<br />
(2.90)<br />
2.64<br />
(.34)<br />
2.43<br />
(.31)<br />
Acquaintances<br />
8.48<br />
(2.78)<br />
11.79<br />
(3.46)<br />
3.02<br />
(.34)<br />
2.72<br />
(.26)<br />
Friends<br />
5.54<br />
(2.32)<br />
8.59<br />
(4.07)<br />
2.67<br />
(.62)<br />
2.44<br />
(.51)<br />
Relatives<br />
6.43<br />
(1.68)<br />
9.58<br />
(2.65)<br />
2.76<br />
(.40)<br />
2.73<br />
(.38)<br />
Social Status*<br />
S(-P) H(+P)<br />
S( + P) H(-P)<br />
Equals<br />
7.77<br />
(2.16)<br />
7.69<br />
(1.60)<br />
4.88<br />
(.78)<br />
10.71<br />
(2.91)<br />
11.30<br />
(3.58)<br />
5.79<br />
(1.04)<br />
2.46<br />
(.35)<br />
2.98<br />
(.31)<br />
4.13<br />
(.79)<br />
2.33<br />
(.36)<br />
2.78<br />
(.34)<br />
4.41<br />
(.77)<br />
Social Obligation<br />
Nonexistent<br />
8.06<br />
(2.47)<br />
li.OO<br />
(3.48)<br />
2.66<br />
(.37)<br />
2.37<br />
(.37)<br />
Explicit<br />
5.26<br />
(.89)<br />
6.27<br />
(1.22)<br />
4.46<br />
(.79)<br />
4.82<br />
(.78)<br />
Implicit<br />
6.24<br />
(1.52)<br />
9.06<br />
(2.63)<br />
2.61<br />
(.32)<br />
2.61<br />
(.34)<br />
*S — speaker. H = hearer, P = (soeial) power.
118 Speech Act Realization<br />
Table 5.6. F(df) Values for the Multivariate Analysis of Variance<br />
Source of variance<br />
Social distance<br />
Distance & group<br />
Social status<br />
Status & group<br />
Social obligation<br />
Obligation & group<br />
*p < .01.<br />
**p < .001.<br />
Length of utterance<br />
14.82**<br />
(3,66)<br />
.14<br />
(3,66)<br />
177.50**<br />
(2,67)<br />
15.52**<br />
(2,67)<br />
84.85**<br />
(2,67)<br />
9.92**<br />
(2,67)<br />
Severity scale<br />
16.16**<br />
(3,66)<br />
1.89<br />
(3,66)<br />
184.83**<br />
(2,67)<br />
2.99<br />
(2,67)<br />
224.33**<br />
(2,67)<br />
5.43*<br />
(2,67)<br />
indicates that when speakers are less certain about roles and relationships with their<br />
interlocutors they negotiate more and accordingly use more words. It seems that<br />
both groups in our study feel the need for such negotiation in the case of mere<br />
acquaintances. On the other hand, with respect to strangers and relatives, they<br />
behave very similarly since here the roles are more fixed—strangers, we may never<br />
see again; relatives are here to stay. In either case there is less need for negotiation,<br />
as was also found by Wolfson (1989). The difference lies in the way Israelis, both<br />
native speakers of Hebrew and learners, react to friends. It seems that friendship is<br />
more binding in Israeli society and therefore, when a friend performs a breach of<br />
confidence, the disappointment is great. A sense of solidarity is well founded<br />
among friends (verbal reports of subjects) and there is no need to negotiate via<br />
lengthy exchanges. An example is situation 13: "A friend who takes the same course<br />
as you do was not willing to share some important material which he was able to get<br />
and which can help you prepare for the exam in that course." The complaints<br />
produced by the respondents in this case were rather brief, expressing deep personal<br />
disappointment.<br />
With respect to social status, Table 5.6 shows that this factor is significant for<br />
both length of utterance and strategy selection on the severity scale for both groups<br />
(p < .001), but more important, there is interaction between status and group with<br />
respect to length of utterance (p < .001). From Table 5.5 it is clear that again<br />
learners always use longer utterances than native speakers, but this is particularly<br />
strong in the case of S(+P) - H(-P), where the speaker is of higher status than the<br />
interlocutor. It is under these conditions that learners use the longest utterance,<br />
negotiating most fervently. Furthermore, the standard deviation (SD) in this case is<br />
extremely high (3.58), indicating great variability among the subjects in the group<br />
of learners.<br />
The strongest interaction effect in our analysis is found between social obligation<br />
and group, as can be seen from Table 5.6. Here the interaction is both for length
The Speech Act of Complaining 119<br />
of utterance (p < .001) and for strategy selection (p < .01). In terms of length of<br />
utterance, both groups use the most intensive negotiation when there is no explicit<br />
social obligation; here again, expectations can be less well predicted and the obligation<br />
is not fixed. In cases like these, negotiation might bring about repair of the<br />
violation, hence if pays for the speaker to negotiate. On the other hand, when there<br />
is explicit obligation, one can simply appeal to it without too much negotiation,<br />
such as in situation 3: "You are driving your new car on a main street. Suddenly a<br />
car crosses in front of you not having stopped at a stop sign." In this case, the<br />
obligation is part of the law, and therefore the only thing subjects need to say is that<br />
one must abide by traffic laws. On the other hand, for situation 10, "You have<br />
returned from your vacation one day later than expected and are unable to hand in a<br />
seminar paper on time. Your professor is not prepared to extend the time," it seems<br />
worthwhile to negotiate and hope that the professor will change his mind. The<br />
greatest difference between natives and learners is found in cases where the obligation<br />
is implicit. Here the learners feel a greater need to negotiate. They also exhibit<br />
greater variability than natives do: on situations with no social obligation, there is a<br />
difference of 1.19 standard deviation between natives and learners on length of<br />
utterance, while the difference is 1.85 when the obligation is implicit (indicating<br />
greater variability for the implicit obligation).<br />
The interaction between social obligation and group (p < .01) on the severity<br />
scale is realized in a rather unexpected manner: for situations with explicit social<br />
obligation, the learners choose more severe strategies than the natives (4.82 as<br />
compared to 4.46; see Table 5.5); for the situations with implicit obligation, they<br />
show exactly the same means; and when there is no obligation whatsoever, the<br />
difference is very small, with the natives scoring only slightly higher than the<br />
learners. Both groups use the more severe strategies when there is social obligation,<br />
but this is the only case when the learners do so even more than the natives,<br />
probably feeling much more confident that in such obvious situations, they have the<br />
right to complain explicitly and boldly.<br />
Cultural and Ethnic Considerations<br />
The last question to be discussed in this chapter relates to culturally specific parameters<br />
derived from context knowledge (Faerch & Kasper, 1984), which might be<br />
perceived differently by learners than by native speakers, thus leading to additional<br />
deviations in speech act performance. This question was investigated with two<br />
groups of immigrants in Israel: Russians and Moroccans (Hauser & Swindler,<br />
1988). There were 20 respondents in each group; all were immigrants who had been<br />
in Israel for at least two years when the questionnaire was administered. The two<br />
groups reacted quite similarly in most situations and indicated a very similar spread<br />
of strategy selection. The differences occurred in two specific situations.<br />
Situation 3 was as follows: "You are nearing an empty spot in a parking place<br />
when another car drives in and takes the place, disregarding you completely." In this<br />
situation, Russians and Moroccans reacted very differently. Russians tended to use<br />
warnings, threats, and even curses, while Moroccans opted out or chose one of the<br />
two softer realizations of complaints, below the level of reproach or disapproval.
120 Speech Act Realization<br />
When questioned in verbal reports, the Russians claimed that people should be<br />
considerate and play a "fair game" in a parking lot, while the Moroccans said that<br />
the offense was not serious; in fact, they might also do something like that.<br />
Situation 8 read: "You are in great need of a money loan. You approach a friend<br />
whom you had helped in the past, but s/he rejects your request." In this situation,<br />
the two groups of speakers reacted quite differently. Russians tended to accept the<br />
rejection and not to complain. If they did voice a complaint, it was "below the level<br />
of reproach." Moroccans, on the other hand, made use of the full scale of strategies.<br />
In verbal reports, the Russians claimed that when it comes to money business, it is<br />
actually better not to borrow from friends, while the Moroccans said that friends<br />
should be there to help you in time of need.<br />
In another study, which compared new immigrants from Romania with longestablished<br />
immigrants from the same country (Hoch-Pasko, 1988), it was found<br />
that some interesting differences occurred in the subjects' consideration of the<br />
contextual features. Thus, newcomers tended to opt out from expressing complaints<br />
in situations that related to a breach of social norms affecting the public in general,<br />
such as throwing litter in the public domain and being late for a movie, resulting in<br />
disturbing the audience. Yet, when the violation affected the individual (e.g., a<br />
neighbor in the building playing very loud music late at night), the newcomers<br />
tended to use more severe complaints than the long-established immigrants.<br />
Conclusion<br />
A number of interlanguage features of speech act behavior have been established in<br />
the series of studies reported in this chapter, two of which seem to be general and<br />
might therefore apply to all speech act behavior of learners and three which might be<br />
specific to the speech act of complaining.<br />
The two general interlanguage features of speech act performance are length of<br />
utterance and variability. Learners at the intermediate to advanced level of second<br />
language acquisition tend to be verbose and use more words than native speakers,<br />
more than they themselves would use in their own language, in order to negotiate<br />
the intentions of their speech acts in the new language. When compared to native<br />
speakers of the target language, this difference is strong.<br />
More pronounced variability in strategy selection, use of intensifiers, softeners,<br />
number of moves, etc. can be seen from the consistently larger standard derivations<br />
exhibited by learners. Learners are less certain about such decisions and therefore<br />
tend to vary more in their choices than native speakers.<br />
The speech act of complaining, being an inherently face-threatening act, exhibits<br />
interesting deviations in learners' interlanguage from native speakers' norms,<br />
which are specific to this speech act. Such deviations seem to focus on the fact that<br />
newcomers to the target community attempt to sound less offensive and less facethreatening.<br />
As a result, their overall choice of realization patterns is consistently<br />
closer to the less severe end of the scale when compared to native speakers. As part<br />
of their variability in choice, they also seem to be more sensitive to differences in<br />
social factors related to their interlocutors. In our study, the unexpected finding
The Speech Act of Complaining 121<br />
related to the notion of social obligation: when the situation was one in which social<br />
obligation was explicit either by law or convention, the learners felt more secure in<br />
their choices and were less concerned with being polite and cautious.<br />
The studies of the interlanguage of speech act behavior presented in this chapter<br />
indicate that contextual parameters such as social factors and situational factors<br />
might affect realization patterns of speech acts produced by learners. Some of these<br />
interlanguage features have the potential to represent any speech act behavior while<br />
others are speech act specific.<br />
Appendix<br />
Sample items from discourse completion questionnaire (translated from Hebrew):<br />
Situation 9<br />
It is not the first time that loud rock music is heard from your neighbor's apartment quite late<br />
at night. You pick up the phone and say:<br />
Situation 13<br />
A friend who takes the same course as you at the university refuses to share some important<br />
material for the next test, which s/he managed to get hold of. In the past, you helped him/her<br />
many times. You see him/her on campus and say:<br />
References<br />
Blum-Kulka, S. (1982). Learning how to say what you mean in a second language: A study<br />
of the speech act performance of learners of Hebrew as a second language. Applied<br />
Linguistics, 3, 29-59.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Studies in Second Language Acquisition, 8, 47-61.<br />
Brown, P., & Levinson, S. (1987). Politeness: Some universah in language usage. Cambridge:<br />
Cambridge University Press.<br />
Faerch, C., & Kasper, G. (1984). Pragmatic knowledge: Rules and procedures. Applied<br />
Linguistics, 5, 214-25.<br />
Hauser, Z., & Swindler, I. (1988). A comparison of realization patterns of complaints among<br />
Israelis from Morroco and from the Soviet Union. Unpublished seminar paper, School<br />
of Education, Tel Aviv University.<br />
Hoch-Pasko, R. (1988). Complaints as expressed by Israelis from Roumanian background.<br />
Unpublished seminar paper, School of Education, Tel Aviv University.
122 Speech Act Realization<br />
Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Eine Untersuchung des<br />
Englischen fortgeschrittener deutscher Lerner. Tubingen, Germany: Narr.<br />
Katz, L. (1987). Face risk in the expression of censure: A cross-cultural study. Unpublished<br />
master's thesis, School of Education, Tel Aviv University.<br />
Leech, G. N. (1983). Principles of pragmatics. London: Longman.<br />
Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In<br />
S. Gass & L. Selinkcr (Eds.), Language transfer and language learning (232-49).<br />
Rowley, MA: Newbury House.<br />
Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to<br />
native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language<br />
acquisition (303-25). Rowley, MA: Newbury House.<br />
Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd<br />
(Eds.), Sociolinguistics and language acquisition (18-36). Rowley, MA: Newbury<br />
House.<br />
Olshtain, E., & Weinbach, L. (1987). Complaints—A study of speech act behavior among<br />
native and non-native speakers of Hebrew. In M. B. Papi & J. Verschueren (Eds.),<br />
The Pragmatic perspective: Selected papers from the 1985 International Pragmatics<br />
Conference (195-208). Amsterdam: Benjamins.<br />
Thomas, J. A. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.<br />
Weinbach, L. (1988). The realization pattern of complaints in Hebrew: Native speakers and<br />
learners. Unpublished master's thesis, School of Education, Tel Aviv University.<br />
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. Cambridge, MA: Newbury<br />
House.
6<br />
Interlanguage Requestive Hints<br />
ELDA WEIZMAN<br />
This chapter explores the use of Hints as a request strategy by learners. It addresses<br />
the question of regularities to be observed in the use of requestive Hints by language<br />
learners with various mother tongues at different proficiency levels, and the similarities<br />
or differences between them and native speakers. This analysis contributes<br />
both to our knowledge of learners' pragmalinguistic behavior, as well as to our<br />
understanding of indirectness. The study was conducted within the framework of<br />
the Cross-Cultural Speech Act Realization Project (CCSARP) (cf. Blum-Kulka,<br />
House, & Kasper, 1989), and focuses on a comparison of Hebrew learners (henceforth<br />
HL) with Hebrew native speakers (henceforth HN), examined in the context of<br />
a larger body of CCSARP data. Drawing on findings concerning the situational<br />
variation and the use of Hint substrategies by native speakers (Blum-Kulka, 1989;<br />
Weizman, 1989), I propose to examine whether the opacity inherent in Hints is<br />
exploited by learners as a strategy of communication, whether the impact of situational<br />
variations marks Hint selection by learners as it does with native speakers,<br />
and whether the use of Hint substrategies is comparable in the two groups.<br />
Defining Requestive Hints<br />
The category of requestive Hints consists of the most nonconventional and indirect<br />
of all request strategies. Typical instances include:<br />
/ have so much work to do (when used as a request to be left alone).<br />
Don't you have anything to do (when used as a request to be left alone).<br />
It's cold in here (when used as a request to close the window).<br />
/' ve just missed my bus (when used as a request for a ride).<br />
Are you going home now (when used as a request for a ride).<br />
Do you have your notebook with you (when used as a request that the hearer<br />
lend the speaker her notebook).<br />
The kitchen is in a bit of a mess (when used as a request that the hearer clean up<br />
the kitchen).<br />
123
124 Speech Act Realization<br />
In Hints, the interpretation of the speaker's intentions is highly contextembedded,<br />
and is not secured either by the sentence meaning of the utterance (i.e.,<br />
by its context-independent, literal meaning), as in direct requests (e.g., "Leave me<br />
alone, please"); or by some grammatical or semantic device, conventionally used to<br />
convey the requestive force, as in indirect, conventional requests (e.g., "Could you/<br />
Would you leave me alone, please")<br />
Initially, the CCSARP coding scheme made a distinction between two types of<br />
Hints—mild Hints and strong Hints—in terms of the amount of contextual knowledge<br />
required for their interpretation. For the purpose of the present analysis, these<br />
two strategies are collapsed, and each Hint is re-analyzed. The proposed analysis<br />
assumes, following Weizman (1985; 1989), that although all Hints are inherently<br />
opaque, they are not all opaque in the same way, or to the same extent. Hence, Hints<br />
vary on two dimensions: degree and type of opacity. In terms of type, it is either<br />
Hints' illocutionary force, or their propositional content, or both, that is obscure,<br />
and that should, therefore, be "explored" and "deciphered" by the interpreter. On<br />
the illocutionary dimension, a Hint is considered opaque when its utterance meaning<br />
does not provide sufficient indication of its intended illocutionary force; consequently,<br />
it carries the potential to perform a number of illocutionary acts. On the<br />
propositional dimension, a Hint is considered opaque when its sentence meaning<br />
does not provide sufficient evidence of the content of the act. Type and degree of<br />
opacity interact, however, since on each of these dimensions Hints may be distinguished<br />
in terms of degree of opacity, that is, the amount of "missing clues"<br />
required for full interpretation. (For further details, refer to "Distribution of Hint<br />
Substrategies in Learners' Data," below.) Obviously, since in real-life situations<br />
disambiguation is highly context-dependent (see Dascal & Weizman, 1987; Weizman<br />
& Dascal, 1991), the distinction by types is not always clear-cut. Generally<br />
speaking, however, it seems safe enough to suggest that while illocutionary opacity<br />
entails propositional opacity, the opposite does not hold, since opacity on the<br />
propositional dimension does not necessarily imply opacity on the illocutionary<br />
scale. For example, the utterance "I have so much work to do" could be interpreted<br />
either as a request (to be left alone) or as an apology (for having forgotten to do the<br />
shopping). Hence, it is considered opaque on both the illocutionary and the propositional<br />
dimensions. The utterance "If only someone helped me," on the other hand,<br />
would most plausibly be interpreted as a request, but the content of the request is<br />
completely obscure, since the speaker does not specify the kind of help needed.<br />
Hence, this Hint is considered relatively transparent on the illocutionary dimension,<br />
and extremely opaque on the propositional one. Note that the two dimensions<br />
interact: since certain propositional components are entailed by the illocutionary<br />
force (Vanderveken, 1984, 190), a certain degree of propositional transparency may<br />
help to disambiguate the illocutionary force. For instance, the illocutionary force of<br />
promises imposes on the propositional content the requirement that a reference be<br />
made to a (semantically) future action (Searle, 1975). Since, however, reference to a<br />
future action is also apt to be made in requests, propositional content in Hints can<br />
only be considered as contributing to the elucidation of illocutionary force. Hence,<br />
it is the balance between the opacity types that changes from one Hint to another,<br />
and creates the enormous variety of requestive Hints. 1<br />
The very essence of Hints cannot be fully grasped unless one attempts to answer
Interlanguage Requestive Hints 125<br />
the question: Why use Hints at all Assuming the speaker to be a rational agent<br />
(Grice, 1975), he or she is supposed to achieve any given end by choosing "that<br />
action which, most effectively, and at least cost, attains that end, ceteris paribus"<br />
(Kasher, 1982, 32). What end could possibly be attained by the use of such a high<br />
degree of opacity Surely, if the end to be achieved is the requestive one—that is, if<br />
the speaker intends to get a requested act carried out by the hearer as a result of the<br />
hearer's recognition of this intent (Searle, 1975)—then the most effective way of<br />
achieving this purpose is not by using an "off-record" strategy (Brown & Levinson,<br />
1978, 210), but by using direct ones. One would tend, then, to associate indirectness<br />
with politeness and tact, following Leech, for instance: "in general, the more<br />
tactful a directive is, the more indirect and circumlocutionary it is" (Leech, 1980,<br />
109); or: "Indirect illocutions tend to be more polite (a) because they increase the<br />
degree of optionality, and (b) because the more indirect an illocution is, the more<br />
diminished and tentative its force tends to be" (Leech, 1983, 108). Along these lines,<br />
it is tempting to suggest that the end to be achieved by the use of Hints is to get a<br />
requested act carried out without threatening the hearer's face. Experiments have<br />
shown, however, that although perceptions of politeness vary across cultures, Hints<br />
are not conceived as the most polite strategy. Within the framework of CCSARP,<br />
House (1986) and Blum-Kulka (1987) have shown that speakers of Hebrew, American<br />
English, British English, and German conceive of Hints as less polite than<br />
conventional indirectness. The same point was made earlier by Walters (1979, 286,<br />
287) about speakers of American English and Puerto Rican Spanish, within the<br />
framework of studies on learners' performance of requests in role-playing situations<br />
(Eraser, Rintell,& Walters, 1980; Walters, 1981; Walters, 1979;Rintell, 1981). 2 Evidence<br />
then seems to indicate that the most effective way to make requests while saving<br />
the hearer's face is not the use of Hints, but the use of conventional indirectness.<br />
If Hints are not the most efficient way to achieve the requestive purpose, or to<br />
save face, what then is the end that can be attained most effectively by using them<br />
In Weizman (1989) I have suggested the following: the large repertoire of plausible<br />
interpretations of a given Hint seems to imply that Hints are the most efficient way<br />
for the requester to make a request while at the same time securing the possibility of<br />
legitimately denying some of its illocutionary and prepositional components. I have<br />
therefore suggested (Weizman, 1989) that Hints be thought of as the only request<br />
strategy that bears a high deniability potential for both parties: the requester may<br />
plausibly deny having made a request (e.g., "I never meant to ask anything of you")<br />
or deny its prepositional content (e.g., "I never asked you to fetch me my coat"; "I<br />
just wanted you to close the window"); the requestee may legitimately ignore the<br />
request or pretend to have misunderstood its content. Since this deniability potential<br />
might be exploited by learners, the recognition of its importance underlies the<br />
questions raised in the following section.<br />
Research Questions<br />
This study raises three main questions, underlined by previously observed data and<br />
some working assumptions. First, in the CCSARP data, native speakers' use of<br />
Hints is usually remarkably low relative to either direct strategies or conventional
126 Speech Act Realization<br />
indirect ones. The analysis of request strategies in five languages combined (Australian<br />
English, Canadian French, German, Argentinian Spanish, and Hebrew) reveals<br />
that native speakers' selection of Hints varies from 1.14% to 7.76%, depending on<br />
situational constraints (Blum-Kulka & House, 1989, 130). Low frequency is also<br />
manifested in Hint selection by native speakers of British English, Danish, and<br />
German in most of the situations (Kasper, 1989; House & Kasper, 1987). A similar<br />
trend is depicted in studies of other data types, e.g., studies of request realization by<br />
native speakers of English and German in role-plays (Kasper, 1981), and by native<br />
speakers of Hebrew in authentic face-to-face interaction, phone calls, and letters to<br />
the editor (Blum-Kulka, Danet, & Gerson, 1985). In terms of the deniabilitypotential<br />
hypothesis, this low frequency seems to indicate either that native speakers<br />
only rarely feel the need to secure a way to opt out of their request; or else, that only<br />
seldom do they consider it worth doing at the expense of efficiency.<br />
As for learners, the deniability potential might appeal to them for other reasons.<br />
More specifically, it might be exploited as a strategy of communication. By virtue of<br />
their nonconventional form, Hints might be employed to avoid the need to cope with<br />
the target-language conventional realizations; and, by virtue of their opacity, they<br />
might be exploited to avoid specific reference to the interlocuter's preceding turns.<br />
Put differently, their high deniability potential might be exploited by way of "risk<br />
avoidance," whereby a learner can "tailor his message to the resources he has<br />
available" (Corder, 1983, 17) in order to cope with "what to an individual presents<br />
itself as a problem in reaching a particular communicative goal" (Faerch & Kasper,<br />
1983, 212).<br />
If Hints are indeed employed as a communication strategy, a higher frequency of<br />
use is likely to be observed in learners, mainly those least exposed to the target<br />
culture and therefore most liable to feel insecure. This raises the questions whether<br />
learners use Hints more frequently than native speakers, and whether the incidence<br />
of Hints decreases with the increase in length of stay in the target culture. Answers<br />
will be suggested in "Proportion of Hints and Situational Variation in Learners"<br />
(below).<br />
Second, observed situational variation in native speakers' use of all request<br />
strategies, including Hints (Blum-Kulka, 1989) and Hint substrategies (Weizman,<br />
1989) suggests powerful situational constraints on their requestive behaviour.<br />
CCSARP questionnaires include eight situations, differentiated in terms of six social<br />
parameters: the degree of (1) dominance and (2) familiarity between the speakers;<br />
the requester's (3) right to make the request and the (4) degree of difficulty<br />
involving in making it; the (5) likelihood of compliance and (6) degree of obligation<br />
on the part of the requestee (Blum-Kulka & House, 1989). The second issue discussed<br />
in "Proportion of Hints and Situational Variation in Learners" (below) concerns<br />
the impact of situational variance on the selection of Hints by learners.<br />
Third, Hints in the CCSARP data were found to consist of several main substrategies,<br />
ranging from extreme opacity to relative transparency. The low incidence<br />
of other substrategies in native speakers' data was interpreted as indicating that<br />
although Hints are nonconventional in form; that is, nonconventional in terms of the<br />
lexical and grammatical items used for their realizations, their use is not entirely<br />
convention-free. The following point has been made (Weizman, 1989): one usually<br />
thinks of Hints, and justly so, as an open-ended category, because utterances of any
Interlanguage Requestive Hints 127<br />
form could be interpreted as carrying a given illocutionary force, depending on the<br />
situation. Nevertheless, it seems to be the case that even the use of Hints obeys a<br />
certain order, since in a given situation a given set of semantic contents is more<br />
likely to be preferred, although an open-ended repertoire of potential candidates is<br />
available. For example, when requesting that the hearer give the speaker a lift, a<br />
Hint would preferably be realized by questioning some constituent of the feasibility<br />
of that action; for example, by asking, "Do you have a car" or "Are you going home"<br />
whereby a positive answer is a necessary, but not sufficient, condition for the<br />
feasibility of the requested action. On the other hand, when requesting a roommate<br />
to clean up the kitchen, a Hint is more likely to be realized by stating some reason<br />
for the request; for example, "the kitchen is in a terrible mess." (For further details,<br />
see "Distribution of Hint Substrategies in Learners' Data," below.) 3 It is therefore<br />
important to emphasize that requestive Hints are viewed here as being accountable<br />
in terms of "conventions of means," specifying "a semantic device by which an<br />
indirect speech act can be performed," as opposed to "conventions of form," that is,<br />
"conventions about the wording of indirect speech acts" (Clark, 1979, 433).<br />
Furthermore, marked preference for the most opaque substrategy has been observed<br />
in native speakers and was interpreted as a confirmation of the hypothesis<br />
that opacity in Hints is intentionally and purposefully exploited (Weizman, 1989).<br />
The questions that arise at this point are whether learners prefer the same substrategies,<br />
and whether they, too, opt for the most opaque one. Answers to the last<br />
two questions will be suggested in "Distribution of Hint Substrategies" (below).<br />
Methodology<br />
CCSARP data, elicited by means of a discourse-completion test, offers the possibility<br />
of exploring the use of requests and apologies by native speakers of Australian,<br />
American and British English, Canadian French, Danish, German, and<br />
Hebrew, as well as by learners of English in Denmark, Germany, and the United<br />
States, of German in Denmark, and of Hebrew in Israel (Blum-Kulka, House, &<br />
Kasper, 1989, chapter 1).<br />
The present study compares learners with native speakers, all students. Table<br />
6.1 indicates the distribution.<br />
A two-step analysis was undertaken. First, 305 Hebrew learners, (henceforth<br />
Table 6.1. Comparison of Learners and Native Speakers<br />
Learners<br />
Native Speakers<br />
Target<br />
language<br />
Country<br />
Mother<br />
tongue<br />
No. of<br />
subjects<br />
Language<br />
No. of<br />
subjects<br />
Hebrew<br />
English<br />
English<br />
English<br />
English<br />
Israel<br />
Germany<br />
Denmark<br />
USA<br />
Israel<br />
Various<br />
German<br />
Danish<br />
Various<br />
Various<br />
305<br />
200<br />
200<br />
35<br />
47<br />
Hebrew<br />
Australian English<br />
American English<br />
German<br />
Canadian French<br />
Argentinian Spanish<br />
173<br />
226<br />
27<br />
200<br />
128<br />
40
128 Speech Act Realization<br />
referred to as HL), were compared with 173 Hebrew native speakers (henceforth<br />
HN). Then, all other learners (n = 482) were compared with the rest of the native<br />
speakers (« = 621). The reason for the distinction is that only Hebrew learner's data<br />
allow for quantitative analysis in terms of length of stay in the target culture, and for<br />
qualitative analysis of Hint substrategies. The rest of the data lends itself mainly to<br />
the study of frequency of use and situational variation.<br />
Data Analysis<br />
Proportion of Hints and Situational Variation in Learners<br />
Of the eight request situations featuring in the CCSARP questionnaire, the following<br />
five are analyzed here:<br />
SI (referred to as "Kitchen"): A student asks his roommate to clean up the<br />
kitchen the latter had left in a mess the night before.<br />
S5 (referred to as "Notes"): A student asks another to lend her some lecture<br />
notes.<br />
S7 (referred to as "Ride"): A student asks people living on the same street for<br />
a ride home.<br />
SI 1 (referred to as "Policeman"): A policeman asks a driver to move her car.<br />
S15 (referred to as "Lecturer"): A university lecturer asks a student to present<br />
his paper a week earlier than scheduled.<br />
Table 6.2 shows the impact of the situational variation on the selection of Hints<br />
by HL as compared to HN. The analysis reveals that in both populations the<br />
frequency of Hints is situationally dependent. In each of them, Hints rarely occur in<br />
the Notes situation (1.7% in learners, 1.2% in native speakers). The Lecturer<br />
situation is also marked by a relatively low incidence of Hints (4.2% and 2.7%,<br />
respectively), while in both groups, Hints most frequently occur in the Policeman<br />
situation (14.6% and 12.8%, respectively). In this respect, no significant differences<br />
are found between HL and HN. 4<br />
Table 6.2. Distribution of Hints by Situation in Hebrew Native Speakers (HN)<br />
and Hebrew Learners (HL)<br />
Learners (HL)<br />
Native speakers (HN)<br />
Situation<br />
N<br />
n<br />
%<br />
N<br />
n<br />
%<br />
p* value<br />
chi-square<br />
1 Kitchen<br />
5 Notes<br />
7 Ride<br />
1 1 Policeman<br />
15 Lecturer<br />
Total<br />
293<br />
297<br />
272<br />
240<br />
191<br />
1293<br />
33<br />
5<br />
28<br />
35<br />
8<br />
109<br />
11.3<br />
1.7<br />
10.3<br />
14.6<br />
4.2<br />
8.4<br />
163<br />
170<br />
165<br />
164<br />
150<br />
812<br />
15<br />
2<br />
23<br />
21<br />
4<br />
65<br />
9.2<br />
1.2<br />
13.9<br />
12.8<br />
2.7<br />
8.0<br />
0.492<br />
# *<br />
0.249<br />
0.611<br />
0.449<br />
0.472<br />
* *<br />
1.324<br />
0.258<br />
0.573<br />
Note: Of A' utterances, n are rcquestivc Hints.<br />
*For learners (HL) versus native speakers (HN), by chi-square analysis.<br />
;!: *The number of Hints is too small to allow for statistical analysis.
Inierlanguage Requestive Hints 129<br />
As concerns native speakers, these findings confirm earlier ones. Blum-Kulka<br />
and House (1989) show the significance of situationally specific features on the<br />
determination of request strategies in general, and Weizman (1989) reports on<br />
situational constraints affecting the selection of Hint substrategies. Blum-Kulka and<br />
House (1989, 128) further indicate that for native speakers of the 5 languages<br />
examined—Australian English, Canadian French, German, Hebrew, and Argentinian<br />
Spanish—the Notes and Lecturer situations attract not only the lowest proportion<br />
of Hints, but also the highest degree of cross-cultural agreement. From the<br />
comparison presented here, the same picture emerges for Hebrew learners: their<br />
sensitivity to situationally specific features is as crucial to Hint selection as it is with<br />
Hebrew native speakers, and they, too, share the agreement concerning the two<br />
situations in question. These findings also seem to agree with Blum-Kulka and<br />
Levenston's (1987) observation, that in four of the five situations examined, the<br />
distribution of strategy types by situation revealed no statistically significant differences<br />
between Hebrew learners and Hebrew native speakers (157).<br />
The question arising at this point is whether this high situational sensitivity in<br />
Hebrew learners reflects a general tendency to be observed in learners of other<br />
languages. To answer this question, different sets of CCSARP data were collapsed<br />
(as explained in "Research Questions," above), so as to allow comparison of Hint<br />
selection by native speakers of various mother tongues with Hint selection by<br />
learners of various target languages, regardless of their mother tongues. We know<br />
by now that situational features affect internal and external modifications of requests<br />
(mitigating devices such as interrogatives, conditionals, hedgers, politeness markers)<br />
(Faerch & Kasper, 1989; House, 1989). Table 6.3 indicates that they also affect<br />
the selection of Hints.<br />
In Table 6.3, two populations are compared: on the one hand, 621 native<br />
speakers (of Australian and American English, German, Canadian French, and<br />
Argentinian Spanish) and on the other, 482 learners (of English). The results show a<br />
marked situational variation in Hint selection by learners, and suggest similarities in<br />
that respect with native speakers. Again we see that in both groups, the Notes<br />
Table 6.3. Distribution of Hints by Situation in Native Speakers<br />
(of Australian and American English, German, Canadian French and Argentinian Spanish)<br />
versus Learners (of English, with Various Mother Tongues)<br />
Learners (HL)<br />
Native speakers (HN)<br />
Situation<br />
N<br />
n<br />
%<br />
N<br />
n<br />
%<br />
p* value<br />
chi-square<br />
1 Kitchen<br />
5 Notes<br />
7 Ride<br />
471<br />
479<br />
474<br />
24<br />
2<br />
25<br />
5.1<br />
0.4<br />
5.3<br />
614<br />
618<br />
614<br />
49<br />
2<br />
38<br />
8.0<br />
0.3<br />
6.2<br />
0.060<br />
*#<br />
0.522<br />
3.535<br />
*#<br />
0.410<br />
1 1 Policeman<br />
473<br />
31<br />
6.6<br />
610<br />
51<br />
8.4<br />
0.265<br />
1.243<br />
15 Lecturer<br />
Total<br />
465<br />
2362<br />
24<br />
106<br />
5.2<br />
4.5<br />
604<br />
3060<br />
31<br />
171<br />
5.1<br />
5.6<br />
0.983<br />
0.00045<br />
Note: Of TV utterances, n are rcquestive Hints.<br />
*For learners (HL) versus native speakers (HN), by chi-square analysis.<br />
* :t The number of Hints is too small to allow for statistical analysis.
130 Speech Act Realization<br />
situation attracts the smallest percentage of Hints (0.5% in learners, 0.19 in native<br />
speakers); the Lecture situation attracts a relatively small proportion of Hints<br />
(4.59% in learners, 5.57% in native speakers); and Hints are most likely to occur in<br />
the Policeman situation (6.9% and 8.87%, respectively). On the whole, observed<br />
situational variation by learners indicates that Hint selection by learners is as situationally<br />
determined as their selection by native speakers.<br />
Looking at the totals in Table 6.3 we can see that, taken together, the proportions<br />
of Hints employed by learners and native speakers (4.5% and 5.6%) are fairly<br />
similar. The similarity between HL and HN, as shown in Table 6.2, is more<br />
pronounced: 8% and 8.4%, respectively. 5 These findings, then, suggest no reason to<br />
believe that Hints are used as a strategy of communication.<br />
This interpretation is further supported if Hint selection is examined in terms of<br />
learners' length of stay in the target community. Table 6.4 compares the use of Hints<br />
by Hebrew learners with different periods of residence in Israel. Table 6.4 reveals<br />
striking similarities between the compared groups of immigrants. These similarities<br />
seem to be particularly interesting since language exposure was shown to affect the<br />
acquisition of target norms, at least as far as verbosity is concerned (Blum-Kulka &<br />
Olshtain, 1986, 158). The lack of effect of length of stay on Hint selection, I<br />
suggest, has to do with two factors: (1) Hint selection by HL matches NH norms<br />
from the very start, and there is therefore no reason to expect any mismatch with<br />
time; (2) The use of Hint differs from that of any other strategy, mainly because<br />
conventions of use, that is, lexical and grammatical conventions, are not at stake<br />
here. This aspect with be further discussed in "Conclusion" (below). 6<br />
Distribution of Hint Substrategies in Learners' Data<br />
As noted above, although Hints are essentially nonconventional in form and hence<br />
highly context-embedded, a qualitative analysis of CCSARP Hints reveals a consistent<br />
semantic pattern of Hint substrategies, differing in terms of opacity types and<br />
ranging in degree of opacity. Type classification is based upon the distinction<br />
between illocution and proposition, and is mainly a matter of focus rather than a<br />
clear-cut distinction. The degree of opacity is a theoretical construct, determined in<br />
terms of the (assumed) disambiguation procedures required for full interpretation of<br />
the speaker's meaning. Weizman (1989) classified and ranked native speakers'<br />
substrategy preferences on both the prepositional scale and the illocutionary scale.<br />
Beginning with the propositional scale, we can rank the native speakers' preferences<br />
from least to most opaque. Least opaque of the substrategies is reference to<br />
the requested act, where S explicitly refers to the act, but there is no reference to the<br />
effect that H is responsible for its performance, as in<br />
I haven't got time to clean up the kitchen.<br />
Slightly more opaque is reference to H's involvement, where S refers indirectly to<br />
H's responsibility, but does not name the requested act, as in<br />
You've left the kitchen in a mess.<br />
The next substrategy is reference to related components, where S refers to some<br />
object semantically related to the requested act, as in<br />
The kitchen is in a terrible mess.
Intel-language Requestive Hints 131<br />
Table 6.4. Distribution of Hints in Hebrew<br />
Learners' Data by Length of Stay in Israel<br />
Length of stay<br />
(years)<br />
N<br />
n<br />
%<br />
10<br />
24<br />
8<br />
33.3<br />
Note: n Hints were used by A' subjects.<br />
Of greatest opaqueness is the substrategy of zero propositional content, where no<br />
reference is made to any of the propositional components related to the requested<br />
act. Examples include:<br />
Are you going to give us a hand<br />
Are you going to do something for me<br />
Attention!<br />
Substrategies on the illocutionary scale can also be ranked from least to most<br />
opaque. Least opaque is questioning H's commitment, where S checks whether H<br />
feels committed to carrying out some act, the beneficiary of which is to be S. The<br />
nature of the requested act, however is not specified, nor is it hinted at. Examples<br />
include:<br />
Are you going to give us a hand<br />
Are you going to do something for me<br />
Somewhat more opaque is questioning feasibility, where a reference is made, indirectly<br />
or partially, to some precondition for the feasibility of the act. Examples<br />
include:<br />
Do you have a car<br />
Are you going directly home<br />
Have you got your note with you<br />
Yet more opaque is stating potential grounders, where a statement made by S is a<br />
potential argument for the performance of the requested act, and may therefore be<br />
interpreted as the reason for that request. Examples include:<br />
I've just missed my bus.<br />
I live near your place.<br />
The kitchen is in a bit of a mess.<br />
Madam, we have an injured person; this car is in the way.<br />
Of greatest opaqueness is the substrategy of zero illocutionary components, where<br />
no reference is made to the nature of the illocutionary act, and hence the interpretation<br />
of both the illocution and the propositional content is entirely contextembedded,<br />
as in<br />
Attention!
132 Speech Act Realization<br />
In this section, learners' use of Hints is compared with native speakers' use in<br />
terms of substrategy selection. Specifically, it raises the questions whether HL and<br />
HN use the same substrategies, and whether they use them with similar frequency.<br />
Table 6.5 presents the selection of Hint substrategies by Hebrew learners. The<br />
learners predominantly opt for the substrategy of Stating potential grounders<br />
(54.2%, « = 52). Their second choice is the combination of Stating potential<br />
grounders and Questioning feasibility (18.8%, n = 18), as in:<br />
Hi there. Are you going home now I've just missed my bus,<br />
Hi. How are you doing I missed yesterday's class. Do you have your notebook<br />
with you<br />
Only in the third place do we find the intermediate Questioning feasibility (12.5%,<br />
n = 12).<br />
The similarities and differences between learners and native speakers are evident<br />
from Table 6.6, which presents the distribution of the three predominant options in<br />
three populations: Hebrew learners (HL), native speakers of Hebrew (HN), and<br />
native speakers of Australian English and Canadian French (NS).<br />
First of all, the most opaque illocutionary substrategy is obviously preferred by<br />
all three groups (63.4% HL, 53.2% HN, 75.7% NS). This preference provides<br />
further support for the high-deniability hypothesis: Hints are exploited precisely by<br />
virtue of their opacity. Thus, if Hints are deemed situationally acceptable, the<br />
requester tends to opt for their most opaque variant, rather than trying to compensate<br />
for their inherent opacity by choosing a relatively transparent substrategy.<br />
However, Table 6.6 also reveals an important difference between learners and<br />
native speakers. While Hebrew learners turn to redundant combinations as their<br />
second choice (22%, n = 18), both HN and NS select them as their third choice<br />
Table 6.5. Distribution of Hint Substrategies in Hebrew Learners' Data 7<br />
Illocutionary<br />
scale<br />
Prepositional<br />
scale (%)<br />
Reference to<br />
requested act<br />
Reference<br />
to H's<br />
involvement<br />
Reference<br />
to related<br />
components<br />
Zero<br />
Various<br />
combinations<br />
Total<br />
Questioning feasibility<br />
Stating potential<br />
grounders<br />
0<br />
2.1<br />
11.5<br />
8.3<br />
0<br />
21.9<br />
0<br />
0<br />
1.0<br />
21.9<br />
12.5<br />
(12)<br />
54.2<br />
(52)<br />
Zero<br />
Other<br />
Potential grounders &<br />
questioning feasibility<br />
Total<br />
0<br />
1.0<br />
0<br />
3.1<br />
(3)<br />
0<br />
7.3<br />
1.0<br />
28.1<br />
(27)<br />
0<br />
2.1<br />
1.0<br />
25<br />
(24)<br />
1.0<br />
0<br />
0<br />
1.0<br />
(1)<br />
0<br />
3.1<br />
16.7<br />
42.7<br />
(41)<br />
1.0<br />
(1)<br />
13.5<br />
(13)<br />
18.8<br />
(18)<br />
100<br />
(96)
Interlanguage Requestive Hints 133<br />
Table 6.6. Distribution of the Three Predominant Hint Substrategies<br />
in Hebrew Learners (HL), Native Speakers of Hebrew (HN) and Native<br />
Speakers of Australian English and Canadian French (NS)<br />
Substrategies<br />
Subjects<br />
HL<br />
(N = 82)<br />
HN<br />
(N = 62)<br />
NS<br />
(N = 111)<br />
Stating potential grounders<br />
Stating potential grounders<br />
and questioning feasibility<br />
Questioning feasibility<br />
63.4%<br />
(n = 52)<br />
22%<br />
(n = 18)<br />
14.6%<br />
(n = 12)<br />
53.2%<br />
(n = 33)<br />
14.5%<br />
(n = 9)<br />
32.3%<br />
(n = 20)<br />
75.7%<br />
(n = 84)<br />
6.3%<br />
(n = 1)<br />
18%<br />
(« = 20)<br />
Note: Of N hints, (n) are of the substrategy under consideration.<br />
(14.5%, n = 9 and 5.2%, n = 7 respectively). These differences, statistically<br />
significant (HL differ from HN at p < 0.05 [P value = 0.0369, chi-square =<br />
6.5965], and from NS at/ < 0.01 [P value = 0.006, chi-square = 10.243], seem to<br />
be related to a most striking interlanguage specific feature: learners' preference for<br />
"verbosity" (Levenston 1971).<br />
In the realization of requests, learners' tendency toward verbosity was mainly<br />
observed in the use of supportive moves (explanations and justifications for the<br />
request) by learners of Hebrew and American English (Blum-Kulka & Olshtain,<br />
1986; Blum-Kulka & Levenston, 1987) and by Danish and German learners of<br />
English (House & Kasper, 1987; Edmondson & House, 1991; Faerch & Kasper,<br />
1989; House, 1989). A typical example for Hebrew speakers in English L2 would<br />
be<br />
I went over the material we will study in the next weeks and I would rather like to<br />
have your lecture notes next week, if it's possible and if you can be ready. (Blum-<br />
Kulka & Olshtain, 1986, 175)<br />
A typical example for German learners of English would be<br />
Good evening. Perhaps you've already seen me once. We're living on the same<br />
street. You know, my bus has just left, and as I- noticed that you have come by car I<br />
was going to ask you whether you could give me a lift. (House & Kasper, 1987,<br />
1283)<br />
Verbosity implies a redundant use of words. Similarly, combinations of Hint<br />
substrategies imply a redundant use of substrategies, since each of the substrategies<br />
may be used in an independent position as a self-sufficient requestive Hint (Weizman,<br />
1989, 88). It is suggested, therefore, that HL preference for redundancy is<br />
comparable with learners' preference for verbosity: if indeed verbosity is exploited<br />
by learners as part of their attempt to cope with communication problems—mainly<br />
insecurity resulting from the learning situation—then so is redundancy: both may be<br />
considered as typically "strategic," that is, related to the context of language<br />
learning.
134 Speech Act Realization<br />
Conclusion<br />
From a theoretical point of view, I suggested that the most striking feature of requestive<br />
Hints is that lack of conventions of form and hence context-embeddedness,<br />
reach a degree so high that deniability potential is always secured. The peculiarity<br />
of this feature, I pointed out, resides in its apparent mismatch with the concept of<br />
communication as a goal-oriented activity carried out by a rational agent. It is<br />
therefore the task of research into nonconventional indirectness to explain what goal<br />
could be efficiently pursued by the use of Hints.<br />
Researchers of interlanguage, I further hypothesized, might find the task more<br />
rewarding, were it to be found that this very deniability potential is liable to be<br />
exploited by learners to achieve a goal specific to the learning situation: the goal<br />
of overcoming unease and insecurity by avoiding unambiguous commitments to<br />
culture-specific meanings of conventions of form.<br />
In light of these assumptions, the findings support the following conclusions:<br />
1. Basically, learners seem well able to use Hints for the realization of requests:<br />
they do not use them in any markedly deviating manner as regards frequency,<br />
dependence upon situationally specific features, and preference for opaque Hint<br />
substrategies. The use of nonconventional indirectness, therefore, seems to be one<br />
of the pragmalinguistic essentials with which learners come to L2 and which,<br />
therefore, they need not acquire anew (Blum-Kulka, 1982).<br />
2. The hypothesis that the very lack of linguistic constraints inherent in Hints is<br />
exploited by learners to avoid risks is, at first sight, not supported by the findings,<br />
since learners do not reveal any marked preference for Hints, not even when they<br />
are first exposed to the target language. The issue, however, is more complicated. In<br />
various areas, learners were observed to exhibit general preference for putting<br />
information "on record." Thus, for instance, intermediate learners of English were<br />
shown to prefer propositional explicitness to ellipsis (Kasper 1981, 1982. For further<br />
discussion, cf. Kasper 1989, 245). If the tendency toward explicitness is indeed<br />
an interlanguage specific feature, one would expect to find learners more reluctant<br />
to use Hints than native speakers. In other words, along this line of argument, Hints<br />
would be less frequent in learners' data than in native speakers'. It could therefore<br />
be the case that similar use of Hints by native speakers and by learners reflects an<br />
interplay between two conflicting tendencies: on the one hand, preference for explicitness,<br />
encouraging a decrease in the frequency of Hints; and on the other, risk<br />
avoidance, entailing an increase in the frequency of Hints. The results, then, could<br />
be interpreted as indicating that these two tendencies counterbalance each other.<br />
This hypothesis should be empirically tested.<br />
3. Learners' tendency toward redundancy is no doubt a strategy of communication.<br />
This peculiarity is related to earlier findings concerning learners' preference<br />
for verbosity; I interpret it, therefore, as a typical feature of the learning situation.<br />
4. The findings have an important implication for the nature of indirectness in<br />
general. The fact that the same Hint substrategies are selected by learners and by<br />
native speakers, and that situational variation is similar in both populations, supports<br />
the assumption put forward in Weizman (1989): namely, that although apparently<br />
"open-ended," even nonconventional indirectness is governed by semantic<br />
norms. Moreover, learners' preference for the most opaque Hint substrategy further
Interlanguage Requestive Hints 135<br />
supports the hypothesis (Weizman, 1989) that opacity in Hints is not used as the last<br />
possible choice. Rather, it is purposely exploited as the most efficient means when<br />
deniability is deemed necessary.<br />
Acknowledgments<br />
I am indebted to Gabi Kasper and Shoshana Blum-Kulka for comments on an earlier<br />
version; to Hadassa Kantor for helping me obtain additional Hebrew learners' data,<br />
incorporated for my purposes into CCSARP learners' data; to Joel Walters for<br />
discussions of aspects of interlanguage; to Tsipi Parnassa for the statistical analysis;<br />
and to Miriam Schlesinger for stylistic comments on the last version of the paper.<br />
Notes<br />
1. This bi-dimensional analysis implies that the category of Hints encompasses several<br />
substrategies. The category includes Ervin-Tripp's (1976) "hints" and "non-explicit question<br />
directives," also labeled "nonconventional instrumental moves" or NCI's (Gordon & Ervin-<br />
Tripp, 1984). It further includes Walters's (1979, 286, 287) SVOs and Donde (where) questions.<br />
2. Cross-cultural variation is highly marked in all the quoted studies. For example,<br />
Walters (1979, 286, 287) notes a cultural difference in the scale value of politeness of<br />
SVOs (e.g., "you have any ball and rope"), which parallel roughly my "Questioning feasibility:<br />
Reference to hearer's involvement and to requested object." In Spanish, SVOs are<br />
located near the politeness end of the scale, that is, near conventional indirectness, while in<br />
American English they are not far from the impoliteness end. Blum-Kulka (1987) indicates<br />
that in American English Hints follow the typically conventional Query preparatories (Can<br />
you . . . Could you . . . ) as the next most polite strategy, while for Hebrew speakers,<br />
Hints are not only less polite than query preparatories but also less polite than performatives<br />
("I want you to . . .").<br />
3. These semantic preferences, I suggested, have to do, indirectly and partially, with<br />
some requirements presupposed by the felicity conditions, or with some constituents thereof<br />
(Weizman, 1985, 1590; whereas in conventional indirectness (e.g., in "can you" questions),<br />
the requester refers to the preparatory conditions explicitly and fully.<br />
4. Situational variation was analyzed for HL and HN separately, by means of the Chocran<br />
test, which takes into account only fully completed questionnaires. On this analysis, S5<br />
was found to be significantly different (alp < 0.05) from SI, S7, and Sll, while S15 was<br />
significantly different (at p < 0.05) from Sll. Differences between both populations were<br />
analyzed for each situation separately, by means of chi-square analysis.<br />
5. The total number of Hints («) is a dependent observation, since a considerable number<br />
of respondents were found to use Hints in at least two situations. In Table 6.2, for example,<br />
15.4% of the 65 HN Hints and 10.1% of the 109 HL Hints are dependent. Consequently, chisquare<br />
analysis is not appropriate for the total.<br />
6. Note, however, that while 17% (n = 8) of the members of the first group (length of<br />
stay up to one year) used hints in two situations, members of the other groups systematically<br />
chose no more than a single Hint each. This might suggest a decrease in Hint selection over<br />
time, but given the relatively small number of group members, further investigation is<br />
required.
136 Speech Act Realization<br />
1. The total number of Hints in HL data as presented in Tables 6.5 and 6.6 (n = 96)<br />
differs slightly from the total number of HL Hints in Table 6.2 (n = 109), due to a difference<br />
of nuance between the initial CCSARP definition of Hints and the one suggested in Weizman<br />
(1985, 1989). Note, however, that this difference has no bearing on the claims made about the<br />
similarities between learners and native speakers in terms of the selection of Hint substrategics.<br />
References<br />
Blum-Kulka, S. (1982). Learning to say what you mean in a second language: A study of the<br />
speech act performance of learners of Hebrew as a second language. Applied Linguistics,<br />
3, 29-60.<br />
Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different Journal of<br />
Pragmatics, 11, 145-60.<br />
Blum-Kulka, S. (1989). Playing it safe: The role of conventionality in indirectness. In<br />
S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Request<br />
and apologies (37-70). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., Danet, B., & Gerson, R. (1985). The language of requesting in Israeli<br />
society. In J. Forgas (Ed.), Language and social situation (113-41). New York:<br />
Springer-Verlag.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests<br />
and apologies. Norwood, NJ: Ablex.<br />
Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requesting<br />
behavior. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (123-54). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., & Levenston, E. (1987). Lexical-grammatical pragmatic indicators. Studies<br />
in Second Language Acquisition, 9, 155-70.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Studies in Second Language Acquisition, 8, 165-79.<br />
Brown, P., & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In<br />
E. A. Goody (Ed.), Questions and politeness: Strategies in social interaction (56—<br />
311). Cambridge: Cambridge University Press.<br />
Clark, H. H. (1979). Responding to indirect speech acts. Cognitive Psychology, 11, 430-77.<br />
Corder, S. P. (1983). Strategies of communication. In C. Faerch & G. Kasper (Eds.),<br />
Strategies in interlanguage communication (15—19). London: Longman.<br />
Dascal, M., & Weizman, E. (1987). Contextual exploitation of interpretation clues in text<br />
understanding: An integral model. In J. Verschueren & M. Bertucceli Papi (Eds.), The<br />
pragmatic perspective: Selected papers from the 1985 International Pragmatics Conference<br />
(31-45). Amsterdam: Benjamins.<br />
Edmondson, W., & House, J. (19191). Do learners talk too much The waffle phenomenon in<br />
interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood<br />
Smith, & M. Swain (Eds), Foreign!second language pedagogy research (273-87).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Ervin-Tripp, S. (1976). Is Sybill there The structure of some American English directives.<br />
Language in Society, 5, 25-66.<br />
Faerch, C., & Kasper, G. (1983). On identifying communication strategies in interlanguage<br />
production. In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication<br />
(210-38). London: Longmans.<br />
Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request
Interlanguage Requestive Hints 137<br />
realization. In S. Blum-Kulka, J. House, and G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (221-47). Norwood, NJ: Ablex.<br />
Eraser, B., Rintell, E., & Walters, J. (1980). An approach to conducting research on the<br />
acquisition of pragmatic competence in a second language. In D. Larsen-Freeman<br />
(Ed.), Discourse analysis in second language research (75-91). Rowley, MA: Newbury<br />
House.<br />
Gordon, D., & Ervin-Tripp, S. (1984). The structure of children's requests. In R. Schieffelbush<br />
& A. Pickar (Eds.), The acquisition of communicative competence (295-<br />
321). Baltimore, MD: Baltimore University Book Press.<br />
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and<br />
semantics (Vol. 3): Speech acts (41-58). New York: Academic Press.<br />
House, J. (1986). Cross-cultural pragmatics and foreign language teaching. In K. R. Bausch,<br />
F. G. Konigs, & R. Kogelheide (Eds.), Probleme und Perspektiven der Sprachlehrforschung<br />
(281-95). Frankfurt: Scriptor.<br />
House, J. (1989). Politeness in English and German: The functions of Please and Bitte. In<br />
S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests<br />
and apologies (96-119). Norwood, NJ: Ablex.<br />
House, J., & Kasper, G. (1987). Interlanguage pragmatics. Requesting in a foreign language.<br />
In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance.<br />
Festschrift for Werner Mullen on the occasion of his 60th birthday (1250-88).<br />
Tubingen: Narr.<br />
Kasher, A. (1982). Gricean inference revisited. Philosophica, 29, 25-44.<br />
Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr.<br />
Kasper, G. (1982). Teaching-induced aspects of interlanguage discourse. Studies in Second<br />
Language Acquisition, 4, 99-113.<br />
Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden,<br />
D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse<br />
and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.<br />
Leech, G. (1980). Explorations in semantics and pragmatics. Amsterdam: Benjamins.<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
Levenston, E. (1971). Over-indulgence and under-representation—aspects of mother tongue<br />
interference. In G. Nickel (Ed.), Papers in contrastive linguistics (115-21). Cambridge:<br />
Cambridge University Press.<br />
Rintell, E. (1981). Sociolinguistic variation and pragmatic ability: A look at learners. International<br />
Journal of the Sociology of Language, 27, 111-34.<br />
Searle, J. R. (1975). Indirect speech acts. In P. Cole & J. L. Morgan (Eds.), Syntax and<br />
semantics (Vol. 3). Speech acts (59-82). New York: Academic Press.<br />
Vanderveken, D. (1984). What is an illocutionary force In M. Dascal (Ed.), Dialogue: An<br />
interdisciplinary approach (185-208). Amsterdam: Benjamins.<br />
Walters, J. (1979). Strategies for requesting in Spanish and English: Structural similarities<br />
and pragmatic differences. Language Learning, 29, 277-93.<br />
Walters, J. (1981). Variation in the requesting behavior of bilingual children. International<br />
Journal of the Sociology of Language, 27, 35-58.<br />
Weizman, E. (1985). Towards an analysis of opaque utterances. Theoretical Linguistics, 12,<br />
153-63.<br />
Weizman, E. (1989). Requestive Hints. In S. Blum-Kulka, J. House, & G. Kasper (Eds.),<br />
Cross-cultural pragmatics: Requests and apologies (71-95). Norwood, NJ: Ablex.<br />
Weizman, E., & Dascal, M. (1991). On clues and cues: Strategies of text-understanding.<br />
Journal of Literary Semantics, XXI1, 18-30.
7<br />
Cross-Linguistic Influence<br />
in the Speech Act of Correction<br />
TOMOKO TAKAHASHI<br />
and LESLIE M. BEEBE<br />
In this chapter, we examine American and Japanese performance of the speech act<br />
of correction by looking at how this speech act is performed with status unequals—<br />
a person of lower status addressing someone of higher status and a person of higher<br />
status addressing someone of lower status. By focusing on correction, we are not<br />
looking at opinions, but rather at situations where one person knows the other has<br />
made a factual error. By looking at situations between a professor and a student, we<br />
are choosing situations where the roles are clearly defined and the status relationships<br />
are relatively clear-cut. The difference in status here is not necessarily socioeconomic,<br />
but rather a matter of the power structure within the classroom.<br />
In previous research, we have investigated American versus Japanese performance<br />
on several face-threatening speech acts such as disagreement (Beebe &<br />
Takahashi, 1989a,b), refusals (Beebe, Takahashi, & Uliss-Weltz, 1990; Takahashi<br />
& Beebe, 1986, 1987), announcing embarrassing information (Beebe & Takahashi,<br />
1989a), and chastisement (Beebe & Takahashi, 1989b). The present study continues<br />
our research project on sociolinguistic variation in face-threatening speech acts.<br />
Face-threatening speech acts have received much attention (e.g., Brown &<br />
Levinson, 1978) and have been found important because they are the source of<br />
cross-cultural miscommunications, or what Thomas (1983) calls "pragmatic failure."<br />
In this domain, we have focused on American and Japanese speakers because<br />
their cultures are vastly divergent; thus they were expected to show a large difference<br />
in the use of politeness strategies (e.g., Ide, 1989; also see Kasper 1990, for a<br />
review of literature on politeness; Blum-Kulka, House, & Kasper, 1989, for research<br />
on requests and apologies).<br />
In our research, we have analyzed each speech act as consisting of a sequence of<br />
"semantic formulas" (Cohen & Olshtain, 1981) such as the positive remark (e.g., "I<br />
wish I could, but . . ."), the expression of regret (e.g., "I'm sorry"), the refusal<br />
itself (e.g., "I can't come"), and so on. In examining the number, order, and/or<br />
138
The Speech Act of Correction 139<br />
content of semantic formulas used in several face-threatening speech acts, we have<br />
found differences in the realization of speech acts by Americans and Japanese.<br />
We shall focus here on the use of positive remarks, softeners, and other similar<br />
formulas. These formulas have been found particularly interesting because they are<br />
frequently used by Americans and/or Japanese in order to make each speech act less<br />
face-threatening. The definition of politeness varies from study to study (e.g.,<br />
Brown & Levinson, 1978, 1987; Ide, 1989; Lakoff, 1973; Leech, 1983). It is,<br />
however, generally agreed that politeness aims at smooth communication (Ide,<br />
1989), seeks to reduce friction (Lakoff, 1973), and to disarm the potential for<br />
aggression (Brown & Levinson, 1987). Along the same line, politeness is viewed<br />
here as the language usage developed in order to make a situation less facethreatening<br />
and to make communication smoother. In this sense, we consider positive<br />
remarks and softeners to be important politeness strategies.<br />
One of the major focuses of our research has been to examine the cross-cultural<br />
influence of Japanese upon English as a second language (ESL) at the discourse<br />
level. In this domain, we have found much evidence of native language (NL)<br />
transfer in Japanese ESL speakers as well as other examples of their efforts (successful<br />
or unsuccessful) to try to sound like English native speakers.<br />
Another focus of our research has-been on how Americans and Japanese perform<br />
such speech acts with status unequals. This question was asked because it is generally<br />
claimed that Japanese are very conscious of social status while Americans<br />
are relatively less status-conscious. Ide (1989), for example, argues that in many<br />
honorific languages such as Japanese "the choice of expression is made according to<br />
the variable of power and distance of the addressees" (Ide, 1989, 238). Similarly,<br />
Matsumoto (1989) argues that in Japanese, social context plays a much larger role<br />
than is generally assumed.<br />
Barnlund and Araki (1985, 23) also support the view that "status relations are<br />
critical in determining the character of communicative behavior in Japan (Nakane,<br />
1970; Nomura & Barnlund, 1983) and are less critical in the United States (de<br />
Tocqueville, 1965; Nomura & Barnlund, 1983)." In their study of compliments by<br />
Japanese and Americans, Barnlund and Araki (1985, 23) have found that "Japanese<br />
exchanged compliments less frequently in close relations and more often in relations<br />
that were more distant," while "with Americans it appeared that compliments might<br />
be exchanged more often in close relationships and less in more distant ones."<br />
It is thus suggested that both Japanese and Americans style-shift according to the<br />
psychological distance of the interlocutors. Based on the view that the Japanese are<br />
more conscious of social status than Americans, it can be further hypothesized that<br />
Japanese style-shift more than Americans according to the social as well as psychological<br />
distance of the addressees.<br />
In the investigation of sociolinguistic rule transfer at the phonological level,<br />
Beebe (1980) and Schmidt (1977) have found that sociolinguistic style-shifting<br />
patterns are transferred from the NL to the second language (L2). By focusing on<br />
the discourse level here, we discuss the differences among native English, Japanese<br />
ESL, and native Japanese responses, and ask whether and to what extent Americans<br />
and Japanese native speakers style-shift and whether to what extent Japanese ESL<br />
learners transfer patterns of sociolinguistic style-shifting from Japanese to English.
140 Speech Act Realization<br />
In sum, we discuss in this paper: (1) the use of politeness strategies that the<br />
speaker uses in order to make each correction less face-threatening (i.e., more<br />
polite)—e.g., positive remarks, softeners, expressions of regret; (2) NL transfer in<br />
the Japanese ESL responses; and (3) style-shifting by the three groups according to<br />
the status of the interlocutor. The data will be presented, for the most part, descriptively<br />
in terms of semantic content in order to give the flavor of responses by the<br />
three groups in an encounter of interlocutors of different status.<br />
Method<br />
The discourse completion questionnaire used for this study consisted of 12 items—<br />
two situations for each of the following: correction, disagreement, chastisement,<br />
announcing embarrassing information, and two other speech acts as controls (see<br />
Appendix for descriptions of the 12 situations). For each pair eliciting one type of<br />
speech act, one situation involved a person of lower status talking to someone in a<br />
higher position and one involved someone in a higher-status position talking to a<br />
person of lower status. The situations were presented in randomized order. In this<br />
paper, we shall discuss only correction.<br />
There were 55 subjects who filled out the questionnaire—15 Americans who<br />
responded in English (AE), 15 Japanese who responded in English (JE), and 25<br />
Japanese who responded in Japanese (JJ). The Japanese subjects tested in English<br />
had high-intermediate to advanced levels of ESL proficiency. All subjects were<br />
college graduates. For the Americans, there were 11 females and 4 males, with an<br />
average age of 33. For the Japanese ESL subjects, there were 8 females and 7 males,<br />
also with an average age of 33. For the Japanese tested in Japanese, there were 11<br />
females and 14 males, with an average age of 32. The American and Japanese ESL<br />
subjects were tested in New York, and the Japanese subjects using Japanese were<br />
tested in Tokyo, Japan.<br />
All the responses were analyzed as consisting of a sequence of "semantic formulas"<br />
(Cohen & Olshtain, 1981) or bits of meaning such as the expression of regret<br />
and the correction statement itself. For each of the subject groups, the total number<br />
of semantic formulas of any kind used for each situation was obtained.<br />
Results and Discussion<br />
Correction Situation 1 (Higher to Lower)<br />
Correction situation 1 read as follows:<br />
You are a professor in a history course. During class discussion, one of your<br />
students gives an account of a famous historical event with the wrong date.<br />
In this situation, we are dealing with an utterance that will be made by a person in a<br />
higher-status position (the professor) to a person in a lower-status position (the<br />
student). What is common to all three groups is that the majority of subjects decided
The Speech Act of Correction 141<br />
to correct the student. One American opted out, saying "thank you." One Japanese<br />
using English said to the class, "Does everyone agree with that date" but this was<br />
clearly an attempt to elicit peer correction. One Japanese using Japanese opted out,<br />
and another wrote "I will wait until other students notice the mistake."<br />
The most noticeable difference is that 9 out of 14 Americans (64%) prefaced<br />
their correction with at least one positive remark: most typically, "That was very<br />
good but I believe that took place in 1945." (Other responses were quite similar,<br />
with a variety of adjectives being used: "excellent," "great," "marvelous," etc.) Two<br />
American respondents used more than one positive remark: "That was a great<br />
account of the attack on Pearl Harbor. Everything was in line except the date. It was<br />
1942, not 1943," and "Excellent description, Henry! I like the way you outlined the<br />
events. Now when did all that take place" We have thus found 11 positive remarks<br />
(79%) used by the American group.<br />
Japanese ESL speakers, on the other hand, never once prefaced their responses<br />
with a wholehearted positive remark. Most typical of their responses was: "I think<br />
that date is not correct," or "The date is wrong," or simply, "In 1972." Only 3 out of<br />
13 Japanese ESL speakers (23%) used something resembling the American positive<br />
remark. But upon reading these remarks to a group of American native speakers, we<br />
were backed up in our intuition that the remarks were so lukewarm that a native<br />
speaker would hardly call them "positive" or feel comfortable with them. The<br />
remarks were: "Well, I'm almost satisfied with your account of that event except the<br />
date of it," "Your ideas is fine, but I think the age you mentioned is something<br />
wrong," "Thank you for bringing up interesting account, although the date you just<br />
mentioned was actually 1919."<br />
Japanese using Japanese used even fewer positive remarks: only 3 out of 23<br />
(13%) had the professor make a positive comment to correct his student's mistake.<br />
A typical Japanese response in Japanese (87% of the data) had no prefacing positive<br />
remark at all (e.g., "It's 1603," or "Wait a second. The date is incorrect," or "The<br />
date you just mentioned is incorrect. Please check it by the next class.") The three<br />
who used positive remarks used ones very similar to American English patterns.<br />
The remarks were: "Very nicely done, but why don't you check the date you just<br />
mentioned," "Your presentation was very good, but it was supposed to have taken<br />
place in 1549." "It was a very good presentation, but just one mistake—the date."<br />
The percentage of positive remarks used by the three groups is presented in<br />
Figure 7.1 The finding is that Americans use by far the highest percentage of<br />
positive remarks and Japanese using Japanese use the lowest with the Japanese<br />
speakers of English in between: AE > JE > JJ. This suggests that the percentage of<br />
positive remarks used by Japanese ESL learners is influenced by Japanese, where<br />
positive remarks are evidently not critical in this particular situation.<br />
Positive remarks (including praise, complimenting, and positive evaluations)<br />
are extremely important prefixes to face-threatening acts in English. In our research<br />
on refusals (Beebe et al., 1990), one of the striking differences between Japanese<br />
and English native speakers was that Americans felt the need to preface their refusals<br />
with what we called "positive adjuncts": for example, "I'd love to, [but] ..." "That<br />
sound wonderful, fbutj . . ."In our studies of disagreement (Beebe & Takahashi,<br />
1989a,b), we also found that Americans often used expressions of "token agree-
142 Speech Act Realization<br />
Fig. 7.1. Percentage of positive remarks in correction situation I, where the professor corrects<br />
the student's mistake (higher to lower). Note: These figures exclude the subjects who<br />
opted out and those who described what they would do instead of writing what they would<br />
say.<br />
Key:<br />
AE—Americans using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese<br />
merit" (Brown & Levinson, 1978): for example, "I agree with you, [but] . . ."<br />
"Yes, [but] ..." (also see LoCastro, 1986).<br />
Similarly, as we have found in this study, Americans use a positive remark to<br />
preface a correction to a lower-status person. This is an "adjunct" to the correcting<br />
act because although it is very common (used by 64% of our American respondents<br />
in this situation), it cannot stand alone and accomplish the speech act of correction.<br />
Just as "I'd love to . . ." is a vital to polite refusal in many situations but unable to<br />
stand alone as a refusal, a positive remark such as "It was a very good account"<br />
softens the upcoming correction but cannot substitute for it in any way. We thus<br />
maintain that it, too, is an "adjunct" to a correction.<br />
In addition to the use of positive remarks, American respondents seem to have<br />
made more efforts to soften their corrections than their Japanese counterparts. They<br />
tried, for instance, to soften the correction and make it the least face-threatening<br />
possible by using the following types of softening devices 71% of the time:<br />
(1) "softeners"—that is, expressions (or hedges) such as "I believe," "1 think,"<br />
"You may have (the wrong date)", (2) questions such as "Did you say . . . "<br />
"When did that happen", and (3) other expressions intended to lighten the gravity<br />
of the interlocutor's mistake or to defend the interlocutor, such as "You made one<br />
small error in the date."<br />
Japanese ESL respondents also used similar formulas: for example, (1) "I<br />
think," "I'm afraid" as softeners; (2) "Isn't it 1945" "That was in 1492, wasn't it"<br />
as questions; and (3) "I guess it's not very important" to lighten the gravity of the
The Speech Act of Correction 143<br />
mistake. However, Japanese respondents using English used softeners only 50% of<br />
the time, as compared with 71% use by Americans. Japanese responding in Japanese<br />
also used the equivalents, yet much less frequently (26% of the time): for<br />
example, (1) "I think"; (2) "Was it 1968"; and (3) "It might be a slip of the tongue."<br />
The differences among the groups in the use of all types of softening devices are<br />
presented in Figure 7.2.<br />
Essentially, the pattern for softeners is the same as that for positive remarks:<br />
AE > JE > JJ. That is, the largest percentage (71%) of softeners was used by the<br />
Americans, who used more than the Japanese using English (50%). And the Japanese<br />
using English used more than the Japanese using Japanese (26%). The Japanese<br />
using English result, being midway between the native speakers of Japanese and<br />
English, suggests cross-linguistic influence from the native language, Japanese, on<br />
the English of Japanese.<br />
As we compare Figure 7.2 with Figure 7.1, we notice that there is a smaller gap<br />
among the three groups in the use of softeners (Figure 7.2) than in the use of<br />
positive remarks (Figure 7.1). It has been pointed out that the Japanese tend to place<br />
value on silence and to be skeptical about the efficacy of communicating delicate<br />
thoughts or feelings through strictly verbal strategies. Clancy (1986), for example,<br />
has noted that "[the] Japanese have little faith in verbal expression or in those who<br />
rely upon it" (214). It is generally observed that the Japanese try to express regret,<br />
for instance, through nonverbal and/or paralinguistic means such as facial expres-<br />
Fig. 7.2. Percentage of softeners in correction situation I,* where the professor corrects tne<br />
student's mistake (higher to lower). Note: *Includes softeners of three types, including<br />
questions and other softening expressions. **These figures exclude the subjects who opted<br />
out and those who described what they would do instead of writing what they would say.<br />
Key: AE—Americans using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese
144 Speech Act Realization<br />
sions, tone of voice, sighs, hesitance, and so forth. Although our discourse completion<br />
questionnaires are not designed to collect such expressions, we have often<br />
found that Japanese native and ESL speakers used pauses to show hesitance or<br />
silence (see, e.g., Beebe et al., 1990). This could be seen as evidence that Japanese<br />
prefer less verbally explicit expressions, especially in face-threatening situations.<br />
Americans, on the other hand, seem to rely on verbal expression and use a<br />
clearly verbal strategy for dealing with a threat to face. American culture promotes<br />
explicit verbal means of correcting, refusing, or disagreeing and thus encourages<br />
explicit verbal means of undoing the threat to face that the explicit correction,<br />
refusal, or disagreement poses. This results in a strong tendency toward a ritualized<br />
use of the positive remark. In other words, the positive remark is a very important<br />
politeness strategy for Americans.<br />
The Japanese rely much less on the positive remark than Americans. Instead,<br />
they use softeners, which Americans also use. The softener is different from the<br />
positive remark in that the former is a down-toning device integrated in the main<br />
body of a speech act (e.g., / think [softener] it's 1942, not 1943 [main body of the<br />
speech act of correction]), while the positive remark is a preceding adjunct that is<br />
phrasal and separate from the main body (e.g., It was a good presentation,<br />
but . . .). It seems, therefore, possible that the Japanese use the positive remark<br />
much less often because of its explicitly verbal nature. It is also possible that the<br />
positive remark is avoided because it may sometimes sound too untruthful for<br />
people who have little faith in verbal expression. For Japanese ESL learners, it is<br />
much easier to transfer the NL knowledge in the use of the softener than that of the<br />
positive remark.<br />
Apart from the use of these formulas, the overall tone of the Japanese responses<br />
in Japanese is quite different from that of American responses. For instance, 6 out of<br />
23 Japanese speakers using Japanese (26%) chose to point out directly that there was<br />
a mistake in the presentation (e.g., "The date you just mentioned is incorrect") and<br />
then to say, for example, "Please check the date" or "Why don't you check the<br />
date" All in all, Japanese responses sounded much more authoritarian than American<br />
responses. They had a ring of superiority that came from the use of a directive to<br />
check the date and from its position following a direct statement that a mistake had<br />
been made.<br />
Correction Situation 2 (Lower to Higher)<br />
Let us now look at situation 2, which reads as follows:<br />
You are a student in a sociology class. During the lecture, the professor quotes a<br />
famous statement attributing it to the wrong scholar.<br />
In this situation, the status of the participants was reversed; it was the professor who<br />
made the mistake. Therefore, any correction attempted would have to be from the<br />
lower-status interlocutor (the student) to the higher-status interlocutor (the professor).<br />
The most striking difference between the two status situations was that no native<br />
Americans or Japanese ESL speakers used a positive remark to a higher-status<br />
person before a correction ("It was a very interesting lecture"), and only one Japa-
The Speech Act of Correction 145<br />
nese using Japanese did so. It was evidently judged totally inappropriate to praise<br />
the professor and then correct. Thus, only one out of 37 subjects (who opted to say<br />
something) used a positive remark (or positive remarks) from student to professor,<br />
whereas in the opposite situation from professor to student, 15 subjects out of 50<br />
(who opted to say something) used a positive remark. This shows that all groups<br />
were highly sensitive to status differences in this regard.<br />
Avoidance or "opting out" (Bonikowska, 1988) was indeed possible since the<br />
overall directions to the questionnaire made it clear that saying nothing was an<br />
option for any of the situations. Two Americans opted out completely. Another<br />
wrote, "Probably nothing, or I would say, 'Are you sure you have the correct name<br />
Isn't it ...'" And another wrote, "Probably nothing, but in a small seminar I<br />
might say something like 'You mean Ellis, don't you' especially if anyone else<br />
looked confused." Still another American commented, "If I did not know the<br />
professor well, I wouldn't say anything. If I did know the professor, I would<br />
approach him after class and say, 'Wasn't it X who said Y'"<br />
Three Japanese ESL speakers opted to say nothing. Another said, "I would not<br />
take prompt action. I would observe others' reactions, and if nobody seemed to have<br />
noticed, I would say something." What was interesting about those who opted out<br />
was that two out of three Japanese ESL subjects who said they would say nothing<br />
specified that they expected the professor to recognize his mistake. Americans did<br />
not seem to have this reason for keeping quiet.<br />
Japanese responding in Japanese were the ones who avoided the difficulty of<br />
correcting the professor most: 10 out of 25 (40%) decided to opt out, whereas 3<br />
Japanese ESL learners (20%) said they would say nothing and 2 Americans (13%)<br />
opted out.<br />
The dominant pattern used by all three groups in this situation (from the student<br />
to the professor) was, again, to use a softener (or softeners) such as "I think that<br />
was . . ." before a correction. The percentage of softeners used by the three different<br />
groups is presented in Figure 7.3.<br />
What is most interesting in the differences between Japanese ESL and native<br />
American responses is not the frequency of softeners used or the sequence of<br />
formulas that the respondents chose to use. It is rather the wording of the softener or<br />
the correction formula that is of particular interest. Books comparing Japanese and<br />
American culture stress that Americans value variety and originality in verbal<br />
expression as a way of demonstrating sincere feeling as opposed to perfunctory<br />
politeness (see Condon, 1984; Goldstein & Tamura, 1975). They warn Americans<br />
not to try to transfer this into their Japanese. Japanese culture prescribes certain<br />
formulaic expressions that can be used over and over again in appropriate situations<br />
without needing to be varied or elaborated upon to convey politeness (see Coulmas,<br />
1981; Loveday, 1982). The American value placed on variety is not shared by the<br />
Japanese. On the contrary, knowing and using the correct formula is what counts in<br />
Japanese politeness (see Beebe & Takahashi, 1989a,b, for similar findings).<br />
Clancy (1986, 216) explains why the Japanese depend on formulaic expressions<br />
as follows:<br />
In Japan, there seems to be an extensive codification of contexts in which particular<br />
feelings are expected; speakers need only indicate, by means of the right
146 Speech Act Realization<br />
Fig. 7.3. Percentage of softeners in correction situation II, where the student corrects the<br />
professor's mistake (lower to higher). Note: *These figures exclude the subjects who opted<br />
out and those who described what they would do instead of writing what they would say.<br />
**Some subjects used softeners more than once.<br />
Key: AE-Americans using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese<br />
formula, that they are experiencing the appropriate reaction, without expressing<br />
any more personal, individualized response. An important goal of socialization in<br />
Japan is to promote the unanimity in feeling that will support the norms of verbal<br />
agreement and empathy.<br />
She further notes that Japanese mothers provide instruction in verbal formulas and<br />
expect from their children much earlier mastery of social courtesy than American<br />
mothers (Clancy, 1986, 236).<br />
Ide (1989) would argue that Japanese has a much higher degree of "discernment"<br />
(obligatory choice) in expressing politeness; that is, the choice of formal<br />
linguistic forms such as honorifics or morphological marking on the verb according<br />
to social conventions. American English, on the other hand, is high in "volition"<br />
(strategic choice); that is, the choice of interactional strategies such as jokes or<br />
questions according to the speaker's intention. The Japanese usually search for the<br />
appropriate linguistic formulas or politeness marking according to the variable of<br />
power and distance of the addressees, while American English speakers are allowed<br />
to make a considerably more active choice from a relatively wider range of possibilities.<br />
In other words, Japanese and American English are different in that the<br />
speaker's focus is placed on the socially prescribed norm in Japanese and on the<br />
speaker's own intention in American English. Japanese ESL learners' use of formulaic<br />
expressions, therefore, can be explained as transfer of a belief in the efficacy of<br />
choosing the appropriate expression according to relative status. This would be<br />
another reason why Japanese prefer formulaic expressions to more personal and<br />
individualized responses.
The, Speech Act of Correction 147<br />
In this research, the difference in softeners used by Japanese and American<br />
respondents may reflect these purported cultural differences. Americans seemed to<br />
be elaborating their softeners, by using more self-deprecating softeners with a<br />
greater variety to wording and a preference for the "[clause] + but ..." structure<br />
such as "I may be wrong/mistaken, but . . ." and "I'm not sure, but . . ." Then,<br />
they said, "I think that . . ." One American used a combination of softeners: "I'm<br />
curious about the quote you read. I was under the impression White said that. Am I<br />
wrong"<br />
Japanese ESL respondents tended to stick to short, formulaic softeners, such as<br />
"I'm afraid," "I think," or "I understand." Their responses might have been shorter<br />
and less varied simply because they were using a second language. Empirical<br />
investigation of non-Japanese learners at the same proficiency levels would be<br />
needed to verify this. However, the apparently formulaic brevity is worth examining<br />
because natural data have also shown that Japanese ESL speakers overuse "I'm<br />
afraid" as a softener. Although it was surely meant to have a polite softening effect<br />
when used by Japanese ESL subjects, it was not used at all by English native<br />
speakers.<br />
Americans might avoid the use of "I'm afraid," especially from student to<br />
professor, because it has a slight ring of superiority. ("I'm afraid I made a mistake"<br />
sounds appropriate but "I'm afraid you made a mistake" sounds patronizing.) It is<br />
highly unlikely that the Japanese ESL speakers meant to convey this tone since they<br />
used softeners much more frequently from student to professor than from professor<br />
to student. As we look at the responses by Japanese using Japanese, the ESL use of<br />
"I'm afraid" becomes a curious case because Japanese native speakers did not use<br />
the equivalent of "I'm afraid" in their Japanese. Instead they used the question form<br />
predominantly (80%) in order to soften corrections: for example, "Isn't it X who<br />
said Y" or "Isn't that quoted by X"<br />
With respect to formulaic expressions used by native speakers, some studies<br />
(Cohen & Olshtain, 1981; Olshtain & Cohen, 1983; Olshtain, 1983; Owen, 1980)<br />
report that English native speakers' choices of apology forms are highly patterned.<br />
In other words, there are major strategies frequently used to express an apology: for<br />
example, "I'm sorry" and "It was my fault" (see Olshtain & Cohen, 1983). Eraser<br />
(1981) has found that the more formal the situation, the longer and more elaborate<br />
the apology. Thus, it can be said that the degree to which major strategies are<br />
actually ritualized may vary in different cultures as well as according to different<br />
speech acts and situations. In Japanese culture, however, it seems that formulaic<br />
expressions are especially favored.<br />
Another interesting finding was that one Japanese ESL respondent used a question<br />
form quite different from the dominant questioning pattern used by all three<br />
groups; namely, "Isn't it X who said Y" The Japanese ESL subject said, "Would<br />
you please tell me who made this statement again" This type of seemingly factual<br />
question, although it has never been the predominant pattern in native Japanese<br />
or Japanese ESL, has cropped up repeatedly in our research on Japanese use of<br />
face-threatening speech acts such as warning, disagreement, and chastisement (see<br />
Beebe & Takahashi, 1989a,b; Beebe & Takahashi, 1988).<br />
On the face of it, this question form could be interpreted as a very polite way of
148 Speech Act Realization<br />
correcting someone, but it is not so polite as it is intended to be. Indirectness is not<br />
always highly valued in American speech. This finding is corroborated by Blum-<br />
Kulka's (1987) finding that indirectness and politeness do not necessarily co-vary.<br />
In her study of English and Hebrew requests, Blum-Kulka (1987) examined different<br />
request categories including hints (e.g., "We don't want any crowding" [as a<br />
request to move the car]) and conventional request forms (e.g., "Would you mind<br />
moving your car"). She has found that American native speakers consider hints<br />
less polite than conventional indirectness although hints are rated more indirect than<br />
conventional indirectness. Similarly, in our research, when told that some Japanese<br />
respondents would say nothing to correct the professor, expecting that the professor<br />
would see his or her own mistake, Americans were surprised and a little amused.<br />
When told that one Japanese said, "Would you please tell me who made this<br />
statement again" a native English speaker immediately piped up, "It sounds as if<br />
you're trying to trap him."<br />
The use of a question to elicit a self-correction is fairly frequent from a teacher<br />
to a student in class, but the request from a student to a professor to repeat something<br />
is likely to be interpreted as a simple request. When the professor discovers<br />
that it is an indirect way of making a correction, he or she is likely to feel embarrassed<br />
or annoyed, particularly if he or she has fallen for it and answered the<br />
question on a literal level (see Beebe & Takahashi, 1989a,b, for a discussion of a<br />
distinctively Japanese questioning strategy found in natural data).<br />
It is generally recognized that the communicative style of the Japanese is indirect<br />
and intuitive, especially compared with that of Americans. In her study of<br />
Japanese mother-child interaction, Clancy (1986) reports that Japanese mothers<br />
frequently use rhetorical questions and hints to express a variety of negative attitudes<br />
and opinions such as skepticism, frustration, and disapproval to their children.<br />
She also reports that Japanese mothers strongly emphasize sensitivity to the needs,<br />
wishes, and feelings of others. Such training serves a purpose in Japanese interaction.<br />
This point is well elucidated by Kasper (1990) as follows: "Rather than<br />
emphasizing distance, indirectness in Japanese culture appears to express empathy<br />
between the participants, symbolizing a high degree of shared presuppositions and<br />
expectancies" (200). Upon receiving an indirect request, for example, the Japanese<br />
feel that the speaker is trying to make the speech act less face-threatening; they also<br />
understand (or at least try to understand) the feelings and needs of the speaker in<br />
making such a request. It is, therefore, possible that in Japanese culture, indirectness<br />
serves to symbolize empathy. It is also likely that a closer relationship exists in<br />
Japanese culture between indirectness and politeness than in American culture.<br />
Two Correction Situations Compared<br />
Now let us turn to the data on style shifting and investigate whether the styleshifting<br />
pattern of the Japanese using English might be transferred from their native<br />
Japanese pattern. As we compare the two status situations, we find a dramatic style<br />
shift in the use of softeners among Japanese subjects. Figure 7.4 presents the<br />
comparison between the two status situations in the use of softeners by the three
The Speech Act of Correction 149<br />
Fig. 7.4 Percentage of softeners in the two correction situations.<br />
Key: AE—American using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese<br />
subject groups. As seen here, the Americans style-shifted the least, while the<br />
Japanese both in Japanese and English displayed great style shifting by the use of<br />
softeners. This pattern found in the use of softeners within each situation again<br />
suggests that there is native language transfer; this time, however, it is not just the<br />
use of softeners that shows cross-linguistic influence, it is actually the style shifting<br />
itself that Japanese ESL speakers are transferring.<br />
This is reminiscent of a finding from our refusal data (Beebe et al., 1990), where<br />
Japanese in both English and Japanese style-shifted more than Americans using<br />
English in the use of expressions of regret (e.g., "I'm sorry"). Figure 7.5 is based on<br />
our refusal data.<br />
Similarly, the use of positive remarks prefacing disagreements suggests similar<br />
style-shifting tendencies, as seen in Figure 7.6. The data on Japanese using Japanese<br />
are not available on disagreements so it is not possible to show in this instance that<br />
JJ > JE > AE; however, the finding that JE > AE is consistent with the other<br />
findings.<br />
Figures 7.4, 7.5, and 7.6 suggest that the lower- to higher-status situation is<br />
more face-threatening than the other status situation, higher to lower, especially for
150 Speech Act Realization<br />
Fig. 7.5. Percentage of expressions of regret in the two refusal situations. (Data from Beebe,<br />
Takahashi, & Uliss-Weltz, 1990.)<br />
Key: AE—Americans using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese<br />
the Japanese in both English and Japanese. More softening devices (softeners,<br />
positive remarks, and expressions of regret) are used with the lower-status person<br />
addressing the higher-status person than vice versa. The degree of shift is greater for<br />
Japanese subjects than for Americans in each case and greater for Japanese using<br />
Japanese than Japanese using English in the corrections and refusals where both sets<br />
of data are available.<br />
In correction, as opposed to disagreement and refusal, there are many situations<br />
where elaboration and multiple prefacing positive remarks are not appropriate. It<br />
would sound very odd for a student to say, for example, to a professor, "You are a<br />
very intelligent person, you know a lot about this subject, and you have certainly<br />
earned your reputation as an expert in this area, but that novel was published in<br />
1972, not 1970." The number and character of prefacing remarks depends upon the<br />
situation, the status of the participants, and the speech act. An ESL teacher, on the<br />
other hand, might compliment a student's book report before correcting a small<br />
error, but would not feel compelled in a drill session to give a positive remark about
The Speech Act of Correction 151<br />
Fig. 7.6. Percentage of positive remarks and token agreement in the two disagreement<br />
situations. (Data from Beebe & Takahashi, 1989a, b; no JJ data are available in this study.)<br />
Key: AE—Americans using English<br />
JE—Japanese using English<br />
every sentence the student utters. Thus, the situation is a key variable in the use of<br />
positive remarks.<br />
Similarly, the social status of the two interlocutors is very important in determining<br />
whether a positive remark should be used. A professor might say something<br />
positive before correcting a factual error by a student ("That was a good report, but<br />
it was in 1942, not 1943") but a student would sound arrogant using a positive<br />
remark with a professor ("That was a fascinating lecture, but it was President<br />
Carter, not President Ford"). The positive remark often used in corrections is praise<br />
of the interlocutor, and such praise usually come from the "superior," not from the<br />
lower-status to the higher-status person. The positive remarks often used in refusals<br />
and disagreement are an expression of the speaker's positive feeling about the offer<br />
or invitation (e.g., "Oh, I wish I could, but . . .") or token agreement with the<br />
interlocutor's opinion (e.g., "I agree with you, but . . ." "Yes, but . . ."). This type<br />
of positive remark is not direct praise of the interlocutor. It seems, therefore,<br />
necessary to distinguish positive remarks in different speech acts.<br />
In this study, we have found that none of the subjects (except for one JJ) used a
152 Speech Act Realization<br />
positive remark in the situation where a student corrects a professor, as seen in<br />
Figure 7.7. Here we see that all three groups converged in the status situation Lower<br />
to Higher. At the same time, Americans style-shifted most in this particular situation.<br />
As previously discussed, the positive remark is one of the most important<br />
politeness strategies Americans employ in face-threatening situations. This formula<br />
was used as often as 79% of the time by Americans in the situation where a higherstatus<br />
person is talking to a lower-status person, while it was used by none in the<br />
other situation. The shift from 79% to zero is much greater than those found in the<br />
performance of the Japanese subjects, who simply do not use the positive remark<br />
very often in any situation.<br />
As mentioned above, in the speech act of correction, the situation, the relative<br />
social status, expertise status, role status or other kinds of status have been found<br />
important factors influencing the way the speaker performs. When we engage in a<br />
speech act like giving embarrassing information, however, it does not seem to<br />
matter so much who has what status. Rather, the prefacing positive remark seems<br />
uniformly out of place. We say, "I think you have something in your teeth," not<br />
"You are really well groomed today and I know you are an avid tooth-brusher, but<br />
you have spinach in your teeth." Some face-threatening acts seem to require more<br />
elaboration and some require simplicity.<br />
The pattern of style shifting may vary according to many factors. In our research,<br />
we have found that, generally, native speakers of Japanese are influenced in<br />
the way they perform the speech act of correction in English by the way they<br />
perform corrections in Japanese; for example, they style-shift in English according<br />
to the status of the interlocutor in a manner similar to the style-shifting typical of<br />
native Japanese.<br />
Lastly, it should be pointed out that our Japanese ESL subjects had highintermediate<br />
to advanced levels of ESL proficiency. In our study of refusals by<br />
Americans, Japanese native speakers, and two Japanese ESL proficiency groups<br />
(Takahashi & Beebe, 1987), we argued that transfer was greater among highproficiency<br />
ESL learners than their low-proficiency counterparts at the discourse<br />
level. The lower-proficiency students do not have the fluency in the target language<br />
to give free rein to pragmatic transfer phenomena. The high-proficiency students,<br />
on the other hand, have the control over English to express Japanese sentiments at<br />
the pragmatic level. In her recent study of politeness in Japanese, Matsumoto (1989)<br />
has argued that in Japanese, "politeness expressions are not simply additions to a<br />
neutral utterance which conveys the prepositional content: rather, the structure of<br />
the language requires some choice of expression that conveys additional information<br />
on the social context" (208). In other words, Japanese utterances are always marked<br />
linguistically for social context (also see Ide, 1989). That is, some utterances may<br />
have exactly the same content but vary in linguistic forms according to the social<br />
contexts in which the utterances were made. Japanese ESL learners may try to<br />
transfer such linguistic markings into English, but are unable to find places to<br />
transfer these features. Instead, their utterances in different social contexts may<br />
sound uniform in English regardless of differences in the social contexts because the<br />
utterances are the same in content and only the content is expressed in English.<br />
Advanced learners, on the other hand, find substitutes for the places to transfer these
The Speech Act of Correction 153<br />
Fig. 7.7. Percentage of positive remarks in the two correction situations.<br />
Key: AE—Americans using English<br />
JE—Japanese using English<br />
JJ—Japanese using Japanese<br />
social differences or at least to express such sentiments. In the present study, we<br />
have found that ESL learners transfer Japanese style-shifting patterns into English.<br />
If we had tested low-proficiency learners, however, we might not have found the<br />
same phenomenon.<br />
Conclusion<br />
In sum, we have uncovered a number of patterns in the speech act of correction.<br />
First, we have demonstrated that it is a typically American pattern to use a positive<br />
remark such as "That was a great account" before saying "but" and making a<br />
correction when speaking to a person of lower status. Only 13% of the Japanese<br />
using Japanese did this whereas 64% of the Americans did. Japanese using English<br />
were in between. This pattern can be schematically represented as AE > JE > JJ. It<br />
provides evidence that there was cross-linguistic influence from Japanese. This<br />
same pattern could be claimed to show incomplete acquisition, however.<br />
Second, we have shown that softeners give evidence of cross-linguistic influ-
154 Speech Act Realization<br />
ence in both the situation with the higher- to lower-status interlocutor and the<br />
situation with the lower- to higher-status interlocutor. With the professor to the<br />
student, we see that softeners were used with correction 71% of the time, and<br />
the pattern is: AE > JE > JJ. The exact opposite pattern exists with the student<br />
speaking to the professor: AE < JE < JJ.<br />
Third, and most important, particularly striking in our data is that style-shifting<br />
patterns are transferred from the native language, Japanese. Whereas transfer has<br />
been well documented on the phonological and syntactic levels, there was traditionally<br />
less interest at the discourse level. Recent studies, however, have shown<br />
that transfer at the discourse level is also widespread (e.g., Beebe et al., 1990;<br />
Blum-Kulka, 1982, 1983; Cohen & Olshtain, 1981; Olshtain, 1983; Olshtain &<br />
Cohen, 1983; Takahashi & Beebe, 1987). What is striking in this study is that it is<br />
not only discourse patterns that transfer, but also the style shift between two different<br />
discourse patterns with interlocutors of different status.<br />
It is the Japanese using Japanese whose style shifting shows the greatest frequency<br />
in certain situations. Americans, in the same situations, show much more use of<br />
softeners, whether they are speaking to someone of higher or lower status. The<br />
style-shifting patterns are important because they are sociolinguistic evidence of a<br />
significant aspect of Japanese and American cultures. Sakamoto and Naotsuka<br />
(1982) have claimed that Japanese and American cultures have different "polite<br />
fictions" upon which they operate. Americans go by the polite fiction that "you and<br />
I are equals," whereas Japanese go by the polite fiction that "you are my superior."<br />
In our research, we have found these stereotypical attitudes to be of some value, but<br />
also limited. Americans are not unconscious of status; rather, they often show how<br />
conscious of status they are by making moves to downplay or cover up status<br />
differences—to deny that status differences do exist (see Figures 7.4, 7.5, and 7.6).<br />
Americans are indeed aware of such differences, and as in the use of positive<br />
remarks in the two correction situations (see Figure 7.7), they sometimes show a<br />
very marked style-shifting pattern. Japanese, on the other hand, are conscious of<br />
status, and do not try to cover it up (see Figures 7.4, 7.5, and 7.6).<br />
Since style shifting according to interlocutor status has been shown to be greater<br />
among Japanese using English than among Americans using English in refusals and<br />
disagreements as well as in corrections, we naturally believe that there is a common<br />
explanation. Where data on Japanese using Japanese are available, that is, on<br />
refusals and corrections, the data are consistent with an explanation of transfer. It is<br />
not only the semantic content that is transferred but, more interesting, the patterns<br />
of style shifting.<br />
Acknowledgments<br />
We would like to express our appreciation to the editors of this volume for their<br />
helpful comments. We also wish to thank the students in the TESOL and Applied<br />
Linguistics programs at Teachers College, Columbia University, for their many<br />
suggestions. We also thank those who took the time to answer the discourse completion<br />
questionnaire for the present research.
The Speech Act of Correction 155<br />
Appendix<br />
The following are the 12 situations used in the questionnaire. After the description of each<br />
situation there was a three-line blank for the respondents to write what they thought they<br />
would say. "Higher to Lower" indicates that the particular speech act was performed by a<br />
higher-status person speaking to a lower-status person, and "Lower to Higher," vice versa.<br />
The situations were presented in randomized order.<br />
[Correction—Higher to Lower] (This categorization did not appear in the questionnaire.)<br />
You are a professor in a history course. During class discussion, one of your students gives an<br />
account of a famous historical event with the wrong date.<br />
[Correction—Lower to Higher]<br />
You are a student in a sociology class. During the lecture, the professor quotes a famous<br />
statement, attributing it to the wrong scholar.<br />
[Disagreement—Higher to Lower]<br />
You are a corporate executive. Your assistant submits a proposal for reassignment of secretarial<br />
duties in your division. Your assistant describes the benefits of this new plan, but you<br />
believe it will not work.<br />
[Disagreement—Lower to Higher]<br />
You work in a corporation. Your boss presents you with a plan for reorganization of the<br />
department that you are convinced will not work. Your boss says: "Isn't this a great plan"<br />
[Chastisement—Higher to Lower]<br />
You are a corporation president and you have asked your assistant to prepare xerox copies of<br />
essential documents for an important press conference. Your assistant arrives at the last<br />
moment with 100 copies of the wrong materials.<br />
[Chastisement—Lower to Higher]<br />
You are a middle manager in a large corporation and your boss hands you a 50-page<br />
document, asking you to make 30 copies of each page. Ten minutes later he comes back to get<br />
the copies because it turns out he only wanted 30 copies of the front page. You have just made<br />
10 copies of the whole packet. Obviously he is angry with you.<br />
[Announcing Embarrassing Information—Higher to Lower]<br />
You are a corporate executive talking to your assistant. Your assistant, who will be greeting<br />
some important guests arriving soon, has some spinach in his/her teeth.<br />
[Announcing Embarrassing Information—Lower to Higher]<br />
You are a student, speaking with your professor to prepare for a three-way meeting you have<br />
with the dean of the college. Your professor has some mustard on his cheek from lunch and<br />
you are aware of it.<br />
[Praise (Control)—Higher to Lower]<br />
You are a teacher trainer and one of your students submits an original paper that gives you<br />
fresh insights into teacher behavior. You run into the student in the hallway.
156 Speech Act Realization<br />
[Praise (Control)—Lower to Higher]<br />
You are a beginning graduate student and you attend a lecture given by a famous scholar in<br />
your field. During the lecture you become aware of some ideas that you have never thought<br />
about before. You are excited about the lecture and approach the professor afterward.<br />
[Persuasion (Control)—Higher to Lower]<br />
You are a corporate executive. You have a plan for an advertising campaign for a new<br />
product. You are enthusiastic about the plan and need the cooperation of your assistants in<br />
implementing it.<br />
[Persuasion (Control)—Lower to Higher)<br />
You are a middle manager in a large corporation. Your company exports its products<br />
to a foreign country where you have lived. You know of a major problem in the<br />
product design in terms of its practicality for that foreign market. You want the<br />
corporation president to order a revision.<br />
References<br />
Barnlund, D., & Araki, S. (1985). Intercultural encounters: The management of compliments<br />
by Japanese and Americans. Journal of Cross-Cultural Psychology, 16, 9-26.<br />
Beebe, L. (1980). The sociolinguistic variation and style shifting in second language acquisition.<br />
Language Learning, 30, 433-47.<br />
Beebe, L., & Takahashi, T. (1988, January). Too many questions: Acquiring the social rules<br />
of speaking in a second language. Paper presented at the 10th Annual Applied<br />
Linguistics Winter Conference, sponsored by NYS TESOL. New York, NY.<br />
Beebe, L., & Takahashi, T. (1989a). Do you have a bag: Social status and patterned<br />
variation in second language acquisition. In S. Gass, C. Madden, D. Preston, and L.<br />
Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics<br />
(103-25). Clevedon and Philadelphia: Multicultural Matters.<br />
Beebe, L., & Takahashi, T. (1989b). Sociolinguistic variation in face-threatening speech acts:<br />
Chastisement and disagreement. In M. Eisenstein (Ed.), The dynamic interlanguage:<br />
Empirical studies in second language variation (199-218). New York: Plenum.<br />
Beebe, L., & Takahashi, T. (1988, January). Too many questions: Acquiring the social rules<br />
of speaking in a second language.<br />
Beebe, L., Takahashi, T, & Uliss-Weltz, R. (1990). Pragmatic transfer in ESL refusals. In<br />
R. C. Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative<br />
competence in a second language (55-73). New York: Newbury House.<br />
Blum-Kulka, S. (1982). Learning to say what you mean in a second language: A study of the<br />
speech act performance of learners of Hebrew as a second language. Applied Linguistics,<br />
3, 29-60.<br />
Blum-Kulka, S. (1983). Interpreting and performing speech acts in a second language—A<br />
cross-cultural study of Hebrew and English. In N. Wolfson & E. Judd (Eds.), Sociolinguistics<br />
and language acquisition (36-55). Rowley, MA: Newbury House.<br />
Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different Journal of<br />
Pragmatics, 11, 131-46.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests<br />
and apologies. Norwood, NJ: Ablex.<br />
Bonikowska, M. (1988). The choice of opting out. Applied Linguistics, 9, 169-84.<br />
Brown, P., & Lcvinson, S. (1978). Universals in language usage: Politeness phenomena. In
The Speech Act of Correction 157<br />
E. W. Goody (Ed.), Questions and politeness (54-310). Cambridge: Cambridge<br />
University Press.<br />
Brown, P., & Levinson, S. (1987). Politeness: Some universals in language use. Cambridge:<br />
Cambridge University Press.<br />
Clancy, P. (1986). The acquisition of communicative style in Japanese. In B. Schieffelin & E.<br />
Ochs (Eds.), Language socialization across cultures (213-50). New York: Academic<br />
Press.<br />
Cohen, A., & Olshtain, E. (1981). Developing a measure of sociocultural competence: The<br />
case of apology. Language Learning, 31, 113-34.<br />
Condon, J. C. (1984). With respect to the Japanese. Tokyo: Yohan.<br />
Coulmas, F. (1981). "Poison to your soul": Thanks and apologies contrastively viewed. In F.<br />
Coulmas (Ed.), Conversational routines (69-91). The Hague: Mouton.<br />
De Tocqueville, A. (1965). Democracy in America. New York: Academic Press.<br />
Fraser, B. (1981). On apologizing. In F. Coulmas (Ed.), Conversational routines (259-71).<br />
The Hague: Mouton.<br />
Goldstein, B. Z., & Tamura, K. (1975). Japan and America: A comparative study in language<br />
and culture. Tokyo: Charles E. Tuttle.<br />
Kasper, G. (1990). Linguistic politeness: Current research issues. Journal of Pragmatics, 14,<br />
193-218.<br />
Ide, S. (1989). Formal forms and discernment: Two neglected aspects of universals of<br />
linguistic politeness. Multilingua, 8, 223-48.<br />
Lakoff, R. (1973). The logic of politeness; or minding your p's and q's. Papers from the Ninth<br />
Regional Meeting of the Chicago Linguistic Society (292-305).<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
LoCastro, V. (1986, November) / agree with you, but . . . Paper presented at JALT '86<br />
Conference, Hamarnatsu, Japan.<br />
Loveday, L. (1982). The sociolinguistics of learning and using a non-native language.<br />
Oxford: Pergamon.<br />
Matsumoto, Y. (1989). Politeness and conversational universals—observations from Japanese.<br />
Multilingua, 8, 207-21.<br />
Nakane, C. (1970). Japanese society. Berkeley and Los Angeles: University of California<br />
Press.<br />
Nomura, N., & Barnlund, D. (1983). Patterns of interpersonal criticism in Japan and the<br />
United States. International Journal of Intercultural Relations, 7, 1-18.<br />
Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In<br />
S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49).<br />
Rowley, MA: Newbury House.<br />
Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd<br />
(Eds.), Sociolinguistics and language acquisition (18-35). Rowley, MA: Newbury<br />
House.<br />
Owen, M. (1980). Apologies and remedial interchanges. The Hague: Mouton.<br />
Sakamoto, N., & Naotsuka, R. (1982). Polite fictions: Why Japanese and Americans seem<br />
rude to each other. Tokyo: Kinseido.<br />
Schmidt, R. W. (1977). Sociolinguistic variation and language transfer in phonology. Working<br />
Papers on Bilingualism, 12, 79-95.<br />
Takahashi, T., & Beebe, L. (1986). ESL teachers' evaluation of pragmatic vs. grammatical<br />
errors. CUNY Forum, 12, 172-203.<br />
Takahashi, T., & Beebe, L. (1987). The development of pragmatic competence by Japanese<br />
learners of English. JALT Journal, 8, 131-55.<br />
Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.
This page intentionally left blank
Ill<br />
DISCOURSE PERSPECTIVES<br />
The chapters in this section expand the perspective of ILP into different directions.<br />
Their common denominator is their focus on interlanguage discourse, rather than on<br />
individual speech acts. As Aston comments in Chapter 11, speech-act-oriented ILP<br />
is characterized by an intra-organism view and a focus on interactional (as opposed<br />
to transactional) goals. The authors of the following chapters have adopted an interorganism<br />
view, with varying emphasis on transactional and interactional discourse<br />
functions.<br />
House (Chapter 8) examines inappropriate responses in learners' conversations<br />
with native speakers. In the second language literature so far, miscommunication in<br />
native-nonnative (and nonnative-nonnative) discourse has primarily been addressed<br />
in research on repair and interactional modification. In this research tradition,<br />
emphasis is given to the interactional treatment of communication problems by the<br />
interlocutors, rather than to locating their contributing sources. House offers a new<br />
perspective on the topic by attempting to account for the complexity of responding<br />
behavior in both native and nonnative talk. Her processes-oriented analysis yields a<br />
threefold classification of problem sources: language-based difficulties, conceptual<br />
and strategic deficiencies related to gaps in pragmatic knowledge, and operational<br />
difficulties, which may lead to inputs being virtually ignored. Whereas the first two<br />
problem sources are likely to typify learners more than native speakers, type 3<br />
problems may well surface in native speaker talk as well. Future research in this<br />
area will have to take into account, as House does, the processing constraints<br />
operating in learners' decoding and encoding of messages in conversational interaction.<br />
Zuengler (Chapter 9) provides a comprehensive review of interlanguage studies<br />
concerned with the effect of conversational topic on learners' variable performance.<br />
The main findings emerging from work in this area are that both expertise in the<br />
topic and affective involvement influence interlocutors' role dynamics and conversational<br />
behavior. Most studies indicate that superior topic expertise, whether real or<br />
subjectively perceived, results in more control of the discourse by the expert participant,<br />
regardless of native or nonnative speaker status. However, expertise may be<br />
neutralized by affective factors such as emotional involvement and test anxiety. One<br />
of the important implications of interlanguage studies on topic is the emphasis on<br />
factors internal to the speech event (rather than adopting the traditional focus on<br />
event-external factors such as learners' age, LI, etc.). This work could thus be the<br />
159
160 Discourse Perspectives<br />
first step toward full ethnographies of communication in the study of interlanguage<br />
variation, with the speech event as the unit of analysis. This development would<br />
imply the inclusion of other event-based parameters of variability, such as the nature<br />
of the setting, relationship between the participants, and message content. Presumably<br />
the quality of learners' speech and conversational control, at least in the case of<br />
advanced nonnative speakers, is subject to complex interactions between eventinternal<br />
and -external factors which are still not well understood.<br />
Blum-Kulka and Sheffer (Chapter 10) expand the ILP research agenda by considering<br />
the potential for ILP phenomena in a first language. The population examined<br />
are American-born longtime immigrants to Israel, who maintain English as the<br />
language of the home. The metapragmatic discourse of these families in their first<br />
language is shown to manifest an intercultural pattern that differs in its style from<br />
both Israeli and American discourse. Traditionally, shifts in LI use among immigrants<br />
would be considered a case of attrition. This study suggests that we can gain<br />
new insights into ILP by placing such phenomena in the classical interlanguage<br />
study paradigm, comparing the first language version spoken by the immigrants to<br />
native speech in both relevant contact cultures. Consequently, in the approach<br />
offered here, ILP phenomena are by no means restricted to second language varieties.<br />
We have chosen Aston's contribution to be the final chapter to this volume<br />
because it presents a more optimistic picture of interlanguage discourse than is<br />
usually painted in ILP. Rather than focusing on pragmatic failure and other sorts of<br />
miscommunication, Aston's concern is with the negotiation of comity, the establishment<br />
and maintenance of friendly relationships in native and native-nonnative talk<br />
exchanges. In order to achieve such interactional goals, participants minimally need<br />
to find common ground for sharing attitudes toward features of the world ("solidarity")<br />
or for showing affiliation with the other ("support"). Applying conversation<br />
analytic methods, Aston carefully demonstrates how the conversational resources<br />
drawn on to establish rapport in native-native discourse are transformed into a<br />
different set of resources in native-nonnative interaction. The important implication<br />
of Aston's work, both for theory construction in ILP and second language pedagogy,<br />
is to extend to ILP the fundamental tenet of interlanguage studies, namely, the<br />
autonomy of learner language. Instead of viewing interlanguage pragmatic knowledge<br />
and behavior as deficient in terms of native norms, we need to consider its<br />
functionality and inner justification.
8<br />
Toward a Model for the Analysis<br />
of Inappropriate Responses in<br />
Native/Nonnative Interactions<br />
JULIANE HOUSE<br />
Why do learners of a second or foreign language often fail to respond in a way<br />
acceptable to their native interlocutors Is it because they "misheard" or simply<br />
didn't hear what was said to them, because they did not understand what was said to<br />
them, or because they did not understand what the interlocutor had meant by what<br />
he or she said to them<br />
In trying to find possible answers to the question of what may go wrong in a<br />
learner response, one notices immediately that response failures stem from many<br />
different and possibly interacting sources: they can be a result of inadequate perception,<br />
of inappropriate comprehension at the syntactic, semantic, pragmatic, and<br />
discourse levels of language or of gaps in the learner's knowledge of the world; they<br />
can result from uncooperativeness on the part of the responder, who in this case may<br />
have understood perfectly well what the interlocutor had tried to communicate but<br />
just wanted to be difficult; or the responder may have perceived, and comprehended<br />
correctly, and intended to cooperate, but cannot manage a response which is, at this<br />
stage of the discourse, expected from his interlocutor.<br />
Given the difficulty of coming to grips with interlanguage responses, it is<br />
understandable that attempts at investigating them are relatively rare (notable exceptions<br />
are Keller-Cohen, 1979; Kasper, 1981, 1984; Thomas, 1983). What is needed<br />
in dealing with deviant learner responses and misunderstandings in interactions<br />
between learners and native speakers is a model comprehensive and powerful<br />
enough to take account of the complexity and ambiguity of the causes of response<br />
failure.<br />
In this chapter I shall (1) set the framework for the eventual development of such<br />
a model by reviewing various theoretical and empirical approaches and their relevance<br />
for explaining the difficulties learners face when responding; (2) attempt to<br />
integrate some of the assumptions and concepts found useful in these approaches<br />
into a discourse-processing model; and (3) illustrate the explanatory power of this<br />
161
162 Discourse Perspectives<br />
model by demonstrating how it can account for a set of exemplary misreponses<br />
taken from role-play-elicited discourse data.<br />
Approaches to Analyzing Responding Behavior<br />
The following research strands are relevant in addressing the issue of interlanguage<br />
responses: (1) social views of language, (2) intercultural miscommunication,<br />
(3) studies of misunderstanding based on pragmatic theories, (4) the psychopathology<br />
of everyday discourse, (5) cognitive-discourse-processing models.<br />
Social Views of Language<br />
Language can be viewed from two perspectives: (a) an intra-organism one, where<br />
the focus is on what goes on in language users' minds; and (b) an inter-organism<br />
(social) one, where what goes on between two or more language users matters most<br />
(Halliday, 1978, 56). Looking upon language as an essentially social phenomenon<br />
has had a long tradition in Europe and North America that may be traced back to<br />
linguists of the Prague school and their work on language functions, functional<br />
styles, and functional sentence perspectives; and to British contextualists such as<br />
Firth and later Halliday, who, influenced by the anthropologist Malinowski, emphasized<br />
the importance of "the context of situation" in establishing meanings in talk.<br />
Firth (1964) suggests, for example, that "the give and take of talk resembles a preordained<br />
ritual . . . the moment a conversation is started whatever is said is a<br />
determining condition for what in any reasonable condition may follow" (94).<br />
In North America, a similar tradition of stressing the embeddedness of language<br />
in context or culture exists in the work of anthropologists and linguists such as Boas,<br />
Sapir, and Whorf, and later sociologists of language and sociolinguists such as<br />
Cicourel, Labov, and Goffman as well as that group of sociologists around Harold<br />
Garfinkel who came to be known as the "ethnomethodologists." On the ethnomethodological<br />
view, it is less useful to judge a response according to whether it<br />
matches the previous speaker's unobservable intentional state; rather, the main<br />
concern is to look at how observed and contextualized stretches of discourse form<br />
coherent wholes. The notion of sequentially organized implications as manifest in<br />
tied pairs or adjacency pairs is relevant here (Schegloff & Sacks, 1973); that is,<br />
when the first part of such a couplet is produced, it sequentially implicates the<br />
production of the second part of the pair.<br />
The close-order organization of talk as evident in the tying of utterances in<br />
adjacency pairs thus enables discourse participants to monitor their conversation<br />
jointly such that the achievement of "understanding" (or "misunderstanding") can<br />
be seen as displayed through the positioning of elements of talk and actions (Benson<br />
& Hughes, 1983, 179); any response is to be seen as "a kind of claim as to what was<br />
meant by the previous utterance" (Bilmes, 1990).<br />
As Goffman (1981, 12) pointed out, one major problem with this view is that the<br />
burden of defining a response is simply thrown back onto the previous speaker's<br />
initiating statement, and from there reference is made back to another initiating
Native I Normative Interactions 163<br />
statement preceding it, and so on; it is thus difficult to know where to start or how to<br />
find proof of the appropriateness of a given response in itself. Another difficulty<br />
with this view is that it lends itself best to a fairly formalistic, micro-level analysis,<br />
which does not take the larger co-text into account.<br />
In his own attempt to define a response, Goffman (1981, 35) states that a<br />
response, which emanates from an individual and is inspired by a prior speaker, tells<br />
us something about that individual's position or alignment in what is occurring. In<br />
the context of an ongoing discourse, "alignment" implies, first, that two interactional<br />
moves are brought into something resembling agreement or partnership with one<br />
another; second, that the response is of local significance, that is, relevant and<br />
comprehensible at the moment of speaking.<br />
Goffman (1981) singles out one type of response which he calls a "reply", that<br />
is, "a response in which the alignment implied and the object to which reference is<br />
made are both conveyed through words or their substitutes" (35). Replies, then,<br />
reach more directly and verbally backwards to the immediately preceding move,<br />
ostensibly achieving cohesion through overt surface connections on the utteranceexpression<br />
level (examples are the ritualized reciprocal pairs in phatic phases of<br />
talk), while responses are often less "tidy" in their back-referencing; they do not<br />
necessarily bear on the immediately preceding surface expression but rather are<br />
coherent on account of a variety of different interpretation schemata available to<br />
speakers. 1 In simplistic terms, if we accept Goffman's distinction, but wish to<br />
consider both "replies" and "responses" in his technical sense, then we are obliged<br />
to investigate under the notion of "responding" (in the nontechnical sense) both the<br />
linguistic links between segments of talk and the interpretation schemata by means<br />
of which talk is achieved.<br />
Intercultural Miscommunication<br />
There is a rich literature on misunderstandings occurring in interactions between<br />
members of different cultures. Gumperz (1982a, b) found that misunderstandings<br />
often result from participants misusing "contextualization cues," that is, aspects of<br />
the surface forms of utterances (prosodic, phonological, lexical choices, routine<br />
formulas that may signal relevant interpretive frames). In Gumperz's view, it is such<br />
(often culture-specific) linguistic cues that are crucial for processes of conversational<br />
inferencing, that is, "the 'situated' or context-bound process of interpretation, by<br />
means of which participants in a conversation assess others' intentions and on which<br />
they base their responses" (1982a, 153).<br />
Tannen (1984a,b, 1986), who has studied both native-native and nativenonnative<br />
talk, explains deviant responses and misunderstandings as a "matter of<br />
framing," of exploring frames, breaking frames, or reframing in Batesort's (1972)<br />
sense. Frames are dynamically linked as an "interpretation and responding chain<br />
where negotiated footings are continually changed." Tannen (1984b) lists eight<br />
levels of differences in the ways speakers signal what they mean: when to talk, what<br />
to say, pacing and pairing, intonation, formularity, indirectness, cohesion, and<br />
coherence—and these eight may lead to differential ways in which conversational<br />
partners tend to assess others' intentions as a basis for making their responses.
164 Discourse Perspectives<br />
Along these eight levels, cross-cultural as well as intra-cultural misunderstandings<br />
can arise.<br />
Research has been done on some of these eight areas. Erickson (1975) and<br />
Erickson and Shultz (1982) analyzed misunderstandings in multiracial consulting<br />
interviews, where they described how subtle differences in interactional style across<br />
racial and ethnic groups can lead to misinterpreted responses. Kochman (1981)<br />
examined black and white interaction styles. Scollon and Scollon (1981), in their<br />
work on Athabaskan Indians, revealed several culturally conditioned differences,<br />
especially in face-presentation and the distribution of talk and silence, where differences<br />
in the lengths of silences tended to be interpreted as challenges or disagreements<br />
by members of the other culture.<br />
Work on communicative style (Lakoff, 1976), politeness phenomena (Brown &<br />
Levinson, 1987), Gricean Implicature (Grice, 1975; Smith, 1982; Sperber & Wilson,<br />
1986) as well as the interpretation of indirect speech acts (e.g., Clark, 1979;<br />
Clark & Schunk, 1980; Clark & Lucy, 1975) have resulted in hypotheses suggesting<br />
that "normal talk" is "indirect"; that is, if they want to respond appropriately,<br />
speakers need to infer their partners' intentions in order to find out how the words<br />
used were "really meant." Indirectness lies at the heart of many if not most misresponses<br />
and misunderstandings in talk, and such alignment failures are of course<br />
much more likely to occur in talk between people from different cultural backgrounds,<br />
where indirectness and politeness conventions often diverge (Blum-Kulka,<br />
1987; House, 1986).<br />
Pragmatic-Theory-Based Studies of Misunderstandings<br />
The third research strand is pragmatic-theory-based and is represented by the work<br />
of Dascal (1977, 1985); Dascal and Katriel (1979), Dascal and Idan (1988), Blum-<br />
Kulka and Weizman (1988) and Weizman and Blum-Kulka (1988). In this line of<br />
inquiry, misunderstandings in "normal talk" between native speakers are analyzed at<br />
different levels of pragmatic meaning. Dascal (1985) introduced the notion of a<br />
"conversational demand" set up by an utterance, the perception of which may differ<br />
for interlocutors. What exactly can go wrong in the perception of a conversational<br />
demand can be described by reference to four questions that need to be answered in<br />
interpreting what speaker S has said (Fillmore, 1976, 78):<br />
1. What did S say (proposition)<br />
2. What was S talking about (what was said plus what was implicated, that is,<br />
the extended semantic meaning)<br />
3. Why did S bother to say it (illocution)<br />
4. Why did 5" say it in the way he said it (key, tone).<br />
On any one of these layers of significance and any combinations thereof, misunderstandings<br />
can arise. 2 Weizman and Blum-Kulka (1988) made a similar distinction of<br />
the potential causes of misunderstandings in terms of the levels of prepositional<br />
content, illocutionary point and mode, and, following Searle (1983, 1992), they<br />
also differentiate between an "individual-I" (a speaker's) meaning and a "collectivewe"<br />
direction pertaining to the three levels along which communication can break<br />
down.
Native I Normative Interactions 165<br />
Psychopathological Studies of Misunderstandings<br />
The fourth approach relevant to construing a theory for the analysis of misunderstandings<br />
is the psychopathological approach, which can be traced back to Freud's<br />
intuitions about the "psychopathology of everyday life," and in particular, slips of<br />
the tongue. This approach is linked to the work of Langer and her associates (e.g.,<br />
Langer, 1978; Langer, Blank, & Chanowitz, 1978; Langer & Imber, 1980), Reason<br />
and Mycielska (1982) as well as Heikkinen and Valo (1985), who related the<br />
findings in this paradigm to interlanguage responses, and most recently Heckhausen<br />
and Beckmann's (1990) work on action slips. The central points made in connection<br />
with this approach are that much of what is traditionally conceived of as "ostensibly<br />
thoughtful action" is in fact characterized by "mindlessness" (Langer, Blank, &<br />
Chanowitz, 1978), and that it is especially in overlearned behaviour, that is, routinized<br />
and automatized actions, that "slips" occur (cf. also Goffman's [1959]<br />
distinction between "knows-better" slips and "does-not-know-better" errors).<br />
Langer's "Non-Thinking Hypothesis" has been given indirect support through<br />
the work by Langer and Abelson (1972) and Abelson (1976) and the many ensuing<br />
studies on the packaging of knowledge in the mind, suggesting that in many types of<br />
behaviour actants rely on scripts to reduce cognitive work, effort, and time. Because<br />
of the imprintedness of larger plans, schemata, scripts, or social episodes in human<br />
memory, we are able to predict upcoming moves and consequently stop paying<br />
attention to the reality of the interlocutor's input. The "illusion of control" is often<br />
only disrupted by the occurrence of "slips," the recognition of which is then equivalent<br />
to a sudden intrusion of reality.<br />
In the case of language learners, slips tend to occur for a different reason:<br />
learners have not yet fully developed and automatized scripts at their disposal.<br />
Because learners often "rehearse" internally what will be their next move, they also<br />
often neglect any real input (Heikkinen & Valo, 1985).<br />
Information-Processing Approaches to Language Understanding<br />
The information-processing approach is closely linked to the psychopathological<br />
one in that here, too, an assumption of scripted behavior is made, in which cognitive<br />
schemata, frames, or scripts are representations of repeated behavioral patterns<br />
designed to reduce cognitive work. The basic idea of organizing knowledge in<br />
packets that function as ready-made subsets goes back to Bartlett's (1932) conception<br />
of human memory as consisting of high-level structures, or schemata, encapsulating<br />
knowledge about particular objects, events, and situations. Such knowledge<br />
structures have also been referred to as frames, that is, networks with slots<br />
which can be filled with certain values (Minsky, 1975), and scripts, which represent<br />
sequences of routine actions, such as the restaurant script (Schank & Abelson,<br />
1977).<br />
Cognitive models that explain the processes operative in discourse comprehension<br />
(Kintsch & van Dijk, 1978; van Dijk & Kintsch, 1983; Kintsch, 1988) are<br />
especially relevant to analyzing interlanguage responses. In the original Kintsch and<br />
van Dijk model (1978), the construction of a prepositional text base was considered<br />
to be the major component in comprehension. In their later (van Dijk & Kintsch,
166 Discourse Perspectives<br />
1983) version, a "situational component" was added. Strategical behavior was<br />
considered to be the driving force in the construction of a propositional text base,<br />
discourse comprehension being regarded as basically a strategic process. In a later<br />
attempt to combine production systems with a connectionist approach, Kintsch<br />
(1988) developed a "construction-integration model" of discourse understanding,<br />
claiming that initially comprehension is bottom-up-guided and it is only in a later<br />
integration process that appropriate meaning is selected. A cluster of propositions is<br />
derived from the discourse via activation of the closest neighbors of the original<br />
text-derived propositions in the general knowledge net. Kintsch assumes a minimally<br />
organized and structured knowledge system, with specific structures being<br />
generated in the context of the task for which they are needed and resulting in dense<br />
associative knowledge nets.<br />
The strongest arguments for a cognitive-processing-model approach to the explication<br />
of inappropriate responses is that such an approach is (potentially) explicit,<br />
and may integrate the different insightful perspectives sketched above. Social perspectives<br />
may be encapsulated as such rules, norms, and conventions must have<br />
psychocognitive correlates inside the individual if they affect conversational behavior.<br />
Further, the very notion of inappropriateness presupposes at least two cognitive<br />
systems in interaction: in principle, an information-processing model should enable<br />
the identification of contextualized misunderstandings by explicating processing<br />
"states" of the interacting systems.<br />
Toward a Model for Analyzing Responses<br />
In building on discourse-processing approaches as sketched above, Edmondson<br />
(1989) set up a discourse comprehension and production model, which I have<br />
adapted and outlined in Figure 8.1. I shall briefly describe this model and then<br />
illustrate its explanatory power by applying it to the analysis of discourse data. As<br />
can be seen, discourse comprehension can hardly be considered independently of<br />
discourse production, especially in the case of dyadic spoken face-to-face discourse.<br />
The model operates on two "levels": a prelinguistic or conceptual level<br />
(made up of the upper boxes A, B, C, F, G, and H) and a lower linguistic level<br />
(consisting of the lower boxes D, E, and I) on which the linguistic decoding of the<br />
input and the encoding of the output occurs. These levels are networked in rather<br />
more complex ways than Figure 8.1 suggests. In briefly elaborating upon the<br />
subcomponents inside this model, I shall restrict myself to characterizations relevant<br />
to my purposes.<br />
Box A: The Operant Discourse Frame<br />
The assumption here is that a copy of some relevant stored situational constellation<br />
or schema is made and "filled in" by knowledge derived from the current discourse<br />
situation. The result is a complex of conceptual and linguistic representations: the<br />
operant discourse frame. The term frame (cf. Bateson, 1972; Goffman, 1974;<br />
Tannen, 1979 and above) is being used here as distinct from schema, which pro-
Native I Normative Interactions 167<br />
Fig. 8.1. Schematic discourse processing model.<br />
vides the skeleton that is fleshed out in the ongoing discourse. In my terminology,<br />
then, a schema is built up into a discourse frame. In ongoing discourse, however,<br />
further or alternative schemata can be called up, leading to revision, switching,<br />
reactivation or indeed coexistence of operant discourse frames (see the notion of<br />
coexistent discourse worlds in Edmondson, 1981). The following elements are<br />
contained in a currently activated discourse frame:<br />
1. Knowledge of the currently relevant discourse topic<br />
2. Knowledge of the currently relevant interactional move<br />
3. Representation of the discourse outcomes resulting from the preceding interaction<br />
4. A primarily linguistic representation of the prepositional content of the last<br />
turn held in short-term memory, which can be accessed from the currently<br />
active frame.<br />
In terms of the distinctions made by Dascal (1977, 1985) and Blum-Kulka and<br />
Weizman (1988) it is the level of aboutness (topic, proposition) which is captured in<br />
"operant discourse frames".<br />
Box B: Interactional Goals<br />
It is assumed that the speaker has or develops a goal for the ongoing encounter,<br />
together with relevant subgoals, which develop as the discourse unfolds. There is
168 Discourse Perspectives<br />
clearly interaction between boxes A and B, in that what has happened so far may<br />
influence what the speaker seeks to get out of the encounter. Similarly the goals of<br />
the speaker system will affect what form the "nonlinguistic reaction" represented in<br />
box F may take, just as long-term goals will impinge on strategic manipulation (box<br />
G). For instance, in the case of the elicited role-play data to be analyzed below, the<br />
role specifications in part predetermine what are here called interactional goals, just<br />
as these role specifications (in part) preselect a particular schema, that is, a set of<br />
expectations concerning what type of discourse will take place. In terms of the<br />
distinctions made by Weizman and Blum-Kulka (1988), we may here include two<br />
types of goals in an interaction: the individual-I point and the collective-we direction/purpose<br />
of the encounter (although it is often difficult to disentangle the two).<br />
Box C: Current Prediction/Expectation<br />
This simply marks a result of ongoing processing: the system expects a more or less<br />
specific type of input.<br />
Box D: Decode<br />
The process of extracting a "discourse meaning" may involve linguistic decoding<br />
(itself a complex of subsystems), discourse interpretation strategies, and various<br />
inferencing procedures as suggested by Gumperz, Tannen, and others. The strategic<br />
knowledge that may be activated in box G may also be called up in this decoding<br />
process.<br />
Box E: Discourse Meaning<br />
As an outcome of the decoding processes in D, a discourse meaning for the current<br />
input is arrived at.<br />
Box F: Nonlinguistic Reaction<br />
An interpreted input leads initially to a nonlinguistic response, a sort of "gut reaction":<br />
the determinants of this response are in large measure personality-based, and<br />
not, strictly speaking, part of a discourse system. In Edmondson (1987), some<br />
"reactive formatives" are posited, and a distinction is made between an interactional<br />
option (which basically concerns whether the hearer/speaker will take a positive or<br />
a negative attitude towards the current discourse move made by the interlocutor),<br />
and an illocutionary option concerning the category of speaker-meaning appropriate<br />
to this purpose.<br />
Box G: Choice of Discourse Strategies<br />
In this phase of the system, the hearer/speaker calls upon strategic or procedural<br />
knowledge, which may lead to the employment of interactional strategies designed<br />
to disguise the reaction in F above. The hearer/speaker may, for instance, decide to
Native I Normative Interactions 169<br />
delay but at the same time anticipate the raw reaction stored in F, or may cloak it via<br />
various politeness strategies. In this phase, which may be likened to a filtering<br />
instance that serves to modulate speech acts under the constraints of politeness<br />
maxims, potential responses of the interlocutor are taken into account, in the anticipation<br />
of which speakers adjust their initial reactions.<br />
Box H: Immediate Communicative Plan<br />
At this junction in the system a prelinguistic communicative plan has resulted,<br />
the scope of which may range from small turn-internal units to longer discourse<br />
stretches, in which case elements of such a plan will be transferred to box B for later<br />
recall.<br />
Box I: Linguistic Encoding<br />
It is at this level that the communicative plan arrived at in H is transformed into<br />
linguistic representation. Box ] then contains at the very least a complex of linguistic<br />
resources, search strategies, prepacked chunks such as routine formulas, and<br />
gambits, acting as discourse lubricants. If linguistic searching is unsuccessful (i.e.,<br />
the communicative plan established in H cannot be realized linguistically), backlooping<br />
may occur, leading, for example, to "reduction strategies" (Kasper 1982;<br />
Faerch & Kasper, 1980, 1983) with which speakers attempt to adjust a communicative<br />
plan in light of the insurmountable complexity of the linguistic realization<br />
necessary to carry it out.<br />
This model does not aspire to completeness, focusing instead on what appear to<br />
be critical points of processing from the learner's perspective. While the pathway<br />
mapped in Figure 8.1, and the numbering adopted for expository purposes, suggest<br />
a straightforward linear-processing procedure, the "boxes" are interrelated and the<br />
whole system operates in parallel, that is, various nodes, items, and paths are<br />
activated at the same time. Further, various shortcuts exist in a fully developed<br />
system (see Edmondson, 1987 on the nature and function of routines). The following<br />
analyses suggest several refinements to the model, in terms of further paths<br />
through the subsystems.<br />
Testing the Model<br />
First of all, we have to clarify what exactly is to be understood by an inappropriate<br />
response. Following Goffman's (1981, 35) characterization of a response as a move<br />
inspired by a prior speaker that is (a) aligned to what is "occurring now" and (b) to<br />
be understood as "relevant now," we can say that an inappropriate response is a<br />
response interpreted not to be in alignment with the preceding discourse and is thus<br />
taken not to be relevant at the moment of speaking. However insightful, such a<br />
definition is not easily operationalized. It will, however, suffice here.<br />
The model's operation will be demonstrated in an analysis of conversational<br />
responses. The data were elicited via role-plays from advanced German learners of
170 Discourse Perspectives<br />
English interacting with English native speakers in a variety of simulated everyday<br />
situations. The dialogues included both immediate contact and telephone conversations,<br />
each requiring centrally a specific speech act such as a request, a complaint,<br />
or a suggestion. Interactants were given brief, oral, and "open" situational instructions,<br />
with roles being varied along the dimensions of dominance and social distance.<br />
3<br />
In cases of inappropriate responding, a retrospective interview was conducted.<br />
In these interviews, subjects listened to a playback of their recorded conversation<br />
and, at the same time, studied the transcription. With respect to the critical (mis)response,<br />
they were then asked to reflect upon how they had "understood" their<br />
interlocutor's ("response-inspiring") utterance and what they (thought they) had<br />
intended with their own ensuing utterance. These retrospective interviews were not<br />
consistently taped as it often transpired that subjects felt much freer to do serious indepth<br />
probing of their own behavior when they were not recorded. Respondents<br />
frequently pointed out that in the retrospective sessions they were (actually often<br />
painfully) confronted with their own "failure," an experience absent in the original<br />
role-play session (the recording of which was never felt as intrusive).<br />
In analyzing the data, I have used the discourse analytical categories provided in<br />
Edmondson's (1981) discourse model and in Edmondson and House's (1981) Interactional<br />
Grammar of English. Instances of inappropriate learner responses will be<br />
related to the Discourse Processing Model outlined above, and to concepts, methods,<br />
and findings adopted from the five research strands described above. At the<br />
same time, detailed analysis, even if at times rather speculative, suggests quite<br />
strongly that things are much more complex than indicated by the model.<br />
The analysis of the interlanguage data examined revealed that there are three<br />
sources for the pragmatic-responding errors found:<br />
1. language-based difficulties, focusing on problems that are apparently located<br />
in linguistic decoding and encoding<br />
2. conceptual and strategic deficiencies, which may result in learners choosing<br />
inappropriate or inadequately mitigated speech acts<br />
3. operational difficulties, which may lead to an input apparently being ignored.<br />
It is conceded immediately that these three distinctions are decidedly fuzzy-edged.<br />
Indeed, it will transpire that taking a discourse-processing perspective on learner<br />
responding behaviors makes neat categorization of responses a difficult if not rather<br />
artificial undertaking.<br />
Language-Based Responding Deficiencies<br />
ENCODING DIFFICULTIES<br />
The most predictable and most common learner response difficulties relate to the<br />
apparent inability to find and employ routine formulas and gambits appropriately<br />
(box I). The following are examples of this type of responding deficit:
Situation 1: Invitation to a Party/Opening Phase<br />
Native I Normative Interactions 171<br />
NS: Hi, how are you<br />
NNS: Oh, just fine, what about yourself<br />
NS: Pretty good, what have you been up to lately<br />
NNS: Well, I worked at home for University, I was really busy.<br />
NS: And what have you been doing nights<br />
NNS: As I told you, I didn't go out much, usually I stay at home and work.<br />
In commenting on the last move in this discourse stretch, the learner reported that<br />
she had intended to express the translation equivalent of the German phrase wie<br />
gesagt, which in German is a rather meaningless formula used to provide coherence<br />
with preceding discourse stretches. In the context above, the English phrase as I told<br />
you carries, however, a rather petulant and needlessly aggressive overtone, which<br />
the learner certainly did not intend to imply. An expression such as "As I say, I've<br />
been working so ..." would have been more appropriate in this context.<br />
Learners admit that they infrequently model their utterances on supposed German<br />
translation equivalents. In situation 2 below, the learner had intended to produce<br />
a final Thanks, but through the inappropriate inclusion of the downtoner<br />
anyway, she produced a distinct "misfire": anyway collocates exclusively with<br />
Thanks in a context in which a Request has not been satisfactorily met, its semantic<br />
meaning being "in spite of everything."<br />
Situation 2: Request for Money/Closing Phase<br />
NS: Just leave me your name and address; I'll write it down and then give you the<br />
money.<br />
NNS: Okay, that's kind of you. Thank you anyway.<br />
In the retrospective interview, the learner indicated that she had considered the token<br />
anyway to be a convenient filler corresponding to the German all-purpose gambit<br />
also, which would be perfectly acceptable in an equivalent German interchange.<br />
In situation 3 below, the learner fails to reply appropriately to a typically vague<br />
and essentially ritual suggestion made by a native speaker in the (phatic) opening<br />
phase of an interaction, thus violating the politeness principle of "leaving options"<br />
(Lakoff 1973).<br />
Situation 3: Phatic Talk/Neighbors<br />
NNS: Hello. I'm your new neighbor. I'd like to visit you because I've just moved in and<br />
don't know you.<br />
NS: Oh, how nice! Well, we'll have to get to know each other.<br />
NNS: Yes, 1 think so as well. Perhaps you should drink a cup of tea.<br />
In this context an appropriate response would have been a vague reciprocating<br />
phrase such as "maybe you'd like to come over for a cup of tea sometime." The<br />
learner explained that she had meant to say something equivalent to the German "ja<br />
naturlich, wir sollten mal auf ein TaBchen Kaffee zusammen kommen," which does<br />
have the casual, nonimposing overtone the supposed English equivalent totally
172 Discourse Perspectives<br />
lacks because of the modal should. The learner initiation is, of course, also lacking<br />
in appropriateness.<br />
In situation 4 below the learner again fails to respond appropriately inside a<br />
phatic phase of an interaction because she lacks the appropriate tokens. In this case,<br />
unsuccessful transfer from the mother tongue was not the culprit; in fact, the token<br />
"Okay," which would have been correct in the following exchange, is very common<br />
in German as well and fulfils the same functions:<br />
Situation 4: Request to Hand Over Lost Property/Closing Phase<br />
NS: What number is your house number<br />
NNS: Eleven Eleven.<br />
NS: Number 11, okay, about two o'clock then.<br />
NNS: Yeah, that's right.<br />
The phrase "that's right" used by the learner in her final move implies a personal<br />
evaluation of the truth of the interlocutor's preceding utterance. What the learner<br />
had meant to realize in this final discourse slot was a ritual move sealing the<br />
agreement reached before, that is, the realization of the ritual speech act Okay<br />
(Edmondson & House, 1981, 60), for which conventional tokens such as okay or<br />
fine would suit. The learner had realized that such an act was called for at that<br />
final stage in the interaction but did not have the appropriate tokens at her disposal.<br />
The next example reveals learners' inability to produce discourse lubricating<br />
devices (gambits) such as Uptakers or Starters in a conventionally accepted way.<br />
Indeed, inappropriate responses that result from the learner's inability to employ an<br />
acceptable gambit token (e.g., the token well) are very frequent in the data. They<br />
originate, as I have written elsewhere (House 1982b; 1984), from the transfer into<br />
English of the German gambit token ja, which fulfils many different discourse<br />
lubricating functions. Here is an example:<br />
Situation 5: Request for a Lift<br />
NS: Why don't you use your own car<br />
NNS: Yes, you know, it's too dirty.<br />
DECODING DIFFICULTIES<br />
I turn now to inappropriate learner responses which seem to relate to difficulties of<br />
decoding (box D). The inappropriate learner responses resulting from decoding<br />
problems are less easily ascertained than were the ones related to the unavailability<br />
and unencodability of routines and gambits; in some cases both decoding and<br />
encoding problems seem to have been operative.<br />
In situation 6 below the learner takes the indirect request "do you mind if 1 do P"<br />
apart and responds separately (and idiosyncratically) to the two chunks "do you<br />
mind" ("no") and "if I do P" ("do it"), suggesting that she did not interpret the<br />
indirect request as an intended entity (see, however, the studies conducted by Clark<br />
and his associates, e.g., Clark and Schunk 1980, which revealed that it is indeed<br />
sometimes the literal meaning of an indirect request that native speakers respond<br />
to.)
Situation 6: Small Talk/Waiting Room<br />
Nativei'Normative Interactions 173<br />
NS: Excuse me, do you mind if I open the window It's so stuffy in here.<br />
NNS: No, do it.<br />
Besides admitting the decoding problems mentioned above, the learner reported that<br />
she had actually wanted to realize an uptaking routine formula equivalent to the<br />
German bitte schon, but was unable to do so and thus resorted to the linguistically<br />
simple imperative form. It had not struck her that the result might violate the<br />
politeness principle.<br />
In situation 7 below, the learner was unable to interpret the discourse stretch<br />
"what was your name, again" correctly; that is, as a polite request for information.<br />
The use of the past tense in particular led to interpretive uncertainty, so the learner<br />
fell back on a simplistic echoic strategy, providing an amusing type of "reply"<br />
(Goffman, 1981). When the native speaker made a joke of the learner's inappropriate<br />
response, the learner might have sought to do likewise, but lacked or at least was<br />
unable to access the set token "I think so," producing the contextually puzzling<br />
response "I think," a not-uncommon mistake for Germans learning English:<br />
Situation 7: Gas Company/ Complaint<br />
NS: Well, let me look at my list. What was your name, again<br />
NNS: It was Brown.<br />
NS: It was Brown. Is it still Brown<br />
NNS: I think.<br />
NS: Well, I do have a Brown on my list.<br />
In situation 8, the learner was unable to use an appropriate discourse interpretation<br />
strategy for the formula "May I help you" (routinely employed in service<br />
encounters), and went on to produce a reply in Goffman's (1981) sense, which in<br />
part reproduced the native speaker's preceding utterance—a defensive habit acquired<br />
in the foreign language classroom. In following this echoic strategy, these<br />
learners produce impolite responses. It is, however, surely part of (common) strategic<br />
knowledge that the deference manifested in the first pair part of a request is<br />
totally inappropriate for use in a positive response to that request: the error, therefore,<br />
appears to be mainly one of decoding. However, a system failure in terms of<br />
strategic knowledge (see box G), or more precisely, its use in discourse interpretation,<br />
may also have been present.<br />
Situation 8: Gas Company/Complaint/Opening Phase<br />
NS: May I help you<br />
NNS: Yes, of course, you may. I want to complain.<br />
In situation 9, the learner revealed in the retrospective interview that she was not<br />
in a position to correctly interpret the routine formula "What have you been doing<br />
lately": the response is plausibly an initiation of her own interactional goals; that<br />
is, box B determines what results in box H. Note that a nonlinguistic reformulation<br />
of the interlocutor's query, roughly of the form "What are you doing when it is late"<br />
is appropriately responded to—in other words, the input is not interpreted, but it is<br />
not ignored:
174 Discourse Perspectives<br />
Situation 9: Invitation to a Party/Opening Phase<br />
NS: Hi, how are you What have you been doing lately<br />
NNS: Well, I want to go to a party tonight.<br />
An interesting example of learners' failure to use appropriate inferencing strategies<br />
(box D) necessary for arriving at a plausible discourse meaning of the interlocutor's<br />
preceding move (box E) is provided in situation 10:<br />
Situation 10: Small Talk/Elevator Got Stuck<br />
NS: I'm not trusting this elevator any more. What about you Are you Are you going<br />
to walk up or are you going to ride the rest of the way<br />
NNS: I, I, the next time I will take the stairs.<br />
My hypothesis here (and it was in fact substantiated in the interview) was that the<br />
learner was unable to derive from the NS utterance the implication that NS had<br />
stopped, or was about to stop the lift. The problem is one of inferencing. Such<br />
inferencing is in this case role-play-specific, in that whereas in real-life situations<br />
speakers make inferences from the enveloping situation to the words uttered by the<br />
interlocutor, it is the other way round in a role-play situation, in which the words<br />
used can redefine the situational context (something similar can, of course, happen<br />
in telephone talk). In situation 10 the learner was unable to infer what situational<br />
redefinition the interlocutor had meant to create through his initiating move; he<br />
couldn't make sense of the utterance, hesitated, and then made a good attempt at<br />
replying, revealing that the propositional content of the input was adequately interpreted.<br />
While this instance of an inferencing problem may be role-play-specific,<br />
interpretive problems based on wrong inferences are not.<br />
Conceptual and Strategic Deficiencies<br />
In many cases, it will be a matter of speculation as to whether perceived impoliteness<br />
is a result of the responding learner's lacking the conventionalized tokens for<br />
performing a strategically modulated plan (box I), or whether the deficiency lies in<br />
the pragmatic knowledge base itself (box G). The following examples of inappropriate<br />
learner responses evidence a lack of relevant culture-specific pragmatic knowledge,<br />
resulting in violations of the politeness principle or one of its maxims (e.g.,<br />
the maxim of tact) to be observed in realizing polite utterances (e.g., Lakoff, 1980;<br />
Leach, 1983; Holmes, 1985).<br />
Situation 11: Landlady<br />
NNS: I intend to go out with a friend tonight, so I won't be home tomorrow morning for<br />
breakfast, Mrs Bennett.<br />
NS: Well, you know, Anita, I bought extra biscuits for your breakfast.<br />
NNS: Well, that doesn't matter, 'cause they will surely be fresh on Tuesday anyway.<br />
In commenting on the pretty rude response "that doesn't matter," the learner in<br />
question said that she had tried to produce a translation equivalent of the German<br />
minimizers das ist dock nicht so schlimm or das macht dock nichts, both of which<br />
carry a much more pacifying and consolatory overtone than the abrupt pseudo-
Native I Normative Interactions 175<br />
equivalent chosen. In fact, the phrase "that doesn't matter" is overrepresented in my<br />
data, and all its users proclaimed ignorance of the inherent impolite connotation.<br />
The point here is that English has no "translation equivalent": an entirely different<br />
speech act is in fact conventionally called for in English; that is, some form of<br />
apology seems to be needed rather than a minimizer as is conventionally employed<br />
in German. In other words, a cross-cultural comparison would need to go "further<br />
back" than the issue of linguistic form. Situation 12 may help establish the point:<br />
Situation 12: Complaint/Messy Kitchen<br />
NS: Look here, Anne, are you responsible for that mess in the kitchen<br />
NNS: Oh, yes, I think, but that doesn't matter.<br />
In situation 13 the learner apparently lacks the pragmatic knowledge that a polite<br />
and inoffensive countersuggestion is called for in this discourse slot, rather than a<br />
grossly impolite rejection, abruptly bringing the talk to a close:<br />
Situation 13: Party/Small Talk<br />
NS: I have never been to Greece before but I always thought it must be beautiful.<br />
NNS: Oh, yes, but there arc too many German tourists there in the summer.<br />
NS: Mmh. Yes, well, what do you think Should we go and have some cheese now<br />
NNS: Oh, no, no, no, I stay here.<br />
NS: Oh, I see (rather nonplussed)<br />
The learner in situation 13 stated that she had not intended to cut off the conversation.<br />
On the contrary, she had just wanted to express her desire to stay exactly where<br />
she was and continue chatting. The response produced is offensively ego-oriented<br />
(for discussion of a pervasive preference for self-reference and a preoccupation with<br />
self rather than the interlocutor in German native speaker talk, see House, 1979,<br />
1982a, b, c, 1984, 1989). Even "Oh, no, no, no, we stay here" would produce a<br />
totally different effect. Hence, I suggest, a deficiency in (culture-specific) pragmatic<br />
knowledge, or a case of "schematic transfer," or at least, a bypassing of such<br />
considerations in discourse production is the problem: the problem is not linguistic.<br />
A similar lack of awareness of the effect of her response is attested to in situation<br />
14. Here both gaps in pragmatic knowledge (box G), namely, that a response in the<br />
form of a (repeated) polite request is called for, and the nonavailability of appropriate<br />
linguistic tokens for making a polite request (box I) were suggested by the<br />
learner as contributing to what she later felt was an insulting response to the nurse's<br />
preceding utterance:<br />
Situation 14: Making a Last-Minute Doctor's Appointment<br />
NNS: Couldn't you make an exception for once<br />
NS: Well, I don't see any way of squeezing you in at the moment.<br />
NNS: Well, you can. Anybody can.<br />
Operational<br />
Difficulties<br />
While the types of inappropriate responses previously characterized could be related<br />
to some malfunctioning inside one or two phases ("boxes") of the proposed
176 Discourse Perspectives<br />
discourse-processing model, the following instances of deviant learner responses<br />
seem rather to be system malfunctions and can, informally, be accounted for as<br />
shortcuts, in which the learner seems to ignore the input (see the above discussion of<br />
the "psychopathological" research paradigms). Something similar, in this case<br />
based on interpretative problems, was suggested in the discussion of situation<br />
10 above, in which, it was hypothesised, interactional goals "took over" the<br />
communicative-planning task. Similar "system leaps" will be suggested in the following.<br />
General cognitive overload, insecurity, chunking problems and so on may<br />
be causes. The role-play situation itself may be an additional causative factor. The<br />
phenomenon to be explicated is familiar to nonexpert speakers of a foreign language,<br />
as when, for example, one requires so much time to rehearse one's own<br />
contribution to a group discussion that the topic has moved on before one is ready to<br />
bid for a turn. In two-party interactions, such operational difficulties may, apparently,<br />
lead to the performance of a rehearsed move, independent of its current<br />
discourse relevance.<br />
The following are examples of such nonpartner-oriented responses:<br />
Situation 15: Invitation to a Party/Opening Phase<br />
NS: Hi, Susanne, how are you doing<br />
NNS: I'd like to invite you to a party I'm doing this night.<br />
Here one might informally say "Susanne didn't listen": all she did was jump from<br />
her goal of inviting the interlocutor to an immediate communicative plan and its<br />
linguistic encoding, bypassing not only the partner's input but also any strategic<br />
considerations (box B feeds directly into H). However, it is again of interest (see the<br />
remarks offered for situation 12) that she picks up her interlocutor's token "doing."<br />
The suggestion here is that this advanced learner is perfectly familiar with the<br />
collocation "give a party": the grammatical "error" is a form of lip-service to the<br />
input received, linguistically decoded, but otherwise ignored. The following three<br />
situations are further examples of this type of self-centered shortcut:<br />
Situation 16: Lost Property/Opening Phase<br />
NS: Goood morning.<br />
NNS: Morning. This is Frank speaking. Thanks a lot for your great party.<br />
NS: Oh, I'm glad you had a good time, but to tell the truth I feel pretty icky today.<br />
NNS: So I just want to. The reason I'm calling you is I'm looking desperately for my bag.<br />
There are two obvious interpretations of the inappropriate learner move in situation<br />
16 above. One is that after the performance of a thank you, the learner feels that he<br />
has engaged himself long enough in small talk, and shifts directly from his interactional<br />
goal (B) to an immediate communicative plan (H). Alternatively, the word<br />
"icky" may have caused interpretive problems, again leading to the same fallback<br />
strategy (see situation 13 above). In the retrospective interview, the learner in<br />
question reported that both interpretations were simultaneously true in his case.<br />
Situation 17: Can't Pay Rent<br />
NS: Well, my dear, if you can't pay for this week and for next week I'm afraid I'll have<br />
to let you go.<br />
NNS: No, you know I have always paid in time before.
Native I Normative Interactions 177<br />
The learner appears to have been working on the additional argument in his favor<br />
that is brought forward, and under stress simply reacts negatively to the proposition<br />
"I shall have to let you go," before producing it.<br />
Situation 18: Can't Pay Rent<br />
NS: Can't you get a loan<br />
NNS: Well, maybe, you can wait till next week because I have an interview then.<br />
This data segment is similar (note that the same stressful situation holds). In both<br />
situations 17 and 18 the actual input is not taken into account by the learner, who<br />
seems to be altogether too busy planning and producing reasons why she shouldn't<br />
be thrown out. Such a phase-shift interpretation is in line with the subjects' own<br />
diagnosis of what went wrong: both learners admitted that they were too busy<br />
mapping out their argumentative plan to be able to listen, let alone respond to the<br />
more "local" utterances produced by their interlocutors.<br />
The response errors to be characterized in what follows are similar to the<br />
"interactional slips" described above: learners anticipate and predict a next move on<br />
the part of the interlocutor, and respond to it, even though it does not occur. The<br />
actually occurring input is then ignored. In terms of our model, the discourse<br />
meaning (box E) of an input is derived directly from the hearer/speaker's current<br />
predictions/expectations (box C). The system "hears" what it wants to hear (the<br />
same phenomenon occurs in native-speaker discourse, and is familiar in written<br />
discourse when one proofreads one's own manuscripts).<br />
Situation 19: Lost Property/Opening Phase<br />
NS: Oh, hello, Mike.<br />
NNS: Oh, fine. You<br />
NS: I'm fine (puzzled pause).<br />
Quite clearly, the student responds to a "How-are-you" move on the part of her<br />
interlocutor, even though it is not there. In fact, "how are you" was a topic of<br />
discussion in several of the communication courses the learner-subjects had attended;<br />
in other words, this learner may have been primed by foreign language<br />
instruction to expect this input. This does not to my mind invalidate the point being<br />
made—that expectation can determine interpretation—but rather reinforces it.<br />
Situation 20: Lost Property/Opening Phase<br />
NS: Hi, Susie, how are you<br />
NNS: Good morning, Kim, I'm fine. How are you<br />
NS: Oh, my head's terribly fuzzy.<br />
NNS: Oh, I'm glad to hear that.<br />
In this situation the student's schema predicted a conventional fine, that is, a<br />
positive response to the conventionalized query, and this was responded to. Expectation<br />
again determined interpretation. There is, additionally, the possibility that<br />
fuzzy was not understood, such that decoding uncertainty may have influenced the<br />
predictive inteipretation.<br />
Situation 21: Gas Company/Closing Talk<br />
NS: I'll phone you right back and we'll arrange this whole thing again.<br />
NNS: Yes, thank you very much.
178 Discourse Perspectives<br />
NS: And once again, we're terribly sorry that we didn't get round to your place today.<br />
NNS: Oh, thank you.<br />
This discourse stretch takes place inside the closing phase of a situation in which the<br />
learner had complained about an appointment not having been kept by an employee<br />
of a gas company. A final ritual apology is offered by the offender, which is,<br />
however, not responded to by a minimizer or some other appropriate move but by a<br />
repetition of a Thanks previously realized by the learner. This learner seems to have<br />
been operating inside a "closing talk" schema, knew that a Thank-you was her<br />
operative response, predicted a repetition of the compensatory offer made before,<br />
and responded to such a move as though it had occurred in reality. In the interview<br />
the learner reported that she had in fact engaged in some "advance planning," and<br />
apparently this advance planning barred her from flexibly reacting to the apology,<br />
which was unexpected as it referred back to an early phase in the encounter.<br />
As a last type of inappropriate response because of "psychopathological" slips, I<br />
would like to cite some instances that cannot be related to any one phase in the<br />
proposed discourse-processing model or even one of the suggested shortcutting<br />
paths: we deal here, rather, with cases in which it appears learners were unable to<br />
handle the cognitive load under the time constraints imposed by fast-moving oral<br />
interaction, and the necessity of chunking and weighting the incoming discourse<br />
stretches. The following example attests to such difficulties experienced by the<br />
learners in my corpus:<br />
Situation 22: Invite/Opening Phase<br />
NS: You've been working! Oh, I didn't know you had a job! What kind of job<br />
NNS: Well, at least I'm trying.<br />
The learner only responds to the first part of the interlocutor's utterance. The case is<br />
complicated by the connotations of the term "work." For some students, "work" is<br />
something one forces oneself to do. It has nothing to do with paid employment.<br />
Thus the response to the first part of the input may be adequate from the learner's<br />
perspective, but ignores the apparent misunderstanding evidenced in NS's further<br />
query. It is possible that this misunderstanding itself influences the content of the<br />
learner's answer (some people feel apologetic about admitting to the fact that one<br />
"works" occasionally), and this complex of factors leads to overload. It is also<br />
possible that the learner was not able to infer the interlocutor's different understanding<br />
of the term "work," and simply bypassed the problem by responding to his own<br />
interpretation of the input.<br />
Conclusion<br />
The focus of this chapter has been on postulating a cognitive model inside which<br />
different kinds of "inappropriate responses" may be located. The major problem<br />
remains the validity of explanatory analyses offered inside this, or indeed any other<br />
framework. How does one "know" of a concrete piece of data which processing or<br />
knowledge constraints or difficulties led to its occurrence Given our current state of<br />
knowledge, it seems to me that one can only plead for rich, context-based interpre-
Native I Normative Interactions 179<br />
live hypotheses, established inside an explicit cognitive framework. I have attempted<br />
such here. Furthermore, this study supports the belief that introspective and<br />
retrospective data may be useful for interpreting observed pragmatic failure.<br />
I have suggested here some of the complexity of responding in a foreign<br />
language, and attempted to specify what problems may underlie instances of<br />
pragmatic-responding failure. Three types of problems were suggested, and an<br />
attempt was made to understand them.<br />
1. Language-based difficulties leading to linguistic decoding and encoding<br />
problems<br />
2. Conceptual and strategic deficiencies related to gaps in culture-specific pragmatic<br />
knowledge<br />
3. Operational difficulties that may lead to interactional slips and inputs apparently<br />
being ignored.<br />
Furthermore, the notion of cognitive overload has been given some illustrative<br />
substance. One pedagogically relevant conclusion (that has been made often enough<br />
in the past (e.g., Edmondson, 1989; Edmondson & House, 1991) is the desirability<br />
of learners' having a thoroughly established battery of routine formulas for more<br />
successful discourse participation in L2. At the same time, the data above suggest<br />
that routines may be too fixed, as it were, in that nonexpected responses to routinized<br />
queries may shortcut a learner-discourse-processing system. The desirability<br />
of adequate strategic repair and elucidation strategies is a further pedagogic<br />
desideratum supported by the data.<br />
I have suggested several research strands might be relevantly searched and<br />
adapted for an eventual integrative-interdisciplinary approach to understanding and<br />
explaining "deviant" interlanguage responses and misunderstandings. I have integrated<br />
some of the insights of the five research strands into the frame provided by a<br />
discourse-processing model, and have supported this model with a number of illustrative<br />
conversations. Broadening our descriptive and explanatory framework is<br />
essential to understanding why learners' responses seem somewhat enigmatic at<br />
times. I conclude with a quote from Wallace Chafe:<br />
Understanding is increased by expanding the descriptive framework to include a<br />
large number of causal relations. Two things are necessary here. One is attention<br />
to a maximally wide range of data. How easy would it be to understand the<br />
workings of the solar system if we restricted ourselves to observing only the sun<br />
and moon, because they are bigger and brighter and thus easier to observe than the<br />
planets The other necessity is having the ability to imagine a maximally wide<br />
range of explanatory models. Someone had to think of planets moving in elliptical<br />
orbits around the sun. . . . These two necessities can be encapsulated in the statement<br />
that understanding increases with an expanded field of vision. (1990, 19)<br />
Notes<br />
1. A similar distinction between responses in which linguistic cohesion is at stake (replies)<br />
and those in which underlying shared knowledge provides for coherence is made by<br />
Pomerantz (1984, 153ff.), who also discusses different types of problems speakers may find<br />
themselves faced with when pursuing a response.
180 Discourse Perspectives<br />
2. Schegloff (1987) lists comparable causes for misunderstandings in talk-in-interaction,<br />
when he makes a distinction between two sources of "trouble": problematic reference ("propositions")<br />
and problematic sequential implicativeness, the latter comprising four distinct<br />
sources of trouble: favoured action interpretation (illocutions), serious/non-serious trouble<br />
sources (key/tone), as well as a so-called constructive/composite distinction and the practice<br />
of "joke first."<br />
3. The following is an example of the type of brief situational and role outline informants<br />
were given at the outset of each recording session:<br />
You, student X, have just discovered that your car doesn't start, which annoys you greatly as<br />
you have a course at the university in twenty minutes. It is already fairly late, and as you're<br />
sure that you won't be able to make it by bike or bus, you decide to phone up a friend of<br />
yours, Y, to ask him for a lift.<br />
An instruction matching K's role would stress the fact that Y had just decided that he would<br />
skip classes this morning or some other complicating factor.<br />
References<br />
Abelson, R. P. (1976). A script theory of understanding attitude and behavior. In J. Carroll &<br />
T. Payne (Eds.), Cognition and social behavior, pp. 33-46 Hillsdale, NJ: Lawrence<br />
Erlbaum.<br />
Bartlett, F. C. (1932). Remembering. Cambridge: Cambridge University Press.<br />
Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine.<br />
Benson, D., & Hughes, J. (1983). The perspective of ethnomethodology. London: Longman.<br />
Bilmes, J. (1990). Mishearings. Unpublished manuscript, University of Hawaii at Manoa.<br />
Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different Journal of<br />
Pragmatics, 11, 145-60.<br />
Blum-Kulka, S., & Weizman, E. (1988). The inevitability of misunderstanding: Discourse<br />
ambiguities. Text, 8, 219-41.<br />
Brown, P., & Levinson, S. (1987). Politeness. Cambridge: Cambridge University Press.<br />
Chafe, W. (1990). Looking ahead. Text, 10, 19-22.<br />
Clark, H. (1979). Responding to indirect speech acts. Cognitive Psychology, II, 430-77.<br />
Clark, H., & Lucy, P. (1975). Understanding what is meant from what is said: A study in<br />
conversationally conveyed requests. Journal of Verbal Learning and Verbal Behavior,<br />
14, 56-72.<br />
Clark, H., & Schunk, D. (1980). Polite responses to polite requests. Cognition, 8, 111-43.<br />
Dascal, M. (1977). Conversational relevance. Journal of Pragmatics, I, 309-28.<br />
Dascal, M. (1985). The relevance of misunderstanding. In M. Dascal (Ed.), Dialogue: An<br />
interdisciplinary approach (441-59). Amsterdam: Benjamins.<br />
Dascal, M., & Idan, A. (1988). From individual to collective action. In F. Vandamme and R.<br />
Pinxten (Eds.), Festschrift fiir Leo Apostelo (133-48). Ghent: Communication and<br />
Cognition.<br />
Dascal, M., & Katriel, T. (1979). Digressions: A study in conversational coherence. PTL: A<br />
Journal for Descriptive Poetics and Theory of Literature, 4, 203-32.<br />
Edmondson, W. J. (1981). Spoken discourse. London: Longman.<br />
Edmondson, W. J. (1987). "Acquiring" and "Learning": The discourse system integration<br />
hypothesis. In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance.<br />
Tubingen: Narr. 1070-89.<br />
Edmondson, W. J. (1989). Discourse production, routines and language learning. In P.<br />
Bierbaumer et al. (Eds.), Eng/isch als Zweitsprache. Tubingen: Narr.
Native I Normative Interactions 181<br />
Edmondson, W. J., & House, J. (1981). Let's talk and talk about it. A pedagogic interactional<br />
grammar of English. Munich: Urban and Schwarzenberg.<br />
Edmondson, W. J., & House, J. (1991). Do learners talk too much The waffle phenomenon<br />
in interlanguage pragmatics. In R. Philippson, E. Kellerman, L. Selinker, M. Sharwood<br />
Smith & M. Swain (Eds.), Foreign language pedagogy research: A commemorative<br />
volume for Claus Faerch (273-87). Clevedon and Philadelphia: Multilingual<br />
Matters.<br />
Erickson, F. (1975). Gatekeeping and the melting pot: Interaction in counseling encounters.<br />
Harvard Educational Review, 45, 44-70.<br />
Erickson, F., & Shultz, J. J. (1982). The counselor as gatekeeper. New York: Academic<br />
Press.<br />
Faerch, C., & Kasper, G. (1980). Processes and strategies in foreign language learning and<br />
communication. Interlanguage Studies Bulletin, 5, 47-118.<br />
Faerch, C., & Kasper, G. (Eds.). (1983). Strategies in interlanguage communication. London:<br />
Longman.<br />
Fillmore, C. (1976). Topics in lexical semantics. In R. Cole (Ed.), Current issues in linguistic<br />
theory. Bloomington: Indiana University Press. 76-138.<br />
Firth, J. (1964). The tongues of men, and speech. London: Oxford University Press.<br />
Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday.<br />
Goffman, E. (1974). Frame analysis. New York: Harper and Row.<br />
Goffman, E. (1981). Forms of talk. Oxford: Basil Blackwell.<br />
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and<br />
semantics (Vol. 3): Speech acts. New York: Academic Press.<br />
Gumperz, J. (1982a). Discourse strategies. Cambridge: Cambridge University Press.<br />
Gumperz, J. (1982b). Language and social identity. Cambridge: Cambridge University Press.<br />
Halliday, M.A.K. (1978). Language as social semiotic. London: Arnold.<br />
Heckhausen, H., & Beckmann, J. (1990). Intentional action and action slips. Psychological<br />
Review, 97, 36-48.<br />
Heikkinen, H., & Valo, M. (1985). Slips in interaction: The psychopathology of everyday<br />
discourse. In J. P. Forgas (Ed.), Language and social situations (213-28). New York:<br />
Springer.<br />
Holmes, J. (1985). Sex differences and miscommunication: Some data from New Zealand. In<br />
J. B. Pride (Ed.), Cross-cultural encounters. Communication and miscommunication<br />
(24-43). Melbourne: River Seine.<br />
House, J. (1979). Interaktionsnormen in deutschen und englischen Alltagsdialogen. Linguistische<br />
Berichte, 59, 76-90.<br />
House, J. (1982a). Conversational strategies in English and German dialogues. In G. Nickel<br />
& D. Nehls (Eds.), Error analysis, contrastive linguistics and second language<br />
learning (135-51). Heidelberg: Groos.<br />
House, J. (1982b). Gambits in deutschen und englischen Alltagsdialogen. Grazer Linguistische<br />
Studien, 17/18, 110-32.<br />
House, J. (1982c). Opening and closing phases in English and German dialogues. Grazer<br />
Linguistische Studien, 16, 52-83.<br />
House, J. (1984). Some methodological problems and perspectives in contrastive discourse<br />
analysis. Applied Linguistics, 5, 245-55.<br />
House, J. (1986). Cross-cultural pragmatics and foreign language teaching. In Seminar fur<br />
Sprachlehrforschung (Ed.), Probleme und Perspektiven der Sprachlehrforschung.<br />
Frankfurt: Scriptor.<br />
House, J. (1989). "Oh, excuse me please": Apologizing in a foreign language. In B. Kettemann,<br />
P. Bierbaumer, A. Fill & A. Karpf (Eds.). Englisch als Zweitsprache.<br />
Tubingen: Narr. 303-28.
182 Discourse Perspectives<br />
Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr.<br />
Kasper, G. (1982). Kommunikationsstrategien in der interimsprachlichen Produktion. Die<br />
Neueren Sprachen, 81, 578-600.<br />
Kasper, G. (1984). Pragmatic comprehension in learner-native speaker discourse. Language<br />
Learning, 34, 1-20.<br />
Keller-Cohen, D. (1979). Repetition in the non-native acquisition of discourse. In R. Freedle<br />
(Ed.), New directions in discourse processing. Norwood, NJ: Ablex. 245-71.<br />
Kintsch, W. (1988). The role of knowledge in discourse comprehension: A constructionintegration<br />
model. Psychological Review, 95, 163-82.<br />
Kintsch, W., & van Dijk, T. (1978). Towards a model of text comprehension and production.<br />
Psychological Review, 85, 363-94.<br />
Kochman, T. (1981). Black and white styles in conflict. Chicago: Chicago University Press.<br />
Lakoff, R. (1973). The logic of politeness, or, minding your p's and q's. Papers from the 9th<br />
Regional Meeting of the Chicago Linguistic Society (292-305). Chicago: Chicago<br />
Linguistic Society.<br />
Lakoff, R. (1976). Why you can't say what you mean [Review of Edwin Newman, Strictly<br />
Speaking}. Centrum, 4, 151-70.<br />
Lakoff, R. (1980). Psychoanalytic discourse and ordinary conversation. In R. Shuy & A.<br />
Schnukel (Eds.), Language use and the uses of language (269-87). Washington DC:<br />
Georgetown University Press. 269-87.<br />
Langer, E. (1978). Rethinking the role of thought in social interaction. In J. H. Harvey, W. J.<br />
Ickes & R. F. Kiddy (Eds.), New directions in attribution research (Vol. 2). Hillsdale,<br />
NJ: Lawrence Erlbaum.<br />
Langer, E., & Abelson, R. P. (1972). The semantics of asking a favor: How to succeed in<br />
getting help without really trying. Journal of Personality and Social Psychology, 24,<br />
26-32.<br />
Langer, E., Blank, A., & Chanowitz, B. (1978). The mindlessness of ostensibly thoughtful<br />
action: The role of "placebic" information in interpersonal interaction. Journal of<br />
Personality and Social Psychology, 36, 635-42.<br />
Langer, E., & Imber, L. (1980). Role of mindlessness in the perception of deviance. Journal<br />
of Personality and Social Psychology, 39, 360-67.<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
Minsky, M. (1975). A framework for representing knowledge. In P. Winston (Ed.), The<br />
psychology of computer vision. New York: McGraw-Hill. 211-77.<br />
Pomerantz, A. (1984). Pursuing a response. In J. M. Atkinson & J. Heritage (Eds.), Structures<br />
of social action. Cambridge: Cambridge University Press. 152-63.<br />
Reason, J., & Mycielska, K. (1982). Absent-minded The psychology of mental lapses and<br />
everyday errors. Englewood Cliffs, NJ: Prentice-Hall.<br />
Schank, R., & Abelson, R. P. (1977). Scripts, plans, goals and understanding. Hillsdale,<br />
NJ: Lawrence Erlbaum.<br />
Schegloff, E. A. (1987). Some sources of misunderstanding in talk-in-interaction. Linguistics,<br />
25, 201-18.<br />
Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 7, 289-327.<br />
Scollon, R., & Scollon, S. (1981). Narrative, literacy and face in interethnic communication.<br />
Norwood, NJ: Ablex.<br />
Searle, J. (1983). Intentionality. Cambridge: Cambridge University Press.<br />
Searle, J. (1992). Conversation. In J. Searle, H. Parret, & J. Verschueren, (pp. 7-29) (On)<br />
Searle on conversation. Amsterdam: Benjamins.<br />
Smith, N. V. (Ed.). (1982). Mutual knowledge. New York: Academic Press.<br />
Sperber, D., & Wilson, D. (1986). Relevance. London: Basil Blackwell.
Native/Normative Interactions 183<br />
Tannen, D. (1979). What's in a frame In R. Freedle (Ed.), New directions in discourse<br />
processing. Norwood, NJ: Ablex. 137-81.<br />
Tannen, D. (1984a). Conversational style: Analyzing talk among friends. Norwood, NJ:<br />
Ablex.<br />
Tannen, D. (1984b). The pragmatics of cross-cultural communication. Applied Linguistics,<br />
5, 189-95.<br />
Tannen, D. (1986). That's not what I meant. New York: Ballantine.<br />
Thomas, J. A. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.<br />
van Dijk, T., & Kintsch, W. (1983). Strategies of discourse comprehension. New York:<br />
Academic Press.<br />
Weizman, E., & Blum-Kulka, W. (1988). Ordinary misunderstanding. Paper delivered at the<br />
Varna Conference on Models of Meaning, Varna, Bulgaria.
9<br />
Explaining NNS Interactional Behavior:<br />
The Effect of Conversational Topic<br />
JANE ZUENGLER<br />
As Kasper (1989) points out, interlanguage (IL) performance is now understood as a<br />
variable phenomenon. Much of the variability is both normal and systematic. IL<br />
variation, in fact, represents a major focus of current second language acquisition<br />
(SLA) research and theory, and is the theme of conferences and books in the field<br />
(e.g., the 1987 Applied Linguistics Conference at Ann Arbor, Michigan; the 1990<br />
Second Language Research Forum [SLRF] at Eugene, Oregon; books by Eisenstein<br />
[ed.], 1989; Gass, Madden, Preston, & Selinker [eds.], 1989; Tarone, 1988). Increasingly,<br />
research has investigated IL performance in native speaker-nonnative<br />
speaker (NS-NNS) interactions. The findings point to a number of factors that cause<br />
IL performance to vary. One such factor emerging now is that of conversational<br />
topic. Though topic has been called "one of the most under-researched areas in the<br />
study of IL variation" (Tarone, 1988, 119), there is a small but growing body of<br />
empirical studies on conversational topic necessary to consider if we are to understand<br />
learner acquisition and performance of conversational skills.<br />
In addressing conversational topic, we are not simply looking at what it is that<br />
interlocutors talk about. Topic must not be viewed as a superficial, categorical<br />
construct independent of the conversation. The fundamental importance of topic is<br />
that it situates a speaker. First of all, speakers have cognitive and/or affective<br />
relationships to particular topics. For example, we may talk about computer software,<br />
and know a lot, a little, or nothing about it. Independent of how much we<br />
know, computer software may be something we are greatly, or not at all, interested<br />
in. Second, topic situates the speaker within the interaction, that is, vis-a-vis<br />
the interlocutor. In other words, it can shape one's conversational role. Taking<br />
computer software as an example, conversational roles are not simply a function of<br />
how much each interlocutor knows, or cares about, computer software in an absolute<br />
sense. How the interlocutors talk about software, and how active or not their<br />
roles, are significantly influenced by what their conversational partner knows and<br />
feels about the topic. Therefore, studying the influence of topic on NS-NNS interactions<br />
will enable us to learn more about conversational dynamics and role-taking in<br />
184
Explaining NNS Interactional Behavior 185<br />
general, and increase our understanding, in particular, of IL conversational performance<br />
and development.<br />
This chapter will first synthesize the research on conversational topic and IL<br />
performance. Following that will be a discussion of the research outcomes, as they<br />
relate to two current theoretical models of IL variation. The two models are Selinker<br />
and Douglas's Discourse Domain Model and Giles's Speech Accommodation Theory<br />
(SAT).<br />
Research on IL and Conversational Topic 1<br />
The research germane to topic and IL has generally taken one of two focuses; it<br />
either compares the same learners' IL performance across various topics, or it<br />
compares performance by different IL learners when a particular topic is held<br />
constant. Measures of IL performance vary considerably among the studies. While<br />
a researcher may analyze the same set of IL features from one investigation to the<br />
next, there is no such uniformity from researcher to researcher. However, the lack of<br />
uniformity does not in itself represent a problem; it simply reflects the fact that in an<br />
emerging research area, there is a wide range of potential measures to consider.<br />
It will be apparent, in the research review, that some studies select traditional<br />
"linguistic" measures (e.g., pronunciation, verb accuracy) while others analyze features<br />
more recognizably pragmatic (e.g., directives, interruption behavior). Still<br />
others analyze both types of features. Whether the units of analysis are "linguistic"<br />
or "pragmatic," the overwhelming majority of the studies to be reviewed are,<br />
nevertheless, germane to the pragmatics of IL performance. The researchers are<br />
investigating whether, and how, topic influences the ways in which IL speakers<br />
perform a conversation, and, specifically, the type of conversational role they assume.<br />
Such research, therefore, has pragmatically relevant outcomes, even though,<br />
in some studies, the measures themselves may be more traditionally "linguistic"<br />
than pragmatic.<br />
Only two of the studies to be reviewed are not clearly pragmatically based<br />
(though they contain pragmatic measures). Cornu and Delahaye (1987) and Smith<br />
(1989) ask acquisitional questions, rather than questions of performance per se.<br />
They look at performance as a means, rather than an end in itself, because they are<br />
primarily interested in determining whether IL acquisition varies across discourse<br />
domains.<br />
Studies of IL Performance Across Topics<br />
Since many IL speakers are university students studying their major field through<br />
the second language (i.e., have a "language for specific purposes" [LSP] orientation),<br />
several studies have compared subjects' IL performance when conversing on a<br />
topic within their major field with their performance on a topic lying outside their<br />
major field. Two of the studies in particular illustrate the impact of topic on the roles<br />
that interlocutors take in the conversations.<br />
Selinker and Douglas (1985) report on an IL speaker (Luis) participating in two
186 Discourse Perspectives<br />
conversations, one concerning civil engineering, his major field, and the other<br />
concerning a life-story topic (how to prepare his native food). Luis's performance<br />
varied from one conversation to the next. Each topic, the authors suggest, led to<br />
different role dynamics between Luis and his interlocutor, resulting in variable IL<br />
performance. Selinker and Douglas point out that when Luis conversed about his<br />
major field, he was the relative "knower" in the interaction, since it was his specialty,<br />
and his interlocutor was a NS who was not in the same field. Therefore,<br />
Louis's role was to make evident his expertise. In conversing about food, on the<br />
other hand, Luis was outside his area of expertise and interest, and the same was<br />
true for his interlocutor, who in this case was a friend of his. In this type of<br />
interaction, there was more role negotiation, partly because the interlocutors were<br />
friends, and partly because neither "knew" the topic.<br />
The topic-influenced role differences led to variation in some of Luis's communication<br />
strategies. In particular, there was variation in the strategy Luis used when<br />
he searched for a word he did not know. When conversing about engineering, Luis<br />
(as "knower") continued talking until he found a synonym his interlocutor could<br />
respond to. However, when he needed vocabulary to discuss food, Luis was more<br />
likely to let the episode break down, saying "forget it" at one point. There was no<br />
impetus for Luis to show that he was the "knower" of food (because he was not),<br />
and he very likely held less interest in that topic than in the major field topic.<br />
Luis also varied the way he corrected his interlocutor. This occurred when the<br />
NS interlocutor misinterpreted the content of what Luis was saying. When discussing<br />
engineering, Luis corrected his interlocutor very directly without mitigating<br />
what he said. When discussing the life-story topic, however, Luis mitigated his<br />
corrections by using politeness strategies. Here again, Selinker and Douglas suggest,<br />
the cross-topic variation in response to interlocutor mistakes may reflect the<br />
relative importance of the topic to the IL speaker. Misinterpretations by the interlocutor<br />
when talking about food preparation may not be as urgent to correct because<br />
that topic is probably much less important to the IL speaker than is his major field.<br />
Results from a larger study by Zuengler and Bent (1991) concur with Selinker<br />
and Douglas (1985) in pointing to the influence of topic on the conversational roles<br />
taken by IL speakers. The Zuengler and Bent research involved, in part, an analysis<br />
of 45 IL speakers conversing on two separate topics in dyads with NS interlocutors.<br />
(See Zuengler 1989b for a subset of the results). All 45 NS-NNS dyads were first<br />
asked to have a conversation about favorite and holiday foods. Then, each dyad<br />
discussed a topic within their shared major field. The researchers looked for patterns<br />
of conversational participation and dominance, employing measures such as amount<br />
of talk, use of pause fillers to retain one's turn, topic moves, and back-channels.<br />
Outcomes revealed variation in IL conversational participation from topic to topic.<br />
When discussing food, the NNSs (relative to their NS partners) produced significantly<br />
more talk and more pause fillers, while their NS partners back-channeled<br />
significantly more, role patterns that Zuengler and Bent labeled NNS "speaker" and<br />
NS "listener." These particular role patterns did not necessarily appear, however,<br />
when the subjects had a conversation concerning their major field. In dyads in<br />
which the NSs and NNSs shared the same major field and were at an equivalent<br />
level in it, the NSs talked significantly more than their NNS partners, while the
Explaining NNS Interactional Behavior 187<br />
NNSs back-channeled significantly more. That is, the role patterns were opposite to<br />
those in the conversations about food. When discussing the major field topic, it was<br />
the NSs who performed as active "speakers" and the NNSs who were the active<br />
"listeners." (In both conversations, the NNSs produced more pause fillers. Topic<br />
moves were nonsignificant with both topics.) The NS as "speaker" in the major field<br />
conversations supports claims by Beebe and Giles (1984), among others, who assert<br />
that NSs tend to dominate interactions with NNSs, perhaps due to feelings of<br />
ethnolinguistic superiority (but see discussion below). However, the food conversations<br />
showed no such NS dominance, and in fact, as pointed out, the NNSs spoke<br />
significantly more than the NSs. Zuengler and Bent (1991) suggest that while<br />
discussing food, the NSs may have conceded more of the talk to the NNSs because<br />
the topic held little importance to the NSs. Or, the NSs may have judged the NNSs<br />
as knowing more than they about the food topic. Since the NNSs resided in the NSs'<br />
setting (namely, the United States), while the NSs had not, correspondingly, resided<br />
in the NNSs' settings, the NNSs tended to be familiar with NS foods, whereas the<br />
NSs did not typically show conversance with NNS foods. Relative knowledge of the<br />
topic was, in fact, an important influence on conversational roles in other comparisons<br />
tested by Zuengler and Bent (1991). (These outcomes will be brought up<br />
below.)<br />
Several additional studies looked at IL performance on major field versus nonmajor<br />
field topics, but focused primarily on specific characteristics of the IL per se,<br />
rather than on interactional dynamics such as role relationships. Cornu and Delahaye<br />
(1987) tested two subjects and report IL variation between major field and<br />
nonmajor field topics. Providing detailed results for one of the two subjects, they<br />
report that when the IL speaker talked about economics, her major field, she<br />
produced more embedded sentences, used fewer nonverbal strategies, performed<br />
more self-correction of grammatical errors, and was less hesitant about and more<br />
appropriate in her lexical choices. Cornu and Delahaye conclude that the subject<br />
was more in control of the major field topic than the nonmajor field topic; she had<br />
more control of the economics topic, they suggest, because her IL contained more<br />
"integration of the target lexical systems" (1987, 150). Greater integration (and,<br />
consequently, more control) can result from a learner's having greater exposure to<br />
and practice in the TL in topic areas such as the major field.<br />
A recent study by Smith (1989) compared IL performance on general topics in<br />
the SPEAK test (a test of spoken language proficiency developed by ETS) versus<br />
performance on major field topics in a specially developed version of the same test.<br />
Thirty-eight graduate students were tested, and global measures were obtained of<br />
their comprehensibility, pronunciation, grammar, and fluency. In contrast to the<br />
preceding studies' outcomes, these subjects showed no significant differences in<br />
their IL performance on major field versus nonmajor field topics. One possible<br />
explanation for the lack of difference in performance is test anxiety as a variable.<br />
Smith (1989) reports that a native-speaker control group that took the test indicated<br />
that they found each test "quite stressful" (160). If native speakers found the task<br />
stressful, it is quite likely that the IL speakers did as well. Further research in this<br />
area is important for determining whether test anxiety was indeed important in<br />
eliciting data such as in Smith (1989).
188 Discourse Perspectives<br />
In addition to looking at IL performance on major field versus nonmajor field<br />
topics, research has investigated performance on topics that differ according to<br />
speaker interest or affect. Eisenstein and Starbuck (1989) asked eight IL speakers<br />
what topics they were interested in as well as topics they considered uninteresting.<br />
(The researchers referred to these as "invested" and "uninvested" topics, respectively.)<br />
Each subject had two conversations with the researchers, one on an invested<br />
topic, and the other on an uninvested topic. What Eisenstein and Starbuck analyzed<br />
was the subjects' targetlike accuracy of their verb system across topics. Overall, the<br />
subjects' IL was significantly less accurate when discussing the invested topic than<br />
when discussing the uninvested topic. Specifically, subjects were less accurate in<br />
lexical choice, tense choice, tense form, and use of repetitions. The researchers<br />
suggest that the differences across topics might be due to monitoring or cognitive<br />
load. When discussing the invested topic, subjects talked more, and this may have<br />
caused them to monitor less. Or, it is possible that the cognitive load was greater in<br />
the invested condition (though the subjects were at an advanced IL level, and the<br />
invested topics did not appear to be cognitively demanding). Additional research<br />
should help clarify whether monitoring or cognitive load may be confounding<br />
variables in invested-uninvested topic comparisons.<br />
Certain topics can trigger positive or negative emotions in IL speakers, and their<br />
performance may vary as a result. A pilot study reported by Zuengler (1982)<br />
consisted of asking 13 IL speakers questions on two different topics, a "neutral" one<br />
(asking them their views on children watching television), and one that was potentially<br />
ethnically-threatening (telling subjects that they should speak English in the<br />
United States, and if they wanted to speak their native language, they should "go<br />
back to [their] own country"; subjects were asked their views on that). Subjects'<br />
responses to the two different topics were compared with respect to their production<br />
of three phonological variables. Zuengler provides descriptive data illustrating that<br />
some of the subjects were indeed emotionally aroused by the ethnically threatening<br />
topic. Those subjects had lower accuracy when responding to the ethnically threatening<br />
question than in responding to the neutral question. On the other hand,<br />
subjects who did not respond negatively to the ethnically threatening question<br />
(perhaps because their ethnic identity was weak) performed more accurately when<br />
answering the ethnically threatening question.<br />
Emotionality of topic was also a major focus in research by Dowd (1984).<br />
However, in contrast to Zuengler (1982), Dowd's research did not concern emotion<br />
attached to ethnic identity. Instead, Dowd asked her subjects to talk about an<br />
unpleasant personal experience they had had. Prior to the "emotional" topic, they<br />
were asked to discuss a "neutral" topic, for example, to describe their favorite food,<br />
their home, family, etc. Subjects were 100 Mexican female IL speakers. Dependent<br />
measures of IL performance were six phonological variables. Results were mixed.<br />
Only two of the six variables, final consonant clusters and /r/, significantly differed<br />
in accuracy from one topic to the next. The pattern of shift in accuracy also differed.<br />
In going from "neutral" to "emotional" topics, subjects significantly decreased in<br />
TL accuracy on final consonant clusters, but they significantly increased in accuracy<br />
on /r/. Though Dowd's (1984) outcomes reveal an (albeit limited) effect of<br />
topic shift, it is not known why the variables shifted in different directions. Such
Explaining NNS Interactional Behavior 189<br />
bidirectional shifts are important for SLA researchers to understand (Dowd,<br />
Zuengler, & Berkowitz, 1990; see also Tarone, 1985). It will be noted that though<br />
the Zuengler (1982) and Dowd (1984) research may have had a general pragmatic<br />
orientation, their measures for analysis were limited to linguistic items. Obviously,<br />
it will be important to analyze pragmatic measures as well as to view IL performance<br />
in such research.<br />
IL Variation Within a Given Topic<br />
Three research studies investigate IL performance when the topic is held constant,<br />
the important variable being IL speakers' knowledge of the topic relative to that of<br />
their (NS) interlocutors. All three studies consider relative topic knowledge as a<br />
definer of one's conversational role (and thus are closely related in conception to<br />
several studies discussed earlier). The present studies asks, implicitly or explicitly:<br />
Can an interlocutor's knowledge of the conversational topic, relative to the partner's<br />
knowledge, explain whether one interlocutor dominates, controls, or directs the<br />
other, or whether both interlocutors play an equal role in building the conversation<br />
As mentioned earlier, there are claims by Beebe and Giles (1984) and others that in<br />
many NS-NNS interactions, the NS tends to dominate, due, perhaps, to feelings of<br />
ethnolinguistic superiority. Investigations of relative topic knowledge provide an<br />
important means of testing such claims.<br />
Woken and Swales (1989), while limiting their study to three female NS-NNS<br />
dyads, provide clear evidence for control in the NNS's performance when she had<br />
greater topic knowledge. Each of the NS-NNS dyads was composed of an NNS<br />
graduate student in computer science who was asked to show her NS partner (who<br />
was not a computer science major) how to use some software. The researchers<br />
report that in all three dyads, the NNSs, due to their topic expertise, clearly controlled<br />
the conversation in talking more (and producing longer t-units), correcting<br />
the NS more than the NS corrected them, and in giving more directions. The NNS<br />
corrections to the NSs were largely unmitigated (as were those of the NNS in<br />
Selinker and Douglas [1985], when discussing his major field knowledge). The<br />
NSs, on the other hand, produced more inquiries, which served as requests for<br />
clarification and for lexical help from the NNSs. Woken and Swales (1989) conclude<br />
that their three dyads illustrate the influence of relative topic expertise on<br />
control and participation in a conversation, regardless of the nativeness or nonnativeness<br />
of the interlocutors. As such, the evidence of NNS control contradicts the<br />
claims of Beebe and Giles (1984) and others regarding NS tendency to dominate (or<br />
at the least, it qualifies any such claims regarding the NS). However, it remains to<br />
be determined whether less fluent NNSs than these would also control such conversations.<br />
Like Woken and Swales (1989), the Zuengler and Bent (1991) study introduced<br />
earlier also addressed relative topic knowledge in NS-NNS interactions, but in a<br />
larger number of dyads. Forty-five male NS-NNS dyads were formed to constitute<br />
three distinct groups. In all dyads, NSs and NNSs shared their major field and were<br />
asked to have a conversation on a specific topic within the field. In 15 of the dyads,<br />
the NNS was the relative "topic expert," in 15 other dyads, the NS was the relative
190 Discourse Perspectives<br />
"topic expert," and in the remaining 15, NS and NNS were relatively equal with<br />
respect to topic knowledge. (Topic expertise was determined by what level the<br />
subject was at in the major field.) Measures of conversational participation and<br />
dominance included amount of talk, pause fillers, back-channels, topic moves, and<br />
interruptions. While results were nonsignificant for interruptions, each of the other<br />
measures revealed some significant patterns. When the NNSs had relative topic<br />
expertise, they produced significantly more talk, more fillers, and more backchannels<br />
than their NS partners. When, on the other hand, the NSs were the "topic<br />
experts," the NSs produced more talk and more fillers. As discussed earlier, when<br />
both NS and NNS were at the same stage in the major field, the NSs talked<br />
significantly more, while the NNSs produced more pause fillers. Interestingly,<br />
regardless of topic expertise, the NNSs produced significantly more back-channels.<br />
Zuengler and Bent (1991) view the results as a clear indication of the influence<br />
of relative topic knowledge on one's role in a conversation; topic expertise led to<br />
greater amounts of talk and more pause-filling. At the same time, the NNSs,<br />
regardless of topic expertise, tended to signal an active listening role as well, by<br />
greater use of back-channels. It was only when the NNSs were the "topic experts"<br />
that they displayed speaker dominance by talking more. As mentioned above in the<br />
discussion about cross-topic variation, the fact that the NSs talked more than the<br />
NNSs in the "equal" dyads perhaps indicates a tendency for NSs to dominate due to<br />
feelings of ethnolinguistic superiority, even though they may be "equal" with respect<br />
to topic knowledge.<br />
A third study illustrates that relative topic expertise can influence 1L performance<br />
even if it is only perceived, rather than actual, expertise. Zuengler (1989a)<br />
administered an art judgment test to 45 dyads of NS-NNS women. She then asked<br />
them to have two conversations about their individual judgments. In the first conversation,<br />
subjects did not know how well they had performed on the test. However,<br />
prior to the second conversation, Zuengler manipulated the scores so that in 15 of<br />
the dyads, the NNS was led to believe that she had performed much better in making<br />
art judgments than her NS partner had (i.e., that the NNS was the relative "topic<br />
expert"), while in another 15 dyads, it was the NS who was led to believe that she<br />
was the relative "topic expert." (The manipulated scores had no relation to subjects'<br />
actual scores.) To establish a control group, in 15 of the dyads neither interlocutor<br />
was told how she had performed. All 45 dyads then had a second conversation about<br />
their judgments. Zuengler (1989a) reports an analysis of amount of talk, interruptions,<br />
and task moves in an effort to determine whether perceived expertise had an<br />
effect on conversational performance.<br />
Outcomes revealed that the NSs consistently dominated in terms of amount of<br />
talk; NSs produced significantly more talk than their NNS partners, across all<br />
conversations. Moreover, topic expertise (in the second conversation) led the NS<br />
perceived experts to increase their already greater amount of talk in the first conversation.<br />
The other measures revealed two patterns: (1) perceived expertise led both<br />
NSs and NNSs to interrupt more successfully and to move the task along more often<br />
than their perceived nonexpert partners; and (2) the NSs tended to dominate in the<br />
control group conversations. They not only produced more talk, but they interrupted<br />
more successfully and moved the task along more frequently than their NNS part-
Explaining NNS Interactional Behavior 191<br />
ners did. According to Zuengler (1989a), the findings show that when topic expertise<br />
differences are not apparent to either interlocutor, it is the NS who will tend to<br />
dominate a conversation with an NNS. The outcomes of Zuengler and Bent (1991),<br />
discussed above, corroborate this. However, when interlocutors are aware of differences<br />
in topic expertise (even though the differences may be apparent, not real), the<br />
relative topic expert, NS or NNS, will tend to dominate the conversation. The only<br />
exception is in amount of talk, which the NSs produced more of overall. Zuengler<br />
(1989a) suggests that the NNSs may have talked less because of limitations in their<br />
language proficiency. While none were IL beginners, it was apparent, in listening to<br />
the conversations, that many of the NNSs showed difficulty in producing fluent<br />
talk. Analysis of speech rate revealed that NNSs consistently spoke significantly<br />
more slowly than NSs. NNSs often signaled difficulty by uttering "I don't know<br />
how to say this" and "this is difficult for me."<br />
To summarize, the research conducted on conversational topic addresses two<br />
questions: Does IL performance vary from one topic to the next And, does IL<br />
performance vary within a given topic The review of research on conversational<br />
topic reveals that both types of variation are evident in IL performance. (Of course,<br />
caveats should be attached, as not all of the studies have statistically tested results,<br />
and one of the studies [Smith, 1989] had nonsignificant outcomes). While it is<br />
obvious that more research must be undertaken in this area, the research conducted<br />
to date already provides evidence that topic does indeed affect IL performance.<br />
Explaining the Outcomes: Looking to SLA Theory<br />
It is important to recognize that it is not topic per se that is of importance to SLA<br />
research; rather, it is what topic reveals about IL development and use. That is, how<br />
can research on topic contribute to theories of second language acquisition, and, at<br />
the same time, how can language acquisition theories help us interpret the outcomes<br />
of the research While some of the research studies discussed in this paper address<br />
theory, either in their conception or in interpreting their outcomes, others do not.<br />
Whether the researchers directly address theory or not in their studies, the majority<br />
of the research outcomes can be shown to relate to either of two current SLA<br />
theories in particular: the Discourse Domain Model and Giles' Speech Accommodation<br />
Theory (SAT). Considering the outcomes that relate to these two theories will<br />
help us understand topic as part of the larger picture of language acquisition and<br />
use, and, at the same time, help us assess the viability of the theories in accounting<br />
for the data. 2<br />
The Discourse Domain Model: IL Development<br />
and Use Varies According to Domain<br />
The Discourse Domain Model (Douglas & Selinker, 1985; Selinker & Douglas,<br />
1985) is a cognitively oriented theory which states that learners develop ILs through<br />
various content areas, or discourse domains, that are important to or needed by<br />
them. 3 Selinker and Douglas claim that because IL development is domain-situated,
192 Discourse Perspectives<br />
a learner's IL may be well-developed, structurally and functionally, in one domain,<br />
but not in the next. Fossilization, backsliding, and disfluencies in IL can occur when<br />
learners are asked to converse within a domain that they have little or no knowledge<br />
of (and/or do not know the forms and functions for conversing about it in the L2).<br />
While Selinker and Douglas conceptualize discourse domain without using the<br />
word "topic," it can be argued that the latter is a performance element of the<br />
former. 4 That is, people do not talk about discourse domains; instead, they talk<br />
about "things" (topics) that engage a domain of discourse. As such, a discourse<br />
domain theory has direct relevance to studies of IL and topic, particularly to some of<br />
the research discussed earlier which illustrates IL variation across topics.<br />
Discourse domain theory can explain the effect of knowing, or being interested<br />
in, a topic. Subjects in Selinker and Douglas (1985) and Cornu and Delahaye (1987)<br />
appeared to speak more fluently and assertively when discussing their major field<br />
than when talking outside their major field. The major field topic engaged a discourse<br />
domain of which the subjects had more cognitive control than the other<br />
topic's domain. (However, test anxiety may undermine the control, and thus explain<br />
Smith's [1989] lack of significant differences.)<br />
Asking subjects to talk about topics of varying interest to them (Eisenstein &<br />
Starbuck, 1989) may also illustrate discourse domain theory (though further research<br />
is needed to establish this). Often, what we have differing interests in we also<br />
have differing amounts of knowledge about. That Eisenstein and Starbuck's (1989)<br />
subjects were less formally accurate in discussing their topic of interest may indicate<br />
that they had engaged a discourse domain in which they were quite comfortable, and<br />
in talking more (which they did), they neglected to monitor their speech. Or, as the<br />
authors also suggest, the cognitive load of the invested (interesting) topic may have<br />
been greater, causing the subjects to be less formally accurate. In either case, it is<br />
discourse domain theory, in suggesting that IL can vary in development and use<br />
from domain to domain, that can provide an explanation for the cross-topic IL<br />
variation observed in the studies just cited.<br />
Speech Accommodation Theory (SAT): IL Use Varies<br />
According to Interactional Dynamics<br />
Topic and discourse domains, nonetheless, are not solely cognitive. When IL speakers<br />
are interacting with someone, the research has shown, how much they know<br />
about the topic will influence their IL use. However, it is not strictly the speaker's<br />
absolute knowledge of the topic that will determine how active a conversational role<br />
she will take; it is the speaker's knowledge compared to the interlocutor's knowledge.<br />
In other words, topic knowledge within an interaction is interactionally defined.<br />
It is in this sense that SAT is relevant for explaining some of the outcomes.<br />
By combining several social psychological theories, Giles's Speech Accommodation<br />
Theory (SAT) explains why speakers adjust (i.e., accommodate) their<br />
speech while interacting with others. Conversations are perceived by participants as<br />
being either inter-individual or inter-group encounters; if the latter, it is often ethnic<br />
differences that become salient. Speakers can show solidarity, create distance, or<br />
maintain their position, by converging toward their partner's speech patterns, by
Explaining NNS Interactional Behavior 193<br />
diverging, or by not making any adjustments. Detailed explanation of SAT, as well<br />
as its application to SLA, is outside the scope of this discussion; the reader is<br />
referred instead to Beebe (1988), Beebe and Giles (1984), and Beebe and Zuengler<br />
(1983).<br />
As a theory of conversational dynamics, SAT can explain some of the IL<br />
variation that is beyond the conceptual scope of the discourse domain model. At the<br />
same time, the research outcomes can offer a needed elaboration of SAT.<br />
First of all, SAT, in emphasizing interactional dynamics, can provide an explanation<br />
for the outcomes in Zuengler (1989a), where perceived, and not actual,<br />
expertise differences led to IL variation. Interlocutors were led to have certain<br />
perceptions of their own and their partners' topic knowledge, and it was interlocutor<br />
perceptions (and not actual knowledge) that influenced IL performance. Even when<br />
actual knowledge levels are involved (see Zuengler & Bent, 1991; also, Woken &<br />
Swales, 1989), topic knowledge is interactionally determined according to comparisons<br />
the interlocutors make of each other. That is, the outcomes show that IL<br />
performance can be a function of the speaker's knowing more, or less, about the<br />
topic than the interlocutor knows, rather than how much absolute knowledge of the<br />
topic the speaker has, divorced from the interaction.<br />
As mentioned above, SAT posits that interactions can be viewed by interlocutors<br />
as being inter-group or inter-individual encounters. Primarily, SAT points to salient<br />
ethnic group differences between interlocutors as illustrating what would be perceived<br />
as an inter-group encounter. When ethnicity (or other group differences) is<br />
not made salient, it is likely, according to SAT, that the encounter will be perceived<br />
as an inter-individual one, with interlocutors viewing each other on an individual,<br />
rather than group, basis.<br />
Research suggests that it is topic that may be an important determinant of<br />
whether the NS-NNS encounter will be an inter-group or inter-individual encounter.<br />
In Zuengler (1982), IL speakers were given a question to respond to that threatened<br />
their ethnic identity. This approach, certainly, made ethnicity salient and a number<br />
of the subjects responded by defending their ethnicity. For these subjects, the topic<br />
focused the interaction on ethnic differences and made the interaction, for them, an<br />
inter-group encounter. Where conversational topics draw on NSs' and NNSs' individual<br />
experiences and training, instead of emphasizing their ethnic, or NS-NNS<br />
differences, we would expect the interlocutors to experience the interaction as an<br />
inter-individual encounter. We can point to the outcomes showing that conversational<br />
performance is a function of relative topic knowledge (where the topic is not<br />
ethnically salient). IL performance varied according to relative topic knowledge<br />
differences in Woken and Swales (1989), Zuengler (1989a), and Zuengler and Bent<br />
(1991). In these conversations, the IL speakers knew (or thought they knew) more<br />
or less than their NS interlocutors; their performance varied according to these<br />
knowledge differences. Clearly, then, what was salient to these interactions was an<br />
inter-individual dynamic (the interlocutor differences in topic knowledge).<br />
However, it is necessary to consider the outcomes of two studies (Zuengler,<br />
1989a; Zuengler & Bent, 1991) that showed that the NSs dominated when the<br />
interlocutors were supposedly equal in topic knowledge. That is, the conversational<br />
performance split along NS-NNS lines, with the NS as dominator and the NNS as
194 Discourse Perspectives<br />
subordinate. This indicates, I would argue, that if there are no apparent topic<br />
knowledge differences between the NSs and the NNSs, the interaction may become<br />
an inter-group encounter, with the NSs, unfortunately, exercising their beliefs of<br />
ethnolinguistic superiority (see Beebe & Giles, 1984).<br />
Conclusion<br />
As we have seen, two current theories of SLA, the Discourse Domain Model and<br />
SAT, can provide explanations for the IL performance variation apparent in the<br />
majority of the studies on conversational topic reviewed here. They provide complementary<br />
contributions, since the Discourse Domain Model is cognitively based and<br />
SAT is a social psychological theory. After all, when looking at the influence of<br />
topic on NS-NNS interactions, we must recognize that "topic" is both a cognitive<br />
and a socially situated construct. Recognizing this will not only help us understand<br />
the present research outcomes but will also enable us to plan the needed future work<br />
in this area.<br />
Notes<br />
1. Because conversational topic is the focus of this paper, research studies involving<br />
written IL and topic will not be included. For a discussion of some of the research on written<br />
IL variation, see Selinker and Douglas (1987).<br />
2. While it is essential, in theory evaluation, to determine the explanatory adequacy of<br />
the theory per se, such an evaluation is beyond the scope of this paper. See McLaughlin<br />
(1987).<br />
3. Tarone (1988) states that the Discourse Domain Model stems from social psychology,<br />
but does not provide support for this assertion. Selinker and Douglas, in their explanation of<br />
the model, clearly present it as a cognitive theory.<br />
4. Therefore, criticisms that a concept such as "discourse domain" is too vague for<br />
methodological purposes (see Tarone, 1988) may be unfounded. One could argue that it is<br />
topic that represents the operationalization of discourse domain.<br />
References<br />
Beebe, L. M. (1988). Five sociolinguistic approaches to second language acquisition. In<br />
L. M. Beebe (Ed.), Issues in second language acquisition: Multiple perspectives<br />
(43-77). New York: Newbury House.<br />
Beebe, L. M., & Giles, H. (1984). Speech-accommodation theories: A discussion in terms of<br />
second-language acquisition. International Journal of the Sociology of Language, 46,<br />
5-32.<br />
Beebe, L. M., & Zuengler, J. (1983). Accommodation theory: An explanation for style<br />
shifting in second language dialects. In N. Wolfson & E. Judd (Eds.), Sociolinguistics<br />
and language acquisition (195-213). Rowley, MA: Newbury House.<br />
Cornu, A. M., & Delahaye, M. (1987). Variability in interlanguagc reconsidered: LSP vs.<br />
non-LSP IL talk. English for Specific Purposes, 6, 145—52.
Explaining NNS Interactional Behavior 195<br />
Douglas, D., & Selinker, L. (1985). Principles for language tests within the "discourse<br />
domains" theory of interlanguage: Research, test construction and interpretation.<br />
Language Testing, 2, 205-21.<br />
Dowd, J. L. (1984). Phonological variation in L2 speech: the effects of emotional questions<br />
and field-dependence/field-independence on second language performance. Unpublished<br />
doctoral dissertation, Teachers College, Columbia University.<br />
Dowd, J., Zuengler, J., & Berkowitz, D. (1990). L2 social marking: Research issues.<br />
Applied Linguistics, 10, 16-29.<br />
Eisenstein, M. R. (Ed.). (1989). The dynamic interlanguage: Empirical studies in second<br />
language acquisition. New York: Plenum.<br />
Eisenstein, M. R., & Starbuck, R. (1989). Investment in topic and verb system accuracy of<br />
advanced L2 learners. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.),<br />
Variation in second language acquisition (Vol. 2): Psycholinguistic issues (125-37).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Gass, S., Madden, C., Preston, D., & Selinker, L. (Eds). (1989). Variation in second<br />
language acquisition (Vol. 1): Discourse and pragmatics (Vol. 2): Psycholinguistic<br />
issues. Clevedon and Philadelphia: Multilingual Matters.<br />
Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden,<br />
D. Preston & L. Selinker (Eds.), Variation in second language acquisition (Vol. 1):<br />
Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.<br />
McLaughlin, B. (1987). Theories of second-language learning. London: Edward Arnold.<br />
Selinker, L., & Douglas, D. (1985). Wrestling with "context" in interlanguage theory. Applied<br />
Linguistics, 6, 190-204.<br />
Selinker, L., & Douglas, D. (Eds.). (1987). Interlanguage [Special issue]. English for Specific<br />
Purposes, 6(1).<br />
Smith, J. (1989). Topic and variation in ITA oral proficiency. English for Specific Purposes,<br />
8, 155-68.<br />
Tarone, E. E. (1985). Variability in interlanguage use: A study of style-shifting in morphology<br />
and syntax. Language Learning, 35, 373-404.<br />
Tarone, E. E. (1988). Variation in interlanguage. London: Edward Arnold.<br />
Woken, M. D., & Swales, J. (1989), Expertise and authority in native-non-native conversations:<br />
The need for a variable account. In S. Gass, C. Madden, D. Preston & L.<br />
Selinker (Eds.), Variation in second language acquisition (Vol. 1): Discourse and<br />
pragmatics (211-27). Clevedon and Philadelphia: Multilingual Matters.<br />
Zuengler, J. (1982). Applying accommodation theory to variable performance data in L2.<br />
Studies in Second Language Acquisition, 4, 181-92.<br />
Zuengler, J. (1989a). Assessing an interaction-based paradigm: How accommodative should<br />
we be In M. R. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in<br />
second language acquisition (49-67). New York: Plenum.<br />
Zuengler, J. (1989b). Performance variation in NS-NNS interactions: Ethnolinguistic difference,<br />
or discourse domain In S. Gass, C. Madden, D. Preston & L. Selinker (Eds.),<br />
Variation in second language acquisition (Vol. 1): Discourse and pragmatics (228-<br />
44). Clevedon and Philadelphia: Multilingual Matters.<br />
Zuengler, J., & Bent, B. (1991). Relative knowledge of content domain: An influence on<br />
native-non-native conversations. Applied Linguistics, 12, 397-415.
10<br />
The Metapragmatic Discourse<br />
of American-Israeli Families at Dinner<br />
SHOSHANA BLUM-KULKA and HADASS SHEFFER<br />
In the broadest sense, interlanguage pragmatics is concerned with the ways in which<br />
nonnatives do things with words in a second language. The phenomena investigated<br />
from this perspective may relate to both pragmatic and discoursal knowledge; in<br />
practice, the emphasis has been on contrasting native and nonnative speech act<br />
performance, on detecting learners' inappropriate speech act realization and accounting<br />
for pragmatic errors or deficits in terms of both source and communicative<br />
effects (see Blum-Kulka, House, & Kasper, 1989 for an overview).<br />
We deviate from this tradition here in several respects. Here we shall be concerned<br />
with interlanguage pragmatics phenomena occurring in the speech of bilingual<br />
speakers, for whom bilingualism is a day-to-day social reality. Ours is not a<br />
learners' population; we shall be looking at families interacting around the dinner<br />
table in three groups of families: native-born Israelis, native-born Americans, and<br />
our target group, American immigrants to Israel. The immigrant group is composed<br />
of middle- to upper-middle-class families living in Israel from 9 to 19 years. The<br />
discoursal styles used by family members at the dinner table in this group have been<br />
compared with the dinner-table talk of socioeconomically similar native Israeli and<br />
native American families. All members of the immigrant families are bilingual<br />
(English-Hebrew); both languages are being used at the dinner table, with English<br />
being dominant.<br />
The adults in these families are hence speaking mostly in their native language<br />
(English) and to a lesser degree in their second language (Hebrew). The reverse is<br />
true for the children: most children partake in family discourse in what for them is a<br />
second language (English), reverting occasionally to Hebrew, their native language.<br />
Is such a peculiar bilingual language situation the relevant context to look for<br />
interlanguage pragmatics phenomena After all, we have been used to thinking<br />
about interlanguage as an L2 (second language) specific notion; this assumption<br />
underlies all studies to date, whether linguistically or pragmatically oriented. Yet<br />
there is another context where systematicity in LI "errors" does come under consideration:<br />
the domain of attrition studies (see Weltens, 1987 for a review). As in our<br />
196
American-Israeli Families at Dinner 197<br />
case, first language attrition studies focus on immigrant populations. But they differ<br />
from ours in two important respects: since they are concerned with changes that<br />
occur over time in a first language, it is this first language that serves as the criterion<br />
norm. As reported by Wei tens, the focus in first language attrition studies is on<br />
language loss as an intergenerational phenomenon over time, as established from<br />
questionnaires, or on particular linguistic phenomenon at a given point in time<br />
(e.g., Sharwood-Smith, 1983). In our case, language use in the immigrant families<br />
is studied at a particular point in time and is compared to both native American and<br />
native Israeli use. Furthermore, differing from that of attrition studies, our focus is<br />
on pragmatic, rather than on linguistic, phenomena.<br />
The bilingual families studied are, in effect, in contact with two incongruent<br />
pragmatic systems, each realized by a culturally specific style of language use. The<br />
possibility of a two-way interference under such conditions was already recognized<br />
by Weinrich (1953, 1): "Those instances of deviation from the norms of either<br />
language which occur in the speech of bilinguals as a result of their familiarity with<br />
more than one language, i.e. as a result of language contact, will be referred to as<br />
interference phenomena." We propose to demonstrate that indeed, as implied by<br />
Weinrich, language contact creates bi-directional effects, both from the first language<br />
to the second and vice versa. The latter type of effect, namely, interference<br />
from the second to the first language, has been noted in the very few studies to date<br />
that addressed this issue. 1 We shall claim that rather than just showing traces of<br />
"interference," the mutual influence between two sociolinguistic systems in the case<br />
studied results in the emergence of a specific, pragmatic interlanguage.<br />
We are interested in use rather than in usage (Widdowson, 1978), specifically in<br />
the pragmatic norms that govern a particular speech community's interactional style<br />
(Hymes, 1974) and its modes of pragmatic socialization. Studies in the ethnography<br />
of speaking and cross-cultural pragmatics have clearly shown that ways of speaking<br />
are culturally determined and that culture-specific interactional styles differ along<br />
dimensions such as levels of directness in speech act realization (e.g., Tannen,<br />
1981; Blum-Kulka, House, & Kasper, 1989), interpretive strategies and signaling<br />
devices (Gumperz, 1982), and conceptions and practices of "polite" behavior<br />
(Basso, 1979; House & Kasper, 1981). Language in use is a powerful medium of<br />
sociocultural socialization. Children acquire tacit knowledge of systems of belief<br />
and cultural ways of speaking through exposure to and participation in languagemediated<br />
interaction in the home (Gleason & Weintraub, 1976; Heath, 1983;<br />
Schieffelin & Ochs, 1986; Ochs, 1988; Schieffelin, 1990). The cross-cultural comparison<br />
of Israeli to American Jewish families in our sample reveals the two groups<br />
differing on several dimensions of language use (Blum-Kulka & Katriel, 1991;<br />
Blum-Kulka & Snow, 1992) and practices of pragmatic socialization (Blum-Kulka,<br />
1990).<br />
In previous work that focused on requestive behavior we found that the families<br />
we observed manifest a bicultural, or intercultural style systematically different<br />
from both the Israeli and American patterns. For example, immigrant families<br />
realize requests less directly than native Israelis, but more directly than Americans<br />
living in the United States. Furthermore, this intercultural style is realized regardless<br />
of the language spoken. Speakers manifest characteristics of this style when
198 Discourse Perspectives<br />
speaking both their first and second languages (Blum-Kulka, 1988; Blum-Kulka,<br />
1991).<br />
In this chapter we shall address these issues from a metapragmatic perspective.<br />
Our underlying assumption is that a speech community's interactional style is indexed<br />
both by the ways in which its members realize different speech acts as well as<br />
by the ways in which they comment on language use: metapragmatic comments<br />
(MCs) can reveal the pragmatic norms governing use and socialization. As in the<br />
case of requests, our findings show the metapragmatic discourse of the immigrant<br />
families to differ meaningfully from parallel discourse in both the American and<br />
Israeli families, whether realized in a first or a second language.<br />
The thrust of our argument is that under certain conditions, interlanguage pragmatics<br />
can be realized in the first language. Furthermore, for bilinguals in conditions<br />
of a twofold cultural contact, pragmatics may be the first area affected by<br />
exposure to the sociocultural influence of the second language.<br />
This argument is developed as follows: we shall (1) describe the background to<br />
the study and our methods of data analysis; (2) present our findings with regard to<br />
inter- and intra-group diversity in metapragmatic discourse; (3) discuss the unique<br />
characteristics of the bilingual families' IL metapragmatic discourse, as manifested<br />
by metacomments on language usage in the course of dinner-table conversations;<br />
and (4) examine the degree of linguistic and pragmatic awareness of bilingual<br />
speakers in general by considering data from interviews with the immigrant families<br />
in our sample.<br />
Background<br />
The 24 families studied are all native-born, middle- and upper-middle-class Jewish<br />
families, with two to three school-age children. Families were selected through the<br />
snowball technique. The criteria for inclusion were: (a) college education for both<br />
parents, (b) professional occupation for both parents, (c) European family origin for<br />
both parents, (d) being religiously nonobservant, (e) being native-born Israeli or<br />
American (for both immigrants to Israel and families taped in the United States). For<br />
American immigrants to Israel, we required a minimum of nine years of residence<br />
in Israel.<br />
As can be seen in Table 10.1, most children (16 out of 29) in the immigrant<br />
families were born in Israel and have not spent more than a month or two in an<br />
English-speaking environment. Six children have spent two years abroad and two<br />
have spent five. Out of the four children born in the United States, two came to<br />
Israel at the age of one, one at the age of two, and one at the age of four and a half.<br />
Yet though Hebrew has become unquestionably the dominant language 2 for all<br />
of these children, English plays an important role. English is the main language<br />
spoken at home, even though it is often mixed with Hebrew (Olshtain & Blum-<br />
Kulka, 1989). Adults and children alike practice their bilingualism not only in<br />
speaking (speaking English at home and Hebrew at work and at school) but also in<br />
reading. The adults all report reading in English for pleasure, and 13 out of the 14<br />
school-age children report reading in both languages.
American-Israeli Families at Dinner<br />
199<br />
Table 10. 1. Children in the Immigrant Families<br />
Israeli born<br />
17<br />
6<br />
2<br />
N = 25<br />
American born<br />
2<br />
1<br />
1<br />
N = 4<br />
N = 29<br />
Time spent abroad (years)<br />
0<br />
2<br />
5<br />
Age of immigration (years)<br />
1<br />
2<br />
4.5<br />
The families 3 were observed in their homes (Americans in the United States,<br />
native Israelis, and immigrants in Israel). We have recorded three dinner-table<br />
conversations (one by video and two by audio) for each family. An observer was<br />
present at all dinners. 4 Three types of analysis have been carried out on these data to<br />
date: first, transcripts were coded for the use of acts of social control (e.g., all<br />
directives) by all participants in the first 20 minutes of all dinners, yielding a corpus<br />
of 4120 control acts. A subset of the 903 control acts (cf. Ervin-Tripp & Gordon,<br />
1986) issued by parents to children was examined for parental style of politeness in<br />
issuing directives. Second, this analysis was complemented by a study of the metapragmatic<br />
comments (cf. Becker, 1988) issued by parents to children during the<br />
same dinners (Blum-Kulka, 1990). A third analysis focused on code-switching and<br />
code-mixing in the immigrant families (Olshtain & Blum-Kulka, 1989). In the<br />
following, we shall draw on results from all three.<br />
The families were also interviewed at length in their homes; the interviews were<br />
meant to probe mainly cultural preferences with regard to issues of children's<br />
pragmatic socialization, extended in the case of the immigrant families to include<br />
questions about language learning and bilingualism. The insights gained from the<br />
interviews with the immigrants are incorporated in the present discussion, having<br />
been found to be particularly helpful in dealing with issues of metapragmatic awareness.<br />
The immigrant families studied present a case par excellence of a bilingual and<br />
bicultural social reality. They maintain in practice and in attitude close contact with<br />
two languages and two cultures, and hence are exposed to two systems of pragmatic<br />
and discoursal rules. There is, of course, individual variation among the families,<br />
some showing a higher degree of convergence (Giles, 1979) to the Israeli culture<br />
than others. Yet as illustrated by the two following extreme cases on the "convergence"<br />
continuum in our sample, even the least "American" family in the group<br />
maintains an intimate link with the language and cultural heritage of their country of<br />
origin. Consider the case of the Bells and the Darnos.<br />
The Bells represent the Anglophone extreme of our bilingual continuum. They<br />
have chosen to live in a Jerusalem neighborhood known for its density of Americans<br />
and Canadians. They claim that at least on the street they live, English is the
200 Discourse Perspectives<br />
dominant language even for the children. And indeed, during the interview we were<br />
suprised to detect a trace of American accent in the Hebrew speech of their Israeliborn,<br />
8-year-old daughter. At work, the husband uses Hebrew only (at a very<br />
advanced level, given his occupation), at home, English only. Working as an instructor<br />
of English, the wife speaks English for most of both her professional and<br />
private life. The family spends most summers in the United States, and all the<br />
children are balanced Hebrew-English bilinguals. While the Bells invest familial<br />
effort in preserving English, the Darnos vacillate between a wish to raise their<br />
children as bilingual speakers and the no-less-strong wish to acculturate to the<br />
Hebrew-dominant Israeli culture. They mostly, but not entirely, speak English at<br />
home and Hebrew at school and work. For the children, Hebrew is the preferred and<br />
unquestionably the better known language; their English is more advanced than that<br />
of other Israeli children their age, but it still definitely remains a second language.<br />
For the parents, mastering Hebrew to the degree that would allow them to read<br />
Hebrew literature in the original presented a challenge they faced within the family<br />
context. Thus the father worked through all his son's literature assignments at high<br />
school, a process he claimed to have proven extremely beneficial for his own<br />
acculturation.<br />
Notions of Interactional Style: Metapragmatic Comments<br />
One way of capturing cross-cultural diversity with regard to interactional style is to<br />
observe cultural differences in actual verbal and nonverbal behavior systematically.<br />
The realization of requests in three groups of families, mentioned above (Blum-<br />
Kulka, 1988; 1990), provides such an example. Cultural notions of style can be<br />
further captured by what people say they do with words, namely, by their metapragmatic<br />
discourse. Metapragmatic comments (cf. Becker, 1988) made in regard to<br />
linguistic and conversational behavior can reveal the pragmatic norms underlying<br />
such behavior. In Silverstein's terms, metapragmatic comments can be seen as<br />
"non-referential indexes" (Silverstein, 1976). Just as language use can index nonreferentially<br />
social dimensions such as degree of deference, comments made in<br />
regard to the perceived violation of a conversational norm will index for the members<br />
of the particular speech community the network of interactional norms that<br />
govern language use in that community. Such comments relate to the smooth flow<br />
of discourse by explicit bids for turn, by bringing to attention breaches of turntaking<br />
rules, or by focusing on perceived violations of conversational maxims, as<br />
laid out by Grice (1975). Metapragmatic comments are one of the explicit ways in<br />
which members discuss criteria for verbal appropriateness. In family discourse,<br />
metapragmatic comments further serve instrumentally in socializing children to the<br />
pragmatic norms of the given culture (Blum-Kulka, 1990).<br />
Research in the development of communicative competence in childhood reveals<br />
that the process is very much dependent on culturally differential adult input.<br />
Caretakers across cultures differ in their preferences in amount and type of talk<br />
expected from children, and in the degree to which they deem it essential to explicitly<br />
teach language to children (Heath, 1983; Schieffelin & Ochs, 1986; Ochs, 1988;
American-Israeli Families at Dinner 201<br />
Schieffelin, 1990). In Western societies it is rather the manner of speaking, namely<br />
politeness, that tends to be attended to explicitly. Thus American middle-class<br />
mothers systematically prompt the use of politeness formulas and sanction perceived<br />
lack of polite behavior (Gleason & Weintraub, 1976; Grief & Gleason, 1980;<br />
Gleason, Perlman, & Grief, 1984; Becker, 1988; Becker, 1990; see also Schmidt,<br />
Chapter 1). Yet "politeness" is a culturally relative notion that serves in folk etymology<br />
as a cover term for social criteria applied to the distinction between socially<br />
appropriate and inappropriate verbal and nonverbal behavior. Considered as a measure<br />
of pragmatic socialization, the analysis of metapragmatic comments can thus<br />
show parents' culturally differentiated notions with regard to the style of linguistic<br />
politeness required from children (Blum-Kulka, 1990). 5<br />
In the following we shall compare the metapragmatic discourse of both adults<br />
and children in the three groups of families studied. Several issues underlie the<br />
discussion. First, what is the degree of inter-group diversity in pragmatic norms, as<br />
revealed by the metapragmatic discourse of adults and children We hypothesize<br />
that Americans and Israelis differ in pragmatic norms, and that immigrant families<br />
differ from both, abiding by an IL system of pragmatics. Second, if our hypothesis<br />
is confirmed, what is the degree of intra-group (child > adult) conformity within<br />
each group Do children in the three groups follow in their parents' footsteps in<br />
types of verbal behavior they sanction We further hypothesize that since children<br />
can be expected to be socialized to culturally determined ways of speaking, the<br />
metapragmatic discourse of the children in all groups will tend to replicate trends<br />
detected in the parent's discourse. Children in the immigrant families will be no<br />
exception, showing IL pragmatics being transmitted across generations. Furthermore,<br />
if our hypotheses are supported, how far are the trends depicted for the<br />
immigrants dependent on the language spoken We shall claim that the IL pragmatics<br />
phenomena depicted for the immigrant families are realized independently of<br />
language choice.<br />
Method<br />
Transcripts of dinner-table conversations were examined for the occurrence of metapragmatic<br />
comments. 6 MCs related to verbal behavior were classified into three<br />
distinct categories (see also Blum-Kulka 1990 for an additional "behavioral" category):<br />
Talk-Regulation. This category includes comments made to regulate the smooth<br />
flow of turn-taking. They include bidding for turn ("can I say something"), allocation<br />
of turns ("Okay, David, your turn"), negating a turn ("wait, Talya is first"),<br />
upholding a turn ("I'm talking now") and checking listener's attention ("are you<br />
listening").<br />
Maxim Violation. This category includes comments signaling perceived violation<br />
of one of the four Gricean (Grice, 1975) maxims (see also Pellegrini, 1987).<br />
Comments in regard to the maxim of relevance prompt the addressee to respond to a<br />
conversational demand ("Beth, there is a question on the floor") or delegitimize<br />
mention ("one should not say that"). Comments in regard to quality cast doubt on<br />
the truth-value of a speaker's proposition (e.g., in response to a child's reporting
202 Discourse Perspectives<br />
having seen a "GIANT turtle," the mother inquires, "How giant is giant Have you<br />
really seen it"). Comments in regard to quantity set limits to degree of informativeness<br />
of stated propositions ("we heard that"), but also elicit information or just talk<br />
when felt lacking ("aren't you participating today"). Comments in regard to manner<br />
prompt the use of politeness formulas ("say 'please'"), correct ungrammatical<br />
language, note improper forms of address and reference (CHILD:" . . . that stupid<br />
teacher" MOTHER: "Who" CHILD: "Varda, the Math teacher"), and sanction the use<br />
of slang and vulgar language.<br />
Metalinguistic Comments. This category captures talk about language; it includes<br />
queries and responses about word meanings, as well as comments topicalizing<br />
language, including cross-linguistic comparisons ("Did you know the Eskimos<br />
have a hundred words for 'snow'").<br />
Findings: The Degree of Inter- and Infra-Group Diversity<br />
Americans show the highest degree of metapragmatic awareness: more than half of<br />
the comments (54%, n = 1052) belong to American speakers, the other half being<br />
divided between Israelis and immigrants. This pattern is repeated for both adults<br />
and children. Through the course of two dinner-table conversations, the average<br />
number of MCs per family member present (including non-MC speakers) is 20 for<br />
American adults, 10 for immigrant adults and 8 for Israelis, and for the children 14<br />
(American), 7 (immigrants) and 6 (Israeli) respectively.<br />
GROUP AND AGE EFFECTS<br />
To test the statistical significance of intra- and inter-group diversity in pragmatic<br />
norms, two independent scores were computed per speaker: (1) a talk-regulation<br />
(TRE) score, computed as the ratio of TR from maxim-violation (MV) and metalinguistic<br />
(ML) MCs; and (2) metalinguistic score (MLI), computed as the ratio of<br />
ML MCs from MV. We shall refer to these scores respectively as TRE and MLI. 7<br />
The mean scores for both categories are presented in Table 10.2. The first point<br />
to note in Table 10.2 is that MC discourse is not equally distributed among the<br />
participants. While among the American and Israeli adults we did not find any<br />
parent present who did not make at least one comment, among the immigrant adults<br />
7 parents made no metapragmatic comments, and among the children 2 Americans,<br />
4 Israeli and 7 children in the immigrant families did not partake in this type of<br />
discourse. Since the TRE and MLI scores reflect preferences for actual use, the<br />
following analyses are based on the subsample of participants found to engage in<br />
metapragmatic discourse. The second point to note is that except for the immigrant<br />
children's score for TRE, all other immigrant scores consistently fall between the<br />
Israeli and American means.<br />
The results show that Americans and Israelis differ significantly on both TRE<br />
and MLI measures, this trend being repeated for adults and children (for adults,<br />
7129] = 3.64, p < .001, for children 7[30] = 2.36, p < .05). For the immigrants,<br />
TR scores for the adults differ significantly from both Israeli and American ones.<br />
Immigrant adults have significantly higher TRE scores than Israelis (T[22] = 3.81,
American-Israeli Families at Dinner 203<br />
Table 10.2. Group) Differences in Talk-Regulation (TR) and Metalinguistic (ML) Scores<br />
for Adults and Children<br />
Category<br />
TRE<br />
Adults<br />
Standard deviation<br />
Children<br />
Standard deviation<br />
MLI<br />
Adults<br />
Standard deviation<br />
Children<br />
Standard deviation<br />
/; < 0.001*.<br />
p < 0.005**.<br />
p < 0.050***.<br />
ISR<br />
Mean [N]<br />
0.06 [16]<br />
0.06<br />
0.42 [15]<br />
0.56<br />
2.70 [14]<br />
3.90<br />
2.60 [13]<br />
3.30<br />
Group<br />
IVTM AM Significance<br />
Mean \N] Mean (N] IS/AM IM/AM IM/IS<br />
0.18 [9] 0.59 [16] * *** *<br />
0.14<br />
0.17 [14] 1.80 [17] *** ** —<br />
0.35<br />
1.50 [8] 0.49 [16] *** * —<br />
0.99<br />
2.40 [11] 0.38 [17] *** *** —<br />
3.40<br />
p < .001) while they score significantly lower than Americans (T[22] = 2.12, p <<br />
.05). Immigrant adult MLI scores are higher for immigrants than for Americans<br />
(7122] = 3.81, p < .001) but similar to those of Israeli adults. On the other hand,<br />
the children in the immigrant families differ significantly from American children<br />
on both measures (for TRE, T[29] = 2.70, p < .01, for MLI, 7"[26] = 2.42, p <<br />
.05), but not from group 1 Israeli children.<br />
The effects of group membership (Israeli/immigrant/American) and Age<br />
(child/adult) on these scores were tested via a series of paired two-way analyses of<br />
variance. Our expectation for cross-cultural diversity was fully confirmed. Israeli<br />
versus American group membership has significant main effects on both MLI and<br />
TRE. For MLI, only group membership is the significant distinguishing factor:<br />
F(56,l) = 12,2, p < .005. Age had no effect, showing adults and children following<br />
similar patterns, as predicted. For TRE, both group membership and age matter<br />
(for group, F(56,l) = \5A,p< .005, for age, F(56,l) = 10.87, p< .05). There<br />
was no interaction effect.<br />
Against this background of cross-cultural diversity, the immigrants are shown to<br />
occupy a peculiar interim position, as predicted. Though they speak mostly English,<br />
Group membership sets them apart from Americans on both MLI and TRE. For<br />
Metalinguistic discourse, the main effect for group is F(48,l) = 11.1, p < .005.<br />
This pattern being similar for both adults and children, we found no effect for age.<br />
On the measure of talk-regulation both group membership and age are important:<br />
for group F(48,l) = 7.3,p< .005), for ageF(48,l) = 4.5,p< .010. There was no<br />
interaction effect.<br />
A different picture emerges in comparing the immigrants to the Israelis. For<br />
both, we find a similarly high degree of metalinguistic awareness for both adults
204 Discourse Perspectives<br />
and children. Hence it is not suprising that we found no effect for Group (nor Age)<br />
on the MLI measure. On the measure of talk-regulation, we found no effect for<br />
group but a nearly significant effect for age, (F[42,1 = 3,7, p = .06). In interpreting<br />
this result it should be borne in mind that immigrant adults (but not children)<br />
score significantly higher on TRE than Israeli adults (Table 10.2).<br />
These findings confirm our earlier hypothesis that Israelis and Americans differ<br />
significantly on both talk-regulation and metalinguistic measures. Against this<br />
background of cross-cultural diversity in pragmatic norms, the immigrants seem to<br />
waver between the two. We expected to find them to differ from both, thereby<br />
indicating an IL system of pragmatics, and this expectation was confirmed for the<br />
adults. Immigrant adults differ significantly from both Israeli and American adults<br />
in their talk-regulation discourse (T-tests), but the difference is not big enough to<br />
show a Group effect (analysis of variance). On the measure of metalinguistic MCs,<br />
cross-cultural diversity is confirmed again, showing Israelis and Americans to have<br />
different attitudes to the domain of language as topic. In this regard, the immigrants<br />
differ from the Americans, resembling the Israelis in their degree of metalinguistic<br />
awareness. As will be elaborated, there are different, group-specific motivations for<br />
this phenomenon.<br />
Our second hypothesis is partially confirmed: children in all three groups conform<br />
to adult patterns in regard to metalinguistic discourse. But they differ from<br />
adults in attitudes to talk-regulation; this could mean that for this aspect of discourse<br />
children either have different notions, or different needs. These possibilities will be<br />
explored below. The children in the immigrant families do not differ in a statistically<br />
significant way from other Israeli children in regard to talk-regulation, as do their<br />
parents. Yet qualitative analysis of these children's talk-regulation behavior, to be<br />
discussed below, does suggest a unique pattern, in line with the "IL replication"<br />
hypothesis.<br />
Finally, in considering the metapragmatic discourse of the immigrants, it is<br />
important to note that the overall majority of MCs (83%) are uttered in English, not<br />
Hebrew. Only 10% of comments are made in Hebrew, with the remaining 7% made<br />
in a mixture of English and Hebrew, manifesting the language choice patterns<br />
typical of this group in general (Olshtain & Blum-Kulka, 1989). Hence the fact that<br />
immigrant adults are found to differ from Americans in their metapragmatic discourse<br />
means that they activate in their first language pragmatic norms adapted from<br />
the second. But native language pragmatic norms seem at work too; in certain<br />
respects, immigrant adults differ from Israelis as well. The metapragmatic discourse<br />
of the bilingual families hence is seen to exhibit the first assumption underlying all<br />
interlanguage studies: it is systematically different from both LI and L2.<br />
ORDER OF PREFERENCES<br />
The "in between" character of the immigrant families' IL pragmatics system is<br />
further highlighted by considering the three groups' order of preference for domains<br />
of discourse attended to. We have analyzed the distribution of the three categories of<br />
MCs, uttered at the dinner table by all members of the family in the three<br />
groups. For this analysis MCs have been seperated by speaker's role (adult or
American-Israeli Families at Dinner 205<br />
Fig. 10.1. Adults' metapragmatic comments in three groups.<br />
child) and divided by type of comment (discourse management/maxim-violation/<br />
metalinguistic).<br />
Adults. The marked difference between American and Israeli adults stems from<br />
the culturally differing emphasis on conversational behavior as compared to language<br />
as topic. In the American adults' discourse, the two conversational behavior<br />
categories (maxim-violation and talk-regulation) take up 82% of all metapragmatic<br />
discourse space; in the Israeli discourse, they constitute less than half of all comments<br />
made (38%) (Fig. 10.1). The relative importance attached to conversational<br />
norms (54%) versus language (45.5%) shows the Immigrants occupying an interim<br />
position between the other two groups: this aspect of discourse is more important for<br />
them than for Israelis, but less important than for Americans.
206 Discourse Perspectives<br />
Fig. 10.2. Children's metapragmatic comments in three groups.<br />
Children. American children share with their parents a very high degree of<br />
metapragmatic awareness in regard to conversational behavior (talk-regulation plus<br />
maxim-violation constitutes 78%), especially when it concerns discourse management<br />
(48.4%), (Fig. 10.2). Israeli children, on the other hand, share with their<br />
parents a relatively high degree of metalinguistic awareness (51%). The children in<br />
the immigrant families reveal a pattern of their own. Similar to Israelis, they find<br />
language important (43%); similar to Americans, they attend to maxim-violation<br />
(31%). Consider the decreasing emphasis put on conversational management by<br />
children in the three groups: Americans 80%, immigrants 52%, Israelis 49%.<br />
Children Compared to Adults. Are children sensitive to the same aspects of<br />
discourse as their parents Are their pragmatic norms similar to those of the adults<br />
Tannen (1981) found that second-generation Greek-Americans who spoke no
American-Israeli Families at Dinner 207<br />
Greek, nevertheless manifested in their conversations in English levels of indirectness<br />
typifying Greek ways of speaking. In a similar vein, we explored the possibility<br />
that the metapragmatic discourse of children born to immigrant parents will<br />
show traces of unique parental input, setting these children apart from other nativeborn<br />
Israelis. Comparing the children's preferences to those of the adults allows us<br />
track the success of pragmatic socialization: children following in their parents'<br />
footsteps in aspects of the discourse attended to may be said to have internalized<br />
their parents' discoursal criteria. Such replication is manifest to some degree in all<br />
groups: as shown by the analysis of variance across the groups, age had a significant<br />
effect only on talk-regulation scores; the lack of such an effect on metalinguistic<br />
scores at least partially supports the replication hypothesis. Comparing adult/child<br />
patterns of distribution lends further support to partial replication:<br />
For the Americans, the children follow in their parents' footsteps in deeming<br />
conversational norms more important than language (see Figures 10.1 and 10.2). A<br />
different trend is depicted for the Israeli children, who vacillate between the domains<br />
more than their parents (but note small number of comments). Interestingly,<br />
replication seems most salient for the Immigrants: in this group, the order of preferences<br />
in matters of pragmatics attended to among the children is the exact mirror<br />
image of the order exhibited by the parents. Metalinguistic comments constitute<br />
almost half of all MCs for both (45.5% for adults, 42.6% for children), followed by<br />
maxim-violation comments (adults 33.3.%, children 38.2%) and talk-regulation<br />
MCs (adults 21.3%, children 19.2%). But in their relatively high metalinguistic<br />
awareness these children also resemble Israeli group 1 children. Hence it may the<br />
case that their metapragmatic discourse is affected by peer-group interaction rather<br />
than by parental input.<br />
The next sections discuss the unique characteristics of this IL metapragmatic<br />
discourse.<br />
Characteristics of Bilingual Metapragmatic Discourse<br />
Talk-Regulation<br />
The adults in the bilingual families pay attention to all three dimensions examined:<br />
discourse management, language, and conversational norms. In regard to discourse<br />
management, the model followed is the mainstream, middle-class American one:<br />
"fair" turn-allocation, discouraging interruptions, and generally orderly systematic<br />
turn-taking seem the discourse corollary of American ideals of individual rights and<br />
equal opportunity for all. American individualism, with all that it entails for mutual<br />
respect for individual space, has been long-remarked upon by outside observers,<br />
such as de Tocqueville, and studied extensively, more recently, from both sociological<br />
(Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985) and sociopsychological<br />
points of view (Brown, 1991). Cross-cultural research has amply shown that the<br />
discoursal manifestation of such ideals, namely, the systematics of turn-taking<br />
(Sacks, Schegloff, and Jefferson, 1974) are by no means universal (Reisman, 1974;<br />
Philips, 1983). Regular turn-taking is not necessarily adhered to even in all domains
208 Discourse Perspectives<br />
of American conversations (Edelsky, 1981). But as the Jewish-American adults'<br />
metapragmatic discourse shows, the rules seems to have an unquestionable normative<br />
status in American society. In this respect, Jewish-Americans seem to approximate<br />
mainstream American ways of speaking; by paying attention to this particular<br />
aspect of discourse, the immigrant adults reveal a sensitivity going back to their own<br />
language socialization. It is this positively evaluated feature of American interaction<br />
that they are trying to preserve in their children's talk. It is important to note that by<br />
its very nature, this feature is not bound to any specific language; potentially it is<br />
equally applicable to conversations in Hebrew or English.<br />
The ethnographic interviews with the families occasioned clear formulations of<br />
the adults' prevalent attitude toward talk-regulation, specifically turn-taking. Asked<br />
whether they discourage interruptions by children, we received from one of the<br />
families the following reply:<br />
FATHER: 8 It's very important that they don't interrupt others. It's a principle here. In<br />
fact, sometimes we say to the other, "Wait, your brother is still talking," or<br />
"Jessica is still talking."<br />
MOTHER: And sometimes it ends up where they have to raise their hands, even (laughter).<br />
FATHER: Oh, yeah.<br />
The findings also show Israeli and American children to differ markedly in<br />
degree of attention paid to talk-regulation, with the Israeli bilingual children in the<br />
immigrant families paying the least attention to this aspect of discourse. This seems<br />
to indicate a failure on the part of the immigrant parents to transmit their basically<br />
American talk-regulation norms to their children. But the analysis of talk-regulation<br />
comments by type, 9 suggests otherwise. Of specific interest are the ways in which<br />
bids for turn are accomplished.<br />
American children make explicit bids for turns: 47.4% (n = 116) of their talkregulation<br />
comments are permission requests for a turn, often combined by a comment<br />
indicating awareness of the rule, such as "Daddy, can I say something Is it<br />
my turn" (Blum-Kulka, 1990). Israeli children bid less for turns (28% of talkregulation<br />
MCs, n = 25) and when they do, instead of asking permission to talk,<br />
they try to secure their interlocutors' attention by uttering a clearly addressed preexchange<br />
move, preparing the ground for the ensuing talk (Edmondson, 1981):<br />
"Daddy, do you know what happened today" The children in the immigrant families<br />
seem to vacillate between the two techniques. Bids for turn constitute 78.9% of<br />
their talk-regulation MCs (n = 19), showing them particularly concerned with<br />
obtaining permission to talk. But they avoid the American style of explicit turntaking<br />
metatalk. Instead, we find attempts to secure attention that look like indirectly<br />
performed permission-requests:<br />
NAVA (age 5): Ima (Mother), now can I tell you what they do<br />
(Talk round continues.)<br />
NAVA: Should 1 tell you what they do, what the GANANOT (nursery-school<br />
teachers) do Should I tell you what they do<br />
Subsequently Nava's attempts act as a pre-exchange move, similar to the Israeli<br />
child's attempts. Yet semantically she is closer to the American child, using "tell" as
American-Israeli Families at Dinner 209<br />
a "linguistic action-verbial" (cf. Verschueren, 1985). Hence the pattern exhibited is<br />
an interim, interlanguage-type one, showing influence from both contact cultures.<br />
Maxim-Violation Comments<br />
Maxim-violation comments can relate to the perceived violation of any one of the<br />
four Gricean maxims: relevance, quality, quantity, and manner. The distribution of<br />
maxim-violation comments by type (Figures 10.3 and 10.4) shows American and<br />
Israeli adults to differ markedly in preferences for the type of maxim violation most<br />
likely to be noticed verbally. The immigrant adults seem to differ more in this case<br />
from the Americans than from the Israelis.<br />
Americans emphasize foremost adherence to the maxim of quality, namely,<br />
factuality of the proposition asserted (60.2%). Both Israelis and immigrants distribute<br />
their attention more equally between all four aspects of discourse. For Israelis,<br />
manner (mainly politeness) is highly attended to (40%), followed by quality<br />
(33.7%) and relevance. Degree of informativeness and volubility (quantity) are<br />
rarely spoken about (6%). The immigrant adults fall between the two other groups<br />
in their order of preference. Quality is very important (45.8%), followed by attention<br />
to violations of manner (31.4%) and relevance (20.4%). Attention to quantity<br />
of talk is negligable (2.1%).<br />
The relative salience of comments in regard to quality in all three groups<br />
deserves special attention. Prompting adherence to the factuality of stated propositions<br />
is by no means a universal norm of pragmatic socialization. Heath (1983)<br />
found an emphasis on this aspect both among the white working-class people of<br />
Roadville and the middle-class "townspeople," but not in the black community of<br />
Trackton. In Roadville, children are expected to "stick to the truth" in retelling<br />
nonfictive stories of shared events that the adults are aware of; middle-class parents<br />
also expect children to share information only the children know, but as in Roadville,<br />
probe their children to recognize contradictions in such factual accounts.<br />
Insistence on factuality is interpreted by Heath as manifesting literate traditions. On<br />
the other hand, in the black working-class community of Trackton, oral traditions<br />
dominate; hence adherence to factuality is suspended in favor of "storyness": the<br />
purpose of narratives above all is to entertain and "to establish the story-teller's<br />
intimate knowledge of truths about life larger than the factual details of real events"<br />
(Heath, 1983, 188).<br />
True to their middle-class literate background, all families in our group require<br />
children to learn to distinguish fact from fiction. In the following example, the<br />
child's report is challenged by both parents, and found counterfactual.<br />
ILANA (age 13.5): Did you have a good time at Leora's<br />
NAOMI (age 5): I wasn't at Leora's.<br />
FATHER: Yes you were, this afternoon=<br />
MOTHER: =You weren't Where did you go from CAN (nursery school)<br />
NAOMI: To a home.<br />
MOTHER: Not to this home.<br />
NAOMI: Yes, to this home.<br />
MOTHER: I wasn't here.
210 Discourse Perspectives<br />
Fig. 10.3. Types of maxim-violation comments: adults.<br />
The immigrant families seem to abide in the respect by the middle-class norm of<br />
Heath's "townspeople." Children are invited to tell A-events (cf. Labov & Fanshel,<br />
1977) that only they know, but must beware of contradictions. In American families<br />
adherence to factuality supersedes all other maxims. The immigrant families, as<br />
well as the Israelis, find other aspects of discourse also worthy of attention. Thus<br />
parents also comment on the relevance of a child's contribution (MOTHER: "I don't<br />
want to hear it, Naomi, I don't want to hear it now"), and on its manner of delivery<br />
(FATHER: "... but don't shout").<br />
A very similar picture of cross-cultural diversity emerges when we compare<br />
American and Israeli children. American children, like their parents, are foremost<br />
concerned with quality (65.6%); Israeli children, like their parents, when they do<br />
comment (n equals only 19) stress manner (45%). But what about the children in the
American-Israeli Families at Dinner 211<br />
Fig. 10.4. Types of maxim-violation comments: children.<br />
immigrant families They share with other Israeli children the same types of input in<br />
Hebrew from peer interaction, school, and the media. If they are found to differ<br />
from other Israeli children, it would mean that the 11 pragmatic norms of the parents<br />
are transmitted through bilingualism to the next generation, signaling the importance<br />
of parents' input as a dominant source of pragmatic socialization. And indeed,<br />
when it comes to Gricean maxims, the children in the immigrant group differ from<br />
the group 1 Israeli children.<br />
Differing from the Israelis, who attend primarily to manner, the bilingual children<br />
stress in their comments the importance of factuality even more than their<br />
parents (72.7% compared to 45.8% in parents' discourse), and resemble American<br />
children in having this aspect of discourse at the top of their list of preferences for<br />
maxim-violation MC comments.
212 Discourse Perspectives<br />
Comments related to quality are made mostly among siblings, challenging the<br />
truth-value of a statement by "lo yaxol lihyot" (can't be) or simply "that's not true":<br />
DAVID (age 7.5): She plays XAMES AVANIM (marbles).<br />
NAOMI (age 12): We don't play. We play after you go to sleep.<br />
DAVID: No, you play before I go to (sleep).<br />
NAOMI: (That's) not true. Last time we didn't.<br />
Comments made to parents are phrased more indirectly:<br />
DANA (age 14): You know you're exaggerating.<br />
FATHER: I'am exaggerating What<br />
Comments on relevance in both the Israeli and the immigrant families mostly<br />
take the form of restricting undesirable topics, but are less frequent in the immigrant<br />
than in the Israeli data. Thus for example, Naomi, age 12, interrupts her brother's<br />
(age 7.5) dramatic account of a thriller with:<br />
NAOMI: Don't say it because maybe they want to see it.<br />
Similar to American children, this group pays little attention to manner. Israeli<br />
children, on the other hand, find manner very important. Typical examples in the<br />
Israeli data are comments to siblings sanctioning "bad" words, such as a 6-year-old<br />
telling her twin sister not to say "mag'il" (disgusting) about food, or a 10-year-old<br />
reacting to her 12-year-old sister's comment "she was a shitty teacher" by "Shlomit,<br />
you don't say words like that here."<br />
Metalinguistic Comments<br />
In their type of metalinguistic awareness the bilingual families manifest a pattern of<br />
their own. Given their active bilingualism and the adults' recent history of second<br />
language learning, it seemed reasonable to expect from these speakers a relatively<br />
high level of explicit attention to matters of language. And indeed, the bilingual<br />
families exceed the Americans by far in their degree of language awareness though<br />
they fall behind the Israeli norm. For both immigrant adults and their children,<br />
metalinguistic comments constitute almost half of all metapragmatic talk, while for<br />
Americans such comments do not exceed a fifth of metapragmatic discourse. For<br />
Israelis, matters of language figure as the most prominent aspect of discourse<br />
commented on, for both adults and children. (Figures 10.1 and 10.2).<br />
The high level of attention to matters of language in the Israeli families confounds<br />
the issue of linguistic awareness as a hallmark of bilingualism. This is due to<br />
the peculiar history of the revival of Hebrew, which provides the sociolinguistic<br />
context for the Israeli preoccupation with matters of "normative" versus "acceptable"<br />
language use (Rabin, 1976). Within the Israeli families, speakers are very<br />
concerned with matters of grammaticality (i.e., how to form the plural of kalba<br />
(bitch): "Did you know one should say [klavot] and not [kalbot]"), but also with<br />
word-meaning (in response to children's querying the meaning of loan-words (dilemma)<br />
or just rare ones, word origins ("where does x comes from"), pronounciation<br />
(how Ixl & Irl are pronounced in different registers), and the acceptibility of
American-Israeli Families at Dinner 213<br />
slang. On the other hand, American speakers, when occasionally they do topicalize<br />
language-related issues, make general comments in regard to the language of others,<br />
not their own ("Did you know the Eskimos have a hundred words for 'snow'").<br />
A close look at the metalinguistic comments made in the bilingual families<br />
reveals several unique features. First, though clarifying word meanings to children<br />
is a favorite pastime, in these families it is carried out in a multilingual fashion. Both<br />
Hebrew and English words may be explained in the language the query was made in<br />
("what's garlic It's a spice"), or explained and translated, with the degree of<br />
translational equivalence between items in a given semantic field becoming the<br />
topic of lengthy discussions. For example, a 13-year-old's test question (Heath,<br />
1983) to her 5-year-old sister "What is OGER (hamster) Naomi, do you know" sets<br />
in motion a discussion of the biological differences between a "hamster" and a<br />
"guinea pig," combined with the linguistic quest for the Hebrew names of these<br />
animals (Olshtain & Blum-Kulka, 1989).<br />
Word-meaning queries in these families serve differential functions for adults<br />
and children. Children's queries concern the meaning of English words, or ask for<br />
equivalents between the two languages.<br />
RIVKA (age 12): What are physicists [FISIKAIM]<br />
MOTHER: Yes.<br />
The mother confirms thatfisikaim, a phonological loan translation, is indeed the<br />
equivalent of "physicists."<br />
RIVKA (age 12): How do you say SOTER, ABA (Father)<br />
FATHER: Contradictory.<br />
The adults, on the other hand, rely on their children's Hebrew expertise for<br />
clarifying the meaning of Hebrew words:<br />
FATHER: What's KRAZA<br />
DORON (age J5): Poster.<br />
We also find an explicit analytical concern (in Bialystok's sense of "analysis";<br />
Bialystok, Chapter 2) with issues of bilingualism, (including code-switching and<br />
mixing), both on the sociopsychological and educational level. Though our presence<br />
at the dinners no doubt affected the prominence given to language-related issues in<br />
the talk agenda, we know from subsequent interviews with the families that the<br />
concerns expressed are very real. A major issue is language choice. All the families<br />
interviewed expressed a clear English-maintainance language policy; for all, speaking<br />
English at home is a matter of deliberate choice. The parents report having<br />
spoken only English with the children from the time of birth, with the children<br />
hearing Hebrew for the first time sometimes only at the nursery school, as late as the<br />
age of three. Yet through schooling, Hebrew became the children's dominant language,<br />
while English remains that of the parents. At the time of the interviews,<br />
English is estimated as being used in the home about 80% of the time, an estimate<br />
confirmed by examination of the dinner-table conversations in these families (Olshtain<br />
& Blum-Kulka, 1989).
214 Discourse Perspectives<br />
Another issue is the degree of mixing and switching, and attitudes toward both.<br />
Both practices are highly common during the dinner-table conversations but, differing<br />
from Gumperz's findings that show bilingual speakers as unaware of the language<br />
they are speaking (Gumperz, 1982), we found indications of our bilingual<br />
speakers being aware of actually using a hybrid language variety, one that draws on<br />
both languages. This variety is alternatively referred to as "Hibbish" or "Hebrish" as<br />
exemplified in the following exchanges:<br />
ILAN (age 8): (to brother) You're talking gibbish.<br />
FATHER: Hibbish.<br />
ILAN: Oh, yeah.<br />
And a few minutes later, Idan is commenting on his mother's code-switching from<br />
English to Hebrew:<br />
ILAN: Those two words, IMA (Mother), were in Hebrish.<br />
MOTHER: You're right. I changed in the middle.<br />
The language game at work is one of naming; by trying to find an appropriate label<br />
for their own mixed dialect, these families indicate their high degree of linguistic<br />
self-awareness.<br />
Metalinguistic awareness also allows for humor:<br />
FATHER: Nisani, show Marit [the observer] how you speak Hebrew with an<br />
American accent.<br />
NISAN (age 6): LO (no).<br />
OBSERVER: Yes, please.<br />
NISAN: What<br />
FATHER: Say something like when you imitate Mommy and me like ANI RAJITI<br />
(I saw).<br />
NISAN: ABA AMAR PAAM [HATARNEJOL] (Daddy once said [hatarnejol]-<br />
[replacing /j/ for /g/ in pronouncing the Hebrew word for "rooster."])<br />
But the children's high proficiency in Hebrew can also lend itself to a power<br />
game. After being reassured by his parents that he speakes both English and Hebrew<br />
well, a child criticizes his mother's Hebrew. Sister and father alike find the comment<br />
face-threatening, to the extent that the father feels it necessary to come to the<br />
mother's defense:<br />
ILAN (age 8): ANI SAMTI LEV, IMA, ET HATAUYOT SHELAX SHE SHEAT<br />
KORET LI MIXTAVIM VEHAKOL, AT KORET LI 1M TAUYOT,<br />
BLI MI VTA IVRIT. (I noticed, Mom, your mistakes, when you read<br />
me letters and everything, you read with mistakes, without a Hebrew<br />
accent.)<br />
MIRA (age 6.5): It's not nice to say that.<br />
MOTHER: I really am pretty bad.<br />
FATHER: But actually that's not true, Ilan, there are some words that you know,<br />
but I think there are some things that IMA (Mother) can express more<br />
clearly, even in Hebrew still.<br />
ILAN: I know ABA (Father), except I am better in reading and talking in<br />
Hebrew.
American-Israeli Families at Dinner 215<br />
And finally, awareness of code-switching and mixing is combined with ambivalent<br />
feelings toward the benefits of such practices:<br />
INTERVIEWER: MA ATEM MEDABRIM BABAJIT BEDEREX KLAL ANGLIT<br />
(What do you normally speak at home English)<br />
FATHER: We try to speak English.<br />
JESSICA (age 9): We talk part English and part Hebrew. AN1/ANI KMO MEDABERET<br />
NAGID BEEMTSA HAMISHPAT AN1 OSA/OVERET MEANGLIT<br />
LEIVRIT UMEIVRIT LEANGLIT. (I like, say in middle of the sentence,<br />
I go from English to Hebrew and from Hebrew to English.)<br />
FATHER: As a matter of fact, I have been puzzling over it for the last year. I used<br />
to think it was cute when they mixed the sentences together.<br />
MOTHER: I still think it's cute. I do it.<br />
FATHER: I don't. I think it's bad for their Hebrew and their English, so I fuss at<br />
them for the last year now, speak either English or Hebrew but not the<br />
two of them together.<br />
Degree of Language Dependency<br />
It is especially noteworthy that MCs are indices of the immigrants' perception of<br />
good discourse regardless of the language they are uttered in, or comment on. This<br />
can be seen by the tacit agreement of participants to ignore the language they speak<br />
completely if the issue at hand concerns pragmatically proper behavior. In the<br />
following case a violation of the maxim of manner is at issue:<br />
SHARON (age 6): Mommy, TAVI'I LI OD (bring me more) fish.<br />
MOTHER: Excuse me.<br />
SHARON: BEVAKASHA, TNI LI OD (please, give me more) fish.<br />
That the child is mixing English and Hebrew singles out this example as typical<br />
of the bilingual family. In sequence, the exchange is paralleled by several examples<br />
in both the American and Israeli families, all parents being concerned with socializing<br />
children to use politeness markers. But the mixing is irrelevant to the main issue<br />
discussed: after first using a colloquial, slangy expression in Hebrew (literally,<br />
"bring me" instead of "get me" or "give me") the child interprets her mother's<br />
reaction as a reprimand for her lack of politeness, though it might have also been<br />
interpreted as a negative reaction to mixing the two languages. Next, Sharon adds<br />
the required "magic word" (bevakasha/please) and shifts register ("tni li" instead of<br />
"tavi'i li"). But the "fish" remains "fish," and the exchange double-coded, the<br />
mother never saying a word in Hebrew. Pragmatic socialization is hence taking<br />
place independently of the language used for its implementation.<br />
Degree of Metapragmatic Awareness<br />
We have seen that the parents in the bilingual families try to socialize their children<br />
according to a normative system of pragmatic rules shaped after the American<br />
example. By attending explicitly to matters of discourse they highlight the salience
216 Discourse Perspectives<br />
of this domain, helping develop their children's overall communicative competence,<br />
regardless of the language used. The comments are made in English, but are equally<br />
valid for talk in both languages. We also have evidence that in actual use, both<br />
adults and children in these families abide by an IL pragmatics system, which<br />
conforms to neither of the two languages involved (Blum-Kulka, 1988). In issuing<br />
directives, bilingual parents' choice of strategies differs significantly from preferences<br />
shown by both Israeli and American parents, the same being true for children<br />
in these families. How far are the members of this group aware (in Schmidt's sense<br />
of the word, Schmidt, Chapter 1) of any incongruence between the two pragmatic<br />
systems represented by the languages they speak How far are they aware of their<br />
own IL pragmatics These questions will be explored below in the wider context of<br />
the issue of metapragmatic awareness in both first and second language (for a<br />
different perspective see also Schmidt, Chapter 1; Bialystok, Chapter 2). We shall<br />
draw here mainly on monolingual and bilingual speakers' explicit metacommunicative<br />
responses to questions on language use during the ethnographic interviews.<br />
The term metapragmatics seems to imply the capacity of speakers to formulate<br />
explicit rules of speaking. But it is a matter of debate, whether the capacity to "do"<br />
reliable metapragmatics is reserved to professional pragmaticists or, indeed, whether<br />
it can be practiced by all. For example, Schmidt (Chapter 1) argues against<br />
Wolfson's (1989) strong position with regard to the fallibility of native speakers'<br />
intuitions about language use. Our data support his argument that both native and<br />
nonnative speakers do have conscious access to such rules, though there are limitations<br />
to this type, of reflexivity.<br />
With regard to native speakers the argument is far from new. Cultural anthropologists<br />
have long argued for the need to understand cultures emicly from the<br />
"native's point of view" (Geertz, 1976), incorporating as valid data cultural members'<br />
"experience near" (cf. Geertz, 1973) accounts of various dimensions of social<br />
realities, including language use. Participants' understandings are a crucial source<br />
of knowledge since "all social research, and indeed all social life, is founded on<br />
participant-observation" (Hammersley & Atkinson, 1983, 235). A case in point are<br />
anthropological studies focused on terms that name a particular way of speaking,<br />
are part of a speech community's metapragmatic vocabulary, indicating a heightened<br />
cultural attention and articulation with respect to a specific sociocultural domain<br />
(Katriel, 1986). The study of such cultural notions proceeds with the help of<br />
cultural members explicating the meaning clusters associated with such terms<br />
(Spradley, 1979), as exemplified by Katriel's study of dugri (straight) talk in Israeli<br />
speech (Katriel, 1986).<br />
Metapragmatics is also part of the discourse of language socialization, though it<br />
is not clear at this stage how much of the acquisition of pragmatics is dependent on<br />
explicit attention drawn by caregivers to this domain. But as demonstrated above,<br />
parental discourse addressed to children is rich with metapragmatic comments with<br />
respect to discourse management, conversational norms, and the use of politeness<br />
formulae.<br />
Even more than native speakers vary in the degree of their knowledge of pragmatic<br />
rules, they also seem to vary in their capacity to formulate such rules explicitly.<br />
Yet if asked, native speakers show high levels of agreement in judging the
American-Israeli Families at Dinner 217<br />
appropriateness of a specific request variant in a given situation (Olshtain & Blum-<br />
Kulka, 1985), or in agreeing on producing culturally determined request-variants<br />
for the same situation (Blum-Kulka & House, 1989). Such results confirm empirically<br />
the cross-culturally varied systematicity of pragmatic systems and the degree<br />
to which the knowledge of such rules is implicitly there for native speakers to draw<br />
on. Sociolinguists would be out of business if the whole intricate network of<br />
relationships governing such choices were available consciously to every native<br />
speaker; yet parts of this knowledge are retrievable through questioning.<br />
In the case of interviews tapping metapragmatic knowledge, the questions asked<br />
cast interviewees in the role of participant-observers, invited to reflect on their own<br />
language experience. Hence the social situation created is what Glaser & Strauss<br />
(1964) call an awareness context. Focus on the issues topicalized may thus transform<br />
tacit knowledge into explicit formulations.<br />
In the course of our ethnographic interviews with the families in the project, we<br />
asked them both to rate the degree of importance attached to socializing children to<br />
various aspects of "appropriate" language use, and to comment on perceived differences<br />
between their own and native Israelis' ways of speaking. It is from responses<br />
to the first type of question that we learned about the bilingual parents' conscious<br />
efforts to socialize their children to systematic turn-taking norms, as discussed<br />
above. The second type of question followed from discussions of the family's<br />
discourse practices and attitudes in a certain domain. Thus, for example, we asked<br />
the families to tell us about their naming practices and attitudes to nicknaming<br />
(Blum-Kulka & Katriel, 1991), and encouraged their attempts to compare their own<br />
practices and attitudes with those of both American and Israeli. The following<br />
extract illustrates well such a comparative, reflexive effort:<br />
MOTHER: 1 grew up in the South where they say "Yes, Ma'am" for a woman.<br />
INTERVIEWER: Any woman<br />
MOTHER: Any woman other than me, anybody other than me—my mother's<br />
friends, to my mother "Yes, Ma'am," and "Yes, Sir" to a man.<br />
FATHER: We had those things too.<br />
MOTHER: And to this day. Now I don't expect anybody to say "Yes, Ma'am" to me.<br />
I don't even want people to call me Mrs. Joss, I like to to be called<br />
Lydia—by everybody. Even Jasson's friends, Jessica's friends, everybody<br />
calls me Lydia. In the States, if I were to call my mother's friend by<br />
her first name, why, there isn't anything like that! So here I (like being<br />
called)<br />
INTERVIEWER: (Still today)<br />
MOTHER: You know that's funny. When I go back, I still call my mother's friends<br />
Mrs. Such-and-such. I think.<br />
FATHER: Here it is different.<br />
MOTHER: =Anyway, here I don't mind being called by my first name by my<br />
children's friends because it's that way here.<br />
Lydia draws a clear distinction between past and present, "here" and "there."<br />
While shunning the habits of her childhood, she accepts the difference in terms of<br />
address between Israel and the United States as given, surprising herself (i.e.,<br />
raising tacit knowledge to the level of consciousness) in realizing that she adapts to
218 Discourse Perspectives<br />
this difference in norms ("you know that's funny . . ."). This metapragmatic awareness<br />
to cross-cultural (and in this case also regional) variability in ways of speaking<br />
is presented in an emotionally favorable light: she "likes" or at least "does not mind"<br />
being addressed by her first name even by status inferiors (children) because "it's<br />
that way here." In Schumann's terms, Lydia is describing her process of acculturation,<br />
motivated by social and affective factors (Schumann, 1978). Other immigrant<br />
speakers go as far as viewing this process of acculturation as a transformation of<br />
personality:<br />
FATHER: Ofl, ofi hishtana, lo rak ha-safa (it's your character that changes, not only the<br />
language)<br />
Yet there are limits to self-awareness and its accuracy on matters linguistic and<br />
pragmatic. In actual use, we found no difference between Israeli and immigrant<br />
adults in the relative salience of comments eliciting politeness formulas from children.<br />
But through the interviews, adults in the immigrant families claimed they do<br />
insist on such formulas, with such insistence occasionally cast in half-apologetic<br />
terms, as a conscious cultural transfer strategy:<br />
FATHER: A request is a request and it should have a "please" joined to it.<br />
MOTHER: It's also still important to me, the "pleases" and "thank yous."<br />
Another area where the bilinguals' awareness becomes inexact is matters involving<br />
the pragmalinguistics of Israeli Hebrew. Talking about the lack of politeness in<br />
public places (a complaint voiced often by Israelis; see Blum-Kulka, 1992) an<br />
argument ensues between a couple stemming from a misunderstanding of a pragmalinguistic<br />
phenomenon:<br />
MOTHER: . . . and "thank you" and "excuse me," SLIXA, which people don't know in<br />
this country.<br />
FATHER: They do, they do.<br />
MOTHER: When they step on your foot, or hit you with an elbow when getting on the bus.<br />
FATHER: Well, I mean SLIXA in the sense of "move out of my way."<br />
Rachel (Mother) is criticizing the Israeli way of speaking for its perceived lack<br />
of IFIDS (Elocutionary Force Indicating Devices) for expressive speech acts: no<br />
marking of gratitude, no expression of apology. But regardless of the degree of<br />
accuracy of her observation, it is flawed by her mention of both slixa and "excuse<br />
me": first, as translation equivalents, and second, as apology IFIDS. Literally slixa<br />
means "pardon"; but its use in apologies is quite limited, competing with several<br />
other lexical items serving this function, such as ani mictaer (I'm sorry) or ani<br />
mitnacel (I apologize), which are much more frequent (unpublished CCSARP data).<br />
Slixa is frequently used as an attention-getter, which in certain situations preempts<br />
forthcoming violations that may need an apology and hence can be seen as a doublefunction<br />
formula. "Excuse me" in English serves the same double function, while<br />
"sorry" is the unmarked apology IFID (see also Bergman & Kasper, Chapter 4).<br />
Rachel is misunderstood by her husband because of this double inaccurracy; he<br />
insists on the use of slixa only as an attention-getter (or preemptive formula) and
American-Israeli Families at Dinner 219<br />
reserves judgment as to the general politeness issue, while she is talking about<br />
speech acts, namely the perceived Israeli absence of polite expressives ("when they<br />
step on your foot"). This example shows the potential fallibility of metapragmatic<br />
observations for a first as well as for a second language.<br />
Finally, we come to the most problematic issue of awareness: Are the adults in<br />
these families aware that their process of acculturation has affected the pragmatic<br />
system of their native English In other words, that both in the way they realize<br />
speech acts and in their metapragmatic discourse in the course of natural conversation<br />
they are instantiating an interlanguage pragmatic system The answer is no. As<br />
illustrated by the example of the use of politeness formulas above, the norms<br />
claimed are not necessarily the norms enacted. On the contrary: metatalk about<br />
pragmatic socialization seems motivated by the underlying assumption that by<br />
maintaining English at the home you also maintain American ways of speaking. In<br />
no part of the interview did we find any indication for a self-awareness related to the<br />
pragmatic aspects of the style used. Self-awareness seems to stop short at the<br />
borderline between linguistics and pragmatics: while both dinner-table conversations<br />
and interviews are particularly rich with metareferences to linguistic issues of<br />
language choice and code-switching, they are remarkably devoid of reference to the<br />
pragmatic issues of how either of the languages used differs in style from native<br />
American or native Israeli usage.<br />
Conclusion<br />
We have seen that immigrant, bilingual speakers manifest through their metapragmatic<br />
discourse adherence to an underlying pragmatic system that differs in important<br />
ways from the systems of both their contact cultures. This unique pragmatic<br />
system is revealed in both first and second language use, regardless of language<br />
choice. Furthermore, we have noted in the immigrant families' discourse a high<br />
degree of awareness of both bilingual usage and cross-cultural differences between<br />
the two relevant contact cultures. Yet there are limitations to this awareness: it is<br />
fallible on contrastive pragmalinguistic matters and lacking in sensitivity with regard<br />
to the interlanguage phenomena manifested in the first language.<br />
There are two general implications to these findings. First, they suggest that in<br />
the case of bilinguals maintaining their first language under conditions of intensive<br />
sociocultural contact with a second language, the pragmatics of the first language<br />
may be the primary domain to be affected by this contact. Ironically, while pragmatic<br />
competence is the most difficult aspect of language to master in learning a second<br />
language, it seems also to be, under certain conditions of bilingualism, to be the<br />
easiest to lose in the first language. Second, the specific case of bilinguals studied<br />
suggests that interlanguage is not necessarily a second-language-use phenomenon.<br />
The discourse style realized by the bilingual families, mostly in their first language,<br />
shows that pragmatic systems can be bi-directionally transferable, both from first to<br />
second language as well as from second to first. The resulting hybrid style of the<br />
bilinguals constitutes a prime example of interlanguage pragmatics.
220 Discourse Perspectives<br />
Notes<br />
This research was funded by Grant No. 87-00167/1 from the United States Israeli Binational<br />
Science Foundation (BSF), Jerusalem, Israel to Shoshana Blum-Kulka and Catherine Snow.<br />
1. Transfer of pragmatic rules from the second to the first language by immigrants is<br />
reported by Clyne (1982, 105) who found for example that German immigrants to Australia<br />
adopt the use of "Ja, danke" (thank you) as a positive reply to an offer, as called for by the<br />
pragmatic rule in English, and in contrast to German, in which "Danke" is a negative reply.<br />
More recently, Tao & Thompson (1991) found that Mandarin speakers, for whom American<br />
English has become their dominant language, when conversing in Mandarin make extensive<br />
use of American English backchannel strategies atypical of Mandarin speakers.<br />
2. We use "dominant" here in the sense that it is the language the children use most, get<br />
their schooling in, and know better. As can be seen in Table 10.1, it is not necessarily the first<br />
language. Some of the children spoke only English at home, up to the age of 3 or 4, when<br />
they began to attend nursery school.<br />
3. The project had two stages: during stage 1 (1985-1988) we collected data from 34<br />
families. The analysis of control acts was based on data derived from transcripts of three<br />
meals from this sample (Blum-Kulka, 1988). For the second stage, (1989-1992), we selected<br />
the 24 best-matched families across the three groups (8 per group). The study of metapragmatic<br />
discourse, reported here, is derived from transcripts of these 24 families, two meals for<br />
each.<br />
4. In reporting findings from the project at conferences we are often challenged on this<br />
point: but what about the presence of the observer Didn't it affect the proceedings The<br />
answer is that of course it did. As it does in all forms and types of participant-observation,<br />
whether sociolinguistic or ethnographic in outlook. Technically the inter-group comparisons<br />
we are doing are still valid since this condition is shared by all families. But beyond that, we<br />
can even say that the presence of the observer actually strengthens the validity of our<br />
findings, since if the families are found to differ under these conditions it means that their<br />
cultural styles transpire even under self-imposed constraints of self-presentation to an "outsider."<br />
And indeed, in the final account, it is their nature as cultural texts of familial selfrepresentation<br />
that makes these dinner conversations so fascinating to study from a crosscultural<br />
perspective.<br />
5. For an analysis of parental MCs in the context of pragmatic socialization see Blum-<br />
Kulka (1990).<br />
6. Since duration of meals varied by 10 to 20 minutes per meal, length was normalized to<br />
1000 transcript lines for each.<br />
7. We are grateful to Dan Davies for suggesting this analysis and to Kobi Yelenik for the<br />
programing.<br />
8. Transcription conventions follow CHILDES (McWhinney & Snow, 1985). Indices<br />
used here are: [ . . . ] = trailing off, [-] unfinished word; [\] = interruption; [=] =<br />
latching; [#] = pause; { ) = overlap. For convenience (differing from CHILDES) we have<br />
reinserted capitals when needed, and used CAPS to denote code-switching from English to<br />
Hebrew.<br />
9. The category of types of talk-regulation (TR) includes the following comments related<br />
to turns (T): bids for T, upholding T, arguing T, allocating T, negating T, prefacing T and<br />
checking for attention. Half of American children's comments fall into the category of "bids<br />
for T," with the rest distributed among the other categories. Israeli children's comments are<br />
distributed among all categories, with no clear preference for any specific type. Children in
American-Israeli Families at Dinner 221<br />
the immigrant families show a high preference (83.3%) for bidding for turns, almost totally<br />
avoiding all other TRs.<br />
References<br />
Basso, K. (1979). Portraits of "the whiteman": Linguistic play and cultural symbols among<br />
the Western Apache. Cambridge: Cambridge University Press.<br />
Becker, J. A. (1988). The success of parent's indirect techniques for teaching preschoolers<br />
pragmatic skills. First Language, 8, 173-81.<br />
Becker, J. A. (1990). Processes in the acquisition of pragmatic competence. In G. Conti-<br />
Ramsden & C. Snow (Eds.), Children's language (Vol. 7, 7-24). Hillsdale, NJ:<br />
Lawrence Erlbaum.<br />
Bellah, R. H., Madsen, R., Sullivan, W. M., Swidler, A., &Tipton, S. M. (1985). Habits of<br />
the heart. New York: Harper and Row.<br />
Blum-Kulka, S. (1988). Cross-cultural differences in interactional styles and the acquisition<br />
of communicative competence. Final Report to the BSF. Institute of Communication,<br />
Hebrew University, Jerusalem.<br />
Blum-Kulka, S. (1990). You don't touch lettuce with your fingers: Parental politeness in<br />
family discourse. Journal of Pragmatics, 14 (2), 259-89.<br />
Blum-Kulka, S. (1991). Interlanguage pragmatics: The case of requests. In R. Phillipsen,<br />
E. Kellerman, L. Selinker, M. Sharwood-Smith, & M. Swain (Eds.), Foreign language<br />
pedagogy research: A commemorative volume for Claus Faerch (255—72).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Blum-Kulka, S., & Snow, C. (1992). Developing autonomy for tellers, tales and telling in<br />
family narrative-events. Journal of Narrative and Life History, 2(3):187-217.<br />
Blum-Kulka, S. (1992). The metapragmatics of politeness in Israeli society. Watts, S. Ide &<br />
K. Ehlich (Eds.), Politeness in Language: Studies in Its His Theory and Practice.<br />
Berlin: Mouton de Gruyter.<br />
Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requesting<br />
behavior. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (123-53). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests<br />
and apologies. Advances in Discourse Processes (Vol. 31). Norwood, NJ:<br />
Ablex.<br />
Blum-Kulka, S. & Katriel, T. (1991). Nicknaming practices in families: A cross-cultural<br />
perspective. In S. Ting-Tooney & F. Korseny (Eds.), Cross Cultural interpersonal<br />
communication: International & intercultural communication manual (Vol. 15, 58-<br />
77). London: Sage.<br />
Brown, R. (1991, September). Language and the relational self. Paper delivered at the<br />
Fourth International Conference on Language and Social Psychology, Santa Barbara,<br />
California.<br />
Clyne, M. G. (1982). Multilingual Australia: Resources, needs, policies. Melbourne: River<br />
Seine.<br />
Edelsky, C. (1981). Who's got the floor Language in Society, 10, 383-421.<br />
Edmondson, W. (1981). Spoken Discourse: A model for analysis. London: Longman.<br />
Ervin-Tripp, S., & Gordon, D. P. (1986). The development of children's requests. In R. E.<br />
Schiefelbusch (Ed.), Communicative competence: Assessment and intervention (61-<br />
96). San Diego, CA: College Hill Press.
222 Discourse Perspectives<br />
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.<br />
Geertz, C. (1976). From the native's point of view: On the nature of anthropological understanding.<br />
In E. Kenneth, K. Basso, & H. A. Selby (Eds.), Meaning in anthropology<br />
(221-31). Albuquerque: University of New Mexico Press.<br />
Giles, H. (1979). Ethnicity markers in speech. In R. Scherer & H. Giles (Eds.), Social<br />
markers in speech (251-91). Cambridge: Cambridge University Press.<br />
Glaser, B., & Strauss, A. (1964). Awareness contexts and social interaction. American<br />
Sociological Review, 24, 669-79.<br />
Gleason, J. B., Perlman, R., & Grief, E. (1984). What's the magic word: Learning language<br />
through politeness routines. Discourse Processes, 7, 493-502.<br />
Gleason, J. B., & Weintraub, S. (1976). The acquisition of routines in child language: "Trick<br />
or Treat." Language in Society, 5, 129-36.<br />
Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday.<br />
Goffman, E. (1967). Interaction ritual: Essays on face to face behavior. New York: Doubleday.<br />
Grice, H. Paul. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and<br />
semantics (Vol. 3): Speech acts (43-59). New York: Academic Press.<br />
Grief, E. & Gleason, J. B. (1980). Hi, thanks, and goodbye: More routine information.<br />
Language in Society, 9, 159-67.<br />
Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press.<br />
Hammersley, M., & Atkinson, P. (1983). Ethnography:Principles andPractice. New York:<br />
Routledge.<br />
Heath, S. B. (1983). Ways with words: Language, life, and work in communities and<br />
classrooms. Cambridge: Cambridge University Press.<br />
House, J., & Kasper, G. (1981). Politeness markers in English and German. In F. Coulmas<br />
(Ed.), Conversational routine (157-85). The Hague: Mouton.<br />
Hymes, D. (1974). Foundations of sociolinguistics. Philadelphia: University of Pennsylvania<br />
Press.<br />
Katriel, T. (1986). Talking straight: Dugri speech in Israeli Sabra culture. Cambridge:<br />
Cambridge University Press.<br />
Labov, W., & Fanshel, D. (1977). Therapeutic discourse. New York: Academic Press.<br />
McWhinney, B., & Snow, C. (1985). The child language data exchange system. Journal of<br />
Child language, 12, 271-96.<br />
Ochs, E. (1988). Culture and language development: Language acquisition and language<br />
socialization in a Samoan village. Cambridge: Cambridge University Press.<br />
Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to<br />
native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language<br />
acquisition (303-29). Rowley, MA: Newbury House.<br />
Olshtain, E., & Blum-Kulka, S. (1989). Happy Hebrish: Mixing and switching in American-<br />
Israeli family interaction. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.),<br />
Variation in second language acquisition (Vol. 1): Discourse and pragmatics (59—<br />
84). Clevedon and Philadelphia: Multilingual Matters.<br />
Pellegrini, A. D., Brody, G., & Stoneman, Z. (1987). Children's conversational competence<br />
with their parents. Discourse Processes, 10, 95-106.<br />
Philips, S. (1983). The invisible culture: Communication in classroom and community on the<br />
Warm Spring Indian Reservation. New York: Longman.<br />
Rabin, C. (1976). Acceptability in a revived language. In S. Greenbaum (Ed.), Acceptability<br />
in language (149-67). The Hague: Mouton.<br />
Rcisman, R. (1974). Contrapuntal conversation in an Antiguan village. In R. Bauman &
American-Israeli Families at Dinner 223<br />
J. Sherzer (Eds.), Explorations in the ethnography of speaking. Cambridge: Cambridge<br />
University Press.<br />
Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simplest systematics for the organization<br />
of turn taking in conversation. Language, 50, 696-735.<br />
Schieffelin, B. (1990). The give and take of everyday life: Language socialization of Kaluli<br />
children. Cambridge: Cambridge University Press.<br />
Schieffelin, B., & Ochs, E. (Eds). (1986). Language socialization across cultures. Cambridge:<br />
Cambridge University Press.<br />
Sharwood-Smith, M. (1983). On first language loss in the second language acquirer. In<br />
S. Gass & L. Selinker (Eds.), Language transfer in language learning. Rowley, MA:<br />
Newbury House.<br />
Schumann, J. (1978). The acculturation model for second language acquisition. In R. Gingras<br />
(Ed.), Second language acquisition and foreign language teaching. Arlington,<br />
VA: Center for Applied Linguistics.<br />
Silverstein, M. (1976). Shifters, linguistic categories and cultural description. In K. Basso &<br />
H. Selby (Eds.), Meaning in anthropology (11-55). Albuquerque: University of New<br />
Mexico Press.<br />
Spradley, J. P. (1979). Ethnographic interview. New York: Holt, Rinehart and Winston.<br />
Tannen, D. (1981). Indirectness in discourse: Ethnicity as conversational style. Discourse<br />
Processes, 3(4), 221-38.<br />
Tao, H., & Thompson, S. A. (1991). English backchannels in Mandarin conversations: A<br />
case study of superstratum pragmatic "interference." Journal of Pragmatics, 16, 209-<br />
25.<br />
Weinreich, U. (1953). Language in contact: Findings and problems. New York: Linguistics<br />
Circle of New York. (Reprinted, The Hague: Mouton, 1974).<br />
Weltens, B. (1987). The attrition of foreign language skills: A literature review. Applied<br />
Linguistics, 8(1), 2-39.<br />
Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University<br />
Press.<br />
Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House.<br />
Verschueren, J. (1985). What people say they do with words. Norwood, NJ: Ablex.
11<br />
Notes on the Interlanguage of Comity<br />
GUY ASTON<br />
I take it that interlanguage pragmatics is concerned with the means by which goals<br />
are (or are not) achieved in contexts of spoken or written discourse where one or<br />
more of the participants is a nonnative speaker—what I shall term, for ease of reference,<br />
NNS discourse. As well as having descriptive and explanatory objectives—<br />
attempting to categorize and account for uses of discourse strategies that contrast<br />
qualitatively or quantitatively with those employed in contexts with exclusively<br />
native speaker participants (NS discourse)—it may also have pedagogic ones,<br />
aiming to facilitate learners' achievement of their goals in interaction. Whichever<br />
objective is primary, we must ask what kinds of goals need to be achieved, and what<br />
constraints on their achievement distinguish NNS from NS discourse. In so doing it<br />
seems important to consider not just the constraints relevant to the achievement of<br />
goals of communicating information but also those relevant to achieving those<br />
social goals of interpersonal rapport that dominate much of our everyday talk.<br />
Two main perspectives on conversational goals and constraints have been<br />
adopted in investigating the pragmatics of NNS discourse. On the one hand, we find<br />
work that takes the point of view of the NNS speaker, focusing on the means by<br />
which he or she attempts to express and convey particular meanings to interlocutors<br />
in reference to the psycholinguistic constraints (those of restricted L2 [second<br />
language] competence) under which such a speaker operates. On the other hand, we<br />
find work that treats the achievement of goals in discourse as a collective concern,<br />
focusing on the interactive procedures by which mutually satisfactory outcomes are<br />
jointly arrived at, in reference to the peculiar constraints of the social context (those<br />
of predictably unshared competence) in which participants in NNS discourse operate.<br />
Both perspectives may be illustrated from the literature on communication<br />
strategies (CS), which grew out of a concern with the means whereby learners can<br />
express an intended meaning given their limited knowledge of the L2. Thus Corder<br />
defined CS as "a systematic technique employed by a speaker to express his meaning<br />
when faced with some difficulty" (1983, 16); Faerch and Kasper, as "potentially<br />
conscious plans for solving what to an individual presents itself as a problem in<br />
reaching a particular communicative goal" (1983, 36; my emphasis). In contrast<br />
with this speaker-oriented, intra-organism view of CS as procedures employed to<br />
224
Notes on the Interlanguage of Comity 225<br />
express a meaning, other researchers in the field instead proposed that they be seen<br />
as "a mutual attempt to agree on a meaning where requisite meaning structures do<br />
not seem to be shared" (Tarone 1983, 65), taking an inter-organism perspective in<br />
which CS are considered to involve procedures of negotiation designed to obtain a<br />
mutually ratified outcome—what Widdowson (1983) has termed the convergence<br />
of participants' worlds.<br />
This second perspective involves a somewhat different view of discourse goals<br />
(perlocutionary ones of "agreeing on" rather than simply illocutionary ones of<br />
"expressing" a given meaning) and of the constraints in terms of which CS are to be<br />
accounted for. Rather than treating the latter as primarily a matter of nonnative<br />
speakers' "lexical deficit" (Bongaerts & Poulisse, 1989), leading to a categorization<br />
of CS in terms of speakers' personal "avoidance" (message abandonment, topic<br />
avoidance) and "achievement" strategies (circumlocution, paraphrase, appeal for<br />
assistance, transfer, mime: Faerch & Kasper, 1983), Tarone's definition stresses the<br />
constraints that derive from the lack of intersubjective resources in NNS discourse.<br />
It evaluates such "resource gaps" (not only lexicogrammatical but also<br />
sociocultural—a lack of shared conventions concerning language use, and of shared<br />
knowledge and experience of the world) not in terms of the "objective" status of<br />
resources as available or otherwise, but in terms of their seeming status—of participants'<br />
reciprocal perceptions and expectations as to the resources shared in the<br />
current context. A presumed lack of intersubjective context for communication may<br />
oblige participants to seek ways of making that context explicit and of establishing<br />
what is in fact shared, rather than simply assuming the "reciprocity of perspectives"<br />
of NS participants where meaning can reliably be expanded on the basis of an<br />
"etcetera principle" (Cicourel, 1973).<br />
This second viewpoint leads to an expansion of taxonomies of CS to include, for<br />
example, (a) strategies of "recipient design," which aim to alleviate the hearer's<br />
eventual problems of understanding rather than just the speaker's problems of<br />
expression—through repetition, explanation, explication, and avoidance of reference<br />
to predictably unshared culture-specific knowledge and experience (Tarone &<br />
Yule, 1987; Aston, 1988c); (b) interactive procedures of comprehension checking<br />
and confirmation, which repair failure and manifest that intersubjectivity has in fact<br />
been achieved (Long, 1983; Varonis & Gass, 1985); (c) preparatory strategies<br />
designed to influence assumptions and expectations preemptively—for instance,<br />
overt disclaimers of competence ("I have a problem. I'm a foreigner, and . . .":<br />
Anderson, 1988b), and covert ones, such as the adoption of a foreign accent and/or<br />
incorrect lexicogrammar (Di Pietro, 1987; Warren-Leubacker & Bohannon, 1982)<br />
—strategies that can be seen as canceling the presupposition that the resources<br />
usually shared in NS discourse are available, activating what has been called a<br />
"NNS set" of expectations (Anderson, 1988b): their utility in facilitating communication<br />
is shown by Warren-Leubacker and Bohannon (1982), who demonstrate how<br />
initial manifestations of incompetence by the NNS give rise to greater repetition and<br />
explanation by the NS participant. The wider view of CS provided by an interorganism<br />
perspective seems pedagogically useful, since in the final analysis language<br />
users, including learners and their interlocutors, are typically concerned with<br />
achieving perlocutionary outcomes of convergence, not merely illocutionary ones of
226 Discourse Perspectives<br />
self-expression. It is clear from comparing the two approaches that the communication<br />
strategies highlighted differ according to whether discourse constraints are<br />
conceived of in intra- or inter-organism terms.<br />
While differing in their perspectives on the goals and constraints involved, both<br />
these approaches to communication strategies share a conception of discourse as<br />
primarily a matter of conveying information, that is, as transactional (Brown &<br />
Yule, 1983). In the intra-organism approach, such strategies as paraphrase, circumlocution,<br />
and mime are designed to convey information successfully by clarifying<br />
intended sense and reference. The inter-organism approach also treats discourse<br />
goals as primarily transactional. Desired outcomes are formulated in terms of participants'<br />
final information states, with convergence in respect of an "agreed meaning":<br />
CS "are used to deal with problems which may arise in the transmission<br />
of ... information" (Tarone & Yule, 1987, 50).<br />
This tendency to conceive goals as transactional (and hence to consider discourse<br />
constraints in relation to the achievement of such goals) glosses over the fact<br />
that much everyday speech use instead has as its primary goals the negotiation of<br />
interpersonal relationships. In such interactional speech, Brown and Yule note<br />
(1983, Iff.), the correct conveyance of information may be very much less important.<br />
Participants seem to be involved in "choric" behaviour, as Firth (1964, 112)<br />
put it, where what is important is not so much the understanding of what is said as<br />
the harmony of the dance of their saying, so that they "end up feeling comfortable<br />
with each other" (Brown & Yule, 1983, 12): the emphasis is on ritual displays of<br />
agreement and mutual appreciation. 1 Rather than a use of language for communication<br />
(if by this term we mean, as is generally the case in applied linguistics, a<br />
convergence of participants' information states), in interactional speech we have a<br />
use of language for what Leech (1983) terms comity—the establishment and maintenance<br />
of friendly relations. What seems primarily at issue is a convergence of<br />
participants' worlds in affective terms—sharing feelings and attitudes rather than<br />
knowledge and ideas (Aston 1988b). An extreme example is Malinowski's (1930)<br />
"phatic communion," which he saw as totally devoid of information content; but<br />
even where some information is conveyed, as in everyday chat between lovers,<br />
friends, and gossips, the primary point of talk still seems to be that of establishing<br />
and maintaining relationships of varying degrees of intimacy: even at academic<br />
congresses, as Firth (1964) noted, much talk has a primary function of this type. As<br />
Coupland et al. (1992) observe, interactional speech may also play a significant role<br />
in what are prima facie purely transactional contexts, for instance, in celebrating<br />
transactional success or making transactional failure acceptable. In service encounters<br />
in shops, we often find assistants and customers engaging in pieces of interactional<br />
chat when the requested goods are not available—or unexpectedly are so<br />
(Aston 1988a; cf. examples, 2, 7, and 15 below). Given the frequency with which<br />
interactional speech occurs in everyday talk, it is relevant for interlanguage pragmatics<br />
to ask whether, and in what way, specific constraints condition the pursuit of<br />
interactional goals in NNS discourse, and what "comity strategies " may be available<br />
to participants in order to meet these.<br />
Turning aside from the literature on communication strategies to other areas of<br />
research into NNS discourse, we find a series of studies that focus on interactional
Notes on the Interlanguage of Comity 227<br />
aspects of such talk. Work in contrast!ve pragmatics, while primarily concerned<br />
with the comparison of speech act realizations in different cultures by native speakers,<br />
has also examined realizations by nonnative speakers, comparing these with<br />
typical realizations in those speakers' Lls and with those of native speakers. The<br />
main speech acts selected for analysis are ones that are either face-threatening—<br />
requests (the CCSARP project: cf. Blum-Kulka et al., 1989a), refusals (Beebe &<br />
Cummings, 1985), disagreements (Beebe & Takahashi, 1989), complaints (Olshtain<br />
& Weinbach, 1987)—or face-supportive—compliments (Wolfson, 1981), thanks<br />
(Eisenstein & Bodman, 1986), apologies (Olshtain, 1983;Trosborg, 1987; CCSARP;<br />
Garcia, 1989)—and their realizations have been compared in terms of the politeness<br />
strategies employed to give and redress face (generally using taxonomies based on<br />
the work of Brown & Levinson, 1978, 1987). For example, in examining requests<br />
and apologies, the CCSARP project has examined ways in which learners' assessment<br />
of face-threat and the strategies they employ to redress that threat appear to<br />
differ from those of native speakers of the L2. In this way, such work highlights<br />
strategies employed by nonnative speakers to achieve interactional goals of satisfactory<br />
rapport.<br />
In terms of the distinction drawn above in discussing communication strategies,<br />
it should however be noted that this line of research adopts an intra-organism<br />
approach: it focuses on nonnative participants' expression of politeness. Limiting<br />
attention to the realization of the single speech act, it does not consider eventual<br />
interactive procedures by which mutually satisfactory interpersonal outcomes are<br />
established, and it views constraints on the achievement of interactional goals in<br />
terms of the NNS's limited competence; that is, strategies of politeness are usually<br />
accounted for in terms of inadequate knowledge of the sociolinguistic conventions<br />
of the L2 (giving rise to transfer from the LI) rather than in terms of a lack of<br />
intersubjective resources (giving rise to negotiation). This limits the value of the<br />
approach as an account of the strategies by which satisfactory rapport is achieved in<br />
NNS discourse. For while the NNS participant's politeness strategies undoubtedly<br />
contribute to the process whereby friendly relations are established and maintained<br />
in such discourse, inasmuch as they indicate that speaker's friendly intent, an<br />
account of such strategies does not explain how both participants 2 effectively come<br />
to feel comfortable with each other as an outcome. (This because it is not necessarily<br />
sufficient to convey an intent to redress face in order to effectively redress it,<br />
and because the redress of the hearer's face by one speaker is clearly only one<br />
element in the process of mutual facework—itself in turn arguably only one aspect<br />
of the comity process.) Nor would it appear that the constraints on the achievement<br />
of interactional goals in NNS discourse are simply those posed by the objective<br />
limits to the NNS' competence; there may also be intersubjective, social constraints<br />
under which NNS interactional speech operates, due to a perceived lack of shared<br />
resources. The CCSARP project, which aims to "investigate the similarities and<br />
differences in the realisation patterns of given speech acts between native and nonnative<br />
speakers of a given language, relative to the same social constraints" (Blum-<br />
Kulka et al., 1989b, 13), thereby assumes that the social constraints in NS and NNS<br />
discourse can be treated as substantially identical rather than as in any way intrinsically<br />
distinct. In discussion of the project data on request realizations, however,
228 Discourse Perspectives<br />
Faerch and Kasper (1989; see also Kasper, 1989) are led to place this assumption in<br />
doubt: noting that the tendency for nonnative speakers to provide warrants for<br />
requests is independent of their LI, and thus cannot be accounted for in terms of<br />
transfer, they suggest that this phenomenon reflects distinctive patterns of expectation<br />
in NNS discourse, so that "for NNS contributions to be successful they may<br />
have to follow different conversational principles" (Kasper, 1989, 54), "for instance,<br />
by placing more emphasis on phatic, metalingual and metacommunicative<br />
functions" (Faerch & Kasper, 1989, 245). Fiksdal (1988), in one of the few studies<br />
in this area to examine the politeness strategies in both NS and NNS in NNS<br />
discourse, similarly suggests that distinctive principles of interaction may be involved.<br />
By arguing in favour of context-specific principles for the conducting of<br />
interactional concerns in NNS discourse, these suggestions go in the direction of an<br />
inter-organism perspective on the constraints to which such discourse may be subject.<br />
It is the nature of such constraints that I wish to consider in this chapter.<br />
Interactional Speech from an Inter-Organism Perspective<br />
A Neglected Issue<br />
It seems useful, as a premise to more detailed consideration, to note some reasons<br />
for the lack of attention paid to interactional speech in NNS discourse from an interorganism<br />
perspective. This may help clarify the theoretical and methodological<br />
stance I want to adopt.<br />
One reason for this apparent neglect is undoubtedly the widespread assumption<br />
of a speech-act-based model of discourse. Placing the emphasis on the isolated<br />
utterance rather than on the jointly-performed dance as a whole, such a model in a<br />
sense presupposes that goals in discourse are the transactional ones of achieving<br />
correct understanding, in the shape of "illocutionary uptake" (Searle, 1969). Even<br />
where descriptions of NNS discourse have ventured beyond the single act, they have<br />
largely considered interactivity as a matter of "negotiating meaning"—of jointly<br />
clarifying and confirming understanding of the information that an individual speaker<br />
intends to convey by a particular act (Long, 1983; Widdowson, 1983). A speechact-based<br />
view entails that politeness, too, constitutes a kind of information, a<br />
"social meaning" conveyed parallel to the main illocution and to be correctly understood<br />
alongside it (Leech, 1983). Rather than conveying information, Brown and<br />
Yule (1983) argue, interactional speech instead involves negotiating shared attitudes<br />
and feelings: to describe this process we need to see talk as progressively bringing<br />
about affective convergence through the joint realization of interactive procedures<br />
(Aston, 1988b).<br />
A second reason is the tendency in applied linguistics to formulate pedagogic<br />
objectives in transactional terms—teaching language for communication (e.g.,<br />
Widdowson, 1978; Johnson & Morrow, 1981; Brown & Yule, 1983). As a result of<br />
this tendency (in part derived from speech act theory), "communicative approaches"<br />
to language teaching have generally treated the achievement of interactional goals as<br />
a matter of the NNS conveying information in a socially acceptable manner (by
Notes on the Interlanguage of Comity 229<br />
employing appropriate politeness strategies: e.g. Littlewood, 1981; Widdowson,<br />
1983). Such an approach to the interactional, as noted above, casts limited light<br />
upon the processes by which mutually satisfactory interpersonal outcomes are collectively<br />
achieved. The pedagogic importance of these processes and outcomes,<br />
however, should not be underestimated. Many learners have goals of using the L2 to<br />
establish interpersonal rapport (Geoghegan, 1983), and it should be borne in mind<br />
that a context of positive rapport also facilitates transactional speech use; for instance,<br />
by rendering operative a "benevolence principle" whereby eventual understanding<br />
failures are more likely to be interpreted as errors rather than offenses<br />
(Thomas, 1983; Slugowski & Turnbull, 1988; Anderson, 1988b). A context of<br />
satisfactory rapport may also facilitate second language acquisition tout court<br />
(Gumperz, 1982; Fillmore, 1979). 3<br />
A third reason for the neglect of interactional speech is the difficulty of collecting<br />
relevant data. For instance, studies that have examined NNS politeness strategies<br />
have for the most part employed discourse completion tasks, where subjects are<br />
required to state what utterance they feel would be appropriate to a specified context<br />
(e.g., Blum-Kulka et al., 1989a). This method, however, discounts interactivity,<br />
with the possibility of the utterance's joint production over time (on the basis of<br />
back-channel feedback or its absence: cf. e.g., Anderson et al., 1988), and of the<br />
realization of the appropriate action across a series of turns (as would appear to be<br />
frequently the case for such face-threatening actions as complaints and apologies:<br />
George, 1990; Owen, 1983). Data where realizations are interactively determined<br />
can be obtained through the use of role-play, where subjects jointly simulate a<br />
situation calling for a given action-type (Beebe & Cummings, 1985; House, 1989),<br />
but it is difficult for this technique to reproduce the interpersonal context of naturally<br />
occurring talk. Role-played interactions are without effective social consequences,<br />
since the relationships between the characters are, in the final analysis, fictional and<br />
temporary; as a corollary, the real-life relationships of the actors (rather than the<br />
characters) may constitute an extraneous influence, revealed in the metadiscursive<br />
nudges, winks, and laughter that accompany so many learner role-plays (Zanca,<br />
1988; Wildner-Bassett, 1989). Experimentally contrived situations in the laboratory<br />
pose similar difficulties. These too are isolated from the real-world context of actual<br />
relationships: situations where goals are transactional, such as the picture description<br />
tasks widely employed in work on communication strategies, are far easier to<br />
set up experimentally than ones where goals are additionally or primarily interactional<br />
(cf. Aston, 1986a, 1988b for further discussion of these issues). To examine<br />
interactional speech from an inter-organism perspective, it seems essential to look<br />
first and foremost at naturally occurring conversational data.<br />
A final reason would appear to be of an ideological nature. While a range of<br />
studies have used naturally occurring data to consider interactional aspects of NNS<br />
discourse, these have typically examined situations in which interpersonal conflict<br />
emerges rather than ones in which rapport is created successfully. Motivated by<br />
social concern at the extent of conflictuality and the abuse of power in cross-cultural<br />
settings, they have stressed how negative affect derives from misunderstanding of<br />
intent (e.g., Gumperz, 1982; Thomas, 1983). Their emphasis has thus been on the<br />
avoidance of interactional failure. If we are to understand how interactional success
230 Discourse Perspectives<br />
can be achieved in NNS discourse, we need to focus on strategies for the establishment<br />
of positive rapport—enhancing friendly relations—rather than just the avoidance<br />
of negative outcomes of conflict and hostility. If interactional speech is seen<br />
not simply as facilitating the achievement of transactional goals by making communication<br />
acceptable, but as having objectives that may be the primary purpose of the<br />
talk, then comity needs to be seen as not simply a matter of mitigating face-threat<br />
and hence avoiding wnfriendly relations, but as a manner of enhancing relations in a<br />
positive sense (Aston, 1989). 4<br />
It is for the reasons outlined, I believe, that we find only passing observations in<br />
the literature which consider comity strategies from an inter-organism perspective.<br />
These limited observations are, however, suggestive, since they indicate that certain<br />
characteristic features of NNS discourse can find explanations in terms of the use of<br />
such strategies. For instance, McCurdy (1980) notes that participants in NNS discourse<br />
may "fake" understanding as a means of avoiding putting transactional<br />
failure, with its negative affective consequences, on record. Tarone (1980) notes the<br />
occurrence of repairs to linguistic form in context where mutual understanding has<br />
already been achieved. Excluding these from her transactionally defined taxonomy<br />
of communication strategies, she suggests instead that they have primarily social<br />
ends of "meta-communication . . . which says, 'I'm a member of your group'"<br />
(1980, 426)—an interpretation that finds echoes in the following remark of Widdowson's:<br />
Learners, like the ordinary people they are, adjust their language to an acceptable<br />
norm for two reasons: either in order to be more effectively communicative or in<br />
order to indicate a sense of identity with a particular group of language users. . . .<br />
It is difficult to correct learner error which has little or no communicative consequence<br />
because to do so is to ask learners to subscribe to an etiquette which may<br />
seem alien to them, quaint, even ridiculous; to conform to standards of behaviour<br />
that represent a code of conduct for particular social groups with which they have<br />
no social connection and no real affinity. (1984, 248-49; my emphasis)<br />
The use in NNS discourse of confirmations of understanding and of repairs to form<br />
may in other words be accounted for in terms of participants' interactional rather<br />
than transactional concerns (Aston, 1986b); that is, they may constitute comity<br />
strategies aiming to maintain and enhance rapport.<br />
Such isolated observations, however, tell us little of the range of comity strategies<br />
that may be employed given the constraints peculiar to NNS discourse. Given<br />
the difficulty of setting up controlled contexts in which negotiation of rapport is at<br />
issue, any attempt to examine such strategies would seem to involve what Faerch<br />
termed "straying around in the forest" (1984, 60), focusing on a range of examples<br />
from various sources in the light of a general theory rather than attempting to deal<br />
with systematically elicited data. As Faerch stressed, such a methodology means<br />
that the investigator has to identify problems before proceeding to a description of<br />
solutions; I shall thus be more concerned here with the nature of those problems—<br />
the constraints on interactional speech in NNS discourse—than with the potential<br />
solutions to them—the precise range of the comity strategies available. I want to<br />
begin by examining the comity process in NS discourse and the intersubjective
Notes on the Interlanguage of Comity 231<br />
resources it draws on, hoping thereby to identify areas that may be problematic in<br />
NNS discourse owing to the absence of shared resources. By corollary, we may also<br />
be able to highlight resources for comity that are, instead, likely to be shared in<br />
contexts of NNS discourse, and hence be potentially available for use.<br />
Comity in NS Discourse<br />
Brown and Yule (1983, 4) argue that interactional speech involves the negotiation of<br />
shared feelings and attitudes by participants, as opposed to the sharing of knowledge<br />
and ideas that is at issue in transactional speech. The sharing of attitudes can<br />
be seen from two perspectives, those of negotiating solidarity and of negotiating<br />
support (Aston, 1988b). Brown and Oilman define solidarity as a matter of "likemindedness<br />
or similar behaviour dispositions" (1972: 258), so that its negotiation<br />
involves participants in finding they share attitudes to features of their world in<br />
common—for example, that both feel the same way about the weather, or about an<br />
Oxford University Press decision as to the print run on a book: 5<br />
(1) A: T's - tsuh beautiful day out isn't it.<br />
B: Yeh it's jus' gorgeous.<br />
(Pomerantz, 1975, 4)<br />
(2) A: Surprised that they: er didn't print eNOUGH, actually.<br />
c: Yes:, T they obviously underestimated it"* didn't they, badly, because I've been to<br />
th- THREE bookshops now, and none of them-<br />
A:<br />
A You'd have thought they'd've judged THAT a bit better, yes:."*<br />
A: ==Sold out<br />
c: Yeah.<br />
(Gavioli & Mansfield, 1990: Lod A-04/a; A = assistance, C = customer)<br />
Such routines of agreement (for a detailed description, cf. Pomerantz, 1984a) seem<br />
characteristic of the negotiation of solidarity. The tendency to shift topics in casual<br />
conversation reflects the search for things to agree on—to identify common concerns<br />
toward which attitudes may be shared (Brown & Yule, 1983, 11).<br />
As well as features of the situation in which the talk takes place (including ones<br />
brought to that situation from past experience), common concerns may include<br />
features of the talk itself. For instance, conversational closings often involve agreement<br />
as to the conversation's satisfactory nature:<br />
(3) A: Nice talking to you ''sir.'*<br />
B:<br />
A Nice •*- talking to you.<br />
(Pomerantz 1975, 4)<br />
Shared attitudes toward the talk may also be manifested without such explicit<br />
agreements, but simply by production of analogous contributions. Joint laughter can<br />
show shared pleasure at repartee (Jefferson et al., 1987; Davies, 1984). Recycling<br />
of structural chunks (Tannen, 1989), for instance, by repeating a prior sequence<br />
with an inversion of the discourse roles involved, can allow participants to show<br />
their mutual acceptance and approval of that prior sequence. Consider the following<br />
extract, again from an encounter in a bookshop:
232 Discourse Perspectives<br />
(4) c: I hesitate to ask you this, but could you tell me why: + "The dancing Wu Li<br />
masters" is under religion<br />
A: + It is: believe it or not, according to Fontana, + an alternative theology book,<br />
c: + Oh:! *(2syll)!+<br />
A:<br />
A So is •* Capra's "Turning point" and "The tao of physics."<br />
c: + "The tao of physics" I could: - I could believe, I mean it doesn't 'actually<br />
MENtion:'* re Ty ligions in there:,**<br />
A: ^According -~*A: AA According -** according to Fontana it's the same kind of<br />
book, so therefore: -<br />
1—»c: =Mhm. I 'see.'* Yeah it is, it's (a group)! They're all about god and TT physics.'**<br />
A: *(2syll)+<br />
*: ^Mm.*« Yeah,<br />
c: Yehhs. T Er- they're- [laughs]-*<br />
2-^ A: A Yeah! Right! [laughs] Lot to do with religion that, yes"* [laughs]<br />
(16)<br />
(Gavioli & Mansfield, 1990: Lod B-07/a)<br />
Here we find a reciprocation of ironic comments followed by self-laughter (arrows).<br />
First the customer offers what he apparently intends as a little witticism ("They're<br />
all about God and physics"). The assistant initially fails to appreciate this irony,<br />
simply acknowledging it as informative ("Mm. Yeah"), whereupon the customer<br />
initiates laughter himself. The assistant then joins in laughing, and goes on to<br />
provide a contribution of her own that repeats the same structure: she produces a<br />
similar ironic comment, and in her turn follows this with self-laughter. Thereby she<br />
shows she has in fact appreciated the customer's previous contribution, and what<br />
appears to have started off as a failure in interactional terms is turned into a successful<br />
piece of repartee, where shared attitudes are displayed to the talk itself.<br />
Solidarity implies that attitudes are shared in the sense that participants' feelings<br />
vis-a-vis some state or event of which they have common experience are the same.<br />
In general terms, it implies that with respect to such experience, A's wants for A<br />
correspond to B's wants for B. However, it is also possible for attitudes to be shared<br />
in another sense. Rather than sharing attitudes to features of common experience<br />
(including their common talk), both may share attitudes to features of the experience<br />
of one of the participants, to the sources of her or his individual joys and griefs,<br />
where A's wants for A correspond to B's wants for A (or vice versa). One cannot<br />
always share another's feelings about such personal matters a triumph and bereavement<br />
in the sense of having the same feeling oneself, feeling as the other, but one<br />
may be able to do so in the sense of sympathising, or feeling for the other. As<br />
Goffman put it, "the needs, desires, conditions, experiences, in short the situation of<br />
one individual, when seen from his own point of view, provides a second individual<br />
with directions for formulating ritual gestures of concern" (1971, 92). Adopting<br />
Goffman's term, I shall refer to this type of affective convergence as support.<br />
If solidarity is largely characterized by routines of agreement, support seems<br />
characterized by routines of affiliation. In the following examples, sympathy is<br />
provided in response to personal disclosures:<br />
(5) E: I hadda liddle operation on my TOE this week, I hadtuh have- TOEnail taken off.<br />
B: (0.5) WHY::hh
Notes on the Interlanguage of Comity 233<br />
E: Oh, I have a fungus'n I had'n infection, (0.3) T's a *HELL of a- 1 *<br />
-^B: A OH::::::-«Emma. Innat awful, (0.7) Well what a SHAME. Didjeh haftuh go in<br />
the hospit'l<br />
E: No::, I dist hadda local deal, en I- id wadn'any fun, but I'm BEtter I wz, lying on<br />
the couch out'n front.<br />
^>B: Oh::: I'm sorry EmMA<br />
(Sacks, 1992, 573)<br />
(6) L: =But you know when you get OUT it's kind of co:ld.<br />
E:<br />
T (Oh: oh) YA:h«<br />
L:<br />
A Well it was,"* two o'clock in the morning and T then last night*<br />
E:<br />
A Huh Haw hawh ha:*w.<br />
—> E: T Oo I(h) bet that was (fu:n)*<br />
L:<br />
A nhhhhh hn-hn-hn with no:"* c- hh T hh clothes on God it's good, hu-uh hUH huhh<br />
hh*<br />
—>E:<br />
A hhh hh aaAAAAAA:::::::: Isn't that exerting,*<br />
L: Uh:*: 1 *<br />
—>E:<br />
A Oh:* that's wonderful,*<br />
L:<br />
A Oh::* God we had, we, I never had so much fun in my li:fe.<br />
(Jefferson, 1984, 200)<br />
As with solidarity, support may also focus on features of the talk itself—where the<br />
latter is viewed as an individual rather than a joint accomplishment. One participant<br />
can show appreciation of another's contributions to the discourse by laughing at her<br />
jokes and anecdotes—and that other can then in turn appreciate that appreciation:<br />
(7) A: Can I help you.<br />
c: Yeah, can you GIVE me a: - a- a prediction on when Jacques Lacan's "Ecrits" is<br />
likely to come back in<br />
A: + Frankly no. I can give you a + er + educated guess, of: i(4syll)."*<br />
—»c:<br />
A Well.* That's [laugh\ - better than T I can* do [laugh}, so: -<br />
—* A: *[laugh }"*<br />
(07)<br />
(Gavioli & Mansfield, 1900: Lod C-30/b)<br />
Routines involving compliments and apologies, which also propose affiliation with<br />
the other's interests, can similarly be seen as means of negotiating support, in<br />
reference to the patterns of "supportive" and "remedial" interchanges described by<br />
Goffman (1971: for examples and discussion, cf. Aston, 1988b).<br />
Grounds for Solidarity and Support<br />
Just as the transfer of information, through reference to maxims of sincerity and<br />
truthfulness, relies on bases of speakers' knowledge and belief to warrant their<br />
factual claims, so the negotiation of solidarity and support appears to rely on there<br />
being affective grounds that warrant the feelings in question. For instance, if they<br />
are to share attitudes in the solidary sense, it seems that both participants must have<br />
personal experience of the features that are assessed (Pomerantz, 1984b). In examples<br />
1-3 above, the weather, the unavailability of a book, and the quality of the
234 Discourse Perspectives<br />
current conversation are matters to which each participant can be assumed to have<br />
independent access. Where such experience is not self-evidently available, participants<br />
may describe it in the talk, telling stories from their past that provide warrants<br />
for their attitudes concerning, say, the poor safety standards of fairgrounds (Ryave,<br />
1978). Where, on the other hand, a participant has no (or less direct) experience of<br />
the features in question to be drawn on, the negotiation of solidarity may be<br />
problematic—B cannot easily claim to know what it's like to travel by British Rail /<br />
live under a Conservative government / write a dissertation if he has never done so.<br />
In such cases, support may be more appropriate than solidarity. Grounds for sharing<br />
attitudes in the supportive sense lie in the nature of the relationship between the<br />
participants rather than in common experience: I can sympathise with you insofar as<br />
I have a personal involvement with you. This point was well made by Sacks, who<br />
noted how in disclosing one's personal experience to others,<br />
it's pretty much teller's business to tell the story with respect to its import for him,<br />
and it is his involvement in it that provides for the story's telling. . . . teller can<br />
tell it to somebody who knows and cares about him, and maybe recipient can tell it<br />
to someone who also knows and cares about initial teller, but it goes very little<br />
further than that. (Sacks, 1978, 261)<br />
This underlines two requirements for the negotiation of support: (a) that the teller be<br />
involved in the story (i.e., he is entitled to have feelings toward the events described<br />
by virtue of direct experience of them), and (b) that the recipient be involved with<br />
("knows and cares about") the teller. It is this latter factor that entitles the recipient<br />
to hold attitudes toward events of which he lacks independent experience.<br />
As in negotiating solidarity, participants may also be called upon to make their<br />
grounds for sharing attitudes in the supportive sense explicit, demonstrating that<br />
they really do "know and care about" each other. Thus in the following example, L<br />
demonstrates that she understands G's general situation (arrow): by citing her<br />
knowledge that G is "tied down" she shows her entitlement to sympathize with G in<br />
her dilemma of the previous night.<br />
(8) o: And Danny didn't get in so I didn't go: typing last ni:ght,<br />
L: =Didn't y T ou::"*<br />
G: A No:^ I T ca- I thought well I c^an't leave him for VT two** hours if I'm if he's<br />
crying when I've left him for one.<br />
L: ±Oh:::::.«<br />
L:<br />
A *n:No.*"<br />
L: + Oh: dear me.<br />
G: So: I euh you know as I say 1 didn't get to t ¥ yping.~*<br />
—* L: A Oh::::: You're"* well tie:d dow:n aren't T you."*<br />
G:<br />
A Well-* I am rea: T lly:« Yah.<br />
L: AYe:h."<br />
(Rahman, 1:4-6, cited in Jefferson, 1984, 204-5)<br />
Clearly, degrees of "knowing and caring about: vary, and this is reflected in the<br />
varying "tellability" of personal information. There are events we disclose and jokes<br />
we tell to close friends but not to casual acquaintances. The loneliness that follows<br />
loss of a loved one can be interpreted in terms of no longer having anyone close
Notes on the Interlanguage of Comitv 235<br />
enough to confide in (Sacks, 1972). What seems at issue in such cases is the<br />
restrictedness of the attitudes concerned, which not anyone may share supportively,<br />
but only someone who is sufficiently close to us. This requirement has been stressed<br />
by work in social psychology showing that growing intimacy correlates with an<br />
increase in the "privateness" of disclosure (Morton, 1978): conversely, Goffman<br />
noted how there is "a potential conflict between the provision of minor services to<br />
unacquainted others and the obligation to keep one's distance" (1971, 120).<br />
The restrictedness of attitudes also appears relevant in the negotiation of solidarity.<br />
Sharing mild feelings about the weather is a rather different matter from<br />
sharing homicidal feelings toward a common boss. While in the former case our<br />
feelings are those that anybody might be expected to have—those of competent<br />
consociates—in the latter they define a limited group, grounded in a highly specific<br />
common experience and perspective. Social psychology has argued that interpersonal<br />
attraction is related to perceived similarity—we like those who are like us—<br />
with "uncommon similarity" being particularly attractive (Lea & Duck, 1982). As<br />
with support, sharing restricted solidary attitudes thus implies a closer relationship,<br />
one between individuals with a peculiar "likemindedness."<br />
In both solidary and supportive senses, participants appear to negotiate progressively<br />
more restricted attitudes as intimacy increases, thereby increasingly defining<br />
themselves as somes rather than anys (Schenkein, 1978). From this perspective,<br />
engaging in interactional speech is not only a matter of identifying potential<br />
focuses of solidarity and/or support for which participants may have warrants, but<br />
also of locating ones where appropriately restricted attitudes may be shared. Resources<br />
for comity are constrained both by what is presumably common ground and<br />
by the presumed intimacy of the relationship. While to a large extent these constraints<br />
will be specific to the individual relationship concerned, we can, I think,<br />
also see ways in which they may generally differ in contexts of NS and NNS<br />
discourse, given the nature of the common ground and of the relationships which<br />
such contexts imply.<br />
Grounds for Solidarity and Support and NNS Discourse<br />
Focusing on what I have termed solidarity, Fiksdal notes that "because rapport is<br />
created when speakers share common interest or common ground . . . NSs may<br />
rely on a different rapport system than NNSs" (1988, 3). It is clear that the common<br />
ground of consociates—what everybody can be expected to have experience of<br />
within a given culture—will not necessarily be available as a resource for solidarity<br />
where cultural backgrounds differ. While members of different cultures may share<br />
access to the weather (though there may well be differences in their feelings about<br />
it), they are unlikely to share similar experience of Mrs. Thatcher or British Rail,<br />
with respect to which the foreigner may lack warrants for protest on a par with those<br />
of the paid-up British native. In consequence, participants may have difficulty<br />
negotiating agreement vis-a-vis such topics. In extreme cases, we can imagine the<br />
absence of common ground emerging in ways like the following:<br />
(9) NNS: British Rail gets worse and worse, doesn't it<br />
NS: I wouldn't have thought the trains were any better where you come from.
236 Discourse Perspectives<br />
(10) NS: British Rail gets worse and worse, doesn't it<br />
NNS: Well, I wouldn't know.<br />
(fabricated)<br />
It is not merely such single domains of experience that cannot be assumed shared,<br />
however: comparing the behavior of interviewers in NS and NNS discourse, Fiksdal<br />
(1988, 8ff.) notes how appeals to generic intersubjective experience (through the<br />
use of expression such as "you know what I mean"), which invoke the interlocutor's<br />
capacity to fill in detail from common knowledge (following an etcetera principle,<br />
are not employed with NNS interviewees. In NS discourse such uses appear to help<br />
create rapport: not only because they assume common ground but because they<br />
propose allusiveness as a satisfactory discourse style. That is, participants can share<br />
attitudes to the fact that they don't need to spell out detail in order to understand<br />
each other. A well-known example of this involve irony, which, by alluding to what<br />
is conventionally said in routine situations, can be an important strategy for achieving<br />
solidarity (Sperber & Wilson, 1981, 1985, 1986). Discussing the utterance "He<br />
almost won, didn't he" said by his wife to a companion when John has just come<br />
off court announcing that he "only just" lost after an ignominious defeat at tennis,<br />
Sperber and Wilson (1985) note that<br />
If they are old friends in the habit of making mild fun of their husbands, then the<br />
correct interpretation . . . would be easily accessible, and would merely confirm<br />
a range of shared assumptions about their relationship to each other and their<br />
outlook on the world. Suppose, however, that they are meeting for the first time.<br />
Then a hearer who penetrated the irony would have learned an enormous amount<br />
about the speaker's attitude to her husband, her outlook on life, all in a very short<br />
time. As a result of this exchange, a degree of intimacy would be achieved<br />
between them that would not have been created to anything like the same extent by<br />
[alternative, explicit examples], (viii: 74)<br />
John's wife's quip is a contribution to the discourse that not just anyone might have<br />
made, and not just anyone understood: its production and grasp reveal the particular<br />
individuals concerned as sharing a whole set of attitudes. Irony is reputed to be less<br />
available as an interactional resource in NNS discourse, where it may be heard as<br />
learner error rather than wit (Harder, 1980; Widdowson, 1984), and the absence of<br />
the shared sociolinguistic competence necessary as a basis for an allusive style may<br />
be one reason for the widely noted long-windedness and prosaicness of NNS discourse<br />
(Blum-Kulka & Olshtain, 1986; Kasper, 1989; Aston, 1988c).<br />
Such absence of culture-specific common ground means that NNS discourse<br />
may lack certain of the resources typically available to consociates in the negotiation<br />
of solidarity. Similarly in negotiating support, with its basis in "knowing and caring<br />
about," it may be difficult for outsiders to the NS culture, who may be stereotyped<br />
victims and/or perpetrators of cultural prejudices, to draw on the neighborly respect<br />
and understanding that is expected of consociates. This might suggest that the<br />
resources for interactional speech in NNS discourse are drastically restricted with<br />
respect to those available in NS discourse, entailing a "reduced personality" for the<br />
learner, as Harder (1980) puts it. However, other areas of common ground, and<br />
bases for mutual respect and understanding, may instead be available. We should
Notes on the Interlanguage of Comity 237<br />
not exclude the possibility that some of these may be characteristic of NNS discourse.<br />
As Borsch (1986) has documented, learners frequently perceive NNS discourse<br />
as expanding rather than restricting their interactional capacity. To appreciate<br />
how comity can be negotiated in such discourse, in other words, we must not only<br />
recognize the limits imposed by participants' nonconsociacy, but also explore the<br />
potential of nonconsociacy for providing interactional resources.<br />
Resources for Comity in NNS Discourse<br />
Negotiating Solidarity: Exploiting the Lack of Consociacy<br />
I have suggested that negotiating solidarity is a matter of sharing attitudes toward<br />
features of common experience. Thereby participants demonstrate their similarities<br />
or likemindedness. While consociates may be able to draw on their "cultural value<br />
similarity, cultural status similarity, cultural linguistic similarity, and perceived<br />
cultural attitudinal similarity: (Ting-Toomey, 1986, 116), these are less likely to be<br />
present in inter-ethnic encounters. To find areas of likemindedness in NNS discourse,<br />
participants arguably need to turn from their identities as representative<br />
members of their cultures of origin to focus on their identities as individuals, and to<br />
their relationships as individuals to those cultures. If both participants manifest a<br />
certain independence with respect to their native cultures, revealing what Ting-<br />
Toomey terms "flexible cultural identities," similarities may be found at a "metacultural"<br />
level. Recognition of the interactional potential of such flexibility may<br />
underlie learners' claims of enjoying "the freedom to act like an outlaw" when using<br />
the L2 (Borsch, 1986, 73).<br />
To exemplify, one way in which participants in NNS discourse seem able to<br />
establish solidarity is by both manifesting critical or ironic attitudes to their respective<br />
cultures of origin. In the following extract, the British NS takes a critical stance<br />
toward his country's role in the birth of America, from which he disaffiliates (first<br />
arrow), and then, by going on to criticise the role of the Italian NNS' compatriots in<br />
this process, seems to invite the latter to do the same in her turn (second arrow):<br />
(11) NNS: Um: someone told me, er that the captain, of Mayflower was from Walton.<br />
T ls it true 4<br />
NS:<br />
A He wasn't 4 FROM Walton he was CHRIStened in Walton.<br />
NNS: Ah.<br />
NS: Yeah. Don't know where he was born. + It was, a little bit before my time.<br />
NNS: [laughs}<br />
NS: =An- and: + but he was christened at All Saints Church. + Er: what was his<br />
name + Something Matthews I think, but anyway.<br />
NNS: Mm.<br />
1—> NS: Yeah. + + He started it all going over in America, he started all the trouble,<br />
you know. He took some of our trouble makers over there T and they started<br />
more trouble, 4<br />
NNS: ^{laughs} 1 *
238 Discourse Perspectives<br />
2-^><br />
NS: and then your lot went over from the mafia and did the - made the rest, you<br />
know, I mean-<br />
NNS: [laughs}<br />
NS: So. T l mean-"*<br />
NNS: ^Yeah."*<br />
NS: + Between them they really carved it up ().<br />
(Recca, 1988, 217)<br />
Further into the same conversation, the NNS does take up a similar opportunity to<br />
disaffiliate from her culture, criticizing her compatriots as relatively humorless and<br />
unduly vociferous:<br />
(12) NS: [talking about his stay in a hotel in Italy] 1 was known to the staff as papa<br />
grappa, ^[laughs \*<br />
NNS: ^[laughs}'* Yeah.<br />
NS: Yeah. Papa grappa.<br />
—» NNS: Mhm. [laughs} + The Italians have a sense of humor sometimes.<br />
NS: A very good sense of T humor. <<br />
NNS: ^[laughs]*<br />
NS: Yeah.<br />
NSS: Mm:<br />
NS: I like the Romans, I like: 'the"* the Italians in Rome.<br />
NNS: ^Mm."*<br />
NNS: Mhm. They are very friendly.<br />
NS: Oh T yes.^<br />
—> NNS: A Some'*times they are too noisy. As Americans.<br />
. NS: =Too noisy<br />
NNS: Yeah.<br />
NS: Yeah. Oh Yeah. And when the family T ARGument started, A:H.*<br />
NNS: *[laughs]*<br />
NS: [laughs]<br />
(Recca, 1988, 242)<br />
Both participants here display critical attitudes toward their respective cultures,<br />
from which they distance themselves. Taking the stance of fellow-outlaws, both<br />
occupy a no-man's-land that becomes a common ground, and whose successful<br />
creation is the focus of mutual appreciation and laughter.<br />
A second, related area vis-a-vis which solidarity may be negotiated concerns the<br />
task of being an NNS—of achieving and maintaining an identity outside the context<br />
of one's own culture. This is clearly a resource for solidarity in NNS-NNS encounters,<br />
but it can also be exploited in NS-NNS interaction if both participants can<br />
claim experience of being nonnative speakers. In the last example, the NS refers to<br />
his own experience as a foreigner, and in another extract from the same conversation,<br />
offers a few words in the language of his NNS interlocutor:<br />
(13) NS: And twice a week, I used to drive<br />
NNS: Mh<br />
NS: to the Vatican.<br />
NNS: Mhm.<br />
—> NS: "Vaticano," T isn't it'*
Notes on the Interlanguage of Comity 239<br />
NNS:<br />
A Mhm.'* You have got a BEAUtiful pronunciation. Congratulations.<br />
NS: Especially when I say "Una birra per favore."<br />
[mutual laughter]<br />
(Recca, 1988, 243)<br />
In this way discourse roles are momentarily inverted: the NS places himself in the<br />
role of an incompetent NNS, and the NNS in that of the NS who can provide<br />
support. This repetition with inversion enables the participants to share attitudes<br />
toward the business of being/interacting with a nonnative, and in this respect again<br />
to show a common flexibility with respect to their native cultures.<br />
In the cases just discussed, participants can be seen as making up for their lack<br />
of a common culture by using their experience as cultural outsiders as a basis for<br />
negotiating solidarity. Nonconsociacy also appears to offer specific grounds for<br />
sharing solidary attitudes with respect to the talk itself, seen as an appreciable joint<br />
accomplishment. While in NS discourse, the satisfactoriness of talk typically lies in<br />
its nonroutine nature, its particular wittiness and harmoniousness, which makes it<br />
the achievement of somes rather than anys, in NNS discourse even the routine may<br />
be problematic, with the result that its successful bringing-off can appear a noteworthy<br />
accomplishment that provides grounds for mutual satisfaction. Jordan and Fuller<br />
(1975) discuss a lingua franca example in which two Americans (n and g) are<br />
talking to two Mayans (J and M):<br />
(14) M: en Maya se dice pel.<br />
G: UH HUH.<br />
j: T iSabes*<br />
G: *Pel.«<br />
N: Pel.<br />
M: Ah.<br />
G: PEL.<br />
j: T Pel.-<<br />
M:<br />
A Pel.«<br />
G: PEL. PEL.<br />
j: Por que pe, ele.<br />
G: PE, ELE.<br />
N: PEL.<br />
j: Pe, e, ele.<br />
o: PE, E, ELE. AHA.<br />
j: iHah<br />
G: PEL.<br />
j: *Hah.-«<br />
M:<br />
A Hah.-*<br />
G: PEL.<br />
r. Pel.<br />
Of it, they observe:<br />
(Jordan & Fuller 1975, 26)<br />
We find that an agreement is constructed over what "this" is called in Maya,<br />
though the original difficulty with what "this" is has in no way been resolved. It is<br />
evident here that speakers go to extraordinary lengths to assure each other that they
240 Discourse Perspectives<br />
know, understand, have heard, and are even able to spell a word the meaning of<br />
which is still problematic. So that what emerges as the achievement of the talk is<br />
not the imparting of information but rather the celebration of agreement, collaboration,<br />
consensus, a "look we can still talk together." (Jordan & Fuller, 1975,<br />
27; my emphasis)<br />
In NS discourse, to repeat and spell another's words successfully would hardly be a<br />
cause for remark (except in cases where linguistic competence cannot be assumed<br />
shared, e.g., with small children). In NNS discourse, on the other hand, insofar as<br />
such ordinary success cannot be presumed on, its occurrence can be treated as<br />
extraordinary. Even the production of formally correct speech can allow participants<br />
to demonstrate their concern with achieving certain standards of talk, and offer<br />
grounds for satisfaction. As well as such linguistic grounds, lack of shared sociocultural<br />
competence may allow the relatively everyday to be treated as extraordinary:<br />
in the next example, the fact that a large city bookshop accepts book tokens<br />
becomes a cause for joint celebration in a way that would hardly seem possible in<br />
the context of NS discourse. The outcome appears noteworthy precisely because the<br />
NNS did not, and was not expected to, assume that it was routinely predictable:<br />
(15) c": You DON'T accept book tokens here, do you<br />
A: Yes, we do.<br />
—> c": =Yes you do!<br />
A: Oh, yes!<br />
(Gavioli & Mansfield, 1990: Lod C-14; c" = non-native customer)<br />
Such potential for "celebration of the ordinary" (Aston, 1988b) would seem to be a<br />
specific resource of NNS discourse. Once again the absence of common ground is<br />
paradoxically what provides participants with a common ground, in that they can<br />
share attitudes toward the fact of its creation.<br />
Negotiating Support: Exploiting Incompetence<br />
NNS discourse also appears to offer particular resources for the negotiation of<br />
support, which, I suggested earlier, is based upon participants' "knowing and caring<br />
about" each other. Support, that is, implies that interlocutors are in some way<br />
defined as specific individuals—as somes, not just anys. The fact that participants<br />
in NNS discourse are not consociates may facilitate such a definition. The NNS'<br />
observable lack of competence sets her apart from the category of consociates, who<br />
can be presumed ordinarily competent other than in special circumstances (for<br />
instance, following injury or temporary distraction). Known incompetence can thus<br />
be a distinctive, individualizing feature that, by warranting support in the face of<br />
inadequate performance (providing grounds for appeals for assistance to, and displays<br />
of benevolent sympathy from, an interlocutor), can constitute an interactional<br />
resource. Thus in the following bookshop example, the nonnative customer's admission<br />
of sociocultural incompetence (first arrow) leads to the assistant's abandoning<br />
his previous animosity and becoming much more helpful, proposing possible<br />
locations for and then looking up the title in question in Books in print:
Notes on the Interlanguage of Comity 241<br />
(16) c" Excuse me, um: a fr- friend of mine said she reserved a BOOK, in er Lovalts,<br />
which is called "Story of the pendulum."<br />
A =Well: - you've already been here haven't you. + T (Befo-)' 4<br />
c" A Not at the-"* not at the THIS desk.<br />
A =Well you asked the girl here. + 'Cos I was standing behind her when you<br />
asked her.<br />
1—» c" =Oh I'm T jusf* being getting lost probably.<br />
A A Yeah."<br />
A Yeah. + You've been to history<br />
c" I have been everywhere.<br />
A + Well- I haven't a clue, I mean you don't know what it's about<br />
C" + Well- they are different stories- a T bout pendulums, yes."<br />
A<br />
A STOries. Well, it- 4 it should be in- in literature then, shouldn't it On: on the<br />
ground floor<br />
c" + I see-1 think she said it was FIRst floor. + Where I can-1 where- who can:<br />
tell me-1 -1 mean I can BUY the book, but which- which depar- who can advise<br />
me what T sort of-"* where can I look for it.<br />
A: ±Er:.«<br />
A: What's it called the: -<br />
c": "Story of the pendulum."<br />
A: ="Story of the pendulum." + Erm: well I MIGHT get some clue from the: er +<br />
looking it up on the microfiche, on the other hand it might- if it just says "Story<br />
of the pendulum," then: er: + you know, (you don't -) hh.<br />
2—> c": =Okay, hh you are very kind.<br />
(65)<br />
(Gavioli & Mansfield, 1990: Lod B-12)<br />
Incompetence may also give grounds for sharing attitudes supportively vis-a-vis the<br />
discourse itself, by providing for the appreciable nature of one or other participant's<br />
contributions. In the last example, the assistant's support for her incompetence is<br />
made the focus of appreciation by the NNS customer (second arrow); incompetence<br />
may similarly legitimate appreciation by the NS of the NNS' contributions to the<br />
talk, along the lines of "You have got a beautiful pronunciation" in example 13<br />
above.<br />
In NS discourse, appreciable contributions tend to be nonroutine ones—that not<br />
just anyone might have made. In NNS discourse, on the other hand, appreciable<br />
contributions may also be routine, "ordinary" ones, that precisely because they<br />
would be routine for consociates, can appear extraordinary where nonconsociates<br />
are involved. What is little more than routine cooperation by a NS can justify a<br />
NNS' effusive gratitude; NNS near-conformity to NS norms of speech can be<br />
remarked upon and applauded. In this light we can perhaps understand the interactional<br />
potential of self-repairs on linguistic form, for instance, noted by Tarone<br />
(1980): competence in those who are expected to be incompetent is potentially a<br />
praiseworthy accomplishment, a focus for the negotiation of support.<br />
Personality Reduction and Expansion<br />
I have suggested that the nonnative's status as an outsider with limited competence<br />
may offer specific resources for the negotiation of solidarity and support in talk. If
242 Discourse Perspectives<br />
this is the case, participants may have an interest in ensuring the activation of an<br />
"NNS set" for the discourse, in order to make such resources available. However, to<br />
take such a line is not without its problems: to assume the status of an incompetent<br />
outsider can clearly appear equivalent to assuming a reduced personality, inviting at<br />
worst dismissive scorn and at best unwelcome paternalism. In the talk that follows<br />
example 16 above, for instance, the assistant at one point suggests that the book the<br />
NNS customer needs might be a simplified reader for foreigners—a potentially<br />
condescending suggestion that the NNS brusquely rejects:<br />
(17) A (Hm.) + Well: - [laughs] Yeah. I don't know, you- you've been to the HIStory<br />
department,<br />
c" Mhm.<br />
A And you've been to languages next door,<br />
c" Well I- it was not a language book.<br />
—> A Well okay, but 1 mean they have: um: simplified ENglish books there for<br />
example. + They have: abridged: easy readers, + in that department,<br />
c" But that was nothing to do with that.<br />
A =It wasn't. Right. And have you tried the Literature department ON the ground<br />
floor.<br />
(Gavioli & Mansfield, 1990: Lod B-12)<br />
There are, however, a series of potentially mitigating factors that seem relevant in<br />
evaluating the NNS' role in such cases. First it seems worth nothing that the NNS<br />
need not explicitly admit incompetence for an "NNS set" to be triggered. Goffman<br />
(1983) notes that while customers in service encounters can pose their requests<br />
immediately if these correspond to what the assistant can reasonably be expected to<br />
have "in mind" (where the accessibility of the referent can be assumed), preparatory<br />
work is called for where requests are predictably more problematic, with preliminary<br />
greetings and apologies orienting the assistant to expect a nonroutine request.<br />
To the nonnative speaker, for whom even a routine request may be intersubjectively<br />
problematic, the use of such a preparatory strategy appears to offer a means of<br />
avoiding the default assumptions of NS discourse, and NS acknowledgment of the<br />
"marked" opening appears to signal the tacit instantiation of an "NNS set." Anderson<br />
(1988a) found that in the PIXI corpus of English bookshop encounters, such<br />
preparatory openings were acknowledged by assistants only when made by NNS<br />
customers, as in the following example:<br />
(18) c": Excuse me.<br />
-» A: Yeah.<br />
c": I'm looking for + this book "Residual and: infancy registration, + reGREssions."<br />
(Gavioli & Mansfield, 1990: Lod B-05)<br />
In this context the strategic use of such "marked" openings may thus be a way of<br />
establishing the limited competence of the NNS tacitly, without explicit selfdenigration.<br />
In any case, the explicit or implicit assumption of incompetent outsider status<br />
needs to be assessed not just in terms of its immediate implications for the non-
Notes on the Interlanguage of Comity 243<br />
native's face, but as a means to wider ends. It constitutes, on the one hand, an<br />
avoidance strategy with respect to possible interactional trouble. By laying claim to<br />
an initial face "lower" than that which is hoped for, learners reduce the risk of faceloss<br />
with respect to too high a claim: they can fail acceptably, with less risk of<br />
misunderstanding being heard as offense. On the other hand, by making available<br />
the sort of interactional resources described above, it can constitute an achievement<br />
strategy by opening up future opportunities for comity, for instance, by facilitating<br />
the "celebration of the ordinary" and laying the basis of a "knowing and caring<br />
about" that can be drawn on in negotiation of support later on in the encounter.<br />
Shared History as an Interactional Resource<br />
We must not forget that the construction of rapport takes place over time. It cannot<br />
be treated as a one-off event, the result of a single instance of negotiation of<br />
affective convergence, but is rather a progressive achievement. I suggested earlier<br />
that comity involves sharing increasingly restricted attitudes, establishing increasingly<br />
"uncommon" common ground and progressively greater intimacy. This process<br />
involves participants in building up, and building off, a joint interactional<br />
history. The importance of such history as a resource in talk was underlined by<br />
Goffman:<br />
In many of the conversations John has with others (whether with familiars or<br />
unfamiliars), he will not merely fall back on what can generally be taken to be<br />
shared, nor, indeed, will he merely tailor his references to the particular mind of<br />
his particular hearer; rather he will be inclined to select from whatever he shares<br />
with this other, just those topics that allow him to employ allusive phrases that<br />
only the recipient would immediately understand. Thus, his talk will not so much<br />
depend on common understanding as seek it out and then celebrate it. Indeed this<br />
gives to ordinary verbal contacts a greater degree of exclusivity and mutual dovetailing<br />
than one might otherwise expect—not only "recipient design," then, but<br />
the celebration of what it implies. (Goffman 1983, 18)<br />
Of the various ways in which prior talk and its outcomes can provide resources in<br />
interactional speech, some seem particularly relevant to NNS discourse. For instance,<br />
the use of irony, traditionally viewed as relatively inaccessible in NNS<br />
discourse owing to the lack of intersubjective competence as a basis for allusion,<br />
seems quite possible if allusion is instead made to features of the shared history of<br />
the interactants. In the next example, again taken from the conversation analysed by<br />
Recca (1988), the male NS makes a joking comment concerning the possible attitudes<br />
of Anglo-Saxon women to rape (first arrow), which obtains appreciative<br />
laughter from the NNS (she makes it clear that she appreciates the joke rather than<br />
the sexism: second arrow):<br />
(19) NS: So they chopped his head off, and they + took it away, and:<br />
NNS: + Mh.<br />
NS: and once more the: land was robbed, and: the houses were burnt, and: the<br />
women were raped, and it was:<br />
NNS: =Again.
244 Discourse Perspectives<br />
1—» NS: Yeah. And the women said "when are you coming back," you know.<br />
NNS: And ya T (/')' 4<br />
NS: ^[laughs]"*<br />
2-^> NNS: That is not very nice of you. [laughs] + Yeah.<br />
NS: But there we are.<br />
(Recca, 1988, 221)<br />
When the topic of women's roles reappears later on in the conversation, the NNS is<br />
able to ironically allude to her interlocutor's sexism as a known fact (arrows):<br />
(10) NS: I used to be on the council in Walton. + And: we had a woman, a very intelligent<br />
woman,<br />
—* NNS: Oh:, how extraordinary.<br />
NS: Oh:, terrible it is.<br />
NNS: [laughs]<br />
NNS: You know. + I like my women to be women, you know, T (), you know. ^<br />
—> NNS: A Yeah, yeah, I know,"* you are very:<br />
NS: Yeah.<br />
1—» NNS: progressive, ^[laughs]*<br />
NS:<br />
A yeah."*<br />
-» NNS: I understood this, [laughs]<br />
(Recca, 1988, 253)<br />
By alluding to an outcome established in prior talk, the NNS seems able to be ironic<br />
here without assuming that a wider, more generic sociocultural background is<br />
shared.<br />
It may be noted that with her final "I understood this," the NNS makes it explicit<br />
that her irony alludes to the prior talk. In the absence of the presumable common<br />
ground and involvement implied by consociacy, there in fact appears to be a greater<br />
need for participants in NNS discourse to make warrants for their attitudes<br />
explicit—to cite grounds for solidarity and support. Such explicitness has the advantage<br />
of also providing resources for further interactional speech. For instance,<br />
participants' descriptions of their grounds for holding attitudes will often involve<br />
disclosure of personal experiences, with the result that they become better known to<br />
each other in a growing intimacy. And while the need to disclose relatively routine<br />
experience may at times seem tedious for participants, who may feel obliged to<br />
provide continuously what, in the terms of NS discourse, is trivial information<br />
(Anderson, 1988b), we again need to bear in mind that this process contributes to<br />
the accumulation of interactional resources on the basis of which rapport may<br />
subsequently be developed further.<br />
Conclusion<br />
In this chapter I have attempted to outline some resources available for comity in<br />
NNS discourse, arguing that these may not merely be a limited set of those available<br />
in NS discourse. If this is the case, morals for both the descriptive and pedagogic<br />
concerns of interlanguage pragmatics can be drawn. Descriptively, it supports the
Notes on the Interlanguage of Comity 245<br />
argument that interlanguage pragmatics should operate with a difference hypothesis<br />
rather than a deficit hypothesis (Faerch & Kasper, 1989, 246), and not simply<br />
analyze NNS discourse in terms of failure to conform to NS conversational norms.<br />
Pedagogically, it implies that the learner's task in developing an ability for interactional<br />
speech using the L2 is not simply one of acquiring nativelike sociolinguistic<br />
competence in the attempt to mimic the behavior of a native speaker, but requires<br />
the development of an ability to use specific comity strategies appropriate to the<br />
context of NNS discourse.<br />
I have not attempted to provide a taxonomy of such strategies, merely to suggest<br />
areas where resources may lie, given the constraints that seem relevant to interactional<br />
goals from an inter-organism perspective. Certain possible maxims for the<br />
pursuit of comity nonetheless appear to emerge: cite warrants for your attitudes;<br />
admit incompetence; celebrate ordinary success; repair actual incompetence where<br />
possible; appreciate benevolence; distance yourself from the stereotypes of your<br />
culture. Just how such maxims can be effectively translated into specific comity<br />
strategies in interactive discourse calls for considerably more detailed study. They<br />
may, however, offer an initial framework for interpreting differences between NS<br />
and NNS discourse from an inter-organism perspective in interactional terms (as<br />
suggested by Aston, 1986b; Ehrlich et al., 1989), rather than in the purely transactional<br />
ones generally employed.<br />
The approach taken here, I would argue, also offers specific implications for<br />
pedagogy. It suggests that we need to help learners (a) to appreciate that there are<br />
differences in the bases for comity in NS and NNS discourse; (b) to view NNS<br />
discourse not just in negative terms of limitation, but as providing characteristic<br />
resources that can be exploited; (c) to investigate this potential by critically reflecting<br />
on the strategies employed in samples of NS and NNS discourse and by experimenting<br />
with the latter in "comity activities" where rapport is negotiated using the<br />
L2 (Aston, 1986a, 1988a,b). It has often been stressed that attempts to get learners<br />
to conform to NS practices only seem likely to succeed up to the point at which their<br />
identity is not thereby put at risk (Thomas, 1983; Litlewood, 1983; Widdowson,<br />
1984): hence the importance of also helping them develop comity strategies of their<br />
own to establish satisfactory interpersonal relationships through the L2.<br />
Notes<br />
1. This is not, of course, to exclude the possibility that interactional speech can in some<br />
cases take a conflictual form, as in the banter of reciprocal insult between close friends<br />
(Slugowski & Turnbull, 1988). Such use is, however, typically keyed as nonserious (Goffman,<br />
1974), and the importance of the harmony of the dance remains, as stressed for example<br />
by Labov's work on ritual insults (1972).<br />
2. For ease of discussion, I adopt a default view of conversation as dyadic in this paper.<br />
3. This claim can be interpreted in terms of a lowering of what Krashen calls the<br />
"affective filter," which, he argues, "is down when the acquirer [. . .] considers himself to be<br />
a potential member of the group speaking the target language" (1985, 3-4).<br />
4. An analogy can be drawn here with the emphasis in work on communication strategies<br />
on achievement rather than avoidance strategies. A focus on achievement seems desirable
246 Discourse Perspectives<br />
from a pedagogic perspective, since if comity is seen solely in terms of avoiding failure—of<br />
limiting conflict—the task of learners becomes truly Herculean, as they are condemned to<br />
wander a minefield of risk wherein they will inevitably always be less competent than a<br />
native speaker. This is hardly a motivating prospect for the learner, or for the pedagogy,<br />
whose role becomes that of trying to limit the extent of the social damage that the learner will<br />
inevitably suffer. If on the other hand comity is approached in the positive terms of seeking<br />
out opportunities for enhancing rapport—in particular those opportunities specifically offered<br />
by NNS discourse—then the learner is offered chances of success, and also, I would argue,<br />
the pedagogy.<br />
5. Transcription conventions used in examples are those employed for the PIXI corpora<br />
(Gavioli & Mansfield, 1990). Extracts from other sources have been retranscribed using the<br />
same conventions.<br />
+ short pause<br />
(n)<br />
longer pause (n seconds)<br />
T texf* spoken in overlap with next speaker's A texf*<br />
TT texf* < * spoken in overlap with next speaker's AA text**<br />
= text latched to previous turn in transcript sequence<br />
= =text latched to previous-but-one turn in transcript sequence<br />
text - previous utterance broken off<br />
text- previous syllable broken off<br />
text:<br />
previous syllable drawn out (number of colons indicates extent of lengthening)<br />
. , ! give a rough guide to intonation<br />
TEXT capitalization indicates marked emphasis<br />
[comment] situational features or nonverbal behavior<br />
(text) speech unclear: tentative transcription<br />
(nsyll) speech unclear: n = approximate number of syllables<br />
() speech unclear<br />
Name speech altered to preserve anonymity is italicized<br />
References<br />
Anderson, L. (1988a). Access routines in service encounters. In G. Aston (Ed.), Negotiating<br />
service: Studies in the discourse of bookshop encounters (99-134). Bologna: Cooperative<br />
Libraria Universitaria Editrice.<br />
Anderson, L. (1988b). Issues in native-nonnative interaction. In G. Aston (Ed.), Negotiating<br />
service: Studies in the discourse of bookshop encounters. (267-86). Bologna: Cooperative<br />
Libraria Universitaria Editrice.<br />
Anderson, L., Aston, G., & Tucker, G. (1988). The joint production of requests in service<br />
encounters. In G. Aston (Ed.), Negotiating service: Studies in the discourse of bookshop<br />
encounters (135-51). Bologna: Cooperativa Libraria Universitaria Editrice.<br />
Aston, G. (1986a). Roleplay and interactional speech. In S. Holden (Ed.), Techniques of<br />
teaching: From theory to practice (20-25). Oxford: Modern English.<br />
Aston, G. (1986b). Trouble-shooting in interaction with learners: The more the merrier<br />
Applied Linguistics, 7, 128-43.<br />
Aston, G. (1988a). An applied linguistic approach to the analysis of naturally occurring<br />
conversation. In C. Cipolli & E. Rigotti (Eds.), Ricerche di semantica testuale, (259-<br />
82). Brescia: La Scuola.
Notes on the Interlanguage of Comity 247<br />
Aston, G. (1988b). Learning comity: An approach to the description and pedagogy of<br />
interactional speech. Bologna: Cooperativa Libraria Universitaria Editrice.<br />
Aston, G. (1988c). Strategic di comunicazione nella scrittura in LI e in L2. In E. Glass,<br />
F. Marroni, G. Micks, & C. Pagetti (Eds.), Metamorfosi: Traduzione-tradizione<br />
(232-42). Pescara: CLUA.<br />
Aston, G. (1989). Solidarity and conflict as metaphors of the discourse process. In L. Curti,<br />
L. Di Michele, T. Frank, & M. Vitale (Eds.), // mum del linguaggio (413-26).<br />
Naples: Istituto Universitario Orientale.<br />
Beebe, L., & Cummings, M. (1985). Speech act performance: A function of the data<br />
collection procedure Paper presented at TESOL '85, New York.<br />
Beebe, L., & Takahashi, T. (1989). Do you have a bag Social status and patterned variation<br />
in second language acquisition. In S. Gass, C. Madden, D. Preston, & L. Selinker<br />
(Eds.), Variation in second language acquisition: Discourse and pragmatics (103-<br />
25). Clevedon and Philadelphia: Multilingual Matters.<br />
Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989a). Cross-cultural pragmatics: Requests<br />
and apologies. Norwood, NJ: Ablex.<br />
Blum-Kulka, S., House, J., & Kasper G. (1989b). Investigating cross-cultural pragmatics:<br />
An introductory overview. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural<br />
pragmatics': Requests and apologies (1-34). Norwood, NJ: Ablex.<br />
Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic<br />
failure. Journal of Pragmatics, 8, 47-61.<br />
Bongaerts, T., & Poulisse, N. (1989). Communication strategies in LI and L2: Same or<br />
different Applied Linguistics, 10, 253-68.<br />
Borsch, S. (1986). Some ideas concerning the emotional dimension of foreign language<br />
teaching. In G. Kasper (Ed.), Learning, teaching and communication in the foreign<br />
language classroom (71—81). Aarhus, Denmark: Aarhus University Press.<br />
Brown, G., & Yule, G. (1983). Teaching the spoken language. Cambridge: Cambridge<br />
University Press.<br />
Brown, P., & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In<br />
E. Goody (Ed.), Questions and politeness (56-289). Cambridge: Cambridge University<br />
Press.<br />
Brown, P., & Levinson, S. (1987). Politeness. Cambridge: Cambridge University Press.<br />
Brown, R., & Oilman, A. (1972). The pronouns of power and solidarity. In P. P. Giglioli<br />
(Ed.), Language and social context (252-82). London: Penguin. (Original work<br />
published 1960.)<br />
Cicourel, A. V. (1973). Cognitive sociology. London: Penguin.<br />
Corder, S. P. (1983). Strategies of communication. In Faerch, C. & Kasar, G., Eds. Strategies<br />
in interlanguage communication (15-19). London: Longman.<br />
Coupland, J., Coupland, N., & Robinson, J. (1992). "How are you": Negotiating phatic<br />
communion. Language in Society, 27, 207-30.<br />
Davies, C. (1984). Joint joking: Improvisational humorous episodes in conversation. Berkeley<br />
Linguistics Society, 10, 360-71.<br />
Di Pietro, R. (1987). Strategic interaction. Cambridge: Cambridge University Press.<br />
Ehrlich, S., Avery. P., & Yorio, C. (1989). Discourse structure and the negotiation of<br />
comprehensible input. Studies in Second Language Acquisition, 11, 397-414.<br />
Eisenstein, M., & Bodman, J. (1986). "I very appreciate": Expressions of gratitude by native<br />
and non-native speakers of American English. Applied Linguistics, 7, 167-85.<br />
Faerch, C. (1984). Strategies in production and reception—Some empirical evidence. In<br />
A. Davies, C. Criper, & A.P.R. Howatt (Eds.), Interlanguage (49-77). Edinburgh:<br />
Edinburgh University Press.<br />
Faerch, C., & Kasper, G. (1983a). Plans and strategies in foreign language communication.
248 Discourse Perspectives<br />
In Faerch, C. & Kasper G. (Eds.) Strategies in Interlanguage Communication (210-<br />
39). London: Longman.<br />
Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request<br />
realization. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics:<br />
Requests and apologies (221-47). Norwood, NJ: Ablex.<br />
Fiksdal, S. (1988). Verbal and nonverbal strategies of rapport in cross-cultural interviews.<br />
Linguistics and Education, I, 3-18.<br />
Fillmore, L. W. (1979). Individual differences in second language acquisition. In C. J.<br />
Fillmore, D. Kempler, & W. Wang (Eds.), Individual differences in language ability<br />
and language behavior (203-28). New York: Academic Press.<br />
Firth, J. R. (1964). The tongues of men, and speech. London: Oxford University Press.<br />
(Original work published 1930/1937.)<br />
Garcia, C. (1989). Apolgising in English: Politeness strategies used by native and non-native<br />
speakers. Multilingua, 8, 3-20.<br />
Gavioli, L., & Mansfield, G. (Eds.). (1990). The PIXI corpora: Bookshop encounters in<br />
English and Italian. Bologna: Cooperativa Libraria Universitaria Editrice.<br />
Geoghegan, G. (1983). Non-native speakers of English at Cambridge University. Cambridge:<br />
Bell Educational Trust.<br />
George, S. (1990). Getting things done in Naples: Action, language and context in discourse<br />
description. Bologna: Cooperativa Libraria Universitaria Editrice.<br />
Goffman, E. (1971). Relations in public. London: Penguin.<br />
Goffman, E. (1974). Frame analysis. London: Penguin.<br />
Goffman, E. (1983). Felicity's condition. American Journal of Sociology, 89, 1-53.<br />
Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press.<br />
Harder, P. (1980). Discourse as self-expression—On the reduced personality of the second<br />
language learner. Applied Linguistics, I, 262-70.<br />
House, J. (1989). Politeness in English and German: The functions of please and bitte. In<br />
S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests<br />
and apologies (96-119). Norwood, NJ: Ablex.<br />
Jefferson, G. (1984). On stepwise transition from talk about a trouble to inappropriately nextpositioned<br />
matters. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action<br />
(346-69). Cambridge: Cambridge University Press.<br />
Jefferson, G., Sacks, H., & Schegloff, E. (1987). Notes on laughter in the pursuit of<br />
intimacy. In G. Button & J. Lee (Eds.), Talk and social organisation (152-205).<br />
Clevedon and Philadelphia: Multilingual Matters.<br />
Johnson, K., & Morrow, K., (Eds.). (1981). Communication in the classroom. London:<br />
Longman.<br />
Jordan, B., & Fuller, N. (1975). On the non-fatal nature of trouble: Sense-making and<br />
trouble-managing in lingua franca talk. Semiotica, 13, 11—31.<br />
Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden,<br />
D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse<br />
and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.<br />
Krashen, S. (1985). The input hypothesis. London: Longman.<br />
Labov, W. (1972). Rules for ritual insults. In D. Sudnow (Ed.), Studies in social interaction<br />
(120-79). New York: Free Press.<br />
Lea, M., & Duck, S. (1982). A model for the role of similarity of values in friendship<br />
development. British Journal of Social Psychology, 21, 301-10.<br />
Leech, G. (1983). Principles of pragmatics. London: Longman.<br />
Littlewood, W. (1981). Communicative language teaching: An introduction. Cambridge:<br />
Cambridge University Press.
Notes on the Interlanguage of Comity 249<br />
Littlewood, W. (1983). Contrastive pragmatics and the foreign language learner's personality.<br />
Applied Linguistics, 4, 200-206.<br />
Long, M. H. (1983). Native speaker/non-native speaker conversation and the negotiation of<br />
comprehensible input. Applied Linguistics, 4, 126-41.<br />
Malinowski, B. (1930). The problem of meaning in primitive languages. In C. Ogden & I. A.<br />
Richards, The meaning of meaning (296-336). London: Routledge and Kegan<br />
Paul.<br />
McCurdy, P. (1980). Talking to foreigners: The role of rapport. Unpublished doctoral dissertation,<br />
University of California at Berkeley.<br />
Morton, T. (1978). Intimacy and reciprocity of exchange: A comparison of spouses and<br />
strangers. Journal of Personality and Social Psychology, 36, 72-81.<br />
Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In<br />
S. Gass & L. Selinker (Lids.), Language transfer in language learning (232-49).<br />
Rowley, MA: Newbury House.<br />
Olshtain, E., & Weinbach, L. (1987). Complaints—A study of speech act behavior among<br />
native and nonnative speakers of Hebrew. In M. Papi & J. Verschueren (Eds.), The<br />
pragmatic perspective (195-208). Amsterdam: Benjamin.<br />
Owen, M. (1983). Apologies and remedial interchanges. The Hague: Mouton.<br />
Pomerantz, A. (1975). Second assessments: A study of some features of agreements/<br />
disagreements. Unpublished doctoral dissertation, University of California at Irvine.<br />
Pomerantz, A. (1984a). Agreeing and disagreeing with assessments: Some features of preferred/dispreferred<br />
turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structures of<br />
social action (152-63). Cambridge: Cambridge University Press.<br />
Pomerantz, A. (1984b). Giving a source or basis: The practice in conversation of telling "how<br />
I know." Journal of Pragmatics, 8, 607-25.<br />
Recca, L. (1988). Negoziazioni di identita fra parlanti nativi e non nativi in conversazioni in<br />
lingua inglese. Unpublished dissertation, Universita di Bologna, Facolta di Lettere e<br />
Filosofia.<br />
Ryave, A. (1978). On the achievement of a series of stories. In J. Scherikein (Ed.), Studies in<br />
the organization of conversational interaction (113-32). New York: Academic Press.<br />
Sacks, H. (1972). An initial investigation of the usability of conversational data for doing<br />
sociology. In D. Sudnow (Ed.), Studies in social interaction (31-74). New York: Free<br />
Press.<br />
Sacks, H. (1978). Some technical considerations of a dirty joke. In J. Schenkein (Ed.),<br />
Studies in the organization of conversational interaction (249-69). New York: Academic<br />
Press.<br />
Sacks, H. (1992). Lectures on conversation H. Oxford: Basil Blackwell.<br />
Schenkein, J. (1978). Identity negotiations in conversation. In Schenkein 1978a. Studies in<br />
the organization of conversation interaction. New York: Academic Press. 57-78.<br />
Searle, J. R. (1969). Speech acts. Cambridge: Cambridge University Press.<br />
Slugowski, B. R., & Turnbull, W. (1988). Cruel to be kind and kind to be cruel: Sarcasm,<br />
banter and social relations. Journal of Language and Social Psychology, 7, 101-21.<br />
Sperber, D., & Wilson, D. (1981). Irony and the use-mention distinction. In P. Cole (Ed.),<br />
Radical pragmatics (295-318). New York: Academic Press.<br />
Sperber, D., & Wilson, D. (1985). Relevance. Unpublished manuscript.<br />
Sperber, D., & Wilson, D. (1986). Relevance. Oxford: Basil Blackwell.<br />
Tannen, D. (1989). Talking voices: Repetition, dialogue and imagery in conversational<br />
discourse. Cambridge: Cambridge University Press.<br />
Tarone, E. (1980). Communication strategies, foreigner talk, and repair in intcrlanguage.<br />
Language Learning, 30, 417—31.
250 Discourse Perspectives<br />
Tarone, E. (1983). Some thoughts on the notion of "communication strategy." In C. Faerch &<br />
G. Kaspcr (Eds.), Strategies in interlanguage communication (61-74). London:<br />
Longman. (Original work published 1981.)<br />
Tarone, E., & Yule, G. (1987). Communication strategies in East-West interactions. In<br />
L. Smith (Ed.), Discourse across cultures (49-65). New York: Prentice-Hall.<br />
Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.<br />
Ting-Toomey, S. (1986). Interpersonal ties in intergroup communication. In W. B.<br />
Gudykunst (Ed.), Intergroup communication (114—26). London: Arnold.<br />
Trosborg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11,<br />
147-67.<br />
Varonis, E., & Gass, S. (1985). Non-native/non-native conversations: A model for the<br />
negotiation of meaning. Applied Linguistics, 6, 71-90.<br />
Warren-Leubackcr, A., & Bohannon, J., Ill (1982). The effects of expectation and feedback<br />
on speech to foreigners. Journal of Psycholinguistic Research, 11, 207-15.<br />
Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University<br />
Press.<br />
Widdowson, H. G. (1983). Learning purpose and language use. Oxford: Oxford University<br />
Press.<br />
Widdowson, H. G. (1984). Explorations in applied linguistics, 2. Oxford: Oxford University<br />
Press.<br />
Wildner-Bassett, M. E. (1989). Coexisting discourse worlds and the study of pragmatic<br />
aspects of learners' interlanguage. In W. Oleksy (Ed.), Contrastive pragmatics (251 —<br />
75). Amsterdam: Benjamin.<br />
Wolfson, N. (1981). Compliments in cross-cultural perspective. TESOL Quarterly, 15, 117-<br />
24.<br />
Zanca, C. (1988). // roleplay nell'insegnaniento della lingua straniera. Unpublished dissertation,<br />
Universita di Bologna, Facolta di lettere e Filosofia.
ABOUT THE AUTHORS<br />
GUY ASTON is associate professor of English at the School for Interpreters and<br />
Translators of the University of Bologna. In the 1980s he was coordinator of the<br />
PIXI research project on the pragmatics of public service encounters in English and<br />
Italian, and has since continued working on the acquisition of oral competence in a<br />
second language, using conversational analytic methods of discourse description.<br />
LESLIE M. BEEBE is professor of linguistics and education in the Applied<br />
Linguistics and TESOL programs at Teachers College, Columbia University. She is<br />
editor of Issues in Second Language Acquisition: Multiple Perspectives and coauthor<br />
of English in the Cross-Cultural Era: The Social Rules of Speaking and<br />
TESOL. She publishes widely in sociolinguistics and second language acquisition.<br />
During 1990-1991, she served as president of the American Association for Applied<br />
Linguistics.<br />
MARC BERGMAN has taught English as a Foreign Language in such places as<br />
Yemen, Holland, and Thailand since 1976. He is currently a lecturer in the ELT<br />
Department of Thames Valley University (London), where he is assistant director of<br />
the M.A. in English Language Teaching program. His research interests include<br />
cross-cultural and interlanguage pragmatics, ethnographic approaches to language<br />
learning/teaching, and media discourse.<br />
ELLEN BIALYSTOK is professor of psychology at York University in Toronto,<br />
Canada. Her research is in cognitive development, language acquisition, and bilingualism.<br />
Current research projects include investigating the development of children's<br />
ability to use symbolic representation in the domains of language and number.<br />
Recent books include Communication Strategies and Language and Processing<br />
in Bilingual Children.<br />
SHOSHANA BLUM-KULKA is associate professor in the departments of Communication<br />
and Education at Hebrew University, Jerusalem, where she teaches<br />
courses in sociolinguistics, discourse analysis, language and education, and crosscultural<br />
pragmatics. Her research is in cross-cultural pragmatics, interlanguage<br />
pragmatics, and family discourse. She is co-author and co-editor of Cross-Cultural<br />
Pragmatics: Requests and Apologies and of Interlingual and Intercultural Communication.<br />
She is currently completing an extensive study of family discourse in<br />
Israeli and American families.<br />
JEAN BODMAN is president of the American Training Institute. She is a senior<br />
consultant for the U.S.I.A. and has written numerous ESL And EEL textbooks.<br />
251
252 About the Authors<br />
MIRIAM EISENSTEIN is associate professor of English Education at New York<br />
University. She is chair of the Second Language Acquisition Circle and author of<br />
The Dynamic Interlanguage.<br />
JULIANE HOUSE is professor of Applied Linguistics and head of the English<br />
Language Program at the University of Hamburg, Germany. She is author of A<br />
Model for Translation Quality Assessment and co-editor of Interlingual and Intercultural<br />
Communication and Cross-Cultural Pragmatics. Her research interests include<br />
translation theory, contrastive and interlanguage pragmatics, and misunderstandings<br />
in everyday life.<br />
GABRIELE <strong>KASPER</strong> is on the faculty of the M.A. in ESL and Ph.D. in Second<br />
Language Acquisition programs at the University of Hawaii. Before she discovered<br />
that she likes the tropics, she taught applied linguistics in Germany and Denmark.<br />
Her research interests are cross-cultural and interlanguage pragmatics, and psycholinguistic<br />
aspects of second language learning and use.<br />
ELITE OLSHTAIN is professor of Language Teaching at Tel Aviv University and<br />
dean of the School of Education. In 1979 she established the M.A. in TEFL<br />
program at Tel Aviv University. She teaches graduate courses in second language<br />
acquisition, discourse analysis, course design and policy making, the structure of<br />
English, and classroom-oriented research. Her research interests are in discourse<br />
analysis, language attrition, factors affecting success in language acquisition, and<br />
curriculum development. She had published articles in various professional journals<br />
and anthologies and is co-author (with Fraida Dubin) of Facilitating Language<br />
Learning, Reading by all Means, Reading on Purpose, and Course Design. She has<br />
also been involved in various materials development projects.<br />
RICHARD SCHMIDT teaches in the M.A. in ESL and Ph.D. in Second Language<br />
Acquisition programs at the University of Hawaii. He has done research and teacher<br />
training in Brazil, Egypt, Japan, Spain, and Thailand. His main research interests<br />
are in the areas of psycholinguistics and sociolinguistics and the intersection of the<br />
two.<br />
TOMOKO TAKAHASHI is associate professor of Languages and Cultures at Soka<br />
University in Tokyo (currently at the Los Angeles campus). She has published in the<br />
areas of second language acquisition, psycholinguistics and sociolinguistics. She<br />
has also written a number of books and textbooks for Japanese learners of English.<br />
LIORA WEINBACH is a teacher of Hebrew as a Second Language at Tel Aviv<br />
University and coordinator of a government project for the development of materials<br />
for the teaching of Hebrew in the (former) Soviet Union. She is co-author of The Art<br />
of Discourse (in Hebrew) and a series of basic Hebrew-foreign language dictionaries<br />
("2000+"). Her publications include textbooks for the teaching of Hebrew as a<br />
Second Language, and a book on the methodology of academic writing. She is<br />
currently researching inter-ethnic variation in the speech act of complaining within<br />
Israeli society.
About the Authors 253<br />
ELDA WEIZMAN is currently senior lecturer at the Bar-Han School for Translators<br />
and Interpreters, and at the Hebrew Department of Ben Gurion University of the<br />
Negev. Her publications include articles on the comparative analysis of journalistic<br />
texts in English, French, and Hebrew, and its implications for translators and<br />
learners; the nature of irony and hints, in a cross-cultural perspective; and the<br />
interpretation (and misinterpretation) of indirect speakers' meanings in written and<br />
oral discourse. She is currently engaged in a research project on discourse patterns<br />
in news interviews, supported by the Israeli Academy of Sciences and Humanities.<br />
JANE ZUENGLER is associate professor in the English Department, University of<br />
Wisconsin-Madison, where she teaches courses in second language acquisition and<br />
conducts research on sociolinguistic aspects of second language use. She has conducted<br />
research on accommodation theory and second language performance, and is<br />
presently researching the influence of content domain on L2 conversational participation.<br />
She has published articles in Studies in Second Language Acquisition,<br />
Applied Linguistics, and anthologies on second language research.