Interlanguage Pragmatics

Interlanguage Pragmatics

This page intentionally left blank

Interlanguage Pragmatics EDITED BY

Gabriele Kasper and Shoshana Blum-Kulka

New York Oxford OXFORD UNIVERSITY PRESS 1993

Oxford University Press Oxford New York Toronto Delhi Bombay Calcutta Madras Karachi Kuala Lumpur Singapore Hong Kong Tokyo Nairobi Dar es Salaam Cape Town Melbourne Auckland Madrid and associated companies in Berlin Ibadan

Copyright © 1993 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 200 Madison Avenue, New York, New York 10016 Oxford is a registered trademark of Oxford University Press, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Interlanguage pragmatics / edited by Gabriele Kasper and Shoshana Blum-Kulka. p. cm. ISBN 0-19-506602-2 1. Interlanguage (Language learning) 2. Pragmatics. 3. Speech acts (Linguistics) I. Blum-Kulka, Shoshana. P53.15193 1993 410.4'4—dc20 92-34791

135798642 Printed in the United States of America on acid-free paper

CONTENTS

Interlanguage Pragmatics: An Introduction, 3 Gabriele Kasper and Shoshana Blum-Kulka I

Cognitive Approaches to Interlanguage Pragmatic Development, 19

1 Consciousness, Learning, and Interlanguage Pragmatics, 21 Richard Schmidt 2 Symbolic Representation and Attentional Control in Pragmatic Competence, 43 Ellen Bialystok II

Speech Act Realization, 59

3 Expressing Gratitude in American English, 64 Miriam Eisenstein and Jean Bodman 4 Perception and Performance in Native and Nonnative Apology, 82 Marc L. Bergman and Gabriele Kasper 5 Interlanguage Features of the Speech Act of Complaining, 108 Elite Olshtain and Liora Weinbach 6 Interlanguage Requestive Hints, 123 Elda Weizman 1 Cross-Linguistic Influence in the Speech Act of Correction, 138 Tomoko Takahashi and Leslie M. Beebe III

Discourse Perspectives, 159

8 Toward a Model for the Analysis of Inappropriate Responses in Native/Nonnative Interactions, 161 Juliane House 9 Explaining NNS Interactional Behavior: The Effect of Conversational Topic, 184 Jane Zuengler

vi

Contents

10 The Metapragmatic Discourse of American-Israeli Families at Dinner, 196 Shoshana Blum-Kulka and Hadass Sheffer 11 Notes on the Interlanguage of Comity, 224 Guy Aston A B O U T THE A U T H O R S , 251

Interlanguage Pragmatics

This page intentionally left blank

Interlanguage Pragmatics: An Introduction GABRIELE KASPER and SHOSHANA BLUM-KULKA

The Scope of Interlanguage Pragmatics Interlanguage pragmatics (ILP) is a second-generation hybrid. As its name betrays, ILP belongs to two different disciplines, both of which are interdisciplinary. As a branch of Second Language Acquisition Research, ILP is one of several specializations in interlanguage studies, contrasting with interlanguage phonology, morphology, syntax, and semantics. As a subset of pragmatics, ILP figures as a sociolinguistic, psycholinguistic, or simply linguistic enterprise, depending on how one defines the scope of "pragmatics." For thorough discussion of definitional issues, see Leech (1983) and Levinson (1983). The perspective on pragmatics we adopt is an action-theoretical one, viewing pragmatics as the study of people's comprehension and production of linguistic action in context. Interlanguage pragmatics has consequently been defined as the study of nonnative speakers' use and acquisition of linguistic action patterns in a second language (L2) (e.g., Kasper, 1989b). Yet tying interlanguage pragmatics to nonnative speakers, or language learners, may narrow its scope too restrictively. As Blum-Kulka (1991; Blum-Kulka & Sheffer, Chapter 10) demonstrates through the case of American immigrants to Israel, speakers fully competent in two languages may create an intercultural style of speaking that is both related to and distinct from the styles prevalent in the two substrata, a style on which they rely regardless of the language being used. The intercultural style hypothesis is supported by many studies of cross-cultural communication, notably interactional sociolinguistics (e.g., Gumperz, 1982; Tannen, 1985) and research into the pragmatic behavior of immigrant populations across generations (e.g., Clyne, 1979; Clyne, Ball, & Neil, 1991). It also receives strong anecdotal support, worthy of systematic investigation, by highly proficient nonnative speakers whose L2 conversational behavior carries interlanguage-specific traits, and who claim at the same time that they do not abide by native norms any more when conversing in their native language. For instance, one of us was told by several of her Chinese students that in response to invitations and offers they wish to accept, they no longer engage 3

4

Interlanguage Pragmatics: An Introduction

in ritual refusal, as required by traditional Chinese culture. Some of her Japanese students claim that they are much more direct in their interaction in Japanese than they used to be before extended exposure to Western ways. Emerging intercultural styles, so prevalent in the international academic community, deserve interlanguage pragmaticists' close attention. Hence, it appears useful to include under ILP the study of intercultural styles brought about through language contact, the conditions for their emergence and change, the relationship to their substrata, and their communicative effectiveness. A look at the literature on ILP (cf. the overview in Kasper & Dahl, 1991), however, suggests that the populations studied have invariably been nonnative speakers, reflecting the status of ILP as a branch of second language research. While the present collection largely follows this line, Blum-Kulka and Sheffer (Chapter 10) extend the perspective to include native speakers' intercultural styles. Furthermore, according to researchers' labeling of their objects of study, ILP predominantly refers to the comprehension and production of linguistic action, including discourse regulation. An area of study that most investigators would clearly recognize as "pragmatic" yet that is not usually included under ILP is communication strategies. The de facto separation of pragmatics and communication strategies in second language studies reflects different alignments chosen by researchers in each area. The study of communication strategies has predominantly been grounded in psycholinguistic models of cognitive processing (e.g., Poulisse, 1990; Bialystok, 1990), whereas ILP has derived its theoretical and empirical foundation from general and especially cross-cultural pragmatics (e.g., Blum-Kulka, House, & Kasper, 1989). For the most part, research on communication strategies has examined learners' solutions to referential problems; ILP has focused on the illocutionary and politeness dimensions of speech act performance. While this division of labor reflects different research traditions, it has little theoretical support. In Bachman's (1990, 84ff.) model of communicative competence, for example, pragmatic competence, a component of language competence, subdivides into sociolinguistic and illocutionary competence, which in turn entails the ability to express a variety of communicative functions, such as making reference. Strategic competence is seen as processing ability, operating on the language competence in its entirety and including "strategic" solutions to comprehension or production problems. In this volume, strategic aspects of speech act performance and discourse participation are examined by Weizman (Chapter 6), Aston (Chapter 11), and House (Chapter 8); processing considerations for pragmatic development are proposed by Schmidt (Chapter 1) and Bialystok (Chapter 2).

Domains of ILP Pragmatic Comprehension Early studies focused on learners' attribution of illocutionary force and perception of politeness. Research on the attribution of illocutionary force has centered on the comprehension of indirect speech acts, factors contributing to ease or difficulty of pragmatic comprehension, the role of linguistic form and context information, and

Interlanguage Pragmatics: An Introduction

5

learner variables influencing force attribution. Carrell (1979) demonstrated that advanced L2 learners have complete access to conversational implicature, and make full use of their inferential ability in the comprehension of indirect speech acts. The only stumbling block for these learners was the "pope answer," a highly culturespecific violation of the maxim of relation. Bouton (1988) was interested in the impact of cultural background on the comprehension of indirect answers. He found a significant difference between six groups of learners from different cultural backgrounds and native speakers of American English. Comparison of the learner groups showed similar perceptions for German, Spanish-Portugese, and Taiwanese learners, differing from those of Korean, Japanese, and Chinese learners from the People's Republic of China. In addition to influence from learners' cultural background, Bouton also established an effect for type of implicature. Comprehension was easiest when the relevance maxim had been violated, whereas understated negative evaluation, a violation of the maxim of quantity, was more difficult to understand, for nonnative speakers as well as native speakers. In examining the relative effect of linguistic form and context information on learners' perception of indirect requests, Carrell (198 la, b) found that learners primarily relied on linguistic utterance features, regardless of their linguistic and cultural background, age, and proficiency. Her findings are at odds with those established by Ervin-Tripp, Strage, Lampert, and Bell (1987), who found that children acquirifig their first (LI) and second language strongly base their pragmatic comprehension on situational information, disregarding form. Other studies in developmental pragmatics (e.g., Reeder & Wakefield, 1987) support Ervin-Tripp et al.'s results (cf. Bialystok, Chapter 2). The apparent conflict in outcomes of Carrell's and Ervin-Tripp et al.'s studies can be reconciled by invoking Gibbs's (e.g., 1984) Conventional Meaning Model. In a series of experiments, Gibbs demonstrated that listeners directly access nonliteral meanings if linguistic forms and situational contexts are conventionalized. Absence of familiar and appropriate contexts and novel, nonconventionalized utterances requires (sequential) processing of literal and nonliteral meaning. Carrell's studies precisely illustrate the second condition, since the highly abstract task her subjects were faced with could only be solved via bottom-up processing. Ervin-Tripp et al.'s task conditions, on the other hand, fullfil the criteria that would bias listeners toward immediate processing of indirect meaning: the indirect requests were embedded in everyday situations familiar to the children, allowing them to apply situational schemata to their understanding of directive utterances. In partial replication of Gibbs's research on the processing of conventionally indirect requests (e.g., 1983), Takahashi and Roitblat (1992) examined through latency measurement whether Japanese learners of English reconstruct implied requestive force by processing both literal and implied meanings or immediately recover the nonliteral meaning (see also Takahashi, 1990, for a review of the literature on nonliteral utterance comprehension). Models of pragmatic comprehension, representing a variety of theoretical orientations, are reviewed by House (Chapter 8), who also offers an analysis of learners' pragmatic comprehension in ongoing interaction. In all of the studies cited above, English figures as L2. The only study on force attribution by learners of a different target language is a small investigation by Koike (1989), who examined how beginning classroom learners of Spanish with

6

Interlanguage Pragmatics: An Introduction

English as LI understood a Spanish request, apology, and command. Learners correctly identified the illocutions in 95% of the cases or better. In seeking cues to decide on force, learners were shown to rely most frequently on formulaic illocutionary force indicators such as "por favor" and "lo siento," and on words with key prepositional meanings. A related line of inquiry has examined how learners assess the politeness value of different speech act realization strategies. While most of the research on force attribution has studied on-line utterance comprehension, politeness perceptions have been investigated through off-line metapragmatic judgment tasks such as card sorting (Carrell & Konneker, 1981; Tanaka & Kawade, 1982), paired comparison (Walters, 1979), multiple choice (Tanaka & Kawade, 1982), and rating scales (Eraser, Rintell, & Walters, 1980; Rintell, 1981; Olshtain & Blum-Kulka, 1985; Kitao, 1990). Results confirm learners' ability to distinguish different degrees of politeness in conventions of means and forms, although their perceptions do not always agree with those of native speakers. Japanese learners of English largely agreed with American native speakers in their relative politeness judgments of syntactic modes (imperative, declarative, interrogative), the politeness marker "please," and of deferential address terms, but they differed in the politeness values attributed to request modification by tense and modals. Whereas American informants perceived positively worded requests as more polite than negatively worded ones, this assessment was not shared by the Japanese raters. Negative politeness strategies (Brown & Levinson, 1987) were rated as more polite by Americans and Japanese in the United States than by Japanese in Japan (Kitao, 1990). Learners' differential politeness perceptions have been attributed to a variety of factors. Learners were found to differ in the extent to which they base their politeness perceptions in L2 on those in LI. Spanish learners of English did not transfer their LI perceptions of formally equivalent requestive strategies to L2, rating their Spanish requests as more deferential than their English counterparts (Eraser et al., 1980). Learners of Hebrew, who appeared to base their politeness perceptions of L2 requests and apologies initially on LI, increased their tolerance for directness and positive politeness (Brown & Levinson, 1987) with length of residence in the target community (Olshtain & Blum-Kulka, 1985). Length of residence rather than L2 proficiency accounted for differential politeness perceptions in learners and L2 native speakers. The rec.eiver's age and sex influenced politeness assessment in Spanish learners' perception of English requests, whereas no such effects were noticeable in suggestions (Rintell, 1981). Comparing the perceptions of politeness in request strategies by Japanese EFL and ESL learners, Kitao (1990) found that exposure to English accounted for different ratings between those groups. In addition to learners' assessments of pragmalinguistic information, a few studies have also examined nonnative speakers' sociopragmatic perceptions. Probing into learners' "universal" and culture-specific assumptions about apology frequency and realization, Olshtain (1983) found that for Russian learners of Hebrew, the event rather than culture and language were the decisive variables. Englishspeaking learners felt less need to apologize in Hebrew than in English, thus suggesting a culture-specific approach. Asked to assess the weight and values of contextual factors in apologizing, German learners of English were found largely to

Interlanguage Pragmatics: An Introduction

7

agree with native speakers of British English, except for degree of imposition involved in the apology, which the German raters found to be consistently higher than the English judges (House, 1988). Ratings of contextual factors in apologies were also provided by Thai learners of English and American native speakers (Bergman & Kasper, Chapter 4). The factor on which these informants differed most was obligation to apologize. Japanese learners of English reported that in American society, refusal was a more socially acceptable act than in Japan, and could therefore appropriately be carried out more directly (Robinson, 1992). There are very few studies that examine learners' sociopragmatic perceptions by direct probing, such as rating tasks (e.g., House, 1988; Bergman & Kasper, Chapter 4; Olshtain & Weinbach, Chapter 5) and self-report in interviews (Cohen & Olshtain 1991; Robinson, 1992). Many more studies infer learners' L2 sociopragmatic knowledge through their production. Thus Beebe and collaborators (e.g., Beebe, Takahashi, & Uliss-Weltz, 1990; Beebe & Takahashi, 1989a, b), focusing on the effect of status on the performance of face-threatening acts by Japanese learners of English, found that the Japanese informants style-shifted more according to interlocutor status than speakers of American English (but see Takahashi & Beebe, Chapter 7, for a counterexample). What research still needs to demonstrate is how learners' sociopragmatic perceptions change over time, and how such change is reflected in their linguistic action patterns.

Production of Linguistic Action The available evidence suggests that regardless of a particular L1 and L2, and of the type of learning context (naturalistic vs. instructed), learners have access to the same range of realization strategies for linguistic action as native speakers, and demonstrate sensitivity to contextual constraints in their strategy choice (BlumKulka, 1982; Kasper, 1989b; Rintell & Mitchell, 1989). The main obstacle to learners' exploiting their "general pragmatic knowledge base" (Blum-Kulka, 1991) appears to be their restricted L2 linguistic knowledge, or difficulty in accessing it smoothly (e.g., Blum-Kulka, 1982; Koike, 1989; Edmondson & House, 1991). But other factors intervene: a lack of L2 pragmalinguistic sophistication, combined with negative transfer of sociopragmatic norms from LI or nonnative perceptions of L2 sociopragmatic norms, or even purposeful loyalty to LI cultural patterns, may yield deviations from native use at high proficiency levels as well (Blum-Kulka, 1991). While ILP research has by now covered a wide variety of typologically different Lls, there is still only a handful of languages studied as L2: in addition to different national varieties of English, a few studies have examined Hebrew (Blum-Kulka, 1982, 1991; Olshtain, 1983; Olshtain & Cohen, 1989; Olshtain & Weinbach, Chapter 5), German (Faerch & Kasper, 1989), Norwegian (Svanes, 1989), Spanish (Koike, 1989), and Japanese (Sawyer, 1992) as targets. Until the scope of target languages has been considerably broadened, universality claims need to be voiced with caution. Learners' distribution patterns of strategies and forms have been shown to vary from those of native speakers. Some studies report that learners prefer more direct modes of conveying pragmatic intent than native speakers; others suggest the oppo-

8

Interlanguage Pragmatics: An Introduction

site. Preference for higher directness in IL than in both or either LI or L2 has been documented in learners' requests (Tanaka 1988; Koike, 1989; Fukushima, 1990), making and rejecting of suggestions (Bardovi-Harlig & Hartford, 1990), refusals (Robinson, 1992), and a variety of conflictive acts (Kasper, 1981). By contrast, nonnative speakers of Hebrew preferred less direct requests (Blum-Kulka, 1982, 1991) and complaints (Olshtain & Weinbach, Chapter 5) than native speakers. Explanations of these findings have been sought in the scope of learners' linguistic IL knowledge, transfer from L I , and perceptions of what is sociopragmatically appropriate in the target community. Some studies noted differences in politeness approach between native speakers and nonnative speakers. Venezuelan Spanishspeaking learners of English systematically used positive politeness strategies when apologizing to a host for not having attended her party, whereas American English native speakers preferred a negative politeness approach (Garcia, 1989). Conversely, in pre-trial interviews, Athabaskan defendants employed negative politeness strategies, whereas white defendants as well as the interviewers deployed more positive politeness (Scollon & Scollon, 1983). In order to repair uncomfortable moments in academic advising sessions at an American university, students who were native speakers of English predominantly relied on positive politeness strategies, while Taiwanese students adopted a negative politeness approach (Fiksdal 1990). Less reliance by nonnative speakers on positive politeness has also been observed in the performance of corrections by Japanese speakers of English (Takahashi & Beebe, Chapter 7). American informants in status-higher positions tended to preface a correction by a positive remark, whereas Japanese nonnative speakers used this solidarity strategy very infrequently. Nonnative speakers' strategy choice is sometimes less responsive to contextual factors than native speakers'. Japanese learners of English used the same (direct, barely mitigated) requestive strategies in conversation with status-unequal and socially distant interlocutors as with status-equal and familiar coparticipants, whereas American English controls varied their request behavior in the two conditions (Tanaka, 1988). Japanese learners of English also underdifferentiated their realizations of offers and requests in three conditions of social distance (Fukushima, 1990). In expressing gratitude, native speakers of American English varied the length of their speech activity according to degree of indebtedness; no such effect was found for nonnative speakers of different linguistic backgrounds (Eisenstein & Bodman 1986, Chapter 3; Bodman & Eisenstein, 1988). The quality and range of linguistic forms by which linguistic action can be implemented and modified has consistently been shown to differ between native speakers and nonnative speakers, the nonnative speakers' repertoire typically being more restricted and less complex than native speakers' (Scarcella, 1979; Kasper, 1981; Blum-Kulka, 1982, 1991; Schmidt, 1983; House & Kasper, 1987; Trosborg, 1987, Faerch & Kasper, 1989; Eisenstein & Bodman, Chapter 3). While some of these findings may simply reflect the state of learners' lexical and syntactic knowledge, the issue becomes a clearly pragmalinguistic one when learners demonstrably "know" a particular lexical item or syntactic structure yet use it in a way that does not convey the intended illocutionary force or politeness value. In the interlanguage

Interlanguage Pragmatics: An Introduction

9

of nonnative speakers of Hebrew, requests are lended an unintended whining emphasis by the use of "bevakasha" (please) in sentence-initial (rather than intra- or postsentential) position, and such pragmalinguistic deviations are singled out by native judges as "nonnative" (Blum-Kulka, 1991). One area where insufficient control of pragmalinguistic knowledge is particularly obvious is that of pragmatic routines. Coulmas's (1981) contention that routine formulas are a serious stumbling block for nonnative speakers has been supported by nearly every ILP study (e.g., Scarcella, 1979; House, 1988; Eisenstein & Bodman, 1986; Kasper, 1989a; Wildner-Bassett, 1984; Fukushima, 1990), yet there has been little systematic investigation of this phenomenon. There is evidence of learners supplying near-literal translations of LI routines: for example, the German, "entschuldigen Sie bitte" (English, "excuse me, please") instead of "I'm sorry" (House, 1988); of using a translation equivalent for an LI routine where none is used in L2, such as prefacing a high-imposition request with "I'm sorry" (from Japanese "sumimasen" or "gomennasai" [Fukushima, 1990]); and of failing to use any kind of routine where one would be required, such as failing to offer an expression of gratitude (Kasper, 1981). Bodman and Eisenstein (1988) observed that in the attempt to express gratitude in English, learners would use literal translations of LI proverbial expressions in written production questionnaires but not in role-plays, which displayed considerable disfluencies but no overt use of LI proverbs (also Eisenstein & Bodman, Chapter 3). At the nonroutinized end of speech act production, learners have been found to engage in more speech activity than native speakers. The "waffle phenomenon" (Edmonson & House, 1991) has been noted in requests (Blum-Kulka & Olshtain, 1986; House & Kasper, 1987; Faerch & Kasper, 1989; Weizman, Chapter 6), apologies (House, 1988, Bergman & Kasper, Chapter 4), and complaints (Olshtain & Weinbach, Chapter 5), as well as in referential communication (Bongaerts, Kellerman, & Bentlage, 1987; Tarone & Yule, 1987; Yule & Tarone, 1990). According to Blum-Kulka and Olshtain (1986), waffling is proficiency-dependent, being strongest at an intermediate stage when learners possess the linguistic means to say as much as they wish, yet at the same time feel more of a need to be explicit about their communicative goals and the reasoning behind them than more acculturated nonnative speakers do. Edmondson and House (1991) point out, however, that the waffling effect in speech act realization is observable only in learners' written responses to production questionnaires, not in role-plays. This observation corroborates Bodman and Eisenstein's findings about the differential use of proverbial material in written and oral-interactive production, and the much shorter contributions made in role-plays by nonnative speakers in comparison with native speakers (Eisenstein & Bodmann, Chapter 3). In Edmondson and House's analysis (see Bergman & Kasper, Chapter 4, for more discussion), learners' extensive use of supportive strategies in the absence of formulaic routines suggests that nonroutinized material functions to compensate for the lack of automatized discourse routines. It will be worthwhile for IL pragmaticists to examine whether this hypothesis bears out across languages and tasks, and whether it interacts with factors such as proficiency and length of residence.

10

Interlanguage Pragmatics: An Introduction

Development of Pragmatic Competence The bulk of ILP research focuses on nonnative speakers' use of pragmatic knowledge in comprehension and production, rather than on development. This focus is also adopted by the data-based studies in this book. Of the available developmental investigations, some cross-sectional studies did not find proficiency effects in learners' strategy selection (Takahashi & Beebe, 1987; Trosborg, 1987; Svanes, 1991) whereas others did (Takahashi & DuFon, 1989; Maeshiba, Yoshinaga, Kasper & Ross, in press). However, developmental effects were observable in learners' repertoires of pragmatic routines and modality markers (Scarcella, 1979; Trosborg, 1987). Possibly, the inconsistent results reflect instrument effects and the difficulty of determining precisely learners' proficiency levels across studies. The few longitudinal studies to date indicate distinct developmental patterns in learners' request realization (Schmidt, 1983; Ellis, 1992) and use of a sentence-final particle in Japanese (Sawyer, 1992). They strongly suggest the need for more longitudinal studies in naturalistic settings, observing learners from the very beginning of their language acquisition process. Equally important as a reliable and valid empirical data base is a theoretical framework to account for pragmatic learning. In this book two such frameworks, grounded in different models of cognitive processing, are proposed by Bialystok (Chapter 1) and Schmidt (Chapter 2; see also Kasper & Schmidt, 1992).

Pragmatic Transfer Influence from learners' native language and culture on their IL pragmatic knowledge and performance has been amply documented. Because of its potential risk to communicative success, the focus has been on negative transfer; that is, the influence of LI pragmatic competence on IL pragmatic knowledge that differs from the L2 target. Positive transfer, that is, pragmatic behaviors or other knowledge displays consistent across LI, IL, and L2, have received less attention. We think this is so because positive transfer usually results in communicative success and therefore is less exciting to study. Furthermore, it is methodologically difficult to disentangle positive transfer from universal pragmatic knowledge and generalization on the basis of available IL pragmatic knowledge. Negative transfer has been attested to at the sociopragmatic level, influencing learners' perception of status relationships (Beebe, Takahashi, & Uliss-Weltz, 1990; Takahashi & Beebe, Chapter 7), of the appropriateness of carrying out refusals (Robinson, 1992), of the need to apologize (Olshtain, 1983) and to express gratitude (Eisenstein & Bodman, Chapter 3); their lack of accommodation to target norms for complimenting, responding to compliments, and negotiating invitations (Wolfson, 1989); their choice of politeness style (Garcia, 1989; Olshtain & Cohen, 1989, Blum-Kulka, 1982) and of particular strategic options (House, 1988; Beebe, Takahashi, & Uliss-Weltz, 1990). Evidence for pragmalinguistic transfer at the level of form-force mapping has been less documented in the literature than anecdotal accounts would suggest (Blum-Kulka, 1982; Faerch & Kasper, 1989 in requesting; House, 1988; Olshtain & Cohen, 1989 in apologizing; Bodman & Eisenstein, 1988,

Interlanguage Pragmatics: An Introduction

11

in thanking); possibly learners' proficiency levels in most ILP studies were too high for negative form-force transfer to occur. Most of the reported pragmalinguistic transfer affects the strategic options and forms that modify the politeness value of a linguistic act. Learners' choices of semantic formulas and linguistic tokens for apologizing were influenced by LI patterns (Olshtain & Cohen, 1989; Trosborg, 1987; House, 1988; Bergman & Kasper, Chapter 4; Maeshiba, Yoshinaga, Kasper & Ross, in press). LI lexical and syntactic material used to mitigate requestive force was transferred to learners' IL request performance (Blum-Kulka, 1982; BlumKulka & Levenston, 1987; Faerch & Kasper, 1989). Questioning patterns employed in Japanese to express disagreement were used as disagreement strategies in Japanese learners' English (Beebe & Takahashi, 1989b). Although the literature abounds in evidence for pragmatic transfer, little has yet been done to investigate more closely the conditions under which pragmatic transfer is or is not operative. Whereas there is some indication from performance data (Kasper, 1981; House & Kasper, 1987; Faerch & Kasper, 1989) and retrospective reports (Olshtain, 1983; Cohen & Olshtain, 1991; Robinson 1992) that learners invoke criteria such as assessment of linguistic and cultural distance, and of the specificity or sameness of pragmalinguistic or sociopragmatic patterns in L1/L2, there is only one study to date that systematically examines pragmatic transferability. Takahashi (1992) found that the transferability of conventionally indirect requests from Japanese to English was highly context-dependent, and varied with learners' proficiency and familiarity with the situation. At a conceptual level, it is important to note that "negative" pragmatic transfer does not necessarily reflect lack of competence in the pragmatics of the target community. When nonnative speakers communicate in a style different from native ways of speaking, it is a matter of attribution if this style is seen as lacking in some way, or just different, and if its maintenance over time is considered negatively, as fossilization, or positively, as a marker of cultural identity. The degree of sociocultural accommodation to the L2 culture may be as well a matter of choice as of ability. A foreign accent, for example, can well shield a nonnative speaker, identifying her as nonnative, and thereby flexing the norms by which she is judged and lending her a certain latitude in choosing her ways of speaking. The desirable goal for the high-proficiency second language speaker, be it in contexts of immigration or in the use of L2 in cross-cultural communication, may well be that of disidentification, rather than absolute convergence. There is thus much room for future research, which will be necessary in order to advance beyond merely ascertaining pragmatic transfer to understanding its differential modes of operation and symbolic functions.

Communicative Effect Grammatical or phonological IL deviations from target language norms have the "advantage" of being easily recognizable by native speakers. A "nonnative" identification also serves to protect such speakers from the risk of the peculiarities of their speech being attributed to flaws in their personality, or ethnocultural origins. Higher levels of L2 proficiency allow for ease of communication, but still leave open the

12

Interlanguage Pragmatics: An Introduction

possibility of pragmatic failure at the pragmalinguistic and sociopragmatic level (Thomas, 1983). Differing from grammatical errors, pragmatic failure is neither easily recognizable by interlocutors without training in pragmatics, nor explained away by recognizing the speaker as nonnative. Blum-Kulka and Olshtain (1986) explain pragmatic failure as linked to cultural variability in the implementation of Grice's (1975) conversational model: certain types of pragmatic deviations from target norms, such as the learner's tendency to verbosity (discussed above) are seen as violations of a cultural norm for the balance required between the maxims of clarity and quantity. Others (e.g., Riley, 1989) are concerned with defining the scope of communicative behavior amenable to pragmatic failure, insisting that both verbal and nonverbal phenomena be included. Interactional sociolinguistics has provided rich evidence for miscommunication resulting from interlocutors' differences in the use of contextualization conventions (Gumperz, 1982) and different conversational styles (Tannen, 1981; 1985). While disturbing in everyday communication, incompatible styles, resulting in mutual misreadings of the other person's intention and attitude, can have serious consequences in gate-keeping encounters, leading to unfavorable outcomes for the client (Erickson & Shultz, 1982; Scollon & Scollon, 1983). There have been three major approaches to the study of pragmatic failure. Miscommunication research as referred to above provides micro-sociolinguistic analyses of naturalistic encounters, minutely identifying problematic features at the levels of prosody, pragmatics, syntax, lexis, discourse organization, conversational management, and nonverbal behavior. Qualitative analysis of performance data is sometimes combined with quantitative measures of particular features and with retrospective interviews of the participants. Micro-sociolinguistic analysis ascertains conversational style differences and identifies instances where such differences become problematic, but does not usually inquire into the origin of different conversational styles. A second approach is contrastive pragmatics, involving the crosscultural and cross-linguistic comparison of speech act realization patterns through identifying similarities and differences between the pairs or groups of languages studied. Such research is purely descriptive, having no predictive power for the study of IL pragmatics and actual communicative practices in cross-cultural encounters but serving an important hypothesis-generating and explanatory role in studies of interlanguage pragmatic performance and knowledge. In order to study the relationship between learners' prior knowledge and pragmatic performance, a third line of investigation needs to be called upon: ILP. In its canonical form, ILP research, following received methodology in interlanguage studies (Selinker, 1972) by comparing learners' IL production and comprehension with parallel LI and L2 data, provides the methodological tool to determine where and how learners' pragmatic performance differs from L2, and to establish where IL specific behaviors appear to be influenced by learners' LI knowledge. (As argued above, determining with any certainty whether or not transfer has actually been operative requires additional measures, such as retrospective reports and transferability studies.) Furthermore, ILP can establish which IL-specific pragmatic behaviors reside in sources other than transfer (e.g., see Kasper, 1981; Blum-Kulka, 1982; Schmidt, 1983; Olshtain & Cohen, 1989).

Interlanguage Pragmatics: An Introduction

13

However, just as contrastive pragmatic study is unable to identify pragmatic transfer, learner-focused ILP, unless supplemented by other measures, such as ratings of learners' performance by native speakers (Fraser et al., 1980; Eisenstein & Bodman, 1986, Chapter 3), cannot make claims about communicative effect. "Negative" pragmatic transfer may or may not result in pragmatic failure. Erickson and Shultz (1982) demonstrate that successful co-membershiping can neutralize communicative style differences. Tannen (1985) cites examples of "negative" transfer that lead to positive attributions (reminiscent of "charming" foreign accents), and styles that, though different, are complementary rather than conflicting, thus leading to successful outcomes. Clyne (1979) concludes that communication conflict arises not so much from local differences in linguistic action patterns but from features that impinge on interlocutors' perceptions of power, trust and solidarity. While there is a legitimate ecological interest in the identification of miscommunication and its causal relationship to LI communicative practices, nonnative communicative styles, whether transfer-induced or not, do not necessarily result in pragmatic failure. The time-honored contrastive equation "difference = negative transfer = error" has proved to be just as little true for pragmatics as for other domains of nonnative language learning and use, though it has an indisputable heuristic value. To summarize, the strength of interactional sociolinguistics is identifying pragmatic failure; of contrastive pragmatics, identifying cross-cultural and crosslinguistic pragmatic differences and similarities; and of ILP, identifying learnerspecific pragmatic behaviors and their relationship to learners' LI and L2. A fullfledged research program that sheds light on the relationship between cross-cultural differences, IL-specific pragmatic features, including transfer, and communicative effects will usefully combine methods from all three areas of investigation. Reflecting the various approaches employed in studying interlanguage pragmatics, this book is organized into three parts: (I) Cognitive approaches to interlanguage pragmatic development, (II) Speech act realization, and (III) Discourse perspectives. For an overview of the chapters, the reader is referred to the introduction to each part.

References Bachman L. F. (1990). Fundamental considerations in language testing. Oxford: Oxford University Press. Bardovi-Harlig, K., & Hartford, B. S. (1990). Congruence in native and nonnative conversations: Status balance in the academic advising session. Language Learning, 40, 467501. Beebe, L. M., & Takahashi, T. (1989a). Do you have a bag? Social status and patterned variation in second language acquisition. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (103-25). Clevedon and Philadelphia: Multilingual Matters. Becbe, L. M., & Takahashi, T. (19895). Sociolinguistic variation in face-threatening speech

14

Interlanguage Pragmatics: An Introduction

acts. In M. Eisenstein (Ed.), The dynamic interlanguage (199-218). New York: Plenum. Beebe, L. M., Takahashi, T., & Uliss-Weltz, R. (1990). Pragmatic transfer in ESL refusals. In R. C. Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative competence in a second language (55-73). New York: Newbury House. Bialystok, E. (1990). Communication strategies. Oxford: Basil Blackwell. Blum-Kulka, S. (1982). Learning how to say what you mean in a second language. Applied Linguistics, 3, 29-59. Blum-Kulka, S. (1991). Interlanguage pragmatics: The case of requests. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds.), Foreign/second language pedagogy research (255-72). Clevedon and Philadelphia: Multilingual Matters. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Blum-Kulka, S., & Levenston, E. (1987). Lexical-grammatical pragmatic indicators. Studies in Second Language Acquisition, 9, 155—70. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Studies in Second Language Acquisition, 8, 47-61. Bodman, J., & Eisenstein, M. (1988). May God increase your bounty: The expression of gratitude in English by native and nonnative speakers. Cross Currents, 15, 1-21. Bongaerts, T., Kellerman, E., & Bentlage, A. (1987). Perspective and proficiency in L2 referential communication. Studies in Second Language Acquisition, 9, 171-99. Bouton, L. (1988). A cross-cultural study of ability to interpret implicatures in English. World Englishes, 17, 183-96. Brown, P., & Levinson, S. C. (1987). Politeness. Cambridge: Cambridge University Press. Carrell, P. L. (1979). Indirect speech acts in ESL: Indirect answers. In C. A. Yorio, K. Perkins & J. Schachter (Eds.), On TESOL '79 (297-307). Washington, DC: TESOL. Carrell, P. L. (1981a). Children's understanding of indirect requests: Comparing child and adult comprehension. Journal of Child Language, 8, 329-45. Carrell, P. L. (1981b). Relative difficulty of request forms in L1/L2 comprehension. In M. Hines & W. Rutherford (Eds.), On TESOL '81 (141-52). Washington, DC: TESOL. Carrell, P. L., & Konneker, B. H. (1981). Politeness: Comparing native and nonnative judgments. Language Learning, 31, 17-31. Clyne, M. (1979). Communicative competences in contact. ITL, 43, 17-37. Clyne, M., Ball, M., & Neil, D. (1991). Intercultural communication at work in Australia: Complaints and apologies in turns. Multilingua, JO, 251-73. Cohen, A. D., & Olshtain, E. (1991, October). The production of speech acts by nonnatives. Paper presented at the Theory construction and research methodology in second language acquisition conference, Lansing, MI. Coulmas, F. (1981). "Poison to your soul." Thanks and apologies contrastively viewed. In F. Coulmas (Ed.), Conversational routine (273—88). The Hague: Mouton. Edmondson, W., & House, J. (1991). Do learners talk too much? The waffle phenomenon in interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds.), Foreign!second language pedagogy research (273—86). Clevedon and Philadelphia: Multilingual Matters. Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by native and non-native speakers of American English. Applied Linguistics, 7, 167-85.

Inlerlanguage Pragmatics: An Introduction

15

Ellis, R. (1992). Learning to communicate in the classroom. Studies in Second Language Acquisition, 14, 1-23. Erickson, F., & Shultz, J. (1982). The counselor as gatekeeper: Social interaction in interviews. New York: Academic Press. Ervin-Tripp, S., Stragc, A., Lampert, M, & Bell, N. (1987). Understanding requests. Linguistics, 25, 107-43. Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request realization. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics (221 -47). Norwood, NJ: Ablex. Fiksdal, S. (1990). The right time and pace: A microanalysis of cross-cultural Datekeeping interviews. Norwood, NJ: Ablex. Fraser, B., Rintell, E., & Walters, J. (1980). An approach to conducting research on the acquisition of pragmatic competence in a second language. In D. Larsen-Freeman (Ed.), Discourse analysis in second language research (75-91). Rowley, MA: Newbury House. Fukushima, S. (1990). Offers and requests: Performance by Japanese learners of English. World Englishes, 9, 317-25. Garcia, C. (1989). Apologizing in English. Politeness strategies used by native and nonnative speakers. Multilingua, 8, 3-20. Gibbs, R. W. (1983). Do people always process the literal meaning of indirect requests? Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 524-33. Gibbs, R. W. (1984). Literal meaning and psychological theory. Cognitive Science, 8, 275304. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics. (Vol. 3): Speech acts (41-58). New York: Academic Press. Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. House, J. (1988). "Oh excuse me please . . .": Apologizing in a foreign language. In B. Kettemann, P. Bierbaumer, A. Fill, & A. Karpf (Eds.), Englisch alsZweitsprache (303-27). Tubingen: Narr. House, J., & Kasper, G. (1987). Interlanguage pragmatics: Requesting in a foreign language. In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance. Festschrift for Werner Mullen (1250-88). Tubingen: Narr. Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr.

Kasper, G. (1989a). Interactive procedures in interlanguage discourse. In W. Oleksy (Ed.), Contrastive pragmatics (189-229). Amsterdam: Benjamins. Kasper, G. (1989b). Variation in interlanguage speech act realization. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters. Kasper, G., & Dahl, M. (1991). Research methods in interlanguage pragmatics. Studies in Second Language Acquisition, 13, 215-47. Kasper, G., & Schmidt, R. (1992, March). Interlanguage pragmatics and language learning. Paper presented at the Sociolinguistics and TESOL Colloquium. TESOL Congress, Vancouver. Kitao, K. (1990). A study of Japanese and American perceptions of politeness in requests. Doshida Studies in English, 50, 178-210. Koike, D. A. (1989). Pragmatic competence and adult L2 acquisition: Speech acts in interlanguage. Modern Language Journal, 73, 79-89. Leech, G. N. (1983). Principles oj pragmatics. London: Longman. Lcvinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press.

16

Interlanguage Pragmatics: An Introduction

Maeshiba, N., Yoshinaga, N., Kaspcr, G., & Ross, S. (in press). Transfer and proficiency in interlanguage apologizing. In J. Neu & S. Gass (Eds.), Speech acts across cultures. Berlin: Mouton de Gruyter. Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49). Rowley, MA: Newbury House. Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language acquisition (303-25). Rowley, MA: Newbury House. Olshtain, E., & Cohen, A. (1989). Speech act behavior across languages. In H. W. Dechert & M. Raupach (Eds.), Transfer in language production (53-67). Norwood, NJ: Ablex. Poulisse, N. (1990). Compensatory strategies in L2 production. Amsterdam: Foris. Reeder, K., & Wakefield, J. (1987). The development of young children's speech act comprehension: How much language is necessary? Applied Psycholinguistics, 8, 1-18. Riley, P. (1989). "Well, don't blame me": On the interpretation of pragmatic errors. In W. Olesky (Ed.), Contrastive pragmatics (231-51). Amsterdam: Benjamins. Rintell E. (1981). Sociolinguistic variation and pragmatic ability: A look at learners. International Journal of the Sociology of Language, 27, 11-34. Rintell, E., & Mitchell, C. (1989). Studying requests and apologies: An inquiry into method. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics (24872). Norwood, NJ: Ablex. Robinson, M. A. (1992). Introspective methodology in interlanguage pragmatics research. In G. Kasper (Ed.), Pragmatics of Japanese as a native and foreign language. Technical Report No. 3 (27-82). Second Language Teaching & Curriculum Center, University of Hawaii at Manoa. Sawyer, M. (1992). The development of pragmatics in Japanese as a second language: The sentence-final particle ne. In G. Kasper (Ed.), Pragmatics of Japanese as a native and foreign language. Technical Report No. 3 (83-125), Second Language Teaching & Curriculum Center, University of Hawaii at Manoa. Scarcella, R. (1979). On speaking politely in a second language. In C. A. Yorio, K. Perkins, & J. Schachter (Eds.), On TESOL '79 (275-87). Washington, DC: TESOL. Schmidt, R. (1983). Interaction, acculturation and the acquisition of communicative competence. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and second language acquisition (137-74). Rowley, MA: Newbury House. Scollon, R., & Scollon, S.B.K. (1983). Face in interethnic communication. In J. C. Richards & R. Schmidt (Eds.), Language and communication (156-88). London: Longman. Selinker, L. (1972). Interlanguage. 1RAL 10, 209-30. Svanes, B. (1989). En undersoekelse av realisasjonsmoensteret for spraakhandlingen "aa be noen om aa gjoere noe." Maal og Minne, 1-2, 89-107. Takahashi, S. (1990). Exploring the comprehension process of nonliteral utterances and some implications for automaticity. University of Hawaii Working Papers in ESL, 9, 67-97. Takahashi, S. (1992). Transferability of indirect request strategies. University of Hawaii Working Papers in ESL, 11, 69-124. Takahashi, S., & Roitblat, H. L. (1992, July). Comprehension of'L2 indirect requests. Paper presented at Pacific Second Language Research Forum, Sydney. Takahashi, S., & DuFon, P. (1989). Cross-linguistic influence in indirectness: The case of English directives performed by native Japanese speakers. Unpublished manuscript, University of Hawaii at Manoa.

Interlanguage Pragmatics: An Introduction

17

Takahashi, T., & Beebe, L. M. (1987). The development of pragmatic competence by Japanese learners of English. JALT Journal, 8, 131-55. Tanaka, N. (1988). Politeness: Some problems for Japanese speakers of English. JALT Journal, 9, 81-102. Tanaka, S., & Kawade, S. (1982). Politeness strategies and second language acquisition. Studies in Second Language Acquisition, 5, 18—33. Tannen, D. (1981). Indirectness in discourse: Ethnicity as conversational style. Discourse Processes, 4, 221-38. Tannen, D. (1985). Cross-cultural communication. In T. A. Van Dijk (Ed.), Handbook of discourse analysis (Vol. 4, 203-15). London: Academic Press. Tarone, E., & Yule, G. (1987). Communication strategies in East-West interactions. In L. Smith (Ed.), Discourse across cultures (49-65). New York: Prentice-Hall. Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112. Trosborg, A. (1987). Apology strategies in natives/nonnatives. Journal of Pragmatics, 11, 147-67. Walters, J. (1979). The perception of deference in English and Spanish. In C. A. Yorio, K. Perkins, & J. Schachter (Eds.), On TESOL '79 (288-96). Washington, DC: TESOL. Wildner-Bassett, M. (1984). Improving pragmatic aspects of learners' interlanguage. Tubingen: Narr. Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House. Yule, G., & Tarone, E. (1990). Eliciting the performance of strategic competence. In R. C. Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative competence (179-94). New York: Newbury House.

This page intentionally left blank

I COGNITIVE APPROACHES TO INTERLANGUAGE PRAGMATIC DEVELOPMENT

Although there is a comprehensive literature on children's development of pragmatic competence, little work has been done on the acquisition of pragmatic ability by adult second language learners. Schmidt (Chapter 1) and Bialystok (Chapter 2) discuss two central issues in adult pragmatic learning, each proposing a different theoretical framework to address the problem. Schmidt explores the role of conscious awareness in the acquisition of pragmatic competence. Based on a critical examination of recent work in experimental psychology and support from diary reports, Schmidt concludes that the necessary condition for pragmatic learning to take place is attention to the pragmalinguistic or sociopragmatic information to be acquired. While implicit and incidental learning seems possible, noticing and generalizing about relevant features in the input is highly facilitative. Schmidt's analysis has important implications for research methodology in pragmatics and language teaching. Theory and cited illustrations encourage the use of self-report as a method of data collection in interlanguage pragmatics. For second language instruction, Schmidt's analysis indicates activities that alert the learner to pragmalinguistic and sociopragmatic features as well as a consciousness-raising approach to the teaching of L2 pragmatics. Bialystok examines the learning task in the pragmatic domain of adult second language learners as opposed to child first language acquirers. Applying her twodimensional model of language learning and use to pragmatics, she concludes that the learning problem for children acquiring their first language and adults learning a second language is quite different. Children primarily need to develop an analyzed knowledge of form-function mappings. For them, acquiring pragmalinguistic resources and their contextual distribution patterns takes precedence over the need to develop the control strategies required for efficient use of pragmatic knowledge. With adults, the order of learning tasks is reversed. While they also have to acquire new pragmalinguistic knowledge, their primary problem appears to be one of control, increasing their ability to process pragmatic information smoothly in contexts and making socially and contextually appropriate selections of linguistic forms. 19

20

Cognitive Approaches to Interlanguage Pragmatic Development

Schmidt's and Bialystok's proposals offer different but compatible approaches to pragmatic interlanguage development: Schmidt is concerned with the conditions for initial intake; Bialystok addresses the cognitive dimensions on which pragmatic interlanguage competence evolves. Both approaches, we hope, will inspire databased studies of adult pragmatic learning.

1

Consciousness, Learning and Interlanguage Pragmatics RICHARD SCHMIDT

During the past decade, the study of interlanguage pragmatics has produced important empirical findings, primarily through the identification and comparison of speech act realization patterns in various languages based on data from both native and nonnative speakers. In addition to this focus on product, some attention has been paid to the processes of comprehension and production in second language pragmatics (Faerch & Kasper, 1984, 1989; Kasper, 1984). In contrast to these concerns, there has been little discussion of how pragmatic abilities are acquired in a second language. This chapter is concerned with the ways in which consciousness may be involved in learning the principles of discourse and pragmatics in a second language.' The role of conscious and nonconscious processes in the acquisition of morphosyntax has been hotly debated within the field of second language acquisition (Krashen, 1981, 1983; Munsell & Carr, 1981; Rutherford & Sharwood Smith, 1985; Seliger, 1983; Sharwood Smith, 1981), but these debates have ignored pragmatic and discoursal abilities. My discussion will of necessity be speculative, drawing on current theories of the role of consciousness in human learning in general, drawn primarily from cognitive science and experimental psychology, with some suggestions for the extension of general principles to the learning of pragmatics. This is an issue with important pedagogical implications. In second language teaching, as Richards (1990) points out, there are currently two major approaches to the teaching of conversation in second language programs. The first is an indirect approach, in which conversational competence is seen as the product of engaging learners in conversational interaction; the underlying assumption is that the ability to carry on conversation (which includes pragmatic ability and other factors as well) is something that is acquired simply in the course of doing it. In practice, this leads to the use of group work activities or other tasks that require interaction. The second, a more direct approach, focuses explicitly on the strategies involved in conversation and emphasizes consciousness-raising concerning these strategies. 21

22

Cognitive Approaches to Interlanguage Pragmatic Development

Is Pragmatic Knowledge Conscious or Unconscious? Wolfson has argued that native speaker knowledge of what she calls rules of speaking (which include both pragmatic and discoursal rules) is mostly unconscious: Rules of speaking and, more generally, norms of interaction are ... largely unconscious. What this means is that native speakers, although perfectly competent in the uses and interpretation of the patterns of speech behavior which prevail in their own communities are, with the exception of a few explicitly taught formulas, not even aware of the patterned nature of their speech behavior. [Native speakers] . . . are not able . . . to describe their own rules of speaking. (Wolfson, 1989, 37)

Wolfson cites several types of evidence in support of her claim that speakers do not have reliable information concerning the ways in which they use language: people who are bilingual or bidialectal may switch from one language or variety to another without being aware of it and cannot accurately report their use of these languages or varieties (Blom & Gumperz, 1972); native speakers often report that they typically use or do not use specific forms, but their descriptions do not match reality (Wolfson, D'Amico-Reisner, & Huber, 1983); even highly trained linguists who rely on intuition to describe such phenomena as the differences between men's and women's speech (e.g., Lakoff, 1973) may find their intuitions proven incorrect; textbook writers, who almost always rely on intuition rather than empirical data, provide information regarding language use that is frequently wrong (Cathcart, 1989; Holmes, 1988; Williams, 1988). There are several reasons why we should expect native speakers' intuitions about these matters to be fallible. First, there is the obvious problem of the intrusion of prescriptive norms, stereotypes, and folk-linguistic beliefs; when asked what they do, informants are likely to report what they think they should do. Second, this kind of introspection violates basic principles distinguishing between potentially accurate and inaccurate verbal reports (Ericsson & Simon, 1984; Nisbett & Wilson, 1977), because such intuitions are general rather than specific, retrospective rather than concurrent, and sometimes call for information that could not be reported even if the other conditions were met. Ericsson and Simon (1984) propose that the only information that is potentially available for accurate self-report is information that is attended to in short-term memory in the performance of a task. In other words, in order to give an accurate report of your own performance, you must have been paying attention and aware of what you were doing at the time. Speech act realizations and other aspects of rules of speaking are often produced by fluent speakers with little conscious reflection or deliberation during their performance, and are therefore not accurately reportable. If accurate self-reports are limited to reporting information that has been stored as a result of one's own conscious thought processes, intuitions about the linguistic behavior of groups are particularly suspect (Cameron, 1985). The evidence cited by Wolfson (1989) shows that native speakers do not necessarily have access to their own rules of speaking, but it fails to show that speakers never have any access to such rules. Blum-Kulka (Chapter 10) and Olshtain and

Consciousness, Learning, and Interlanguage Pragmatics

23

Blum-Kulka (1989) have argued that Hebrew-English bilinguals in Israel exhibit heightened metapragmatic awareness and are aware of their code-switching behavior. Odlin suggests that linguistic forms that are important for communicative competence are, in general, highly salient and accessible to awareness, which may be why the metalanguage observed in anthropological linguistics tends to describe linguistic functions more accurately than linguistic form (Odlin, 1986). The fact that communicative behavior is sometimes accurately reportable is also compatible with the principle that accurate self-report depends on information that is attended to during performance. Pragmatic and discoursal knowledge is not always used automatically and unreflectively. Conversations vary a great deal in terms of spontaneity and planning (Ochs, 1979). Some people preplan telephone conversations, and writing involves a great deal of conscious deliberation and choices in discourse organization. There are many occasions on which particular care is given to producing appropriately polite language. Students may worry about how to address professors, and many aspects of the use of personal address are not unreflecting responses to a determining context but represent strategic and sometimes manipulative choices (Kendall, 1981). Pragmatic knowledge therefore seems to be partly conscious and partly accessible to consciousness, although it cannot be the case that all pragmatic knowledge is accessible to consciousness. Just as linguists seek to discover general principles of language that are reflected in the effortless control of grammar by native speakers but of which they have no conscious awareness, research in pragmatics seeks to identify patterns and general principles that native speakers are equally unable to articulate based on introspection. However, even if a great deal of pragmatic knowledge is held implicitly and cannot be articulated, this does not tell us how such knowledge was established. Skillful performance that currently relies on automatic processing and makes little demand on either attention or consciousness may have originated from conscious declarative knowledge (Lewis & Anderson, 1985). General principles, patterns, and rules of pragmatics may be beyond the reach of introspection, but this does not inform us of the possible role that awareness of crucial features of language rules, however incomplete and transitory, may play in the establishment of such knowledge (Munsell & Carr, 1981).

Consciousness and Principles of Language Learning Our ordinary language use of words like conscious, consciousness and consciously is ambiguous. This is one reason why theorists in psychology and applied linguistics have preferred to use related technical terms such as explicit versus implicit knowledge (Bialystok, 1979, 1981;Krashen, 1981; Odlin, 1986; Sharwood Smith, 1981), controlled versus automatic processing (Bialystok, Chapter 2; Bialystok & Bouchard-Ryan, 1985; Carroll, 1981; McLaughlin, Rossman & McLeod, 1983; Posner & Klein, 1973; Shiffrin & Schneider, 1977); declarative versus procedural knowledge (Anderson, 1982; Ellis, 1989b; Faerch & Kasper, 1984; O'Malley, Chamot & Walker, 1987), serial versus parallel processing (McClelland, Rumelhart, & the POP Research Group, 1986), and so on. Unfortunately, the use of

24

Cognitive Approaches to Interlanguage Pragmatic Development

technical terms does not by itself eliminate the ambiguities. Odlin (1986) has discussed the various ways in which the contrast between explicit and implicit knowledge has been understood, and Norman and Shallice (1986) have identified ambiguities inherent in the concept of automatic processing, some of which are exact parallels to the ambiguities of consciousness. Since a great deal of debate about conscious and unconscious processes has been fueled by conceptual and definitional disagreements (Bowers, 1984; White, 1980), it is preferable to grapple with these issues directly, rather than masking them with alternative terms. It seems to me that when we speak of having been conscious of something, we most often mean that we were aware of it, that we subjectively experienced it as part of the "stream" of consciousness (Battista, 1978; James, 1890; Natsoulis, 1987). However, when we speak of having done something consciously, we may mean either that we did it with awareness of what we were doing or that we did it deliberately. This is one of the main ambiguities involved in most discussions of consciousness: consciousness as awareness versus consciousness as intent (Ceci & Howe, 1982). When we speak of consciousness as awareness, there is also a question of the degree or level of our awareness. We may mean that we simply noticed the occurrence of something or that we had a more abstract understanding of it (Bowers, 1984). Therefore, when we speak of language learning as being conscious or unconscious, we might be thinking of several distinct aspects of the problem of consciousness in learning, including at least the following: whether a learner is trying to learn something; whether the learner is aware that he or she is learning; whether the target language forms that are learned are consciously noticed or picked up through some kind of subliminal perception; whether learners acquire general rules or principles on the basis of conscious understanding and insight or more intuitively; or whether learners are able to give an accurate account of the rules and principles that seem to underlie the construction of utterances. There is experimentally based literature from psychology that bears on all of these issues, along with a small amount of evidence from second language acquisition studies. It is useful to summarize the relevant research in terms of three principal distinctions. Conscious Perception versus Subliminal Influences in Learning My personal choice of a label for the key concept here is noticing, although there are a variety of technical terms for this, including focal awareness (Atkinson & Shiffrin, 1968), episodic awareness (Allport, 1979), conscious perception (Dixon, 1971) and apperceived input (Gass, 1988). Each of these constructs presupposes the allocation of attentional resources to some stimulus and identifies the level at which perceived events are subjectively experienced and are reportable by the person who experiences them.2 Events may remain unnoticed for several reasons—because attention is directed elsewhere, because the information is too complex to be processed, or because it is presented too quickly or too softly to be consciously seen or heard. While it is virtually impossible when observing naturalistic language learning to know exactly

Consciousness, Learning, and Intel-language Pragmatics

25

what the learner has or has not noticed, the existence of unnoticed information can be established under experimental conditions by the failure of subjects to report their awareness of a stimulus if asked immediately following its presentation. This criterion of subjective awareness can be contrasted with an objective measure of perception, which various experimenters have argued is best established by a subject's ability to discriminate among two or more alternative stimuli in a forced choice task (Cheesman & Merikle, 1986; Eriksen, 1960; Moore, 1988). Although many theorists believe that unconscious learning (in some sense) predominates in second language learning, it is very unlikely that what language learners consciously perceive or notice in input is unimportant for learning. A more difficult question is whether it is necessary to notice what is said in a language in order for that information to be stored in memory and to play a role in language learning, or whether it is also possible for some learning to be based on unnoticed information, information that is perceived at some level and perhaps processed subliminally without being consciously registered. There is a widespread belief (at least in North America) that the existence of subliminal learning of some kind has been established for decades. In the 1950s, Packard objected to the covert manipulation of consumers through the use of subliminal messages in advertising (Packard 1957), a theme expanded upon by Key (1973). Beginning in the 1980s, subliminal audiocassettes were aggressively marketed that promised everything from cures for obesity and drug addiction to enhanced visual acuity, improvement in examination performance, and more effective language learning. However, there seems to be virtually no scientific support for claims of behavior modification through subliminal messages. Moore has reviewed the research on subliminal techniques in advertising, concluding that the advertising stories everyone has heard about (such as the stimulation of movie theater patrons to buy popcorn or softdrinks through subliminal messages) are apocryphal. Such techniques probably never were used, and even if they were, "there is no evidence that subliminal messages can influence motivation or complex behavior" (Moore, 1988, 293). Merikle has examined commercially distributed "subliminal" audiotapes and subjected them to both psychophysical experimentation and spectographic analysis, reporting that the cassettes analyzed contained no embedded subliminal messages whatsoever that could conceivably influence behavior (Merikle, 1988, 355). There is a well-attested phenomenon of subliminal perception. Stimuli that are presented too rapidly for conscious detection or in competition with tasks that are assumed to consume all attentional resources may activate existing memory structures and associations (Dixon, 1971, 1981; Marcel, 1983). Eich (1984) has reported experiments in which pairs of words were both presented to the unattended channel in a shadowing task, one of which was ambiguous (e.g., fair or fare), while the other word biased its less common interpretation (e.g., taxi). Recognition of both members of such pairs was poor but in a spelling test subjects were biased in the direction of the disambiguated meaning. These and other similar demonstrations show that words that are not consciously perceived or noticed can be processed to the level of word meaning. However, all demonstrations of subliminal perception so far have involved subtle effects resulting from the unconscious detection and pro-

26

Cognitive Approaches to Interlanguage Pragmatic Development

cessing of very familiar stimuli. Such effects do not imply the creation of new memory structures, the establishment of new associations, or the learning of new concepts (Ericsson & Simon, 1984; Underwood, 1976, 1982), and certainly nothing remotely analogous to learning a second language. At the present time, the available evidence is compatible with the strong assertion that there is no such thing as subliminal language learning or any other kind of subliminal learning. Second language forms that are not noticed do not affect learning. This allows the concept of intake in second language learning to be defined in terms of what the learner attends to and notices (Schmidt, 1990).

Explicit versus Implicit Learning The contrast between subliminal learning and implicit learning, or learning without understanding, has to do with the level of awareness involved. I use noticing to mean registering the simple occurrence of some event, whereas understanding implies recognition of a general principle, rule, or pattern. For example, a second language learner might simply notice that a native speaker used a particular form of address on a particular occasion, or at a deeper level the learner might understand the significance of such a form, realizing that the form used was appropriate because of status differences between speaker and hearer. Noticing is crucially related to the question of what linguistic material is stored in memory (Atkinson & Shiffrin 1968; Kihlstrom, 1984); understanding relates to questions concerning how that material is organized into a linguistic system. Implicit learning refers to nonconscious generalization from examples. The general phenomenon of implicit learning has been well established in the psychological literature and is viewed as a natural product of attending to structured input (Hartman, Knopman, & Nissen, 1989; Reber, 1989). There is a gathering consensus within psychology that the mechanisms of implicit learning probably involve the strengthening and weakening of connections between nodes in complex networks as the result of experience, rather than through the unconscious induction of rules abstracted from data. An example of this recent shift in perspective can be seen in the work of Reber, who has carried out numerous experiments involving exposing subjects to strings of letters generated by an artificial grammar. After training, subjects were able to make accurate judgments about the well-formedness of novel strings, without being able to articulate the rules of well-formedness (Reber, 1976; Reber, Allen & Regan, 1985; Reber, Kassin, Lewis & Cantor, 1980). Until recently, Reber (1976) argued that knowledge resulting from implicit learning was encoded in the form of unconscious abstract representations. In a more recent publication, Abrams and Reber (1988) have suggested that implicit learning as demonstrated in these experiments probably rests upon some kind of covariation counter, a system that logs both event frequencies and event co-occurrences. One model that simulates the mechanisms currently believed to underlie implicit learning is Parallel Distributed Processing (PDP). PDP has been used to model the acquisition of the German definite article (MacWhinney et al., 1989), the past tense in English (Rumelhart & McClelland, 1986), the development of visual word recognition skills (Seidenberg & McClelland, 1989), and the acquisition of gender in French (Sokolik & Smith, 1989).

Consciousness, Learning, and Interlanguage Pragmatics

27

Explicit learning, that is, conscious problem solving, relies on different mechanisms, including attempts to form mental representations, searching memory for related knowledge and forming and testing hypotheses (Mathews, Buss, Stanley, Blanchard-Field, Cho & Druhan, 1989; Johnson-Laird, 1983). Both implicit learning and explicit learning have particular strengths. Implicit learning appears to be superior for the learning of fuzzy patterns based on perceptual similarities and the detection of nonsalient covariance between variables, while explicit learning is superior when a domain contains rules that are based on logical relationships rather than perceptual similarities (Mathews et al., 1989).

Intentional versus Incidental Learning Whereas the concepts of subliminal and implicit learning are both related to the consciousness as awareness, incidental learning refers to consciousness as intent. If, as I have claimed, it is necessary to notice the occurrence of linguistic forms in order for them to serve as intake for learning, is it also necessary to deliberately pay attention to such features in order to notice them? More generally, is it necessary to want to learn in order to learn? This is not so difficult a question as the others I have raised. In many cases, it does not matter if a language learner intends to pay attention or not. A language learner's limited processing abilities may make it impossible to notice something regardless of an intent to do so. There are other cases in which some task to be performed forces the learner's attention to be focused on some pieces of information rather than others, and in such cases, what is stored in memory is the information that must be attended to in order to complete the task (Ericsson & Simon, 1984); the learner's intention to learn is irrelevant (Anderson, 1985). On the other hand, there are many situations in which a language learner is free to opt in and out of learning contexts and to pay attention or not, depending on one's personal hierarchy of deep goals and momentary dispositions (Baars, 1988; Kahneman, 1973; Kihlstrom, 1984); in such cases paying attention is crucial.

Extensions to the Learning of Pragmatics and Discoursal Rules I have argued that linguistic forms can serve as intake for language learning only if they are noticed by learners; that paying attention to such forms is certainly helpful, but not necessary if other factors in the learning context focus attention on them so that they are noticed; and that general principles of the organization of language may be discovered through the use of either explicit or implicit learning mechanisms. I have also suggested that even in cases where what native speakers "know" about the pragmatic principles of their language is inaccessible to consciousness, such knowledge may nevertheless be based on insights and understanding at the time of learning. What evidence is there that these claims are relevant for the learning of pragmatics?

28

Cognitive Approaches to Interlanguage Pragmatic Development

First Language Learning of Pragmatics Research on the acquisition of first language pragmatics suggests that both noticing and some level of understanding are important in such learning. Clark (1978) has observed that the types of metalinguistic abilities shown by preschool children are primarily related to communicative interaction rather than grammatical form. The ethnographic literature on language socialization shows that an important childrearing goal is to develop the child's communicative competence. Demuth (1986) has reported on the prompting routines for appropriate verbal behavior that play an active role in the social development of Basotho children. Ochs (1986) has described the ways in which Samoan caregivers use prosodic strategies for teaching children how to encode affect-laden utterances. Clancy (1986) has shown how Japanese mothers interweave questions and declarative hints to socialize children in the use of indirectness. Watson-Gegeo and Gegeo (1986) describes how Kawara'ae caregivers use repeating routines to teach children what to say and when to say it. These and many similar reports suggest that while parents and other caregivers use different socialization strategies in different cultures, there is probably universal validity to the observation of Gleason and Perlmann: Unlike the acquisition of syntax, semantics, and even some sociolinguistic rules, when it conies to speaking politely adults do not leave it to the child to construct the rules on his or her own. Here, they take an active, even energetic part in directly instructing their children in the use of the various politeness devices. (Gleason & Perlmann, 1985, 102)

Snow, Perlmann, Gleason and Hooshyar (1990) have examined parent-child interactions in 110 families in order to see what kinds of information concerning politeness strategies are made available to children from their interaction with parents. Assuming that the basic dimensions of power, social distance, and degree of imposition underlie the general rule system for politeness, Snow et al. looked for evidence for three types of information that might be made available to children: direct teaching of general rules of politeness, manipulation of the dimensions of politeness so that the relevant covariations were made more salient, and information about the use of specific forms. Snow et al. found that the first type of information was rare, but that there was plentiful evidence in their data that children were explicitly told what forms to use in particular situations, and that correlations between forms and the dimensions of politeness were made salient in interaction. These findings suggest (though they do not prove this point) that children are not only exposed to but also notice surface forms. Children are also presented with information that could be used to induce more general principles (through either implicit or explicit learning mechanisms), but are not taught the underlying principles directly.

Second Language Learning of Pragmatics Since adults can report their understandings much more readily than children, it ought to be possible to examine the role of noticing and understanding in the development of pragmatic ability by adult second language learners directly, by

Consciousness, Learning, and Interlanguage Pragmatics

29

asking learners to report their experiences. Even so, the relevant data are difficult to obtain, requiring both a sound methodology for eliciting self-reports (Faerch & Kasper, 1987) and opportunities to catch learners in the actual process of learning, rather than simply performing their current competence. Unfortunately, there have been few studies of any aspect of the phenomenology of second language learning and no studies at all which have attempted systematically to ascertain what learners have been conscious of as pragmatic principles were learned. Anecdotally, there is evidence for a relationship between what learners notice and understand about pragmatics and discourse and what is learned. The following six examples are from my own experience, either as a language learner or from interacting in English with second language speakers with different linguistic and cultural backgrounds; they concern interactions about which I wrote brief notes to myself shortly after they happened. The first four examples represent the coincidence of recognition and insight with rapid learning; the last two represent instances of less successful learning. (1) In the course of a 22-week stay in Brazil, during which I progressed from no proficiency at all in Portuguese to the S-2 level on the FSI scale (see Schmidt & Frota, 1986, for details), I kept a language learner's diary. Several entries illustrate the phenomenon of being told about some aspect of the pragmatics of Brazilian Portuguese in class and then almost immediately noticing it in input, such as the following: Journal entry, Week 6. This week we were introduced to and drilled on the imperfect. . . . The basic contrast seems straightforward enough: ontem eufui ao clube ["yesterday 1 went to the club"] vs. antigamente eu ia ao clube ["formerly I used to go to the club"]. L gave us a third model: ontem eu ia ao clube ["yesterday I was going to the club"], which L says is a common way of making excuses. . . . Wednesday night Amos came over to play cards, and the first thing he said was eu ia telefonar para voce ["I was going to call you"], exactly the kind of excuse L had said we could expect.

(2) I noted in my diary several times the difficulties I had with telephone conversations, especially in knowing when and how to end a conversation (Schmidt & Frota, 1986, 276). I knew that with friends the closing move would be for both parties to say ciao, but I could never identify the point at which I could say it, so I would often stand holding the phone waiting patiently for the other person to say it first. Finally, during the last week of my stay, a friend came to my apartment and used my telephone to make several calls. I listened carefully, and noticed that in two successive calls, shortly before saying ciao, my friend said the phrase entdo td, which means no more than "so, then." Suspecting that this might be a preclosing formula, I immediately called another friend and after a few minutes of talk, said entdo td, paused briefly and plunged ahead with ciao in the same turn. It worked, and after that I had no trouble at all getting off the phone efficiently. I subsequently asked several native speakers how to close a telephone conversation. None could tell me, but when 1 suggested the use of entdo td, they agreed that was right. (3) Midway through my stay in Brazil, I took a trip to another city for several days, and later wanted to send postcards to people I had met there. I wrote a few cards, and then asked a native speaker to rewrite one for me. I noticed that he began

30

Cognitive Approaches to Interlanguage Pragmatic Development

rephrasing my message with the expression E ai, como estdo? ("So, how are you?"), so I did the same with each subsequent card. A week after sending the cards, I got a call from one of the recipients (a native speaker of English who was a long-term resident in Brazil) who began the conversation by commenting that my Portuguese must be improving rapidly, given the colloquially appropriate style of my postcard. (4) On the first day of a 2-week trip to Thailand, I presented a paper at the end of the day at a national conference. After the lecture, several Thais with whom I would be working for the following week approached me and made some brief remarks in English (I know no Thai) and then slipped away. I found myself standing by myself much quicker than I expected, and had the unsettling feeling that my talk must have been very poorly received. I returned to my hotel feeling quite depressed about this. That evening, I looked over some materials that I had collected during the day, including an article by Sukwiwat and Fieg (1987) on greeting and leave-taking in Thai. Sukwiwat and Fieg pointed out that conversations are closed quickly in Thai but tend to be drawn gradually to a close in English, so that Americans are often taken aback by what appear to be abrupt, brusque, and sometimes rude departures. Thais, on the other hand, think that American leave-takings drag on excessively and involve unnecessary verbiage. I immediately realized that I might have misinterpreted the significance of what had happened earlier. For the remainder of my stay, I tried my best to beat the Thais at their own game by closing conversations faster than they could, for example, by suddenly announcing, "well, I'm leaving now." I never succeeded in getting away faster than they did, but my disquiet at this aspect of Thai behavior evaporated and I suffered no discomfort from behaving in a way that would be rude by my own cultural norms. (5) Between the early 1960s and mid-1970s, I lived mostly in Arabic-speaking countries and became fairly proficient in both Egyptian and Lebanese Arabic. In some varieties of Arabic, parents and other relatives may address children with what Ayoub (1964) has called bipolar kin terms, ego addressing alter with the term that in its literal sense would be appropriate for alter addressing ego; e.g., a grandfather may address his granddaughter with a term equivalent to "grandpa." This occurs when the senior wants the junior to do something, but chooses a conciliatory request form, metaphorically reversing the power relationship between the two—what Brown and Levinson (1987) would call a point-of-view operation. I knew of this phenomenon only from Ayoub's article, however. I never noticed parents using it with their children in either Egypt or Lebanon, although I often observed parents and other family members interacting with children. Years later, I noticed the use of such a form when visiting friends from Lebanon in California. Playing in the swimming pool, the mother said to her son, in English, "Okay, Baron, swim down the other end of the pool now, Mommy." I have since been assured by speakers of both Egyptian and Lebanese Arabic that they do use such forms, but second language speakers of Arabic whom I have asked have reported that, like me, they have never noticed it being used. (6) In several publications (Schmidt, 1983, 1984), I reported on a case study of a Japanese learner of English whose overall level of communicative competence was superior to his rather rudimentary control of English grammar. In looking at the development of pragmatic ability by my subject, Wes, I found that he often used

Consciousness, Learning, and Interlanguage Pragmatics

31

hints that native speakers of English, including myself, did not realize were intended as directives. For example, once in a theater, Wes turned to me and asked me if I liked my seat. I responded that my seat was fine, not realizing at all that he was indirectly requesting that we change places. After many years of interacting with Japanese speakers of English, I think that I now recognize such hints on most occasions, but this learning process has been slow. All of these anecdotes indicate an apparently very close connection between noticing what was present in input and learning. Each case of successful learning also involved more than just noticing the forms used, but also an appreciation of their functional meaning: that an imperfective signaled an excuse, that entao td was a preclosing device, that e af was useful for greetings, that an abrupt departure did not necessarily imply a problem. Two of my Portuguese examples also illustrate intentional rather than incidental learning—I was deliberately seeking speech routines for openings and closings, and discovered them. By contrast, in the case of abrupt Thai departures, the learning was incidental; I had no prior awareness of a learning problem or intent to learn anything. 1 noticed the behavior and may or may not have carried on some conscious inferencing in arriving at my conclusion that my poor lecture lay behind it (those thought processes are not recoverable), but my corrected understanding of the significance of such behavior was fortuitous and the information was externally provided. In the case of the Portuguese imperfect used for excuses, explicit information about pragmatic function seems to have made the input more salient, though it is virtually certain that such forms were in input all along. In contrast to the Thai example, externally provided information about Arabic bipolar kin terms had no effect on my learning. I never noticed their occurrence in the dialects to which I was exposed, and they never became part of my competence in Arabic. This example suggests some of the difficulty in accounting for what becomes conscious and what does not. This is a complex issue beyond the scope of this paper, but part of such an account would include Baars's observation that events remain unnoticed if they are either uninterpretable in context or so stable as to be part of the context (Baars, 1983). The Arabic use of bipolar kin terms seems to be especially opaque to native speakers of English, who find them nearly uninterpretable.3 In the case of both Thai departures and Wes's hints in English, the problem lay not in noticing what was said but in understanding what was intended. The interpretation of hints is problematic for native speakers as well as learners (Ervin-Tripp, 1972), but it is not clear to me why externally provided information was sufficient to block future inferences from abrupt departures to perception of a problem in the case of my Thai example, whereas knowledge about Japanese speech behavior at a similar level of generality did not lead quickly to the establishment of the appropriate inferencing behavior.

Explicit and Implicit Learning of General Principles While all of my examples involve understanding in the sense of matching surface forms with meaning, none of them are good examples of generalization from specific examples to more general principles. However, there are cases in which the

32

Cognitive Approaches to Interlanguage Pragmatic Development

learning of pragmatics and discourse must involve such generalization, for example, not just the recognition and use of frozen routines such as entao td and e ai, como estdo, but learning less frozen formulas, as well as fully productive structures for speech act realizations. A good example of the involvement of consciousness in generalizing a formula has been provided by Ferguson (1976) in recalling his learning of Arabic root-echo responses. There are numerous adjacency pairs in Arabic in which a greeting, compliment, or other initiating utterance requires a formulaic response that contains a lexical item (usually a verb) derived from the triconsonantal root of the most important lexical item in the initiating utterance. On one occasion, Ferguson bought an article of clothing in a market, and when the purchase was complete, the seller said to him "mabruuk" (congratulations). He did not know the response formula for this, but did know that an appropriate response form would be one which contained the root BRK from the first part of the adjacency pair. By analogy with several other response formulas that he did know, he guessed what the root-echo response form might be. Ferguson comments: Probably 'alia ybaarikfiik was the root-echo response to BRK. I tried it, and the smile showed I had given the right reply. The whole analysis took only a split second, and was just like getting an instance of grammatical concord or case government right. (141)

This is an example of conscious problem solving or explicit learning, but I have indicated that implicit language learning is also possible. It may be useful, therefore, to spell out in some detail how more general principles of pragmatics might be acquired without being conscious of them. Following Fox (1987) and Spolsky (1989), I suggest that some pragmatic and discoursal principles are better represented as associative networks rather than as prepositional rules, and that connectionist models are promising in accounting for those aspects of pragmatic knowledge that do appear to be unconscious. This may involve less of a paradigm shift in the areas of pragmatics and discourse than in syntax, since researchers in pragmatics have had a less fixed notion of what is meant by a "rule" of pragmatics than have syntacticians, and relatively little attention has been given to consideration of how such rules might be represented psychologically. Some ethnomethodologists have rejected the concept of rules as conceived in formal logic as a model of social action (Mehan & Wood, 1975). Probabilistic network approaches have been suggested for the analysis of code-switching (Dearholt & Valdes-Falles, 1978); and Pomerantz (1978) has described compliment responses as the result of the cooperation of multiple constraints. Each of these approaches is compatible with a connectionist interpretation. There is one type of representation of pragmatic rules that I think is psychologically implausible, but that can also be recast in network form. The distribution of address forms has been represented by both Geoghegan (1971) and Ervin-Tripp (1972) in the form of a flowchart, illustrating decision points in the form of serially ordered binary selectors. For American English, Ervin-Tripp indicates that the first question to be asked is whether or not the addressee is a child or adult; then, if an adult, whether the interaction takes place within a status-marked setting; if the setting is not status-marked, whether the addressee's name is known; if the name is

Consciousness, Learning, and Interlanguage Pragmatics

33

known, whether the addressee is kin; if not kin, whether the addressee is a friend or colleague; if so, whether the addressee is of higher rank; and so on, finally exiting the system with an appropriate address form. Ervin-Tripp (1972) and Geoghegan (1971) state explicitly that while paths through the flowchart represent rules, such flowcharts are like a formal grammar in representing a logical model and are not intended as psychological models of decision making. There are various problems with these models. Kendall (1981) has pointed out that they are too deterministic, and that a factor called "dispensation," meaning essentially to disregard all other factors, is introduced to get around the problem of variability. Positing serially ordered selectors also implies complete scalability (each selector must be listed only once, and selectors encountered first in the flowchart must outweigh all subsequent selectors in their influence), which cannot be empirically supported. For the present discussion, the most important drawback to such models is that they are unlikely to have any psychological reality. While conscious choices of which address forms in unclear situations might indeed involve sequential consideration of the types of selectors contained within flowcharts, there is little reason to suggest that automatic choices are made on the basis of speeded-up serial processing; most psychological accounts of automatic processing assume that parallel processes dominate. However, flowcharts such as those suggested by ErvinTripp (1972) and Geoghegan (1971) can easily be restructured into connectionist architecture, and the choice of address forms can be reconceptualized as a network of unordered connections between features of social context (addressee age, rank, marital status, etc.) and linguistic outputs. Some connections between social context features and address forms may be so heavily weighted that the connection is almost categorical, while others may be very weak, leading to fuzzier, less determinate outcomes. Other kinds of pragmatic knowledge that may be similarly represented include the complex patterns of covariation among features of social context and the linguistic realizations of speech acts that have been empirically documented by analysis of CCSARP data (Blum-Kulka, House & Kasper, 1989). The theoretical framework proposed by Brown and Levinson (1987), which attempts to relate a very wide range of pragmatic realizations to variation in three basic contextual features—social distance, power, and culture-specific evaluations of threat to face—is a similar case since these cooperating (or conflicting) constraints exert probabilistic influences. If we assume that associative network models have some face validity as a model of implicit pragmatic knowledge, we may then ask how such knowledge may be acquired, and specifically the role that consciousness is likely to play in the establishment of a network. There is some evidence from experimental psychology that bears on several aspects of this question. (1) Do learners have to keep track (consciously, by counting) of the frequency with which contextual or pragmatic features occur? The answer to this is almost certainly no. Learners may be able to make reasonably accurate estimates of the relative frequency of such things, but they do not do so by counting, and it is widely accepted that attention to a stimulus event is sufficient to trigger the automatic (effortless and unintended) encoding of its frequency of occurrence (Hasher & Zacks, 1984). (2) Do learners need to notice the specific relevant pragmalinguistic or contex-

34

Cognitive Approaches to Interlanguage Pragmatic Development

tual features of an event in order to trigger such encoding of frequency? This question is somewhat controversial, but the answer is probably yes. Hanson and Hirst (1988) point out that an event may be thought of as a cluster of attributes. They report experiments supporting the hypothesis that attention to specific stimulus attributes is necessary in order to encode frequency information for those attributes. (3) Do learners need to understand the significance of co-occurring linguistic and social context features in order to acquire a network of complex covariations? This is perhaps the most interesting question, and strikes to the heart of what is meant by implicit learning. Experiments in implicit learning suggest that implicit learning may be self-organizing, and that it is not necessary to realize the significance of one event for another in order to establish connections. Lewicki (1986) has reported a series of experiments designed to demonstrate the nonconscious detection of covariations involving social stimulus material. Subjects were presented with a series of descriptions of persons which mentioned a number of psychological and social characteristics. Some of these traits were manipulated by the experimenter, either to confirm or disconfirm preexisting stereotypes. After a learning phase, subjects rated new stimulus material. The experiments showed that correlations built into the personality descriptions during the learning phase influenced judgments in the testing phase. By running different versions of the same basic experiment, Lewicki was able to assess subject awareness at various points in learning. Subjects did notice the manipulated traits (as intended by the experimenter) and were momentarily aware of their co-occurrence in single stimulus descriptions. That is, they were able to recall both of them when questioned immediately after exposure.4 However, the subjects were unaware of any systematic relationship between the manipulated traits. When told that some traits (out of a large number used as descriptors) had been systematically manipulated by the experimenter, subjects were unable to identify which ones had been manipulated. Analogously, we can specify the minimum requirements of learning an address system (or any other system of complex covariations) in a second language. If the task is to acquire an address system in which the ingroup/outgroup distinction is relevant or in which address forms systematically vary by sex of addressee, learners must attend to and notice in input both the linguistic forms and the relevant contextual features. This may mean attending to features of context that either are not relevant or are defined differently in the native language, so that learning a new pragmatic system often entails learning how to make new interpretative assessments of the world. However, it does not seem to be necessary for learners to make any conscious connection between the address forms encountered and the contextual factors that are correlated with such forms. For example, in learning the address system of Japanese, when you hear someone you know as Mr. Morita addressed as Morita-kun, where kun is an address form, you must notice both the form and the relevant contextual factors (these include sex, age, and rank of both speaker and addressee, intimacy, tone, and setting) if this is to be intake for learning, but need not draw the conclusion that Morita was addressed that way because of any of these factors. Nevertheless, it would certainly be extremely helpful to be consciously aware of such connections. It is sometimes argued that implicit learning is superior to conscious problem solving (Krashen, 1981), but this seems to be true for only some

Consciousness, Learning, and Intel-language Pragmatics

35

types of learning tasks. Reber has reported several times that subjects learning artificial grammars under an implicit learning condition (subjects were told to memorize examples, which presumably interfered with any attempt to analyze the input) were better able to recognize valid new strings generated by the grammar than those subjects who were told to try to figure out the rules of the underlying grammar. However it cannot be assumed that subjects who attempted to discover the rules succeeded in doing so. In a recent publication, Reber (1989) makes exactly this point. Arguing that the particular artificial grammar to be learned was constructed in such a way that subjects were unlikely to be able to find the rules they were searching for, Reber now argues that "looking for rules will not work if you cannot find them," but "looking for rules will work if you can find them" (Reber, 1989, 223). McLeod and McLaughlin (1986) report anecdotal evidence for the frequent occurrence of rapid restructuring following "clicks of comprehension."

Conclusions The data from experimental psychology clearly support a conservative hypothesis that whatever learning might result from unattended processing is insignificant compared to the results of attended processing. The data seem to me to be also compatible with two much stronger hypotheses, that attention to input is a necessary condition for any learning at all, and that what must be attended to is not input in general, but whatever features of the input play a role in the system to be learned. For the learning of pragmatics in a second language, attention to linguistic forms, functional meanings, and the relevant contextual features is required. I also claim that learners experience their learning, that attention is subjectively experienced as noticing, and that the attentional threshold for noticing is the same as the threshold for learning. Finally, I argue that, while incidental and implicit learning are both possible, consciously paying attention to the relevant features of input and attempting to analyze their significance in terms of deeper generalizations are both highly facilitative. I do not claim that the anecdotal examples from my own language learning experiences prove these points, because the most that language learner diary reports can establish is that learners have noticed crucial facts about language use. What is needed is much more systematically gathered data on what learners notice (and are able to report) and what they do not notice (are unable to report) as they are learning. 5 Suspicions have been voiced that "it is doubtful that [introspection] can shed light on how the learner moves from one state to another, i.e., how input becomes intake" (Ellis, 1989a). I think that investigation of the learner's thoughts at such points of change is just what needs to be investigated. Even the harshest critics of reliance upon introspective methods agree that individuals do know the focus of their attention at any given point in time, as well as the content of their current thoughts, emotions, evaluations, and plans (Nisbett & Wilson, 1977) and that these conscious thought processes can be reported. A priori conclusions that there will be no relationships between such phenomena and language development are unwarranted. No strong prescriptions for the teaching of second language pragmatics can be

36

Cognitive Approaches to Interlanguage Pragmatic Development

drawn from this discussion, but some general observations seem in order. Simple exposure to sociolinguistically appropriate input is unlikely to be sufficient for second language acquisition of pragmatic and discoursal knowledge because the linguistic realizations of pragmatic functions are sometimes opaque to language learners and because the relevant contextual factors to be noticed are likely to be defined differently or may be nonsalient for the learner. Second language learners may fail to experience the crucial noticings for years. The fact that this does not seem to happen in first language learning is attributable not to any sort of pragmatics acquisition device, but to the efforts that parents and other caregivers make in order to teach communicative competence to children, using a variety of strategies. Motivation is an important determinant of the allocation of attentional resources (Crookes & Schmidt, 1991). Because of the close connections among pragmatic realization strategies, assessments of role and status relationships between speaker and hearer, and the expression of personality, it is likely that there is a stronger relationship between motivation, acculturation and other affective factors in the development of pragmatic and discoursal ability than in other aspects of language learning, such as syntax (Schmidt, 1983). Those who are concerned with establishing relationships with target language speakers are more likely to pay close attention to the pragmatic aspects of input and to struggle to understand than those who are not so motivated. But since intentional learning is unnecessary when some task causes attention to be focused on what is to be learned, one way to develop pragmatic competence in classroom contexts could be through task-based language teaching (Long, in press). Tasks can be selected that focus the learner's attention on pragmatic forms, functions and co-occurring features of social context. Explicit teacher-provided information about the pragmatics of the second language can also play a role in learning, provided that it is accurate and not based solely on fallible native speaker intuitions. Explicit teaching is often more efficient than attention to input for identifying the pragmalinguistic forms of the target language. The understanding of general rules and patterns may be unnecessary for learning, but Grossberg (1988) has argued that the learning mechanisms modeled in connectionist networks are slow because they result only from gradual changes in the bottom-up adaptive weights of the network, whereas top-down processes such as focused attention and expectations greatly speed up and actively reorganize the way in which input is processed. This is not to claim that explicit knowledge somehow "becomes" implicit knowledge, but to recognize a synergistic relationship between the mechanisms of implicit and explicit learning (Mathews et al. 1989), which justifies a consciousness-raising approach to the teaching of pragmatics.

Notes 1. I am grateful to Michael Long, Paul Munsell, and Danny Steinberg for very helpful comments on an earlier draft of this paper. 2. I owe the distinction made here between information that is perceived and information that is noticed to Bowers (1984), who argues that information becomes conscious when it is processed to the level of short-term memory and selectively attended to. Bowers also distin-

Consciousness, Learning, and Interlanguage Pragmatics

37

guishes between two senses of unconscious, referring to information that is unnoticed and information that is unappreciated or uncomprehended, and I have drawn upon his model in my description of subliminal versus implicit learning. 3. When I have presented this example to native speakers of English, they have often assumed that the mother must have meant for her son to "swim to Mommy," but this is not a correct interpretation; that is, this is not an example of a missing preposition. 4. In another series of experiments, Lewicki (1986) attempted to demonstrate that information which is presented subliminally or which is not attended to may also lead to learning. These experiments did not successfully demonstrate learning, but some interesting subtle effects were found. Subjects responded more slowly to questions mentioning those stimulus traits that had been presented subliminally. Lewicki argues that this demonstrates the internalization of weak processing algorithms, which could eventually result in more demonstrable learning effects. Such an experimental demonstration would disprove my claim that there is no subliminal learning whatsoever. Baars (1988) has claimed that this zero-point question is essentially unanswerable and has obscured the more important and answerable question of whether more conscious involvement is needed to learn more information, the answer to which is clearly affirmative. 5. Michael Long (personal communication) has pointed out to me that language learners may sometimes produce a vocabulary item in a second language that they did not know they knew until that moment, not being sure that it is right and certainly not knowing how it ever got into the mental lexicon. However, the issue of whether the learner noticed such a lexical item in input (which must have occurred, if my account is correct) is quite separate from the question of whether the learner will be able to say much later when it was encountered. We know all sorts of things without being able to recall the circumstances under which we acquired that knowledge.

References Abrams, M., & Reber, A. (1988). Implicit learning: Robustness in the face of psychiatric disorders. Journal of Psycholinguistic Research, 17(5), 425-39. Allport, D. (1979). Conscious and unconscious cognition: A computational metaphor for the mechanism of attention and integration. In L. Nilsson (Ed.), Perspectives on memory research. Hillsdale, NJ: Lawrence Erlbaum, pp. 61-89. Anderson, J. (1982). Acquisition of cognitive skill. Psychological Review, 89, 369-406. Anderson, J. (1985). Cognitive psychology and its implications (2nd ed.). New York: W. H. Freeman. Atkinson, R., & Shiffrin, R. (1968). Human memory: A proposed system and its control processes. In K. Spense (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. 2). New York: Academic Press, pp. 89-195. Ayoub, M. (1964). Bi-polarity in Arabic kinship terms. In H. Lunt (Ed.), Proceedings of the ninth international congress of linguistics. The Hague: Mouton, pp. 1100-1106. Baars, B. (1983). Conscious contents provide the nervous system with coherent, global information. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.), Consciousness and self-regulation: Advances in research and theory (Vol. 3). New York: Plenum Press, pp. 41-79. Baars, B. (1988). A cognitive theory of consciousness. New York: Cambridge University Press. Battista, J. (1978). The science of consciousness. In K. Pope & .1. Singer (Eds.), The stream

38

Cognitive Approaches to Interlanguage Pragmatic Development

of consciousness: Scientific investigations into the flow of human experience. New York: Plenum Press, pp. 57-87. Bialystok, E. (1979). Explicit and implicit judgements of L2 grammaticality. Language Learning, 29, 81-103. Bialystok, E. (1981). The role of linguistic knowledge in second language use. Studies in Second Language Acquisition, 4, 31-45. Bialystok, E., & Bouchard-Ryan, E. (1985). A metacognitive framework for the development of first and second language skills. In D. Forest-Prcssley, G. MacKinnon, & T. Waller (Eds.), Metacognition, cognition, and human performance. Orlando, FL: Academic Press, pp. 207-52. Blom, J., & Gumperz, J. (1972). Social meaning in linguistic structures: Code-switching in Norway. In J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics (407-34). New York: Holt, Rinehart & Winston, pp. 407-34. Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Bowers, K. (1984). On being uncousciously influenced and informed. In K. Bowers & D. Meichenbaum (Eds.), The unconscious reconsidered. New York: Wiley, pp. 227-72. Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Cameron, D. (1985). Feminism and linguistic theory. London: Macmillan. Carroll, J. (1981). Conscious and automatic processes in language learning. Canadian Modern Language Review, 37, 462-74. Cathcart, R. L. (1989). Authentic discourse and the survival English curriculum. TESOL Quarterly, 23(1), 105-26. Ceci, S., & Howe, M. (1982). Metamemory and the effects of intending, attending and intending to attend. In G. Underwood (Ed.), Aspects of consciousness (Vol. 3): Awareness and self-awareness. London: Academic Press, pp. 147-64. Cheesman, J., & Merikle, P. M. (1986). Distinguishing conscious from unconscious perception. Canadian Journal of Psychology, 40, 343-67. Clancy, P. (1986). The acquisition of communicative style in Japanese. In B. Scheiffelin & E. Ochs (Eds.), Language socialization across cultures. Cambridge: Cambridge University Press, pp. 213-49. Clark, E. (1978). Awareness of language: Some evidence from what children say and do. In A. Sinclair, R. Jarvella, & W. Levelt (Eds.), The child's conception of language. Berlin: Springer-Verlag, pp. 17-43. Crookes, G., & Schmidt, R. (1991). Motivation: Reopening the research agenda. Language Learning, 41(4), 469-512. Dearholt, D., & Valdes-Fallis, G. (1978). Toward a probabilistic automata model of some aspects of code-switching. Language in Society, 7(3), 411-19. Demuth, K. (1986). Prompting routines in the language socialization of Basotho children. In B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures. Cambridge: Cambridge University Press, pp. 51-79. Dixon, N. (1971). Subliminal perception: The nature of a controversy. London: McGrawHill. Dixon, N. (1981). Preconscious processing. New York: Wiley. Eich, E. (1984). Memory for unattended events: Remembering with and without awareness. Memory and Cognition, 12, 105-11. Ellis, R. (1989a). Review of C. Facrch&G. Kasper (Eds.), Introspection in second language research. Applied Linguistics, 10(3), 360-64.

Consciousness, Learning, and Interlanguage Pragmatics

39

Ellis, R. (1989b). The role of practice in classroom learning. Revue de L'AILAIAILA Review, 5, 20-39. Ericsson, K., & Simon, H. (1984). Protocol analysis: Verbal reports as data. Cambridge, MA: MIT Press. Eriksen, C. W. (1960). Discrimination and learning without awareness: A methodological survey and evaluation. Psychological Review, 67, 279-300. Ervin-Tripp, S. (1972). On sociolinguistic rules: Alternation and co-occurrence. In J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics. New York: Holt, Rinehart & Winston (pp. 213-50). Faerch, C., & Kasper, G. (1984). Pragmatic knowledge: Rules and procedures. Applied Linguistics, 5(3), 214-25. Faerch, C., & Kasper, G. (Eds.). (1987). Introspection in second language research. Clevedon and Philadelphia: Multilingual Matters. Faerch, C., & Kasper, G. (1989). Interlanguage request modification. In S. Blum-Kulka, J. House & G. Kasper (Eds.), Cross-cultural pragmatics. Norwood, NJ: Ablex, pp. 221-47. Ferguson, C. (1976). The structure and use of politeness formulas. Language in Society, 5, 137-51. Fox, B. (1987). Variation and patterning in discourse and grammar. In K. M. Denning, S. Inkeles, F. C. McNair-Knox, & J. R. Rickford (Eds.), Variation in language: NWAVE-XV at Stanford. Stanford: Stanford University, Department of Linguistics. Gass, S. (1988). Integrating research areas: A framework for second language studies. Applied Linguistics, 9, 198-217. Geoghegan, W. (1971). Information processing systems in culture. In P. Kay (Ed.), Explorations in mathematical anthropology. Cambridge, MA: MIT Press, pp. 3-35. Gleason, J., & Perlmann, R. (1985). Acquiring social variation in speech. In H. Giles & R. St. Clair (Eds.), Recent advances in language, communication, and social psychology. London: Lawrence Erlbaum. Grossberg, S. (1988). Nonlinear neural networks: Principles, mechanisms, and architectures. Neural Networks, I , 17-61. Hanson, C., & Hirst, W. (1988). Frequency encoding of token and type information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14(2), 289-97. Hartman, M., Knopman, D. S., & Nissen, M. J. (1989). Implicit learning of new verbal associations. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(6), 1070-82. Hasher, L., &Zacks, R. (1984). Automatic processing of fundamental information: The case of frequency of occurrence. American Psychologist, 39, 1372-88. Holmes, J. (1988). Doubt and uncertainty in ESL textbooks. Applied Linguistics, 9(1), 2144. James, W. (1890). The principles of psychology. New York: Holt. Johnson-Laird, P. (1983). Mental models. Cambridge, MA: Harvard University Press. Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall. Kasper, G. (1984). Pragmatic comprehension in learner-native speaker discourse. Language Learning, 34(4), 1-20. Kendall, M. (1981). Toward a semantic approach to terms of address: A critique of deterministic models in sociolinguistics. Language & Communication 1(2/3), 237—54. Key, W. B. (1973). Subliminal seduction. Englewood Cliffs, NJ: Signet. Kihlstrom, J. (1984). Conscious, subconscious, unconscious: A cognitive perspective. In K. Bowers & D. Meichcnbaum (Eds.), The unconscious reconsidered. New York: Wiley, pp. 149-211.

40

Cognitive Approaches to Intel-language Pragmatic Development

Krashcn, S. (1981). Second language acquisition and second language learning. Oxford: Pergamon. Krashen, S. (1983). Newmark's "ignorance hypothesis" and current second language acquisition theory. In S. Gass & L. Selinker (Eds.), Language transfer in language learning. Rowley MA: Newbury House, pp. 135-53. Lakoff, R. (1973). Language and woman's place. Language in Society, 2(1), 45-80. Lewicki, P. (1986). Nonconscious social information processing. New York: Academic Press. Lewis, M., & Anderson, J. (1985). Discrimination of operator schemata in problem solving: Learning from examples. Cognitive Psychology, 17, 26-65. Long, M. H. (in press). Task-based language teaching. Oxford: Basil Blackwell. MacWhinncy, B., Leinbach, J., Taraban, R., & MacDonald, J. (1989). Language learning: Cues or rules? Journal of Memory and Language, 28, 255-77. McClelland, J., Rumelhart, D., & the POP Research Group. (1986). Parallel distributed processing: Explorations in the microstructure of cognition (Vol. 2): Psychological and biological models. Cambridge: MA: MIT Press. McLaughlin, B., Rossman, T., & McLeod, B. (1983). Second language learning: An information-processing perspective. Language Learning, 33, 135-57. McLeod, B., & McLaughlin, B. (1986). Restructuring or automaticity? Reading in a second language. Language Learning, 36(2), 109-23. Marcel, A. (1983). Conscious and unconscious perception: An approach to the relations between phenomenal experience and cognitive processes. Cognitive Psychology, 15, 238-300. Mathews, R. C., Buss, R. R., Stanley, W. B., Blanchard-Fields, F., Cho, J. R., & Druhan, B. (1989). Role of implicit and explicit processes in learning from examples: A synergistic effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(6), 1083-1100. Mehan, H., & Wood, H. (1975). The reality of ethnomethodology. New York: Wiley. Merikle, P. M. (1988). Subliminal auditory messages: An evaluation. Psychology and Marketing, 5(4), 355-72. Moore, T. E. (Ed.). (1988). Subliminal influences in marketing (Special issue]. Psychology and Marketing, 5(4), 291-372. Munsell, P., & Carr, T. (1981). Monitoring the monitor: Review of Second language acquisition and second language learning. Language Learning, 31, 493-502. Natsoulis, T. (1987). The six basic concepts of consciousness and William James' stream of thought. Imagination, Cognition and Personality, 6, 289-319. Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231-59. Norman, D., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.), Consciousness and selfregulation: advances in research and theory (Vol. 4). New York: Plenum Press, pp. 1-18. Ochs, E. (1979). Planned and unplanned discourse. In T. Givon (Ed.), Syntax and semantics, (Vol. 12): Discourse and syntax. New York: Academic Press, pp. 51-80. Ochs, E. (1986). From feelings to grammar: A Samoan case study. In B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures (251-272). Cambridge: Cambridge University Press. Odlin, T. (1986). On the nature and use of explicit knowledge. IRAL, 24(2), 123-44. Olshtain, E., & Blum-Kulka, S. (1989). Happy Hebrish: Mixing and switching in AmericanIsraeli family interactions. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.),

Consciousness, Learning, and Interlanguage Pragmatics

41

Variation in second language acquisition: Discourse and pragmatics. Clevedon & Philadelphia: Multilingual Matters, pp. 59-83. O'Malley, J. M., Chamot, A. U., & Walker, C. (1987). Some applications of cognitive theory to second language acquisition. Studies in Second Language Acquisition, 9, 287-306. Packard, V. (1957). The growing power of admen. Atlantic Monthly (September), 55-59. Pomerantz, A. (1978). Compliment responses: Notes on the co-operation of multiple constraints, (n J. Schenkein (Ed.), Studies in the organization of conversational interaction. New York: Academic Press, pp. 79-110. Posner, M., & Klein, R. (1973). On the functions of consciousness. In S. Kornblum (Ed.), Attention and performance IV. London: Academic Press, pp. 21-35. Reber, A. (1976). Implicit learning of synthetic languages: The role of instructional set. Journal of Experimental Psychology: Human Learning and Memory, 2, 88-94. Reber, A. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology: General, 118(3), 219-35. Reber, A., Allen, R., & Regan, S. (1985). Syntactical learning and judgment, still unconscious and still abstract: Comment on Dulany, Carlson, and Dcwey. Journal of Experimental Psychology: General, 114, 17-24. Reber, A., Kassin, S., Lewis, S., & Cantor, G. (1980). On the relationship between implicit and explicit modes in the learning of a complex rule structure. Journal of Experimental Psychology: Human Learning and Memory, 6, 492-502. Richards, J. C. (1990). The language teaching matrix: Curriculum, methodology, and materials. New York: Cambridge University Press. Rumelhart, D., & McClelland, J. (1986). On learning the past tense of English verbs. In McClelland, Rumelhart & the POP Research Group, 216-71. Rutherford, W., & Sharwood Smith, M. (1985). Consciousness raising and universal grammar. Applied Linguistics, 6, 274-82. Schmidt. R. (1983). Interaction, acculturation, and the acquisition of communicative competence: A case study of an adult. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition. Rowley, MA: Newbury House, pp. 237-74. Schmidt, R. (1984). The strengths and limitations of acquisition: A case study of an untutored language learner. Language Learning and Communication, 3(1), 1-16. Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 17-46. Schmidt, R., & Frota, S. (1986). Developing basic conversational ability in a second language: A case study of an adult learner of Portuguese. In R. Day (Ed.), Talking to learn: Conversation in second language acquisition. Rowley, MA: Newbury House, pp. 137-326. Seidenberg, M., & McClelland, J. (1989). A distributed, developmental model of word recognition and naming. Psychological Review, 96(4), 523-68. Seliger, H. (1983). The language learner as linguist: Of metaphors and realities. Applied Linguistics, 4, 179-91. Sharwood Smith, M. (1981). Consciousness-raising and the second language learner. Applied Linguistics, 2, 159-68. Shiffrin, R., & Schneider, W. (1977). Controlled and automatic human information processing. II. Perceptual learning, automatic attending, and a general theory. Psychological Review, 84, 127-90. Snow, C. E., Perlman, R. Y., Gleason, J. B., & Hooshyar, N. (1990). Developmental perspectives on politeness: Sources of children's knowledge. Journal of Pragmatics, 14(2), 289-305.

42

Cognitive Approaches to Interlanguage Pragmatic Development

Sokolik, M., & Smith, M. (1989). French gender recognition: A network model and implications for second language acquisition. Paper presented at the Second Language Research Forum, Los Angeles. Spolsky, B. (1989). Communicative competence, language proficiency, and beyond. Applied Linguistics, 10(2), 138-56. Sukwiwat, M., & Fieg, J. (1987). Greeting and leave-taking. PASAA, 17(2), 1-12. Underwood, G. (1976). Attention and memory. Oxford: Pergamon Press. Underwood, G. (1982). Attention and awareness in cognitive and motor skills. In G. Underwood (Ed.), Aspects of consciousness (Vol. 3): Awareness and self-awareness. London: Academic Press, pp. 111-45. Watson-Gegeo, K., & Gegeo, D. (1986). Calling-out and repeating routines among Kwara'ae children. In B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures (17-50). Cambridge: Cambridge University Press. White, P. (1980). Limitations on verbal reports of internal events: A refutation of Nisbett and Wilson and of Bern. Psychological Review, 57(1), 105-12. Williams, M. (1988). Language taught for meetings and language used in meetings: Is there anything in common? Applied Linguistics, 9(1), 45-58. Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House. Wolfson, N., D'Amico-Reisner, L., & Huber, L. (1983). How to arrange for social commitments in American English: The invitation. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition. Rowley, MA: Newbury House, pp. 116-28.

2 Symbolic Representation and Attentional Control in Pragmatic Competence ELLEN BIALYSTOK

Pragmatic competence entails a variety of abilities concerned with the use and interpretation of language in contexts. It includes speakers' ability to use language for different purposes—to request, to instruct, to effect change. It includes listeners' ability to get past the language and understand the speaker's real intentions, especially when these intentions are not directly conveyed in the forms—indirect requests, irony and sarcasm are some examples. It includes command of the rules by which utterances are strung together to create discourse. This apparently simple achievement to produce coherent speech itself has several components—turntaking, cooperation, cohesion. These three aspects of pragmatic competence converge most prominently on the ability to use and interpret nonliteral forms, such as metaphorical uses of language and indirect requests. Native speakers are expected to see through the forms to retrieve the speaker's intentions. When listeners fail to do this, conversation (and sometimes more than just conversation) breaks down. All three aspects of pragmatic competence are implicated in some respect for all conversation, but the challenge is greater when the conversation includes nonliteral forms to signal indirect requests. It is at these times that the ability to perform according to pragmatic rules is more rigorously tested. To participate successfully in such conversations, speakers need to have mastered the three aspects of language listed above. First, speakers must be capable of using language for different purposes so that the speech act of requesting is properly distinguished from other intended effects of language use. Second, speakers must be capable of modifying the form of the request to reflect social aspects of the context, embedding the request in the forms necessary to indicate the socially expected degree of politeness or deference. Finally, speakers must be capable of participating in an interaction following the conventions of conversation in order to make the request. Descriptions of how speakers come to use and understand indirect and meta43

44

Cognitive Approaches to Interlanguage Pragmatic Development

phoric speech provide an important component of the explanation of the development of language proficiency both for children learning a first language and adults learning a second language. For children, this skill is reflected in their ability to understand the intention of forms used in the following types of conversations (examples taken from Shatz, 1983, 843): (1)

MOTHER: MOTHER: CHILD: (2) EXPERIMENTER: CHILD:

He's three and three-quarters. (Child looks puzzled). How old are you? Four and two dollars. (inquiring about a 3-year-old's ability): Can you tie your shoes? They are tied.

This chapter will examine the nature and development of the linguistic representations that underlie pragmatic competence, specifically the competence involved in the use and interpretation of indirect and metaphoric speech. Knowledge for rules of use must be learned, represented, and transformed in the same way as the knowledge that controls other, more formal, aspects of the linguistic system. How is this knowledge represented, how do these representations change for speakers of a second language, and how do these representations support the use and interpretation of nonliteral speech? This chapter will begin with a brief summary of the descriptions offered for children's development of pragmatic abilities. I shall organize this summary around children's developing competence with the three aspects of pragmatic ability identified above: increase in the variety of speech acts, greater modification of speech according to the context, and mastery of conversation and discourse. In the next section, I outline a model of language processing. Finally, I shall apply this model to the problem of adult second language learners' development of pragmatic competence.

Children's Pragmatic Competence The most frequently studied aspect of pragmatic competence is the ability to make use of a variety of language functions. These functions have been identified and classified in various taxonomies. Searle's (1979) five-part classification system provides a moderate number of distinctions and is widely used. He includes the following as a finite and exhaustive list of possible speech acts: representatives, directives, commissives, expressives, and declarations. Other systems differ in number and detail. Halliday (1975), for example, considers the relevant distinctions to be among the ideational, interpersonal, and textual functions of language. The point, however, remains essentially the same, namely, that these categories constitute different linguistic functions that native speakers would normally be expected to master. How do children master the variety of language functions that native speakers of a language routinely express? According to Clark and Clark (1977), the speech of young children learning their first language at the earliest stages is mostly confined to assertions (representatives) and requests (directives). Early assertions generally indicate the existence of objects, describe their location, or comment on their

Symbolic Representation and Attentional Control

45

properties. Requests are typically motivated by needs or wants, asking for food, drink, toys, and the like. Within this limited number of categories and primitive syntactical resources for language structure, children manage nonetheless to express a range of semantic relations. These semantic relations are perhaps a better estimate of the children's ability to manipulate language to express a variety of intentions. Bloom (1973) astutely pointed out that when her daughter Kathryn said "mommy sock" on two different occasions, her intention in each was quite different. In one case, Kathryn was describing the spatial relation between an object and a location, and in the other, the interactive relation between an actor and an object. Similarly, a young child's utterance of "apple" can serve as a request, a declaration, or a question in different circumstances. In the preschool years, therefore, the range of pragmatic functions that children typically engage in is more limited than the range of semantic relations they express through these functions. Nonetheless, these semantic relations still constitute a finite set, including such intentions as presence of object, location of object, request for action, request for information (Bloom, 1973; Bowerman, 1973; Schlesinger, 1971). In spite of these limitations, Shatz and McCloskey (1984) argue that young preschool children know a great deal about different speech acts. They point out that by age 2 children are able to distinguish between such speech acts as questions and nonquestions regardless of how they are formed. There were even some children under 2 years who were found to discriminate between types of questions—yes/no questions from wh- questions. Shatz and McCloskey claim that these achievements indicate that "young children's responses are not governed by a literal interpretation strategy" and that they have an "early understanding of the constraints of discourse" (25). These two insights correspond to the second and third aspects of pragmatic competence identified here. As children enter the school years, there is an expansion in the range of language functions available to them. This increase is achieved in two ways: by adding new speech acts to the repertoire, such as commissives and expressives, and by learning to use a wider range of formal structures to express the speech acts they have already been using. This addition of new and more flexible resources to express old intentions enhances the child's pragmatic competence, even without the addition of new speech acts. By the early school years, children's control over a variety of speech acts is fairly impressive, although the use of declarations is still rare. The second aspect of pragmatic competence is the ability to adjust the form of speech to comply with contextual factors. This manipulation is most clearly signaled by children's ability to use and interpret such forms as indirect requests. As this aspect of competence is developing, young children often make conspicuous errors, such as the frequently cited example of the following telephone conversation: (3) CALLER: Is your mother home? CHILD: Yes.

In this case the child has failed to identify the question as an indirect request for action.

46

Cognitive Approaches to Interlanguage Pragmatic Development

The development of children's use of requests and the variety of direct and indirect means for expressing those requests is an impressive achievement in the preschool years. Becker (1982) documents the changes that occur in children's formulation of requests from the earliest attempts based only on gestures, sounds, and primitive utterances such as "more," to requests based on well-formed syntactic structures, and by about age 3 to the inclusion of some indirect requests. By age 5 children's command of requests is quite sophisticated, including a variety of indirect forms, hints and utterances that involve sequencing several steps to achieve the requested goal. The proportion of indirect to direct requests increases between the ages of about 3 and 5 years, but then remains relatively constant until about 8 years. Ervin-Tripp (1977) documents as well the development of children's ability to both produce and interpret indirect speech. She points out that there is considerable variability among children in this development and identifies some of the social and contextual factors that account for that variability. Perhaps more surprising is variability in the interpretation of indirect speech that she attributes to linguistic differences, even by the same children (Ervin-Tripp, 1989). She reports that FrenchEnglish bilingual children in Switzerland interpreted indirect speech correctly in French, accounting for such devices as irony and sarcasm, but incorrectly in English, taking the speech for its literal rather than its intended meaning. The third aspect of pragmatic competence is understanding the structure of conversation: What are the rules that govern how to structure discourse and participate in conversations? Knowing how to carry on a conversation includes knowing the rules for mechanical aspects, such as turn-taking, as well as linguistic aspects, such as coherence and knowledge of adjacency pairs, or what kinds of utterances are expected to follow others. McTear (1984) summarizes these achievements by claiming that the important feature of children's developing conversation is their increased awareness of and use of the contingency relations that characterize normal conversational interactions. Bruner (1983) claims that a great deal of the knowledge about conversations is learned in infancy. Adults and infants engage in conversation-type activities that include turn-taking, focusing on common objects, and vocalization (simply cooing or gurgling for the infant). From this basis, he argues, the essential dynamics of conversation are learned and easily transferred to linguistic interactions. Even with the prelinguistic experiences in which Bruner invests so much explanatory power, children must still learn the rules that govern linguistic conversations. Bloom, Rocissano, and Hood (1976) studied the development of discourse coherence by children to determine the extent to which utterances were appropriately related to preceding utterances. They found that children were contributing to conversations with considerable accuracy by the age of two. Shatz and McCloskey (1984) point to the need for specific knowledge of conversational rules; such knowledge is learned from early on but is certainly not complete in the preschool years: Speaking appropriately, then, is not just a matter of speaking grammatically or matching forms in initiation with forms in response. Appropriate language behaviour also involves knowledge of the conversational rules, conventions, and social circumstances governing linguistic interactions. (19)

Symbolic Representation and Attentional Control

47

There is some dispute about when these rules are learned. McTear (1984) attributes knowledge of them to preschool children but Ackerman (1981) defers their emergence until the school years. In summary, children continue to advance in their mastery of pragmatic uses of language through the preschool and early school years. They learn to use a wider range of speech acts, they learn to modify the forms of their utterances to accommodate social aspects of language use and to interpret indirect forms, and they learn the conventions for participating in conversations. Although the rudiments of all these abilities are evident from a very early age, possibly from the child's first words, real competence with these pragmatic functions of language cannot be claimed to be complete until the school years. Adult second language learners must also gain control over the pragmatic uses of language in contexts, but they certainly do not begin with a childlike naivete about the social uses of language. What sort of explanation, then, could account for development of pragmatic competence both by children who are learning their first language and by adults who already have mastered another linguistic system?

Learning and Using Language The analysis of pragmatic competence presented here is based on a model of language processing that we have been using as the framework for research into a variety of aspects of language acquisition and use. The model has two sides to it and requires two kinds of descriptions. One of these is the description of learner competence. For this purpose, the goal of the model is to describe the processing ability of language learners in terms of the cognitive mechanisms responsible for learning and using language. The second is the description of task demands. For this purpose, language functions are analyzed for the cognitive demands they place upon learners. Different uses of language involve different processing abilities of language learners; the assumption is that these demands can be systematically determined and analyzed for their difficulty. Language proficiency, then, is considered in terms of the fit between the processing abilities of the learner and the task demands imposed by a specific language use situation. Where the two are congruent, learners will perform well; where the task demands are excessive relative to the learner's ability, learners will struggle. The descriptions that are applied to these two sides of the problem (namely, speakers learning and using a language and the tasks they are mastering) are based on two cognitive components of language processing. These processing components are the basis for both assessing learners' abilities and determining task difficulty. The processing components are called analysis of knowledge and control of processing. Each component is specialized for a different aspect of the complex process of learning and using a language. Moreover, each processing component develops with experience and maturity on its own course. The development of these two processing components is normally correlated, as children generally move in a roughly synchronous manner toward greater mastery of each. Some situations,

48

Cognitive Approaches to Interlanguage Pragmatic Development

however, accelerate the development of one of them and create an imbalance in the learner's ability to process language. These imbalances would either enable children to succeed or prevent them from performing in language use situations that would normally be expected for children whose development followed the more usual course. Analysis of knowledge is the process of making explicit, or analyzing, a learner's implicit knowledge of a domain. As a result of the process of analysis, mental representations of a domain of knowledge become both more explicit and more organized around formal (abstract) categories. For language, knowledge of the linguistic system becomes explicit in terms of the constituents and categories that make up the language (words, letters, phonemes) and in terms of the rules that govern the structure and combination of those categories (subject-verb-object— SVO, negation). None of this explicitness is necessary for the uses of language undertaken by young children. One does not need explicit access to these concepts and rules in order to produce grammatical sentences or understand conversation. Once a domain has been analyzed, that knowledge can be used for functions not supported by implicit representations. The behavioral outcome of high levels of analysis is the ability to articulate structural principles of organization for the domain. At the same time, new operations become possible only with greater levels of analysis. Advanced or specialized uses of language, such as literacy skills, are based on more analyzed representations of language than are necessary for children's earlier uses of the system. The construct of analysis is compatible with theoretical notions of linguistic structuring by children posited by Berman (1986), Bowerman (1982), Karmiloff-Smith (1986), and others as well as the explication of implicit representations in other cognitive domains, such as classification, as described by Gelman and Baillargeon (1983). Control of processing is the process of controlling attention to relevant and appropriate information and integrating those forms in real time. Language presents multiple sources of information, both linguistic and nonlinguistic, and part of effective language processing is being able to attend to the required information without being distracted by irrelevant or misleading cues. Oral language, for example, provides potential information about meanings as well as formal information about the words and structures chosen, the style of speech used, and features of the pronunciation. Carrying on a conversation, however, requires that attention be directed primarily to the meanings. Donaldson (1978) contends that children need to develop control over attentional procedures to succeed in school and to progress in many of the tasks usually associated with Piagetian operational thought. These include such tasks as classification and conservation. The construct is similar to the executive control schemes, such as those posited by Case (1985), which regulate the selection and coordination of lower level schemes responsible for problem solving. The difference between executive schemes and control in the present system is that control of processing is not hierarchically related to other schemes. Rather, it exists at the same level as a different but equivalently powerful aspect of processing (cf. Jackendoff, 1987). In this sense, control of processing is the selective attention to the relevant aspects of a representation for the purpose of carrying out a specific task.

Symbolic Representation and Attentional Control

49

In my research into children's language development, I have investigated the role of the two processing components in the development of children's language proficiency. In this way, I have shown that children's progress with such activities as learning to read and developing metalinguistic awareness can be traced directly to their progress in mastering each of these component processes (e.g., Bialystok 1986, 1988). When tasks (such as understanding written texts or performing specific metalinguistic judgments) are analyzed for the demands they place upon each of the processing components of analysis and control, the result is generally that children who have been shown by means of independent measures to possess sufficient levels of the relevant processing component are the children who are likely to succeed on the language task. Consider, then, the development of each of the processing components in somewhat more detail. Regarding the process of analysis, the changes in representation of language brought about through analysis of the linguistic system can be divided into three levels of representation (Bialystok, 1991), I have called these conceptual representation, formal representation, and symbolic representation. In conceptual representation, language is organized only around the meanings it represents. Language is transparent and representations of meanings are based on semantic and conceptual information carried through the language. This kind of representation is children's first access to language. They understand language in context almost regardless of the specific forms being used. Similarly, children are not aware of the words they use to express their intentions, and the categories they form are built out of the semantic properties of those intentions, not out of the formal means for expressing them. In this view, oral uses of language can proceed from conceptual representations. Formal representations are coded in terms of the structure of language. The categories that organize the representations are the formal categories of language itself: language is made of words and sounds, rules of order must be observed when forming sentences, and the like. Children have formal knowledge of language structure when they can identify letters and tell you what sound a letter makes, count the number of words in a sentence, and detect and possibly correct grammatical violations. This level, then, refers to explicit knowledge of language structure and corresponds most closely to what most investigators mean by metalinguistic knowledge. In symbolic representation, the coding is a relation between a form and a referent. Symbolic representation is an explicit accounting of the way in which language refers. The knowledge of letters coded at this level is a symbolic relation between a form (the letter) and the sound it signifies (the referent). The representation is in the form of the relation stands for. Words are symbols that stand for meanings, so a symbolic coding of a word explicitly includes this relational property. Symbolic representation involves understanding that letters and written words do not have meaning in themselves but rather signify words and concepts that do have meaning. It is an advance for children to move from understanding letters and sounds as objects (the letter B makes the sound "b") to understanding letters as symbols (the letter B stands for the sound "b"). Symbolic representation deals directly with the problem of how language refers. In these terms, children may

50

Cognitive Approaches to Interlanguage Pragmatic Development

become literate only when their representations of linguistic knowledge have become analyzed to the point that the written forms for the language are understood as symbols and not merely as formal objects. Control of processing develops as well throughout childhood. As children move from using language as a strictly oral system to becoming acquainted with its written and formal properties, new attentional strategies are required. Reading, for example, demands a delicate balance between attention to the semantic and formal properties of the written text. Children's competence in early reading can be traced in part to their level of control of processing (Bialystok & Mitterer, 1987). If this model of language processing is applied to the development of pragmatic competence, then the abilities subsumed under descriptions of pragmatic competence must be stated in terms of their basis in levels of analysis and control. If these two processing components are responsible for language performance, then they must also be responsible for the pragmatic abilities of language users. Moreover, both components must be integrated into the account of pragmatic competence, even if it turns out that there is an imbalance in the levels of each that speakers must attain. Accordingly, two issues must be addressed. First, it is necessary to determine the ways in which representations of language must change to support the pragmatic functions of language and the level of analysis necessary for these uses. Second, it is necessary to determine the attentional strategies that must be developed to use language appropriately in contexts and the way in which these strategies relate to the ones normally used during conversation. These analyses will allow one to determine the level of proficiency within these two processing components necessary to demonstrate pragmatic competence and also permit a more detailed description of the processes that adult second language learners must master in order to achieve this competence.

Processes in Pragmatic Competence Throughout the descriptions of children's mastery of the three aspects of pragmatic competence described in the first section, two factors were consistently cited in one way or another. First is the need for children to expand their linguistic resources. Shatz and McCloskey (1984), for example, attribute the improved ability of older children to interpret speech acts to their increased vocabulary. These children have developed a range of possible acceptable responses to various forms, and these forms are selected more appropriately on the basis of contextual cues. Second is the ability to select between two possible interpretations of an utterance. In the case of the indirect request given in (3), for example, the choice is between the literal meaning (Is your mother home?) and the intended meaning (Call her to the telephone). Shatz and McCloskey (1984) make the following claim for development: "Children's earliest appreciation of speech acts, then, depends on their ability to attend to either contextual or surface features of an utterance. Their responses grow more sophisticated as their capacity to process information selectively from additional sources develops" (30).

Symbolic Representation and Attentional Control

51

These two aspects of development correspond more or less to advances in analysis and control respectively. Shatz (1983) similarly points to this kind of dual achievement for an explanation of the development of pragmatic competence: "A theory of the acquisition of communicative competence, then, must make reference to children's representational and processing capabilities as well as to the content of their social and linguistic knowledge" (844). Consider, then, some of the processing demands made upon analysis and control by pragmatic competence. For pragmatic knowledge, conceptual representation is the stage in which utterances can be formed to satisfy specific speech acts but the speaker is focusing on the intended meaning and not on the forms being selected to express that intention. Formal representations are organized around categories. At this stage, speakers are able to classify a group of utterances as being requests or assertions and are aware of the relation between direct and indirect speech acts. The representation of utterances, that is, can be in terms of the illocutionary force of those utterances. For symbolic representation there is additionally an explicit representation for the relation between the pragmatic intention and the forms used to achieve that intention. Representations at this level could include the relation between explicit devices such as politeness markers and their modifying effect on the interpretation of utterances in specific contexts. At this explicit relational level of representation, the learner would also have direct access to a repertoire of alternative forms of expression that indicate more or less the same intentional content. The critical aspect of the representation at this level is not the explicit forms or speech acts known by the learner as was the case for the formal level, but rather the relation between these forms and the meaning they convey in contexts. Such representations would make explicit the way in which different forms, such as politeness markers, modify the social interpretation of the utterance. Pragmatic competence is an achievement that depends minimally on formal representations, but ideally on symbolic representations. The interpretation of meaning in contexts is inherently a relational problem. Specifically, it involves the relation between a set of linguistic forms and the meanings intended by those forms in specific contexts. The representation that underlies this performance consists of a relation between a given meaning and a range of possible forms that give rise to that meaning. Selecting the appropriate form requires an assessment of contextual and social factors. Thus the mapping is not between form and meaning—the usual problem in semantics—but between form and social context, with meaning held constant across intentions within a socially defined situation. Put this way, the problem for pragmatics is to develop the resource of equivalents from which selections can occur. The richer the repertoire, the greater would be the pragmatic competence. Children need to learn all of this. Children acquire pragmatic competence by building up their representations of the linguistic system through all three stages. First they simply understand meaning in contexts. Macnamara (1972) cited this ability to understand utterances in context regardless of forms as the very basis of language acquisition. Next they learn about different speech acts and different forms and realize that these signal different meanings. But the relation between these

52

Cognitive Approaches to Interlanguage Pragmatic Development

forms and contexts is not worked out until this knowledge is represented symbolically. At this stage, selection from the formal alternatives can be deliberate and effective, and comprehension can be more accurate. Becker (1982) offers a description of pragmatic development based primarily on children's development of requests that is largely compatible with the view being described here. She points out that children's first requests are "unanalysed wholes" that eventually become decomposed to create new forms. Children build up more complete representations that allow them to manipulate the components of these wholes: "Toddlers can vary their requests systematically in several ways, which is only possible once they have acquired a more abstract system" (30). This "more abstract system," I would argue, is the formal or symbolic representation that is the result of the process of linguistic analysis. Children younger than 5 years old, according to Gordon and Ervin-Tripp (1984) do not, theoretically, have the processing ability to manage indirect requests, such as "It's cold in here." Yet, children under 5 years of age routinely produce and understand these forms correctly. How is this possible? The solution offered by Gordon and Ervin-Tripp is that these children have learned direct mappings between forms and functions and can use and understand them in familiar activities and interactions. They argue that for these children, the form "it's cold in here" is not an indirect request at all, but a perfectly direct means of achieving a goal. Shatz (1978) addresses the same question and suggests that 2-year-olds' (appropriate) responses to such forms as indirect requests may be based more on primitive response strategies than on inferential analysis and propositional content. Children may respond on the basis of some prototypical meanings (Shatz & McCloskey, 1984) that are suggested by the situation and not respond to the forms of language being used. Children may have heard sentences like "Can you talk on the telephone" in contexts in which action rather than a verbal answer was expected, creating a bias toward that response. This bias overestimates children's ability to interpret the illocutionary force of such indirect requests. On both these views, children's language competence is consistent with the argument that their knowledge of the pragmatic functions of language is based on conceptual representation, not formal or symbolic representation. Therefore, while children may appear to function appropriately, they cannot be described as having pragmatic competence. With conceptual representations, children are not even aware that the form is functioning as a request. When children come to understand that this form is one way of achieving a request, then their representation of that relation can be said to be formal. Symbolic representation would be signaled by the speaker's ability to deliberately select unusual or indirect forms, to manipulate aspects of the request such as the politeness or deference implied, or to understand the relation between the formal features of the utterance and its illocutionary function. Adult second language learners generally begin the task of mastering the pragmatic structure of the new language at the second (formal) level of representation and attempt to develop a symbolic representation from that system. Children begin with the problem of "how to mean," how to interpret indirect requests, how to string together discourse into conversation. Adults learning a second language have al-

Symbolic Representation and Attentional Control

53

ready sorted out the nature of meaning and already have explicit formal categories for concepts corresponding to speech acts. They may even have explicit categories for the formal pragmatic markers, such as politeness terms, used in the second language. The problem for adults is to learn the symbolic relation between forms and contexts appropriate to the second language. This may also entail expanding formal linguistic resources by learning new forms of expression or new sets of equivalents for given expressions in much the same way that vocabulary is learned. Some forms can even be learned in terms of a social context: when in a restaurant we say it this way. The claim here is strong regarding the amount of learning that is a precondition to pragmatic competence for children. In this view, children's knowledge of language must be drastically restructured and organized around a new set of formal explicit categories before children can even develop the symbolic representations that underlie real pragmatic competence. Adult learners of a second language also need to develop more analyzed representations of the language. Culturally specific forms and rules for pragmatic language use may require some specific analysis by adult learners both in the sense of organizing implicit knowledge into new explicit categories and in the sense of increasing the repertoire of language structure by learning new forms. New categories may be necessary because the second language may make social distinctions not followed in the speakers' first language, distinctions relating, for example, to social status, age, or sex of the listener. More forms may be necessary because the second language may include a richer variety of alternatives for expressing the same intentions. Conversational features such as rules for turn-taking, interrupting, and opening and closing conversations may also require explicit learning and analysis by adult second language learners. At the same time, some languages may be more indirect, and speech acts (such as requests) may be more buried. Second language learners would have to learn all these pragmatic conventions and the forms for conveying them. The control issue for pragmatic competence is the problem of attention. In addition to the usual problem of a formal (surface structure) and semantic (meaning) representation of an utterance competing for attention, the difficulty of deciding between literal and metaphorical (or indirect) semantic interpretations must also be resolved. How do children learn to ignore the literal meaning and respond by attending to the intention of utterances? In Clark's (1979) model for the interpretation of indirect meanings, indirect utterances give rise to two representations in listeners corresponding to the literal and intended meanings of the utterance. Adults have both interpretations simultaneously available and solve the problem of meaning by attending to one of them as a function of contextual factors. This selective attention may be especially difficult to execute in a second language where the conventions are less familiar. Cultural differences such as the use of indirect requests for action, the use of sarcasm and irony, and the forms for signaling politeness and deference may make the decision regarding the intended meaning difficult. Thus for adults the problem of correct interpretation is the responsibility of control of processing. Children, according to Shatz and McCloskey (1984), do not have two represen-

54

Cognitive Approaches to Interlanguage Pragmatic Development

tations available for the interpretation of indirect speech. They rely exclusively on contextual conditions and do not, in general, form a representation for the literal meaning. This direct mapping between the utterance and a unique meaning in context is a consequence of the language being represented at the conceptual level. De Villiers (1984) puts the case somewhat more tentatively but makes essentially the same point. She argues that children do not have multiple mappings between forms and meanings and that their interpretation of meanings is based on their knowledge of familiar contexts and familiar utterances. Thus the control problem is reduced, as there are not two interpretations competing for attention. Intended meaning is resolved more on the basis of social-contextual circumstances than on the basis of linguistic forms. In summation, the problem presented to children learning their first language and adults learning their second language regarding pragmatic competence is different. For children, the primary task is the process of analysis. Children need to explicate and expand their linguistic resources to cope with the demands of using language for different purposes, in different contexts, and to different effects. Adults need to learn some social conventions for the new language and to solve some mapping problems between forms and social conditions, but the task of organizing a linguistic system along the lines of speech functions so that the system is operable in contexts has been solved for a first language. Adjustments in the organizing structures in which the first language is already represented as a formal or symbolic system carrying pragmatic functions are a relatively small problem. For adults, the problem to be solved for pragmatic competence is essentially to develop the control strategies to attend to the intended interpretations in contexts and to select the forms from the range of possibilities that satisfy the social and contextual needs of the communicative situation. Adults make pragmatic errors, not only because they do not understand forms and structures, or because they do not have sufficient vocabulary to express their intentions, but because they choose incorrectly. They attend to the wrong possibility for meaning, they fail to attend to a social distinction that needs to be marked linguistically, or they select the incorrect politeness marker for the situation or the listener. Many examples of the kinds of new selection strategies adults must learn when speaking a second language and the kinds of social and contextual errors they make while learning them are described in the research reported in the collection by Blum-Kulka, House, and Kasper (1989a). Pragmatic competence is achieved when control of processing is mastered for a richly analyzed representation of the language. Adults do, of course, still need to worry about the analysis problem. They need to continue to build up their repertoire of formal linguistic resources and to verify that their organization of the system has followed the correct categories. In my view, this problem is relatively minor; other views, however, lead to different conclusions. Blum-Kulka, House, and Kasper (1989b) describe two opposing views on this issue. Eraser (1985) argues that speech act markers are universal but have different distributions across languages and cultures, while Wierzbicka (1985) argues speech act markers are unique realizations created by languages and cultures. The prevailing opinion in the literature appears to be that of Fraser, and it is from that assumption that my argument, which attributes little analysis difficulty to adult

Symbolic Representation and Attentional Control

55

second language learners, follows. If Wierzbicka is correct, the analysis problem for adults learning a second language is considerably exacerbated. Similarly, while children's greatest obstacle for mastering pragmatic competence is in developing analyzed representations of language, they still need to pay attention to control. Errors such as the one in (3) do occur, and children have to learn the processes for attending to the appropriate meaning possibility. Children use and understand indirect speech from an early age, and so the role of control in early language processing for pragmatic proficiency is tangible. The argument presented here, however, is that the greatest challenge for each group, children learning their first language and adults learning a second language, lies in a different component. This division, in which children face a more serious barrier from analysis and adults from control, is not accidental. The two processing components are intricately connected to each other, but are nonetheless ordered in their priority. They are connected because each entails the other. Analyzed representations carry with them the prospect of selective attention to aspects of a representation or deciding between competing representations. Yet attention is a vacuous concept in the absence of content, so control of processing presupposes representations. This is not a hopeless tangle; in my empirical work I have been able to measure learners' levels of mastery of these two processing components independently and show their separate involvement in language tasks. Thus, although proficient reading, for example, depends on both analysis and control, analysis is prior (Bialystok, 1988). Further research with the processing abilities of children and adults as they master the pragmatic functions of language will address these claims directly.

References Ackerman, B. P. (1981). The understanding of young children and adults of the deictic adequacy of communication. Journal of Experimental Child Psychology, 31, 256-70. Becker, J. A. (1982). Children's strategic use of requests to mark and manipulate social status. In S. A. Kuczaj (Ed.), Language development (Vol. 2): Language, thought and culture (1-35). Hillsdale, NJ: Erlbaum. Berman, R. A. (1986). A step-by-step model of language learning. In I. Levin (Ed.), Stage and structure: Reopening the debate (191-219). Norwood, NJ: Ablex. Bialystok, E. (1986). Factors in the growth of linguistic awareness. Child Development, 57, 498-510. Bialystok, E. (1988). Aspects of linguistic awareness in reading comprehension. Applied Psycholinguistics, 9, 123-139. Bialystok, E. (1991). Letters, sounds, and symbols: Changes in children's understanding of written language. Applied Psycholinguistics, 12, 75—89. Bialystok, E., & Mitterer, J. (1987). Metalinguistic differences among three kinds of readers. Journal of Educational Psychology, 79, 147-53. Bloom, L. (1973). One word at a time: The use of single word utterances before syntax. The Hague: Mouton. Bloom, L., Rocissano, L., & Hood, L. (1976). Adult-child discourse: Developmental interaction between information processing and linguistic knowledge. Cognitive Psychology, 8, 521-52.

56

Cognitive Approaches to Interlanguage Pragmatic Development

Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989a). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Blum-Kulka, S., House, J., & Kasper, G. (1989b). Investigating cross-cultural pragmatics: An introductory overview. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural pragmatics: Requests and apologies (1-34). Norwood, NJ: Ablex. Bowerman, M. (1973). Early syntactic development: A cross-linguistic study with special reference to Finnish. Cambridge: Cambridge University Press. Bowerman, M. (1982). Reorganizational processes in lexical and syntactic development. In E. Wanner & L. R. Gleitman (Eds.), Language acquisition: The state of the art (319346). Cambridge: Cambridge University Press. Bruner, J. S. (1983). Child's talk: Learning to use language. New York: Norton. Case, R. (1985). Intellectual development: Birth to adulthood. New York: Academic Press. Clark, H. H. (1979). Responding to indirect speech acts. Cognitive Psychology, 11, 430-77. Clark, H. H., & Clark, E. V. (1977). Psychology and language. New York: Harcourt Brace Jovanovich. de Villiers, J. (1984). Form and force interactions: The development of negatives and questions. In R. L. Schiefelbusch & J. Pickar (Eds.), The acquisition of communicative competence (193-236). Baltimore, MD: University Park Press. Donaldson, M. (1978). Children's minds. Glasgow: Fontana. Ervin-Tripp, S. (1977). Wait for me, roller skate! In S. Ervin-Tripp & C. Mitchell-Kernan (Eds.), Child discourse (165-188). New York: Academic Press. Ervin-Tripp, S. (1989). Children's social concepts in language development. Invited address to the Society for Research in Child Development, Kansas City, MO. Fraser, B. (1985). On the universality of speech act strategies. In S. George (Ed.), From the linguistic to the social context (43-69). Bologna, Italy: CLUEB. Gelman, R., & Baillargeon, R. (1983). A review of some Piagetian concepts. In J. H. Flavell & E. M. Markman (Eds.), Handbook of child psychology, (Vol. 3): Cognitive development (167-230). New York: Wiley. Gordon, D., & Ervin-Tripp, S. (1984). The structure of children's requests. In R. L. Schiefelbusch & J. Pickar (Eds.), The acquisition of communicative competence (83-106). Baltimore, MD: University Park Press. Halliday, M.A.K. (1975). Learning how to mean. London: Arnold. Jackendoff, R. (1987). Consciousness and the computational mind. Cambridge, MA: MIT Press. Karmiloff-Smith, A. (1986). From metaprocess to conscious access: Evidence from children's metalinguistic and repair data. Cognition, 28, 95-147. Macnamara, J. (1972). The cognitive basis of language learning in infants. Psychological Review, 79, 1-13. McTear, M. F. (1984). Structure and process in children's conversational development. In S. A. Kuczaj (Ed.), Discourse development: Progress in cognitive development research (37-76). New York: Springer-Verlag. Schlesinger, I. M. (1971). Production of utterances and language acquisition. In D. I. Slobin (Ed.), The ontogenesis of grammar: A theoretical symposium (63-101). New York: Academic Press. Searle, J. R. (1979). A taxonomy of illocutionary acts. In J. R. Searle (Ed.), Expression and meaning (1-27). Cambridge: Cambridge University Press. Shatz, M. (1978). On the development of communicative understandings: An early strategy for interpreting and responding to messages. Cognitive Psychology, JO, 271-301. Shatz, M. (1983). Communication. In J. H. Flavell & E. M. Markman (Eds.), Handbook of Child Psychology, (Vol. 3): Cognitive Development. New York: Wiley, pp. 84-89.

Symbolic Representation and Attentional Control

57

Shatz, M., & McCloskey, L. (1984). Answering appropriately: A developmental perspective on conversational knowledge. In S. A. Kuczaj (Ed.), Discourse development: Progress in cognitive development research. New York: Springer-Verlag, pp. 19-36. Wierzbicka, A. (1985). Different cultures, different languages, different speech acts. Journal of Pragmatics, 9, 145-78.

This page intentionally left blank

II SPEECH ACT REALIZATION This section includes five studies of interlanguage speech act performance: thanking, apologizing, complaining, requesting, and correcting. While there is a large literature on requests and only somewhat less on apologies, the other three speech acts have not yet been studied much. Consequently, for the investigation of requests and apologies, there is a much richer empirical and theoretical foundation for interlanguage pragmaticists to work from. According to various semantic criteria, the five speech acts form intersecting groups. In Searle's (1976) typology, thanking, apologizing, and complaining represent expressive speech acts; requesting is a directive act; and correcting, a representative act. Following Leech's (1983) functional classification, thanking and apologizing are convivial, while complaining, requesting, and correcting are competitive acts. While these groupings may not seem to be of much interest beyond speech act theory, they have behavioral correlates in speech act realization patterns, suggesting that there is indeed a relationship between speech act theorists' analyses and language users' communicative practice. Because illocutionary and social goals are compatible in the case of convivial acts, thanking and apologizing tend to be aggravated in force. Complaining, requesting, and correcting tend to be mitigated, reflecting the tension between illocutionary and social purpose. However, these are rough generalizations: as has frequently been pointed out in the literature, not only would it depend on the context of a request whether and how much it needs to be mitigated, it might even be perceived as quite consistent with the hearer's facewants; for instance, if the very fact of being asked to do something, even if involving some effort on the part of the speaker, conveys appreciation of the hearer's competence or attributes. Indirectness and conventionalized means of implementation are common features of most linguistic action, yet speech acts differ in the extent to which fixed conventionalized expressions are routinely used. Since thanking and apologizing are regularly implemented by means of standardized routines (though, obviously, not only by them), learners not only have to find out when those acts are appropriate or even required in the target community but also what formulas need to be used, and in conjunction with what other means of expressing gratitude or regret. In case of acts that are contextually perceived by the speaker as posing a threat to the hearer's face, such as the competitive acts under study in the following chapters, the main decisions for learners (as well as native speakers) to make are the ones formulated 59

60

Speech Act Realization

by Brown and Levinson (1987): whether to carry out the face-threatening act at all, and if so, whether to do it in an on-record or some kind of off-record fashion. The literature documents considerable cross-cultural variability and learner-specific performance patterns on these issues (see the Introduction). Eisenstein and Bodman (Chapter 3) examine the expression of gratitude by native and nonnative speakers of English, representing a variety of linguistic and cultural backgrounds. Their study combines data from a number of sources: observation of naturalistically occurring events, role-plays, oral and written production questionnaires, and retrospective interviews. LI data were obtained using written questionnaires. Learners' responses were rated for appropriateness, using native speakers' thanking performance as baseline data. Differences between learners' and native speakers' expressions of gratitude were noticeable at the sociopragmatic and pragmalinguistic level. Sociopragmatic differences involved the kinds of events and interlocutor relationships that require thanking, as well as the learners' familiarity with the situational context. Thanking behavior was further influenced by the culture-specific prominence of values such as modesty, gratefulness, and indebtedness. While the learners' performance was pragmalinguistically successful in contexts that required simple, ritualized responses, they did much less well when more complex, creative expressions were called for. Even the advanced learners often did not fluently command the range of conventions of means and forms by which the native speakers expressed gratitude. Because Eisenstein and Bodman used a variety of data types, the impact of these instruments on the subjects' responses could be assessed. Questionnaires, roleplays, and naturalistic data exhibited the same types and wording of semantic formulas. However, not only did the interactive conditions produce more linguistic activity in terms of restatements and reformulations, they also emphasized the role of the giver in the speech event: it was through her conversational contributions that the receiver's choice of specific semantic formulas was prompted. Thanking thus appeared as a jointly developed event, involving mutually adjusted contributions by both participants. One of the issues demonstrated by Eisenstein and Bodman's chapter is the crosscultural variability of contextual factors in the event of thanking. Bergman and Kasper (Chapter 4) focus on the assessment of contextual factors in situations where the speaker has committed some offense, and the way in which contextual assessment influences informants' selection of apology strategies. Informants were Thai learners of English and native speakers of American English and Thai. Data were collected by means of assessment and production questionnaires. Informants largely agreed in their perception of the relationship between context-external factors (social distance and dominance) and context-internal factors (severity of the offense, offender's obligation to apologize, the likelihood for the offended party to accept the apology, and offender's loss of face). However, when it came to assessing individual offense contexts, cross-cultural differences prevailed, Thai and American respondents differing most in their perception of the offender's obligation to apologize. Of the apology strategies, aggravating the apology was most sensitive to contextinternal factors. Respondents assumed more explicit responsibility for the offense the closer they perceived the relationship between the interlocutors. Social distance

Speech Act Realization

61

thus was important in apology performance. Status, on the other hand, was unrelated to learners' and native speakers' strategy choice. While learners and English native speakers displayed the same frequency patterns in expressing explicit apology and responsibility as well as in their upgrading of apologetic force, the learners made more use of context-dependent strategies such as downgrading responsibility and severity, offering repair, and different forms of verbal redress. In more than half of the cases, the difference between learners' and American English native speakers' strategy suppliance could be traced to sociopragmatic transfer from Thai. Bergman and Kasper discuss the learners' oversupplying, relative to the target norm, of context-dependent strategies as yet another instance of the verbosity phenomenon noted in some of the interlanguage pragmatics literature. Evidence is accumulating that verbosity is very likely to be instrument-induced. While learners' responses to (written) production questionnaires, such as in Bergman and Kasper's study, are typically characterized by verbosity, their contributions to interactive tasks such as role-plays are not. In fact, Eisenstein and Bodman's (Chapter 3) study shows that learners' expressions of gratitude in role-plays are 50% shorter than those of native speakers of English. These output differences reflect differential cognitive demands placed on learners in the two conditions, and highlight the research purposes to which the two types of data collection instruments might best be put. Production questionnaires serve to elicit knowledge displays without making demands on learners' fluency or interactional skills. Role-plays, on the other hand, require the ability to compute contextual factors and assemble relevant linguistic material in a highly automatized fashion. Because they make simultaneous demands on learners' comprehension and production systems, they are useful tools for probing learners' ability to instantiate sociopragmatic and pragmalinguistic knowledge in interaction. Apologies and complaints may occur in response to the same event: apologies as remedial activity produced by the offender, and complaints as expression of disapproval voiced by the offended party. Olshtain and Weinbach (Chapter 5) examine cross-cultural and situational variability in the choice of complaining strategies, and the interaction of contextual assessment with strategy choice. Data were collected by means of the same types of instrument as in Bergman and Kasper's apology study: production questionnaires to elicit responses to offensive events, and assessment questionnaires to probe for informants' perceptions of contextual factors. Olshtain and Weinbach's chapter reports on three studies: a study of complaint performance by native speakers of Hebrew, a cross-cultural comparison of complaining by native speakers of Hebrew and of British and American English, and an interlanguage study, comparing complaint realization by nonnative speakers of Hebrew at intermediate and advanced proficiency levels with that of native speakers. The responses by the native speakers of Hebrew were found to cluster around three strategies: expressing disapproval, voicing an explicit complaint (such as "you should have done x"), and giving a warning. Strategy choice was determined by interlocutors' status relationship, low-high constellations prompting less severe complaints; equal and high-low relationships, more severe ones. The cross-cultural comparison yielded highly consistent response patterns across the three native speaker groups. An important feature to note is that although informants were given

62

Speech Act Realization

the choice to opt out, two-thirds of the respondents did choose to complain. It will be a matter for future studies to examine whether members of more interdependence-oriented cultures choose to abstain from complaining more frequently. Learners' complaints differed from those of the target group on all selected measures. The nonnative speakers produced longer complaint utterances, chose more severe complaining strategies, and used both more softeners and intensifiers. Strategy choice was influenced by interlocutors' relative status, social distance, and the hearer's obligation to have avoided the offensive act. Both groups produced the most talk when no explicit hearer obligation existed, and the least when there was a clear contractual violation. Increasing obligation resulted in more severe strategies in both groups. However, the learners produced longer utterances when the hearer's obligation was implicit, and they opted for more severe strategies than the native speakers when an explicit obligation had been violated. Just as in Eisenstein and Bodman's findings, the learner responses in this study, too, displayed more variability, suggesting that these nonnative speakers were not quite accustomed to target conventions of means and forms. Conventionality of means and forms has predominantly been studied in the context of requests. In a variety of request contexts, native and nonnative speakers of different languages were shown to prefer conventionalized indirect request strategies. In interlanguage pragmatics in particular, little attention has been given to indirectness in requests that is not conventionalized in terms of particular illocutionary force indicating frames (though conventionality of means is extant in indirect requests, too). Weizman (Chapter 6), therefore, in her study on requestive hints, makes an important contribution to understanding interlanguage requesting. Since previous research has shown hints to be less polite than conventionally indirect requests, their prime function is not one of politeness. Rather, their inherent opacity (opacity of illocution, proposition, or both) enables speaker and hearer to exploit their deniability potential. Based on this functional analysis, Weizman hypothesized that learners make more use of hints than native speakers, employing them as a communication strategy in compensation for the lack of conventionalized forms. Comparison of the hinting patterns in native and nonnative speakers' responses to eight request contexts in a discourse completion questionnaire did not reveal greater reliance on hinting by learners than by native speakers, hence no support for the communication strategy hypothesis. Length of residence in the target community had no effect on the amount of hints used, suggesting that learners possess the ability to employ nonconventional indirectness from early on. The nonnative speakers matched native speaker options in the contextual distribution of hints and preference for substrategies. Both groups preferred hints with the greatest illocutionary opacity, confirming the deniability potential hypothesis. The nonnative speakers, however, employed more redundant combinations, suggesting that learners invest more linguistic activity in reaching their communicative goals even when hinting. This part concludes with a study by Takahashi and Beebe (Chapter 7) on corrections performed by Japanese learners of English in comparison with native speakers of Japanese and American English. In their previous studies of face-threatening acts carried out by the same groups of native and nonnative speakers, the authors had found distinctive patterns of style shifting according to interlocutor status. Focusing

Speech Act Realization

63

on the modification of corrections by means of positive remarks and softeners, the Japanese learners' style-shifting patterns were clearly influenced by transfer from Japanese. While Japanese learners, reflecting native sociopragmatic norms, styleshifted more than American respondents when refusing, contradicting and disagreeing, this study indicated dramatic style-shifting in the American speakers' use of positive remarks. Their prevalent use of positive remarks in the high-low condition, which was not matched by the Japanese learners or native speakers, provides more evidence of a positive politeness orientation in American interaction, and greater emphasis on status congruence in Japanese conversational behavior. The study also confirms Beebe and Takahashi's earlier claim that pragmatic transfer prevails in higher proficiency learners. The five chapters on interlanguage speech act realization highlight a number of important issues: (1) even quite proficient learners tend to have less control over the conventions of forms and means used by native speakers in the performance of linguistic action; (2) differences between learners' and native speakers' sociopragmatic perceptions of comparable speech events are systematically related to differences in their speech act performance; (3) transfer at the pragmalinguistic and sociopragmatic level persists at higher levels of proficiency; (4) learners produce more speech than native speakers when the task is less demanding on their control skills; and (5) researchers must pay close attention to the constraints of different data collection instruments on learners' performance.

References Brown, P., & Levinson, S. C. (1987). Politeness. Cambridge: Cambridge University Press. Leech, G. (1983). Principles of pragmatics. London: Longman. Searle, J. (1976). A classification of illocutionary acts. Language in Society, 5, 1-23.

3 Expressing Gratitude in American English MIRIAM EISENSTEIN and JEAN BODMAN

Expressing gratitude has important social value in American English. One indication of its importance is that it is one of the few functions that most speakers can remember being explicitly taught as children. Used frequently in a wide range of interpersonal relationships, this function, when appropriately expressed, can engender feelings of warmth and solidarity among interlocutors. Conversely, the failure to express gratitude adequately can have negative consequences for the relationship of speaker and listener. Expressions of gratitude can range from simple, phatic utterances to lengthy communicative events mutually developed by both the giver and the recipient of a gift, favor, reward, or service. Most native speakers of English on a conscious level associate the expression of gratitude with the words "thank you"; however, they are unaware of the underlying complex rules and the mutuality needed for expressing gratitude in a manner satisfying to both the giver and recipient. Similarly, second and foreign language learners are unaware of the underlying rules for expressing gratitude in English; in fact, they usually assume that the expression of gratitude is universal and remain unaware of significant differences in its cross-cultural realization (Eisenstein & Bodman, 1986). Because of this, the function of expressing gratitude is particularly difficult for learners to perform successfully.

Background The importance of social interaction rituals, such as expressing gratitude, was addressed by Goffman (1967). He explained that every individual has a potential emotional response to others which is related to his or her "face," defined as "an image of self, delineated in terms of approved social attributes" (5). Violation of social norms may result in loss of face with accompanying negative feelings on the part of one or both participants. Brown and Levinson (1987) identify the desire for approval as "positive face" and define "negative face" as "the desire to be unim64

Expressing Gratitude in American English

65

peded in one's actions" (13). Some acts, verbal and nonverbal, may be counter to the "face wants" of speaker or hearer and, therefore, are "face-threatening acts" (65). Brown and Levinson categorize expressing thanks as a face-threatening act in which the speaker acknowledges a debt to the hearer—thus threatening the speaker's negative face. Searle (1969) stresses the positive aspects of thanking, which he defines as an illocutionary act performed by a speaker based on a past act performed by the hearer that was beneficial. Leech (1983) describes thanking as a convivial function whose goal of stating appreciation helps maintain a polite and friendly social atmosphere. Thus, it is apparent that expressing gratitude is a complex act potentially involving both positive as well as negative feelings on the part of giver and receiver. Thomas (1983) identifies the difficulty encountered by nonnative speakers in the cross-cultural realization of speech acts. She notes that misunderstandings can arise not only from language limitations (pragmalinguistic failure) but also from inadequate utilization of social conventions and values in the target culture (sociopragmatic failure). Coulmas (1981) views thanking in cross-cultural perspective and underscores the challenge for speakers of European and Asian languages to express thanking adequately to each other. He posits a useful distinction between thanks that entail indebtedness to the addressee and thanks that imply no indebtedness. It is useful to keep in mind that not all expressions using the words "thank you" refer to gratitude. Rubin (1983) collected natural data on uses of the words "thank you" that referred not only to gratitude but also to other language functions, such as compliments and closings. In fact, Hymes (1971) states that "thank you" as it is used in British English is often more of a formal marker than an expression of gratitude, which is its more common function in American English.

Experiment 1 We began investigating how gratitude is expressed by observing its use in natural contexts by native speakers of American English. Data were either audiotaped and transcribed at a later time or written down as field notes. We focused on only those utterances whose illocutionary force was that of gratitude in response to receiving a gift, favor, reward or service.1 Fifty situations were identified in which expressions of gratitude occurred. These ranged from short, formulaic expressions that appeared to be highly ritualized to intricate and lengthy interchanges conveying deeply felt emotions. As a result of a series of pilot studies, we constructed a questionnaire containing fourteen of these situations designed to elicit expressions of gratitude. The pilot studies revealed the importance of describing the roles and relationships of the interlocutors in the questionnaire in detail as well as carefully describing specific contexts for the situations; we found that small changes in these variables significantly affected the nature of the responses. Fifty-six native speakers of American English, representing both males and females, a range of ages, diverse social backgrounds, and natives of a variety of regions with the United States, were asked to write responses to each of the fourteen situations. The written data elicited by the questionnaire were analyzed by coding each

66

Speech Act Realization

utterance in terms of its underlying speech act. For example, on leaving a dinner party, a guest said: Thank you for inviting me. I had a great time.

We coded this utterance as "Thanking + Expressing Pleasure." Wherever possible, we used functional categories described in the literature (van Ek, 1976, Searle, 1969); however, in some instances we had to create our own tentative terminology where appropriate descriptors had not been previously identified. For example, after opening a gift, a subject responded: Oh, how beautiful. How did you know? It's just what I wanted!

The italicized utterances were clearly not intended to express the functions of Asking for Information or Expressing Need. We tentatively coded them as "Expressing Intimacy: Mind Reading," an indirect compliment acknowledging the accuracy of the giver's understanding of the receiver's unexpressed desires. The fourteen questionnaire items are briefly summarized in Appendix 1. Some of these items produced phatic, ritualized responses, such as brief comments to a bus driver (Item 2), a cashier (Item 5), a garage attendant (Item 13), and a friend handing over a newspaper (Item 12). In post hoc interviews, native American English speakers characterized utterances, such as Thanks.

Thank you. Have a nice day.

as virtually automatic. The expression of gratitude in situations such as these appears to be a social amenity rather than a genuine expression of gratitude and has been identified by Rubin (1983) as the "bald thank-you." Some other items on the questionnaire produced relatively short, but more creative responses than those described above. Swift thanking followed by a single brief comment was typical for Item 6 (a friend bringing attention to a bit of food on a diner's face) and Item 8 (thanking a spouse for spontaneously helping around the house). Typically, our respondants felt (although a great more could have been said) that comments like Thanks. You're a sweetheart. (coded as Thanking + Expressing Affection) and Thanks. That was really nice of you. (coded as Thanking + Complimenting the Giver) were sufficient in recognizing a spouse's thoughtfulness. The remaining items elicited and seemed to require much more complex and lengthy expressions of gratitude. These were most successfully analyzed as speech act sets (Cohen & Olshtain, 1981), groups of semantic formulae that together achieve the appropriate language for a particular situation. In appreciation for a generously offered $500 loan, a characteristic response was: You're a lifesaver. Thanks. I'll never forget it. You really can't imagine what this means to me.

Expressing Gratitude in American English

67

(Coded as Complimenting the Person/Action + Thanking + Expressing Indebtedness + Expressing an Inability to Articulate Deep Feelings). Our respondents, when interviewed later, indicated that in these situations gratitude was much more challenging to express. It was not uncommon for the respondents to state with humility that their linguistic skills were inadequate for the task. American native speakers, rating the appropriateness of the utterances, found humbling admissions such as this not only expressed the depth of the feelings in a satisfactory manner, but also adequately expressed gratitude. It is useful to consider two characteristics that Goffman (1967) describes that are operant in social interactions—"demeanor" and "deference"—in order to understand the complex linguistic task speakers face. "Demeanor" refers to the social desirability of an individual reflected through his or her appearance and behavior. "Deference" is the appreciation an individual exhibits to another through his or her words and actions. Expressing gratitude requires that the recipient of a gift, favor, or service exhibit both proper demeanor and proper deference in situations in which he or she is feeling especially vulnerable. The recipient must show humility and gratitude without losing dignity and control. The giver must remain sensitive to the needs of the receiver and also behave with adequate deference and demeanor. The struggle to find the words and exhibit acceptable demeanor and deference that are mutually satisfactory is highly challenging. In addition, the difficulty of the task is further compounded by the fact that both demeanor and deference in expressing gratitude are culturally bound and, hence, difficult to translate from one sociolinguistic context to another. The lengthiest speech act sets were produced by situations that caused the recipient to feel unusually grateful or indebted to the giver. None of the individual semantic formulas constituting a set could be identified as more salient or of a higher order than the others. Instead, the members of each set interacted synergistically to express gratitude appropriately. Furthermore, the functions within the speech act set did not appear in a fixed order. The direct expression of thanking, for example, could be stated at the beginning, the middle, or the end of the set. While greater emotion sometimes provoked longer speech act sets, this did not occur when there was considerable social distance between interlocutors. Item 4 (a vice president of personnel offers a relatively new employee a raise) elicited surprisingly brief expressions of gratitude. This confirms one aspect of Wolfson's Bulge Theory (1989), identifying brevity in communications between socially distant interlocutors. However, we did not find in all cases the terseness that Wolfson (1989) identified in communications among intimates. While the situation in which a spouse is helpful around the house did produce a brief response, the situations involving gift giving and loaning a large amount of money evinced quite lengthy responses despite the intimacy of the interlocutor relationship. Our data revealed that, although speakers are free to say anything they wish in expressing gratitude (and there were occasional examples of highly creative speech act sets), most speakers seemed to draw from a finite pool of conventionalized expressions and ideas. Within the speech act set of expressing gratitude for a gift, many native respondants referred to the lack of necessity for such generosity by using expressions such as "Oh, you shouldn't have" and "You didn't have to."

68

Speech Act Realization

In accepting the loan, native speakers stated their inability to express their appreciation sufficiently: God, I don't know how to thank you. I can't tell you how much I appreciate this. 1 can't thank you enough. I can't tell you what this means to me.

Experiment 2 The same questionnaire that was administered to the native speakers of American English was given to 67 nonnative college students in advanced-level ESL classes. They were middle-class and represented fifteen language backgrounds—the largest groups were Chinese, Korean, Japanese, Russian, and Spanish. The majority of these learners had been in the United States for approximately two years, although the range was as low as three months and as high as five years. The nonnative speakers were asked first to respond to the items in the questionnaire in English. Subsequently, they were given the opportunity to respond to the same situations in their native languages if they wished. Twenty-five of the subjects provided us with LI responses to the situations. (Bilingual speakers were asked to provide literal and figurative translations of the LI responses.) This allowed us to consider instances of cross-linguistic influence. Using native speaker data as a baseline, we rated nonnative written responses. Rating descriptors can be found in Table 3.1. The nonnatives were successful on the whole in responding in a nativelike manner to situations requiring simple, phatic, ritualized expressions of gratitude. However, items requiring complex speech act sets (Items 1, 3, 4, 7, 9, 10, 14) were problematic for many of our nonnative subjects. They were not able to approximate acceptable native speech acts in a significant number of cases. In fact, their difficulties in adequately expressing gratitude in a nativelike manner were extensive and severe. (See Table 3.2 for a summary of ratings on nonnative responses.) Table 3.1. Acceptability Scale for Nonnative Responses Not acceptable

A violation of a social norm, a faux pas, a likely instance of sociopragmatic failure.

Problematic

An error that might cause misunderstanding, but of a less serious nature. Language so strange, unexpected, or garbled that interpretation is difficult. Instances of pragmalinguistic and/or sociopragmatic failure.

Acceptable

Clear and appropriate language, but containing small errors that do not interfere seriously with native speakers' understanding.

Nativelike perfect

Close to native responses in content, syntax, and lexicon.

Not comprehensible

An utterance that is extremely hard, if not impossible, to comprehend. Often an instance of pragmalinguistic failure.

Resistant

Nonnative participants, although finding it possible to answer some items, refused to answer others, or gave reasons why they could not or would not answer particular items.

69

Expressing Gratitude in American English Table 3.2. Summary of Results of All Students on Individual Questions Question 1 3

Question topic $5.00 Sweater

4

Raise

7

$500.00

9

Lunch

No response

Not acceptable

Problematical

Acceptable

Perfect

NC*

5 7.5%

3 4.5%

9 13.4%

9 13.4%

39 58.2%

1 1.5%

1 1.5%

2 3.0%

3 4.5%

19 28.4%

14 20.9%

29 43.3%

0 0%

0 0%

5 7.5%

10 14.9%

16 23.9%

14 20.9%

18 26.9%

1 1.5%

3 4.5%

9 13.4%

3 4.5%

14 20.9%

15 22.4%

21 31.3%

1 1.5%

4 6.0%

5 3.0%

11 16.4%

22 32.8%

9 13.4%

14 20.9%

1 1.5%

5 7.5%

Resistant

10

Farewell

14 20.9%

7 10.4%

21 31.3%

11 16.4%

13 19.4%

0 0%

0 0%

14

Dinner

7 10.4%

2 3.0%

24 35.8%

23 34.3%.

11 16.4%

0 0%

0

0%

*Not comprehensible.

In post hoc interviews, some nonnative speakers expressed a belief that they did not do well in responding to some situations. They attributed their shortcomings to a lack of real-life experience with some of the situations either because of their limited social roles (e.g., homemakers who had no experience in the business world) or because our situations lacked social reality in their native cultures (e.g., individuals who felt very uncomfortable with the situation involving the large loan). Although, in post hoc interviews, some native speakers expressed the same reservations, all were, nevertheless, willing and able to find the appropriate language to respond to the situations. Analysis showed that ease or difficulty in responding appropriately to particular items stemmed from linguistic factors as well as cultural ones. An analysis of poor responses indicated that nonnatives lacked the words and the syntax to work their way through culturally familiar and unfamiliar situations. Our data contained many examples of syntactic errors (e.g., "I consider change a job") and lexical problems including difficulties with intensifiers (e.g., "I very appreciate"), word order (e.g., "I'll pay back you"), idioms (e.g., "Thank you. Sound is good."), prepositions (e.g., "I hope to see you by us"), and choice of words (e.g., "It is so glad to me that I have such kind of good friend"). Table 3.3 shows the percentages of students whose responses were scored "Acceptable" and "Nativelike/Perfect." While Japanese speakers of English are generally thought to be very polite, they performed poorly on paper compared to the other groups in our study. Possible explanations for this include the lack of cultural congruity between Japanese and American cultures and the fact that written data do not reflect nonverbal cues and prosodic features which might serve to soften responses in real-life situations. The difficulty in cross-cultural communication between Japanese and English speakers is well documented in the research of Beebe and Takahashi (1989a,b). A revealing study by Coulmas (1981) shows that Japanese speakers have particular difficulty in English with those expressions of gratitude that

70

Speech Act Realization Table 3.3. "Acceptable" and "Native-like/Perfect" Ratings by Language Group Language background

Acceptable and nativelike/perfect (%)

Nativelike/perfect (%)

Japanese

30.2

14.8

Spanish

34.0

20.0

Korean

36.0

23.0

Chinese

48.5

28.1

Russian

66.6

58.3

imply indebtedness. In Japanese, Coulmas explains, an appropriate expression of gratitude acknowledges this debt by using the same expressions as would be required for an apology (e.g., Sore-wa kyoshuku desu, which literally means "I feel ashamed," can function as "thank you" or "I'm sorry" [89]). Brown and Levinson (1987) explain that, unlike England and the United States, Japan is a "debt-sensitive culture" in which thanks can be expressed, in effect, by saying, "I am humiliated, so awful is my debt" (247). By contrast, the relative success of the Russian participants in the study does not reflect the general American perception that Russians often express themselves inappropriately. Perhaps expressing gratitude is not one of the problem areas for them, or it is possible that judgments in this study reflect the limitations in the written channel, as might be the case with the Japanese participants. Interestingly, the Russian participants were most resistant in expressing themselves when they felt insecure or when they felt that the situations were not to be taken seriously. It may be that maintaining adequate demeanor played an important part in the behavior of these respondants. (See Eisenstein & Bodman, 1986, for more information on these data.)

Experiment 3 In our earlier work, discussed above, we noted that naturally occurring expressions of gratitude were different in some respects from our written data.2 In an attempt to include the influence of intonation and prosodic features in our data, we administered the questionnaire orally by reading each situation to ten native speakers and taping their responses. Transcripts were almost identical to the language of the written questionnaires. Some prosodic features seemed to add to the perceived sincerity and effectiveness of the gratitude expressed. Emotion was conveyed by tone as well as words. The range of emotion was more evident in the audiotapes. Gratitude expressed to socially distant individuals was clearly reserved in tone in comparison with effusive warmth and enthusiasm evidenced between friends and intimates. In post hoc interviews, the native speakers stated that they had found the process awkward and had ambivalent feelings about the naturalness of their responses. They felt that they would say much more in a natural situation. In fact, this was corroborated by the frequent uneasy pauses in their responses. It was as though the subjects

Expressing Gratitude in American English

71

were waiting for someone to reply to their initial expression of gratitude before they could continue. Dubin (personal communication) has noted similar difficulties with people leaving messages on telephone answering machines. This led us to postulate that expressing gratitude might not be merely a language function that existed as a response to a beneficial event, but might, instead, be a negotiated and interactive event that had greater social significance.

Experiment 4 In light of the importance to our subjects of interacting with the giver in order to express themselves fully and naturally in expressing gratitude, we used the same situations to set up role-plays that we hoped would reveal the interactive nature of how gratitude is expressed. In addition, during this period, we continued to collect naturally occurring examples. Some samples of role-plays and natural dialogues appear in Appendix 2. Our corpus consisted of 98 role-plays. Thirty-four were performed by native pairs, 40 by nonnative pairs and 24 by natives paired with nonnatives. The participants were literate, middle-class adults from a variety of social and ethnic backgrounds. Role-plays were recorded and transcribed for subsequent analysis. The procedures described in the earlier experiments were then followed. We found that the role-plays and natural situations incorporated the same words and semantic formulas that appeared in the written data. In this sense, our written data were representative of certain aspects of natural language use. For example, the following expressions occurred under all conditions in response to receiving a gift: It's beautiful. It's just what I needed. How did you know? You shouldn't have.

Yet the role-plays went beyond the written data by providing us with additional insights into the functions of thanking and responding to thanks as they exist in a conversational interaction. Similar findings comparing oral and written data have been reported by Beebe and Cummings (1985) and DeCapua and Dunham (1989). Just as role-play data gave us additional information through the inclusion of conversational turns, natural data was even richer and more revealing because it evidenced restatements of the same information using slightly different language and showed that the same function recurred over time. In analyzing the role-plays, we found that the language expressed by the giver (of the gift, favor, reward, or service) is crucial to enabling the receiver to convey gratitude successfully. The giver prompts and comments throughout the development of the speech act set. Prompts appear to function as linguistic enabling devices, allowing the receiver to reassure the giver of his or her gratitude. Examples of prompts are: 1 hope it fits. 1 hope you like it. How was your lunch . . . did you like it? Are you sure it's enough?

72

Speech Act Realization

Comments seem to keep the conversation moving by filling a turn-taking opportunity and giving the thanker additional cues on how to thank the giver. That'll keep you warm. It should be washable. Machine washable. We're going to miss you. It was nice having your.

There is evidence in our data that the act of giving may be perceived as disrupting the social equilibrium. Giving places the receivers in a position of obligation. The givers can mitigate the disequilibrium by downplaying the importance of the gift, favor, reward, or service (e.g., "Here's a little something I picked up for you"). Some givers also attempt to downplay the seriousness of the event by introducing humor. They can also reassure receivers by stating that the act of giving brings them pleasure and by claiming that the gesture has caused them no inconvenience. The giver can have an important role in reassuring the receiver that it is all right to take. Note the following interchange from natural data regarding the $5.00 loan: 1: Oh, no, I don't want to bother you. 2: Really, I have plenty. Believe me, you can have it.

Likewise, to restore the balance of the social relationship, the receiver needs to express the appropriate amount of appreciation to the giver. In our natural data, we noted that the receiver continued to thank the giver using a variety of strategies until the giver indicated that enough had been said. The giver often indicated his or her intent to end the thanking episode by abruptly nominating a change of topic. The nonnative speakers had difficulty replicating several aspects of the native model. Nonnative role-plays were often as much as fifty percent shorter than those of natives. In some instances, even the advanced nonnatives in our study lacked the language to express themselves more fully. (See Appendix 2, Dialogues 6 and 7.) This is in contrast to results reported by Blum-Kulka and Olshtain (1986). In attempting to find the right words, the advanced learners they studied said too much and were perceived as verbose. While our data showed that some individual turns of nonnatives were longer because of hesitations, cirumlocutions, and searches for the right words, the conversations taken as a whole were not long enough for natives to feel that gratitude had been adequately conveyed. Another aspect that nonnatives failed to replicate was the warm and sincere tone that natives conveyed when they expressed gratitude. In our first study, we questioned whether the nonnatives would be able to compensate for their lexical and syntactic difficulties by sounding and acting very appreciative. We found, unfortunately, that those students who lacked fluency and had to search continuously for words also spoke with an unconvincing tone. When native speakers were asked to judge these nonnative samples, they stated that the nonnatives sounded insincere. Those students who were more fluent, by contrast, had a more nativelike tone and sounded truly grateful. In our native data, there was a definite rhythm in the interactions of the interlocutors. In order for the gifts, services, rewards, or favors to be accepted successfully, a mutually agreed-upon "script" needed to be followed. As a result, there

Expressing Gratitude in American English

73

was more frequent communication breakdown when two nonnatives conversed than when one member of the pair was a native. The nonnatives' language flowed better when they had the support of a native speaker. When a favor or a gift needed to be repaid, native speakers often explicitly stated their intention to do so as a part of their thanks. For example, in expressing gratitude for the small loan, a considerable number of natives were very specific about repayment: "Thanks a million. I'll pay you back next Monday." In some instances the intention to repay was stated in more general terms: "We hope you'll come over to our place sometime soon." Nonnatives sometimes did not express reciprocity or did so in an inappropriate manner; e.g., "Thank you very much for dinner. You will come to our house next week?" In terms of the time mentioned, natives characterized this offer of reciprocity as being offered too specifically. They felt that the nonnatives were in too much of a hurry to pay them back. The overall offer, they felt, sounded too abrupt and demanding. Natives felt that they were being "put on the spot" to obligate themselves. In studying the groups and subgroups in our sample, it became apparent that cross-cultural differences in language values and customs all contributed to potential misunderstanding. These issues are addressed by Thomas (1983) in her discussion of pragmalinguistic and sociopragmatic failure. Furthermore, Blum-Kulka and Olshtain (1986) note that pragmatic principles are subject to both intercultural and intracultural variation. Some Anglo-American subjects said that they could not conceive of revealing to their friends that they were in need of money. Russian, Ukranian, Chinese, and German subjects stated that they, too, would find lack of money a source of shame and would have serious difficulty borrowing it from a friend. In receiving a gift, several groups indicated that they would probably put the gift aside and open it at a later time to avoid social embarrassment. Chinese participants explained that their cultural values demanded that an emphasis be placed on the presence of the guest rather than on a material object presented by the guest. Other Asian, Greek, and Anglo-American subjects indicated that they would probably say little about the gift and open it later when others were not present. However, some Americans and Hispanics informed us that it was customary for them to open a gift in front of the giver to share the enjoyment. Expressions of gratitude occur at the time of giving and, in our natural data, are found to reoccur later. For Anglo-American subjects, this reentry of thanks is as important as the original thanks, which would have to be performed out of politeness. A written note or phone call at a later date is not only highly valued but often expected. However, in traditional Jewish-American culture, important interactions tend to occur face to face; writing or calling are considered less meaningful. Our Polish and Puerto Rican subjects also preferred thanks to be conveyed in person. The cross-cultural complexity of this language function was also revealed by subjects who stated that expressions of gratitude offered in a context they considered inappropriate could be perceived as distancing, insulting, or rude. In many cultures, the words "thank you" are not commonly used to express appreciation to family members for acts of kindness considered part of their social roles. A Puerto Rican informant who had lived for many years in the United States described how hurt and angry her father became when she thanked him for helping her take care of her son, his grandchild. Her mother called berating her:

74

Speech Act Realization How could you have been so thoughtless? You thanked your father. He was happy to take care of Johnnie. Have you forgotten how to behave? He's your father and he loves you. How could you be so cold—to thank him?

Compliments relating to the father's kindness would have been better received. The opposite occurred in the case of an American family who welcomed a distant relative from Argentina into their home. After a month had passed during which the family had extended themselves personally and financially for their guest, they became increasingly annoyed that the relative rarely expressed appreciation for their efforts with the exception of an occasional terse "thank you." In Argentina, favors for family members are expected as part of the warmth and bonding of the family unit. While this is also true in American culture, the explicit statement of gratitude is required after each event. In many cultures it is felt that gratitude and the intent to reciprocate need not be expressed overtly. In fact, it might be considered insulting to do so. Among friends and relatives, reciprocity is taken for granted. In the Republic of Croatia in Yugoslavia, for example, thanking behavior must be judicious. If people are thanked excessively (more times than necessary or with inappropriate length or enthusiasm), this behavior is interpreted not as expressing gratitude but as an indirect solicitation for more gifts or favors. Another area of cross-cultural difference arose from the high value placed on modesty and humility by several cultures in our sample. A Chinese informant indicated that modesty is required in responding to the offer of a raise: Thank you very much. But I think I have not done so well to get a raise. Anyway, I'd try to do better. Also, as mentioned above, apologizing appears to be associated with expressions of gratitude in some cultures. A Japanese participant responded to the offer of the raise: I'm sorry. 1 will try harder in the future. Another Japanese, in response to the $500 loan, said: I'm sorry. I'll always remember the debt of gratitude. Americans find these kinds of utterances difficult to interpret and find them uncomfortable and confusing.

Conclusion Our data show that thanking is a speech act that is mutually developed. It can involve a complex series of interactions and encodes cultural values and customs. Both the giver and the thanker collaborate in the development of a successful thanking episode. Because expressing gratitude involves "face wants" in the sense that the speaker expresses both approval of and indebtedness to the receiver of thanks, the feelings of the interlocutors are closely tied to this interaction. Our study and those of others reveal that, while thanking appears to be a universal function, it is accomplished differently in contrasting cultures, whose values may focus differentially on the various components that constitute expressions of gratitude.

Expressing Gratitude in American English

75

In methodological terms, our studies reveal that, while written and oral questionnaire data miiTor the words and expressions used in conveying gratitude, roleplays reveal the interactive aspects of the function more fully. Although natural data are most indicative of the whole process, role-plays have the advantage of providing enough control to allow meaningful comparisons to be made. An integration of all four kinds of data was effective in helping us understand how native and nonnatives express gratitude in English. Finally, it was evident that even advanced learners of English have considerable difficulty adequately expressing gratitude. They need information on the nature of what to say, the language used to express it, and the context in which it is needed, "rom a pedagogical perspective, role-plays can serve not only as a diagnostic tool for teachers, but can also provide much needed practice for learners. Ideally, with their heightened awareness, learners can also observe outside the classroom the use of pragmatic functions in social interaction and vehicles that depict such interaction (e.g., dramas, soap operas). It may also be helpful for students to compare English models to their own speech in order to enhance their awareness of the pragmalinguistic and sociopragmatic rules of English.

Appendix 1: The Questionnaire Please read the following short descriptions of situations in which you might find yourself. Think of what you might say in response to this situation. Write your response (if any) in the space provided. Say as much or as little as you wish—you may choose to say nothing in several circumstances. 1. It's Friday. You look in your wallet and notice that you only have $2.00. Your good friend at work notices this and hears you say, 'Darn, I'll have to go to the bank.' Your friend asks if you need money, and you say that you forgot to go to the bank. Your friend says, 'I have plenty. How much do you need?' You say, 'Could you lend me $5.00? I'll pay you back on Monday.' Your friend says, 'Sure. Are you sure you don't need more than that?' You say you don't. Your friend gives you the $5.00. 2. You board the bus, pay your money and take a seat near the front of the bus. Just before your stop, you signal the driver to stop. You move to the front, the bus comes to a stop, and the doors open. 3. It's your birthday, and you're having a few people over for dinner. A friend brings you a present. You unwrap it and find a blue sweater. 4. You work for a large company. The Vice-President of Personnel calls you into his office. He tells you to sit down. You feel a little nervous, because you have only been working there for six months. The Vice-President says, 'You're doing a good job. In fact, we are so pleased with you that I'm going to give you a $20.00 a week raise.' 5. In the supermarket, the cashier puts your groceries in bags and turns to begin checking out the next customer. You pick up your bags to leave. 6. At the table in a restaurant a friend says, 'You have something on your face.' You ask where. Your friend tells you. You rub your face and ask, 'Is it off?" Your friend says that it is. 7. You find yourself in sudden need of money—$500.00. You mention this to a friend.

76

Speech Act Realization

8.

9.

10. 11. 12.

13. 14.

Your friend immediately offers to lend it to you. You are surprised and very grateful. Your friend writes out a check for $500.00 and gives it to you. At first you say, 'Oh no, I didn't mean for you to lend it to me. I couldn't take it.' Your friend says, 'Really, it's all right. What are friends for?' After your friend insists again, you take the check. You are married. Both you and your spouse work. You come home late from work and find that your spouse has done some work around the house that you had promised to do, but had not had a chance to do. Your friend suggests going out to lunch. You say that you'd like to go, but you only have $2.00. Your friend says, 'Ah, don't worry. I'll take you today.' Your friend takes you to a very nice restaurant—a much more expensive one than the ones you usually go to. You have a wonderful meal. Your friend pays, and you get up to leave. You have just gotten a new and better job. A friend at the office tells you she has organized a farewell party for you. You have just gotten your hair cut in a new style, and you like it better than the old way. Your friend sees you and says, 'Hey, you've got a new haircut. It looks nice.' You are sharing an apartment with a friend. You're both sitting and relaxing in the living room. You ask your friend to hand you the newspaper which is nearby. Your friend gives you the newspaper. You pick up your car in a parking garage. As the attendant who drove up your car walks past you to get the next person's car, you hand him a tip. You have been invited to the home of a rather new friend. You have dinner with him and his wife and a few other friends of theirs. The food was great, and you really enjoyed the evening. As you leave, your hosts accompany you to the door.

Appendix 2 Dialogue 1

(A role-play between two native speakers)

WOMAN 1: I've got something for you. WOMAN 2: You do? (takes package and unwraps) Oh, my goodness. Why did you do this? . . . a blue sweater, (giggles) How did you know this is just what I needed? It's wonderful. Thank you. I'm so surprised! WOMAN 1: (laughing) Price tag in it? WOMAN 2: I can really use this . . . WOMAN 1: Good. I hope it fits. WOMAN 2: Oh, I'm sure it will. Thank you. (goes to try it on) It's medium, I'm sure it will. TA-DA! WOMAN 1: (laughing) Yes, it does fit nicely. WOMAN 2: Yes, it's just what I needed. That's really nice. WOMAN 1: Wear it in good health. WOMAN 2: Thank you. WOMAN 1: So, how's everything? (topic changes.) Dialogue 2

(A role-play between two native speakers)

WOMAN 1: I've only got $2.00. What am I going to do now? I guess I'll have to go to the bank.

Expressing Gratitude in American English WOMAN WOMAN WOMAN WOMAN WOMAN WOMAN WOMAN Dialogue 3

2: 1: 2: 1: 2: 1: 2:

77

Oh, you don't have to go to the bank. I have money. I can lend you some. Oh no, I don't want to bother you. Thanks. Really I have plenty. Believe me, you can have it. Can you lend me $5.00, do you think? Yeah. Sure. Here. Oh, I really appreciate this. Thanks. You're welcome.

(Natural data—a conversation between two native speakers)

MOTHER: Another box, huh? Oooo . . . what did you buy? Oh, I shouldn't . . . I thought when she went out I should tell her not to buy one thing. DAUGHTER: Us. This is for you. MOTHER: What? DAUGHTER: And this one is for me. MOTHER: This is for me? Ohhhhh. DAUGHTER: That's for you. MOTHER: Ohhh, Jeanie, what'd you do that for now? DAUGHTER: (giggles) MOTHER: That's too much, (opening the gift box) Oooo! Oh, my, that's lovely. Lovely. Oh my, Jeanie, that's pretty. DAUGHTER: Open it up. MOTHER: Ooo. Beautiful. Oh, yes. DAUGHTER: See? . . . That'll keep you warm. MOTHER: Yes, I should say so. That's beautiful. DAUGHTER: Good. I'm glad you like it. It should be washable. Machine washable. MOTHER: Is it? DAUGHTER: I think so. Just toss it in the machine. MOTHER: Ohh. Oh, my, that's lovely. Thank you. DAUGHTER: Oh. you're welcome, (they kiss and laugh.) What are we having for dinner? Dialogue 4

(Natural data—a conversation between two native speakers)

MOTHER: DAUGHTER: MOTHER: DAUGHTER: MOTHER: DAUGHTER: MOTHER: DAUGHTER: MOTHER: DAUGHTER: Dialogue 5 AUNT: UNCLE: ERIK: UNCLE: ERIK:

(looking in a gift box of assorted pastries) Oooo. Oo. Oo. Oo. You've got a choice. Ohhh. You spend too much money. No. Oh, dear. No. Yes. No, it was fine. You shouldn't. I'll have to pay you. No, no, no. (switches topic.)

(Natural data—A conversation among native speakers, a 12-year-old boy, two men, and two women) Why don't you give Erik his? Erik? Here's . . . Oh, thank you. Your Christmas present. Thank you. (opening present) Oh! Excellent!

Speech Act Realization

78

AUNT: Do you like that? ERIK: I ... I've always wanted a good sweater. We were going to get one. Now I don't need one. Excellent, thanks. AUNT: Good. UNCLE: Good. AUNT: I don't know. It may be a little big. DAD: Ooo. That looks great. MOM: Oo-la-la! DAD: Hey, Joe College. MOM: Ooo. And that's a real wool one. Aww. That looks like an Eddie Bauer. ERIK: Thank you. MOM: Is it wool? AUNT: Yeah. It may be a little big. But I figure that . . . Erik . . . any minute now, it'll fit. MOM: Yeah. Oh ... handsome! AUNT: It's a little big. MOM: He was just talking about how he wanted some good-looking sweaters to wear to school. That's what the kids are wearing. Oh, boy. AUNT: (looking at the television) Hey, look, the Superbowl's coming on. Dialogue 6

(A role-play between a Turkish man and a Brazilian woman)

MAN: This is very kind of you inviting . . . eh ... to you birthday . . . eh ... party . . . eh ... may I introduce you . . . eh ... angora sweater if you mind . . . eh ... if you like a sweater. WOMAN: Ay! Are you give me a sweater . . . a blue sweater? MAN: A blue sweater . . . yes. WOMAN: Ah ... Oh, that's wonderful. I really do like it. MAN: I'm happy you like it. WOMAN: I will. And . . . eh ... MAN: That's fine. And . . . eh ... hoping to celebrate you next birthday all together again. WOMAN: And so you can give me another sweater! (laughing) This time red, okay? MAN: Yes, okay. Dialogue 7 WOMAN: MAN: WOMAN: MAN: WOMAN: MAN:

(A role-play between a French man and a Malaysian Chinese woman)

I know today's a special day. I brought something for you. Why is this a special day? Is a surprise? It's your birthday. Oh, is birthday. Are you sure? How did you know that? (long pause) I'm really surprised, but . . . uh ... I'm really at ease, but thank you very much because that's very nice and I didn't know you learned it was my birthday. I appreciate it. WOMAN: Don't mention it. (gives him the present) MAN: Wow! My God! But I like! Ah ... that's . . . oh ... I like that. Thank you very much. WOMAN: Don't mention it.

Expressing Gratitude in American English Dialogue 8

79

(A role-play between a Chinese woman and a Russian woman)

WOMAN 1: Since you have been work here . . . uh ... very good and very diligent. I like you very much. From next week I will raise your salary $15.00 every week. Would you like this? WOMAN 2: Yes. Thanks a lot. Thanks a lot. WOMAN 1: Don't mention it. . . . And . . . uh ... that's all. Would you please go back to your work? Dialogue 9 1: 2: 1: 2: 1: 2: 1: 2:

(A role-play between native English speakers)

How about going out to lunch this afternoon? Are you free? I'm free, but I ... I don't know. 1 just didn't take enough with me. . . Well, don't worry about it, I'll treat you. No, I don't want you to do that. It's okay. It's my pleasure. (silence) Next time you can treat me. Okay? Well, all right.

Dialogue 10

(A role-play between a Spanish man and a Hispanic woman)

WOMAN: Joe, 1 have a little problem. 1 just realize that I don't have any money with me and the banks are closed now. Do you think is possible for you to lend me at least ten dollars for the weekend? MAN: Ten dollars? Do you need that money for what? To eat or . . . ? WOMAN: No . . . just to have some money with me for the weekend in case any emergency comes up ... MAN: Yes, I can. I can borrow . . . WOMAN: I can lend you . . . MAN: I can lend you, I can lend you ten dollars, but don't forget next Thursday . . . when you . . . Yes, when you get paid, give me back ten dollars and if you want to give me some interests . . . well . . . WOMAN: (laughing) Oh ... thank you very much. I really appreciate that.

Appendix 3 The Role of the Giver Offer Here. I'll take you. How much do you need? Downplay the offer Here's a little something. Really, I have plenty. I hate to eat alone. Comment It's washable.

Role of the Thanker Hesitate Oh, I couldn't. Are you sure it's all right? Oh, come on. No! Express surprise For me? You're kidding! Compliment It's lovely. This is nice.

80

Speech Act Realization Prompt I hope it fits. Do you like it? Did you have a good time? Reassure Really, it's all right. What are friends for? React Oh, good. I'm glad you like it. It was nice to see you. Close Well, good. Don't worry about it. It is/was my pleasure. Don't mention it. (Change topic)

You're so nice. You're the best husband/wife in the whole world. Offer or suggestion of repayment I'll pay you back on ... Next time, I'll take you. I hope we can do this again sometime (soon). If there is anything I can ever do for you . . . Re-thank I love the sweater you gave me. That was such a nice evening. Thanks again for helping me out. I don't know what I would have done without you.

Notes 1. We noted that utterances containing words such as "thank you" or "thanks a lot" did not always express gratitude. For example, in the exchange: 1: That haircut doesn't do a thing for you. 2: Thanks a lot! You really know how to make me feel good!

the illocutionary force of "thanks" is irritation rather than gratitude. 2. Further information on our findings regarding cross-linguistic and cross-cultural elements in the written data will be integrated with our discussion on role-plays below since many of the same elements appeared in the written data as well as the oral data.

References Beebe, L. M., & Cummings, M. C. (1985). Speech act performance: A function of the data collection procedure? Paper presented at the Sixth Annual Sociolinguistics Colloquium, TESOL Convention, New York. Beebe, L. M., & Takahashi, T. (1989a). Do you have a bag? Social status and patterned variation in second language acquisition. In S. Gass, C. Madden, D. Preston & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (103-125). Clevedon and Philadelphia: Multilingual Matters. Beebe, L. M., & Takahashi, T. (1989b). Sociolinguistic variation in face-threatening speech acts: Chastisement and disagreement. In M. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in SL variation (199-218). New York: Plenum. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Studies in Second Language Acquisition, 8, 165-79. Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Cohen, A., & Olshtain, E. (1981). Developing a measure of sociocultural competence: The case of apology. Language Learning, 31, 113-34.

Expressing Gratitude in American English

81

Coulmas, F. (1981). Poison to your soul: Thanks and apologies contrastively viewed. In F. Coulmas (Ed.), Conversational routine (66-91). The Hague: Mouton. DeCapua, A., & Dunham, J. F. (1989). Interaction strategies in the requesting and giving of advice. Paper presented at the Second Language Acquisition Circle, October 13, 1989, New York University, New York. Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by native and non-native speakers of American English. Applied Linguistics, 7, 167-85. Goffman, E. (1967). Interaction rituals. New York: Pantheon Books. Hymes, D. H. (1971). Sociolinguistics and the ethnography of speaking. In E. Ardener (Ed.), Social anthropology and language (47-93). London: Tavistock. Leech, G. N. (1983). Principles of pragmatics. London: Longman. Rubin, J. (1983). The use of "Thank you." Paper presented at the Sociolinguistics Colloquium, TESOL Convention, Toronto, Canada. Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112. van Ek, J. A. (1976). The threshold level for modern language learning in schools. London: Longman. Wolfson, N. (1989). The social dynamics of native and nonnative variation in complimenting behavior. In M. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in SL variation (219-36). New York: Plenum.

4 Perception and Performance in Native and Nonnative Apology MARC L. BERGMAN and GABRIELE KASPER "I'm sorry, so sorry, Please accept my apology . . . " BRENDA LEE, "I'm Sorry" (1958)

In terms of their actional characteristics, apologies relate in important ways to other frequently studied expressive speech acts. Like the speech acts of thanking, complimenting, and complaining, apologies occur post-event. Complimenting and thanking involve events deemed praiseworthy by prevailing social norms; complaints and apologies refer back to events that constitute norm infringements. By complimenting and thanking, an event is made into a praiseworthy occasion; by complaining about or apologizing for its occurrence, an event is made into a transgression. In terms of value attribution, the relationship between event and speech act is reflexive rather than unilateral. Complaints, thanks, and compliments commonly hold the addressee to be causally involved in the preceding event; the agent assuming responsibility for the event necessitating the apology is the speaker. Compliments differ from thanks in that thanks require that the addressee's preceding action be beneficial to the speaker; compliments refer to addressee-related events, which do not need to be beneficial to anybody in particular. The actional descriptors referred to above allow us to distinguish compliments, thanks, complaints, and apologies schematically (Table 4.1). Apologies can be defined as compensatory action to an offense in the doing of which S was causally involved and which is costly to H. This conceptualization is supported by Goffman's (1971) view of apologies as remedial interchanges, remedial work serving to reestablish social harmony after a real or virtual offense. Following Goffman's distinction of ritual and substantive compensation, apologies can be classified into (1) those redressing virtual offenses, which are remedied by the sole offering of an apologetic formula, and (2) those redressing actual damage inflicted on the addressee, sometimes including an offer of material compensation. Both kinds of apology have been demonstrated to vary cross-culturally. 82

83

Apology, Perception, and Performance Table 4.1. Actional Features of Four Expressives

Event

e-Evaluation*

e-Actor

e-Recipient

Compliment

post

good

H

H

Thanking

post

good

H

S

Complaint

post

bad

H

S

Apology

post

bad

S

H

*e = event.

Ritualistic apologies are sometimes distinguished from substantive ones by different formulae. Thus in (American) English, ritualistic apologies with the formula "excuse me" are offered as territory invasion signals when addressing strangers (e.g., prior to asking direction), as announcements of temporary absence from ongoing interaction (e.g., in order to answer the phone), or upon virtual or real intrusion of another person's physical space (e.g., passing somebody in a narrow hallway). Borkin and Reinhart (1978) define the function of "excuse me" as "a formula to remedy a past or immediately forthcoming breach of etiquette or other light infraction of a social rule" (61). "I'm sorry," in their analysis, is used in a wider range of contexts, especially "in remedial interchanges when a speaker's main concern is about a violation of another person's right or damage to another person's feelings" (61). Borkin and Reinhart's analysis is given strong empirical support by House (1988). She found that in seven apology contexts involving substantive offenses, native speakers of British English (n = 100) used "I'm sorry" up to 80% of the time. In six of these contexts, "excuse me" did not figure at all, in one of them it was used with negligible frequency (3%). The ritualistic function of much apologetic behavior, and its cross-cultural variability, has also been noted by Coulmas (1981) in his analysis of expressions of gratitude and indebtedness in a number of Western languages as opposed to Japanese. Coulmas notes that in many contexts requiring expressions of gratitude in Western cultures, such as upon receiving a gift, Japanese requires an apologetic formula such as "sumimasen." The function he ascribes to ritualistic apologies in Japanese concurs with the functional properties of "excuse me" described by Borkin and Reinhart (1978); namely, indicating "the speaker's willingness to conform to conventional rules and social expectations. . . . Verbal apology occurs even if there was no serious or real offence as a precaution against inadvert misconduct or unanticipated negative interpretation of one's performance" (Coulmas, 1981,84). While Borkin and Reinhart's (1978) analysis suggests that acquiring appropriate formulas for ritualistic apology is problematic for nonnative speakers (NNS), substantive apologies confront the learner with a more complex learning task. First (as with ritualistic apologies), she has to identify the occurrence of an event that requires apology. This may require restructuring her cultural knowledge as such events have been shown to vary cross-culturally (Olshtain, 1983). Second, the severity of the offense and the weights of contextual variables such as power and distance need to be assessed—another potential trap as perceptions of these social variables are also subject to cross-cultural variation (House, 1988; Vollmer & Olshtain, 1989). Finally, appropriate output strategies have to be selected. While all

84

Speech Act Realization

available evidence points to a universally valid apology speech act set (Olshtain, 1989), preferences for strategy choice are contextually and cross-culturally at variance (Olshtain, 1989). As Garcia (1989) has demonstrated, cultural differences do not only obtain in preferences for local strategic choices but for global approaches to the speech event, such as opting for a deference versus a solidarity style. While we do not believe that language users actually go through the motions of planning linguistic action in the serial fashion outlined above (Schmidt, Chapter 1), appropriate apologizing requires the specified knowledge components, and the language user's ability to access this knowledge fast and flexibly. In interlanguage (IL) pragmatics, only a few studies have addressed the issue of how NNS apologize in ongoing interaction with a native (NS) or NNS interlocutor, and the available studies (Rasper, 1981; Trosborg, 1987; Garcia, 1989) have analyzed NNS's performance in terms of their pragmatic knowledge (what semantic formulas do they use) rather than examining how pragmatic knowledge is accessed under immediate processing constraints. Studies of IL apologizing have essentially addressed the same research question—the accessibility of apology strategies to NNS—yet they have examined this issue by means of different data-gathering procedures. Rintell and Mitchell (1989) compared NS and NNS use of apology strategies in their responses to written and oral Discourse Completion questionnaires and found only slight differences between the two conditions. In order to review previous substantive findings and assess further instrument effects, the results of five studies are summarized in Table 4.2. Holmes (1989) investigated the realization of apologies in an ethnographically collected corpus of 183 remedial interchanges produced by adult NS of New Zealand English (NS NZE). Hers is not an interlanguage study but has been included because it allows us to compare elicited data with naturally occurring speech. Olshtain (1983) examined apology performance in role-plays, enacted by NS of American English (NS AE), Russian (NS Rus), and Hebrew (NS Heb), and by American and Russian learners of Hebrew (IL E-Heb, IL Rus-Heb). Role-play was also employed by Trosborg (1987) in a study involving NS of British English (NS BE), NS of Danish (NS Dan), and three groups of Danish learners of English at different levels of proficiency. Table 4.2 includes the figures for the lower advanced level (IL Dan-En). House (1988) used a Discourse Completion Task (DCT) to study apology realization by NS of British English (BE NS), NS of German (NS Ger), and German learners of English (IL Ger-En). By means of the same DCT questionnaire, Kasper (1989) looked at the apology responses provided by NS of Danish (NS Dan) in comparison with Danish learners of English (IL Dan-En) and Danish learners of German (IL Dan-Ger). Throughout the five studies, apology realizations were coded according to the semantic formulae identified as constituting the apology speech act set (Olshtain & Cohen, 1983; Blum-Kulka, House & Kasper, 1989). Figures reported in the studies have been summarized according to these categories. Most subjects apologized explicitly by means of an Elocutionary Force Indicating Device (IFID) such as "I'm sorry" and stated whether they assumed responsibility for the offense (responsibility): e.g., by blaming a third party ("the kitchen must have mixed up the orders"), admitting the offensive act ("I forgot to bring your

85

Apology, Perception, and Performance Table 4.2. Selection of Apology Strategies (%) by Fifteen Groups of Informants Account

Minimization

Repair

Verbal redress

9

23

NR

5

2

NR NR NR NR NR

100 58 92 71

42 33 33 23 21

NR NR NR NR NR

0 0 8 0 0

0 0 0 0 7

7 6

NR NR

24 33 41

22 18 13

21 25 14

22 12 21

5 6 2

13 11 16

15 13 14

9 4 8

9 16 12

13 16 23

8 5 1

IFID

Intensifier

ethno. (Holmes) NS NZE

61

NR

RP (Olshtain) NS AE NS Rus NS Heb IL E-Heb IL Rus-Heb

92 66 66 69 64

RP (Trosborg) NS BE NS Dan IL Dan-E

Responsibility

75

8

NR

DCT (House) NS BE NS Ger IL Ger-E

80 69 73

40 31 48

70

80

3 5 6

DCT (Kasper) NS Dan IL Dan-E IL Dan-Ger

72 75 68

22

49 82 75

18 5 10

27

15

66

book"), or by self-blame ("it's my fault"). The extremely low frequencies of IFID provided by the British-English and Danish speakers in Trosborg's role-plays do not seem to be the result of a language or instrument effect since high frequencies of IFID were supplied by speakers of the same languages in Kasper's (1989) DCT study and by Olshtain's (1983) role-players. This suggests differential contextual effects in Trosborg's data as opposed to the other studies. Intensification of apology ("Im very/terribly sorry") was only reported in the DCT studies, where the English NS and German-English IL users upgraded their apologies more than the Danes and the German NS. Except for the Danish NS, who provided the same amount of explanations or accounts for the offense in the role-play and DCT, the interactive conditions elicited considerably more accounts than the noninteractive DCT. This instrument effect may be indicative of different psychological and discoursestructural conditions. Subjects may feel more psychologically pressed to account for purported misdeeds in a direct encounter with the offended party, and in ongoing discourse where they can distribute various apologetic acts over several turns. Accounts may even be requested by the interlocutor, a possibility precluded by the one-turn response in a DCT. The British-English and Danish NS tended to minimize the severity of the offense (e.g., "it's only a tiny scratch (on your car)") more in the role-plays than in the DCTs, which suggests a similar effect as in the case of accounts. The greater effort of these speakers to minimize the offense is consistent with their reluctance to explicitly apologize and take on responsibility. Repair, or compensation for the incurred damage ("I'll bring it in tomorrow"), was offered with about the same frequencies as minimization. Olshtain's (1983) role-plays and

86

Speech Act Realization

Holmes' (1989) ethnographic data provided very few instances of repair. Expressions of concern for the hearer ("I hope you didn't wait long") or promises of forebearance ("it won't happen again"), conflated as "verbal redress," were used extremely rarely by all groups across conditions. For the majority of NS and IL users, expressing an explicit apology and making a responsibility statement were the essential components of apology, whereas providing explanations, minimizing the offense, and offering repair and verbal redress were optional and, as demonstrated by Olshtain (1989), context-dependent strategies. This summary thus extends Olshtain's findings (1983, 1989) to a wider range of languages. The variation in supplied apology strategies raises a number of questions. For instance, why do Holmes's (1989) informants make considerably fewer responsibility statements than any other group? Our supposition is that this is due to the type of apology included in Holmes's corpus. Even though Holmes does not comment on this, her offense categories "space" and "talk" offense in particular suggest that she included ritualistic apologies in her corpus. The apologetic formulas by which ritualistic apology is performed very often do not co-occur with other redressive activity. As the role-plays and DCT questionnaires exclusively provide contexts for substantive apologies, the naturalistic data set does not provide an adequate baseline for the two types of elicited data. Furthermore, the conspicuous difference between Trosborg's (1987) role-plays and the remaining elicited data suggests that close attention be given to instrument effects in the study of apology. Research to date has examined apology behavior in a variety of Western cultures and languages. While these studies have been important in providing preliminary evidence for a universally valid apology speech act set, and the differential selections from this set according to contextual factors, it is requisite to extend the scope of study to non-Western languages and cultures to advance the fundamental issue in cross-cultural pragmatics; namely, the universality and specificity of linguistic action. One of the methodological problems in cross-cultural and cross-linguistic comparison is to determine whether the contextual conditions in which the speech act behavior under study occurs are perceived as the same or different by the groups to be compared (Arndt & Janney, in press; Blum-Kulka & House, 1989). With specific reference to apologies, Wolfson, Marmor, and Jones (1989) comment that "a cross-linguistic study of apologies may well reveal that the notions of offense and obligation are culture-specific and must, therefore, become an object of study in themselves" (180). Brown and Levinson's (1987) politeness theory predicts that the weightiness of face-threatening acts (FTAs), computed by adding the values of social distance, dominance, and degree of imposition as perceived by actors in a given context, determines the kind and amount of redress afforded in the performance of FTAs (76). According to theory and empirical evidence (Kasper, 1990 for overview), both the weights and values of the contextual factors vary cross-culturally. In the cross-cultural study of apology, it is therefore essential to establish what constitutes an offense, how members of different cultures perceive offense contexts, and how these perceptions are reflected in output strategies.

Apology, Perception, and Performance

87

The Present Study The central research questions we wish to examine in this study are: (1) How are contextual factors in a variety of offense contexts perceived by Thai and American informants? (2) How is the selection of apology strategies determined by contextual factors? and (3) What patterns of intracultural and intercultural variability are observable in the selection of apology strategies by Thai NNS of English as compared to NS of Thai and American English? Informants in this study were 423 Thai graduate students at Chulalongkorn University, Bangkok, at an intermediate level of proficiency in English, and 30 NS of American English who were students at the University of Hawaii at Manoa. Data were collected by means of two questionnaires. The Assessment questionnaire included 20 items, each of which specified a different offense context (see Appendix 1 for sample item). Informants were asked to rate these contexts on a 5-point rating scale for 4 context-internal factors (Severity of offense, offender's Obligation to apologize, Likelihood for the apology to be accepted, offender's Face-loss) and 2 context-external factors (social Distance and Dominance) (for the distinction between context-internal and context-external factors, see Brown & Fraser, 1979; Blum-Kulka & House, 1989). The selection of the context-external factors Distance and Dominance (power) follows Brown and Levinson's (1987) weightiness formula (76) and their argument that Distance and Dominance constitute "very general pancultural dimensions which nevertheless probably have 'emic' correlates" (76). The context-internal factors are assumed to function as components of Brown and Levinson's dimension "degree of imposition," specified for the speech event of apologizing. Severity of offense and Obligation to apologize were found to distinguish different types of offense by Olshtain (1989). Likelihood for the apology to be accepted was added in view of an analysis of apology responses; however, it is included in the present study to examine possible interdependence between it and other contextual factors, and effects on the selection of apology strategies. Offender's Face-loss was included since, as Vollmer and Olshtain (1989) note, "the expression of responsibility seems to be linked directly to the S's cost and loss of face which results from performing the speech act of apologizing" (198), and which consequently can be expected to have an impact on the redress afforded in carrying out an apology. The Dialog Construction (DC) questionnaire included the same offense contexts as the Assessment questionnaire. Informants were asked to supply the offender's and the offended person's turn (see Appendix 2 for sample item). Since the apology response was elicited from the informants rather than being provided as part of the questionnaire items, the instrument differed from the standard Discourse Completion questionnaire (Blum-Kulka, House & Kasper, 1989). For this study, only the first pair parts were analyzed. The two questionnaires included the following contexts: 1. A and B are friends. A has had an accident with a car borrowed from B. (Damaged Car)

88

Speech Act Realization

2. A and B are friends. A borrowed a magazine from B and poured coffee over it. (Ruined Magazine) 3. At a staff meeting, teacher A contradicts teacher B. (Contradiction) 4. At a staff meeting, teacher A accuses teacher B of being a poor teacher. (Poor Teacher) 5. At an office, a junior colleague forgets to pass on a private message to a senior colleague. (Private Message Low-High) 6. At an office, a senior colleague forgets to pass on a private message to a junior colleague. (Private Message High-Low) 7. At an office, a junior colleague forgets to pass on an important business message to a senior colleague. (Business Message Low-High) 8. At a office, a senior colleague forgets to pass on an important business message to a junior colleague. (Business Message High-Low) 9. At a restaurant, a customer changes her mind after the order has already been taken. (Order Change) 10. At a restaurant, a waiter spills food on a customer's clothes. (Food on Customer) 11. At a restaurant, a waiter brings the wrong order. (Wrong Order) 12. At a restaurant, a customer spills food on a waiter. (Food on Waiter) 13. At Bangkok Airport, a customs official messes up a traveler's suitcase. (Messed-up Bag) 14. At Bangkok Airport, a traveler is caught trying to smuggle a Buddha out of the country. (Smuggled Buddha) 15. At Bangkok Airport, a customs official breaks a legally purchased statue when searching a traveler's suitcase. (Broken Status) 16. At Bangkok Airport, a traveler is unable to produce a customs form. (Customs Form) 17. A professor has not yet graded a term paper that a student was supposed to pick up. (Ungraded Paper) 18. A student forgets a book she was supposed to return to her professor. (Borrowed Book) 19. A professor misplaces a student's term paper and fails the student. (Failed Student) 20. A student plagiarizes from a published book and is found out by a professor. (Cheating Student) Items in both questionnaires were randomized. The Assessment questionnaire was filled out by 30 American NS of English and by 30 Thai NNS of English. Both versions were in English; however, for the Thai study, conceivably difficult vocabulary items were glossed in Thai. The DC questionnaire was prepared in an English and in a Thai version. The English version was filled out by 30 NS of English and by 288 Thai NNS of English. The Thai version was filled out by 136 Thai NS.

The Assessment Study The mean ratings and standard deviations of contextual factors in each offense situation are included in Appendix 3. Calculation of Pearson Product Moment

Apology, Perception, and Performance

89

Coefficients revealed no statistically significant correlation between the contextexternal factors Distance and Dominance, nor between any of the context-external and the context-internal factors. This finding lends support to Brown and Levinson's (1987) thesis of the mutual independence of contextual factors (80f.). However, we hesitate to conclude that such independence would continue to obtain in ongoing interaction. While a static, snapshotlike view of an interactional configuration, as represented by the questionnaire items, suggests the integrity of Distance and Dominance relative to each other and to the interactional content, conversational interchanges where apologies are jointly negotiated may well evidence a dynamic and interdependent fluctuation of participants' ascriptions of Distance and Dominance (e.g., Aronsson & Satterlund-Larsson, 1987), and of these context-external with context-internal factors. The ratings of context-internal factors by both Thai and American informants yielded high correlations between Severity and Obligation (American r = .7210, p < .000, Thai r = .6815, p < .001), Severity and Likelihood (American r = - .6382, p < .002, Thai r = - .7505, p < .000), Severity and Face-loss (American r = .7564, p < .000, Thai r = .5848, p < .007), and Obligation and Face-loss (American r = .8528, p < .000, Thai r = .7401, p < .000). Overall, then, it can be concluded that Severity of offense is systematically related to the offender's Obligation to apologize (corroborating Olshtain's (1989) finding about the covariance of Severity and Obligation [160]), the Likelihood for the apology to be accepted—the more serious the infringement, the less the likelihood for the apology to be successful—and the extent to which the offender's (positive) Face is adversely affected. Furthermore, Face-loss and Obligation can be viewed as codeterminant: more Face-loss requires more Obligation to apologize, and more Obligation entails more Face-loss. The demonstrated interrelatedness of context-internal factors can be understood as an explication of Brown and Levinson's (1987) contextual dimension "degree of imposition" (76) for the speech act of apology. What is noteworthy is the strong congruence by which Thai and American informants perceived the relationships between the context-internal factors, and the lack of interrelation expressed in the ratings of the context-external factors. However, rather than drawing hasty conclusions about congruent social perceptions by members of widely different cultures, we feel that two caveats are in order. First, the Thai questionnaire was administered in an English version. The assessments given by the Thai informants may therefore represent their perceptions of offense contexts in English-speaking cultures rather than in their native culture. Their approximation of the American ratings might be taken as an expression of intercultural competence, a construct parallel to interlanguage competence. To what extent the Thais' intercultural perceptions are reflective of, or different from, their social perceptions in their native culture can only be examined by comparative Thai material, which we unfortunately were unable to collect. Second, the noted correlations obtained across the 20 contexts represented in the questionnaire. They do not allow predictions of the values assigned to contextual factors in individual contexts. In the remainder of this section, we shall take a closer look at the contextual assessments of different offense situations. Statistically signif-

90

Speech Act Realization

icant differences between the Thai and American ratings were determined by multivariate analysis (MANOVA). In order to facilitate the following presentation, we shall divide the continuous ratings into three categories: low (1-2.3), medium (2.43.7), and high (3.8-5).

SEVERITY OF OFFENSE Only Order Change and (according to the American raters) Borrowed Book were perceived as light offenses. Medium Severity was assigned to offenses involving low material costs (Ruined Magazine), inconveniences rather than infringements of legal rights or entitlements (Ungraded Paper, Contradiction, Messed-up Bag, Private Messages), or mishaps occurring as part of a job (Food on Waiter, Wrong Order). These offenses represent minor impositions on somebody's time, money, physical space, energy, or face-wants; they warrant no legal claims to redress. High Severity offenses have major real-life consequences rather than involving easily repairable inconveniences. They may constitute illegal action (Cheating Student), high material costs (Broken Statue), potential or real obstruction of regular procedures and negligence of professional obligations (Business Messages, Failed Student), or violation of a person's physical integrity without this being an accepted job or task risk (Food on Customer). Offenses that were rated as less severe by American than by Thai informants were Borrowed Book, a routine fact of American academic life, and Smuggled Buddha. For cultural outsiders, taking a Buddha statue out of Thailand constitutes no more than a petty offense, in legal terms. For Thais, the act is not just illegal but sacrilegious, expressing contempt of religious beliefs and of the most significant symbol of Thai culture. Americans gave higher Severity ratings than the Thais to Wrong Order, Private Message High-Low, Poor Teacher, and Cheating Student. Whereas the Thais assigned contradicting a colleague and questioning his or her professional competence the same medium Severity rating, the Americans perceived the "Poor Teacher" allegation as much more serious.

OBLIGATION TO APOLOGIZE Except for Order Change, which was rated low on Obligation by the Thais, both groups of informants perceived the offender's Obligation to apologize as medium or high in all contexts. However, the American raters gave as many as 12 offenses higher scores on Obligation than the Thais. A consistent relationship was discernable between Obligation and Severity in the ratings by both groups. First, Offense contexts received the same ratings on Obligation as on Severity. Thais and Americans perceived 13 contexts as similar on these dimensions. This finding is consistent with the assessment of seven offense contexts by Israeli raters, reported by Olshtain (1989). Second, Offenses received a categorically higher rating on Obligation than on Severity. In the American ratings, the low Severity offenses Order Change and Borrowed Book scored medium on Obligation, and five medium Severity offenses (Ruined Magazine, Wrong Order, Food on Waiter, and the Private Messages) registered high on Obligation. The Thai raters assessed six medium Severity items as high on Obligation (Private Message Low-High, Wrong Order, Food on Waiter, Borrowed Book, Damaged Car, Messed-up Bag). This relationship between Severi-

Apology, Perception, and Performance

91

ty and Obligation has not been reported in previous apology assessment studies (House, 1988; Olshtain, 1989; Vollmer & Olshtain, 1989). The ratings of individual offense items on Obligation thus permit us to elaborate upon the correlational finding, "the more severe an offense, the more it warrants apology": Either the obligation to apologize is directly proportionate to the severity of the offense, or the perceived need to apologize categorically exceeds the degree of the norm infringement. This latter finding strongly suggests the precedence of relational over transactional concerns in much human interaction: in the interest of restoring social harmony, remedial action in the form of apology is seen as required even if the offense is a minor matter. Building on Goffman's (1971) discussion of ritual and substantive apology (also Owen, 1983), the following analogy comes to mind: ritual apology is offered as redress to virtual offenses, that is, to events that might, but did not, cause offense. By the same token, factual minor offenses can be magnified into virtual major infringements through disproportionate offer of redress. The benefit accruing from a disproportionate (and acted-upon) need to apologize is to reestablish the implied parties—offender and offended person—as fully competent and responsible participants in the business of social interaction. There was, however, one exception to the patterned relationship between Severity and Obligation. In the Thai ratings, Smuggled Buddha was assessed as the highest offense of all (4.7) yet assigned to medium Obligation (3.5). There were no indications from the ratings of the other context-internal factors to explain this irregularity; although Smuggled Buddha registered low on Likelihood and high on Face-loss, so did other items that were assessed as strongly requiring apology. We think that perhaps the informants reasoned strategically on this item: since an apology is concomitant with admitting to an offense that is a major cultural and legal infringement, an escape route for the offender might be to pretend ignorance by not apologizing. We shall see whether this speculation bears out in the Thais' choice of apology strategies. LIKELIHOOD OF APOLOGY ACCEPTANCE The Thai raters did not perceive Likelihood for the apology to be accepted as low in any context, whereas the American raters felt that acceptance was unlikely in Broken Statue, Cheating Student, and Smuggled Buddha. These three contexts, which (except for Smuggled Buddha in the American ratings) registered highest on Severity in Thai and American assessments, were also given the lowest ratings in the medium category by the Thais. Likewise, Thai and American raters perceived the same seven offenses as having a high Likelihood for apology acceptance. All of these offenses had been given low ratings on Severity. OFFENDER'S FACE-LOSS All offense contexts were perceived as involving medium or high damage to the offender's face by both Thai and American raters. However, the Thais considered only four contexts—Contradiction, Poor Teacher, Order Change, and Private Message High-Low—as less than highly face-threatening. The Americans, on the other hand, felt in as many as nine contexts that the offender's face was not severely threatened. In addition to the four contexts also rated as medium for Face-loss by the Thais, the Americans attached only minor Face-loss to low and medium Severity

92

Speech Act Realization

offenses that were due to an oversight rather than to an intentional norm infringement, and that did not involve high costs to the offended party (Ruined Magazine, Messed-up Bag, Customs Form, Ungraded Paper, Borrowed Book). For the Thais, minor Severity of an offense did not entail minor Face-loss. Out of twelve medium Severity offenses, the Thais rated nine high on Face-loss. The highest Severity offenses were also given the highest ratings on Face-loss by Thais and Americans. The emerging pattern is thus that for the Americans, offender's Face-loss and Severity of offense were codeterminant. In the Thai informants' perception, by contrast, high Severity and high Face-loss were interrelated, but lower Severity might be outweighed by other factors that could make an event highly threatening to the offender's face even in the absence of much damage in objective terms. Such factors seem to include demonstration of undesirable personal attributes such as clumsiness, carelessness, or forgetfulness. We are reluctant, however, to draw further inferences from these differential ratings, since they may well be indicative of conceptual differences between the notion of "face" in Thai and American culture. Comparative ethnosemantic study of the concept efface" in both cultures and languages is needed, exploring its meaning and function within each cultural context (cf. Hu 1944 for an examination of the concept of face in Chinese culture). Of the context-external factors Distance and Dominance, we shall only examine Distance, as no effects were found for Dominance on any of the other contextual factors, nor on the selection of apology strategies.

DISTANCE With few exceptions, the Thai and American ratings of Distance reflected the social role relationship between offender and offended party. Thai and American informants agreed in perceiving the closest relationships between friends, and the most distant relationships between strangers in service and administrative encounters. Differences are noticeable in the two groups' perceptions of student-professor and work relationships. Both groups assessed the relationships between student and professor as medium Distance, corresponding to "acquaintance" as opposed to familiars or strangers. Yet for the Thais, student-professor tended more towards the low end of the medium Distant category (> 3), whereas for the Americans, it approximated closer to the high end of medium Distance (< 3). Possibly, Thais perceive the prototypical relationship of students and professors more similar to that of distant family members, while Americans regard it more as a work relationship involving participants on different levels of positional hierarchies. The latter view is supported by the ratings of relationships between colleagues, which the Thais perceived as more distant than the Americans in four out of six cases. Previous studies with different populations are consistent with either the Thai or the American assessments of social Distance in the two types of role relationships. German and British raters assessed relationships between colleagues as less distant than student-professor relationships (House, 1988, 305), whereas Israeli informants perceived more Distance between colleagues than between student and professor (Olshtain, 1989,160). Neither the Thai nor the American ratings indicated any evidence of differential perceptions of Distance between the relationships of teachers and office colleagues on the one hand, and between coworkers at different ranks on the other hand.

93

Apology, Perception, and Performance

Our findings on the Assessment study can be summed up as follows: 1. Thai and American raters perceive context-external and context-internal factors as unrelated. 2. Context-internal factors in offense contexts are highly interrelated. Severity covaries with Obligation to apologize, Likelihood for the apology to be accepted, and offender's Face-loss, lending support to Olshtain's (1989) hypothesis that "severity of offense is the representative contextual factor in the sociopragmatic set of the apology" (160). 3. Thai and American raters consistently perceive Obligation, Likelihood, and Face-loss as higher than Severity of offense, suggesting the primacy of interpersonal concerns over transactional goals in much remedial exchange. 4. Despite the overall consistent relationship of contextual factors in the American and Thai assessments, individual offense contexts may be characterized by constellations of contextual factors that are not predictable from the general pattern, and that vary cross-culturally. In each of the 20 situations, Thais and Americans differed in their perception of at least one contextual variable, most on Obligation, least on Likelihood. By way of summary, Table 4.3 presents the two language groups' differential assessments. Table 4.3. Differences in American and Thai Assessments of Contextual Factors Obligation Borrowed Book Messed-up Bag Poor Teacher Private Mess. H-L Wrong Order Smuggled Buddha Ruined Magazine Cheating Student Private Mess. L-H Business Mess. L-H Business Mess. H-L Food on Customer Food on Waiter Customs Form Contradiction Damaged Car Order Change Failed Student Ungraded Paper Broken Statue

Face-loss

Distance

Dominance

Severity

Likelihood

94

Speech Act Realization

The Dialog Construction Study The DC data were coded into the following major categories: IFID—Illocutionary Force Indicating Device, specifying the force of apology (I'm sorry, I'm afraid) Upgrader—Element increasing apologetic force (I'm terribly sorry, I really didn't mean to hurt you) Taking on Responsibility—speaker admitting the offense, including self-blame (How stupid of me), lack of intent (I didn't mean to do this), and admission of fact (I haven't graded it yet) Downgrading Responsibility or Severity of offense—(a) utterance reducing speaker's accountability for the offense, including excuse (My watch had stopped), justification (I was suddenly called to a meeting), claiming ignorance (1 didn't know you were expecting me), problematizing a precondition (we weren't supposed to meet before 12:00), or denial (I didn't do it); (b) utterance reducing severity of offense (I'm only 10 minutes late) Offer of Repair—speaker offering to remedy damage inflicted on offended party by an action to restitute H's entitlements (I'll pay for the damage, I'll have it marked tomorrow) Verbal Redress—speaker showing concern for offended party (I hope you weren't offended), efforts to appease (let me buy you a drink) or promise of forebearance (it won't happen again) Interrater reliability was established through consensus coding by three raters (Thai and English NS data) and two raters (Thai-English interlanguage data). The distribution of apology strategies across offense contexts and language groups is included in Appendix 4. The overall strategy distribution by language groups is displayed in Figure 4.1. Consistent with previous studies, IFID and Taking on Responsibility were used in more than half of the possible cases. Upgrading

Fig. 4.1. Overall use of apology strategies (%) by NS of English (English, N = 30), NS of Thai (Thai, N = 136), and Thai NNS of English (IL, N = 288). IFID = Illocutionary Force Indicating Device, UG = Upgrader, TR = Taking on Responsibility, DG = Downgrading, RE = Offer of Repair, VR = Verbal Redress.

Apology, Perception, and Performance

95

apologetic force, Downgrading Responsibility, and offer of Repair were supplied in one third of all contexts, and Verbal Redress was expressed on about every tenth occasion or less. The contextual distribution of strategies was examined by comparing strategy use in each context between pairs of language groups (NS English-NS Thai, NS English-IL users, NS Thai-IL users). IFID Overall, correlational analysis revealed no effects from the contextual factors on the selection of IFIDs, the exception being Obligation, which was found to covary with the use of IFIDs by the IL group (r = .6588, p < .002). In 16 out of the 20 situations, the three language groups agreed in their frequency of IFID suppliance. All groups used few IFIDs (between 14% and 50%) in low Severity contexts such as Order Change and Customs Form, as well as in high Severity situations such as Cheating Student and Smuggled Buddha. Thus the Thais matched their medium rating of Smuggled Buddha on Obligation by providing few apologetic formulas. Conversely, IFIDs were supplied with high frequency (80%96%) in response to high Severity offenses such as Broken Statue and Food on Customer, which received identical IFID frequencies as the medium Severity offense Food on Waiter. The inverse relationship, for some offenses, between Severity and IFID was especially pronounced in the English data: in the lowest Severity context, Borrowed Book, 83% IFIDs were supplied, as in the medium Severity contexts Ungraded Paper and Wrong Order (86%), while the high Severity item Damaged Car only elicited 47% IFIDs. Contradiction and Poor Teacher were assigned to different severity categories (medium and high) in the American Assessments, yet no significant difference obtained in the frequencies of IFIDs in these contexts. These two items were both assessed as medium offenses by the Thai raters, which is reflected in the frequencies of IFIDs provided by Thai NS and IL users. All three language groups used the same frequencies of IFIDs in response to the Messages offenses (60%-81%). The difference in Severity between Private and Business messages was not matched by IFID suppliance. The theoretically interesting point emerging from the highly variable use of IFIDs is its intricate relationship with Severity of offense. By responding to low Severity offenses with an explicit apologetic formula, the offender symbolically emphasizes her eagerness to repair whatever minor norm infringement has occurred. Not using IFIDs to remedy high Severity offenses may have either of two functions. First, an offender may avoid admitting responsibility for the offense committed, which would be concomitant with an explicit apology. Second, and conversely, where the offender is prepared to assume responsibility, an all-purpose apologetic formula, which is also used for ritualistic apology, might not be felt to convey adequately a substantive apology for a major offense. An expression of apology propositionally related to the specific offense might be more apt to convey the sincerity of the speaker's regret.

UPGRADING THE APOLOGY Upgrading of apology correlated highly with Obligation (r = .7628, p < .000) and Face-loss (r = .7248, p < .000) in the English data, and with Obligation in the Thai-English IL data (r = .6887, p < .001). Except for Obligation and IFID, which

96

Speech Act Realization

covaried in the IL data, Upgrading, remarkably, was the only strategy that correlated with context-internal factors; it reveals that for the English NS, Obligation to apologize and Face-loss have their behavioral correlates not in the use of apologetic formulas, nor in expressing responsibility, but in the intensification of either or both of these strategies. This intensification makes good sense, considering the routinized nature of apologetic formulas: in order for them to count as sincere apology, they warrant Upgrading, which serves to emphasize their substantive nature. The IL users, while sharing with the English NS the preference to express Obligation through Upgrading, also increased their use of IFIDs with greater Obligation. If this increase can be taken to indicate that the NNS regard IFIDs just as apt a means to express Obligation as Upgraders, it might reveal a tendency for the NNS to underdifferentiate the apologetic function of IFIDs, compared to English NS use. In all three groups, Upgrading correlated highly with IFID (English: r = .6906, p < .001, Thai: r = .7636, p < .000, IL: r = .8014, p < .000). To the extent Upgrading was used to intensify an IFID, it could be claimed that the covariance expresses the relationship between an independent (IFID) and a dependent variable (Upgrader). Yet it must be remembered that Upgrading was coded when it operated on other apology strategies as well, for instance, on Lack of Intent as a subcategory of Responsibility (e.g., I really didn't mean to hurt you). However, no covariance could be established between Upgrading and any other apology strategy. Least Upgrading (0%-20%) was afforded across language groups to the low Obligation and low Face-loss context Order Change, the medium Obligation offenses Customs Form, Contradiction, and Smuggled Buddha, and to Poor Teacher, which had been rated high on Obligation by the American and medium by the Thai informants. Consistent Upgrading was furthermore provided in Business Message Low-High. In half of the offense contexts, Upgrading was supplied differentially by the three groups. In four situations, the Thai NS provided fewer Upgrading than the English NS, which is reflective of their lower ratings of these offenses for Obligation (Failed Student and the restaurant situations Wrong Order, Food on Waiter, and Food on Customer). The reverse—the Thais providing more Upgrading than the English NS—was true in Messed-up Bag, where the English informants did not upgrade at all, and Ungraded Paper. The IL users closely approximated target norms in all but one context (Messed-up Bag).

TAKING ON RESPONSIBILITY Taking on Responsibility was the only apology strategy that covaried with a contextexternal factor (Distance) in all three groups (English: r = — .6514,p < .001, Thai r = -.6418, p < .001, IL: r = -.7165, p < .000). English and Thai NS and Thai-English NNS thus agreed in expressing more responsibility for the offensive act, the closer they were to the offended person; conversely, the more distant the relationship to the offended party, the less they were likely to admit accountability. Consistent with the ratings for Distance, in Damaged Car and Ruined Magazine, involving two friends, all groups of informants expressed Responsibility in 89%98% of the possible cases. In the four Message offenses, taking place between coworkers, offenders assumed Responsibility with a frequency of 66%-93%. The

Apology, Perception, and Performance

97

Thai NS and IL users afforded the same amount of Responsibility in the four Message contexts, regardless of status relationship (senior to junior or vice versa) or Severity of offense (private versus business message). The Thai NS consistently provided very high Responsibility frequencies (91%-93%), the IL users offering less (76%-84%). Curiously, the English NS expressed more Responsibility for failing to pass on a private message than for a business message. Two of the studentprofessor contexts also registered high for Responsibility: Borrowed Book and Failed Student. These items clearly illustrate that Severity of offense is not operative in determining whether or not an offender explicitly assumed Responsibility: Borrowed Book, which ranged lowest on Severity in the American ratings, elicited the highest amount of Responsibility (I forgot to take it along); Failed Student, one of the highest Severity offenses, was afforded almost as much Responsibility. High Distance relationships such as all waiter-customer interactions registered low on Responsibility (3%—47%). The failure of Severity to predict whether or not offenders take on Responsibility was also manifest at the lower end of the Responsibility scale: the lowest frequencies of Responsibility were observed in the lowest Severity contexts Order Change as well as in the high Severity context Smuggled Buddha. However, the Thais' reluctance to assume Responsibility for the Smuggled Buddha reflected its medium rating on Obligation. In seven contexts, more Thai NS than English NS assumed Responsibility, but only in two of these situations (Food on Customer and Broken Status) did the IL users follow their native pattern.

DOWNGRADING RESPONSIBILITY OR SEVERITY All three groups used considerably fewer strategies to downgrade responsibility for the offense than to explicitly assume responsibility. The high Severity offense Smuggled Buddha was downgraded by far the most frequently by all groups. This supports further our earlier speculation about the low rating of Smuggled Buddha on Obligation to apologize by the Thai informants. Cheating Student, another high Severity offense, was heavily downgraded by the Thai informants, whereas the English NS downgraded this offense in less than half of the possible cases. Contexts in which high Responsibility corresponded to low Downgrading were Damaged Car, Ruined Magazine, Borrowed Book, Failed Student, and Customs Form. Moreover, informants downgraded responsibility less in some contexts with high occurrence of IFIDs (Food on Customer, Wrong Order, Food on Waiter, Broken Statue). Yet some offenses, notably the undelivered Messages, were redressed both by highly frequent use of Responsibility and fairly frequent Downgrading strategies (all groups between 25% and 52%). With only 6 out of 20 situations in which the three language groups did not differ in downgrading responsibility, this category displayed the most intergroup variability. In most of these contexts, the Thai NS downgraded more than either or both English NS and IL users. The IL users downgraded responsibility more than did the English NS in three of the restaurant situations and in the two teachers' meetings. OFFER OF REPAIR In the American data, Repair correlated negatively with Downgrading (r = — .6824, p < .001): offenders will be less prone to offer Repair the more they downgrade the offense. The same logic did not obtain for the Thai data. Neverthe-

98

Speech Act Realization

less, for more than half of the offense contexts, English NS, Thai NS, and IL users agreed on whether or not they offered Repair. Least Repair (0%-10%) was offered in compensation for verbal or verbally conveyed offenses (Contradiction and Poor Teacher, the Private Message contexts, Order Change). Offenses that elicited Repair offers in more than half of the possible cases were those in which material compensation could be provided (Broken Statue, Ruined Magazine, Damaged Car, Food on Customer) or a neglected obligation honored (Ungraded Paper, Borrowed Book, Wrong Order, Failed Student). The language groups offered Repair differentially in three of the Message situations. Very few Repair offers were made by the Thai NS in all of the Message contexts, whereas the English NS and IL users offered Repair more frequently to compensate for the undelivered Business Messages. The IL users provided more Repair than the English NS in one type of high Distance context: customs official-traveler (Messedup Bag, Customs Form, Smuggled Buddha).

VERBAL REDRESS This strategy ranged lowest in frequency across language groups and offense contexts. By far the most frequently chosen subcategory of Verbal Redress was Concern for Hearer. Groups agreed in affording very little Verbal Redress in the four custom official-traveler contexts, three of the waiter-customer contexts and of the student-professor interactions, and in all Message contexts (0%-19%). Only one offense, Contradiction, elicited more than 20% Verbal Redress from all groups, which is consistent with previous findings for a comparable offense in English and German (House, 1988,310) and Hebrew (Olshtain, 1989,161). In three contexts, only one of the groups supplied over 20% Verbal Redress: the Thai NS in Failed Student, and the IL users in Poor Teacher and Food on Waiter. In 11 contexts, either or both Thai NS and IL users provided more Verbal Redress than the English NS. Expressing concern for the offended party was thus a more common strategy for the Thais, regardless of the language they used. Table 4.4 summarizes the contexts in which the IL users differed from the English NS in their use of apology strategies. No differences were obtained for Ruined Magazine, Private Message High-Low, and Failed Student. Of the six apology strategies, learners and English NS differed least in Upgrading apologetic force and in the canonical strategies IFID and taking on Responsibility, whereas they differed most in the context-dependent strategies Repair, Verbal Redress, and Downgrading responsibility or severity of offense. If we apply Selinker's (1969) operational definition for determining effects of language transfer to these data, the white dots indicate effects of pragmatic transfer in the Thai-English IL users' apology performance. Fifty-five percent of the differences in the use of apology strategies can thus be attributed to pragmatic transfer. We are not at all convinced that this is an appropriate way to determine pragmatic transfer; however, since we are not aware of a comparably rigorous method that might be more in accordance with current thinking on cross-linguistic influence, we shall accept this procedure, and the obtained results, until further notice. To summarize our findings on the contextual distribution of apology strategies across the three languages:

Apology, Perception, and Performance

99

Table 4.4. Differences in Apology Strategies Used between NS of English and Thai-English IL Users UG

TR

IFID

RE

VR

DG

Messed-up Bag (13) Business Mess. L-H (7) Cheating Student (20) Ungraded Paper (17) Broken Statue (15) Borrowed Book (18) Food on Customer (10) Food on Waiter (12) Order Change (9) Poor Teacher (4) Customs Form (16) Smuggled Buddha (14) Wrong Order (11) Private Mess. H-L (6) Business Mess. H-L (8) Damaged Car (1) Contradiction IFID = Illocutionary Force Indicating Device, UG = Upgrader, TR = Taking on Responsibility, DG — Downgrading, RE = Offer of Repair, VR = Verbal Redress. = IL users differ from English NS atp < 0.05. = IL users differ from English NS at/? < 0.05; English NS differ from Thai NS at/) < 0.05; no difference between Thai NS and IL users at p < 0.05.

1. Contextual factors operated differentially in the strategy selection. While Obligation had been found to determine the choice of IFID by NS of German and British English (House, 1988), in the present study it was only the IL users who made their choice of IFID contingent on perceived Obligation to apologize. 2. The strategy most sensitive to contextual factors was Upgrading: the more Obligation and Face-loss involved in an offense, the more Upgrading of apology would be provided. Previous studies had shown NS of Hebrew to increase apologetic force in direct proportion to Severity of offense and in indirect proportion to the offender's status vis-a-vis the offended party (Olshtain, 1989); the latter finding was also obtained for NS of British English and German (House, 1988). These concurring results provide consistent evidence for the underspecification of IFID as an apology strategy: since routine formulas are used both in ritual and substantive function, sincere expression of regret warrants specific marking through intensification (House, 1988; Vollmer & Olshtain, 1989). 3. Informants were more prone to explicitly express Responsibility for the offense the closer the relationship between the offender and the offended

100

Speech Act Realization person. This finding appears to contradict Wolfson's (1989) bulge hypothesis, according to which less redress is offered in interaction between very low and very high distance participants, whereas the most redress is provided in medium distance relationships. The contradiction is resolved if we reconsider the relationships comprised by our questionnaire contexts. Whereas they do include total strangers and acquaintances such as colleagues and friends, they do not extend to intimate relationships. Because the relationships spectrum covered in this study does not include the lowest end of the social distance continuum, our finding does not present counter-evidence to the bulge hypothesis. 4. No effect of contextual factors was found on Downgrading Responsibility, Repair, and Verbal Redress. All three of these context-dependent strategies tended to be used more frequently by the IL users than by the English NS. 5. The Thai-English IL users differed least from the English NS in their suppliance of Upgrading and the canonical strategies IFID and taking on Responsibility. Most differences occurred in the context-dependent strategies. 6. More than half of the differences in apology suppliance can tentatively be attributed to pragmatic transfer from Thai apology patterns.

By examining Thai learners' contextual perceptions and choices of apology strategies, we have shed light on their sociopragmatic knowledge of apology. We have not analyzed their actual wordings of apology strategies, that is, their pragmalinguistic knowledge of apologizing in English, which would require another study. How can we account for the relative oversuppliance of context-dependent strategies by the Thai learners? Our finding is consistent with House's (1988) observation that NNS tend to do "too much of a good thing," a phenomenon she labelled "gushing," or, less benevolently, "waffling" (Edmondson & House, 1991). Waffling had been reported to characterize request realization by learners of different native and target languages (e.g., Blum-Kulka & Olshtain, 1986; Kasper, 1989). House (1988) was the first to draw our attention to a similar phenomenon in NNS apologizing, demonstrating that German learners of English provided more apology upgrading and expressions of responsibility than NS of British English. Edmondson and House (1991) point out that all the studies in which waffling is evident are written DCTs, and remind us that waffling was not present in the interactional, negotiated discourse elicited by open-ended role-plays (Kasper, 1981; see also Eisenstein & Bodman, Chapter 3). Therefore, waffling may well be an instrument effect of DCT—both of the Discourse Completion and Dialog Construction variety. But Edmondson and House (1991) go further in their account of waffling. In both roleplay data (Kasper, 1981; Edmondson, House, Kasper, & Stemmer, 1984) and DCT data (House, 1988, 1989), German learners of English were consistently shown to use fewer conventionalized requestive and apologetic forms than British NS. At the same time, waffling was extant in the DCT though not in the role-play data. In order to reconcile these seemingly contradictory findings, Edmondson and House refer to the differential cognitive demands of discourse production in face-to-face interaction (role-play) and written questionnaires. They argue that paucity of adequate

101

Apology, Perception, and Performance

requestive and apologetic routines characterizes learners' speech act realization in both production tasks because such routines, though available in IL, are not "integrated into learners' discourse production systems" (285). Yet when learners' planning and execution of linguistic action does not operate under the pressures of conversational turn-taking, they will compensate for the lack of pragmatic routines by oversupplying nonconventionalized speech act realization strategies. Our study provides partial support for Edmondson and House's (1991) hypothesis. The Thai learners did indeed waffle, but they did not do so in compensation for lacking apology routines. Rather, our findings suggest that DCT, in whatever version, provides learners with an opportunity for knowledge display that is precluded for many NNS by the cognitive demands of face-to-face interaction. Comparison of learners' pragmatic performance in interactive discourse and DCT may thus throw light on the state of learners' discourse production systems, or the extent to which available pragmatic knowledge is readily accessible in conversation.

Acknowledgments The research reported in this paper was supported in part by the University of Hawaii Project Development Fund, grant R-90-866-F-034-B-101. We gratefully acknowledge the assistance we received from colleagues and students in carrying out the study. Staff and students at the Language Institute, Chulalongkorn University, Bangkok, helped us collect the Thai data. Naoko Maeshiba and Naoko Yoshinaga made their English NS data available to us. Peggy DuFon and Satomi Takahashi have a major share in the data analysis. Special thanks are more than due J. D. Brown, Thorn Hudson, and Steven Ross for their assistance with the statistical processing of the data during different phases of the project.

Appendix 1 Sample item from Assessment questionnaire At a friend's home John and Paul are good friends. John borrowed Paul's car for the weekend. Unfortunately, when he was backing up to park, he didn't see a lamppost. He hit it and damaged the rear of the car. He is now returning the car to Paul. 1. How close are John and Paul in this situation? 1 very close

2

3

4

5 very distant

4

5 John lower than Paul

2. What is the status relationship between John and Paul? 1 John higher than Paul

2

3 John = Paul

102

Speech Act Realization 3. How serious is John's offense? 1 not serious at all

2

3

4

5 very serious

3

4

5 absolutely

4

5 very unlikely

4

5 loses face

4. Does John need to apologize? 1 not at all

2

5. How likely is Paul to accept John's apology? 1 very likely

2

3

6. Does John gain or lose face in this situation? 1 gains face

2

3

Appendix 2 Sample item from Discourse Completion questionnaire At a friend's home John and Paul are good friends. John borrowed Paul's car for the weekend. Unfortunately, when he was backing up to park, he didn't see a lamppost. He hit it and damaged the rear of the car badly. He is now returning the car to Paul. PAUL: Is everything okay? JOHN:

PAUL:

Appendix 3: Mean Ratings (X) and Standard Deviations of Contextual Factors by American (n = 30) and Thai Informants (n = 30) Severity Item 1 X SD 2X SD 3X SD 4X SD

5 X SD

6 X SD IX SD

8X SD

9X

SD

10 X SD

11 X SD

12 X SD

13 X SD

14 X SD

15 X

SD 16 X SD 17 X SD 18 X SD 19 X SD 20 X SD

Obligation

Likelihood

Face-loss

Distance

Dominance

Am

Th

Am

Th

Am

Th

Am

Th

Am

Th

Am

Th

4.10 0.99 2.60 1.05 2.90 1.12 4.50* 0.56 3.60 0.96 3.50* 1.02 4.60 0.48 4.40 0.80 2.10 1.14 4.60 0.56 3.30* 0.89 3.50 1.28 3.20 1.19 3.60* 1.30 4.60 0.55 2.90 1.33 2.80 1.10 1.00* 0.91 4.70 0.53 4.80* 0.37

3.70 0.97 2.50 0.89 3.40 0.89 3.60* 1.05 3.20 1.02 2.90* 1.03 4.50 0.72 4.10 0.76 1.80 1.05 4.30 0.59 2.60* 1.05 3.00 1.05 3.40 1.09 4.70* 0.65 4.50 0.85 2.80 1.24 3.20 1.07 3.10* 1.20 4.50 0.96 4.50* 0.67

4.90* 0.30 4.70* 0.47 3.10 1.31 4.20* 0.87 4.50* 0.56 4.30* 0.97 4.90* 0.30 4.90* 0.40 3.40* 1.26 5.00* 0.18 4.80* 0.42 4.80* 0.50 3.50* 1.09 3.40 1.43 4.90 0.25 3.30 1.24 3.70 1.21 3.80 0.86 4.90* 0.25 4.60 0.80

4.30* 0.95 3.50* 0.97 3.00 1.25 3.10* 1.15 3.90* 1.00 3.50* 0.92 4.50* 0.62 4.10* 0.76 2.30* 0.96 4.60* 0.71 4.00* 1.03 4.10* 0.75 4.10* 0.73 3.50 1.63 4.70 0.59 3.10 1.16

3.60 1.08 4.50 0.88 3.30 0.82 2.70 0.99 3.80 0.76 3.90 0.96 3.00 1.06 3.60 0.99 4.10 1.09 2.90 1.06 2.90* 1.45 4.00 1.13 3.10* 1.11 2.30 1.23 2.20 1.08 3.10 1.14 4.10 0.81 4.50* 0.56 3.60 0.98 2.20 0.99

3.60 1.02 4.40 0.67 3.00 0.96 2.90 1.19 4.00 0.87 4.20 0.82 3.40 1.05 3.70 0.85 4.10 1.15 3.10 1.12 4.00* 0.94 4.00 0.98 3.70* 0.91 2.50 1.26 2.70 1.30 3.50 1.12 4.40 1.03 4.00* 0.91 3.50 1.20 2.40 1.02

4.30 0.83 3.50 1.15 2.60 0.96 3.70* 1.06 4.00 1.02 3.60 0.87 4.50 0.76 3.90 0.94 2.90 1.20 4.80* 0.40 4.20 0.84 4.50* 0.88 2.50* 1.23 3.80* 1.39 4.30 0.86 3.60* 0.95 3.30* 0.97 3.20* 1.12 4.60 0.84 4.70 0.53

4.30 0.80 3.90 0.80 2.60 1.07 2.60* 1.14 3.80 0.86 3.70 0.83 4.10 0.88 4.10 0.91 3.00 0.41 4.40* 0.80 3.80 0.79 3.80* 0.82 3.80* 0.82 4.60* 0.66 4.70 0.59 4.10* 0.81 3.90* 0.77 4.10* 0.60 4.50 0.67 4.70 0.57

1.50 0.62 1.80* 0.90 2.80* 0.95 3.10 1.24 3.10* 0.69 3.20* 0.73 3.10 0.75 3.00* 0.91 4.50 0.76 4.70 0.87 4.40 0.84 4.50 0.96 4.80* 0.54 3.90 1.36 4.80 0.56 4.60 0.72 3.20 0.78 3.10* 0.75 3.10 0.79 3.50* 0.85

1.40 0.66 2.80* 0.30 3.90* 0.86 3.50 1.33 3.60* 0.76 3.70* 0.86 2.90 0.87 3.70* 0.94 4.60 0.72

2.80 0.56 3.00* 0.55 2.90* 0.36 3.10 0.63 1.70 0.94 4.40 0.91 1.30* 0.54 4.40 0.88 4.20 1.01

4.60

1.60

0.60 4.60 0.66 4.60 0.72 4.30* 0.83 4.40 0.80 4.60 0.80 4.20 0.83 2.90 1.00 2.50* 1.12 2.80 1.01 2.90* 1.09

1.14 1.70 0.88 4.00 1.28 3.50 1.41 2.30* 1.42 3.40* 1.52 2.00* 1.10 4.60 0.76 1.90 1.11 4.30 1.04 1.40 0.92

2.70 0.85 4.00* 1.17 2.20* 1.31 3.10 0.50 1.40 1.05 4.60 0.84 1.90* 0.98 4.50 0.96 3.90 1.37 1.50 0.81 2.00 0.87 3.60 1.17 2.90 0.50 3.30* 0.94 2.30* 0.94 2.90* 0.44 4.40 0.99 1.60 0.87 4.40 0.84 1.60 0.80

3.20 1.26 4.10 0.93 4.40* 0.72 4.60 0.61

Item numbers in the leftmost column refer to the following offense contexts: 1 — Damaged car, 2 = Ruined magazine, 3 — Contradiction, 4 = Poor teacher, 5 = Private message low-high, 6 = Private message high-low, 7 = Business message low-high, 8 = Business message high-low, 9 = Order change, 30 = Food on customer, 11 = Wrong order, 12 = Food on waiter, 13 = Messedup bag, 14 = Smuggled Buddha, 15 = Broken statue, 16 = Customs form, 17 = Ungraded paper, 18 - Borrowed book, 19 = Failed student, 20 ~ Cheating student.

* = p < 0.05.

Appendix 4: Use of Apology Strategies (%) by NS of English (EN, n = 30), NS of Thai (TN, n = 136), and Thai NNS of English (IL, n = 228) UG

IFID

Item

1 n % 2n % 3n % 4n % 5n %

6n % 7n % 8n

% 9n %

10 n

DG

TR

EN

TN

IL

EN

TN

IL

EN

TN

IL

14 47

82 60 85 66 79 59 79 63 104 78 98 73 107 81 92 70 31 23

185 64 211 75 185 65 180 64 219 77 217 76 219 77 225 78 41 14 273

11 37 8 27 4 14 5

35 27

83 29 66 23 41 14 49 18 99 35C 102 36 153 54 117 41 4 1 206

27 90 27 90 11 37" 19 63

117 89 117 91 94 71" 81 65 121 91 125 93 122 92" 121 92 21 16a 62

262 91 276 98 139 49 136 49 217 76 236 81 238 84 238 81 25 9 108

20 67 18 60 20 67 23 77 19 63 22 76

18 60 8 27 28

115

17

14 47

10 33 16 55 16 53 2 7 23

21

16 17 13 22 18 73 55C 63 47 73 55 58 44 4 3 64

27

90 24 80 19 66» 22 73 1 3» 2

RE

VR

EN

TN

IL

EN

TN

IL

EN

TN

IL

1

29 22a 28

44

22"=

2' 49

20 67 18 60 0 0 1

94 72 88 68 10 8 9 7 7 5 11 8 3

217 75 196 70 16 6 11

2 7 1 3 7 23 2 8" 4 13 2 7" 1 5ab 1

18 14 12 9

31 11 13 5 74 26 58 21b 50 18 53 19b 38 13" 34

3.b

1 3a 13 43b

15 50b 12 40 13 43 12 41" 8 27 12 40ab

3

72 54= 62 50C 63 47 70

52= 42 32 44 34 9 7ac

22

15" 7

[7bc

66 24bc 86 30 85 30» 68 24b 68 26 0 Qbc

7

3 1

3 0 0 10 35* 8 21' 1 3 25

2ac

9 7»c 14 10' 84

4 14 5

13 5 37 I3bc

51 18'' 3 lc

188

3ab

3 [Jab

1

41

31 20

16 15 11 14 10 18 [4a

22 17" 2 2a 10

12"

3 lb 10

% 11 n % 12 n % 13 n % 14 n % 15 n % 16 n % 17 n % 18 n % 19 n %

20 n %

93 24 80 28

93 10 33»b 11 39 23 79 7 24 25 83 ab

25 83b 19 63 7 24b

87 85 64 115 87 76 60= 39 30 108 82 49 37 75 57= 93 70 98 73 39 31 =

96 236 84 263 92 186 65b 105 37 236 83 108 38 152 53" 166 59b 221 78 135 48 be

77= 13 43* 22 73= 0

48 »= 26 20» 57 43= 41

73= 84 30 177 63 75 26"

Qab

32'

7

21 16 61 46 24

44 15 177

18 60

21 71

46 ac

25= 64 23 83 29 52 18

25 11 38 5 17 4 13=

10 33 13 43= 4 14

50 38 29 22" 17 13

62 60

7ab

9 30 6 20= 14 47

3 11 = 7 24 =b 26 87 28 93 29 97 27 90 10 35

47 a

41 30

57 43= 66 52 32 25==

91 69= 106 80 94 71 124

93 120

89 59 46

38b 78 28 99 35 133 47 28 10= 140 49b 231 81 212 74

212 81 246 86 122 43

10b

4 13 9 30" 4

13=b 22 79 2 7 7 24 14 47= 2 7ab

2 7a 14 48 ab

16C 24 18 28 21

42 33= 105 80 15 11

28 21 106 80== 40 30= 36 27ac

92 72*

2bc

31 11 36 13b 79 28b 260 91 13 5 48 17 114 40=

114 ll b 7 2= 201 71"

83 19 63 10 33 1 3b 4

14b 13 45 2 7ab

12 40ab

17

57

21 70 3 10

63 79 60 50 38 12 9 21 16= 75 57 23 17»c

84 64= 78 59 85 63 22 17

66 230 82 124 44 54 19b 146 5]bc

181 64 123 44 be

216

76h 216 65 197 69 45 16

2 0 lb 1 4»b

0 0 0 0 0b

l=

0 0 0

1=

1 3" 2 7= 0

The leftmost column specifies offense contexts, cf. Appendix 3. IFID = Illocutionary Force Indicating Device, UG = Upgrader, TR ~ Taking on Responsibility, DC = Downgrading, RE = Offer of Repair, VR = Verbal Redress. a

= EN and TN different at p < 0.05.

>> = EN and 1L different at p < 0.05. = = TN and IL different at p < 0.05.

]ab

8

4

1 1

16 6" 82 29b 10 4 3 1 12

23 17= 5 4 4 3 8 6" 0 0 7 5= 18 14== 32 24a 6 5=

4"

5 1 7 2 7

6= 36 13 21 7b

106

Speech Act Realization

References Arndt, H., & Janney, R. W. (in press). Universality and relativity in cross-cultural politeness research: A historical perspective. Multilingua. Aronsson, K., & Satterlund-Larsson, U. (1987). Politeness strategies and doctor-patient communication: On the social choreography of collaborative thinking. Journal of Language and Social Psychology, 6, 1-27. Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requestive behaviour in five languages. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural pragmatics: Requests and apologies (123-54). Norwood, NJ: Ablex. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Appendix: The CCSARP coding manual. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (273-94). Norwood, NJ: Ablex. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Studies in Second Language Acquisition, 8, 65-119. Borkin, A., & Reinhart, S. M. (1978). "Excuse me" and "I'm sorry." TESOL Quarterly, 12, 57-79. Brown, P., & Eraser, C. (1979). Speech as a marker of situation. In K. Scherer & H. Giles (Eds.), Social markers in speech (33-63). Cambridge: Cambridge University Press. Brown, P., & Levinson, S. D. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Coulmas, F. (1981). "Poison to your soul." Thanks and apologies contrastively viewed. In F. Coulmas (Ed.), Conversational routine (69-91). The Hague: Mouton. Edmondson, W., & House, J. (1991). Do learners talk too much? The waffle phenomenon in interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds.), Foreign/second language pedagogy research (273-86). Clevedon and Philadelphia: Multilingual Matters. Edmondson, W., House, J., Kasper, G., & Stemmer, B. (1984). Learning the pragmatics of discourse. Applied Linguistics, 5, 113-27. Eisenstein, M., & Bodman, J. W. (1986). "I very appreciate": Expressions of gratitude by native and non-native speakers of American English. Applied Linguistics, 7, 167-85. Garcia, C. (1989). Apologizing in English: Politeness strategies used by native and nonnative speakers. Multilingua, 8, 3-20. Goffman, E. (1971). Relations in public. Harmondsworth: Penguin. Holmes, J. (1989). Sex differences and apologies: One aspect of communicative competence. Applied Linguistics, 10, 194-213. House, J. (1988). "Oh excuse me please . . .": Apologizing in a foreign language. In B. Kettemann, P. Bierbaumer, A. Fill & A. Karpf (Eds.), Englisch ah Zweitsprache (303-27). Tubingen: Narr. House, J. (1989). Politeness in English and German: The functions of "please" and "bitte." In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (96-119). Norwood, NJ: Ablex. House, J., & Kasper, G. (1981). Politeness markers in English and German. In F. Coulmas (Ed.), Conversational routine (157-85). The Hague: Mouton. Hu, H. C. (1944). The Chinese concepts of "face." American Anthropologist, 46, 45-64. Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr. Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters.

Apology, Perception, and Performance

107

Kasper, G. (1990). Linguistic politeness: Current research issues. Journal of Pragmatics, 14, 193-218. Kasper, G. (1989). [Native and nonnative apologies]. Unpublished manuscript. Leech, G. (1983). Principles of pragmatics. London: Longman. Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49). Rowley, MA: Newbury House. Olshtain, E. (1989). Apologies across languages. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (155-73). Norwood, NJ: Ablex. Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and second language acquisition (18-35). Rowley, MA: Newbury House. Olshtain, E., & Weinbach, L. (1987). Complaints: A study of speech act behavior among native and non-native speakers of Hebrew. In J. Verschueren & M. Bertucelli-Papi, (Eds.). The pragmatic perspective (195-208). Amsterdam: Benjamins. Owen, M. (1983). Apologies and remedial interchanges. Berlin: Mouton. Rintell, E., & Mitchell, C. L. (1989). Studying requests and apologies: An inquiry into method. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (248-72). Norwood, NJ: Ablex. Selinker, L. (1969). Language transfer. General Linguistics, 9, 67-92. Trosborg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11, 147-67. Vollmer, H. J., & Olshtain, E. (1989). The language of apologies in German. In S. BlumKulka, J. House & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (197-218). Norwood, NJ: Ablex. Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House. Wolfson, N., Marmor, T., & Jones, S. (1989). Problems in the comparison of speech acts across cultures. In S. Blum-Kulka, J. House & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (174-96) Norwood, NJ: Ablex.

5 Interlanguage Features of the Speech Act of Complaining ELITE OLSHTAIN and LIORA WEINBACH

In the speech act of complaining, the speaker (S) expresses displeasure or annoyance—censure—as a reaction to a past or ongoing action, the consequences of which are perceived by S as affecting her unfavorably. This complaint is usually addressed to the hearer (H) whom the S holds, at least partially, responsible for the offensive action. For the purpose of this study, censure will be assumed to have been expressed whenever S chooses to verbalize her disapproval of the violation. From the speaker's point of view, the following preconditions need to be fulfilled in order for the speech act of complaining to take place: 1. H performs a socially unacceptable act (SUA) that is contrary to a social code of behavioral norms shared by S and H. 2. S perceives the SUA as having unfavorable consequences of herself, and/or for the general public. 3. The verbal expression of S relates post facto directly or indirectly to the SUA, thus having the illocutionary force of censure. 4. S perceives the SUA as: (a) freeing S (at least partially) from the implicit understanding of a social cooperative relationship with H; S therefore chooses to express her frustration or annoyance, although the result will be a "conflictive" type of illocution in Leech's terms (Leech, 1983, 104); and (b) giving S the legitimate right to ask for repair in order to undo the SUA, either for her benefit or for the public benefit. It is the latter perception that leads to instrumental complaints aimed at "changing things" that do not meet with our standards or expectations. The main goal of such instrumental complaints is to ensure that H performs some action of repair as a result of the complaint. Since complaining or expressing censure is a face-threatening act directed toward H, and since its conflictive nature might result in a breach of the social goal of maintaining comity and harmony between S and H (Leech, 1983), S is faced with a series of "payoff" considerations before the actual realization of censure can take 108

The Speech Act of Complaining

109

place. Obviously, the annoyed and frustrated S needs to air her feelings in order to regain emotional balance or in order to get "things done properly." But how does this compare to the risk of losing a friend or causing another person considerable embarrassment and possibly even anger which might affect his willingness to carry out repair? In other words, to what extent might S be willing to risk certain social advantages in order to rid herself of the frustration caused by the SUA or in order to improve matters and correct the situation (in case of bad service, bureaucracy, etc.)? In cases where the damage done could be repaired by H, S needs to consider the most effective route for such a remedy to occur; a straightforward reproach may not ensure the repair process, while the avoidance of complaining might.

Payoff Considerations According to Brown and Levinson (1987, 60) the realization of a face-threatening act (FTA), such as the speech act of censure, follows a certain path of decision making. We contend that with respect to the speech act of complaining this series of decision-making steps is really a sequence of payoff considerations. There are at least three main junctures of decision making: 1. At the first juncture of decision making, the speaker has the possibility of completely opting out from performing the act; in other words, refraining from expressing censure altogether. Such opting out may have certain payoffs from S's point of view: she can get credit for being tactful and polite; she can avoid the consequences of potentially damaging H's face; and she can give H the feeling that S is considerate and cares for H (who is responsible for the violation). On the other hand, S might remain frustrated since she has not expressed her feelings of annoyance and disappointment and has not reprimanded H for the SUA. Furthermore, in cases of instrumental complaints (leading to expected repair) mentioned earlier, performing the act may risk the potential for repair and at times might endanger S's position with respect to H. Such considerations of risk will affect S's decision to carry out the act or opt out. 2. At the second node, S has the possibility of choosing to carry out the act "on" or "off" record. Off record in the case of censure would be some hint related to the inconvenience that resulted from the SUA, without explicitly mentioning either the SUA or H. Such an evasion of the clear and obvious expression of censure would present similar payoffs for S as in step 1 above, but if H chooses to pick up the hint, then some reaction from H, either an apology or an offer of repair, might set things straight and S would no longer feel annoyed. For example, if in the situation where there is a line of people waiting to get to a counter a person suddenly tries to push in front of S, the latter might say something like, "I'm the last in line." No real facedamage has been committed since there was no direct mention of the offender or the offense, yet the offender will, most probably, go and stand behind S and thus resolve the problem. On the other hand, if S decides to express censure "on" record, she moves on to the next juncture of decision making and payoff considerations. 3. At the third step of Brown and Levinson's path, S has a choice to realize the speech act with or without redress. If she chooses not to use redress, then censure is

110

Speech Act Realization

expressed unmitigatedly as a statement or request that explicitly mentions the SUA and/or H as violator. In this case, payoffs have been considered irrelevant or less important and S chooses to express her frustration without any mitigation. In fact, there is a possibility that S will not even refrain from a full realization of the complaint and will indeed use a personal threat or insult. The risk that S is willing to take is open conflict with H. In turn, H might react to such a complaint with an open attack on S, serving as retaliation for the face-damage suffered by H. On the other hand, if S chooses to use redress as part of the realization of the speech act of complaining, then two options are open: the positive politeness orientation or the negative politeness orientation. In the case of censure, a speaker opting for a positive politeness orientation would probably still prefer the explicit mention of both the SUA and H but with some expression of mutual concern and understanding, which creates mitigation. The payoff for such a choice would be the fact that S would fully air her frustration but at the same time would express personal interest or understanding concerning H. In an in-group context, this orientation will not necessarily lead to real conflict and the risk will be relatively small. Thus, if a mother tells her son, "You have left some dirty dishes in the sink again," and there is an understanding between them that he should place the dishes in the dishwasher, such a complaint will not greatly damage the relations between the two. If the decision is negative politeness oriented, the strategy to perform the act of censure might take the form of a mitigated expression. In the case of censure, such mitigated strategies might be realized as a conventional request for repair, where applicable, or as a statement relating to the SUA but not directly to H. Thus in the case of pushing into the line, S might say "Excuse me, sir. There's a line here," or in the case of a latecomer to a meeting, "I was really hoping to start our meeting on time." It seems that in the negative politeness orientation, even when the complaint is explicitly expressed, mitigation might lower the risk and thus lessen the effect on payoff.

A Cross-Cultural Comparison of Native Speakers' Norms In order to investigate features of interlanguage with respect to the speech act of complaining, it was necessary to first establish baseline intra-cultural and crosscultural norms. We started with a study of the realization of complaints in Hebrew (Olshtain & Weinbach, 1987; Weinbach, 1988). Data were collected via responses to a discourse completion questionnaire, consisting of 20 situations. Each situation presented respondents with a detailed description of the context, the social status and the social distance between the interlocutors, and details on whether there exists mutual commitment or an explicit/implicit contract (social obligation) that requires H to be responsible for the SUA (see Appendix). The data collected from 35 Israeli university students led to the development of a scale for the perception of the severity of the complaint. This scale enabled us to categorize complaint realizations on five points along a continuum. The categories were defined in terms of the degree of face-threat that S was willing to undertake when expressing censure or, in other words, in terms of payoff considerations as described above. The following five realization patterns were established.

The Speech Act of Complaining

111

1. Below the level of reproach are various realizations that enable S to avoid explicit mention of the offensive event or direct focus on S. In fact, this category deliberately leaves room for H's interpretation as to whether or not a complaint was actually voiced. A typical reaction in this category would be a general remark: for example, if H spilled something and damaged the tablecloth or some papers, S might say: "Such things happen," or "Don't worry about it, there's no real damage." In terms of payoff considerations, S makes sure (in this realization) that there is no real breach in social harmony as a result of the SUA and that H does not feel reprimanded too strongly. 2. Expression of annoyance or disapproval encompasses various realizations that are vague and indirect and do not explicitly mention either the SUA or H, but do express general annoyance at the violation. S still tries to avoid open confrontation with H, but makes it clear that there was a violation of some sort. Thus, a realization of a complaint in a statement such as: "Such lack of consideration!" or "This is really unacceptable behavior" does not specify exactly what was wrong and who was responsible. H can choose to interpret it as a complaint but he can also ignore the illocutionary intent. 3. Explicit complaint refers to realizations where S has made the decision to use an open face-threatening act toward H, but to instigate no sanctions. In this category S explicitly refers to either the SUA, or H, or both. Thus, any of the following realizations of complaints in a case where a doctor postponed an operation without informing the patient would be considered an "explicit complaint": "You're inconsiderate!" "One should not postpone this type of operation," or "You should not have postponed such an operation." 4. Accusation and warning is expressed as a complaint when S chooses to perform an open face-threatening act and further implies potential sanctions against H. An example of this type of complaint addressed toward someone who was very late for a meeting is, "Next time I'll let you wait for hours." 5. Immediate threat is expressed when S chooses to openly attack H. A realization of this strategy is, "You'd better pay the money right now"; "I'm not moving one inch before you change my appointment." This strategy can also consist of curses and direct insults, such as "You're an idiot." Olshtain and Weinbach (1987) suggested that these five categories of realization patterns of complaining make up the speech act set for this speech act (Olshtain & Cohen, 1983). Table 5.1 indicates that speakers of Hebrew tend to cluster around the three central strategies: disapproval, complaint, and warning. When the speaker is of lower status than the hearer, the tendency is to opt for less severe complaints (disapproval and complaint); when the interlocutors are equals or the speaker has higher social status, the tendency is stronger for the two more severe realizations (complaint and warning). Few respondents choose the softest strategy, "below level of reproach"—mostly in cases where the speaker is of lower status than the hearer; even fewer make use of "immediate threat," which seerns to increase with social power. The second study to be reported here was designed to compare different cultural groups with respect to a set of situations in terms of their preference for the optingout strategy, the direct strategy, the unmitigated, and the mitigated strategies (Katz, 1987). In this study the realizations of complaining clustered for all three groups in

112

Speech Act Realization Table 5.1. Distribution of Strategies by Social Status among Native Speakers of Hebrew (% of actual responses of each strategy in the particular social status context)

Strategy

S(-P) H ( + P) (%)

Equals (%)

S(+P) H(-P) (%)

Below level of reproach

18

8

6

Disapproval

28

15

13

Complaint

38

48

43

Warning

14

25

32

2

4

6

100

100

100

Threat

S — speaker, H = hearer, P = (social) power.

three major categories: the unmitigated form, comparable to what we called earlier the "warning"; the mitigated strategy, in fact, the conventional complaint; and the indirect strategy, which resembles implicit disapproval. Unique to this study is that it also allowed for the strategy of opting out in cases where the respondents chose to indicate that they would prefer not to offer a verbal reaction to the situation given. Subjects in this study consisted of two English-language groups—one British (n = 27) and one American (n = 23)—and a Hebrew-speaking group (n = 25). The English-speaking informants were students and staff from Lancaster and Birmingham Universities in England, and from Columbia and New York Universities in New York City; the Hebrew speakers were from Tel Aviv University and Beit Berl Teachers' Training College. Their age ranged from 19 to 55. The instrument used to collect the data was a Reaction Elicitation Questionnaire, which is in fact a modified version of Blum-Kulka's (1982) Discourse Completion Questionnaire. The Modified Elicitation Questionnaire was designed to elicit expressions of censure from the three groups of respondents in their country of domicile. The purpose of modification was to obviate, as much as possible, constraints made on responses with respect to tone or content, and furthermore, to allow the option of zero response leading to the opting-out strategy. The questionnaire consisted of two main parts (beyond the section dealing with personal information): (1) 25 situations to which respondents were asked to react in writing; and (2) an evaluation sheet on which respondents were to indicate the degree of perceived severity of the SUA provided in the questionnaire. The 25 situations provided in the questionnaire consisted of 20 descriptions related to acts expected to elicit expressions of censure and 5 distractor situations placed at randem among the 20 other situations. The acts described in the 20 censure situations refer to manifestations of socially unacceptable behavior in a Western culture. These situations were divided into 5 main content categories: 1. littering 2. noise making 3. unpunctuality

113

The Speech Act of Complaining Table 5.2. The Option to Express or Refrain from from Expressing Censure by Cultural Group British 533 tokens (%)

American 431 tokens (%)

Israeli 491 tokens (%)

Censure

65

61

64

Avoidance

35

39

36

100

100

100

4. queue jumping 5. petty stealing The first important finding of this study is that respondents from all three cultures opted out less than they chose to express censure, for the given situations. Table 5.2 presents these data. About two-thirds of the respondents in each group chose to realize the speech act of censure, while about one-third opted out by indicating on the questionnaire that they would prefer "to say nothing" in the given situation. Table 5.3 presents strategy preferences among those who decided to carry out the act of censure. From considering the data in both tables 5.2 and 5.3, it becomes apparent that the differences among the three groups were negligible. This was contrary to our expectations: from other studies we had come to perceive Israeli society as much more direct and positive politeness oriented and the British culture as much more indirect and negative politeness oriented, with the American culture falling somewhere in between. It seems, therefore, that the chosen situations had a much stronger impact on strategy choice, and this impact was similar in all three cultures. It is conceivable that the seemingly "universal" situations selected for the instrument contributed to similar strategy choice in the three cultural groups. Figure 5.1 shows the variation of responses by the three groups of respondents according to the five situations used in this study. The distribution of responses further reinforces that the three culture groups differ in their speech act behavior of complaining only slightly with respect to three situations (noise, unpunctuality, line), while showing almost identical realization patterns for the other two. The overall picture emphasizes the similarity of performance among the three culture groups.

Table 5.3. Strategies to Express Censure by Cultural Group British 347 tokens (%)

American 264 tokens (%)

Israeli 313 tokens (%)

Indirect

20

14

13

Mitigated

15

17

16

Unmitigated

65

69

71

100

100

100

114

Speech Act Realization

Fig. 5.1. Variation of responses by three groups of respondents.

The Intel-language of Complaints In this section we are concerned with the pragmalinguistic and sociopragmatic features of the interlanguage of complaining as exhibited by intermediate and advanced learners of Hebrew as a second language. We have tried to focus specifically on deviations from native norms that might lead to sociopragmatic failure (Thomas, 1983; Blum-Kulka & Olshtain, 1986). In previous studies (Kasper, 1981; Olshtain, 1983; Olshtain & Cohen, 1983; Olshtain & Blum-Kulka, 1985; Blum-Kulka & Olshtain, 1986) it has been established that nonnative speech act behavior can deviate from native behavior: in strategy selection, in utterance length, in the consideration of social and pragmatic features, in carrying out or opting out from performing a speech act, and in varying the degree of external and internal modification. It is our objective to describe some of the potential deviations that might arise particularly during the acquisition of the speech act of complaining. The research presented in this section will focus on one large-scale study comparing native and nonnative realizations of complaints with respect to five measures: utterance length as expressed in number of words, utterance length as expressed in number of moves, position on the severity scale, use of softeners, and use of intensifiers. Additionally, two shorter studies dealing with cultural and ethnic preferences for strategy selection, within the nonnative group, will be discussed as an expansion of our understanding of the various factors that might affect interlanguage behavior. Native and Nonnative Realizations of Complaints in Hebrew The main goal of this investigation was to identify factors that distinguish native from nonnative realizations of complaints. Two main hypotheses were formulated to underlie the various studies reported here.

The Speech Act of Complaining

115

1. According to previous work on interlanguage pragmatics (Blum-Kulka & Olshtain, 1986), we would expect nonnative speakers to produce longer utterances (as measured by number of words or by number of moves) than native speakers, given the same descriptions of situations on the discourse completion test (see Appendix). 2. Although the cross-cultural comparison above demonstrated striking similarities in the performance of the three "central" strategies on the severity scale (Tables 5.2 and 5.3), learners will tend to move more toward the less severe strategies than do native speakers; new immigrants or newcomers to any culture might try to avoid straightforward face-threatening interactions at all costs. Strategy selection for complaining can be ranked on a 5-point scale, where 5 is the strongest, most severe complaint, and 1 is slightly below the level of reproach. Native speakers can be expected to have a higher average along this scale (tendency toward stronger complaint) than nonnative speakers. Furthermore, in their attempt to sound less aggressive and more polite, learners might tend to use more softeners, which lower the face-threatening nature of complaints. Accordingly, they will use fewer intensifiers, since such intensifiers aggravate complaining force and create even stronger conflict. Two groups of subjects—35 native speakers of Hebrew and 35 learners (nonnatives)—were compared on five separate measures in order to test the above two hypotheses. Length of utterance was tested on two measures—number of words and number of moves within the realization; severity of complaint was measured by the average position along the severity scale (from 1 to 5), and softeners and intensifiers were compared across all situations. A multivariate analysis of variance (MANOVA) was conducted and the differences between the two groups were found significant (F[5,64j = 25,7; p < .001). The results of a univariate analysis for the differences between the two groups on the five measures are presented on Table 5.4. The results indicate strong group differences for the first three measures: total length of utterance, number of moves, and the severity scale (strategy selection). The learners, as expected in hypothesis 1, use considerably more words in order to express the speech act of complaining. The number of moves may be related to the overall length of utterance since more moves would have to be realized in more words, but it is obvious that independent of the number of moves, the overall length of utterance indicates a strong difference between the groups. This result further reinforces the findings presented in Blum-Kulka and Olshtain (1986) and strengthens our belief that advanced learners tend to make use of more words in order to negotiate their intentions expressed via speech act realizations. The results presented in Table 5.4 also confirm hypothesis 2, at least partially. According to hypothesis 2 we expected native speakers to prefer more severe strategies in comparison to the learners. The average on the severity scale is 2.79 for the Hebrew speakers and only 2.61 for learners, which fully confirms this part of hypothesis 2. Furthermore, Table 5.4 also indicates that, as expected, learners tend to use more softeners (.68)' in comparison to natives (.54). However, the last part of hypothesis 2 was not confirmed, as it seems that learners use more intensifiers (.77) than natives (.68). It may well be that in their attempt to negotiate their intentions, the use of more words overcomes any other considerations, intensifiers being used mainly for message clarification rather than for intensification.

116

Speech Act Realization

Table 5.4. Univariate F-Tests for Differences between Native Speakers and Learners on Five Measures (averaged over 20 situations) Native Ss

Learners

X

SD

X

SD

Total no. of words

7.05

1.03

10.14

1.84

74.86***

No. of moves

1.67

.18

2.09

.30

50.03***

Severity scale

15.24***

Measure

F

2.79

.21

2.61

.18

Total softeners

.54

.16

.68

.24

8.05**

Total intensifiers

.68

.16

.77

.18

4.67*

*p < .05.

**p < .01.

***p < .001.

Some situations were particularly sensitive to the differences between the two groups. Consider situation 11: "Your car needs some urgent repair work and you take it to the garage close to closing time. There are two cars in line before yours. You ask nicely to have your car fixed in spite of the late hour. The owner of the garage apologizes for not being able to fix it and rejects any possibility of even looking at it at this late hour." The native speakers made relatively short complaints in this situation, such as: "I won't come here again"; "Okay, I'll do without it"; "You just lost a client." The learners, on the other hand, had much longer utterances: "I was hoping for better treatment since I have been your client for quite some time. What can I do now?"; "Sir, I have to have my car. You know Mr. X, he recommended that I come here and I don't want to complain to him about you—it might give your garage a bad name. So, why don't you fix it for me?" These examples show how learners feel the need to "say more" in the hope that they might get the car fixed, after all. Situation 9 was particularly sensitive to the differences in strategy selection on the severity scale: "It is not the first time that loud rock music is heard from your neighbor's apartment quite late at night. You pick up the phone and you say . . ." In this situation the native speakers tended to opt for the most severe strategies, many using threats and ultimatums such as: "If you don't turn off your stereo soon I'll break your door down," "If you don't stop this loud music immediately I'll call the police." The learners were much more hesitant in their selection of strategies, opting for complaints rather than threats, such as: "I'm trying very hard to sleep but it is impossible with all this noise," "Some people would like a little peace and quiet but you are completely inconsiderate." Following the analysis above, we raised a number of additional questions. Given the high significance of length of utterance and selection of strategy on the severity scale as measures distinguishing the two groups, might particular aspects of the situation interact with these differences? We were concerned, primarily, with features of social status, social distance, and the existence of some kind of a contract or social obligation on H's part to have avoided the reason for complaining. These three independent variables: social status, social distance, and social obligation

The Speech Act of Complaining

111

were embedded in the description of the situations on the discourse completion questionnaire. Table 5.5 presents the means and standard deviations of utterance length and severity scale produced by the two groups, by social distance, social status and social obligation, while Table 5.6 gives the F values for the multivariate analysis of variance for these data. From the F values presented in Table 5.6 we see that social distance is a significant factor for both groups in terms of both length of utterance and severity of complaint (p < .001). It does not, however, interact with group and is nonsignificant as far as group distinctions are concerned. It is interesting to look at the means presented in Table 5.5 for the two groups: although length of utterance is always higher for learners, as was to be expected based on the main effect established earlier, it is surprisingly similar in its distribution according to different types of social distance. Thus, both groups use more words with acquaintances, as if there seems to be a need to negotiate more with interlocutors of this type. This finding goes along with Wolfson's (1989) description of "the bulge" phenomenon, which Table 5.5. Means and Standard Deviations (in parentheses) for Two Dependent Variables According to Groups by Social Distance, Social Status, and Social Obligation Number of words

Severity scale

Natives

Learners

Natives

Learners

Strangers

6.74 (1.61)

9.57 (2.90)

2.64 (.34)

2.43 (.31)

Acquaintances

8.48 (2.78)

11.79 (3.46)

3.02 (.34)

2.72 (.26)

Friends

5.54 (2.32)

8.59 (4.07)

2.67 (.62)

2.44 (.51)

Relatives

6.43 (1.68)

9.58 (2.65)

2.76 (.40)

2.73 (.38)

S(-P) H(+P)

7.77 (2.16)

10.71 (2.91)

2.46 (.35)

2.33 (.36)

S( + P) H(-P)

7.69 (1.60)

11.30 (3.58)

2.98 (.31)

2.78 (.34)

Equals

4.88 (.78)

5.79 (1.04)

4.13 (.79)

4.41 (.77)

8.06 (2.47)

li.OO (3.48)

2.66 (.37)

2.37 (.37)

Explicit

5.26 (.89)

6.27 (1.22)

4.46 (.79)

4.82 (.78)

Implicit

6.24 (1.52)

9.06 (2.63)

2.61 (.32)

2.61 (.34)

Social factor Social Distance

Social Status*

Social Obligation Nonexistent

*S — speaker. H = hearer, P = (soeial) power.

118

Speech Act Realization Table 5.6. F(df) Values for the Multivariate Analysis of Variance Source of variance

Length of utterance

Severity scale

Social distance

14.82** (3,66)

16.16** (3,66)

Distance & group

.14 (3,66)

1.89 (3,66)

177.50** (2,67)

184.83** (2,67)

Status & group

15.52** (2,67)

2.99 (2,67)

Social obligation

84.85** (2,67)

224.33** (2,67)

Obligation & group

9.92** (2,67)

5.43* (2,67)

Social status

*p < .01. **p < .001.

indicates that when speakers are less certain about roles and relationships with their interlocutors they negotiate more and accordingly use more words. It seems that both groups in our study feel the need for such negotiation in the case of mere acquaintances. On the other hand, with respect to strangers and relatives, they behave very similarly since here the roles are more fixed—strangers, we may never see again; relatives are here to stay. In either case there is less need for negotiation, as was also found by Wolfson (1989). The difference lies in the way Israelis, both native speakers of Hebrew and learners, react to friends. It seems that friendship is more binding in Israeli society and therefore, when a friend performs a breach of confidence, the disappointment is great. A sense of solidarity is well founded among friends (verbal reports of subjects) and there is no need to negotiate via lengthy exchanges. An example is situation 13: "A friend who takes the same course as you do was not willing to share some important material which he was able to get and which can help you prepare for the exam in that course." The complaints produced by the respondents in this case were rather brief, expressing deep personal disappointment. With respect to social status, Table 5.6 shows that this factor is significant for both length of utterance and strategy selection on the severity scale for both groups (p < .001), but more important, there is interaction between status and group with respect to length of utterance (p < .001). From Table 5.5 it is clear that again learners always use longer utterances than native speakers, but this is particularly strong in the case of S(+P) - H(-P), where the speaker is of higher status than the interlocutor. It is under these conditions that learners use the longest utterance, negotiating most fervently. Furthermore, the standard deviation (SD) in this case is extremely high (3.58), indicating great variability among the subjects in the group of learners. The strongest interaction effect in our analysis is found between social obligation and group, as can be seen from Table 5.6. Here the interaction is both for length

The Speech Act of Complaining

119

of utterance (p < .001) and for strategy selection (p < .01). In terms of length of utterance, both groups use the most intensive negotiation when there is no explicit social obligation; here again, expectations can be less well predicted and the obligation is not fixed. In cases like these, negotiation might bring about repair of the violation, hence if pays for the speaker to negotiate. On the other hand, when there is explicit obligation, one can simply appeal to it without too much negotiation, such as in situation 3: "You are driving your new car on a main street. Suddenly a car crosses in front of you not having stopped at a stop sign." In this case, the obligation is part of the law, and therefore the only thing subjects need to say is that one must abide by traffic laws. On the other hand, for situation 10, "You have returned from your vacation one day later than expected and are unable to hand in a seminar paper on time. Your professor is not prepared to extend the time," it seems worthwhile to negotiate and hope that the professor will change his mind. The greatest difference between natives and learners is found in cases where the obligation is implicit. Here the learners feel a greater need to negotiate. They also exhibit greater variability than natives do: on situations with no social obligation, there is a difference of 1.19 standard deviation between natives and learners on length of utterance, while the difference is 1.85 when the obligation is implicit (indicating greater variability for the implicit obligation). The interaction between social obligation and group (p < .01) on the severity scale is realized in a rather unexpected manner: for situations with explicit social obligation, the learners choose more severe strategies than the natives (4.82 as compared to 4.46; see Table 5.5); for the situations with implicit obligation, they show exactly the same means; and when there is no obligation whatsoever, the difference is very small, with the natives scoring only slightly higher than the learners. Both groups use the more severe strategies when there is social obligation, but this is the only case when the learners do so even more than the natives, probably feeling much more confident that in such obvious situations, they have the right to complain explicitly and boldly.

Cultural and Ethnic Considerations The last question to be discussed in this chapter relates to culturally specific parameters derived from context knowledge (Faerch & Kasper, 1984), which might be perceived differently by learners than by native speakers, thus leading to additional deviations in speech act performance. This question was investigated with two groups of immigrants in Israel: Russians and Moroccans (Hauser & Swindler, 1988). There were 20 respondents in each group; all were immigrants who had been in Israel for at least two years when the questionnaire was administered. The two groups reacted quite similarly in most situations and indicated a very similar spread of strategy selection. The differences occurred in two specific situations. Situation 3 was as follows: "You are nearing an empty spot in a parking place when another car drives in and takes the place, disregarding you completely." In this situation, Russians and Moroccans reacted very differently. Russians tended to use warnings, threats, and even curses, while Moroccans opted out or chose one of the two softer realizations of complaints, below the level of reproach or disapproval.

120

Speech Act Realization

When questioned in verbal reports, the Russians claimed that people should be considerate and play a "fair game" in a parking lot, while the Moroccans said that the offense was not serious; in fact, they might also do something like that. Situation 8 read: "You are in great need of a money loan. You approach a friend whom you had helped in the past, but s/he rejects your request." In this situation, the two groups of speakers reacted quite differently. Russians tended to accept the rejection and not to complain. If they did voice a complaint, it was "below the level of reproach." Moroccans, on the other hand, made use of the full scale of strategies. In verbal reports, the Russians claimed that when it comes to money business, it is actually better not to borrow from friends, while the Moroccans said that friends should be there to help you in time of need. In another study, which compared new immigrants from Romania with longestablished immigrants from the same country (Hoch-Pasko, 1988), it was found that some interesting differences occurred in the subjects' consideration of the contextual features. Thus, newcomers tended to opt out from expressing complaints in situations that related to a breach of social norms affecting the public in general, such as throwing litter in the public domain and being late for a movie, resulting in disturbing the audience. Yet, when the violation affected the individual (e.g., a neighbor in the building playing very loud music late at night), the newcomers tended to use more severe complaints than the long-established immigrants.

Conclusion A number of interlanguage features of speech act behavior have been established in the series of studies reported in this chapter, two of which seem to be general and might therefore apply to all speech act behavior of learners and three which might be specific to the speech act of complaining. The two general interlanguage features of speech act performance are length of utterance and variability. Learners at the intermediate to advanced level of second language acquisition tend to be verbose and use more words than native speakers, more than they themselves would use in their own language, in order to negotiate the intentions of their speech acts in the new language. When compared to native speakers of the target language, this difference is strong. More pronounced variability in strategy selection, use of intensifiers, softeners, number of moves, etc. can be seen from the consistently larger standard derivations exhibited by learners. Learners are less certain about such decisions and therefore tend to vary more in their choices than native speakers. The speech act of complaining, being an inherently face-threatening act, exhibits interesting deviations in learners' interlanguage from native speakers' norms, which are specific to this speech act. Such deviations seem to focus on the fact that newcomers to the target community attempt to sound less offensive and less facethreatening. As a result, their overall choice of realization patterns is consistently closer to the less severe end of the scale when compared to native speakers. As part of their variability in choice, they also seem to be more sensitive to differences in social factors related to their interlocutors. In our study, the unexpected finding

The Speech Act of Complaining

121

related to the notion of social obligation: when the situation was one in which social obligation was explicit either by law or convention, the learners felt more secure in their choices and were less concerned with being polite and cautious. The studies of the interlanguage of speech act behavior presented in this chapter indicate that contextual parameters such as social factors and situational factors might affect realization patterns of speech acts produced by learners. Some of these interlanguage features have the potential to represent any speech act behavior while others are speech act specific.

Appendix Sample items from discourse completion questionnaire (translated from Hebrew): Situation 9 It is not the first time that loud rock music is heard from your neighbor's apartment quite late at night. You pick up the phone and say:

Situation 13 A friend who takes the same course as you at the university refuses to share some important material for the next test, which s/he managed to get hold of. In the past, you helped him/her many times. You see him/her on campus and say:

References Blum-Kulka, S. (1982). Learning how to say what you mean in a second language: A study of the speech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-59. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Studies in Second Language Acquisition, 8, 47-61. Brown, P., & Levinson, S. (1987). Politeness: Some universah in language usage. Cambridge: Cambridge University Press. Faerch, C., & Kasper, G. (1984). Pragmatic knowledge: Rules and procedures. Applied Linguistics, 5, 214-25. Hauser, Z., & Swindler, I. (1988). A comparison of realization patterns of complaints among Israelis from Morroco and from the Soviet Union. Unpublished seminar paper, School of Education, Tel Aviv University. Hoch-Pasko, R. (1988). Complaints as expressed by Israelis from Roumanian background. Unpublished seminar paper, School of Education, Tel Aviv University.

122

Speech Act Realization

Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Eine Untersuchung des Englischen fortgeschrittener deutscher Lerner. Tubingen, Germany: Narr. Katz, L. (1987). Face risk in the expression of censure: A cross-cultural study. Unpublished master's thesis, School of Education, Tel Aviv University. Leech, G. N. (1983). Principles of pragmatics. London: Longman. Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In S. Gass & L. Selinkcr (Eds.), Language transfer and language learning (232-49). Rowley, MA: Newbury House. Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language acquisition (303-25). Rowley, MA: Newbury House. Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition (18-36). Rowley, MA: Newbury House. Olshtain, E., & Weinbach, L. (1987). Complaints—A study of speech act behavior among native and non-native speakers of Hebrew. In M. B. Papi & J. Verschueren (Eds.), The Pragmatic perspective: Selected papers from the 1985 International Pragmatics Conference (195-208). Amsterdam: Benjamins. Thomas, J. A. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112. Weinbach, L. (1988). The realization pattern of complaints in Hebrew: Native speakers and learners. Unpublished master's thesis, School of Education, Tel Aviv University. Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. Cambridge, MA: Newbury House.

6 Interlanguage Requestive Hints ELDA WEIZMAN

This chapter explores the use of Hints as a request strategy by learners. It addresses the question of regularities to be observed in the use of requestive Hints by language learners with various mother tongues at different proficiency levels, and the similarities or differences between them and native speakers. This analysis contributes both to our knowledge of learners' pragmalinguistic behavior, as well as to our understanding of indirectness. The study was conducted within the framework of the Cross-Cultural Speech Act Realization Project (CCSARP) (cf. Blum-Kulka, House, & Kasper, 1989), and focuses on a comparison of Hebrew learners (henceforth HL) with Hebrew native speakers (henceforth HN), examined in the context of a larger body of CCSARP data. Drawing on findings concerning the situational variation and the use of Hint substrategies by native speakers (Blum-Kulka, 1989; Weizman, 1989), I propose to examine whether the opacity inherent in Hints is exploited by learners as a strategy of communication, whether the impact of situational variations marks Hint selection by learners as it does with native speakers, and whether the use of Hint substrategies is comparable in the two groups. Defining Requestive Hints The category of requestive Hints consists of the most nonconventional and indirect of all request strategies. Typical instances include: / have so much work to do (when used as a request to be left alone). Don't you have anything to do? (when used as a request to be left alone). It's cold in here (when used as a request to close the window). /' ve just missed my bus (when used as a request for a ride). Are you going home now? (when used as a request for a ride). Do you have your notebook with you? (when used as a request that the hearer lend the speaker her notebook). The kitchen is in a bit of a mess (when used as a request that the hearer clean up the kitchen). 123

124

Speech Act Realization

In Hints, the interpretation of the speaker's intentions is highly contextembedded, and is not secured either by the sentence meaning of the utterance (i.e., by its context-independent, literal meaning), as in direct requests (e.g., "Leave me alone, please"); or by some grammatical or semantic device, conventionally used to convey the requestive force, as in indirect, conventional requests (e.g., "Could you/ Would you leave me alone, please?") Initially, the CCSARP coding scheme made a distinction between two types of Hints—mild Hints and strong Hints—in terms of the amount of contextual knowledge required for their interpretation. For the purpose of the present analysis, these two strategies are collapsed, and each Hint is re-analyzed. The proposed analysis assumes, following Weizman (1985; 1989), that although all Hints are inherently opaque, they are not all opaque in the same way, or to the same extent. Hence, Hints vary on two dimensions: degree and type of opacity. In terms of type, it is either Hints' illocutionary force, or their propositional content, or both, that is obscure, and that should, therefore, be "explored" and "deciphered" by the interpreter. On the illocutionary dimension, a Hint is considered opaque when its utterance meaning does not provide sufficient indication of its intended illocutionary force; consequently, it carries the potential to perform a number of illocutionary acts. On the propositional dimension, a Hint is considered opaque when its sentence meaning does not provide sufficient evidence of the content of the act. Type and degree of opacity interact, however, since on each of these dimensions Hints may be distinguished in terms of degree of opacity, that is, the amount of "missing clues" required for full interpretation. (For further details, refer to "Distribution of Hint Substrategies in Learners' Data," below.) Obviously, since in real-life situations disambiguation is highly context-dependent (see Dascal & Weizman, 1987; Weizman & Dascal, 1991), the distinction by types is not always clear-cut. Generally speaking, however, it seems safe enough to suggest that while illocutionary opacity entails propositional opacity, the opposite does not hold, since opacity on the propositional dimension does not necessarily imply opacity on the illocutionary scale. For example, the utterance "I have so much work to do" could be interpreted either as a request (to be left alone) or as an apology (for having forgotten to do the shopping). Hence, it is considered opaque on both the illocutionary and the propositional dimensions. The utterance "If only someone helped me," on the other hand, would most plausibly be interpreted as a request, but the content of the request is completely obscure, since the speaker does not specify the kind of help needed. Hence, this Hint is considered relatively transparent on the illocutionary dimension, and extremely opaque on the propositional one. Note that the two dimensions interact: since certain propositional components are entailed by the illocutionary force (Vanderveken, 1984, 190), a certain degree of propositional transparency may help to disambiguate the illocutionary force. For instance, the illocutionary force of promises imposes on the propositional content the requirement that a reference be made to a (semantically) future action (Searle, 1975). Since, however, reference to a future action is also apt to be made in requests, propositional content in Hints can only be considered as contributing to the elucidation of illocutionary force. Hence, it is the balance between the opacity types that changes from one Hint to another, and creates the enormous variety of requestive Hints. 1 The very essence of Hints cannot be fully grasped unless one attempts to answer

Interlanguage Requestive Hints

125

the question: Why use Hints at all? Assuming the speaker to be a rational agent (Grice, 1975), he or she is supposed to achieve any given end by choosing "that action which, most effectively, and at least cost, attains that end, ceteris paribus" (Kasher, 1982, 32). What end could possibly be attained by the use of such a high degree of opacity? Surely, if the end to be achieved is the requestive one—that is, if the speaker intends to get a requested act carried out by the hearer as a result of the hearer's recognition of this intent (Searle, 1975)—then the most effective way of achieving this purpose is not by using an "off-record" strategy (Brown & Levinson, 1978, 210), but by using direct ones. One would tend, then, to associate indirectness with politeness and tact, following Leech, for instance: "in general, the more tactful a directive is, the more indirect and circumlocutionary it is" (Leech, 1980, 109); or: "Indirect illocutions tend to be more polite (a) because they increase the degree of optionality, and (b) because the more indirect an illocution is, the more diminished and tentative its force tends to be" (Leech, 1983, 108). Along these lines, it is tempting to suggest that the end to be achieved by the use of Hints is to get a requested act carried out without threatening the hearer's face. Experiments have shown, however, that although perceptions of politeness vary across cultures, Hints are not conceived as the most polite strategy. Within the framework of CCSARP, House (1986) and Blum-Kulka (1987) have shown that speakers of Hebrew, American English, British English, and German conceive of Hints as less polite than conventional indirectness. The same point was made earlier by Walters (1979, 286, 287) about speakers of American English and Puerto Rican Spanish, within the framework of studies on learners' performance of requests in role-playing situations (Eraser, Rintell,& Walters, 1980; Walters, 1981; Walters, 1979;Rintell, 1981).2 Evidence then seems to indicate that the most effective way to make requests while saving the hearer's face is not the use of Hints, but the use of conventional indirectness. If Hints are not the most efficient way to achieve the requestive purpose, or to save face, what then is the end that can be attained most effectively by using them? In Weizman (1989) I have suggested the following: the large repertoire of plausible interpretations of a given Hint seems to imply that Hints are the most efficient way for the requester to make a request while at the same time securing the possibility of legitimately denying some of its illocutionary and prepositional components. I have therefore suggested (Weizman, 1989) that Hints be thought of as the only request strategy that bears a high deniability potential for both parties: the requester may plausibly deny having made a request (e.g., "I never meant to ask anything of you") or deny its prepositional content (e.g., "I never asked you to fetch me my coat"; "I just wanted you to close the window"); the requestee may legitimately ignore the request or pretend to have misunderstood its content. Since this deniability potential might be exploited by learners, the recognition of its importance underlies the questions raised in the following section.

Research Questions This study raises three main questions, underlined by previously observed data and some working assumptions. First, in the CCSARP data, native speakers' use of Hints is usually remarkably low relative to either direct strategies or conventional

126

Speech Act Realization

indirect ones. The analysis of request strategies in five languages combined (Australian English, Canadian French, German, Argentinian Spanish, and Hebrew) reveals that native speakers' selection of Hints varies from 1.14% to 7.76%, depending on situational constraints (Blum-Kulka & House, 1989, 130). Low frequency is also manifested in Hint selection by native speakers of British English, Danish, and German in most of the situations (Kasper, 1989; House & Kasper, 1987). A similar trend is depicted in studies of other data types, e.g., studies of request realization by native speakers of English and German in role-plays (Kasper, 1981), and by native speakers of Hebrew in authentic face-to-face interaction, phone calls, and letters to the editor (Blum-Kulka, Danet, & Gerson, 1985). In terms of the deniabilitypotential hypothesis, this low frequency seems to indicate either that native speakers only rarely feel the need to secure a way to opt out of their request; or else, that only seldom do they consider it worth doing at the expense of efficiency. As for learners, the deniability potential might appeal to them for other reasons. More specifically, it might be exploited as a strategy of communication. By virtue of their nonconventional form, Hints might be employed to avoid the need to cope with the target-language conventional realizations; and, by virtue of their opacity, they might be exploited to avoid specific reference to the interlocuter's preceding turns. Put differently, their high deniability potential might be exploited by way of "risk avoidance," whereby a learner can "tailor his message to the resources he has available" (Corder, 1983, 17) in order to cope with "what to an individual presents itself as a problem in reaching a particular communicative goal" (Faerch & Kasper, 1983, 212). If Hints are indeed employed as a communication strategy, a higher frequency of use is likely to be observed in learners, mainly those least exposed to the target culture and therefore most liable to feel insecure. This raises the questions whether learners use Hints more frequently than native speakers, and whether the incidence of Hints decreases with the increase in length of stay in the target culture. Answers will be suggested in "Proportion of Hints and Situational Variation in Learners" (below). Second, observed situational variation in native speakers' use of all request strategies, including Hints (Blum-Kulka, 1989) and Hint substrategies (Weizman, 1989) suggests powerful situational constraints on their requestive behaviour. CCSARP questionnaires include eight situations, differentiated in terms of six social parameters: the degree of (1) dominance and (2) familiarity between the speakers; the requester's (3) right to make the request and the (4) degree of difficulty involving in making it; the (5) likelihood of compliance and (6) degree of obligation on the part of the requestee (Blum-Kulka & House, 1989). The second issue discussed in "Proportion of Hints and Situational Variation in Learners" (below) concerns the impact of situational variance on the selection of Hints by learners. Third, Hints in the CCSARP data were found to consist of several main substrategies, ranging from extreme opacity to relative transparency. The low incidence of other substrategies in native speakers' data was interpreted as indicating that although Hints are nonconventional in form; that is, nonconventional in terms of the lexical and grammatical items used for their realizations, their use is not entirely convention-free. The following point has been made (Weizman, 1989): one usually thinks of Hints, and justly so, as an open-ended category, because utterances of any

127

Interlanguage Requestive Hints

form could be interpreted as carrying a given illocutionary force, depending on the situation. Nevertheless, it seems to be the case that even the use of Hints obeys a certain order, since in a given situation a given set of semantic contents is more likely to be preferred, although an open-ended repertoire of potential candidates is available. For example, when requesting that the hearer give the speaker a lift, a Hint would preferably be realized by questioning some constituent of the feasibility of that action; for example, by asking, "Do you have a car?" or "Are you going home?" whereby a positive answer is a necessary, but not sufficient, condition for the feasibility of the requested action. On the other hand, when requesting a roommate to clean up the kitchen, a Hint is more likely to be realized by stating some reason for the request; for example, "the kitchen is in a terrible mess." (For further details, see "Distribution of Hint Substrategies in Learners' Data," below.)3 It is therefore important to emphasize that requestive Hints are viewed here as being accountable in terms of "conventions of means," specifying "a semantic device by which an indirect speech act can be performed," as opposed to "conventions of form," that is, "conventions about the wording of indirect speech acts" (Clark, 1979, 433). Furthermore, marked preference for the most opaque substrategy has been observed in native speakers and was interpreted as a confirmation of the hypothesis that opacity in Hints is intentionally and purposefully exploited (Weizman, 1989). The questions that arise at this point are whether learners prefer the same substrategies, and whether they, too, opt for the most opaque one. Answers to the last two questions will be suggested in "Distribution of Hint Substrategies" (below).

Methodology CCSARP data, elicited by means of a discourse-completion test, offers the possibility of exploring the use of requests and apologies by native speakers of Australian, American and British English, Canadian French, Danish, German, and Hebrew, as well as by learners of English in Denmark, Germany, and the United States, of German in Denmark, and of Hebrew in Israel (Blum-Kulka, House, & Kasper, 1989, chapter 1). The present study compares learners with native speakers, all students. Table 6.1 indicates the distribution. A two-step analysis was undertaken. First, 305 Hebrew learners, (henceforth Table 6.1. Comparison of Learners and Native Speakers Learners

Native Speakers Mother tongue

No. of subjects

Israel

Various

305

Hebrew

173

Germany

German

200

Australian English

226

Denmark

Danish

200

American English

27

English

USA

Various

35

German

200

English

Israel

Various

47

Canadian French

128

Target language

Country

Hebrew English English

Language

Argentinian Spanish

No. of subjects

40

128

Speech Act Realization

referred to as HL), were compared with 173 Hebrew native speakers (henceforth HN). Then, all other learners (n = 482) were compared with the rest of the native speakers (« = 621). The reason for the distinction is that only Hebrew learner's data allow for quantitative analysis in terms of length of stay in the target culture, and for qualitative analysis of Hint substrategies. The rest of the data lends itself mainly to the study of frequency of use and situational variation.

Data Analysis Proportion of Hints and Situational Variation in Learners Of the eight request situations featuring in the CCSARP questionnaire, the following five are analyzed here: SI (referred to as "Kitchen"): A student asks his roommate to clean up the kitchen the latter had left in a mess the night before. S5 (referred to as "Notes"): A student asks another to lend her some lecture notes. S7 (referred to as "Ride"): A student asks people living on the same street for a ride home. SI 1 (referred to as "Policeman"): A policeman asks a driver to move her car. S15 (referred to as "Lecturer"): A university lecturer asks a student to present his paper a week earlier than scheduled. Table 6.2 shows the impact of the situational variation on the selection of Hints by HL as compared to HN. The analysis reveals that in both populations the frequency of Hints is situationally dependent. In each of them, Hints rarely occur in the Notes situation (1.7% in learners, 1.2% in native speakers). The Lecturer situation is also marked by a relatively low incidence of Hints (4.2% and 2.7%, respectively), while in both groups, Hints most frequently occur in the Policeman situation (14.6% and 12.8%, respectively). In this respect, no significant differences are found between HL and HN. 4 Table 6.2. Distribution of Hints by Situation in Hebrew Native Speakers (HN) and Hebrew Learners (HL) Native speakers (HN)

Learners (HL) Situation

N

n

%

N

n

%

p* value

chi-square 0.472

1 Kitchen

293

33

11.3

163

15

9.2

0.492

5 Notes

297

5

1.7

170

2

1.2

#*

7 Ride

272

28

10.3

165

23

13.9

1 1 Policeman

240

35

14.6

164

21

15 Lecturer

191

8

4.2

150

4

1293

109

8.4

812

65

8.0

Total

Note: Of A' utterances, n are rcquestivc Hints. *For learners (HL) versus native speakers (HN), by chi-square analysis. ;!:

*The number of Hints is too small to allow for statistical analysis.

**

0.249

1.324

12.8

0.611

0.258

2.7

0.449

0.573

Inierlanguage Requestive Hints

129

As concerns native speakers, these findings confirm earlier ones. Blum-Kulka and House (1989) show the significance of situationally specific features on the determination of request strategies in general, and Weizman (1989) reports on situational constraints affecting the selection of Hint substrategies. Blum-Kulka and House (1989, 128) further indicate that for native speakers of the 5 languages examined—Australian English, Canadian French, German, Hebrew, and Argentinian Spanish—the Notes and Lecturer situations attract not only the lowest proportion of Hints, but also the highest degree of cross-cultural agreement. From the comparison presented here, the same picture emerges for Hebrew learners: their sensitivity to situationally specific features is as crucial to Hint selection as it is with Hebrew native speakers, and they, too, share the agreement concerning the two situations in question. These findings also seem to agree with Blum-Kulka and Levenston's (1987) observation, that in four of the five situations examined, the distribution of strategy types by situation revealed no statistically significant differences between Hebrew learners and Hebrew native speakers (157). The question arising at this point is whether this high situational sensitivity in Hebrew learners reflects a general tendency to be observed in learners of other languages. To answer this question, different sets of CCSARP data were collapsed (as explained in "Research Questions," above), so as to allow comparison of Hint selection by native speakers of various mother tongues with Hint selection by learners of various target languages, regardless of their mother tongues. We know by now that situational features affect internal and external modifications of requests (mitigating devices such as interrogatives, conditionals, hedgers, politeness markers) (Faerch & Kasper, 1989; House, 1989). Table 6.3 indicates that they also affect the selection of Hints. In Table 6.3, two populations are compared: on the one hand, 621 native speakers (of Australian and American English, German, Canadian French, and Argentinian Spanish) and on the other, 482 learners (of English). The results show a marked situational variation in Hint selection by learners, and suggest similarities in that respect with native speakers. Again we see that in both groups, the Notes Table 6.3. Distribution of Hints by Situation in Native Speakers (of Australian and American English, German, Canadian French and Argentinian Spanish) versus Learners (of English, with Various Mother Tongues) Learners (HL) Situation

N

n

Native speakers (HN) %

N

n

%

p* value

chi-square

1 Kitchen

471

24

5.1

614

49

8.0

5 Notes

479

2

0.4

618

2

0.3

0.060 *#

3.535 *#

7 Ride

0.410

474

25

5.3

614

38

6.2

0.522

1 1 Policeman

473

31

6.6

610

51

8.4

0.265

1.243

15 Lecturer

465

24

5.2

604

31

5.1

0.983

0.00045

2362

106

4.5

3060

171

5.6

Total

Note: Of TV utterances, n are rcquestive Hints. *For learners (HL) versus native speakers (HN), by chi-square analysis. *:tThe number of Hints is too small to allow for statistical analysis.

130

Speech Act Realization

situation attracts the smallest percentage of Hints (0.5% in learners, 0.19 in native speakers); the Lecture situation attracts a relatively small proportion of Hints (4.59% in learners, 5.57% in native speakers); and Hints are most likely to occur in the Policeman situation (6.9% and 8.87%, respectively). On the whole, observed situational variation by learners indicates that Hint selection by learners is as situationally determined as their selection by native speakers. Looking at the totals in Table 6.3 we can see that, taken together, the proportions of Hints employed by learners and native speakers (4.5% and 5.6%) are fairly similar. The similarity between HL and HN, as shown in Table 6.2, is more pronounced: 8% and 8.4%, respectively.5 These findings, then, suggest no reason to believe that Hints are used as a strategy of communication. This interpretation is further supported if Hint selection is examined in terms of learners' length of stay in the target community. Table 6.4 compares the use of Hints by Hebrew learners with different periods of residence in Israel. Table 6.4 reveals striking similarities between the compared groups of immigrants. These similarities seem to be particularly interesting since language exposure was shown to affect the acquisition of target norms, at least as far as verbosity is concerned (Blum-Kulka & Olshtain, 1986, 158). The lack of effect of length of stay on Hint selection, I suggest, has to do with two factors: (1) Hint selection by HL matches NH norms from the very start, and there is therefore no reason to expect any mismatch with time; (2) The use of Hint differs from that of any other strategy, mainly because conventions of use, that is, lexical and grammatical conventions, are not at stake here. This aspect with be further discussed in "Conclusion" (below).6

Distribution of Hint Substrategies in Learners' Data As noted above, although Hints are essentially nonconventional in form and hence highly context-embedded, a qualitative analysis of CCSARP Hints reveals a consistent semantic pattern of Hint substrategies, differing in terms of opacity types and ranging in degree of opacity. Type classification is based upon the distinction between illocution and proposition, and is mainly a matter of focus rather than a clear-cut distinction. The degree of opacity is a theoretical construct, determined in terms of the (assumed) disambiguation procedures required for full interpretation of the speaker's meaning. Weizman (1989) classified and ranked native speakers' substrategy preferences on both the prepositional scale and the illocutionary scale. Beginning with the propositional scale, we can rank the native speakers' preferences from least to most opaque. Least opaque of the substrategies is reference to the requested act, where S explicitly refers to the act, but there is no reference to the effect that H is responsible for its performance, as in I haven't got time to clean up the kitchen. Slightly more opaque is reference to H's involvement, where S refers indirectly to H's responsibility, but does not name the requested act, as in You've left the kitchen in a mess.

The next substrategy is reference to related components, where S refers to some object semantically related to the requested act, as in The kitchen is in a terrible mess.

131

Intel-language Requestive Hints Table 6.4. Distribution of Hints in Hebrew Learners' Data by Length of Stay in Israel Length of stay (years) <1

N

n

%

163

55 14

33.7 32.6

11 3

35.5 33.3

1-2 2-5 5-8

43 31 9

8-10

6 24

>10

2

33.3

8

33.3

Note: n Hints were used by A' subjects.

Of greatest opaqueness is the substrategy of zero propositional content, where no reference is made to any of the propositional components related to the requested act. Examples include: Are you going to give us a hand? Are you going to do something for me? Attention!

Substrategies on the illocutionary scale can also be ranked from least to most opaque. Least opaque is questioning H's commitment, where S checks whether H feels committed to carrying out some act, the beneficiary of which is to be S. The nature of the requested act, however is not specified, nor is it hinted at. Examples include: Are you going to give us a hand? Are you going to do something for me?

Somewhat more opaque is questioning feasibility, where a reference is made, indirectly or partially, to some precondition for the feasibility of the act. Examples include: Do you have a car? Are you going directly home? Have you got your note with you?

Yet more opaque is stating potential grounders, where a statement made by S is a potential argument for the performance of the requested act, and may therefore be interpreted as the reason for that request. Examples include: I've just missed my bus. I live near your place. The kitchen is in a bit of a mess. Madam, we have an injured person; this car is in the way.

Of greatest opaqueness is the substrategy of zero illocutionary components, where no reference is made to the nature of the illocutionary act, and hence the interpretation of both the illocution and the propositional content is entirely contextembedded, as in Attention!

132

Speech Act Realization

In this section, learners' use of Hints is compared with native speakers' use in terms of substrategy selection. Specifically, it raises the questions whether HL and HN use the same substrategies, and whether they use them with similar frequency. Table 6.5 presents the selection of Hint substrategies by Hebrew learners. The learners predominantly opt for the substrategy of Stating potential grounders (54.2%, « = 52). Their second choice is the combination of Stating potential grounders and Questioning feasibility (18.8%, n = 18), as in: Hi there. Are you going home now? I've just missed my bus, Hi. How are you doing? I missed yesterday's class. Do you have your notebook with you?

Only in the third place do we find the intermediate Questioning feasibility (12.5%, n = 12). The similarities and differences between learners and native speakers are evident from Table 6.6, which presents the distribution of the three predominant options in three populations: Hebrew learners (HL), native speakers of Hebrew (HN), and native speakers of Australian English and Canadian French (NS). First of all, the most opaque illocutionary substrategy is obviously preferred by all three groups (63.4% HL, 53.2% HN, 75.7% NS). This preference provides further support for the high-deniability hypothesis: Hints are exploited precisely by virtue of their opacity. Thus, if Hints are deemed situationally acceptable, the requester tends to opt for their most opaque variant, rather than trying to compensate for their inherent opacity by choosing a relatively transparent substrategy. However, Table 6.6 also reveals an important difference between learners and native speakers. While Hebrew learners turn to redundant combinations as their second choice (22%, n = 18), both HN and NS select them as their third choice Table 6.5. Distribution of Hint Substrategies in Hebrew Learners' Data7 Illocutionary scale Reference to requested act

Reference to H's involvement

Reference to related components

Zero

Questioning feasibility

0

11.5

0

0

1.0

12.5 (12)

Stating potential grounders

2.1

8.3

21.9

0

21.9

54.2 (52)

Zero

0

0

0

1.0

0

Other

1.0

7.3

2.1

0

3.1

1.0 (1) 13.5 (13)

Potential grounders & questioning feasibility

0

1.0

1.0

0

16.7

18.8 (18)

1.0 (1)

42.7 (41)

100 (96)

Prepositional scale (%)

Total

3.1 (3)

28.1 (27)

25 (24)

Various combinations

Total

133

Interlanguage Requestive Hints Table 6.6. Distribution of the Three Predominant Hint Substrategies in Hebrew Learners (HL), Native Speakers of Hebrew (HN) and Native Speakers of Australian English and Canadian French (NS) HL (N = 82)

HN (N = 62)

NS (N = 111)

Stating potential grounders

63.4% (n = 52)

53.2% (n = 33)

75.7% (n = 84)

Stating potential grounders and questioning feasibility

22% (n = 18)

14.5% (n = 9)

(n = 1)

Questioning feasibility

14.6% (n = 12)

32.3% (n = 20)

18% (« = 20)

Substrategies Subjects

6.3%

Note: Of N hints, (n) are of the substrategy under consideration.

(14.5%, n = 9 and 5.2%, n = 7 respectively). These differences, statistically significant (HL differ from HN at p < 0.05 [P value = 0.0369, chi-square = 6.5965], and from NS at/? < 0.01 [P value = 0.006, chi-square = 10.243], seem to be related to a most striking interlanguage specific feature: learners' preference for "verbosity" (Levenston 1971). In the realization of requests, learners' tendency toward verbosity was mainly observed in the use of supportive moves (explanations and justifications for the request) by learners of Hebrew and American English (Blum-Kulka & Olshtain, 1986; Blum-Kulka & Levenston, 1987) and by Danish and German learners of English (House & Kasper, 1987; Edmondson & House, 1991; Faerch & Kasper, 1989; House, 1989). A typical example for Hebrew speakers in English L2 would be I went over the material we will study in the next weeks and I would rather like to have your lecture notes next week, if it's possible and if you can be ready. (BlumKulka & Olshtain, 1986, 175)

A typical example for German learners of English would be Good evening. Perhaps you've already seen me once. We're living on the same street. You know, my bus has just left, and as I- noticed that you have come by car I was going to ask you whether you could give me a lift. (House & Kasper, 1987, 1283)

Verbosity implies a redundant use of words. Similarly, combinations of Hint substrategies imply a redundant use of substrategies, since each of the substrategies may be used in an independent position as a self-sufficient requestive Hint (Weizman, 1989, 88). It is suggested, therefore, that HL preference for redundancy is comparable with learners' preference for verbosity: if indeed verbosity is exploited by learners as part of their attempt to cope with communication problems—mainly insecurity resulting from the learning situation—then so is redundancy: both may be considered as typically "strategic," that is, related to the context of language learning.

134

Speech Act Realization

Conclusion From a theoretical point of view, I suggested that the most striking feature of requestive Hints is that lack of conventions of form and hence context-embeddedness, reach a degree so high that deniability potential is always secured. The peculiarity of this feature, I pointed out, resides in its apparent mismatch with the concept of communication as a goal-oriented activity carried out by a rational agent. It is therefore the task of research into nonconventional indirectness to explain what goal could be efficiently pursued by the use of Hints. Researchers of interlanguage, I further hypothesized, might find the task more rewarding, were it to be found that this very deniability potential is liable to be exploited by learners to achieve a goal specific to the learning situation: the goal of overcoming unease and insecurity by avoiding unambiguous commitments to culture-specific meanings of conventions of form. In light of these assumptions, the findings support the following conclusions: 1. Basically, learners seem well able to use Hints for the realization of requests: they do not use them in any markedly deviating manner as regards frequency, dependence upon situationally specific features, and preference for opaque Hint substrategies. The use of nonconventional indirectness, therefore, seems to be one of the pragmalinguistic essentials with which learners come to L2 and which, therefore, they need not acquire anew (Blum-Kulka, 1982). 2. The hypothesis that the very lack of linguistic constraints inherent in Hints is exploited by learners to avoid risks is, at first sight, not supported by the findings, since learners do not reveal any marked preference for Hints, not even when they are first exposed to the target language. The issue, however, is more complicated. In various areas, learners were observed to exhibit general preference for putting information "on record." Thus, for instance, intermediate learners of English were shown to prefer propositional explicitness to ellipsis (Kasper 1981, 1982. For further discussion, cf. Kasper 1989, 245). If the tendency toward explicitness is indeed an interlanguage specific feature, one would expect to find learners more reluctant to use Hints than native speakers. In other words, along this line of argument, Hints would be less frequent in learners' data than in native speakers'. It could therefore be the case that similar use of Hints by native speakers and by learners reflects an interplay between two conflicting tendencies: on the one hand, preference for explicitness, encouraging a decrease in the frequency of Hints; and on the other, risk avoidance, entailing an increase in the frequency of Hints. The results, then, could be interpreted as indicating that these two tendencies counterbalance each other. This hypothesis should be empirically tested. 3. Learners' tendency toward redundancy is no doubt a strategy of communication. This peculiarity is related to earlier findings concerning learners' preference for verbosity; I interpret it, therefore, as a typical feature of the learning situation. 4. The findings have an important implication for the nature of indirectness in general. The fact that the same Hint substrategies are selected by learners and by native speakers, and that situational variation is similar in both populations, supports the assumption put forward in Weizman (1989): namely, that although apparently "open-ended," even nonconventional indirectness is governed by semantic norms. Moreover, learners' preference for the most opaque Hint substrategy further

Interlanguage Requestive Hints

135

supports the hypothesis (Weizman, 1989) that opacity in Hints is not used as the last possible choice. Rather, it is purposely exploited as the most efficient means when deniability is deemed necessary.

Acknowledgments I am indebted to Gabi Kasper and Shoshana Blum-Kulka for comments on an earlier version; to Hadassa Kantor for helping me obtain additional Hebrew learners' data, incorporated for my purposes into CCSARP learners' data; to Joel Walters for discussions of aspects of interlanguage; to Tsipi Parnassa for the statistical analysis; and to Miriam Schlesinger for stylistic comments on the last version of the paper.

Notes 1. This bi-dimensional analysis implies that the category of Hints encompasses several substrategies. The category includes Ervin-Tripp's (1976) "hints" and "non-explicit question directives," also labeled "nonconventional instrumental moves" or NCI's (Gordon & ErvinTripp, 1984). It further includes Walters's (1979, 286, 287) SVOs and Donde (where) questions. 2. Cross-cultural variation is highly marked in all the quoted studies. For example, Walters (1979, 286, 287) notes a cultural difference in the scale value of politeness of SVOs (e.g., "you have any ball and rope?"), which parallel roughly my "Questioning feasibility: Reference to hearer's involvement and to requested object." In Spanish, SVOs are located near the politeness end of the scale, that is, near conventional indirectness, while in American English they are not far from the impoliteness end. Blum-Kulka (1987) indicates that in American English Hints follow the typically conventional Query preparatories (Can you . . . ? Could you . . . ?) as the next most polite strategy, while for Hebrew speakers, Hints are not only less polite than query preparatories but also less polite than performatives ("I want you to . . ."). 3. These semantic preferences, I suggested, have to do, indirectly and partially, with some requirements presupposed by the felicity conditions, or with some constituents thereof (Weizman, 1985, 1590; whereas in conventional indirectness (e.g., in "can you" questions), the requester refers to the preparatory conditions explicitly and fully. 4. Situational variation was analyzed for HL and HN separately, by means of the Chocran test, which takes into account only fully completed questionnaires. On this analysis, S5 was found to be significantly different (alp < 0.05) from SI, S7, and S l l , while S15 was significantly different (at p < 0.05) from S l l . Differences between both populations were analyzed for each situation separately, by means of chi-square analysis. 5. The total number of Hints («) is a dependent observation, since a considerable number of respondents were found to use Hints in at least two situations. In Table 6.2, for example, 15.4% of the 65 HN Hints and 10.1% of the 109 HL Hints are dependent. Consequently, chisquare analysis is not appropriate for the total. 6. Note, however, that while 17% (n = 8) of the members of the first group (length of stay up to one year) used hints in two situations, members of the other groups systematically chose no more than a single Hint each. This might suggest a decrease in Hint selection over time, but given the relatively small number of group members, further investigation is required.

136

Speech Act Realization

1. The total number of Hints in HL data as presented in Tables 6.5 and 6.6 (n = 96) differs slightly from the total number of HL Hints in Table 6.2 (n = 109), due to a difference of nuance between the initial CCSARP definition of Hints and the one suggested in Weizman (1985, 1989). Note, however, that this difference has no bearing on the claims made about the similarities between learners and native speakers in terms of the selection of Hint substrategics.

References Blum-Kulka, S. (1982). Learning to say what you mean in a second language: A study of the speech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-60. Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal of Pragmatics, 11, 145-60. Blum-Kulka, S. (1989). Playing it safe: The role of conventionality in indirectness. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Request and apologies (37-70). Norwood, NJ: Ablex. Blum-Kulka, S., Danet, B., & Gerson, R. (1985). The language of requesting in Israeli society. In J. Forgas (Ed.), Language and social situation (113-41). New York: Springer-Verlag. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requesting behavior. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (123-54). Norwood, NJ: Ablex. Blum-Kulka, S., & Levenston, E. (1987). Lexical-grammatical pragmatic indicators. Studies in Second Language Acquisition, 9, 155-70. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Studies in Second Language Acquisition, 8, 165-79. Brown, P., & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In E. A. Goody (Ed.), Questions and politeness: Strategies in social interaction (56— 311). Cambridge: Cambridge University Press. Clark, H. H. (1979). Responding to indirect speech acts. Cognitive Psychology, 11, 430-77. Corder, S. P. (1983). Strategies of communication. In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication (15—19). London: Longman. Dascal, M., & Weizman, E. (1987). Contextual exploitation of interpretation clues in text understanding: An integral model. In J. Verschueren & M. Bertucceli Papi (Eds.), The pragmatic perspective: Selected papers from the 1985 International Pragmatics Conference (31-45). Amsterdam: Benjamins. Edmondson, W., & House, J. (19191). Do learners talk too much? The waffle phenomenon in interlanguage pragmatics. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds), Foreign!second language pedagogy research (273-87). Clevedon and Philadelphia: Multilingual Matters. Ervin-Tripp, S. (1976). Is Sybill there? The structure of some American English directives. Language in Society, 5, 25-66. Faerch, C., & Kasper, G. (1983). On identifying communication strategies in interlanguage production. In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication (210-38). London: Longmans. Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request

Interlanguage Requestive Hints

137

realization. In S. Blum-Kulka, J. House, and G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (221-47). Norwood, NJ: Ablex. Eraser, B., Rintell, E., & Walters, J. (1980). An approach to conducting research on the acquisition of pragmatic competence in a second language. In D. Larsen-Freeman (Ed.), Discourse analysis in second language research (75-91). Rowley, MA: Newbury House. Gordon, D., & Ervin-Tripp, S. (1984). The structure of children's requests. In R. Schieffelbush & A. Pickar (Eds.), The acquisition of communicative competence (295321). Baltimore, MD: Baltimore University Book Press. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and semantics (Vol. 3): Speech acts (41-58). New York: Academic Press. House, J. (1986). Cross-cultural pragmatics and foreign language teaching. In K. R. Bausch, F. G. Konigs, & R. Kogelheide (Eds.), Probleme und Perspektiven der Sprachlehrforschung (281-95). Frankfurt: Scriptor. House, J. (1989). Politeness in English and German: The functions of Please and Bitte. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (96-119). Norwood, NJ: Ablex. House, J., & Kasper, G. (1987). Interlanguage pragmatics. Requesting in a foreign language. In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance. Festschrift for Werner Mullen on the occasion of his 60th birthday (1250-88). Tubingen: Narr. Kasher, A. (1982). Gricean inference revisited. Philosophica, 29, 25-44. Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr. Kasper, G. (1982). Teaching-induced aspects of interlanguage discourse. Studies in Second Language Acquisition, 4, 99-113. Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters. Leech, G. (1980). Explorations in semantics and pragmatics. Amsterdam: Benjamins. Leech, G. (1983). Principles of pragmatics. London: Longman. Levenston, E. (1971). Over-indulgence and under-representation—aspects of mother tongue interference. In G. Nickel (Ed.), Papers in contrastive linguistics (115-21). Cambridge: Cambridge University Press. Rintell, E. (1981). Sociolinguistic variation and pragmatic ability: A look at learners. International Journal of the Sociology of Language, 27, 111-34. Searle, J. R. (1975). Indirect speech acts. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics (Vol. 3). Speech acts (59-82). New York: Academic Press. Vanderveken, D. (1984). What is an illocutionary force? In M. Dascal (Ed.), Dialogue: An interdisciplinary approach (185-208). Amsterdam: Benjamins. Walters, J. (1979). Strategies for requesting in Spanish and English: Structural similarities and pragmatic differences. Language Learning, 29, 277-93. Walters, J. (1981). Variation in the requesting behavior of bilingual children. International Journal of the Sociology of Language, 27, 35-58. Weizman, E. (1985). Towards an analysis of opaque utterances. Theoretical Linguistics, 12, 153-63. Weizman, E. (1989). Requestive Hints. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (71-95). Norwood, NJ: Ablex. Weizman, E., & Dascal, M. (1991). On clues and cues: Strategies of text-understanding. Journal of Literary Semantics, XXI1, 18-30.

7 Cross-Linguistic Influence in the Speech Act of Correction TOMOKO TAKAHASHI and LESLIE M. BEEBE

In this chapter, we examine American and Japanese performance of the speech act of correction by looking at how this speech act is performed with status unequals— a person of lower status addressing someone of higher status and a person of higher status addressing someone of lower status. By focusing on correction, we are not looking at opinions, but rather at situations where one person knows the other has made a factual error. By looking at situations between a professor and a student, we are choosing situations where the roles are clearly defined and the status relationships are relatively clear-cut. The difference in status here is not necessarily socioeconomic, but rather a matter of the power structure within the classroom. In previous research, we have investigated American versus Japanese performance on several face-threatening speech acts such as disagreement (Beebe & Takahashi, 1989a,b), refusals (Beebe, Takahashi, & Uliss-Weltz, 1990; Takahashi & Beebe, 1986, 1987), announcing embarrassing information (Beebe & Takahashi, 1989a), and chastisement (Beebe & Takahashi, 1989b). The present study continues our research project on sociolinguistic variation in face-threatening speech acts. Face-threatening speech acts have received much attention (e.g., Brown & Levinson, 1978) and have been found important because they are the source of cross-cultural miscommunications, or what Thomas (1983) calls "pragmatic failure." In this domain, we have focused on American and Japanese speakers because their cultures are vastly divergent; thus they were expected to show a large difference in the use of politeness strategies (e.g., Ide, 1989; also see Kasper 1990, for a review of literature on politeness; Blum-Kulka, House, & Kasper, 1989, for research on requests and apologies). In our research, we have analyzed each speech act as consisting of a sequence of "semantic formulas" (Cohen & Olshtain, 1981) such as the positive remark (e.g., "I wish I could, but . . ."), the expression of regret (e.g., "I'm sorry"), the refusal itself (e.g., "I can't come"), and so on. In examining the number, order, and/or 138

The Speech Act of Correction

139

content of semantic formulas used in several face-threatening speech acts, we have found differences in the realization of speech acts by Americans and Japanese. We shall focus here on the use of positive remarks, softeners, and other similar formulas. These formulas have been found particularly interesting because they are frequently used by Americans and/or Japanese in order to make each speech act less face-threatening. The definition of politeness varies from study to study (e.g., Brown & Levinson, 1978, 1987; Ide, 1989; Lakoff, 1973; Leech, 1983). It is, however, generally agreed that politeness aims at smooth communication (Ide, 1989), seeks to reduce friction (Lakoff, 1973), and to disarm the potential for aggression (Brown & Levinson, 1987). Along the same line, politeness is viewed here as the language usage developed in order to make a situation less facethreatening and to make communication smoother. In this sense, we consider positive remarks and softeners to be important politeness strategies. One of the major focuses of our research has been to examine the cross-cultural influence of Japanese upon English as a second language (ESL) at the discourse level. In this domain, we have found much evidence of native language (NL) transfer in Japanese ESL speakers as well as other examples of their efforts (successful or unsuccessful) to try to sound like English native speakers. Another focus of our research has-been on how Americans and Japanese perform such speech acts with status unequals. This question was asked because it is generally claimed that Japanese are very conscious of social status while Americans are relatively less status-conscious. Ide (1989), for example, argues that in many honorific languages such as Japanese "the choice of expression is made according to the variable of power and distance of the addressees" (Ide, 1989, 238). Similarly, Matsumoto (1989) argues that in Japanese, social context plays a much larger role than is generally assumed. Barnlund and Araki (1985, 23) also support the view that "status relations are critical in determining the character of communicative behavior in Japan (Nakane, 1970; Nomura & Barnlund, 1983) and are less critical in the United States (de Tocqueville, 1965; Nomura & Barnlund, 1983)." In their study of compliments by Japanese and Americans, Barnlund and Araki (1985, 23) have found that "Japanese exchanged compliments less frequently in close relations and more often in relations that were more distant," while "with Americans it appeared that compliments might be exchanged more often in close relationships and less in more distant ones." It is thus suggested that both Japanese and Americans style-shift according to the psychological distance of the interlocutors. Based on the view that the Japanese are more conscious of social status than Americans, it can be further hypothesized that Japanese style-shift more than Americans according to the social as well as psychological distance of the addressees. In the investigation of sociolinguistic rule transfer at the phonological level, Beebe (1980) and Schmidt (1977) have found that sociolinguistic style-shifting patterns are transferred from the NL to the second language (L2). By focusing on the discourse level here, we discuss the differences among native English, Japanese ESL, and native Japanese responses, and ask whether and to what extent Americans and Japanese native speakers style-shift and whether to what extent Japanese ESL learners transfer patterns of sociolinguistic style-shifting from Japanese to English.

140

Speech Act Realization

In sum, we discuss in this paper: (1) the use of politeness strategies that the speaker uses in order to make each correction less face-threatening (i.e., more polite)—e.g., positive remarks, softeners, expressions of regret; (2) NL transfer in the Japanese ESL responses; and (3) style-shifting by the three groups according to the status of the interlocutor. The data will be presented, for the most part, descriptively in terms of semantic content in order to give the flavor of responses by the three groups in an encounter of interlocutors of different status. Method The discourse completion questionnaire used for this study consisted of 12 items— two situations for each of the following: correction, disagreement, chastisement, announcing embarrassing information, and two other speech acts as controls (see Appendix for descriptions of the 12 situations). For each pair eliciting one type of speech act, one situation involved a person of lower status talking to someone in a higher position and one involved someone in a higher-status position talking to a person of lower status. The situations were presented in randomized order. In this paper, we shall discuss only correction. There were 55 subjects who filled out the questionnaire—15 Americans who responded in English (AE), 15 Japanese who responded in English (JE), and 25 Japanese who responded in Japanese (JJ). The Japanese subjects tested in English had high-intermediate to advanced levels of ESL proficiency. All subjects were college graduates. For the Americans, there were 11 females and 4 males, with an average age of 33. For the Japanese ESL subjects, there were 8 females and 7 males, also with an average age of 33. For the Japanese tested in Japanese, there were 11 females and 14 males, with an average age of 32. The American and Japanese ESL subjects were tested in New York, and the Japanese subjects using Japanese were tested in Tokyo, Japan. All the responses were analyzed as consisting of a sequence of "semantic formulas" (Cohen & Olshtain, 1981) or bits of meaning such as the expression of regret and the correction statement itself. For each of the subject groups, the total number of semantic formulas of any kind used for each situation was obtained. Results and Discussion Correction Situation 1 (Higher to Lower) Correction situation 1 read as follows: You are a professor in a history course. During class discussion, one of your students gives an account of a famous historical event with the wrong date.

In this situation, we are dealing with an utterance that will be made by a person in a higher-status position (the professor) to a person in a lower-status position (the student). What is common to all three groups is that the majority of subjects decided

The Speech Act of Correction

141

to correct the student. One American opted out, saying "thank you." One Japanese using English said to the class, "Does everyone agree with that date?" but this was clearly an attempt to elicit peer correction. One Japanese using Japanese opted out, and another wrote "I will wait until other students notice the mistake." The most noticeable difference is that 9 out of 14 Americans (64%) prefaced their correction with at least one positive remark: most typically, "That was very good but I believe that took place in 1945." (Other responses were quite similar, with a variety of adjectives being used: "excellent," "great," "marvelous," etc.) Two American respondents used more than one positive remark: "That was a great account of the attack on Pearl Harbor. Everything was in line except the date. It was 1942, not 1943," and "Excellent description, Henry! I like the way you outlined the events. Now when did all that take place?" We have thus found 11 positive remarks (79%) used by the American group. Japanese ESL speakers, on the other hand, never once prefaced their responses with a wholehearted positive remark. Most typical of their responses was: "I think that date is not correct," or "The date is wrong," or simply, "In 1972." Only 3 out of 13 Japanese ESL speakers (23%) used something resembling the American positive remark. But upon reading these remarks to a group of American native speakers, we were backed up in our intuition that the remarks were so lukewarm that a native speaker would hardly call them "positive" or feel comfortable with them. The remarks were: "Well, I'm almost satisfied with your account of that event except the date of it," "Your ideas is fine, but I think the age you mentioned is something wrong," "Thank you for bringing up interesting account, although the date you just mentioned was actually 1919." Japanese using Japanese used even fewer positive remarks: only 3 out of 23 (13%) had the professor make a positive comment to correct his student's mistake. A typical Japanese response in Japanese (87% of the data) had no prefacing positive remark at all (e.g., "It's 1603," or "Wait a second. The date is incorrect," or "The date you just mentioned is incorrect. Please check it by the next class.") The three who used positive remarks used ones very similar to American English patterns. The remarks were: "Very nicely done, but why don't you check the date you just mentioned," "Your presentation was very good, but it was supposed to have taken place in 1549." "It was a very good presentation, but just one mistake—the date." The percentage of positive remarks used by the three groups is presented in Figure 7.1 The finding is that Americans use by far the highest percentage of positive remarks and Japanese using Japanese use the lowest with the Japanese speakers of English in between: AE > JE > JJ. This suggests that the percentage of positive remarks used by Japanese ESL learners is influenced by Japanese, where positive remarks are evidently not critical in this particular situation. Positive remarks (including praise, complimenting, and positive evaluations) are extremely important prefixes to face-threatening acts in English. In our research on refusals (Beebe et al., 1990), one of the striking differences between Japanese and English native speakers was that Americans felt the need to preface their refusals with what we called "positive adjuncts": for example, "I'd love to, [but] . . . " "That sound wonderful, fbutj . . ."In our studies of disagreement (Beebe & Takahashi, 1989a,b), we also found that Americans often used expressions of "token agree-

142

Speech Act Realization

Fig. 7.1. Percentage of positive remarks in correction situation I, where the professor corrects the student's mistake (higher to lower). Note: These figures exclude the subjects who opted out and those who described what they would do instead of writing what they would say. Key: AE—Americans using English JE—Japanese using English JJ—Japanese using Japanese

merit" (Brown & Levinson, 1978): for example, "I agree with you, [but] . . ." "Yes, [but] . . . " (also see LoCastro, 1986). Similarly, as we have found in this study, Americans use a positive remark to preface a correction to a lower-status person. This is an "adjunct" to the correcting act because although it is very common (used by 64% of our American respondents in this situation), it cannot stand alone and accomplish the speech act of correction. Just as "I'd love to . . ." is a vital to polite refusal in many situations but unable to stand alone as a refusal, a positive remark such as "It was a very good account" softens the upcoming correction but cannot substitute for it in any way. We thus maintain that it, too, is an "adjunct" to a correction. In addition to the use of positive remarks, American respondents seem to have made more efforts to soften their corrections than their Japanese counterparts. They tried, for instance, to soften the correction and make it the least face-threatening possible by using the following types of softening devices 71% of the time: (1) "softeners"—that is, expressions (or hedges) such as "I believe," "1 think," "You may have (the wrong date)", (2) questions such as "Did you say . . . ?" "When did that happen?", and (3) other expressions intended to lighten the gravity of the interlocutor's mistake or to defend the interlocutor, such as "You made one small error in the date." Japanese ESL respondents also used similar formulas: for example, (1) "I think," "I'm afraid" as softeners; (2) "Isn't it 1945?" "That was in 1492, wasn't it?" as questions; and (3) "I guess it's not very important" to lighten the gravity of the

The Speech Act of Correction

143

mistake. However, Japanese respondents using English used softeners only 50% of the time, as compared with 71% use by Americans. Japanese responding in Japanese also used the equivalents, yet much less frequently (26% of the time): for example, (1) "I think"; (2) "Was it 1968?"; and (3) "It might be a slip of the tongue." The differences among the groups in the use of all types of softening devices are presented in Figure 7.2. Essentially, the pattern for softeners is the same as that for positive remarks: AE > JE > JJ. That is, the largest percentage (71%) of softeners was used by the Americans, who used more than the Japanese using English (50%). And the Japanese using English used more than the Japanese using Japanese (26%). The Japanese using English result, being midway between the native speakers of Japanese and English, suggests cross-linguistic influence from the native language, Japanese, on the English of Japanese. As we compare Figure 7.2 with Figure 7.1, we notice that there is a smaller gap among the three groups in the use of softeners (Figure 7.2) than in the use of positive remarks (Figure 7.1). It has been pointed out that the Japanese tend to place value on silence and to be skeptical about the efficacy of communicating delicate thoughts or feelings through strictly verbal strategies. Clancy (1986), for example, has noted that "[the] Japanese have little faith in verbal expression or in those who rely upon it" (214). It is generally observed that the Japanese try to express regret, for instance, through nonverbal and/or paralinguistic means such as facial expres-

Fig. 7.2. Percentage of softeners in correction situation I,* where the professor corrects tne student's mistake (higher to lower). Note: *Includes softeners of three types, including questions and other softening expressions. **These figures exclude the subjects who opted out and those who described what they would do instead of writing what they would say. Key: AE—Americans using English JE—Japanese using English JJ—Japanese using Japanese

144

Speech Act Realization

sions, tone of voice, sighs, hesitance, and so forth. Although our discourse completion questionnaires are not designed to collect such expressions, we have often found that Japanese native and ESL speakers used pauses to show hesitance or silence (see, e.g., Beebe et al., 1990). This could be seen as evidence that Japanese prefer less verbally explicit expressions, especially in face-threatening situations. Americans, on the other hand, seem to rely on verbal expression and use a clearly verbal strategy for dealing with a threat to face. American culture promotes explicit verbal means of correcting, refusing, or disagreeing and thus encourages explicit verbal means of undoing the threat to face that the explicit correction, refusal, or disagreement poses. This results in a strong tendency toward a ritualized use of the positive remark. In other words, the positive remark is a very important politeness strategy for Americans. The Japanese rely much less on the positive remark than Americans. Instead, they use softeners, which Americans also use. The softener is different from the positive remark in that the former is a down-toning device integrated in the main body of a speech act (e.g., / think [softener] it's 1942, not 1943 [main body of the speech act of correction]), while the positive remark is a preceding adjunct that is phrasal and separate from the main body (e.g., It was a good presentation, but . . .). It seems, therefore, possible that the Japanese use the positive remark much less often because of its explicitly verbal nature. It is also possible that the positive remark is avoided because it may sometimes sound too untruthful for people who have little faith in verbal expression. For Japanese ESL learners, it is much easier to transfer the NL knowledge in the use of the softener than that of the positive remark. Apart from the use of these formulas, the overall tone of the Japanese responses in Japanese is quite different from that of American responses. For instance, 6 out of 23 Japanese speakers using Japanese (26%) chose to point out directly that there was a mistake in the presentation (e.g., "The date you just mentioned is incorrect") and then to say, for example, "Please check the date" or "Why don't you check the date?" All in all, Japanese responses sounded much more authoritarian than American responses. They had a ring of superiority that came from the use of a directive to check the date and from its position following a direct statement that a mistake had been made.

Correction Situation 2 (Lower to Higher) Let us now look at situation 2, which reads as follows: You are a student in a sociology class. During the lecture, the professor quotes a famous statement attributing it to the wrong scholar.

In this situation, the status of the participants was reversed; it was the professor who made the mistake. Therefore, any correction attempted would have to be from the lower-status interlocutor (the student) to the higher-status interlocutor (the professor). The most striking difference between the two status situations was that no native Americans or Japanese ESL speakers used a positive remark to a higher-status person before a correction ("It was a very interesting lecture"), and only one Japa-

The Speech Act of Correction

145

nese using Japanese did so. It was evidently judged totally inappropriate to praise the professor and then correct. Thus, only one out of 37 subjects (who opted to say something) used a positive remark (or positive remarks) from student to professor, whereas in the opposite situation from professor to student, 15 subjects out of 50 (who opted to say something) used a positive remark. This shows that all groups were highly sensitive to status differences in this regard. Avoidance or "opting out" (Bonikowska, 1988) was indeed possible since the overall directions to the questionnaire made it clear that saying nothing was an option for any of the situations. Two Americans opted out completely. Another wrote, "Probably nothing, or I would say, 'Are you sure you have the correct name? Isn't it . . . ? ' " And another wrote, "Probably nothing, but in a small seminar I might say something like 'You mean Ellis, don't you?' especially if anyone else looked confused." Still another American commented, "If I did not know the professor well, I wouldn't say anything. If I did know the professor, I would approach him after class and say, 'Wasn't it X who said Y?'" Three Japanese ESL speakers opted to say nothing. Another said, "I would not take prompt action. I would observe others' reactions, and if nobody seemed to have noticed, I would say something." What was interesting about those who opted out was that two out of three Japanese ESL subjects who said they would say nothing specified that they expected the professor to recognize his mistake. Americans did not seem to have this reason for keeping quiet. Japanese responding in Japanese were the ones who avoided the difficulty of correcting the professor most: 10 out of 25 (40%) decided to opt out, whereas 3 Japanese ESL learners (20%) said they would say nothing and 2 Americans (13%) opted out. The dominant pattern used by all three groups in this situation (from the student to the professor) was, again, to use a softener (or softeners) such as "I think that was . . ." before a correction. The percentage of softeners used by the three different groups is presented in Figure 7.3. What is most interesting in the differences between Japanese ESL and native American responses is not the frequency of softeners used or the sequence of formulas that the respondents chose to use. It is rather the wording of the softener or the correction formula that is of particular interest. Books comparing Japanese and American culture stress that Americans value variety and originality in verbal expression as a way of demonstrating sincere feeling as opposed to perfunctory politeness (see Condon, 1984; Goldstein & Tamura, 1975). They warn Americans not to try to transfer this into their Japanese. Japanese culture prescribes certain formulaic expressions that can be used over and over again in appropriate situations without needing to be varied or elaborated upon to convey politeness (see Coulmas, 1981; Loveday, 1982). The American value placed on variety is not shared by the Japanese. On the contrary, knowing and using the correct formula is what counts in Japanese politeness (see Beebe & Takahashi, 1989a,b, for similar findings). Clancy (1986, 216) explains why the Japanese depend on formulaic expressions as follows: In Japan, there seems to be an extensive codification of contexts in which particular feelings are expected; speakers need only indicate, by means of the right

146

Speech Act Realization

Fig. 7.3. Percentage of softeners in correction situation II, where the student corrects the professor's mistake (lower to higher). Note: *These figures exclude the subjects who opted out and those who described what they would do instead of writing what they would say. **Some subjects used softeners more than once. Key: AE-Americans using English JE—Japanese using English JJ—Japanese using Japanese formula, that they are experiencing the appropriate reaction, without expressing any more personal, individualized response. An important goal of socialization in Japan is to promote the unanimity in feeling that will support the norms of verbal agreement and empathy.

She further notes that Japanese mothers provide instruction in verbal formulas and expect from their children much earlier mastery of social courtesy than American mothers (Clancy, 1986, 236). Ide (1989) would argue that Japanese has a much higher degree of "discernment" (obligatory choice) in expressing politeness; that is, the choice of formal linguistic forms such as honorifics or morphological marking on the verb according to social conventions. American English, on the other hand, is high in "volition" (strategic choice); that is, the choice of interactional strategies such as jokes or questions according to the speaker's intention. The Japanese usually search for the appropriate linguistic formulas or politeness marking according to the variable of power and distance of the addressees, while American English speakers are allowed to make a considerably more active choice from a relatively wider range of possibilities. In other words, Japanese and American English are different in that the speaker's focus is placed on the socially prescribed norm in Japanese and on the speaker's own intention in American English. Japanese ESL learners' use of formulaic expressions, therefore, can be explained as transfer of a belief in the efficacy of choosing the appropriate expression according to relative status. This would be another reason why Japanese prefer formulaic expressions to more personal and individualized responses.

The, Speech Act of Correction

147

In this research, the difference in softeners used by Japanese and American respondents may reflect these purported cultural differences. Americans seemed to be elaborating their softeners, by using more self-deprecating softeners with a greater variety to wording and a preference for the "[clause] + but . . . " structure such as "I may be wrong/mistaken, but . . ." and "I'm not sure, but . . ." Then, they said, "I think that . . ." One American used a combination of softeners: "I'm curious about the quote you read. I was under the impression White said that. Am I wrong?" Japanese ESL respondents tended to stick to short, formulaic softeners, such as "I'm afraid," "I think," or "I understand." Their responses might have been shorter and less varied simply because they were using a second language. Empirical investigation of non-Japanese learners at the same proficiency levels would be needed to verify this. However, the apparently formulaic brevity is worth examining because natural data have also shown that Japanese ESL speakers overuse "I'm afraid" as a softener. Although it was surely meant to have a polite softening effect when used by Japanese ESL subjects, it was not used at all by English native speakers. Americans might avoid the use of "I'm afraid," especially from student to professor, because it has a slight ring of superiority. ("I'm afraid I made a mistake" sounds appropriate but "I'm afraid you made a mistake" sounds patronizing.) It is highly unlikely that the Japanese ESL speakers meant to convey this tone since they used softeners much more frequently from student to professor than from professor to student. As we look at the responses by Japanese using Japanese, the ESL use of "I'm afraid" becomes a curious case because Japanese native speakers did not use the equivalent of "I'm afraid" in their Japanese. Instead they used the question form predominantly (80%) in order to soften corrections: for example, "Isn't it X who said Y?" or "Isn't that quoted by X?" With respect to formulaic expressions used by native speakers, some studies (Cohen & Olshtain, 1981; Olshtain & Cohen, 1983; Olshtain, 1983; Owen, 1980) report that English native speakers' choices of apology forms are highly patterned. In other words, there are major strategies frequently used to express an apology: for example, "I'm sorry" and "It was my fault" (see Olshtain & Cohen, 1983). Eraser (1981) has found that the more formal the situation, the longer and more elaborate the apology. Thus, it can be said that the degree to which major strategies are actually ritualized may vary in different cultures as well as according to different speech acts and situations. In Japanese culture, however, it seems that formulaic expressions are especially favored. Another interesting finding was that one Japanese ESL respondent used a question form quite different from the dominant questioning pattern used by all three groups; namely, "Isn't it X who said Y?" The Japanese ESL subject said, "Would you please tell me who made this statement again?" This type of seemingly factual question, although it has never been the predominant pattern in native Japanese or Japanese ESL, has cropped up repeatedly in our research on Japanese use of face-threatening speech acts such as warning, disagreement, and chastisement (see Beebe & Takahashi, 1989a,b; Beebe & Takahashi, 1988). On the face of it, this question form could be interpreted as a very polite way of

148

Speech Act Realization

correcting someone, but it is not so polite as it is intended to be. Indirectness is not always highly valued in American speech. This finding is corroborated by BlumKulka's (1987) finding that indirectness and politeness do not necessarily co-vary. In her study of English and Hebrew requests, Blum-Kulka (1987) examined different request categories including hints (e.g., "We don't want any crowding" [as a request to move the car]) and conventional request forms (e.g., "Would you mind moving your car?"). She has found that American native speakers consider hints less polite than conventional indirectness although hints are rated more indirect than conventional indirectness. Similarly, in our research, when told that some Japanese respondents would say nothing to correct the professor, expecting that the professor would see his or her own mistake, Americans were surprised and a little amused. When told that one Japanese said, "Would you please tell me who made this statement again?" a native English speaker immediately piped up, "It sounds as if you're trying to trap him." The use of a question to elicit a self-correction is fairly frequent from a teacher to a student in class, but the request from a student to a professor to repeat something is likely to be interpreted as a simple request. When the professor discovers that it is an indirect way of making a correction, he or she is likely to feel embarrassed or annoyed, particularly if he or she has fallen for it and answered the question on a literal level (see Beebe & Takahashi, 1989a,b, for a discussion of a distinctively Japanese questioning strategy found in natural data). It is generally recognized that the communicative style of the Japanese is indirect and intuitive, especially compared with that of Americans. In her study of Japanese mother-child interaction, Clancy (1986) reports that Japanese mothers frequently use rhetorical questions and hints to express a variety of negative attitudes and opinions such as skepticism, frustration, and disapproval to their children. She also reports that Japanese mothers strongly emphasize sensitivity to the needs, wishes, and feelings of others. Such training serves a purpose in Japanese interaction. This point is well elucidated by Kasper (1990) as follows: "Rather than emphasizing distance, indirectness in Japanese culture appears to express empathy between the participants, symbolizing a high degree of shared presuppositions and expectancies" (200). Upon receiving an indirect request, for example, the Japanese feel that the speaker is trying to make the speech act less face-threatening; they also understand (or at least try to understand) the feelings and needs of the speaker in making such a request. It is, therefore, possible that in Japanese culture, indirectness serves to symbolize empathy. It is also likely that a closer relationship exists in Japanese culture between indirectness and politeness than in American culture.

Two Correction Situations Compared Now let us turn to the data on style shifting and investigate whether the styleshifting pattern of the Japanese using English might be transferred from their native Japanese pattern. As we compare the two status situations, we find a dramatic style shift in the use of softeners among Japanese subjects. Figure 7.4 presents the comparison between the two status situations in the use of softeners by the three

The Speech Act of Correction

149

Fig. 7.4 Percentage of softeners in the two correction situations. Key: AE—American using English JE—Japanese using English JJ—Japanese using Japanese

subject groups. As seen here, the Americans style-shifted the least, while the Japanese both in Japanese and English displayed great style shifting by the use of softeners. This pattern found in the use of softeners within each situation again suggests that there is native language transfer; this time, however, it is not just the use of softeners that shows cross-linguistic influence, it is actually the style shifting itself that Japanese ESL speakers are transferring. This is reminiscent of a finding from our refusal data (Beebe et al., 1990), where Japanese in both English and Japanese style-shifted more than Americans using English in the use of expressions of regret (e.g., "I'm sorry"). Figure 7.5 is based on our refusal data. Similarly, the use of positive remarks prefacing disagreements suggests similar style-shifting tendencies, as seen in Figure 7.6. The data on Japanese using Japanese are not available on disagreements so it is not possible to show in this instance that JJ > JE > AE; however, the finding that JE > AE is consistent with the other findings. Figures 7.4, 7.5, and 7.6 suggest that the lower- to higher-status situation is more face-threatening than the other status situation, higher to lower, especially for

150

Speech Act Realization

Fig. 7.5. Percentage of expressions of regret in the two refusal situations. (Data from Beebe, Takahashi, & Uliss-Weltz, 1990.) Key: AE—Americans using English JE—Japanese using English JJ—Japanese using Japanese

the Japanese in both English and Japanese. More softening devices (softeners, positive remarks, and expressions of regret) are used with the lower-status person addressing the higher-status person than vice versa. The degree of shift is greater for Japanese subjects than for Americans in each case and greater for Japanese using Japanese than Japanese using English in the corrections and refusals where both sets of data are available. In correction, as opposed to disagreement and refusal, there are many situations where elaboration and multiple prefacing positive remarks are not appropriate. It would sound very odd for a student to say, for example, to a professor, "You are a very intelligent person, you know a lot about this subject, and you have certainly earned your reputation as an expert in this area, but that novel was published in 1972, not 1970." The number and character of prefacing remarks depends upon the situation, the status of the participants, and the speech act. An ESL teacher, on the other hand, might compliment a student's book report before correcting a small error, but would not feel compelled in a drill session to give a positive remark about

The Speech Act of Correction

151

Fig. 7.6. Percentage of positive remarks and token agreement in the two disagreement situations. (Data from Beebe & Takahashi, 1989a, b; no JJ data are available in this study.) Key: AE—Americans using English JE—Japanese using English

every sentence the student utters. Thus, the situation is a key variable in the use of positive remarks. Similarly, the social status of the two interlocutors is very important in determining whether a positive remark should be used. A professor might say something positive before correcting a factual error by a student ("That was a good report, but it was in 1942, not 1943") but a student would sound arrogant using a positive remark with a professor ("That was a fascinating lecture, but it was President Carter, not President Ford"). The positive remark often used in corrections is praise of the interlocutor, and such praise usually come from the "superior," not from the lower-status to the higher-status person. The positive remarks often used in refusals and disagreement are an expression of the speaker's positive feeling about the offer or invitation (e.g., "Oh, I wish I could, but . . .") or token agreement with the interlocutor's opinion (e.g., "I agree with you, but . . ." "Yes, but . . ."). This type of positive remark is not direct praise of the interlocutor. It seems, therefore, necessary to distinguish positive remarks in different speech acts. In this study, we have found that none of the subjects (except for one JJ) used a

152

Speech Act Realization

positive remark in the situation where a student corrects a professor, as seen in Figure 7.7. Here we see that all three groups converged in the status situation Lower to Higher. At the same time, Americans style-shifted most in this particular situation. As previously discussed, the positive remark is one of the most important politeness strategies Americans employ in face-threatening situations. This formula was used as often as 79% of the time by Americans in the situation where a higherstatus person is talking to a lower-status person, while it was used by none in the other situation. The shift from 79% to zero is much greater than those found in the performance of the Japanese subjects, who simply do not use the positive remark very often in any situation. As mentioned above, in the speech act of correction, the situation, the relative social status, expertise status, role status or other kinds of status have been found important factors influencing the way the speaker performs. When we engage in a speech act like giving embarrassing information, however, it does not seem to matter so much who has what status. Rather, the prefacing positive remark seems uniformly out of place. We say, "I think you have something in your teeth," not "You are really well groomed today and I know you are an avid tooth-brusher, but you have spinach in your teeth." Some face-threatening acts seem to require more elaboration and some require simplicity. The pattern of style shifting may vary according to many factors. In our research, we have found that, generally, native speakers of Japanese are influenced in the way they perform the speech act of correction in English by the way they perform corrections in Japanese; for example, they style-shift in English according to the status of the interlocutor in a manner similar to the style-shifting typical of native Japanese. Lastly, it should be pointed out that our Japanese ESL subjects had highintermediate to advanced levels of ESL proficiency. In our study of refusals by Americans, Japanese native speakers, and two Japanese ESL proficiency groups (Takahashi & Beebe, 1987), we argued that transfer was greater among highproficiency ESL learners than their low-proficiency counterparts at the discourse level. The lower-proficiency students do not have the fluency in the target language to give free rein to pragmatic transfer phenomena. The high-proficiency students, on the other hand, have the control over English to express Japanese sentiments at the pragmatic level. In her recent study of politeness in Japanese, Matsumoto (1989) has argued that in Japanese, "politeness expressions are not simply additions to a neutral utterance which conveys the prepositional content: rather, the structure of the language requires some choice of expression that conveys additional information on the social context" (208). In other words, Japanese utterances are always marked linguistically for social context (also see Ide, 1989). That is, some utterances may have exactly the same content but vary in linguistic forms according to the social contexts in which the utterances were made. Japanese ESL learners may try to transfer such linguistic markings into English, but are unable to find places to transfer these features. Instead, their utterances in different social contexts may sound uniform in English regardless of differences in the social contexts because the utterances are the same in content and only the content is expressed in English. Advanced learners, on the other hand, find substitutes for the places to transfer these

The Speech Act of Correction

153

Fig. 7.7. Percentage of positive remarks in the two correction situations. Key: AE—Americans using English JE—Japanese using English JJ—Japanese using Japanese

social differences or at least to express such sentiments. In the present study, we have found that ESL learners transfer Japanese style-shifting patterns into English. If we had tested low-proficiency learners, however, we might not have found the same phenomenon.

Conclusion In sum, we have uncovered a number of patterns in the speech act of correction. First, we have demonstrated that it is a typically American pattern to use a positive remark such as "That was a great account" before saying "but" and making a correction when speaking to a person of lower status. Only 13% of the Japanese using Japanese did this whereas 64% of the Americans did. Japanese using English were in between. This pattern can be schematically represented as AE > JE > JJ. It provides evidence that there was cross-linguistic influence from Japanese. This same pattern could be claimed to show incomplete acquisition, however. Second, we have shown that softeners give evidence of cross-linguistic influ-

154

Speech Act Realization

ence in both the situation with the higher- to lower-status interlocutor and the situation with the lower- to higher-status interlocutor. With the professor to the student, we see that softeners were used with correction 71% of the time, and the pattern is: AE > JE > JJ. The exact opposite pattern exists with the student speaking to the professor: AE < JE < JJ. Third, and most important, particularly striking in our data is that style-shifting patterns are transferred from the native language, Japanese. Whereas transfer has been well documented on the phonological and syntactic levels, there was traditionally less interest at the discourse level. Recent studies, however, have shown that transfer at the discourse level is also widespread (e.g., Beebe et al., 1990; Blum-Kulka, 1982, 1983; Cohen & Olshtain, 1981; Olshtain, 1983; Olshtain & Cohen, 1983; Takahashi & Beebe, 1987). What is striking in this study is that it is not only discourse patterns that transfer, but also the style shift between two different discourse patterns with interlocutors of different status. It is the Japanese using Japanese whose style shifting shows the greatest frequency in certain situations. Americans, in the same situations, show much more use of softeners, whether they are speaking to someone of higher or lower status. The style-shifting patterns are important because they are sociolinguistic evidence of a significant aspect of Japanese and American cultures. Sakamoto and Naotsuka (1982) have claimed that Japanese and American cultures have different "polite fictions" upon which they operate. Americans go by the polite fiction that "you and I are equals," whereas Japanese go by the polite fiction that "you are my superior." In our research, we have found these stereotypical attitudes to be of some value, but also limited. Americans are not unconscious of status; rather, they often show how conscious of status they are by making moves to downplay or cover up status differences—to deny that status differences do exist (see Figures 7.4, 7.5, and 7.6). Americans are indeed aware of such differences, and as in the use of positive remarks in the two correction situations (see Figure 7.7), they sometimes show a very marked style-shifting pattern. Japanese, on the other hand, are conscious of status, and do not try to cover it up (see Figures 7.4, 7.5, and 7.6). Since style shifting according to interlocutor status has been shown to be greater among Japanese using English than among Americans using English in refusals and disagreements as well as in corrections, we naturally believe that there is a common explanation. Where data on Japanese using Japanese are available, that is, on refusals and corrections, the data are consistent with an explanation of transfer. It is not only the semantic content that is transferred but, more interesting, the patterns of style shifting.

Acknowledgments We would like to express our appreciation to the editors of this volume for their helpful comments. We also wish to thank the students in the TESOL and Applied Linguistics programs at Teachers College, Columbia University, for their many suggestions. We also thank those who took the time to answer the discourse completion questionnaire for the present research.

The Speech Act of Correction

155

Appendix The following are the 12 situations used in the questionnaire. After the description of each situation there was a three-line blank for the respondents to write what they thought they would say. "Higher to Lower" indicates that the particular speech act was performed by a higher-status person speaking to a lower-status person, and "Lower to Higher," vice versa. The situations were presented in randomized order. [Correction—Higher to Lower] (This categorization did not appear in the questionnaire.) You are a professor in a history course. During class discussion, one of your students gives an account of a famous historical event with the wrong date. [Correction—Lower to Higher] You are a student in a sociology class. During the lecture, the professor quotes a famous statement, attributing it to the wrong scholar. [Disagreement—Higher to Lower] You are a corporate executive. Your assistant submits a proposal for reassignment of secretarial duties in your division. Your assistant describes the benefits of this new plan, but you believe it will not work. [Disagreement—Lower to Higher] You work in a corporation. Your boss presents you with a plan for reorganization of the department that you are convinced will not work. Your boss says: "Isn't this a great plan?" [Chastisement—Higher to Lower] You are a corporation president and you have asked your assistant to prepare xerox copies of essential documents for an important press conference. Your assistant arrives at the last moment with 100 copies of the wrong materials. [Chastisement—Lower to Higher] You are a middle manager in a large corporation and your boss hands you a 50-page document, asking you to make 30 copies of each page. Ten minutes later he comes back to get the copies because it turns out he only wanted 30 copies of the front page. You have just made 10 copies of the whole packet. Obviously he is angry with you. [Announcing Embarrassing Information—Higher to Lower] You are a corporate executive talking to your assistant. Your assistant, who will be greeting some important guests arriving soon, has some spinach in his/her teeth. [Announcing Embarrassing Information—Lower to Higher] You are a student, speaking with your professor to prepare for a three-way meeting you have with the dean of the college. Your professor has some mustard on his cheek from lunch and you are aware of it. [Praise (Control)—Higher to Lower] You are a teacher trainer and one of your students submits an original paper that gives you fresh insights into teacher behavior. You run into the student in the hallway.

156

Speech Act Realization

[Praise (Control)—Lower to Higher] You are a beginning graduate student and you attend a lecture given by a famous scholar in your field. During the lecture you become aware of some ideas that you have never thought about before. You are excited about the lecture and approach the professor afterward. [Persuasion (Control)—Higher to Lower] You are a corporate executive. You have a plan for an advertising campaign for a new product. You are enthusiastic about the plan and need the cooperation of your assistants in implementing it. [Persuasion (Control)—Lower to Higher)

You are a middle manager in a large corporation. Your company exports its products to a foreign country where you have lived. You know of a major problem in the product design in terms of its practicality for that foreign market. You want the corporation president to order a revision.

References Barnlund, D., & Araki, S. (1985). Intercultural encounters: The management of compliments by Japanese and Americans. Journal of Cross-Cultural Psychology, 16, 9-26. Beebe, L. (1980). The sociolinguistic variation and style shifting in second language acquisition. Language Learning, 30, 433-47. Beebe, L., & Takahashi, T. (1988, January). Too many questions: Acquiring the social rules of speaking in a second language. Paper presented at the 10th Annual Applied Linguistics Winter Conference, sponsored by NYS TESOL. New York, NY. Beebe, L., & Takahashi, T. (1989a). Do you have a bag?: Social status and patterned variation in second language acquisition. In S. Gass, C. Madden, D. Preston, and L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (103-25). Clevedon and Philadelphia: Multicultural Matters. Beebe, L., & Takahashi, T. (1989b). Sociolinguistic variation in face-threatening speech acts: Chastisement and disagreement. In M. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in second language variation (199-218). New York: Plenum. Beebe, L., & Takahashi, T. (1988, January). Too many questions: Acquiring the social rules of speaking in a second language. Beebe, L., Takahashi, T, & Uliss-Weltz, R. (1990). Pragmatic transfer in ESL refusals. In R. C. Scarcella, E. Andersen, & S. D. Krashen (Eds.), Developing communicative competence in a second language (55-73). New York: Newbury House. Blum-Kulka, S. (1982). Learning to say what you mean in a second language: A study of the speech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-60. Blum-Kulka, S. (1983). Interpreting and performing speech acts in a second language—A cross-cultural study of Hebrew and English. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition (36-55). Rowley, MA: Newbury House. Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal of Pragmatics, 11, 131-46. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Bonikowska, M. (1988). The choice of opting out. Applied Linguistics, 9, 169-84. Brown, P., & Lcvinson, S. (1978). Universals in language usage: Politeness phenomena. In

The Speech Act of Correction

157

E. W. Goody (Ed.), Questions and politeness (54-310). Cambridge: Cambridge University Press. Brown, P., & Levinson, S. (1987). Politeness: Some universals in language use. Cambridge: Cambridge University Press. Clancy, P. (1986). The acquisition of communicative style in Japanese. In B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures (213-50). New York: Academic Press. Cohen, A., & Olshtain, E. (1981). Developing a measure of sociocultural competence: The case of apology. Language Learning, 31, 113-34. Condon, J. C. (1984). With respect to the Japanese. Tokyo: Yohan. Coulmas, F. (1981). "Poison to your soul": Thanks and apologies contrastively viewed. In F. Coulmas (Ed.), Conversational routines (69-91). The Hague: Mouton. De Tocqueville, A. (1965). Democracy in America. New York: Academic Press. Fraser, B. (1981). On apologizing. In F. Coulmas (Ed.), Conversational routines (259-71). The Hague: Mouton. Goldstein, B. Z., & Tamura, K. (1975). Japan and America: A comparative study in language and culture. Tokyo: Charles E. Tuttle. Kasper, G. (1990). Linguistic politeness: Current research issues. Journal of Pragmatics, 14, 193-218. Ide, S. (1989). Formal forms and discernment: Two neglected aspects of universals of linguistic politeness. Multilingua, 8, 223-48. Lakoff, R. (1973). The logic of politeness; or minding your p's and q's. Papers from the Ninth Regional Meeting of the Chicago Linguistic Society (292-305). Leech, G. (1983). Principles of pragmatics. London: Longman. LoCastro, V. (1986, November) / agree with you, but . . . Paper presented at JALT '86 Conference, Hamarnatsu, Japan. Loveday, L. (1982). The sociolinguistics of learning and using a non-native language. Oxford: Pergamon. Matsumoto, Y. (1989). Politeness and conversational universals—observations from Japanese. Multilingua, 8, 207-21. Nakane, C. (1970). Japanese society. Berkeley and Los Angeles: University of California Press. Nomura, N., & Barnlund, D. (1983). Patterns of interpersonal criticism in Japan and the United States. International Journal of Intercultural Relations, 7, 1-18. Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In S. Gass & L. Selinker (Eds.), Language transfer in language learning (232-49). Rowley, MA: Newbury House. Olshtain, E., & Cohen, A. (1983). Apology: A speech act set. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition (18-35). Rowley, MA: Newbury House. Owen, M. (1980). Apologies and remedial interchanges. The Hague: Mouton. Sakamoto, N., & Naotsuka, R. (1982). Polite fictions: Why Japanese and Americans seem rude to each other. Tokyo: Kinseido. Schmidt, R. W. (1977). Sociolinguistic variation and language transfer in phonology. Working Papers on Bilingualism, 12, 79-95. Takahashi, T., & Beebe, L. (1986). ESL teachers' evaluation of pragmatic vs. grammatical errors. CUNY Forum, 12, 172-203. Takahashi, T., & Beebe, L. (1987). The development of pragmatic competence by Japanese learners of English. JALT Journal, 8, 131-55. Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112.

This page intentionally left blank

Ill DISCOURSE PERSPECTIVES The chapters in this section expand the perspective of ILP into different directions. Their common denominator is their focus on interlanguage discourse, rather than on individual speech acts. As Aston comments in Chapter 11, speech-act-oriented ILP is characterized by an intra-organism view and a focus on interactional (as opposed to transactional) goals. The authors of the following chapters have adopted an interorganism view, with varying emphasis on transactional and interactional discourse functions. House (Chapter 8) examines inappropriate responses in learners' conversations with native speakers. In the second language literature so far, miscommunication in native-nonnative (and nonnative-nonnative) discourse has primarily been addressed in research on repair and interactional modification. In this research tradition, emphasis is given to the interactional treatment of communication problems by the interlocutors, rather than to locating their contributing sources. House offers a new perspective on the topic by attempting to account for the complexity of responding behavior in both native and nonnative talk. Her processes-oriented analysis yields a threefold classification of problem sources: language-based difficulties, conceptual and strategic deficiencies related to gaps in pragmatic knowledge, and operational difficulties, which may lead to inputs being virtually ignored. Whereas the first two problem sources are likely to typify learners more than native speakers, type 3 problems may well surface in native speaker talk as well. Future research in this area will have to take into account, as House does, the processing constraints operating in learners' decoding and encoding of messages in conversational interaction. Zuengler (Chapter 9) provides a comprehensive review of interlanguage studies concerned with the effect of conversational topic on learners' variable performance. The main findings emerging from work in this area are that both expertise in the topic and affective involvement influence interlocutors' role dynamics and conversational behavior. Most studies indicate that superior topic expertise, whether real or subjectively perceived, results in more control of the discourse by the expert participant, regardless of native or nonnative speaker status. However, expertise may be neutralized by affective factors such as emotional involvement and test anxiety. One of the important implications of interlanguage studies on topic is the emphasis on factors internal to the speech event (rather than adopting the traditional focus on event-external factors such as learners' age, L I , etc.). This work could thus be the 159

160

Discourse Perspectives

first step toward full ethnographies of communication in the study of interlanguage variation, with the speech event as the unit of analysis. This development would imply the inclusion of other event-based parameters of variability, such as the nature of the setting, relationship between the participants, and message content. Presumably the quality of learners' speech and conversational control, at least in the case of advanced nonnative speakers, is subject to complex interactions between eventinternal and -external factors which are still not well understood. Blum-Kulka and Sheffer (Chapter 10) expand the ILP research agenda by considering the potential for ILP phenomena in a first language. The population examined are American-born longtime immigrants to Israel, who maintain English as the language of the home. The metapragmatic discourse of these families in their first language is shown to manifest an intercultural pattern that differs in its style from both Israeli and American discourse. Traditionally, shifts in LI use among immigrants would be considered a case of attrition. This study suggests that we can gain new insights into ILP by placing such phenomena in the classical interlanguage study paradigm, comparing the first language version spoken by the immigrants to native speech in both relevant contact cultures. Consequently, in the approach offered here, ILP phenomena are by no means restricted to second language varieties. We have chosen Aston's contribution to be the final chapter to this volume because it presents a more optimistic picture of interlanguage discourse than is usually painted in ILP. Rather than focusing on pragmatic failure and other sorts of miscommunication, Aston's concern is with the negotiation of comity, the establishment and maintenance of friendly relationships in native and native-nonnative talk exchanges. In order to achieve such interactional goals, participants minimally need to find common ground for sharing attitudes toward features of the world ("solidarity") or for showing affiliation with the other ("support"). Applying conversation analytic methods, Aston carefully demonstrates how the conversational resources drawn on to establish rapport in native-native discourse are transformed into a different set of resources in native-nonnative interaction. The important implication of Aston's work, both for theory construction in ILP and second language pedagogy, is to extend to ILP the fundamental tenet of interlanguage studies, namely, the autonomy of learner language. Instead of viewing interlanguage pragmatic knowledge and behavior as deficient in terms of native norms, we need to consider its functionality and inner justification.

8 Toward a Model for the Analysis of Inappropriate Responses in Native/Nonnative Interactions JULIANE HOUSE

Why do learners of a second or foreign language often fail to respond in a way acceptable to their native interlocutors? Is it because they "misheard" or simply didn't hear what was said to them, because they did not understand what was said to them, or because they did not understand what the interlocutor had meant by what he or she said to them? In trying to find possible answers to the question of what may go wrong in a learner response, one notices immediately that response failures stem from many different and possibly interacting sources: they can be a result of inadequate perception, of inappropriate comprehension at the syntactic, semantic, pragmatic, and discourse levels of language or of gaps in the learner's knowledge of the world; they can result from uncooperativeness on the part of the responder, who in this case may have understood perfectly well what the interlocutor had tried to communicate but just wanted to be difficult; or the responder may have perceived, and comprehended correctly, and intended to cooperate, but cannot manage a response which is, at this stage of the discourse, expected from his interlocutor. Given the difficulty of coming to grips with interlanguage responses, it is understandable that attempts at investigating them are relatively rare (notable exceptions are Keller-Cohen, 1979; Kasper, 1981, 1984; Thomas, 1983). What is needed in dealing with deviant learner responses and misunderstandings in interactions between learners and native speakers is a model comprehensive and powerful enough to take account of the complexity and ambiguity of the causes of response failure. In this chapter I shall (1) set the framework for the eventual development of such a model by reviewing various theoretical and empirical approaches and their relevance for explaining the difficulties learners face when responding; (2) attempt to integrate some of the assumptions and concepts found useful in these approaches into a discourse-processing model; and (3) illustrate the explanatory power of this 161

162

Discourse Perspectives

model by demonstrating how it can account for a set of exemplary misreponses taken from role-play-elicited discourse data.

Approaches to Analyzing Responding Behavior The following research strands are relevant in addressing the issue of interlanguage responses: (1) social views of language, (2) intercultural miscommunication, (3) studies of misunderstanding based on pragmatic theories, (4) the psychopathology of everyday discourse, (5) cognitive-discourse-processing models.

Social Views of Language Language can be viewed from two perspectives: (a) an intra-organism one, where the focus is on what goes on in language users' minds; and (b) an inter-organism (social) one, where what goes on between two or more language users matters most (Halliday, 1978, 56). Looking upon language as an essentially social phenomenon has had a long tradition in Europe and North America that may be traced back to linguists of the Prague school and their work on language functions, functional styles, and functional sentence perspectives; and to British contextualists such as Firth and later Halliday, who, influenced by the anthropologist Malinowski, emphasized the importance of "the context of situation" in establishing meanings in talk. Firth (1964) suggests, for example, that "the give and take of talk resembles a preordained ritual . . . the moment a conversation is started whatever is said is a determining condition for what in any reasonable condition may follow" (94). In North America, a similar tradition of stressing the embeddedness of language in context or culture exists in the work of anthropologists and linguists such as Boas, Sapir, and Whorf, and later sociologists of language and sociolinguists such as Cicourel, Labov, and Goffman as well as that group of sociologists around Harold Garfinkel who came to be known as the "ethnomethodologists." On the ethnomethodological view, it is less useful to judge a response according to whether it matches the previous speaker's unobservable intentional state; rather, the main concern is to look at how observed and contextualized stretches of discourse form coherent wholes. The notion of sequentially organized implications as manifest in tied pairs or adjacency pairs is relevant here (Schegloff & Sacks, 1973); that is, when the first part of such a couplet is produced, it sequentially implicates the production of the second part of the pair. The close-order organization of talk as evident in the tying of utterances in adjacency pairs thus enables discourse participants to monitor their conversation jointly such that the achievement of "understanding" (or "misunderstanding") can be seen as displayed through the positioning of elements of talk and actions (Benson & Hughes, 1983, 179); any response is to be seen as "a kind of claim as to what was meant by the previous utterance" (Bilmes, 1990). As Goffman (1981, 12) pointed out, one major problem with this view is that the burden of defining a response is simply thrown back onto the previous speaker's initiating statement, and from there reference is made back to another initiating

Native I Normative Interactions

163

statement preceding it, and so on; it is thus difficult to know where to start or how to find proof of the appropriateness of a given response in itself. Another difficulty with this view is that it lends itself best to a fairly formalistic, micro-level analysis, which does not take the larger co-text into account. In his own attempt to define a response, Goffman (1981, 35) states that a response, which emanates from an individual and is inspired by a prior speaker, tells us something about that individual's position or alignment in what is occurring. In the context of an ongoing discourse, "alignment" implies, first, that two interactional moves are brought into something resembling agreement or partnership with one another; second, that the response is of local significance, that is, relevant and comprehensible at the moment of speaking. Goffman (1981) singles out one type of response which he calls a "reply", that is, "a response in which the alignment implied and the object to which reference is made are both conveyed through words or their substitutes" (35). Replies, then, reach more directly and verbally backwards to the immediately preceding move, ostensibly achieving cohesion through overt surface connections on the utteranceexpression level (examples are the ritualized reciprocal pairs in phatic phases of talk), while responses are often less "tidy" in their back-referencing; they do not necessarily bear on the immediately preceding surface expression but rather are coherent on account of a variety of different interpretation schemata available to speakers.1 In simplistic terms, if we accept Goffman's distinction, but wish to consider both "replies" and "responses" in his technical sense, then we are obliged to investigate under the notion of "responding" (in the nontechnical sense) both the linguistic links between segments of talk and the interpretation schemata by means of which talk is achieved.

Intercultural Miscommunication There is a rich literature on misunderstandings occurring in interactions between members of different cultures. Gumperz (1982a, b) found that misunderstandings often result from participants misusing "contextualization cues," that is, aspects of the surface forms of utterances (prosodic, phonological, lexical choices, routine formulas that may signal relevant interpretive frames). In Gumperz's view, it is such (often culture-specific) linguistic cues that are crucial for processes of conversational inferencing, that is, "the 'situated' or context-bound process of interpretation, by means of which participants in a conversation assess others' intentions and on which they base their responses" (1982a, 153). Tannen (1984a,b, 1986), who has studied both native-native and nativenonnative talk, explains deviant responses and misunderstandings as a "matter of framing," of exploring frames, breaking frames, or reframing in Batesort's (1972) sense. Frames are dynamically linked as an "interpretation and responding chain where negotiated footings are continually changed." Tannen (1984b) lists eight levels of differences in the ways speakers signal what they mean: when to talk, what to say, pacing and pairing, intonation, formularity, indirectness, cohesion, and coherence—and these eight may lead to differential ways in which conversational partners tend to assess others' intentions as a basis for making their responses.

164

Discourse Perspectives

Along these eight levels, cross-cultural as well as intra-cultural misunderstandings can arise. Research has been done on some of these eight areas. Erickson (1975) and Erickson and Shultz (1982) analyzed misunderstandings in multiracial consulting interviews, where they described how subtle differences in interactional style across racial and ethnic groups can lead to misinterpreted responses. Kochman (1981) examined black and white interaction styles. Scollon and Scollon (1981), in their work on Athabaskan Indians, revealed several culturally conditioned differences, especially in face-presentation and the distribution of talk and silence, where differences in the lengths of silences tended to be interpreted as challenges or disagreements by members of the other culture. Work on communicative style (Lakoff, 1976), politeness phenomena (Brown & Levinson, 1987), Gricean Implicature (Grice, 1975; Smith, 1982; Sperber & Wilson, 1986) as well as the interpretation of indirect speech acts (e.g., Clark, 1979; Clark & Schunk, 1980; Clark & Lucy, 1975) have resulted in hypotheses suggesting that "normal talk" is "indirect"; that is, if they want to respond appropriately, speakers need to infer their partners' intentions in order to find out how the words used were "really meant." Indirectness lies at the heart of many if not most misresponses and misunderstandings in talk, and such alignment failures are of course much more likely to occur in talk between people from different cultural backgrounds, where indirectness and politeness conventions often diverge (Blum-Kulka, 1987; House, 1986).

Pragmatic-Theory-Based Studies of Misunderstandings The third research strand is pragmatic-theory-based and is represented by the work of Dascal (1977, 1985); Dascal and Katriel (1979), Dascal and Idan (1988), BlumKulka and Weizman (1988) and Weizman and Blum-Kulka (1988). In this line of inquiry, misunderstandings in "normal talk" between native speakers are analyzed at different levels of pragmatic meaning. Dascal (1985) introduced the notion of a "conversational demand" set up by an utterance, the perception of which may differ for interlocutors. What exactly can go wrong in the perception of a conversational demand can be described by reference to four questions that need to be answered in interpreting what speaker S has said (Fillmore, 1976, 78): 1. What did S say? (proposition) 2. What was S talking about? (what was said plus what was implicated, that is, the extended semantic meaning) 3. Why did S bother to say it? (illocution) 4. Why did 5" say it in the way he said it? (key, tone). On any one of these layers of significance and any combinations thereof, misunderstandings can arise.2 Weizman and Blum-Kulka (1988) made a similar distinction of the potential causes of misunderstandings in terms of the levels of prepositional content, illocutionary point and mode, and, following Searle (1983, 1992), they also differentiate between an "individual-I" (a speaker's) meaning and a "collectivewe" direction pertaining to the three levels along which communication can break down.

Native I Normative Interactions

165

Psychopathological Studies of Misunderstandings The fourth approach relevant to construing a theory for the analysis of misunderstandings is the psychopathological approach, which can be traced back to Freud's intuitions about the "psychopathology of everyday life," and in particular, slips of the tongue. This approach is linked to the work of Langer and her associates (e.g., Langer, 1978; Langer, Blank, & Chanowitz, 1978; Langer & Imber, 1980), Reason and Mycielska (1982) as well as Heikkinen and Valo (1985), who related the findings in this paradigm to interlanguage responses, and most recently Heckhausen and Beckmann's (1990) work on action slips. The central points made in connection with this approach are that much of what is traditionally conceived of as "ostensibly thoughtful action" is in fact characterized by "mindlessness" (Langer, Blank, & Chanowitz, 1978), and that it is especially in overlearned behaviour, that is, routinized and automatized actions, that "slips" occur (cf. also Goffman's [1959] distinction between "knows-better" slips and "does-not-know-better" errors). Langer's "Non-Thinking Hypothesis" has been given indirect support through the work by Langer and Abelson (1972) and Abelson (1976) and the many ensuing studies on the packaging of knowledge in the mind, suggesting that in many types of behaviour actants rely on scripts to reduce cognitive work, effort, and time. Because of the imprintedness of larger plans, schemata, scripts, or social episodes in human memory, we are able to predict upcoming moves and consequently stop paying attention to the reality of the interlocutor's input. The "illusion of control" is often only disrupted by the occurrence of "slips," the recognition of which is then equivalent to a sudden intrusion of reality. In the case of language learners, slips tend to occur for a different reason: learners have not yet fully developed and automatized scripts at their disposal. Because learners often "rehearse" internally what will be their next move, they also often neglect any real input (Heikkinen & Valo, 1985).

Information-Processing Approaches to Language Understanding The information-processing approach is closely linked to the psychopathological one in that here, too, an assumption of scripted behavior is made, in which cognitive schemata, frames, or scripts are representations of repeated behavioral patterns designed to reduce cognitive work. The basic idea of organizing knowledge in packets that function as ready-made subsets goes back to Bartlett's (1932) conception of human memory as consisting of high-level structures, or schemata, encapsulating knowledge about particular objects, events, and situations. Such knowledge structures have also been referred to as frames, that is, networks with slots which can be filled with certain values (Minsky, 1975), and scripts, which represent sequences of routine actions, such as the restaurant script (Schank & Abelson, 1977). Cognitive models that explain the processes operative in discourse comprehension (Kintsch & van Dijk, 1978; van Dijk & Kintsch, 1983; Kintsch, 1988) are especially relevant to analyzing interlanguage responses. In the original Kintsch and van Dijk model (1978), the construction of a prepositional text base was considered to be the major component in comprehension. In their later (van Dijk & Kintsch,

166

Discourse Perspectives

1983) version, a "situational component" was added. Strategical behavior was considered to be the driving force in the construction of a propositional text base, discourse comprehension being regarded as basically a strategic process. In a later attempt to combine production systems with a connectionist approach, Kintsch (1988) developed a "construction-integration model" of discourse understanding, claiming that initially comprehension is bottom-up-guided and it is only in a later integration process that appropriate meaning is selected. A cluster of propositions is derived from the discourse via activation of the closest neighbors of the original text-derived propositions in the general knowledge net. Kintsch assumes a minimally organized and structured knowledge system, with specific structures being generated in the context of the task for which they are needed and resulting in dense associative knowledge nets. The strongest arguments for a cognitive-processing-model approach to the explication of inappropriate responses is that such an approach is (potentially) explicit, and may integrate the different insightful perspectives sketched above. Social perspectives may be encapsulated as such rules, norms, and conventions must have psychocognitive correlates inside the individual if they affect conversational behavior. Further, the very notion of inappropriateness presupposes at least two cognitive systems in interaction: in principle, an information-processing model should enable the identification of contextualized misunderstandings by explicating processing "states" of the interacting systems.

Toward a Model for Analyzing Responses In building on discourse-processing approaches as sketched above, Edmondson (1989) set up a discourse comprehension and production model, which I have adapted and outlined in Figure 8.1. I shall briefly describe this model and then illustrate its explanatory power by applying it to the analysis of discourse data. As can be seen, discourse comprehension can hardly be considered independently of discourse production, especially in the case of dyadic spoken face-to-face discourse. The model operates on two "levels": a prelinguistic or conceptual level (made up of the upper boxes A, B, C, F, G, and H) and a lower linguistic level (consisting of the lower boxes D, E, and I) on which the linguistic decoding of the input and the encoding of the output occurs. These levels are networked in rather more complex ways than Figure 8.1 suggests. In briefly elaborating upon the subcomponents inside this model, I shall restrict myself to characterizations relevant to my purposes.

Box A: The Operant Discourse Frame The assumption here is that a copy of some relevant stored situational constellation or schema is made and "filled in" by knowledge derived from the current discourse situation. The result is a complex of conceptual and linguistic representations: the operant discourse frame. The term frame (cf. Bateson, 1972; Goffman, 1974; Tannen, 1979 and above) is being used here as distinct from schema, which pro-

Native I Normative Interactions

167

Fig. 8.1. Schematic discourse processing model. vides the skeleton that is fleshed out in the ongoing discourse. In my terminology, then, a schema is built up into a discourse frame. In ongoing discourse, however, further or alternative schemata can be called up, leading to revision, switching, reactivation or indeed coexistence of operant discourse frames (see the notion of coexistent discourse worlds in Edmondson, 1981). The following elements are contained in a currently activated discourse frame: 1. Knowledge of the currently relevant discourse topic 2. Knowledge of the currently relevant interactional move 3. Representation of the discourse outcomes resulting from the preceding interaction 4. A primarily linguistic representation of the prepositional content of the last turn held in short-term memory, which can be accessed from the currently active frame. In terms of the distinctions made by Dascal (1977, 1985) and Blum-Kulka and Weizman (1988) it is the level of aboutness (topic, proposition) which is captured in "operant discourse frames". Box B: Interactional Goals It is assumed that the speaker has or develops a goal for the ongoing encounter, together with relevant subgoals, which develop as the discourse unfolds. There is

168

Discourse Perspectives

clearly interaction between boxes A and B, in that what has happened so far may influence what the speaker seeks to get out of the encounter. Similarly the goals of the speaker system will affect what form the "nonlinguistic reaction" represented in box F may take, just as long-term goals will impinge on strategic manipulation (box G). For instance, in the case of the elicited role-play data to be analyzed below, the role specifications in part predetermine what are here called interactional goals, just as these role specifications (in part) preselect a particular schema, that is, a set of expectations concerning what type of discourse will take place. In terms of the distinctions made by Weizman and Blum-Kulka (1988), we may here include two types of goals in an interaction: the individual-I point and the collective-we direction/purpose of the encounter (although it is often difficult to disentangle the two).

Box C: Current Prediction/Expectation This simply marks a result of ongoing processing: the system expects a more or less specific type of input.

Box D: Decode The process of extracting a "discourse meaning" may involve linguistic decoding (itself a complex of subsystems), discourse interpretation strategies, and various inferencing procedures as suggested by Gumperz, Tannen, and others. The strategic knowledge that may be activated in box G may also be called up in this decoding process.

Box E: Discourse Meaning As an outcome of the decoding processes in D, a discourse meaning for the current input is arrived at.

Box F: Nonlinguistic Reaction An interpreted input leads initially to a nonlinguistic response, a sort of "gut reaction": the determinants of this response are in large measure personality-based, and not, strictly speaking, part of a discourse system. In Edmondson (1987), some "reactive formatives" are posited, and a distinction is made between an interactional option (which basically concerns whether the hearer/speaker will take a positive or a negative attitude towards the current discourse move made by the interlocutor), and an illocutionary option concerning the category of speaker-meaning appropriate to this purpose.

Box G: Choice of Discourse Strategies In this phase of the system, the hearer/speaker calls upon strategic or procedural knowledge, which may lead to the employment of interactional strategies designed to disguise the reaction in F above. The hearer/speaker may, for instance, decide to

Native I Normative Interactions

169

delay but at the same time anticipate the raw reaction stored in F, or may cloak it via various politeness strategies. In this phase, which may be likened to a filtering instance that serves to modulate speech acts under the constraints of politeness maxims, potential responses of the interlocutor are taken into account, in the anticipation of which speakers adjust their initial reactions.

Box H: Immediate Communicative Plan At this junction in the system a prelinguistic communicative plan has resulted, the scope of which may range from small turn-internal units to longer discourse stretches, in which case elements of such a plan will be transferred to box B for later recall.

Box I: Linguistic Encoding It is at this level that the communicative plan arrived at in H is transformed into linguistic representation. Box ] then contains at the very least a complex of linguistic resources, search strategies, prepacked chunks such as routine formulas, and gambits, acting as discourse lubricants. If linguistic searching is unsuccessful (i.e., the communicative plan established in H cannot be realized linguistically), backlooping may occur, leading, for example, to "reduction strategies" (Kasper 1982; Faerch & Kasper, 1980, 1983) with which speakers attempt to adjust a communicative plan in light of the insurmountable complexity of the linguistic realization necessary to carry it out. This model does not aspire to completeness, focusing instead on what appear to be critical points of processing from the learner's perspective. While the pathway mapped in Figure 8.1, and the numbering adopted for expository purposes, suggest a straightforward linear-processing procedure, the "boxes" are interrelated and the whole system operates in parallel, that is, various nodes, items, and paths are activated at the same time. Further, various shortcuts exist in a fully developed system (see Edmondson, 1987 on the nature and function of routines). The following analyses suggest several refinements to the model, in terms of further paths through the subsystems.

Testing the Model First of all, we have to clarify what exactly is to be understood by an inappropriate response. Following Goffman's (1981, 35) characterization of a response as a move inspired by a prior speaker that is (a) aligned to what is "occurring now" and (b) to be understood as "relevant now," we can say that an inappropriate response is a response interpreted not to be in alignment with the preceding discourse and is thus taken not to be relevant at the moment of speaking. However insightful, such a definition is not easily operationalized. It will, however, suffice here. The model's operation will be demonstrated in an analysis of conversational responses. The data were elicited via role-plays from advanced German learners of

170

Discourse Perspectives

English interacting with English native speakers in a variety of simulated everyday situations. The dialogues included both immediate contact and telephone conversations, each requiring centrally a specific speech act such as a request, a complaint, or a suggestion. Interactants were given brief, oral, and "open" situational instructions, with roles being varied along the dimensions of dominance and social distance.3 In cases of inappropriate responding, a retrospective interview was conducted. In these interviews, subjects listened to a playback of their recorded conversation and, at the same time, studied the transcription. With respect to the critical (mis)response, they were then asked to reflect upon how they had "understood" their interlocutor's ("response-inspiring") utterance and what they (thought they) had intended with their own ensuing utterance. These retrospective interviews were not consistently taped as it often transpired that subjects felt much freer to do serious indepth probing of their own behavior when they were not recorded. Respondents frequently pointed out that in the retrospective sessions they were (actually often painfully) confronted with their own "failure," an experience absent in the original role-play session (the recording of which was never felt as intrusive). In analyzing the data, I have used the discourse analytical categories provided in Edmondson's (1981) discourse model and in Edmondson and House's (1981) Interactional Grammar of English. Instances of inappropriate learner responses will be related to the Discourse Processing Model outlined above, and to concepts, methods, and findings adopted from the five research strands described above. At the same time, detailed analysis, even if at times rather speculative, suggests quite strongly that things are much more complex than indicated by the model. The analysis of the interlanguage data examined revealed that there are three sources for the pragmatic-responding errors found: 1. language-based difficulties, focusing on problems that are apparently located in linguistic decoding and encoding 2. conceptual and strategic deficiencies, which may result in learners choosing inappropriate or inadequately mitigated speech acts 3. operational difficulties, which may lead to an input apparently being ignored. It is conceded immediately that these three distinctions are decidedly fuzzy-edged. Indeed, it will transpire that taking a discourse-processing perspective on learner responding behaviors makes neat categorization of responses a difficult if not rather artificial undertaking.

Language-Based Responding Deficiencies ENCODING DIFFICULTIES The most predictable and most common learner response difficulties relate to the apparent inability to find and employ routine formulas and gambits appropriately (box I). The following are examples of this type of responding deficit:

Native I Normative Interactions

171

Situation 1: Invitation to a Party/Opening Phase NS: NNS: NS: NNS: NS: NNS:

Hi, how are you? Oh, just fine, what about yourself? Pretty good, what have you been up to lately? Well, I worked at home for University, I was really busy. And what have you been doing nights? As I told you, I didn't go out much, usually I stay at home and work.

In commenting on the last move in this discourse stretch, the learner reported that she had intended to express the translation equivalent of the German phrase wie gesagt, which in German is a rather meaningless formula used to provide coherence with preceding discourse stretches. In the context above, the English phrase as I told you carries, however, a rather petulant and needlessly aggressive overtone, which the learner certainly did not intend to imply. An expression such as "As I say, I've been working so . . . " would have been more appropriate in this context. Learners admit that they infrequently model their utterances on supposed German translation equivalents. In situation 2 below, the learner had intended to produce a final Thanks, but through the inappropriate inclusion of the downtoner anyway, she produced a distinct "misfire": anyway collocates exclusively with Thanks in a context in which a Request has not been satisfactorily met, its semantic meaning being "in spite of everything." Situation 2: Request for Money/Closing Phase NS: Just leave me your name and address; I'll write it down and then give you the money. NNS: Okay, that's kind of you. Thank you anyway.

In the retrospective interview, the learner indicated that she had considered the token anyway to be a convenient filler corresponding to the German all-purpose gambit also, which would be perfectly acceptable in an equivalent German interchange. In situation 3 below, the learner fails to reply appropriately to a typically vague and essentially ritual suggestion made by a native speaker in the (phatic) opening phase of an interaction, thus violating the politeness principle of "leaving options" (Lakoff 1973). Situation 3: Phatic Talk/Neighbors NNS: Hello. I'm your new neighbor. I'd like to visit you because I've just moved in and don't know you. NS: Oh, how nice! Well, we'll have to get to know each other. NNS: Yes, 1 think so as well. Perhaps you should drink a cup of tea.

In this context an appropriate response would have been a vague reciprocating phrase such as "maybe you'd like to come over for a cup of tea sometime." The learner explained that she had meant to say something equivalent to the German "ja naturlich, wir sollten mal auf ein TaBchen Kaffee zusammen kommen," which does have the casual, nonimposing overtone the supposed English equivalent totally

172

Discourse Perspectives

lacks because of the modal should. The learner initiation is, of course, also lacking in appropriateness. In situation 4 below the learner again fails to respond appropriately inside a phatic phase of an interaction because she lacks the appropriate tokens. In this case, unsuccessful transfer from the mother tongue was not the culprit; in fact, the token "Okay," which would have been correct in the following exchange, is very common in German as well and fulfils the same functions: Situation 4: Request to Hand Over Lost Property/Closing Phase NS: NNS: NS: NNS:

What number is your house number? Eleven Eleven. Number 11, okay, about two o'clock then. Yeah, that's right.

The phrase "that's right" used by the learner in her final move implies a personal evaluation of the truth of the interlocutor's preceding utterance. What the learner had meant to realize in this final discourse slot was a ritual move sealing the agreement reached before, that is, the realization of the ritual speech act Okay (Edmondson & House, 1981, 60), for which conventional tokens such as okay or fine would suit. The learner had realized that such an act was called for at that final stage in the interaction but did not have the appropriate tokens at her disposal. The next example reveals learners' inability to produce discourse lubricating devices (gambits) such as Uptakers or Starters in a conventionally accepted way. Indeed, inappropriate responses that result from the learner's inability to employ an acceptable gambit token (e.g., the token well) are very frequent in the data. They originate, as I have written elsewhere (House 1982b; 1984), from the transfer into English of the German gambit token ja, which fulfils many different discourse lubricating functions. Here is an example: Situation 5: Request for a Lift NS: Why don't you use your own car? NNS: Yes, you know, it's too dirty.

DECODING DIFFICULTIES I turn now to inappropriate learner responses which seem to relate to difficulties of decoding (box D). The inappropriate learner responses resulting from decoding problems are less easily ascertained than were the ones related to the unavailability and unencodability of routines and gambits; in some cases both decoding and encoding problems seem to have been operative. In situation 6 below the learner takes the indirect request "do you mind if 1 do P" apart and responds separately (and idiosyncratically) to the two chunks "do you mind" ("no") and "if I do P" ("do it"), suggesting that she did not interpret the indirect request as an intended entity (see, however, the studies conducted by Clark and his associates, e.g., Clark and Schunk 1980, which revealed that it is indeed sometimes the literal meaning of an indirect request that native speakers respond to.)

Nativei'Normative Interactions

173

Situation 6: Small Talk/Waiting Room NS: Excuse me, do you mind if I open the window? It's so stuffy in here. NNS: No, do it.

Besides admitting the decoding problems mentioned above, the learner reported that she had actually wanted to realize an uptaking routine formula equivalent to the German bitte schon, but was unable to do so and thus resorted to the linguistically simple imperative form. It had not struck her that the result might violate the politeness principle. In situation 7 below, the learner was unable to interpret the discourse stretch "what was your name, again" correctly; that is, as a polite request for information. The use of the past tense in particular led to interpretive uncertainty, so the learner fell back on a simplistic echoic strategy, providing an amusing type of "reply" (Goffman, 1981). When the native speaker made a joke of the learner's inappropriate response, the learner might have sought to do likewise, but lacked or at least was unable to access the set token "I think so," producing the contextually puzzling response "I think," a not-uncommon mistake for Germans learning English: Situation 7: Gas Company/ Complaint NS: NNS: NS: NNS: NS:

Well, let me look at my list. What was your name, again? It was Brown. It was Brown. Is it still Brown? I think. Well, I do have a Brown on my list.

In situation 8, the learner was unable to use an appropriate discourse interpretation strategy for the formula "May I help you" (routinely employed in service encounters), and went on to produce a reply in Goffman's (1981) sense, which in part reproduced the native speaker's preceding utterance—a defensive habit acquired in the foreign language classroom. In following this echoic strategy, these learners produce impolite responses. It is, however, surely part of (common) strategic knowledge that the deference manifested in the first pair part of a request is totally inappropriate for use in a positive response to that request: the error, therefore, appears to be mainly one of decoding. However, a system failure in terms of strategic knowledge (see box G), or more precisely, its use in discourse interpretation, may also have been present. Situation 8: Gas Company/Complaint/Opening Phase NS: May I help you? NNS: Yes, of course, you may. I want to complain.

In situation 9, the learner revealed in the retrospective interview that she was not in a position to correctly interpret the routine formula "What have you been doing lately?": the response is plausibly an initiation of her own interactional goals; that is, box B determines what results in box H. Note that a nonlinguistic reformulation of the interlocutor's query, roughly of the form "What are you doing when it is late" is appropriately responded to—in other words, the input is not interpreted, but it is not ignored:

174

Discourse Perspectives

Situation 9: Invitation to a Party/Opening Phase NS: Hi, how are you? What have you been doing lately? NNS: Well, I want to go to a party tonight.

An interesting example of learners' failure to use appropriate inferencing strategies (box D) necessary for arriving at a plausible discourse meaning of the interlocutor's preceding move (box E) is provided in situation 10: Situation 10: Small Talk/Elevator Got Stuck NS: I'm not trusting this elevator any more. What about you? Are you? Are you going to walk up or are you going to ride the rest of the way? NNS: I, I, the next time I will take the stairs.

My hypothesis here (and it was in fact substantiated in the interview) was that the learner was unable to derive from the NS utterance the implication that NS had stopped, or was about to stop the lift. The problem is one of inferencing. Such inferencing is in this case role-play-specific, in that whereas in real-life situations speakers make inferences from the enveloping situation to the words uttered by the interlocutor, it is the other way round in a role-play situation, in which the words used can redefine the situational context (something similar can, of course, happen in telephone talk). In situation 10 the learner was unable to infer what situational redefinition the interlocutor had meant to create through his initiating move; he couldn't make sense of the utterance, hesitated, and then made a good attempt at replying, revealing that the propositional content of the input was adequately interpreted. While this instance of an inferencing problem may be role-play-specific, interpretive problems based on wrong inferences are not.

Conceptual and Strategic Deficiencies In many cases, it will be a matter of speculation as to whether perceived impoliteness is a result of the responding learner's lacking the conventionalized tokens for performing a strategically modulated plan (box I), or whether the deficiency lies in the pragmatic knowledge base itself (box G). The following examples of inappropriate learner responses evidence a lack of relevant culture-specific pragmatic knowledge, resulting in violations of the politeness principle or one of its maxims (e.g., the maxim of tact) to be observed in realizing polite utterances (e.g., Lakoff, 1980; Leach, 1983; Holmes, 1985). Situation 11: Landlady NNS: I intend to go out with a friend tonight, so I won't be home tomorrow morning for breakfast, Mrs Bennett. NS: Well, you know, Anita, I bought extra biscuits for your breakfast. NNS: Well, that doesn't matter, 'cause they will surely be fresh on Tuesday anyway. In commenting on the pretty rude response "that doesn't matter," the learner in question said that she had tried to produce a translation equivalent of the German minimizers das ist dock nicht so schlimm or das macht dock nichts, both of which carry a much more pacifying and consolatory overtone than the abrupt pseudo-

Native I Normative Interactions

175

equivalent chosen. In fact, the phrase "that doesn't matter" is overrepresented in my data, and all its users proclaimed ignorance of the inherent impolite connotation. The point here is that English has no "translation equivalent": an entirely different speech act is in fact conventionally called for in English; that is, some form of apology seems to be needed rather than a minimizer as is conventionally employed in German. In other words, a cross-cultural comparison would need to go "further back" than the issue of linguistic form. Situation 12 may help establish the point: Situation 12: Complaint/Messy Kitchen NS: Look here, Anne, are you responsible for that mess in the kitchen? NNS: Oh, yes, I think, but that doesn't matter.

In situation 13 the learner apparently lacks the pragmatic knowledge that a polite and inoffensive countersuggestion is called for in this discourse slot, rather than a grossly impolite rejection, abruptly bringing the talk to a close: Situation 13: Party/Small Talk NS: NNS: NS: NNS: NS:

I have never been to Greece before but I always thought it must be beautiful. Oh, yes, but there arc too many German tourists there in the summer. Mmh. Yes, well, what do you think? Should we go and have some cheese now? Oh, no, no, no, I stay here. Oh, I see (rather nonplussed)

The learner in situation 13 stated that she had not intended to cut off the conversation. On the contrary, she had just wanted to express her desire to stay exactly where she was and continue chatting. The response produced is offensively ego-oriented (for discussion of a pervasive preference for self-reference and a preoccupation with self rather than the interlocutor in German native speaker talk, see House, 1979, 1982a, b, c, 1984, 1989). Even "Oh, no, no, no, we stay here" would produce a totally different effect. Hence, I suggest, a deficiency in (culture-specific) pragmatic knowledge, or a case of "schematic transfer," or at least, a bypassing of such considerations in discourse production is the problem: the problem is not linguistic. A similar lack of awareness of the effect of her response is attested to in situation 14. Here both gaps in pragmatic knowledge (box G), namely, that a response in the form of a (repeated) polite request is called for, and the nonavailability of appropriate linguistic tokens for making a polite request (box I) were suggested by the learner as contributing to what she later felt was an insulting response to the nurse's preceding utterance: Situation 14: Making a Last-Minute Doctor's Appointment NNS: Couldn't you make an exception for once? NS: Well, I don't see any way of squeezing you in at the moment. NNS: Well, you can. Anybody can.

Operational

Difficulties

While the types of inappropriate responses previously characterized could be related to some malfunctioning inside one or two phases ("boxes") of the proposed

176

Discourse Perspectives

discourse-processing model, the following instances of deviant learner responses seem rather to be system malfunctions and can, informally, be accounted for as shortcuts, in which the learner seems to ignore the input (see the above discussion of the "psychopathological" research paradigms). Something similar, in this case based on interpretative problems, was suggested in the discussion of situation 10 above, in which, it was hypothesised, interactional goals "took over" the communicative-planning task. Similar "system leaps" will be suggested in the following. General cognitive overload, insecurity, chunking problems and so on may be causes. The role-play situation itself may be an additional causative factor. The phenomenon to be explicated is familiar to nonexpert speakers of a foreign language, as when, for example, one requires so much time to rehearse one's own contribution to a group discussion that the topic has moved on before one is ready to bid for a turn. In two-party interactions, such operational difficulties may, apparently, lead to the performance of a rehearsed move, independent of its current discourse relevance. The following are examples of such nonpartner-oriented responses: Situation 15: Invitation to a Party/Opening Phase NS: Hi, Susanne, how are you doing? NNS: I'd like to invite you to a party I'm doing this night.

Here one might informally say "Susanne didn't listen": all she did was jump from her goal of inviting the interlocutor to an immediate communicative plan and its linguistic encoding, bypassing not only the partner's input but also any strategic considerations (box B feeds directly into H). However, it is again of interest (see the remarks offered for situation 12) that she picks up her interlocutor's token "doing." The suggestion here is that this advanced learner is perfectly familiar with the collocation "give a party": the grammatical "error" is a form of lip-service to the input received, linguistically decoded, but otherwise ignored. The following three situations are further examples of this type of self-centered shortcut: Situation 16: Lost Property/Opening Phase NS: NNS: NS: NNS:

Goood morning. Morning. This is Frank speaking. Thanks a lot for your great party. Oh, I'm glad you had a good time, but to tell the truth I feel pretty icky today. So I just want to. The reason I'm calling you is I'm looking desperately for my bag.

There are two obvious interpretations of the inappropriate learner move in situation 16 above. One is that after the performance of a thank you, the learner feels that he has engaged himself long enough in small talk, and shifts directly from his interactional goal (B) to an immediate communicative plan (H). Alternatively, the word "icky" may have caused interpretive problems, again leading to the same fallback strategy (see situation 13 above). In the retrospective interview, the learner in question reported that both interpretations were simultaneously true in his case. Situation 17: Can't Pay Rent NS: Well, my dear, if you can't pay for this week and for next week I'm afraid I'll have to let you go. NNS: No, you know I have always paid in time before.

Native I Normative Interactions

177

The learner appears to have been working on the additional argument in his favor that is brought forward, and under stress simply reacts negatively to the proposition "I shall have to let you go," before producing it. Situation 18: Can't Pay Rent NS: Can't you get a loan? NNS: Well, maybe, you can wait till next week because I have an interview then.

This data segment is similar (note that the same stressful situation holds). In both situations 17 and 18 the actual input is not taken into account by the learner, who seems to be altogether too busy planning and producing reasons why she shouldn't be thrown out. Such a phase-shift interpretation is in line with the subjects' own diagnosis of what went wrong: both learners admitted that they were too busy mapping out their argumentative plan to be able to listen, let alone respond to the more "local" utterances produced by their interlocutors. The response errors to be characterized in what follows are similar to the "interactional slips" described above: learners anticipate and predict a next move on the part of the interlocutor, and respond to it, even though it does not occur. The actually occurring input is then ignored. In terms of our model, the discourse meaning (box E) of an input is derived directly from the hearer/speaker's current predictions/expectations (box C). The system "hears" what it wants to hear (the same phenomenon occurs in native-speaker discourse, and is familiar in written discourse when one proofreads one's own manuscripts). Situation 19: Lost Property/Opening Phase NS: Oh, hello, Mike. NNS: Oh, fine. You? NS: I'm fine (puzzled pause).

Quite clearly, the student responds to a "How-are-you" move on the part of her interlocutor, even though it is not there. In fact, "how are you?" was a topic of discussion in several of the communication courses the learner-subjects had attended; in other words, this learner may have been primed by foreign language instruction to expect this input. This does not to my mind invalidate the point being made—that expectation can determine interpretation—but rather reinforces it. Situation 20: Lost Property/Opening Phase NS: NNS: NS: NNS:

Hi, Susie, how are you? Good morning, Kim, I'm fine. How are you? Oh, my head's terribly fuzzy. Oh, I'm glad to hear that.

In this situation the student's schema predicted a conventional fine, that is, a positive response to the conventionalized query, and this was responded to. Expectation again determined interpretation. There is, additionally, the possibility that fuzzy was not understood, such that decoding uncertainty may have influenced the predictive inteipretation. Situation 21: Gas Company/Closing Talk NS: I'll phone you right back and we'll arrange this whole thing again. NNS: Yes, thank you very much.

178

Discourse Perspectives NS: And once again, we're terribly sorry that we didn't get round to your place today. NNS: Oh, thank you.

This discourse stretch takes place inside the closing phase of a situation in which the learner had complained about an appointment not having been kept by an employee of a gas company. A final ritual apology is offered by the offender, which is, however, not responded to by a minimizer or some other appropriate move but by a repetition of a Thanks previously realized by the learner. This learner seems to have been operating inside a "closing talk" schema, knew that a Thank-you was her operative response, predicted a repetition of the compensatory offer made before, and responded to such a move as though it had occurred in reality. In the interview the learner reported that she had in fact engaged in some "advance planning," and apparently this advance planning barred her from flexibly reacting to the apology, which was unexpected as it referred back to an early phase in the encounter. As a last type of inappropriate response because of "psychopathological" slips, I would like to cite some instances that cannot be related to any one phase in the proposed discourse-processing model or even one of the suggested shortcutting paths: we deal here, rather, with cases in which it appears learners were unable to handle the cognitive load under the time constraints imposed by fast-moving oral interaction, and the necessity of chunking and weighting the incoming discourse stretches. The following example attests to such difficulties experienced by the learners in my corpus: Situation 22: Invite/Opening Phase NS: You've been working! Oh, I didn't know you had a job! What kind of job? NNS: Well, at least I'm trying.

The learner only responds to the first part of the interlocutor's utterance. The case is complicated by the connotations of the term "work." For some students, "work" is something one forces oneself to do. It has nothing to do with paid employment. Thus the response to the first part of the input may be adequate from the learner's perspective, but ignores the apparent misunderstanding evidenced in NS's further query. It is possible that this misunderstanding itself influences the content of the learner's answer (some people feel apologetic about admitting to the fact that one "works" occasionally), and this complex of factors leads to overload. It is also possible that the learner was not able to infer the interlocutor's different understanding of the term "work," and simply bypassed the problem by responding to his own interpretation of the input.

Conclusion The focus of this chapter has been on postulating a cognitive model inside which different kinds of "inappropriate responses" may be located. The major problem remains the validity of explanatory analyses offered inside this, or indeed any other framework. How does one "know" of a concrete piece of data which processing or knowledge constraints or difficulties led to its occurrence? Given our current state of knowledge, it seems to me that one can only plead for rich, context-based interpre-

Native I Normative Interactions

179

live hypotheses, established inside an explicit cognitive framework. I have attempted such here. Furthermore, this study supports the belief that introspective and retrospective data may be useful for interpreting observed pragmatic failure. I have suggested here some of the complexity of responding in a foreign language, and attempted to specify what problems may underlie instances of pragmatic-responding failure. Three types of problems were suggested, and an attempt was made to understand them. 1. Language-based difficulties leading to linguistic decoding and encoding problems 2. Conceptual and strategic deficiencies related to gaps in culture-specific pragmatic knowledge 3. Operational difficulties that may lead to interactional slips and inputs apparently being ignored. Furthermore, the notion of cognitive overload has been given some illustrative substance. One pedagogically relevant conclusion (that has been made often enough in the past (e.g., Edmondson, 1989; Edmondson & House, 1991) is the desirability of learners' having a thoroughly established battery of routine formulas for more successful discourse participation in L2. At the same time, the data above suggest that routines may be too fixed, as it were, in that nonexpected responses to routinized queries may shortcut a learner-discourse-processing system. The desirability of adequate strategic repair and elucidation strategies is a further pedagogic desideratum supported by the data. I have suggested several research strands might be relevantly searched and adapted for an eventual integrative-interdisciplinary approach to understanding and explaining "deviant" interlanguage responses and misunderstandings. I have integrated some of the insights of the five research strands into the frame provided by a discourse-processing model, and have supported this model with a number of illustrative conversations. Broadening our descriptive and explanatory framework is essential to understanding why learners' responses seem somewhat enigmatic at times. I conclude with a quote from Wallace Chafe: Understanding is increased by expanding the descriptive framework to include a large number of causal relations. Two things are necessary here. One is attention to a maximally wide range of data. How easy would it be to understand the workings of the solar system if we restricted ourselves to observing only the sun and moon, because they are bigger and brighter and thus easier to observe than the planets? The other necessity is having the ability to imagine a maximally wide range of explanatory models. Someone had to think of planets moving in elliptical orbits around the sun. . . . These two necessities can be encapsulated in the statement that understanding increases with an expanded field of vision. (1990, 19)

Notes 1. A similar distinction between responses in which linguistic cohesion is at stake (replies) and those in which underlying shared knowledge provides for coherence is made by Pomerantz (1984, 153ff.), who also discusses different types of problems speakers may find themselves faced with when pursuing a response.

180

Discourse Perspectives

2. Schegloff (1987) lists comparable causes for misunderstandings in talk-in-interaction, when he makes a distinction between two sources of "trouble": problematic reference ("propositions") and problematic sequential implicativeness, the latter comprising four distinct sources of trouble: favoured action interpretation (illocutions), serious/non-serious trouble sources (key/tone), as well as a so-called constructive/composite distinction and the practice of "joke first." 3. The following is an example of the type of brief situational and role outline informants were given at the outset of each recording session: You, student X, have just discovered that your car doesn't start, which annoys you greatly as you have a course at the university in twenty minutes. It is already fairly late, and as you're sure that you won't be able to make it by bike or bus, you decide to phone up a friend of yours, Y, to ask him for a lift.

An instruction matching K's role would stress the fact that Y had just decided that he would skip classes this morning or some other complicating factor.

References Abelson, R. P. (1976). A script theory of understanding attitude and behavior. In J. Carroll & T. Payne (Eds.), Cognition and social behavior, pp. 33-46 Hillsdale, NJ: Lawrence Erlbaum. Bartlett, F. C. (1932). Remembering. Cambridge: Cambridge University Press. Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine. Benson, D., & Hughes, J. (1983). The perspective of ethnomethodology. London: Longman. Bilmes, J. (1990). Mishearings. Unpublished manuscript, University of Hawaii at Manoa. Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal of Pragmatics, 11, 145-60. Blum-Kulka, S., & Weizman, E. (1988). The inevitability of misunderstanding: Discourse ambiguities. Text, 8, 219-41. Brown, P., & Levinson, S. (1987). Politeness. Cambridge: Cambridge University Press. Chafe, W. (1990). Looking ahead. Text, 10, 19-22. Clark, H. (1979). Responding to indirect speech acts. Cognitive Psychology, I I , 430-77. Clark, H., & Lucy, P. (1975). Understanding what is meant from what is said: A study in conversationally conveyed requests. Journal of Verbal Learning and Verbal Behavior, 14, 56-72. Clark, H., & Schunk, D. (1980). Polite responses to polite requests. Cognition, 8, 111-43. Dascal, M. (1977). Conversational relevance. Journal of Pragmatics, I , 309-28. Dascal, M. (1985). The relevance of misunderstanding. In M. Dascal (Ed.), Dialogue: An interdisciplinary approach (441-59). Amsterdam: Benjamins. Dascal, M., & Idan, A. (1988). From individual to collective action. In F. Vandamme and R. Pinxten (Eds.), Festschrift fiir Leo Apostelo (133-48). Ghent: Communication and Cognition. Dascal, M., & Katriel, T. (1979). Digressions: A study in conversational coherence. PTL: A Journal for Descriptive Poetics and Theory of Literature, 4, 203-32. Edmondson, W. J. (1981). Spoken discourse. London: Longman. Edmondson, W. J. (1987). "Acquiring" and "Learning": The discourse system integration hypothesis. In W. Lorscher & R. Schulze (Eds.), Perspectives on language in performance. Tubingen: Narr. 1070-89. Edmondson, W. J. (1989). Discourse production, routines and language learning. In P. Bierbaumer et al. (Eds.), Eng/isch als Zweitsprache. Tubingen: Narr.

Native I Normative Interactions

181

Edmondson, W. J., & House, J. (1981). Let's talk and talk about it. A pedagogic interactional grammar of English. Munich: Urban and Schwarzenberg. Edmondson, W. J., & House, J. (1991). Do learners talk too much? The waffle phenomenon in interlanguage pragmatics. In R. Philippson, E. Kellerman, L. Selinker, M. Sharwood Smith & M. Swain (Eds.), Foreign language pedagogy research: A commemorative volume for Claus Faerch (273-87). Clevedon and Philadelphia: Multilingual Matters. Erickson, F. (1975). Gatekeeping and the melting pot: Interaction in counseling encounters. Harvard Educational Review, 45, 44-70. Erickson, F., & Shultz, J. J. (1982). The counselor as gatekeeper. New York: Academic Press. Faerch, C., & Kasper, G. (1980). Processes and strategies in foreign language learning and communication. Interlanguage Studies Bulletin, 5, 47-118. Faerch, C., & Kasper, G. (Eds.). (1983). Strategies in interlanguage communication. London: Longman. Fillmore, C. (1976). Topics in lexical semantics. In R. Cole (Ed.), Current issues in linguistic theory. Bloomington: Indiana University Press. 76-138. Firth, J. (1964). The tongues of men, and speech. London: Oxford University Press. Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday. Goffman, E. (1974). Frame analysis. New York: Harper and Row. Goffman, E. (1981). Forms of talk. Oxford: Basil Blackwell. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and semantics (Vol. 3): Speech acts. New York: Academic Press. Gumperz, J. (1982a). Discourse strategies. Cambridge: Cambridge University Press. Gumperz, J. (1982b). Language and social identity. Cambridge: Cambridge University Press. Halliday, M.A.K. (1978). Language as social semiotic. London: Arnold. Heckhausen, H., & Beckmann, J. (1990). Intentional action and action slips. Psychological Review, 97, 36-48. Heikkinen, H., & Valo, M. (1985). Slips in interaction: The psychopathology of everyday discourse. In J. P. Forgas (Ed.), Language and social situations (213-28). New York: Springer. Holmes, J. (1985). Sex differences and miscommunication: Some data from New Zealand. In J. B. Pride (Ed.), Cross-cultural encounters. Communication and miscommunication (24-43). Melbourne: River Seine. House, J. (1979). Interaktionsnormen in deutschen und englischen Alltagsdialogen. Linguistische Berichte, 59, 76-90. House, J. (1982a). Conversational strategies in English and German dialogues. In G. Nickel & D. Nehls (Eds.), Error analysis, contrastive linguistics and second language learning (135-51). Heidelberg: Groos. House, J. (1982b). Gambits in deutschen und englischen Alltagsdialogen. Grazer Linguistische Studien, 17/18, 110-32. House, J. (1982c). Opening and closing phases in English and German dialogues. Grazer Linguistische Studien, 16, 52-83. House, J. (1984). Some methodological problems and perspectives in contrastive discourse analysis. Applied Linguistics, 5, 245-55. House, J. (1986). Cross-cultural pragmatics and foreign language teaching. In Seminar fur Sprachlehrforschung (Ed.), Probleme und Perspektiven der Sprachlehrforschung. Frankfurt: Scriptor. House, J. (1989). "Oh, excuse me please": Apologizing in a foreign language. In B. Kettemann, P. Bierbaumer, A. Fill & A. Karpf (Eds.). Englisch als Zweitsprache. Tubingen: Narr. 303-28.

182

Discourse Perspectives

Kasper, G. (1981). Pragmatische Aspekte in der Interimsprache. Tubingen: Narr. Kasper, G. (1982). Kommunikationsstrategien in der interimsprachlichen Produktion. Die Neueren Sprachen, 81, 578-600. Kasper, G. (1984). Pragmatic comprehension in learner-native speaker discourse. Language Learning, 34, 1-20. Keller-Cohen, D. (1979). Repetition in the non-native acquisition of discourse. In R. Freedle (Ed.), New directions in discourse processing. Norwood, NJ: Ablex. 245-71. Kintsch, W. (1988). The role of knowledge in discourse comprehension: A constructionintegration model. Psychological Review, 95, 163-82. Kintsch, W., & van Dijk, T. (1978). Towards a model of text comprehension and production. Psychological Review, 85, 363-94. Kochman, T. (1981). Black and white styles in conflict. Chicago: Chicago University Press. Lakoff, R. (1973). The logic of politeness, or, minding your p's and q's. Papers from the 9th Regional Meeting of the Chicago Linguistic Society (292-305). Chicago: Chicago Linguistic Society. Lakoff, R. (1976). Why you can't say what you mean [Review of Edwin Newman, Strictly Speaking}. Centrum, 4, 151-70. Lakoff, R. (1980). Psychoanalytic discourse and ordinary conversation. In R. Shuy & A. Schnukel (Eds.), Language use and the uses of language (269-87). Washington DC: Georgetown University Press. 269-87. Langer, E. (1978). Rethinking the role of thought in social interaction. In J. H. Harvey, W. J. Ickes & R. F. Kiddy (Eds.), New directions in attribution research (Vol. 2). Hillsdale, NJ: Lawrence Erlbaum. Langer, E., & Abelson, R. P. (1972). The semantics of asking a favor: How to succeed in getting help without really trying. Journal of Personality and Social Psychology, 24, 26-32. Langer, E., Blank, A., & Chanowitz, B. (1978). The mindlessness of ostensibly thoughtful action: The role of "placebic" information in interpersonal interaction. Journal of Personality and Social Psychology, 36, 635-42. Langer, E., & Imber, L. (1980). Role of mindlessness in the perception of deviance. Journal of Personality and Social Psychology, 39, 360-67. Leech, G. (1983). Principles of pragmatics. London: Longman. Minsky, M. (1975). A framework for representing knowledge. In P. Winston (Ed.), The psychology of computer vision. New York: McGraw-Hill. 211-77. Pomerantz, A. (1984). Pursuing a response. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action. Cambridge: Cambridge University Press. 152-63. Reason, J., & Mycielska, K. (1982). Absent-minded? The psychology of mental lapses and everyday errors. Englewood Cliffs, NJ: Prentice-Hall. Schank, R., & Abelson, R. P. (1977). Scripts, plans, goals and understanding. Hillsdale, NJ: Lawrence Erlbaum. Schegloff, E. A. (1987). Some sources of misunderstanding in talk-in-interaction. Linguistics, 25, 201-18. Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 7, 289-327. Scollon, R., & Scollon, S. (1981). Narrative, literacy and face in interethnic communication. Norwood, NJ: Ablex. Searle, J. (1983). Intentionality. Cambridge: Cambridge University Press. Searle, J. (1992). Conversation. In J. Searle, H. Parret, & J. Verschueren, (pp. 7-29) (On) Searle on conversation. Amsterdam: Benjamins. Smith, N. V. (Ed.). (1982). Mutual knowledge. New York: Academic Press. Sperber, D., & Wilson, D. (1986). Relevance. London: Basil Blackwell.

Native/Normative Interactions

183

Tannen, D. (1979). What's in a frame? In R. Freedle (Ed.), New directions in discourse processing. Norwood, NJ: Ablex. 137-81. Tannen, D. (1984a). Conversational style: Analyzing talk among friends. Norwood, NJ: Ablex. Tannen, D. (1984b). The pragmatics of cross-cultural communication. Applied Linguistics, 5, 189-95. Tannen, D. (1986). That's not what I meant. New York: Ballantine. Thomas, J. A. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112. van Dijk, T., & Kintsch, W. (1983). Strategies of discourse comprehension. New York: Academic Press. Weizman, E., & Blum-Kulka, W. (1988). Ordinary misunderstanding. Paper delivered at the Varna Conference on Models of Meaning, Varna, Bulgaria.

9 Explaining NNS Interactional Behavior: The Effect of Conversational Topic JANE ZUENGLER

As Kasper (1989) points out, interlanguage (IL) performance is now understood as a variable phenomenon. Much of the variability is both normal and systematic. IL variation, in fact, represents a major focus of current second language acquisition (SLA) research and theory, and is the theme of conferences and books in the field (e.g., the 1987 Applied Linguistics Conference at Ann Arbor, Michigan; the 1990 Second Language Research Forum [SLRF] at Eugene, Oregon; books by Eisenstein [ed.], 1989; Gass, Madden, Preston, & Selinker [eds.], 1989; Tarone, 1988). Increasingly, research has investigated IL performance in native speaker-nonnative speaker (NS-NNS) interactions. The findings point to a number of factors that cause IL performance to vary. One such factor emerging now is that of conversational topic. Though topic has been called "one of the most under-researched areas in the study of IL variation" (Tarone, 1988, 119), there is a small but growing body of empirical studies on conversational topic necessary to consider if we are to understand learner acquisition and performance of conversational skills. In addressing conversational topic, we are not simply looking at what it is that interlocutors talk about. Topic must not be viewed as a superficial, categorical construct independent of the conversation. The fundamental importance of topic is that it situates a speaker. First of all, speakers have cognitive and/or affective relationships to particular topics. For example, we may talk about computer software, and know a lot, a little, or nothing about it. Independent of how much we know, computer software may be something we are greatly, or not at all, interested in. Second, topic situates the speaker within the interaction, that is, vis-a-vis the interlocutor. In other words, it can shape one's conversational role. Taking computer software as an example, conversational roles are not simply a function of how much each interlocutor knows, or cares about, computer software in an absolute sense. How the interlocutors talk about software, and how active or not their roles, are significantly influenced by what their conversational partner knows and feels about the topic. Therefore, studying the influence of topic on NS-NNS interactions will enable us to learn more about conversational dynamics and role-taking in 184

Explaining NNS Interactional Behavior

185

general, and increase our understanding, in particular, of IL conversational performance and development. This chapter will first synthesize the research on conversational topic and IL performance. Following that will be a discussion of the research outcomes, as they relate to two current theoretical models of IL variation. The two models are Selinker and Douglas's Discourse Domain Model and Giles's Speech Accommodation Theory (SAT).

Research on IL and Conversational Topic1 The research germane to topic and IL has generally taken one of two focuses; it either compares the same learners' IL performance across various topics, or it compares performance by different IL learners when a particular topic is held constant. Measures of IL performance vary considerably among the studies. While a researcher may analyze the same set of IL features from one investigation to the next, there is no such uniformity from researcher to researcher. However, the lack of uniformity does not in itself represent a problem; it simply reflects the fact that in an emerging research area, there is a wide range of potential measures to consider. It will be apparent, in the research review, that some studies select traditional "linguistic" measures (e.g., pronunciation, verb accuracy) while others analyze features more recognizably pragmatic (e.g., directives, interruption behavior). Still others analyze both types of features. Whether the units of analysis are "linguistic" or "pragmatic," the overwhelming majority of the studies to be reviewed are, nevertheless, germane to the pragmatics of IL performance. The researchers are investigating whether, and how, topic influences the ways in which IL speakers perform a conversation, and, specifically, the type of conversational role they assume. Such research, therefore, has pragmatically relevant outcomes, even though, in some studies, the measures themselves may be more traditionally "linguistic" than pragmatic. Only two of the studies to be reviewed are not clearly pragmatically based (though they contain pragmatic measures). Cornu and Delahaye (1987) and Smith (1989) ask acquisitional questions, rather than questions of performance per se. They look at performance as a means, rather than an end in itself, because they are primarily interested in determining whether IL acquisition varies across discourse domains.

Studies of IL Performance Across Topics Since many IL speakers are university students studying their major field through the second language (i.e., have a "language for specific purposes" [LSP] orientation), several studies have compared subjects' IL performance when conversing on a topic within their major field with their performance on a topic lying outside their major field. Two of the studies in particular illustrate the impact of topic on the roles that interlocutors take in the conversations. Selinker and Douglas (1985) report on an IL speaker (Luis) participating in two

186

Discourse Perspectives

conversations, one concerning civil engineering, his major field, and the other concerning a life-story topic (how to prepare his native food). Luis's performance varied from one conversation to the next. Each topic, the authors suggest, led to different role dynamics between Luis and his interlocutor, resulting in variable IL performance. Selinker and Douglas point out that when Luis conversed about his major field, he was the relative "knower" in the interaction, since it was his specialty, and his interlocutor was a NS who was not in the same field. Therefore, Louis's role was to make evident his expertise. In conversing about food, on the other hand, Luis was outside his area of expertise and interest, and the same was true for his interlocutor, who in this case was a friend of his. In this type of interaction, there was more role negotiation, partly because the interlocutors were friends, and partly because neither "knew" the topic. The topic-influenced role differences led to variation in some of Luis's communication strategies. In particular, there was variation in the strategy Luis used when he searched for a word he did not know. When conversing about engineering, Luis (as "knower") continued talking until he found a synonym his interlocutor could respond to. However, when he needed vocabulary to discuss food, Luis was more likely to let the episode break down, saying "forget it" at one point. There was no impetus for Luis to show that he was the "knower" of food (because he was not), and he very likely held less interest in that topic than in the major field topic. Luis also varied the way he corrected his interlocutor. This occurred when the NS interlocutor misinterpreted the content of what Luis was saying. When discussing engineering, Luis corrected his interlocutor very directly without mitigating what he said. When discussing the life-story topic, however, Luis mitigated his corrections by using politeness strategies. Here again, Selinker and Douglas suggest, the cross-topic variation in response to interlocutor mistakes may reflect the relative importance of the topic to the IL speaker. Misinterpretations by the interlocutor when talking about food preparation may not be as urgent to correct because that topic is probably much less important to the IL speaker than is his major field. Results from a larger study by Zuengler and Bent (1991) concur with Selinker and Douglas (1985) in pointing to the influence of topic on the conversational roles taken by IL speakers. The Zuengler and Bent research involved, in part, an analysis of 45 IL speakers conversing on two separate topics in dyads with NS interlocutors. (See Zuengler 1989b for a subset of the results). All 45 NS-NNS dyads were first asked to have a conversation about favorite and holiday foods. Then, each dyad discussed a topic within their shared major field. The researchers looked for patterns of conversational participation and dominance, employing measures such as amount of talk, use of pause fillers to retain one's turn, topic moves, and back-channels. Outcomes revealed variation in IL conversational participation from topic to topic. When discussing food, the NNSs (relative to their NS partners) produced significantly more talk and more pause fillers, while their NS partners back-channeled significantly more, role patterns that Zuengler and Bent labeled NNS "speaker" and NS "listener." These particular role patterns did not necessarily appear, however, when the subjects had a conversation concerning their major field. In dyads in which the NSs and NNSs shared the same major field and were at an equivalent level in it, the NSs talked significantly more than their NNS partners, while the

Explaining NNS Interactional Behavior

187

NNSs back-channeled significantly more. That is, the role patterns were opposite to those in the conversations about food. When discussing the major field topic, it was the NSs who performed as active "speakers" and the NNSs who were the active "listeners." (In both conversations, the NNSs produced more pause fillers. Topic moves were nonsignificant with both topics.) The NS as "speaker" in the major field conversations supports claims by Beebe and Giles (1984), among others, who assert that NSs tend to dominate interactions with NNSs, perhaps due to feelings of ethnolinguistic superiority (but see discussion below). However, the food conversations showed no such NS dominance, and in fact, as pointed out, the NNSs spoke significantly more than the NSs. Zuengler and Bent (1991) suggest that while discussing food, the NSs may have conceded more of the talk to the NNSs because the topic held little importance to the NSs. Or, the NSs may have judged the NNSs as knowing more than they about the food topic. Since the NNSs resided in the NSs' setting (namely, the United States), while the NSs had not, correspondingly, resided in the NNSs' settings, the NNSs tended to be familiar with NS foods, whereas the NSs did not typically show conversance with NNS foods. Relative knowledge of the topic was, in fact, an important influence on conversational roles in other comparisons tested by Zuengler and Bent (1991). (These outcomes will be brought up below.) Several additional studies looked at IL performance on major field versus nonmajor field topics, but focused primarily on specific characteristics of the IL per se, rather than on interactional dynamics such as role relationships. Cornu and Delahaye (1987) tested two subjects and report IL variation between major field and nonmajor field topics. Providing detailed results for one of the two subjects, they report that when the IL speaker talked about economics, her major field, she produced more embedded sentences, used fewer nonverbal strategies, performed more self-correction of grammatical errors, and was less hesitant about and more appropriate in her lexical choices. Cornu and Delahaye conclude that the subject was more in control of the major field topic than the nonmajor field topic; she had more control of the economics topic, they suggest, because her IL contained more "integration of the target lexical systems" (1987, 150). Greater integration (and, consequently, more control) can result from a learner's having greater exposure to and practice in the TL in topic areas such as the major field. A recent study by Smith (1989) compared IL performance on general topics in the SPEAK test (a test of spoken language proficiency developed by ETS) versus performance on major field topics in a specially developed version of the same test. Thirty-eight graduate students were tested, and global measures were obtained of their comprehensibility, pronunciation, grammar, and fluency. In contrast to the preceding studies' outcomes, these subjects showed no significant differences in their IL performance on major field versus nonmajor field topics. One possible explanation for the lack of difference in performance is test anxiety as a variable. Smith (1989) reports that a native-speaker control group that took the test indicated that they found each test "quite stressful" (160). If native speakers found the task stressful, it is quite likely that the IL speakers did as well. Further research in this area is important for determining whether test anxiety was indeed important in eliciting data such as in Smith (1989).

188

Discourse Perspectives

In addition to looking at IL performance on major field versus nonmajor field topics, research has investigated performance on topics that differ according to speaker interest or affect. Eisenstein and Starbuck (1989) asked eight IL speakers what topics they were interested in as well as topics they considered uninteresting. (The researchers referred to these as "invested" and "uninvested" topics, respectively.) Each subject had two conversations with the researchers, one on an invested topic, and the other on an uninvested topic. What Eisenstein and Starbuck analyzed was the subjects' targetlike accuracy of their verb system across topics. Overall, the subjects' IL was significantly less accurate when discussing the invested topic than when discussing the uninvested topic. Specifically, subjects were less accurate in lexical choice, tense choice, tense form, and use of repetitions. The researchers suggest that the differences across topics might be due to monitoring or cognitive load. When discussing the invested topic, subjects talked more, and this may have caused them to monitor less. Or, it is possible that the cognitive load was greater in the invested condition (though the subjects were at an advanced IL level, and the invested topics did not appear to be cognitively demanding). Additional research should help clarify whether monitoring or cognitive load may be confounding variables in invested-uninvested topic comparisons. Certain topics can trigger positive or negative emotions in IL speakers, and their performance may vary as a result. A pilot study reported by Zuengler (1982) consisted of asking 13 IL speakers questions on two different topics, a "neutral" one (asking them their views on children watching television), and one that was potentially ethnically-threatening (telling subjects that they should speak English in the United States, and if they wanted to speak their native language, they should "go back to [their] own country"; subjects were asked their views on that). Subjects' responses to the two different topics were compared with respect to their production of three phonological variables. Zuengler provides descriptive data illustrating that some of the subjects were indeed emotionally aroused by the ethnically threatening topic. Those subjects had lower accuracy when responding to the ethnically threatening question than in responding to the neutral question. On the other hand, subjects who did not respond negatively to the ethnically threatening question (perhaps because their ethnic identity was weak) performed more accurately when answering the ethnically threatening question. Emotionality of topic was also a major focus in research by Dowd (1984). However, in contrast to Zuengler (1982), Dowd's research did not concern emotion attached to ethnic identity. Instead, Dowd asked her subjects to talk about an unpleasant personal experience they had had. Prior to the "emotional" topic, they were asked to discuss a "neutral" topic, for example, to describe their favorite food, their home, family, etc. Subjects were 100 Mexican female IL speakers. Dependent measures of IL performance were six phonological variables. Results were mixed. Only two of the six variables, final consonant clusters and /r/, significantly differed in accuracy from one topic to the next. The pattern of shift in accuracy also differed. In going from "neutral" to "emotional" topics, subjects significantly decreased in TL accuracy on final consonant clusters, but they significantly increased in accuracy on /r/. Though Dowd's (1984) outcomes reveal an (albeit limited) effect of topic shift, it is not known why the variables shifted in different directions. Such

Explaining NNS Interactional Behavior

189

bidirectional shifts are important for SLA researchers to understand (Dowd, Zuengler, & Berkowitz, 1990; see also Tarone, 1985). It will be noted that though the Zuengler (1982) and Dowd (1984) research may have had a general pragmatic orientation, their measures for analysis were limited to linguistic items. Obviously, it will be important to analyze pragmatic measures as well as to view IL performance in such research.

IL Variation Within a Given Topic Three research studies investigate IL performance when the topic is held constant, the important variable being IL speakers' knowledge of the topic relative to that of their (NS) interlocutors. All three studies consider relative topic knowledge as a definer of one's conversational role (and thus are closely related in conception to several studies discussed earlier). The present studies asks, implicitly or explicitly: Can an interlocutor's knowledge of the conversational topic, relative to the partner's knowledge, explain whether one interlocutor dominates, controls, or directs the other, or whether both interlocutors play an equal role in building the conversation? As mentioned earlier, there are claims by Beebe and Giles (1984) and others that in many NS-NNS interactions, the NS tends to dominate, due, perhaps, to feelings of ethnolinguistic superiority. Investigations of relative topic knowledge provide an important means of testing such claims. Woken and Swales (1989), while limiting their study to three female NS-NNS dyads, provide clear evidence for control in the NNS's performance when she had greater topic knowledge. Each of the NS-NNS dyads was composed of an NNS graduate student in computer science who was asked to show her NS partner (who was not a computer science major) how to use some software. The researchers report that in all three dyads, the NNSs, due to their topic expertise, clearly controlled the conversation in talking more (and producing longer t-units), correcting the NS more than the NS corrected them, and in giving more directions. The NNS corrections to the NSs were largely unmitigated (as were those of the NNS in Selinker and Douglas [1985], when discussing his major field knowledge). The NSs, on the other hand, produced more inquiries, which served as requests for clarification and for lexical help from the NNSs. Woken and Swales (1989) conclude that their three dyads illustrate the influence of relative topic expertise on control and participation in a conversation, regardless of the nativeness or nonnativeness of the interlocutors. As such, the evidence of NNS control contradicts the claims of Beebe and Giles (1984) and others regarding NS tendency to dominate (or at the least, it qualifies any such claims regarding the NS). However, it remains to be determined whether less fluent NNSs than these would also control such conversations. Like Woken and Swales (1989), the Zuengler and Bent (1991) study introduced earlier also addressed relative topic knowledge in NS-NNS interactions, but in a larger number of dyads. Forty-five male NS-NNS dyads were formed to constitute three distinct groups. In all dyads, NSs and NNSs shared their major field and were asked to have a conversation on a specific topic within the field. In 15 of the dyads, the NNS was the relative "topic expert," in 15 other dyads, the NS was the relative

190

Discourse Perspectives

"topic expert," and in the remaining 15, NS and NNS were relatively equal with respect to topic knowledge. (Topic expertise was determined by what level the subject was at in the major field.) Measures of conversational participation and dominance included amount of talk, pause fillers, back-channels, topic moves, and interruptions. While results were nonsignificant for interruptions, each of the other measures revealed some significant patterns. When the NNSs had relative topic expertise, they produced significantly more talk, more fillers, and more backchannels than their NS partners. When, on the other hand, the NSs were the "topic experts," the NSs produced more talk and more fillers. As discussed earlier, when both NS and NNS were at the same stage in the major field, the NSs talked significantly more, while the NNSs produced more pause fillers. Interestingly, regardless of topic expertise, the NNSs produced significantly more back-channels. Zuengler and Bent (1991) view the results as a clear indication of the influence of relative topic knowledge on one's role in a conversation; topic expertise led to greater amounts of talk and more pause-filling. At the same time, the NNSs, regardless of topic expertise, tended to signal an active listening role as well, by greater use of back-channels. It was only when the NNSs were the "topic experts" that they displayed speaker dominance by talking more. As mentioned above in the discussion about cross-topic variation, the fact that the NSs talked more than the NNSs in the "equal" dyads perhaps indicates a tendency for NSs to dominate due to feelings of ethnolinguistic superiority, even though they may be "equal" with respect to topic knowledge. A third study illustrates that relative topic expertise can influence 1L performance even if it is only perceived, rather than actual, expertise. Zuengler (1989a) administered an art judgment test to 45 dyads of NS-NNS women. She then asked them to have two conversations about their individual judgments. In the first conversation, subjects did not know how well they had performed on the test. However, prior to the second conversation, Zuengler manipulated the scores so that in 15 of the dyads, the NNS was led to believe that she had performed much better in making art judgments than her NS partner had (i.e., that the NNS was the relative "topic expert"), while in another 15 dyads, it was the NS who was led to believe that she was the relative "topic expert." (The manipulated scores had no relation to subjects' actual scores.) To establish a control group, in 15 of the dyads neither interlocutor was told how she had performed. All 45 dyads then had a second conversation about their judgments. Zuengler (1989a) reports an analysis of amount of talk, interruptions, and task moves in an effort to determine whether perceived expertise had an effect on conversational performance. Outcomes revealed that the NSs consistently dominated in terms of amount of talk; NSs produced significantly more talk than their NNS partners, across all conversations. Moreover, topic expertise (in the second conversation) led the NS perceived experts to increase their already greater amount of talk in the first conversation. The other measures revealed two patterns: (1) perceived expertise led both NSs and NNSs to interrupt more successfully and to move the task along more often than their perceived nonexpert partners; and (2) the NSs tended to dominate in the control group conversations. They not only produced more talk, but they interrupted more successfully and moved the task along more frequently than their NNS part-

Explaining NNS Interactional Behavior

191

ners did. According to Zuengler (1989a), the findings show that when topic expertise differences are not apparent to either interlocutor, it is the NS who will tend to dominate a conversation with an NNS. The outcomes of Zuengler and Bent (1991), discussed above, corroborate this. However, when interlocutors are aware of differences in topic expertise (even though the differences may be apparent, not real), the relative topic expert, NS or NNS, will tend to dominate the conversation. The only exception is in amount of talk, which the NSs produced more of overall. Zuengler (1989a) suggests that the NNSs may have talked less because of limitations in their language proficiency. While none were IL beginners, it was apparent, in listening to the conversations, that many of the NNSs showed difficulty in producing fluent talk. Analysis of speech rate revealed that NNSs consistently spoke significantly more slowly than NSs. NNSs often signaled difficulty by uttering "I don't know how to say this" and "this is difficult for me." To summarize, the research conducted on conversational topic addresses two questions: Does IL performance vary from one topic to the next? And, does IL performance vary within a given topic? The review of research on conversational topic reveals that both types of variation are evident in IL performance. (Of course, caveats should be attached, as not all of the studies have statistically tested results, and one of the studies [Smith, 1989] had nonsignificant outcomes). While it is obvious that more research must be undertaken in this area, the research conducted to date already provides evidence that topic does indeed affect IL performance.

Explaining the Outcomes: Looking to SLA Theory It is important to recognize that it is not topic per se that is of importance to SLA research; rather, it is what topic reveals about IL development and use. That is, how can research on topic contribute to theories of second language acquisition, and, at the same time, how can language acquisition theories help us interpret the outcomes of the research? While some of the research studies discussed in this paper address theory, either in their conception or in interpreting their outcomes, others do not. Whether the researchers directly address theory or not in their studies, the majority of the research outcomes can be shown to relate to either of two current SLA theories in particular: the Discourse Domain Model and Giles' Speech Accommodation Theory (SAT). Considering the outcomes that relate to these two theories will help us understand topic as part of the larger picture of language acquisition and use, and, at the same time, help us assess the viability of the theories in accounting for the data.2

The Discourse Domain Model: IL Development and Use Varies According to Domain The Discourse Domain Model (Douglas & Selinker, 1985; Selinker & Douglas, 1985) is a cognitively oriented theory which states that learners develop ILs through various content areas, or discourse domains, that are important to or needed by them. 3 Selinker and Douglas claim that because IL development is domain-situated,

192

Discourse Perspectives

a learner's IL may be well-developed, structurally and functionally, in one domain, but not in the next. Fossilization, backsliding, and disfluencies in IL can occur when learners are asked to converse within a domain that they have little or no knowledge of (and/or do not know the forms and functions for conversing about it in the L2). While Selinker and Douglas conceptualize discourse domain without using the word "topic," it can be argued that the latter is a performance element of the former.4 That is, people do not talk about discourse domains; instead, they talk about "things" (topics) that engage a domain of discourse. As such, a discourse domain theory has direct relevance to studies of IL and topic, particularly to some of the research discussed earlier which illustrates IL variation across topics. Discourse domain theory can explain the effect of knowing, or being interested in, a topic. Subjects in Selinker and Douglas (1985) and Cornu and Delahaye (1987) appeared to speak more fluently and assertively when discussing their major field than when talking outside their major field. The major field topic engaged a discourse domain of which the subjects had more cognitive control than the other topic's domain. (However, test anxiety may undermine the control, and thus explain Smith's [1989] lack of significant differences.) Asking subjects to talk about topics of varying interest to them (Eisenstein & Starbuck, 1989) may also illustrate discourse domain theory (though further research is needed to establish this). Often, what we have differing interests in we also have differing amounts of knowledge about. That Eisenstein and Starbuck's (1989) subjects were less formally accurate in discussing their topic of interest may indicate that they had engaged a discourse domain in which they were quite comfortable, and in talking more (which they did), they neglected to monitor their speech. Or, as the authors also suggest, the cognitive load of the invested (interesting) topic may have been greater, causing the subjects to be less formally accurate. In either case, it is discourse domain theory, in suggesting that IL can vary in development and use from domain to domain, that can provide an explanation for the cross-topic IL variation observed in the studies just cited. Speech Accommodation Theory (SAT): IL Use Varies According to Interactional Dynamics Topic and discourse domains, nonetheless, are not solely cognitive. When IL speakers are interacting with someone, the research has shown, how much they know about the topic will influence their IL use. However, it is not strictly the speaker's absolute knowledge of the topic that will determine how active a conversational role she will take; it is the speaker's knowledge compared to the interlocutor's knowledge. In other words, topic knowledge within an interaction is interactionally defined. It is in this sense that SAT is relevant for explaining some of the outcomes. By combining several social psychological theories, Giles's Speech Accommodation Theory (SAT) explains why speakers adjust (i.e., accommodate) their speech while interacting with others. Conversations are perceived by participants as being either inter-individual or inter-group encounters; if the latter, it is often ethnic differences that become salient. Speakers can show solidarity, create distance, or maintain their position, by converging toward their partner's speech patterns, by

Explaining NNS Interactional Behavior

193

diverging, or by not making any adjustments. Detailed explanation of SAT, as well as its application to SLA, is outside the scope of this discussion; the reader is referred instead to Beebe (1988), Beebe and Giles (1984), and Beebe and Zuengler (1983). As a theory of conversational dynamics, SAT can explain some of the IL variation that is beyond the conceptual scope of the discourse domain model. At the same time, the research outcomes can offer a needed elaboration of SAT. First of all, SAT, in emphasizing interactional dynamics, can provide an explanation for the outcomes in Zuengler (1989a), where perceived, and not actual, expertise differences led to IL variation. Interlocutors were led to have certain perceptions of their own and their partners' topic knowledge, and it was interlocutor perceptions (and not actual knowledge) that influenced IL performance. Even when actual knowledge levels are involved (see Zuengler & Bent, 1991; also, Woken & Swales, 1989), topic knowledge is interactionally determined according to comparisons the interlocutors make of each other. That is, the outcomes show that IL performance can be a function of the speaker's knowing more, or less, about the topic than the interlocutor knows, rather than how much absolute knowledge of the topic the speaker has, divorced from the interaction. As mentioned above, SAT posits that interactions can be viewed by interlocutors as being inter-group or inter-individual encounters. Primarily, SAT points to salient ethnic group differences between interlocutors as illustrating what would be perceived as an inter-group encounter. When ethnicity (or other group differences) is not made salient, it is likely, according to SAT, that the encounter will be perceived as an inter-individual one, with interlocutors viewing each other on an individual, rather than group, basis. Research suggests that it is topic that may be an important determinant of whether the NS-NNS encounter will be an inter-group or inter-individual encounter. In Zuengler (1982), IL speakers were given a question to respond to that threatened their ethnic identity. This approach, certainly, made ethnicity salient and a number of the subjects responded by defending their ethnicity. For these subjects, the topic focused the interaction on ethnic differences and made the interaction, for them, an inter-group encounter. Where conversational topics draw on NSs' and NNSs' individual experiences and training, instead of emphasizing their ethnic, or NS-NNS differences, we would expect the interlocutors to experience the interaction as an inter-individual encounter. We can point to the outcomes showing that conversational performance is a function of relative topic knowledge (where the topic is not ethnically salient). IL performance varied according to relative topic knowledge differences in Woken and Swales (1989), Zuengler (1989a), and Zuengler and Bent (1991). In these conversations, the IL speakers knew (or thought they knew) more or less than their NS interlocutors; their performance varied according to these knowledge differences. Clearly, then, what was salient to these interactions was an inter-individual dynamic (the interlocutor differences in topic knowledge). However, it is necessary to consider the outcomes of two studies (Zuengler, 1989a; Zuengler & Bent, 1991) that showed that the NSs dominated when the interlocutors were supposedly equal in topic knowledge. That is, the conversational performance split along NS-NNS lines, with the NS as dominator and the NNS as

194

Discourse Perspectives

subordinate. This indicates, I knowledge differences between an inter-group encounter, with ethnolinguistic superiority (see

would argue, that if there are no apparent topic the NSs and the NNSs, the interaction may become the NSs, unfortunately, exercising their beliefs of Beebe & Giles, 1984).

Conclusion As we have seen, two current theories of SLA, the Discourse Domain Model and SAT, can provide explanations for the IL performance variation apparent in the majority of the studies on conversational topic reviewed here. They provide complementary contributions, since the Discourse Domain Model is cognitively based and SAT is a social psychological theory. After all, when looking at the influence of topic on NS-NNS interactions, we must recognize that "topic" is both a cognitive and a socially situated construct. Recognizing this will not only help us understand the present research outcomes but will also enable us to plan the needed future work in this area.

Notes 1. Because conversational topic is the focus of this paper, research studies involving written IL and topic will not be included. For a discussion of some of the research on written IL variation, see Selinker and Douglas (1987). 2. While it is essential, in theory evaluation, to determine the explanatory adequacy of the theory per se, such an evaluation is beyond the scope of this paper. See McLaughlin (1987). 3. Tarone (1988) states that the Discourse Domain Model stems from social psychology, but does not provide support for this assertion. Selinker and Douglas, in their explanation of the model, clearly present it as a cognitive theory. 4. Therefore, criticisms that a concept such as "discourse domain" is too vague for methodological purposes (see Tarone, 1988) may be unfounded. One could argue that it is topic that represents the operationalization of discourse domain.

References Beebe, L. M. (1988). Five sociolinguistic approaches to second language acquisition. In L. M. Beebe (Ed.), Issues in second language acquisition: Multiple perspectives (43-77). New York: Newbury House. Beebe, L. M., & Giles, H. (1984). Speech-accommodation theories: A discussion in terms of second-language acquisition. International Journal of the Sociology of Language, 46, 5-32. Beebe, L. M., & Zuengler, J. (1983). Accommodation theory: An explanation for style shifting in second language dialects. In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition (195-213). Rowley, MA: Newbury House. Cornu, A. M., & Delahaye, M. (1987). Variability in interlanguagc reconsidered: LSP vs. non-LSP IL talk. English for Specific Purposes, 6, 145—52.

Explaining NNS Interactional Behavior

195

Douglas, D., & Selinker, L. (1985). Principles for language tests within the "discourse domains" theory of interlanguage: Research, test construction and interpretation. Language Testing, 2, 205-21. Dowd, J. L. (1984). Phonological variation in L2 speech: the effects of emotional questions and field-dependence/field-independence on second language performance. Unpublished doctoral dissertation, Teachers College, Columbia University. Dowd, J., Zuengler, J., & Berkowitz, D. (1990). L2 social marking: Research issues. Applied Linguistics, 10, 16-29. Eisenstein, M. R. (Ed.). (1989). The dynamic interlanguage: Empirical studies in second language acquisition. New York: Plenum. Eisenstein, M. R., & Starbuck, R. (1989). Investment in topic and verb system accuracy of advanced L2 learners. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition (Vol. 2): Psycholinguistic issues (125-37). Clevedon and Philadelphia: Multilingual Matters. Gass, S., Madden, C., Preston, D., & Selinker, L. (Eds). (1989). Variation in second language acquisition (Vol. 1): Discourse and pragmatics (Vol. 2): Psycholinguistic issues. Clevedon and Philadelphia: Multilingual Matters. Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden, D. Preston & L. Selinker (Eds.), Variation in second language acquisition (Vol. 1): Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters. McLaughlin, B. (1987). Theories of second-language learning. London: Edward Arnold. Selinker, L., & Douglas, D. (1985). Wrestling with "context" in interlanguage theory. Applied Linguistics, 6, 190-204. Selinker, L., & Douglas, D. (Eds.). (1987). Interlanguage [Special issue]. English for Specific Purposes, 6(1). Smith, J. (1989). Topic and variation in ITA oral proficiency. English for Specific Purposes, 8, 155-68. Tarone, E. E. (1985). Variability in interlanguage use: A study of style-shifting in morphology and syntax. Language Learning, 35, 373-404. Tarone, E. E. (1988). Variation in interlanguage. London: Edward Arnold. Woken, M. D., & Swales, J. (1989), Expertise and authority in native-non-native conversations: The need for a variable account. In S. Gass, C. Madden, D. Preston & L. Selinker (Eds.), Variation in second language acquisition (Vol. 1): Discourse and pragmatics (211-27). Clevedon and Philadelphia: Multilingual Matters. Zuengler, J. (1982). Applying accommodation theory to variable performance data in L2. Studies in Second Language Acquisition, 4, 181-92. Zuengler, J. (1989a). Assessing an interaction-based paradigm: How accommodative should we be? In M. R. Eisenstein (Ed.), The dynamic interlanguage: Empirical studies in second language acquisition (49-67). New York: Plenum. Zuengler, J. (1989b). Performance variation in NS-NNS interactions: Ethnolinguistic difference, or discourse domain? In S. Gass, C. Madden, D. Preston & L. Selinker (Eds.), Variation in second language acquisition (Vol. 1): Discourse and pragmatics (22844). Clevedon and Philadelphia: Multilingual Matters. Zuengler, J., & Bent, B. (1991). Relative knowledge of content domain: An influence on native-non-native conversations. Applied Linguistics, 12, 397-415.

10 The Metapragmatic Discourse of American-Israeli Families at Dinner SHOSHANA BLUM-KULKA and HADASS SHEFFER

In the broadest sense, interlanguage pragmatics is concerned with the ways in which nonnatives do things with words in a second language. The phenomena investigated from this perspective may relate to both pragmatic and discoursal knowledge; in practice, the emphasis has been on contrasting native and nonnative speech act performance, on detecting learners' inappropriate speech act realization and accounting for pragmatic errors or deficits in terms of both source and communicative effects (see Blum-Kulka, House, & Kasper, 1989 for an overview). We deviate from this tradition here in several respects. Here we shall be concerned with interlanguage pragmatics phenomena occurring in the speech of bilingual speakers, for whom bilingualism is a day-to-day social reality. Ours is not a learners' population; we shall be looking at families interacting around the dinner table in three groups of families: native-born Israelis, native-born Americans, and our target group, American immigrants to Israel. The immigrant group is composed of middle- to upper-middle-class families living in Israel from 9 to 19 years. The discoursal styles used by family members at the dinner table in this group have been compared with the dinner-table talk of socioeconomically similar native Israeli and native American families. All members of the immigrant families are bilingual (English-Hebrew); both languages are being used at the dinner table, with English being dominant. The adults in these families are hence speaking mostly in their native language (English) and to a lesser degree in their second language (Hebrew). The reverse is true for the children: most children partake in family discourse in what for them is a second language (English), reverting occasionally to Hebrew, their native language. Is such a peculiar bilingual language situation the relevant context to look for interlanguage pragmatics phenomena? After all, we have been used to thinking about interlanguage as an L2 (second language) specific notion; this assumption underlies all studies to date, whether linguistically or pragmatically oriented. Yet there is another context where systematicity in LI "errors" does come under consideration: the domain of attrition studies (see Weltens, 1987 for a review). As in our 196

American-Israeli Families at Dinner

197

case, first language attrition studies focus on immigrant populations. But they differ from ours in two important respects: since they are concerned with changes that occur over time in a first language, it is this first language that serves as the criterion norm. As reported by Wei tens, the focus in first language attrition studies is on language loss as an intergenerational phenomenon over time, as established from questionnaires, or on particular linguistic phenomenon at a given point in time (e.g., Sharwood-Smith, 1983). In our case, language use in the immigrant families is studied at a particular point in time and is compared to both native American and native Israeli use. Furthermore, differing from that of attrition studies, our focus is on pragmatic, rather than on linguistic, phenomena. The bilingual families studied are, in effect, in contact with two incongruent pragmatic systems, each realized by a culturally specific style of language use. The possibility of a two-way interference under such conditions was already recognized by Weinrich (1953, 1): "Those instances of deviation from the norms of either language which occur in the speech of bilinguals as a result of their familiarity with more than one language, i.e. as a result of language contact, will be referred to as interference phenomena." We propose to demonstrate that indeed, as implied by Weinrich, language contact creates bi-directional effects, both from the first language to the second and vice versa. The latter type of effect, namely, interference from the second to the first language, has been noted in the very few studies to date that addressed this issue. 1 We shall claim that rather than just showing traces of "interference," the mutual influence between two sociolinguistic systems in the case studied results in the emergence of a specific, pragmatic interlanguage. We are interested in use rather than in usage (Widdowson, 1978), specifically in the pragmatic norms that govern a particular speech community's interactional style (Hymes, 1974) and its modes of pragmatic socialization. Studies in the ethnography of speaking and cross-cultural pragmatics have clearly shown that ways of speaking are culturally determined and that culture-specific interactional styles differ along dimensions such as levels of directness in speech act realization (e.g., Tannen, 1981; Blum-Kulka, House, & Kasper, 1989), interpretive strategies and signaling devices (Gumperz, 1982), and conceptions and practices of "polite" behavior (Basso, 1979; House & Kasper, 1981). Language in use is a powerful medium of sociocultural socialization. Children acquire tacit knowledge of systems of belief and cultural ways of speaking through exposure to and participation in languagemediated interaction in the home (Gleason & Weintraub, 1976; Heath, 1983; Schieffelin & Ochs, 1986; Ochs, 1988; Schieffelin, 1990). The cross-cultural comparison of Israeli to American Jewish families in our sample reveals the two groups differing on several dimensions of language use (Blum-Kulka & Katriel, 1991; Blum-Kulka & Snow, 1992) and practices of pragmatic socialization (Blum-Kulka, 1990). In previous work that focused on requestive behavior we found that the families we observed manifest a bicultural, or intercultural style systematically different from both the Israeli and American patterns. For example, immigrant families realize requests less directly than native Israelis, but more directly than Americans living in the United States. Furthermore, this intercultural style is realized regardless of the language spoken. Speakers manifest characteristics of this style when

198

Discourse Perspectives

speaking both their first and second languages (Blum-Kulka, 1988; Blum-Kulka, 1991). In this chapter we shall address these issues from a metapragmatic perspective. Our underlying assumption is that a speech community's interactional style is indexed both by the ways in which its members realize different speech acts as well as by the ways in which they comment on language use: metapragmatic comments (MCs) can reveal the pragmatic norms governing use and socialization. As in the case of requests, our findings show the metapragmatic discourse of the immigrant families to differ meaningfully from parallel discourse in both the American and Israeli families, whether realized in a first or a second language. The thrust of our argument is that under certain conditions, interlanguage pragmatics can be realized in the first language. Furthermore, for bilinguals in conditions of a twofold cultural contact, pragmatics may be the first area affected by exposure to the sociocultural influence of the second language. This argument is developed as follows: we shall (1) describe the background to the study and our methods of data analysis; (2) present our findings with regard to inter- and intra-group diversity in metapragmatic discourse; (3) discuss the unique characteristics of the bilingual families' IL metapragmatic discourse, as manifested by metacomments on language usage in the course of dinner-table conversations; and (4) examine the degree of linguistic and pragmatic awareness of bilingual speakers in general by considering data from interviews with the immigrant families in our sample.

Background The 24 families studied are all native-born, middle- and upper-middle-class Jewish families, with two to three school-age children. Families were selected through the snowball technique. The criteria for inclusion were: (a) college education for both parents, (b) professional occupation for both parents, (c) European family origin for both parents, (d) being religiously nonobservant, (e) being native-born Israeli or American (for both immigrants to Israel and families taped in the United States). For American immigrants to Israel, we required a minimum of nine years of residence in Israel. As can be seen in Table 10.1, most children (16 out of 29) in the immigrant families were born in Israel and have not spent more than a month or two in an English-speaking environment. Six children have spent two years abroad and two have spent five. Out of the four children born in the United States, two came to Israel at the age of one, one at the age of two, and one at the age of four and a half. Yet though Hebrew has become unquestionably the dominant language2 for all of these children, English plays an important role. English is the main language spoken at home, even though it is often mixed with Hebrew (Olshtain & BlumKulka, 1989). Adults and children alike practice their bilingualism not only in speaking (speaking English at home and Hebrew at work and at school) but also in reading. The adults all report reading in English for pleasure, and 13 out of the 14 school-age children report reading in both languages.

American-Israeli Families at Dinner

199

Table 10. 1. Children in the Immigrant Families Israeli born 17 6 2 N = 25

American born 2 1 1 N = 4

Time spent abroad (years) 0 2 5

Age of immigration (years) 1 2 4.5

N = 29

The families3 were observed in their homes (Americans in the United States, native Israelis, and immigrants in Israel). We have recorded three dinner-table conversations (one by video and two by audio) for each family. An observer was present at all dinners.4 Three types of analysis have been carried out on these data to date: first, transcripts were coded for the use of acts of social control (e.g., all directives) by all participants in the first 20 minutes of all dinners, yielding a corpus of 4120 control acts. A subset of the 903 control acts (cf. Ervin-Tripp & Gordon, 1986) issued by parents to children was examined for parental style of politeness in issuing directives. Second, this analysis was complemented by a study of the metapragmatic comments (cf. Becker, 1988) issued by parents to children during the same dinners (Blum-Kulka, 1990). A third analysis focused on code-switching and code-mixing in the immigrant families (Olshtain & Blum-Kulka, 1989). In the following, we shall draw on results from all three. The families were also interviewed at length in their homes; the interviews were meant to probe mainly cultural preferences with regard to issues of children's pragmatic socialization, extended in the case of the immigrant families to include questions about language learning and bilingualism. The insights gained from the interviews with the immigrants are incorporated in the present discussion, having been found to be particularly helpful in dealing with issues of metapragmatic awareness. The immigrant families studied present a case par excellence of a bilingual and bicultural social reality. They maintain in practice and in attitude close contact with two languages and two cultures, and hence are exposed to two systems of pragmatic and discoursal rules. There is, of course, individual variation among the families, some showing a higher degree of convergence (Giles, 1979) to the Israeli culture than others. Yet as illustrated by the two following extreme cases on the "convergence" continuum in our sample, even the least "American" family in the group maintains an intimate link with the language and cultural heritage of their country of origin. Consider the case of the Bells and the Darnos. The Bells represent the Anglophone extreme of our bilingual continuum. They have chosen to live in a Jerusalem neighborhood known for its density of Americans and Canadians. They claim that at least on the street they live, English is the

200

Discourse Perspectives

dominant language even for the children. And indeed, during the interview we were suprised to detect a trace of American accent in the Hebrew speech of their Israeliborn, 8-year-old daughter. At work, the husband uses Hebrew only (at a very advanced level, given his occupation), at home, English only. Working as an instructor of English, the wife speaks English for most of both her professional and private life. The family spends most summers in the United States, and all the children are balanced Hebrew-English bilinguals. While the Bells invest familial effort in preserving English, the Darnos vacillate between a wish to raise their children as bilingual speakers and the no-less-strong wish to acculturate to the Hebrew-dominant Israeli culture. They mostly, but not entirely, speak English at home and Hebrew at school and work. For the children, Hebrew is the preferred and unquestionably the better known language; their English is more advanced than that of other Israeli children their age, but it still definitely remains a second language. For the parents, mastering Hebrew to the degree that would allow them to read Hebrew literature in the original presented a challenge they faced within the family context. Thus the father worked through all his son's literature assignments at high school, a process he claimed to have proven extremely beneficial for his own acculturation.

Notions of Interactional Style: Metapragmatic Comments One way of capturing cross-cultural diversity with regard to interactional style is to observe cultural differences in actual verbal and nonverbal behavior systematically. The realization of requests in three groups of families, mentioned above (BlumKulka, 1988; 1990), provides such an example. Cultural notions of style can be further captured by what people say they do with words, namely, by their metapragmatic discourse. Metapragmatic comments (cf. Becker, 1988) made in regard to linguistic and conversational behavior can reveal the pragmatic norms underlying such behavior. In Silverstein's terms, metapragmatic comments can be seen as "non-referential indexes" (Silverstein, 1976). Just as language use can index nonreferentially social dimensions such as degree of deference, comments made in regard to the perceived violation of a conversational norm will index for the members of the particular speech community the network of interactional norms that govern language use in that community. Such comments relate to the smooth flow of discourse by explicit bids for turn, by bringing to attention breaches of turntaking rules, or by focusing on perceived violations of conversational maxims, as laid out by Grice (1975). Metapragmatic comments are one of the explicit ways in which members discuss criteria for verbal appropriateness. In family discourse, metapragmatic comments further serve instrumentally in socializing children to the pragmatic norms of the given culture (Blum-Kulka, 1990). Research in the development of communicative competence in childhood reveals that the process is very much dependent on culturally differential adult input. Caretakers across cultures differ in their preferences in amount and type of talk expected from children, and in the degree to which they deem it essential to explicitly teach language to children (Heath, 1983; Schieffelin & Ochs, 1986; Ochs, 1988;

American-Israeli Families at Dinner

201

Schieffelin, 1990). In Western societies it is rather the manner of speaking, namely politeness, that tends to be attended to explicitly. Thus American middle-class mothers systematically prompt the use of politeness formulas and sanction perceived lack of polite behavior (Gleason & Weintraub, 1976; Grief & Gleason, 1980; Gleason, Perlman, & Grief, 1984; Becker, 1988; Becker, 1990; see also Schmidt, Chapter 1). Yet "politeness" is a culturally relative notion that serves in folk etymology as a cover term for social criteria applied to the distinction between socially appropriate and inappropriate verbal and nonverbal behavior. Considered as a measure of pragmatic socialization, the analysis of metapragmatic comments can thus show parents' culturally differentiated notions with regard to the style of linguistic politeness required from children (Blum-Kulka, 1990).5 In the following we shall compare the metapragmatic discourse of both adults and children in the three groups of families studied. Several issues underlie the discussion. First, what is the degree of inter-group diversity in pragmatic norms, as revealed by the metapragmatic discourse of adults and children? We hypothesize that Americans and Israelis differ in pragmatic norms, and that immigrant families differ from both, abiding by an IL system of pragmatics. Second, if our hypothesis is confirmed, what is the degree of intra-group (child > adult) conformity within each group? Do children in the three groups follow in their parents' footsteps in types of verbal behavior they sanction? We further hypothesize that since children can be expected to be socialized to culturally determined ways of speaking, the metapragmatic discourse of the children in all groups will tend to replicate trends detected in the parent's discourse. Children in the immigrant families will be no exception, showing IL pragmatics being transmitted across generations. Furthermore, if our hypotheses are supported, how far are the trends depicted for the immigrants dependent on the language spoken? We shall claim that the IL pragmatics phenomena depicted for the immigrant families are realized independently of language choice.

Method Transcripts of dinner-table conversations were examined for the occurrence of metapragmatic comments.6 MCs related to verbal behavior were classified into three distinct categories (see also Blum-Kulka 1990 for an additional "behavioral" category): Talk-Regulation. This category includes comments made to regulate the smooth flow of turn-taking. They include bidding for turn ("can I say something?"), allocation of turns ("Okay, David, your turn"), negating a turn ("wait, Talya is first"), upholding a turn ("I'm talking now") and checking listener's attention ("are you listening?"). Maxim Violation. This category includes comments signaling perceived violation of one of the four Gricean (Grice, 1975) maxims (see also Pellegrini, 1987). Comments in regard to the maxim of relevance prompt the addressee to respond to a conversational demand ("Beth, there is a question on the floor") or delegitimize mention ("one should not say that"). Comments in regard to quality cast doubt on the truth-value of a speaker's proposition (e.g., in response to a child's reporting

202

Discourse Perspectives

having seen a "GIANT turtle," the mother inquires, "How giant is giant? Have you really seen it?"). Comments in regard to quantity set limits to degree of informativeness of stated propositions ("we heard that"), but also elicit information or just talk when felt lacking ("aren't you participating today?"). Comments in regard to manner prompt the use of politeness formulas ("say 'please'"), correct ungrammatical language, note improper forms of address and reference (CHILD:" . . . that stupid teacher" MOTHER: "Who?" CHILD: "Varda, the Math teacher"), and sanction the use of slang and vulgar language. Metalinguistic Comments. This category captures talk about language; it includes queries and responses about word meanings, as well as comments topicalizing language, including cross-linguistic comparisons ("Did you know the Eskimos have a hundred words for 'snow'?").

Findings: The Degree of Inter- and Infra-Group Diversity Americans show the highest degree of metapragmatic awareness: more than half of the comments (54%, n = 1052) belong to American speakers, the other half being divided between Israelis and immigrants. This pattern is repeated for both adults and children. Through the course of two dinner-table conversations, the average number of MCs per family member present (including non-MC speakers) is 20 for American adults, 10 for immigrant adults and 8 for Israelis, and for the children 14 (American), 7 (immigrants) and 6 (Israeli) respectively.

GROUP AND AGE EFFECTS To test the statistical significance of intra- and inter-group diversity in pragmatic norms, two independent scores were computed per speaker: (1) a talk-regulation (TRE) score, computed as the ratio of TR from maxim-violation (MV) and metalinguistic (ML) MCs; and (2) metalinguistic score (MLI), computed as the ratio of ML MCs from MV. We shall refer to these scores respectively as TRE and MLI.7 The mean scores for both categories are presented in Table 10.2. The first point to note in Table 10.2 is that MC discourse is not equally distributed among the participants. While among the American and Israeli adults we did not find any parent present who did not make at least one comment, among the immigrant adults 7 parents made no metapragmatic comments, and among the children 2 Americans, 4 Israeli and 7 children in the immigrant families did not partake in this type of discourse. Since the TRE and MLI scores reflect preferences for actual use, the following analyses are based on the subsample of participants found to engage in metapragmatic discourse. The second point to note is that except for the immigrant children's score for TRE, all other immigrant scores consistently fall between the Israeli and American means. The results show that Americans and Israelis differ significantly on both TRE and MLI measures, this trend being repeated for adults and children (for adults, 7129] = 3.64, p < .001, for children 7[30] = 2.36, p < .05). For the immigrants, TR scores for the adults differ significantly from both Israeli and American ones. Immigrant adults have significantly higher TRE scores than Israelis (T[22] = 3.81,

American-Israeli Families at Dinner

203

Table 10.2. Group) Differences in Talk-Regulation (TR) and Metalinguistic (ML) Scores for Adults and Children Group ISR Mean [N]

IVTM

AM

Mean \N]

Mean (N]

Adults Standard deviation

0.06 [16] 0.06

0.18 [9] 0.14

0.59 [16]

Children Standard deviation

0.42 [15] 0.56

0.17 [14] 0.35

1.80 [17]

Adults Standard deviation

2.70 [14] 3.90

1.50 [8] 0.99

0.49 [16]

Children Standard deviation

2.60 [13] 3.30

2.40 [11] 3.40

0.38 [17]

Category

Significance IS/AM

IM/AM

IM/IS

TRE * ***

*** **

* —

MLI *** ***

* ***

— —

/; < 0.001*. p < 0.005**. p < 0.050***.

p < .001) while they score significantly lower than Americans (T[22] = 2.12, p < .05). Immigrant adult MLI scores are higher for immigrants than for Americans (7122] = 3.81, p < .001) but similar to those of Israeli adults. On the other hand, the children in the immigrant families differ significantly from American children on both measures (for TRE, T[29] = 2.70, p < .01, for MLI, 7"[26] = 2.42, p < .05), but not from group 1 Israeli children. The effects of group membership (Israeli/immigrant/American) and Age (child/adult) on these scores were tested via a series of paired two-way analyses of variance. Our expectation for cross-cultural diversity was fully confirmed. Israeli versus American group membership has significant main effects on both MLI and TRE. For MLI, only group membership is the significant distinguishing factor: F(56,l) = 12,2, p < .005. Age had no effect, showing adults and children following similar patterns, as predicted. For TRE, both group membership and age matter (for group, F(56,l) = \5A,p< .005, for age, F(56,l) = 10.87, p< .05). There was no interaction effect. Against this background of cross-cultural diversity, the immigrants are shown to occupy a peculiar interim position, as predicted. Though they speak mostly English, Group membership sets them apart from Americans on both MLI and TRE. For Metalinguistic discourse, the main effect for group is F(48,l) = 11.1, p < .005. This pattern being similar for both adults and children, we found no effect for age. On the measure of talk-regulation both group membership and age are important: for group F(48,l) = 7.3,p< .005), for ageF(48,l) = 4.5,p< .010. There was no interaction effect. A different picture emerges in comparing the immigrants to the Israelis. For both, we find a similarly high degree of metalinguistic awareness for both adults

204

Discourse Perspectives

and children. Hence it is not suprising that we found no effect for Group (nor Age) on the MLI measure. On the measure of talk-regulation, we found no effect for group but a nearly significant effect for age, (F[42,1 = 3,7, p = .06). In interpreting this result it should be borne in mind that immigrant adults (but not children) score significantly higher on TRE than Israeli adults (Table 10.2). These findings confirm our earlier hypothesis that Israelis and Americans differ significantly on both talk-regulation and metalinguistic measures. Against this background of cross-cultural diversity in pragmatic norms, the immigrants seem to waver between the two. We expected to find them to differ from both, thereby indicating an IL system of pragmatics, and this expectation was confirmed for the adults. Immigrant adults differ significantly from both Israeli and American adults in their talk-regulation discourse (T-tests), but the difference is not big enough to show a Group effect (analysis of variance). On the measure of metalinguistic MCs, cross-cultural diversity is confirmed again, showing Israelis and Americans to have different attitudes to the domain of language as topic. In this regard, the immigrants differ from the Americans, resembling the Israelis in their degree of metalinguistic awareness. As will be elaborated, there are different, group-specific motivations for this phenomenon. Our second hypothesis is partially confirmed: children in all three groups conform to adult patterns in regard to metalinguistic discourse. But they differ from adults in attitudes to talk-regulation; this could mean that for this aspect of discourse children either have different notions, or different needs. These possibilities will be explored below. The children in the immigrant families do not differ in a statistically significant way from other Israeli children in regard to talk-regulation, as do their parents. Yet qualitative analysis of these children's talk-regulation behavior, to be discussed below, does suggest a unique pattern, in line with the "IL replication" hypothesis. Finally, in considering the metapragmatic discourse of the immigrants, it is important to note that the overall majority of MCs (83%) are uttered in English, not Hebrew. Only 10% of comments are made in Hebrew, with the remaining 7% made in a mixture of English and Hebrew, manifesting the language choice patterns typical of this group in general (Olshtain & Blum-Kulka, 1989). Hence the fact that immigrant adults are found to differ from Americans in their metapragmatic discourse means that they activate in their first language pragmatic norms adapted from the second. But native language pragmatic norms seem at work too; in certain respects, immigrant adults differ from Israelis as well. The metapragmatic discourse of the bilingual families hence is seen to exhibit the first assumption underlying all interlanguage studies: it is systematically different from both LI and L2.

ORDER OF PREFERENCES The "in between" character of the immigrant families' IL pragmatics system is further highlighted by considering the three groups' order of preference for domains of discourse attended to. We have analyzed the distribution of the three categories of MCs, uttered at the dinner table by all members of the family in the three groups. For this analysis MCs have been seperated by speaker's role (adult or

American-Israeli Families at Dinner

205

Fig. 10.1. Adults' metapragmatic comments in three groups. child) and divided by type of comment (discourse management/maxim-violation/ metalinguistic). Adults. The marked difference between American and Israeli adults stems from the culturally differing emphasis on conversational behavior as compared to language as topic. In the American adults' discourse, the two conversational behavior categories (maxim-violation and talk-regulation) take up 82% of all metapragmatic discourse space; in the Israeli discourse, they constitute less than half of all comments made (38%) (Fig. 10.1). The relative importance attached to conversational norms (54%) versus language (45.5%) shows the Immigrants occupying an interim position between the other two groups: this aspect of discourse is more important for them than for Israelis, but less important than for Americans.

206

Discourse Perspectives

Fig. 10.2. Children's metapragmatic comments in three groups.

Children. American children share with their parents a very high degree of metapragmatic awareness in regard to conversational behavior (talk-regulation plus maxim-violation constitutes 78%), especially when it concerns discourse management (48.4%), (Fig. 10.2). Israeli children, on the other hand, share with their parents a relatively high degree of metalinguistic awareness (51%). The children in the immigrant families reveal a pattern of their own. Similar to Israelis, they find language important (43%); similar to Americans, they attend to maxim-violation (31%). Consider the decreasing emphasis put on conversational management by children in the three groups: Americans 80%, immigrants 52%, Israelis 49%. Children Compared to Adults. Are children sensitive to the same aspects of discourse as their parents? Are their pragmatic norms similar to those of the adults? Tannen (1981) found that second-generation Greek-Americans who spoke no

American-Israeli Families at Dinner

207

Greek, nevertheless manifested in their conversations in English levels of indirectness typifying Greek ways of speaking. In a similar vein, we explored the possibility that the metapragmatic discourse of children born to immigrant parents will show traces of unique parental input, setting these children apart from other nativeborn Israelis. Comparing the children's preferences to those of the adults allows us track the success of pragmatic socialization: children following in their parents' footsteps in aspects of the discourse attended to may be said to have internalized their parents' discoursal criteria. Such replication is manifest to some degree in all groups: as shown by the analysis of variance across the groups, age had a significant effect only on talk-regulation scores; the lack of such an effect on metalinguistic scores at least partially supports the replication hypothesis. Comparing adult/child patterns of distribution lends further support to partial replication: For the Americans, the children follow in their parents' footsteps in deeming conversational norms more important than language (see Figures 10.1 and 10.2). A different trend is depicted for the Israeli children, who vacillate between the domains more than their parents (but note small number of comments). Interestingly, replication seems most salient for the Immigrants: in this group, the order of preferences in matters of pragmatics attended to among the children is the exact mirror image of the order exhibited by the parents. Metalinguistic comments constitute almost half of all MCs for both (45.5% for adults, 42.6% for children), followed by maxim-violation comments (adults 33.3.%, children 38.2%) and talk-regulation MCs (adults 21.3%, children 19.2%). But in their relatively high metalinguistic awareness these children also resemble Israeli group 1 children. Hence it may the case that their metapragmatic discourse is affected by peer-group interaction rather than by parental input. The next sections discuss the unique characteristics of this IL metapragmatic discourse.

Characteristics of Bilingual Metapragmatic Discourse Talk-Regulation The adults in the bilingual families pay attention to all three dimensions examined: discourse management, language, and conversational norms. In regard to discourse management, the model followed is the mainstream, middle-class American one: "fair" turn-allocation, discouraging interruptions, and generally orderly systematic turn-taking seem the discourse corollary of American ideals of individual rights and equal opportunity for all. American individualism, with all that it entails for mutual respect for individual space, has been long-remarked upon by outside observers, such as de Tocqueville, and studied extensively, more recently, from both sociological (Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985) and sociopsychological points of view (Brown, 1991). Cross-cultural research has amply shown that the discoursal manifestation of such ideals, namely, the systematics of turn-taking (Sacks, Schegloff, and Jefferson, 1974) are by no means universal (Reisman, 1974; Philips, 1983). Regular turn-taking is not necessarily adhered to even in all domains

208

Discourse Perspectives

of American conversations (Edelsky, 1981). But as the Jewish-American adults' metapragmatic discourse shows, the rules seems to have an unquestionable normative status in American society. In this respect, Jewish-Americans seem to approximate mainstream American ways of speaking; by paying attention to this particular aspect of discourse, the immigrant adults reveal a sensitivity going back to their own language socialization. It is this positively evaluated feature of American interaction that they are trying to preserve in their children's talk. It is important to note that by its very nature, this feature is not bound to any specific language; potentially it is equally applicable to conversations in Hebrew or English. The ethnographic interviews with the families occasioned clear formulations of the adults' prevalent attitude toward talk-regulation, specifically turn-taking. Asked whether they discourage interruptions by children, we received from one of the families the following reply: FATHER: 8 It's very important that they don't interrupt others. It's a principle here. In fact, sometimes we say to the other, "Wait, your brother is still talking," or "Jessica is still talking." MOTHER: And sometimes it ends up where they have to raise their hands, even (laughter). FATHER: Oh, yeah.

The findings also show Israeli and American children to differ markedly in degree of attention paid to talk-regulation, with the Israeli bilingual children in the immigrant families paying the least attention to this aspect of discourse. This seems to indicate a failure on the part of the immigrant parents to transmit their basically American talk-regulation norms to their children. But the analysis of talk-regulation comments by type,9 suggests otherwise. Of specific interest are the ways in which bids for turn are accomplished. American children make explicit bids for turns: 47.4% (n = 116) of their talkregulation comments are permission requests for a turn, often combined by a comment indicating awareness of the rule, such as "Daddy, can I say something? Is it my turn?" (Blum-Kulka, 1990). Israeli children bid less for turns (28% of talkregulation MCs, n = 25) and when they do, instead of asking permission to talk, they try to secure their interlocutors' attention by uttering a clearly addressed preexchange move, preparing the ground for the ensuing talk (Edmondson, 1981): "Daddy, do you know what happened today?" The children in the immigrant families seem to vacillate between the two techniques. Bids for turn constitute 78.9% of their talk-regulation MCs (n = 19), showing them particularly concerned with obtaining permission to talk. But they avoid the American style of explicit turntaking metatalk. Instead, we find attempts to secure attention that look like indirectly performed permission-requests: NAVA (age 5): Ima (Mother), now can I tell you what they do? (Talk round continues.) NAVA: Should 1 tell you what they do, what the GANANOT (nursery-school teachers) do? Should I tell you what they do?

Subsequently Nava's attempts act as a pre-exchange move, similar to the Israeli child's attempts. Yet semantically she is closer to the American child, using "tell" as

American-Israeli Families at Dinner

209

a "linguistic action-verbial" (cf. Verschueren, 1985). Hence the pattern exhibited is an interim, interlanguage-type one, showing influence from both contact cultures.

Maxim-Violation Comments Maxim-violation comments can relate to the perceived violation of any one of the four Gricean maxims: relevance, quality, quantity, and manner. The distribution of maxim-violation comments by type (Figures 10.3 and 10.4) shows American and Israeli adults to differ markedly in preferences for the type of maxim violation most likely to be noticed verbally. The immigrant adults seem to differ more in this case from the Americans than from the Israelis. Americans emphasize foremost adherence to the maxim of quality, namely, factuality of the proposition asserted (60.2%). Both Israelis and immigrants distribute their attention more equally between all four aspects of discourse. For Israelis, manner (mainly politeness) is highly attended to (40%), followed by quality (33.7%) and relevance. Degree of informativeness and volubility (quantity) are rarely spoken about (6%). The immigrant adults fall between the two other groups in their order of preference. Quality is very important (45.8%), followed by attention to violations of manner (31.4%) and relevance (20.4%). Attention to quantity of talk is negligable (2.1%). The relative salience of comments in regard to quality in all three groups deserves special attention. Prompting adherence to the factuality of stated propositions is by no means a universal norm of pragmatic socialization. Heath (1983) found an emphasis on this aspect both among the white working-class people of Roadville and the middle-class "townspeople," but not in the black community of Trackton. In Roadville, children are expected to "stick to the truth" in retelling nonfictive stories of shared events that the adults are aware of; middle-class parents also expect children to share information only the children know, but as in Roadville, probe their children to recognize contradictions in such factual accounts. Insistence on factuality is interpreted by Heath as manifesting literate traditions. On the other hand, in the black working-class community of Trackton, oral traditions dominate; hence adherence to factuality is suspended in favor of "storyness": the purpose of narratives above all is to entertain and "to establish the story-teller's intimate knowledge of truths about life larger than the factual details of real events" (Heath, 1983, 188). True to their middle-class literate background, all families in our group require children to learn to distinguish fact from fiction. In the following example, the child's report is challenged by both parents, and found counterfactual. ILANA (age 13.5): NAOMI (age 5): FATHER: MOTHER: NAOMI: MOTHER: NAOMI: MOTHER:

Did you have a good time at Leora's? I wasn't at Leora's. Yes you were, this afternoon= =You weren't? Where did you go from CAN (nursery school)? To a home. Not to this home. Yes, to this home. I wasn't here.

210

Discourse Perspectives

Fig. 10.3. Types of maxim-violation comments: adults.

The immigrant families seem to abide in the respect by the middle-class norm of Heath's "townspeople." Children are invited to tell A-events (cf. Labov & Fanshel, 1977) that only they know, but must beware of contradictions. In American families adherence to factuality supersedes all other maxims. The immigrant families, as well as the Israelis, find other aspects of discourse also worthy of attention. Thus parents also comment on the relevance of a child's contribution (MOTHER: "I don't want to hear it, Naomi, I don't want to hear it now"), and on its manner of delivery (FATHER: " . . . but don't shout"). A very similar picture of cross-cultural diversity emerges when we compare American and Israeli children. American children, like their parents, are foremost concerned with quality (65.6%); Israeli children, like their parents, when they do comment (n equals only 19) stress manner (45%). But what about the children in the

American-Israeli Families at Dinner

211

Fig. 10.4. Types of maxim-violation comments: children.

immigrant families? They share with other Israeli children the same types of input in Hebrew from peer interaction, school, and the media. If they are found to differ from other Israeli children, it would mean that the 11 pragmatic norms of the parents are transmitted through bilingualism to the next generation, signaling the importance of parents' input as a dominant source of pragmatic socialization. And indeed, when it comes to Gricean maxims, the children in the immigrant group differ from the group 1 Israeli children. Differing from the Israelis, who attend primarily to manner, the bilingual children stress in their comments the importance of factuality even more than their parents (72.7% compared to 45.8% in parents' discourse), and resemble American children in having this aspect of discourse at the top of their list of preferences for maxim-violation MC comments.

212

Discourse Perspectives

Comments related to quality are made mostly among siblings, challenging the truth-value of a statement by "lo yaxol lihyot" (can't be) or simply "that's not true": DAVID (age 7.5): NAOMI (age 12): DAVID: NAOMI:

She plays XAMES AVANIM (marbles). We don't play. We play after you go to sleep. No, you play before I go to (sleep). (That's) not true. Last time we didn't.

Comments made to parents are phrased more indirectly: DANA (age 14): You know you're exaggerating. FATHER: I'am exaggerating? What?

Comments on relevance in both the Israeli and the immigrant families mostly take the form of restricting undesirable topics, but are less frequent in the immigrant than in the Israeli data. Thus for example, Naomi, age 12, interrupts her brother's (age 7.5) dramatic account of a thriller with: NAOMI: Don't say it because maybe they want to see it.

Similar to American children, this group pays little attention to manner. Israeli children, on the other hand, find manner very important. Typical examples in the Israeli data are comments to siblings sanctioning "bad" words, such as a 6-year-old telling her twin sister not to say "mag'il" (disgusting) about food, or a 10-year-old reacting to her 12-year-old sister's comment "she was a shitty teacher" by "Shlomit, you don't say words like that here."

Metalinguistic Comments In their type of metalinguistic awareness the bilingual families manifest a pattern of their own. Given their active bilingualism and the adults' recent history of second language learning, it seemed reasonable to expect from these speakers a relatively high level of explicit attention to matters of language. And indeed, the bilingual families exceed the Americans by far in their degree of language awareness though they fall behind the Israeli norm. For both immigrant adults and their children, metalinguistic comments constitute almost half of all metapragmatic talk, while for Americans such comments do not exceed a fifth of metapragmatic discourse. For Israelis, matters of language figure as the most prominent aspect of discourse commented on, for both adults and children. (Figures 10.1 and 10.2). The high level of attention to matters of language in the Israeli families confounds the issue of linguistic awareness as a hallmark of bilingualism. This is due to the peculiar history of the revival of Hebrew, which provides the sociolinguistic context for the Israeli preoccupation with matters of "normative" versus "acceptable" language use (Rabin, 1976). Within the Israeli families, speakers are very concerned with matters of grammaticality (i.e., how to form the plural of kalba (bitch): "Did you know one should say [klavot] and not [kalbot]?"), but also with word-meaning (in response to children's querying the meaning of loan-words (dilemma) or just rare ones, word origins ("where does x comes from"), pronounciation (how Ixl & Irl are pronounced in different registers), and the acceptibility of

American-Israeli Families at Dinner

213

slang. On the other hand, American speakers, when occasionally they do topicalize language-related issues, make general comments in regard to the language of others, not their own ("Did you know the Eskimos have a hundred words for 'snow'?"). A close look at the metalinguistic comments made in the bilingual families reveals several unique features. First, though clarifying word meanings to children is a favorite pastime, in these families it is carried out in a multilingual fashion. Both Hebrew and English words may be explained in the language the query was made in ("what's garlic? It's a spice"), or explained and translated, with the degree of translational equivalence between items in a given semantic field becoming the topic of lengthy discussions. For example, a 13-year-old's test question (Heath, 1983) to her 5-year-old sister "What is OGER (hamster) Naomi, do you know?" sets in motion a discussion of the biological differences between a "hamster" and a "guinea pig," combined with the linguistic quest for the Hebrew names of these animals (Olshtain & Blum-Kulka, 1989). Word-meaning queries in these families serve differential functions for adults and children. Children's queries concern the meaning of English words, or ask for equivalents between the two languages. RIVKA (age 12): What are physicists? [FISIKAIM]? MOTHER: Yes.

The mother confirms thatfisikaim, a phonological loan translation, is indeed the equivalent of "physicists." RIVKA (age 12): How do you say SOTER, ABA (Father)? FATHER: Contradictory.

The adults, on the other hand, rely on their children's Hebrew expertise for clarifying the meaning of Hebrew words: FATHER: What's KRAZA? DORON (age J5): Poster.

We also find an explicit analytical concern (in Bialystok's sense of "analysis"; Bialystok, Chapter 2) with issues of bilingualism, (including code-switching and mixing), both on the sociopsychological and educational level. Though our presence at the dinners no doubt affected the prominence given to language-related issues in the talk agenda, we know from subsequent interviews with the families that the concerns expressed are very real. A major issue is language choice. All the families interviewed expressed a clear English-maintainance language policy; for all, speaking English at home is a matter of deliberate choice. The parents report having spoken only English with the children from the time of birth, with the children hearing Hebrew for the first time sometimes only at the nursery school, as late as the age of three. Yet through schooling, Hebrew became the children's dominant language, while English remains that of the parents. At the time of the interviews, English is estimated as being used in the home about 80% of the time, an estimate confirmed by examination of the dinner-table conversations in these families (Olshtain & Blum-Kulka, 1989).

214

Discourse Perspectives

Another issue is the degree of mixing and switching, and attitudes toward both. Both practices are highly common during the dinner-table conversations but, differing from Gumperz's findings that show bilingual speakers as unaware of the language they are speaking (Gumperz, 1982), we found indications of our bilingual speakers being aware of actually using a hybrid language variety, one that draws on both languages. This variety is alternatively referred to as "Hibbish" or "Hebrish" as exemplified in the following exchanges: ILAN (age 8): (to brother) You're talking gibbish. FATHER: Hibbish. ILAN: Oh, yeah.

And a few minutes later, Idan is commenting on his mother's code-switching from English to Hebrew: ILAN: Those two words, IMA (Mother), were in Hebrish. MOTHER: You're right. I changed in the middle.

The language game at work is one of naming; by trying to find an appropriate label for their own mixed dialect, these families indicate their high degree of linguistic self-awareness. Metalinguistic awareness also allows for humor: FATHER: Nisani, show Marit [the observer] how you speak Hebrew with an American accent. NISAN (age 6): LO (no). OBSERVER: Yes, please. NISAN: What? FATHER: Say something like when you imitate Mommy and me like ANI RAJITI (I saw). NISAN: ABA AMAR PAAM [HATARNEJOL] (Daddy once said [hatarnejol][replacing /j/ for /g/ in pronouncing the Hebrew word for "rooster."])

But the children's high proficiency in Hebrew can also lend itself to a power game. After being reassured by his parents that he speakes both English and Hebrew well, a child criticizes his mother's Hebrew. Sister and father alike find the comment face-threatening, to the extent that the father feels it necessary to come to the mother's defense: ILAN (age 8): ANI SAMTI LEV, IMA, ET HATAUYOT SHELAX SHE SHEAT KORET LI MIXTAVIM VEHAKOL, AT KORET LI 1M TAUYOT, BLI MI VTA IVRIT. (I noticed, Mom, your mistakes, when you read me letters and everything, you read with mistakes, without a Hebrew accent.) MIRA (age 6.5): It's not nice to say that. MOTHER: I really am pretty bad. FATHER: But actually that's not true, Ilan, there are some words that you know, but I think there are some things that IMA (Mother) can express more clearly, even in Hebrew still. ILAN: I know ABA (Father), except I am better in reading and talking in Hebrew.

American-Israeli Families at Dinner

215

And finally, awareness of code-switching and mixing is combined with ambivalent feelings toward the benefits of such practices: INTERVIEWER: MA ATEM MEDABRIM BABAJIT BEDEREX KLAL? ANGLIT? (What do you normally speak at home? English?) FATHER: We try to speak English. JESSICA (age 9): We talk part English and part Hebrew. AN1/ANI KMO MEDABERET NAGID BEEMTSA HAMISHPAT AN1 OSA/OVERET MEANGLIT LEIVRIT UMEIVRIT LEANGLIT. (I like, say in middle of the sentence, I go from English to Hebrew and from Hebrew to English.) FATHER: As a matter of fact, I have been puzzling over it for the last year. I used to think it was cute when they mixed the sentences together.

MOTHER: I still think it's cute. I do it. FATHER: I don't. I think it's bad for their Hebrew and their English, so I fuss at them for the last year now, speak either English or Hebrew but not the two of them together.

Degree of Language Dependency

It is especially noteworthy that MCs are indices of the immigrants' perception of good discourse regardless of the language they are uttered in, or comment on. This can be seen by the tacit agreement of participants to ignore the language they speak completely if the issue at hand concerns pragmatically proper behavior. In the following case a violation of the maxim of manner is at issue: SHARON (age 6): Mommy, TAVI'I LI OD (bring me more) fish. MOTHER: Excuse me. SHARON: BEVAKASHA, TNI LI OD (please, give me more) fish.

That the child is mixing English and Hebrew singles out this example as typical of the bilingual family. In sequence, the exchange is paralleled by several examples in both the American and Israeli families, all parents being concerned with socializing children to use politeness markers. But the mixing is irrelevant to the main issue discussed: after first using a colloquial, slangy expression in Hebrew (literally, "bring me" instead of "get me" or "give me") the child interprets her mother's reaction as a reprimand for her lack of politeness, though it might have also been interpreted as a negative reaction to mixing the two languages. Next, Sharon adds the required "magic word" (bevakasha/please) and shifts register ("tni li" instead of "tavi'i li"). But the "fish" remains "fish," and the exchange double-coded, the mother never saying a word in Hebrew. Pragmatic socialization is hence taking place independently of the language used for its implementation.

Degree of Metapragmatic Awareness We have seen that the parents in the bilingual families try to socialize their children according to a normative system of pragmatic rules shaped after the American example. By attending explicitly to matters of discourse they highlight the salience

216

Discourse Perspectives

of this domain, helping develop their children's overall communicative competence, regardless of the language used. The comments are made in English, but are equally valid for talk in both languages. We also have evidence that in actual use, both adults and children in these families abide by an IL pragmatics system, which conforms to neither of the two languages involved (Blum-Kulka, 1988). In issuing directives, bilingual parents' choice of strategies differs significantly from preferences shown by both Israeli and American parents, the same being true for children in these families. How far are the members of this group aware (in Schmidt's sense of the word, Schmidt, Chapter 1) of any incongruence between the two pragmatic systems represented by the languages they speak? How far are they aware of their own IL pragmatics? These questions will be explored below in the wider context of the issue of metapragmatic awareness in both first and second language (for a different perspective see also Schmidt, Chapter 1; Bialystok, Chapter 2). We shall draw here mainly on monolingual and bilingual speakers' explicit metacommunicative responses to questions on language use during the ethnographic interviews. The term metapragmatics seems to imply the capacity of speakers to formulate explicit rules of speaking. But it is a matter of debate, whether the capacity to "do" reliable metapragmatics is reserved to professional pragmaticists or, indeed, whether it can be practiced by all. For example, Schmidt (Chapter 1) argues against Wolfson's (1989) strong position with regard to the fallibility of native speakers' intuitions about language use. Our data support his argument that both native and nonnative speakers do have conscious access to such rules, though there are limitations to this type, of reflexivity. With regard to native speakers the argument is far from new. Cultural anthropologists have long argued for the need to understand cultures emicly from the "native's point of view" (Geertz, 1976), incorporating as valid data cultural members' "experience near" (cf. Geertz, 1973) accounts of various dimensions of social realities, including language use. Participants' understandings are a crucial source of knowledge since "all social research, and indeed all social life, is founded on participant-observation" (Hammersley & Atkinson, 1983, 235). A case in point are anthropological studies focused on terms that name a particular way of speaking, are part of a speech community's metapragmatic vocabulary, indicating a heightened cultural attention and articulation with respect to a specific sociocultural domain (Katriel, 1986). The study of such cultural notions proceeds with the help of cultural members explicating the meaning clusters associated with such terms (Spradley, 1979), as exemplified by Katriel's study of dugri (straight) talk in Israeli speech (Katriel, 1986). Metapragmatics is also part of the discourse of language socialization, though it is not clear at this stage how much of the acquisition of pragmatics is dependent on explicit attention drawn by caregivers to this domain. But as demonstrated above, parental discourse addressed to children is rich with metapragmatic comments with respect to discourse management, conversational norms, and the use of politeness formulae. Even more than native speakers vary in the degree of their knowledge of pragmatic rules, they also seem to vary in their capacity to formulate such rules explicitly. Yet if asked, native speakers show high levels of agreement in judging the

American-Israeli Families at Dinner

217

appropriateness of a specific request variant in a given situation (Olshtain & BlumKulka, 1985), or in agreeing on producing culturally determined request-variants for the same situation (Blum-Kulka & House, 1989). Such results confirm empirically the cross-culturally varied systematicity of pragmatic systems and the degree to which the knowledge of such rules is implicitly there for native speakers to draw on. Sociolinguists would be out of business if the whole intricate network of relationships governing such choices were available consciously to every native speaker; yet parts of this knowledge are retrievable through questioning. In the case of interviews tapping metapragmatic knowledge, the questions asked cast interviewees in the role of participant-observers, invited to reflect on their own language experience. Hence the social situation created is what Glaser & Strauss (1964) call an awareness context. Focus on the issues topicalized may thus transform tacit knowledge into explicit formulations. In the course of our ethnographic interviews with the families in the project, we asked them both to rate the degree of importance attached to socializing children to various aspects of "appropriate" language use, and to comment on perceived differences between their own and native Israelis' ways of speaking. It is from responses to the first type of question that we learned about the bilingual parents' conscious efforts to socialize their children to systematic turn-taking norms, as discussed above. The second type of question followed from discussions of the family's discourse practices and attitudes in a certain domain. Thus, for example, we asked the families to tell us about their naming practices and attitudes to nicknaming (Blum-Kulka & Katriel, 1991), and encouraged their attempts to compare their own practices and attitudes with those of both American and Israeli. The following extract illustrates well such a comparative, reflexive effort: MOTHER: 1 grew up in the South where they say "Yes, Ma'am" for a woman. INTERVIEWER: Any woman? MOTHER: Any woman other than me, anybody other than me—my mother's friends, to my mother "Yes, Ma'am," and "Yes, Sir" to a man. FATHER: We had those things too. MOTHER: And to this day. Now I don't expect anybody to say "Yes, Ma'am" to me. I don't even want people to call me Mrs. Joss, I like to to be called Lydia—by everybody. Even Jasson's friends, Jessica's friends, everybody calls me Lydia. In the States, if I were to call my mother's friend by her first name, why, there isn't anything like that! So here I (like being called) INTERVIEWER: (Still today?) MOTHER: You know that's funny. When I go back, I still call my mother's friends Mrs. Such-and-such. I think. FATHER: Here it is different. MOTHER: =Anyway, here I don't mind being called by my first name by my children's friends because it's that way here.

Lydia draws a clear distinction between past and present, "here" and "there." While shunning the habits of her childhood, she accepts the difference in terms of address between Israel and the United States as given, surprising herself (i.e., raising tacit knowledge to the level of consciousness) in realizing that she adapts to

218

Discourse Perspectives

this difference in norms ("you know that's funny . . ."). This metapragmatic awareness to cross-cultural (and in this case also regional) variability in ways of speaking is presented in an emotionally favorable light: she "likes" or at least "does not mind" being addressed by her first name even by status inferiors (children) because "it's that way here." In Schumann's terms, Lydia is describing her process of acculturation, motivated by social and affective factors (Schumann, 1978). Other immigrant speakers go as far as viewing this process of acculturation as a transformation of personality: FATHER: Ofl, ofi hishtana, lo rak ha-safa (it's your character that changes, not only the language)

Yet there are limits to self-awareness and its accuracy on matters linguistic and pragmatic. In actual use, we found no difference between Israeli and immigrant adults in the relative salience of comments eliciting politeness formulas from children. But through the interviews, adults in the immigrant families claimed they do insist on such formulas, with such insistence occasionally cast in half-apologetic terms, as a conscious cultural transfer strategy: FATHER: A request is a request and it should have a "please" joined to it. MOTHER: It's also still important to me, the "pleases" and "thank yous."

Another area where the bilinguals' awareness becomes inexact is matters involving the pragmalinguistics of Israeli Hebrew. Talking about the lack of politeness in public places (a complaint voiced often by Israelis; see Blum-Kulka, 1992) an argument ensues between a couple stemming from a misunderstanding of a pragmalinguistic phenomenon: MOTHER: . . . and "thank you" and "excuse me," SLIXA, which people don't know in this country. FATHER: They do, they do. MOTHER: When they step on your foot, or hit you with an elbow when getting on the bus. FATHER: Well, I mean SLIXA in the sense of "move out of my way."

Rachel (Mother) is criticizing the Israeli way of speaking for its perceived lack of IFIDS (Elocutionary Force Indicating Devices) for expressive speech acts: no marking of gratitude, no expression of apology. But regardless of the degree of accuracy of her observation, it is flawed by her mention of both slixa and "excuse me": first, as translation equivalents, and second, as apology IFIDS. Literally slixa means "pardon"; but its use in apologies is quite limited, competing with several other lexical items serving this function, such as ani mictaer (I'm sorry) or ani mitnacel (I apologize), which are much more frequent (unpublished CCSARP data). Slixa is frequently used as an attention-getter, which in certain situations preempts forthcoming violations that may need an apology and hence can be seen as a doublefunction formula. "Excuse me" in English serves the same double function, while "sorry" is the unmarked apology IFID (see also Bergman & Kasper, Chapter 4). Rachel is misunderstood by her husband because of this double inaccurracy; he insists on the use of slixa only as an attention-getter (or preemptive formula) and

American-Israeli Families at Dinner

219

reserves judgment as to the general politeness issue, while she is talking about speech acts, namely the perceived Israeli absence of polite expressives ("when they step on your foot"). This example shows the potential fallibility of metapragmatic observations for a first as well as for a second language. Finally, we come to the most problematic issue of awareness: Are the adults in these families aware that their process of acculturation has affected the pragmatic system of their native English? In other words, that both in the way they realize speech acts and in their metapragmatic discourse in the course of natural conversation they are instantiating an interlanguage pragmatic system? The answer is no. As illustrated by the example of the use of politeness formulas above, the norms claimed are not necessarily the norms enacted. On the contrary: metatalk about pragmatic socialization seems motivated by the underlying assumption that by maintaining English at the home you also maintain American ways of speaking. In no part of the interview did we find any indication for a self-awareness related to the pragmatic aspects of the style used. Self-awareness seems to stop short at the borderline between linguistics and pragmatics: while both dinner-table conversations and interviews are particularly rich with metareferences to linguistic issues of language choice and code-switching, they are remarkably devoid of reference to the pragmatic issues of how either of the languages used differs in style from native American or native Israeli usage. Conclusion We have seen that immigrant, bilingual speakers manifest through their metapragmatic discourse adherence to an underlying pragmatic system that differs in important ways from the systems of both their contact cultures. This unique pragmatic system is revealed in both first and second language use, regardless of language choice. Furthermore, we have noted in the immigrant families' discourse a high degree of awareness of both bilingual usage and cross-cultural differences between the two relevant contact cultures. Yet there are limitations to this awareness: it is fallible on contrastive pragmalinguistic matters and lacking in sensitivity with regard to the interlanguage phenomena manifested in the first language. There are two general implications to these findings. First, they suggest that in the case of bilinguals maintaining their first language under conditions of intensive sociocultural contact with a second language, the pragmatics of the first language may be the primary domain to be affected by this contact. Ironically, while pragmatic competence is the most difficult aspect of language to master in learning a second language, it seems also to be, under certain conditions of bilingualism, to be the easiest to lose in the first language. Second, the specific case of bilinguals studied suggests that interlanguage is not necessarily a second-language-use phenomenon. The discourse style realized by the bilingual families, mostly in their first language, shows that pragmatic systems can be bi-directionally transferable, both from first to second language as well as from second to first. The resulting hybrid style of the bilinguals constitutes a prime example of interlanguage pragmatics.

220

Discourse Perspectives

Notes This research was funded by Grant No. 87-00167/1 from the United States Israeli Binational Science Foundation (BSF), Jerusalem, Israel to Shoshana Blum-Kulka and Catherine Snow. 1. Transfer of pragmatic rules from the second to the first language by immigrants is reported by Clyne (1982, 105) who found for example that German immigrants to Australia adopt the use of "Ja, danke" (thank you) as a positive reply to an offer, as called for by the pragmatic rule in English, and in contrast to German, in which "Danke" is a negative reply. More recently, Tao & Thompson (1991) found that Mandarin speakers, for whom American English has become their dominant language, when conversing in Mandarin make extensive use of American English backchannel strategies atypical of Mandarin speakers. 2. We use "dominant" here in the sense that it is the language the children use most, get their schooling in, and know better. As can be seen in Table 10.1, it is not necessarily the first language. Some of the children spoke only English at home, up to the age of 3 or 4, when they began to attend nursery school. 3. The project had two stages: during stage 1 (1985-1988) we collected data from 34 families. The analysis of control acts was based on data derived from transcripts of three meals from this sample (Blum-Kulka, 1988). For the second stage, (1989-1992), we selected the 24 best-matched families across the three groups (8 per group). The study of metapragmatic discourse, reported here, is derived from transcripts of these 24 families, two meals for each. 4. In reporting findings from the project at conferences we are often challenged on this point: but what about the presence of the observer? Didn't it affect the proceedings? The answer is that of course it did. As it does in all forms and types of participant-observation, whether sociolinguistic or ethnographic in outlook. Technically the inter-group comparisons we are doing are still valid since this condition is shared by all families. But beyond that, we can even say that the presence of the observer actually strengthens the validity of our findings, since if the families are found to differ under these conditions it means that their cultural styles transpire even under self-imposed constraints of self-presentation to an "outsider." And indeed, in the final account, it is their nature as cultural texts of familial selfrepresentation that makes these dinner conversations so fascinating to study from a crosscultural perspective. 5. For an analysis of parental MCs in the context of pragmatic socialization see BlumKulka (1990). 6. Since duration of meals varied by 10 to 20 minutes per meal, length was normalized to 1000 transcript lines for each. 7. We are grateful to Dan Davies for suggesting this analysis and to Kobi Yelenik for the programing. 8. Transcription conventions follow CHILDES (McWhinney & Snow, 1985). Indices used here are: [ . . . ] = trailing off, [-] unfinished word; [\] = interruption; [=] = latching; [#] = pause; { ) = overlap. For convenience (differing from CHILDES) we have reinserted capitals when needed, and used CAPS to denote code-switching from English to Hebrew. 9. The category of types of talk-regulation (TR) includes the following comments related to turns (T): bids for T, upholding T, arguing T, allocating T, negating T, prefacing T and checking for attention. Half of American children's comments fall into the category of "bids for T," with the rest distributed among the other categories. Israeli children's comments are distributed among all categories, with no clear preference for any specific type. Children in

American-Israeli Families at Dinner

221

the immigrant families show a high preference (83.3%) for bidding for turns, almost totally avoiding all other TRs.

References Basso, K. (1979). Portraits of "the whiteman": Linguistic play and cultural symbols among the Western Apache. Cambridge: Cambridge University Press. Becker, J. A. (1988). The success of parent's indirect techniques for teaching preschoolers pragmatic skills. First Language, 8, 173-81. Becker, J. A. (1990). Processes in the acquisition of pragmatic competence. In G. ContiRamsden & C. Snow (Eds.), Children's language (Vol. 7, 7-24). Hillsdale, NJ: Lawrence Erlbaum. Bellah, R. H., Madsen, R., Sullivan, W. M., Swidler, A., &Tipton, S. M. (1985). Habits of the heart. New York: Harper and Row. Blum-Kulka, S. (1988). Cross-cultural differences in interactional styles and the acquisition of communicative competence. Final Report to the BSF. Institute of Communication, Hebrew University, Jerusalem. Blum-Kulka, S. (1990). You don't touch lettuce with your fingers: Parental politeness in family discourse. Journal of Pragmatics, 14 (2), 259-89. Blum-Kulka, S. (1991). Interlanguage pragmatics: The case of requests. In R. Phillipsen, E. Kellerman, L. Selinker, M. Sharwood-Smith, & M. Swain (Eds.), Foreign language pedagogy research: A commemorative volume for Claus Faerch (255—72). Clevedon and Philadelphia: Multilingual Matters. Blum-Kulka, S., & Snow, C. (1992). Developing autonomy for tellers, tales and telling in family narrative-events. Journal of Narrative and Life History, 2(3):187-217. Blum-Kulka, S. (1992). The metapragmatics of politeness in Israeli society. Watts, S. Ide & K. Ehlich (Eds.), Politeness in Language: Studies in Its His Theory and Practice. Berlin: Mouton de Gruyter. Blum-Kulka, S., & House, J. (1989). Cross-cultural and situational variation in requesting behavior. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (123-53). Norwood, NJ: Ablex. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests and apologies. Advances in Discourse Processes (Vol. 31). Norwood, NJ: Ablex. Blum-Kulka, S. & Katriel, T. (1991). Nicknaming practices in families: A cross-cultural perspective. In S. Ting-Tooney & F. Korseny (Eds.), Cross Cultural interpersonal communication: International & intercultural communication manual (Vol. 15, 5877). London: Sage. Brown, R. (1991, September). Language and the relational self. Paper delivered at the Fourth International Conference on Language and Social Psychology, Santa Barbara, California. Clyne, M. G. (1982). Multilingual Australia: Resources, needs, policies. Melbourne: River Seine. Edelsky, C. (1981). Who's got the floor? Language in Society, 10, 383-421. Edmondson, W. (1981). Spoken Discourse: A model for analysis. London: Longman. Ervin-Tripp, S., & Gordon, D. P. (1986). The development of children's requests. In R. E. Schiefelbusch (Ed.), Communicative competence: Assessment and intervention (6196). San Diego, CA: College Hill Press.

222

Discourse Perspectives

Geertz, C. (1973). The interpretation of cultures. New York: Basic Books. Geertz, C. (1976). From the native's point of view: On the nature of anthropological understanding. In E. Kenneth, K. Basso, & H. A. Selby (Eds.), Meaning in anthropology (221-31). Albuquerque: University of New Mexico Press. Giles, H. (1979). Ethnicity markers in speech. In R. Scherer & H. Giles (Eds.), Social markers in speech (251-91). Cambridge: Cambridge University Press. Glaser, B., & Strauss, A. (1964). Awareness contexts and social interaction. American Sociological Review, 24, 669-79. Gleason, J. B., Perlman, R., & Grief, E. (1984). What's the magic word?: Learning language through politeness routines. Discourse Processes, 7, 493-502. Gleason, J. B., & Weintraub, S. (1976). The acquisition of routines in child language: "Trick or Treat." Language in Society, 5, 129-36. Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday. Goffman, E. (1967). Interaction ritual: Essays on face to face behavior. New York: Doubleday. Grice, H. Paul. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and semantics (Vol. 3): Speech acts (43-59). New York: Academic Press. Grief, E. & Gleason, J. B. (1980). Hi, thanks, and goodbye: More routine information. Language in Society, 9, 159-67. Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Hammersley, M., & Atkinson, P. (1983). Ethnography:Principles andPractice. New York: Routledge. Heath, S. B. (1983). Ways with words: Language, life, and work in communities and classrooms. Cambridge: Cambridge University Press. House, J., & Kasper, G. (1981). Politeness markers in English and German. In F. Coulmas (Ed.), Conversational routine (157-85). The Hague: Mouton. Hymes, D. (1974). Foundations of sociolinguistics. Philadelphia: University of Pennsylvania Press. Katriel, T. (1986). Talking straight: Dugri speech in Israeli Sabra culture. Cambridge: Cambridge University Press. Labov, W., & Fanshel, D. (1977). Therapeutic discourse. New York: Academic Press. McWhinney, B., & Snow, C. (1985). The child language data exchange system. Journal of Child language, 12, 271-96. Ochs, E. (1988). Culture and language development: Language acquisition and language socialization in a Samoan village. Cambridge: Cambridge University Press. Olshtain, E., & Blum-Kulka, S. (1985). Degree of approximation: Nonnative reactions to native speech act behavior. In S. Gass & C. Madden (Eds.), Input in second language acquisition (303-29). Rowley, MA: Newbury House. Olshtain, E., & Blum-Kulka, S. (1989). Happy Hebrish: Mixing and switching in AmericanIsraeli family interaction. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition (Vol. 1): Discourse and pragmatics (59— 84). Clevedon and Philadelphia: Multilingual Matters. Pellegrini, A. D., Brody, G., & Stoneman, Z. (1987). Children's conversational competence with their parents. Discourse Processes, 10, 95-106. Philips, S. (1983). The invisible culture: Communication in classroom and community on the Warm Spring Indian Reservation. New York: Longman. Rabin, C. (1976). Acceptability in a revived language. In S. Greenbaum (Ed.), Acceptability in language (149-67). The Hague: Mouton. Rcisman, R. (1974). Contrapuntal conversation in an Antiguan village. In R. Bauman &

American-Israeli Families at Dinner

223

J. Sherzer (Eds.), Explorations in the ethnography of speaking. Cambridge: Cambridge University Press. Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simplest systematics for the organization of turn taking in conversation. Language, 50, 696-735. Schieffelin, B. (1990). The give and take of everyday life: Language socialization of Kaluli children. Cambridge: Cambridge University Press. Schieffelin, B., & Ochs, E. (Eds). (1986). Language socialization across cultures. Cambridge: Cambridge University Press. Sharwood-Smith, M. (1983). On first language loss in the second language acquirer. In S. Gass & L. Selinker (Eds.), Language transfer in language learning. Rowley, MA: Newbury House. Schumann, J. (1978). The acculturation model for second language acquisition. In R. Gingras (Ed.), Second language acquisition and foreign language teaching. Arlington, VA: Center for Applied Linguistics. Silverstein, M. (1976). Shifters, linguistic categories and cultural description. In K. Basso & H. Selby (Eds.), Meaning in anthropology (11-55). Albuquerque: University of New Mexico Press. Spradley, J. P. (1979). Ethnographic interview. New York: Holt, Rinehart and Winston. Tannen, D. (1981). Indirectness in discourse: Ethnicity as conversational style. Discourse Processes, 3(4), 221-38. Tao, H., & Thompson, S. A. (1991). English backchannels in Mandarin conversations: A case study of superstratum pragmatic "interference." Journal of Pragmatics, 16, 20925. Weinreich, U. (1953). Language in contact: Findings and problems. New York: Linguistics Circle of New York. (Reprinted, The Hague: Mouton, 1974). Weltens, B. (1987). The attrition of foreign language skills: A literature review. Applied Linguistics, 8(1), 2-39. Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University Press. Wolfson, N. (1989). Perspectives: Sociolinguistics and TESOL. New York: Newbury House. Verschueren, J. (1985). What people say they do with words. Norwood, NJ: Ablex.

11 Notes on the Interlanguage of Comity GUY ASTON

I take it that interlanguage pragmatics is concerned with the means by which goals are (or are not) achieved in contexts of spoken or written discourse where one or more of the participants is a nonnative speaker—what I shall term, for ease of reference, NNS discourse. As well as having descriptive and explanatory objectives— attempting to categorize and account for uses of discourse strategies that contrast qualitatively or quantitatively with those employed in contexts with exclusively native speaker participants (NS discourse)—it may also have pedagogic ones, aiming to facilitate learners' achievement of their goals in interaction. Whichever objective is primary, we must ask what kinds of goals need to be achieved, and what constraints on their achievement distinguish NNS from NS discourse. In so doing it seems important to consider not just the constraints relevant to the achievement of goals of communicating information but also those relevant to achieving those social goals of interpersonal rapport that dominate much of our everyday talk. Two main perspectives on conversational goals and constraints have been adopted in investigating the pragmatics of NNS discourse. On the one hand, we find work that takes the point of view of the NNS speaker, focusing on the means by which he or she attempts to express and convey particular meanings to interlocutors in reference to the psycholinguistic constraints (those of restricted L2 [second language] competence) under which such a speaker operates. On the other hand, we find work that treats the achievement of goals in discourse as a collective concern, focusing on the interactive procedures by which mutually satisfactory outcomes are jointly arrived at, in reference to the peculiar constraints of the social context (those of predictably unshared competence) in which participants in NNS discourse operate. Both perspectives may be illustrated from the literature on communication strategies (CS), which grew out of a concern with the means whereby learners can express an intended meaning given their limited knowledge of the L2. Thus Corder defined CS as "a systematic technique employed by a speaker to express his meaning when faced with some difficulty" (1983, 16); Faerch and Kasper, as "potentially conscious plans for solving what to an individual presents itself as a problem in reaching a particular communicative goal" (1983, 36; my emphasis). In contrast with this speaker-oriented, intra-organism view of CS as procedures employed to 224

Notes on the Interlanguage of Comity

225

express a meaning, other researchers in the field instead proposed that they be seen as "a mutual attempt to agree on a meaning where requisite meaning structures do not seem to be shared" (Tarone 1983, 65), taking an inter-organism perspective in which CS are considered to involve procedures of negotiation designed to obtain a mutually ratified outcome—what Widdowson (1983) has termed the convergence of participants' worlds. This second perspective involves a somewhat different view of discourse goals (perlocutionary ones of "agreeing on" rather than simply illocutionary ones of "expressing" a given meaning) and of the constraints in terms of which CS are to be accounted for. Rather than treating the latter as primarily a matter of nonnative speakers' "lexical deficit" (Bongaerts & Poulisse, 1989), leading to a categorization of CS in terms of speakers' personal "avoidance" (message abandonment, topic avoidance) and "achievement" strategies (circumlocution, paraphrase, appeal for assistance, transfer, mime: Faerch & Kasper, 1983), Tarone's definition stresses the constraints that derive from the lack of intersubjective resources in NNS discourse. It evaluates such "resource gaps" (not only lexicogrammatical but also sociocultural—a lack of shared conventions concerning language use, and of shared knowledge and experience of the world) not in terms of the "objective" status of resources as available or otherwise, but in terms of their seeming status—of participants' reciprocal perceptions and expectations as to the resources shared in the current context. A presumed lack of intersubjective context for communication may oblige participants to seek ways of making that context explicit and of establishing what is in fact shared, rather than simply assuming the "reciprocity of perspectives" of NS participants where meaning can reliably be expanded on the basis of an "etcetera principle" (Cicourel, 1973). This second viewpoint leads to an expansion of taxonomies of CS to include, for example, (a) strategies of "recipient design," which aim to alleviate the hearer's eventual problems of understanding rather than just the speaker's problems of expression—through repetition, explanation, explication, and avoidance of reference to predictably unshared culture-specific knowledge and experience (Tarone & Yule, 1987; Aston, 1988c); (b) interactive procedures of comprehension checking and confirmation, which repair failure and manifest that intersubjectivity has in fact been achieved (Long, 1983; Varonis & Gass, 1985); (c) preparatory strategies designed to influence assumptions and expectations preemptively—for instance, overt disclaimers of competence ("I have a problem. I'm a foreigner, and . . .": Anderson, 1988b), and covert ones, such as the adoption of a foreign accent and/or incorrect lexicogrammar (Di Pietro, 1987; Warren-Leubacker & Bohannon, 1982) —strategies that can be seen as canceling the presupposition that the resources usually shared in NS discourse are available, activating what has been called a "NNS set" of expectations (Anderson, 1988b): their utility in facilitating communication is shown by Warren-Leubacker and Bohannon (1982), who demonstrate how initial manifestations of incompetence by the NNS give rise to greater repetition and explanation by the NS participant. The wider view of CS provided by an interorganism perspective seems pedagogically useful, since in the final analysis language users, including learners and their interlocutors, are typically concerned with achieving perlocutionary outcomes of convergence, not merely illocutionary ones of

226

Discourse Perspectives

self-expression. It is clear from comparing the two approaches that the communication strategies highlighted differ according to whether discourse constraints are conceived of in intra- or inter-organism terms. While differing in their perspectives on the goals and constraints involved, both these approaches to communication strategies share a conception of discourse as primarily a matter of conveying information, that is, as transactional (Brown & Yule, 1983). In the intra-organism approach, such strategies as paraphrase, circumlocution, and mime are designed to convey information successfully by clarifying intended sense and reference. The inter-organism approach also treats discourse goals as primarily transactional. Desired outcomes are formulated in terms of participants' final information states, with convergence in respect of an "agreed meaning": CS "are used to deal with problems which may arise in the transmission of ... information" (Tarone & Yule, 1987, 50). This tendency to conceive goals as transactional (and hence to consider discourse constraints in relation to the achievement of such goals) glosses over the fact that much everyday speech use instead has as its primary goals the negotiation of interpersonal relationships. In such interactional speech, Brown and Yule note (1983, Iff.), the correct conveyance of information may be very much less important. Participants seem to be involved in "choric" behaviour, as Firth (1964, 112) put it, where what is important is not so much the understanding of what is said as the harmony of the dance of their saying, so that they "end up feeling comfortable with each other" (Brown & Yule, 1983, 12): the emphasis is on ritual displays of agreement and mutual appreciation.1 Rather than a use of language for communication (if by this term we mean, as is generally the case in applied linguistics, a convergence of participants' information states), in interactional speech we have a use of language for what Leech (1983) terms comity—the establishment and maintenance of friendly relations. What seems primarily at issue is a convergence of participants' worlds in affective terms—sharing feelings and attitudes rather than knowledge and ideas (Aston 1988b). An extreme example is Malinowski's (1930) "phatic communion," which he saw as totally devoid of information content; but even where some information is conveyed, as in everyday chat between lovers, friends, and gossips, the primary point of talk still seems to be that of establishing and maintaining relationships of varying degrees of intimacy: even at academic congresses, as Firth (1964) noted, much talk has a primary function of this type. As Coupland et al. (1992) observe, interactional speech may also play a significant role in what are prima facie purely transactional contexts, for instance, in celebrating transactional success or making transactional failure acceptable. In service encounters in shops, we often find assistants and customers engaging in pieces of interactional chat when the requested goods are not available—or unexpectedly are so (Aston 1988a; cf. examples, 2, 7, and 15 below). Given the frequency with which interactional speech occurs in everyday talk, it is relevant for interlanguage pragmatics to ask whether, and in what way, specific constraints condition the pursuit of interactional goals in NNS discourse, and what "comity strategies " may be available to participants in order to meet these. Turning aside from the literature on communication strategies to other areas of research into NNS discourse, we find a series of studies that focus on interactional

Notes on the Interlanguage of Comity

227

aspects of such talk. Work in contrast!ve pragmatics, while primarily concerned with the comparison of speech act realizations in different cultures by native speakers, has also examined realizations by nonnative speakers, comparing these with typical realizations in those speakers' Lls and with those of native speakers. The main speech acts selected for analysis are ones that are either face-threatening— requests (the CCSARP project: cf. Blum-Kulka et al., 1989a), refusals (Beebe & Cummings, 1985), disagreements (Beebe & Takahashi, 1989), complaints (Olshtain & Weinbach, 1987)—or face-supportive—compliments (Wolfson, 1981), thanks (Eisenstein & Bodman, 1986), apologies (Olshtain, 1983;Trosborg, 1987; CCSARP; Garcia, 1989)—and their realizations have been compared in terms of the politeness strategies employed to give and redress face (generally using taxonomies based on the work of Brown & Levinson, 1978, 1987). For example, in examining requests and apologies, the CCSARP project has examined ways in which learners' assessment of face-threat and the strategies they employ to redress that threat appear to differ from those of native speakers of the L2. In this way, such work highlights strategies employed by nonnative speakers to achieve interactional goals of satisfactory rapport. In terms of the distinction drawn above in discussing communication strategies, it should however be noted that this line of research adopts an intra-organism approach: it focuses on nonnative participants' expression of politeness. Limiting attention to the realization of the single speech act, it does not consider eventual interactive procedures by which mutually satisfactory interpersonal outcomes are established, and it views constraints on the achievement of interactional goals in terms of the NNS's limited competence; that is, strategies of politeness are usually accounted for in terms of inadequate knowledge of the sociolinguistic conventions of the L2 (giving rise to transfer from the LI) rather than in terms of a lack of intersubjective resources (giving rise to negotiation). This limits the value of the approach as an account of the strategies by which satisfactory rapport is achieved in NNS discourse. For while the NNS participant's politeness strategies undoubtedly contribute to the process whereby friendly relations are established and maintained in such discourse, inasmuch as they indicate that speaker's friendly intent, an account of such strategies does not explain how both participants2 effectively come to feel comfortable with each other as an outcome. (This because it is not necessarily sufficient to convey an intent to redress face in order to effectively redress it, and because the redress of the hearer's face by one speaker is clearly only one element in the process of mutual facework—itself in turn arguably only one aspect of the comity process.) Nor would it appear that the constraints on the achievement of interactional goals in NNS discourse are simply those posed by the objective limits to the NNS' competence; there may also be intersubjective, social constraints under which NNS interactional speech operates, due to a perceived lack of shared resources. The CCSARP project, which aims to "investigate the similarities and differences in the realisation patterns of given speech acts between native and nonnative speakers of a given language, relative to the same social constraints" (BlumKulka et al., 1989b, 13), thereby assumes that the social constraints in NS and NNS discourse can be treated as substantially identical rather than as in any way intrinsically distinct. In discussion of the project data on request realizations, however,

228

Discourse Perspectives

Faerch and Kasper (1989; see also Kasper, 1989) are led to place this assumption in doubt: noting that the tendency for nonnative speakers to provide warrants for requests is independent of their LI, and thus cannot be accounted for in terms of transfer, they suggest that this phenomenon reflects distinctive patterns of expectation in NNS discourse, so that "for NNS contributions to be successful they may have to follow different conversational principles" (Kasper, 1989, 54), "for instance, by placing more emphasis on phatic, metalingual and metacommunicative functions" (Faerch & Kasper, 1989, 245). Fiksdal (1988), in one of the few studies in this area to examine the politeness strategies in both NS and NNS in NNS discourse, similarly suggests that distinctive principles of interaction may be involved. By arguing in favour of context-specific principles for the conducting of interactional concerns in NNS discourse, these suggestions go in the direction of an inter-organism perspective on the constraints to which such discourse may be subject. It is the nature of such constraints that I wish to consider in this chapter.

Interactional Speech from an Inter-Organism Perspective A Neglected Issue? It seems useful, as a premise to more detailed consideration, to note some reasons for the lack of attention paid to interactional speech in NNS discourse from an interorganism perspective. This may help clarify the theoretical and methodological stance I want to adopt. One reason for this apparent neglect is undoubtedly the widespread assumption of a speech-act-based model of discourse. Placing the emphasis on the isolated utterance rather than on the jointly-performed dance as a whole, such a model in a sense presupposes that goals in discourse are the transactional ones of achieving correct understanding, in the shape of "illocutionary uptake" (Searle, 1969). Even where descriptions of NNS discourse have ventured beyond the single act, they have largely considered interactivity as a matter of "negotiating meaning"—of jointly clarifying and confirming understanding of the information that an individual speaker intends to convey by a particular act (Long, 1983; Widdowson, 1983). A speechact-based view entails that politeness, too, constitutes a kind of information, a "social meaning" conveyed parallel to the main illocution and to be correctly understood alongside it (Leech, 1983). Rather than conveying information, Brown and Yule (1983) argue, interactional speech instead involves negotiating shared attitudes and feelings: to describe this process we need to see talk as progressively bringing about affective convergence through the joint realization of interactive procedures (Aston, 1988b). A second reason is the tendency in applied linguistics to formulate pedagogic objectives in transactional terms—teaching language for communication (e.g., Widdowson, 1978; Johnson & Morrow, 1981; Brown & Yule, 1983). As a result of this tendency (in part derived from speech act theory), "communicative approaches" to language teaching have generally treated the achievement of interactional goals as a matter of the NNS conveying information in a socially acceptable manner (by

Notes on the Interlanguage of Comity

229

employing appropriate politeness strategies: e.g. Littlewood, 1981; Widdowson, 1983). Such an approach to the interactional, as noted above, casts limited light upon the processes by which mutually satisfactory interpersonal outcomes are collectively achieved. The pedagogic importance of these processes and outcomes, however, should not be underestimated. Many learners have goals of using the L2 to establish interpersonal rapport (Geoghegan, 1983), and it should be borne in mind that a context of positive rapport also facilitates transactional speech use; for instance, by rendering operative a "benevolence principle" whereby eventual understanding failures are more likely to be interpreted as errors rather than offenses (Thomas, 1983; Slugowski & Turnbull, 1988; Anderson, 1988b). A context of satisfactory rapport may also facilitate second language acquisition tout court (Gumperz, 1982; Fillmore, 1979).3 A third reason for the neglect of interactional speech is the difficulty of collecting relevant data. For instance, studies that have examined NNS politeness strategies have for the most part employed discourse completion tasks, where subjects are required to state what utterance they feel would be appropriate to a specified context (e.g., Blum-Kulka et al., 1989a). This method, however, discounts interactivity, with the possibility of the utterance's joint production over time (on the basis of back-channel feedback or its absence: cf. e.g., Anderson et al., 1988), and of the realization of the appropriate action across a series of turns (as would appear to be frequently the case for such face-threatening actions as complaints and apologies: George, 1990; Owen, 1983). Data where realizations are interactively determined can be obtained through the use of role-play, where subjects jointly simulate a situation calling for a given action-type (Beebe & Cummings, 1985; House, 1989), but it is difficult for this technique to reproduce the interpersonal context of naturally occurring talk. Role-played interactions are without effective social consequences, since the relationships between the characters are, in the final analysis, fictional and temporary; as a corollary, the real-life relationships of the actors (rather than the characters) may constitute an extraneous influence, revealed in the metadiscursive nudges, winks, and laughter that accompany so many learner role-plays (Zanca, 1988; Wildner-Bassett, 1989). Experimentally contrived situations in the laboratory pose similar difficulties. These too are isolated from the real-world context of actual relationships: situations where goals are transactional, such as the picture description tasks widely employed in work on communication strategies, are far easier to set up experimentally than ones where goals are additionally or primarily interactional (cf. Aston, 1986a, 1988b for further discussion of these issues). To examine interactional speech from an inter-organism perspective, it seems essential to look first and foremost at naturally occurring conversational data. A final reason would appear to be of an ideological nature. While a range of studies have used naturally occurring data to consider interactional aspects of NNS discourse, these have typically examined situations in which interpersonal conflict emerges rather than ones in which rapport is created successfully. Motivated by social concern at the extent of conflictuality and the abuse of power in cross-cultural settings, they have stressed how negative affect derives from misunderstanding of intent (e.g., Gumperz, 1982; Thomas, 1983). Their emphasis has thus been on the avoidance of interactional failure. If we are to understand how interactional success

230

Discourse Perspectives

can be achieved in NNS discourse, we need to focus on strategies for the establishment of positive rapport—enhancing friendly relations—rather than just the avoidance of negative outcomes of conflict and hostility. If interactional speech is seen not simply as facilitating the achievement of transactional goals by making communication acceptable, but as having objectives that may be the primary purpose of the talk, then comity needs to be seen as not simply a matter of mitigating face-threat and hence avoiding wnfriendly relations, but as a manner of enhancing relations in a positive sense (Aston, 1989).4 It is for the reasons outlined, I believe, that we find only passing observations in the literature which consider comity strategies from an inter-organism perspective. These limited observations are, however, suggestive, since they indicate that certain characteristic features of NNS discourse can find explanations in terms of the use of such strategies. For instance, McCurdy (1980) notes that participants in NNS discourse may "fake" understanding as a means of avoiding putting transactional failure, with its negative affective consequences, on record. Tarone (1980) notes the occurrence of repairs to linguistic form in context where mutual understanding has already been achieved. Excluding these from her transactionally defined taxonomy of communication strategies, she suggests instead that they have primarily social ends of "meta-communication . . . which says, 'I'm a member of your group'" (1980, 426)—an interpretation that finds echoes in the following remark of Widdowson's: Learners, like the ordinary people they are, adjust their language to an acceptable norm for two reasons: either in order to be more effectively communicative or in order to indicate a sense of identity with a particular group of language users. . . . It is difficult to correct learner error which has little or no communicative consequence because to do so is to ask learners to subscribe to an etiquette which may seem alien to them, quaint, even ridiculous; to conform to standards of behaviour that represent a code of conduct for particular social groups with which they have no social connection and no real affinity. (1984, 248-49; my emphasis)

The use in NNS discourse of confirmations of understanding and of repairs to form may in other words be accounted for in terms of participants' interactional rather than transactional concerns (Aston, 1986b); that is, they may constitute comity strategies aiming to maintain and enhance rapport. Such isolated observations, however, tell us little of the range of comity strategies that may be employed given the constraints peculiar to NNS discourse. Given the difficulty of setting up controlled contexts in which negotiation of rapport is at issue, any attempt to examine such strategies would seem to involve what Faerch termed "straying around in the forest" (1984, 60), focusing on a range of examples from various sources in the light of a general theory rather than attempting to deal with systematically elicited data. As Faerch stressed, such a methodology means that the investigator has to identify problems before proceeding to a description of solutions; I shall thus be more concerned here with the nature of those problems— the constraints on interactional speech in NNS discourse—than with the potential solutions to them—the precise range of the comity strategies available. I want to begin by examining the comity process in NS discourse and the intersubjective

231

Notes on the Interlanguage of Comity

resources it draws on, hoping thereby to identify areas that may be problematic in NNS discourse owing to the absence of shared resources. By corollary, we may also be able to highlight resources for comity that are, instead, likely to be shared in contexts of NNS discourse, and hence be potentially available for use.

Comity in NS Discourse Brown and Yule (1983, 4) argue that interactional speech involves the negotiation of shared feelings and attitudes by participants, as opposed to the sharing of knowledge and ideas that is at issue in transactional speech. The sharing of attitudes can be seen from two perspectives, those of negotiating solidarity and of negotiating support (Aston, 1988b). Brown and Oilman define solidarity as a matter of "likemindedness or similar behaviour dispositions" (1972: 258), so that its negotiation involves participants in finding they share attitudes to features of their world in common—for example, that both feel the same way about the weather, or about an Oxford University Press decision as to the print run on a book:5 (1) A: T's - tsuh beautiful day out isn't it. B: Yeh it's jus' gorgeous. (Pomerantz, 1975, 4) (2) A: Surprised that they: er didn't print eNOUGH, actually. c: Yes:, Tthey obviously underestimated it"* didn't they, badly, because I've been to th- THREE bookshops now, and none of themA: A You'd have thought they'd've judged THAT a bit better, yes:."* A: ==Sold out? c: Yeah. (Gavioli & Mansfield, 1990: Lod A-04/a; A = assistance, C = customer)

Such routines of agreement (for a detailed description, cf. Pomerantz, 1984a) seem characteristic of the negotiation of solidarity. The tendency to shift topics in casual conversation reflects the search for things to agree on—to identify common concerns toward which attitudes may be shared (Brown & Yule, 1983, 11). As well as features of the situation in which the talk takes place (including ones brought to that situation from past experience), common concerns may include features of the talk itself. For instance, conversational closings often involve agreement as to the conversation's satisfactory nature: (3) A: Nice talking to you ''sir.'* B: A Nice •*- talking to you. (Pomerantz 1975, 4)

Shared attitudes toward the talk may also be manifested without such explicit agreements, but simply by production of analogous contributions. Joint laughter can show shared pleasure at repartee (Jefferson et al., 1987; Davies, 1984). Recycling of structural chunks (Tannen, 1989), for instance, by repeating a prior sequence with an inversion of the discourse roles involved, can allow participants to show their mutual acceptance and approval of that prior sequence. Consider the following extract, again from an encounter in a bookshop:

232

Discourse Perspectives (4) c: I hesitate to ask you this, but could you tell me why: + "The dancing Wu Li masters" is under religion? A: + It is: believe it or not, according to Fontana, + an alternative theology book, c: + Oh:! *(2syll)!+ A: ASo is •* Capra's "Turning point" and "The tao of physics." c: + "The tao of physics" I could: - I could believe, I mean it doesn't 'actually MENtion:'* reTyligions in there:,** A: ^According -~*A:AAAccording -** according to Fontana it's the same kind of book, so therefore: 1—»c: =Mhm. I 'see.'* Yeah it is, it's (a group)! They're all about god andTTphysics.'** A: *(2syll)+ *: ^Mm.*« Yeah, c: Yehhs. TEr- they're- [laughs]-* 2-^ A: AYeah! Right! [laughs] Lot to do with religion that, yes?"* [laughs] (16) (Gavioli & Mansfield, 1990: Lod B-07/a)

Here we find a reciprocation of ironic comments followed by self-laughter (arrows). First the customer offers what he apparently intends as a little witticism ("They're all about God and physics"). The assistant initially fails to appreciate this irony, simply acknowledging it as informative ("Mm. Yeah"), whereupon the customer initiates laughter himself. The assistant then joins in laughing, and goes on to provide a contribution of her own that repeats the same structure: she produces a similar ironic comment, and in her turn follows this with self-laughter. Thereby she shows she has in fact appreciated the customer's previous contribution, and what appears to have started off as a failure in interactional terms is turned into a successful piece of repartee, where shared attitudes are displayed to the talk itself. Solidarity implies that attitudes are shared in the sense that participants' feelings vis-a-vis some state or event of which they have common experience are the same. In general terms, it implies that with respect to such experience, A's wants for A correspond to B's wants for B. However, it is also possible for attitudes to be shared in another sense. Rather than sharing attitudes to features of common experience (including their common talk), both may share attitudes to features of the experience of one of the participants, to the sources of her or his individual joys and griefs, where A's wants for A correspond to B's wants for A (or vice versa). One cannot always share another's feelings about such personal matters a triumph and bereavement in the sense of having the same feeling oneself, feeling as the other, but one may be able to do so in the sense of sympathising, or feeling for the other. As Goffman put it, "the needs, desires, conditions, experiences, in short the situation of one individual, when seen from his own point of view, provides a second individual with directions for formulating ritual gestures of concern" (1971, 92). Adopting Goffman's term, I shall refer to this type of affective convergence as support. If solidarity is largely characterized by routines of agreement, support seems characterized by routines of affiliation. In the following examples, sympathy is provided in response to personal disclosures: (5) E: I hadda liddle operation on my TOE this week, I hadtuh have- TOEnail taken off. B: (0.5) WHY::hh

Notes on the Interlanguage of Comity

233

E: Oh, I have a fungus'n I had'n infection, (0.3) T's a *HELL of a-1* -^B: AOH::::::-«Emma. Innat awful, (0.7) Well what a SHAME. Didjeh haftuh go in the hospit'l? E: No::, I dist hadda local deal, en I- id wadn'any fun, but I'm BEtter I wz, lying on the couch out'n front. ^>B: Oh::: I'm sorry EmMA? (Sacks, 1992, 573) (6) L: E: L: E: —> E: L: —>E: L: —>E: L:

=But you know when you get OUT it's kind of co:ld. (Oh: oh) YA:h« A Well it was,"* two o'clock in the morning and Tthen last night* A Huh Haw hawh ha:*w. T Oo I(h) bet that was (fu:n)* A nhhhhh hn-hn-hn with no:"* c- hhThh clothes on God it's good, hu-uh hUH huhh hh* A hhh hh aaAAAAAA:::::::: Isn't that exerting,* Uh:*:?1* A Oh:* that's wonderful,* A Oh::* God we had, we, I never had so much fun in my li:fe. (Jefferson, 1984, 200) T

As with solidarity, support may also focus on features of the talk itself—where the latter is viewed as an individual rather than a joint accomplishment. One participant can show appreciation of another's contributions to the discourse by laughing at her jokes and anecdotes—and that other can then in turn appreciate that appreciation: (7) A: Can I help you. c: Yeah, can you GIVE me a: - a- a prediction on when Jacques Lacan's "Ecrits" is likely to come back in? A: + Frankly no. I can give you a + er + educated guess, of: i(4syll)."* —»c: AWell.* That's [laugh\ - better than TI can* do [laugh}, so: —* A: *[laugh }"* (07) (Gavioli & Mansfield, 1900: Lod C-30/b)

Routines involving compliments and apologies, which also propose affiliation with the other's interests, can similarly be seen as means of negotiating support, in reference to the patterns of "supportive" and "remedial" interchanges described by Goffman (1971: for examples and discussion, cf. Aston, 1988b).

Grounds for Solidarity and Support Just as the transfer of information, through reference to maxims of sincerity and truthfulness, relies on bases of speakers' knowledge and belief to warrant their factual claims, so the negotiation of solidarity and support appears to rely on there being affective grounds that warrant the feelings in question. For instance, if they are to share attitudes in the solidary sense, it seems that both participants must have personal experience of the features that are assessed (Pomerantz, 1984b). In examples 1-3 above, the weather, the unavailability of a book, and the quality of the

234

Discourse Perspectives

current conversation are matters to which each participant can be assumed to have independent access. Where such experience is not self-evidently available, participants may describe it in the talk, telling stories from their past that provide warrants for their attitudes concerning, say, the poor safety standards of fairgrounds (Ryave, 1978). Where, on the other hand, a participant has no (or less direct) experience of the features in question to be drawn on, the negotiation of solidarity may be problematic—B cannot easily claim to know what it's like to travel by British Rail / live under a Conservative government / write a dissertation if he has never done so. In such cases, support may be more appropriate than solidarity. Grounds for sharing attitudes in the supportive sense lie in the nature of the relationship between the participants rather than in common experience: I can sympathise with you insofar as I have a personal involvement with you. This point was well made by Sacks, who noted how in disclosing one's personal experience to others, it's pretty much teller's business to tell the story with respect to its import for him, and it is his involvement in it that provides for the story's telling. . . . teller can tell it to somebody who knows and cares about him, and maybe recipient can tell it to someone who also knows and cares about initial teller, but it goes very little further than that. (Sacks, 1978, 261)

This underlines two requirements for the negotiation of support: (a) that the teller be involved in the story (i.e., he is entitled to have feelings toward the events described by virtue of direct experience of them), and (b) that the recipient be involved with ("knows and cares about") the teller. It is this latter factor that entitles the recipient to hold attitudes toward events of which he lacks independent experience. As in negotiating solidarity, participants may also be called upon to make their grounds for sharing attitudes in the supportive sense explicit, demonstrating that they really do "know and care about" each other. Thus in the following example, L demonstrates that she understands G's general situation (arrow): by citing her knowledge that G is "tied down" she shows her entitlement to sympathize with G in her dilemma of the previous night. (8) o: And Danny didn't get in so I didn't go: typing last ni:ght, L: =Didn't yTou::"* G: ANo:^ I Tca- I thought well I c^an't leave him for VTtwo** hours if I'm if he's crying when I've left him for one. L: ±Oh:::::.« L: A*n:No.*" L: + Oh: dear me. G: So: I euh you know as I say 1 didn't get to t ¥ yping.~* —* L: AOh::::: You're"* well tie:d dow:n aren't Tyou."* G: AWell-* I am rea:Tlly:« Yah. L: AYe:h." (Rahman, 1:4-6, cited in Jefferson, 1984, 204-5)

Clearly, degrees of "knowing and caring about: vary, and this is reflected in the varying "tellability" of personal information. There are events we disclose and jokes we tell to close friends but not to casual acquaintances. The loneliness that follows loss of a loved one can be interpreted in terms of no longer having anyone close

Notes on the Interlanguage of Comitv

235

enough to confide in (Sacks, 1972). What seems at issue in such cases is the restrictedness of the attitudes concerned, which not anyone may share supportively, but only someone who is sufficiently close to us. This requirement has been stressed by work in social psychology showing that growing intimacy correlates with an increase in the "privateness" of disclosure (Morton, 1978): conversely, Goffman noted how there is "a potential conflict between the provision of minor services to unacquainted others and the obligation to keep one's distance" (1971, 120). The restrictedness of attitudes also appears relevant in the negotiation of solidarity. Sharing mild feelings about the weather is a rather different matter from sharing homicidal feelings toward a common boss. While in the former case our feelings are those that anybody might be expected to have—those of competent consociates—in the latter they define a limited group, grounded in a highly specific common experience and perspective. Social psychology has argued that interpersonal attraction is related to perceived similarity—we like those who are like us— with "uncommon similarity" being particularly attractive (Lea & Duck, 1982). As with support, sharing restricted solidary attitudes thus implies a closer relationship, one between individuals with a peculiar "likemindedness." In both solidary and supportive senses, participants appear to negotiate progressively more restricted attitudes as intimacy increases, thereby increasingly defining themselves as somes rather than anys (Schenkein, 1978). From this perspective, engaging in interactional speech is not only a matter of identifying potential focuses of solidarity and/or support for which participants may have warrants, but also of locating ones where appropriately restricted attitudes may be shared. Resources for comity are constrained both by what is presumably common ground and by the presumed intimacy of the relationship. While to a large extent these constraints will be specific to the individual relationship concerned, we can, I think, also see ways in which they may generally differ in contexts of NS and NNS discourse, given the nature of the common ground and of the relationships which such contexts imply.

Grounds for Solidarity and Support and NNS Discourse Focusing on what I have termed solidarity, Fiksdal notes that "because rapport is created when speakers share common interest or common ground . . . NSs may rely on a different rapport system than NNSs" (1988, 3). It is clear that the common ground of consociates—what everybody can be expected to have experience of within a given culture—will not necessarily be available as a resource for solidarity where cultural backgrounds differ. While members of different cultures may share access to the weather (though there may well be differences in their feelings about it), they are unlikely to share similar experience of Mrs. Thatcher or British Rail, with respect to which the foreigner may lack warrants for protest on a par with those of the paid-up British native. In consequence, participants may have difficulty negotiating agreement vis-a-vis such topics. In extreme cases, we can imagine the absence of common ground emerging in ways like the following: (9) NNS: British Rail gets worse and worse, doesn't it? NS: I wouldn't have thought the trains were any better where you come from.

236

Discourse Perspectives (10)

NS: British Rail gets worse and worse, doesn't it? NNS: Well, I wouldn't know.

(fabricated) It is not merely such single domains of experience that cannot be assumed shared, however: comparing the behavior of interviewers in NS and NNS discourse, Fiksdal (1988, 8ff.) notes how appeals to generic intersubjective experience (through the use of expression such as "you know what I mean"), which invoke the interlocutor's capacity to fill in detail from common knowledge (following an etcetera principle, are not employed with NNS interviewees. In NS discourse such uses appear to help create rapport: not only because they assume common ground but because they propose allusiveness as a satisfactory discourse style. That is, participants can share attitudes to the fact that they don't need to spell out detail in order to understand each other. A well-known example of this involve irony, which, by alluding to what is conventionally said in routine situations, can be an important strategy for achieving solidarity (Sperber & Wilson, 1981, 1985, 1986). Discussing the utterance "He almost won, didn't he?" said by his wife to a companion when John has just come off court announcing that he "only just" lost after an ignominious defeat at tennis, Sperber and Wilson (1985) note that If they are old friends in the habit of making mild fun of their husbands, then the correct interpretation . . . would be easily accessible, and would merely confirm a range of shared assumptions about their relationship to each other and their outlook on the world. Suppose, however, that they are meeting for the first time. Then a hearer who penetrated the irony would have learned an enormous amount about the speaker's attitude to her husband, her outlook on life, all in a very short time. As a result of this exchange, a degree of intimacy would be achieved between them that would not have been created to anything like the same extent by [alternative, explicit examples], (viii: 74)

John's wife's quip is a contribution to the discourse that not just anyone might have made, and not just anyone understood: its production and grasp reveal the particular individuals concerned as sharing a whole set of attitudes. Irony is reputed to be less available as an interactional resource in NNS discourse, where it may be heard as learner error rather than wit (Harder, 1980; Widdowson, 1984), and the absence of the shared sociolinguistic competence necessary as a basis for an allusive style may be one reason for the widely noted long-windedness and prosaicness of NNS discourse (Blum-Kulka & Olshtain, 1986; Kasper, 1989; Aston, 1988c). Such absence of culture-specific common ground means that NNS discourse may lack certain of the resources typically available to consociates in the negotiation of solidarity. Similarly in negotiating support, with its basis in "knowing and caring about," it may be difficult for outsiders to the NS culture, who may be stereotyped victims and/or perpetrators of cultural prejudices, to draw on the neighborly respect and understanding that is expected of consociates. This might suggest that the resources for interactional speech in NNS discourse are drastically restricted with respect to those available in NS discourse, entailing a "reduced personality" for the learner, as Harder (1980) puts it. However, other areas of common ground, and bases for mutual respect and understanding, may instead be available. We should

Notes on the Interlanguage of Comity

237

not exclude the possibility that some of these may be characteristic of NNS discourse. As Borsch (1986) has documented, learners frequently perceive NNS discourse as expanding rather than restricting their interactional capacity. To appreciate how comity can be negotiated in such discourse, in other words, we must not only recognize the limits imposed by participants' nonconsociacy, but also explore the potential of nonconsociacy for providing interactional resources.

Resources for Comity in NNS Discourse Negotiating Solidarity: Exploiting the Lack of Consociacy I have suggested that negotiating solidarity is a matter of sharing attitudes toward features of common experience. Thereby participants demonstrate their similarities or likemindedness. While consociates may be able to draw on their "cultural value similarity, cultural status similarity, cultural linguistic similarity, and perceived cultural attitudinal similarity: (Ting-Toomey, 1986, 116), these are less likely to be present in inter-ethnic encounters. To find areas of likemindedness in NNS discourse, participants arguably need to turn from their identities as representative members of their cultures of origin to focus on their identities as individuals, and to their relationships as individuals to those cultures. If both participants manifest a certain independence with respect to their native cultures, revealing what TingToomey terms "flexible cultural identities," similarities may be found at a "metacultural" level. Recognition of the interactional potential of such flexibility may underlie learners' claims of enjoying "the freedom to act like an outlaw" when using the L2 (Borsch, 1986, 73). To exemplify, one way in which participants in NNS discourse seem able to establish solidarity is by both manifesting critical or ironic attitudes to their respective cultures of origin. In the following extract, the British NS takes a critical stance toward his country's role in the birth of America, from which he disaffiliates (first arrow), and then, by going on to criticise the role of the Italian NNS' compatriots in this process, seems to invite the latter to do the same in her turn (second arrow): (11) NNS: Um: someone told me, er that the captain, of Mayflower was from Walton. T ls it true?4 NS: AHe wasn't 4 FROM Walton he was CHRIStened in Walton. NNS: Ah. NS: Yeah. Don't know where he was born. + It was, a little bit before my time. NNS: [laughs} NS: =An- and: + but he was christened at All Saints Church. + Er: what was his name? + Something Matthews I think, but anyway. NNS: Mm. 1—> NS: Yeah. + + He started it all going over in America, he started all the trouble, you know. He took some of our trouble makers over there Tand they started more trouble, 4 NNS: ^{laughs}1*

238

Discourse Perspectives 2-^>

NS: and then your lot went over from the mafia and did the - made the rest, you know, I meanNNS: [laughs} NS: So. Tl mean-"* NNS: ^Yeah."* NS: + Between them they really carved it up (??). (Recca, 1988, 217)

Further into the same conversation, the NNS does take up a similar opportunity to disaffiliate from her culture, criticizing her compatriots as relatively humorless and unduly vociferous: (12)

NS: [talking about his stay in a hotel in Italy] 1 was known to the staff as papa grappa, ^[laughs \* NNS: ^[laughs}'* Yeah. NS: Yeah. Papa grappa. —» NNS: Mhm. [laughs} + The Italians have a sense of humor sometimes. NS: A very good sense of Thumor.< NNS: ^[laughs]* NS: Yeah. NSS: Mm: NS: I like the Romans, I like: 'the"* the Italians in Rome. NNS: ^Mm."* NNS: Mhm. They are very friendly. NS: Oh Tyes.^ —> NNS: ASome'*times they are too noisy. As Americans. . NS: =Too noisy? NNS: Yeah. NS: Yeah. Oh Yeah. And when the family TARGument started, A:H.* NNS: *[laughs]* NS: [laughs]

(Recca, 1988, 242) Both participants here display critical attitudes toward their respective cultures, from which they distance themselves. Taking the stance of fellow-outlaws, both occupy a no-man's-land that becomes a common ground, and whose successful creation is the focus of mutual appreciation and laughter. A second, related area vis-a-vis which solidarity may be negotiated concerns the task of being an NNS—of achieving and maintaining an identity outside the context of one's own culture. This is clearly a resource for solidarity in NNS-NNS encounters, but it can also be exploited in NS-NNS interaction if both participants can claim experience of being nonnative speakers. In the last example, the NS refers to his own experience as a foreigner, and in another extract from the same conversation, offers a few words in the language of his NNS interlocutor: (13)

NS: And twice a week, I used to drive NNS: Mh? NS: to the Vatican. NNS: Mhm. —> NS: "Vaticano," T isn't it?'*

Notes on the Interlanguage of Comity

239

NNS: AMhm.'* You have got a BEAUtiful pronunciation. Congratulations. NS: Especially when I say "Una birra per favore." [mutual laughter] (Recca, 1988, 243)

In this way discourse roles are momentarily inverted: the NS places himself in the role of an incompetent NNS, and the NNS in that of the NS who can provide support. This repetition with inversion enables the participants to share attitudes toward the business of being/interacting with a nonnative, and in this respect again to show a common flexibility with respect to their native cultures. In the cases just discussed, participants can be seen as making up for their lack of a common culture by using their experience as cultural outsiders as a basis for negotiating solidarity. Nonconsociacy also appears to offer specific grounds for sharing solidary attitudes with respect to the talk itself, seen as an appreciable joint accomplishment. While in NS discourse, the satisfactoriness of talk typically lies in its nonroutine nature, its particular wittiness and harmoniousness, which makes it the achievement of somes rather than anys, in NNS discourse even the routine may be problematic, with the result that its successful bringing-off can appear a noteworthy accomplishment that provides grounds for mutual satisfaction. Jordan and Fuller (1975) discuss a lingua franca example in which two Americans (n and g) are talking to two Mayans (J and M): (14) M: en Maya se dice pel. G: UH HUH. j: TiSabes?* G: *Pel.« N: Pel. M: Ah. G: PEL. j: TPel.-< M: APel.« G: PEL. PEL. j: Por que pe, ele. G: PE, ELE. N: PEL. j: Pe, e, ele. o: PE, E, ELE. AHA. j: iHah? G: PEL. j: *Hah.-« M: AHah.-* G: PEL. r. Pel. (Jordan & Fuller 1975, 26)

Of it, they observe: We find that an agreement is constructed over what "this" is called in Maya, though the original difficulty with what "this" is has in no way been resolved. It is evident here that speakers go to extraordinary lengths to assure each other that they

240

Discourse Perspectives know, understand, have heard, and are even able to spell a word the meaning of which is still problematic. So that what emerges as the achievement of the talk is not the imparting of information but rather the celebration of agreement, collaboration, consensus, a "look we can still talk together." (Jordan & Fuller, 1975, 27; my emphasis)

In NS discourse, to repeat and spell another's words successfully would hardly be a cause for remark (except in cases where linguistic competence cannot be assumed shared, e.g., with small children). In NNS discourse, on the other hand, insofar as such ordinary success cannot be presumed on, its occurrence can be treated as extraordinary. Even the production of formally correct speech can allow participants to demonstrate their concern with achieving certain standards of talk, and offer grounds for satisfaction. As well as such linguistic grounds, lack of shared sociocultural competence may allow the relatively everyday to be treated as extraordinary: in the next example, the fact that a large city bookshop accepts book tokens becomes a cause for joint celebration in a way that would hardly seem possible in the context of NS discourse. The outcome appears noteworthy precisely because the NNS did not, and was not expected to, assume that it was routinely predictable: (15) c": A: —> c": A:

You DON'T accept book tokens here, do you? Yes, we do. =Yes you do! Oh, yes! (Gavioli & Mansfield, 1990: Lod C-14; c" = non-native customer)

Such potential for "celebration of the ordinary" (Aston, 1988b) would seem to be a specific resource of NNS discourse. Once again the absence of common ground is paradoxically what provides participants with a common ground, in that they can share attitudes toward the fact of its creation.

Negotiating Support: Exploiting Incompetence NNS discourse also appears to offer particular resources for the negotiation of support, which, I suggested earlier, is based upon participants' "knowing and caring about" each other. Support, that is, implies that interlocutors are in some way defined as specific individuals—as somes, not just anys. The fact that participants in NNS discourse are not consociates may facilitate such a definition. The NNS' observable lack of competence sets her apart from the category of consociates, who can be presumed ordinarily competent other than in special circumstances (for instance, following injury or temporary distraction). Known incompetence can thus be a distinctive, individualizing feature that, by warranting support in the face of inadequate performance (providing grounds for appeals for assistance to, and displays of benevolent sympathy from, an interlocutor), can constitute an interactional resource. Thus in the following bookshop example, the nonnative customer's admission of sociocultural incompetence (first arrow) leads to the assistant's abandoning his previous animosity and becoming much more helpful, proposing possible locations for and then looking up the title in question in Books in print:

Notes on the Interlanguage of Comity

241

(16) c" Excuse me, um: a fr- friend of mine said she reserved a BOOK, in er Lovalts, which is called "Story of the pendulum." A =Well: - you've already been here haven't you. + T(Befo-)'4 c" ANot at the-"* not at the THIS desk. A =Well you asked the girl here. + 'Cos I was standing behind her when you asked her. 1—» c" =Oh I'm Tjusf* being getting lost probably. A A Yeah." A Yeah. + You've been to history? c" I have been everywhere. A + Well- I haven't a clue, I mean you don't know what it's about? C" + Well- they are different stories- aTbout pendulums, yes." A ASTOries. Well, it-4 it should be in- in literature then, shouldn't it? On: on the ground floor? c" + I see-1 think she said it was FIRst floor. + Where I can-1 where- who can: tell me-1 -1 mean I can BUY the book, but which- which depar- who can advise me what Tsort of-"* where can I look for it. A: ±Er:.« A: What's it called the: c": "Story of the pendulum." A: ="Story of the pendulum." + Erm: well I MIGHT get some clue from the: er + looking it up on the microfiche, on the other hand it might- if it just says "Story of the pendulum," then: er: + you know, (you don't -) hh. 2—> c": =Okay, hh you are very kind. (65) (Gavioli & Mansfield, 1990: Lod B-12)

Incompetence may also give grounds for sharing attitudes supportively vis-a-vis the discourse itself, by providing for the appreciable nature of one or other participant's contributions. In the last example, the assistant's support for her incompetence is made the focus of appreciation by the NNS customer (second arrow); incompetence may similarly legitimate appreciation by the NS of the NNS' contributions to the talk, along the lines of "You have got a beautiful pronunciation" in example 13 above. In NS discourse, appreciable contributions tend to be nonroutine ones—that not just anyone might have made. In NNS discourse, on the other hand, appreciable contributions may also be routine, "ordinary" ones, that precisely because they would be routine for consociates, can appear extraordinary where nonconsociates are involved. What is little more than routine cooperation by a NS can justify a NNS' effusive gratitude; NNS near-conformity to NS norms of speech can be remarked upon and applauded. In this light we can perhaps understand the interactional potential of self-repairs on linguistic form, for instance, noted by Tarone (1980): competence in those who are expected to be incompetent is potentially a praiseworthy accomplishment, a focus for the negotiation of support. Personality Reduction and Expansion I have suggested that the nonnative's status as an outsider with limited competence may offer specific resources for the negotiation of solidarity and support in talk. If

242

Discourse Perspectives

this is the case, participants may have an interest in ensuring the activation of an "NNS set" for the discourse, in order to make such resources available. However, to take such a line is not without its problems: to assume the status of an incompetent outsider can clearly appear equivalent to assuming a reduced personality, inviting at worst dismissive scorn and at best unwelcome paternalism. In the talk that follows example 16 above, for instance, the assistant at one point suggests that the book the NNS customer needs might be a simplified reader for foreigners—a potentially condescending suggestion that the NNS brusquely rejects: (17) A (Hm.) + Well: - [laughs] Yeah. I don't know, you- you've been to the HIStory department, c" Mhm. A And you've been to languages next door, c" Well I- it was not a language book. —> A Well okay, but 1 mean they have: um: simplified ENglish books there for example. + They have: abridged: easy readers, + in that department, c" But that was nothing to do with that. A =It wasn't. Right. And have you tried the Literature department ON the ground floor. (Gavioli & Mansfield, 1990: Lod B-12)

There are, however, a series of potentially mitigating factors that seem relevant in evaluating the NNS' role in such cases. First it seems worth nothing that the NNS need not explicitly admit incompetence for an "NNS set" to be triggered. Goffman (1983) notes that while customers in service encounters can pose their requests immediately if these correspond to what the assistant can reasonably be expected to have "in mind" (where the accessibility of the referent can be assumed), preparatory work is called for where requests are predictably more problematic, with preliminary greetings and apologies orienting the assistant to expect a nonroutine request. To the nonnative speaker, for whom even a routine request may be intersubjectively problematic, the use of such a preparatory strategy appears to offer a means of avoiding the default assumptions of NS discourse, and NS acknowledgment of the "marked" opening appears to signal the tacit instantiation of an "NNS set." Anderson (1988a) found that in the PIXI corpus of English bookshop encounters, such preparatory openings were acknowledged by assistants only when made by NNS customers, as in the following example: (18) c": Excuse me. -» A: Yeah. c": I'm looking for + this book "Residual and: infancy registration, + reGREssions." (Gavioli & Mansfield, 1990: Lod B-05)

In this context the strategic use of such "marked" openings may thus be a way of establishing the limited competence of the NNS tacitly, without explicit selfdenigration. In any case, the explicit or implicit assumption of incompetent outsider status needs to be assessed not just in terms of its immediate implications for the non-

Notes on the Interlanguage of Comity

243

native's face, but as a means to wider ends. It constitutes, on the one hand, an avoidance strategy with respect to possible interactional trouble. By laying claim to an initial face "lower" than that which is hoped for, learners reduce the risk of faceloss with respect to too high a claim: they can fail acceptably, with less risk of misunderstanding being heard as offense. On the other hand, by making available the sort of interactional resources described above, it can constitute an achievement strategy by opening up future opportunities for comity, for instance, by facilitating the "celebration of the ordinary" and laying the basis of a "knowing and caring about" that can be drawn on in negotiation of support later on in the encounter. Shared History as an Interactional Resource We must not forget that the construction of rapport takes place over time. It cannot be treated as a one-off event, the result of a single instance of negotiation of affective convergence, but is rather a progressive achievement. I suggested earlier that comity involves sharing increasingly restricted attitudes, establishing increasingly "uncommon" common ground and progressively greater intimacy. This process involves participants in building up, and building off, a joint interactional history. The importance of such history as a resource in talk was underlined by Goffman: In many of the conversations John has with others (whether with familiars or unfamiliars), he will not merely fall back on what can generally be taken to be shared, nor, indeed, will he merely tailor his references to the particular mind of his particular hearer; rather he will be inclined to select from whatever he shares with this other, just those topics that allow him to employ allusive phrases that only the recipient would immediately understand. Thus, his talk will not so much depend on common understanding as seek it out and then celebrate it. Indeed this gives to ordinary verbal contacts a greater degree of exclusivity and mutual dovetailing than one might otherwise expect—not only "recipient design," then, but the celebration of what it implies. (Goffman 1983, 18)

Of the various ways in which prior talk and its outcomes can provide resources in interactional speech, some seem particularly relevant to NNS discourse. For instance, the use of irony, traditionally viewed as relatively inaccessible in NNS discourse owing to the lack of intersubjective competence as a basis for allusion, seems quite possible if allusion is instead made to features of the shared history of the interactants. In the next example, again taken from the conversation analysed by Recca (1988), the male NS makes a joking comment concerning the possible attitudes of Anglo-Saxon women to rape (first arrow), which obtains appreciative laughter from the NNS (she makes it clear that she appreciates the joke rather than the sexism: second arrow): (19)

NS: So they chopped his head off, and they + took it away, and: NNS: + Mh. NS: and once more the: land was robbed, and: the houses were burnt, and: the women were raped, and it was: NNS: =Again.

244

Discourse Perspectives 1—»

NS: Yeah. And the women said "when are you coming back," you know. NNS: And yaT(/'?)'4 NS: ^[laughs]"* 2-^> NNS: That is not very nice of you. [laughs] + Yeah. NS: But there we are. (Recca, 1988, 221)

When the topic of women's roles reappears later on in the conversation, the NNS is able to ironically allude to her interlocutor's sexism as a known fact (arrows): (10) —*

—> 1—» -»

NS: I used to be on the council in Walton. + And: we had a woman, a very intelligent woman, NNS: Oh:, how extraordinary. NS: Oh:, terrible it is. NNS: [laughs] NNS: You know. + I like my women to be women, you know, T (??), you know. ^ NNS:AYeah, yeah, I know,"* you are very: NS: Yeah. NNS: progressive, ^[laughs]* NS: Ayeah."* NNS: I understood this, [laughs] (Recca, 1988, 253)

By alluding to an outcome established in prior talk, the NNS seems able to be ironic here without assuming that a wider, more generic sociocultural background is shared. It may be noted that with her final "I understood this," the NNS makes it explicit that her irony alludes to the prior talk. In the absence of the presumable common ground and involvement implied by consociacy, there in fact appears to be a greater need for participants in NNS discourse to make warrants for their attitudes explicit—to cite grounds for solidarity and support. Such explicitness has the advantage of also providing resources for further interactional speech. For instance, participants' descriptions of their grounds for holding attitudes will often involve disclosure of personal experiences, with the result that they become better known to each other in a growing intimacy. And while the need to disclose relatively routine experience may at times seem tedious for participants, who may feel obliged to provide continuously what, in the terms of NS discourse, is trivial information (Anderson, 1988b), we again need to bear in mind that this process contributes to the accumulation of interactional resources on the basis of which rapport may subsequently be developed further.

Conclusion In this chapter I have attempted to outline some resources available for comity in NNS discourse, arguing that these may not merely be a limited set of those available in NS discourse. If this is the case, morals for both the descriptive and pedagogic concerns of interlanguage pragmatics can be drawn. Descriptively, it supports the

Notes on the Interlanguage of Comity

245

argument that interlanguage pragmatics should operate with a difference hypothesis rather than a deficit hypothesis (Faerch & Kasper, 1989, 246), and not simply analyze NNS discourse in terms of failure to conform to NS conversational norms. Pedagogically, it implies that the learner's task in developing an ability for interactional speech using the L2 is not simply one of acquiring nativelike sociolinguistic competence in the attempt to mimic the behavior of a native speaker, but requires the development of an ability to use specific comity strategies appropriate to the context of NNS discourse. I have not attempted to provide a taxonomy of such strategies, merely to suggest areas where resources may lie, given the constraints that seem relevant to interactional goals from an inter-organism perspective. Certain possible maxims for the pursuit of comity nonetheless appear to emerge: cite warrants for your attitudes; admit incompetence; celebrate ordinary success; repair actual incompetence where possible; appreciate benevolence; distance yourself from the stereotypes of your culture. Just how such maxims can be effectively translated into specific comity strategies in interactive discourse calls for considerably more detailed study. They may, however, offer an initial framework for interpreting differences between NS and NNS discourse from an inter-organism perspective in interactional terms (as suggested by Aston, 1986b; Ehrlich et al., 1989), rather than in the purely transactional ones generally employed. The approach taken here, I would argue, also offers specific implications for pedagogy. It suggests that we need to help learners (a) to appreciate that there are differences in the bases for comity in NS and NNS discourse; (b) to view NNS discourse not just in negative terms of limitation, but as providing characteristic resources that can be exploited; (c) to investigate this potential by critically reflecting on the strategies employed in samples of NS and NNS discourse and by experimenting with the latter in "comity activities" where rapport is negotiated using the L2 (Aston, 1986a, 1988a,b). It has often been stressed that attempts to get learners to conform to NS practices only seem likely to succeed up to the point at which their identity is not thereby put at risk (Thomas, 1983; Litlewood, 1983; Widdowson, 1984): hence the importance of also helping them develop comity strategies of their own to establish satisfactory interpersonal relationships through the L2.

Notes 1. This is not, of course, to exclude the possibility that interactional speech can in some cases take a conflictual form, as in the banter of reciprocal insult between close friends (Slugowski & Turnbull, 1988). Such use is, however, typically keyed as nonserious (Goffman, 1974), and the importance of the harmony of the dance remains, as stressed for example by Labov's work on ritual insults (1972). 2. For ease of discussion, I adopt a default view of conversation as dyadic in this paper. 3. This claim can be interpreted in terms of a lowering of what Krashen calls the "affective filter," which, he argues, "is down when the acquirer [. . .] considers himself to be a potential member of the group speaking the target language" (1985, 3-4). 4. An analogy can be drawn here with the emphasis in work on communication strategies on achievement rather than avoidance strategies. A focus on achievement seems desirable

246

Discourse Perspectives

from a pedagogic perspective, since if comity is seen solely in terms of avoiding failure—of limiting conflict—the task of learners becomes truly Herculean, as they are condemned to wander a minefield of risk wherein they will inevitably always be less competent than a native speaker. This is hardly a motivating prospect for the learner, or for the pedagogy, whose role becomes that of trying to limit the extent of the social damage that the learner will inevitably suffer. If on the other hand comity is approached in the positive terms of seeking out opportunities for enhancing rapport—in particular those opportunities specifically offered by NNS discourse—then the learner is offered chances of success, and also, I would argue, the pedagogy. 5. Transcription conventions used in examples are those employed for the PIXI corpora (Gavioli & Mansfield, 1990). Extracts from other sources have been retranscribed using the same conventions. + (n) T texf* TT texf*<* = text = =text text texttext: . ,? ! TEXT [comment] (text) (nsyll) (??) Name

short pause longer pause (n seconds) spoken in overlap with next speaker's Atexf* spoken in overlap with next speaker's AAtext** latched to previous turn in transcript sequence latched to previous-but-one turn in transcript sequence previous utterance broken off previous syllable broken off previous syllable drawn out (number of colons indicates extent of lengthening) give a rough guide to intonation capitalization indicates marked emphasis situational features or nonverbal behavior speech unclear: tentative transcription speech unclear: n = approximate number of syllables speech unclear speech altered to preserve anonymity is italicized

References Anderson, L. (1988a). Access routines in service encounters. In G. Aston (Ed.), Negotiating service: Studies in the discourse of bookshop encounters (99-134). Bologna: Cooperative Libraria Universitaria Editrice. Anderson, L. (1988b). Issues in native-nonnative interaction. In G. Aston (Ed.), Negotiating service: Studies in the discourse of bookshop encounters. (267-86). Bologna: Cooperative Libraria Universitaria Editrice. Anderson, L., Aston, G., & Tucker, G. (1988). The joint production of requests in service encounters. In G. Aston (Ed.), Negotiating service: Studies in the discourse of bookshop encounters (135-51). Bologna: Cooperativa Libraria Universitaria Editrice. Aston, G. (1986a). Roleplay and interactional speech. In S. Holden (Ed.), Techniques of teaching: From theory to practice (20-25). Oxford: Modern English. Aston, G. (1986b). Trouble-shooting in interaction with learners: The more the merrier? Applied Linguistics, 7, 128-43. Aston, G. (1988a). An applied linguistic approach to the analysis of naturally occurring conversation. In C. Cipolli & E. Rigotti (Eds.), Ricerche di semantica testuale, (25982). Brescia: La Scuola.

Notes on the Interlanguage of Comity

247

Aston, G. (1988b). Learning comity: An approach to the description and pedagogy of interactional speech. Bologna: Cooperativa Libraria Universitaria Editrice. Aston, G. (1988c). Strategic di comunicazione nella scrittura in LI e in L2. In E. Glass, F. Marroni, G. Micks, & C. Pagetti (Eds.), Metamorfosi: Traduzione-tradizione (232-42). Pescara: CLUA. Aston, G. (1989). Solidarity and conflict as metaphors of the discourse process. In L. Curti, L. Di Michele, T. Frank, & M. Vitale (Eds.), // mum del linguaggio (413-26). Naples: Istituto Universitario Orientale. Beebe, L., & Cummings, M. (1985). Speech act performance: A function of the data collection procedure? Paper presented at TESOL '85, New York. Beebe, L., & Takahashi, T. (1989). Do you have a bag? Social status and patterned variation in second language acquisition. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (10325). Clevedon and Philadelphia: Multilingual Matters. Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989a). Cross-cultural pragmatics: Requests and apologies. Norwood, NJ: Ablex. Blum-Kulka, S., House, J., & Kasper G. (1989b). Investigating cross-cultural pragmatics: An introductory overview. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Crosscultural pragmatics': Requests and apologies (1-34). Norwood, NJ: Ablex. Blum-Kulka, S., & Olshtain, E. (1986). Too many words: Length of utterance and pragmatic failure. Journal of Pragmatics, 8, 47-61. Bongaerts, T., & Poulisse, N. (1989). Communication strategies in LI and L2: Same or different? Applied Linguistics, 10, 253-68. Borsch, S. (1986). Some ideas concerning the emotional dimension of foreign language teaching. In G. Kasper (Ed.), Learning, teaching and communication in the foreign language classroom (71—81). Aarhus, Denmark: Aarhus University Press. Brown, G., & Yule, G. (1983). Teaching the spoken language. Cambridge: Cambridge University Press. Brown, P., & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In E. Goody (Ed.), Questions and politeness (56-289). Cambridge: Cambridge University Press. Brown, P., & Levinson, S. (1987). Politeness. Cambridge: Cambridge University Press. Brown, R., & Oilman, A. (1972). The pronouns of power and solidarity. In P. P. Giglioli (Ed.), Language and social context (252-82). London: Penguin. (Original work published 1960.) Cicourel, A. V. (1973). Cognitive sociology. London: Penguin. Corder, S. P. (1983). Strategies of communication. In Faerch, C. & Kasar, G., Eds. Strategies in interlanguage communication (15-19). London: Longman. Coupland, J., Coupland, N., & Robinson, J. (1992). "How are you?": Negotiating phatic communion. Language in Society, 27, 207-30. Davies, C. (1984). Joint joking: Improvisational humorous episodes in conversation. Berkeley Linguistics Society, 10, 360-71. Di Pietro, R. (1987). Strategic interaction. Cambridge: Cambridge University Press. Ehrlich, S., Avery. P., & Yorio, C. (1989). Discourse structure and the negotiation of comprehensible input. Studies in Second Language Acquisition, 11, 397-414. Eisenstein, M., & Bodman, J. (1986). "I very appreciate": Expressions of gratitude by native and non-native speakers of American English. Applied Linguistics, 7, 167-85. Faerch, C. (1984). Strategies in production and reception—Some empirical evidence. In A. Davies, C. Criper, & A.P.R. Howatt (Eds.), Interlanguage (49-77). Edinburgh: Edinburgh University Press. Faerch, C., & Kasper, G. (1983a). Plans and strategies in foreign language communication.

248

Discourse Perspectives

In Faerch, C. & Kasper G. (Eds.) Strategies in Interlanguage Communication (21039). London: Longman. Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request realization. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (221-47). Norwood, NJ: Ablex. Fiksdal, S. (1988). Verbal and nonverbal strategies of rapport in cross-cultural interviews. Linguistics and Education, I , 3-18. Fillmore, L. W. (1979). Individual differences in second language acquisition. In C. J. Fillmore, D. Kempler, & W. Wang (Eds.), Individual differences in language ability and language behavior (203-28). New York: Academic Press. Firth, J. R. (1964). The tongues of men, and speech. London: Oxford University Press. (Original work published 1930/1937.) Garcia, C. (1989). Apolgising in English: Politeness strategies used by native and non-native speakers. Multilingua, 8, 3-20. Gavioli, L., & Mansfield, G. (Eds.). (1990). The PIXI corpora: Bookshop encounters in English and Italian. Bologna: Cooperativa Libraria Universitaria Editrice. Geoghegan, G. (1983). Non-native speakers of English at Cambridge University. Cambridge: Bell Educational Trust. George, S. (1990). Getting things done in Naples: Action, language and context in discourse description. Bologna: Cooperativa Libraria Universitaria Editrice. Goffman, E. (1971). Relations in public. London: Penguin. Goffman, E. (1974). Frame analysis. London: Penguin. Goffman, E. (1983). Felicity's condition. American Journal of Sociology, 89, 1-53. Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Harder, P. (1980). Discourse as self-expression—On the reduced personality of the second language learner. Applied Linguistics, I , 262-70. House, J. (1989). Politeness in English and German: The functions of please and bitte. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-cultural pragmatics: Requests and apologies (96-119). Norwood, NJ: Ablex. Jefferson, G. (1984). On stepwise transition from talk about a trouble to inappropriately nextpositioned matters. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action (346-69). Cambridge: Cambridge University Press. Jefferson, G., Sacks, H., & Schegloff, E. (1987). Notes on laughter in the pursuit of intimacy. In G. Button & J. Lee (Eds.), Talk and social organisation (152-205). Clevedon and Philadelphia: Multilingual Matters. Johnson, K., & Morrow, K., (Eds.). (1981). Communication in the classroom. London: Longman. Jordan, B., & Fuller, N. (1975). On the non-fatal nature of trouble: Sense-making and trouble-managing in lingua franca talk. Semiotica, 13, 11—31. Kasper, G. (1989). Variation in interlanguage speech act realization. In S. Gass, C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second language acquisition: Discourse and pragmatics (37-58). Clevedon and Philadelphia: Multilingual Matters. Krashen, S. (1985). The input hypothesis. London: Longman. Labov, W. (1972). Rules for ritual insults. In D. Sudnow (Ed.), Studies in social interaction (120-79). New York: Free Press. Lea, M., & Duck, S. (1982). A model for the role of similarity of values in friendship development. British Journal of Social Psychology, 21, 301-10. Leech, G. (1983). Principles of pragmatics. London: Longman. Littlewood, W. (1981). Communicative language teaching: An introduction. Cambridge: Cambridge University Press.

Notes on the Interlanguage of Comity

249

Littlewood, W. (1983). Contrastive pragmatics and the foreign language learner's personality. Applied Linguistics, 4, 200-206. Long, M. H. (1983). Native speaker/non-native speaker conversation and the negotiation of comprehensible input. Applied Linguistics, 4, 126-41. Malinowski, B. (1930). The problem of meaning in primitive languages. In C. Ogden & I. A. Richards, The meaning of meaning (296-336). London: Routledge and Kegan Paul. McCurdy, P. (1980). Talking to foreigners: The role of rapport. Unpublished doctoral dissertation, University of California at Berkeley. Morton, T. (1978). Intimacy and reciprocity of exchange: A comparison of spouses and strangers. Journal of Personality and Social Psychology, 36, 72-81. Olshtain, E. (1983). Sociocultural competence and language transfer: The case of apology. In S. Gass & L. Selinker (Lids.), Language transfer in language learning (232-49). Rowley, MA: Newbury House. Olshtain, E., & Weinbach, L. (1987). Complaints—A study of speech act behavior among native and nonnative speakers of Hebrew. In M. Papi & J. Verschueren (Eds.), The pragmatic perspective (195-208). Amsterdam: Benjamin. Owen, M. (1983). Apologies and remedial interchanges. The Hague: Mouton. Pomerantz, A. (1975). Second assessments: A study of some features of agreements/ disagreements. Unpublished doctoral dissertation, University of California at Irvine. Pomerantz, A. (1984a). Agreeing and disagreeing with assessments: Some features of preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action (152-63). Cambridge: Cambridge University Press. Pomerantz, A. (1984b). Giving a source or basis: The practice in conversation of telling "how I know." Journal of Pragmatics, 8, 607-25. Recca, L. (1988). Negoziazioni di identita fra parlanti nativi e non nativi in conversazioni in lingua inglese. Unpublished dissertation, Universita di Bologna, Facolta di Lettere e Filosofia. Ryave, A. (1978). On the achievement of a series of stories. In J. Scherikein (Ed.), Studies in the organization of conversational interaction (113-32). New York: Academic Press. Sacks, H. (1972). An initial investigation of the usability of conversational data for doing sociology. In D. Sudnow (Ed.), Studies in social interaction (31-74). New York: Free Press. Sacks, H. (1978). Some technical considerations of a dirty joke. In J. Schenkein (Ed.), Studies in the organization of conversational interaction (249-69). New York: Academic Press. Sacks, H. (1992). Lectures on conversation H. Oxford: Basil Blackwell. Schenkein, J. (1978). Identity negotiations in conversation. In Schenkein 1978a. Studies in the organization of conversation interaction. New York: Academic Press. 57-78. Searle, J. R. (1969). Speech acts. Cambridge: Cambridge University Press. Slugowski, B. R., & Turnbull, W. (1988). Cruel to be kind and kind to be cruel: Sarcasm, banter and social relations. Journal of Language and Social Psychology, 7, 101-21. Sperber, D., & Wilson, D. (1981). Irony and the use-mention distinction. In P. Cole (Ed.), Radical pragmatics (295-318). New York: Academic Press. Sperber, D., & Wilson, D. (1985). Relevance. Unpublished manuscript. Sperber, D., & Wilson, D. (1986). Relevance. Oxford: Basil Blackwell. Tannen, D. (1989). Talking voices: Repetition, dialogue and imagery in conversational discourse. Cambridge: Cambridge University Press. Tarone, E. (1980). Communication strategies, foreigner talk, and repair in intcrlanguage. Language Learning, 30, 417—31.

250

Discourse Perspectives

Tarone, E. (1983). Some thoughts on the notion of "communication strategy." In C. Faerch & G. Kaspcr (Eds.), Strategies in interlanguage communication (61-74). London: Longman. (Original work published 1981.) Tarone, E., & Yule, G. (1987). Communication strategies in East-West interactions. In L. Smith (Ed.), Discourse across cultures (49-65). New York: Prentice-Hall. Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4, 91-112. Ting-Toomey, S. (1986). Interpersonal ties in intergroup communication. In W. B. Gudykunst (Ed.), Intergroup communication (114—26). London: Arnold. Trosborg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11, 147-67. Varonis, E., & Gass, S. (1985). Non-native/non-native conversations: A model for the negotiation of meaning. Applied Linguistics, 6, 71-90. Warren-Leubackcr, A., & Bohannon, J., Ill (1982). The effects of expectation and feedback on speech to foreigners. Journal of Psycholinguistic Research, 11, 207-15. Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University Press. Widdowson, H. G. (1983). Learning purpose and language use. Oxford: Oxford University Press. Widdowson, H. G. (1984). Explorations in applied linguistics, 2. Oxford: Oxford University Press. Wildner-Bassett, M. E. (1989). Coexisting discourse worlds and the study of pragmatic aspects of learners' interlanguage. In W. Oleksy (Ed.), Contrastive pragmatics (251 — 75). Amsterdam: Benjamin. Wolfson, N. (1981). Compliments in cross-cultural perspective. TESOL Quarterly, 15, 11724. Zanca, C. (1988). // roleplay nell'insegnaniento della lingua straniera. Unpublished dissertation, Universita di Bologna, Facolta di lettere e Filosofia.

ABOUT THE AUTHORS

GUY ASTON is associate professor of English at the School for Interpreters and Translators of the University of Bologna. In the 1980s he was coordinator of the PIXI research project on the pragmatics of public service encounters in English and Italian, and has since continued working on the acquisition of oral competence in a second language, using conversational analytic methods of discourse description. LESLIE M. BEEBE is professor of linguistics and education in the Applied Linguistics and TESOL programs at Teachers College, Columbia University. She is editor of Issues in Second Language Acquisition: Multiple Perspectives and coauthor of English in the Cross-Cultural Era: The Social Rules of Speaking and TESOL. She publishes widely in sociolinguistics and second language acquisition. During 1990-1991, she served as president of the American Association for Applied Linguistics. MARC BERGMAN has taught English as a Foreign Language in such places as Yemen, Holland, and Thailand since 1976. He is currently a lecturer in the ELT Department of Thames Valley University (London), where he is assistant director of the M.A. in English Language Teaching program. His research interests include cross-cultural and interlanguage pragmatics, ethnographic approaches to language learning/teaching, and media discourse. ELLEN BIALYSTOK is professor of psychology at York University in Toronto, Canada. Her research is in cognitive development, language acquisition, and bilingualism. Current research projects include investigating the development of children's ability to use symbolic representation in the domains of language and number. Recent books include Communication Strategies and Language and Processing in Bilingual Children. SHOSHANA BLUM-KULKA is associate professor in the departments of Communication and Education at Hebrew University, Jerusalem, where she teaches courses in sociolinguistics, discourse analysis, language and education, and crosscultural pragmatics. Her research is in cross-cultural pragmatics, interlanguage pragmatics, and family discourse. She is co-author and co-editor of Cross-Cultural Pragmatics: Requests and Apologies and of Interlingual and Intercultural Communication. She is currently completing an extensive study of family discourse in Israeli and American families. JEAN BODMAN is president of the American Training Institute. She is a senior consultant for the U.S.I.A. and has written numerous ESL And EEL textbooks. 251

252

About the Authors

MIRIAM EISENSTEIN is associate professor of English Education at New York University. She is chair of the Second Language Acquisition Circle and author of The Dynamic Interlanguage. JULIANE HOUSE is professor of Applied Linguistics and head of the English Language Program at the University of Hamburg, Germany. She is author of A Model for Translation Quality Assessment and co-editor of Interlingual and Intercultural Communication and Cross-Cultural Pragmatics. Her research interests include translation theory, contrastive and interlanguage pragmatics, and misunderstandings in everyday life. GABRIELE KASPER is on the faculty of the M.A. in ESL and Ph.D. in Second Language Acquisition programs at the University of Hawaii. Before she discovered that she likes the tropics, she taught applied linguistics in Germany and Denmark. Her research interests are cross-cultural and interlanguage pragmatics, and psycholinguistic aspects of second language learning and use. ELITE OLSHTAIN is professor of Language Teaching at Tel Aviv University and dean of the School of Education. In 1979 she established the M.A. in TEFL program at Tel Aviv University. She teaches graduate courses in second language acquisition, discourse analysis, course design and policy making, the structure of English, and classroom-oriented research. Her research interests are in discourse analysis, language attrition, factors affecting success in language acquisition, and curriculum development. She had published articles in various professional journals and anthologies and is co-author (with Fraida Dubin) of Facilitating Language Learning, Reading by all Means, Reading on Purpose, and Course Design. She has also been involved in various materials development projects. RICHARD SCHMIDT teaches in the M.A. in ESL and Ph.D. in Second Language Acquisition programs at the University of Hawaii. He has done research and teacher training in Brazil, Egypt, Japan, Spain, and Thailand. His main research interests are in the areas of psycholinguistics and sociolinguistics and the intersection of the two. TOMOKO TAKAHASHI is associate professor of Languages and Cultures at Soka University in Tokyo (currently at the Los Angeles campus). She has published in the areas of second language acquisition, psycholinguistics and sociolinguistics. She has also written a number of books and textbooks for Japanese learners of English. LIORA WEINBACH is a teacher of Hebrew as a Second Language at Tel Aviv University and coordinator of a government project for the development of materials for the teaching of Hebrew in the (former) Soviet Union. She is co-author of The Art of Discourse (in Hebrew) and a series of basic Hebrew-foreign language dictionaries ("2000+"). Her publications include textbooks for the teaching of Hebrew as a Second Language, and a book on the methodology of academic writing. She is currently researching inter-ethnic variation in the speech act of complaining within Israeli society.

About the Authors

253

ELDA WEIZMAN is currently senior lecturer at the Bar-Han School for Translators and Interpreters, and at the Hebrew Department of Ben Gurion University of the Negev. Her publications include articles on the comparative analysis of journalistic texts in English, French, and Hebrew, and its implications for translators and learners; the nature of irony and hints, in a cross-cultural perspective; and the interpretation (and misinterpretation) of indirect speakers' meanings in written and oral discourse. She is currently engaged in a research project on discourse patterns in news interviews, supported by the Israeli Academy of Sciences and Humanities. JANE ZUENGLER is associate professor in the English Department, University of Wisconsin-Madison, where she teaches courses in second language acquisition and conducts research on sociolinguistic aspects of second language use. She has conducted research on accommodation theory and second language performance, and is presently researching the influence of content domain on L2 conversational participation. She has published articles in Studies in Second Language Acquisition, Applied Linguistics, and anthologies on second language research.