Speech Acts Across Cultures

Speech Acts Across Cultures

WDE

G

Studies on Language Acquisition 11

Editor

Peter Jordens

Mouton de GruyterBerlin New York

Susan M. Gass and Joyce Neu(Editors)

Speech Acts Across CulturesChallenges to Communicationin a Second Language

Mouton de GruyterBerlin New York 1996

Mouton de Gruyter (formerly Mouton, The Hague)is a Division of Walter de Gruyter & Co., Berlin.

The series Studies on Language Acquisition was formerlypublished by Foris Publications, Holland.

@ Printed on acid-free paper which falls within the guidelinesof the ANSI to ensure permanence and durability.

The Library of Congress lists the hardcover edition as follows:

Library of Congress Cataloging-in-Publication Data

Speech acts across cultures ; challenges to communication in asecond language / Susan M. Gass and Joyce Neu (editors).

p. em. - (Studies on language acquisition; 11)Includes bibliographical references and index.ISBN 3-11-014082-9 (alk. paper) ISBN 978-3-11-019125-71. Second language acquisition. 2. Speech acts (Linguistics)

3. Intercultural communication. I. Gass, Susan M. II. Neu,Joyce, 1950- III. Series.PI18.2.S67 1995303.48'2-dc20 95-40820

CIP

© Copyright 1995 by Walter de Gruyter & Co., D-I0785 BerlinAll rights reserved, including those of translation into foreign languages. No part of this bookmay be reproduced or transmitted in any form or by any means, electronic or mechanical,including photocopy, recording or any information storage and retrieval system, withoutpermission in writing from the publisher. Printed in Germany.

Contents

Susan M. GassIntroduction .

Part IMethodological issues

Andrew CohenInvestigating the production of speech act sets.

Noel Houck - Susan M. GassNon-native refusals: A methodological perspective.

1

21

45

Leslie M. Beebe - Martha Clark CummingsNatural speech act data versus written questionnaire data:How data collection method affects speech act performance . 65

Part IISpeech acts in a second language

Initiating and maintaining solidarity

Miriam Eistenstein Ebsworth - Jean "W: Bodman Mary ~arpenter

Cross-cultural realization of greetings in American English. 89

Gayle L. Nelson - Waguida El Bakary - Mahmoud Al BatalEgyptian and American compliments: Focus on second languagelearners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Michael L. Geis - Linda L. HarlowPoliteness strategies in French and English . . . . . . . . . . . . . 129

VI Contents

Naoko Maeshiba - Naoko Yoshinaga - Gabriele Kasper Steven RossTransfer and proficiency in interlanguage apologizing . . . . . . . 155

Face-threatening acts

Beth Murphy - Joyce NeuMy grade's too low: The speech act set of complaining. . . . . . . 191

Diana BoxerEthnographic interviewing as a research tool in speech act analysis:The case of complaints". . . . . . . . . . . . . . . . . . . . . . . 217

Myra GoldschmidtFrom the addressee's perspective: Imposition in favor-asking. . . . 241

Dale April KoikeTransfer of pragmatic competence and suggestions in Spanishforeign language learning. . . . . . . . . . . . . . . . . . . . . . 257

Part IIIApplications

Richard Schmidt - Akihiko Shimura - Zhigang Wang Hy-sook JeongSuggestions to buy: Television commercials from the U. S., Japan,China and Korea. . . . . . . . . . . . . . . . . . . . . . . . . . 285

John L. GrahamCulture, negotiations and international cooperative ventures. . . . 317

Subject index .

Author index .

343

347

Introduction

Susan M. Gass

1. Introduction

This book investigates the notion speech act from a cross-cultural perspective. That is, the starting point for this book is the assumption thatspeech acts are realized from culture to culture in different ways and thatthese differences may result in communication difficulties that range fromthe humorous to the serious.

Early studies in speech acts stem from the field of philosophy (e. g.,Austin 1962; Grice 1957, 1975; Habermas 1979, 1991 and Searle 1969,1975, 1979, 1983, 1986, 1991) and have been extended and amplified onby scholars from a number of different fields (e. g., linguistics - Sadock1974; anthropology - Hymes 1974; Gumperz 1982; child language Ochs - Schiefflin 1979). What these studies have in common is theassumption that fundamental to human communication is the notion ofa speech act, that is, the performance of a certain act through words (e. g.,requesting something, refusing, thanking, greeting someone, complimenting, complaining)l. Not only does the linguistic realization of the samespeech act differ, but the force of a speech act might differ. For example,in some cultures to refuse an offer of something may necessitate much"hedging" or "beating around the bush" before an actual refusal mightbe made. In other cultures, a refusal may not necessitate as much mitigation. The result may, in some cases, be a misinterpretation of whether ornot an actual refusal has been made, but may also be a misunderstandingof the intentionality of the refuser. In these latter instances, an individualmay be labelled as "rude", not because of the fact of refusal, but becauseof the way the refusal was executed. Olshtain and Cohen cite the following example of a misunderstanding due to the realization of the speechact of apology:

One morning, Mrs. G, a native speaker of English now living in Israel, wasdoing her daily shopping at the local supermarket. As she was pushing hershopping cart she unintentially bumped into Mr. Y, a native Israeli. Her

2 Susan M. Gass

natural reaction was to say "I'm sorry" (in Hebrew). Mr. Y turned to herand said, "Lady, you could at least apologize." On another occasion thevery same Mr. Y arrived late for a meeting conducted by Mr. W (a nativespeaker of English) in English. As he walked into the room he said, "Thebus was late," and sat down. Mr. W, obviously annoyed, muttered to himself, "These Israelis, why don't they ever apologize!" (Olshtain - Cohen1989:53)

In other instances, cross-cultural differences (and cross-genderdifferences - cf., Tannen 1982, 1986, 1990) may reflect the degree ofindirectness (cf., Brown - Levinson 1978). For example, when someone says "I'm hungry", it often means something more than a merestatement of fact. It can serve as a suggestion (let's go get something toeat); or it can serve as an exultation (hurry up and finish so we can eat);it can serve as a request for information (when will dinner be ready,I'm hungry). While it may be the case that all languages/cultures havethe means to express a suggestion, an exultation or a request forinformation, it is not necessarily the case that a statement of fact suchas "I'm hungry" will serve all of these functions. This book is dedicatedto the empirical study of a variety of speech acts in diverse culturalsettings and to the implications and applications of empirical speech actdata.

In this book we deal with three major areas of Speech Act research: 1)Methodological Issues, 2) Speech Acts in an L2, and 3) Applications. Inthe first section we deal with issues of methodology. As in any field andclearly in all areas of second language research, issues of methodology arecentral to an understanding of the phenomenon in question. A majorquestion is: to what extent can different methodologies contribute todifferential results? As Tarone - Gass - Cohen (1994: xiii) state: "Thevalidity of any discipline is predicated on the assumption that theresearch methods used to gather data are sufficiently understood andagreed upon." Speech act research is no exception, as all three chaptersin this section aptly illustrate. Human behavior and human interactionare complex phenomena and are subject to many intervening variables.Hence, any attempt to examine data and draw conclusions has to do sofully aware of the multi-faceted nature of the data.

Introduction 3

2. Section One

Leading off in this section, Cohen considers both theoretical and appliedaspects of speech act research. He notes that a first step in speech actresearch is a description of the sociocultural and sociolinguistic abilitiesneeded to produce a given speech act. A second step is the determinationof the research methodology. A third area that needs to be addressed(in those instances when the data are from non-native speakers) is theidentification of interlanguage features in the data.

Most relevant to this section is his discussion of research methodology.Many earlier discussions of various approaches to research methods havefocused on the advantages and/or disadvantages of one method overanother. Cohen argues against this approach and in favor of one thatcombines different research methods. He argues that research methodsplay different roles in the cycle of generating hypotheses, manipulatingvariables, determining the range of speech acts and validation. Forexample, ethnographic data are most important in generating initialhypotheses; they are also useful when dealing with some speech acts,particularly those that occur naturally in discourse. Ethnographic dataare less useful when investigating speech acts that do not occur frequentlyand/or that are so sensitive to sociocultural constraints that the constraining variables could not be controlled. Similarly, role-plays, writtentests, verbal report data are all relevant, but all come with their ownbaggage of advantages/disadvantages and appropriate and inappropriateuses.

In addition to his discussion of theoretical issues surrounding the useof various methods, Cohen considers more practical applications ofvarious methodologies drawn from his own data based on role plays andfollow-up interviews (using videos of the role play) of apologies,complaints and requests. He makes the important point that particularlywhen dealing with non-native speakers, as he was, it is crucial to separatea learner's adeptness with the situation from his/her adeptness with thelanguage. For example, it may be the case in a role-play situation, that thesituation itself is foreign to the subject, making the linguistic productionmore "unnatural" than it would be in a situation in which the learner feltcomfortable.

In his chapter, Cohen discusses some of the pros and cons of hisparticular methodology. In addition, he focuses on a number ofaspects of the retrospective comments of his subjects. He points out thatthrough these comments we are able to gain additional information

4 Susan M. Gass

on what learners are doing when confronted with producing speech actsin a second language. Among the issues discussed are the "din in thehead" phenomenon, self-debate, afterthoughts, formulaic speech,omission, avoidance and simplification. It is through verbal self-reportdata that we are able to learn about the options available to learners andabout the choices they make. However researchers select methodology,Cohen reminds us that it is through triangulation that a more comprehensive picture is able to emerge.

In Chapter Three, Houck and Gass respond to the fact that a significant amount of research into speech acts performed by non-nativespeakers uses discourse completion tests as a means of data elicitation.They point out the well-known limitations of this methodology. Primaryamong the limitations is the fact that the format used on discoursecompletion tests constrains the type and amount of talk. A secondapproach to the investigation of speech acts has been ethnographic innature. Within this framework, data are collected in naturally occurringsituations. This methodology, while alleviating certain problems thathave been apparent in discourse completion methodology, brings with itother problems, namely the difficulty in controlling contextual variablesand the unpredictability of the occurrence of a particular speech act.

In their chapter they consider speech act research from a methodological and substantive perspective. In particularly, they focus on thequestion of an adequate methodology for eliciting spoken speech acts andprovide detail on the ways in which research results may be dependent ondata collection procedures. The specific area of focus is refusals.Refusals are a highly complex speech act primarily because they mayinvolve lengthy negotiations as well as face-saving maneuvers. Becauserefusals normally function as second pair parts, they preclude extensiveplanning on the part of the refuser. Following the work of Beebe Takahashi - Uliss-Weltz (1990), they investigate refusals to 1) invitations, 2) suggestions, 3) offers and 4) requests. They depart fromprevious work on speech acts in two important ways: 1) by using videotaped data and 2) by basing their eliciting instrument on Scarcella'sconceptualization of socio-dramas (1978). Thus, the responses that aregiven are not confined by either the printed page (e. g., the amount ofspace provided on the page, the number of turns that the respondent isexpected to take) or by the closing response of the initiator of the interaction which, in many discourse completion tests, directs the refusal by"sandwiching" it between a given opening remark and the subsequentclosing comment.

Introduction 5

The data-base consists of English responses by native speakers ofJapanese to 8 situations designed to evoke refusals. The subjects of thestudy were given the contextual information surrounding each situation.Following this introduction, each subject role-played the part with anative speaker who had been instructed not to give up too easily in casesin which the non-native speaker initially refused. What resulted wereoften lengthy discussions in which each person negotiated his/her waythrough to a final resolution.

The analysis of the data focuses on such aspects of the discourse as 1)semantic tactics (sequencing and range), 2) turn length, 3) quantity andquality of negotiations needed to effect the refusal or to abandon theattempt to refuse, 4) amount of elaboration and repetition and 5) nonverbal elements such as laughter and pausing. Their data· reveal theexistence of a richer variety of semantic and pragmatic maneuvers thanhas been documented in previous literature. Not only is there a differencein maneuver types, but the methodology used allows for an analysis ofthe discourse structure given the extensive negotiation which takes place.Their chapter deals with substantive findings as well as the methodological implications of the differing results.

In a similar vein to that taken in the Houck and Gass article, Beebeand Cummings question the use of more traditional speech act elicitation(Discourse Completion Test) by comparing the results of speech act databased on two different methodologies. They are quick to point out, ashave the authors of the other two chapters in this section, that eachmethodology brings with it advantages and disadvantages in terms of theactual collection of data and in the analysis of those data. For example,naturalistic data or "notebook data" are valuable in that none of theartificial constraints of data collection are present, but the data thatresult are unconstrained in terms of the many variables that are knownto affect speech acts, such as status, socioeconomic background, age andso forth.

In a particularly ingenious design, Beebe and Cummings set out todirectly compare comparable refusal data from Discourse CompletionTests with those from naturally occurring recorded data. All subjectswere English as a Second Language teachers and were presented with thesame situation. The annual TESOL (Teachers of English to Speakers ofother Languages) conference was about to be held in New York City,where all of the teachers worked. Because it was anticipated that theconference turnout would be large, there was a need to get as manyvolunteers as possible to help with local arrangements. Each participant

6 Susan M. Gass

was given either a written version of the request or was called on thetelephone.

The analysis of the data centers around word-counts as well as semanticformulas. What they find is that the amount of talk is far greater inthe spoken refusals than in the written responses. Furthermore, in theoral data, there is a greater display of elaboration on an excuse (a typicalpart of a refusal). In face-to-face interactions, or, as in this case, telephoneinteractions, it is sociolinguistically inappropriate to flatly refuse withoutoffering some sort of excuse. The written data do not display elaborations in the same way as the oral data. Furthermore, the oral data allowfor learners to "negotiate" their way to the end of a refusal, rather than"packing" their responses into the first turn after the request.

The results are related to Wolfson's "bulge theory". The responseswere less typical of strangers (even though in the case of the telephoneconversations, they were strangers) since both the requester and the refuser shared a common profession and were members of the same professional organization.

Interestingly, while the oral data reveal longer, more repetitive, andmore elaborated responses, the content of the semantic formulas usedwas surprisingly similar (e. g., excuses, negative ability/willingness,apologies). Thus, the value of Discourse Completion Tests in speech actresearch can be validated. Nonetheless, Beebe and Cummings are quickto point out that they are not a substitute for naturally occurring data.

3. Section Two

The first chapter in Section Two deals with greetings. EisensteinEbsworth, Bodman and Carpenter point out the importance of greetings,both in terms of the sociocultural significance as well as their timing inmost language classes. Despite their deceptive simplicity, they arecomplex speech acts. Following the discussion in Cohen concerning theneed for multiple measures, Eisenstein Ebsworth, Bodman and Carpentercollect data on greetings in more than one way. They begin throughobservation of greetings by both native and non-native speakers ofEnglish in naturally occurring situation, noting the kinds of greetings thatoccurred in these situations. The observational data led to the creation ofa questionnaire to elicit data. As in the Beebe and Cummings chapter,they find that while the data elicited from the questionnaire were more

Introduction 7

limited, they were similar in many respects to the naturally occurringdata.

The data base for their study is rich both quantitatively andqualitatively. Their subject pool consisted of 50 native speakers ofAmerican English and 100 non-native speakers of American English. Thislatter group represented a wide range of native languages. All subjectscreated dialogues for pre-specified greeting situations. The non-nativespeakers created dialogues for the same situations in their nativelanguages. A second type of data came from (videotaped) role-plays ofthe same situations. Finally, a subset of the subjects participated in openended interviews following the role-plays.

In their chapter, Eisenstein Ebsworth, Bodman and Carpenterchallenge existing interpretations of greetings (in particular, the lack ofsincerity noted by Searle and others). They categorize and exemplifygreetings by native speakers into various types (greetings on the run,speedy greetings, long greetings, intimate greetings, all-business greetings,introductory greetings and re-greetings). Through a comparison of thedata gathered from non-native speakers in English with that of the samespeakers in their native language, many instances of native languageinfluences can be found. Not only can the native language influences benoted from word-by-word translations of greetings (e.g., incorrect/inappropriate use of titles, incorrect word choice, incorrect prosody), butalso from an understanding of the cultural norms or the context of thegreeting. Further, the authors discuss particular greeting types that areproblematic for learners and the resultant feelings and interpretationsthat come from the different cultural backgrounds of speakers. Theauthors also include in their discussion comments about pedagogicalissues and in particular make a plea for the inclusion of this complexspeech act in teaching materials.

Another speech act that indicates solidarity is that of compliments.This is the focus of Chapter Six by Nelson, El Bakary and Al Batal.Compliments vary considerably cross-culturally, not only in the wordschosen, but also in the context and frequency with which they areoffered. In fact, Nelson, El Bakary and Al Batal note that Egyptians areoften uncomfortable and embarrassed by the frequency with whichcompliments are given in English. As an example, they note the importance of the concept of the "evil eye" in Arabic, a concept that relatesto envy and potential harm coming to the individual whose person orproperty is the object of the compliment. The study reported on in thischapter considers American and Egyptian compliments focusing on the

8 Susan M. Gass

form of the compliment, the object of the compliment, the gender of thecompliment giver and the frequency with which compliments are given.The data come from 20 American university students and 20 Egyptianuniversity students. Each was interviewed and asked to tell the mostrecent compliment they had given, the most recent compliment that theyhad received and the most recent compliment that they had observed.

The resulting analysis revealed a number of differences includinglength of compliments (American compliments are considerably shorter)and the use of comparatives (Egyptians use more similes and metaphors).The syntactic patterns are limited in both American and Egyptiancompliments although the patterns are not identical. The analysis alsoconsiders the different attributes that are complimented and therelationship of the compliment giver in terms of gender to the attributesthey compliment. A discussion of the implications of this study forclassroom practices is included. Because the differences between the compliments of the two cultures studied are slight and because the differencesmay cause embarrassment and discomfort, it is crucial that the differences be presented to learners.

Chapter Seven, the third chapter in the section on solidarity, by Geisand Harlow, is concerned with the use of politeness strategies in a secondlanguage. As in many other chapters in this volume, the authors point outthe importance of learning sociocultural aspects of language along withlinguistic ones. Their chapter investigates the pragmatic conditions (withparticular focus on politeness conditions) affecting how requests andoffers are communicated in French and English with a view to formulating these conditions in such a way as to allow them to be taught explicitlyto learners of French. Geis and Harlow's proposals are based onexperimental determination of how native English and French speakersaccomplish requests and offers and how these are done by learners ofFrench.

Drawing from previous work by Geis and his colleagues, the authorsnote that information exchange in conversation occurs, not at the level ofliteral meaning, but at the level of "gist" (which consists primarily of theillocutionary force of an utterance). The form an utterance communicating a particular gist will take (i. e., syntax, morphology, etc.) is then determined by discourse context and by register, style, and politeness features.

The study consists of paired oral interactions - paired interactionsbetween native French speakers, paired interactions between nativeEnglish speakers, and paired interactions between English-speakinglearners of French. The 2 subjects in each experiment had to solve simple

Introduction 9

children's jigsaw puzzles for which each subject controlled pieces ofhis/her partner's puzzle, necessitating verbal interaction to secure neededpuzzle parts. The authors were able to elicit natural language use in theexperiments, and argue that the insights gained from this experimentalcontext are generalizable to natural speech contexts.

The data indicate that native speakers of French and English tend toframe requests somewhat differently, and that English-speaking learnersof French tend to fall somewhere in between, favoring pragmaticstrategies in their native language.

With a focus on politeness, Geis and Harlow present and defend atheory of the politeness features for the English language and Frenchlanguage cultures, show what the syntactic, morphological, and prosodicconsequences of these features are for the two languages and then discusshow non-native speakers might be taught to communicate requests bothas and when French speakers do.

Yet another means of maintaining solidarity is through apologies. Thisis the subject of Chapter Eight by Maeshiba, Yoshinaga, Kasper andRoss. Non-native speakers have been noted to produce second pair partsthat lack coherence and/or cohesion with preceding first pair parts. Whilelack of cohesion manifests itself in the choice of textually inappropriateutterance structures, failure to establish coherence may affect the propositional content, the illocutionary force, the politeness value of theresponding act, or a combination of the above (cf., Kasper 1984).

The realization patterns of the speech acts studied so far in interlanguage pragmatics, notably requests, apologies, complaints, refusals,compliments, and expressions of gratitude, have been shown to dependon such extralinguistic contextual factors as social distance anddominance, and on factors pertaining to the act itself, for instance thedegree of imposition or offense involved in the act (e. g., Brown Levinson 1978/1987, Blum-Kulka - House - Kasper 1989). While allof these factors will also determine the structure of responding acts, theproperties of the initiating act must be assumed to exert furtherconstraints on the function and form of the responding act.

Chapter Eight investigates responses to apologies, a speech act that hasbeen particularly well researched by scholars such as Cohen (Olshtain Cohen 1983), Olshtain (1983, 1989) and Holmes (1989). Specifically,the way chosen apology patterns condition the functional and formalproperties of the responses is examined. The empirical material for thischapter consists of responses to a Dialog Construction Questionnaire bytwo groups of Japanese learners of English (intermediate and advanced)

10 Susan M. Gass

compared to responses by native speakers of English and Japanese. Inaddition, information was gathered from the native speaker groups onthe likelihood of an apology being necessary and the likelihood of anapology being accepted.

Their results suggest that as a function of proficiency, learners are lesslikely to rely on their native speaker "guidelines" for transfer. However,when faced with situations or circumstances with which they have littleexperience, the advanced learners tend not to rely on their nativestrategies, thereby supporting contentions by Kellerman (1979) regardinglanguage distance and the use of first language forms and/or functions.

The study of apologies, because they are a remedial verbal action to athreatening (face or bodily) act serve as an appropriate bridge to thefollowing section which deals with face-threatening acts.

The first two chapters in this section by Murphy and Neu and byBoxer deal with complaints. Murphy and Neu have a double purpose totheir study. The first, is to determine how native speakers of English andKorean learners of English produce complaints and the second is tounderstand how native speakers judge the speech act set of complaints.Data were collected from native speakers of American English andKorean learners of English through an oral Discourse Completion Task.Subjects were asked to imagine themselves in a situation in which theyhad to complain about a grade to a professor. First, the non-nativespeakers of English were tape-recorded giving their response to thesituation. When these data were analyzed, Murphy and Neu discoveredthat most of the Korean learners of English were producing criticismsrather than complaints. To investigate the salience of these different typesor response, twenty-seven native speakers of American English wereasked to listen to a sample of the complaint and of the criticism speechact produced by the non-native speakers and evaluate the content of theresponse. The Americans judged the "criticizer" to be aggressive, disrespectful, and lacking credibility. The "complainer" was judged to berespectful, credible, and not aggressive. This perception, that Koreanlearners of English are placing the blame for a bad grade on the shouldersof their professor, is certain to complicate the academic lives of Koreannon-native speakers of English.

The finding that non-native speakers of English may construct aspeech act so incorrectly that it becomes another speech act intirely mayhelp us gain a clearer handle on the distinctive features of a speech act.

Boxer, in her chapter, uses ethnographic interviews as a means ofcorroborating data elicited through observation. She is concerned with

Introduction 11

indirect complaints as opposed to direct complaints. In her analysis,Boxer argues that while complaints may be viewed as face-threateningacts, indirect complaints may at times be a form of solidarity and mayinvolve rapport-building.

Boxer leads the reader through an open and frank discussion of ethnographic interview techniques and provides specific suggestions as to howinterviews of this sort may result in a productive use of researchers' andinformants' times. Her informants provided remarkably similar perceptions of the differences between direct and indirect complaints. However,gender differences did emerge when dealing with responses to indirectcomplaints: men tend to offer advice, women tend to commiserate. Inaddition, Boxer focuses on ethnic issues noting that within Jewishculture, complaining appears to be widespread, particularly indirect complaining.

As in other chapters in this volume, Boxer makes the important pointthat accurate descriptions and functions of speech acts are crucial asa basis for providing information (in the form of pedagogical intervention) to language learners. Boxer aptly shows that complaints are notnecessarily what they seem to be on the surface. They serve an importantsocial function. It is therefore crucial that non-native speakers learn whatthat social function is and how to interpret and respond to indirect complaints.

The next chapter by Goldschmidt investigates the variables that determine how people ask favors of each other. Goldschmidt further attemptsto ascertain if the asking of favors in American English is a strategydominated speech act that manifests the social structure of relationships.In particular, she addresses the metalanguage used in favor-asking sincethis speech act is often counter-intuitive, functioning either as a requestor as a directive. Asking a favor of someone is a potential impositiondepending crucially on the relationship of the asker and the asked.Variables such as gender, interpersonal relationships, age, status anddegree of imposition are all important to an understanding of how peoplerespond to favor-asking.

The data for Goldschmidt's chapter come from a survey administeredto 200 people, varying in age, status (student vs. non-student) and gender.The survey consisted of five favor situations. Subjects were asked to rateeach of these as to the degree of imposition involved. Her resultssuggest that imposition is perceived to be great in instances when familyprivacy is compromised and in situations in which a great deal of timeand/or effort may be involved. On the other hand, all participants

12 Susan M. Gass

(whether student or non-student, whether male or female and regardlessof age) similarly perceived the degree of imposition in the various situations.

A recurring theme in this volume has to do with the need to verify theform, the function and the constraining variables of speech acts as a prerequisite for dealing with them in the classroom. It is not enough toprovide practice on "complaining" in a language classroom if we do nothave appropriate and accurate baseline data on which to base ourdescriptions. Goldschmidt continues this line of argument by uncoveringthe rules, patterns, and strategies of favor-asking and by arguing that thisis essential before we can teach non-native speakers to perform in thetarget culture.

The final chapter in this section on face-threatening acts is by Koikewho investigates the speech act of suggesting by English speakers learningSpanish. In particular, Koike questions 1) the extent to which suggestionsare understood as suggestions by learners at different levels ofproficiency, 2) the potential for misunderstanding and the resultantpossibility of negative reaction toward the suggester, 3) the types ofresponses made to suggestions and 4) the degree of understanding of asuggestion as a function of proficiency level.

The data for Koike's study come from responses by three groups oflearners of Spanish (native speakers of English) ranging from beginningto advanced. Each subject was presented with a context and thenwatched a videotape of a native speaker making a suggestion. The taskwas to 1) respond to the suggestion as if the suggestion was beingaddressed toward them, 2) identify the type of speech act and 3) evaluatethe speakers on a variety of personal characteristics. In general, Koikefound that speakers did rely on native language speech act patterns ininterpreting second language speech acts. For example, when the form ofan L2 speech act was similar to the form of the same speech act in the L1,learners were more likely to understand the speech act. It was also notedthat misunderstandings frequently resulted in negative reactions.Interestingly, negative elements in the form of the suggestion often yieldeda negative interpretation. Responses were given to suggestions by manyof the students even in those instances in which misinterpretation hadoccurred. Even when confronted with a negative linguistic element orwith a misinterpretation, learners did not respond in a negative manner.The fact that responses were given and the fact that those responses werenot negative leads to speculation that through responses and continuednegotiation learners will eventually work out the intended meaning. This

Introduction 13

study provides evidence, then, of the considerable use of the nativelanguage in interpreting and responding to speech acts in a secondlanguage. This chapter leads into the following section, the first chapterof which also deals with suggestions, albeit suggestions of a differentsort.

4. Section Three

The last section of this volume deals with applications of speech acts. Inparticular, the chapter by Schmidt, Shimura, Wang and Jeong illustratesthe way suggestions are made in different cultures in terms of TVcommercials. Graham similarly takes us into the business world byconsidering international cooperative ventures.

Schmidt, Shimura, Wang and Jeong consider TV commercials withinthe category of suggestions - that is, suggestions to viewers to buy aparticular product. Their rich data base comes from four countries:United States, Japan, The People's Republic of China and South Korea,all of which view the purpose of commercials in a slightly differentway.

They found that suggestions were more frequent in American television commercials than in the other countries and that the preferredlinguistic mode of making a suggestion was the imperative. This was thecase more in American commercials than in the commercials of othercountries. Thus, American commercials tend to be more overtlysuggestive than those in the three Asian countries under investigation.What is interesting is the difference among the three Asian countries.Japanese and American advertising appear to be the most divergent withKorean and Chinese commercials somewhere in the middle.

By considering the speech act of suggestion in commercial settings,Schmidt, Shimura, Wang and Jeong propose that the language of advertising is a result of a number of factors, among them are universalpragmatic principles, cultural norms, market economy, and arbitraryconventions established by the advertising industry. Other factors (suchas the use of comparative advertising) are determined by governmentregulation. The cross-cultural study of commercials is thus a complexendeavor. As Schmidt, Shimura, Wang and Jeong point out, it is perhapsbest to view the TV commercial not as containing the pure speech act ofa suggestion, but as being a hybrid of requests and suggestions. This

14 Susan M. Gass

chapter clearly shows that the production of speech acts in someinstances takes the perspective of the hearer into account to an evengreater extent than the perspective of the speaker.

Graham, in the final chapter, looks at cultural differences in businessrelationships, considering in particular how cultural differences can causeserious difficulties between or among participants. Graham's specificfocus is the investigation of differences in cultural styles of businessnegotiations. His data base comes from videotaped simulated intracultural negotiations involving business people from thirteen countries.As in the study by Schmidt, Shimura, Wang and Jeong, Graham notesdifferences in the negotiation behavior among the Asian countries in hisstudy. Differences can be found in the amount of use of "no", "you",silent periods, interruptions and so forth.

As has been pointed out earlier in this chapter, baseline data areimportant before designing and implementing pedagogical programs.Similar comments are made by Graham although the context is different.Before being able to design a program for training business people on thedifferences in negotiation styles, one must first have accurate and detaileddescriptions of negotiations styles in the respective cultures.

5. Conclusion

Through data from a wide range of languages and through a wide rangeof speech acts, this volume has set out to describe the ways in whichspeech acts are similar and differ across languages and across cultures.While this goal has been accomplished, there are a number of contexts inwhich this work can be and needs to be extended. In particular, we hopethat the work presented herein will provide the impetus for pedagogicalmaterials and for training within different contexts.

Notes

1. In this introduction we do not deal with the issue of intentionality andits relationship to the speech act itself (cf., Apel 1991, DeMulder 1993,Habermas 1991, Leilich 1993 and Searle 1991). We refer the reader to theworks cited here for further elaboration on the theoretical underpinnings ofthis issue.

Introduction 15

References

Apel, Karl-Otto1991 "Is intentionality more basic than linguistic meaning?", in: Ernest

Lepore - Robert Van Gulick (eds.), 31-55.Austin, John

1962 How to do things with words. Oxford, England: Calderon Press.Bauman, Richard - Joel Sherzer (eds.)

1974 Explorations in the ethnography ofspeaking. New York: CambridgeUniversity Press.

Beebe, Leslie - Tomoko Takahashi - Robin Uliss-Weltz.1990 "Pragmatic transfer in ESL refusals", in: Robin Scarcella - Elaine'

Andersen - Stephen Krashen (eds.), 55-73.Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper (eds.)

1989 Cross-cultural pragmatics: Requests and apologies. Norwood, NJ:Ablex.

Brown, Penelope - Stephen Levinson1978 "Universals of language usage: Politeness phenomena", in: Esther

Goody (ed.), 56-324.Brown, Penelope - Stephen Levinson

1978/87 Politeness: Some universals in language usage. Cambridge:Cambridge University Press.

Cole, Peter - Jerry Morgan (eds.)1975 Syntax and Semantics 3: Speech Acts. New York: Academic Press.

Dechert, Hans - Manfred Raupach (eds.)1989 Transfer in production. Norwood, NJ: Ablex.

De Mulder, Walter1993 "Intentionality and meaning: A reaction to Leilich's 'intentionality,

speech acts and communicative action"', Pragmatics 3: 171180.

Gass, Susan - Larry Selinker (eds.)1983 Language transfer in language learning. Rowley, MA: Newbury

House.Goody, Esther (ed.)

1979 Questions and politeness. Cambridge: Cambridge University Press.Grandy, Richard E. - Richard Warner

1986 Philosophical grounds of rationality. Oxford: Clarendon Press.Grice, H. Paul

1957 "Meaning", Philosophical Review 66: 377-388.Grice, H. Paul

1975 "Logic and conversation", in: Peter Cole and Jerry Morgan (eds.),41-58.

Gumperz, John1982 Discourse strategies. Cambridge: Cambridge University Press.

Habermas,Jurgen1979 Communication and the evolution of society. Boston: Beacon

Press.

16 Susan M. Gass

lIabermas, Jurgen1991 "Comments on John Searle: 'Meaning, communication, and repre-

sentation''', in: Ernest Lepore - Robert Van Gulick (eds.), 17-29.lIolmes, Janet

1989 "Sex differences and apologies: One aspect of communicativecompetence", Applied Linguistics 10: 194-213.

IIymes, Dell1974 "Ways of speaking", in: Richard Bauman - Joel Sherzer (eds.),

433-451.Kasper, Gabriele

1984 "Pragmatic comprehension in learner-native speaker discourse",Language Learning 34: 1-20.

Leilich, Joachim1993 "Intentionality, speech acts and communicative action: A defense of

J. lIabermas' & K. o. Apel's criticism of Searle", Pragmatics 3:155-170.

Lepore, Ernest - Robert Van Gulick (eds.)1991 John Searle and his Critics. Oxford: Basil Blackwell.

Gumperz, John1982 Language and Social Identity. Cambridge: Cambridge University

Press.Ochs, Elinor - Bambi Schiefflin (eds.)

1979 Developmental Pragmatics. New York: Academic Press.Olshtain, Elite

1983 "Sociocultural competence and language transfer: The case ofapology", in: Susan Gass - Larry Selinker (eds.), 232-249.

Olshtain, Elite1989 "Apologies across languages", in: Shoshana Blum-Kulka - Juliane

lIouse - Gabriele Kasper (eds.), 155-173.Olshtain, Elite - Andrew Cohen

1983 "Apology: A speech act set", in: Nessa Wolfson - Elliot Judd (eds.),18-35.

Olshtain, Elite - Andrew Cohen1989 "Speech act behavior across languages", in: lIans Dechert

Manfred Raupach (eds), 53-67.Sadock, Jerrold

1974 Toward a linguistic theory of speech acts. New York: AcademicPress.

Scarcella, Robin1978 "Socio-drama for social interaction", TESOL Quarterly 12: 41-46.

Searle, John1969 Speech Acts. Cambridge: Cambridge University Press.

Searle, John1975 "Indirect speech acts", in: Peter Cole - Jerry Morgan (eds.), 59-82.

Searle, John1979 Expression and meaning: Studies in the theory of speech acts.

Cambridge: Cambridge University Press.

Introduction 17

Searle, John1983 Intentionality: An essay in the philosophy of mind. Cambridge:

Cambridge University Press.Searle, John

1986 "Meaning, communication and representation", In: Richard E.Grandy - Richard Warner (eds.), 209-226.

Searle, John1991 "Response: meaning, intentionality, and speech acts", In: Ernest

Lepore - Robert Van Gulick (eds.), 81-102.Tannen, Deborah

1982 "Ethnic style in male-female conversation", in: John Gumperz (ed.),217-231.

Tannen, Deborah1986 That's not what I meant!: How conversational style makes or breaks

relationships. New York: Ballantine Books.Tannen, Deborah

1990 You just don't understand. New York: Ballantine Books.Tarone, Elaine - Susan Gass - Andrew Cohen (eds.)

1994 Research methodology in second language acquisition. Hillsdale, NJ:Lawrence Erlbaum Associates.

Wolfson, Nessa - Elliot Judd (eds.)1983 Sociolinguistics and language acquisition. Rowley, MA: Newbury.

Part IMethodological Issues

Investigating the production of speech act sets 1

Andrew Cohen

This chapter will discuss both theoretical and applied issues regarding theresearching of speech acts, drawing in part from a recent research effortto describe the processes involved in producing speech act utterances(Cohen - Olshtain 1993). The chapter will end with some illustrativefindings from that study.

1. Theoretical issues

Speech acts have been investigated and described from a variety ofperspectives: Philosophical, social, linguistic and cultural. An effort hasbeen made to identify universal norms of speech behavior and to distinguish these from language-specific norms in order to better understandand evaluate interlanguage behavior.

Given a speech act such as apologizing, requesting, complimenting, orcomplaining, the first concern of the researcher is to arrive at the set ofpotentially universal realization patterns, anyone of which would berecognized as the speech act in question, when uttered in the appropriatecontext. We have referred to this set of strategies as the speech act set ofthe specific speech act (Olshtain - Cohen 1983). In order to arrive at thespeech act set, it is necessary to define the goals of the speech act inquestion and to identify performative and semantic prerequisites for therealization of these goals.

As an example, consider the speech act of requesting. It was necessaryto present a scale of impositives (i. e., degree of imposition; Olshtain Blum-Kulka 1984; Blum-Kulka 1989; Weizman 1989), moving from themost direct to the most indirect request. For apologies, it was necessaryto separate the performative verbs from other semantic preconditionsthat could result in acceptable apology realizations, such as an explanation and justification for the offense, an offer of repair, and so forth.

22 Andrew Cohen

Hence, each speech act presents its unique set of preconditions and interactional goals which have to be addressed in the realization patterns thatcan act as the materialization of the speech act.

As of the 1980s it became clear that in order to adequately define anddescribe such speech act sets, considerable empirical investigation bothwithin and across languages would be needed. Fortunately, over the lastdecade there has been a wide range of empirical studies on speech actbehavior. One of the most comprehensive empirical studies of speech actbehavior, both for its breadth and depth, has been that of the CrossCultural Speech Act Research Project (CCSARP) (Blum-Kulka - House- Kasper 1989), which compared speech act behavior of native speakersof a number of different languages with the behavior of learners of thoselanguages. The CCSARP project also produced useful instruments fordata collection and a coding scheme that has been widely replicated inother speech act studies. Along with the empirical studies, severalexcellent surveys of the research literature have appeared which help todefine and shape the field of investigation with respect to speech actresearch (e.g., Wolfson 1989; Kasper - Dahl 1991).

In this first section, let us consider three theoretical areas of concernwith regard to speech act research. The first concerns the description ofthe sociocultural and sociolinguistic abilities needed to perform a givenspeech act. The second concerns the selection of research method for usein gathering the speech act data. And in cases where the respondents arenon-natives, there is also a need to deal with the interlanguage featurespresent in the speech act data.

1.1. Sociocultural and sociolinguistic abilities

What has emerged both from the large-scale empirical studies and fromthe comprehensive reviews of the literature is that successful planningand production of speech act utterances depend on the sociocultural andsociolinguistic abilities of the speaker. Speakers and hearers are successfulspeech act users when they have control over the speech act sets for agiven speech act in the language in which they converse. Such controlcalls for the ability to provide both socioculturally and sociolinguisticallyappropriate behavior.

Sociocultural ability refers to the respondents' skill at selecting speechact strategies which are appropriate given (1) the culture involved, (2) theage and sex of the speakers, (3) their social class and occupations, and (4)


their roles and status in the interaction. For example, in some cultures(such as in the United States) it may be appropriate for speakers to use arepair strategy by suggesting to the boss when to reschedule a meetingthat they had missed through their own negligence; however, in othercultures (such as Israel), a repair strategy might be considered out of placein that it would most likely be the boss who determines what happensnext. Thus, the sociocultural ability is what determines whether a speechact set is used and which members of the set are selected for use.

Sociolinguistic ability refers to the respondents' skill at selectingappropriate linguistic forms to express the particular strategy used torealize the speech act (e.g., expression of regret in an apology, registration of a grievance in a complaint, specification of the objective of arequest, or the refusal of an invitation). Sociolinguistic ability is thespeakers' control over the actual language forms used to realize thespeech act (e. g., "sorry" vs. "excuse me", "really sorry" vs. "verysorry"), as well as their control over register or formality of the utterancefrom most intimate to most formal language. For example, when studentsare asked to dinner by their professor and they cannot make it, the reply"No way!" would be a phrase for use with an appropriate semantic formula, namely, refusal. The problem is that sociolinguistically, this phrasewould constitute an inappropriate refusal, unless the students had anespecially close relationship with their professor and the utterance weremade in jest.

1.2. Research methodology relating to speech act description andstrategy selection

The complexity of speech act realization and of strategy selection requirescareful development of research methodology in this area. Rather thanchoosing between ethnographic and elicited data methods, the combiningof different approaches to studying the same speech act may best enablethe researcher to reach useful and reliable descriptions of speech actbehavior. An ideal cycle of data collection could be perceived of asfollowing the different collection techniques presented in Olshtain andBlum-Kulka (1985).

The researchers would start with the generation of initial hypothesesbased on ethnographic data collection of natural speech. Then theywould continue to simulate speech such as role-plays which can serve totest the initial hypotheses. From there, they could go to a paper-and-

24 Andrew Cohen

pencil written completion test in order to focus on specific realizationsand manipulate the social and situational variables. If they are concernedwith the effect of the speech act on the listener, they might want to useacceptability tests in order to validate the range of acceptability within aspeech community. Finally, it is advisable to validate findings by means offurther ethnographic data.

~ ethnographicacceptability~ -----.... role-play

written completion .-----------

Each of these data collection techniques has its own merits but it is theuse of more than one that provides us with important triangulation.

Ethnographic observation involves the collecting of naturallyoccurring data. This method has proven effective in collecting data oncertain speech acts, such as compliments (see, for example, Wolfson1989). Yet for other speech acts, such as apologizing, it may be extremelytime-consuming and not very productive. Aguilar Murillo, Aguilar, andMeditz (1991), for example, found that even when they planted someonecrouched behind a door and videotaped the door being opened and theperson getting whacked by the door, the apology events were limited andthe data themselves not very useful.

When comparing native and non-native apologies, complaints orother complex speech acts across a variety of situations, it would beexceedingly time-consuming to gather natural data in all the desiredcategories. It would also be virtually impossible to control all thevariables that role-play and written completion tasks can build into theirdesign - e. g., severity of the offense, familiarity/age/relative status/sex ofinterlocutors, and so forth.

If a role-play situation is acted out theatrically (e. g., the respondentbumps into another person), this would constitute genuine role-play.Another, more popular format is that of a role-play interview in which therespondents are requested to respond as they would in the given situation,without acting it out. These two versions of role-play have been referredto as semi-ethnographic in that they require the participants to take onroles that are not always their own (Olshtain - Blum-Kulka 1985). If theinterlocutor involved is not aware that the event has been contrived for thepurposes of collecting data or is aware and agrees to cooperate (e.g., theactual owner of a store hearing a complaint about merchandise from asubject in a study), then the situation would be called real-play.


Such role-play would most likely be audio- or video-taped. The tapingitself may introduce problems, depending on how intrusively it is done.Even if the taping is relatively unobtrusive, it may still make somerespondents uncomfortable, at least for the first few minutes. Such tapingmay even engender certain reactive effects; Stubbs (1983: 225) hassuggested that respondents might develop special verbal strategies fordealing with tape-recorders.

The role-play may consist of a description of the situation, written inthe native or target language and/or read aloud, a prompt by the interlocutor (depending on the situation), and then the response. It can also bespecified that the interlocutor is to provide one or more rejoinders, toturn the role-play into an interactive event.

There are two options for a written completion task. In both cases, asituation is briefly described in writing, either in the target language or inthe native language. In the first type, that of open-ended elicitations, thereis a written prompt followed by a space for the respondent to provide awritten response. The second option is for what has been referred to asthe Discourse Completion Test (DCT) (Blum-Kulka 1982), whereby thediscourse is structured - part of it left open and part closed, providingboth for the speech act and a rejoinder. In fact, the rejoinder helps to cuethe respondent as to the appropriate nature of the speech act realization - i. e., the level of formality, and a description of the roles andrelationships of the interlocutors.

The written approaches save enumerable time in data collection andhave been seen to provide reasonable projective measures of the samekinds of data collected from oral role-play (Beebe - Cummings, thisvolume; Cohen - Olshtain - Rosenstein 1986). In comparing talk overthe telephone to written questionnaire responses, Beebe and Cummings(this volume) found there was four times as much spoken output as thanwritten. All the same, the results indicated that discourse completion testsare an effective means of gathering a large amount of data quickly,creating an initial classification of semantic formulas, and ascertainingthe structure of speech act(s) under consideration.

A problem with written responses is that certain kinds of informationare not collected this way, such as the prosodic and nonverbal features oforal interaction. Furthermore, the respondent usually has more time torespond when doing so in writing than when doing so orally. Also, thevery act of responding in writing as if speaking may inhibit therespondent, producing a shorter response than would be the case inspeaking. One advantage of the discourse completion test over the open-

26 Andrew Cohen

ended format is that the former indicates the expected length of theutterances while the latter does not.

Acceptability ratings as another means of testing for control ofspeech act behavior involves the obtaining of respondents' judgments asto how appropriate certain responses are for a given situation (Olshtain- Blum-Kulka 1985). In this technique, a series of possible responsesare presented and the respondent has to select the most appropriateof them for the given situation. Usually the responses are scaled on acontinuum according to some dimension. In the case of apology research,the scale could go from the least intensified to the most intensifiedapology.

In addition to the above-mentioned techniques which are useful for thedescription of speech act behavior within a group, we can use researchtechniques such as verbal report to give us insights regarding the choicesmade by individuals in their speech behavior. By now it has been clearlydemonstrated that verbal report is not one measure, but rather encompasses a variety of measures, intended to provide mentalistic, dataregarding cognitive processing (Afflerbach - Johnston 1984; Olson Duffy - Mack 1984; Faerch - Kasper 1987). Such verbal reports include data that reflect self-report (learners' descriptions of what they do,characterized by generalized statements about learning behavior), selfobservation (the inspection of specific, not generalized language behaviorintrospectively or retrospectively), self-revelation (think-aloud, stream-ofconsciousness disclosure of thought processes while the information isbeing attended to), or some combination of these (Cohen - Hosenfeld1981; Cohen 1987).

Given the intrusive nature of verbal report techniques, it would beunreasonable to ask speakers to provide such data while they are engagedin the communicative act.2 Yet once the interaction is over, subjects maynot be able to retrospect fully as to the strategy selection that they carriedout a few minutes prior to the intervention. For this reason, in the Cohenand Olshtain study (1993), subjects were videotaped interacting in roleplay situations and then viewed the videotapes (one or more times) as ameans of jogging their memory as to their thought processes during theinteractions. Some illustrative findings from the study will be presentedlater in the chapter.


1.3. The study of speech act interlanguage

In second-language acquisition research, there is a concern for the way inwhich learners learn and produce speech acts as part of the sociolinguisticcomponent of their communicative competence. It has been established inprevious studies that in speech act behavior, as in other language areas,there is a discrepancy between a learner's receptive and productiveabilities. Thus, in a study done with immigrants in Israel, it was foundthat while it might take as long as eight years to acquire native-like reception of speech acts, one may never truly acquire native-like production(Olshtain - Blum-Kulka 1985).

When dealing with the production of speech acts the immediateproblem is the evaluation of interlanguage speech act behavior. Thequestions that could be asked with regard to interlanguage featuresinclude the following:

a. To what extent have learners acquired the sociocultural and sociolinguistic abilities needed to realize the particular speech act?

b. To what extent is the learner's speech act behavior similar to ordifferent from a native speaker's behavior under the same circumstances?

c. What compensation strategies do learners use when their language isinadequate?

d. What is the learners' selection route and decision making process withrespect to strategy preference, modification preference, content limitation, and illocutionary intent?

While the first two questions relate to the evaluation of product, the lasttwo are concerned with process and require introspection and reflection.There is a need to probe the actual decision-making and selection processthat learners at different levels of proficiency go through in order toidentify strategies that lead to the successful production of speech acts inthe target language.

2. Research design issues

Now that we have discussed some basic theoretical issues, let us look atspecific issues in research design. For the most part, the discussion will be

28 Andrew Cohen

based on issues that came up in the design and execution of the Cohenand Olshtain (1993) study of speech act production.

The study sought to describe ways in which non-native speakers planand execute speech act utterances, and the relationship between choiceof processing strategies and successful execution of the utterance. Thesubjects were fifteen advanced English foreign-language learners, allHebrew University undergraduates, eleven of whom were native speakersof Hebrew and the remaining four advanced non-natives - nativespeakers of French, Spanish, Portuguese, and Arabic.

The subjects were given six speech act situations (two apologies, twocomplaints, and two requests) in which they were to role play alongwith a native speaker. The interactions were videotaped, and after eachset of two situations of the same type, the tape was played back and therespondents were asked both fixed and probing questions regarding thefactors contributing to the production of their response to that situation.The subjects were interviewed in three sessions - after the apology, complaint, and request situations respectively - instead of waiting untilafter all six speech act situations, in order to obtain a more accurateretrospective report of behavior. It was feared that the delaying of theverbal report would reduce the reliability of the protocols, even using thevideotaped behavior as a memory aid.

2.1. The role-play interview as a research tool

The first issue to consider is the use of a semi-oral, role-play interview(i. e., written situation and then role-play) as a simulation of actualbehavior. The question is whether such an elicitation technique is reallysemi-ethnographic, as suggested in Olshtain and Blum-Kulka (1985).What is the effect of having respondents take on a role they would notassume in real life? In some instances in the Cohen and Olshtain study(1993), respondents remarked that a given situation happened to them allthe time. In several cases, the respondents commented that they hadperformed that speech act the previous day - e. g., requesting a neighborto turn down loud music late at night. In other cases, respondents madeit clear that it never happened to them.

In instances where the respondent had never had to react in such asituation (e.g., apologizing for keeping a classmate's book two weeksbeyond the agreed date), it could be argued that the instrument forcesunnatural behavior and that if the respondent were not a good actor, the


results might be problematic. The researcher's task would be todistinguish respondents' language proficiency from their situationaladeptness. In the research study under discussion, the respondents werenot given the choice to opt out of the speech act. If they deflected thestimulus, the interlocutor pursued the issue. This is not necessarily thecase in the real world, where a person may opt not to apologize, complain, or request something (Bonikowska 1988).

Another effect of the situation might be the degree of planning itactivates in the speaker. In other words, the situation itself may haveproperties that stimulate planning more than do other situations, regardless of the personal characteristics of the speaker. So, for example, if therespondents feel that they are in the right, as in a complaint situation,they may not plan as much as if they feel the need to, say, make amends,as in an apology situation. This observation was made by several of therespondents in the Cohen and Olshtain study.

Furthermore, an assumption was made in the Cohen and Olshtainstudy that a sampling of three speech acts (apology, complaint, request)in six situations could give a fair idea of how non-natives prepare andexecute utterances. It is possible that this was too small a sampling ofspeech act production behavior through role-play. The study alsorevealed that the speech act behavior was conditioned by the nature ofthe situation. For example, a student's asking his/her teacher for a lifthome - where the inequality of status was found to play an importantrole in the mind of the respondent - usually prompted a style shift, atleast after the interlocutor, playing the role of the teacher, replied,"What?" in response to the student's initial request.

In addition, whereas an effort was made to select situations that werecross-culturally appropriate - i. e., that had the same cultural weight indifferent cultures (such as a neighbor playing loud music late at night), itis possible that one or another of the situations could still have beenviewed by a respondent as not constituting an infraction. For example,the situation of "being half an hour late to meet a friend to study for anexam" may not be considered a serious offense in Latin America, and, infact, one of the respondents was originally from Argentina. In this case,however, the respondent had lived most of her life in Israel.

Finally, the situations were written in the foreign language, English,which thus provided clues for how to respond - for those respondentswho picked up on this. From time to time respondents lifted languageforms out of the text which described the situation - language formsthat were only partially or not at all in their productive knowledge. For

30 Andrew Cohen

example, in the "lift" situation, a respondent named Hava noted that shelifted "my bus has just left" out of the text. Also, whereas she wouldsimply say "token," she requested a "phone token" in the "token"situation because that was written in the text. Wassim also indicatedtaking the expression "phone token" from the text. In that same situation, Yaakov said he had used the word "urgent" because the wordappeared in the description of the situation - that he would not haveused it otherwise. Likewise, Shlomit said she also used "urgent" because"it was included in the situation." Finally, there was an instance of therespondent's combining his own material with that contained in the text.So, in the "lift" situation, Yaakov described how he arrived at askingDebbie, "Can I come by your car?":

First I thought "with your car, with you" and that I would not mention thecar because I didn't know how to indicate hamixonit she/ax 'your car.' Iworried that she would think I wanted to go for a ride with her. "To get aride with you" would be an expression I wouldn't know how to use. "CanI come" are words that I know how to use. After I heard Debbie read "bycar," I said "by your car."

The reverse was also true. There were numerous cases where respondentsdid not make use of clues that were in the written descriptions of thesituation. For example, in the situation calling for a request from theteacher for a "lift" home, there were respondents who disregarded thisclue and had difficulties finding a word in English for this request. TheSemi-Direct Oral Proficiency Interview (SOPI) (Stansfield - Kenyon Paiva - Doyle - DIsh - Cowles 1990), for example, gives the promptsin English L1 rather than in the foreign language being assessed. This wayno clues are given concerning the response.

2.2. The role-play interview: Data collection issues

There are variables operating in the collection of role-play interview datathat can have considerable bearing on the reliability and validity of theresults. We will use the Cohen and Olshtain study (1993) to help illustrate these variables, the choices that were made, and the possible orprobable effects of these choices.

In administering the role-play interview in that study, it was decidedthat the interlocutor would give the respondents an opportunity to readthe descriptions of two brief role-play situations at a time (two apologies,two complaints, and two requests in all). Then she slowly read each situ-


ation out loud to the respondents, giving them time to think of aresponse, and then gave her opener and had the respondent role-playwith her. The interaction was videotaped and audiotaped as well.

The native English-speaking interlocutor determined whether theinteraction had reached its natural and logical end - usually after fouror five exchanges. Thus, in this case, flexible structuring of the role-playwas used. In a previous study that had been recorded only on audiotape,the structure had been fixed - i. e., an opener was followed by a singleresponse from the respondent (Cohen - Olshtain 1981). It was felt thatthis earlier approach put limitations on the depth and breadth of the dataavailable for analysis purposes. However, a problem that arose in the useof flexibly structured role-play was that the interlocutor was not necessarily consistent from one respondent to the next. During the pilot sessionsand in one or two instances at the beginning of the data collection, therole-play interlocutor was perhaps too easy on the respondent. Forexample, when the respondent was slow at making the request for a rideexplicit, the interlocutor offered, "Yeah, do you want a lift?" At othertimes, the interlocutor was perhaps too tough on the respondents. Theywould apologize, for instance, and she would not accept their apology.Perhaps it could be argued that in one or two of these cases a nativespeaker in a natural setting would accept the apology more readily.

The probing interviews conducted in this study were designed toobtain retrospective self-observational data about the cognitive processesinvolved in the production of speech act realizations. The interviewer'sprobes were conducted in what was the native language for eleven of therespondents, and a language of greater proficiency than English for theother four respondents. Effort was made to have the respondents beprecise, and to give examples where possible. When the respondents werenot sure as to what they did and why, the interviewers played the relevantportion of the videotaped session a second or even a third time. Thisusually helped to jog the respondent's memory. In working with verbalreport, there is always the danger that if the interviewers suggest toomuch, the respondents may fabricate inaccurate descriptions of what theydid to produce utterances. Another problem associated with the power ofsuggestion in verbal report is that continued mention of a particularbehavior may do more than simply heighten awareness regarding it. Suchmention may actually cause the behavior to take place.

It is possible, for example, that when respondents were asked toindicate the language in which they were thinking, it may have stimulatedthem - especially the trilingual ones - to think in a language in which

32 Andrew Cohen

they had not been thinking. The question is the extent to which an effortto heighten awareness about behavior that is taking place but is notattended to, may inadvertently trigger that behavior. The Cohen andOlshtain (1993) study seemed to reveal the three most common patternsfor language of thought in planning and responding to be "planning inEnglish and responding in English," "planning in Hebrew and translatingfrom Hebrew to English in the response," and "planning in Hebrew withthe response in English."

Furthermore, whereas the French, Portuguese, and Arabic speakersreported that they tended to think in Hebrew rather than in their nativelanguage, they indicated that they thought in their native language in oneor two situations: The French speaker for planning and producing hisrequest to his teacher for a lift home, the Portuguese speaker for planningan apology after forgetting to return a book to a classmate and a complaint after a peer's refusal to let her use her notes, and the Arabic speakerfor planning in the same "notes" situation. In the case of the Spanishspeaker whose English was weak, the patterns were reported to be mostcomplex, involving planning in Hebrew and Spanish simultaneously or ina staggered fashion, and then translating from Spanish and/or Hebrew toEnglish. In reporting these cases of language choice, the researcher mustbear in mind the possible intrusive effects of the verbal report methodhere.

There is always the danger that if interviewers make leading suggestions in their efforts to elicit verbal report, the respondents mayfabricate inaccurate descriptions of what they did to produce utterances(Cohen 1991). In addition, there is the possibility that the interviewermight make false assumptions based on intuitions regarding the verbalreport and might put words in the respondent's mouth, as in the following case from Cohen and Olshtain (1993): "I could see you were focusingon grammar." In this instance, the informant indicated that he was notdoing so. On the plus side, verbal report interviews provide feedbackfrom respondents regarding aspects of their behavior that would otherwise be left to the intuitions and speculations of the investigator.

Then there is still the issue as to the relationship between the reportedbehavior and actual behavior. The use of immediate retrospection (immediate playback of the tape after two situations) was intended to diminishthe likelihood of the retrospections being fabricated, but the possibilitystill exists. No effort was made to investigate the relationship between therespondents' report of planning and actual evidence of their planning(e.g., pauses in delivery).


In this study the respondent informants were not trained in givingverbal report nor in the dynamics of speech act production. They werethus naive informants - at least at the outset. It is likely that some, if notmany of them, became more aware of the phenomena being investigatedas they progressed from the first response session to the second and thethird. Hence, it is possible that our data were impaired by a lack oftraining. Had we taken the measures of speech act production on whichwe wanted verbal report and trained the respondents to pay particularattention to them, perhaps the results would have been more informative.Of course, there is then the risk that the training itself will implant certainnotions about "appropriate" behavior in the heads of the respondentssuch that they no longer behave the way that they would have.

2.3. The use of multiple measures of speech act production

In the field of language assessment, there is a current emphasis on themulti-method approach. The attitude is that anyone method would notbe assessing the entirety of the behavior in question. Would this also betrue with respect to determining speech act production behavior in oralcommunication? The ethnographic approach would be difficult toemploy, unless respondents were somehow to record their productionstrategies (e.g., in a journal) just after performing speech acts. Likewise,a written discourse completion task would be at best a projective measureof speaking.3 It would be possible to gather acceptability data, both fromnon-native peers and from natives. Such ratings could help to determinethe extent to which the speech act utterances themselves are appropriatefOf the given situations.

3. Search, retrieval, and selection of language forms

Let us now take a look at some of the data produced by respondentsin the Cohen and Olshtain study, bearing in mind the methodologicalproblems raised regarding the collection of such data. We will considercommunication strategies and the concerns that one or more speakersreported in searching for, retrieving, or selecting language forms to use intheir speech act utterances. These examples represent all the instances

34 Andrew Cohen

that were identified in the analysis of the verbal protocols for these fifteenspeakers. Eight of the categories reflect areas that have been much discussed in the communication strategy literature: din in the head, monitor,use of formulaic speech, message omission or abandonment, lexicalavoidance or simplification, and approximation. The other four categories reflect insights gained from the use of verbal report protocols: Selfdebate, afterthoughts, partial delivery of a thought, and delivery of adifferent thought.

3.1. Retrieval process - "din in the head"

Ricki noted after completing the first two situations that she had difficultyin speaking English because of a long period of non-use: "When I startspeaking English after not speaking it for a long time, my vocabulary isweak and it is hard to retrieve words from memory." Krashen (1985:40-41) has called attention to the "din in the head" phenomenon whereby the "din," or sense of having the language available for use, may takeanywhere from one to two hours of good input and may wear off after afew days. In certain oral elicitation tasks, there may be a warm-up period,but often this period is not long enough to activate the din in the head.

3.2. Self-debate before selection

In the "lift" situation, Hava debated between "to get a ride" and "to givea lift," and finally asked whether she "could get a lift." Shalom debatedamong "drive," "come," and "go," .and ended up with, "Can I come withyou?" Galit wanted to make a polite request and was uncertain as towhether she could ask, "Do you have any room in the car?" As she putit: "It has a lot of meanings and I wasn't sure that it was correct, so Ichanged my tactic, and decided she would understand better if I said, 'Iwant to drive with you.' I thought of 'lift,' but didn't know how to useit in a sentence so I left it out." In the same situation, Lily debatedamong three expressions, "in the same neighborhood/your sameneighborhood/in your neighborhood." She was translating from Spanishand felt that the result was not good. Also with regard to the "lift" situation, Yaakov debated how to address Debbie - "Debbie," "Teacher,""Gveret ('lady')," or "Gveret Teacher." He decided to address her theway he would in a high school class in Israel.


3.3. Afterthoughts

In the "meeting" situation, Ricki used "very" as the intensifier in herexpression of apology, "very sorry," but reported thinking to herselfafterwards that she could have said "terribly sorry." She also used"stopped" in that situation ("I'm very sorry, but I-I met some friendsand they stopped me and I couldn't go on ... ") and, as she put it, "I knewit wasn't the correct word but I was already in the middle of things."Sometimes the afterthoughts respondents have during a given speakingtask can, in fact, cause later communicative failure in that their minds stillengaged in some previous language form while they are being called uponto perform a new task.

3.4. Awareness of using the monitor

Four of the respondents referred to their use or nonuse of monitoring.With regard to the "meeting" situation, Lily commented, "I always thinkabout grammar and so my pace is so slow. I think about how to structurethe sentence correctly, verb tenses and other aspects. E. g., 'I haven't sleepgood' ~ 'I didn't sleep good.' I thought the first form wasn't correct." Inthe "music" situation, Lily erroneously said, "you have listened to themusic very loud last night" and noted, "With this confusion, I wonderedwhether to continue with the mistake or correct myself. I decided that itwas important to correct myself because if I am aware of an error and itis possible to correct it, I want to do it." Ricki could also be viewedas a consistent monitor user. With respect to the "music" situation, shecommented, "I am always thinking about grammar ... When I haveproblems like 'not/don't,' I correct them. 'I was yesterday awake -' justcame out that way and I noted that it was not correct."

Hagar on the other hand would be viewed as an underuser of themonitor. With regard to the same situation, she remarked, "I don't effortat grammar. I am aware that it is bad. I focus on the idea, the message.Grammar gets me stuck. I prefer not to know how grarpmatical I sound.I depend on the listeners to see if they understand me, using facial expressions and letting them complete my sentences for me." Wassim onlythought about grammar extensively in the "notes" situation in which itwas not spontaneous in that he was translating from Arabic. In the"meeting" and the "book" situations, he reported: "When I first read thesituations, I thought that it would be good to think about my grammar,

36 Andrew Cohen

but I then forgot about it because it was more important for me thatDebbie understand me."

3.5. Use of formulaic speech

In the "lift" situation, Nogah used "I would love to-" in requesting aride, which sounded peculiar for the requesting party to use. Nogahnoted that she had heard this expression a lot and that is why it poppedup in her utterance. Although this was the only reported instance of anunanalyzed phrase appearing in the respondent's data, it is likely thatsuch formulaic speech occurs with some regularity in the output of nonnatives (Ellis 1985).

3.6. Omission, avoidance, or simplification

There were also examples of respondents not saying what was intendedfor lack of the appropriate forms or lack of certainty about them.

3.6.1. Omission

Two cases of omission of an utterance occurred in the data. In the"meeting" situation, Lily thought of saying that she was late because ofa problem at home, but decided that it would be too difficult for her tosay it in English. Instead she chose to say that she usually comes late. Shealso indicated that in general she chooses the easiest utterance - the onefor which she knows the verbs and the sentence structure, and can say itdirectly "without having to express it in a round-about way." In the "lift"situation, Shlomit debated whether she should address her teacher byname, and then chose instead to say, "Excuse me, are you going home?"because, as she put it, "it was a bit more formal - in general, when Iaddress a lecturer in Hebrew, I do it this way."

3.6.2. Abandoning a word or expression

Five instances of breakdown were identified in the data. In the "meeting"situation, Galit said, "I really don't have any exc-" and stopped there. She


said she got stuck because of the x. In the "book" situation, Shalomasked, "Anything I can do to comp - something?" He said that he sortof knew the word "compensate" receptively. In the "music" situation,Hagar started the utterance, "Can't you just -" and stopped. She felt thatwhat she was starting to say was inappropriate and did not know how toconvey the correct message in English. In the same situation, Lilyproduced, "I want you to - that -" and, in explanation, noted, "Iwanted to say that I didn't want that to happen again but stopped in themiddle because it was too complicated for me." In the "notes" situation,Nogah wanted to indicate that she always gave her friend class notes ifshe wanted them, but did not know how to say it: "I debated between'often' and 'always' and I couldn't remember it, so I let it go." She simplysaid, "When you need things I al - I give you" and made no furtherattempt to supply the adverb.

3.6.3. Partial delivery of a thought

Two instances of partial delivery of an utterance were identified.In the "notes" situation, Hagar was not sure whether she should justcontinue requesting the notes or whether she should simply saythat she did not need any favors from her friend and thank heranyway. She chose to be angry but commented that "anger doesn'tcome out well in English." As she put it, "I started and got stuckbecause of my English and so I chose a compromise." Her compromisewas to be sarcastic: "Well, you're very kind to me. I mean I gaveyou in the past things and it's - uhm - alright, no thank you." In thesame situation, Nogah wanted to use strong language but did not knowhow to say it in English in a way that would not sound too exaggerated,so instead of saying the English equivalent of tov lada'at 'it's good toknow' or ani ezkor 'et ze 'I'll remember this,' she simply said, "I needthem too."

3.6.4. Delivery of a different thought

There were two examples found of a different thought being delivered.4 Inthe "meeting" situation, Hava wanted to indicate that the bus did notcome, but she reported that she did not find the words in English, soinstead she said, "I missed the bus." Galit, in looking for a reason that

38 Andrew Cohen

she needed a ride, said, "My bus is very late," which she saw right awayto be incorrect. As she explained it, "I meant that it wouldn't be leavinguntil later in the evening, but grammatically the sentence was OK so I leftit. I let it go because it wasn't so bad - she would understand what Imeant."

3.6.5. Lexical avoidance or simplification

There was one identifiable instance of lexical avoidance and one ofsimplification in the data. In the "music" situation, Shlomit wanted tosay that her neighbor's music was "too loud" but avoided the equivalentEnglish forms by saying, "Your music is - uhm - and I can't sleep withyour music." In the "notes" situation, Yaakov simplified his utterance inorder to execute it, "I really don't like - this." He explained as follows:"I searched for something else like, "the way you act/your behavior," butit didn't come to mind when I was answering. I used the easiest way outat the moment."

3.6.6. Approximation

In five instances the word search ended in an approximation as thespeaker felt or knew the word was incorrect but could not come up withan alternative. In the "book" situation, Jackie was looking for a word toindicate repair but did not find it. He said, "I'm shocked, I'm sorry,"but he was looking for lefatsot 'to compensate' and, in his words,"had a blackout." Also in the "book" situation, Galit wanted to say theEnglish equivalent of xomer 'material,' and could not find a wordlike "notebook," so she said "stuff": "I didn't find the - stuff." Inthe "music" situation, she asked the neighbor to "reduce" the volume.Her retrospective comment was as follows: "I had my doubts aboutthe word 'reduce'; it seemed like a literary word to me." When it wasnoted that the interlocutor (Debbie) had in fact supplied the phrasewhen she said, "I would have turned it down," Galit replied, "I was moreinto my own words than into listening to Debbie's." In the same situation, Jackie wanted to ask that the neighbor "turn it down," and insteadhe got stuck with "put it lower." Finally, in the "token" situation, Rickisaid she used "Listen -" as an opener "because I didn't have anythingelse to use."


4. Discussion and Conclusions

This chapter has covered both theoretical and applied issues with regardto researching the production of speech act behavior. The chapter beganby calling attention to the sociocultural and sociolinguistic abilities necessary for the production of speech acts. It was noted that the selection ofthe appropriate speech act strategy is conditioned by a host of social,cultural, situational, and personal factors. Then, the research cycle ofethnography, role-play research, written completion tests, and acceptability checks was presented. It was indicated that each of these datacollection techniques has its own merits, but that it is the use of morethan one that provides us with important triangulation. It was suggestedthat in addition to considering the above-mentioned techniques which areuseful for the description of speech act behavior within a group, theresearcher of speech act behavior also needs to better understand thechoices made by individuals and that here is where verbal reports can bemost valuable. Next, there was discussion of research design issuesrelating to the role-play interview as a research tool and to specific datacollection issues. Then some findings generated by speech act productionresearch were presented. These related to the language of thought and tothe search, retrieval, and selection of language forms. Perhaps two ofthese areas, namely, the debate before selection and afterthoughts,warrant extra comment in that they especially provide us with a windowinto the speech act production process.

The debate that goes on in the speaker's mind before selection, whichemerged from the verbal report data, seems to suggest that when facedwith role-play situations - and by extension, in the real world as well learners have their own individual spectrum of options from which tochoose. Some of these options relate to semantic elements, some togrammatical features, and others to illocutionary intent. Among theselearner options, some would lead to appropriate responses while othersmay lead to inappropriate ones. Learners make decisions based on thoseoptions available to them at the moment, without knowing which maylead to inappropriate results, while the native speaker makes choicesbased only on acceptable realizations. It is only through verbal reportthat we are able to tap these kinds of decision-making processes.

With regard to afterthoughts, it became apparent through the verbalreport protocols in the Cohen and Olshtain study that after completing aspeech act situation, learners continued to think about the degree towhich they were successful in their performance in that situation. Often

40 Andrew Cohen

these afterthoughts can lead to selfawareness and may affect future interactions both for better and for worse. Furthermore, the verbal reportprocess itself can unintentionally trigger learners' awareness as to theirspeech act performance. Whereas learners may be mistaken in how theyassess their speech act behavior, this extra awareness may, in fact, directthem to proper use as well. Clearly more work will need to be done tobetter understand the reactive effects of verbal report techniques inspeech production research.

5. Implications for the language learner and the languageclassroom

At a time when teachers have been encouraged to give attention in theirinstruction to communicative language which includes speech acts, thereis evidence that learners of a language may lack even partial mastery ofthese speech acts and that this lack of mastery may cause breakdowns incommunication, much to the consternation of the speaker and hearer(Wolfson 1989). The role of the researcher can be to determine the degreeof control that learners have over different speech acts through themultiple measures suggested above - ethnography, role play, writtencompletion, and acceptability ratings. They can couple with this information verbal report data which add insights regarding the cognitiveprocesses and conscious strategies used to interpret their role in an interaction and to produce appropriate speech act utterances.

Ideally, this information could then be used to prepare a course ofinstruction that would teach to the gaps in language knowledge, and alsogive tips as to strategies that may be useful for producing utterances. Atpresent there are only a few published studies dealing with the teachingof speech act behavior, but the findings seem promising. For example, astudy of advanced English as a Foreign Language learners in Israel wouldsuggest that the fine points of speech act behavior such as (1) types ofintensification and downgrading, (2) subtle differences between speechact strategy realizations, and (3) consideration of situational features, canbe taught in the foreign-language classroom (Olshtain - Cohen 1990;1991).

Likewise, a study by Billmyer (1990) found that tutored English as aSecond Language learners produced a greater number of norm-appropriate compliments, produced spontaneous compliments (which the un-


tutored group did not), used a more extensive repertoire of semanticallypositive adjectives, and deflected many more compliments in their replythan did untutored learners. Her conclusion was that formal instructionconcerning the social rules of language use given in the classroom canassist learners in communicating more appropriately with natives outsideof the classroom.

Notes

1. Parts of this chapter are based on Cohen and Olshtain 1993.2. On the other hand, verbal report data have been collected while informants

have been engaged in the other language skills - reading, writing, andlistening.

3. Robinson (1991) had twelve native Japanese speakers fill out a discoursecompletion test of the ability to refuse requests or invitations, and had themthink aloud while they were doing so. She then played back their think alouddata in an interview to get them to explain retrospectively their thoughts at thetime of completing the task. The verbal report was conducted in the targetlanguage, English, and not in their native language. Although this study hadsome innovative methods in it, such as having the respondents write downtheir reactions to each situation at the time of the response, the languagerelated data were somewhat limited. Respondents had more to say aboutpersonality matters and about reactions to the given situations.

4. Note that this is more than simply omission because an alternate thought issupplied.

References

Afflerbach, Peter - Peter Johnston1984 "On the use of verbal reports In reading research" , Journal of

Reading Behavior 16: 307-322.Aguilar Murillo, Evelyn - Hellen Aguilar - Aimee Meditz

1991 "Teaching speech act behavior through video: Apologies", Athens,OH: Linguistics Department, Ohio University. (Paper presented atthe Ohio TESOL Fall Conference, Ohio University, Athens, OH,November 8-9, 1991.)

Billmyer, Kristine1990 "'I really like your lifestyle', ESL learners learning how to compli

ment", Penn Working Papers in Educational Linguistics 6.2: 31-48.Blum-Kulka, Shoshana

1982 "Learning to say what you mean in a second language: A studyof the speech act performance of learners of Hebrew as a secondlanguage", Applied Linguistics 3: 29-59.

42 Andrew Cohen

Blum-Kulka, Shoshana1989 "Playing it safe: The role of conventionality in indirectness", in:

Shoshana Blum-Kulka - Juliana House - Gabriele Kasper (eds.),37-70.

Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper (eds.)1989 Cross-cultural pragmatics: Requests and apologies. Norwood, NJ:

Ablex.Bonikowska, Malgorzata P.

1988 "The choice of opting out", Applied Linguistics 9: 169-181.Celce Murcia, Marianne (ed.)

1991 An introduction to teaching English as a second or foreign language.2nd ed. Cambridge, MA: Newbury House/HarperCollins.

Cohen, Andrew D.1987 "Using verbal reports in research on language learning", in: Claus

Faerch - Gabriele Kasper (eds.), 82-95.Cohen, Andrew D.

1991 "Feedback on writing: The use of verbal report", Studies in SecondLanguage Acquisition 13: 133-159.

Cohen, Andrew D. - Carol Hosenfeld1981 "Some uses of mentalistic data in second-language research",

Language Learning 31: 285-313.Cohen, Andrew D. - Elite Olshtain

1981 "Developing a measure of sociocultural competence: The case ofapology", Language Learning 31: 113-134.

Cohen, Andrew D. - Elite Olshtain - David S. Rosenstein1986 "Advanced EFL apologies: What remains to be learned?"

International Journal of the Sociology of Language 62: 51-74.Cohen, Andrew D. - Elite Olshtain

1993 "The production of speech acts by EFL learners", TESOL Quarterly27: 33-56.

Cohen, Andrew D. - Elite Olshtain1994 "Researching the production of L2 speech acts''', in: Elaine E. Tarone

- Susan M. Gass - Andrew D. Cohen (eds.), 143-156.Faerch, Claus - Gabriele Kasper

1987 "From product to process-introspective methods in secondlanguage research", in: Claus Faerch - Gabriele Kasper (eds.),5-23.

Faerch, Claus - Gabriele Kasper (eds.)1987 Introspection in second language research. Clevedon, England:

Multilingual Matters.Kasper, Gabriele (ed.)1991 Pragmatics of Japanese as a native and target language. Technical

Report No.3. Honolulu, HI: Second Language Teaching andCurriculum Center, University of Hawaii.

Kasper, Gabriele - Merete Dahl1991 "Research methods in interlanguage pragmatics", Studies in Second

Language Acquisition 13: 215-247.


Kieras, David E. - Marcel A. Just (eds.)1984 New methods in reading comprehension research. Hillsdale, NJ:

Lawrence Erlbaum Associates.Mac Mathuna, Liam - David Singleton (eds.)

1984 Languages across cultures. Dublin: Irish Association for AppliedLinguistics.

Olshtain, Elite - Andrew D. Cohen1983 "Apology: A speech act set", in: Nessa Wolfson - Elliot Judd (eds.),

18-35.Olshtain, Elite - Shoshana Blum-Kulka

1984 "Cross-linguistic speech act studies: Theoretical and empiricalissues", in: Liam Mac Mathuna - David Singleton (eds.), 235-248.

Olshtain, Elite - Shoshana Blum-Kulka1985 "Crosscultural pragmatics and the testing of communicative

competence", Language Testing 2: 16-30.Olshtain, Elite - Andrew D. Cohen

1990 "The learning of complex speech act behavior", TESL CanadaJournal 7: 45-65.

Olshtain, Elite - Andrew D. Cohen1991 "Teaching speech act behavior to nonnative speakers", in: Marianne

Celce Murcia (ed.), 154-165.Olson, Gary M. - Susan A. Duffy - Robert L. Mack

1984 "Thinking-out-Ioud as a method for studying real-time comprehension processes", in: David E. Kieras - Marcel A. Just (eds.),253-286.

Robinson, Mary1991 "Introspective methodology in interlanguage pragmatics research",

in: Gabriele Kasper (ed.), 29-84.Stansfield, Charles - Dorry M. Kenyon - Ricardo Paiva - Fatima Doyle Ines DIsh - Maria A. Cowles

1990 "The development and validation of the Portuguese speaking test",Hispania, 73: 641-651.

Stubbs, Michael1983 Discourse analysis. Oxford: Basil Blackwell.

Tarone, Elaine E. - Susan M. Gass - Andrew D. Cohen (eds.)1994 Research methodology in second-language acquisition. Hillsdale,

NJ: Lawrence Erlbaum Associates.Weizman, Elda

1989 "Requestive hints", in: Shoshana Blum-Kulka - Juliane House Gabriele Kasper (eds.), 71-95.

Wolfson, Nessa1989 Perspectives: Sociolinguistics and TESOL. Cambridge: Newbury

House.Wolfson, Nessa - Elliot Judd (eds.)

1983 Sociolinguistics and language acquisition. Rowley, MA: NewburyHouse.

Non-native refusals: A methodological perspective~1-

Noel Houck - Susan M. Gass

1. Introduction

Much research in recent years has been carried out in the area of interlanguage pragmatics. In this chapter we focus on one area of this researchdomain, that of refusals, in the hopes of elucidating methodologicalissues in non-native speech act research. A number of papers in this areaemphasize the point, with which we concur, that methodological issuescannot be ignored, for it is not clear to what extent differences inmethodology yield differences in results (see, for example, Cohen thisvolume; Cohen - Olshtain 1994).

Wolfson (1981) and Wolfson, Marmor and Jones (1989) have arguedthat ethnographic data collection is the most reliable means of learningabout the social and linguistic constraints on a particular speech act. Thismethodology allows for observation of naturally occurring speech eventswith precise recording about the social setting, location, and the participants, thereby providing information about the linguistic and socialconstraints on the use of a given speech act (cf., Watson-Gegeo 1988 fora discussion of methodological issues).1

However, as has been pointed out by a number of researchers (e.g.,Rintell - Mitchell 1989; Kasper - Dahl 1991), there are limitations.Not only can contextual variables not be controlled, but also the occurrence of a particular speech act cannot be predicted. If one is truly tounderstand a given speech act, many occurrences are needed; this, ofcourse, is difficult when one must rely on instances when a particularspeech act is used by speakers who are unaware of being observed. 2

In the most detailed treatment to date on the issue of methodology insecond language speech act research, Kasper and Dahl (1991) review 39studies of interlanguage pragmatics. They characterize the methods usedalong two dimensions: 1) By the constraints each imposes on the data and2) By the degree to which production or comprehension is studied. Forour purposes, we focus only on production data, although we bear in

46 Noel Houck - Susan M. Gass

mind Kasper's (1984) caution that many apparent production problemsare a result of a non-native speaker's inadequate comprehension of previous parts of the discourse.

Focussing on production data, Kasper and Dahl describe two majordata elicitation measures, discourse completion and role-play. DiscourseCompletion Tests are written questionnaires consisting of a brief description of a situation followed by dialogue with a blank line where the subject is to put in what s/he believes to be an appropriate response. Theother major type of production data comes from role-play, both open andclosed. An example of a closed role-play comes from Rintell and Mitchell(1989), in which subjects were given an oral version of the DiscourseCompletion Test. In open role-plays, on the other hand, an entire dialogue is observed and recorded.

1.1. Discourse completion tests

There are advantages and disadvantages to each of these types of datacollection. With regard to the Discourse Completion Test, perhaps themost widely used methodology in interlanguage pragmatics, the advantages are clear: Large amounts of data can be collected in a relatively shortamount of time. Furthermore, because of the consistency of the situation,responses can be compared along a number of dimensions (e. g., age,gender, ethnicity). On the other hand, there is the question of the extentto which the data collected are a reflection of the sociolinguisticconstraints that operate on the speech act in question. This is similarlyargued by Wolfson, Marmor and Jones (1989), who point out that "shortdecontextualized written segments" may not be comparable to whattakes place in actual interaction.

In fact, recent research on the comparability of Discourse CompletionTest data with data collected using other techniques has revealed someimportant differences. Rose (1992) has shown that the frequency ofdifferent types of response varies with the instrument. Rose comparedrequests elicited by a Discourse Completion Test with responses to amultiple choice questionnaire, both of which were administered in Englishto native speakers of English and in Japanese to native speakers ofJapanese. He reports that while the most frequent response to all situationson the Discourse Completion Test was conventionally indirect requests,responses to the multiple choice questionnaire exhibited more contextualvariation, with respondents often choosing to opt out or to hint.

Non-native refusals 47

The richness of naturally occurring refusals cannot be adequatelycaptured with a formalized structure such as that represented by theDiscourse Completion Test. As we will show below, refusals are veryoften filled with multi-turn responses involving negotiation, hedging andeven reversal. This has been further pointed out by Bardovi-Harlig andHartford (1990) and Hartford and Bardovi-Harlig (1992). Their work,based on naturalistic data, focussed on rejections of advice in academicadvising sessions. In their data they found different strategies for refusals/rejections than had been found in the Beebe, Takahashi, and Uliss-Weltz(1990) study. For example, in Bardovi-Harlig and Hartford's data, thesubjects exhibited what they called verbal avoidance, evidenced in theform of postponement ("Can I think about it?"), requests for repetition,or requests for additional information. Hartford and Bardovi-Harlig(1992) concluded that the Discourse Completion Test results in limiteddata; specifically, their results show a more limited range of semantic formulas, fewer status-preserving strategies, and none of the extended negotiations which occurred in the natural data. Clearly, written responses,especially those that are "sandwiched" between an opening statementand a follow-up statement (as in Discourse Completion Tests), do notallow a speaker to exhibit the strategies found in naturalistic data.

1.2. Role-play

Role-plays have the advantage of providing data in an oral mode ratherthan a written mode (although below we will deal with differences inthese two channels). In a closed role-play (e.g., Walters 1980), subjectsare given a situation and are asked how they would respond. In Walters'study children were asked to make a request to a puppet. These puppetsvaried in age, sex, and race. However, as pointed out above, any typeof data that is "closed", in that it does not allow a free range of answers,will suffer from the possibility of non-symmetry with naturally occurringdata.

Of the common data elicitation methods, open role-plays are theclosest to what we might expect to reflect naturally occurring speechevents. They have the advantage of allowing the researcher to set upsituations in which the occurrence of a particular speech act is likely incircumstances in which the occurrence can be recorded and/or videotaped, thus making possible the close analysis of long interactionsequences of comparable data.


However, they are not problem free. They are cumbersome toadminister and time-consuming in both their administration andanalysis. Furthermore, role-plays are just that, role-plays, so again weare left with the question of the degree to which they mirror the linguistic behavior of individuals in the particular setting established by theresearcher.

With these caveats in mind, in this study we have adopted open roleplays in order to study refusals. We have done this fully aware of the disadvantages, but also aware of the advantages particularly with regard torefusals. As Edmondson (1981 cited in Beebe - Cummings 1985) pointsout, "some speech acts are the result not of a single utterance but of anegotiation, a cooperative achievement, or a conversational outcomebetween two speakers." It is clear that to see this type of negotiation inrefusals, we must use an open format to elicit data.

2. Background

2.1. Methodological issues

We turn now to a discussion of the literature in which methodologyis the focus. In addition to Kasper and Dahl (1991), who present a surveyon methodological issues, there are two notable papers that presentresults based on a comparison of methodologies. One is by Rintelland Mitchell (1989) and the other by Beebe and Cummings (1985; thisvolume).

Rintell and Mitchell used written and oral versions of a DiscourseCompletion Test (eliciting apologies and requests) that were given to lowadvanced learners of English and to native English speakers. They wereformulated as role-plays with a variety of social roles and situationsrepresented. Clearly some differences did exist between the two modalities of elicitation. In particular, for the second language speakers, the oraldata were longer than the written data. This difference was not apparentin the native speaker responses, leading the authors to conclude that itwas not so much the methodology that resulted in different responses,but rather the way in which the two groups approached the tasks. Ingeneral they found that the "language elicited ... is very similar whethercollected in written or oral form" (1989: 270). They argue that the


Discourse Completion Test is in actuality a role-play. That is, both thewritten and spoken forms provide data that resemble spoken languagerather than written language.

Beebe and Cummings (1985; this volume) study refusals using twotypes of data for their analysis: The Discourse Completion Test and telephone requests. The data they collected were only from native speakersof English. In both the written and the oral tasks, subjects were asked ifthey would be willing to help with the local arrangements for the TESOLconvention in New York.

What they found was that in the oral data, there was more elaborationof the refusal; in the written data, the layout on the page allowed for onlya minimum amount of data to be produced. Elaborations come as a resultof the "requester's" response. If the requester, upon hearing a refusal,responds "all right, thank you" and then hangs up, there will be nofurther need for elaboration. But if, on the other hand, there is silence orsome other attempt to keep the conversation going, the refuser will feel aneed to elaborate so as not to be offensive to the requester. Goffman(1971) points out that the offending person (in this case the refuser) needsto receive reassurance from the addressee that his/her offending remark isnot taken as a serious offense. Elaboration is what restores the offenderto his/her proper place in the eyes of the addressee. Beebe and Cummingspoint out that the written test biases "the response toward less negotiation, less hedging, less repetition, less elaboration, less variety andultimately less talk" (this volume: 71).

Beebe and Cummings analyzed the written and oral data in terms ofthe types of responses given, finding that the written data reflect thecontent of oral data (e.g., the use of "I'm sorry"; the frequency withwhich excuses were offered; the frequency with which willingness orability was offered). Where the two modes differed was in what they callthe "psychosocial" domain. That is, when one refuses, one needs to takea cue from the requester as to how offensive or how important the refusalis. This will then dictate the degree to which further elaboration, hedging,or apologizing is necessary.

2.2. Refusals

Refusals are a highly complex speech act primarily because they ofteninvolve lengthy negotiations as well as face-saving maneuvers to accommodate the noncompliant nature of the speech act. Because refusals


normally function as second pair parts, they preclude extensive planningon the part of the refuser.

2.3. Second language refusals

Two studies on second language refusals are relevant for our purposes.The first is a study by Beebe, Takahashi and Uliss-Weltz (1990) in whichthe major concern was the existence of pragmatic transfer. Four groupsof native speakers of Japanese and English (two native speaker controlsand two second language groups) filled out a Discourse Completion Testinvolving twelve situations including refusals of requests, refusals ofinvitations, refusals of suggestions and refusals of offers. Each situationinvolved an initial segment of written speech followed by a blank andthen followed by a rejoinder that forced the subjects to write a refusalin the preceding blank. In analyzing the results, the authors consideredthe order of semantic formulas. Their preliminary conclusions suggestevidence of pragmatic transfer although they are quick to caution usregarding the limitations of the data elicitation methods used.

The second study we mention is that of Kinjo (1987), who examinedrefusals to invitations and requests in English and Japanese. Data werecollected orally, with subjects responding to a taped invitation or request.In her analysis, Kinjo considered the degree to which mitigators played arole and the degree to which directness/indirectness reflected thestereotypical notions one has of these two cultures. As with the Beebeet al. study, Kinjo warns that the results that come about as a result ofthis modified role-play method may not reflect naturally occurringspeech.

3. The study

In our attempt to investigate interlanguage refusals, we were primarilyconcerned with the interaction involved in the refusal itself. Refusals areplayed out events, rather than instances characterized by a brief exchangeor single utterance. That is, we begin with the notion that the modifiedrole-play, a typical means of gathering data, is insufficient to an understanding of the complete speech event of refusing.


3.1. Method

Following the work of Beebe et aI., we investigate refusals to four typesof situations: 1) Suggestions, 2) Offers, 3) Invitations and 4) Requests.We depart from previous studies of refusals in two ways: 1) We use videotaped data, and 2) We use full role-play situations rather than modifiedrole-plays, with the eliciting instrument based on Scarcella's (1978)conceptualization of socio-drama. This allows participants to carry outthe refusal to its logical conclusion. Thus, the responses that are given arenot confined by either the printed page (e.g., the amount of spaceprovided on the page, the number of turns that the respondent is expectedto take) or by the closing response of the initiator of the interactionwhich, in many Discourse Completion Tests, directs the refusal by "sandwiching" it between a given opening remark and the subsequent closingcomment.

Two situations requiring refusals were created for each of the fourrefusal types so that in all eight situations existed (see Appendix). Thesetting for each was the home of an American host family who asked theguest to do something undesirable and quite unusual, such as 1) get astrange haircut, 2) pierce their ears, 3) go skydiving, or 4) give a speechat church.3

3.2. Subjects

Our data-base consists of an interaction involving a native speaker ofEnglish who was the person making the request, invitation, suggestion, oroffer and Japanese English as a second language (ESL) students at twolevels of proficiency. The subjects of the study were given the contextualinformation surrounding each situation. Following this introduction,each subject "role-played" the part with a native speaker who had beeninstructed not to give up too easily in cases in which the non-nativespeaker initially refused. We made certain that each subject understoodthe situation before the session began. All sessions were videotaped anda subset was transcribed. For each of the eight situations, data from fournon-native speaker-native speaker (two lower-level proficiency and twohigher-level proficiency) pairs were gathered. In analyzing the results forthis study, we focussed on a subset of the data collected. 4 In Houck andGass (in press) other data are analyzed from the perspective of non-verbalcommunication.


4. Findings

Data collected using an open role-play differ from data collectedusing a written or tape-recorded elicitation instrument in a number ofsignificant ways. The most obvious is that a real face-to-face encounterresults in a dynamic interaction. It is one thing to formulate a refusalon paper; it is quite another to deliver that refusal to a person whowill respond to it. Not once in our data did the refusal interactionterminate with the subject's initial response. 5 The role-plays resultedin what were often lengthy interactions in which the participantsnegotiated their way to a resolution. During this time, speakers hemmedand hawed, cut each other off, requested clarification, self corrected,modified and elaborated their positions, and generally became involvedin negotiating semantic, pragmatic, and social meaning. 6 Thus, ourrole-play data differ from other data on refusals both quantitatively andqualitatively.

4.1. Quantitative analysis

One quantitative consequence of using an open role-play is that the dataconsist not of one response, but rather of a series of turns. To obtain aquantitative measure of the data, we considered not only turn length, butalso the number of turns. Turn length often varied according to level ofEnglish ability (Blum-Kulka - Olshtain 1986). In our data as well, lessproficient subjects had shorter turns. After eliminating back channels,such as "mm" and "oh", and pause fillers, such as "uh", we found thatsubjects with lower English proficiency averaged 3.5 words per turn.Higher proficiency subjects were much more prolific, averaging 10.7words per turn.

The total number of turns from the triggering speech act to the end ofthe role-play varied from 7 to 18. On the average, subjects required 9.8turns at talk to reach a resolution.7 As might be expected with realnegotiations, the outcomes differed considerably. Resolution wasachieved when the participants reached agreement, and: 1) The nativespeaker accepted the non-native's refusal; 2) The native speaker and nonnative speaker reached a compromise; or 3) The non-native speakeraccepted the native speaker's offer, request, invitation, or suggestion, andthe role play ended with a few final comments or plans.


4.2. Qualitative analysis: Classifying the data

We also analyzed the refusal sequences, categorizing the responses madeby the non-native speakers. As a starting point, we applied a comprehensive classification system of refusals, developed by Beebe, Takahashiand Uliss-Weltz (1990) to each non-native speaker response. Severalcategories in this system accounted for approximately 2/3 of the responses.

- Conventional nonperformative refusals (e. g., "I can't" "No")- Statements of regret (e.g., "I'm sorry")- Excuses/reasons/ explanations (e.g., "But 1 don't know yoU")8- Proposals of alternatives (e.g., "Please wait in your car if you want to

meet him")

These responses also predominate in data from other research onrefusals, such as Kinjo (1987) and Bardovi-Harlig and Hartford (1990).

However, our data also contained non-native speaker responses thatdid not correspond neatly with Beebe et al.'s classes. Three of thesewere linguistic responses: Confirmations, requests for clarification/information and agreements. An additional category we labelled nonverbal responses.

4.2.1. Confirmations

Confirmations occurred frequently in the conversations of lowerproficiency non-native speakers. When a non-native speaker begangroping for words or exhibiting signs of linguistic distress, the nativespeaker often leapt in, checking assumptions and elaborating on minimalutterances. The non-native speaker could then respond with a singleword, indicating that the native speaker was correct. The non-nativespeaker was thus able to get away with a minimum of speech, as in (1)where he is a guest at a weekend homestay. At breakfast, the nativespeaker is inviting him to go skydiving with the family that day.

(1) Confirmation (skydiving)

1. NS: Do you like to skydive?2. NNS: No3. NS: No?


4. ~ NNS: ((nods)) Yes5. NS: Why6. NNS: Vh, I head(eh) headek headache7. NS: Y, have ... headache?8. ~ NNS: Headache9. NS: Oh you have a headache oh no he has a headache

10. ~ NNS: Headache

In line 2 the non-native speaker indicates his lack of interest in skydiving,and in line 4 he confirms it with a single word ("Yes") and no explanation. It is the native speaker who requests an explanation (line 5) and,having received one (line 6), repeats the non-native speaker's explanation(lines 7 and 9), to which the non-native speaker offers little support(a confirmatory repetition of his excuse in lines 8 and 10). While thissegment can be seen as an instance of negotiated meaning, it alsoprovides opportunities for the non-native speaker to elaborate on hisexcuse or to add an apology. The non-native speaker seems to recognize that a contribution is called for, but limits it to the single-wordrepetitions in lines 8 and 10; the native speaker interprets the minimalresponses and adopts an appropriate attitude (line 9, "Oh no he has aheadache").

In this exchange, the native speaker and non-native speaker work outthe non-native speaker's excuse together, with the native speaker askingquestions and reacting to the information provided, while the non-nativespeaker provides minimal answers and confirms the native speaker'srestatements.

4.2.2. Request for clarification

On the other hand, some non-native speakers formulated their ownrequests for clarification, as in example (2).

(2) Request for information/clarification (skydiving)

1. ~ NNS: What means2. NS: What is skydiving

In this example the non-native speaker has just been informed that shewill be going skydiving with the host family that day; she requests anexplanation of the term "skydiving" ("what means").


Thus, non-native speakers in these open role plays often spend someof their time ostensibly in the negotiation of meaning - with confirmations and requests for information, although it is likely that the clarification requests were actually serving the function of verbal avoidance asdiscussed by Bardovi-Harlig and Hartford (1990).

4.2.3. J\greennent

A third response type that differed from responses in most previousstudies of refusals is agreement. While Imai (1981), Rubin (1983) andBeebe et al. mention general or unenthusiastic acceptance as types of refusals, acceptances that evolve from initial refusals are not discussed.

In several of the role plays, when faced with a persistent nativespeaker, the non-native speaker abandoned her attempt to refuse andaccepted. In example (3), the non-native speaker's hostess at a weekendhomestay has offered to give the non-native speaker a punk-style haircutlike her children's. This interaction occurs after the non-native speakerhas given two explicit refusals and a reason (only his barber, who is a hairspecialist, can cut his hair):

(3) Agreement (haircut)

1. NNS:2. NS:3. NNS:4. NS:5. NNS:6. NS:

7. ~NNS:

Ummmm ((laugh)) I like this barberUhhuhYeahBut but you like my children's haircuts, right?UmmmSo I c'n I can cut your hair and you can feelcomfortable and cool?Yeah please

In line 7, the non-native speaker abruptly changes his stance and agreesto let the native speaker cut his hair. When asked afterwards if theywould really have agreed to having their hair cut, approximately half thenon-native speakers polled said that they would, because she was theirhostess.


4.2.4. Nonverbal responses

In addition to those responses in which the non-native speaker used, ifnot propositions, at least lexical items from which a reasonable proposition could be inferred, non-native speakers expressed their reactions andresponses through nonverbal signals. We contend that these nonverbalsigns often performed the same functions as turns with recoverablepropositions. In fact, they were often used to confirm a native speakerstatement or to request clarification or information.

Our transcripts contain a number of non-native speaker nonverbalsignals, such as the nod in (4) and raised eyebrows in (5), which clearlycarry intended communicative content. And, indeed, they can function bythemselves as a turn, performing an interactive function.

(4) Nonverbal: Confirmation (skydiving)

The non-native speaker has expressed fear of skydiving.

1. NS: ((to others )) She's afraid2. NNS: ((nodding)) Mm3. NS: But you're sure you don't want to go skydiving4. ~ NNS: ((nods))

In line 1, the native speaker informs others that the non-native speaker isafraid; and in line 2 the non-native speaker confirms this with an "Mm"accompanied by nods. In line 3, the native speaker states an implicationof the non-native speaker's fear, i. e., that she does not wish to skydive, asa request for confirmation. The non-native speaker confirms the nativespeaker's understanding with a (nonverbal) nod (line 4). Thus, in line 4the nod signals confirmation.

In example (5), as in example (4), the native speaker attempts a restatement of the non-native speaker's previous statement, but in this case shemisinterprets the non-native speaker's meaning.

(5) Nonverbal: Request for information/clarification (speech atchurch)

As the host family and the non-native speaker prepare to goto church, the non-native speaker is informed that she hasbeen requested to give a talk about herself and her life at theuniversity.


1. NS:

2. NNS:3. NS:4. ~NNS:

... they want you to give a speech to everybody inthe church. Is that OK?Urn urn it's it's very short time for (me)Oh OK you do not want to give a long speech((raises eyebrows))

In this interaction, the non-native speaker apparently intended theutterance in line 2 ("it's it's a very short time for (me)") to mean that shewould not have enough time to prepare a speech. However, the nativespeaker understands her to be saying that she is willing to give a speechif it is a short one (line 3). Thus, the native speaker's paraphrase does notcorrespond to the non-native speaker's intentions.' The non-nativespeaker's turn in line 4 is an opportunity to confirm or disconfirm thenative speaker's interpretation. Her raised eyebrows convey her uncertainty and function as a question/request for further clarification.

Thus, in face-to-face interactions, non-native speakers may call on anumber of resources in negotiating refusals. And they may employ theseresources to convey different meanings, depending on the context.

5. Discussion and conclusion

To summarize, the use of open role plays illustrates that refusals oftenrequire a number of turns to effect a response. The number of turnsrequired may reflect the natural need for conversationalists to interact tosolve a problem - e. g., through negotiation and elaboration of meaning.The negotiation/elaboration may necessitate a greater number of turnswhen a non-native speaker is involved than when only native speakers areconversing. Or, it may indicate the persistence or stubbornness of theindividual native speaker interlocutor and the non-native speakerrespondent. A non-native speaker may also need to tryout more ploys toresolve disharmony (see Bardovi-Harlig - Hartford 1990).

The use of open role plays has also shown that the performance of actssuch as refusals involves the use of resources not required or even appropriate in noninteractional role play. Thus, we identified three acts confirmation, request for clarification, and agreement - which have notbeen included in most previous classification schemes.

These new classes of acts are special in that, unlike the acts conventionally associated with refusals, they are characteristic of dynamic inter-


action in general. However, their occurrence in stressful negotiations isespecially appropriate and plays a crucial role in the non-native speaker'snegotiation of a response.

Additionally, these three classes are particularly effective because inaddition to their obvious speech act function, individual instances of theacts can represent a discourse tactic or social maneuver designed to softenthe unpleasantness of a refusal. For instance, a refusal that develops intoa series of non-native speaker confirmations, as in example (1), mayallow the non-native speaker to build up solidarity with the nativespeaker in a face-threatening situation.

It is to be further noted that a request for information or clarificationmay function as an avoidance tactic. Bardovi-Harlig and Hartford (1990)have pointed out that requests for information are employed by nonnative speakers as an indirect means of avoiding a refusal, in their case arefusal of a suggestion. The interaction in example (2) took place after themeaning of skydiving had been carefully explained and after a previousrole play on skydiving had been acted out in front of this particular nonnative speaker.

Interestingly, another subject, who had observed two skydiving roleplays also requested an explanation of the term "skydiving" during hisown skydiving role play, supporting the contention that these requests forclarification may reflect the speaker's wish to avoid direct refusal ratherthan a real need to negotiate meaning.

Finally, the fact that a change of heart took place and agreementultimately occurred in a number of cases represents the ultimate in refusalalleviation. The fact that agreement occurred in refusal negotiations, allof which began with a non-native speaker's clear disinclination tocomply, points to the interesting question of which contextual factorsfacilitate agreement and which mitigate against it. Under certain conditions, non-native speakers gave up refusing in favor of agreement; underother conditions, non-native speakers were unmovable. For instance, inthose situations in which the offer or request was not dangerous or potentially painful, or when subjects might be seen as disappointing theirhostess if they refused, they often eventually agreed. However, if a positive response meant that the subject or the host or host's family might beput in a dangerous situation, the subject continued to refuse, no matterhow tricky a linguistic feat, and even at the risk of appearing ungraciousand impolite.

Thus, the inclusion of confirmations, requests for clarification andagreements in a classification system of refusals is indicated as soon as we


consider a refusal not as a simple response to a static situation but as adynamic negotiated achievement.

A third practical methodological implication resulted from our use ofvideotapes, which enabled us to capture the use of nonverbal resources tonegotiate meaning. Most importantly, under the appropriate circumstances a head movement or a raised eyebrow can clearly perform thesame function as a verbal "yes" or "oh?" (see Houck - Gass in press forfurther discussion). And yet these gestures are available for incorporationinto analyses of speech act performance only when the researcher isdealing with observed interactions. By including nonverbal signals asintentional speech act moves, we recognize the wide range of resourcesavailable for communicating a message.

Our data reveal the existence of a richer variety of meaningfulresources and maneuvers than has generally been documented in discussions of non-native refusals. The negotiations we have described go farbeyond the notion of a simple response consisting of linguistically analyzable units. They involve art interaction not only between what the nonnative speaker wants to say and what her interlocutor wants her to say,but how to say it - what grammar, gesture and discourse tactics to use tocarry out both her social obligations and her personal wishes effectivelyin a particular situation.

Important work has been done collecting data on the selection andrealization of linguistic acts across cultures using written discoursecompletion tests and closed role plays. The addition of interactionaldata from open role plays can only enrich our understanding of speechacts.

Appendix

Situations used

1. You are ready to leave the house to go to a party with the children of your hostfamily - Nathan, age 21, and Jennifer, age 23. They are telling you abouttheir friends and the things they usually do at parties. The more they talk, themore you realize that everyone at the party will be using dangerous drugs.Nathan picks up his car keys and starts for the door. [Invitation]

2. You are at your host family's home. Your host family, the Quentins, has goneto a neighbor's house to discuss a business matter. They have left you at homewith specific instructions not to let anyone in the house, no matter what theysay. It could be dangerous. About 5 minutes after they leave, the doorbell


rings. It's a woman who says that she is Mr. Quentin's cousin from Detroit.She is just passing through Lansing and says, "Can I come in and wait?"[Request]

3. It is Saturday morning at your host family's home. At breakfast the family tellsyou that they have made reservations at the airport for all of you to go skydiving this morning. The whole family - Mr. and Mrs. Cousins, Meg, andTim - are all getting ready to go. They ask you if you have ever gone skydivingbefore. When you say no, they say, "Don't worry! It's easy!" [Invitation]

4. It is Sunday morning and you have agreed to attend church services with yourhost family, the Jarvises. As you are getting ready to leave the house forchurch, Mrs. Jarvis informs you that there are plans for you to give a shortspeech about university life in Japan after the services. She says, "I hope youwon't mind." [Request]

5. It is 11 :OOam Saturday morning at the home of your host family, the Larsons.You arrived at the Larsons' home last night at about 8:00pm. You thoughtthat you would be having dinner with them, but they thought you had eaten,so you had no dinner. This morning you had only a piece of toast and coffee.You are now very hungry. Mrs. Larson walks into the room and tells you thatyou will be going to an early barbecue for dinner. She suggests that becauseyou will be eating at about 5:00pm, you skip lunch today. But you are reallyhungry. [Suggestion]

6. You are at the home of your host family, the Sumners. Both the children,Charlie and Karen Sumner, have short, very ugly haircuts. At one point, theyask you how you like their hair. You answer politely that it looks very cool andcomfortable. Mrs. Sumner announces proudly that she cuts their hair herself.And because you like the style, she will be glad to cut your hair to look liketheirs. "Now where are my scissors ... ?", she asks. [Offer]

7. You are watching MTV with your host family on Saturday. You notice thatboth men and women rock stars have at least 4 earrings in their ears. Youcomment that this style is very interesting. Your host family's son Bob, age 22,says, "Oh, I'm glad you like it. My girlfriend pierced my ears. Why don't youget yours done, too? I'll call her right now, and she can be here in 20 minutesto pierce your ears." Bob goes to the telephone to call. [Suggestion]

8. You are at your host family's home. Your host mother, Mrs. Boulware, isadmiring the expensive new pen that your family gave you before you leftJapan. Mrs. Boulware sets the pen down on a low table, and you and she gointo the backyard to look at her flowers. When you return to the room, theBoulware's pet dog, Ruffy, is happily chewing on your pen. When Mrs.Boulware gets the pen out of Ruffy's mouth, it is ruined. Mrs. Boulware says,"Oh, I am so sorry. I'll buy you a new one." [Offer]

Notes

* We are grateful to Joyce Neu for helpful comments on an earlier version of thischapter. We also thank the panelists and participants of the TESOLSociolinguistics Colloquium (1991), where these data were originally presented, for their comments and suggestions.


1. An alternative to detailed ethnographic data is what Beebe (1993) refers to as"notebook data". This data type consists of memorizing the core act (e. g.,refusal) when it occurs, as well as any supporting moves that the researchercan commit to memory; writing down immediately everything that she canremember precisely, as well as any partially recalled speech or additionalinformation that may be relevant to a description of the interaction; andmarking notes ruthlessly to reflect which dialogue was recalled verbatim andwhich was reconstructed. Although it is limited to capturing short interactions, the method allows an alert observer to gather large amounts of dataon particuar types of acts in relatively short spans of time.

2. Bardovi-Harlig - Hartford (1992) analyze the advantages and limitationsof naturally occurring data collected in an institutional setting, in whichthe interactants and situations are relatively invariant, and in which theinteractions are videotaped (and participants are aware of being observed).

3. One could argue that it is highly unlikely that a non-native speaker wouldencounter situations such as the ones used in this study. However, that appearsnot to be the case. The data for this study were collected immediately following a home-stay weekend in which these students had visited an Americanfamily. Some of these situations had quite coincidentally been encountered ashad even more bizarre ones, such as a suggestion to go to the morgue to see adead body. What is interesting and what will be discussed later in this chapteris the extent to which the non-native speakers gave in to a request. Thisappeared to be in large part determined by the extent to which the guest couldcomply with the native speaker's request, offer, suggestion, or invitation without putting herself at risk.

4. For the lower proficiency students we analysed the data from two of the skydiving situations, two of the speech at church situations and two of the haircut situations. For the higher proficiency students, we analysed data from twoof the visiting cousin situations.

5. This may, of course, be due to the instructions given to them, but it also maybe a result of the methodology used, which did not allow for a comfortableclosure early on in the interaction.

6. The result is, of course, messy. In our data, the researcher determined thenature of the initiating speech act, but had no real effect on the remainingspeech; in pure observational research, the researcher, of course, controls evenless (cf., Beebe 1993).

7. In addition to the native speaker's persistence, it might be expected that theamount of negotiation reflected a japanese reluctance to refuse directly.However, a look at the non-native speakers' use of the most direct linguisticrefusal "no" indicates that they were often willing to state refusals directly.(In five of the eight role plays, the japanese subjects indicated refusal atleast once with "No.") This is corroborated by the data from Kinjo, whofound that her japanese subjects were more open and direct than herAmerican subjects.The data from this study give a preliminary indication that japanese subjectswill give direct negative responses, at least in some situations. However, a


study of the factors responsible for determining the level of directness was notwithin the scope of this project.

8. As Joyce Neu points out, there are other interpretations possible to a phrase,such as "But I don't know you." For example, it may be a challenge to thespeaker's right to ask the hearer to do X. While this might be the case in interactions between two native speakers, we are confident that in our data, giventhe participants, given the intonation and given the body and facial expressions, these are truly expressions of refusal.

References

Bardovi-Harlig, Kathleen. - Beverly S. Hartford1990 "Learning to say 'no': Native and non-native rejections in English",

Paper presented at the conference on Pragmatics and LanguageLearning. University of Illinois, April.

1992 "Natural conversations, institutional talk, and interlanguage prag-matics", Paper presented at the Pacific Second Language ResearchForum. University of Sydney, July.

Beebe, Leslie1993 "Rudeness: The neglected side of communicative competence?",

Paper presented at the Fourteenth Annual TESOL and Sociolinguistics Colloquium. TESOL National Convention, Atlanta,April.

Beebe, Leslie - Martha C. Cummings1985 "Speech act performance: A function of the data collection proce

dure", Paper presented at the Sixth Annual TESOL and Sociolinguistics Colloquium. TESOL, New York, 1985.

Beebe, Leslie - Tomoko Takahashi - Robin Uliss-Weltz1990 "Pragmatic transfer in ESL refusals", in: Robin Scarcella -Elaine

Andersen - Stephen Krashen (eds.), 55-73.Blum-Kulka, Shoshana - ]uliane House - Gabriele Kasper (eds.)

1989 Cross-cultural pragmatics: Requests and apologies. Norwood, N]:Ablex.

Blum-Kulka, Shoshana - Elite Olshtain1986 "Too many words: Length of utterance and pragmatic failure",

Studies in Second Language Acquisition 8: 165-180.Bouton, Lawrence - Yamuna Kachru (eds.)

1992 Pragmatics and language learning 3. Urbana-Champaign, IL:University of Illinois.

Cohen, Andrew - Elite Olshtain1994 "Researching the production of speech acts", in: Elaine Tarone

Susan Gass - Andrew Cohen (eds.), 143-156.Edmondson, Willis

1981 Spoken discourse. London: Longman.Goffman, Erving

1971 Relations in public: Microstudies of the public order. New York:Basic Books.


Hartford, Beverly - Kathleen Bardovi-Harlig1992 "Experimental and observational data in the study of interlanguage

pragmatics", in: Lawrence Bouton - Yamuna Kachru (eds.), 3352.

Houck, Noel - Susan Gassin press "Nonverbal communication In non-native refusals", in: Adam

Jaworski (ed.).Imai, Masaki

1981 Sixteen Ways to Avoid Saying No. Tokyo: Nihon Keizai Shinbun.Jaworski, Adam (ed.)

in press Silence: Interdisciplinary perspectives.Kasper, Gabriele

1984 "Pragmatic comprehension in learner-native speaker discourse",Language Learning 34: 1-20.


Language Acquisition 13.2: 215-247.Kinjo, Hiromi

1987 "Oral refusals of invitations and requests in English and Japanese",Journal of Asian Culture 1: 83-106.

Rintell, Ellen M. - Candace J. Mitchell1989 "Studying requests and apologies: An inquiry into method", in:

Shoshana Blum-Kulka - Juliane House - Gabriele Kasper (eds.),248-272.

Rose, Kenneth1992 "Speech acts and DCTs: How reliable is questionnaire data?" Paper

presented at TESOL, Vancouver, March.Rubin, Joan

1983 "How to tell when someone is saying 'no' revisited", In: NessaWolfson - Elliot Judd (eds.), 10-17.

Scarcella, Robin1978 "Socio-drama for social interaction", TESOL Quarterly 12.1:

41-46.Scarcella, Robin - Elaine Andersen - Stephen Krashen (eds.)

1990 Developing communicative competence in a second language. NY:Newbury House.

Tarone, Elaine - Susan Gass - Andrew Cohen (eds.)1994 Research methodology in second-language acquisition. Hillsdale,

NJ: Lawrence Erlbaum Associates.Walters, Joel

1980 "Grammar, meaning and sociocultural appropriateness in secondlanguage acquisition", Canadian Journal of Psychology 34:337-345.

Watson-Gegeo, Karen Ann1988 "Ethnography in ESL: Defining the essentials", TESOL Quarterly

22.4: 575-592.


Wolfson, Nessa1981 "Compliments in cross-cultural perspective", TESOL Quarterly

15.2: 117-124.Wolfson, Nessa - Elliot Judd (eds.)

1983 Sociolinguistics and language acquisition. Cambridge, MA:Newbury House.

Wolfson, Nessa - Thomas Marmor - Steve Jones1989 "Problems in the comparison of speech acts across cultures", in:


Natural speech act data versus writtenquestionnaire data: How data collection methodaffects speech act performance ~~

Leslie M. Beebe - Martha Clark Cummings

1. Introduction

In the early 1980's, when cross-cultural speech act research wasbeginning to take hold (see Kasper 1992 for a complete review of thisliterature), there was a debate raging about the preferred way to collectdata on speech acts. Manes and Wolfson (1980) claimed that the bestapproach was to collect samples of spontaneous speech in natural settingswhere none of the participants was aware of being observed or studied.However, written role play questionnaires (called Discourse CompletionTests) had been and continue to be used extensively to elicit speech actdata across different languages (e.g., Blum-Kulka 1982; Olshtain 1983;Olshtain - Cohen 1983; Blum-Kulka - House - Kasper 1989; Beebe- Takahashi - Uliss-Weltz 1990).

Since the present study was first presented in 1985, studies of crosscultural speech act realization, also known as interlanguage pragmatics(Kasper - Dahl 1991), have still relied heavily on Discourse CompletionTests to collect data. Kasper and Dahl rate discourse completion on thelower end of production tasks used to collect such data, pointing out thatthey are "a much used and much criticized elicitation format in crosscultural and IL [interlanguage] pragmatics" (1991: 221). Nevertheless,they have been used exclusively to collect data in ten studies mentionedin the Kasper and Dahl (1991) review, to wit: studies of requests (BlumKulka 1982; Blum-Kulka - Olshtain 1986; House - Kasper 1987;Faerch - Kasper 1989), complaints (Olshtain - Weinbach 1987),refusals (Takahashi - Beebe 1987; Beebe - Takahashi - Uliss-Weltz1990), corrections (Takahashi - Beebe 1993) and suggestions (Banerjee- Carrell 1988).

As Kasper and Dahl (1991) point out, up to now, few attempts havebeen made to compare data collection techniques. Rintell and Mitchell(1989) compared data collected with Discourse Completion Tests andclosed role plays and found that they yielded very similar data. Kasper

66 Leslie M. Beebe - Martha Clark Cummings

and Dahl (1991) suggest that this is because neither data collectionprocedure is interactive.

Bodman and Eisenstein (1988) compared data collected throughDiscourse Completion Tests, open-ended role plays and field notes onnaturalistic data. The data differed in length and complexity, withDiscourse Completion Tests being shortest and least complex andnaturalistic data most complex.

Beebe and Takahashi (1989a; 1989b) used natural data as a supplement to written discourse completion data in discussing performance offace threatening acts between interlocutors of different status: disagreement, chastisement, and giving embarrassing information, such as tellingothers they have spinach in their teeth.

In an effort to establish the reliability of Discourse Completion Tests,Rose (1992a) compared Discourse Completion Tests with multiplechoice questionnaire data and found significantly fewer hints on theDiscourse Completion Tests than on the Multiple Choice Questionnaires.He did not compare questionnaire data and natural data. In a secondstudy, Rose (1992b) collected data using two types of DiscourseCompletion Tests. "One form included hearer response, while the otherform did not. The two forms were identical in all other respects. Theresults showed that although responses on the No Hearer ResponseDiscourse Completion Test tended to be slightly longer and use slightlymore supportive moves and downgraders, inclusion of hearer responsedid not have a significant effect on requests elicited" (1992 b: 49).

Finally, Dahl (in progress) compared authentic discussions with openended role plays and found, as reported in Kasper - Dahl (1991:244):

The most important features that distinguished between authentic androle play productions across discourse types were amount of talk anddirectness in the performance of face-threatening acts. Amount of talk alsodistinguished the two types of role plays from each other, with the interactive role plays producing less talk and the monologic role plays more talkthan their authentic counterparts.As amount of talk typically distinguishes between different interlocutorrelationships (cf., Wolfson's [1989] bulge hypothesis), and directness interacts with contextual factors in conveying politeness (see Kasper 1990, foran overview), the discomforting conclusion suggested by Dahl's study isthat role plays are not representative of authentic interaction on thesemeasures.However, Dahl emphasized that the way the role plays were elicitedimplied a number of constraints that might have seriously reduced thegeneralizability of her study. Moreover, she warned that the circumstances

Natural speech act data versus written questionnaire data 67

of the data collection might have introduced some extraneous factors thatcould have impaired the validity of the role plays.

This brief review of the literature to date indicates that the debatecontinues over the reliability and validity of Discourse Completion Testdata and that oral role plays, closed or open, do not solve all theproblems inherent in the collection of speech act data. We present ourstudy in support, with certain caveats, of the continuation of DiscourseCompletion Test data collection.

We support the continuation of Discourse Completion Test datacollection, while in fact, we have reservations about all the methods thathave been used to collect data on speech act performance. As we see it,each approach to data collection has strengths and weaknesses. Sincethere have been concerns about the naturalness of discourse completiondata, we offer this chapter as a qualitative and quantitative analysis of thedifferences between written Discourse Completion Test data and naturalspoken data from telephone conversations. In the end, we would like tosuggest that naturalness is only one of many criteria for good data andthat other approaches featuring natural data have drawbacks of otherkinds.

Beebe (1992) discusses the strengths and weaknesses of natural data("ethnographic" and "notebook" data) in a paper on "questionablequestions" - expressions which are syntactically yes-no questions butwhich function as criticisms, topic nominations, complaints, complimentresponses, suggestions, etc. Beebe argues that the weaknesses of writtenquestionnaire data have been widely discussed, but that less attention hasbeen paid to the problems that exist with "ethnographic" data.Ethnographic data may be natural, and natural data may be good in thatthey represent spontaneous natural speech as it really is. But ethnographic data and notebook data are often unsystematic. The socialcharacteristics (e. g., age, socioeconomic status, ethnic group) of the informants are frequently unreported and often unknown. There are vastlydifferent numbers of informants in each social category. The data areunsystematically collected as well. Most, but not all, examples tend tocome from an undefined target population, and the sample population aswell, is often undefined.

The stated goal of ethnographic research on speech act performanceand social rules of speaking is to characterize the sociolinguistic norms ofa "speech community" (in the sense of Hymes 1972a; 1972b; e.g., inWolfson 1983; Daikuhara 1986). However, the family, colleagues,friends, and acquaintances, not to mention the associated strangers,


around a researcher are not necessarily a "speech community." In a largeurban center, the population tends to be very mobile - geographically andsocially - and the circle of friends and colleagues of the researcher willnot necessarily share a speech variety. Furthermore, when the researcher'sgraduate students participate in the data collection (e.g., Manes Wolfson 1981; Holmes 1988), the target and sample populations canbecome even more problematic to define. It is circular to argue that agroup is a speech community because it shares a linguistic variety, and itshares a linguistic variety because it is a speech community.

This concern is voiced in addition to the criticism that the so-called"ethnographic" data, though natural, are not truly ethnographic and thatin the field of sociolinguistics, we have only begun to investigate thesocial rules of speaking in their societal context. In our field, researchersare currently studying spoken data and speculating about the ways theyreflect societal values, but we are not really investigating societal valuesas anthropologists might.

To add to these difficulties, there are problems with tape-recording.What can be taped with approval is a biased subset of the natural speechthat is spoken. Writing down data in a notebook solves these dilemmasto some extent, but presents accuracy problems. Reconstructed dialogand even memorized or immediately recalled data are more likely to beaccurately recorded in terms of pragmatic force than in terms of actuallinguistic structure. At best, with training and practice we can memorizethe core speech act and perhaps a few supporting expressions. The largerlinguistic context must be reconstructed approximately. (See Beebe 1994for a complete discussion of this data collection method.)

2. Method

2.1. Data collection procedures

In the present study, our purpose was to compare data from writtenDiscourse Completion Tests and telephone conversation data taperecorded and transcribed with permission from the respondents. Once thedata were collected, we counted the number of words used by respondents filling out two-turn Discourse Completion Tests and talking on thetelephone, then identified the semantic formulas they used according tothe categories developed in Beebe, Takahashi, and Uliss-Weltz (1990).


2.2. Subjects

There were 22 subjects in the sample reported here. Eleven nativeEnglish-speaking teachers of English as a second language (ESL) in theTESOL Program at Teachers College, Columbia University wereapproached individually and asked to fill out a single written discoursecompletion item. Eleven other native English-speaking teachers of Englishas a second language, all members of New York State TESOL, werecalled on the telephone and asked the same question. Some of these, itturned out, were also students at Teachers College. All 22 subjects werefemale and American.

2.3. The request

In both the Discourse Completion Test and the telephone call, therequester asked the ESL teacher if she would be willing to help out at theTESOL '85 convention in New York. On the Discourse Completion Test,the hypothetical request read:

My name is Susan Miller. I'm calling on behalf of Jim Jenkins and the LocalCommittee for TESOL '85 ... and we're really sort of desperate for volunteers to help out on-site at the convention here in New York. I waswondering, if you haven't already volunteered, if you would like to now ... ?

On the telephone, the requester said the same thing, using her own nameand the name of Jim Lydon, the Local Committee Chair. If the recipientof the call volunteered (approximately twenty did), Cummings gave theirnames and phone numbers to the TESOL '85 Local Committee. If therecipient of the call refused, the requester informed her that she had taperecorded the call for TESOL study and a asked permission to use thedata. The requester continued to make telephone calls until 11 refusalswere collected. All eleven refusers gave their permission.

3. Results

3.1. Amount of talk

Table 1 indicates the differences in the amount of talk between thewritten role plays and spoken responses. (See Appendix for classification


Table 1. Amount of talk in female native speakers' refusals to volunteer atTESOL '85

Context of Request

Characteristicsof Refusals

Discourse CompletionTest (Written)n= 11

Telephone Conversation (Oral)n =11

Total Frequencywords 611 2719sentences 60 229semantic formulas 61 103semantic repetitions 1 31semantic elaborations 1 19turns 20 85

Frequency in First Turnwords 497 386sentences 39 26semantic formulas 42 25semantic repetitions 1 3semantic elaborations 0 3

Average Frequency per Refusalwords 55.54 247.18sentences 5.45 20.87semantic formulas 5.54 9.36semantic repetitions 0.09 2.81semantic elaborations 0.09 1.72turns 1.81 7.72

of refusals). On every measure the total amount of talk on the telephonefar exceeded the amount on the questionnaire. There were more thanfour times as many words spoken, 3.8 times as many sentences spoken,almost twice as many semantic formulas used, and more than four timesas many turns taken. A semantic formula, described by Fraser (1980) andcited by Olshtain and Cohen (1983: 20) "consists of a word, phrase orsentence which meets a particular semantic criterion or strategy,and ... can be used to perform the act in question." Whereas repetitionsand elaborations were extremely infrequent in the written data (oneeach), they were very common on the telephone (31 repetitions andelaborations).


One speaker, for instance, began by saying that the only reason shecouldn't help out was that she had made other plans. On her next turnshe said that the convention was scheduled at a bad time. Finally she indicated that the organization might be at fault, scheduling conferenceswhen certain members were celebrating their religious holidays.

There were no such elaborations on excuses in the written data. Onthe Discourse Completion Test, the layout of the written questionnaireencouraged the respondents to imagine a conversation in which theywould have only two turns. Therefore, it was necessary to include everything of importance they had to say on the imaginary "first turn."Teachers filling out the questionnaire used more words, sentences, andsemantic formulas on turn one than did those responding over thetelephone. It could be argued that the lower overall amount of talkis a function of the smaller number of turns and that if there were anequal number of turns, there would be more comparable levels of talk.On the other hand, it could be argued that the Discourse CompletionTest not only biases the respondent toward packing the whole refusalinto the first turn, but also that the written nature of the task, plus thefact that it is test-like, and the fact that it is imagined, biases respondentstoward an answer that summarizes rather than elaborates and thatresponds definitively rather than hedges and negotiates. _Thus, thekey formulas that are needed to fill the social requirements of the participants are generally stated at once in the Discourse Completion Testresponse.

There is evidence to suggest that the second analysis is the correctone. Written role plays bias the response toward less negotiation,less hedging, less repetition, less elaboration, less variety and ultimatelyless talk. For one thing, the amount of talk (measured in numberof words, sentences, and formulas) sharply fell off in the secondturn of the Discourse Completion Test. There were no repetitions.Only one elaboration occurred. And, out of the average of 55.5words total per refusal, an average of 45.1 of them were used in the firstturn.

The evidence seems to point toward the testing instrument as a biasingfactor. Only one of the eleven Discourse Completion Test respondentsused all of the space provided on the questionnaire. This Teachers Collegestudent might have been eliminated from the sample since she was a closepersonal friend of the data collector and knew that the telephone studywas being conducted. She was kept in the study because she had not seenor heard the data, nor was she acquainted with the classification of


semantic formulas. Interestingly, however, her response was both the onlyquestionnaire to use up all the space and the only one to use a joke foravoidance. (The requester's second turn reads: "Oh, OK. I understand.Well, thanks anyway. Maybe another time." To which the friend replied:"Another time? Hmm. When is the next TESOL in New York? Will Ihave reached retirement?") The friend's response was the only one torequest empathy from the hypothetical caller, and it was the only one torepeat the requester's words as an avoidance strategy. Thus, it seems thatknowing the data collector and knowing the hypothetical item on thequestionnaire was in fact a real question being asked on the telephone byher friend on behalf of the TESOL '85 Committee, the subject made amore realistic response.

This leads us to discuss the Discourse Completion Test in terms ofwhat Wolfson (1981; 1985) called her "extremes follow similar pattern"theory or her "bulge theory." Wolfson (1983: 125-126) (using the termsof Brown and Gilman 1960) in her study of invitations, found that"power, on the one hand, or inequality of status, favors direct invitationsand disfavors attempts at negotiation or expressions of good intent." Shehad "no examples at all of ambiguous invitations given to a superior."Her data showed that "solidarity which leads to reciprocity is, indeed, aprerequisite to the initiation of invitation negotiations." She found thatanother dimension - intimacy - was also important. For example, "incases where participants are intimates who share the same social status,fear of rejection is minimized, and as a consequence, negotiation is oftenunnecessary." Finally, it is with "nonintimates of approximately equalsocial status" - exactly our situation on the telephone - that mostnegotiation takes place. Wolfson (1983; 1985; 1988) claimed that intimates and strangers pattern similarly in all of her work on speech actsbetween native speakers of English.

Our field notes on naturally occurring data support Wolfson'shypothesis. Strangers are brief. If they want to say "no," they do so.Intimates are also brief. It is friends and other acquaintances whoare most likely to get involved in long negotiations with multiplerepetitions, extensive elaborations, and a wide variety of semanticformulas. In the telephone conversations reported here, although theinterlocutors were strangers, there was not one "no" refusal. Only oneoccurred in the questionnaire data for the same request. We wouldcontend they were not strangers in the usual sense, because they sharedwhat Goffman (1967: 109) calls "equal and joint membership in a largeorganization" .


With regard to our data (see Table 1), we wish to make three claims:

1) The discourse Completion Test as a data collection method disfavorsthe long negotiated sequences which occur in natural conversation.

2) Common membership in a known social network (such as New YorkState TESOL) reduces social distance and lends the type of instabilityto the relationship that acquaintances have. It thereby leads them innatural conversation to negotiate in a long sequence of turns and totalk the way they would in conversation with friends and otheracquaintances.

3) If subjects filling out a Discourse Completion Test substitute in theirimagination a known interlocutor for the stranger in the test situation(as did the data collector's friend [see pages 71-72]), this will affectthe length, tone, and other features of the response.

Tables 2 and 3 lead us to further generalizations about the content andtone of refusals in questionnaire versus telephone conversation data.Analyzing the number of subjects who resorted to each formula (orstrategy), we see from Table 2 that the similarities are in many ways morestriking than the differences. No one used the performative "I refuse,"and this, by the way, mirrors our natural participant observation datawhere the performative verb, "refuse", is rare. In addition, there was onlyone instance of a direct "no" in the entire corpus. The adjunct of positivefeeling, the expression of regret, the statement of negative ability orwillingness, and the excuse were the four formulas that both groups usedfour or more times. A very brief response using all four of these formulaswould be, "I'd really love to help out [adjunct] but I'm sorry [regret] Ican't [negative ability] because my family and I are going upstate theweek of the convention [excuse]." These formulas seem to fulfill thestereotypical American requirements for politeness and clarity in situations where specificity is needed and one wants to establish or maintainsome level of rapport. And, these four categories are the ones that arerelatively frequent (used by 1/3 or more of the subjects) for bothquestionnaire and telephone conversations. Thus, the similarities betweennatural spoken refusals and written questionnaire refusals are quitestrong - strong enough to suggest that Discourse Completion Tests area good way to discover what semantic formulas are frequently used (orexpected) in performance of a speech act.

Table 2 also leads us to see differences between questionnaireand natural responses to a request, however. From the complete list ofsemantic formulas, thirteen were never used by a single subject in writing;


Table 2. Number of female native speakers using semantic formulas in refusals tovolunteer at TESOL '85

Context of Request

Semantic Formulas and Discourse Completion Telephone Con-Strategies for Refusal Test (Written) versations (Oral)

n = 11 n= 11

Directperformative verb 0 0no 1 0negative ability/willingness 9 8

Indirectregret 8 4wish 0 2excuse/reason 11 9

alternativeoffer alternative 0 1suggest alternative 1 2

condition for acceptance 2 3promise 1 0principle 0 2philosophy 0 1

attempt to dissuadeguilt trip 0 1criticism 0 3request for empathy 1 4

avoidancenonverbal avoidance 0 1topic switch 0 0joke for avoidance 1 0repetition of request 1 0postponement 1 2hedging 0 7

Adjuncts to Refusalspositive feeling/opinion 6 5empathy 0 2gratitude 1 0self-defense 1 1


only eight were never used by a subject on the telephone. Thus telephoneresponses were not only longer, more repetitive, and more elaborated,they were more varied in terms of the number of different formulas andstrategies resorted to. Houck and Gass (this volume) similarly found thatoral role plays of refusals showed more complex responses thanDiscourse Completion Test data.

Whereas Table 2 gives us the number of subjects who resorted to usinga formula, Table 3 indicates the number of times each formula was used bythe eleven subjects in both the written role play and the actual telephoneconversation settings. Again as in Table 2, the similarities are reassuring forresearchers who use Discourse Completion Tests. The frequency counts forall formulas or strategies, with all the subcategories included, were verysimilar. That is, in only 5 out of 27 formulas, strategies, or even subcategories was there a difference of three or more tokens.

This shows us that in many respects, written questionnaires accuratelyreflect the content expressed in natural speech. Questionnaires yielded 17excuses; telephone conversations contained 16. Questionnaires had 12statements of negative ability/willingness; telephone conversationscontained 14. Questionnaires said "I'm sorry" 11 times; telephoneresponses used it 9 times. The content was in many important ways verysimilar.

We would like to argue that these findings legitimize the use ofDiscourse Completion Test data for certain purposes in sociolinguisticresearch. They indicate that native speakers of a language are in fact ableto write stereotypical responses that reflect the values of the nativeculture. They write refusals which contain an almost formulaic core ofsemantic content that meets the basic social requirements of politenessand clarity. Tannen (1982: 9), in her comparison of recorded spontaneousconversation and transcribed narratives, also found that spoken narrativewas more elaborated, giving more background information, and that"the most striking difference is the increased integration or compactnessof the written text". Furthermore she found that the spoken versionsshowed the speaker's attitude, not explicitly but through paralinguisticcues and repetition, whereas the written texts tended to remain unevaluated and content-focused. This was also true in our data. Speakers tendedto repeat the same phrase four or five times, such as "That's the onlyproblem," or "I don't even know if I'll be here" whereas in writing, eachidea was stated only once.

We are not claiming, however, that Discourse Completion Test data arein any way a substitute for data on natural speech. They are not the same.


Table 3. Total frequencya of semantic formulas used by female native speakers inrefusals to volunteer at TESOL '85

Context of Request

Semantic Formulas and Discourse Completion Telephone Con-Strategies for Refusal Test (Written) versations (Oral)

n =11 n =11

Directperformative verb 0 0no 1 0negative ability/willingness 12 14

Indirectregret 11 9wish 0 2excuse/reason 17 16

alternativeoffer alternative 0 1suggest alternative 1 3

condition for acceptance 2 3promise 1 0principle 0 2philosophy 0 1

attempt to dissuadeguilt trip 0 1criticism 0 6request for empathy 1 6

avoidancenonverbal avoidance 0 1topic switch 0 1joke for avoidance 2 0repetition of request 1 0postponement 2 3hedging 0 15

Adjuncts to Refusalspositive feeling/opinion 7 12empathy 0 3gratitude 1 0self-defense 1 1

a Elaborations and repetitions excluded because there were none in the writtendata.


Nor are oral role plays or even spontaneous responses to a preplannedquestion exactly the same.

We must consider the psychological domain in addition to the strictlysocial or situational setting. It is particularly in the psychological domainwhere the results of this study show differences between writtenquestionnaire and telephone data. Analysis of Table 3 indicated thatalthough most categories of refusal had very similar frequency counts,there were five categories where the questionnaire group versus telephone group displayed a difference in frequency of three to fifteentokens:

1) Avoidance by hedging (15 telephone/ 0 Discourse Completion Test)Example: "I don't know what you mean by volunteering."

2) Request for empathy (6 telephone/ 1 Discourse Completion Test)Example: "I'm very, very tired. I really, really am. I drink a gallon oforange juice a day, I get so thirsty from saying, 'This is a book! This isa book!'"

3) Expression of empathy (3 telephone/ 1 Discourse Completion Test)Example: "This makes your job twice as hard."

4) Expression of positive feeling (12 telephone/ 7 Discourse CompletionTest)Example: "I am a gung-ho proponent of ESL."

5) Criticism (6 telephone/ 0 Discourse Completion Test)Example: "It seems to me we're dealing with so many differentcultures but we're really overlooking our own."

The differences are admittedly small (except in the case of hedging), butthe findings seem important for other reasons. They reflect the psychological (as opposed to the social or situational) domain. That is, they areclosely related to feelings. We would like to claim that the main reasonthe spoken data are different from the Discourse Completion Test data isthat the Discourse Completion Test, a written hypothetical exercise, doesnot bring out the "psycho-social" dynamics of an interaction betweenmembers of a group.

The literature on apologies and remedial exchanges (e.g., Goffman1971; Olshtain 1983; Owen 1983; Olshtain - Blum-Kulka 1985; Trosborg 1987; House 1988; Olshtain - Cohen 1989; Rintell - Mitchell1989; Bergman - Kasper 1993) shows us that the refusals we collectedfrom the telephone were also examples of what Goffman (1971: 139)calls "remedial work," "transforming what could be seen as offensiveinto what can be seen as acceptable". As Goffman (1971: 119) puts it:


"When the individual provides an account or makes an apology, hebecomes needful of the addressee's providing a comment of some kind inreturn; for only in this way can he be sure that his corrective message hasbeen received and that it has been deemed sufficient to re-establish himas a proper person."

Let us look more closely at the categories that were used primarily inthe telephone conversations as opposed to the questionnaires. First of all,hedging - a type of verbal avoidance - occurred 15 times on thetelephone, but never once on the questionnaire. In our data, hedgingappeared to be an avoidance of saying "I don't want to" or "I can't."Seven teachers hedged, most claiming they didn't know where theywould be.

The expression of empathy occurred six times in the telephone data.As one subject said, "Oh, Martha, I really appreciate your fix! (laughs) Ireally do!" Then she proceeded to request empathy by saying she was"over her head" in two similar situations. Later she expressed moreempathy, "I don't envy your task. I commiserate with you." Colorfullanguage, such as "I really appreciate your fix," is typical of real interaction. Only one person expressed empathy on the Discourse CompletionTest and that was the data collector's close friend. It wasn't seen as necessary to express empathy toward a fictional character on paper. However,the more formulaic expressions of positive opinion do occur, particularlycompliments about how wonderful other conventions were. These expressions are also more frequent with real interlocutors where thepsychological dynamics make the refuser want to re-establish rapport.

Although positive feelings seemed appropriate for both questionnaireand real settings, criticism and "guilt tripping" (Example: "You caughtme at a bad time. It's Saturday night, you know. I'm trying to get readyto go out.") occurred only in spontaneous telephone responses. Goffman(1967) points out that when remedial work is attempted and no acknowledgement seems to be forthcoming, the guilty party has no alternativebut to express indignation. One teacher, after insisting three times thather only problem was that she had already made plans to go upstate,decided to let out her real feelings. She argued for approximately sixminutes that TESOL had been insensitive in the scheduling of state andnational events on Jewish holidays. In one short excerpt from hercriticism, she said:

I think it's disgusting. I really do. I mean it's supposed to be this unitedorganization and it's turning ... Whatchamacallit ... It's alienating a wholepart of its constituency. I think that's sad. I really do. I, for one, am turned


off by it myself and I'm not very religious, but I feel for my friends who are,who are constantly being discriminated against. And I'll tell you I think Iwould be more active myself if it wasn't for the way I'm feeling. I just havethis really indifferent feeling now and it's sad.

It is extremely unlikely that a hypothetical situation could evoke suchstrong emotion as the actual scheduling of TESOL during Passover in1985.

Finally, we feel that the telephone conversation data may have inadvertently been biased by us. At the outset of the study we decided to remainneutral by interacting as little as possible with the telephone respondents,giving only the minimal responses of "mm hm," "uh-huh," etc. However,this was not the expected response in a remedial exchange. Owen (1983:57) found that "if one speaker merely acknowledges that remedial workhas been performed, rather than accepting it, it ... has the effect of 'eliciting' further remedial work," which he refers to as "elaborations" or"recyclings. "

Schegloff (1982: 74), too, points out that if we make the sounds "uhhuh" and "mm hm," commonly referred to as "accompaniment signals"or "backchannel actions" which he calls "continuers," the speakerunderstands that we expect her to go on. Not only that, but "in passingthe opportunity to do a fuller turn," we are also "passing the opportunityto do something in particular - the opportunity to do whatever mighthave been relevantly done at that point" (1982: 87).

Tannen's (1982) investigations of different conversational stylesfurther support the notion that by saying "uh-huh" instead of some alternative, expected response, such as "Oh, that's OK," or "That's all right,"or "It's not your fault," we may have inadvertently indicated that wewere not satisfied with the remedial work and wanted more - moreexcuses, more hedging, more elaboration and justification. As Owen(1983: 104) puts it, a pause immediately following an apology or excuseleads the listener to feel "that an acknowledgement or acceptance is beingwithheld, and the inference may be drawn that the remedial work offeredis being rejected." Our telephone respondents, then, may have spoken foran unnaturally long time, or may have used a wider array of excuses thanwould have been necessary had they received the expected response totheir remedial work.


4. Conclusions

Returning to our original research questions, we asked whether questionnaire data were an accurate reflection of spoken data or a useful researchmethod in other respects. In this chapter we argue that DiscourseCompletion Tests are a highly effective research tool as a means of:

1) Gathering a large amount of data quickly;2) Creating an initial classification of semantic formulas and strategies

.that will likely occur in natural speech;3) Studying the stereotypical, perceived requirements for a socially

appropriate response;4) Gaining insight into social and psychological factors that are likely to

affect speech and performance; and5) Ascertaining the canonical shape of speech acts in the minds of speakers

of that language.

However, they are not intended to give us natural speech and they do not"accurately reflect natural speech or even unselfconscious, elicited speechwith respect to:

1) Actual wording used in real interaction;2) The range of formulas and strategies used (some, like avoidance, tend

to be left out);3) The length of response or the number of turns it takes to fulfill the

function;4) The depth of emotion that in turn qualitatively affects the tone,

content, and form of linguistic performance;5) The number of repetitions and elaborations that occur;6) The actual rate of occurrence of a speech act - e. g., whether or not

someone would refuse at all in a given situation.

Thus, we support the continued use of Discourse Completion Tests, whileacknowledging their many weaknesses. They do not give us naturalspeech, nor do they claim to do so. To date, however, many studies ofnatural speech have not given us scientifically collected speech samplesthat represent the speech of any identifiable group of speakers. They donot give us situational control, despite the fact that situation is known tobe one of the most influential variables in speech act performance.Discourse Completion Test data do not have the repetitions, the numberof turns, the length of responses, the emotional depth, or other featuresof natural speech, but they do seem to give us a good idea of the stereo-


typical shape of the speech act - at least in this case of refusals. Since thedata from speech act studies are generally used by teachers andresearchers in TESOL and, more generally, cross-cultural communication, we believe that native speaker perceptions of what constitutes anappropriate refusal, apology, or request is valuable information. We didnot discover a single semantic formula to amplify the classification ofsemantic formulas as a result of collecting natural data. All the semanticformulas had been found in earlier questionnaire data, though not allwere found in the questionnaire data for this study.

In the end, we advocate the comparison of data collected by differentdata collection procedures, and we urge researchers of interlanguage andnative speaker pragmatics to gather data through multiple approachessince each approach has its own strengths and weaknesses.

Appendix

Classification of refusals

I. DirectA. Performative (e.g., "I refuse")B. Non-performative statement

1. "No"2. Negative willingness/ability ("I can't." "I won't." "I don't think so.")

II. IndirectA. Statement of regret (e. g., "I'm sorry "; "I feel terrible ... ")B. Wish (e.g., "I wish I could help you ")C. Excuse, reason, explanation (e.g., "My children will be home that night."; "I

have a headache ... ")D. Statement of alternative

1. I can do X instead of Y (e.g., "I'd rather ... " "I'd prefer ... ")2. Why don't you do X instead of Y (e.g., "Why don't you ask someone else?")

E. Set condition for future or past acceptance (e. g., "If you had asked me earlier, Iwould have ... ")

F. Promise of future acceptance (e.g., "I'll do it next time"; "I promise I'll ... " or"Next time I'll ... " -using "will" of promise or "promise")

G. Statement of principle (e.g., "I never do business with friends.")H. Statement of philosophy (e.g., "One can't be too careful.")I. Attempt to dissuade interlocutor

1. threat or statement of negative consequences to the requester (e. g., "I won'tbe any fun tonight" to refuse an invitation)

2. guilt trip (e.g., waitress to customers who want to sit a while -"I can't makea living off people who just order coffee.")

3. criticize the request/requester, etc. (statement of negative feeling or opinion);insult/attack (e.g., "Who do you think you are?"; "That's a terrible idea.")


4. request for help, empathy, and assistance by dropping or holding the request5. let interlocutor off the hook (e. g., "Don't worry about it." "That's okay."

"You don't have to.")6. self defense (e.g., "I'm trying my best." "I'm doing all 1 can do. " "1 no do

nutting wrong."J. Acceptance which functions as a refusal

1. unspecific or indefinite reply2. lack of enthusiasm

K. Avoidance1. nonverbal

a. silenceb. hesitationc. do nothingd. physical departure

2. verbala. topic switchb. joke

. c. repetition of part of request, etc. ("Monday?")d. postponement (e.g., "I'll think about it.")e. hedging (e.g., "Gee, 1 don't know." "I'm not sure.")

Adjuncts to refusals

1. Statement of positive opinion/feeling or agreement ("That's a good idea ... "; "I'dlove to ... ")

2. Statement of empathy (e. g., "I realize you are in a difficult situation."),3. Pause fillers (e.g., "uhh"; "well"; "oh"; "uhm")

Note: This appendix was originally Appendix C in Beebe, L. M., T. Takahashi andR. Uliss-Weltz, 1990. "Pragmatic transfer in ESL refusals". In Scarcella, R. C.,E. Andersen, and S. C. Krashen (eds.), Developing Communicative Competence ina Second Language. Rowley, MA: Newbury House Publishers, Inc.

Notes

* The authors gratefully acknowledge the assistance of Frances Williams, astudent at Teachers College Columbia University, who helped us with thetabulation of the data, and Ximena Waissbluth, a student at the MontereyInstitute of International Studies, and Naomi Fujita, a student at TeachersColumbia University, who helped us with the preparation of the final draft ofthe manuscript. We also thank Heinle and Heinle for their permission to use theclassification of semantic formulas first published as Appendix C in Beebe,L.M., T. Takahashi, and R. Uliss-Weltz, 1990, "Pragmatic transfer in ESL refusals", in Scarcella, R. C., E. Andersen and S. C. Krashen, (eds.), DevelopingCommunicative Competence in a Second Language. Finally, we thank TomokoTakahashi and Robin Uliss-Weltz who contributed to the development of asystem of classification of semantic formulas for refusals.


References

Banerjee, Janet - Patricia L. Carrell1988 "Tuck in your shirt, you squid: Suggestions In ESL", Language

Learning 38: 313-364.Beebe, Leslie M.

1992 "Questionable questions", Paper presented at the TESOLConvention, Vancouver, B. C.

Beebe, Leslie M.1994 "Notebook data on power and the power of notebook data", Paper

presented at the TESOL Convention, Baltimore.Beebe, Leslie M. - Martha Clark Cummings

1985 "Speech act performance: A function of the data collectionprocedure?" Paper presented at the TESOL Convention, New York.

Beebe, Leslie M. - Tomoko Takahashi1989 a "Do you have a bag? Social status and patterned variation in second

language acquisition", in: Susan Gass - Carolyn Madden - DennisPreston - Larry Selinker (eds.), 103-125.

1989 b "Sociolinguistic variation in face-threatening speech acts", in:Miriam Eisenstein (ed.), 199-218.

Beebe, Leslie M. - Tomoko Takahashi - Robin Uliss-Weltz1990 "Pragmatic transfer in ESL refusals", in: Robin Scarcella - Elaine

Andersen - Stephen Krashen (eds.), 55-73.Bergman, Marc - Gabriele Kasper

1993 "The interlanguage of apologizing: cross-cultural evidence", In:Gabriele Kasper - Shoshana Blum-Kulka (eds.).

Blum-Kulka, Shoshana1982 "Learning how to say what you mean in a second language: A study

of speech act performance of learners of Hebrew as a secondlanguage", Applied Linguistics 3: 29-59.


Ablex.Bodman, Jean - Miriam Eisenstein

1988 "May God increase your bounty: The expression of gratitude inEnglish by native and non-native speakers", Cross Currents 15:1-21.

Brown, Roger - Albert Gilman1960 "The pronouns of power and solidarity", In: Thomas A. Sebeok

(ed.),253-276.Coulmas, Florian (ed.)

1981 Conversational routine: Explorations in standardized communication situations and prepatterned speech. Berlin: Mouton deGruyter.

Dahl, Meretein progress Authentic and role play interaction: A study of method.

[Unpublished Ph.D. dissertation, University of Copenhagen.]


Daikuhara, Midori1986 "A study of compliments from a cross-cultural perspective: Japanese

vs American English", The Penn Working Papers in EducationalLinguistics. 2: 103-134.

Dechert, Hans W. - Manfred Raupach (eds.)1989 Transfer in language production. Norwood, NJ: Ablex.

Eisenstein, Miriam (ed.)1989 The dynamic interlanguage. New York: Plenum Press.

Faerch, Claus - Gabriele Kasper1989 "Internal and external modification in interlanguage request realiz-

ation", in: Shoshana Blum-Kulka - Juliane House - GabrieleKasper (eds.), 221-247.

Fine, Jonathan (ed.) .1988 Second language discourse: A textbook of current research. Vol.

XXV. Norwood, NJ: Ablex.Fraser, Bruce

1981 "On apologizing", in: Florian Coulmas (ed.), 259-271.Gass, Susan - Carolyn Madden (eds.)

1985 Input in second language acquisition. Rowley, MA: Newbury House.Gass, Susan - Carolyn Madden - Dennis Preston - Larry Selinker (eds.)

1989 Variation in second language acquisition: Discourse and pragmatics.Clevedon: Multilingual Matters.


House.Goffman, Erving

1967 Interaction ritual: Essays on face to face behavior. Garden City:Anchor Books.

Holmes, Janet1988 "Paying compliments: A sex-preferential politeness strategy",

Journal of Pragmatics 12: 445 -465.House, Juliane

1988 "'Oh excuse me please ... ' apologizing in a foreign language", in:Bernhard Kettemann - Peter Bierbaumer - Alwin Fill Annemarie Karpf (eds.), 303-327.

House, Juliane - Gabriele Kasper1981 "Politeness markers in English and German", in: Florian Coulmas

(ed.),157-185.1987 "Interlanguage pragmatics: Requesting in a foreign language", in:

Wolfgang Lorscher - Rainer Schulze (eds.), 1250-1288.Hymes, Dell

1972a "Models of interaction of language and social setting", in: JohnMacnamara (ed.), 8-28.

1972 b "On communicative competence", in: John B. Pride - John Holmes(eds.),269-293.

Kettemann, Bernard - Peter Bierbaumer - Alwin Fill - Annemarie Karpf (eds.)1988 Englisch als Zweitsprache. Tiibingen: Narr.


Kasper, Gabriele1990 "Linguistic politeness: Current research issues", Journal of

Pragmatics 14: 193-218.1992 "Pragmatic transfer", Second Language Research 8: 203-231.

Kasper, Gabriele - Shoshana Blum-Kulka (eds.)1993 Interlanguage pragmatics. New York: Oxford University Press.


Language Acquisition 13: 215-247.Lorscher, Wolfgang - Rainer Schulze (eds.)

1987 Perspectives on language in performance: Festschrift for WernerHullen: Tiibingen: Narr.

Macnamara, John (ed.)1972 "Problems of Bilingualism", Journal of Social Issues.

Manes, Joan - Nessa Wolfson1981 "The compliment formula", in: Florian Coulmas (ed.), 115-132.

oIshtain, Elite1983 "Sociocultural competence and language transfer: The case of

apology", in: Susan Gass - Larry Selinker (eds.), 232-249.Olshtain, Elite - Shoshana Blum-Kulka

1985 "Degree of approximation: Non-native reactions to native speech actbehavior", in: Susan Gass - Carolyn Madden (eds.), 18-35.

Olshtain, Elite - Andrew Cohen1989 "Speech act behavior across languages", in: Hans W. Dechert

Manfred Raupach (eds.), 53-68.Olshtain, Elite - Liora Weinbach

1987 "Complaints: A study of speech act behavior among native and nonnative speakers of Hebrew", in: Jef Verschueren - MarcellaBertuccelli-Papi (eds.), 195-208.

Owen, Marion1983 Apologies and remedial interchanges. Berlin: Mouton Publishers.

Pride, John B. - Janet Holmes (eds.)1972 Sociolinguistics. Harmondsworth: Penguin.

Richards, Jack - Richard Schmidt (eds.)1983 Language and communication. London: Longman.

Rintell, Ellen - Candace J. Mitchell1989 "Studying requests and apologies: An inquiry into method", in:


Rose, Kenneth1992a "Speech acts and DCTs: How reliable is questionnaire data?" Paper

presented at the TESOL Convention, Vancouver, B. C.1992 b "Speech acts and questionnaires: The effect of hearer response",

Journal of Pragmatics 17: 49-62.Scarcella, Robin - Elaine Andersen - Stephen Krashen (eds.)

1990 Developing communicative competence in a second language. NewYork: Newbury House.


Schegloff, Emanuel1982 "Discourse as an interactional achievement: Some uses of 'uh huh'

and other things that come between sentences", in: Deborah Tannen(ed.), 71-93.

Sebeok, Thomas (ed.)1960 Style in language. Cambridge, MA: MIT Press.

Takahashi, Tomoko - Leslie M. Beebe1987 "The development of pragmatic competence by Japanese learners of

English", JALT Journal 8: 131-155.1993 "Cross-linguistic influence in the speech act of correction", In:

Gabriele Kasper - Shoshana Blum-Kulka (eds.), 138-157.Tannen, Deborah

1982 "Oral and literate strategies in spoken and written narratives",Language 58: 1-21.

1985 "Silence: Anything but", in: Deborah Tannen - Muriel SavilleTroike (eds.), 93-111.

Tannen, Deborah (ed.)1982 Analyzing discourse: Text and talk: Georgetown University round

table on language and linguistics, 1981. Washington, D.C:Georgetown University Press.

Tannen, Deborah - Muriel Saville-Troike (eds.)1985 Perspectives on silence. Norwood, NJ: Ablex.

Trosberg, Anna1987 "Apology Strategies in natives/non-natives", Journal of Pragmatics

11: 147-167.Verschueren, Jef - Marcella Bertuccelli-Papi (eds.)

1987 The pragmatic perspective. Amsterdam: John Benjamins.Wolfson, Nessa

1981 "Invitations, compliments and the competence of native speakers",International Journal of Psycholinguistics 24. 7-22.

1983 "Rules of speaking", in: Jack Richards - Richard Schmidt (eds.),61-87.

1985 Research methodology and the question of validity. Paper presentedat the TESOL Convention, New York.

1988 "The Bulge: A theory of speech behavior and social distance", in:Jonathan Fine (ed.), 21-38.

1989 Perspectives: Sociolinguistics and TESOL. New York: NewburyHouse.

Part IISpeech Acts in a Second Language

Initiating and maintaining solidarity

Cross-cultural realization of greetings inAmerican English

Miriam Eisenstein Ebsworth - Jean W. BodmanMary Carpenter

1. Introduction

Greetings are among the first speech acts that are learned by children intheir native languages. Dogancay (1990) identifies greetings among theroutines explicitly taught to children. Greetings commonly appear in thespeech of American English-speaking children between the ages of ninemonths to eighteen months. Greeting is important in developing andmaintaining social bonds in all age groups. In fact, greeting rituals havebeen found in nearly all cultures (Levinson 1983). In our study we investigated the use of greetings by native and non-native English speakers inan attempt to better understand their common function.

Simplified greetings are introduced early in most second languagecourses and are often included in texts on cross-cultural communication(Chan 1991; jupp - Hodlin 1983; Morgan 1990). Our research showsthat greetings are complex, involving a wide range of behaviors and asensitivity to many situational and psychosocial variables. Greetings inAmerican English are made up of a range of linguistic and non-verbalchoices which may include a simple wave or smile, a single utterance ora lengthy speech act set which can involve complex int~ractional rulesand take place over a series of conversational turns. This study revealsthat non-natives have significant difficulty in performing greetings in amanner that is acceptable to native speakers of American English.

2. Background

With increased mobility of peoples throughout the world and the breakdown of small, egalitarian face-to-face societies (Gumperz 1982),communicative conventions have become more important in establishingunderstanding and acceptance. Greeting is one of the functions in

90 Miriam Eisenstein Ebsworth - Jean w: Bodman - Mary Carpenter

language that establishes a platform for acceptance creating a positivesocial bond between interlocutors. When it is not performed well, it canresult in confusion, awkwardness and hostility.

Greetings have a high perceptual saliency because they often openconversations. However, most speakers are consciously aware of only asmall number of high-frequency, ritualized semantic formulas thatcontribute to greetings. Despite the fact that some descriptions ofgreetings exist in the literature, there is a significant need for research onhow greetings are truly performed. Scarcella (1979) studied how Anglosversus Hispanics get to know each other and Nine-Curt (1977) reportedon greeting behavior among Puerto Ricans. Scarcella found that topicsdiffered between the two groups she studied, and Nine-Curt found somepatterns that could cause potential misunderstanding. Searle (1969)defines greeting as an illocutionary act, simpler than many other functions because it lacks propositional content and thus has no requirementfor sincerity. Searle is not alone in viewing greeting as a simple utteranceexpressing recognition of another person's presence, which, in English,may also be followed by a phatic expression relating to health and wellbeing. In their 1985 book, TESOL Techniques and Procedures, Bowen,Madsen and Hilferty claim that greetings are part of phatic communication. "Phatic expressions are mostly frozen sentence partials" (p.102)."They communicate attitudes rather than just bare facts" (p. 102). Here,as in other cases in the literature, greeting is discussed as a functionlimited to the act of recognizing someone while making a comment whichis in fact devoid of content. Van Ek (1975) characterizes greeting as asocializing function, but he limits his list to expressions such as "goodmorning," "hello," "how are you?" Wilkins (1976) sees greeting asstraight-forward and highly routine. Bowen and his colleagues go so faras to state that teaching greetings "involves little more than modeling andpractice" (p. 103).

3. Methodology

In the area of cross-cultural pragmatics, there exists a challenge for researchers to capture the authenticity, creativity and richness of naturalspeech while attempting to control the many variables inherent inlanguage use so that data from different individuals can be meaningfullycompared. Although we do not yet have an ideal solution to meet this

Cross-cultural realization ofgreetings in American English 91

challenge, we have come up with an approach that combines naturalobservation and elicited data (Ebsworth 1992; Bodman - Eisenstein1988).

We began by observing greetings among natives and non-natives asthey occurred in natural discourse. On the basis of these natural data, weidentified a great many situations in which different kinds of greetingstypically occurred.

We used this natural discourse as a resource for the creation of anopen-ended questionnaire in which we asked informants to constructdialogues based on some of the situations we identified. Eisenstein andBodman (1986) and Bodman and Eisenstein (1988) demonstrated thatsemantic formulas used in the realization of speech acts could be successfully elicited through carefully constructed role-plays and questionnaires.While elicited data were more limited in demonstrating interactionalpatterns, they did accurately represent some of the langu'age used in natural discourse.

Fifty native speakers of American English provided baseline data forthis study. Second language data came from two groups: One group wascomposed of twenty bilingual graduate students, and the other was composed of eighty adult, advanced level English as a Second Language (ESL)students in a language program at the American Language Institute atNew York University. The first group was composed of individuals whospoke English and one of the following languages: Spanish (two), French(one), Hindi (one), Japanese (six), Mandarin (two), Taiwanese (three),Greek (two), Hebrew (one), Malay (one), Arabic (one). Of the lattergroup, there were twenty Japanese, thirteen Chinese, twelve Russian,eight Spanish, seven Korean, five French/Creole, four Greek and elevenspeakers representing a variety of additional languages. The median ageof the informants (native and non-native) was twenty-three with a rangeof nineteen to sixty-five years of age. The informants were predominantlymiddle class.

Both the bilingual and the non-native English speaking groups wereasked to construct dialogues for seven typical greeting situations inEnglish (see Appendix A); then they were asked to construct dialogues forthe same situations in their native languages and to provide literaltranslations into English. A total of 283 dialogues were collected. Finally,we had native and non-native subjects role play the same situations onvideotape. A total of thirty sets of role-plays were transcribed: Ten dyadsconsisted of native/native pairs; ten dyads were composed of native/nonnative pairs; and ten were non-native/non-native pairs.

92 Miriam Eisenstein Ebsworth - Jean m Bodman - Mary Carpenter

We conducted twenty open-ended post-hoc interviews with participants from a subset of the populations described above to help providean informed interpretation of the data gathered.

4. Results

4.1. Natural data: Native English speakers

Our research leads us to question some of the assumptions in theliterature. While Searle states that there is no sincerity requirement forgreetings, we have found that greetings often convey feelings which arereflected either in the words themselves or the tone of voice (e.g., "Oh,it's nice to see you." or "Hi, how are you?" [warm tone]).

We found many examples in which greetings exhibited attitudes both sincere and insincere. We think the sincerity principle is operativebecause the unmarked case presumes the speaker to be sincere. Thehearer commonly assumes that the speaker wants to be polite inrecognizing him or her. Greetings are often accompanied by a smileand a warm and friendly tone. Yet, sometimes the feelings of the speakersare not what they appear to be. Occasionally, because good manners inpublic require politeness, cordiality will be feigned with a person who isdisliked.

Furthermore, our research does not uphold Searle and Bowen et al.'scontention that greetings communicate "attitudes rather than just barefacts" (p. 103). We have found that greetings contain both attitudes andfacts. As an example of their view, Bowen et al. point out that whensomeone says "How are you?" an honest answer is not expected. But ourobservations show that among friends and even acquaintances, anhonest answer is often given. When two professors encountered eachother in an elevator they greeted each other by Professor A saying,"Hi, how's it going?" and Professor B responding, "Oh, this is myworst day!" In a similar manner, when returning from a trip, a travelerwas greeted by a friend, saying, "Hi! How was your trip?" Thetraveler responded, "Terrible." While the questions opening thegreetings in these exchanges mayor may not have been intended to bephatic, the answers appear genuine since they convey both feeling andcontent.


Many of the writers we have quoted have suggested that greeting issimple compared to other functions; yet, in looking at natural discourse,we find that what people actually say is not always routine. While somegreetings that we observed were simple, short speech acts, many wereextensive and complex speech act sets. These observations wereanalogous to those described for apologies by Cohen and Olshtain(1981), and expressions of gratitude by Eisenstein and Bodman (1986)and Bodman and Eisenstein (1988).

An analysis of our data has led us to identify a number of categoriesof greetings. One style of common American greeting among intimatesand acquaintances is what we have called greetings on the run. Twopeople see each other and exchange brief phatic statements or questionswhich do not necessarily require responses. An example of this wasobserved in the hallway of New York University between a male and afemale student who appeared to be in their late teens. This exchangebegan as the speakers approached each other from different directionsand made eye contact. They exchanged smiles and began talking.Throughout the entire exchange they continued moving past each otheruntil each turned and moved away. The female student began thegreeting by saying, "Hi, how ya doin'?" And the young studentresponded, "Hi! Gotta run. I'm late for class." She smiled at him and saidcheerfully, "Okay!" As in the example above, greetings on the runoften contain an indirect excuse or apology for the shortness of thecommunication. Greetings on the run may also contain a short utterancethat conveys reassurance that no slight is intended or that morelengthy contact is hoped for in the future. Expressions like, "I'll callyou," "See you" or "Talk to you soon" are examples of this kind ofreassurance.

A second greeting that begins and ends abruptly, we have named thespeedy greeting. We differentiate it from greeting on the run becauseinformation is exchanged. This dialogue was recorded between a femaleand male who were colleagues. They were middle class professionals intheir 40's who had known each other for some years. The conversationtook place in an elevator on the way to their offices. The man greeted thewoman by saying,

He: Hi, how've you been?She: Not bad. 'N you?He: Oh, can't complain. Busy.She: I know. Me, too.

94 Miriam Eisenstein Ebsworth - Jean ~ Bodman - Mary Carpenter

He: Oh well, gotta take off. See ya. *

She: Bye. Take care.

The third type of greeting that we identify is called the chat. This greetingstarts off like the speedy greeting, but includes a short discussion on atopic or two before either leave-taking or the real purpose of the communication is introduced. This conversation took place between twofemale friends in their early 20's in an office lounge area.

Female 1: Hi!Female 2: Hi.Female 1: Howa ya doin'?Female 2: All right - comfortable - pretty good. Oh! Got that letter, by

the way, that I said I was waiting for. I finally got it.Female 1: Wow! That's great. That's pretty good.Female 2: Look, I'll see you later.Female 1: Okay. Bye.

The next kind of greeting that we observed we characterized as the longgreeting. One type of this greeting involves re-establishing bonds betweentwo people after a period of separation. This greeting is characterized bya number of greetings separated by narrations of events that occurredwhile the individuals were apart. In this example, two middle-aged,female neighbors greet each other on the sidewalk between their housesafter about a month's separation.

Michelle:Bea:Michelle:Bea:

Michelle:

Bea:Michelle:

Bea:

Bea!Michelle!Where've you been? I haven't seen you around.We were away. We just got back. What's new with you? Whathave you been up to?(Michelle reports on neighborhood news in detail.) We missedyou. How are you? It's so nice to see you. Where'd you go?(Bea describes her vacation in detail.)Well, I'm glad you're back. It's so nice to see you. I missedtalking to you.Aw. Well, we're back! How have you been doing?

* In this and other examples the function of leave-taking is included to show thebrevity and context of the interaction.


An interesting form of greeting occurs between people who know eachother well and speak frequently. We call this the intimate greeting. Thespeakers may know each other so well that they may leave a great dealimplied or unsaid. Sometimes the greeting itself is omitted except for nonverbal gestures. In one example a professor entered the office, walkedover to his secretary, leaned over and said to her, "Should I ask?" Thesecretary shook her head "No." In another example, the husband walkedin, kissed his wife and said, "Well?" She responded, "Yes." He smiled andsaid, "Great. What else did you do today?" She then began to narrate theevents of her day.

Another form of greeting that is sometimes characterized with avery short preliminary greeting or, sometimes, no greeting at all, wehave labelled the all-business greeting. This typically occurs betweenAmericans who have a non-social relationship and, because they believethat the other person's time is limited, show respect and considerationby opening the conversation by immediately stating their business orneeds.

Client: Mr. Matone?Joe Matone: Yes?Client: I want to talk to you about Puerto Rico.Joe Matone: ah? Come in. What about Puerto Rico?

The introductory greeting involves greeting between two people who aremeeting for the first time. It can be said that this greeting is one in whichthe primary function of the interactions is to allow the parties to find aconnection (mutual friends, mutual experiences) or a topic of mutualinterest. The opening is not always a greeting or a formal introduction. Itcan start with a comment.

A: Nice party.B: Yes.A: Who do you know here?B: Bill. I work with him.A: ah. Are you an accountant, too?B: No, I'm in public relations.A: ah. Well, I'm an old friend of Bill's.

There are many greetings that follow predictable patterns for specialoccasions. These greetings may be highly ritualized in routine, frequently


repeated contexts. In service encounters greetings may take the form of"Yes? Can I help you?" (store); "Who's next?", or "Next?" (fast-foodrestaurant); "Fill it up?" (gas station); "Good morning. AT!. [companyname] How may I help you?" (telephone receptionist).

A special category we call the re-greeting, involves acknowledgingsomeone you have greeted earlier or see repeatedly during the day. Thisform of the re-greeting can involve a nonverbal gesture (a nod or a wave)or a few quick words which refer to a shared topic. In this example, oneco-worker, having learned earlier in the day that her co-worker wasnot feeling well, greets another by saying, "Mary? Feel better?" Maryresponds, "Yes. Thanks!"

4.2. Non-native results

We analyzed and reviewed written dialogues constructed by Englishlearners. In these we found that greeting poses many problems for nonnative speakers. A comparison of these data with translations of nativelanguage dialogues of greetings showed evidence of transfer as wellas developmental problems and confusion. Role-plays were viewed andtranscribed. These data were consistent with the written data, but included additional interactive and non-verbal aspects of the greetingroutines. This was also found in the study on expressions of gratitude(Bodman - Eisenstein 1988) and in Beebe and Cummings' (1985) workon refusals. Post hoc interviews with non-natives together with elicitedand observed data reflected many examples of cross-cultural dissonancein the pragmatic systems of American English and other languages andcultures.

The following are some general findings typical of the Englishgreetings of non-natives from a variety of cultures. Non-natives were ableto perform greetings in a manner that was often acceptable to nativelisteners. However, sociopragmatic failure sometimes occurred. The factthat we sometimes observed non-natives performing greetings in Englishsuccessfully may be due in part to the perceptual saliency ofgreetings because they occur at the beginnings of interactions.Slobin (1985) notes that the beginnings of utterances are easier toremember; a parallel may exist between the beginnings of discourse.Furthermore, greetings take place frequently, so there may be moreopportunities for learning how to perform them than other lessubiquitous speech acts.

Cross-cultural realization of greetings in American English 97

Despite these successes there were many examples of greetings madeby non-natives that seemed a bit strange, unusual or "foreign". Alsosome types of English greetings were not well understood by the nonnatives and, as a result, they did not know how to respond. This strangeor unusual "interlanguage-talk" was sometimes a result of the fact thatthe other speaker said something that violated the conventions of thenon-native speaker's first language or surprised them in some manner. Asa result the non-native speaker chose to be creative, idiosyncratic orunusual. In addition, native speakers of English exhibited a greatervariety in the types of greetings and creative language used in producingthe greeting than we might have expected from the literature on thesubject. This was particularly evident when native speakers engaged ininformal or intimate exchanges. Non-natives tended to follow ratherritualized routines and remain formal; apparently, they lacked the repertoire for imitating informal repartee so common in native greetings.

Videotaped role-plays revealed that this verbal formality among nonnatives carried over into their non-verbal behavior. Non-natives tookfixed physical positions in role-plays. On the other hand, native speakersvaried their stance and gestures depending on the level of formality orintimacy required by the setting and topic.

Sometimes the interactions of non-natives during role-plays werejudged by native speakers to be more like interrogations when commentsand expansions on the other speaker's utterances would have been moreappropriate. One non-native speaker tended to take the role of theconversation leader; this speaker asked a question, the other speakeranswered briefly. After a short pause, while the first speaker thought upanother question, the interaction continued. Another non-native featureof the role-plays was the presence of abrupt topic changes. The sum totalwas that the interaction was awkward and unnatural. Wolfson (1989)has commented on the same interactional problem among non-nativeswho have difficulty responding to compliments; Rintell (1989) has madea similar observation concerning the expression of emotion. The sameconversational feature can appear in the speech of native children. Ellis(1984) noted that children exhibited avoidance behavior when theyfound they were unable to keep a conversation going. They changed thetopic abruptly. This made their conversations sound like interrogations.

Two categories of greeting appear to present non-native speakers withspecial problems. The non-natives found both the speedy greeting andgreeting on the run almost impossible to perform. They were unable tomake their greetings short enough. Even when they were given instruc-


tion to make their greetings shorter, they were unable to do so. Here is anexample from our role-play data between two non-native speakers:

A:B:A:B:A:A and B:

I'll come in, President.Yes, please.By the way, what is you have something to discuss with me?Before that, why don't you have a seat.Oh. Thank you.(perform a full greeting)

As receivers of these speedy greetings, they reported feeling that they werebeing treated badly or rudely. Several informants pointed to these typesof greeting as evidence of Americans' lack of politeness, sensitivity andinterest in others. Although we reported above that there were fewerinstances than we expected in our data of the non-natives saying thingsthat were judged to be examples of sociopragmatic failure when theyexpressed or responded to greetings in English, some types of Englishgreetings are received by non-native English speakers as a sociopragmaticfailure on the part of the native English speaker.

There were a number of instances of pragmalinguistic failure (Thomas1983). In a few cases, students used the phrase "How do you do?" whenthey meant to say "How are you?" An inappropriate use of titles wasalso a problem. One example was a non-native speaker saying, "Hi,President" to the head of a company; in another case, a non-nativespeaker greeted a woman whose name was unknown by saying, "Hello,Lady." Interactions in the workplace with persons in authority revealedthat native speakers chose language that was characterized by emotionalrestraint. Occasionally, however, the non-natives who had not yet learnedthe proper register used highly informal language. In several instances wehad reason to believe that the non-natives were trying to lighten a stressful moment by joking, but did not have the sociolinguistic control to doso successfully. Some subtle differences in the data showed approximation but not mastery of conventionalized language. An example of thiswas, "Hi. I do not know you. My name is (name)." Our native informants felt that, "Hi, I don't think we've met. I'm (name)" would havebeen more appropriate. In another example, a native speaker of Japanesewrote the following dialogue:

A: Hello, my fellows. How's your work going?B: Oh. President. So far, so good.


Videotaped role-plays revealed that, in addition to the non-verbalproblems noted above, non-native speech often lacked appropriateprosodic features. In particular, non-natives had difficulty approximatingthe right tone and intensity. They often spoke in a monotone and lackedthe ability to project enthusiasm and warmth.

Our data, as interpreted with the help of native informants, indicatesome problems particular to certain linguistic and cultural groups inlearning to approximate American English greetings. In Puerto Rico,greeting a friend or acquaintance is of such a high priority that on-goingconversations are often interrupted to greet passers-by. Nine-Curt (1977)elaborates on this by noting that speakers are continually looking awayfrom each other to notice others in their vicinity. She refers to this as the"rubber neck" syndrome of Puerto Ricans. When this behavior is transferred to American English, it can be distracting to the American orperceived as impolite to the person with whom one is engaged inconversation. In Puerto Rico, however, it would be socially inappropriateif a person passing by was not acknowledged by his or her friend. AngloAmericans feel their first obligation is to the person who is speaking tothem and not in greeting the friend who is walking by; therefore, theymust maintain eye contact with the speaker in order to show interest andbe polite. Passers-by are expected to understand that if a friend is engagedin a conversation, the friend and the conversation are not to be interrupted. The friend, if he or she wishes to greet this person, will positionhimself or herself just outside of listening distance and in the person'sline of sight and wait to be recognized when there is a pause in theconversation.

As mentioned earlier, our data also demonstrate that the speedygreeting as performed in American English presents difficulties for manyother cultures including Hispanics. When Americans must choosebetween the competing obligations of greeting a friend or leaving the areato do something else, they are able to abbreviate their greeting in amanner that is acceptable to a person who is a member of the samespeech community. This short-cut is understood by both parties and aslight is neither intended or, usually, felt. If the friend demands time totalk or greet, the other interlocutor will usually quickly negotiate a timeto talk at greater length later that day or week. Hispanics appear to findthe speedy greeting next to impossible to perform. For them, friends inone's presence take priority over other obligations, such as imminentappointments (Ebsworth 1992). By extension a third party who is keptwaiting is expected to understand when the excuse is presented that the


person was delayed beca~se they had to say hello to a friend. AmericanEnglish speakers are surprised that when they use the speedy greetingwith Hispanics, the Hispanics can experience it as disrespectful.

Just as length of greeting can vary cross-culturally, the choice of anappropriate topic can be a source of difficulty. Certain topics that arefreely raised in American English greetings have different rules of use inother cultures and languages. It is common to ask about the well-being ofthe person being greeted as well as that person's family members inEnglish and many other languages. Among Arabs, Iranians and Afghans,however, men may ask about the well-being of other male familymembers, but are not as free to inquire about females as is commonlydone in the United States. Our Russian, Ukrainian and Georgian informants stated that greetings among co-workers and acquaintances did notusually contain inquiries about well-being. When Americans greetedthem with expressions like, "Hi. How are you?" they sometimes foundthe question so unexpected and startling that they responded with silenceand an embarrassed expression on their faces. They informed us that theydid not understand why the person wanted to know about their healthwhen they did not know this person well.

As with the Russians, Ukrainians and Georgians who are startled byquestions about well-being, speakers of American English sometimes findthemselves speechless when they are greeted in a number of countries inAsia by two common greetings that translate, "Have you eaten?" and"Where are you going?" One response we observed is that the nativespeaker of English misinterpreted the greeting as an invitation. In theother instance, American English speakers mistook the intent of thegreeting and, when spoken by a non-intimate, felt that the question wasinappropriate and a violation of their privacy.

Among some American subcultures, a frequent strategy for opening aconversation with a stranger is not to greet them, but to state a potentially-shared complaint. Indirect complaints have been identified byBoxer (1993) as performing a bonding function for some Americans. Thisstrategy is shocking to members of other speech communities for whomthis alternative is not available.

The way that greetings are performed can vary from culture to culture.In English, greeting usually involves serial turn-taking. In Afghanistan,both parties often begin greeting each other simultaneously. Questionsabout well-being are frequently not answered, but are overlapped by theother speaker with a similar question about well-being. When a responseis given to "How are you?", it is often simply, "Thank you." English


speakers find this confusing. For Afghans, these ritualized greetings areobligatory each time two friends meet. While English speakers will perform the greeting on the run, Afghans almost always stop walking andperform a full greeting each time they encounter each other during theday. Americans typically wave, nod or say a word or two in subsequentgreetings unless they have a subject to discuss.

Complex rules of non-verbal behavior that accompany greeting suchas bowing, kissing, handshaking and touching vary from culture toculture. Our Japanese informants confided that they were often highlyembarrassed when enthusiastically hugged or kissed by Americans. Thesenon-verbal behaviors often distinguish social classes and generationaldistinctions within a particular culture.

A number of our non-native informants expressed anxiety aboutgreeting people in social settings. They often stated that they did notknow what to say. In examining the models that non-natives are typicallygiven in language classrooms, we found that the ritual of beginning witha "hello", exchanging names and following with asking about well-beingwas universally presented as a model of greeting. However, in our data,much more variety occurs. For example, comments on the food, thepeople, and/or the drinks often precede an exchange of names or personalinformation. Little is available in textbook materials to show learnershow a topic of conversation is mutually developed or how nativespeakers ease into formal introductions. There is not always a good fitbetween American greeting rituals at parties and those common in othercultures. For example, in Swedish, Bratt Paulston (1990) reports that aguest is expected to go around and formally introduce him/herself to allthe other guests while in an American context introductions oftenproceed under the guidance of the host or hostess or are an option, butnot a requirement, for the individual.

5. Conclusion

We have shown that greeting is a socially significant event in universalterms, and that like other major speech acts its realization is languagespecific. Greeting can consist of a single speech act or a speech act set.Successful greetings may be simple or complex, phatic or meaningful,formulaic or creative. Our data show that even relatively advanced nonnative English speakers experience difficulty with various aspects of


American greetings on both productive and receptive levels. Challengesfor cross-cultural communication range from lexical choices to substantial differences in cultural norms and values; thus, pragmalinguistic orsociopragmatic failure may occur in cross-cultural greeting encounters.

A major implication for second language pedagogy is that models forlearning must be based on research into how greetings are actually performed. Regrettably, few current texts for English as a second or foreignlanguage meet this criterion. Furthermore, the complexity and interactivenature of greetings and how they are realized in different languages andcultures must be considered.

In the area of research methodology, the mixed approach to datagathering and analysis taken here is consistent with previous studiesindicating that the semantic formulas that appear in elicited written dataare parallel to those used in role-plays and found in natural observation.The interactive nature of greetings and the combination of associatedverbal and nonverbal elements is most clearly illustrated in naturallanguage use but is effectively mirrored in role-plays. Written data arehelpful in determining the degree to which individuals are aware oflanguage and content appropriate for greeting others in particular sociolinguistic contexts. The interpretation of data through interviews withresearch participants representing the various cultural and linguisticgroups considered is extremely helpful in reaching an understanding ofpotentially problematic group differences and in identifying areas ofshared values and perspectives. It is through such research that we willcontinue to expand our awareness of language functions and improve ourunderstanding of cross-cultural communication.

Appendix A

Questionnaire

For each one of the situations below, write a short dialogue that represents typicallanguage that you and an English speaker would use if you found yourselves inthe situations described.

Situation #1: (Peers greeting each other) Two people who are friends are walkingtoward each other. They are both in a hurry to keep appointments. They see eachother and say:

Situation #2: (Peers greeting each other) Two people who are friends see eachother. They are on their way to other places but are not in a hurry. They have aminute to chat. They see each other and say:


Situation #3: (Peers greeting each other) One friend has been invited to a dinnerparty by another. One friend opens the door. They see each other and say:

Situation #4: (Higher and lower status) An employee has been asked to come tothe boss' office for an unknown reason. The employee knocks and enters theroom. They see each other and say:

Situation #5: (Higher and lower status) A boss happens to be passing by anemployee's work space. With no ulterior motive, the boss stops in. They see eachother and say:

Situation #6: (Stranger to stranger) Two students are sitting next to each other onthe first day of class. One turns to the other, and they say:

Situation #7: (Stranger to stranger) At a party, one person sees another person (ofthe same sex) who looks friendly. They walk toward each other and say:

After you have completed these dialogues in English, please re-do the same dialogues in your native language. Then, if possible, provide a literal translation ofthese dialogues in English.Thank you.

Appendix B

Sample Written Responses

Non-native, Puerto Rican

A: Hi, I don't know you, but my name is Provy.B: Are you enjoying yourself? Very well organized party, isn't it?

Native English speaker

A: Hi!B: Hi, how you doin'?A: Okay. You?B: Okay. Catch you later.

Native English speaker

A: Ah, hello, (name of boss). You wanted to see me?B: Yes. Come in. How are you?A: Fine, thanks. How are you?B: Fine. Have a seat.

Non-native, Israeli

A: Hi, Batya. How's school?B: Baruch Hashem. How's your husband feeling?


A 42-year old native speaker of English:

A: Great party!B: Yeah. The food is terrific.A: The Smith's always go all out.B: Oh? Is this an annual event?A: Sort of. Are you new in the neighborhood?B: Yeah. We just bought the house next door.A: Well, welcome to the neighborhood. My name is Steve. I live at 677.

Sample natural data

Two students, native English speakers

A: How are you?B: Fine.A: Take care.B: Okay.

Clerk, Customer, (strangers), native English speakers

A: Hi, how are you?B: Good. It's nice to see the sunshine at last.A: Right.

Sample transcripts of videotaped role-plays

Two Japanese women, college graduate students, non-native speakers

A: Hi.B: Hi.A: How're ya doing?B: Thank you, I'm fine. How are you?A: Yeah, I'm pretty good.B: I haven't seen you around two years, right?A: Is that so?B: Yeah, that long. (they giggle) Where are you going?A: I'm going to the Metropolitan Museum ...

Israeli, Jamaican women, non-native speakers

A: Hi. I never thought you have time for those romantic restaurants.B: Aha - Oh, I love to eat. I love to go to restaurants. Fancy seeing you here.A: I love eating, too. That's one of my main problems. You know that.B: Yes, I do. Oh, this is my husband. Jack, this is (name).


A: It's nice meeting you. That's my boyfriend.B: Oh, nice meeting you.A: That's my boss.B: I see they're summoning us to the table.A: Enjoy your meal.B: Bye.A: Bye.

Japanese, Israeli women, non-native speakers

A: Excuse me, ma'am, but do you have the time?B: I see that you have the time.A: (Looking at watch on her own wrist) Why, it's broken.B: Oh ... It's broken. Well, it's seven past two.A: Thank you very much. Do you like the picture, Picasso?B: I like the picture, but it's not Picasso, I'm very sorry to tell you. It's Van Gogh.A: Oh, I didn't know that.B: There is a little difference between them.A: I didn't recognize this picture. I like the picture of impressionist.B: But this is not impressionist, I'm sorry to tell you. This is not, I'm sorry to tell

you. This is a classical picture from the nineteenth century.A: Thank you very much for your kind lecture.

Native man and woman

A: Hi, Steve.B: How are you doing, Adina?A: Okay.B: What's happening?A: Vh ... vacation was really good.B: Oh, yeah? Where'd you go?A: Vh, skiing.B: Lucky you!A: Yeah, and you?B: Just stayed in town, bored to tears.A: That's too bad.B: I know. I gotta get out of teaching. There's more money elsewhere. Oh, are you

taking a class this semester?A: Oh, yeah, four ...B: Four! Are you out of your mind?A: Yep.B: And working, too? I don't know ...A: Yep.B: It's crazy.A: Yeah. Tell me about it.


References

Beebe, Leslie - Martha C. Cummings1985 "Speech act performance: A function of the data collection

procedure?" Paper presented at the Sixth Annual SociolinguisticsColloquium, TESOL Convention, New York City.

Bodman, Jean - Miriam Eisenstein1988 "May God increase your bounty: The expression of gratitude in

English by native and non-native speakers", Cross Currents 15:1-21.

Bowen, J. Donald - Harold Madsen - Ann Hilferty1985 TESOL Techniques and Procedures. Rowley, MA: Newbury House.

Boxer, Diana1993 Complaining and commiserating: A speech act view of solidarity in

spoken American English. New York: Peter Lang Publishing.Bratt Paulston, Christina

1990 "Linguistic and communicative competence", in: Robin Scarcella Elaine Andersen - Stephen Krashen (eds.), 287-302.

Chan, .Carole1991 Latinos in the work force: Diversity and tradition. Los Angeles, CA:

Los Angeles County Commission on Human Relations.Cohen, Andrew - Elite Olshtain

1981 "Developing a measure of sociocultural competence: The case ofapology", Language Learning 31.1: 113-34.

Dogancay, Seran1990 "Your eye is sparkling: Formulaic expressions and routines in

Turkish", Pennsylvania Working Papers in Educational Linguistics6.2: 49-64.

Ebsworth, Timothy J.1992 Appropriateness of Puerto Rican/American cross-cultural

communication. [Unpublished Ph.D. dissertation, New YorkUniversity.]

Eisenstein, Miriam (ed.)1989 The dynamic interlanguage. New York: Plenum Press.

Eisenstein, Miriam - Jean W. Bodman1986 "I very appreciate: Expressions of gratitude by native and non

native speakers of American English", Applied Linguistics 7.2:167-85.

Ellis, Rod1984 Understanding Second Language Acquisition. New York: Oxford

University Press.Gumperz, John J.

1982 Discourse Strategies. Cambridge: Cambridge University Press.JuPP, Thomas C. - Susan Hodlin

1983 Industrial English. London: Heinemann Educational Books Ltd.Levinson, Stephen C.

1983 Pragmatics. New York: Cambridge University Press.


Morgan, Graham1990 "Anti-racism and language diversity: Raising metalinguistic aware

ness to combat racism", Reading 24.3: 192-203.Nine-Curt, Carmen J.

1977 Nonverbal communication. Cambridge, MA: National Assessmentand Dissemination Center for Bilingual Education, Lesley College.

Rintell, Ellen1989 "The use of language to express emotion by second language

learners and native speakers", in: Miriam Eisenstein (ed.), 237-260.Scarcella, Robin

1979 "On speaking politely in a second language", in: Carlos Yorio et al.(eds.),275-287.

Scarcella, Robin - Elaine Andersen - Stephen Krashen (eds.)1990 Developing communicative competence in a second language. New

York: Newbury House.Searle, John R.

1969 Speech acts: An essay in the philosophy of language. Cambridge:Cambridge University Press.

Slobin, Daniel1985 The cross-linguistic study of language acquisition. Hillsdale, NJ:

Lawrence Erlbaum Associates.Thomas, Jenny

1983 "Cross-cultural pragmatic failure", Applied Linguistics 4.2: 91-112.van Ek, Jan A.

1975 The threshold level. London: Longman.Wilkins, David A.

1976 Notional syllabuses. Oxford: Oxford University Press.Wolfson, Nessa

1989 Perspectives: Sociolinguistics and TESOL. New York: NewburyHouse.

Yorio, Carlos - Kyle Perkins - Jacquelyn Schachter1979 On TESOL C79: The learner in focus. Washington, DC: TESOL.

Egyptian and Anterican contplintents: Focus onsecond language learners)f-

Gayle L. Nelson - Waguida £1 Bakary Mahmoud Al Batal

1. Introduction

Not only did I need to know the right words. I needed to know the appropriate situations in which blessings were given and the appropriatemoments at which to give them (Spindel 1989: 215).

Carol Spindel lived in a small African village in the Ivory Coast and studied the Dyula language in order to interact with those around her. Theview she presents, that language learners need to know more about thetarget language than the phonological, morphological, and syntacticrules, more than the "right words", is well known to applied linguists andlanguage teachers. In large part due to the work of Canale andSwain (1980: 30-34), Hymes (1972: 269-288; 1974: 10-24), Wolfson(1981: 117; 1983: 82-83; 1989: 32-53), and language teachers havebecome aware of the notions of communicative competence, thatlanguage must be not only linguistically accurate, but socially appropriate, and that sociolinguistic rules sometimes need to be explicitlytaught. This realization has led to the study of language in use, to thestudy of the sociolinguistic rules as well as the linguistic rules of language.One method of investigating the sociolinguistic rules of a given speechcommunity is to identify and study specific speech acts within thatcommunity.

The term "speech act" has been defined as a minimal unit of discourse,a basic unit of communication (Searle 1969: 16). Examples of speech actsinclude giving compliments, making statements, asking questions,apologizing, leave-taking, making introductions, making requests,expressing gratitude, making refusals, and, as illustrated above, givingblessings. This chapter presents the results of a cross-cultural study on thespeech act of complimenting, comparing various aspects of AmericanEnglish and Egyptian Arabic compliments. With a focus on secondlanguage learners, it extends an earlier discussion of these data (Nelson- £1 Bakary - Al Batal 1993).

110 Gayle L. Nelson - Waguida El Bakary - Mahmoud Al Batal

2. Rationale

Within the communicative competence paradigm, a primary rationalefor studying speech acts is to obtain sociolinguistic knowledge of the rulesof the target language. According to this view, the study of speechacts, such as compliments, is of interest to language teachers in order forthem to instruct students in the socially appropriate uses of complimentsin the target language. It may be difficult, however, for students toachieve communicative competence in a second language due to thetransfer of sociolinguistic rules from their first language to their secondlanguage.

The notion of transfer or interference, originally used to describe thephenomena of phonological and syntactic transfer from a first languageto a second language, was adopted by sociolinguists as it became clearerthat "rules governing speech events may differ substantially from onelanguage group to another, thus leading to different rules and norms forturn taking, amount of talking, speech act realizations, etc." (Schmidt Richards 1980: 146). Using the "rules governing speech events" fromone's first language speech community when interacting with members ofa second language speech community is referred to as pragmatic transfer.Pragmatic transfer may lead to pragmatic failure, to a negative judgmentabout a speaker such as his or her being impolite or uncooperative (Leech1983: 281). It is hoped that by comparing speech acts across cultures,miscommunication resulting from the pragmatic transfer of first languagerules to second language speech situations can be predicted and prevented(Wolfson 1989: 140).

Speech act studies have been criticized as being ethnocentric in thatmost have investigated variations of English (Blum-Kulka - House Kasper 1989: 10). The present study is valuable, in part, because it wasconducted in Arabic as well as English and the results contribute to anunderstanding of the sociolinguistic rules of Arabic, an understudied areaof speech act research.

3. Compliments

The speech act of complimenting was selected for cross-cultural study fortwo reasons. First, American compliments tend to be a "troublesomeaspect of English for learners from different cultural backgrounds"

Egyptian and American compliments 111

(Holmes - Brown 1987: 525). For example, non-native speakers ofEnglish are often embarrassed by the frequency with which Americanscompliment (Holmes - Brown 1987: 525; Wolfson 1981: 123). Asecond reason is that, although extensive research has been conducted onAmerican compliments, few cross-cultural studies have investigatedcomplimenting (exceptions include Barnlund - Araki 1985; Wolfson1981). For the purpose of this study, a compliment is defined broadly asan expression of praise or positive regard.

Wolfson and Manes (1980), using ethnographic methodology, havecollected over 1000 American compliments in a wide range of situations.They (Wolfson 1981: 122; Wolfson 1983: 85; Wolfson - Manes 1980:402-403) found that approximately 80 % of American compliments fallinto three syntactic patterns:

Syntactic PatternsNP is/looks (intensifier) AD]I like/love NPPRO is AD] NP

ExamplesYour shoes are great.I love your perm.These are great cookies.

Two of these three patterns depend on adjectives for their positivesemantic value and two thirds of the adjectival compliments use one offive adjectives: nice, good, beautiful, pretty and great. With regard toattributes praised, Americans most frequently compliment personalappearance and ability.

Wolfson (1981: 120) has also noted cultural differences incomplimenting and observed that Iranian and Arabic speakers tend to useproverbs and other precoded ritualized expressions when complimenting.She gives the example of an Arabic speaker complimenting a friend·'schild. The English equivalent is: "She [the child] is like the moon and shehas beautiful eyes".

Holmes and Brown (1987), also using ethnographic methodology,collected 200 compliments in New Zealand. Their results were similar tothose of Wolfson and Manes. Almost 80 % of the compliments belongedto one of the above three syntactic patterns; two thirds used one of fiveadjectives: nice, good, beautiful, lovely and wonderful; and the mostfrequently praised attributes were personal appearance and skill.

Barnlund and Araki (1985), using interviews and questionnaires,investigated Japanese and American compliments. Interviewees wereasked to describe 1) the most recent compliment they had given andreceived, 2) the relationship between the giver and recipient of the

112 Gayle L. Nelson - Waguida EI Bakary - Mahmoud Al Batal

compliment, 3) the attribute praised, 4) the exact words used in thecompliment, and 5) the day the compliment was given. Interview dataindicated that the japanese compliment much less frequently thanAmericans. The compliments reported by the japanese occurred, on theaverage, 13 days before the interview, whereas the compliments reportedby Americans occurred, on the average, 1.6 days earlier. Their findingsalso indicate that japanese and Americans tend to compliment fiveattributes: appearance, work and study, personal traits, skill and taste,but with varying frequencies. japanese most frequently praise skill (31 %)and work and study (19 0/0), whereas Americans most frequently praiseappearance (34 %) and personal traits (33 0/0).

Knapp, Hopper and Bell (1984) used a similar method andasked subjects to describe recent compliments given and received andalso to provide information related to the compliment giver andrecipient. Data were analyzed for 1) attributes praised, 2) complimentforms, and 3) relationships between giver and recipient of compliments.Appearance, attire and performance were the most frequently praisedattributes. Compliment forms were analyzed according to syntacticpatterns and also along four dimensions: direct/indirect, specific/general, comparison/no comparison and normal/amplified. Seventyfive per cent of the compliments followed one of the three patternsidentified by Wolfson and Manes (1980: 402-403). Complimentstended to be direct, general, non-comparative, l and normal (withoutintensifier). In their analysis of the relationships between the giversand recipients of compliments, Knapp et al. (1984: 26) found thatcompliments are likely to be exchanged between individuals of thesame sex, and between individuals in close, rather than distant, relationships.

Before describing the present study, it is important to discuss twoimportant facets of complimenting in Arabic. The first concerns the beliefin the evil eye. The evil eye refers to the "belief that someone can projectharm by looking at another's property or person" (Maloney 1976: v).Frequently, the evil eye relates to "envy in the eye of the beholder" and ismost dangerous to pregnant women, children, and anyone who isbeautiful (Spooner 1976: 77). For example, if a person compliments amother on her child, the compliment, by causing the evil eye to notice thechild, may cause harm to visit the child. To counteract this effect, thegiver of the compliment invokes Allah to protect the child, saying, AllaahyiHfazu ('May God protect him') or maa shaa'a Allaah! ('What God haswilled!'). In a study of pregnant women at the American University in


Beirut hospital, Harfouche (1981: 87) found that 54.9 % believed in theharmful effects of the evil eye. 2

Another important facet of Arabic complimenting is the practice ofoffering the object of the compliment to the person who complimented.For example, if Mohammed compliments Sami on his new cassette tape,Sami might say basiita; itfaDDal! ('This is nothing; please take it!')or m~addam ('It is offered'). This practice of offering the object of thecompliment to the giver of the compliment seems, however, to be moreproblematic for non-native Arabic speakers than native Arabic speakers.Arabic speakers recognize this offering as a ritual and do not take itliterally whereas non-native Arabic speakers or English speakers frequently accept the literal meaning and thus are either reluctant tocompliment, or embarrassed when the Arabic-speaker offers them theobject of the compliment.

4. The study

This study investigated Egyptian and American compliments todetermine similarities and differences in 1) compliment form, 2) attributes praised, 3) gender of the compliment giver and recipient, and 4)compliment frequency.

4.1. Method

An expanded version of Barnlund and Araki's (1985) procedure forobtaining compliment data was used. Audiotaped interviews wereconducted with 20 Egyptian university students in Egypt and 20American students in the United States. All students were between 18 to25 years of age; half were males and half females. Interviewees wereasked to describe the most recent compliment they had given, received,and observed, the relationship between the complimenter and therecipient, the attribute praised, the exact words used in the compliment,and the day the compliment was given. Interviews were conducted in thenative language of the interviewees. This procedure provided a corpus of60 American and 60 Egyptian compliments.

This method of obtaining data has three advantages. First, the corpusof compliments represents actual compliments given. Second, the


interviewee defines what constitutes a compliment. The interviewee maybe in a better position to define a compliment within a particular sociocultural context than researchers or coders. For example, one complimentreported by an interviewee in this study, Show off macho, may have beenperceived as a criticism by researchers, but within the subculture of whiteAmerican male university students, this utterance was given as a compliment. One male was complimenting another on his waterskiing.

The third advantage is that the interviewee, as the giver, receiver, andobserver of the compliment, decides when the compliment ends. Thisending point is not as apparent as one might expect. For example, anAmerican male reported the compliment in (1).

(1) You-'re really sweet and nice. I really appreciate your help.(AM8)

The first part of this utterance seems an obvious compliment, but thesecond part (I really appreciate your help) could be viewed as an expression of thanks. The interviewee, however, perceived the second segmentas part of the total compliment.

At the completion of the interviews, the audiotapes were transcribed;the American compliments were transcribed in English and the Egyptianin Arabic. The Arabic transcripts were also translated into English for theanalysis of attributes praised, gender of compliment giver and recipient,and compliment frequency. The analysis of the form of the Arabiccompliments, however, was based on the Arabic transcripts, not theEnglish translations.

4.2. Results

4.2.1. Compliment form

Compliment form refers to the language used to express the compliment.The compliments differed in their length, use of metaphor and comparatives, and to some extent, syntactic structure. They were also similar inthat both Egyptian and American compliments were primarily adjectivalin that an adjective was responsible for their positive meaning.


The American compliments tended to be short, as in (2) through (4).

(2) You look great. (AF4)

(3) Your perm is nice. (AF7)

(4) Good job. (AM29)

In this particular corpus of American compliments, the average numberof words in the compliments reported by American females was 6.2 andthe average for American males was 4.3.

The Egyptian compliments were longer, as in (5).

(5) eeh sh-shiyaaka di! eeh l-fustaan da! bass, iHna ma-ntdarsh.(EF21)'What is all this chicness! What is this dress! Stop, we cannot[take all that].'

The average compliment reported by Egyptian females contained 10.7Arabic words and by Egyptian males 8.7 words.

The length of the Egyptian compliments appears to be related to twofeatures of Arabic discourse: 1) Repetition of almost the same idea witha change in words and 2) The use of several adjectives in a series.Examples of repeating a similar sentence are in (6) and (7).

(6) eeh l-Halaawa di! eeh sh-shiyaaka di! (EMI0)'What is all this beauty! What is all this chicness!'

(7) waliid SaaHbi w-akhuuya. waliid ana caarfa akhlaa~u. waliidakhlaa'u kwayyisa w-mafiish aHsan minnu. shaab akhlaa'ukwayyisa, sum'a kwayyisa, mafiish Hadd byitkallim caleeghalaT. huwwa kamaan biy'aamil n-naasmu'amla SaHH.mafihuush Haaga tit'aayib. l-'eeb l-waHiid illi fiih innu gaayizyi'mill-'amal wi ma y caddarsh eeh illi yiHSal. (EM17)'Walid is my friend and [like] my brother. Walid, I know hismanners well. Walid has good manners and there is no onebetter than him. He is a young man whose manners are good,and no one says anything bad about him. He also treats peoplewell. He does not have any shortcomings. The only fault he hasis that he may do something and not consider the consequences.'

116 Gayle L. Nelson - Waguida El Rakary - Mahmoud Al Ratal

An example of an Egyptian compliment that used a series of adjectivesis in (8).

(8) inti insaana 'add eeh kwayyisa wi zarifa wi 'a~la wi Hakiima wirazina wi zakiyya wi labiqa wi mish na~Sik Haaga. (EF8)'You are such a good and nice and sensible and wise andserious and intelligent and diplomatic person and don't lack athing.'

In 15 out of 60 or 25 % of the Egyptian compliments, the speakerrepeated the same idea or used a series of adjectives when praisinganother person. Only 5 % of American compliments used such patterns.

American and Egyptian compliments also differ in their use ofcomparatives, particularly similes and metaphors; 11 % of the Egyptiancompliments used comparatives whereas comparatives were not used atall in the American compliments.

Examples of Egyptian compliments using comparisons include (9)through (11).

(9) huwwa zayy akhuuya. (EM17)'He is like a brother to me.'

(10) A: fustaanik Hilwawi.B: el-Ha~ii~a tasriHtik aHla. (EF30)A: 'Your dress is very nice.'B: 'Really, your hair style is nicer.'

(11) amiiSak Hilw zayy bitaa~ MiHammad Fu~aad. (EF2)'Your shirt is as pretty as [the singer] Mohammed Fouad's.'

Marriage is a common metaphor in Egyptian compliments as illustratedby examples (12) through (14).

(12) shaklak 'ariis in-naharda. (EM8)'You look like a bridegroom today.'

(13) eeh sh-shiyaaka di! iHna Hlawweena awi. fii miin fi-TTarit?(EF4)'What is all this chicness! We have become very pretty! Who ison the way? [on the horizon]?'


(14) mulfita li-n-naZar giddan! eeh sh-shiyaaka di! eeh l-fustaan da!bass! iHna ma-ni'darsh (;ala kida. shaklik in-naharda zayyl-'aruusa. (EF20)'[You look] very attractive! What is all this chicness! What isthis dress?!Stop! We cannot take this. You look like a bride today.'

Seven percent of the Egyptian compliments used a marriage metaphor.No American compliments referred to marriage. An analysis of thesyntactic structure of the compliments indicated both similarities anddifferences. Sixty-six per cent of the American compliments used one ofthe three syntactic patterns identified by Wolfson and Manes in theirresearch on American compliments (NP is/looks AD]; I like/love NP;PRO is AD] NP). Similarly, a limited number of syntactic patternsaccount for the majority of the Egyptian compliments:

Syntactic PatternsNP AD] (intensifier)

NP VP (intensifier) (AD])

eeh NP DEM. PRO.

Examplesinti shiik awi.'You are very chic.'(;ineeki Hlawwit awi.'Your eyes have become very beautiful.'eeh sh-shiyaaka di!'What is all this chicness.'

Fifty percent of the Egyptian compliments used the syntactic pattern: NPAD] (intensifier). This pattern is similar to the American pattern: NPis/looks (intensifier) AD]. The Arabic version contains no verb "to be"because in Arabic, the verb "to be" is not used in a nominal sentence;thus, inti Hilwa means "you are beautiful". Fourteen percent of thecompliments used the second pattern: NP VP (intensifier) (AD]).

The third pattern (eeh NP Dem. Pro.), used in 14 % of the Egyptiancompliments, does not appear in the American data. It represents onetype of a precoded set of phrases that is used in particular situations thatdemand complimenting; these phrases or formulas cannot be changed.

Three of the Egyptian compliments made reference to Allah (e.g., maashaa'a Allaah (;aleek; 'God's grace be upon you'), whereas no Americancompliments referred to God.

The Egyptian and American compliments shared one major similarity:they were primarily adjectival. Seventy per cent of the Egyptian compliments and 73 % of the American compliments used adjectives. Four


Arabic adjectives, Hilw ('pretty'), Kwayyis ('good'), shiik ('chic'), andTayyib ('kind'), accounted for 66 % of the Arabic adjectives used. Themost frequent was Hilw which was used in 34 % of the adjectival compliments. Three English adjectives, great, good, and nice were used in74 % of the American adjectival compliments.

4.2.2. Attributes praised

In order to develop a means of classifying the compliments according tothe attributes praised, the researchers read the compliments and notedpossible classifications. Based on the researchers' notes, a classificationscheme with four categories (appearance, traits, skill, and work) waspiloted by two coders. The coders, however, were, at times, unable todistinguish between the two categories "skill" and "work". Asa result ofthe pilot test, the classification scheme was modified. The categories"skill" and "work" were collapsed into one category, "skiIVwork". Next,compliments were classified independently by two coders who were different from the first two coders. This second pair of coders classified thecompliments as belonging to one of the three categories: appearance,traits, or skill/work.

The category, "appearance", referred to one's looks and included haircuts, eyes, and clothing. "Skill/work" referred to the "quality of something produced through ... skill or effort: a well-done job, a skillfullyplayed game, a good meal" (Manes 1983: 101). "Traits" referred topersonality characteristics such as loyalty, kindness, maturity and intelligence. An intercoder reliability of 97 % was determined by comparing theclassifications of the coders.

As shown in Table 1, personal appearance was frequently praised inboth cultures; it was praised in SO % of the Egyptian compliments and43 % of the American. The attribute of physical appearance is, however,not as straightforward as it first appears. Manes (1983: 99) found thatAmericans tend to compliment "aspects of personal appearance whichare the result of deliberate effort, not simply natural attractiveness." OurAmerican data support Manes; only one compliment praised a naturalattribute. A male praised a female friend, saying, You have nice-lookinglegs.


Table 1. Attributes praised in Egyptian and American compliments

Egyptian

Appearance(Natural)30*(15)

50%*(250/0)

American

Appearance(Natural)26*(1)43%**(1.6%)

Traits

20

33%

Traits

6

10%

SkillslWork

7

12%

SkillslWork

28

47%

Other

3

5%

Other

o

0%

Total

60

100%

Total

60

100%

* number of occurrences out of 60 compliments** percentage (based on 60 compliments)

However, half of our Egyptian compliments on appearance praised natural attributes. Some of the compliments are exemplified in (15) and (16).

(15) sha'rik gamiil. (EF25)'Your hair is beautiful.'

(16) ~ineeki Hilwa awi. (EM12)'Your eyes are very pretty.'

Consistent with the work of Manes (1983: 98), Americans in the presentstudy tended to compliment clothes and hair. Egyptians also complimented clothes and hair, but in addition, they complimented eyes, skin,and general attractiveness.

The second largest category of Egyptian compliments was "traits"which accounted for 33 % of the Egyptian corpus. Examples are in (17)through (19).

(17) shakhSiyyitik kwayyisa. (EMll)'Your personality is good.'

(18) inti ~a'la, wi zakiyya wi diblumasiyya. (EF8)'You are mature and intelligent and diplomatic.'


(19) inti Tayyiba wi shakhSiyyitik kwayyisa. (EF16)'You are very kind and your personality is good.'

"Traits" accounted for only 10 % of American compliments. The largestcategory of American compliments was "skills/work" which accountedfor 47% of the American corpus. Examples of American compliments on"skill/work" include examples (20) through (23).

(20) You did a great job. (AF19)

(21) This paper is really good. (AM12)

(22) Nice catch. (AM21)

(23) You guys are doing a great job. (AM13)

Twelve percent of the Egyptian compliments praised skills or work.

4.2.3. Gender of compliment giver and recipient

As shown in Table 2, of the compliments reported, 61 % of those givenby Egyptian females and 62 % given by American females praised appearance. Forty-three per cent of the compliments given by Egyptian malespraised appearance, whereas only 29 % of those given by American malespraised appearance. Both Egyptian and American males praised femaleson appearance more frequently than they praised males on appearance.Egyptian males praised personality traits in 41 % of the compliments andAmerican males praised skill or work in 590/0. American males weretwice as likely to praise males on "skill/work" as females.

4.2.4. Compliment frequency

The interview data indicate that the Americans complimented morefrequently than the Egyptians. Americans reported compliments thatwere given, on the average, 1.6 days before the interview. Egyptiansreported compliments that were given, on the average, 8.6 days beforethe interview.


Table 2. Attributes praised according to nationality and gender of giver andreceiver of compliments

Egyptian

Females

Appearance Traits SkillslWork Other Total

Female to 8 (35%) 4 (180/0) 3 (13%) 1 (4%) 16 (690/0)Female

Female to 6 (260/0) 1 (40/0) 0 (00/0) o (00/0) 7 (31%)Male

Total 14 (61 %) 5 (22%) 3 (13%) 1 (4%) 23 (1000/0)

Males

Male to 12 (32 %) 8 (220/0) 0 (0%) 2 (5%) 22 (59%)Female

Male to 4 (11 %) 7 (19%) 4 (11 %) 0(0%) 15 (41 %)Male

Total 16 (430/0) 15 (41 0/0) 4 (11 0/0) 2 (50/0) 37 (100%)

American

Females

Appearance Traits SkillslWork Other Total

Female to 14 (54%) 0 (0%) 4 (15%) 0(0%) 18 (690/0)Female

Female to 2 (80/0) 2 (8%) 4 (15%) o (00/0) 8 (31 %)Male

Total 16 (620/0) 2 (8%) 8 (300/0) 0(0%) 26 (100%)

Males

Male to 6 (18%) 2 (6%) 6 (180/0) o (00/0) 14 (42%)Female

Male to 4 (120/0) 2 (6%) 14 (41 0/0) o (00/0) 20 (59%)Male

Total 10 (290/0) 4 (12%) 20 (590/0) o (00/0) 34 (100%)


5. Discussion

In order for students to become communicatively competent in a secondlanguage, they need to learn both the linguistic and sociolinguistic rulesof conversational discourse. However, achieving communicative competence may, at times, be complicated due to the transfer of rules fromstudents' first language to their second language. One of the goals ofcross-cultural studies such as this one is to predict the inappropriatetransfer of first language rules to second language situations and toprovide "a basis for determining which areas are most vulnerable to miscommunication and should therefore be focused on" (Wolfson 1989:140).

The results of this study provide linguistic and sociolinguistic information about Arabic compliments that can be helpful to both instructors ofEnglish as a second language who teach Arabic speakers, and to studentsand teachers of Arabic as a second language. Linguistic and sociolinguistic knowledge of Arabic compliments can help teachers of Englishas a second language 1) predict areas where Arabic speakers may havedifficulty, 2) recognize when Arabic speakers are transferring nativelanguage conventions to their use of English, and 3) understand thereasons why Arabic speakers make certain linguistic and sociolinguisticerrors. Conversely, linguistic and sociolinguistic knowledge of Englishcompliments can help teachers of Arabic as a second language 1) predictareas where English speakers may have difficulty, 2) recognize whenEnglish speakers are transferring first language conventions to secondlanguage situations, and 3) understand the reasons why English speakersmake certain linguistic errors or act inappropriately in some sociolinguistic contexts. Such sociocultural knowledge may also result in lesspragmatic failure, fewer cross-cultural misunderstandings, and improvedintercultural communication.

It seems reasonable to assume that areas of similarity betweenAmerican and Arabic compliments will cause language learners less difficulty than areas of difference. Similarities include the speech act itself;complimenting seems common in both cultures. Both cultures also sharesimilarities in compliment form and attributes praised. Over 70 % ofboth the Arabic and English compliments were adjectival; they dependedon an adjective for their positive semantic value, and both used a limitednumber of adjectives. Consistent with the findings of other researchers(Wolfson - Manes 1980: 400), this study suggests that Americansfrequently use the adjectives good, nice, and great. Common Arabic


adjectives include Hilw, Kwayyis, shiik and Tayyib. Both Egyptian andAmerican compliments also tended to use a limited number of syntacticpatterns, and the most frequent pattern in both sets of data was similar:NP is/look (intensifier) AD] for American compliments and NP AD](intensifier) for Egyptian compliments. Finally, in both cultures, malesand females tended to praise females on their personal appearance,particularly on their clothes and hair. Students of English and Arabic canuse these similarities between Egyptian and American compliments totheir advantage by learning the limited number of adjectives used in complimenting, the common syntactic patterns that are similar in bothlanguages, and the attributes that are praised in both cultures.

Although the two cultures appear to share some similarities in compliment form and attributes praised, they also seem to differ in these twoareas. With regard to compliment form, the Egyptian compliments in thisstudy tended to be longer than the American compliments. This length isrelated in part, to repetition, a feature of Arabic discourse (Suleiman1973: 292). Arabic speakers use repetition to express their feelings; themore something is valued, the more the repetition. With compliments, themore the repetition, the better the compliment. Other facets of Arabiccontributing to compliment length are long arrays of adjectives (Shouby1951: 291) and elaboration (Almaney - Alwan 1982: 83-84).

The findings of this study suggest that Egyptians also tend to use moresimiles and metaphors than Americans. Again, similes and metaphors arenot particular to compliments; they are common in Arabic (Shouby 1951:298-299). In this particular corpus of Egyptian compliments, marriage isa frequent metaphor, perhaps suggesting not only the importance ofmarriage in Egyptian society, but the centrality of the family.

The American compliments, on the other hand, are more likely to beshort and less often include metaphors and similes. For Arabic-speakingstudents learning English, it may appear that American complimentswould be simple and easy to learn, but although the syntax seems simple,such plain, unelaborate utterances may be difficult for an Arabic speakerbecause they seem inadequate; they do not seem to express what he or shewants to say. Compliments such as Nice shirt may be perceived by Arabicspeakers as flat and relatively meaningless. If Arabic speakers, in anattempt to make compliments sound sincere to their own ears, use morewords than would a native English speaker, "pragmatic failure mightresult from overindulgence in words," causing native speakers to sense alack of appropriateness (Blum-Kulka - Olshtain 1986: 175). Englishspeakers, on the other hand, may have difficulty with the repetition and


metaphor in Arabic compliments, feeling phony and insincere whenrepeating the same compliment in several ways. In this case, pragmaticfailure may result not from too many words, but from too few.

Wolfson (1981: 19) notes that Arabs compliment in the form of"proverbs and other precoded ritualized phrases." She uses a proverb asan illustration:

Speaker 1: X is a nice girl and beautiful.

Speaker 2: Where is the soil compared with the star?

In line with Wolfson's findings, it was expected that the Egyptian data inthis study would contain a few proverbs. This was not the case. The datadid, however, illustrate the frequency of precoded set formulas in Arabiccompliments (e. g., eeh il-Halaawa di! eeh sh-shiyaaka di! 'What is all thisbeauty! What is all this chicness.'). These qualitative findings are consistent with the quantitative findings of Nelson, EI Bakary and Al Batal(1993) in which the third most popular compliment form chosen byEgyptians was "use a formulaic expression."

The results of this study also suggest sociolinguistic differencesbetween Egyptian and American compliments in factors such as attributes praised, gender, and frequency of complimenting. Consistent withother studies (Barnlund - Araki: 1985: 13-14; Knapp et al. 1984:17-18; Manes 1983: 98-102), these findings suggest that Americanstend to compliment skills and personal appearance, and are likely to compliment someone of the same gender (Knapp et al. 1984: 26). This studyalso suggests a relationship between the gender of the compliment giverand recipient and the attribute being complimented. For example, theAmerican males tended to compliment other males on skills and work,whereas American females tended to compliment other females onappearance. Like Americans, Egyptians tended to compliment personalappearance, but unlike Americans, Egyptians, particularly males, tendedto compliment males and females on personality traits. Neither Americanmales nor females were particularly likely to compliment individuals ontheir personalities. Both second language Arabic and English teachers canhelp non-native speakers become more communicatively competent byassisting them in becoming more aware of appropriate attributes topraise and the role of gender in complimenting.

A final difference between American and Egyptian complimentsappears to be the frequency with which individuals compliment. In this


study, Americans seemed to compliment more frequently than Egyptians.This tendency of Americans to compliment frequently is problematic tomany non-native speakers of English who are embarrassed by what theyperceive as excessive complimenting. If Egyptians compliment less oftenthan Americans, they may perceive American compliments to beinsincere. Such an interpretation may lead to a communication breakdown. Let's suppose, for example, that an American student or teacher,in an attempt to be friendly and open a conversation, repeatedly compliments an Egyptian in class. If the Egyptian student perceives the compliment as insincere, he or she may feel uncomfortable and pull awayfrom the person who complimented. The result may be that, instead ofbecoming friendly, the Egyptian becomes distrustful of the American.Second language English teachers can help prevent such misunderstandings by structuring classroom activities so that students learn thatAmericans compliment frequently and that compliments serve functionssuch as maintaining social harmony, opening conversations, andgreetings (Wolfson 1983: 87-90).

Ideally, speech acts should be studied in their natural context usingethnomethodology (Wolfson 1983: 94). Ethnomethodology is difficultfor cross-cultural studies due to problems of comparability (Blum-Kulka- House - Kasper 1989: 13) and a lack of ethnographers who belongto speech communities other than English-speaking ones. Although thecompliments in this study were not observed under natural conditions,our findings for native speakers of English are similar to complimentscollected by Wolfson and Manes in natural settings. This similaritysuggests the reliability of this study's methodology and data. However,the method has limitations. Because these compliments were not collectedin a natural setting, it is difficult, if not impossible, to answer thequestion: What functions do compliments serve in Arabic? Other questions also exist - How does power distance and status interact withcomplimenting behavior? How does familiarity interact with complimenting? What are common responses to compliments?

In addition to the above questions, two other facets of Arabic complimenting did not appear in the data. When an Arabic speakercompliments an individual, particularly a pregnant woman or child, thespeaker often uses an invocation such as Allaah yiHfazu ('May Godprotect him') to protect the individual from bad luck. For non-nativespeakers of Arabic, the omission of such an invocation may producemisunderstanding, and for non-native speakers of English, the transfer ofthis rule to English may also result in misunderstandings. An ethno-


graphic study of compliments in various contexts would provideinformation on this practice that would be helpful to both secondlanguage Arabic and English students.

A second facet of Arabic complimenting that did not appear in thedata relates to the practice of offering the object of a compliment to theperson who complimented. Although it is assumed that this is a commonpractice in Arabic-speaking countries (Almaney - Alwan 1982: 98-99),little is known about the conditions surrounding its use. The ethnographic study of Arabic compliments in context would contribute to ourunderstanding of these situations and would help students of Arabic learnwhat these situations are, what to say, and how to respond.

These questions suggest fruitful areas of additional inquiry. This studyrepresents a first step in the study of complimenting in Arabic, butadditional steps are needed to paint a more complete picture of Arabiccomplimenting, to understand the appropriate situations in whichcompliments are given and the appropriate moments at which to givethem.

Notes

* This chapter is based on data from an ongoing research project comparingArabic and English speech acts such as compliments and complimentresponses. The authors would like to thank Joyce Neu and John Murphy fortheir helpful comments on versions of this chapter.

1. Examples of comparative compliments are You play better than Jerry andYou ere prettier than ever.

2. We also have the evil eye phenomenon in the U. S. (and other Western cultures)when we say "Knock on wood" to maintain good luck.

References

Almaney, A.J. - A.J. Alwan1982 Communicating with Arabs. Prospect Heights, IL: Waveland Press,

Inc.Barnlund, Dean C. - Shoko Araki

1985 "Intercultural encounters: The management of compliments byJapanese and Americans". Journal of Cross-Cultural Psychology 16:9-26.

Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper1989 "Investigating cross-cultural pragmatics: An introductory over

view", in: Shoshana Blum-Kulka - Juliane House - GabrieleKasper (eds.), 1-34.


"The influence of the Arabic language on the psychology of theArabs", Middle East Journal 5: 284-302.

1974


Ablex.Blum-Kulka, Shoshana - Elite Olshtain

1986 "Too many words: Length of utterance and pragmatic failure",Studies in Second Language Acquisition 8: 165-179.

Canale, Michael - Merrill Swain1980 "Theoretical bases of communicative approaches to second language

teaching and testing", Applied Linguistics 1: 1-47.Dundes, Alan (ed.)

1981 The evil eye: A folklore casebook. New York: Garland Publishing.Harfouche, Jamal Karam

1981 "The evil eye and infant health in Lebanon", in: Alan Dundes (ed.),86-106.

Holmes, Janet - Dorothy F. Brown1987 "Teachers and students learning about compliments", TESOL

Quarterly 21: 523-546.Hymes, Dell

1972 "On communicative competence", in John B. Pride - Janet Holmes(eds.),269-293.Foundations in sociolinguistics: An ethnographic approach.Philadelphia: University of Pennsylvania Press.

Knapp, Mark L. - Robert Hopper - Robert A. Bell1984 "Compliments: A descriptive taxonomy", Journal of Communica

tion 34: 19- 31.Leech, Geoffrey N.

1983 Principles of pragmatics. London and New York: Longman.Manes, Joan

1983 "Compliments: A mirror of cultural values", in: Nessa Wolfson Elliot Judd (eds.), 96-102.

Maloney, Clarence (ed.)1976 The evil eye. New York: Columbia University Press.

Nelson, Gayle L. - Waguida El Bakary - Mahmoud Al Batal1993 "Egyptian and American compliments: A cross-cultural study",

International Journal of Intercultural Relations, 17: 293 -313.Pride, John - Janet Holmes (eds.)

1972 Sociolinguistics. Harmondsworth, Middlesex: Penguin.Prosser, Michael H. (ed.)

1973 Intercommunication among nations and people. New York: Harperand Row.

Schmidt, Richard W. - Jack C. Richards1980 "Speech acts and language learning", Applied Linguistics 1: 129-157.

Searle, John R.1969 Speech acts. Cambridge: Cambridge University Press.

Shouby, E.1951


Spindel, Carol1989 In the shadow of the sacred grove. New York: Vintage Books.

Spooner, Brian1976 "The evil eye in the Middle East'" in: Clarence Maloney (ed.),

76-84.Suleiman, Michael W.

1973 "The Arabs and the West: Communication gap", in: Michael H.Prosser (ed.), 278-303.

Wolfson, Nessa1981 "Compliments in cross-cultural perspective", TESOL Quarterly 15:

117-124.1983 "An empirically based analysis of complimenting in American

English", in: Nessa Wolfson - Elliot Judd (eds.), 82-95.1989 Perspectives: Sociolinguistics and TESOL. New York: Newbury

House/Harper & Row Publishers.Wolfson, Nessa - Elliot Judd (eds.)


Wolfson, Nessa - Joan Manes1980 "The compliment as social strategy", Papers in Linguistics:

International Journal of Human Communication 13: 391-410.

Politeness strategies in French and English

Michael L. Geis - Linda L. Harlow

1. Introduction

Anyone who wishes to learn to speak a second language faces twoproblems. The first is to learn the language system. The second is to learnto use the language. Since learning a language is a precondition onlearning to use it, much of the energy in second language education isdevoted to the former task. However, unless students are also taught howto use the second language, we may expect that they will fall back onlanguage patterns acquired in the course of learning to use their firstlanguage to guide how they use the second. The inevitable result is thatstudents will commonly employ forms that are contextually inappropriate in that they differ in style, politeness, and register from what nativespeakers would employ.

The problem language learners face is that there is an extraordinaryamount of stylistic, politeness, and register variation in any language,and, therefore, in any speech context, there will be a large number ofutterances available to him or her which communicate what he or shewants to communicate, but which will differ in style, politeness, andregister, and therefore differ interactionally.

In this chapter, we report on a pilot study we have done concerninghow English, French, and English-speaking French learners do requestingin an informal experimental conversational context (in which subjectsengage cooperatively in assembling jigsaw puzzles). The goal of the studywas to determine which English and French forms are politeness equivalents and to what degree French learners employ forms used by Frenchspeakers. Before discussing the study, and its implications for teachingFrench to speakers of English, it will be necessary to present our theoretical assumptions in some detail.

In order to understand the dimensions of the problem secondlanguage learners face, it will be useful to consider a hypotheticalexample drawn from English. Suppose John goes to a store to

130 Michael L. Geis - Linda L. Harlow

purchase "widgets," where he encounters Sue, who is a clerk, and theconversation in (1) ensues:

(1) John: Ya got any widgets?Sue: Yeah.John: Why don~t you give me three?Sue: Sure.

Now, at turn T1, John could have said any of the utterances of (2) by wayof determining whether or not Sue has widgets for sale.

(2) a. Have you got any widgets?b. Ya got any widgets?c. Got any widgets?

Or, he could have said any of the sentences of (3).

(3) a. Do you have any widgets?b. Ya have any widgets?c. Have any widgets?

The same is true of (4), and stylistic variants thereof.

(4) a. Do you carry widgets? (Ya carry widgets?, Carry widgets?)b. Do you stock widgets? (Ya stock widgets?, Stock widgets?)

All of these sentences ask whether or not Sue has widgets available forsale, and will be said therefore to be transactionally equivalent. However,they differ significantly from each other in style and register, and someone learning English must learn when any particular form will and willnot be contextually felicitous.

The three utterances of (2) do not appear to differ at all in literalmeaning (L-Meaning), as is characteristic of stylistic variants. We mightrepresent the L-Meaning of these sentences informally as in (5), where"Sp" stands for the speaker and "Ad" for the addressee.

(5) Sp inquires of Ad whether or not Ad possesses widgets.

The sentences of (3) also do not differ from each other in L-Meaning, andarguably do not differ in L-Meaning from those of (2) either, which is to

Politeness strategies in French and English 131

say not only that the variation in (3) is stylistic, but that the choicebetween using have and got is also stylistic.

Proposition (5), though it accounts for the literal meanings of (2) and(3), does not adequately represent how they will be interpreted in context, for John is not asking simply whether Sue possesses widgets butwhether she possesses them for sale. If one goes into a store that has lightbulbs illuminating the store, but does not sell them, and asks whetherthey have lightbulbs, the normal answer will be, "No," even though thatis technically false. 1 The reason is that clerks hear utterances like (2)and (3) as having a meaning in context or S-Meaning (for contextualsignificance), something like that represented in (6).

(6) Sp inquires of Ad whether or not Ad possesses widgets for sale.

In short, clerks hear utterances like (2) and (3) as having essentially thesame meaning in context as the utterances of (4), forms that are restrictedin use to the commercial register, despite the obvious differences in literalmeaning. 2

One of the reasons for positing the two levels of discourse meaning,L-Meaning and S-Meaning, is that we cannot speak meaningfully oflinguistic variation unless we can say that we have something that variesand something that stays the same. In the case of stylistic variation, whatstays the same is L-Meaning and what varies is linguistic form. However,register variants clearly do not have the same L-Meaning (have and gotdo not have the same L-Meaning as carry and stock) and so we must lookto a different level of meaning to express how register variants are thesame, namely S-Meaning. The same is true when we come to politenessvariation as we shall shortly show.

What was said of turn T1 goes equally for T3. Sp might have said anyof the sentences of (7) at T3 by way of signaling that it is three widgetsthat Sp desires.

(7) a. I'll take three.b. Give me three.c. Why don't you give me three?d. How about three?e. Could I have three?f. Could you give me three?

The utterances of (7) clearly differ from each other in L-Meaning.However, they would do precisely the same work at turn T

3, namely


communicate that the number of widgets desired is three. As such, theycould be argued to have the same significance in context or S-Meaning(i. e., are transactionally equivalent) were they to be uttered at turn T

3.3

The forms in (7) are neither stylistic, nor register variants. Instead,they differ in politeness. Utterances (7a) and (7e) differ from (7b)-(7d)and (7f) in that the former are egocentric (the subject is I) and the latterexocentric (the subject is you - an understood you in the case of (7b)).All things being equal, exocentric forms are more polite than are egocentric forms. Utterances like (7c)-(7f) can be said to show deference tothe addressee in that they suggest that the addressee has an option as towhether or not he or she will comply with the request, even though he orshe very well may not have such an option. These utterances are therefore more polite than utterances like (7a) and (7b) that do not showdeference. In English, utterances displaying deference are interrogative(cf., Can I have three?), while those that do not display deference areeither imperative (Give me three) or declarative (I~ll have three).

The politeness interrogatives of (7) are worthy of special note.Questions can be used to request information, of course, but there arevery clear cases, such as these, in which they are being used actually toprovide information (namely, that it is three widgets that John wants),not request it.4 Suppose, as another example, that John goes into a frozenyoghurt store and notes that the store offers large dishes of pineappleyoghurt for sale at a price he can afford. At such a point, he could useany of the sentence forms in (7) to frame his request (substituting "a largepineapple" for "three," of course). In this case, it would be clear thatnone of the interrogative utterances would be used to request information- John has all the information he needs to place his order. Instead, he isusing these utterances to provide information - the information that hewants a large pineapple yoghurt. It is when we find interrogatives beingused to convey information, rather than request it, that we have clearcases of what we might call "politeness interrogatives".

As we have seen, someone learning English has to cope with a verygreat deal of stylistic, register, and politeness variation. He or she mustlearn what register, style, and politeness options (sociopragmatic options)are available, how these options are realized morphosyntactically andprosodically, and when any given option will be appropriate. The samewould be true, of course, in learning another language. To complicatematters, different language cultures 5 may vary in what sociopragmaticoptions are available, and even when the same options are available, theymay take different morphosyntactic and prosodic forms.


Suppose, for instance, that we have two students who are studyingtogether in informal circumstances and one wants to borrow a piece ofpaper from the other. A French student might use either a negative orpositive declarative intonation question like those in (8) by way ofmaking a polite, deferential request, whereas an American student wouldmost likely use the interrogative form (9).6

(8) a. Tu n~as pas de papier??'You don't have any paper?'b. Tu as du papier?'You have some paper?'

(9) Do you have any paper?

This contrast between French and English is quite striking, for though theFrench do have an inverted interrogative form like (9), namely (10), it isvery rarely used in informal contexts (Desirat - Harde 1988; Gadet1989; Di Vito 1991, Valdman 1967), and the literal English equivalent to(8a), namely (11), would never be used to make a polite request.

(10) As-tu du papier?

(11) You don't have any paper?

Instead it would be used to confirm a speaker's belief (presupposition)that the addressee does not in fact have any paper. 8 Confusion may resultwhen an English-speaking learner of French is the recipient of a negativedeclarative intonation question like (8 a). He or she may transfer nativelanguage sociopragmatic knowledge and take it to be a confirmation ofthe speaker's belief that he or she does not have the item in question(which might be quite baffling) or, it may be interpreted as an abrupt,somewhat impolite request (You don~t have any paper to give me? Wellyou should!), rather than the polite request that it is in French.

Accordingly, if we are to teach students how to converse in a secondlanguage we must teach them not only how to produce and understandutterances in that language that are linguistically well-formed, but alsothat are contextually felicitous. That is, students must learn to produceutterances that are consistent with the discourse context and that employstylistic, register, and politeness features that are consistent with thesocial context and the relative statuses and roles of participants. But


before we do this, we must attempt to determine which forms in the firstand second languages are stylistic, politeness, and register equivalents.Empirical studies produced by native speakers in context are necessary inorder to complement theoretical studies based on intuition (Blum-Kulka- House - Kasper 1989: 3), for native speakers' intuitive understandingof how speech acts function within their native languages have not alwaysbeen proven to reflect observed speech behaviors (Wolfson 1983: 83).

In our pilot study we were specifically interested in finding out whichsyntactic patterns are employed in making requests by native French andEnglish speakers in an attempt to determine which features of context(especially politeness features) seem most to govern utterance construction in the two languages in this context. We were then interested incomparing utterances of native French speakers with those of Englishspeakers learning French to determine to what degree any differencesfound could be attributed to sociolinguistic factors. Accordingly we wereless interested in what we might call "mechanical" errors of phonology,morphology, and syntax than in differences of sociolinguistic interest,especially cases of contextual infelicity.

In order to compare the productions of speakers of different languages it is necessary to have some means of determining when we havecross-linguistic functional equivalents. In general, two utterances will bewhat we might call "conversational equivalents" in two languages if theyhave the same significance when used in the same context (S-Meaning).We must therefore be clear on what it means for utterances to beS-Meaning equivalents.

In many cases, especially when asking a straight-forward factual question (What is the capital of Ohio?) or making a straight-forward assertion (Columbus is the capital of Ohio), the significance or S-Meaningof the utterance is essentially the same as its L-Meaning. Cases like thesewill rarely present significant problems to language learners. The problemarises in the case of utterances exhibiting significant stylistic, politeness,or register influences. In what follows, we shall focus on one such classof utterances, namely, utterances used in requesting that can be said tohave illocutionary significance. One way in which an utterance can haveillocutionary significance is by inquiring as to whether some condition ona speech act is satisfied (Would you like a glass of water?) or assertingthat it is satisfied (I would like a glass of water). We shall refer toillocutionary significance in this sense as "transactional illocutionarysignificance". Another way in which an utterance can have illocutionarysignificance is by doing"face work" (Goffman 1967), perhaps by redres-


sing a face-threat (Brown - Levinson 1978) associated with the action.We shall refer to illocutionary significance in this sense as "interactionalsignificance". Let us briefly discuss these two notions.

As Searle (1969) observed, for a request to be felicitous, certain conditions must be satisfied. His conditions on requesting (cf., (12)) will befamiliar to those who have read in the speech act literature.

(12) Searle's Conditions on Requests (H is the hearer and S is thespeaker)Propositional Content Condition: Future act A of H.Preparatory Conditions:

H is able to do A. S believes H is able to do A.9It is not obvious to both Sand H that H will do A in thenormal course of events of his own accord.

Sincerity Condition: S wants H to do A.Essential Condition: Counts as an attempt to get H to do A.

And, we would expect utterances used in requesting normally to addressthese conditions in one way or another. Those that do can be said to havetransactional illocutionary significance.

Geis (1995) has argued that in order to apply speech acttheory to the analysis of how we do requesting in conversation,including, in particular, how we do requesting in multi-turn interactions like (1), it is necessary to revise Searle's speech act theory inone quite critical way. Instead of saying what speech act a specificutterance performs, a practice that sheds little light on multi-turninteractions, we shall say what condition it addresses (if any). Geis alsoargues that it is necessary to revise Searle's speech act structures in certainways.

Since multi-turn conversational sequences do not have a (single) propositional content, we must abandon the propositional content conditionin favor of a statement of the domain of the request, which consists of aspecification of the properties of some requested thing or action. Asecond major revision concerns Searle's sincerity condition. In somecases, it identifies the psychological state from which the initiator of therequest launches an action, as is true of requests, but in other cases it doesnot. Thus, the sincerity condition on asserting something is that thespeaker believe that what is being asserted is true, rather than that thespeaker desires to cause the addressee to believe that the assertion is true.We shall adopt the view that speech acts 10 are subject to an initial state


condition that always specifies the psychological state from which theinitiator initiates the action.

Searle's preparatory conditions are defective in a different way. Notethat Searle includes both a condition that the hearer be able to performthe desired action and that the speaker believe the hearer to be able toperform the action. The former condition is what we might call a "satisfaction" condition, for it is concerned with speech act success; the latteris a "felicity" condition in that it is concerned more with speech actappropriateness. We shall adopt the view that speech act structuresshould be restricted to satisfaction conditions, with the issue of felicitybeing reserved to the issue of utterance planning (i. e., to the issue whetheror not the initiator of an act should initiate the act). Additionally, as iscommon in speech act theoretic work, we shall add to the ability condition identified by Searle, a willingness condition stipulating that theresponder must be willing to perform an action before the request can besuccessful.

One last revision is in order. Searle's essential condition identifies thedesired effect of a request - to cause the addressee to perform someaction. However, the actions we perform have other effects as well, including, in particular, what we might call the interactional effects of theacts. Brown and Levinson (1978) have argued that people have two facewants, a negative-face want not to have one's freedom of action impededand a positive-face want to be valued and to have what one values bevalued. They also argue that many types of speech acts are face threatening (FTA), and that when performing an FTA we have the option toredress the face threat. Thus, requesting someone to do somethingthreatens their negative face. It also threatens the initiator's positive facebecause of the possibility that the request be denied. We shall refer to theface threats associated with an act as the interactional effects of the act.

We may now summarize our revisions of Searle's statement of the conditions on requesting, where "Init" is the initiator 11 (i. e., the beneficiary)of the request and "Resp" is the responder, and the action"A" involvesthe responder's transferring some object to the initiator.

(13) Init Requests Resp to Do AInitial State Condition: Init desires that Resp do A C 1

Illocutionary Effects:Transactional Effect: To cause Resp to do A.Interactional Effect: A threatens Resp's negative face andInit's positive face


Satisfaction Condition:Resp is able to do A C2

Subcondition:Res!) Possess X C3

Resp is :J:illing to do A C4

Domain (A):' Resp transfers X to Init Domain (X): ..... 12

The basic idea is thaf any utterance that occurs in a request sequence,however short or long'it may be, will be expected to address one or moreelements of such a ~tructure. Utterances in different languages thataddress the same elements of such a structure in the same way (includinginteractionally) will be: said to be conversational equivalents.

Brown and Levinson (1978: 273) noted that theories of indirect speechacts "appear to have lpissed the extremely systematic way in which therational strategies of face redress, like pessimism and hedging, are able topredict the internal structure of polite indirect requests." Brown andLevinson do not provJlde a formal account of how what we might call"politeness features" determine the internal structures of utterances.What we shall argue is that the internal structures of utterances involvedin requesting, like those in (7), are a compositional function of whatthe speaker wishes to communicate - which elements of speech actstructures the utterance addresses - and politeness features. We shallrestrict our attention here to the morphological, syntactic, and lexicalproperties of utterances.

What politeness features operate in English and French is a largelyopen question. We shall assume the system in (14).

(14) A. Orientation Features: determine the subject of the utterance.1) Init-oriented. For cases of utterances instantiating the initial

state condition, of which there are two subtypes in English.a) [nit-specific. Utterances that are desire-specific will be

those like [ want a corner piece, [ would like a cornerpiece, and [ need a corner piece, that fairly literallyinstantiate the initial-state condition. They are egocentricif the initiator is speaking (and are therefore not verypolite) and are exocentric if the responder is speaking(Do you want a corner piece?)

b) Object-specific. Utterances (which in English employeither have or take) which communicate that the speakeris selecting something from available alternatives. On this


view, ['11 have a hot chocolate. S-Means "I hereby select ahot chocolate." These forms are also egocentric if utteredby the initiator and are not considered especially polite.

c) Object-only. This feature is used for utterances with nosubject (cf., How about a corner piece?). We do not havea good understanding of this class of cases. We treat theseforms as init-oriented because they are restricted in use toidentifying what the initiator desires.

2) Resp-oriented. These are utterances that instantiate eitherthe willingness condition or the ability condition. Respwillingness forms are more polite than resp-ability forms, allother things being equal, but are a problem for the initiatorin that a resp-willingness rejection (I won't do that) moregreatly threatens the initiator's positive face than does aresp-ability rejection ([ can't do that).a) Resp-ability. These are utterances that address the ability

satisfaction condition and are relatively polite if utteredby the initiator - cf., Can you give me that corner piece?

b) Resp-willingness. These are utterances that address thewillingness satisfaction condition and are relatively politeif uttered by the initiator. Imperatives like Give me thatcorner piece will be treated as nondeferential respwillingness utterances (as is consistent with the traditionalview that they have understood you subjects and arevolitional - cf., Give me that corner piece, won't you?)

3) Domain-oriented. This is for utterances that address condition C3 such as Do you have a corner piece?

B. Power and Deference Features:1) Power features. These forms are involved in lexical selection

for modal verbs and main verbs.a) [nit-up. This is for utterance forms that suggest that the

initiator enjoys greater social power over the responder.It is for contrasts between cases like ['11 take a cornerpiece (init-up) versus ['11 have a corner piece (no-init-up)or [ want a corner piece (init-up) versus [ need a cornerpiece (no-init-up).13

b) No-init-up. This feature pairs with"init-up" to define thecontrasts just discussed.

2) Deference Proper. These features determine sentence type inconcert with other features.


a) Deference. (Can I have a corner piece?:J Would you give mea corner piece?)

b) No-Deference. (I:Jll have a corner piece:J Give me a cornerpiece)

These features do not exhaust the set of politeness features, nor are wefully confident of their being correct.

Now, if we wanted to teach someone to use English felicitously, wecould try to pair up sentences with contexts with instructions like "Saysuch-and-such type of sentence if you are in a context having property Cand have social relationship of type R with your addressee by way ofmaking a request for something of type X." However, given the ratherlarge array of utterance types available,14 this would be a daunting task.We believe, instead, that it would make more sense to teach studentswhat sociopragmatic distinctions are made in a language culture and howto map sociopragmatic features into morphosyntactic and prosodicfeatures. One might tell students of English that in such-and-suchcontext, one needs to show deference to an addressee and one shouldtherefore use an interrogative sentence. Or that in another context, onemust show deference and have an ability-condition-specific, addresseeoriented utterance, the result being an interrogative sentence with you asthe subject and can or could as the modal (Can you give me a cornerpiece). Obviously, if we are to provide instruction like this, we must findout what sociopragmatic features are operative in the native languageand in the target language and what their morphosyntactic and prosodicimplications are.

2. Pilot study

In this chapter, we report on a pilot study we have done on hownative French- and English-speaking subjects do requesting in a specificexperimental context and how English-speaking learners of French dorequesting in French in the same context. Our objective was to determinewhat politeness features seem to be at work in the two language culturesand how the various politeness features found are realized morphosyntactically and prosodically in the two languages.

Three groups of subjects participated in the experiment: 26 pairs ofFrench native speakers from the University of Nantes, who were spending


a quarter at The Ohio State University (OSU) to complete a certificate inthe Business program, 23 pairs of English native speakers drawn largelyfrom an introductory linguistics course at OSU, and 27 pairs of Englishspeaking learners of French enrolled in intermediate conversation coursesat OSU.15

For each run of the experiment, two subjects within the same languagegroup were seated at opposite ends of a small table. Each subject wasasked to put together a children's jigsaw puzzle. Each was given a set ofpieces to his or her puzzle (at the subject's right hand side) and a set ofpieces to his or her partner's puzzle (at the subject's left hand side).A "starter piece" was set before each subject. Two pieces to each puzzlewere hidden in a box to be retrieved when it became clear that neithersubject held the piece. In one experimental context, a barrier was placedbetween the subjects to make it impossible for subjects to see the piecescontrolled by the other subject. This was done to encourage multi-turnsequences (as was our hiding two pieces of each puzzle from view).Subjects were instructed to assist each other by handing over neededpieces, subject to the constraint that the exchange had to result fromsome sort of verbal interaction. (We were careful not to use the wordsrequest or offer or other language (e.g., ask your partner for a piece) inour instructions that might bias the specific language forms subjectsused.) In general, subjects seemed to find the task to be fun, and relativelycolloquial language resulted. All experimental runs were recorded onvideotape and audiotape and were transcribed, including relevant nonverbal behavior.

3. Results and discussion

After completion of the experiment, ten experimental runs from eachgroup of subjects were randomly selected for analysis, with the onlystipulation being that within each language group, there would be approximately ten females and ten males, divided equally between thebarrier and non-barrier conditions. Demographic data on the 60 pairedsubjects (see Table 1 16) revealed that the groups were evenly matched forsex. Although the age of the subjects ranged from 18-69, each group wasfairly even in age, considering that the median age for each group was 21and that 80 % of the English group, 90 % of the French learners group,and 100 % of the French natives were in the age range of 18-29.


Table 1.

Demographic data English French FrenchLearners

Sex 2 FF 2FF 3FF2MM 2MM 2MM6FM 6FM 5FM

Age Range 18-69 20-22 18-45

Mean Age 28 21 23

Median Age 21 21 21

F = Female; M = Male

It was a given of this experiment that subjects were to assist eachother. Put in satisfaction condition terms, condition C4 on requestingwas understood to be satisfied or true at the outset. One of thepredictions we made, therefore, was that there would be relatively fewinterrogative utterances that instantiate this condition (i. e., whoseS-Meaning can be expressed in terms of this condition). (If a conditionis understood to be satisfied, there is no need to address it further.)In fact, of the 245 English request initiations (105 in the barrier-down(NB) condition and 139 in the barrier-up (B) condition), we find onlythose in (15).

(15) No BarrierFl: Will you hand me that piece that's closest to you?F2: ((Points))Fl: Right there. YeahF2: ((Laughs and passes piece)) Let's see here.Fl: And that corner piece. ((Points))F2: ((Picks up piece))Fl: Umhum

We counted this as two instances of a willingness-condition specificutterance, but it should be clear that these did not represent genuinelyindependent instantiations of this condition. Among the French requestinitiations, only 7 out of the 243 spoke to the willingness of the addressee


to hand over pieces. These were all made by the same speaker within thesame conversation, and were of the variety found in (16).

(16) Tu veux me passer celui ou il yale plus de beige, enfin, d'ocre?Celui-la. Voila, exactement ~a.

'Do you want to pass me the one where there is the most beige,well, dark yellow? That one. That's it, that's it exactly.'

On the other hand, imperative utterances, which, as we noted, we aretreating as no-deference, resp-willingness forms, can occur when thewillingness of the addressee to perform the desired action can beassumed. Indeed, this is a precondition on their use.

We were interested in evoking two kinds of utterances - utterancesemployed to identify whether or not some desired piece was held by theother party and utterances employed to indicate a desire for some piece.What we found was that there was variation in the types of utterancesused in both circumstances. Thus, in cases in which the existence of adesired piece was unknown (sometimes because the initiator was unclearhow to describe it), we would get domain-oriented forms like Do youhave a corner piece? (English) or Tu n'as pas de coin? (French), andvariants thereof (utterances addressing condition C3 of the ability condition C2 ) but we also got utterances addressing other conditions a.s well,such as, init-oriented forms like I need a corner piece (English) or ]'aibesoin d'un coin (French). Thus, even in this case subjects made politeness choices in how they framed these request initiations.

In circumstances in which the initiator could easily identify a desiredpiece (as when the barrier was down), it would be apparent that theresponder had the piece (C3 was true) and therefore the responder wasable to provide the piece (C2 was true). Since C4 was given as true, all theresponder-specific satisfaction conditions would be true. At such a point,the only thing that is not known is what specific piece the initiator wants.It is at that point that virtually all request forms (all of the forms of (7),and more) become available to the speaker and speakers are forced tomake politeness choices. This corresponds to a real world situation suchas we find in a frozen yoghurt store where the sizes and flavors offeredand the prices charged are posted in easy view. In such a circumstance theonly thing that is not known is what the customer wants and, therefore,any of the forms of (7), and more, become available.

We organized the data obtained according to the orientation anddeference choices the speaker made, with the orientation features


representing the highest level of organization (since they equate to thesatisfaction condition the utterance addresses). In what follows, weprovide a summary of our findings. 17 In all cases, we give forms utteredby initiators, for we were interested in requesting, not offering.

I. Init-oriented forms (Utterances addressing the initial-state conditionC1 ) - See Table 2.18

A. Init-specific forms: 19 I need a corner piece/J~ai besoin d~un coin 20

B. Object-specific: Forms (whose S-Meanings are "I hereby select X")are restricted in use to cases in which the initiator can see that the responder has the desired piece.

1. No-init-up forms:a. No-deference: I'll have that corner piece/(No French Equivalent)b. Deference: Can I have a corner piece?/Est-ce que je peux avoir un coin?

2. Init-up forms:a. No-deference: I'll take a corner piece/Ie prendrai un coinb. Deference: Can I take a corner piece?/Est-ce que je peux prendre un

coin?

Table 2.

I. Init-Oriented forms

A 41 39 80 (32.7) 40 28 68 (28)B.l.a 0 0 o (0) 0 0 o (0)B.l.b 1 20 21 (8.6) 0 1 1 (0.4)B.2.a 0 2 2 (0.8) 1 0 1 (0.4)B.2.b 0 0 o (0) 0 0 o (0)

Totals 42 61 103 (42.1) 41 29 70 (28.8)

English

B NB T(%)

French

B NB T(%)

French Learners

B NB T(%)

57 11 68 (28)0 0 o (0)0 9 9 (3.7)0 0 o (0)0 0 o (0)

57 20 77 (31.7)

B = Barrier; NB = No Barrier; T = TotalA = init-specific formsB.l.a = Object-specific, no init-up, no deferenceB.l.b = Object-specific, no init-up, deferenceB.2.a = Object-specific, init-up, no deferenceB.2.b = Object-specific, init-up, deference


The primary init-oriented forms are the LA forms (cf., Table 2), for theyemploy some verb that means "desire" or "need," which is cognate withthe predicate of the initial-state condition. We find init-specific forms inboth the "barrier up" (B) and "no-barrier" (NB) conditions. One majordifference between English and French is that what we are calling objectspecific, init-oriented forms (cf., I. B) were virtually never used by Frenchnatives in our experiment.

II. Resp-oriented forms (Utterances addressing the willingness andability conditions.)

A. Resp-willingness.1. No-deference: Give/Hand me a corner piece/Donne/Passe-moi un coin2. Deference: Will you give me a corner piece?/Tu me passeras un coin?

B. Resp-ability.1. No-deference: You could give me a corner piece (now)/Tu peux/

pourrais me donner un coin2. Deference: Could you give me a corner piece?/Est-ce que tu

peux/pourrais me donner un coin?

There was a striking contrast between the English and French speakerswith respect to the use of imperative utterances (see Table 3). Over 100/0

Table 3

II. Resp-oriented forms

English French French Learners

B NB T(%) B NB T(%) B NB T(%)

A.l 0 1 1 (0.4) 2 24 26 (10.7) 0 6 6 (2.5)A.2 0 2 2 (0.8) 0 7 7 (2.9) 0 1 1 (0.4)B.l 0 0 o(0) 0 0 0(0) 0 0 o(0)B.2 0 5 5 (2) 3 9 12 (4.9) 0 7 7 (2.9)

Totals 0 8 8 (3.2) 5 40 45 (28.5) 0 14 14 (5.8)

B =Barrier; NB =No Barrier; T =TotalA.l = Resp-willingness, no deferenceA.2 =Resp-willingness, deferenceB.l =Resp-ability, no deferenceB.2 =Resp-ability, deference


of the French natives' requests were in the imperative form, comparedwith just 0.4 % (1 request) of the English speakers. It seems that in aninformal situation such as the one in the current study (helping someoneperform some task in an informal atmosphere), using the direct form ofthe imperative is not considered impolite in French. The addition oflexical softeners such as s:lil te plait ('please'), tiens ('well'), and done('so') in more than one third of the imperative requests also served toreduce the abruptness of the imperative form. The fact that the Frenchlearners patterned themselves more after the American English speakersprobably reflects the English taboo against the use of the imperative. Thissupports research reported by Kellerman (1979; 1983) which indicatesthat error due to interference from the native language is more likely theless the perceived distance between the target language and the nativelanguage. It is difficult to extrapolate the use of the imperative to a largenumber of other French contexts without further study, but certainlyteachers could encourage students to use it in similar informal contexts,particularly in conjunction with lexical softeners.

There were no particularly striking differences among the languagegroups in regard to resp-ability utterances, utterances addressing condition C2• They were relatively rare, we think, largely because utterancesaddressing its precondition, C

3, were so common, as we shall shortly see.

A positive response to C3

utterances implies a positive response to C2,

given the conditions of the experiment, and so such utterances willdiminish in frequency as C

3utterances increase (and conversely, of

course).The last class of utterances we shall consider are domain-oriented

utterances, so-called because they inquire as to whether some desiredthing is available (in requests for things) or whether someone is availableto do something (in requests for actions). In English, the principledomain-oriented utterances are Yes-No-Questions like those in (2)employing have in the case of requests for things and questions like Youdoing anything? or Whateha doing?, and the like, for requests for actions.

As we noticed earlier, French domain-oriented utterances commonlyconsist of positive and negative declarative intonation questions, as wellas inverted questions like those used in English. Negative declarativeintonation questions seem to be more polite than the various types ofsemantically positive French question forms. The explanation for thisseems to be that they are being used to confirm a speaker presuppositionthat the addressee does not have the desired thing and therefore providethe addressee with a "built-in" excuse for rejecting the request. In Table


4, we combine French Est-ce que questions, declarative intonationquestions, and interrogative (inverted) sentences in row A.1 for theyeither do not presuppose the existence of the desired thing or express apositive presupposition. They are contrasted, then, with the negativedeclarative intonation questions (A.2), which convey a negative presupposition. We then go on to show the distribution of the variousquestion forms in Table 5.

III. Domain-oriented forms. (Utterances addressing Subcondition C3

ofthe ability condition.) - See Table 4.

A. Deference:1) No-Presupposition/Pos-Presupposition: Do you have a corner

piece?/Tu as un coin?, Est-ce que tu as un coin?, As-tu un coin?2) Neg-Presupposition: You wouldn't have a corner piece?/Tu n'as pas

de coin? 21

B. No-Deference: You (must) have a corner piece/Tu dais avoir un coin.

As Table 4 indicates, the contrasts between English and French in regardto C3 utterances were quite striking. As noted in Table 5 and in connection with examples (8) and (11), French speakers rarely use the invertedinterrogative form in making information questions, preferring, instead,to use Est-ce que questions or either positive (Tu as ... ?) or negative (Tun'as pas ... ?) declarative intonation questions. Interestingly, the French

Table 4.

III. Domain-Oriented Forms

English French French Learners

B NB T(%) B NB T(%) B NB T(%)

A.1 51 17 68 (27.8) 41 12 53 (21.8) 86 46 132 (54.3)A.2 0 8 8 (3.3) 32 24 56 (23) 1 1 2 (0.8)B 0 0 0 (0) 0 9 9 (3.7) 0 1 1 (0.4)

Totals 51 25 76 (31.1) 73 45 118 (48.5) 87 48 135 (55.5)

B =Barrier; NB =No Barrier; T =TotalA.l = Deference, no-presupposition/possible presuppositionA.2 = Deference, negative-presuppositionB =No deference


Table 5.

French question forms

French Natives French Learners

B NB T (%) B NB T (%)

Est-ce que tu as ... 43 20 63 (47.4) 50 42 92 (59.0)Tu as ... 33 36 69 (51.9) 3 2 5 (3.2)As-tu ... 1 0 1 (0.8) 39 20 59 (37.8)

Totals 77 56 133 (100) 92 64 156 (100)

B =Barrier; NB =No Barrier; T =Total

were just as likely to inquire about the nonexistence of a puzzle piece thanthe affirmative existence of it. We speculate that in using the negativeform, French speakers were providing a built-in excuse for the addresseeto not provide the piece, that is to say, a face-saving device for bothspeakers even though the speaker clearly expected to receive the piece.

Quite strikingly, the French learners virtually exclusively employedEst-ce que questions and inverted interrogative forms, almost totallyignoring positive declarative intonation forms and both negative Est-ceque and negative declarative intonation forms. We offer the followingexplanation for the learners' behavior. Regarding the excessive use ofinversion for question format, we feel that this may be due to a specialemphasis on it in the language classroom. Since intonation per se is soeasy to use, teachers tend to concentrate their teaching efforts on Est-ceque questions. They also focus on the inverted forms for they aregrammatically more difficult. In addition, studies have found that Frenchtextbooks frequently do not clarify sufficiently differences in style andfrequency of use regarding Est-ce que, intonation, and inversion (Di Vito,1991; 1992; Walz 1986). Even if inversion is described as useful inprimarily written contexts, most textbooks have exercises that requirestudents to practice using inversion for forming questions in spoken contexts as well. To make matters worse, inversion is often used in textbooksto provide directions for oral and written exercises and activities.Although the use is correct since the textbook represents written languageand students do need to become familiar with the form, it may nonetheless give students the impression that inversion is used more frequentlythan it really is.


Moreover, English speakers usually do not use declarative intonationquestions as "ordinary" questions. They tend to be relatively contextbound (e.g., used when the speaker believes that the propositionalcontent expressed by the question may be false or is surprising - Yousaw John? How odd. I thought he was in Chile!) Thus, it seems thatlearners were transferring their native language speech act strategies inthis case and thereby avoided the commonly used intonation patternpreferred by the French.

With regard to the lack of negative request forms, English speakers donot employ negative declarative intonation questions simply to requestinformation (or, as in this experiment, to request things). Instead they areemployed to confirm a speaker belief that the propositional contentexpressed by the proposition is true. (You didn't see John? He was here!I saw him several times!) Moreover, as Fraser and Nolen (1981: 103)point out, negative interrogatives are perceived as less deferential byAmericans than positive interrogatives, a view that we can expect wouldbe carried over to a second language.

4. Conclusions / Implications

Our study of American English, French, and French learner requestingbehavior is unusual, we believe, in that it was based on how eachlanguage group did requesting in precisely the same experimentalconversational context. This method gives us a reliable measure, webelieve, of what counts as French and English functional equivalents fordoing requesting in relatively informal conversational contexts, and ameans of assessing possible sociopragmatic interference in languagelearning by our French learners. Further research is needed to confirm orrevise our findings in this pilot study. Larger sample sizes, different contexts, and eventually comparisons between English and French and otherlanguages will shed further light on the nature of pragmatic languageinterference.

There is an important theoretical conclusion to be drawn from thisstudy, and this is that different language cultures can vary along twodimensions sociopragmatically. First, different language cultures maydiffer in the sociopragmatic distinctions they make. Thus, whereas inEnglish, colloquial domain-oriented utterances tend to be restricted topositive inverted forms like Do you have a corner piece? and stylistic


variations thereof (cf., our discussion of example (1)), the French havethree forms: inverted questions, positive and negative declarative intonation questions, and Est-ce que questions. This strongly suggests that theFrench language culture makes more sociopragmatic distinctions thandoes the American English language culture, with stylistic and politenessconsiderations each playing a role. Second, even when two languagecultures make the same politeness distinctions, the forms they use in aparticular context may be quite different syntactically and semantically.We are inclined to think, for instance, that the French declarative intonation question, Tu as un coin?, is the closest politeness equivalent to theEnglish form, Do you have a corner piece?, though our study is mute onthe point.

The results of this study suggest several implications for the teachingof French. First, although questions formed by inversion need to betaught in the classroom along with Est-ce que and intonation, the data inthis study suggest that the importance of the use of simple intonationquestions in conversation is not at all clear to the learner and needs to bemade so. Language textbooks need to reduce the amount of invertedforms used, clarify the limited use of inversion, and refrain from obligingstudents to use inversion in oral activities.

Second, in addition to engaging in a certain amount of sociopragmaticteaching that stresses that the declarative intonation question is one of themost commonly used French devices for requesting in informal contexts,the data suggest that teachers stress that the negative intonation form isespecially polite because it allows the addressee a face-saving basis forrejection of the request in that it provides a built-in excuse because ofthe negative presupposition it communicates (You don't have a cornerpiece -, and so, therefore you can't pass it over.) This is as true, ofcourse, of negative Est-ce que questions as of negative declarativeintonation questions. In teaching the negative declarative intonation,teachers might note that they seem to parallel in use the Englishform You wouldn't have a green piece?, which, as line (A.2) ofTable 4 shows, does occur in English in roughly the same context as theFrench negative declarative.

Third, the imperative form, in conjunction with the use of lexicalsofteners, was preferred in over 10 % of the French natives' requests andyet the French learners eschewed the imperative form, perhaps becausethere is a taboo against using it in English in most circumstances. Thissuggests that learners should be taught that the imperative form is not asimpolite in French as it is in English, at least in some informal contexts.


Further studies will indicate whether the imperative form can be appliedappropriately in polite requests in other contexts.

Finally, this study, has attempted to add to the growing body ofresearch into sociopragmatic choices available to speakers of languages. Itis becoming more and more clear that teachers need to focus their effortsactively not only on teaching contrasting linguistic structures in English andthe second language, but also on how to make the correct sociopragmaticchoices in conversation (the correct style, politeness, and register choices).In so doing, language teachers will greatly facilitate learners' successfulinteractions with native speakers of the second language.

Notes

1. We have actually carried out this experiment informally in a few stores to seewhat would happen. The answers, so far, have all been negative, except forone case in which the clerk said, Do you mean "ever?"

2. Note, for instance, that while any of the utterance-types in (2) and (3) mightbe used between friends in requesting a beer (Do you have any beer?), theutterances of (4) could not (Do you carry beer?).

3. We would expect that these utterances could have different significances orS-Meanings were they to be uttered in different contexts.

4. We shall take the position here that there are just two basic things thatutterances do: provide information and request information. Such actions asrequesting, suggesting, inviting are treated as being epiphenomenal incharacter - as the product of implicature (Grice 1975), if one likes.

5. The notion "language culture" is defined as a pattern of social expectationsassociated with a particular, homogeneous subgroup of speakers of alanguage. Such a notion is required if we are to understand how differentsubgroups of those speaking a single language differ in how they uselanguage.

6. Since "intonation questions" of the sort that we find in French aredeclarative rather than interrogative in form, in that they do not exhibitinversion, we shall refer to them as "declarative intonation questions."

7. The"?" at the end of this and other declarative (in form) intonation questionsin French and English is intended to represent "question intonation."

8. In French one can also use the negative interrogative form in a presupposition-checking manner. The difference in the languages is that French, unlikeEnglish, makes this form available as a polite form.

9. Searle's providing alternate characterizations of this condition - one in- "objective" terms and one in terms of speaker beliefs - will not do. Geis

(1995) argues that felicity conditions should be stated objectively, with theissue of speaker beliefs arising only in utterance planning.

10. Geis (1995) demonstrates that so-called speech acts are social, as opposed tolinguistic, actions in that they can commonly be performed nonverbally and


the factors that distinguish them are invariably social or psychological, rather than linguistic in nature. With the understanding that speech acts areactually communicative social actions, not linguistic actions, there is noharm in continuing to use the older term.

11. The initiator of an act is the one whose desires or needs are the cause of theinitiation of the act and is not to be equated with the person who speaks first.We would say that the cases of a customer who goes to a store to purchasesomething (an attempt to buy) and of a door-to-door salesperson who triesto sell something to a customer (an attempt to sell) are speech acts theoretically equivalent in that in both cases the buying-selling actions will besuccessful just in case the initiator (buyer) comes to desire the thing beingoffered for sale and the responder is able and willing to sell it, that is, has thething and finds the price the initiator is willing to pay acceptable.

12. Here would be spelled out the properties of the thing desired. If the initiatorwishes to purchase one pint jar of blackberry jam, the Domain would be asin (1).(1) Domain(x): blackberry-jam(x) & size(x, pint) and number(x, one)

13. Note that people who are socially inferior to others (children versus parentsor employees versus employers) tend to use the less "pushy" forms I need anew bike/I need a raise over the more pushy forms I want a new bike/I wanta raise, whereas superiors have no difficulty using the pushy forms I wantyou to clean up your room/I want you to finish the report before you gohome.

14. The examples of (7) barely touch the surface of the problem.15. A minimum of 5 quarters of French study were a prerequisite for entering the

course. The average student had taken the equivalent of 8 or 9 quartercourses prior to this one.

16. Only 5 FM pairs in the learner group participated in the entire experiment.Thus, an additional pair of FF had to be used.

17. The percentages given in the tables below are based on 245 English, 243French, and 245 French learner request initiations.

18. In the tables that follow, we shall use the roman numerals employed in thisclassification to identify forms.

19. The deference/no-deference contrast does not apply to init-specificutterances, for we never make requests saying, Do I want a hot chocolate?

20. In what follows, we collapse all instances of an instantiation of a particularcondition into a single type. So, in the case of (LA), we collapse Ineed/want/would like P together and, looking ahead to (LB.1.b), collapsesuch variants as Could/Can/May I have P into a single type of utterance.Obviously, there are politeness differences among these variants.

21. We place the English conditional negative declarative intonation questionYou wouldn't have a corner piece? alongside the French simple negativedeclarative form. It must be noted that these are not literal equivalents.Instead, they are functional (politeness) equivalents.


References


view", in: Shoshana Blum-Kulka - Juliane House - GabrieleKasper (eds.), 1-34.

Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper (eds.)1989 Cross-Cultural pragmatics: Requests and apologies. Norwood, NJ:

Ablex.Brown, Penelope - Stephen C. Levinson

1978 "Universals in language usage: Politeness phenomena", in: EstherGoody, (ed.), 56-311.

Desirat, Claude - Tristan Harde1988 La langue franfaise au 20e siecle. Paris: Bordas.

Cole, Peter - Jerry L. Morgan (eds.)1975 Syntax and semantics III: Speech acts. New York: Academic Press.

Di Vito, Nadine O.1991 "Incorporating native-speaker norms in second language materials",

Applied Linguistics 12: 383-396.1992 "Sensitizing teaching assistants to native-speaker norms in the

communicative classroom", in: Joel C. Walz (ed.), 171-189.Fraser, Bruce - William Nolen

1981 "The association of deference with linguistic form", InternationalJournal of the Sociology of Language 17: 93-109.

Gadet, Fran~oise1989 Le franfais ordinaire. Paris: Armand Colin.


House.Geis, Michael L.

1995 Speech acts and social actions: Toward a theory of conversationalcompetence. Cambridge: Cambridge University Press.

Goffman, Ervin1967 Interaction ritual: Essays on face-to-face behavior. Garden City, NY:

DoubledayGoody, Esther (ed.)

1978 Questions and politeness: Strategies in social interaction.Cambridge: Cambridge University Press.

Grice, H. Paul1975 "Logic and conversation", in: Peter Cole - Jerry L. Morgan (eds.),

41-58.Kellerman, Eric

1979 "Transfer and non-transfer: Where we are now", Studies in SecondLanguage Acquisition 2: 37-57.

1983 "Now you see it, now you don't", in: Susan M. Gass - LarrySelinker (eds.), 112-134.


Searle, John R.1969 Speech acts: An essay in the philosophy of language. Cambridge:

Cambridge University Press.Valdman, Albert

1967 "Norme pdagogique: Les structures interrogatives du fran~ais",

International Review of Applied Linguistics 5: 3-10.Walz, Joel C.

1986 "Is oral proficiency possible with today's French textbooks?",Modern Language Journal 70: 13-20.

Walz, Joel C. (ed.)1992 Development and supervision of teaching assistants in foreign

languages. Boston: Heinle & Heinle.Wolfson, Nessa

1983 "An empirically based analysis of complimenting in AmericanEnglish", in: Nessa Wolfson - Elliot Judd (eds.), 82.-95.

Wolfson, Nessa - Elliot Judd (eds.)1983 Sociolinguistics and language acquisition. Rowley, MA: Newbury

House.

Transfer and proficiency in interlanguageapologizing

Naoko Maeshiba - Naoko Yoshinaga - Gabriele Kasper- Steven Ross

1. Pragmatic transfer

The influence of non-native language users' linguistic and cultural background on their performance of linguistic action in a second language hasbeen a focal concern in interlanguage pragmatics (Kasper 1992, forreview). Transfer effects have been noted at the sociopragmatic andpragmalinguistic level (cf., Leech 1983; Thomas 1983, for the distinctionbetween sociopragmatics and pragmalinguistics). Sociopragmatic transferhas been found to operate in learners' perceptions of contextual factors,such as interlocutors' relative social status (e.g., Beebe - Takahashi Uliss-Weltz 1990; Takahashi - Beebe 1993); assessment whethercarrying out a particular linguistic action is socially appropriate (e. g.,Robinson 1992), and the overall politeness style adopted in an encounter(e.g., Blum-Kulka 1982; Garcia 1989; Olshtain - Cohen 1989).Pragmalinguistic transfer has been noted in learners' use of conventionsof means and form, affecting the illocutionary force and politeness valueof interlanguage utterances (e.g., House - Kasper 1987; Bodman Eisenstein 1988; House 1988; Beebe - Takahashi - Uliss-Weltz 1990).Because of its potential for miscommunication, focus has been given tonegative transfer, the projection of first language-based sociopragmaticand pragmalinguistic knowledge onto second language contexts wheresuch projections result in perceptions and behaviors different from thoseof second language users. Yet positive transfer, the projection of firstlanguage-based sociopragmatic and pragmalinguistic knowledge wheresuch projections result in perceptions and behaviors consistent with thoseof second language users, have also been attested. For instance,conventionally indirect forms for requesting (e. g., Blum-Kulka 1982;House - Kasper 1987; Frerch - Kasper 1989; Takahashi - DuFon 1989)were successfully transferred to English from Hebrew, German, Danish,and Japanese. Meaning conventions were transferred in the apologystrategies used in German-English (House 1988) and Thai-English

156 N. Maeshiba -N. Yoshinaga - G. Kasper - S. Ross

(Bergman - Kasper 1993) interlanguage. In all of the above cases, we arejustified to assume that positive transfer from first language to secondlanguage pragmatic knowledge was operative because the matchingpatterns apply to specific pairs of native and target languages andcultures and not to others. In most instances, however, it is difficult todisentangle positive transfer from learners having recourse to universalpragmatic knowledge and inferencing strategies (Blum-Kulka 1991).

In addition to ascertaining where pragmatic transfer occurs andwhether it leads to perceptions and behaviors divergent from orconsistent with second language users', interlanguage pragmaticists haveattempted to identify the conditions for transfer to occur, and the factorswhich mediate its operation. Transferability constraints posited in theliterature include learners' psychotypology in the sense of Jordens (1977)and Kellerman (1977), and their perceptions of sociopragmatic andpragmalinguistic knowledge as specific for a given culture and languageor as culturally and linguistically "neutral". An example of differentialpsychotypologies is seen in the request patterns of Danish learners, whomade freer use of their first language when requesting in German than inEnglish (House - Kasper 1987). Universal versus culture-specific perceptions distinguished the need to apologize as expressed by Russian andEnglish-speaking learners of Hebrew, the Russians perceiving the need toapologize as determined by the nature of the committed act, whereas theEnglish-speaking learners made apologizing contingent on cultural context (Olshtain 1983). Japanese female informants stated that refusingoffers, requests, or invitations was much more acceptable in Americanthan in Japanese society; hence they felt that transfer from their Japanesenorms of interaction would not be successful in an American context(Robinson 1992). At the pragmalinguistic level, German learners ofEnglish avoided the use of the mitigator 'I mean' because they consideredthe German equivalent ich meine as language-specific (Kasper 1982). Thetransferability of conventionally indirect request strategies from Japaneseto English was shown to be highly context-dependent, and varied withlearner factors such as proficiency and familiarity with the situation(Takahashi 1992).

Non-structural factors interacting with pragmatic transfer includelearner-external factors such as learning context and length of residencein the target community, and learner-internal factors such as attitudetowards the native and target community, and second languageproficiency. Takahashi and Beebe (1987) found that transfer of Japaneserefusal strategies, while occurring in the refusal patterns of English as a

Transfer and proficiency in interlanguage apologizing 157

second language and English as a foreign language learners, was moreprevalent in the English as a foreign language than in the English as asecond language learners' production. According to informants' selfreports, their culture- and language-specific perceptions of refusalstrategies and pragmalinguistic function, noted by Robinson (1992) andKasper (1982) (see above), could partly be attributed to explicit teaching.We are not aware of any studies which specifically examine the effect oflength of residence on pragmatic transfer. However, since some studiessuggest that length of stay influences second language pragmatic behaviorin a non-linear fashion (Olshtain - Blum-Kulka 1985; Blum-Kulka Olshtain 1986), it can be assumed that this factor has an impact onpragmatic transfer also. What exactly this influence is remains to bestudied. A quantitative measure like amount of exposure alone or, forthat matter, proficiency cannot account for the fact that highly proficientlong-term residents often preserve some of their first language communicative style, and even pass it on to the next generations of immigrants(e.g., Clyne 1979; Blum-Kulka - Sheffer 1993). Such an "interculturalstyle", seen, for instance, in American immigrants to Israel, hardlyreflects "deficient" second language communicative competence (BlumKulka 1991). Rather, it appears to express language users' need fordisidentification, or maintaining their cultural identity as separate fromthe community at large. Pragmatic divergence of this kind can best beaccounted for in an accommodation-theoretical framework (e.g., Giles Johnson 1987). Of the learner-internal factors, then, social-psychologicalorientation is a potential determinant of pragmatic transfer. The other,"cognitive" factor, second language proficiency, has been found toconstrain pragmatic transfer in requesting (Blum-Kulka 1982) andapologizing (Olshtain - Cohen 1989). In both studies, it was found thatlearners' limited second language knowledge prevented them fromtransferring complex conventions of means and form from their firstlanguage. These findings are thus consistent with Takahashi and Beebe's(1987) hypothesis that second language proficiency is positively correlated with pragmatic transfer. While their own study on refusals performedby Japanese learners of English at two different proficiency levels did notdemonstrate the predicted proficiency effect, Blum-Kulka's (1991) andOlshtain and Cohen's (1989) studies support Takahashi and Beebe's view.However, since these studies do not look at the performance of learnersat different proficiency levels, they do not provide conclusive evidence foror against the effect of proficiency on transfer. In this chapter, therefore,we shall put Takahashi and Beebe's transfer hypothesis to another test.


2. Native and non-native apology

Second only to requests, apologies are the next-best studied speech act indescriptive, cross-cultural, and interlanguage pragmatics. This is so forgood reasons. In any speech community, participants need to be able toengage in remedial verbal action upon committing an offense, that is, toapologize. While the speech act of apologizing can thus be regarded as apragmatic universal, the conditions which call for apology are clearly not.Speech communities differ in what counts as an offense, the severity ofthe same offensive event, and appropriate compensation. These perceptions will in turn be mediated by social factors such as the interlocutors'relative status and familiarity. Non-native speakers have to learn what thespecific conditions for apology are in the target community, what thestrategies and linguistic means are by which apology can be implemented,and how to make contextually appropriate choices from the apologyspeech act set.

The supposition of an apology speech act set is supported by a largebody of studies examining native and non-native speakers' apologizingpatterns. This notion, first proposed by Olshtain and Cohen (1983) andempirically sustained in a series of studies by these authors (Olshtain1983; 1989; Olshtain - Cohen 1989) entails that apologies can becarried out by a finite set of "conventions of means", or strategies, all ofwhich are related to the offensive act and serve as the speaker's attemptto "make it go away": either by conveying regret and proposing remedy,or by diminishing the offense or the speaker's responsibility for it. Twostrategies, offering an explicit apology and assuming responsibility forthe offense, were found to be used in remedy of most offenses, irrespective of the specific contextual circumstances. The remaining strategies,upgrading apologetic force, downgrading the severity of the offense orspeaker's responsibility, offering repair, and placating the offended partyby different kinds of verbal redress are clearly cross-culturally available,yet their use is highly sensitive to contextual conditions, and subject tocross-cultural variation (Olshtain 1989; Bergman - Kasper 1993).

Selections from the apology speech act set are determined by a varietyof context-internal and context-external factors. One of the contextinternal factors is the nature of the offense. Borkin and Reinhart (1978)suggested that "excuse me" is used to remedy "a breach of etiquette orother light infraction of a social rule", whereas "I'm sorry" is used as anexpression of dismay or regret about "a violation of another person'sright or damage to another person's feelings" (1978: 61). Their observa-


tion compares well to Goffman's (1971) distinction between ritual andsubstantive apology.

Within the category of substantive apology, the offender's obligationto apologize affects the choice of apologetic formula (House 1988) andintensification of apologetic force (Cohen - Olshtain 1981; House 1988;Vollmer - Olshtain 1989; Bergman - Kasper 1993). The factor that hasbeen shown to have the strongest effect on apology realization is theseverity of the infraction. Comparison of apology in Hebrew, AustralianEnglish, Canadian French, and German with assessments of contextualfactors in different offense contexts suggests that "severity of offense isthe representative contextual factor in the socio-pragmatic set of theapology" (Olshtain 1989: 160). In the case of significant injury orinconvenience, Fraser (1980) observed a shift from the pattern apology +account to apology + offer of compensation. Tanaka (1991) noted thatboth native speakers of Japanese and of Australian English increasedapology intensification and formality of apologetic formulae with higheroffense severity. Holmes (1989) noted that severity of offense has adifferential impact on female and male offenders' use of apologystrategies. Female New Zealand offenders apologized most to lightoffenses whereas men apologized most to infractions of medium severity.Whereas native speakers of English intensified apologetic force dependenton severity of offense, non-native speakers were found to take lessaccount of severity in their choice of apology intensification (Bergman Kasper 1993).

According to most studies, apology performance is affected by thecontext-external factors social power and social distance. The lower theoffender's status vis-a-vis the offended person, the more the perpetratoris prone to apologize by means of an explicit apologetic formula (Vollmer- Olshtain 1989), intensify apologetic force (Fraser 1980; Olshtain1989; Vollmer - Olshtain 1989), and choose a more formal apologystrategy (Cohen - Olshtain 1981; Olshtain - Cohen 1983). However,Holmes (1989) found a non-linear relationship between social power andapology in her New Zealand data: most apology was offered in equalstatus relationships, lower status offenders apologizing second most, andhigher status offenders least frequently. In American and Thai apologizing, social power did not influence offender's selection of apologystrategy (Bergman - Kasper 1993). Barnlund and Yoshioka (1990)found that Japanese offenders varied forms of apologizing moreaccording to participants' status than American perpetrators did incomparable contexts.


The impact of social distance on apology behavior varies acrossstudies. Except for a limited tendency towards a negative correlationbetween social distance and use of an explicit apology formula, Olshtain(1989) did not establish any relationship between social distance and useof apology strategy. Bergman and Kasper (1993) found that the closer theinterlocutors, the more likely the offender was to expressly assumeresponsibility for the offensive act. This finding is contrary to that ofWolfson, Marmor, and Jones (1989), in whose study most responsibilitywas expressed between acquaintances and equally little at the two opposite ends of the social distance continuum. This finding was interpreted asfurther evidence in support of Wolfson's bulge hypothesis (1989).However, neither Wolfson et al. (1989) nor Bergman and Kasper's (1993)investigation include intimate interlocutor relationships. Hence neither ofthe two studies has demonstrated evidence for or against the bulgehypothesis.

Studies of interlanguage .apologies include the language pairs Hebrewfirst language-English second language (Cohen - Olshtain 1981),English and Russian first language-Hebrew second language (Olshtain1983), Danish first language-English second language (Trosborg 1987),German first language-English second language (House 1988), Spanishfirst language-English second language (Garcia 1989), and Thai firstlanguage-English second language (Bergman - Kasper 1993). Of these,only Trosborg's (1987) study examined proficiency effects on learners'performance of apology, and therefore had potential implications for atheory of pragmatic development in adult second language learners.However, the only developmental effect she found was that the use ofmodality markers increased with higher proficiency. It is difficult to saywhether this pattern truly reflects a development of pragmalinguisticcompetence or merely an extension of the learners' lexical repertoire. Thelearners' strategy use differed in some respects from that of the nativespeakers of Danish (first language) and British English (second language),but there were no proficiency effects on the learners' use of apologystrategy.

Previous research has offered descriptive accounts of transfer andproficiency in interlanguage users' speech act performance. Consideringthe effects of contextual factors on strategy selection as reported in theliterature, it seems plausible to assume a relationship between contextualfactors and transfer of apology strategies. This study will thereforeexamine whether pragmatic transfer is constrained by contextual factors,and whether it is affected by learners' proficiency level.


3. Method

3.1. Subjects

Four groups of subjects participated in this study:

1) 30 Japanese learners of English (Intermediate) (JEI)2) 30 Japanese learners of English (Advanced) (JEA)3) 30 Native speakers of English (E)4) ,30 Native speakers of Japanese (J)

JEI were students enrolled in the English Foundation Program at theHawai'i Pacific University (HPU) at the time of the study. Their averageage was 22.8. Their TOEFL scores ranged between approximately 400and 500. JEA were undergraduate or graduate students enrolled at theUniversity of Hawai i at Manoa (UHM) at the time of the study. Theiraverage age was 27.5. Their TOEFL scores ranged from 510 to 627 (average 579.2). E and J were undergraduate and graduate students at UHM(average age NSE: 30.4, NSJ: 25.9). J participated as non-native speakersin JEI or JEA.

3.2. Materials

A 20-item Dialog Construction Questionnaire was prepared in Englishand Japanese. The items represented different social domains and interlocutor role relationships in terms of gender, social distance and relativesocial status, and differing degrees of severity of the committed offense.The content of the items is listed in (5) through (24).

(5) A and B are friends. A damaged B's car while backing up.(Damaged Car)

(6) A and B are friends. A borrowed a magazine from B and spilledcoffee over it. (Ruined Magazine)

(7) At a staff meeting, teacher A contradicted teacher B.(Contradiction)


(8) At a staff meeting, teacher A accused teacher B of being a poorteacher. (Poor Teacher)

(9) At an office, a junior colleague forgot to pass on a personalmessage to a senior colleague. (Personal Message Low-High)

(10) At an office, a senior colleague forgot to pass on a personalmessage to a junior colleague. (Personal Message High-Low)

(11 ) At an office, a junior colleague forgot to pass on an importantbusiness message to a senior colleague. (Business MessageLow-High)

(12) At an office, a senior colleague forgot to pass on an importantbusiness message to a junior colleague. (Business MessageHigh-Low)

(13) At a restaurant, a customer changed her mind after the orderhad already been taken. (Order Change)

(14) At a restaurant, a waiter spilled food on a customer's clothes.(Food on Customer)

(15) At a restaurant, a waiter brought the wrong order. (WrongOrder)

(16) At a restaurant, a customer spilled food on a waiter. (Food onWaiter)

(17) At the airport, a customs official has messed up a traveller'ssuitcase. (Messed-up Bag)

(18) At the airport, a traveller has been caught trying to smuggle aBonsai tree into Japan. (Bonsai Tree)

(19) At the airport, a customs official has broken a legally purchasedstatue when searching a traveller's suitcase. (Broken Statue)

(20) At the airport, a traveller is unable to produce a customs form.(Customs Form)


(21) A professor has not yet graded a term paper which a studentwas supposed to pick up. (Ungraded Paper)

(22) A student forgot a book she was supposed to return to herprofessor. (Borrowed Book)

(23) A professor misplaced a student's term paper and failed thestudent. (Failed Student)

(24) A student plagiarized from a published book and is found outby the professor. (Cheating Student)

Subjects were asked to supply both the offender's and the offended party'sturn (see Appendix 1 for a sample item). For this study, only the first pairparts were analyzed.

In order to examine the relationship between contextual factors andstrategy use, an Assessment Questionnaire was prepared, including thesame offense contexts as the Dialog Construction Questionnaire. Eachcontext was rated on a five-point scale for five context-internal factors(severity of offense, offender's obligation to apologize, likelihood for theapology to be accepted, offender's face loss, offended party's face loss)and two context-external factors (social distance and dominance) (seeAppendix 2 for sample item). Since both questionnaires were adaptedfrom a previous study (Bergman & Kasper 1993), the japanese version ofthe questionnaires was prepared by first translating the Englishquestionnaires into japanese and then back into English. Adjustments tothe japanese version were made based on a comparison of the originaland translated English versions. The translations were provided by agraduate student who is a native speaker of Japanese, bilingual injapanese and English, and not an author of this chapter.

3.3. Procedure

Items in all questionnaires were randomized. For the Dialog ConstructionQuestionnaire, subjects were instructed to fill in what they would say ineach of the twenty contexts. The intermediate and advanced learners (jEIand JEA) filled in the English and japanese version of the DialogConstruction Questionnaire in counterbalanced order. At least one weekelapsed before the second Dialog Construction Questionnaire was


administered. The Japanese and English version of the AssessmentQuestionnaire was completed by the native speakers of Japanese (J) andEnglish (E), respectively. No time limits were imposed on completing theDialog Construction and Assessment questionnaires.

3.4. Analysis

The Dialog Construction data were coded into the major categories (fromBergman - Kasper 1993) listed in (25) through (29). IFID: IllocutionaryForce Indicating Device, specifying the force of apology ("I'm sorry","I'm afraid").

(25) Upgrader: Element increasing apologetic force ("I'm terriblysorry", "I really didn't mean to hurt you").

(26) Taking on responsibility: Speaker admitting the offense, includingself-blame ("How stupid of me"), lack of intent ("I didn't mean todo this"), and admission of fact ("I haven't graded it yet").

(27) Downgrading responsibility or severity of offense: (a) utterancereducing speaker's accountability for the offense, including excuse("My watch had stopped"), justification ("I was suddenly called toa meeting"), claiming ignorance ("I didn't know you were expecting me"), problematizing a precondition ("we weren't supposed tomeet before 12"), or denial ("I didn't do it"); (b) utterance reducing severity of offense ("I'm only ten minutes late").

(28) Offer of repair: Speaker offering to remedy damage inflicted onoffended party by specific compensation for the offense ("I'll payfor the damage", "I'll have it marked tomorrow").

(29) Verbal redress: Speaker showing concern for offended party ("Ihope you weren't offended"), efforts to appease ("Let me buy youa drink") or promise of forbearance ("It won't happen again").

Interrater-reliability was established through consensus coding by threeraters (E, JEI, JEA data) and two raters (J data). For the analyses inSection 4, differences between the four groups are reported in percentagesof positive and negative transfer in the JEI and JEA groups. Since thisphase of the study was essentially exploratory, no inferential statisticswere applied. Statistical tests of the transfer analyses in Section 5 werebased on total strategy frequencies tallied for the intermediate and


advanced ESL speakers, and the American baseline groups. All Chisquare tests were calculated with a correction for continuity.

4. Results

4.1. Contextual effects on pragmatic transfer

Comparison of the contextual assessments provided by the nativespeakers of English and japanese showed that there was strong agreement in the two groups' perception of status, obligation to apologize, andlikelihood of apology acceptance. On each of these factors, only two contexts received different assessments. The assessment of likelihood ofapology acceptance parallels the one found by Bergman and Kasper(1993). In their study, ratings of the same contexts obtained from nativespeakers of American English and Thai revealed likelihood of apologyacceptance to be the factor on which both groups agreed most, only threecontexts receiving diverging ratings. In contrast, the findings for statusand obligation to apologize in the present study deviate considerablyfrom previous research. Beebe and Takahashi's studies of face-threateningacts in japanese and English (e.g., Beebe - Takahashi 1989; Takahashi - Beebe 1993) consistently demonstrated a much stronger differentiation of status-relationships in japanese than in American speakers'performance of such acts. We had therefore suspected that a similardifference would show up in japanese and American raters' perceptionsof status in offense contexts. Obligation to apologize was the factor onwhich Thais and Americans differed most - eleven out of the twentyoffense contexts received different scores from these groups (Bergman Kasper 1993). In the present study, the most different ratings were givenon offenders' face loss (7 contexts), offended party's face loss (9 contexts), and social distance (11 contexts).

Because previous research had demonstrated that context assessmentaffects the selection of apology strategies, we reasoned that pragmatictransfer can preliminarily be predicted from similarities and differences ofnative speakers' contextual assessments. Thus, it was assumed thatsimilar native speaker ratings predict positive transfer of apologypatterns, whilst different ratings predict negative transfer. A context was

Table 1. Transfer predictions based on contextual assessment ~

0\0\

Context Social Social Severity Obligation Likelihood Off. H'S TransferDistance Power Face-Loss Face-Loss Prediction ~

~Damaged Car + - + + - - fZJ

~+ ~

~

Ruined Mag. - + - + + + - fZJ;::s--~

Contradiction + + + + + - + positive ~

Poor Teacher + + - + + + + positive IPers.Mess.L-H - + + + + + - positive ~

Pers.Mess.H-L + + + + + + + positive 0<Bus.Mess L-H positive

~

+ + + + + - - ;::s--

Bus.Mess H-L - + + + + - - fZJ~.

~

Order Change - - + + - + + fZJ ~

Food on Cust. - + + + + + + positive IWrong Order - + + + - + + positive 0Food on Wait. + + + + + + + positive

~Messed Up Bag + + + - + + - fZJ ~

~

Bonsai Tree - + + + + + + positive ~""'t

Broken Statue + + + + + - + positive ICustoms Form - + + + + + + positive ~

Ungraded Pap. - + + + + + + positive ~0

Borrowed Book + + - - + - - negative ~~

Failed Student - - + + + + - fZJ

Cheating Student - + - + + + + positive

+ =same rating by J and E- =different rating by J and Eoff. face-loss =offender's face-loss


categorized as "similar" when ratings did not differ on five or morefactors, and as "different" when more than four factors were rateddifferently. The categorizations were based on separate multivariateanalyses of variance for each of the twenty contexts. The dependentvariables in these analyses were the seven contextual factors, and theindependent variable was membership in the native Japanese or Americangroups. Table 1 shows for each context whether it was rated the same ordifferent on each of the seven factors, and the transfer predictions basedon these ratings.

Since the American and Japanese raters in this study generally showedmore agreement than disagreement in their context assessment, predictions of positive transfer by far outnumbered those of negative transfer.Only Borrowed Book received different ratings on four factors, and wastherefore expected to elicit negative transfer. For six contexts, same anddifferent ratings counterbalanced each other; hence no transfer predictions could be made. Thirteen contexts were rated similar and thuspredicted to elicit positive transfer of apology strategies.

4.2. Transfer of apology strategies

In order to determine whether pragmatic transfer was operative, amodified version of Selinker's (1969) operational definition of languagetransfer was adopted from Kasper (1992). According to this definition,lack of statistically significant differences in the frequencies of a pragmatic feature in the first language, second language, and interlanguagecan be operationally defined as positive transfer. Statistically significantdifferences in the frequencies of a pragmatic feature between interlanguage-second language and first language-second language and lack ofstatistically significant difference between interlanguage and firstlanguage can be operationally defined as negative transfer. Appliedto the identification of transfer of apology strategies, positive transferobtains when there is no statistically significant difference in the useof an apology strategy between E and J, E and JEIIJEA, and J andJEIIJEA. Negative transfer requires statistically significant differencesin strategy use between E-J and E-JEIIJEA and no statistically significantdifferences between J-JEIIJEA. Because in this study, J was a subset ofJEI and JEA (i.e., the same subjects served as learners and first languagenative speakers), comparisons between J and JEIIJEA were not carriedout.


Table 2. Predicted positive transfer of apology strategies

Context JEI% JEA%

Contradiction 67 67Poor Teacher 83 67Personal Message L-H 83 83Personal Message H-L 50 67Business Message L-H 83 100Food on Customer 50 67Wrong Order 83 83Food on Waiter 17 33Bonsai Tree 83 100Broken Statue 100 83Customs Form 100 100Ungraded Paper 17 67Cheating Student 83 83

The transfer predictions established through comparison of contextualfactors in the previous section were matched against the actual occurrenceof transfer on each apology strategy in each offense context. There washigh correspondence between the predictions of positive transfer and itsoccurrence. Table 2 displays the percentage of strategies which were transferred positively according to the context-based transfer predictions.

In the majority of contexts where positive transfer had been predicted,the converging social perceptions of Japanese and American subjects wasreflected in the same use of apology strategies by both learner groups.This match was even greater in the case of the advanced learners, whooutperformed the intermediate learners in six contexts. The intermediatelearners, however, did better than the advanced learners in Poor Teacherand Broken Statue. There were only two contexts where the prediction ofpositive transfer was not borne out. In Food on Waiter, both learnergroups apologized differently from the American native speakers on allbut one measure (JEI: Taking on Responsibility) and two measures,respectively (JEA: IFID and Taking on Responsibility). In UngradedPaper, the only strategy which the intermediate learners used in the sameway as the Americans was upgrading apologetic force. Table 3summarizes how the prediction of positive transfer was matched by theactual use of apology strategies.

Again, the advanced learners' performance compared better to thetransfer predictions than the intermediate learners' in their choice of four


Table 3. Positively transferred apology strategies (all contexts)

IFID 0/0 UG% TR% DG% Rep 0/0 VR 0/0

JEI 77 77 85 54 62 62JEA 77 85 92 54 77 85

IFID = Illocutionary Force Indicating DeviceUG = UpgradingTR = Taking on ResponsibilityDG = DowngradingRep = Offer of RepairVR = Verbal Redress

apology strategies. The strategy where both learner groups displayed leastpositive transfer was Downgrading apologetic force.

Overall, native speakers' social perceptions proved to be an excellentpredictor of positive pragmatic transfer: Where Japanese and Americans'contextual assessments converged, Japanese learners of English wereprone to use the same strategies in their interlanguage apologizing as bothnative speaker groups. This was even more true of the advanced learnersthan of the intermediate learners, which makes sense in light of theassumption that advanced learners are likely to be more acculturatedthan intermediate learners, and have the linguistic facility to transferpragmatic strategies from their native language where they perceive thisas consistent with target use.

A requirement for positive transfer is obviously that the native speakergroups display the same kind of behavior. In the instances of positivetransfer noted above, the similarity of native speakers' strategy use couldbe seen as the behavioral correlate to their converging social perceptionsof contextual factors. Interestingly, native speakers also preferred thesame apology strategies where they did not agree in their context assessment, and in these contexts, the learners displayed the same strategychoices as the native speakers did. In other words, positive pragmatictransfer of apology strategies occurred even in contexts where this wasnot predicted by the contextual assessment. Table 4 summarizes the contexts in which apology strategies were positively transferred contrary tothe transfer prediction.

Curiously, Borrowed Book, the only context for which negativetransfer had been predicted, achieved the highest scores on actual positive


Table 4. Positive transfer contrary to prediction

Context JEI% JEA%

Damaged Car 83 67Ruined Magazine 67 33Business Message H-L 83 83Order Change 67 83Messed-up Bag 67 50Borrowed Book 83 100Failed Student 83 83

transfer. In Ruined Magazine and Messed-Up Bag, the advanced learners'strategy choices differed most from the native speakers' apology patternsand were thus more in accordance with the prediction of zero transfer.

The advanced learners followed the prediction of zero transfer intheir selection of IFIDs, the intermediate learners on Repair. On all otherstrategies, the learners converged in their strategy selection with thenative speakers and did not differ in terms of proficiency. Since only thenative speakers' sociopragmatic perceptions were elicited, we have noway of knowing whether the learners viewed the seven contexts similarlyto the American native speakers and selected their apology strategies inaccordance with their contextual assessment, or whether they assessedthese contexts differently from the American judgements but nonethelessfollowed the same pattern of strategy selection.

The same mismatch between transfer predictions and transferoccurrence was observable in the few instances of negative transfer. Just

Table 5. Unpredicted positive transfer of apology strategies (all contexts)

JEIJEA

IFID%

8629

UG%

7186

TR%

8686

DG%

10086

Rep 0/0

2971

VR 0/0

7171

IFID =Illocutionary Force Indicating DeviceUG =UpgradingTR =Taking on ResponsibilityDG =DowngradingRep =Offer of RepairVR =Verbal Redress


as the only prediction of negative transfer (in the Borrowed Book context,see above) contrasted with actual positive transfer of strategy choice,negative transfer occurred in contexts were positive transfer had beenpredicted. Table 6 lists the strategies which were transferred negativelyfrom Japanese, and the frequencies by which these strategies were usedby the learners and native speakers.

Each strategy was thus transferred negatively at least once. In theUngraded Paper context, a professor promised to have a student's papergraded but has not done it yet when the student comes to pick it up. Inresponse to the student's request "Can I have my paper back now?", theintermediate learners apologized less by means of an explicit apology(IFID) than the American native speakers and advanced learners. Typicalresponses given by the four language groups are listed in (31) through (37).

Table 6. Negative transfer of strategy choice

IFID in Ungraded PaperJ%53

JEI%57

A%83 JEI < A

Upgrading in Food on CustomerJ% JEl%50 48

Upgrading in Food on WaiterJ% JEI%20 23J % JEA 0/020 43

Taking on Responsibility in ContradictionJ% JEI%63 66

A 0/077

A%73A%73

A%37

JEI < A

JEI < A

JEA<A

JEI > A

JEI < AA%83

Downgrading Responsibility or Severity in Cheating StudentJ % JEI % A 0/083 71 48 JEI > AJ% JEA% A%83 79 48 JEA > A

Offer of Repair in Food on CustomerJ% JEI%3.3 55

Verbal Redress in ContradictionJ% JEI%30 31

A%70 JEI < A


(31) J: Chotto matte, sugu tsukeru kara"Wait just a minute. I will mark your paper right away"JEI: Not yet. Have a seat and wait a minute.A & JEA: Sorry, I haven't quite finished it. Could you comeback tomorrow?

In Food on Customer, a waiter spills food on a customer's dress. Whilethe Americans and advanced learners intensified the waiter's apology tothe customer, the intermediate learners followed the Japanese nativespeakers by upgrading apologetic force less in this context, as isillustrated in (32). Mooshiwakearimasen here might be taken as an upgraded apology since its literal meaning is 'inexcusable'. However,mooshiwakearimasen is the appropriate form for a waiter at an expensive restaurant to apologize to the customer. In upgraded responses,intensifiers such as taihen 'very/ awfully' or hontooni 'truly/really' areexplicitly added to mooshiwakearimasen. Therefore, the expressionmooshiwakearimasen without any intensifiers was counted as non-upgraded form.

(32) J: Mooshiwakearimasen, okyakusama. "It IS inexcusable,sir/ma'am" .JEI: Oh, I'm sorry.A & JEA: Oh no! I'm so incredibly sorry!

The same offense in reverse role relationships is represented in Food onWaiter, where a guest in a restaurant knocks off a waiter's tray whengetting up and the food spills all over the waiter. While the Americansubjects had the guest apologize to the waiter just as profusely as thewaiter to the guest in the previous situation, most respondents in bothlearner groups and the Japanese native speakers apologized less stronglyby not intensifying their expression of regret.

(33) J: Gomenasai. "I'm sorry."JEI & JEA: I'm sorry. Are you okay?A: Oh, my God! I'm terribly sorry. I'm such a klutz.

However, almost twice as many of the advanced learners as the intermediate learners upgraded apologetic force, suggesting that they are on theirway to abandoning the native Japanese pattern of differentiating apologyintensification according to interlocutor status in favor of the moreegalitarian target usage.

Transfer and proficiency in interlanguage apologizing 173 '

In Contradiction, a teacher contradicts something that another teachersaid at a staff meeting, and hurt his colleague's feelings. Most of theintermediate learners and Japanese native speakers redressed this offenseby explicitly assuming responsibility for it. The American respondentspreferred to offer sympathy in this situation, as will be seen below. Thecontrasting patterns are illustrated by the responses in (34).

(34) J: Konoaida wa gomen. Warugi wa nakattan da."I am sorry for the other day. I didn't mean it".JEI: I'm sorry. I hurt your feelings, I think.A & JEA: I hope you didn't take what I said personally.

In the role of a student who plagiarized for a term paper and is found outby his professor (Cheating Student), the learners and Japanese nativespeakers downplayed the offense by finding excuses or claimingignorance. The Americans used this strategy less, probably on theassumption that attempts to downplay the offense would make their caseworse rather than better, cf., (35).

(35) J: Sumimasen. Tesuto toka iroiro to isogashikute, shimekiri nimaniawazu sono tame ni shikata ga nakatan desu. Hi woaratamete teishutsu shitemo ii desu ka."I sorry. I was busy preparing tests, so in order to prepare themin time it was the only thing I could do. May I submit it later?"JEI: I'm sorry for copying, but I was busy.JEA: I'm sorry. I tried to write the essay myself, but the state

ment in the text was so nicely written, and I didn't knowyou would consider if we copy from the text.

A: Well, actually, I did get some of my ideas from a book.

In the Food on Customer context, the American and advancedrespondents had the waiter offer repair to the customer, such aspromising to have her dress cleaned. Half of the intermediate learners didnot offer repair, and thus assumed a medium position between Americanand Japanese usage. Only a single Japanese respondent offered repair tothis offense. Typical responses are given in (36).

(36) J: Taihen mooshiwake gozaimasen. "I am very sorry"JEI: Oh, I'm sorry. Are you all right?A & JEA: Oh, I'm terribly sorry. We'll have the suit cleaned foryou.


Finally, the intermediate learners followed the Japanese pattern of notexpressing much concern for the insulted teacher in Contradiction. Mostof the American respondents offered tokens of concern for the offendedcolleague's feelings, cf., (37). Half of the advanced learner groupexpressed concern; the advanced learners thus placed themselves betweenthe Japanese/intermediate and American respondents in this context.

(37) J: Warukatta keredo are wa boku no sochokku na iken de atte,kimi wo kizutsukeru tsumori wa nakatta."Sorry, but that was my honest opinion and I didn't mean tohurt you."JEI: I know that I hurt you, but that which I said was myopinion. I think I was right.A: I'm sorry, Jennifer. Did my words upset you?

Most of the negative transfer occurred in contexts with a high powerdifferential between the interlocutors, regardless whether the offenderwas the higher status participant (Food on Waiter, Ungraded Paper) or inthe lower status position (Food on Customer, Cheating Student). In Foodon Customer and in both of the student-professor contexts, socialdistance was perceived differently by Japanese and American raters. It ispossible, therefore, that despite the overall agreement in context perception, the diverging assessment of social distance is contributive to thedifferences in strategy use. This interpretation is consistent with previousstudies, which demonstrated a complex interaction of contextual factorsand choice of apology strategies. As noted above, each context factor hasa different weight as trigger of strategy choice, and the same factor affectsthe selection of some strategies but not others. In previous studies, socialdistance was shown to affect offender's assumption of responsibility forthe offense (Bergman - Kasper 1993) and, more tentatively, their choiceof IFID (Olshtain 1989). This study suggests that diverging perceptionsof social distance can account for different choices of IFID, Upgrading ofapologetic force, and offer of Repair. It does not explain, however, whydifferent assessments of social distance affect some but not all contexts.A closer look at two pairs of contexts which differ only in the interlocutors' power relationship indicates that the direction of the status differential is the crucial factor. Japanese and American raters did not differin their assessment of social power in Food on Customer / Food onWaiter and Borrowed Book / Ungraded Paper. However, in both of thesepaired contexts, Japanese and Americans differed in their actual use of


the IFIDs in the student-professor situations and of Upgrading in thewaiter-customer contexts. Americans used IFIDs equally often in thestudent-professor situation no matter whether the offender was theprofessor or the stuent (83 0/0), and they upgraded their apologies in thecustomer-waiter contexts to the same extent regardless of offender'sstatus (75 %). The japanese respondents, by contrast, made their selection of both strategies contingent on the direction of the status relationship. In the high to low contexts, they apologized less by means of anexplicit formula (professor ~ student (53 0/0)) and intensified apologeticforce less (customer~ waiter (20 % )) than in the corresponding low tohigh contexts (student~ professor: 87 % IFID; waiter ~ customer: 50 %upgrading). This contrast in power differentiation is consistent withBarnlund and Yoshioka's (1990) observation that japanese offenders aremore status-sensitive in their choice of apology strategy than Americans.The learners who followed the native model thus transferred the statusdifferential apology pattern from japanese to English in these contexts.As an instance of sociopragmatic transfer, this finding fits in well withTakahashi and Beebe's (e.g., 1993) work on face-threatening acts injapanese-English interlanguage. Their studies showed that compared toAmericans, native speakers of japanese employ a more distinctly statusdifferentiating approach to corrections and refusals. Mitigators such assofteners and expressions of regret were used more frequently by thestatus lower to the status higher interlocutor than vice versa by bothAmericans and japanese; however, the japanese respondents accentuatedthe status difference more than the Americans did. just as the japaneselearners of English transferred the status-differential patterns of mitigation to their performance of face-threatening acts in English, the intermediate learners in this study aggravated apologetic force according tothe status-differential first language model.

5. Extensions to recent studies of Japanese-Americanapology

The foregoing discussion of transferred apology strategies was based onDialog Construction codings devised by Bergman and Kasper (1993). Inorder to examine the generality of pragmatic transfer, however, it isimportant to compare the results of the present study with other cross-


cultural studies of apology in japanese and American contexts. Barnlundand Yoshioka (1990) found, for example, that japanese were more likelyto offer several types of apology for a transgression than Americansrating the same situation. They were also more likely to offer significantlymore repair by suggesting some form of compensation for their transgressions than the Americans. In contrast, Americans were found to bemore likely to provide a rationale - more excuses, justifications anddowngrading of the severity of the transgression. japanese appear to beequally direct as Americans in apologizing for perceived wrongdoings. AsBarnlund and Yoshioka used different units of measurement (ratings) anda different classificatory system, we will equate the Dialog ConstructionQuestionnaire results with the most obvious of the Barnlund andYoshioka findings in order to establish comparability between the crosscultural contrasts identified by Barnlund and Yoshioka and instances ofnegative transfer in our study.

In the Barnlund and Yoshioka study, the variance in preferredstrategies is taken to represent major pragmatic differences between thetwo cultural norms. Such differences can potentially result in negativepragmatic transfer, as defined above.

Four apology strategies involving pragmatic contrasts in the Barnlundand Yoshioka research were matched with the most comparable fivestrategy types in the present study. These were "explaining the situation"/downgrading, "saying directly 'I am very sorry'''/IFID and upgrading, "offering to do something for the other person"/ repair, "apologizing directly; several ways several times"/ (multiple apologies for asingle transgression).

Instances of apology strategies for the same three groups JEI, JEA, andA were reanalyzed for the twenty independent Dialog ConstructionQuestionnaire contexts. As the preceding section of this study hassuggested, there should be a greater likelihood that the lEIs would usejapanese strategies than the JEAs. The results of the comparisons arepresented and discussed individually below.

5.1. "Explaining the situation"/ downgrading

In Barnlund and Yoshioka, "explaining the situation" is used lessfrequently by japanese than by American offenders. Contrary toBarnlund and Yoshioka's findings, the present study revealed that in onlyone context out of 20 is there a significant difference between Americans


and Japanese in the downgrading/rationalizing category. Specifically, theJEIs used the preferred American strategy of rationalizing the transgression more than the American group. There was no such differencebetween the JEAs and the native speakers of English. The sole differencewas for Poor Teacher.

In this situation teacher A accuses teacher B of being a poor teacherat a staff meeting (Chi-square = 6.38; p = .0118). Such a differencecould indicate a transitional period during which the Japanese learnersover-accommodate toward what they perceive the American normto be, and in a sense 'out-American' the Americans. In no context dothe advanced learners differ from the Americans in terms of thefrequency of downgrading/rationalization, a finding not in agreementwith the pragmatic contrasts established by Barnlund and Yoshioka.This discrepancy may stem from the fact that all of the Japaneserespondents were residents of Honolulu, and had perhaps hadsufficient exposure to the Hawaiian-American norms of downgradingand reference to circumstantial causes of the transgression instead oftaking responsibility, whether such taking of responsibility is warrantedor not.

5.2. "Offering to do something for the other person"/repair

According to Barnlund and Yoshioka, Japanese offenders are more proneto offer compensation for an infraction than Americans. We thereforepredicted that the learners would offer more repair than the Americannative speakers. However, with one exception, the learners differed fromAmericans in their repair offers by providing less rather than more repair.The intermediate learners were much more prone to undersupply repairthan the advanced learners were. The contexts in which the JEIs differedfrom the Americans in their offers of repair are listed in (38) through(44).

(38) Damaged Car: A has had an accident with a car borrowedfrom B.Chi-square = 4.31 p = 0.037

(39) Ruined Magazine: A borrowed a magazine from B and spilledcoffee over it.Chi-square =4.31 p =0.037


(40) Business Message Low to High: At an office, a junior colleagueforgot to pass on an important business message to a seniorcolleague.Chi-square = 7.18 p = 0.007

(41) Business Message High to Low: At an office, a senior colleagueforgot to pass on an important business message to a juniorcolleague.Chi-square = 7.06 p = 0.007

(42) Food on Customer: At a restaurant, a waiter spills food on acustomer's clothes.Chi-square =3.88 p =0.048

(43) Food on Waiter: At a restaurant, a customer spills food on awaiter.Chi-square = 7.12 p = 0.007

(44) Failed Student: A professor misplaced a student's term paperand failed the student.Chi-square = 5.42 P = 0.019

Since the contexts in which the intermediate learners offer less repairinvolve different status relationships and degrees of social distance, theirdivergent apology behavior cannot be explained in terms of contextexternal factors. However, with the exception of Ruined Magazine, theoffenses are all high severity infractions. For the American subjects,repair offers are appropriate ways of redress to these kinds of transgressions. The learners' failure to offer compensation on a regular basisin these contexts suggests that they underdifferentiate their selection ofrepair offer according to high versus low severity offenses.

In contrast to the intermediate learners, the frequency of repair offersby the advanced ESL speakers is very similar to the native speakers ofEnglish. Only in Damaged Car (Chi-square = 4.13; p = 0.037), Food onWaiter (Chi-square =7.12; p =0.007), and Ungraded Paper (Chi-square =4.31; p = 0.037) do the advanced learners differ from native speakers.

In Damaged Car and Food on Waiter, like their intermediateproficiency counterparts, the advanced learners provide significantly lessapology than the native speakers for the transgressions. This may indicatethat while advanced learners are in general familiar with American


apology strategies, in highly marked and unfamiliar contexts such asDamaged Car, they know they cannot revert to japanese strategies but donot have the experiential basis to extrapolate from their repertoire ofsecond language pragmatic strategies. Familiarity with social contexts hasbeen shown to influence interlanguage pragmatic performance generally(Eisenstein - Bodman 1986), and pragmatic transfer specifically(Takahashi 1992).

Food on Waiter represents the influence of role differentiation awareness between Americans and japanese. An infraction such as the one inFood on Waiter does not warrant an offer of compensation from thejapanese because of the role/status differential implicit in customer/service employee relations. Ungraded Paper is the only instance ofnegative transfer (Chi-square 4.31; p = 0.037) in the lEAs' repair offers.

We note that the lEIs did not differ from the native speakers of Englishin this context (a professor failing to grade a student paper on time).Here, the advanced japanese are more inclined to see repair offer fromthe professor as appropriate, whereas their low-proficiency counterpartsdo not. What may appear to be negative transfer by the advancedjapanese in this context may actually be indicative of more subtle andcomplex pragmatic influences. They may, for instance, realize that thesecond language status differential does not require high to low repair.The advanced japanese would therefore be less inclined to considerno repair as appropriate. They might not however realize thatthe American norm is not built on a single egalitarian principle,and that it might be subject to real world constraints such as the factthat professors are notoriously tardy and expect that students understandthis.

5.3. "Direct apology"/IFID and upgrading

This category of apology strategy involves the speaker providing a clearand direct apology for an infraction. The directness is most overtlyexpressed as some variant of the IFID "I am sorry", and is hereconsidered distinct from an indirect form of apology such as "it is ashame it had to turn out that way". Barnlund and Yoshioka find that forboth Americans and japanese, direct apology is the most highly preferredform of redress. japanese are even more inclined to employ directapology. Learners can therefore be expected to use equal or surpassAmericans in their use of direct apology.


This expectation was confirmed. There was only one context, Food onWaiter, in which the JEIs offered direct apology less often than theAmerican native speakers (Chi-square =7.32; p =0.006). The statusdifferential between customer and waiter does not seem to require directapology in the view of the intermediate learners.

5.4. Multiple apologies

The Barnlund and Yoshioka study found that the Japanese are more inclined to provide multiple apologies or apologetic paraphrases thanAmericans. The function of the multiple apology in the Japanese milieu isto demarcate the speaker's responsibility for the transgression, and toprovide signals of sincerity for the apology. Multiple apology shouldtherefore be a prime candidate for negative transfer in the dialog construction tasks used in this study. Indeed, the JEI group should transfermultiple apologies more than the advanced learners. However, the resultssuggest that differences in the frequency of multiple apology is not ascommon as expected. For the JEIs, multiple apology was significantly different from the native English speakers in only two contexts. In MessedUp Bag, the JEIs provided more multiple apologies than the Americans(Chi-square = 4.27; p = 0.038), but in Ungraded Paper, they chose thisstrategy significantly less than the American NS (Chi-square = 4.56;p = 0.017).

For the advanced speakers, the likelihood of negative transfer of themultiple apology strategy can be considered less than that for the intermediate learners. This fact is borne out in the present study. Only in onecontext did the JEAs differ from the native speaker of English. Here, theJapanese provided significantly more multiple apologies (Chi-square =5.82; p =0.015). The transgression in Ruined Magazine is one that canbe considered avoidable. Presumably, since the perpetrator was remissin not preventing the mishap, the severity of the transgression becomesmore acute in the minds of the advanced English as a second languagespeakers.

6. Discussion

Compared to the pervasive effect of positive transfer, negative transferwas infrequently at work in the learners' apology performance. Yet, two


important findings emerge from the analysis of negative transfer. For onething, in only two instances did the advanced learners transfer theirapology behavior from japanese to English when japanese and Americanapology patterns differed, whereas the intermediate group did the samein six instances. From this, it follows that the advanced learners have abetter ability to emulate American apology behavior than the intermediate learners. Secondly, except for taking on responsibility inContradiction, the intermediate learners differed from the Americans inthat they provided fewer instances of the apology strategy in question.Their negative transfer of apology strategies thus consisted in adopting aless elaborated, first language-based approach to redress offenses than theAmerican native speakers and the advanced learners. This study, then,does not lend support to Takahashi and Beebe's (1987; 1993) contentionthat advanced learners display more negative pragmatic transfer because'they have the rope to hang themselves'. Rather, when advanced japaneselearners provide responses to exceptional situations for which they havelittle experience to rely on, they are inclined not to transfer first languagestrategies which they suspect to be insufficient for the context. Theadvanced learners may still not have developed the pragmatic wherewithal to provide the same responses to the subtleties of such situationsas do the native speakers of American English.

There are noteworthy similarities and differences in the transferbehavior of the intermediate learners in this study and the intermediateThai learners of English in Bergman and Kasper (1993). The Thailearners' performance suggested negative transfer of Upgrading andRepair in only one context and of IFID and Taking on Responsibility intwo contexts. Their patterns of negative pragmatic transfer on thesestrategies was thus quite similar to that of the japanese intermediatelearners. However, the Thai learners transferred negatively on their use ofDowngrading in three contexts where the japanese intermediate learnersdid not transfer negatively at all and on Verbal Redress in as many as sixcontexts. Furthermore, these negative transfers were the result of oversupplying the strategy in question, rather than under-using it, which iswhat the japanese learners were inclined to do.

Previous research has demonstrated that negative pragmatic transfer ismore prevalent in foreign language contexts than in second languagecontexts (Takahashi - Beebe 1987). This difference in learning contextscan partly account for the variance in negative transfer between the Thaiand japanese learners: the Thai learners were in an English as a foreignlanguage context, whereas the japanese learners were in an English as a


second language environment. This generalization is also borne out in thecomparison of the Barnlund and Yoshioka predictions with the secondlanguage acquisition patterns observed here. The richer opportunities forinput and productive use of English in the English as a second languagecontext surely put the japanese learners at an advantage. Furthermore,the variety of English which served as target norm in both the Thai andjapanese studies was Hawai'i Standard English, which was consistentwith the input variety received by the japanese in Honolulu, but not bythe Thais in Bangkok. Another reason for the differences between theBarnlund and Yoshioka conclusions and those of the present study relatesto differences in the sociolinguistic norms of the three speech communities. Since the Americans in the Barnlund and Yoshioka study werepresumably from two distinct homogeneous populations in japanand on the U. S. mainland, whereas the Americans and japanese in thepresent study were from a single heterogeneous speech community inHonolulu, we might surmise that there was a greater likelihood forexposure, accommodation and convergence in the Hawaiian milieu.

Given that the focus of this study is apology, it is perhaps most appropriate to apologize for the obvious limitations of the study itself. With theact of apology as the center of much cross-cultural miscommunication, andits status in perceptions of duty, responsibility and liability in American andjapanese societies, it is of particular importance to continue the investigation of apology across a wide variety of communicative domains.

Appendix 1

A sample item from the Dialog Construction Questionnaire (Ruined Magazine)

At a friend's homeAnn and Bill are both 35 years old and are good friends. Ann borrowed acomputer magazine from Bill. Unfortunately, Ann spilled coffee on the magazineand damaged it. She is now returning it to Bill.

Bill: What happened to my magazine?Ann:Bill:

Appendix 2

A sample item from the Assessment Questionnaire (Ruined Magazine)

At a friend's home


Ann and Bill are both 35 years old and are good friends. Ann borrowed acomputer magazine from Bill. Unfortunately, Ann spilled coffee on the magazineand damaged it. She is now returning it to Bill.

1. How CLOSE are Ann and Bill in this situation?1 2 3 4very close

5very distant

2. What is the STATUS RELATIONSHIP between Ann (A) and Bill (B)?1 2 3 4 5A>B A=B A<B

3. How SERIOUS is Ann's OFFENSE?1 2 3 4 5very serious not serious

4. Does Ann have the OBLIGATON to apologize?1 2 3 4 5strong obligation no obligation

5. Is Bill LIKELY to ACCEPT Ann's apology?1 2 3 4 5very likely unlikely

6. Is this situation EMBARRASSING to Ann?1 2 3 4 5not embarrassing very embarrassing

7. Is this situation EMBARRASSING to Bill?1 2 3 4 5not embarrassing very embarrassing

References

Barnlund, Dean C. - Miho Yoshioka1990 "Apologies: Japanese and American styles", International Journal of

Intercultural Relations 14: 193-206.Beebe, Leslie M. - Tomoko Takahashi

1989 "Sociolinguistic variation in face-threatening speech acts", in:Miriam Eisenstein (ed.), 199-218.

Beebe, Leslie M. - Tomoko Takahashi - Robin Uliss-Weltz1990 "Pragmatic transfer in ESL refusals", In: Stephen D. Krashen

Robin Scarcella - Elaine Andersen (eds.), 55 -73.Bergman, Marc L. - Gabriele Kasper

1993 "Perception and performance in native and nonnative apology", in:Gabriele Kasper - Shoshana Blum-Kulka (eds.), 82-107.


Blum-Kulka, Shoshana1982 "Learning how to say what you mean in a second language: A study

of speech act performance of learners of Hebrew as a secondlanguage", Applied Linguistics 3: 29-59.

1991 "Interlanguage pragmatics: The case of requests", in: RobertPhillipson - Eric Kellerman - Larry Selinker - Michael SharwoodSmith - Merrill Swain (eds.), 255-272.

Blum-Kulka, Shoshana - Elite Olshtain1986 "Too many words: Length of utterance and pragmatic failure",

Studies in Second Language Acquisition 8: 165-180.Blum-Kulka, Shoshana - Hadass Sheffer

1993 "The metapragmatic discourse of American-Israeli families atdinner", in: Gabriele Kasper - Shoshana Blum-Kulka (eds.),196-223.

Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper (eds.)1989 Cross-cultural pragmatics. Norwood, NJ: Ablex.

Bodman, Jean - Miriam Eisenstein1988 "May God increase your bounty: The expression of gratitude in

English by native and nonnative speakers", Cross Currents 151:1-21.

Borkin, Ann - Susan M. Reinhart1978 '''Excuse me' and 'I'm sorry"', TESOL Quarterly 12: 57-70.

Clyne, Michael1979 "Communicative competence in contact", ITL 43: 17-37.

Cohen, Andrew D. - Elite Olshtain1981 "Developing a measure of sociocultural competence: The case of

apology"; Language Learning 31: 113-134.Dechert, Hans W. - Manfred Raupach (eds.)

1989 Transfer in language production. Norwood, NJ: Ablex.Eisenstein, Miriam (ed.)

1989 The dynamic interlanguage. New York: Plenum.Eisenstein, Miriam - Jean W. Bodman

1986 '''I very appreciate': Expressions of gratitude by native and nonnative speakers of American English", Applied Linguistics 7:167-185.

Frerch, Claus - Gabriele Kasper1989 "Internal and external modification in interlanguage request realiza

tion", in: Shoshana Blum-Kulka - Juliane House - GabrieleKasper (eds.), 221-247.

Fraser, Bruce1980 "Conversational mitigation", Journal of Pragmatics 4: 341-350.

Garcia, Carmen1989 "Apologizing in English: Politeness strategies used by native and

non-native speakers", Multilingua 8: 3-20.Gass, Susan - Carolyn Madden (eds.)

1985 Input in second language acquisition. Rowley, MA: NewburyHouse.



House.Giles, Howard - Patricia Johnson

1987 "Ethnolinguistic identity theory: A social psychological approach tolanguage maintenance", International Journal of the Sociology ofLanguage 68: 69-99.

Goffman, Erving1971 Relations in public. Harmondsworth: Penguin.

Holmes, Janet1989 "Sex differences and apologies: One aspect of communicative

competence", Applied Linguistics 10: 194-213.House, Juliane

1988 "'Oh excuse me please ... ': Apologizing in a foreign language", in:Bernhard Kettemann - Peter Bierbaumer - Alwin Fill Annemarie Karpf (eds.), 303-327.

House, Juliane - Gabriele Kasper1987 "Interlanguage pragmatics: Requesting in a foreign language", in:

Wolfgang Lorscher - Rainer Schulze (eds.), Perspectives onlanguage in performance 1250-1288.

Jordens, Peter1977 "Rules, grammatical intuitions, and strategies in foreign language

learning", Interlanguage Studies Bulletin 2: 5 -76.Kasper, Gabriele

1982 "Teaching-induced aspects of interlanguage discourse", Studies inSecond Language Acquisition 4: 99-113.

1992 "Pragmatic transfer", Second Language Research 8: 203-231.Kasper, Gabriele (ed.)

1992 Pragmatics of Japanese as native and target language. (SecondLanguage Teaching & Curriculum Center Technical Report #3).Honolulu, HI: University of Hawai'i Press.

Kasper, Gabriele - Shoshana Blum-Kulka (eds.)1993 Interlanguage pragmatics. New York: Oxford University Press.

Kellerman, Eric1977 "Toward a characterization of the strategy of transfer in second

language learning", Interlanguage Studies Bulletin 2: 58-145.Kettemann, Bernhard - Peter Bierbaumer - Alwin Fill - Annemarie Karpf

(eds.).1988 Englisch als Zweitsprache [English as a second language}.

Tiibingen: Narr.Krashen, Stephen - Robin Scarcella - Elaine Andersen (eds.)

1990 On the development of communicative competence in a secondlanguage. Cambridge, MA: Newbury House.

Leech, Geoffrey N.1983 Principles of pragmatics. London: Longman.

Lorscher, Wolfgang - Rainer Schulze (eds.)1987 Perspectives on language in performance. Tiibingen: Narr.


Olshtain, Elite1983 "Sociocultural competence and language transfer: The case of

apology", in: Susan Gass - Larry Selinker (eds.), Language transferin language learning~ 232-249.

1989 "Apologies across languages", in: Shoshana Blum-Kulka - JulianeHouse - Gabriele Kasper (eds.), 155-173.

Olshtain, Elite - Shoshana Blum-Kulka1985 "Degree of approximation: Nonnative reactions to native speech

act behavior", in: Susan Gass - Carolyn Madden (eds.), 303325.

Olshtain, Elite - Andrew Cohen1983 "Apology: A speech act set", in: Nessa Wolfson - Elliot Judd (eds.),

18-35.1989 "Speech act behavior across languages", in: Hans W. Dechert

Manfred Raupach (eds.), 53-67.Phillipson, Robert - Eric Kellerman - Larry Selinker - Michael Sharwood

Smith (eds.)1991 Foreign/second language pedagogy research. Clevedon: Multilingual

Matters.Robinson, Mary Ann

1992 "Introspective methodology in interlanguage pragmatics research",in: Gabriele Kasper (ed.), 27-82.

Selinker, Larry1969 "Language transfer", General Linguistics 9: 67-92.

Takahashi, Satomi - Margaret A. DuFon1989 Cross-linguistic influence in indirectness: The case of English direc

tive performed by native Japanese speakers. [Unpublishedmanuscript, University of Hawaii at Manoa.]

Takahashi, Satomi1992 "Transferability of indirect request strategies", University of Hawaii

Working Papers in ESL 11.1 : 69-124.Takahashi, Tomoko - Leslie M. Beebe

1987 "The development of pragmatic competence by Japanese learners ofEnglish", JALT Journal 8: 131-155.

Takahashi, Tomoko - Leslie M. Beebe1993 "Cross-linguistic influence in the speech act of correction", in:

Gabriele Kasper - Shoshana Blum-Kulka (eds.), 138-157.Tanaka, Noriko

1991 "An investigation of apology: Japanese in comparison withAustralian", Meikai Daigaku Gaikokugo Gakubu Ronshu 4:35-53.

Thomas, Jenny1983 "Cross-cultural pragmatic failure", Applied Linguistics 4: 91-112.

Trosborg, Anna1987 "Apology strategies in natives/non-natives", Journal of Pragmatics

11: 147-167.


Vollmer, Helmut J. - Elite Olshtain1989 "The language of apologies in German", in: Shoshana Blum-Kulka

- Juliane House - Gabriele Kasper (eds.), 197-218.Wolfson, Nessa

1989 Perspectives. Cambrigde, MA: Newbury House.Wolfson, Nessa - Elliot Judd (eds.)


Wolfson, Nessa - Thomas Marmor - Steve Jones1989 "Problems in the comparison of speech acts across cultures", in:


Face-threatening acts

My grade's too low:The speech act set of complaining

Beth Murphy - Joyce Neu

1. Background

Since Hymes' (1972) introduction of the concept of communicativecompetence, there has been a heightened awareness that one of the mostimportant goals in acquiring a second language is learning the appropriate use of linguistic forms. That is, the acquisition of communicativecompetence involves knowledge of the form as well as the appropriatecontext in which to say it. Yet although many second language researchers recognize the importance of communicative competence insecond language acquisition, research on the specific components of communicative competence remains scarce. In fact, Hymes (1972) called fora broader framework for linguists to examine the social aspects oflanguage as well as the structural. In his definition of communicativecompetence, Hymes suggested that we look not only at grammaticality,but also at the feasibility and the appropriateness of utterances. Throughhis investigation of speech acts, Searle (1971; 1975) also demonstratedthe importance of context in examining the relationship betweenlanguage and action. Speech acts are the acts we perform when we speak:e. g., congratulating, thanking, requesting. These acts, in order to be performed satisfactorily, must take place in a specified context of situation.Thus, research on speech acts is crucial in that it can provide us with thesocial context and the appropriate sociocultural rules surrounding nativespeaker utterances.

This chapter is concerned with one aspect of communicativecompetence - the performance of the speech act set l of complaint.Performing this speech act set, which usually involves a face-threateningact, appears to be challenging even for native speakers of English whooften pre-plan how they will go about making a complaint. Given thedelicate nature of this speech act set, non-native speakers of English mayunintentionally express inappropriate complaints. These complaints, inturn, may not only sound non-native, but may serve to alienate the inter-

192 Beth Murphy - Joyce Neu

locutor. The objectives of the study reported in this chapter are to isolatethe components of the complaint speech act set as performed byAmerican native speakers of English and by Korean non-native speakersof English and to analyze how American native speakers perceive Koreannon-native speakers' performance of complaint speech act sets.

2. Speech act production

The foundation for the study of speech acts was laid by Austin (1962)and Searle (1971; 1975). A departure from Chomskyan linguistics, theirwork situated language within a social context, providing us with agreater awareness of the importance of sociolinguistic knowledge in theproduction of speech. By focusing on speech acts rather than on isolatedsentences, Austin found that a class of verbs, called performative verbs,function as the accomplishment of an action by their being spoken. Thatis, by uttering "I apologize," the act of apologizing is performed.Contributing to the development of speech act theory, Searle (1971)defined speech acts as the smallest units of rule-governed meaningfulcommunication.

Researchers such as Manes (1983) and Wolfson (1983) have drawnfrom Austin and Searle's development of speech act theory and appliedit to the analysis of a specific speech act - the compliment. The studiesby Manes and Wolfson reveal that American English speakers compliment on appearance, new acquisitions, and effort. These serve thefunctions of maintaining solidarity and reinforcing social values.Wolfson's 1983 study further demonstrates that the status relationshipbetween the participants plays an important role in the topic of theoffered compliment.

Wolfson, D'Amico-Reisner, and Huber (1983) investigated the sociocultural rules of invitations in middle-class white American society. Theirstudy demonstrates that interpreting this speech act may be troublesomefor non-native speakers since invitations are ambiguous much of the time.Unambiguous invitations are produced when the speaker refers to a timeand/or place/activity as well as requests a response from the addressee.The invitation, "Do you want to go to the movies tomorrow night?"contains all three components that make up an unambiguous invitation.However, Wolfson, D'Amico-Reisner, and Huber found that these unambiguous invitations occurred in only one-third of the data. Ambiguous

The speech act set of complaining 193

invitations, on the other hand, which provide for negotiation betweeninterlocutors, were found to be more representative of how nativespeakers of English arrange for social commitments. These invitationscontain a "lead" that is a question or comment that opens up thepossibility for an unambiguous invitation to follow. For example, "Areyou busy tomorrow night?" is a lead that serves to establish the availability of the person before the issuance of an unambiguous invitation.Thus, although leads often precede invitations, they do not in themselvesconstitute an invitation. Consequently, the distinction between leads andfull invitations may result in misinterpretations between native speakersand non-native speakers.

Cohen and Olshtain (1981) expanded the concept of the speech act intheir analysis of apologies. They found that semantic formulas, whetherin combination or alone, can be used to perform an act of apology. Forexample, a speaker may express an apology, "I'm sorry"; acknowledgeresponsibility for a perceived wrong, "It's my fault"; offer a repair for thewrong, "I'll pay for it"; promise forbearance, "It won't happen again";or explain the situation, "There was a traffic jam." Because each of theseformulas is in itself a speech act, they make up the speech act set ofapology. In another study, Olshtain (1983) used this apology speech actset as a framework for her intercultural research. She discovered thatsome cultures preferred one or another formula, or combination orformulas, to express an apology. American English speakers, for example,tend to express an apology and follow it with an explanation of the situation whereas Hebre'Y speakers tend to give an explanation only.Furthermore, Olshtain noted that these language-specific preferences maycause a second language learner to sound inappropriate in the targetlanguage. By providing just an explanation and no apology, Hebrewspeakers who transfer this formula will undoubtedly sound rude inEnglish.

While Cohen and Olshtain referred to the speech act set, Ferrara(1985), drawing on van Dijk (1977) explained the need to talk aboutmacro speech acts. Although these studies concentrated on a singlespeech act, Ferrara (1985) has argued that speech act theory must beextended to capture' the core action of discourse. He claims that there isa distinction between understanding the text semantically (what the talkmeans) and understanding the text pragmatically (what the talk does).According to Ferrara, capturing "what the talk does" involves identifying the set of macro speech acts that "underlies the entire text and insuresits pragmatic coherence" (1985: 149). Although the macro speech act is


composed of myriad single speech acts, it can only be determined byreference to the dominant speech acts in the text. Ferrara (1985) thusargues for a broader unit of analysis, the macro speech act, as a way ofmore effectively investigating the relationship between language andaction. Second language investigators have more commonly referred tothe "speech act set," a term that appears synonymous with Ferrara'smacro speech act and van Dijk's 1977 use of "macrostructure."

3. Speech act acceptability judgments

To ascertain whether the language-specific preferences noted by Olshtainresult in socially inappropriate utterances, native speaker judgments areneeded. To concentrate only on the productive aspect leaves the researcher with a partial picture of the consequences of speech act performance. In fact, Olshtain and Cohen (1983) highlight the importanceof sociolinguistic acceptability judgments in their discussion of methodological issues concerning the study of speech acts. They argue thatnative speakers' judgments of non-native speakers' performance areneeded to determine whether or not communication has been successful.In a later article, Cohen and Olshtain (1985) again focus on the use ofacceptability judgments as one way of capturing and examining speechact behavior more effectively. The study reported in this chapter has usedacceptability judgments as a critical component necessary to expand ourunderstanding of how and when non-native speakers fail to communicateeffectively. Cohen and Olshtain discuss ways in which non-nativespeakers' performances of the apology speech act set may be deviant.They say that this deviance may be due to "a lack of compatibilitybetween [the] speaker's intent and [the] hearer's standards of acceptability" (1985: 178). The conclusion is that we must investigate performance both from the speaker's perspective and from the listener's aswell.

Although few in number, some studies have concentrated on how nonnative speakers' production is perceived by native speakers. Olshtain andBlum-Kulka (1985), for example, conducted a study in which 172 nativespeakers of English, 160 native speakers of Hebrew, and 124 non-nativespeakers of Hebrew judged the appropriateness of request and apologystrategies in Hebrew. They found that as the length of stay in the targetspeech community increased, non-native speakers' acceptability judg-


ments became increasingly similar to native speakers' judgments. Carrelland Konneker (1980) compared politeness judgments of AmericanEnglish native speakers and non-native speakers of English from differentlanguage backgrounds. The subjects judged and ranked eight differentrequest strategies in English in three specially contextualized situations interms of levels of politeness. Their study revealed that although there wasa high correlation in their politeness judgments of native speakers andnon-native speakers, the non-native speakers of English perceived agreater number of levels of politeness than did the native speakers. Theresearchers concluded that the English as a second language learners'greater number of politeness distinctions may be a result of their oversensitivity to syntactic-semantic features.

Few studies appear to have examined speech acts both from thespeakers' and the listeners' perspectives. The present study examines theproductive aspect of the complaint speech act set as performed byAmerican native speakers and Korean non-native speakers of English.The listener perspective was investigated through the use of Americannative speaker acceptability judgments of the Koreans' performances.

4. Rationale and research questions

The present study has two objectives: First, to compare the componentsof the speech act set produced by American native speakers and Koreannon-native speakers of English; and second, to ascertain how these speechact sets were judged by native speakers on a number of factors (listedbelow). The questions are as listed in (1 a)-(8).

Production:

(1 a) Given the context of expressing disapproval to a professor (i. e.,someone of higher status), will Ameri~an native speakers ofEnglish produce a complaint speech act set?

(1 b) Give the same context, will Korean non-native speakers ofEnglish produce a complaint speech act set?

(2a) If American native speakers produce a complaint speech act set,what are the components of this set?

(2b) If Korean non-native speakers of English produce a complaintspeech act set, what are the components of this set?


(3) Do the speech act sets of complaint produced by Americannative speakers of English and Korean non-native speakers ofEnglish differ? If so, how?

Judgments:

(4) Will American native speakers of English judge the Korean nonnative speakers' productions as aggressive?

(5) Will American native speakers of English judge the Korean nonnative speakers' productions as respectful?

(6) Will American native speakers of English judge Korean nonnative speakers' productions as credible?

(7) Will American native speakers of English judge the Korean nonnative speakers' productions as appropriate?

(8) Will American native speakers of English say that, given thesame situation, they would not produce the same kind ofcomplaint as the Korean non-native speakers of English?

5. Methodology

To answer the questions (1a) through (8), a two-part study was undertaken to analyze both the production of the complaint speech act set bynative speakers and by non-native speakers and to examine nativespeakers' judgments of the non-native speakers' productions.

5.1. Subjects

For the productive part of the study, 14 male American graduate studentsand 14 male Korean graduate students participated (hence the use of thepronoun he/his throughout). All were graduate students in variousdepartments at Penn State University. Twenty-seven American studentsrecruited from an advanced undergraduate course in American-Englishphonetics participated in the receptive part of the study-the acceptability judgments. Twenty three of these subjects were undergraduates; fourwere graduate students. All of these subjects received extra credit in theircourse for their participation in the study and subjects in both parts ofthe study signed informed consent forms agreeing to participate in thestudy.


5.2. Instruments

The speech act data were collected via an oral discourse completion taskthat consisted of a hypothetical situation typed on a sheet of paper. Thesituation placed the subject in the position of a student whose paper hadbeen unfairly marked by a professor. The subject was directed to "talk"to the professor (see Appendix A for the situation).

The instrument in the acceptability judgments part of the studyconsisted of a questionnaire composed of ten "yes-no" questions and oneopen-ended question. Of the ten "yes-no" questions, five were distractoritems. The five "yes-no" questions that were analyzed measured thenative speakers' perceptions of the Korean non-native speakers' responsesin the context of a student expressing disapproval about an undeservedgrade (see Appendix B for the questionnaire and Appendix C for thewritten transcripts of the complainer and the criticizer). The open-endedquestion asked the subjects to explain their answers to one question: "Ifyou were the student in this situation, would your approach be differentfrom the student you've just heard? Please explain your answer for bothspeaker-student 1 and speaker-student 2."

5.3. Procedures

For the first component of the study, the subjects were instructed to readthe instrument carefully and to voice their reaction into a tape recorder.They were told to give as complete a response as if they were actuallyconversing with the professor. All of the subjects indicated that they hadeither experienced the situation or could easily imagine it. Each subjectwas alone during the discourse completion task. The speech data werethen transcribed into written form.

In part two of the study, the questionnaire was distributed to the 27American students. The subjects were told they were going to listen totwo non-native speakers of English role-playing as students who areexpressing their disapproval to a professor about an undeserved grade.Because it was found that most Korean non-native speakers produced acriticism rather than a complaint speech act set, to determine the salienceof this difference for native speakers of English, one complaint and onecriticism were selected for native speaker judgment. For ease of referencein this chapter, the Korean non-native speakers who complained and whocriticized are referred to as "the complainer" and "the criticizer,"


respectively. To validate the researchers' assumption that the complaint and the criticism were indeed distinct speech act sets, the 27American students were asked to evaluate the Korean nonnative speakers' responses. Seventy-four percent of the Americansjudged the complainer's response to be a complaint and the criticizer'sresponse to be a criticism. The American subjects were then instructedto answer the questionnaire by judging only the content of the responsein terms 'of the situation, not the grammar or tone of voice. Theyfirst listened to and judged the complainer's response. They then heardand judged the criticizer's performance. The American native speakerswere given a written transcript of each speaker to distract themfrom focusing on the Korean non-native speakers' grammar or pronunciation problems. Both of the non-native speakers' responses wererandomly selected from the speech data base obtained in the first part ofthe study.

5.4. Methods for data analysis

The speech data elicited for the first part of the study were examinedusing Cohen and Olshtain's (1981) definition of speech act set. WhileCohen and Olshtain investigated the semantic formulas of the speech actset of apology, this study attempted to identify the semantic formulas ofthe speech act set of complaint.

For the second part of the study, a .05 alpha level of significance wasset for a Chi-square analysis of the American students' responses to thefive "yes-no" questions. For all differences between responses analyzed,the Yate's Correction for Continuity was used.

The free responses to the open-ended question were compiled toexamine the reasons why the subjects' approaches would or would notdiffer from those taken by the two Korean non-native speakers.

6. Results

The first research question asked if American native speakers, given thesituation presented in this study of a student expressing disapprovalabout a grade received on a paper to a professor, would produce a com-


plaint speech act set. The answer appears to be yes. A description of thisspeech act set is in section 6.1.

The second research question asked if Korean non-native speakers,given the same context, would also produce a complaint speech act set.Only three of the 14 Korean non-native speakers-produced speech actsets contained a component that could be termed a complaint. Therefore,most Korean non-native speakers do not produce a complaint speech actset in this situation.

6.1. Semantic components of the speech act set of complaint/criticism

The speech act sets produced by the American native speakers and theKorean non-native speakers contained certain core components, with thenotable difference between American native speakers and Korean nonnative speakers occurring in the complaint element of the speech act set.

An explanation of purpose was provided by each of the 14 Americanmale subjects before actually stating the complaint. That is, eachAmerican subject explained the purpose of his presence to the interlocutor (the professor). Give the context of venturing into the office of hisprofessor, the student must explain why he is there. Thus, a student maysay, "Excuse me, I just dropped by to talk about my grade." Through theexplanation of purpose, the student sets the stage and gives cause for hispresence. Common examples from the Americans' responses includesamples (a) and (b).

(a) Subject A52: Hello, Professor Filano. Vh, I got my paper back hereand after looking through it ...

(b) Subject A7: Vh, Dr. Smith, I just came by to see if I could talk aboutmy paper.

An explanation of purpose was produced by 13 out of the 14 Koreansubjects. Their responses paralleled those of the American subjects in thatthe focus is also on the grade/paper as the cause of their presence. Thus,the Korean non-native speakers appear to express the explanation ofpurpose appropriately and in much the same manner as the Americannative speakers, as shown in samples (c) and (d).

(c) Subject K1: Good afternoon Professor. Vh, I have something to talkto you about my paper ...


(d) Subject K4: Hello, Dr. Brown. I came by to ask if some questionabout the papers I handed-I hand in to you.

A complaint was produced by each of the American subjects only afterthey explained their purpose. A complaint is initiated when the speakerperceives he has been treated unfairly by the instructor. Moreover, in thiscontext of a student-professor relationship, the Americans are keenlyaware of the status difference and are thus constrained from producing acriticism. Thus, to express disapproval, it appears that the most sociallyappropriate option available to the Americans is to complain. The topicof all of the Americans' complaints concerns the grade of the paper/test,as shown in samples (e) and (f).

(e) Subject A1: I think, uh, it's my opinion maybe the grade was a littlelow.

(f) Subject AS: I was kind of upset with my grade. Uh, I know thata lot of problems are mine but there are certain areas that I uhwasn't totally in agreement with what you said were my problemshere.

A complaint seems to occur in only three of the 14 Korean subjects'responses. These three complaints are similar to the Americans' in thatthey reflect concern and discontent with the grade, as shown in samples(g) and (h).

(g) Subject KS: Professor, I little bit disappointed uh in my grade ...(h) Subject K9: I was very upset at uh seeing my grade on my research

paper.

However, the other 11 Korean non-native speakers appear to haveproduced not complaints, but criticisms. This appears to be a seriousdeviation from the speech data obtained from the American nativespeakers. Thus, the criticisms may indicate inappropriate behavior on thepart of the Korean non-native speakers. Examples of criticisms producedby the 11 Korean subjects are shown in samples (i) and (j).

(i) Subject K6: But you just only look at your point of view and uh youjust didn't recognize my point.

(j) Subject K8: ... but I think my point of view is more relevant to thesubject and that has to be emphasized ...


A justification occurs in each of the American native speakers responses.The justification demonstrates that the complainer can support hisclaims, that he can explain the reasons why he is complaining. In thissense, the justification lends credence to the complaint. It appears that ifa student complains to a professor, he must justify it by providingwell-developed reasons as support. In this situation, the topic of 13out of 14 of the Americans' justifications focused on the time, work,or effort that went into producing the paper, as shown in samples (k)and (1).

(k) Subject A2: I put a lot of time and effort in this ...(1) Subject A3: Uh, but a great deal of work went into that ...

A justification was similarly given by all of the Korean non-nativespeakers and like the American native speakers, most of the Korean nonnative speakers' justifications referred to the amount of time, effort, orwork put into the paper. Examples of the Korean non-native speakers'responses include samples (m) and (n).

(m) Subject K5: ... and urn I read my articles and readings and also I uhput down all your lectures and so I tried to answer to my best ...

(n) Subject K6: ... I-I spent a lot of times to do research on thispaper ...

A candidate solution: request appears to be the last speech act producedby all of the American native speakers. A candidate solution: request is aproposed course of action offered by the complainer as a way to resolvethe problem. From the data, it is also evident that the American nativespeakers propose the candidate solution in the form of a request ratherthan as a demand. Thus, an asymmetrical status relationship between thecomplainer and the recipient appears to constrain how American nativespeakers can offer a solution to a professor. In this situation, all of theAmericans' candidate solutions involved requesting the professor toreconsider an/or discuss the grade/paper. Representative examples areshown in samples (0) and (p).

(0) Subject A9: I would appreciate it if you would reconsider my grade.(p) Subject A12: ... so, I'd like to perhaps set up a time when we can get

together and discuss in detail what you felt- what was missing orwhat was lacking.


A candidate-solution: request was produced by 12 of the Korean nonnative speakers. These requests are similar to those of the Americannative speakers in that the focus is on asking the professor to reconsiderthe grade/paper, as shown in samples (q) and (r).

(q) Subject K9: And I hope you can reconsider my grade and urn I hopeyou can reevaluate my research paper.

(r) Subject K10: So, I honestly ask you to reconsider my paper ...

Two of the Korean non-native speakers, however, demanded a candidatesolution instead of requesting one, as shown in samples (s) and (t).

(s) Subject K7: Your grading is not fair and uh so it must be changed.(t) Subject K8: I hope you give me some plausible explanation where

defects or handicaps that the points that I missed that should becovered.

Tables 1 and 2 represent a summary of the complaint and criticism speechact sets produced by American native speakers of English and Koreannon-native speakers of English, respectively.

Thus, the first research question is answered affirmatively: Americanmale native speakers of English produce a complaint speech act set inresponse to the context of expressing disapproval to a professor. All 14

Table 1. Complaint speech act set produced by American NS of English

Components

Explanation ofpurpose (EOP)

Complaint(CaMP)

Criticism (CRIT)

Justification(JUST)

Candidatesolution: request

Candidatesolution: demand

Frequency Examplesof Use

100 % I just came by to see if I could talk about mypaper.

100 % I think, uh, it's in my opinion maybe the gradewas a little low.

0%

100 % I put a lot of time and effort into it.

100 % I would appreciate it if you would reconsidermy grade.

0%


Table 2. Criticism speech act set produced by Korean NNS of English

Components Frequency Examplesof Use

EOP 93% I have something to talk to you about mypaper.

CaMP 21 0/0 I little bit disappointed uh in my grade.

CRIT 79% But you just only look at your point of viewand uh you didn't recognize the my point.

JUST 1000/0 I spent a lot of times to do research on thispaper.

CS: request 79% I honestly ask you to reconsider my paper.

CS: demand 21 0/0 Your grading is not fair and uh so it must bechanged.

American native speakers produced a complaint speech act set. The dataindicate that the complaint speech act set consists of four basic components: Explanation of purpose, complaint, justification, and candidatesolution: request.

The second research question that asks whether Korean non-nativespeakers of English will produce a complaint speech act set given thesame context is not answered in the affirmative. Rather, the findings ofthis study indicate that 11 out of 14 Korean non-native speakers (79 %)produce a criticism in the slot where the American native speakersproduce a complaint. This deviation is a critical one in that the complaintspeech act is a fundamental element of the entire speech act set. Excludingthe complaint component and substituting a criticism component maychange the illocutionary force of the set into a criticism. In the followingsection, we will describe the linguistic formulas that differentiate thecomplaint from the criticism component.

6.2. Linguistic features of the speech act set of complaint/criticism

When faced with expressing disapproval to a professor about an undeserved grade, each of the 14 American native speakers produced a


complaint. A complaint serves the function of assuming some of theresponsibility of the perceived wrongdoing and may contain any of thecharacteristics listed in I-V.

I. Use of pronoun "we" in two primary ways: 1) to indicate that bothparties share the blame: e. g., "I know we have a different point ofview on this subject" (Al), "I know through class discussions we'vedisagreed with each other ... " (A2); and 2) as a way of negotiatingthe problem (the grade): e.g., "I hope that we could sit down anddiscuss the paper ... " (A12), "I'd like to perhaps set up a time whenwe can get together and discuss in detail what you felt wasmissing ... " (A12).

II. Use of questioning to ask for advice, for permission to explain oneself, or to get the listener to reconsider or discuss the problem.Thirteen of the 14 American native speakers used the modals"could" or "would" in their questions, thereby indicating politenessor hesitation, for example, "I'd like any advice you could give me ofthe problems I have ... " (AS), "Do you have a minute so that wecould go over the paper together?" (A14).

III. Depersonalization of the problem, transferring blame from theprofessor to the problem. That is, the focus centers on "the paper,""it," "the grade" rather than on a person. Examples include "I feelthis grade may reflect a difference of opinion" (A9), " ... but it seemslike the grade and the effort kind of don't mix" (A8).

IV. Use of mitigators to soften the complaint. All of the American nativespeakers used one or more of the following mitigators: "maybe,""just," "a little," "kind of," "perhaps," "really," "you know,""I don't know." In the complaint "I think uh it's just in myopinion maybe the grade was a little low," virtually every word is amitigator.

V. Acceptance of partial responsibility for the problem: " ... and uhperhaps it wasn't quite as polished as both of us would have liked,but the content is there ... " (A3), "Uh, I know that a lot of theproblems are mine ... " (AS), "Uh, I'm wondering whether it was justlack of explanation on my part or if you had overlooked a few thingswhich I had presented" (A8).

As stated earlier, most of the Korean non-native speakers did not producea complaint but rather a criticism. A criticism abdicates responsibility forthe problem and places the blame on another party, in this situation, the


professor. The data reveal the formulas I - III used in the Korean nonnative speakers' criticisms.

I. Use of second person + modal "should" that indicates that thespeaker is in a position to dictate the behavior of the listener (theprofessor). Examples include" uh you should take account uh intouh I'm a foreigner ... " (K7), " you should be equal to everybody"(K11).

II. Personalization of the problem, placing the blame on the other, e. g.,"But you urn decreased my grade ... " (K4), " ... but you just onlylook at your point of view and you just didn't recognize the my uhpoint" (K6).

III. Refusal to accept responsibility for the problem. For example, "Mypoint of view is more relevant to the subject and that has to beemphasized than yours" (K8), "I don't think that uh in that in mypaper there is any fault or any errors" (K11), "I think my thoughtmy idea is totally correct" (K13).

From the data, then, complaint and criticism appear to be two distinctspeech act sets that contain different components. The speech act set ofcomplaint, as produced by American native speakers, allows the speakerto assume partial responsibility for the problem. In fact, given theasymmetrical status and power relationships in the situation analyzed inthis study, the American native speaker appears to be constrained toaccept responsibility as a matter of politeness and concern for theongoing relationship. By assuming partial responsibility, the complaintfacilitates a negotiation between the student and the professor. Thecomplaint speech act set, given an asymmetrical situation, exhibitsfeatures such as 1) acceptance of partial responsibility, 2) depersonalization of the problem, 3) use of questions, 4) use of mitigators, and 5) useof the pronoun "we."

On the other hand, the speech act set of criticism functions to deny theresponsibility of the speaker for the problem. Consequently, the criticismsgiven by 11 of the Korean non-native speakers place the blame for thepoor grade on the professor. The speech act set of criticism is characterized in this instance by 1) abdication of responsibility, 2) personalizationof the problem, and 3) use of second person + the modal "should." Whilesharing many components, these two speech act sets contain suchdifferent features that they appear to be two distinct speech act sets (seeTable 3).


Table 3. Complaint and criticism speech act sets: A comparis<;>n

Complaint

1. acceptance of partial responsibility2. depersonalization of the problem3. questioning techniques that use

modals "would" and/or "could"4. use of mitigators5. use of pronoun "V\Te"

Criticism

1. denial of responsibility2. personalization of the problem3. use of the modal "should"

7. Sociolinguistic acceptability judgments

This section presents the results of the American English native speakers'acceptability judgments of a complaint produced by a Korean non-nativespeaker (referred to as "the complainer") and a criticism produced by aKorean non-native speaker (referred to as "the criticizer").

Analysis of the acceptability judgments data revealed significantdifferences in the Americans' perceptions of the complainer and thecriticizer in terms of 1) the speakers' aggressiveness (Research Question3); 2) the speakers' respectfulness (Research Question 4); the speakers'credibility (Research Question 5); and the speakers' appropriateness(Research Question 6) (see Table 4).

7.1. Aggressiveness

Significantly more Americans perceived the crItIcIzer to be aggressive(X 2 =12.96, df =1, P < .01). Over 65 % of the Americans judged thespeaker who criticized the professor to be aggressive, whereas only 150/0judged the speaker who complained to be aggressive. The answer toresearch question 3 is yes: Korean non-native speakers who producea criticism in lieu of a complaint in this setting are perceived asmore aggressive than the Korean non-native speaker who produced acriticism.

7.2. Respectfulness

A significant difference was found in native speaker perceptions betweenthe complainer and the criticizer (X2 = 14.84, df = 1, p < .01). While over


Table 4. American NS acceptability judgments of the complainer and the criticizer

QUESTION Com- Criti- X 2

(N=27) plainer ClzerYES NO YES NO

1. Is the student aggressive? 4 23 18 9 12.9150/0 850/0 670/0 330/0 6**

2. Is the student respectful? 23 4 8 19 14.885 0/0 15% 30% 700/0 4**

3. Does the student present a credible 19 8 8 19 7.41case for obtaining his goal? 70% 30% 30% 70% **

4. Is the student's response appropriate 20 7 11 16 4.85for the situation? 74% 26% 41% 59% *

5. If you were the student in this 20 7 20 7 0.00situation, would your approach be 74% 26% 74% 26% nsdifferent from the student you havejust heard?

* p<0.05** p < 0.01

85 % of the American native speakers judged the complaint as respectful,about 30 % rated the criticism as respectful. Over 70 % judged thecriticizer as disrespectful, thus answering the fifth research question in thenegative: American native speakers do not perceive Korean non-nativespeakers' criticisms as respectful in this setting.

7.3. Credibility

When asked, "Does the student present a credible case for obtaining hisgoal?," the American native speakers' perceptions towards the complainer and the criticizer were again significantly different (X 2 = 7.41,df = 1, P < .01). Over 70 % of the native speakers judged the complaint ascredible and the criticism as not credible. The answer to the sixth researchquestion is negative: American native speakers do not perceive most ofKorean non-native speakers' productions as credible.


7.4. Appropriateness

A significant difference was also found in the native speakers' judgmentstowards the complainer and the criticizer (X 2 = 4.85, df = 1, p < .05). Aconsiderably greater number of American native speakers rated the complaint as more appropriate (74.1 %) than the criticism (40.70/0). Thus,research question 6 is negative:. Most Korean non-native speakers'responses to the situation were not perceived as appropriate by Americannative speakers.

7.5. Differences in approach to the situation

The last research question regards whether the American native speakerjudges would take a different approach to the situation than had theKorean non-native speakers they had just heard. No significant differencewas found in the American native speakers' perceptions of the two nonnative speakers, but they did respond that their approach would bedifferent: 74 % of the American native speakers said their approachwould be different; 26 % responded that their approach would not bedifferent.

7.6. Results of the open-ended question

The questionnaire asked the American native speaker judges to explaintheir answers for one of the yes-no questions, the only one, in fact, thatturned out not to be significant in the difference between the complainerand the criticizer.

For the complainer, numerous answers were given. No general themeappeared in the responses. However, for the criticizer, 21 of the 27American judges responded that if they were the student in that setting,they would not have been so aggressive and critical of the professor. Nineof the American native speakers selected one sentence produced by thecriticizer as highly inappropriate: " ... my point of view is more relevantto the subject and that has to be emphasized than yours in this area- inthis specific area ... " One American wrote, "I would be much morecareful not to attack the prof or accuse him/her of grading (sic). It'sdangerous to say, 'my point of view is more relevant!' All these thingsevoke a reaction that is negative, if it wasn't there before!" Another judge


wrote, "No prof would consider regrading his paper - he's too rude."Ten of the responses indicated that the criticism given by the non-nativespeakers not only challenged the professor, but would put him/her on thedefensive. One subject wrote, "His (the non-native speaker) approachwill certainly result with the professor responding in a defensive way; thisis going to hurt him in obtaining his goal." Another native speaker judgestated, "He straight out challenges the prof which is not the right thingto do in order to get a better grade."

8. Discussion

The findings of this study are that when placed in the position of expressing disapproval to a professor about a grade, American native speakersproduce a complaint. Korean non-native speakers in this same situationproduce a criticism, not a complaint. Although the components of explanation of purpose, justification, and candidate solution: request areparallel in the American native speakers' and the Korean non-nativespeakers' performances, the component of criticism instead of complaintis produced in the Korean non-native speakers' utterances. An earlierstudy of American native speakers (Murphy 1989) revealed that an asymmetrical status relationship seems to prevent American students fromcriticizing a professor face-to-face. In the context of a student complaining to a professor about a poor grade, the criticisms produced by theKorean non-native speakers demonstrate inappropriate sociolinguisticbehavior.

This study's analysis of the characteristics of the complaint andcriticism speech act sets provides some insights into why the Korean nonnative speakers' productions are a serious deviation from appropriatebehavior. The American native speakers, by presenting a complaint, notonly accept partial responsibility for the grade, but also facilitate apotential negotiation between themselves and the professor. To achievetheir goal of getting the professor to reconsider the grade, the Americannative speakers depersonalized the problem, incorporated politeness andhesitation markers such as the use of modals and mitigators, and used theinclusive pronoun "we." Using these t:lative speaker strategies as a basisfor comparison, the source of inappropriateness of the Korean non-nativespeakers' productions becomes more apparent. By personalizing theproblem, by not accepting any part of the responsibility for the problem


and telling the professor what action he/she "should" take, the Koreannon-native speakers place the blame on the professor. Criticizing asuperior, when a complaint is called for, was perceived as aggressive,challenging, lacking credibility, and inappropriate. It is unlikely that theKorean non-native speaker who produces a criticism in this context willachieve his/her goals. Given a setting in which a student disagrees with aprofessor's evaluation, American native speakers judged the complaintspeech act to be acceptable; the criticism was perceived not only as unacceptable, but as a dangerous deviation from the accepted norm.

9. Limitations

The study focused on the tape-recorded production of a response to onewritten situation by a small number (28) of male Korean non-nativespeakers. While we believe that this open-ended type of oral discoursecompletion task allowed the subjects to respond as they wished, therewas no interaction between speakers, as would happen in a normalconversation. Ideally, data would be collected in settings in which complaints would be naturally elicited. As others have noted, the difficultylies in how much data one would have to collect to assure the productionof an adequate number of the speech act under analysis. Another concernis that we focused on only one situation - that of a student complainingto a professor. This situation places the subject in an asymmetrical statusrelationship where the subject is of lower status than the person she isasked to complain to. Further research should investigate if there aredifferences in complaints lodged against peers and persons of higher orlower status. We would assume there would be. More data, usingdifferent status situations, males and females, and non-native speakersfrom different language backgrounds, would help to clarify the findingsof this study.

For the second part of the study that focused on American nativespeaker subject judgments of the Korean non-native speakers' productions, all of the American subjects were from the same class atPennsylvania State University (Speech Communication 410 - a course inIntroductory Phonetics). This course is a requirement for all speechcommunication majors at Pennsylvania State University and students inthe course are sensitized to non-native discourse in a way that mostAmericans would not be. This may have led the American subjects to


distinguish between the non-native speakers' complaints and criticisms ina way that others might not have. Another factor in the American nativespeakers' judgments is that although the American native speakers weretold to judge only the content of the Korean non-native speakers'responses, and not the grammatical accuracy or suprasegmentals such asintonation, these may have influenced the American native speakers'attitudes about the appropriateness of the non-native speakers' responses.One alternative would be to give the American native speaker judgestranscriptions of the non-native speakers' responses rather than havethem listen to the taped responses.

10. Implications and conclusion

This study has implications for both the methodology of second languageresearch and English as a second/foreign language instruction. First, thisstudy demonstrates the importance of getting native speaker acceptabilityjudgments of non-native speaker speech data. Native speaker perceptionsabout non-native speakers' productions provide support for researchers'intuitions about appropriateness and acceptability.

The study also provides insight into what American students considerappropriate behavior in an academic setting. Specifically, the speech dataproduced by the Americans as well as the American native speakers'judgments about non-native speakers' productions indicate that to complain to a professor is acceptable and appropriate behavior; however, tocriticize a professor to her face is considered aggressive and disrespectful.

The distinction between a complaint and a criticism is clearly a salientdistinction for American native speakers although clearly a difficult onefor Korean non-native speakers to make. These "distinctive features"make up speech acts and for speakers of different languages, differentfeatures are salient. Where there is a disjuncture between what is salientin one language and what is not in another, the non-native speaker either"hears" a difference that doesn't exist in the target language or doesn't"hear" a difference that native speakers do hear.

For teachers of English as a second/foreign language, these distinctivefeatures need to be taught. In the case of our study, the findings indicatethat international students (Koreans only?) need to learn how to asserttheir rights in an academic setting without being perceived as threateningor disrespectful. To assist international students in achieving this goal,


and by doing so, to avoid potential misunderstanding and even censure,this study has isolated several appropriate native speaker strategies available to English as a second/foreign language learners who find themselvesneeding to complain to a professor.

Speech act research has shown us how difficult it is for a non-nativespeaker to acquire communicative competence in a second language. Outof necessity do international students try to learn how to be perceived asappropriate by their American peers and professors. In our "globalvillage," perhaps it is past time to also argue for the instruction ofAmerican native speakers in the understanding of, and tolerance for, nonnative speaker productions.

Appendix A: Oral discourse completion task

Instrument A: Low status situation

You are handed back a paper by your professor. However, you are startled byyour grade and feel that you have been marked down for disagreeing with theprofessor's point of view rather than on any flaws in your content and analysis.You are particularly upset since you have spent weeks researching this paper andfeel the professor has ignored your effort through simple bias. You decide youmust speak to him/her about this. So, after class, you go to the professor duringoffice hours and say:


Appendix B: American NS Acceptability Judgment Questionnaire

Note: Please judge only the content of the Student Studentresponse, not the grammar or tone. 1 2Remember to answer in the context of therole situation (see instructions).

YES NO YES NO

1. Is the student aggressive (pushy)?

2. Is the student friendly?

3. Is the student respectful?

4. Is the student intelligent?

5. Is the student informal?

6. Is the student to the point?

7. Is the student confident?

8. Does the student present a crediblecase for obtaining his goal?

9. Is the student's response appropriatefor the situation?

10. If YOU were the student in thissituation, would your approach be differentfrom the student you have just heard?

Please explain your answer to question 10, for both student 1 and student 2.Continue on the reverse side if extra room is needed.

Student 1:

Student 2:

#1 #2

11. In an overall assessment of the speaker's CRIT COM CRIT COMresponse, is the student criticizing orcomplaining to the professor?


Appendix C: The Complainer and the Criticizer

The Complainer:

Hello, Dr. My name is ad uh I was very upset at uh seeing your mygrade on my research paper. I read that the I didn't fully follow uh your directionsduring class but you know this is graduate level class and I think the creacreativity is as important as the other things like I um I gotta follow uh yourresearch directions.

The Criticizer:

Professor, can I ask you there ask you why I got this grade for this paper? I spentseveral weeks at least several weeks to think about and planning and researchingfor this paper and I devoted myself wholly to this project and I don't-don't understand why you gave me the lower grade than I expected. I hope you give me someplausible explanation or some defects or handicaps that the points I missed thatshould be covered. (pause) I know you taughted the class a little bit differentpoint of views than uh but I think the my point of view is more relevant to thesubject and that has to be emphasized than yours in this area this specific area.So, I took this um point and developed it to the fullest length and elaborated onit. (pause) Thanks, I understand. But I don't think my work because a little bitdifferent from yours my point and my work has to graded lower than I expectedonly because of-of the fact that uh my point of views is different from yours. Ithink this a little bit unfair.

Notes

1. A speech act set is a combination of speech acts that, taken together, make upa complete speech act. That is, it is frequently the case that one utterance alonedoes not perform a speech act. Some examples are apologies and invitationswhere several utterances are necessay for the intended illocutionary act to beaccomplished.

2. The coding system used was A1-14 for the 14 American subjects and K1-14for the 14 Korean subjects.

References

Austin, John L.1962 How to do things with words. Oxford: Oxford University Press.

Black, Max (ed.)1965 Philosophy in America. London: Allen & Unwin Ltd.

Carrell, Patricia - Beverly Konneker1981 "Politeness: Comparing native and nonnative judgments", Language

Learning 31.1: 17-30.


Cohen, Andrew - Elite Olshtain1981 "Developing a measure of sociocultural competence: The case of

apology", Language Learning 31.1: 113-134.1985 "Comparing apologies across languages," in: Kurt Jankowsky (ed.),

175-184.Cole, Peter - Jerry Morgan (eds.)

1975 Syntax and semantics. New York: Academic Press.Ferrara, Alessandro.

1985 "Pragmatics", in: Teun van Dijk (ed.), 137-157.Gass, Susan - Carolyn Madden (eds.)

1985 Input in second language acquisition. Rowley, MA: NewburyHouse.


House.Hymes, Dell

1972 "On communicative competence," in: John B. Pride - Janet Holmes(eds.),269-293.

Jankowsky, Kurt (ed.)1985 Scientific and humanistic dimensions of language. Amsterdam: John

Benjamins.Manes, Joan

1983 "Compliments: A mirror of cultural values", in: Nessa Wolfson Elliot Judd (eds.), 96-102.

Murphy, Beth1989 Influence of status on the elicitation of two speech act sets:

Complaint and criticism. [Unpublished seminar paper, PennsylvaniaState University.]

Ochs, Elinor1979 "Transcription as theory", in: Elinor Ochs - Bambi Schieffelin

(eds.),43-72.Ochs, Elinor - Bambi Schieffelin (eds.)

1979 Developmental pragmatics. New York: Academic Press.Olshtain, Elite

1983 "Sociocultural competence and language transfer: The case ofapology", in: Susan Gass - Larry Selinker (eds.), 232-249.

Olshtain, Elite - Shoshana Blum-Kulka1985 "Degree of approximation: Nonnative reactions to native speech act

behavior", in: Susan Gass - Carolyn Madden (eds.), 303-323.Olshtain, Elite - Andrew Cohen

1983 "Apology: A speech act set", in: Nessa Wolfson - Elliot Judd (eds.),18-35.

Pride, John B. - Janet Holmes (eds.)1972 Sociolinguistics. Harmondsworth, England: Penguin Books.

Searle, John1965 "What is a speech act?", in: Max Black (ed.), 221-239.1975 "Indirect speech acts", in: Peter Cole - Jerry Morgan (eds.), 59-82.


van Dijk, Teun (ed.)1977 Text and context: Explorations in the semantics and pragmatics of

discourse. London: Longman.Wolfson, Nessa

1983 "An empirically based analysis of complimenting in AmericanEnglish", in: Nessa Wolfson - Elliot Judd (eds.), 82-95.

Wolfson, Nessa - Elliot Judd (eds.)1983 Sociolinguistics and language acquisition. Rowley, MA: Newbury

House.Wolfson, Nessa - Lynne D'Amico-Reisner - Lisa Huber

1983 "How to arrange for social commitments in American English: Theinvitation", in: Nessa Wolfson - Elliot Judd (eds.), 116-128.

Ethnographic interviewing as a researchtool in speech act analysis: The case of complaints

Diana Boxer

The ethnographic interview as a research tool has been virtually ignoredas a means of tapping the norms of communities both in research onspeech act usage among native speakers of particular languages andresearch on non-native speaker pragmatic transfer. With the exception ofvery few such studies (e. g., Katriel 1985), scholars have either concentrated on 1) analyzing spontaneous speech without the corroboration ofinsight from members of the speech community; 2) analyzing speechbehavior based on only native speaker impressions derived from discourse completion questionnaires (Discourse Completion Tests) withoutthe corroboration of spontaneous speech; or 3) analyzing speech behavior through some combination of 1) and 2). However, DiscourseCompletion Tests or more traditional survey techniques uncover onlyintuitions that native speakers have about how they should speak. Theydo not allow the researcher to discover the actual speech patterns ofnative speakers. By combining the use of Discourse Completion Testquestionnaires with the analysis of spontaneous speech, scholars cancompare the use of a specific speech act with the canonical shape of thepatterning of that speech act in the minds of community members (Beebe1985; Beebe - Cummings this volume). But while such a combinationallows for a system of checks on whether people indeed speak as theythink they ought to, it falls short of uncovering the tacit knowledge thatwe all have about how and why we speak as we do.

Research into the speech behavior of native speakers of English isimportant not only for establishing descriptions of how we performverbally in our day-to-day interactions with other native speakers, butalso for the purpose of making use of this baseline information in secondlanguage learning. From the viewpoint of ethnolinguistic study, it isimportant to discover what we as native speakers are doing when we usecertain speech acts in order to learn about our values and social system.From the viewpoint of second language acquisition, such knowledge is anecessary underpinning for the successful teaching of sociolinguistic

218 Diana Boxer

rules. It is now true that there are approximately as many non-nativespeakers of English in the world as native speakers. We must haveeffective communication among all English speakers if we are to avoidunwanted misunderstandings. Thus the need for sociolinguistic description and the application of sociolinguistic findings to TESOL is morecritical than ever. As TESOL becomes a burgeoning endeavor internationally, both in English as a second language settings and English as aforeign language settings, the sociolinguistic component of languagelearning must be attended to. It is now imperative that descriptiveanalyses of the sociolinguistic rules of English-speaking speech communities be carried out and widely replicated for increased generalizability andapplicability.

While it is true that the debate continues as to whether rules ofspeaking can or should be taught in English as a second language/Englishas a foreign language (Kachru 1988; Widdowson 1988), it is clear thatwhere the English as a second language is not a non-native institutionalized variety, the ability to communicate with native speakers accordingto native norms is important. Native speakers often forgive phonological, syntactic and lexical errors as clear signs that a speaker does nothave native control of a language. Native speakers, however, typicallyinterpret sociolinguistic errors as rudeness rather than as the transfer ofdifferent sociolinguistic rules (Ervin-Tripp 1972; Thomas 1983; Wolfson1981; 1983; 1989).

There is a rapidly increasing population of adult learners studyingEnglish as a second language in the United States. For these learners theacquisition of sociolinguistic competence is at least as important aslinguistic competence for successful language learning. These adultlearners are finding themselves ever more involved in communicationwith native speakers in settings including the academic, business,diplomatic, scientific and technical spheres. Since the potential for miscommunication is very great, it is of extreme timeliness that descriptiveanalyses of native speaker sociolinguistic behavior be undertaken andapplied.

This chapter reports on baseline findings on rules for the realizationand underlying social strategies of a specific speech act sequence.Corroborating evidence in the form of ethnographic interviews is discussed in depth. Described here are the results of two sets of interviews,one structured and one open-ended, that were conducted with the aim oftapping the knowledge of native informants about a speech event termed"troubles telling" (Jefferson - Lee 1981; Jefferson 1984), "troubles

Ethnographic Interviewing 219

talk" (Tannen 1990) or "troubles-sharing" (Hatch 1992). More specifically, it delves into what is frequently the initiating move of the speechevent, namely the speech act of "griping" or what is termed here "indirectcomplaining." The term "indirect complaint" is taken from the work ofD'Amico-Reisner (1985) on disapproval exchanges. Indirect complaintsare juxtaposed by D'Amico-Reisner with instances of direct complaint ordisapproval. Indirect complaints differ from instances of direct complaintin that the addressee is not held responsible for a perceived offense.Indirect complaint will be defined here as the expression of dissatisfactionto an interlocutor about oneself or someone/something that is notpresent. The exchange in (1) will illustrate the nature of thegriping/troubles-telling exchange:

(1) Two female graduate students in a course they both dislike.a: I sat through yesterday's class with total non-comprehen

sion!b: Oh, yesterday was the worst!

In the exchange in (1) the speaker signals to the addressee her feelings byusing an indirect complaint about a class they are both taking. By agreeing, the addressee demonstrates to the speaker a mutual sentiment. Onthis basis alone, an opening for further conversation and relationshipbuilding is provided in which they might go on to discover precisely theircommon areas of interests, purposes or sympathies.

The spontaneous speech data of this study consists of 533 troublestelling exchanges that were tape-recorded or recorded in the form of fieldnotes. The methodological basis of data analysis was that of the ethnography of speaking (Hymes 1962). Categories emerged from the analysis.Six categories of complaint responses emerged as major categories ofways in which people in this community respond to indirect complaints:1) Response or topic switch; 2) Questions; 3) Contradiction; 4) Joke/teasing; 5) Advice/lecture; and 6) Agreement/commiseration.

Almost half of the responses to the initiating complaint moves fell intoone of six possible types of responses, that which was termed "commiseration." The apparent preponderance of responses of this one type wascause to speculate that the underlying social strategy of much complainingbehavior in the community is not negative but positive in nature. In orderto check this suspicion it seemed necessary to conduct an informal interview with members of the speech community in order to tap their nativespeaker tacit knowledge on the functions of this speech act/event.

220 Diana Boxer

Table 1. Indirect complaint responses

Category % of corpus Comments

1. 0 or topic switch 10.190/0 rhetorical, deliberate topicswitch, repeatedbackchanneling

2. Question 11.700/0 request for elaboration;challenge question

3. Contradiction 14.720/0 intimates, status unequals,addressee wants distance

4. Joke/teasing 6.23% making light of situation

5. Advice/lecture 13.58 % platitudes, specific advice,moralizing

6. Commiseration 43.58% agreement, reassurance,exclamations, finishingspeaker's sentence

The interviews that are reported on here do indeed illustrate that complaint sequences can often work toward establishing solidarity when theyare part of the troubles-talk event. Of course, this fact is easily apparentfor other more positively evaluated speech acts (e.g., compliments, offers,invitations). However, the speech act with the semantic label "complaint"is not typically thought of as rapport-inspiring speech behavior. Gripingsequences are examined here precisely because, while it is not intuitivelyobvious, they do often fit into the category of speech behaviors that canlead to the establishment of relationships between interlocutors.

2. Ethnographic interviewing as corroborating data forspeech act/event analysis

Since the ethnographic interview is a method of getting people to talkabout what they know - of discovering what human behaviors mean tothe individuals participating in those behaviors - it differs greatly fromthe traditional interview or questionnaire in that it seeks to uncover notonly knowledge that is explicit but also knowledge that is tacit. The tacitknowledge that informants have about behavior is brought out only aftera rapport has been established between the researcher and the informant.


Because ideal informants in studies of speech acts/events are sociolinguistically naive, it is often possible to bring their tacit knowledge to astate of explicitness through gentle questioning by the researcher withinthe setting/context in which the speech behavior typically occurs. Thereverse is also true. Researchers themselves who are native speakers of thecommunity they are studying also possess knowledge that is tacit, and byinterviewing other native speakers, their own tacit knowledge can bemade explicit. Hence, by combining the researcher's own analysis ofspontaneous speech with information gleaned from native informantsthrough an ethnographic interview, a more complete analysis of thespecific speech behavior can be made than that which results from areliance on more traditional interviews or questionnaires.

The knowledge of how to go about conducting an ethnographicinterview cannot be gained simply by reading about how these are bestcarried out. Although there are numerous sources available that dealdirectly with ethnographic interviewing (e.g., Spradley 1979; Briggs1986), the best way to learn how to conduct such an interview is bydoing one. I include here a description of a first attempt at such aninterview and how it went awry. Although as researchers we rarely hearabout what goes wrong in the process of collecting data, I believe it isfruitful for methodological papers to detail such trials and errors. Thebenefit accrues to those who are interested in employing new tools suchas this type of interview, but who do not have firsthand experience withtheir use.1

The first attempt at an ethnographic interview in this study wasinitiated by careful thought about who the best informants would be,what questions would be most fruitful, and how long each interviewshould last. First, given that the best informants are those with whom theethnographer has built a rapport, ten people who were acquaintances ofthe researcher but who knew little about the research under investigationwere chosen as informants. After locating ten informants who met thesecriteria (five males and five females), interview times were scheduled withthe plan of asking eight structured questions of each informant. This wasfollowed by having the informants listen to and comment on a reading offourteen different sequences taken from the data. The intention was tofind out what the informants thought about various aspects of troublestelling as well as to get them to discuss what they saw as the social functions within the actual data sequences.

Asking the very same questions of each informant turned out to be amistake. Although there seemed in theory to be nothing wrong with

222 Diana Boxer

conducting the interview in a structured manner, adhering to a fixedagenda of questions led to an inflexibility that defeated the purpose of theinterview: To uncover tacit as well as explicit knowledge about troublestalk. The fixed agenda did not allow the informants to lead the researcherto unplanned questions based on their answers. A second mistake was intrying to achieve too much within a relatively brief time allocation: Fortyfive minutes to one hour. Attempting to have each informant listen tofourteen sequences of data as well as answer eight structured questionsimposed a heavy burden on the informants. One probable reason for thefeeling of imposition might have been precisely the structured nature ofthe interviews. Apparently, the fact that they were more like interviewsand less like conversations caused the informants to want to "get it overwith." Many of the informants, knowing that they would also have togive their impressions on spontaneous speech data, seemed to want toquickly finish with the structured questions. Indeed, several rushedthrough their responses, giving them less thought than they might havehad the demands of the interview been different. Another drawback ofthis first attempt at interviewing was the desire to get brief, conciseanswers to the structured questions so that they could be easily tabulated.Because of this, the interviewees were less likely to elaborate. As a result,their responses offered very little insight on the various issues broughtup.

Because there was no follow-up on issues brought up, the informants'short replies were accepted as satisfactory. This resulted in an inability toinfer much about what they really knew. While their explicit knowledgewas tapped, a point was never reached in being able to ascertain whatthey tacitly knew about indirect complaining. Spradley, in discussingethnography states: " ... a large part of any culture consists of tacit knowledge. We all know things that we cannot talk about or express in directways. The ethnographer must then make inferences about what peopleknow by listening carefully to what they say, by observing their behavior,and by studying artifacts and their use" (Spradley 1979: 9).

Given that the results of this interview were less than satisfactory, itwas apparent that a second interview would have to be conducted in amore informal, open-ended manner. Of course, this would mean findingten new informants, as those who participated in the first interview wereno longer naive. Before re-doing the interview some very good advicecame forth from an ethnographer who had just been through a similarprocess of trial and error. Micheau (personal communication) had thefollowing guidelines to offer: 1) Do as little talking as possible; 2) Pick


three or four areas to cover; introduce the issues and let the informantstalk; 3) Get the informants to do narratives and then chain theirnarratives together.

Micheau's advice was followed in planning for the second interview.The ten new informants were offered a short introduction on the purposeof the interview. The introduction was an invitation to the informants totalk about an instance/instances in which they were either the speaker oraddressee in a troubles telling exchange. The informants' tales were interspersed with the researcher's questions on some of the issues underinvestigation in this study. Thus the informants talked while their cueswere taken to ask gently probing questions at appropriate junctures intheir narratives. In such a manner the researcher is able to touch upon thevarious issues of concern. One of the drawbacks in such an approach isthe inability to ask the very same questions of each informant. In thiscase, not all ten informants addressed each issue. One of the advantages,on the other hand, is that each informant is able to become introspectivein response to the questions asked of them - questions that emergedirectly from their narratives. Moreover, informants were more engagedin their talk and hence more willing to carryon with their narratives.

Six key issues on which it seemed important to discover communitynorms had emerged from the analysis of spontaneous speech data: 1) Thefirst issue dealt with perceptions of the delineation between direct andindirect complaining. While a direct complaint is a face-threateningactivity (Brown - Levinson 1978), an indirect complaint does not holdthe addressee responsible for a perceived offense. Why is it, then, thatthese two speech behaviors share the same semantic label? It would seemthat their underlying social functions differ greatly. 2) Given that thereappear to be various options in responding to the indirect complaint(e.g., contradiction, topic switch, advice, commiseration), it seemedimportant to gather informants' perceptions on how these differingresponses function to bring about differing outcomes in the troublestelling speech event; 3) The analysis of the spontaneous speech data indicated vast differences in the way intimates respond to indirect complaintsas opposed to non-intimates. Therefore, it seemed important to gatherperceptions on how the social distance variable affects responses withinthe troubles-telling exchange; 4) Gender emerged as a very strongindicator of the propensity of addressees to respond with a reply that insome way enhanced solidarity between the interlocutors: women weretwice as likely to commiserate as men; men were twice as likely to giveadvice (this finding corroborates those of Tannen 1990). This difference

224 Diana Boxer

indicated a call to tap informants' perceptions on the gender variable introubles talk; 5) Related to the gender difference is perception on howindirect complaints are used to open and support interactions; 6) A sixthissue that emerged had to do with ethnicity and the propensity to useindirect complaints as an initiating move in troubles talk. This sixth issuewas not previously dealt with in the research, but which appeared, froman initial analysis of the speech data, to require attention. It was notimmediately apparent in the analysis of the data for this study that theethnic makeup of the 295 interlocutors was heavily Jewish. This ethnicityfactor gradually emerged throughout the data analysis and became aclearer variable only after the quantitative analysis of the study was completed. One of the disadvantages of conducting interviews after undertaking a quantitative analysis is precisely the possibility of missing animportant variable when looking at social distribution of any speech act.Thus while ethnicity was not taken as an independent variable in thequantitative analysis of the larger study (Boxer 1991; 1993a), it appearedto have important implications in complaining behavior. These implications were discussed in some detail by the informants in the course of ourconversations, serving to shed light on the ethnicity issue from the nativeinformant's viewpoint.

Each of these issues was discussed with most of the informants duringour open-ended ethnographic interview.· Two additional issues emergedfrom the informants' narratives, and these were added to the above six.They were: 7) Troubles talk in academia; and 8) The community's imageof troubles talk behavior. Table 2 indicates the various issues discussed,the number of informants who commented on them, and a summary oftheir comments. It also demonstrates how the results of such qualitativeinformation can be quantified.

The ten informants were either students, staff or faculty members at alarge university in the northeastern U. S. or alumni of this universityworking outside of the campus setting. These were individuals withwhom the researcher had some kind of acquaintance or friendship. Thegroup consisted of: 1) Two graduate students, one male and onefemale; 2) Three faculty members, all male; 3) Three staff members,all female; and 4) Two alumni, one male and one female. Because of thepreponderance of faculty, staff and alumni, the average age of theinformants is estimated to be late thirties to early forties. Informants wereinterviewed either in their offices on campus or were invited to theresearcher's home. A detailed examination of the results of the interviewsfollows.


Table 2. Informants' views on troubles-telling issues

Issues # of informants Number in Commentsquestioned agreement

1. D/IC distinction 9 9 All 9 said ICs aremore supportive2 out of 9 said ICsare less constructivein problem solving

2. ResponsesAgreeing 10 10 ICs seek agreement

o response 6 5 5 of 6 claim to ignorechronic complainers

Disagreeing 6 5 Less likely withstrangers, friends, andacquaintances

Advice 6 5 From men and amongintimates

3. Social Distance 7 7 All said they are moreag~eeablew/friendsand strangers thanintimates

4. Gender 8 8 Women complainmore and both sexesuse ICs more to otherwomen. Women aremore supportive

5. Ethnicity 10 9 9 strongly agreed thatJews complain more

6. ICs as openers 8 8 ICs used as openersw/strangers

7. ICs in academia 4 4 Academics like to talkand ICs playa part

8. Image of complaint 8 8 All agree on negativeimage, but oftenpositive reality

(IC: Indirect complaint; D: Direct complaint)

226 Diana Boxer

3. Perceptions of direct and indirect complaining

The informants were in general agreement that people within the community hesitate to participate in confrontational activities such as directcomplaining. Although most thought immediately of direct complainingwhen hearing the term 'complaint', the majority later indicated that griping, grumbling, or indirect complaining is a more commonly occurringactivity than direct complaining within the community. However, onlytwo of the nine informants who addressed this issue even thought of indirect complaints upon being presented with the term "complaint." Onefemale informant, a departmental administrative assistant stated: "Ithought of direct complaints only. But I guess they're much more rare.And in fact most people won't confront directly. I'm timid myself." Amale graduate student stated a similar idea: "One tends to think of thecomplaint department' or the complaint box.'"

In sum, nine informants addressed the issue of the distinction betweendirect and indirect complaints. Among these nine all agreed that indirectcomplaining is less confrontational as a verbal activity. The informantsindicated that indirect complaints offer the complainer a way to let offsteam, a means of leaking the complaint to the responsible party, andthe possibility of establishing a common bond with the recipient of theindirect complaint.

4. Perception of indirect complaint responses

4.1. Agreeing/commiserating as an indirect complaint response

The possibility of establishing a commonality through troubles-tellingexchanges is realized only if agreement is in some way obtained in theresponse to the initial indirect complaint. Indeed the seeking of agreementappears to be a widespread goal of indirect complaints. All ten informants mentioned either agreement as a goal of indirect complaintutterances or the fact that indirect complaint/ agreement exchanges typically serve to forge a common bond between speakers and addressees.Some of their thoughts on the subject are illustrated in examples (1)through (3).


(1) I find a lot of people complain to me about other peoplebecause they know I agree with them.

(2) I guess what causes the camaraderie among people IS theempathy in agreeing.

(3) There's a common bond. That we're both subjected to the sameobstacles and by dealing with it we share something.

Community members perceived the expression of shared sentiments to beone of the important elements of indirect complaint exchanges. Sharedfeelings forge a common bond between interlocutors that may have thepotential for fostering feelings of closeness - feelings that may eventuallylead to the establishment or the deepening of a friendship between thespeaker and addressee.

Immediately striking is the question of why it is that we employnegative evaluations to seek agreement or forge a common bond. Thenotion that complaints are used as strategically positive speech acts iscounterintuitive. Several informants touched on this issue. Example (4) isfrom one of them" a woman.

(4) If you walk over to someone and say "Gee, I'm so lucky" or"What a wonderful day", they'd think you're weird.

4.2. Responding to chronic complainers

One of the several possible types of responses to indirect complaints isoffering no response or changing the subject. Slightly more than 10 % ofthe responses in the corpus fell into this category. One situation in whichthis response type frequently occurred was in reply to chronic complainers.

Six of the informants brought up the subject of chronic complaining,and in so doing were able to shed some light on how they respond tochronic complainers, offering insight into the response type that is herereferred to as 0 response/topic switch. Of these six informants, fiveagreed that they tend to ignore or shun complaints from those they knowto overuse such speech behavior. Examples are given in (5) and (6).

(5) My mother was a constant complainer. I'd ignore it. My dadtunes it out because all she does is complain, but nobody listensto her.

228 Diana Boxer

(6) You let it roll off your back - ignore it. Which can lead to asituation of "crying wolf."

Although this question was not asked of all ten informants, five out of thesix who did speak about this subject agreed they would typically respondto a person whom they knew to be a chronic complainer with either noresponse or an attempt to change the subject. It may indeed be the casethat there is a critical level of complaints that is surpassed by thoseperceived as chronic complainers.

4.3. Advice as an indirect complaint response

The data from face-to face interactions in troubles talk indicated thatmen gave advice almost three times as often as did women; women commiserated approximately twice as much as did men. There was generalagreement among the informants that advice is largely a male response toindirect complaints. Whereas women are more likely to commiserate,men are more likely to give advice, especially in response to indirectcomplaints by female speakers. Of the six informants who specificallyaddressed this issue, five thought that they tend to give advice as a complaint response to intimates, particularly males to female intimates. Bothmale and female informants indicated that men give advice because theyhave been conditioned to think "more logically" and tend to want to"solve problems." While women want to provide emotional support,men want to get to the root of the problem and make it better.

5. The social distance variable

5.1. Disagreeing!contradicting as an indirect complaint response

The findings from the analysis of the spontaneous speech data indicatedthat contradiction is a rare response to a troubles talk opener amongpeople who are not well acquainted. While approximately 15 % of theresponses overall fell into this category, among strangers the incidence ofcontradiction responses was less than 10 % as compared with over 30 0/0among intimates. Six of the informants discussed this response type, and


of these six, five agreed that they would be most likely to contradictindirect complaints by intimates. Examples are given in (7) through (9).

(7) My husband and I disagree about everything. His typicalresponse is, "Oh, quit worrying.'" It's so unsupportive. It ticksme off.

(8) With your spouse or kids. I might contradict my wife on something like the salad dressing in a restaurant, but I'd probablyagree with the guy at the next table.

(9) With intimates you're more likely to disagree, with friends youwant to be liked. With intimates your relationship is moreestablished, you can vent more honestly.

To summarize the informants' thoughts on disagreeing/contradicting,there was general consensus that in this community the tendency is fordisagreement to occur as responses to indirect complaints from intimates.The reason for such a disparity seems evident, and has been put forth inWolfson's Bulge theory (1988). It has to do with the relative certainty ofour relationships with our intimates as compared with the uncertainty ofour relationships with more distant friends and acquaintances. With thislatter group we are more conscious of being inoffensive (see Boxer 1993 bfor further discussion of how the "Bulge" is skewed).

The social distance variable was discussed with seven of the teninformants. All seven agreed that they tend to behave differentlywith strangers than with close friends or intimates. An example is givenin (10).

(10) Some of this stuff goes on in transient conversation, in a store,with a salesperson, with a parking attendant. Particularly whenthe weather is bad. It's just a way of communicating and youmight start with something like an indirect complaint. I thinkpeople just want to be nice to each other - just want to talk toeach other.

Community members intuited that they behave differently with intimatesthan they do with friends, acquaintances and strangers. They claimed tobe more agreeable with the latter than with those people closest to them.One informant's statement summarizes: "On almost every level you'dspeak differently with people you know very well."

230 Diana Boxer

6. The gender variable

The eight community members questioned for this portion of the studyindicated that men and women behave very differently with respect toboth complaining and responding to complaints. There was total agreement among the eight informants that women commiserate much morethan men and thus tend to be more supportive of speakers' complaints.Examples are given in (11) through (13).

(11) I would tend to be less tolerant of a male that complained asmuch as my wife does. I'm much more likely to give advice thanto commiserate. (male)

(12) Women listen better in general and have more concern, thatkind of thing. I would say men are more stubborn and "knowit all" (not me, of course) being dominant, you know. I thinkwomen are more sensitive in general and therefore not asconcerned with being in control, like cut-throat. (male)

(13) Among women there are certain things we have in common,like raising children, which is not always joyous. They changeyour life and it's not all pleasant. (female)

Informants intuited that women participate more in indirect complainingthan men and individuals of both sexes complain more to other womenthan men. This is due to the increased likelihood of obtaining a satisfactory response, that is, one of agreement or commiseration. Women notonly hesitate to complain to men, but find more common ground withother women about which to complain. One reason for this is thatwomen are the principal caregivers of children, a common cause of complaint. This is true even for women of diverse ethnic and racial backgrounds. Women always have the children or husband issue in common,and these provide fertile subjects for indirect complaints. Whereas mentend to want to tell the complainer how to go about solving the problem,women just try to "be there" for the person complaining, providing thefeeling that they understand because they have had similar experiences.Thus, for women, indirect complaining provides one important way forfemale bonding to take place (Boxer 1993c).


7. Ethnicity

Recall that during the analysis of data for this project it became evidentthat a large majority of the interlocutors recorded during the data collection were Jewish. As such, ethnicity gradually emerged as a factoraffecting the way indirect complaints are used in the community. Thus itseemed clear that the ethnicity variable could not be ignored. Given thisethnic slant to the data, the question that arises is the following: Isindirect complaining/troubles-telling a result of the ethnicity of thesegment of the community studied? It seemed that the best way toascertain the effect of Jewish ethnicity on complaining behavior would beto tap into the perceptions of the members of the community who servedas informants for this study.

Informants of mixed ethnic/religious backgrounds were chosen to getto the heart of the interaction of Jewish ethnicity and complaining, justas a mixed group of male and female informants were selected in orderto more clearly understand how gender and indirect complaints interact.Some of their comments are in (14) through (16).Jewish informants:

(14) I think complaining is part of the Jewish culture, but it'spartially true also that Jews are generally taught to, if they findsomething wrong, to express that. Certain other ethnic groupsare more restrained.

(15) Maybe it has something to do with not getting a kinehura [anon-literal translation of this Yiddish term means 'let the evileye not be cast']. I remember my mother and father would neversay "things are wonderful." It's been passed on in our culture.My father in business would never tell anyone how well thingswere going. It's definitely a Jewish thing, this complaining.Particularly in the generation before. How much has beenpassed on, I don't know.

Non-Jewish informant:

(16) My girlfriend [who is Jewish] complains a lot ... There's thereligious perspective too. There's sort of a saintly attitude likeyou don't complain. I know tons of people like that in theChristian religion, particularly women who take all kinds of

232 Diana Boxer

abuse and turn the other cheek. They have very difficult lives,some of these women I teach, but they never complain.

There was 100 % agreement among the informants that Jews participateheavily in complaining in general, and indirect complaining/troublestelling in particular. Some insights were given as to why this may be true.It was stated by more than one informant that Jews are more vocal thansome other groups (e.g., White Anglo-Saxon Protestants) and tend toexpress their feelings and emotions more openly. One informant, whenspeaking about her friend of Italian origin, expressed the perception thatother ethnic groups may also be equally prone to complaining. Becauseof the relative paucity of data from that particular ethnic group in thecorpus, no tentative statements can be made. Perhaps a replication of thisstudy in an ethnic Italian neighborhood would bear out the validity ofthis intuition.

It may well be the case that indirect complaint/troubles tellingbehavior is not merely an ethnic/religious issue but a regional one as well.As the data for the present study is limited in regional scope, replicationof the study might be fruitful in different regions of the U. S. as well asother ethnic communities.

8. Indirect complaints as conversational openers

The subject of how indirect complaints are used to open the troublestelling event and thus to open and support conversations and interactionscame up either directly or indirectly with eight out of the ten informantsin response to several different questions and in various parts of ourconversations. All eight informants said that they often use indirect complaints to open conversations with strangers or little-acquainted addressees. Some of their comments, in (17) and (18) illustrate.

(17) You see those innocuous ones [indirect complaints] in theelevator all the time. It's a tremendous entree for moreconversation which IS a very safe entree to really expressfeelings somehow.

(18) If you're waiting for service, for example. Someone behind mecomplained and I sort of agreed, I was waiting too. She wasupset enough to say something to me and I agreed. It doeshappen that way ... It's sort of an opener.


Informants repeatedly brought up the issue of creating a common bondthrough the use of indirect complaints, and they noted that with strangersparticularly indirect complaints are frequently used to open and sustainconversations. With interlocutors of extreme social distance (i. e.,strangers, casual bystanders) it is not the serious complaints that are usedas openers but the small innocuous ones. These encompass such complaints as those about the weather, the bus or train being late, or waitingin lines. With little-acquainted interlocutors these small negative evaluations serve to find some point of common interest. The informants indicated that negative evaluations appear to be more commonly used asopeners than positive evaluations. It may be the case that positiveevaluations, such as compliments, threaten the negative face (Brown Levinson 1978) of addressees of extreme social distance more thannegative evaluations.

9. Troubles-telling and the academic community

There is indication that talk about troubles, or complaining, is aprevalent verbal activity within the specific speech community studiedhere. As stated earlier, indirect complaints were found to be ubiquitousin spontaneous social conversations among students, faculty and staffon campus. The ethnic and regional factors were discussed by someof the informants in the preceding sections as possible contributingfactors. In addition to these factors, however, several of the informantsindicated that academia is fertile ground for indirect complaints dueto its inherent intellectual atmosphere. Examples are given in (19)through (21).

(19) People at [this University] are great talkers and it goes alongwith the intellectual atmosphere, leading to talking more thanaction. There's an atmosphere of that.

(20) Students always complain about professors. Students bond bygiving each other the feeling that we're all in this together ... It'scooler and more fashionable to be cynical. It's more hip.

(21) I have to watch myself, I think academics do because we tendto be argumentative and at dinner parties where other peoplearen't that way you can get into arguments. Other people aren'tas fond of that.

234 Diana Boxer

Indirect complaining may in fact be a commonly occurring speechbehavior in many communities; however, in the particular communitythat is the focus of study here several factors combine to increase theirincidence of occurrence. The ethnic variable has been discussed as oneimportant contributing factor. The regional factor, that the University isa large, urban one in the northeastern part of the country, may in fact bea second contributing factor. Last, but perhaps not least in importance,the simple fact that it is an academic community may be of significancein the ubiquity of indirect complaints.

10. The negative image of complaining

The information gleaned from the ten informants offered strong indication that they perceive indirect complaints as a normal part of everydaycommunication. These community members indicated that indirect complaints serve to open and support conversations, that indirect complaintsare often used with the purpose of seeking agreement, and that inobtaining agreeable responses indirect complaints aid in forging acommon bond between speakers and addressees. These perceptions,however, run counter to the image community members have of "complaining." In order to grapple with this contradiction, eight of the informants were asked to give their views on their image of complaining.Examples are given in (22) through (25).

(22) Everybody complains. I can't imagine not being able to do that,how you feel inside. But when you say the word 'complain' itseems so negative.

(23) Oh [it's] negative, very negative. But what we're talking aboutisn't so negative. You're bonded with the other person whoagrees with you.

(24) ... It's not a good thing to do. But people do it all the time.Especially when you look at the small gripes... Mostcomplaints serve a different purpose from what you traditionally think of.


Can you think of a better term for it?

(25) "Complain" may be a little too strong for it. Maybe just like"venting. "

Of the eight informants directly questioned about the image that complaint conjures, all were in agreement that it definitely has a negativeimage. Notwithstanding this image, each had something to say on howwhat we had been discussing did not seem so negative. Part of theproblem seems to lie in the fact that the semantic label "complain" coversa broad spectrum of negative evaluation. The term "complaint" or"indirect complaint" may be somewhat misleading. Troubles-telling,troubles-talk, or troubles-sharing all refer to the larger speech event forwhich indirect complaints is typically the initiating speech act. Theseterms seem to be more adequately descriptive and less connotationallynegative than "indirect complaint."

What is so ubiquitous in the ordinary social conversation within thecommunity are not those serious complaints about things that adverselyaffects people's lives but the commonplace small gripes about things notworking right, the weather not being fair, or the professor not teachingwell. It is these innocuous, everyday indirect complaints that are sofrequently used to make small talk, that aid to establish common ground,and that may eventually evolve into getting to know another personbetter.

11. Conclusions

The preceding discussion has had a two-fold focus. First, a comparisonhas been drawn between two interviews undertaken for this study, onestructured and one open-ended in nature. The results of these have beencontrasted in order to illuminate the benefits of the ethnographic interview in uncovering both tacit and explicit knowledge on communitynorms regarding speech behavior. Second, the results of the ethnographicinterview were discussed in detail. The information gleaned frominformally interviewing ten members of the community yielded data thatserved to elucidate several issues emerging from the analysis of spontaneous indirect complaint data.

The issue of how to go about uncovering the norms of a speechcommunity is exceedingly important in studies of the analysis of speech

236 Diana Boxer

in face-to-face interaction. Interviewing native speakers can yield datathat either corroborates or disconfirms the researcher's own analysis ofspontaneous speech data. However, the manner in which informants areinterviewed bears significantly on the information they offer to the research. Informants should be relatively naive about the nature of thestudy so that they can offer a fresh perspective on the issues in question.Moreover, informants should be individuals with whom the researcherhas developed a rapport that enables them to speak openly and at lengthon the subject. Individuals who are comfortable doing much of thetalking make the best informants. Specific questions should emanatenaturally from the narratives of the informants rather than be preconceived by the researcher. Clearly, the researcher will have to decide on theissues of importance during his/her discussions with informants; however,cues about other issues typically come from what the informants have tosay. In such a way and only in such a way can tacit knowledge be uncovered.

Ethnographic interviewing techniques can sometimes be learned onlyby trial and error. The trials and errors reported in the above discussionenabled a comparison to be made between the results of two interviewsthat were very different in nature. By letting informants speak freely inthe second, informal interview and allowing the questions to emerge fromthe informants' ideas, information was gleaned that added greatly to theanalysis of troubles talk and its functions in the community.

Whereas most members of the community immediately perceived"complaint" as direct complaint, they indicated that indirect complainingis a more prevalent and positive activity. The informants were able toexpound upon their perceptions about indirect complaints by talkingabout how indirect complaints were employed in their own conversations. They indicated that indirect complaints are typically used as animportant part of "small talk" and with the underlying strategy ofobtaining agreement. With strangers and non-intimates, agreement orcommiseration is the preferred response, with the end of establishingsome kind of commonality, albeit brief, that makes encounters morepleasant.

Women participate more in troubles-talk than men and are therecipients of more indirect complaints, since they are perceived to bemore supportive in general than men. Men tend to give advice as anindirect complaint response to both male and female speakers.Informants generally agreed that it is a male tendency to want to takecontrol of such situations and solve the problem. Religion/ethnicity is


perceived to playa role in indirect complaint behavior, with some religious/ethnic groups participating in indirect complaint exchanges morethan others (e.g., Jews, Italians). Academia is also perceived to be fertileground for indirect complaint exchanges, as people involved in philosophical issues of all kinds are more likely to take part in the sort of debatethat is sometimes characteristic of indirect complaining.

The community norm is that complaining is a negative speechbehavior. Notwithstanding this explicit perception, the tacit knowledgeof the informants is that indirect complaints more often than not have apositive function in the everyday social conversation of members of thecommunity. Thus given certain requirements that are met by setting andinterlocutor characteristics, what is explicitly thought of as negative istacitly known to be a positive speech behavior.

These baseline findings on the underlying social strategies of troublestelling among native speakers of u.s. English have important implicationsfor applied linguists. Before we can go about teaching rules of speakingto language learners, we need to have information on what nativespeakers do with regard to specific speech acts and events. Research inapplied linguistics over the past twenty years has given us much information on speech act realization and pragmatic transfer in rules ofspeaking. What is suggested here is to go a step beyond the gathering ofdata through questionnaires and/or recorded spontaneous speech. It ispossible to add a deeper dimension to our insight about speech behaviorthrough corroborating evidence gleaned from ethnographic interviews ofmembers of the target language speech community.

If indirect complaints function to open and sustain troubles-tellingsequences, and if, as indicated here, such speech behavior has a positiveunderlying social strategy, this information can be extremely importantfor language learners whose rules for the realization of the speechact sequence differ from our own. As we have seen, the ethnographicinterview has an important role to play in the analysis of speechbehavior. Before we can apply findings on speech act/speech eventpatterning, we need to tap into the tacit knowledge of native speakersof the language. Only in this way can we hope to uncover the truefunctions of language forms as they are used among members of a speechcommunity.

238 Diana Boxer

Notes

1. Details of the initial trial at ethnographic interviews for this project werepresented at the 1994 Conference on Pragmatics and Language Learning atthe University of Illinois (parasession on research methodology). The audiencewelcomed remarks on what went wrong, indicating that such a descriptioncould be very helpful to researchers who plan to use this method in their futurework.

References

Atkinson, Jean - John Heritage (eds.)1984 Structures of social action: Studies in conversation analysis.

Cambridge: Cambridge University Press.Beebe, Leslie - Martha Clark Cummings

1985 Speech act performance: A function of the data collection procedure?Paper presented at the 18th Annual TESOL Convention, New York,April.

Boxer, Diana1991 A descriptive analysis of indirect complaint sequences among

speakers of American English. [Unpublished Ph. D. dissertation.,University of Pennsylvania].

1993 a Complaining and commiserating: A speech act view of solidarity inspoken American English. New York: Peter Lang Publishing.

1993 b "Social distance and speech behavior: The case of indirect complaints", Journal of Pragmatics 19: 103-125.

1993c "Complaining and commiserating: Exploring gender issues", Text13.3: 371-395

Briggs, Charles1986 Learning how to ask: A sociolinguistic appraisal of the role of the

interview in social science research. Cambridge: CambridgeUniversity Press.

Brown, Penelope - Stephen Levinson1978 "Universals in language usage: Politeness phenomena", in: Esther

Goody (ed.), 56-289.D'Amico-Reisner, Lynne

1985 An ethnolinguistic study of disapproval exchanges. [UnpublishedPh. D. dissertation, University of Pennsylvania].

Dascal Marcelo (ed.)1985 Dialogue in interdisciplinary approaches. Amsterdam: John Benjamins.

Ervin-Tripp, Susan1972 "On sociolinguistic rules: Alternation and co-occurrence", in: John

Gumperz - Dell Hymes (eds.), 213-250.Fine, Jonathan (ed.)

1988 Second language discourse: A textbook of current research.Norwood, NJ: Ablex.

Ethnographic Intervi~wing 239

Gladwin, Thomas - William C. Sturtevant (eds.)1962 Anthropology and human behavior. Washington, DC:

Anthropological Society of Washington.Goody, Esther (ed.)

1978 Questions and politeness. Cambridge: Cambridge University Press.Gumperz, John - Dell Hymes (eds.)

1972 Directions in sociolinguistics. New York: Holt, Reinhart andWinston.

Hatch, Evelyn1992 Discourse analysis and language education. Cambridge: Cambridge

University Press.Hymes, Dell

1962 The Ethnography of Speaking, in: Thomas Gladwin and William C.Sturtevant (eds.), 13-53.

Jefferson, Gail1984 "On stepwise transition from talk about a trouble to inappropritely

next-positioned matters", in: Jean Atkinson and John Heritage(eds.), 191-222.

Jefferson, Gail - John R. E. Lee1981 "The rejection of advice: Managing the problematic convergence of

a "troubles-telling" and a "service encounter", Journal ofPragmatics 5.5: 399-423.

Kachru, Braj1988 "Teaching world Englishes", ERIC/CLL Bulletin 12.1: 1-4.

Katriel, Tamar1985 "Griping as a verbal ritual in some Israeli discourse", in: Marcelo

Dascal (ed.), 367-381.Spradley, John

1979 The ethnographic interview. New York: Holt, Reinhart andWinston.

Tannen, Deborah1990 You just don~t understand: Women and men in conversation. New

York: William Morrow and Company.Thomas, Jenny

1983 "Cross-cultural pragmatic failure", Applied Linguistics 4.2: 91-109.Widdowson, Henry

1988 "Language, context and culture in the classroom", ERIC/CLLNews Bulletin 12.1: 6-7.

Wolfson, Nessa1981 "Invitations, compliments and the competence of the native

speaker", International Journal of Psycholinguistics 24.4: 7-22.1988 "The Bulge: A theory of speech behavior and social distance", in:

Jonathan Fine (ed.), 21-38.1989 Perspectives: Sociolinguistics and TESOL. Rowley, MA: Newbury

House.

Frolll the addressee's perspective: Illlposition infavor-asking

Myra Goldschmidt

1. Introduction

It has been well documented by now that conversation is more than staticdiscourse. Indeed, since Hymes published The Ethnography of Speaking(1962) there has been a burgeoning interest in the way verbal discourseis organized in particular speech communities (Gumperz 1971: 219-231;Hymes 1974: 47; Labov 1972: 120-121). More and more, empiricalresearch is uncovering and making available patterns of speech behaviorwithin target language communities. However, immersed within verbalrepertoire is not only the speech behavior of a speaker, but also theimpact this speech behavior has on an addressee in terms of action andreaction. In speech act research particularly, researchers are now lookingat the ensuing talk after the head act and finding that both the speaker'sperspective and the hearer's perspective (Blum-Kulka - Levenston 1987:156-158) are important with regard to gaining insight into the rules ofspeaking (Hymes 1972: 35-71).

It is the intent of this chapter to focus on a particular speech act, thatof favor-asking, and the impact this act has on an addressee in terms of aspecific reaction - imposition. This will be done by looking at the resultsof a survey given to two hundred speakers of American English betweenthe ages of 17 and 69, in order to determine, first of all, how imposedupon people feel when certain types of favors are asked, and second, whatthe pedagogical implications for non-native learners inherent in such ananalysis might be. The major reason for devising a survey on impositionwas to address the "person spoken to" who is as important as the speakerin this dynamic speech act. Very little research has been done from thisperspective, yet it is one that cannot be overlooked. As in the case of otherspeech acts, favor-asking involves negotiated segments of conversationwhich go beyond the philosophic tradition put forth by Austin (1962:150-163) and Searle (1969: 23-24) where talk was dealt with as a staticproduct, rather than a dynamic process.

242 Myra Goldschmidt

With regard to the objectives being sought in doing this survey, theywere threefold. The first objective was to establish an underlying valuesystem in terms of what causes people to feel "put upon" in a particulartarget community. The second objective was to see if the social variablesof role (in terms of student vs. non-student), gender, and age conditionimposition. These three variables were chosen based on the investigator'snative speaker intuition that any differences in the findings may beevidenced by the role, age, and gender of the subjects. Specifically, it wasthought that men and women would perceive imposition differentlybecause they often perceive favors, themselves, differently (Goldschmidt1993: 140). In addition, the age of the individuals, as well as whether ornot they were students, was intuited to condition perceived imposition.Finally, it was hoped that the results from the first two objectives of thedetermination of an underlying value system and the influence of socialvariables would help to clarify any implications for TESOL regardinghow imposed upon people may feel when asked certain types of favors.

Favor-asking denotes a speech act in which the motive or purposebehind the act itself is to get an addressee to do a specific task for thespeaker. In doing so, it encompasses different social role relations between individuals as well as varying intensities of volition. Specifically,favor-asking is characterized by four defining features:

1. Favor-asking is a speech act which involves asking for something"outside" of the addressee's daily routine.

2. Favor-asking entails doing activities that require some time and/oreffort on the part of the addressee or involves a good belonging to theaddressee.

3. Favor-asking entails no role-related obligation on the part of theaddressee to fulfill the task.

4. Favor-asking implies the notion of reciprocity in terms of a returnfavor

(Goldschmidt 1993: 157-169).Favor-asking also imposes (to some extent)1 on an interlocutor to ful

fill a future act which is potentially costly to the addressee, thus is facethreatening (Brown and Levinson 1987: 13-15) and calls for redressiveaction. This redressive action is often in the form of mitigation in orderto compensate for any imposition. Interestingly, Ervin-Tripp and Gordon(1986) show that sensitivity to degree of imposition or intrusiveness inrequesting is often learned by school age, to the extent that childrenbecome aware of what might be disruptive or difficult from the hearer's

Imposition in favor asking 243

point of view and vary their language accordingly. Furthermore, Brownand Levinson (1987: 74- 84) state that imposition is ranked in proportion to the expenditure of service (in the case of favor-asking), includingthe provision of time. However, they caution that the ranking is determined by a whole set of variables which can alter the degree of imposition, including whether actors have specific rights or obligations toperform the act, or whether they have specific reasons for not performingit, or whether the actors are known to enjoy being imposed upon in someway.

Thus, it can be said that favor-asking, like most speech acts, iscomprised of participants engaged in a particular speech situation as ameans to an end. It is also one in which Hymes' (1972: 35 - 71)mnemonic SPEAKING (S =Situation; P =Participants; E =Ends; A =Actsequence; K =Keys; I =Instrumentalities; N =Norms; and G =Genres)would condition every favor-asking situation. In other words, one's moodand physical circumstances at the time of asking could heavily impact onhow an addressee might react to the favor (Ervin-Tripp 1976: 50-63).

The data reported on in this chapter consider the extent to whichimposition is perceived depending on the gender, age and student/nonstudent status of individuals.

2. Method

A survey was given to two hundred subjects from the Philadelphia area(one hundred students at Villanova University between the ages of 17 and21 and one hundred "non-students" between the ages of 20 and 69), whorated the amount of imposition in five data-generated situations. (See the"Results" section for an account of the five situations).2

The selection of the participants in the survey questionnaire wasaccording to the criteria of age (adult speakers) and native language, andwas accomplished in the following ways. First, for the student participants, colleagues of the investigator in the English department atVillanova University were asked if they would administer the surveys totheir English classes during the 1991 spring and fall semesters. Theprofessors agreed to do so and obtained permission from the students toadminister the surveys. The completed surveys were collected andreturned to the investigator.


Selection of the non-student participants was carried out over a periodof eight months. The investigator distributed the surveys to people(friends, acquaintances and strangers) at various places in and around thePhiladelphia area. After a brief explanation about the survey, permissionwas secured and the questionnaires were completed (out of the presenceof the investigator), collected and analyzed.

The favor situations used in the survey were chosen from a corpus of200 favors. In deciding which tokens would be used, the investigatorpicked those examples from the data which seemed to represent favorstypically asked and which seemed to depict a range in terms of degree ofimposition. She then showed the data to another sociolinguist who alsochose those tokens which seemed within the norm of favors she has askedor has heard asked. The questions which were decided upon were thosewhich overlapped between the two linguists.

Each of the five favor situations required that the respondents rateeach favor according to imposition in terms of great imposition (GI),some imposition (51), little imposition (LI), and no imposition (NI). Theterm imposition, itself, was not defined but left to the discretion of eachinformant as to what it meant. Imposition literally means "the act ofputting a burden on"; however, the definition varies from person toperson in terms of what it constitutes. In fact, some folk definitionscollected by the investigator included, "When you ask someone to dosomething they're not ready to do. It can require time, m<:>ney, and/oreffort - though effort is open to interpretation"; "When I feel I'm beingtaken advantage of, or when it's over and above a 'normal' favor in'normal' circumstances, excluding an emergency, or if it were neverreciprocated"; and finally, "When I feel obligated to do something that Idon't really wish to do." In view of these differences of opinion, it wasdecided not to impose a definition, but to leave it open.

The questionnaires were analyzed according to the student and nonstudent populations, age and gender since it was thought that thesevariables could be determinants with regard to the responses. A statisticalanalysis for chi square and probability was undertaken for each of thefive situations (the alpha level was set at 0.05). After analyzing the resultswhich emerged from the data, reply profiles (or the proportion of repliesthat fell into the four given categories) were established in terms ofimposition - what types of favors seem to incite the perceived feeling ofgreat imposition, some imposition, little imposition and no imposition?From these profiles, tables were established to determine if there were anyrecurring patterns with regard to imposition in favor-asking. Following


are the results of the survey which provide insight from the perspective ofthe addressee in order to understand more fully the process of favorasking with a focus on imposition.

3. Results

The respondents to the questionnaire were divided into two groups, nonstudents (N =100) and students (N =100). The age distribution of thesetwo groups is given in Table 1. For analytical purposes the non-studentgroup was further divided into three groups: < 32 years (N =26), 32-45years (N =52), and> 45 years (N =22). Smaller ranges would have givensmaller numbers per group, making statistical evaluation difficult. Thestudent population was not further subdivided because of the very smallage range of 17- 21 years.

The gender of the two groups was also determined and is given inTable 2. When comparing the two groups there were significantly more(p > 0.001) female non-students compared to male non-students(X 2 =19.3; df =1).

A statistical analysis of the five situations for chi square andprobability (p) was then undertaken. The following are the results:

Table 1. Age distribution of survey respondents

Mean AgeStd. DeviationMedian AgeRange

Non-Student (N = 100)

39.4 years10.5 years39.0 years20-69 years

Student (N = 100)

18.4 years0.9 years18.0 years17-21 years

Table 2. Gender distribution of survey respondents

Non-student Student Total

Female 78% (N=78) 48 % (N = 48) 63 % (N = 126)Male 22% (N=22) 52% (N=52) 370/0 (N = 74)

Total N= 100 N= 100 N= 100


1. Favor situation: Husband to wife (who was home from work becauseof a snow storm) on his way to work: "If you have some time today, dome a favor and shovel the driveway" (90 feet long). (Table 3.1)

Table 3.1. Ql by student

NI LI SI GI Total

Student 10(10%)Non-student 17 (17%)

23 (23 %) 32 (32 %) 35 (35 %)17(17%) 38(38%) 28(28%)

100(50%)100(50%)

Total 27(13.5%) 40(20%) 70(35%) 63(31.5%) 200(100%)

NI = No imposition, LI = Little imposition, SI = Some imposition, GI = Greatimposition. Xl= 4.0; df =3; p =0.26

No significant differences were found between the responses of thestudents and non-students in this situation. (Table 3.1)

Table 3.2. Ql by gender

NI LI SI GI Total

Female 20 (15.9%) 24 (19.1 %) 46 (36.5 %) 36 (28.5%) 126 (63 %)Male 7 (9.5%) 16 (21.6%) 24 (32.4%) 24 (36.5 %) 74 (37%)

Total 27 (13.5%) 40 (20%) 70 (35%) 63 (31.5%) 200 (100%)

Xl= 2.72; df = 3; p = 0.43

No significant differences were found between the responses of femalesand males in this situation. (Table 3.2)

Table 3.3. Ql by age

NI LI SI GI Total

Students 10 (10%) 23 (23 %) 32 (32 %) 35 (35%) 100(50%)< 32 yrs 7 (26.9%) 4 (15.4%) 7 (26.9%) 8 (30.8 %) 26 (13%)32-45 yrs 6 (11.5 %) 12 (23.1 %) 20 (38.5%) 14 (26.9%) 52 (26%)> 45 yrs 4 (18.2 %) 1 (4.6%) 11 (50%) 6 (27.3 %) 22 (11 %)

Total 27 (13.5%) 40 (20%) 70 (35%) 63 (31.5 %) 200 (100%)

Xl= 11.68; df =9; p =0.23


No significant differences were found between the students and the threenon-student age groups in this situation (Table 3.3).

Both the non-student and student populations considered this situation to be imposing [SI + GI > NI + LI]. Similarly, females, males, and allage groups considered this to be an imposing situation.

2. Favor situation: One person to a good friend who is not working outside the home: "Could you please watch my children (3) tomorrow for afew hours? I have to go to work, and they'd rather be with your kids thana babysitter."

Table 4.1. Q2 by student

N1 L1 SI G1 Total

Student 13 (13 %) 37 (370/0) 43 (43 %) 7(7%) 100(50%)Non-Student 14 (140/0) 18 (18 %) 42 (42%) 26 (260/0) 100(50%)

Total 27 (13.5%) 55 (27.5%) 85 (42.50/0) 33 (16.5%) 200 (100%)

X 2= 17.55; df = 3; p = 0.001

There is a considerable statistically significant difference between thetwo groups. The non-students found this to be a more imposing situationthan did the students (GI = 26 % vs. 70/0). (Table 4.1)

Table 4.2. Q2 by gender

N1 L1 SI G1 Total

Female 21 (16.7%) 33 (26.2%) 47 (37.3%) 25 (19.8%) 126 (630/0)Male 6 (8.1%) 22 (29.7%) 38 (51.40/0) 8 (10.8%) 74 (370/0)

Total 27 (13.5%) 55 (27.50/0) 85 (42.5%) 33 (16.5%) 200 (100%)

X2= 7.21; df = 3; p = 0.065

No significant differences were found between the females and malesin this situation. Thus, gender was not a significant variable in this situation. (Table 4.2)


Table 4.3. Q2 by age

NI LI SI GI Total

Students 13 (13 %) 37 (370/0) 43 (43 %) 7(7%) 100(50%)< 32 yrs 4 (15.30/0) 6 (23.1 0/0) 10 (38.5 0/0) 6 (23.1 0/0) 26 (130/0)32-45 yrs 7 (13.5%) 7 (13.5%) 25 (28 %) 13 (25%) 52 (26%)> 45 yrs 3 (13.7%) 5 (22.7%) 7 (31.8 %) 7 (31.8%) 22 (11 %)

Total 27 (13.5%) 55 (27.5%) 85 (42.5%) 33 (16.5 %) 200 (100%)

X 2= 20.09; df = 9; p = 0.017

There is a statistically significant difference among the various agegroups (Table 4.3). The non-students considered this to be a more imposing situation compared to the students. In order to determine if anyoneof the non-student age groups considered this a more imposing situation,these three groups [< 32 years, 32 - 45 years, and> 45 years] were analyzed separately for chi square and p values (Xl= 2.76; df =6; p = 0.84).After further analysis, the age of the non-student respondent is not of significance in this situation.

Both the student and non-student groups considered this an imposingsituation (SI + GI > NI + LI). Similarly, both females and males and all theage groups considered this an imposing situation.

3. Favor situation: One neighbor asks another neighbor who is a goodfriend: "If I can't find anyone else, do you think I can count on you totake care of our animals while we are away? It would just be Saturdayafternoon and evening, and Sunday morning for you to feed them and letthem out for a few minutes."


NI LI SI GI Total

Student 21 (21 %) 43 (43 %) 25 (25 %) 11 (11 %) 100(50%)Non-Student 33 (33 %) 26 (26%) 28 (28 %) 13 (13 %) 100(50%)

Total 54 (27%) 69 (34.5%) 53 (26.5%) 24 (12 %) 200 (100%)

X2= 7.19; df = 3; p = 0.07


No statistically significant differences were found between theresponses of the students and non-students in this situation. (Table 5.1)


NI LI SI GI Total

Female 45 (35.7 0/0) 38 (30.2 0/0) 31 (24.6%) 12 (9.5 0/0) 126 (63%)Male 9 (12.2 %) 31 (41.9%) 22 (29.70/0) 12 (16.2 %) 74 (370/0)

Total 54 (27%) 69 (34.5 0/0) 53 (26.5 0/0) 24 (120/0) 200 (100%)

X 2= 13.64; dt = 3; p = 0.003

Males found this situation to incur somewhat more imposition thandid the females. This might suggest that males are asked to do this typeof task less often than are females, thus find it to be more imposing. Also,males may perceive this type of task to require an inordinate amount oftime and effort, which again, may contribute to their belief that thisparticular task constitutes more imposition. (Table 5.2)


NI LI SI GI Total

Students 21 (10.5 0/0) 43 (21.5 %) 25 (12.5 % ) 11 (6.50/0) 100 (50%)< 32 yrs 13(50%) 6 (23.1 0/0) 6 (23.1 %) 1 (3.8%) 26 (13 0/0)32-45 yrs 14 (26.9 0/0) 15 (28.9%) 14 (26.9 0/0) 9 (17.30/0) 52 (26%)>45 yrs 6 (27.3%) 5 (22.70/0) 8 (36.4%) 3 (13.6%) 22 (11 0/0)

Total 54 (27%) 69 (34.5 0/0) 53 (26.5%) 24 (120/0) 200 (100%)

X2= 9.75; dt = 9; p = 0.37

No significant differences were found between the students and thethree non-student age groups in this situation (Table 5.3). Overall, thestudent and non-student groups considered this an unimposing situation(NI+LI>SI+GI).

Question 3 also may be considered to be a "fuzzy" situation in thatthe data showed no strong reaction in anyone of the four types of


imposition, though the trend is certainly towards little and no imposition(61.5 %) as opposed to some and great imposition (38.5 %).

What makes question 3 so interesting, however, is not necessarily thenumbers, but rather the situation itself (caring for animals) comparedwith question 2 (caring for children). Twice as many people thought thatlooking after animals incurred no imposition (27 %) when comparedwith looking after children (13.5 %). This is particularly noticeable in thenon-student group, 33 % of whom said looking after animals incurred noimposition while only 14 % said looking after children incurred noimposition. Similarly, 21 % of the student group believed that lookingafter animals incurred no imposition, but only 13 % said that lookingafter children incurred no imposition. These percentages suggestthat people appear to be more willing to look after animals thanchildren. Perhaps most people have had personal experience with petsrather than children to account for these findings. Or perhaps, morepeople believe that there is less time and effort involved with looking afterpets.

Furthermore, three times as many females than males found this situation to incur no imposition. Again, perhaps this is a situation that morewomen do for others with regard to favor-asking. In terms of age, theprimary group to consider this situation as causing no imposition was thenon-student group less than 32 years of age, especially when comparedwith their looking after children (50 % and 15 % respectively).

4. Favor situation: An acquaintance asks the following of someone shehasn't seen for several months, and who just got out of the hospitalbecause of leg surgery. It is 3pm on Saturday, Christmas Eve: "I called toask you a favor. We're being kicked out of our house tonight becauseDavid (the eldest of 3 sons) wants to cook dinner for his girlfriend and


NI LI 51 GI Total

Student 1 (1 %) 5 (5%) 14 (14%) 80(80%) 100(50%)Non-Student 1 (1 0/0) 2 (20/0) 7 (70/0) 90 (900/0) 100 (500/0)

Total 2 (1 %) 7 (3.5%) 21 (10.5 %) 170 (85%) 200 (100%)

Xl= 4.21; df =3; p =0.24


we need some place to go. Can we (4 people) come to your house for theevening and for dinner? We can pick up something like a pizza."

There are several groups in this analysis [Student/No imposition andNon-student/No imposition/Little imposition] that have small numbers«5). This may invalidate the chi square test. The two groups NI and LIfor both students and non-students were combined to increase the groupsize, and the chi square and p values were recalculated (X 2 = 3.92; df = 2;p = 0.41).

No significant differences were found between the responses of thestudents and non-students in this situation. (Table 6.1)


NI LI 51 GI Total

Female 0(0%) 4 (3.2%) 9 (7.1%) 113 (89.7%) 126 (63%)Male 2 (2.7%) 3 (4.1%) 12 (16.2%) 57 (77%) 74 (37%)

Total 2 (1%) 7 (3.5%) 21 (10.50/0) 170(85%) 200 (100%)

X 2= 8.04; df = 3; p = 0.05

As stated above, the chi square and p values were recalculated becauseof small numbers (X 2 = 5.86; df = 2; p = 0.05).

Gender was found to be a statistically significant factor in the way therespondents viewed this situation. Females found this situation to incurgreat imposition more than did the males. This might suggest thatfemales, especially, consider "private family time" to be extremelyimportant and something that is not to be imposed upon. (Table 6.2)


NI LI SI GI Total

Students 1 (0.5%) 5 (2.5%) 14 (7%) 80 (40%) 100 (500/0)< 32 yrs 1 (3.9 0/0) 0 3 (11.5 %) 22 (84.6%) 26 (13%)32-45 yrs 0 2 (3.9%) 4 (7.7%) 46 (88.5%) 52 (260/0)>45 yrs 0 0 0 22 (100%) 22 (11 %)

Total 2 (1%) 7 (3.5%) 21 (10.50/0) 170 (85%) 200 (100%)

X2= 10.03; df = 9; P = 0.35


As stated above the chi square and p values were recalculated becauseof small numbers (X 2 =8.41; df =6; p =0.38).

Although there is no statistically significant difference between thestudents and the three non-student age groups, there is a trend evident(Table 6.3). With increasing age, the degree of imposition increases; allrespondents over 45 years of age considered this situation to incur greatimposition. All groups considered this a very imposing situation withmore than 85 % of the respondents deeming this a great imposition.

5. Favor situation: One graduate student asks another graduate studentknown only by sight: "I was wondering. Well, 1have these questionnairesfor people doing a study of students here, and 1 was wondering if you'dmind filling one out?"


NI LI SI GI Total

Student 41 (41 0/0) 41 (41 0/0) 12 (12 %) 6 (60/0) 100 (500/0)Non-Student 48 (48 %) 32 (32 %) 16(16%) 4 (4%) 100(50%)

Total 89 (44.5%) 73 (36.5%) 28 (14%) 10 (5%) 200 (100%)

X 2= 2.63; df = 3; p = 0.45

No significant differences were found between the responses of thestudents and the non-students in this situation. (Table 7.1)


NI LI SI GI Total

Female 64 (50.8 %) 38 (30.2 %) 18 (14.2 %) 6 (4.8%) 126 (63%)Male 25 (33.8 %) 35 (47.3 %) 10 (13.5 %) 4 (5.4%) 74 (37%)

Total 89 (44.5%) 73 (36.5%) 28 (14%) 10 (5%) 200 (100%)

X2= 6.84; df = 3; p = 0.08


No significant differences were found between the responses offemales and males in this situation. (Table 7.2)


NI LI 51 GI Total

Students 41 (41 %) 41 (41 %) 12 (12 %) 6 (60/0) 100(50%)< 32 yrs 13 (500/0) 5 (19.2 %) 5 (19.20/0) 3 (11.6%) 26 (13%)32-45 yrs 24 (46.2 %) 18 (34.6 0/0) 9(17.3%) 1 (1.9%) 52 (26%)>45 yrs 11 (50%) 9 (40.9%) 2 (9.1 %) 0 22 (11 %)

Total 89 (44.5%) 73 (36.5 0/0) 28 (14%) 10 (5%) 200 (100%)

Xl= 9.57; df = 9; p = 0.39

No significant differences were found between the students and thethree non-student age groups in this situation (Table 7.3).

Both the student and non-student groups considered this an unimposing situation (NI + LI > SI + GI). Similarly, both females and malesconsidered this an unimposing situation.

Question 5 showed a strong trend towards 'little' and 'no' imposition(81 0/0). This is a situation which appears to entail little time and/or effort.It is also one that students do for other students.

4. Discussion and implications of the results

From the results of the survey, it can be said that the three objectivessought were, in fact, achieved. First of all, with regard to an underlyingvalue system in the targeted speech community, people, in general, feelimposed upon in the following favor-asking situations:

1) In situations where family privacy (or special times relating to thefamily) is intruded upon.

2) In situations involving a great deal of time and/or effort.

In the first type of situation, most of the subjects vehemently felt imposed upon and objected to the situation concerning Christmas Eve sincethis is a special "family" time for most people. Not only were the majo-


rity of great impositions given for this situation, but also a large numberof comments were elicited as well. These ranged from, "This is outrageous!" to "Who needs friends like this?" Clearly, the most indignationand annoyance surfaced in this situation.

In the second type of situation, people generally felt imposed uponwhenever they were asked to do something that required a great deal oftime and/or effort. Though favors, in general, require varying degrees oftime and/or effort, it seems that the greater the amount of perceived timeor effort involved in a favor, the greater is the amount of perceivedimposition. Certainly, time is one of the most precious commoditiespeople have today, and they don't like to spend too much of it doingfavors for others. Likewise, any situation which infringes on a person'stime, also entails some effort on that person's part. Therefore, thequestions about shoveling a 90 foot driveway and taking care of childrengarnered a number of great and some imposition responses.

With regard to the results of the survey for the second objective, itbecame very clear that the investigator's preconceived hypothesesconcerning the social variables of age and gender and role as conditioningfactors in imposition were not fully substantiated. Not only were theoverall results between the males and females similar with respect toimposition intensity, but also the results between the students and thenon-students, as well as the age of the individuals proved to be, for themost part, comparable. What this points to is that there exist unwritten,but understood "parameters" within speech communities concerningwhat is and what is not thought of as imposing. These parameters seemto transcend all other variables and are important to learn as a newmember of a particular speech community.

Finally, the results confirm implications for TESOL. Clearly, speakersmust be cognizant of the rights and obligations they have towards otherswhich include the notion of imposition. In other words, learners shouldrecognize that since favor-asking requires some time and/or effort on thepart of an addressee, the speaker, by asking the favor, is impinging on therights of the addressee. Also, learners of a language need to know whatconstitutes imposition, how people in a particular culture define the levelor degree of imposition, and when and how it is acceptable to imposeupon someone. These are very important to communicate to learnersbecause if favor-asking is defined as a speech act where people are notobligated by their role or status to do the act, then degree of impositionor what can be asked or expected of an addressee is critical for learnersto know about.


This type of information is important for learners to be trulycompetent members of the target language community. Interestingly, nonnative learners often know parts of social routines, but fail in the overalldelivery of them which may result in an undesirable impact on an addressee. Example (1) illustrates an utterance by a non-native learner to ateacher he had had three years earlier.

(1) NNSNS

==> NNS

I have a favor to ask you.Sure, what can I do for you?You need to write a recommendation for me.

This student's asking of the actual favor was inappropriate and couldhave caused a breakdown in communication with a teacher (or otheraddressee) who was not aware of the student's lack of understanding ofthis type of speech behavior. A typical reaction and/or response mighthave been, "I don't need to do anything for you," which would havecaused ill-feelings on the part of both the student and the teacher.

By focusing on both the speaker and the addressee in certain conversational situations, teachers do a great service to new learners of alanguage. Everyone needs help at times, and the best way of ensuring thatthis help is attended to is by asking favors in a way which presents thespeaker in a good light. Since deviation from social norms can be interpreted as impertinent, sarcastic, or rude, learners need to be taught thesocial norms for a particular culture in order to comprehend both how toask a favor on the one hand, and what the listener expects to hear, on theother. In other words, learners need to be sensitive, not only to thespeaker's situation, but to the addressee's situation as well.

Notes

1. Perceived imposition varies from person to person. Whereas one person mayperceive a favor as causing great imposition, another person may perceive thesame favor as causing little or no imposition. Each favor-asking situation isconditioned by several variables which could affect the overall amount ofperceived imposition.

2. The original survey had twelve situations; however, for the purpose of thischapter, a representative sampling was used.


References

Austin, John L.1962 How to do things with words. Oxford: Clarendon Press.

Blum-Kulka, Shoshana - Edward Levenston1987 "Lexical-grammatical pragmatic indicators", Studies in Second

Language Acquisition 9: 155-170.Brown, Penelope - Stephen Levinson

1978 Politeness: Some universals in language usage. Cambridge:Cambridge University.

Ervin-Tripp, Susan1976 "Is Sybil there? The structure of American English directives",

Language in Society 5.1: 25-66.Ervin-Tripp, Susan - David Gordon

1986 "The development of requests", in: R.L. Schiefelbusch (ed.), 61-95.Fishman, Joshua A. (ed.)

1968 Readings in the sociology of language. The Hague: Mouton.Gladwin, Thomas - William C. Sturtevant (eds.)

1962 Anthropology and human behavior. Washington, DC:Anthropological Society of Washington.

Goldschmidt, Myra1993 For the favor of asking: A sociolinguistic analysis. [Unpublished

Ph. D. dissertation, University of Pennsylvania.]Gumperz, John

1971 Language in social groups. Palo Alto, CA: Stanford University Press.Gumperz, John - Dell Hymes (eds.)

1972 Directions in sociolinguistics. New York: Holt, Rinehart andWinston.

Hymes, Dell1962 "The ethnography of speaking", in: Thomas Gladwin and William

C. Sturtevant (eds.), 15-53.[1968] [Reprinted in: Joshua A. Fishman (ed.), Readings in the sociology of

language. The Hague: Mouton, 1968, 99-138.]1972 "Models of the interactions of language and social life", in: John

Gumperz - Dell Hymes (eds.), 35-71.Labov, William

1972 Sociolinguistic patterns. Philadelphia: University of PennsylvaniaPress.

Schiefelbusch, Richard L. (ed.)1986 Language competence: Assessment and intervention. San Diego, CA:

College-Hill.Searle, John R.

1969 Speech acts. Cambridge: Cambridge University Press.

Transfer of pragmatic competence and suggestionsin Spanish foreign language learning

Dale April Koike

1. Introduction

The transfer of knowledge of a first language in the learning of a foreignlanguage is a common process. This knowledge ~an reflect aspects of anycomponent of language, including syntax, phonology, morphology,semantics, and pragmatics. The basic question that is addressed in thisstudy is the following: Do foreign language learners use strategies oftransfer to help them comprehend information? The context of learningexamined here is that of the aural comprehension of the speech act of suggestions.

In a previous study, I discuss the notion of a "pragmatic competence"in a learner's interlanguage, involving knowledge and use of rules ofappropriateness and politeness that dictate the way the learner understands and formulates utterances, such as requests (see Koike1989: 279-281). Data obtained from several experiments suggest thatfirst language pragmatic knowledge transferred to the foreign languagespeech act situation causes learners to attempt to produce utterances thatthey believe would be pragmatically appropriate to the context. Whenfaced with producing a speech act that is more difficult than students feelcompetent in formulating, however, many choose to employ a less appropriate but syntactically simpler form, showing a change of strategy inactual production.

This pragmatic competence is part of interlanguage, a system thatrepresents dynamic stages in the learning process and that are subject tocontinual change and modification (see Selinker 1972: 229-230; Corder1975: 410-411). One may ask whether changes in the way knowledgefrom the first language is transferred to the foreign language can beobserved at different stages of learning. The present study examines thisquestion through data from native English speakers who are learners ofSpanish to see (1) whether adult foreign language learners of variouslevels of proficiency can recognize the speech act of suggestions in the

258 Dale April Koike

foreign language; and (2) to what extent first language pragmaticinformation is transferred in the comprehension of a foreign languagespeech act, and what variables can affect this transfer.

The data used in this investigation come from responses to a questionnaire by adult U. S. native speakers of English who are Spanish learnersat varying levels of study at the University of Texas at Austin. On thebasis of the results of this questionnaire, I suggest that students' levels ofproficiency affect the way in which they transfer first language speech actknowledge to understand a complex foreign language speech act. Whenlearners are advanced enough to begin to analyze the components ofcomplex speech act utterances, they can sometimes misunderstand thespeaker's intent, especially if the speech act utterance in question hasdifferent connotations in the two languages. In other words, learners firstneed to be proficient enough to understand the overall pragmatic intentof complex utterances. When such is the case and they can also payattention to individual elements in the utterance, they may transferknowledge from the first language concerning these elements at thatpoint. In the following sections, I will discuss transfer in foreignlanguage learning, the speech act of suggestions in both Spanish andEnglish, the questions to be addressed, the experiment, and the resultsand conclusions.

2. Transfer in foreign language learning

Many studies have addressed the use of transfer as a strategy in foreignlanguage learning, examining evidence of the phenomenon largely in oralproduction (see especially Gass - Selinker 1983; 1992). All agree thattransfer from the first language does occur in foreign language learning,under various constraints. Corder (1983: 95 and 1992: 29) states that thefirst language acts as a tool in the discovery of the formal properties ofthe foreign language, facilitating especially in the learning of thosefeatures that resemble those of the first language. Corder, qS well asKellerman (1977; 1983), stress that the greater the similarity perceived bythe learner' between the first language and the foreign language, thegreater the likelihood that transfer will occur in foreign language acquisition. Kellerman (1983: 117) claims a first language structure will betreated as language-specific and not transferable to a foreign language, orlanguage-neutral and thus transferable. Blum-Kulka, House, and Kasper

Transfer of pragmatic competence 259

(1989: 26-27) found that learners avoid transfer of language-specificstructures, revealing an awareness of transferability of pragmatic constraints.

Schachter (1983: 104; 1992:38-39) outlines the possible outcomes oflearner hypotheses about the target language based on first languageinformation. One possibility she describes is that the learner may choosethe correct domain or abstract category within the target language (myexample, reflexive pronouns) but the wrong hypothesis about the input,either because of a somewhat mistaken analysis of the input or becausethe learner correctly equates the domains of the two languages butassumes a hypothesis that is appropriate for the first language and not forthe foreign language. In this case, a transfer error is produced. The otherpossibilities are, of course, that the transfer will cause no problem, or willbe understood in spite of the application of the hypothesis to the wrongdomain.

Since they were looking for evidence of transfer, all of the studies citedabove were focused on foreign language or second language production,and not on a receptive skill such as listening comprehension. One cannotdoubt that transfer, a cognitive process, also occurs in listening comprehension and reading activities. A study by Gass (1989: 196) examinedthe extent of transfer in the reading interpretation of simple sentences.One of her conclusions was that second language learners seek an initialhypothesis regarding the second language on the basis of their firstlanguage. When there is an incongruity between the two languages, however, learners may resort to fundamental universal properties of language,such as that of canonical word order. The question arises whether theseclaims and findings about foreign language transfer also apply ~n thesame way to the listening comprehension skill. If they do, then one mustask what the consequences are for the foreign language learner who, asSchachter describes, makes a wrong hypothesis about the input in acorrect domain in attempting to understand the foreign language.

I am assuming, as does Gass (1989: 183), that syntax, semantics, andpragmatics simultaneously interact in the way a learner interprets input,along with other features of the grammar. As Gass points out, there aremany linguistic elements available to cue a learner to interpret input, andbecause of factors such as frequency, ease of interpretation, and information value, not all of the elements are equally used during on-line processing. She adds that if there are cross-linguistic differences in cue usage,then a learner must know not only the appropriate cues of the targetlanguage (e. g., word order) but also the strengths of those cues. In


learning a foreign language, a learner may learn the new cues or rely onknowledge of the first language.

3. Spanish and English suggestions

Suggestions are speech acts that are made presumably in the best interestof the listener, usually to help the listener toward some goal that the latterdesires or is assumed to desire. Suggestions are sometimes made in thebest interest of the speaker as well. As a type of directive, or a way to getsomeone to do something (Searle 1979: 13), suggestions require a futureeffort by the listener (Blum-Kulka - House - Kasper 1989: 12), and callfor strategies to mitigate their force. They represent a complex speech actwhose intent is sometimes misunderstood even by native speakers.

Many forms of suggestions in Spanish and English are alike and easilytransfer between the two languages, such as in (1) through (3).

(1) Sugiero que leas este libro.'I suggest that you read this book'

(2) 2No puedes leer este?'Can't you read this one?'

(3) 2Por que no lees este?'Why don't you read this one?'

There are differences, however, in the formulation of some suggestions inSpanish as opposed to English, particularly in the use of negatives ininterrogative suggestions, for example as in (4) through (9).

(4) Have you thought about reading this book?

(5) 2No has pensado en leer este libro?'Haven't you thought about reading this book?'l

(6) #2Has pensado en leer este libro?'Have you thought about reading this book?'

(7) Should you read this book?


(8) 2No deberias/tienes que leer este libro?'Shouldn't you/don't you have to read this book?'2

(9) #2 Deberias/tienes que leer este libro?'Should you/do you have to read this book?'

To convey the illocutionary force of a suggestion in Spanish, the Spanishequivalents of (4) and (7) must be expressed negatively, as in (5) and (8).The utterances in (6) and (9) convey purely yes-no confirmationquestions. To an English speaker, however, the English translations of (5)and (8) convey a much stronger force than their Spanish counterparts; theforce can come across as almost a reproach to the listener.

Aside from the question of transfer of speech act knowledge, there aredifferences in the grammatical formulation of suggestions in English asopposed to Spanish that can confound the comprehension process. Suchis the use of the negative. I base my ideas on data drawn from two studiesof speech acts in Spanish, specifically, of directives (Hobbs 1990; Koike1994). In the two studies, a total of 83 native Spanish speakers fromMexico were asked what they would say in a potentially face-threateningsituation in which they had to ask a child, a peer, or a non-intimate,

DeclarativeDeberias 'You should/ought to'Puedes/Podrias 'You can/could'No estaria mal si 'It wouldn't hurt if you'Mejor/Seria mejor (mas rapido, buena idea, etc.) si 'It would be better(faster, a good idea, etc.) if'Sugiero/creo (que) 'I suggest/think (that)'Si fuera Ud. 'If I were you' (I would)Vamos a 'Let's'

Interrogativei,Por que no 'Why don't youlWhy not'i,No deberias 'Shouldn't/Should you'i, Que tal/Que te parece si 'How about/What do you think (if)'i,No has pensado en 'Ever think about/Have you thought about/considered'i,No puedes 'Can/Can't you'

ImperativeTrata de 'Try to'

Figure 1. Commonly used formulaic expressions in Spanish suggestions


higher-ranking stranger to move from a chair reserved for someone else.Figure 1 displays the suggestion forms used by the informants in the twostudies, as divided into three types according to syntactic and lexicalconsiderations. The types are declarative, interrogative, and imperative.Of the declarative type of suggestions, seven forms were used, three ofwhich include a si- 'if' clause, conveying an irrealis condition. Five interrogative forms were employed, four of which are negated, and one includes a si- clause.

One must remember that a suggestion is a kind of directive in whichthe speaker normally wants to minimize to the greatest extent the possibility that the listener will be offended at the suggestion. The data showthat Spanish usually requires the negative when formulating interrogativesuggestions. English, on the other hand, allows both negated and nonnegated forms in conveying a suggestion intent, as in examples (10) and(11) spoken in response to someone who said she was running out of timeto turn in an article.

(10) Have you thought about asking for an extension?

(11) Haven:lt you thought about asking for an extension?

Example (10) is ambiguous in that it can have three implications: (a)The speaker is asking a simple information question to find outif the listener has thought about X; (b) The speaker is asking to findout the listener's reaction to X; or (c) As a suggestion, the speaker isletting the listener know s/he (the speaker) believes X should be doneby asking about it. Example (11), however, implies that the speakerbelieves the listener should think about X, assumes that the listenerhas not thought about it, and questions that assumption. Thus, theforce of the utterance for English speakers is rather strong, since thespeaker is not allowing the listener "freedom of action unhindered," asBrown and Levinson (1987: 129-130) describe negative politeness(for further discussion, see Koike 1992: 21-31).3 Use of the negativein English suggestions is not common, especially in interrogativesuggestions. The negative sentence can sound like an insult or a reproach,as if the listener overlooked the option mentioned by the speaker. Insum, the differences between Spanish and English interrogativesuggestions lead to the possibility of problems of miscommunicationand misunderstanding between the English and Spanish languages andcultures.


4. The experiment

4.1. Questions and hypotheses

The purpose of this study is to examine the comprehension and reactionsof English-speaking students of Spanish at different levels of foreignlanguage proficiency upon hearing these negated suggestions. The specific questions to be addressed are:

(a) Do English-speaking students of different levels of Spanishlanguage proficiency understand the intent of the suggestion when theSpanish form is similar but different from the English one, and expressesa different intent than that of the English form?

The question implies that learners may transfer their speech act knowledge from the first language to the foreign language. In the case ofnegated interrogative suggestions, one may ask whether learners perceivethe difference in the forms regarding the negative element. If they dounderstand the suggestion intent and do not notice the negative element,they will attach the suggestion illocutionary force to the utterance, relying on the context and their knowledge of suggestion forms from the firstlanguage. On the other hand, if they do perceive the difference in form,it is hypothesized that learners may simply understand the utterance as asuggestion formulated according to Spanish constraints (as opposed toEnglish ones), or misunderstand the intent.

(b) Do perceptions of the differences in Spanish and English formscause misunderstanding and negative reactions to the speaker?

If learners do perceive the differences in form, they may misunderstandthe intent and react negatively to the speech act, based on first languagespeech act knowledge. This interpretation could account for some miscommunication and negative attitudes toward native Spanish speakers.

(c) How do the learners respond verbally to the speech act?If the students misunderstand the intent and react negatively, they may

respond verbally in a negative manner, thereby causing further miscommunication.

(d) Is there a marked difference in understanding of the speech actaccording to level of foreign language proficiency?

It is hypothesized that there is an increase in ability to identify thespeech act and to reproduce it correlated with higher levels of Spanishforeign language proficiency. It is predicted that lower-level students willunderstand the global intent alone, if they understand at all, and will not


hear the negation. It is also predicted that higher-level students willunderstand the intent and that some will hear the negation and react in anegative manner to the utterance.

In order to seek answers to the research questions posed above, anexperiment was carried out with English-speaking or bilingual students atdifferent levels of Spanish language study during a summer session in auniversity program.

4.2. Subjects, instrumentation, and procedure

The 114 students who participated in this study were of three differentlevels of language study, as follows: 46 first-year students in intensiveSpanish classes at the beginning of the equivalent of the second semester;34 second-year students, at the end of their third semester of Spanish; and34 advanced students, most in their third or fourth year of universitySpanish courses. The advanced students included both non-native (68 0/0)and bilingual Spanish-English Chicano students (32 %). The number ofbilingual speakers in the other groups was negligible. Two classes taughtby different non-native instructors for each level were involved. Theinformants were given a context for each situation, and were then askedto watch a very short portion of a videotaped speech act by a nativespeaker. There were seven speech acts in total (see Appendix One). Theseven native speakers were given a script from which they could varyexcept for the "key" sentence that explicitly expressed the speech act inquestion. 4 They were asked to speak directly to the camera, and try to actas naturally as possible. A videotape was used instead of an audiocassette tape because it was believed that the communication of speechacts is much more effective with "holistic" information, including notonly the actual utterance but also body movements and facial expressions. Included in the tape were one rebuke, one request, one informationquestion, and four suggestions. One of the suggestions contained anirrealis clause in declarative form (e. g., "If I were you, I would ... "), andtwo others represented negated interrogative suggestions. Another was anon-negated interrogative suggestion, which does not convey a suggestion intent.

Students watched each of the seven situations and, immediatelyfollowing each one, answered three basic questions on the questionnaireshown in Appendix Two. First, they were asked to respond, in English orSpanish, to the speaker they heard and saw, as if they were answering


back to the speaker. Second, informants were to identify the type ofspeech act expressed (e. g., suggestion, apology) and, if possible, to reproduce how it was expressed. Third, students were asked to evaluate thespeakers, using a Likert-scale, in terms of degrees of aggressive/passive,rude/polite, non-communicative/communicative, strong/weak, and unfriendly/friendly.

4.3. Results

4.3.1. Listening comprehension and responses

The results from the experiment were tabulated and run through theStatistical Analysis System (SAS) for analysis. Table 1 reveals the resultsin percentages of frequency of occurrence for all groups in all situationsregarding the targeted identification of the speech act.5

In general, only about one third of the first-year and one fourth of thesecond-year students understood the intent of the speech acts (exceptSituation Five), revealing that it is a skill that relatively few students atthese levels can perform. On the average, a little over half of the advancedstudents understood the intent. The results were also run through Tukey'sStudentized Range (HSD) Test to compare the performances of the threegroups. The comparisons reveal that the first- and second-year studentswere not significantly different but the advanced group was significantly

Table 1. % of correct answers for each group by situation

Group

First yearSecond yearAdvancedAverage scoreof all groups

Situation

1 2 3 4 5 6 7

33 28 37 22 17 43 2647 21 29 21 6 32 2956 53 50 62 3 59 68

44 33 39 33 10 44 40

(df =2; Sum of Squares =56.37; Mean Square =28.18; F =14.65; P <0.001)n sizes:First year = 46Second year =34Advanced = 34


Table 2. % of answers by each group for each option in identification of speechact (correct answers underlined)

Option Situation

1 2 3 4 5 6 7

First year classes (n =46)Invitation 2 0 0 17 13 20 0Apology 2 7 0 7 0 0 7Request 2 4 9 15 0 43 4Suggestion 33 35 37 22 22 0 26Order 0 7 2 2 0 2 0Information Question 41 2 20 11 17 22 9Mild rebuke 9 28 17 2 4 4 17Other 2 2 7 2 30 4 20NR 9 15 9 22 14 5 17

Second year classes (n =34)Invitation 9 0 0 9 6 29 0Apology 0 0 3 3 3 0 3Request 0 9 0 9 12 32 9Suggestion 47 32 29 21 12 6 29Order 0 6 0 0 0 0 0Information Question 32 12 6 15 2 24 6Mild rebuke 6 21 18 0 6 3 15Other 0 3 24 21 15 3 18NR 6 18 21 22 41 3 21

Advanced classes (n =34)Invitation 9 0 0 9 15 0 0Apology 0 0 3 6 0 3 0Request 0 6 0 9 0 59 0Suggestion 56 15 50 62 53 0 68Order 0 0 0 0 3 0 9Information Question 21 15 12 9 3 32 3Mild rebuke 9 53 6 3 .3. 0 12Other 0 3 6 3 15 6 6NR 6 9 23 0 9 0 2


different from the other two (mean scores first year = 2.07; secondyear =1.85; advanced =3.50).

The results, however, should be examined with consideration offactors in the situations themselves. Table 2 shows the answers given ineach situation by each of the three groups of students. I will focus on theresults of Situations One, Three, Four, and Five, which involve negatedand non-negated interrogative suggestions. In Situation One, manystudents thought that a simple information question was being asked in2No has pensado en estudiar con tus coLegas de La cLase ahora? 'Have(n't)you thought about studying with your colleagues in the class now?' Whilethe intent of the speaker is clearly that of a suggestion, learners sometimesmisunderstand speech acts, especially when they are not prefaced by suggestion formulaic expressions, such as "Why don't you" and "Howabout." Only a few students in each of the three groups thought theutterance was a mild rebuke. Thus, the data indicate that most of thestudents did not hear the negative element in this utterance, or if they did,they did not associate it with the English interpretation.

The frequencies of appropriate responses to the speech act itself, displayed in Table 3, show that several first- and second-year students andabout half of the advanced group produced a logical and appropriateanswer to the utterance, in spite of the fact that relatively few couldreproduce the utterance accurately, as shown in Table 4.

This pattern, in which many informants could respond to the speechact appropriately but could reproduce the utterance only on occasion,held in all the situations. In general, this indicates that even though manystudents understood only partially or not at all, they could many timessay something that would encourage further interaction and more inputfrom the speaker, which presumably would lead them to understand theintent.

Table 3. % of logical and appropriate responses to the speech act by all threegroups in all situations

Group Situation

1 2 3 4 5 6 7

First year 13 11 22 33 46 63 22Second year 21 18 18 12 3 32 18Advanced 56 76 32 71 65 79 82


Table 4. % of students who reproduced the targeted utterance verbatim

Group Situation

1 2 3 4 5 6 7

First year 2 4 0 4 2 0 9Second year 0 3 0 3 0 0 0Advanced 29 65 9 41 21 29 29

The data from Situation Three are interesting for their variety (seeColumn 3, Table 2). Some students of the first year (20 %) and advanced(12 %) levels thought that the speaker was asking a simple informationquestion with 2No deberias hablar con la profesora y pedir que te decredito extra por ese trabajo? 'Shouldn't you talk with the teacher and askthat she give you extra credit for that work?' Some first-year (17 %) andsecond-year (18 %) students thought they heard a mild rebuke, which wasthe predicted outcome from a transfer of the English interpretation of thetarget utterance. However, 24 % of the second-year students believed theyheard another kind of speech act, usually labelled "Complaint" or"Statement of a Problem." Twenty-one percent of the advanced studentscould not identify the speech act at all. The responses to the speaker andthe reproductions of what the speaker said in Tables 3 and 4, respectively,show that most students of all levels could not reproduce the utteranceaccurately, and only about 20 % of the first- and second-year group and32 % of the advanced students responded appropriately.

Situation Four, with the utterance 2Por que no copias las hojas deRaquel? 'Why don't you copy Rachel's pages?' also revealed some misunderstanding, especially by the first- and second-year students. Severalfirst-year students thought they heard an invitation to have coffee (17 0/0),while others believed they heard a request to borrow some notes (150/0)(see Table 3). Some second-year informants thought the speaker asked aninformation question (15 0/0) or thought they heard another speech act,most commonly labelled "Giving Information" (21 %). These misinterpretations indicate that the students probably inferred the speech actfrom the context when they were unsure of what they had heard. Hearingsomeone talk about coffee and the class homework, and having twentyminutes before class could lead some to think they were invited to have acup, or asked to lend their notes. Almost half the advanced students(41 %) were able to reproduce the utterance exactly as in the situation,


which was prefaced by the 2Por que no? expression, as seen in Table 4.This time, however, the expression did lead more of them who relied onit to interpret the speech act correctly (62 %). So, as in Situation Two, itappears that many students rely on the formulaic expression to help themin speech act comprehension.

In Situation Five, the speaker asked 2Deberias conocer a su mejoramigo, Francisco? 'Should you meet his best friend, Francisco?' whichrepresents a yes-no question and not a suggestion in Spanish. The speakermade an effort to use question intonation in expressing this question, butthe rise in pitch seems to have been too subtle for all levels. Moreover, thesituation lends itself to a suggestion, since the speaker is saying that sheknows the listener is shy and wants to talk to Eduardo. Suddenly, she asksif the listener should perhaps get to know his best friend, Francisco, whois very nice. Without the question intonation, this utterance expresses asuggestion, and the speaker did not make the interrogative intonationclear enough, as verified by some native speakers of Spanish who alsosaw the videotape. Some students in all three groups heard the utteranceas a suggestion (first-year: 22 0/0; second-year: 12 %; advanced: 53 % see Table 3). Thirty percent of the first-year students heard anotheract, usually labelled "description/information" or "opinion," which indicates that they comprehended the statements about the friend morethan the suggestion to meet him. Almost half of the second-year informants did not understand at all. Some of the advanced students (15 0/0)also heard the same speech act as some of the first-year group, labelled"description" or "opinion." In Column 5 of Table 4, the 21 % exactreproduction of the utterance is misleading, since most of these informants did not use any punctuation at all in writing their replies.Therefore, they were probably reporting a suggestion rather than aquestion.

In general, the data suggest that the more advanced students, who aremore proficient in listening comprehension, are better at understandingthe global intent of the speech act, as expected. As far as noticing thenegative element in some of the interrogative suggestions, it appears thatonly some of the advanced students notice this element, or at least onlysome can notice it and retain it long enough in their memory to write itdown. These students were most competent in reproducing the mildrebuke of Situation Two, which was also the shortest speech act, followedby the suggestion of Situation Four. Both of these utterances werepreceded by the 2Por que no? expression. This indicates that students relyheavily on these expressions to help them understand utterances. It is


interesting that in Situation Two, many students reproduced the utterance exactly ( CC

2Por que no viniste a verme antes?~~), and four of thosestudents still identified it as a suggestion, showing that they ignored theverb tense cue.

4.3.2. Judgments of the speaker

Table 5. Mean reaction by group in each situation

Situation 1 aggr rude .,.com strong -friendFirst year 2.3 2.1 2.5 3.1 2.3Second year 2.6 2.2 2.5 3.5 2.3Advanced 2.2 1.6 1.9 3.3 1.7Situation 2First year 3.2 3.0 2.8 3.4 3.2Second year 3.2 3.0 2.7 3.6 3.1Advanced 3.7 3.5 2.8 3.9 3.8Situation 3First year 3.1 2.8 2.9 3.7 3.0Second year 3.5 2.9 3.1 3.5 3.0Advanced 2.7 2.0 2.2 3.2 2.1Situation 4First year 2.2 1.9 2.2 3.1 2.0Second year 2.4 1.6 2.3 3.1 2.0Advanced 1.9 1.6 1.8 2.8 1.6

Situation 5First year 2.2 1.8 2.3 2.9 1.7Second year 2.4 2.0 2.9 2.6 1.9Advanced 2.2 1.6 1.7 3.5 1.5Situation 6First year 3.2 2.6 2.3 3.5 2.3Second year 3.3 2.5 2.1 3.6 2.0Advanced 3.5 2.3 2.0 3.6 1.9

(n =85; First yr. =37; Second yr. =25; Adv. =23)Key:aggr =aggressiverude =rude-com =non-communicativestrong =strong (personality)-friend =unfriendly


Table 5 shows the mean score given to each speaker by the three groupsof students based on the five characteristics of aggressiveness, rudeness,communicativity, strength, and friendliness. The top score was 5,denoting the negative characteristic, and the bottom score was 1,denoting the most positive. Only those students who rated all sixspeakers (Situation Seven was not rated) were included in the tabulationsof results (n =85). In a general overview of the results, it appears that thefirst- and second-year students had a similar perception of the speakers,and the advanced students differed from them, at times, significantly. Ifthis is true, then it indicates that people's opinions of a speaker changewhen they understand the linguistic intent of the speaker's message.When they do not understand or misunderstand the message, they areprobably reacting to what they thought they heard or are basing theirjudgments on other features of the utterance, such as intonation, or thespeaker's physical appearance or gestures.

Returning to the question of how students reacted to the Spanishnegative interrogative suggestions when they heard them, the answers ofthe advanced students were examined in detail since these students weremore likely to have understood the intended speech act. Of the tenadvanced students who correctly reproduced the negative interrogativesuggestion in Situation One, three were non-native Spanish learners andsix were bilingual Chicano Spanish speakers. Two of the non-nativestudents rated the speaker toward the positive end of the scale with lowscores, except for the "strong" category. These two students gave thespeaker a 4, tending toward "strong." The mean score by the three nonnatives for this item was 3.7. Three of the six Chicano students also ratedthe speaker with a 4 in the same item. Two students also rated the speakerwith 4's in the non-communicative category, and one found the speakervery unfriendly. The mean score for these Chicano students in the"strong" item was 3.0. The mean for all advanced students was 3.3. Theresults may reflect some variance in interpretation of the "strong/weak"characteristic. That is, some students may have believed that it is morepositive to have a "strong" as opposed to a "weak" image, and thereforerated the speaker as strong when they wanted to convey a positive rating.If this were the case, then the scores for this category should reflect ascore on one end of the scale when most of the other categories reflectedscores on the opposing end. While the scores for the "strong" item aresometimes higher than all. the others, they are not consistently so, i. e.,when students seemed to be reacting negatively towards a speaker, the"strong" scores would tend toward the higher end of the scale. Another


interpretation of the results is that the students may have reacted to thenegative interrogative suggestion, and found the speaker to be strong inthe sense of "strong-willed" or domineering. This interpretation seems tobe more viable for these informants.

None of the ten advanced students who understood the utterance inSituation One and also noticed the negative element responded in anegative way, and in fact, they generally agreed with the speaker. Fourfirst-year students thought they heard a rebuke, and two of themresponded with answers that can be described as defensive ("The last testwas not that easy"; and "I attend class and study very hard.") but notnegative. The two second-year and two advanced students who thoughtthe speaker was expressing a rebuke did not react negatively. Of course,it is not expected that students would respond to a teacher's suggestionor rebuke in an openly negative way, given the deference factor andpower differential. In general, then, the responses were not negative, evenwhen the students believed they were being scolded.

Those questionnaires that indicated that the speech act in SituationThree (2No deberias hablar con la profesora ... ?) was considered a mildrebuke were examined. Of the eight first-year students who heard a mildrebuke, two wrote no response. The others did not write negativeresponses (e.g., "I have a class"; "I will work harder."). Six second-yearstudents also interpreted the utterance as a mild rebuke, and two had noresponse. The other answers were not negative in nature, ranging from arequest to repeat the utterance to statements regarding the teacher andthe way she helps the students. The two advanced students, however, didreact somewhat negatively, responding with "I don't feel sorry for you forgetting in trouble with the teacher" (a response that is inappropriate forthe utterance given) and "Hey, I wanted sympathy, not a reprimand."Thus, the more proficient, advanced students seem more likely to respondverbally to a peer in a negative manner if they misunderstand theutterance to convey a negative meaning, probably because they are morecapable of responding. Of those three students who did understand theutterance correctly and also noticed the negative element in the suggestion, none gave a negative response.

A Repeated Measures Analysis of Variance was performed, examiningfactors of the situation and the characteristic in question, and theperformance of the three groups. The results, seen in Table 6, show thatthere is a significant difference in how the informants of the three groupsrated the situations and characteristics, but the difference is correlated tothe particular situation and the characteristic in question.


Table 6. Source of variance (Repeated Measures Analysis of Variance - Tests ofhypotheses)

df 5S MS F P

Within groupsSituation 5 370.30 74.06 54.21 0.0001Character 4 362.71 90.68 113.86 0.0001

Between groupsGroup 2 22.22 11.11 3.82 0.03Error 82 238.74 2.91

5. Conclusions

Regarding the research questions posed at the beginning of this study, thedata show the following:

(a) Do English-speaking students of different levels of Spanishlanguage proficiency understand the intent of the suggestion when theSpanish form is similar but different from the English one:} and expressesa different intent than that of the English form?

The first- and second-year students had difficulty in comprehending the global intent of the speech acts in these situations whilemore of the advanced students were more successful. The advancedstudents performed better in those situations in which the headact was expressed in a form using the same formulaic expressionsas those in English, namely "Why don't you" and "If I were you, I(would)." They were slightly less successful in Situations One andThree in which the forms were similar but different from English inthat they were negated in Spanish. All groups performed poorly incomprehending the intent of Situation Five, in which the head act wasuttered with interrogative instead of declarative intonation, and represented a speech act other than one anticipated from the context. Thefault, however, probably lies in the fact that the speaker did not make theintonational difference clear enough. All the results point to the fact thatstudents of all three groups transferred their first language speech actknowledge in differing degrees in understanding the foreign languagespeech act.

(b) Do differences in form cause misunderstanding and negative reactions to the speaker?


The data showed that there was misunderstanding of the intent of thespeech act by about half of the advanced students and about 60 to 75 %of the first- and second-year students, who performed similarly.

In Situation Three, the first- and second-year students tended to judgethe speaker as more aggressive, non-communicative, and strong. It shouldbe remembered that some of the informants in these two groups alsoheard the speech act (i. e., 2No deberias hablarle a la profesora ... ?) as amild rebuke (18 % - see Table 2) or a complaint (24 %). In this situation,some correlation could be drawn between the misinterpretation of thespeech act and the tendency to react negatively to the speaker for somecharacteristics.

As for negative reactions to speakers due to the negative element ininterrogative suggestions, the data are too few to draw any definitive conclusions. The preliminary finding in this study, however, is that advancednon-native speakers who do hear the negative element do seem to reactmore negatively than the bilingual Chicano students who also hear thenegation. If this finding is valid, then there is a possible source of negativereaction to Spanish speakers based on this lexical difference in interrogative suggestions.

(c) How do the students respond to the speech act?In general, many students of all levels were able to say something in

response to all the speech acts that would lead to further interaction andinput from the speaker, which would probably then lead to an understanding of the speaker's true intent through negotiation for meaning.Even when learners noticed the negative element in the negative interrogative suggestions or misinterpreted the speech act according to anEnglish interpretation, they were not likely to respond verbally in anegative manner and thereby cause further miscommunication andmisunderstanding.

(d) Is there a marked difference in understanding of the speech actaccording to the level of foreign language proficiency?

The data indicate that the advanced students were much morecompetent in understanding the true intent of the speech acts than thefirst- and second-year students.

In conclusion, the results of this limited experiment indicate, on thepositive side, that learners usually try to keep the interaction going withthe speaker through their responses to speech acts even when they onlypartially understand or really don't understand at all. In all probability,this continued interaction would eventually lead to comprehension of thespeaker's intent. On the negative side, the results suggest that speech act


comprehension is difficult for first- and second-year language students,even when the utterances are formulated in the same way in bothlanguages.6 Comprehension is more likely to occur at the higher levels oflanguage study when students are able to understand larger units of discourse in a context. A few of the responses to the negated interrogativesuggestions, however, suggest that some non-native advanced studentsmisunderstand the utterance due to an analysis of the components of anutterance and a transfer of first language speech act knowledge and thelexical and syntactic elements associated with particular speech acts inthe first language.

Regarding transfer in the listening comprehension skill, these datareveal that the transfer strategy is applied by learners at different levels ofproficiency, leading to some correct and some incorrect hypotheses aboutthe input. The students seem to transfer pragmatic knowledge inmatching what they can understand of the utterance to the context andother cues such as intonation. The more advanced learners, who canbegin to analyze the input more closely (i. e., pay attention to details oflexicon and structure) can sometimes make incorrect hypotheses aboutthose details they find if the structural/lexical items match those of thefirst language and if they transfer their first language knowledge of thoseitems to the foreign language situation. Thus, a secondary type of transferis demonstrated.

Finally, on a pedagogical note, foreign language educators should beaware that possibilities for misunderstanding and miscommunicationexist for students, even more so in the advanced levels of language studywhen students are more likely to analyze the components of utterancesthey hear. Many times it is assumed that if the message looks and soundsthe same grammatically and lexically in both languages, then studentsshould be able to understand through transfer and analogy. The results inthis study indicate that the comprehension process is much morecomplex, especially in the case of speech acts, since they are so contextdependent for meaning. That is, learners must not only learn to decodeat the utterance level, but also learn to read the context and its relationship to the utterance. They must also learn differences between the firstand target language formulations of the speech acts, such as the use of thenegative in interrogative suggestions. This process implies a knowledge ofthe target language speech acts at both the grammatical/lexical level aswell as the pragmatic level of use. This in turn implies an ability to makepragmatic adjustments in situations of language contact (Blum-Kulka1989: 65; Weizman - Blum-Kulka 1987: 65), whether they be in the


classroom or in a "real life" language interaction. Students of Spanishneed to be exposed not only to the language itself, but to a contextualized, interactive language, such as through videotapes. Only throughexposure to contextualized language at all levels will students becometruly proficient in language use, and understand the target language waysof speaking.7

Appendix one

Note: All scripts below represent what the native speakers actually said in thevideotapes.

Situacion 1: (suggestion)Professor sitting in office. (female - Colombia)Professor: jHola! Pasa. Si, tengo tiempo para hablarte ahora. ~Esta preocupadocon la nota que sacaste en el examen parcial? Pero no fue un examen muy dificil.Em-~no has pensado en estudiar junto con tus colegas de la clase ahora? 'Hi!Come in. Yes, 1 have time to talk to you now. Are you worried about the gradethat you got on the midterm exam? But it wasn't a very hard exam. Urn - haveyou thought about studying together with your fellow students in the class now?'

Situacion 2: (rebuke)Professor in office. (male - Cuba)Professor: No se si tu problema tiene que ver con ellibro de texto, que probablemente te cuesta mucho trabajo entender. Bueno, ~por que no viniste a hablarmeantes? 'I don't know if your problem has to do with the textbook, which probablyis very difficult for you to understand. Well, why didn't you come to speak to meearlier?'

Situacion 3: (suggestion)Young man at table. (male - Spain)Chico: Si, yo tambien creo que es injusto que la profesora nos de tanto trabajo.Pero es mas injusto que ella te de un trabajo extra. No puedo creerlo. 2Nodeberias ir a hablarle y pedir que te de credito extra por ese trabajo? Asi tal vezsuba tu nota en la clase. 'Yes, I, too, believe that it's unfair for the teacher to giveus so much work. But it's even more unfair for her to give you extra work. 1 can'tbelieve it. Shouldn't you go and talk to her and ask her to give you extra creditfor that work? That way maybe your class grade will go up.'

Situacion 4: (suggestion)Young man in room drinking coffee. (male - Puerto Rico)Chico: jHola! Estoy aqui tomando un cafecito antes de ir a la clase. Me dices quenecesitas las hojas de tarea que la profesora nos dio la semana pasada, pero notengo las mias. Por que no copias las hojas de Raquel? Ella esta por aqu t cerca.(Looks at watch). Tenemos unos veinte minutos. 'Hi! I'm here drinking some


coffee before going to class. So you say that you need the worksheets that theteacher gave us last week, but I don't have mine. Why don't you copy Raquel's?She's around here. We have about twenty minutes.'

Situacion 5: (question)Girl in room. (female - Guatemala)Chica: Mira~ no te pongas triste. Se que tienes muchas ganas de conocer aEduardo~ pero que eres demasiado timida para hablarle primero. ~Deberias tratarde conocer a su mejor amigo~ Francisco? Es muy simpatico. 'Look, don't be sad.I know you really want to meet Eduardo, but you're too shy to say something tohim first. Should you try to meet his best friend, Francisco? He's really nice.'

Situacion 6: (request)Girl coming into room. (female - Peru)Chica: lMaria! lHola! Oye~ te vengo a devolver ellapicero que dejastes en mi casaanoche. iQue bonita tu blusa! (No podrias prestarmela para esta noche? 'Maria!Hi! Hey, I've come to return your pen, which you left at my house last night.What a pretty blouse! Could you lend it to me for tonight?'

Situacion 7: (suggestion)Young woman at table. (female - Panama)Chica: Si~ yo se que es muy dificil~ pero tienes que hacer algo. Marta no te puedetratar asi. Si yo fuera tu~ Ie diria que tienes los mismos derechos que ella~ losmismos derechos que cualquier ser humano. 'Yes, I know it's hard, but you haveto do something. Marta can't treat you like this. If I were you, I would tell herthat you have the same rights as she does, the same rights as any other humanbeing.'

Appendix two

CuestionarioDirections: You will see a series of videotaped interactions, in which a speaker issupposedly talking to you. After each interaction, the tape will be stopped andyou will have a few minutes to answer the questions below. The first questionasks you to write the first thing you would say to the speaker on the tape inresponse to what he or she says. The next question asks you to identify the maingist of what the speaker said, and if possible, reproduce it, and the last questionasks you to react to the speaker.

Situation 1: You go to see your Spanish instructor in her office because you arehaving some trouble.1. Please write down a response in English or Spanish to what the speaker has

just said to you.2. What was the main gist of what the speaker said? (Circle one)

A. InvitationB. Apology'


C. RequestD. SuggestionE. OrderF. Information questionG. Mild rebukeH.Other--------------------------

If you can remember how the speaker said the last sentence, please write it herein English or Spanish.3. Please rate the speaker in terms of each characteristic, which you can base on

your reaction to the speaker. Circle one number for each item.a. aggressive passive/gentle5 4 3 2 1b. rude polite54321c. non-communicative communicative5 4 3 2 1d. strong weak5 4 3 2 1e. unfriendly friendly5 4 3 2 1

Situation 2: You go to see your other Spanish instructor, because you are alsohaving problems in that class. (Same Questions 1-3)

Situation 3: You are in the library and run into one of your classmates and complain to him. (Same Questions 1-3)

Situation 4: You go to the Union and see one of your classmates and ask him aquestion about the work.

Situation 5: You (a young woman) sit down to chat with one of your friendsabout another person.

Situation 6: You (a young woman) are in your room when one of your friendsdrops in.

Situation 7: You sit down to chat with one of your friends about another person.

Notes

1. Investigations by Schumann (1979) and Zobl (1980a; 1980b) found the useof preverbal negation by Spanish-speaking learners of English (e. g., "I nowant that.") as evidence for an order of acquisition of morphemes in thesecond language or foreign language, based on transfer from the first language.I mention these studies here because they are relevant to the topic of theacquisition of negation. The negation in the sentences examined in this study,


however, does not change the truth value of the speech act utterances. Thisdifference could be crucial in the way learners deal with negation in acquisition.

2. Two forms in Spanish are offered here, corresponding to what appears to bedialectal variation. I thank Ninfa Burgos-Kohler for pointing this out to me.The symbol # is used to denote a sentence that, pragmatically speaking, is notappropriate for the context.

3. The difference discussed here is really a difference between polite and deferentinteraction, discussed in Koike (1992: 30-34). Briefly, in deferent interaction,social distance to the point of little or no direct interaction is emphasized. Inpolite interaction, the element of respect while maintaining direct interactionis present.

4. Although they were allowed this freedom in expressing themselves, thespeakers actually strayed very little from the script given to them. Their speechacts were checked for extra softeners of illocutionary force, but none werefound. Two elements that could not be controlled, however, were intonationand facial expression. All were asked, however, to be serious and not to smile.

S. Since the purpose of this study is to examine the suggestions in the data as awhole, the tables do not show categories of the data according to register.

6. These findings contrast to those of a previous study (Koike 1989: 282-283)in which students showed that almost all students of first, second, andadvanced levels demonstrated an understanding of speech acts. It should benoted, however, that (a) those speech acts were simpler in language; (b) theywere read to the students by their own teachers, instead of by a series of unfamiliar native speakers; and (c) the rate of speaking was probably slower thanin normal speech.

7. I thank the graduate assistant instructors and undergraduate students of theSpanish classes at the University of Texas at Austin who graciously agreed toparticipate in this experiment.

References

Blum-Kulka, Shoshana1989 "Playing it safe: The role of conventionality in indirectness", in:



view", in: Shoshana Blum-Kulka-Juliane House-Gabriele Kasper(eds.), 1-34.


Ablex.Brown, Penelope - Stephen Levinson

1987 Politeness: Some universals In language usage. Cambridge:Cambridge University Press.


Larry

Larry Selinker

Larry Selinker

1983

1992

1992

1983

1992

1994

1992

Corder, S. Pit1975 "The language of second-language learners: The broader issues",

Modern Language Journal 59: 409-413."A role for the mother tongue", in: Susan Gass{eds.),85-97."A role for the mother tongue", in: Susan Gass(eds.), 18-31.

Gass, Susan1989 "How do learners resolve linguistic conflicts?" in: Susan Gass

Jacquelyn Schachter (eds.), 183-199.Gass, Susan - Jacquelyn Schachter (eds.)

1989 Linguistic perspectives on second language acquisition. Cambridge:Cambridge University Press.


House.Language transfer in language learning. Amsterdam: John Benjamins.

Hatch, Evelyn (ed.)1979 Second language acquisition: A book of readings. Rowley, MA:

Newbury House.Hobbs, Dianne

1990 Gender-based strategies in issuing directives in Mexican Spanish.[Unpublished Ph. D. dissertation, The University of Texas at Austin.]

Kellerman, Eric1977 "Towards a characterization of the strategy of transfer in second

language learning", Interlanguage Studies Bulletin 2.1: 58-145."Now you see it, now you don't", in: Susan Gass - Larry Selinker(eds.), 112-134.

Koike, Dale1989 "Pragmatic competence and adult L2 acquIsItIon: Speech acts in

interlanguage", Modern Language Journal 73.3: 279-289.Language and social relationship in Brazilian Portuguese: The pragmatics of politeness. Austin, TX: University of Texas Press."Negation in Spanish and English suggestions and requests:Mitigating effects?" Journal of Pragmatics 21: 513-526.

Schachter, Jacquelyn1983 "A new account of language transfer", In: Susan Gass

Selinker (eds.), 98-111."A new account of language transfer", In: Susan Gass - LarrySelinker (eds.), 32-46.

Schumann, John1979 "Second language acquisition: The pidginization hypothesis", in:

Evelyn Hatch (ed.), 256-271.Searle, John

1979 Expression and meaning: Studies in the theory of speech acts.Cambridge: Cambridge University Press.


Selinker, Larry1972 "Interlanguage", IRAL 10: 209-231.

Weizman, Elda, - Shoshana Blum-Kulka1987 "Identifying and interpreting translated texts: On the role of prag

matic adjustment", Indian Journal of Applied Linguistics, 13.2:61-73.

Zobl, Helmut1980a "The formal and developmental selectivity of L1 influence on L2

acquisition", Language Learning 30.1: 43-57.1980b "Developmental and transfer errors: Their common bases and

(possibly) differential effects on subsequent learning", TESOLQuarterly 14.4: 469-479.

Part IIIApplications

Suggestions to buy: Television commercials fromthe U. S., Japan, China, and Korea

Richard Schmidt - Akihiko Shimura - Zhigang Wang Hy-sook Jeong

1. Introduction

This chapter deals with television commercials as suggestions to viewersto buy consumer products, comparing television commercials from theUnited States, Japan, the People's Republic of China, and South Koreafrom this perspective.1 It is intended to complement and augment otherstudies of differences in the realization of speech acts across cultures, butis different from most other studies in several respects.

1) It is often assumed that the most problematic issue both in speechact theory and in cross-cultural communication is the determination ofwhat is meant by what is said. It is commonly pointed out that our goalin conversation is to convey our intentions in socially appropriate ways(Aston 1993; Grice 1975; Searle 1969; 1976) and that failure to conveyor interpret intentions may be the most important source of cross-culturalcommunication breakdown (Gumperz, 1982).

However, the goal of a television commercial is obvious; it is topersuade consumers to buy a specific product. Successful communicationis less a matter of getting television viewers to recognize this illocutionaryforce (indeed, advertisers may attempt to mask this goal) than a functionof the persuasive impact of the commercial, including the linguistic andnonlinguistic strategies used to persuade.2

2) Speech act analyses have been based on several kinds of data,including native speaker intuitions, interviews, naturally occurringutterances in face-to-face interaction, role play, spoken or written language elicited through discourse completion tests, and various perceptiontests such as card sorts, paired comparisons and rating scales (see Kasper- Dahl 1991 for review). However, we are unaware of any other studyof speech acts in different cultures that is based on data from televisioncommercials. Our data are not just unique but also represent artful ratherthan naturally occurring discourse, carefully scripted by professionalwriters, although the television commercial is naturally occurring lan-

286 R. Schmidt - A. Shimura - Z. Wang - H. Jeong

guage in another sense, not experimentally elicited for the purpose oflinguistic analysis.

3) Very useful work has been done in the cross-cultural comparison ofspeech act behavior by focusing on patterned variation in speech actrealizations. The distribution of linguistic strategies for performingparticular speech acts has been matched with speaker and hearer variables such as age, sex, social distance, and relative power, together withthe degree of imposition of the speech act involved, factors which are,according to the theory of politeness (Brown - Levinson 1987), theprimary determinants of linguistic choices in speech act realization. In thecase of the television commercial, we might expect similar variation inspeech act realization when characters on screen talk to one another orwhen particular segments of the viewing audience are targeted (e.g.,children, adolescent males, etc.), but the central notions of speaker andhearer are problematic when applied to the language of advertising.

Goffman (1981) has criticized the commonsense notion of speaker,pointing out that a speaker may be the one who speaks the message, theone who has encoded it, or the one who is committed to the beliefs expressed. In ordinary conversation, these three typically coincide; in roleplay, there may be no committed speaker; and in commercials, these rolesare distributed among actors, copywriters, producers and directors, andthe product manufacturer. As for the hearer, Lakoff suggests that inadvertising language and in persuasive language in general, there is noaddressee, but only an audience (Lakoff 1982:31). In addition, whilecommercials may use a particular kind of language in order to influencespecific target audiences and may exploit such roles as celebrity/fan, thesedevices are often used strategically, creating situations and relationshipsrather than being sensitive to them.

4) Many studies of cross-cultural variation in speech act realizationhave direct relevance for second language learners whose goal is to interact effectively with native speakers of the target language under differentsocial constraints. We do not claim such implications for our work, butwe do see this study as basic research for the teaching of language inbusiness contexts, a field for which there is great demand and littleempirical research (Johns 1986).

Television commercials from the u. S.~ Japan~ China~ and Korea 287

2. Television commercials as suggestions

Television commercials provide easily obtainable data that are relevantfor a number of sociolinguistic concerns. Commercials have beenanalyzed as expressions of cultural codes and mythologies (Barthes 1972;Hall - Saracino - Resh 1979; Mueller 1987; Sherry - Camargo 1987);as a medium-specific example of the register of advertising language witha focus on novel uses of language (Agoston - von Raffler-Engel 1979;Bhatia 1987; Leech 1966; Masavisut - Sukwiwat - Wongmontha1987; Moeran 1985; O'Barr 1979); as n:anipulative or deceptive language (Bolinger 1973; 1980; Coleman 1990; Harris 1983; Vestergaard Schreder 1985); and as an exemplar of the broader category of persuasivediscourse (Lakoff 1982; Schmidt - Kess 1985; 1987). Geis (1982) hasperceptively analyzed a number of the pragmatic aspects of Americantelevision commercials, including the ways in which product claims areinterpreted through reference to conversational maxims, but no study todate has dealt with television commercials from a speech act perspective.

As a speech act, the television commercial is clearly some sort of directive (Searle 1976) or impositive (Leech 1983). The essential point of atelevision commercial, the reason an advertiser purchases time, is that itis an attempt to get some hearer or audience, viewers in their role asconsumers, to perform some future action, that is, to buy a product. Wepropose that television commercials are best viewed as suggestions tobuy, however, rather than some other species of directive, such asrequests, orders, or hints. Commercials do not seem to be requests,because they rarely attempt to engage the hearer's compliance on theground that the speaker wants or needs the act to be done. They arenot orders, because advertisers cannot expect consumers to buy a productas a consequence of the advertiser's or manufacturer's authority. They arenot hints, because their illocutionary force is transparent (Weizman1993). A commercial can only suggest or recommend, persuading theviewer "to consider the merits of taking the action in virtue of thespeaker's belief that there is sufficient reason to act" (Fraser 1983: 40).Geis has argued that syntactic imperatives, observed to be common inadvertising (Leech 1966), are to be viewed as suggestions rather thanorders (Geis 1982:19). However, in analyzing commercials assuggestions, we are somewhat hampered by the lack of detailed studies ofthis speech act, particularly from a cross-cultural perspective. Requestshave been investigated extensively (see Blum-Kulka - House - Kasper1989 for summary), but the speech act of suggestion, a cousin of the


request, has been much less studied. We have located only two data-basedcross-cultural studies, Rintell's (1979) brief comparison of suggestions inSpanish and English and Banerjee and Carrell's (1988) comparison of suggestions by native and non-native speakers of English.

Although our discussion of television commercials so far has assumedthat a commercial as a whole is a speech act with a unifying illocutionarypoint, it may be preferable to view the commercial as a whole as a speechevent, the internal structure of which consist of a sequence of utterancesthat may differentially support such functions as suggesting, informing,entertaining, and the like. Our analysis will be based on a distinctionbetween head acts within the discourse of television commercials, thoseutterances or parts of utterances that directly realize the act of suggesting,and various supporting moves that provide grounds or reasons for something to be done or remove objections to the proposed action (BlumKulka - House - Kasper 1989; Blum-Kulka - Olshtain 1984;Edmondson - House 1981). We view the underlying discourse structureof a commercial as in example (1).

(1) Head actConsumer should buy,use (etc.) the product

+because

Supporting movesProduct is effective, stylish(etc.) Product willmake consumerhappy, healthy, young(etc.)

Commenting on the application of the distinction between head act andsupporting moves to data derived from discourse completion tests, BlumKulka and Olshtain (1984) and Edmondson and House (1981) have observed that distinguishing between the two is difficult and subjective,because what may be a supporting move in one case may function as thehead act elsewhere, for example when a hint does not preface a request butby itself conveys the force of requesting. We have attempted to minimizethis problem by providing as strict a separation as possible between headacts and supporting moves or reasons. Operationally, we have defined thehead act of suggestion in television commercials as any utterance or part ofan utterance that linguistically refers to the viewer or some other consumerbuying an advertised product or interacting with the product in some otherway, such as using it or enjoying the benefits of owning it.

We therefore allow the possibility of commercials that have more thanone utterance classified as a head act, as well as commercials that have no

Television commercials from the U.S., Japan, China, and Korea 289

head act as we have defined it, while admitting that much of theremaining language in a commercial may be suggestive in a broadersense. Even this relatively strict definition, however, does not result in theexclusive assignment of each turn or utterance to only one category,either head act of suggestion or supporting move. Consider, for example,E37 from our sample of English commercials:

E37 !'m gonna take what doctors would take, wouldn't you?

In this example, the only linguistic material that directly refers to adesired action to be taken by the viewer is the question tag wouldn't you?However, by our definition, there are two other utterance parts weidentify as additional head acts. The actor uttering the lines of thecommercial, acting as a surrogate consumer, states that he is going to takethe product (!'m gonna take) and will presumably have to buy it in orderto do so. He also asserts that other consumers would do the same (whatdoctors would take). These different strategies for suggesting arecommon in our data, and we therefore code for three instances of thehead act in this case, although the claim that doctors would use theproduct is also clearly presented as a reason (supporting move) for theviewer to act.

3. The study

In this chapter, we look at American television commercials and comparethem with commercials from three Asian countries. Japan, the People'sRepublic of China, and South Korea. There are reasons to think that acomparison of American and Asian commercials may uncover someinteresting differences. A great deal of advertising research supports theclaim that American advertising (in all media) is primarily persuasiverather than informative (Dowling 1980; Hong - Muderrisoglu Zinhan 1987; Hunt 1976; Kaynak - Mitchel 1981; Laczniak 1979;Madden - Caballero - Matsukubo 1986; Resnik - Stern 1977;Stern - Krugman - Resnik 1981). But the view from Asia is ratherdifferent.

In China, advertising was banned during the Cultural Revolution(1966-1976), and a modern advertising infrastructure has beendeveloped only since 1978, as part of a rapid shift from Marxist socialism


to market socialism (Rice - Lu 1988; Tse - Belk - Zhou 1989). Theofficial view is that the functions of advertising are to promote production, invigorate the economy, increase consumer convenience and guideconsumption, develop international economic activities, and promotesocialist moral standards (Central Administration for Industry andCommerce, cited in Rice - Lu 1988). Ho and Sin (1986) foundthat Chinese managers hold that the main purpose of advertising is toinform.

In both japan and Korea, Miracle (1987) claims that advertisers relyon a feel-do-learn strategy, in which the primary goal of advertisers is toentertain and establish feelings and moods that are transferred to theproduct, as opposed to the predominant Western learn-do-feel sequence,in which advertising presents reasons for buying, with positive feelingsthe result rather than the precursor of purchase. japanese advertisingpractices have been quite thoroughly studied, and from a number ofimpressionistic studies (Fields 1983; Miracle 1985; 1987; Yamaki n. d.) aconsistent picture has emerged. japanese television commercials are saidto be evocatively filmed, but extremely indirect in approach. In a study ofprint advertising, Sherry and Camargo report that most japanese ads"neither preach, promise nor praise; some don't even portray productattributes" (Sherry - Camargo 1987: 181).

Based on the literature on advertising in the U. S., China, japan andKorea, our initial hypothesis was the following:

H1 Suggestions will be more frequent in commercials from the u.S.than in those from any of the three Asian countries.

We are equally interested in knowing what linguistic strategies are usedto convey suggestions in commercials. As noted above, previous studieshave commented on the high frequency of imperatives in English advertising, linking the syntactic imperative to the function of suggesting, butwhy this should be so is not entirely clear. Both Rintell (1979) andBanerjee and Carrell (1988) observe that suggestions (for the benefit ofthe hearer)are more likely to be expressed directly than are requests (forthe benefit of the speaker). If one salient feature of suggestions is that themaker of a suggestion assumes no special authority over the addressee(Green 1975), then perhaps there is little potential loss of face involvedin the use of the imperative, which is normally viewed as not allowingoptions However, Banerjee and Carrell found that imperatives were notcommon among suggestions elicited by discourse completion question-

Television commercials from the U. S., Japan, China, and Korea 291

naires, accounting for fewer than 10 % of the responses offered by bothnative and non-native speakers (Banerjee - Carrell 1988: 331) and wereused only in situations requiring immediate attention. Leech (1966) suggested that the high frequency of imperatives in British advertising mightbe a function of several factors: the fact that commercials are necessarilyobvious in declaring their intention, plus the need for attention value,memorability, and selling power.

Whatever principle may explain the occurrence of imperatives inAmerican and British advertising best, we do not expect imperatives to befound as often in Asian television commercials. Even though languagemay function pragmatically in rather different ways in persuasivediscourse than in ordinary conversation (Lakoff 1982), there is everyreason to expect cross-cultural differences in persuasive language, withconsiderable influence form the norms of face-to-face interaction. Forjapanese, it has been said that the bald, unadorned imperative is hardlyever used, and that "the form is considered even too forceful for recipesand other regular instruction" (Matsumoto 1988: 420). Takahashi (1987)developed a taxonomy. for comparing levels of directness in japanese andEnglish directives (using role plays to elicit data), finding that japaneseemployed a more indirect approach in general and were especially likelyto employ hinting strategies. While there is clearly a danger in exaggerating the stereotype of direct American usage as opposed to indirectjapanese speech act performance (Beebe - Takahashi 1989), observationsthat japanese advertising is more evocative than persuasive also suggestthat imperatives will be infrequent in japanese television commercials andthat less direct strategies for making suggestions will be used.

For Korean and Chinese commercials, we have less information uponwhich to base our hypothesis, since there have been no comparativestudies of suggestions or other directives in these languages that we areaware of. Both Koreans and Chinese are often observed to be moreforthright and direct than the japanese, but Koreans and Chinese, likejapanese, are said to emphasize harmonious social relationships and todisprefer any appearance of presumptuousness or over-eagerness,preferences sometimes linked to discourse strategies such as a steadybuildup of information before making a request or arriving at the important message (Kaplan 1966). We therefore hypothesize the following:

H2 Imperatives will be the most frequent form used to realizesuggestions in American television commercials, but not in anyof the three Asian languages.


3.1. Method

3.1.1. Data

The data for this study consist of one full evening of television broadcasting in Honolulu (NBC), Tokyo (Fuji), Seoul (Korean Broadcasting),and Beijing (CCTV) during the first week of March, 1988. Commercialswere recorded from the most watched station in each location at peakaudience times.

These video types yielded slightly more than one hour of commercialsin each language, from which a sample of 50 commercials was drawn foreach country. Our original intent was to balance the samples for productsadvertised, but this proved impossible, because the most commonlyadvertised products are not the same in each country. The Chinese datacontained numerous advertisements for washing machines and televisionsets, not found in any of the other language databases. The English andjapanese tapes both contained a high proportion of automobile ads(March is the traditional season for car sales in both countries), while theKorean database contained fewer commercials for consumer durables ofall types than the other three language samples (Keown - Schmidt jacobs - Ghymn 1992). Because we were interested only in consumerproduct advertising, the sample analyzed in this chapter consist of thefirst 50 commercials from each database, after elimination of repeats,public service announcements, commercials clearly of local rather thannational origin, film trailers, and promotional spots advertising othertelevision programming. We also eliminated from the Chinese sample anumber of commercials showing heavy industrial equipment, on thegrounds that few viewers could be considered potential consumers ofsuch products.

The samples also vary in length. Commercials in the U. S. vary between10 and 30 seconds (occasionally longer), in five second increments.japanese and Korean commercials tend to be shorter than those from theU.S., while Chinese commercials are longer than those from the U.S. andmore than twice as long as those from japan. For the samples analyzedhere, the mean length of commercials for each language was as follows:japanese, 13 sec.; Korean, 18 sec.; English, 22 Sec.; Chinese, 28 sec.

We believe that our samples are reasonably representative of primetime consumer advertising on major channels in the four countries, butdo not claim that our sample is representative of all television advertisingin any of them. For example, we would expect some important

Television commercials from the u. S.~ Japan~ China~ and Korea 293

differences in the language of American television commercials broadcastat different times of the day, such as Saturday mornings or late at night.We would also expect differences in commercials that are not nationallydistributed (e.g., local used car commercials) and for products soldthrough mail-order and advertised primarily on cable stations.

3.1.2. Analysis

All commercials were transcribed, including spoken, printed and sunglanguage, and three translations were produced: a morpheme-bymorpheme translation, a literal translation into English, and an idiomaticor free translation. Except in cases where linguistic form is at issue,examples are presented only in romanized transcription and free translation.

Each of the four authors of this chapter was responsible for theanalysis of commercials in his or her native language and the initialassignment of utterances to categories, after which the examples werediscussed by all four researchers until consensus was reached on eachcategorization.3

Chi-square analysis was used to test hypotheses, and the alpha levelfor significance was set at .05. Reported frequencies represent the numberof instances of a category in the whole set of 50 commercials for eachlanguage (which may include more than one instance from a singlecommercial), not the number or percentage of commercials in whichutterances of a particular type occurred. In addition to results addressingour specific hypotheses, we will also present a number of post-hoc andqualitative analyses, together with examples illustrating the tone ofcommercials from each country.

3.2. Results and discussion

As indicated in Table 1, the hypothesis that suggestions would be morecommon in American commercials than in those from any of the threeAsian countries was supported. Considering all types of suggestionscoded from our data - including suggestions made to either viewers oron screen characters, testimonials and reported behavior referring toother consumers - the set of English commercials contained a significantly higher number of suggestions than the Japanese, Chinese or


Table 1. Frequency of suggestions in the four language samples

Eng ]pns Chns Kor

Total number of suggestions: 78 35 35 53

All: X 2 = 24.73 df = 3 p<0.05Ex] X2= 15.86* df = 1 p<0.05ExC X2 = 15.86* df = 1 p<0.05ExK X2= 4.27* df = 1 p<0.05] xC X2= 0.00 df = 1 n.s.]xK X2= 4.18* df = 1 p<0.05CxK X2= 4.18 df = 1 p<0.05

Suggestions to viewer only: 54 27 26 30

All: X2 = 15.44 df = 3 p < 0.05Ex] X2= 8.50* df = 1 p<0.05ExC X2 = 9.30* df = 1 p<0.05ExK X2 = 6.36* df = 1 p<0.05] xC X2 = 0.52* df = 1 n.s.] xK X2 = 0.66* df = 1 n.s.CxK X2 = 0.78* df = 1 n.s.

* Corrected value, because df =1 (Hatch - Lazaraton 1991: 405 -406)

Korean samples. Korean commercials contained more suggestions thaneither Chinese or Japanese commercials.

The distribution of linguistic forms in suggestions is shown in Table 2,comparing imperatives (including those with please or a tag questionappended) to all other forms, but only including those suggestionsaddressed directly to the viewer. We expect that suggestions addressed byon screen characters to one another might exhibit some interestingdifferences, possibly varying as a function of speaker and hearer sex, age,status and role, but suggestions to on screen characters occurred tooinfrequently in our data to permit meaningful analysis. Some types ofsuggestions (reports of what other consumers have done) would notpermit the imperative at all.

As shown in Table 2, the hypothesis that imperatives would be thepreferred form for suggestions in American television commercials andwould be used less frequently in commercials from the three Asiancommercials was also supported. Japanese commercials also containedsignificantly fewer imperatives than either Chinese or Korean, whichwere not significantly different from each other.


Table 2. Suggestions to the viewer, Imperatives vs. other forms

Eng Jpns Chns Kor

ImperativesOther formsTotalImperatives, All:Imperatives, E x JImperatives, E x CImperatives, E x KImperatives, J x CImperatives, J x KImperatives, C x K

* Corrected values

Xl =19.25Xl = 24.50*Xl = 10.50*Xl =10.80*Xl = 4.07*Xl = 4.27*Xl = 0.55*

df =3df =1df = 1df =1df =1df = 1df = 1

p<0.05p<0.05p<0.05p<0.05p<0.05p<0.05n.s.

332154

32427

111526

102030

These two measures indicate that American television commercialsare more overtly suggestive than commercials from japan, Korea, andChina. One measure, the overall frequency of suggestions ofall types (Table 1), indicates that Korean ads are somewhat moresuggestive than commercials from japan or China, whereas thefrequency of imperatives (Table 2) indicates that japanese commercialsare the least suggestive in that respect. One possible interpretationof these findings is that the persuasive function of television isemphasized in American ads, whereas other functions are emphasizedin Asian commercials, such as simply providing information toconsumers. Another possible interpretation is that the function ofsuggestion is accomplished just as effectively though less directly in Asiancommercials.

A caveat must be raised regarding the identification of syntacticimperatives as a basic measure of directness in suggestions. Althoughthe syntactic imperative is the most direct possible way to make suggestions in each of the four languages, we cannot assume functionalequivalence across languages; in fact, we know that the japanese imperative is virtually a tabu form. Moreover, indirection in suggestions is notone-dimensional. While attempts have been made to rank directiveutterances along a single scale of directness or politeness (Takahashi1987), we find that in television commercials there are at least four waysin which suggestions can be conveyed indirectly. We have alreadyrecognized that what we consider supporting moves (reasons to buy, such


as product attributes) can function indirectly as suggestions in a broadsense, just as hints function as indirect requests in face-to-face interaction. In addition, we find examples in our data of what we will callindirection by participant shift, indirection by action shift, and linguisticindirection.

3.2.1. Indirection by participant shift

Suggestions in commercials that are addressed directly to the viewer as apotential consumer and that refer to the desirability of the viewer buyingthe product or doing something with it (direct) can be distinguishedfrom suggestions that are directed at some other addressee orsimply report what other consumers have done (indirect). As alreadyindicated (see Table 1), suggestions to the viewer were the most commontype in our samples. In each language, we find suggestions addressedto the viewer such as in sample E49, E36, J40, J8, C23, C31, K16,and K42

E49 Consider the Acura Legend Coupe. (male announcer)

E36 So don~t get confused. Shop where you like~ but start at yourBuick dealer. (on screen actor, female)

J40 Ima~ Pipp Erekiban 0 kau to~ 18-kin to daia de dekita sutekinapuchi pendanto ga chuusen de atarimasu.'If you buy Pipp's Erekiban now, you may win a petite pendantmade of 18K gold and diamonds.' (male announcer)

J8 Gankai no shiji ni shitagai tadashiku goshiyoo kudasai.'Please follow your eye doctor's directions and use themproperly.' (printed message)

C23 Xinqiu yinxiang~ nin de lixiang.'Xinqiu stereo system, your ideal choice.' (male announcer)

C31 Qing nin fuyong tongrentang shengchan de kanglaoyannianwan.'Please use the Kanglao Yannian produced by TongrentangMedicine Factory.' (female announcer)


K16 Hyokwalul senthaykhaseyyo.'Choose the effect.' (on screen actor, female)

K42 Philyohal ttayman cokumssik ccaseyyo.'When needed, press a little.' (on screen actor, male)

It is interesting to note who makes the suggestion to the viewer. InEnglish, such suggestions may be made by an on screen character,through printed messages, or through song lyrics, but the most frequentlyused strategy by far is the device of using the voice of an off screen announcer to make the suggestion (34 of 54 suggestions to viewers).Overwhelmingly (in 33 of 34 cases), the off screen announcer is an adultmale. Although the observed frequencies are too small to permit tests ofstatistical significance, interesting patterns can be seen in each of theother languages. Only the commercials from China are like the Americanones in strongly preferring the off screen announcer as the primary sourceof suggestions (18 of 26 suggestions), and in the Chinese commercialstwo thirds of those suggestions were made by female announcers. In thejapanese commercials, the most common way to present suggestions wasthrough printed rather than spoken messages (13 of 27 suggestions to theviewer). On screen actors made suggestions to the viewer as often as didan off screen announcer (7 cases each). However, when an off screenannouncer makes the suggestion, japanese commercials are just asgender-biased as American ones. All off screen voices except two weremale in our japanese sample. One was a child's voice, and onecommercial used a female off screen voice to represent the thoughts of acharacter. The Korean commercials did not use printed messages for anysuggestions, used an off screen announcer for just over half of allsuggestions (16 of 30), and - like the Chinese commercials - did notshow gender bias; 9 off screen announcers were female and 7 were male.The remaining suggestions to the viewer in the Korean commercials wereeither spoken by on screen actors or presented in song lyrics.

A less direct way to ~ake a suggestion is to make it to an on screencharacter, an actor appearing as a surrogate consumer. Such suggestionsmight be made by an off screen voice, but we found no examples of thisstrategy in any of our language samples. In each of the language samples,there are examples of actors making suggestions to other actors, such assamples E42, j2, C33 and Kll.

E42 Here, try this one. (male actor to female actor)


J2 Ojii-chan, Tansu ni gon katte kite kudasai na?'Grandpa, please go and buy Tansu ni Gon, won't you?'(daughter-in-law to father-in-Iaw 4 )

C33 Ni chi dian jianpixiaoshi wan jiu hao le, haoma?'You will feel better after taking Jianpi Xiaoshi Wan, OK?'(mother to son)

Kll Enni twuthongyakul tusil ilici.'Sister, why don't you take a headache medicine?' (female actorto female)

It would be interesting to compare commercials from these differentcountries to see which role relationships are used most often in such casesand who gives advice to whom, but it would take a larger sample thanours to identify patterns.

Alternatively, still in their role as surrogate consumers, on screenactors can indirectly convey a suggestion to the viewer or another onscreen actor by reporting that they have used a product and benefitedfrom it. We found examples of such testimonials from each country, suchas in E36, J37, Cl, and K32.

E36 Thank goodness for Pine Sol. (female actor)

J37 Boku wa itsumo hakuchuu doodoo suwan nan desu yo.'I always drink Swan boldly in broad daylight.' (male celebrity)

Cl Wo young guo, xiaoguo hai bu chuone.'I've tried it and found its effects not bad.' (female actor)

K32 Yocum daewoo patko salayo.'We live now receiving good treatment' (male actor)

Finally, instead of having an on screen character report his or herpurchase or satisfaction with the product, advertisers may suggest indirectly by reporting what other consumers have done. Again, we findexamples in each language sample as in E17, J41, C7, and K23.

E17 Since the ]ohnsons got their Mitsubishi Mirage, they~ve beendriving a lot more. (male announcer)


J41 Katakori no tonari no Gen-san Pipp katte, nonde, kiita.'Gen-san, our next door neighbor with the stiff shoulder,bought Pipp and took it.' (on screen male actor to female)

C7 Luotuo jin wanjia, wan ja huanle duo.'Camel comes to thousands of families and brings them morehappiness.' (female announcer)

K23 I taykeyto ce taykeyto soykoki Masna.'At this home and that home too, Soykoki Masna.' (song lyrics)

Between-language differences in the distribution of suggestions to theviewer, suggestions made to on screen characters, testimonials, and reportswere not significant. However, there is one additional strategy for conveying suggestions indirectly that we have found only in the U. S. commercials,a variant of the testimonial. This is modeled behavior, in which the onscreen "consumer" does not report past actions but is shown purchasingthe product or indicates that he or she is going to buy or use the product.Five U.S. commercials used this strategy, as in samples E6 and E25.

E6 Oh! I'll take it home now. (female actor)

E25 Wear-dated, please. (female actor)

3.2.2. Indirection by action shift

Any suggestion involves some threat to an addressee's face, becausepeople do not, in general, want to be told what to do. The suggestion tobuy involves a literal cost as well. In advertising, various strategies areused to present messages that emphasize the benefits and minimize thecosts to the buyer. Price mayor may not be mentioned. If it is, the pricemay be presented as a reason for buying, either because the price is lowerthan that of the competition or because it will be higher after a limitedoffer expires. Other strategies involve the formulation of price ($ 39.95instead of $ 40.00) or stating a base price in large print while mentioningrestrictions and exclusions in small print. (Both of these strategies werefound only in the U. S. commercials).

The suggestion that the consumer is to buy something, i. e., part withmoney, can also be conveyed less directly by referring to consumer


actions other than buying itself. This is done in commercials from allfour countries. We identified the following categories of suggestedaction:

Suggestion to buy: An utterance that lexically refers to buying, selling,or ordering the product, or a related financial transaction, as in samples£25, J5, C38, and K20.

£25 If you~re one of the thousands of people who asked for a freesample of Wear-Dated carpet with Stainblocker and tried it:lyou~d probably like to know what steps to take to buy it. (maleannouncer)

J5 Nyuu konseputo gokoinyuu no kata ni Goto Kumiko orijinaruterehon kaado purezento chuu.'Giving away original Kumiko Goto telephone cards to thosewho purchase New Concept.' (printed message)

C38 Dianhua dinghu:l shonghuo shangmen.'Order by phone, delivered to your home.' (printed message)

K20 Ne fin Ramyon hana te saollay?'Would you go buy one more Jin Ramyon?' (on screen actor,father to daughter)

Suggestion to get: An utterance that lexically refers to some actionwhich implies the possibility or likelihood of purchase, for example,going to a store or calling for information, as in samples £25, C9,and K44.

£25 Call for the dealer near you. (male announcer)

C9 Quing jizhu wo de shangbiao:l weili pai xiyiji.'Please remember my brand, Weili washing machine.' (cartooncharacter)

K44 Niksaykpyengul chacuseyyo.'Look for the green bottle.' (female announcer)

Suggestion to use: An utterance that refers to a consumer using the product in some way, as in samples £25, J39, Ci, and K12.


£25 Throw in your dirtiest clothes, then toss in a Fab i-Shot pack.(male announcer)

J39 Poora no Dei-ando-dei, Massaaji Kuriimu wa mizu de arainagaseru kara kantan ni tsukaemasu.'As Pola's Day And Day massage cream can be washed off withwater, you can use it easily.' (off screen voice of femalecharacter)

C1 Ni key fangxin de yong.'You can use it without worry.' (on screen female actor, to female)

K12 Wuli emanun Heinz.'Our mother uses Heinz.' (song lyrics)

Suggestion to enjoy benefits: An utterance that refers to the consumerexperiencing the benefits of owning or using the product, as in samples£43, J13, C35, and K24.

£43 Instead of spending your evening creating a classic, you canspend it enjoying one. (male announcer)

J13 Saa, hajimemasen ka, ii iki no shuukan?'Let's start the habit of new breath, shall we?' (on screen malecelebrity)

C35 Ji jiang jianmei you xiang koufu.'You can keep fit and enjoy gourmet's luck at the same time.'(male announcer)

K24 Santtushan masulo kiekhaseyyo.'Remember as a fresh taste.' (female announcer)

Suggestion with unspecified action: An utterance that implies that aviewer will interact with the product in some way, but leaves the desiredaction unspecified or makes a metaphorical suggestion, as in samples£41, J17, and K27.

£41 Listen to the heartbeat of America. (song lyric)


J17 Kotoshi wa fain na nama biiru.'For this year, fine draft beer.' (on screen male celebrity andprinted message)

K27 Memohaseyyo.'Do the Memo.' (female announcer). [Note: the product nameis "Memobis"]

Table 3 shows the distribution of suggestions to buy versus all others.Explicit references to buying (ordering, purchasing, etc.) or to selling (thereciprocal of buying) are not favored in any of the four languages investigated. In each sample, reference to other actions such as getting, using, orbenefiting from owning the product are emphasized. Korean televisioncommercials appear to represent the extreme in avoiding mention ofbuying and selling, while Japanese commercials are apparently the leastreluctant to mention financial transactions. However, between-languagedifferences in the frequencies of suggestions to buy are not statisticallysignificant. It is also worth noting that all but one of the Chinese ads inthis category simply referred to where products are sold, which mightbe taken as a simple statement of availability rather than a suggestionto buy.

Table 3. Content of suggestions

Eng Jpns Chns Kor

Suggestions to buy 7 10 7 2All other actions 71 25 28 51Total 78 35 35 53

Category x Language, All: X 2 = 14.18 df = 3 p<0.05Category x Language, English X2 = 52.05* df = 1 p<0.05Category x Language, Japanese X2= 5.92* cif = 1 p<0.05Category x Language, Chinese X2 = 12.10* df = 1 p<O.OSCategory x Language, Korean X2 = 44.80* cif = 1 p<0.05

* Corrected values


3.2.3. Linguistic indirection and politeness phenomena

As was indicated in Table 2, syntactic imperatives were the most commonform of suggestions in the American television commercials but wereuncommon in Japanese, Chinese, and Korean commercials. Table 4presents an expanded tabulation of linguistic forms found in our data,again limited to suggestions directed at the viewer. The observedfrequencies are too small to permit statistical analysis and are reportedhere simply to illustrate the range of forms found. We were also interestedin seeing whether television commercials contain other suggestion formssuch as those identified for face-to-face communication in English byEdmondson and House (1981), as in example (2).

(2) Why not ...?I suggest that you ...You should/ought to/must ...Maybe you could .The thing to do is .

No examples of expressions such as you should, you ought to or I suggestthat were found in any of the four language samples, though in Americancommercials a number of somewhat similar forms were found:

Table 4. Linguistic form of suggestions to viewer

Eng ]pns Chns Kor

Bare imperative 33 0 7 9Imperative + please or tag 0 3 4 1Elliptical imperative (no verb) 4 14 0 11Negative question 1 2 0 0N ominalization 4 4 3 0Embedded suggestion 4 1 0 0Conditional 2 1 1 0Consumer as object 3 0 6 0Ability/possibility statements 1 2 1 0Passive 0 0 4 0Propositives (e. g., let's go) 0 0 0 4Other forms 2 0 0 5

Total 54 27 26 30


E21 Why not the best?

E16 Why cook in oil, margarine, or butter?

E43 ... you can spend it enjoying one.

E25 You-'d probably like to know ...

E13 Isn-'t it nice to know ...

E9 It-'s a good time for the great taste at MacDonald-'s.

E5 It-'s gotta be a Dodge.

Other forms found in our data that have not been previously identified inthe literature as forms for suggestions included E17 and E30.

E17 Suddenly, the obvious choice. (nominalization)

E30 This one-'s gonna turn your head around. (consumer as object).

We suspect that a larger sample of commercials would yield moresuggestion forms. American commercials occasionally contain needstatements, both those referring to hearer-need (If you-'re a frequent flyer,maybe you need a new credit card [hypothetical example]) and speakerneed (At Friendly Auto Sales, we have to get rid of 200 cars and trucksthis weekend [hypothetical example]), neither of which were found in oursample. As indicated in Table 4, a range of suggestion forms was alsofound in each of the other languages investigated, and presumably alarger sample would also increase these inventories.

There are some major problems in attempting to compare the incidenceof such forms across languages. We think it unlikely that an etic gridcould be devised that would include all forms and provide a universallyvalid ranking of forms by level of directness and politeness. However,three minimal assumptions seem reasonable: (1) imperatives are moredirect and less polite than all other forms; (2) imperatives with tagquestions or overt politeness markers such as please or polite addressforms are more polite (though no less explicit) than bare imperatives; and(3) suggestions in which the addressee is the linguistic subject and inwhich the verb represents the action recommended are more direct thanthose in which the recommended action is omitted or nominalized or sug-


gestions cast as passives. These assumptions are insufficient to rank commercials from the four languages in terms of directness or politeness in acompletely precise way, but some interesting patterns do emerge for eachlanguage.

American television commercials are, as we have already noted, mostdirect according to principle (1) and are also (as seen in Table 4) the leastpolite by principle (2). English advertising is alone in its extremepreference for bare imperatives, with no mitigating devices. However, ourEnglish commercials also contain less explicit forms such as thosementioned in principle (3) and (as discussed previously) often makesuggestions less explicit through participant shift and action shift, so it isnot the case that the U. s. commercials are unremittingly aggressive.There is also an apparent interaction in the u.s. commercials betweenthese different ways of being indirect. Although English prefers the bareimperative for suggestions in commercials and although nearly 10 % ofall suggestions are suggestions to buy, we do not have a single case inwhich either the viewer or an on screen character is bluntly told to buythe product. Explicit mention of the exchange of money appears torequire either participant shift (actors as surrogate consumers may saythat they bought a product or some other consumer did) or linguisticindirection (viewers may be told that a product is a better buy or toldwhat steps to take to buy it.5

This tradeoff between the different dimensions of indirectness in theU. S. commercials was not found in our samples from Asia. The Chineseand Korean commercials in our sample contained few suggestions to buy(regardless of the type of linguistic encoding), and the japanese samplecontained very few imperatives of any sort. However, it is interesting tonote that one of the two examples of imperatives in the japanese samplewas an imperative to buy:

J9 Minna katte ne?'Everybody buy, won't you.' (child's voice)

However, this utterance occurred in a commercial directed at childrenand a child's voice was used to make the utterance, conveying anintimate, cute tone in japanese.

In the Chinese television commercials, the most striking aspect of theform of suggestions is the use of please and polite pronouns. In Chinese,the tone of an imperative may be softened by adding particles ba, Ie, orla at the end of a sentence (none of which are found in our data), usingverbs with less directness of action (as discussed under indirection by


action shift), or by adding please to the imperative. Our Chinesesample contains four instances of imperative plus please, of which C33 istypical.

C33 Nin yao xiangyao nin de xiao baobao huobokeai ma? Na qingnin gei tamen fuyong tongrentangzhiyaochang shenchan dejianpixiaoshiwan.'You want your children to be as healthy and lovely [as these]?Please give them jianpi Xiaoshi Wan produced by TongrentangMedicine Factory.' (female announcer)

Pronouns of address also indicate politeness in Chinese. For the addressee, there is a choice between ni (regular, informal) and nin (polite,honorific). In our data, nin is always used when the suggestion isaddressed to the viewer (as in C33 above, for example), although thesample contains several commercials in which on screen actors use the niform to each other. The use of polite pronouns gives suggestions inChinese commercials a polite and formal tone.

Korean commercials do not favor imperatives with please, butjapanese commercials do. Of the three imperatives in our japanesesample, one uses a question tag, and the other two are printed messageswith kudasai 'please', again sounding more like requests (by Englishnorms) than suggestions. Suggestions in japanese commercials also contain negative questions (two to the viewer and two to on screen actors),which have been identified as preferred japanese request forms(Takahashi 1987).

For both japanese and Korean, the most interesting finding withrespect to linguistic form is the very high incidence of elliptical imperatives, related to the strategy of indirection through action shift. Severalexamples of elliptical imperatives also occur in English, as in E16 andE24.

E16 Pam cooking spray, because how you cook is as important aswhat you cook. (male announcer)

E24 Pine Sol~ because you care about clean. (male announcer)

Elliptical imperatives have been noted for English requests, e. g., salt,meaning 'pass the salt' when said at a dinner table, and are generallyconsidered to be at the explicit, direct end of the continuum of directive

Television commercials from the U. S.~ Japan~ China~ and Korea 307

types (Ervin-Tripp 1976). Elliptical imperatives such as those in E16 andE24 are not direct in this context, however, but are ambiguous. Theutterance Pine Sol, because you care about clean could be paraphrased as'We created Pine soil because you care about clean' or as 'Buy Pine Sol,because you care about clean'.

What is a relatively minor suggestion strategy in English commercialsturns out to be a major strategy in Japanese and Korean. Examples fromthe Japanese sample include J14, J16, and J43.

J14 Kondo no do nichi wa ochikaku no matsuda e.'Next Saturday and Sunday, to Mazda (stores) near (you).'(male announcer)

J16 Odekake mae ni ichi kapuseru.'Before going out, one capsule.' (printed message)

J43 Migaki-arai wa Kaneyon.'For polishing, Kaneyon.' (female actor and printed message)

Korean examples include K9, K25, and K40.

K9 Ttenaki cen nal pamey Kwimitheytlul.'At night before leaving, Kwimithey.' (male announcer)

K25 Iceypwuthen, Shiny Fresh Brown.'From now on, Shiny Fresh Brown.' (song lyrics)

K40 Ismomi nappulttayn, Insatol.'When gums are bad, Insatol.' (male announcer)

Native speakers of Japanese and Korean judge most of the examples ofelliptical imperatives in our data to be typical of advertising language, notlanguage that would be used in face-to-face interaction. What isinteresting about this is that there is a connection between this conventionalized use of language in advertising and grammatical and pragmaticprinciples that function more generally in these languages. Both Koreanand Japanese are discourse sensitive languages that permit the deletion ofany constituent if it is recoverable from context, although verbs are lesslikely than nouns to be deleted. Takahashi (1987) has reported thatJapanese subjects produced directives in discourse completion tasks that


did not refer explicitly to the action to be taken, making the recipient ofthe directive responsible for guessing what was wanted.

4.0. Conclusions

Previous studies of cross-cultural advertising practices have reported thatAmerican advertising is essentially persuasive in nature, while Asianadvertising emphasizes other functions, informativeness in the case ofChinese advertising and entertainment value and the establishment ofpositive feelings in the case of Japanese and Korean advertising.Advertising research of this type is typically based on the intuitive reactions of native speaker judges, and results are reported without referenceto the language used in commercials or other forms of advertising. Thisstudy has shown that prime-time consumer product advertising on television in the U. S. is more overtly persuasive than similar advertising in threeAsian countries in terms of the frequency of suggestions and the frequencywith which imperatives are used to make such suggestions. Indeed, asLeech has pointed out, statements such as "Asian advertising is less persuasive than American advertising" only make sense if they can be relativized in terms of the pragmalinguistic strategies used in differentcommunities and situations (Leech 1983: 231). We have provided somepragmalinguistic evidence that Japanese advertising practices are at theopposite pole from those in the u.s. with respect to the speech act ofsuggesting, while Korean and Chinese commercials are nearer the middleof a continuum.

This study has uncovered a number of other intriguing facts about theways suggestions are made in television commercials in these fourcountries. Perhaps the most interesting general question that can be askedis whether the language of advertising is a function of universal pragmaticprinciples, a reflection of cultural norms, the result of the requirements ofselling in a market economy (Hall - Saracino-Resh 1979), or justa reflection of arbitrary conventions established by the advertisingindustry.

While our evidence is fragmentary, we have reason to think that eachof these forces plays a role and that no one of them is entirely responsible.Our hypotheses were based on the assumption that general culturalnorms regarding directness in language would be reflected in advertisinglanguage. At the same time, these cultural differences are manifestedagainst a background of universal principles. The suggestion to buy

Television commercials from the Uo So;, Japan;, China;, and Korea 309

implies a cost to the viewer, and we find that Leech's maxim of tact, tominimize cost and minimize benefits to a hearer (Leech 1983: 132) isreflected in commercials from all four countries. Suggestions referring topurchase are less common than those that stress benefits to the consumerand avoid mention of the actual exchange of money for goods. In the u. S.data, we also found an apparent trade-off between the use of theimperative, the most direct linguistic form, and the choice of the actionrecommended, a nice illustration of one of the corollaries of Leech's tactmaxim, that the more transparent the cost to the hearer in terms ofprepositional content the greater will be the need for optionality andindirectness in expression of the impositive (Leech 1983: 126).

There are a number of ways in which the commercials in our samplesreflect the economies in which they are embedded, most obviously in thedistribution of products that are advertised. Other differences among thecommercials from the four countries that we have not mentioned so farin this chapter reflect governmental regulation. These differences are lessnoticeable with respect to the head act of suggestion than to the types ofsupporting moves used, the reasons presented to the viewer to buy aproduct. One example is the use of comparative advertising, which isproscribed (more by culture and tradition than by law) in both Japan andKorea, but not in China. Only the u.S. sample contained comparativestatements in which competing brands were mentioned by name (in 6 of50 commercials). However, a wide variety of other comparative deviceswas found, including strategies that can be paraphrased as better thanothers (unspecified), the best~ nothing better, unique~ and winner. Onlythe Japanese commercials are non-comparative in this wider sense, whileChinese commercials are quite comparative, frequently mentioning thenumber of prizes a product has won in government sponsored qualitycompetitions (14 out of 50 commercials). Another aspect of commercialssubject to governmental supervision is the type of claims allowed. In theU. S., FTC regulations prohibit the making of statements that cannot besubstantiated (Geis 1982). This may account for the fact that U. S.commercials make very few explicit claims and are full of puffery, statements that sound important but that actually make few if any claims.Chinese commercials, by contrast, make very strong claims that wouldnot be allowed under U. S. advertising regulations (e. g., promising that aparticular medicinal product will cure a long list of diseases, restoreyouth, and bring the user success and prosperity). By this measure, Chineseadvertising is much more direct than advertising in the U. S., Japan, orKorea. However, the language of advertising cannot be completely a


function of the requirements of selling in a particular economy under aparticular set of government regulations. The u.s. and Japanese commercials were most different with respect to the aspects of languageexamined in this chapter, although the economies in which they aresituated are the most similar in many respects.

There are several ways in which the realization of suggestions to buyin commercials reflects the potential of the medium of television. Theclearest example of this is the use of indirection by participant shift,making suggestions directly to a viewer through an off screen voice, printed message or song lyric, having on screen actors make suggestions toeach other, and so forth. The technique of having on screen charactersdirectly model the action of purchase, found only in the U. S.commercials, may be simply an innovation in the use of the mediumwhich may spread across national and linguistic borders in time. Incommercials from each country, we also find some apparent examples ofregister-specific conventionalized language. For U. S. advertising, the highfrequency of imperatives may be partly viewed as a convention ofadvertising, since it cannot be attributed to universal requirements ofselling (imperatives are infrequent in commercials from the othercountries) or the function of suggesting versus requesting (imperatives arenot the most common form for suggestions in face-to-face interaction).The finding that the formal pronoun nin is used to address the viewer inChinese commercials conflicts with the claim of Fang and Heng (1983)that nin has been replaced almost completely by ni since the CulturalRevolution, so this may be a Chinese advertising convention.(Alternatively, the claim by Fang and Heng, for which no empiricalsources were cited, may be incorrect). Probably the best example ofconventionalized language in advertising we have found in our data isthe use of elliptical imperatives in Korean and Japanese commercials,though as we have noted, this draws upon both the linguistic resources ofthose languages (topic-comment structure, with optional deletion ofconstituents) as well as the pragmatic preference in commercials foravoiding reference to the desired action.

One of the most interesting issues for continued cross-cultural study ofthe pragmatics of advertising and the language of business in general isthe need to further elucidate relationships among the language used,universal pragmatic principles, cultural norms, and the more strictlyconventionalized aspects of advertising register. Another line of fruitfulinquiry raised by the data examined here concerns the nature of speechacts such as suggestions and requests. We do not have a fully satisfactory

Television commercials from the U. S., ]apan,China, and Korea 311

explanation for the occurrence of request-like forms in commercials fromChina, Japan, and (to a lesser extent) Korea, especially those with pleaseand formal politeness markers that are not found in U. S. commercials atall. One possible line of explanation could be that requests and suggestions may be less differentiated speech act categories in Japanese andKorean than in English. An analysis along these lines is suggested byCoulmas' (1981) observation that apologies and expressions of gratitudeare less differentiated as speech act categories in Japanese than in manyother languages. Banerjee and Carrell (1988) reported that their nonnative speakers of English (of Chinese and Malay language backgrounds)sometimes used requesting strategies inappropriately for suggestions inEnglish. However, even in English there is great overlap among the linguistic forms used for requesting and suggesting, and these categories ofspeech acts are not completely distinct at a conceptual level. There areclear cases of requests directed at actions desired by a speaker with nobenefits for a hearer, as well as suggestions in cases where only thehearer's benefit is at issue and the speaker making the suggestion has nointerest at all in whether or not the hearer carries out the action. However,the business concept of a free transaction between a willing seller and awilling buyer implies a balance between the costs and benefits accruing toboth parties. Perhaps our initial analysis, that commercials should beunderstood as suggestions, was incorrect. If commercials are moreproperly analyzed as hybrids containing elements of both request andsuggestion, then the U. S. preference for the linguistic forms typical ofsuggestions and avoidance of those typical of requests might be seen notas a reflection of the fact that commercials are suggestions in essence butas another manifestation of the maxim of tact, a manipulative strategydesigned to minimize apparent costs by using linguistic forms that areappropriate when recommending actions that are clearly to a hearer'sbenefit.

Notes

1. The videotapes of Chinese, Korean and Japanese commercials used in thisresearch were provided by Charles Keown and Lawrence Jacobs. Initial transcription of the English commercials was done by Johanna Guth. An information content analysis of a larger sample of commercials drawn from the samedatabase appears in Keown, Jacobs, Schmidt, and Gymn (1992). Usefulcomments on an earlier draft of this paper were provided by Carl James,Gabriele Kasper, Rajendra Singh, Peter Schmidt, and Keiichi Morita, none ofwhom are responsible for any weaknesses.


2. We do not discuss visual strategies in this chapter. While television is oftenconsidered to be primarily a visual medium in which language plays only asecondary role, Geis (1982) has provided ample evidence that television is noless an auditory medium than is radio.

3. An alternative to using consensus among the authors for coding decisionswould have been to have more than one rater categorize examples from eachlanguage and compute inter-rater reliability coefficients. While this wouldenhance confidence in the analysis of each individual language sample, thismethod would not guarantee comparability of coding across languages, whichwe believe was achieved better using discussion and consensus.

4. It is common for a Japanese women to address her father-in-law as ojiichan'grandpa', especially in front of her children. He is not her grandfather, but heis grandfather to her children and grandfather is seen as his primary role withrespect to the family unit.

5. An exception to the generalization that imperatives are not used in U. S.advertising with explicit reference to the exchange of money occurs inadvertisements for mail-order products (none in the sample analyzed in thischapter), at the end of which a viewer might be told: Call 1-800-xxx-xxxx,Have your credit card ready, Or send $ 9.95 plus $ 3 shipping and handlingto... [hypothetical example]. A possible explanation for these formsmay be that unless the viewer takes note of the address or phone numberwhen the commercial is being broadcast (or, as a minimum, is primed to do thiswhen next hearing the same commercial), the commercial will not fulfill itsgoaL

References

Agoston, Thomas - Walburga von Raffler-Engel1979 "A linguistic analysis of some commercial television advertisements" ,

in: Robert St. Clair (ed.), 224-242.Alatis, James E. - G. Richard Tucker

1979 Georgetown University round table on languages and linguistics1979. Washington, DC: Georgetown University Press.

Aston, Guy1993 "Notes on the interlanguage of comity", in: Gabriele Kasper

Shoshana Blum-Kulka (eds.), 224-250.Banerjee, Janet - Patricia L. Carrell

1988 "Tuck in your shirt, you squid: Suggestions In ESL", LanguageLearning 38: 313-364.

Beebe, Leslie M. - Tomoko Takahashi1989 Do you have a bag?: Social status and patterned variation in second

language acquisition. In: Susan Gass - Carolyn G. Madden Dennis R. Preston - Larry Selinker (eds.), 103-125.

Barthes, Roland1972 Mythologies. New York: Hill and Wang.


Bhatia, Tej.K.1987 "English in advertising: Multiple mixing and media", World Englishes

6: 33-48.Blum-Kulka, Shoshana - Juliane House - Gabriele Kasper (eds.)

1989 Cross-cultural pragmatics: Requests and apologies. Norwood, NJ:ABLEX.

Blum-Kulka, Shoshana - Elite Olshtain1984 "Requests and apologies: A cross-cultural study of speech act

realization patterns (CCSARP)", Applied Linguistics 5: 196-213.Bolinger, Dwight

1973 "Truth is a linguistic function", Language 49: 539-550.1980 Language: The loaded weapon. London: Longman.

Brown, Penelope - Stephen C. Levinson1987 Politeness. Cambridge: Cambridge University Press.

Cole, Peter - Jerry L. Morgan (eds.)1977 Speech acts (Syntax and semantics 3). New York: Academic Press.

Coleman, Linda1990 "The language of advertising" (review article), Journal of Pragmatics

14: 137-145Coulmas, Florian

1981 "Poison to your soul. Thanks and apologies contrastively viewed",in: Florian Coulmas (ed.), 69-91.

Coulmas, Florian (ed)1981 Conversational routine. The Hague: Mouton.

Dowling, Graham R.1980 "Information content in U. S. and Australian television advertising",

Journal of Marketing 44 (Fall): 34-37.Edmondson, Willis - Juliane House

1981 Lefs talk and talk about it. Munich: Urban & Schwarzenberg.Fang, Hanquan - J.H. Heng

1983 "Social changes and changing address norms in China", Language inSociety 12: 495-507.

Feasely, F. (ed.)1987 Proceedings of the 1987 conference of the American Academy of

Advertising.Fields, George

1983 From bonsai to Levi:Js. New York: New Yotk Library.Fraser, Bruce

1983 "The domain of pragmatics", in: Jack C. Richards - Richard W.Schmidt (eds.), 29-59.

Gass, Susan M. - Carolyn G. Madden - Dennis R. Preston - Larry Selinker (eds.)1989 Variation in second language acquisition: Discourse and pragmatics.

Clevedon: Multilingual Matters.Geis, Michael

1982 The language of television advertising. New York: Academic Press.Goffmann, Erving

1981 Forms of talk. Philadelphia: University of Pennsylvania Press.


Green, Georgia M.1975 "How to get people to do things with words: The whimperative

question", in: Peter Cole - Jerry L. Morgan (eds.), 107-141.Grice, H. Paul

1975 "Logic and conversation", in: Peter Cole - Jerry L. Morgan (eds.)41-58.

Gumperz, John J.1982 Discourse strategies. Cambridge: Cambridge University Press.

Hall, Dennis R. - Linda Saracino-Resh1979 "Advertising as cultural language",in: Robert St. Clair (ed.) 191-211.

Harris, Richard J. (ed.)1983 Information processing research. Hillsdale, NJ: Erlbaum.

Hatch, Evelyn - Anne Lazaraton1991 The research manual: Design and statistics for applied linguistics.

New York: Newbury House.Ho, Suk-ching - Yat-ming Sin

1986 "Advertising in China: Looking back and looking forward",International Journal of Advertising 5: 307-316.

Hong, Jae W. - Aydin Muderrisoglu - George M. Zinhan1987 "Cultural differences and advertising expression: A comparative

content analysis of Japanese and u.S. magazine advertising", Journalof Advertising 16: 55-68.

Hunt, Shelby D.1976 "Informational vs. persuasive advertising: An appraisal", Journal of

Advertising 5 (Summer): 5-8.Johns, Ann M.

1986 "The language of business", Annual Review ofApplied Linguistics 7:3-17.

Kaplan, Robert B.1966 "Cultural thought patterns in inter-cultural education", Language

Learning 16: 1-20.Kasper, Gabriele - Shoshana Blum-Kulka (eds.)

1993 Interlanguage pragmatics. New York: Oxford University Press.Kasper, Gabriele - Merete Dahl

1991 "Research methods in interlanguage pragmatics", Studies in SecondLanguage Acquisition 13: 215-247.

Kaynak, Erdener - Lionel A. Mitchel1981 "Analysis of marketing strategies in diverse cultures", Journal of

Advertising Research, 213: 25-32.Keown, Charles F. - Lawrence W. Jacobs - Richard W. Schmidt Kyung-Il Gymn

1992 "Information content of advertising in the United States, Japan,South Korea, and the People's Republic of China", InternationalJournal of Advertising 11: 257-267.

Laczniak, Gene R.1979 "Information content in print advertising", Journalism Quarterly 56:

324-327, 345.


Lakoff, Robin Tolmach1982 "Persuasive discourse and ordinary conversation, with examples

from advertising", in: Deborah Tannen (ed.) 25 -42.Leech, Geoffrey

1966 English in advertising. London: Longman.1983 Principles of pragmatics. London: Longman.

Madden, Charles S. - Marjorie Caballero - Sinya Matsukubo1986 "Analysis of information content in U. S. and Japanese magazine

advertising", Journal of Advertising 15.3: 38-45.Masavisut, Nitaya - Mayuri Sukwiwat - Seri Wongmontha

1987 "The power of the English language in Thai media", World Englishes5: 197-207.

Matsumoto, Yoshiko1988 "Reexamination of the universality of face: Politeness phenomena in

Japanese", Journal of Pragmatics 12: 403-426.Miracle, Gordon E.

1985 "Advertising regulations in Japan and the USA: An introductorycomparison", Waseda Business and Economic Studies 21: 335 -69.

1987 "Feel-do-Iearn: An alternative sequence underlying Japanesecommercials", in: F. Feasely (ed.) R73-R78.

Moeran, Brian1985 "When the poetics of advertising become the advertising of poetics.

Syntactical and semantic parallelism in English and Japanese advertising", Language and Communication 5: 29-44.

Mueller, Barbara1987 "Reflections of culture: An analysis of Japanese and American

advertising appeals", Journal of Advertising Research (June/July):51-59.

O'Barr, William M.1979 "Language and advertising", in: James E. Alatis - G. Richard

Tucker (eds.) 272-286.Resnik, Alan - Bruce L. Stern

1977 "An analysis of information content in television advertising",Journal of Marketing 41: 50-53.

Rice, Marshall D. - Zaiming Lu1988 "A content analysis of Chinese magazine advertisements", Journal of

Advertising 17.4: 43-48.Richards, Jack C. - Richard W. Schmidt (eds)

1983 Language and communication. London: Longman.Rintell, Ellen

1979 "Getting your speech act together: The pragmatic ability of secondlanguage learners", Working Papers on Bilingualism 17: 97-106.

Schmidt, Rosemarie - Joseph F. Kess1985 "Persuasive language in the television medium: Contrasting adver

tising and televangelism", Journal of Pragmatics 9: 287-308.1987 Television advertising and televangelism: Discourse analysis of per

suasive language. Amsterdam: Benjamins.


Searle, John R.1969 Speech acts. Cambridge: Cambridge University Press.1976 "A classification of illocutionary acts", Language in Society 5: 1-23.

Sherry, John F. - Eduardo G. Camargo1987 "May your life be marvelous: English language labeling and the

semiotics of Japanese promotion", Journal of Consumer Research14: 174-188.

St. Clair, Robert (ed)1979 Perspectives on applied sociolinguistics. Lawrence, Kansas:

Coronado Press.Stern, Bruce L. - Dean M. Krugman - Alan Resnik

1981 "Magazine advertising: Analysis of its information content", Journalof Advertising Research 21.2: 39-44.

Takahashi, Satomi1987 A contrastive study of indirectness exemplified in Ll directive speech

acts performed by American and Japanese. [Unpublished master'sthesis, The University of Illinois at Urbana-Champaign, Urbana IL]

Tannen, Deborah (ed)1982 Analyzing discourse: Text and talk~ Georgetown University round

table on languages and linguistics. Washington DC: GeorgetownUniversity Press.

Tse, David K. - Russel W. Belk - Nan Zhou1989 "Becoming a consumer society: A longitudinal and cross-cultural

content analysis of print ads from Hong Kong, the People's Republicof China, and Taiwan", Journal of Consumer Research 15: 457-472.

Vestergaard, Torben - Kim Schreder1985 The language of advertising. Oxford: Basil Blackwell.

Weizman, Elda1993 "Interlanguage requestive hints", in: Gabriele Kasper - Shoshana

Blum-Kulka (eds.) 123-137.Yamaki, Toshio

n. d. International comparison of television commercials [UnpublishedM.S.]

Culture, negotiations and international cooperativeventures

John L. Graham

1. Introduction

"You've all heard the story about the invention of copper wire - twoDutchmen got a hold of a penny." This anecdote was served up during adinner speech by the American president of a joint venture owned by anAmerican multinational company and a comparable Dutch firm. At onelevel the story is a friendly gibe, although the professor from a Dutchuniversity sitting nearby did not appreciate the American's remarks ingeneral or the ethnic joke in particular. Indeed, at another level the storyis stereotyping of the worst sort.

However, at an even deeper level there is an important lesson here forall managers and students of joint ventures and international cooperativearrangements in general. Culture can get in the way. The Americanpresident was in his "humorous" way attributing part of the frictionbetween him and his Dutch associates to differences in cultural values. Hemight have blamed personality differences or clashing "corporate"cultures, but instead he identified national cultural barriers to be a majordifficulty in joint venture management. And although I (also) do notappreciate his humor, I certainly agree that cultural differences betweenjoint venture partners and managers can cause divisive, even decisiveproblems.

Harrigan (1987) suggests that a crucial aspect of joint ventures is thenegotiation of the original agreement. The seeds of success or failure areoften sown at the negotiation table where not only financial and legaldetails are agreed to, but perhaps more importantly, the ambiance ofcooperation is established. Indeed, as Harrigan indicates, the legal detailsand the structure of joint ventures are almost always modified over time,and usually through negotiations. But the atmosphere of cooperationestablished at the negotiation table persists or the venture fails.

The purpose of this chapter is to present selected results from aprogram of research investigating differences in cultural styles of business

318 John L. Graham

negotiations. The study has involved more than 1000 business people inseventeen countries and cultures. The analyses reported below comprisesome of the most interesting findings of the project. Other results arereported in a series of complementary articles (cf., Graham 1980; 1983;1985 a; 1985 b; 1992). Here six business people from each of fourteencountries were videotaped during simulated intracultural negotiations.The content of their negotiation strategies and linguistic structuralaspects of their conversations were analyzed. Our findings suggest thatsubstantial differences exist in negotiation styles across the thirteencountries. Further, it is our supposition that such differences can causefriction, suspicion, and even failure in otherwise mutually beneficial international joint ventures.

The remainder of this chapter is organized into three sections. First,the theoretical perspective is briefly discussed. Next, methods of datacollection are described. Finally, the results are summarized, conclusionsare drawn, and hypotheses for future testing are suggested.

2. Theoretical perspective

Despite the increasing importance of cross-national commercial relationships, business negotiations in different countries have received littleattention. During the 1970s, a few articles appeared in business journals(for example, Jastram 1974; Kapoor 1974; Van Zandt 1970; Wells1972), but they were primarily descriptive and often anecdotal. Recently,more systematic studies of negotiations in foreign countries have beenundertaken. Tung (1982) considered business negotiations betweenAmerican and Chinese executives. Harnett and Cummings (1980) compared bargainers' characteristics and behaviors across several culturalgroups. Graham, Mintu and Rodgers (1994), investigated the determinants of business negotiation outcomes in the United States, and tenforeign countries. Weiss has provided in-depth reports of case studies ofmajor international business negotiations (1987; 1990). Francis (1991)has considered the importance of adaptation in international businessnegotiations. These studies have proven valuable, but most are limited intheir use of questionnaire items as measures of negotiation processes.

Most recently, the outcomes of business negotiations have beenhypothesized to be the result of several factors that can be classified intothree categories or kinds of theoretical constructs - individual

Culture, negotiation~ and international cooperative ventures 319

characteristics, situational constraints, and process measures (see Rubin~ Brown 1975; Sawyer - Guetzkow 1965). Many empirical measuresof both individual characteristics and situational constraints have beentested in previous research. Examples of such individual characteristicsmight be intelligence, self-esteem, credibility, attractiveness, and culturalbackground. Examples of situational constraints might include companygoals, time limitations, or unequal power relations.

2.1. Process measures

A few studies have fo'cused on the process of business negotiations (forexample, Dwyer - Walker 1981; Lewis - Fry 1977; Pennington 1968;Pruitt - Lewis 1975). Graham (19.83: 82) has defined process measuresas "qualitative and quantitative descriptions of the activities involved ina business negotiation for example, bargaining strategies." Based on anextensive review of the negotiation literature, Rubin and Brown (1975)conclude that the behaviors of bargainers during the negotiation processaffect negotiation outcomes. The kinds of behaviors they list are openingmoves, countermoves, types of appeals, demands, and the like. But littlework has been done to investigate relationships among process measuresand negotiation outcomes, individual characteristics and situational constraints.

Moreover, because researchers have only recently turned their attention to process measures, concepts and operational definitions remainvague and 'relations not adequately specified. The present study focuseson the development of operational definitions of process measures usingobservational methods. Further, special attention is given to the influenceof national culture on these process measures.

2.2. Cultu~al differences in business negotiation processes

Culture has been a difficult concept to deal with in any consistent way.Anthropologists and sociologists have been arguing over definitions foryears. Culture has appeared in the marketing literature primarily as adeterminant of consumer behavior (for example, Engel - Blackwell1982)~ but operational definitions seem to have varied from study tostudy. Perhaps the most widely accepted definition is that professed byLinton: "A culture is a configuration of learned behaviors and results of

320 John L. Graham

behavior whose component parts are shared and transmitted bymembers of a particular society" (Linton 1945: 5). The important partof the definition for the present research is the idea that behaviors areshared by members of a particular culture. Or as Spiro put it, "membersof a given society behave in uniform and predictable ways" (Spiro 1950:20). A central goal of the study is to discover what shared behaviorsmanifest themselves during business negotiations in the thirteencountries.

In addition to the bargaining behaviors being consistent withincultures, several authors have suggested that negotiation processes differacross cultures (for example, Condon 1974; Frake 1972; Kay 1970;Sawyer - Guetzkow 1965; Van Zandt 1970). Therefore, a secondpurpose of this work is to identify how bargaining processes in severalcountries might differ from one another.

2.3. Content versus context

Aside from the cultural differences in negotiation behaviors discovered,perhaps the most important implication of the study regards the "contentversus context" issue. Social psychologists have focused on the verbalcontent of negotiation in their research. Alternatively, linguistic theoryholds that consideration of only verbal content yields inadequate understandings of interpersonal interactions. Sociolinguists emphasize theimportance of the context of communication of nonverbal and structuralaspects of language. Our results suggest that the linguists are correct.

Simply stated, the content of conversation is what is said, while thecontext is how it is said. The distinction is both theoretically andpractically a "fuzzy" one. Several researchers have developed schemesfor categorizing the what aspects of negotiations (e.g., Angelmar - Stern1978; Bales 1950; Bonoma - Felder 1977; Pennington 1968; PruittLewis 1975; Walton - McKersie 1965), and used these schemes toanalyze the verbal content of bargaining interactions. More recently, thehow of meaning has also been considered. Take, for example, the interaction described and interpreted in The Wall Street Journal:

The japanese executive sucks in air through his teeth and exclaims, "Sa!That will be very difficult!" What he really means is just plain "no." Butthe japanese consider an absolute "no" to be offensive and usually seek aeuphemistic term. That's why in japan, the "difficult" really may be impossible. The American on the other side of the negotiation table knows none

Culture, negotiations and international cooperative ventures 321

of this and presses ahead to resolve the "difficulty." The Japanese finds thisinexplicable persistence to be abnormally pushy. The atmosphere deteriorates, and sure enough, the big deal falls through (Ricklefs 1978: 4).

How do we understand the meaning of the word "difficult?" Does itmean "no" or "maybe?" Ethnomethodologists emphasize the importanceof context as well as content for establishing a shared meaning ofcommunication. The idea is that communication must be considered asan integrated whole, content and context; and context has often been"taken for granted" in previous negotiation studies.

Certainly the reality of any particular situation provides much of thecontext for making decisions about meaning. And so does all previouscommunication between actors. Gumperz (1979) has posited thathumans, in the course of interaction, also indicate context for interpretation of verbal communications through the use of contextualizationcues. He explains:

Our hypothesis is that conversational inference, i. e., the process by whichspeakers interpret what is intended by a conversational contribution, is inpart determined by a system of conventional discourse-level verbal andnon-verbal signals. These signals, termed "contextualization cues," serve tosignal the way in which any conversational contribution is to be understood, in light of the participants' expectations and the situation at hand(Gumperz 1979: 2).

An example of a contextualization cue might be a rise in tone of voice toindicate or underline an important point. Gumperz and his associateshave also found that contextualization cues vary across cultures. They arebehaviors learned in the course of the individuals' socialization. Further,he suggests that these differences are often the cause of misunderstandings that can have serious consequences (e.g., failed negotiations) incross-cultural interactions.

Authors in other fields also emphasize the importance of context. Forexample, Bonoma and Felder (1977) and Soldow and Thomas (1984)offer alternative definitions of context, non-verbal behaviors and relational communication, respectively. Hall (1976), Cateora (1983), andGraham (1987) argue that the influence of context varies across cultures.Cateora states, "communication in a high-context culture dependsheavily on the context or non-verbal aspects of communication, whereasthe low-context culture depends more on explicit, verbally expressedcommunication" (1983: 133). In the present study both content andcontext are considered.

322 John L. Graham

3. Methods

3.1. Sample

The participants in the experiment are six business people from each ofthe fourteen cultures (japanese, Korean, Taiwanese, Chinese, Russian,German, English, French, Spanish, Brazilian, Mexican, French Canadian,English Canadian, and American (D. S. )). All were participating inexecutive or Master's of Business Administration (MBA) programs andall volunteered for this study. All received college educations and all havehad at least two years' business experience in their respective countries.The sample was limited to experienced business people because Fourakerand Siegel (1963) reported differences in the bargaining behavior ofstudents and businesspeople.

3.2. Laboratory setting

The negotiation simulation, developed by Kelley (1966) and used byPruitt and Lewis (1975) and Lewis and Fry (1977) involves bargainingover three related issues. Differing amounts and types of backgroundinformation can be included with the basic pay-off matrices, dependingon the focus of the research. The simulation is simple enough to belearned quickly, but complex enough to provide usually one half hour offace-to-face interaction. Forty-two negotiations were conducted - threefor each group..

3.3. Data collection

The forty-two interactions were videotaped using a wide-angle perspective to capture postures, body movements, and interpersonal distances.Participants were asked to evaluate the obtrusiveness .of the setting onquestionnaires followirig the .negotiation game and they reported a minimum of discomfort.

3.4. Verbal behaviors

As mentioned, a primary purpose in this exploratory work is the identification and clarification of process measures. Consequently, the discussion


in the sections to follow are organized as a "list" of process measures.Associated with each item on the "list" are operational definitions, a briefaccount of the method of measurement, and mention of apparentdifferences among the thirteen cultural groups.

The first step in the measurement and analysis of verbal behaviorsduring the business negotiations is the transcription of the audio portionof the videotapes. This is a potential source of error in measurement. Acomplete check of the transcript revealed some minor mistakes, and thesewere corrected.

The second step in the measurement and analysis process consisted oftranslation of the non-English interactions. With one exception nativespeakers of the foreign languages were instructed to "make the translations as literal as possible while still communicating the intendedmeaning." The quality of the English grammar, etc. was not the primaryconsideration. The exception to the native speaker rule regarded theRussian translation, wherein a 20 year American resident of Moscow didthe work.

3.5. Content analysis

Angelmar and Stern (1978) have described a content analysis schemedeveloped specifically for the analysis of bargaining communications inbusiness ·settings. Utterances by participants are classified into twelvecategories. The categories and definitions are listed in Table 1. Angelmarand Stern report positive results from a reliability and validity assessmentof the system applied to written communications. The present study isone of the few to apply the scheme to transcripts of conversations.Coding transcribed conversations is a more difficult undertaking; spokenwords are the only channel of communication. Transcripts do not include

. information communicated through other channels such a proxemics,prosody, kinetics, or facial expression. Theory indicates that thesechannels also may be important for accurate interpretation and measurement of conversational contributions.

Two coders were employed in classifying segments of the conversationinto twelve bargaining categories. The author coded all forty-two interactions and research assistants (ignorant of the theory and hypothesesinvolved in the study) coded three interactions to provide a reliabilitycheck. The author is cognizant of the possible biases involved in usingcoders informed about the theory applied in the research. However,

324 John L. Graham

Table 1. Verbal negotiation tactics (The "what" of communications)

Bargaining Behaviors and Definitions Cultures(Anglemar and Stern, 1978) (in each group, n = 6)

JPN KOR TWN

Promise. A statement in which the source indicated his 7* 4 9intention to provide the target with a reinforcingconsequence which source anticipates target willevaluate as pleasant, positive, or rewarding

Threat. Same as promise, except that the reinforcing 4 2 2consequences are thought to be noxious, unpleasant,or punishing.

Recommendation. A statement in which the source 7 1 5predicts that a pleasant environmental consequencewill occur to the target. Its occurrence is not undersource's control

Warning. Same as recommendation, except that the 2 0 3consequences are thought to be unpleasant.

Reward. A statement by the source that is thought to 1 3 2create pleasant consequences for the target.

Punishment. Same as reward, except that the 1 5 1consequences are thought to be unpleasant.

Positive normative appeal. A statement in which the 1 1 0source indicates that the target's past, present, or futurebehavior was or will be in conformity with social norms.

Negative normative appeal. Same as positive normative 3 2 1appeal except that the target's behavior is in violation ofsocial norms.

Commitment. A statement by the source to the effect 15 13 9that its future bids will not go below or above acertain level.

Self-disclosure. A statement in which the source reveals 34 36 42information about itself.

Question. A statement in which the source asks thetarget to reveal information about itself. 20 21 14

Command. A statement in which the source suggests thatthe target perform a certain behavior. 8 13 11

)!- Read "7% of the statements made by Japanese negotiators were promises."a northern China (Tianjin and environs)

Culture~ negotiations and international cooperative ventures 325

Table 1 (contd.)

CHNa RUSS GRM UK FRN SPN BRZ MEX FCAN ECAN USA

6

1

2

5

3

4

7

3

5

11 5

3 5

6 3

11

2

4

3

2

5

7

1

8

8

3

5

6

o

4

8

4

4

1 0 1 1 3 1 1 2 5 0 1

1 3 4 5 3 3 2 1 1 3 2

0 1 2 0 3 2 3 0 2 1 3

1 0 0 0 0 0 0 0 1 0 1

0 0 1 1 0 1 1 1 2 1 1

10 11 9 13 10 9 8 9 8 14 13

36 40 47 39 42 34 39 38 42 34 36

34 27 11 15 18 17 22 27 19 26 20

7 7 12 9 9 17 14 7 5 10 6

326 John L. Graham

resource constraints necessitate this less-than-ideal state. Significantly,analysis of discrepancies in coding between the two coders revealed thissource of bias to be minimal. Intercoder reliability was 63 %, comparableto Angelmar and Stern (1978) wherein they report 66 % agreement forcoding written negotiations.

The data presented in Table 1 represent the percentage of eachcategory of behavior used averaged across the six negotiators ineach country. For example, the number of promises used by a singleAmerican negotiator was divided by the total behavior coded forthat negotiator, then the average percentage of promises acrossthe six American negotiators was calculated and reported in the upperleft corner of Table 1. Such a procedure allows for comparisonsacross the cultural groups controlling for differences in time spentnegotiating.

3.6. Structural Aspects ("no" and "you")

Graham (1985b) suggests that the simple counting of these two wordsmay shed light on subtle differences in cultural styles of persuasion. Hefound substantial differences between the frequency of the use of theword "no" by Brazilian bargainers as opposed to American andJapanese. Several authors (e.g., Nakane 1970; Veda 1974; Van Zandt1970) indicate that Japanese negotiators seldom use the word "no"during negotiations. Graham (1985 b) also notes a Brazilian propensity tospeak more frequently in the second person using the pronoun "you."The number of times each word was used was tallied for each negotiatorthen multiplied by the time of the negotiation in minutes and then dividedby thirty minutes to provide a frequency measure which might be compared across the various groups. Intercoder reliability (calculatedusing Guetzkow's 1960 formula for marginal reliability, the differencein the number of units between coders as a percentage of the sumof the units) was calculated for three of the interactions and foundto be 1 %.'

3.7. Nonverbal behaviors

In this section of explorations into the process of buyer-seller negotiations' nonverbal aspects of the videotaped interactions were considered.


First the rhythm of the conversations will be discussed, specificallyexamining silent periods and conversational overlaps. Next, gaze direction of the participants will be considered. Lastly, findings related totouching during negotiations are presented. All measurements in thissection have been derived irrespective of the verbal content of the interactions to avoid potential bias, that is, the tapes could be coded withoutknowledge of the language spoken. Reliability of the coding wascalculated for all the nonverbal behaviors by having a secondassistant code three interactions using a marginal reliability approach(difference in the tallies of the two coders divided by the sum of thetwo coders). These numbers are reported at the end of each section tofollow.

3.8. Conversational coordination

Communication theory suggests that when two people are effectivelysharing ideas, their communication behaviors - both verbal and nonverbal-will be rhythmically coordinated (Condon 1968; Erickson 1976;Gumperz 1979). Here two measures of conversational coordination,"silent periods" and "conversational overlaps," are operationallydefined, and findings are reported below.

(1) Silent Periods. Silent periods are defined as gaps in conversationsten seconds or more in duration. The time period of ten secondswas selected somewhat arbitrarily, but it is a long enough period ofsilence to appear unnatural to most American observers. The tapeswere searched for gaps in conversations of 10 seconds or more, andthese gaps were noted on the transcripts and tallied (see Table 2).Once again the frequency of occurrence was calculated by multiplyingthe number of silent periods by the duration at each negotiation dividedby thirty minutes (marginal reliability < 1 0/0).

(2) Conversational Overlaps. The concept of "interactional sychrony,"the unconscious coordination of verbal and nonverbal behaviors of twoor more participants in a conversation, is discussed at length by Graham(1980). One possible measure of this construct is the number ofconversational overlaps or interruptions during a conversation.Conversational overlaps are defined here as periods when both speakersare talking simultaneously, or when the conversational contribution ofone speaker overlaps that of the other speaker. Identification of suchoverlaps is independent of the verbal content of the interactions. In the

328 John L. Graham

Table 2. Linguistic aspects of language and nonverbal behaviors ("How" thingsare said)

Bargaining Behaviors (per 30 minutes) Cultures(in each group, n =6)

JPN KOR TWN

Structural Aspects

"No's." The number of times the word "no" wasused by each negotiator.

"You's." The number of times the word "you"was used by each negotiator.

Nonverbal Behaviors

Silent Periods. The number of conversational gapsof 10 seconds or longer.

Conversational Overlaps. Number of interruptions.

Facial Gazing. Number of minutes negotiatorsspent looking at opponent's face.

Touching. Incidents of bargainers touching one another(not including handshaking).

a northern China (Tianjin and environs)

1.9

31.5

2.5

6.2

3.9

o

7.4

34.2

o22.0

9.9

o

5.9

36.6

o12.3

19.7

o

present work, the videotapes were searched for overlaps, and such interruptions in the flow of conversation were noted on the transcripts.Frequencies were calculated as above and reported in Table 2 (marginalreliability =10%).

3.9. Facial gazing

The third nonverbal variable to be considered is facial gazing. Otherresearchers have found significant relationships between facial gazing andoutcomes of negotiations (Lewis - Fry 1977). Moreover, several authorshave suggested differences in facial gazing behavior across cultures(Argyle - Cook 1976).

In this study, facial gazing is defined as the percentage of timea bargainer gazes at the face of his opponent. Ten-minute videotape excerpts of each of the forty-two interactions served as data


Table 2 (contd.)

CHNa RUSS GRM UK FRN SPN BRZ MEX FCAN ECAN USA

1.5 2.3 6.7 5.4 11.3 23.2 41.9 4.5 7.0 10.1 4.5

26.8 23.6 39.7 54.8 70.2 73.3 90.4 56.3 72.4 64.4 54.1

2.3

17.1

3.7 0

13.3 20.8

2.5 1.0 0 0 1.1

5.3 20.7 28.0 14.3 10.6

0.2

24.0

2.9

17.0

1.7

5.1

11.1 8.7 10.2 9.0 16.0 13.7 15.6 14.7 18.8 10.4 10.0

o o o o 0.1 0 4.7 o o o o

here. Using a stopwatch, two observers recorded the time eachparticipant spent gazing at his opponent's face. The method usedwas very similar to that reported by Lewis and Fry (1977), except thathere videotapes were reviewed rather than real-time interactions. Usingvideotapes is a more reliable technique, allowing reviews and reliabilitychecks.

3.10. Touching

Finally, the number of times a negotiator touched a partner (excludingbeginning and ending handshaking) was recorded for each interaction(marginal reliability = 0 %).

330 John L. Graham

4. Discussion

4.1. Results

The results from the analyses are presented in Tables 1 and 2. The unexpected similarities among the fourteen groups are perhaps more strikingthan the differences. Particularly with regard to Anglemar and Stern's(1978) content analysis scheme, negotiation styles appear to be surprisingly consistent across the fourteen cultural groups. Negotiations inall cultures studied are composed primarily of information exchangetactics - questions and self-disclosures.

4.2. Nonverbal behaviors

Reported in Table 2 are the analyses of some linguistic aspects and nonverbal behaviors for the fourteen videotaped groups, as in Graham(1985 b). While our efforts here merely scratch the surface of these kindsof behavioral analyses, they still provide indications of substantialcultural differences. Note that the japanese are at, or next to, the end ofalmost every dimension of the behaviors listed in Table 2. Their facialgazing and touching are the least among the fourteen groups. Only thenorthern Chinese used the words "no" less frequently and only theRussians used more silent periods than did the japanese.

A broader examination of the data in Tables 1 and 2 reveals a moremeaningful conclusion. That is, the variation across cultures is greaterwhen comparing linguistic aspects of language and nonverbal behaviorsthan when the verbal content of negotiations is considered. For example,notice the great differences between japanese and Brazilians in Table 1vis-a-vis Table 2.

Following are further descriptions of the distinctive aspects of each ofthe fourteen cultural groups we have videotaped. Certainly, we cannotdraw conclusions about the individual cultures from an analysis of onlysix business people in each, but the suggested cultural differences areworthwhile to consider briefly.

Japan. Consistent with most descriptions of japanese negotiationbehavior in the literature, the results of this analysis suggest their style ofinteraction to be the least aggressive (or most polite). Threats, commands, and warnings appear to be deemphasized in favor of the more


posItIve promises, recommendations, and commitments. Particularlyindicative of their polite conversational style was their infrequent use of"no" and "you" and facial gazing, as well as more frequent silentperiods.

Korea. Perhaps one of the more interesting aspects of this study is thecontrast of the Asian styles of negotiations. Non-Asians often generalizeabout the Orient. Our findings qemonstrate that this is a mistake. Koreannegotiators used considerably more punishments and commands than didthe japanese. Koreans used the word "no" and interrupted more thanthree times as frequently as the japanese. Moreover, no silent periodsoccurred between Korean negotiators.

China (northern). The behaviors of the negotiators from northernChina (i. e., in and around ·Tianjin) are most remarkable in the emphasison asking questions (34 0/0). Indeed, 70 % of the statements made by theChinese negotiators were classified as information exchange tactics.Other aspects of their behavior were quite similar to the japanese - theuse of "no" and "you" and silent periods.

Taiwan. The behavior of the business people in Taiwan was quitedifferent from that in China and Japan but similar to that in Korea. TheChinese on Taiwan were exceptional in the time of facial gazing, on theaverage almost 20 'out of 30 minutes. They asked fewer questions andprovided more information (self-disclosures) than did any of the otherAsian groups.

Russia. The Russians' style was quite different from that of any otherEuropean group, and, indeed, was quite similar in many respects to thestyle of the Japanese. They used "no" and "you" infrequently and usedthe most silent periods of any group. Only the japanese did less facialgazing, and only the Chinese asked a greater percentage of questions.

Germany. The behaviors of the western Germans are difficult tocharacterize because they fell toward the center of almost all of thecategories. However, the Germans were exceptional in the high percentage of self-disclosures at 47% and the low percentage of questions at11 0/0.

United Kingdom. The behaviors of the British negotiators are remarkably similar to those of the Americans in all respects.

I Spain. Diga is perhaps a good metaphor for the Spanish approach tonegotiations evinced in the data. When you make a phone call in Madrid,the usual greeting on the other end is not hola ('hello') but is, instead,diga ('speak'). The Spaniards likewise used the h~ghest percentage ofcommands (17 %) of any of the groups and gave comparatively little

332 John L. Graham

information (self-disclosures, 34 0/0). Moreover, they interrupted oneanother more frequently than any other group, and they used the term"no" and "you" very frequently.

France. The style of the French negotiators is perhaps the most aggressive of all the groups. In particular, they used the highest percentage ofthreats and warnings (together, 80/0). They also used interruptions, facialgazing and "no" and "you" very frequently compared to the othergroups, and one of the French negotiators touched his partner on the armduring the simulation.

Brazil. The Brazilian businesspeople, like the French and Spanish,were quite aggressive. They used the highest percentage of commands ofall the groups. On average, the Brazilians said the word "no" 42 times,"you" 90 times, and touched one another on the arm about 5 timesduring 30 minutes of negotiation. Facial gazing was also high.

Mexico. The patterns of Mexican behavior are good reminders of thedangers of regional or language-group generalizations. Both verbal andnonverbal behaviors are quite different than those of their LatinAmerican (Brazilian) or continental (Spanish) cousins. Indeed, Mexicansanswer the telephone with the much less demanding bueno. In manyrespects, the Mexican behavior is very similar to that of the negotiatorsfrom the United States.

Francophone Canada. The French-speaking Canadians behavequite similarly to their continental cousins. Like the negotiators fromFrance, they, too, used high percentages of threats and warnings, andeven more interruptions and eye contact. Such an aggressive interactionstyle would not mix well with some of the more low-key styles of someof the Asian groups or with English speakers, including AnglophoneCanadians.

Anglophone Canada. The Canadians who speak English as their firstlanguage used the lowest percentage of aggressive persuasive tactics (thatis, threats, warnings and punishments totaled only 1 %) of all fourteengroups. Perhaps, as communications researchers suggest, such stylisticdifferences are the seeds of interethnic discord as witnessed in Canadaover the years. With respect to international negotiations, theAnglophone Canadians used noticeably more interruptions and "no's"than negotiators from either of Canada's major trading partners, theUnited States and Japan.

United States. Like the Germans and the British, the Americans fell inthe middle of most categories. They did interrupt one another less frequently than all the others, but that was their sole distinction.


These differences across cultures are quite complex. The key here is tobe aware of these kinds of differences so one doesn't misinterpret theJapanese silence, the Brazilian "no, no, no ... ", or the French threat.

4.3. Content versus context

The findings of this study clearly suggest that our understanding ofnegotiation processes is incomplete and perhaps inadequate if we relysolely on analyses of verbal content. As Gumperz (1979) and otherssuggest, the context of communication is crucial. Indeed, at the level ofAngelmar and Stern's (1978) content analysis, it is most difficult todistinguish between cultural styles of bargaining. However, considerationof both the structural aspects of language and nonverbal behaviors yieldssubstantial differences among the groups. That is, cultural background ofthe negotiators affects the "contextual" more than the content-relatedaspects of the negotiation process. This is consistent with the findingsof Neu and Graham (in press) who report that context variables havestronger influences on negotiation outcomes than do contentvariables.

5. Conclusions

These results, of course, are not definitive. The small sample sizes do notallow for tests of statistical significance. How representative the participants are is problematic. They are all experienced business people andcitizens and permanent residents in each of their countries, which is animprovement over most other business negotiation research wherestudents are used as surrogates for bargainers. But how well six businesspeople represent a "cultural style" cannot be determined. Finally,external validity of the experimental setting is questionable. However,this work represents an improvement over most other business negotiation research by the use of face-to-face communication instead of writtenor electronic means. The value and strength of this study are theobservational methods used to measure the negotiation process.Videotaping allows for multiple observers and multiple observationsconcurrent with the bargaining process. Thus, the reliability and validityof the process measures developed does not depend on a priori

334 John L. Graham

experimental manipulations or post hoc partIcIpant self-reports. Themethods developed in this study are time consuming and expensive but,as these findings suggest, potentially fruitful.

Cross-cultlJ.ral interactions were not explicitly considered here.Graham (1980; 1985b) and Adler and Graham (1989) report findingsfrom such studies. Generally, the findings indicate that such culturaldifferences in bargaining process as described above are potential sourcesfor friction alfd misunderstandings between bargainers that often resultin increased transaction costs in international commercial relationships.For example, frequent interruptions of American negotiators by Braziliancounterparts can lead to irritation and to inaccurate attributions ofrudeness when Brazilian executives are just conforming to Braziliannorms for interactions. Likewise, lack of eye contact from Japanesepartners during negotiations' may lead to Americans' suspicions andattributions of Japanese secrecy or even dishonesty. And such problemscan destroy cooperative relationships and preclude otherwise mutuallybeneficial commercial agreements.

Perhaps the most worrisome of our findings is that greater differenceswere related to how things were said than to what was said. That is, thenegotiation styles of the fourteen cultural groups are surprisingly similarat the level of content: The clearest contrasts ~etween groups were foundin structural and nonverbal aspects of conversational styles. Suchdifferences are generally not consciously perceived by negotiators. These"hidden" problems lead not only to ethnic jokes, but worse yet to crosscultural disharmony, prejudices, and perceptions and feelings of ill will.Thus, that necessary condition for joint venture success·- an ambianceof cooperation - can be lost for no apparent reason other than culturalmisunderstanding.

The findings of this study suggest that substantial differences in bargaining styles exist across cultures. Reliable and valid measures ofnegotiation processes have been developed. This exploratory work deserves follow-up research with larger sample sizes. Increased statisticalpower associated with larger samples would allow for investigations ofnot only the culture ---7 process relationship but also more complex relations such as culture ---7 process ---7 negotiation outcomes. The findings ofsuch studies will hold important implications for training businessexecutives and students to manage more efficiently the internationalrelationships of the future, increasingly taking place in a more globalmarketplace.


References

Adler, Nancy J. - John L. Graham1989 "Cross-cultural interaction: The international comparison fallacy",

Journal of International Business Studies Fall: 515-537.Angelmar, Reinhard - Louis W. Stern

1978 "Development of a content analytic system for analysis of bargaining communication in marketing", Journal of MarketingResearch 15: 93-102.

Archibald, Kathleen (ed.)1966 Strategic interaction and conflict. Berkeley Institute of International

Studies: University of California Press.Argyle, Michael - Mark Cook

1976 Gaze and mutual gaze. Cambridge: Cambridge University Press.Bales, Robert F.

1950 Interaction process analysis: A method for the study ofsmall groups.Cambridge, MA: Addison-Wesley Press.

Bonoma, Thomas V. - Leonard C. Felder1977 "Nonverbal communication in marketing: Toward a communica

tional analysis", Journal of Marketing Research 14: 169-180.Campbell, Nigel - Alain Jolibert - Hans-Gunther Meissner - John L. Graham

1988 "A comparison of marketing negotiations in France, Germany, theUnited Kingdom, and the United States", Journal of Marketing 52:49-62.

Cateora, Philip R.1983 International marketing. (5th edition.) Homewood, IL: Richard D.

Irwin.Condon, John C.

1968 "Linguistic-kinesic research and dance therapy", A.D.1:A. convention proceedings.

1974 "Introduction: A perspective for the conference", in: John C.Condon - M. Saito (eds.), 3-14.

Condon, John C. - Mitsuko Saito (eds.)1974 Intercultural encounters in Japan. Tokyo: Simul Press.

Dwyer, Robert F. - Orville C. Walker1981 "Bargaining in an asymmetrical power structure", Journal of

Marketing 45: 104-115.Ehlich, Konrad - Johannes Wagner

in press The discourse of international negotiations. Berlin: Mouton deGruyter.

Engel, James F. - Roger D. Blackwell1982 Consumer behavior. Chicago: The Dryden Press.

Erickson, Fred1976 "Talking down and giving reasons: Hyper-explanation and listening

behavior in interracial interviews", paper delivered at the International Conference on non-verbal behavior, Ontario Institute forStudies in Education, Toronto, Canada, May 11, 1976.

336 John L. Graham

Fouraker, Lawrence E. - Sidney Siegel1963 Bargaining behavior. New York: McGraw-Hill.

Frake, Charles o.1972 "Struck by speech: The Yakan concept of litigation", in: John J.

Gumperz - Dell Hymes (eds.), 106-129.Francis, June N.P.

1991 "When in Rome? The effects of cultural adaptation on interculturalbusiness negotiations", Journal of International Business Studies22.3: 403-428.

Graham, John L.1980 Cross-cultural sales negotIatIons: A multilevel analysis. [Un

published Ph. D. dissertation, School of Business, University ofCalifornia, Berkeley.]

1983 "Brazilian, Japanese, and American business negotiations",Journal of International Business Studies Spring/Summer:81-96.

1985 a "The influence of culture on the process of business negotiations: Anexploratory study", Journal of International Business Studies Spring:81-94.

1985 b "Cross-cultural sale negotiations: A laboratory experiment",Marketing Science 4: 130-146.

1987 "Deference given the buyer: Variations across twelve cultures", inPeter Lorange - Farok Contractor (eds.), 473-486.

Guetzkow, Harold1960 "Unitizing and categorizing problem in coding qualitative data",

Journal of Clinical Psychology 6: 45 -53.Gumperz, John J.

1979 "Sociocultural knowledge in conversational inference", 28th AnnualRoundtable Monograph Series on Languages and Linguistics,Georgetown University.

Gumperz, John J. - Dell Hymes (eds.)1972 Directions in sociolinguistics. New York: Holt, Rinehart and

Winston.Hall, Edward T.

1976 Beyond culture. Garden City, NY: Anchor Press/Doubleday.Harrigan, Katherine R.

1987 "Why joint ventures fail", Euro-Asia Business Review 6.3:20-26.

Harnett, Donald L. - Larry L. Cummings1980 Bargaining behavior: An international study. Houston, TX: Dane

Publications, Inc.Jastram, Rothdow W.

1974 "The Nakado negotiators", California Management Review 17.2:86-90.

Kapoor, Ashok1974 "MNC negotIatIons: Characteristics and planning implications",

Columbia Journal of World Business 9: 121-130.


Kay, Paul1970 "Some theoretical implications of ethnographic semantics", Current

Directions in Anthropology Bulletins of the American Anthropological Association 3:2.

Kelley, Harold H.1966 "A classroom study of the dilemma in interpersonal negotiations",

in: Kathleen Archibald (ed.), 49-73.Kelman, Herbert C. (ed.)

International behavior: A social-psychological analysis. New York:Holt, Rinehart and Winston.

Lewis, Steven A. - William R. Fry1977 "Effects of visual access and orientation on the discovery of inte

grative bargaining alternatives", Organizational Behavior andHuman Performance 20: 75-92.

Linton, Ralph1945 The cultural background of personality. New York: Appleton

Century-Crofts.Lorange, Peter - Farok Contractor (eds.)

1987 Cooperative strategies in international business. Lexington, MA:Lexington Books.

Nakane, Chie1970 Japanese society. Berkeley, CA: University of California Press.

Neu, Joyce - John L. Grahamin press "An analysis of language use in negotiations: The role of context and

content", in: Konrad Ehlich - Johannes Wagner (eds.).Pennington, Allan L.

1968 "Customer-salesman bargaining behavior in retail transactions",Journal of Marketing Research 5: 255-262.

Pruitt, Dean G. - Steven A. Lewis1975 "Development of integrative solutions in bilateral negotia

tion", Journal of Personality and Social Psychology 31.4:621-633.

Ricklefs, Richard1978 "For a businessman headed abroad, some basic lessons", The Wall

Street Journal. January 16: 4.Rubin, Jeffrey Z. - Bert R. Brown

1975 The social psychology of bargaining and negotiation. New York:Academic Press.

Sawyer, Jack - Harold Guetzkow1965 "Bargaining and negotiation in international relations", in: Herbert

C. Kelman, (ed.), 464-520.Soldow, Gary F. - Gloria Penn Thomas

1984 "Relational communication: Form versus content in the sales interaction", Journal of Marketing. 48: 84-93.

Spiro, Melford E.1950 "Culture and personality: The natural history of a false dichotomy",

Psychiatry 14.1: 19-46.

338 John L. Graham

Tung, Rosalie L.1982 "D. S.-China trade negotiations", Journal of International Business

Studies Fall: 25 -38.Deda, Keiko

1974 "Sixteen ways to avoid saying 'no' in japan", in: john C. Condon Mitsuko Saito (eds.), 185-192.

Van Zandt, Howard F.1970 "How to negotiate in japan", Harvard Business Review November

December: 45-56.Walton, Richard E. - Robert B. McKersie

1965 A behavioral theory of labor negotiations. New York: McGraw-Hill.Weiss, Stephen E.

1987 "Creating the GM Toyota joint venture: A case in complex negotiation", Columbia Journal of World Business 22: 23-37.

1990 "The long path to the IBM Mexico agreement An analysis of themicrocomputer investment negotiations, 1983-86", Journal ofInternational Business Studies 21.4: 565 -96.

Wells, Louis T.1972 "Negotiating with third world governments", Harvard Business

Review january-February: 72-80.

List of Contributors

Mahmoud Al BatalDepartment of Near Eastern StudiesEmory UniversityAtlanta, GA 30322

Leslie M. BeebeProgram in Applied LinguisticsTeachers College525 West 120the StreetColumbia UniversityNew York, NY 10027

Jean "U7. BodmanAmerican Training Institute and School of EducationDepartment of Teaching and Learning239 Green StreetRoom 635 East BuildingWashington SquareNew York UniversityNew York, NY 10003-6674

Diana BoxerProgram in LinguisticsUniversity of Florida112 Anderson HallPO Box 115454Gainesville, FL 32611

Mary CarpenterSchool of EducationDepartment of Teaching and Learning239 Green StreetRoom 635 East BuildingWashington SquareNew York UniversityNew York, NY 10003-6674

340 List of Contributors

Andrew CohenInstitute of Linguistics & Asian & Slavic Languages & Literatures196 Klaeber CourtUniversity of Minnesota320 16th Avenue SEMiineapolis, MN 55455

Martha Clark CummingsMonterey Institute of International StudiesDepartment of TESOL425 Van Buren StreetMonterey, CA 93940

Miriam Eisenstein EbsworthSchool of EducationDepartment of Teaching and Learning239 Green StreetRoom 635 East BuildingWashington SquareNew York UniversityNew York, NY 10003-6674

Waguida El BakaryAmerican University in CairoPO Box 2511Cairo Egypt

Susan M. GassEnglish Language CenterMichigan State UniversityEast Lansing, MI48824-1035

Michael L. GeisDepartment of Linguistics222 Oxley Hall1712 Neil AvenueOhio State UniversityColumbus, OH 43210

Myra GoldschmidtEnglish DepartmentSt. Augustine CenterVillanova UniversityVillanove, PA 19085

John L. GrahamGraduate School of ManagementUniversity of California at IrvineIrvine, CA 92717

Linda L. HarlowDepartment of French and Italian248 Cunz Hall1841 Millikin RoadOhio State UniversityColumbus, OH 43210

Noel HouckTemple University of JapanNagaoka Building #14-40-12 TakadanobabaShinjuku-kuTokyo 169Japan

Hy-sook Jeong61 Tong Eui DongJong Ro GuSeoul, Korea

Gabriele KasperDepartment of English as a Second Language1890-East-West RoadUniversity of Hawai'i at ManoaHonolulu, HI 96822

Dale April KoikeSpanish and Portuguese DepartmentBatts Hall 110University of Texas at AustinAustin, TX 78712

Naoko MaeshibaDepartment of Theatre and DanceUniversity of Hawai'i at Manoa1170 East-West RoadHonolulu, HI 96822

Beth Murphy1305 Redbud Hill ApartmentsBloomington, IN 47406

List of Contributors 341

342 List of Contributors

Gayle NelsonDepartment of Applied Linguistics and English as a Second LanguageGeorgia State UniversityUniversity PlazaAtlanta, GA 30303

Joyce NeuConflict Resolution ProgramThe Carter CenterOne CopenhillAtlanta, GA 30307

Steven RossSchool of Policy Studies2-1 GakuenSanda, Hyogo 669-13Japan

Richard SchmidtDepartment of English as a Second Language1890-East-West RoadUniversity of Hawai'i at ManoaHonolulu, HI 96822

Akihiko Shimura3-26-23-208 MinowaKohoku, Yokohama 233Japan

Zhigang Wang3123 Herbert St. #A3Honolulu, HI 96815

Naoko YoshinagaDepartment of Linguistics1890 East-West RoadUniversity of Hawai'i at ManoaHonolulu, HI 96822

Subject index

addiction of responsibility 205addressee 241-242, 245, 254, 255

see also speech act, receptionadvice/lecture 219afterthough~ 34-35,40age 242,244,248,252,254agreement/commiseration 219, 226,

230Americans 95-101,109,111-112,

120,124-125,165,167,171,172-177,178-181,192-193,195 -196, 198-211

Anglos 90apology

form of 164see also speech act, apology

approximation 34, 38Arabs 100,111, 124Australian English 159avoidance 74, 76-77, 97

verbal 47, 55, 58, 74, 78

blessingssee speech act, blesings

British English 160bulge theory 6, 72, 160, 229

Canadian French 159communication 102, 285communicative competence 109-110,

122, 191communicative conventions 89complaint

see speech act, complaintcompliment form 114

Arabic 112-113, 124Egyptian 114-120, 123

confirmationsee speech act confirmation

context 321

continuers 79contradictions 219, 228-229criticism

see speech act, criticismCross-Cultural Speech Act Project 22cross-cultural 1-2, 65, 90, 100,

109-110,122,158,175-176,308,310,334differences 2, 13, 286, 291, 317,

320dissonance 96

cultural 21values 317

culture 1,12,22-23,75,101-102,285,317,319

Danish 160data collection 4,65-81,90-91,

102,161,197,210,221,292declarative intonation 148 -149

negative 133, 146, 148positive 133, 146-147

delivery of a different thought 34, 37"din in the head" 4, 34Discourse Completion Test 4-5,

46-51,59,65-69,71-78,80,217see also role play

Egyptian Arabic 109,113,124elaboration 49, 52-54, 57, 70-72,

79-80empathy 78English 134,137,139,142-149,

161,163-165,181-182,194-195,260-264,267,273as a foreign language 40, 51, 91,

157ethnicity 224, 231ethnographic data collection 45

344 Subject index

ethnographic interview 217,220-221,224,235-236

ethnographic observation 24"evil eye" 7, 112-113explanation

see speech act, explanationexplanation of purpose 195, 203, 209

face-saving 49, 134face-threatening 12, 58, 65, 136, 165,

175,191,223,242,299facial gazing 328, 331-332favor-asking

see speech act, favor-askingformulaic speech 4

use of 34, 36French 133-134,137,139,

141-149

gender 223-224, 230, 242,244-245, 251

Georgians 100German 156, 159

head act 288-289hearer 286

see also speech act, receptionHebrew 159-160, 193-194hedging 1, 77-79Hispanics 90,99-100

I. F. I. D. (Illocutionary Force IndicatingDevice) 164,170-171,174-176,179, 181

illocutionary 134-135, 155, 203,261,263,285,288

immediate retrospection 32imperatives 290-291, 294-295,

303-308imposition 241, 244, 250, 252-254interlanguage 257interlanguage pragmatics 45 -46,

65intimacy (of participants) 72invitations

see speech act, invitationIranians 100, 111

Japanese 111-112, 124-125, 156,159,161,163-165,167, 169,172-177,179,180-181

Jewish 224, 231-232see also ethnicity

joke/teasing 219justification 203, 209, 210

Korean 192,195-211

language habits 129language of thought 32language system 129lexical avoidance/simplification 34,

38linguistic form 131, 191, 203

variation 131

mesage omission/abandonment 34,36

mitigators 50, 175monitor 34-35multiple choice questionnaire 46

native speaker 7, 10, 46, 52, 56-58,69,72,81,89,91,97-99,101,125,158-159,165,169-170,178-179,192-196,199,206,208,211,217-218,242,291,323

natural data 5,66-67,91-92,102

negative 262, 267reaction 274use of 260-261

negotiation 14,47-49,52,57-59,71-72,193,274,319-320,322,327,330-332,334

New Zealand 111, 159non-native speaker 3, 5, 7, 46,

52-59,89,91,96-99,101,111,125,155,158-159,170,191-196,199-200,208,211-212,217-218,241,255,271,275,291

non-verbal 5,51,53-56,59,95-96,99,101,140,320-321,326-327,330,332-334

open-ended elicitations 25overlaps 387

partial delivery of a thought 34, 37personalization of the problem 205phatic expression 90, 92-93, 101politeness 8, 75, 137, 139, 142,

149-150,155,195,257,303-304,306,311

pragmalinguistic 102, 155 -157, 160,308failure 98

pragmatic competence 257Puerto Ricans 90, 97

qeustions 219

real play 24response/topic switch 219, 227requests

see speech act, requestsrole 242, 298role play 3, 5, 24-25, 29, 31, 46-49,

51-52,65-67,69,91,96,98-99,102closed 46-47, 59interview 28, 30open 46-48,52,55,59,66oral 75, 77written 71, 75

Russians 100

"satisfaction" conduction 136, 142second language

acquisition 191learners 40, 109, 129, 149-150,

157, 160pedagogy 102problems of 129profic~ncy 157, 160

self-control 52self-debate 34Semi-Direct Oral Proficiency Interview

30semi-ethnographic 24, 28silent periods 327, 330-331sincerity principle 92, 135

Subject index 345

social differentiation 174social distance 159-160, 174, 178,

223,228-229social power 159, 175, 272sociocultural

ability 3, 22, 27knowledge 122, 124, 192

sociolinguisticability 23, 27, 98, 110, 155acceptability 194, 206behavior 209errors 218knowledge 122, 124, 192rules 109-110, 122, 218

sociopragmatic 96, 98, 102, 132,139,148-150,155,156,170

solutioncandidate request 201-203, 209

Spanish 160, 260'-264, 273, 276speaker 286

see also speech act, productionspeech act 1-5,7,13,21-22,24,29,

33,40,45,52,58-59,67,72,80,91, 101, 125, 134-137, 191-194,217,220,242,257-258,260,263-265,267,273-275,285,287-288agreements 52, 55, 57-58alternatives 53apologies/apologizing 9-10, 21,

28-29,48-49,54,77-79,93,156,158-160,165,167-172,174-176,178-182,192-194

direct 179ritual 159substantive 159blessings 109complaints 1, 28-29, 100, 191,

196-211,223,226,235-236indirect 219, 223-224,

226-227,230,232-236compliments 1,7-8,21,109-120,

122-126,192,220confirmations 53-58criticism 197-198, 200,202-211explanations 53-54

346 Subject index

favor-asking 11, 242-244,253-254see also speech act, requests

goals 21-22greetings 6-7, 89-92, 94-102

all-business 95chat 94intimate 95introductory 95long 94on the run 93,9~ 101re-greeting 96speedy 93,97-100

griping 220, 226invitations 50-52, 72, 100, 192,

220ambiguous 192unambiguous 192-193

offers 51-52, 220production 27-29,33,45,102,

109-110,194provide information 132reasons 53reception 27, 102refusals 4, 45, 47, 49-53, 57-58,

65,70,73-75,77,157repetitions 70, 72, 80restatement 56requests 1, 48, 50-52, 55, 58,

69-70, 77, 129, 132-133,135-137, 140, 142, 145, 150,155,158,201,310-311

suggestions 2, 12-13, 50-52, 58,260-264,267,271-272,285,287-290,293-299,303,305,310to buy 300, 302to enjoy 300to get 300, 302to use 300unspecified 300

thanking 1speech act behavior 22, 27, 39-40,

194,217,241,286speech act event 288

speech act set 10-11, 21-22, 93,158,193,195,198-199,205

speech act strategies 22-23, 39,74-76,80,158,163,175

speech community 67-68,217,219,236-237,241,253-254

spontaneous speech 221, 235 -237status

non-student 242, 244-245, 248,254

student 242, 245, 248, 254stylistic 130-132, 134

see also linguistic variationssupporting 288-289Swedish 101

TESOL 49, 69, 72-73, 78-79, 81,218~242,254

Thai 160, 165, 181touching 329, 332transfur 10, 110, 156, 160, 168,

257-259,261,268,275error 259negative 110, 165, 167, 169-171,

174,176,179-181positive 167-169,180pragmatic 110,155-157,165,

167,175,217,237,257,259strategies 257-258

see also speech act, strategieszero 170

triangulation 4, 39"troubles telling" (aka talk or sharing)

218-219,222-224,226,228,232,235,237

turns 52, 56-57, 2\89length 5, 52number of 52, 70

Ukrainians 100use of second person \ 205

verbal report 26, 28, f2-33, 39-40

written discourse completion 33

Author index

Adler, Nancy J. 334Afflerbach, Peter 26Agoston, Thomas 287Aguilar Murillo, Evelyn 24Al Batal, Mahmoud 109, 124Almancy, A.J. 123, 126Alwan, A.J. 123, 126Angelman, Reinhard 320, 323, 326,

330,333Araki, Shoko 111, 124Argyle, Michael 328Aston, Guy 285Austin,John 1,192,241

Bales, Robert F. 320Banerjee, Janet 65, 288, 290-291,

311Bardovi-Harlig, Kathleen 47, 53, 55,

57-58Barnlund, Dean C. 111, 124, 159,

175-177,179-180,182Barthes, Roland 287Beebe, Leslie 4, 47-51, 53, 55,

65-68,96,155-157,165,175,181,217,291

Belk, Russel W. 290Bell, Robert A. 112Bergman, Marc 77, 155, 158-160,

163-165,174-175,181Bhatia, T. K. 287Billmyer, Kristine 40Blackwell, Roger D. 319Blum-Kulka, Shoshana 9,21-28,

52,65,77,110,123,125,134,155-157,194,241,258,260,275,287-288

Bodman, Jean 66, 91, 93, 96, 155,179

Bolinger, Dwight 287Bonoma, Thomas V. 320-321

Borkin, Ann 158Bowen, J. Donald 90, 92Boxer, Diana 100, 224,

229-230Bratt Paulson, Christian 101Brown, Bert R. 319Brown, Dorothy F. 111Brown, Penelope 2, 9, 135 -137,

223,233,242-243,262,286

Brown, Roger 72

Caballero, Marjorie 289Camargo, Eduardo G. 287, 290Canale, Michael 109Carrell, Patricia 65, 195, 288,

290-291,311Cateora, Philip R. 321Chan, Carole 89Clyne, Michael 157Cohen, Andrew 1-2, 9, 21,

25-26,28-32,40,45,65,70,77, 93, 155, 157-160,193-194,198

Coleman, Linda 287Condon,JohnC.320,327Cook, Mark 328Corder, S. Pit 257, 258Coulmas, Florian 311Cowles, Maria A. 30Cummings, Larry L. 318Cummings, Martha C. 48-49, 96,

217

D'Amico-Reisner, Lynne 192, 219Dahl, Merete 22, 45 -46, 48, 65 -66,

285Daikuhara, Midori 67Di Vito, Nadine O. 133, 147Dogancay, Seran 89

348 Author index

Dowling, Graham R. 289Doyle, Fatima 30Du Fon, Margaret A. 155Duffy, Susan A. 26Dsirat, Claude 137Dwyer, Robert F. 319

Ebsworth, Timothy J. 91, 99Edmondson, Willis 48, 288, 303Eisenstein, Miriam 66, 91, 93, 96,

155,179El Bakary, Waguida 109, 124Ellis, Rod 97Engel, James F. 319Erickson, Fred 327Ervin-Tripp, Susan 218, 242-243,

307

Faerch, Claus 26, 155Fang, Hanguan 310Felder, Leonard C. 320-321Ferrara, Alessandro 193-194Fields, George 290Fouraker, Lawrence E. 322Frake, Charles o. 320Francis, June N.P. 318Fraser, Bruce 70, 148, 159, 287Fry, William R. 319,322,328-329

Gadet, Fran~oise 133Garcia, Carmen 155, 160Gass, Susan 2, 51, 59, 258-259Geis, Michael L. 135, 287, 309Giles, Howard 157Gilman, Albert 72Goffman, Erving 72, 77-78, 134,

159,286Goldschmidt, Myra 242Gordon, David 242Graham, John L. 318-319, 321,

326-327,330,333-334Green, Georgia M. 290Grice, H. Paull, 285Guetzkow, Harold 319-320, 326Gumperz, John 1, 87, 241, 285, 321,

327,333Gymn, Kyng-Il 292

Habermas, Jiirgen 1Hall, Dennis R. 287, 308, 321Hard, Tristan 133Harnett, Donald L. 318Harrigan, Katherine R. 317Harris, Richard J. 287Hartford, Beverly 47, 53, 55,

57-58Hatch, Evelyn 219Heng, J.H. 310Hilferty, Ann 90Ho, Suk-ching 290Hobbs, Dianne 261Hodlin, Susan 89Holmes, Janet 9, 68, 111, 159Hong, Jae W. 289Hopper, Robert 112Houck, Noel 51, 59House, Julianne 9,22,65,77,110,

125,134,155-156,159-160,258,260,287-288,303

Huber, Lisa 192Hunt, Shelby D. 289Hymes, Dell 1, 67, 109, 191, 219,

241,243

Jacobs, Lawrence W. 292Jastram, Rothdow W. 318Jefferson, Gail 218Johns, Ann M. 286Johnson, Patricia 157Johnston, Peter 26Jones, Steve 45-46,160Jordens, Peter 156JuPP, Thomas C. 89

Kachru, Yamuna 218Kaplan, Robert B. 291Kapoor, Ashok 318Kasper, Gabriele 9, 22, 26, 45 -46,

48,65-66,77,110,125,134,155-160,163-165,167,174-175,181,258,260,285,287-288

Katriel, Tama 217Kay, Paul 320Kaynak,Erdenor 289

Kellerman, Eric 145, 156, 258Kelley, Harold H. 322Kenyon, Dorry M. 30Kess, Joseph F. 287Keown, Charles F. 292Kinjo, Hiromi 50, 53Knapp, Mark L. 112, 124Koike, Dale 257, 261-262Konneker, Beverly 195Krugman, Dean M. 289

Labov, William 241Laczniak, Gene R. 289Lakoff, Rubin Tolmach 286-287,

291Lee,JohnR.E.218Leech, Geoffrey N. 110, 155, 287,

291,308-309Levenston, Edward 241Levinson, Stephen 2,9,89,135-137,

223,233,242-243,262,286Lewis, Steven A. 319-320, 322,

328-329Linton, Ralph 319-320Lu, Zaiming 290

Mack, Robert L. 26Madden, Carolyn 289Maloney, Clarence 112Manes, Joan 65,68,111-112,

117-119, 122, 124, 192Marmoor, Thomas 45-46, 160Masavisut, Nitaya 287Matsukubo, Sinya 289Matsumoto, Yoshiko 291McKersie, Robert B. 320Miracle, Gordon E. 290Mitchel, Lionel A. 289Mitchell, Candace J. 45-46, 48, 65,

77Moeran, Brian 287Morgan, Graham 89Muderrisoglu, Aydin 289Murphy, Beth 209

Nakane, Chie 326Nelson, Gayle L. 109, 124

Author index 349

Neu, Joyce 333Nine-Curt, Carmen J. 90, 99Nolen, William 148

O'Barr, William M. 287Ochs, Elinor 1Olshtain, Elite 1-2, 9, 21, 23-32,

40,45,52,65,70,77,93,123,155-160,174,193-194,198,288

Olson, Gary M. 26Owen, Marion 77,79

Paiva, Riccardo 30Pennington, Allan L. 319, 326Pruitt, Dean G. 319-320, 322

Reinhart, Susan M. 158Resnik, Alan 289Rice, Marshall D. 290Richards, Jack 110Ricklefs, Richard 321Rintell, Ellen M. 45 -46, 48, 65, 77,

97,288,290Robinson, Mary 155 -157Rose, Kenneth 46, 66Rubin, Jeffrey Z. 319Rubin, Joan 55

Sadock, Jerrold 1Saracino-Resh, Linda 287, 308Sawyer, Jack 319-320Scarcella, Robin 4, 51, 90Schachter, Jacquelyn 259Schegloff, Emanuel 79Schiefflin, Bambi 1Schmidt, Richard W. 110Schmidt, Rosemarie 287, 292Schroeder, Kim 287Searle, John 1, 90, 92, 109, 135 -136,

191-192,241,260,285,287Selinker, Larry 167, 257-258Sheffer, Hadass 157Sherry, John F. 287, 290Shouby, E. 123Siegel, Sidney 322Sin, Yat-ming 290Slobin, Daniel 92

350 Author index

Soldow, Gary F. 321Spindel, Carol 109Spiro, Melford E. 320Spradley, John 221, 222Stansfield, Charles 30Stern, Bruce L. 289, 320, 323, 326Stern, Louis W. 330, 333Stubbs, Michael 25Suleiman, Michael W. 123Sukwiwat, Mayor 287Swain, Merrill 109

Takahashi, Satomi 155-156, 179,291,295,306-307

Takahashi, Tomoko 4, 47, 50, 53,65-66,68,155-157,165,175,181,291

Tanaka, Noriko 159Tannen,Deborah 2,75,79,219,223Tarone, Elaine 2Thomas, Gloria Penn 321Thomas, Jenny 98, 155, 218Trosberg, Anna 77, 160Tse, David K. 290Tung, Rosalie L. 318

Ueda, Keiko 326Uliss-Weitz, Robin 4, 47, 50, 53, 65,

68, 155Ulsh, Ines 3°

Valdman, Albert 133Van Dijk, Teun 193-194Van Zandt, Howard F. 318, 320,

326Vestergaard, Torben 287Vollmer, Helmut J. 159Von Raffler, Engel 287

Walker, Orville C. 319Walters, Joel 47Walton, Richard E. 320Walz, Joel C. 147Watson Gegeo, Karen Ann 45Wells, Louis T. 318Weiss, Stephen E. 318Weizman, Elda 21, 275, 287Widdowson, Henry 218Wolfson, Nessa 22, 24, 40, 45 -46,

65 -68, 72, 97, 109-112, 117,122, 124-125, 134, 160, 192,218,229

Wongmontha, Seri 287

Yamaki, Toshio 290Yoshioka, Miho 159,175-177,

179-180,182

Zhou, Nan 290Zinhan, George M. 289

Speech Acts Across Cultures

Documents

Transcript of Speech Acts Across Cultures