Contrastive analysis of speech acts: what do we do with the research findings?

"Thanks so much for the lovely meal." may work as an adequate acknowledgement of a fine meal just consumed at a colleague's home in another country. The utterance constitutes one of the semantic formulas or strategies for expressing gratitude. Expressions of gratitude, along with other illocutionary acts (i.e., speech with a social function) have been referred to as speech acts. The choice of the speech act of expressing gratitude may be totally appropriate for the above after-eating-a-dinner context. On the other hand, it may be the wrong speech act or at least the wrong semantic formula to use within that speech act set. Instead of expressing gratitude, perhaps a profuse round of apologies is in order. After all, the hosts had to invite the colleague, prepare the meal, serve it, make sure it was properly consumed, and then as if adding insult to injury, will have to wash the dishes after the colleague has departed. Thus, perhaps the speech act of apology is considered by natives of the language and culture to be more fitting for such a situation, as it might be in some contexts in Japan. To make matters more confusing, it is possible that the social function of the apology is not only that of apologizing but also of thanking the hosts, albeit indirectly.

This article will briefly describe the dimensions of speech acts, call attention to the empirical studies that have been conducted to further our understanding of how speech acts function among natives and nonnatives, address the need for speech act instruction, look at research on the teaching of speech acts, consider what needs to be taught, and finally, discuss possible teaching techniques for speech act lessons.

1. Describing speech acts

There appear to be two kinds of abilities associated with successful production of speech acts such as that of apologizing - what I have termed sociocultural and sociolinguistic abilities. Sociocultural ability refers to the respondents' skill at selecting speech act strategies which are appropriate given (1) the culture involved, (2) the age 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 U.S.) it may be appropriate for speakers who have missed a meeting with their boss through their own negligence to use a repair strategy by suggesting to the boss when to reschedule the meeting. In other cultures (such as Israel), however, such a repair strategy might be considered out of place in that it would most likely be the boss who determines what happens next. Thus, sociocultural knowledge is called for in determining whether a speech act set is appropriate to use and if so which members of the set are s elected for use.

Sociolinguistic ability, in the context of speech act analysis, refers to the respondents' skill at selecting appropriate linguistic forms in order to express the particular strategy used to realize the speech act (e.g., expression of regret in an apology, registration of a grievance in a complaint, specification of the objective of a request, or the refusal of an invitation). Sociolinguistic ability constitutes the speakers' control over the actual language forms used to realize the speech act (e.g., "sorry" vs. "excuse me" "really sorry" vs. "very sorry"), as well as their control over register or formality of the utterance from most intimate to most formal language. For example, when students are asked to dinner by their professor and they cannot make it, while it may well be socioculturally appropriate to decline the invitation, the reply "No way!" would probably constitute an inappropriate choice of forms for realizing the speech act set of refusal. The problem is that sociolinguistically, this phrase wo uld be interpreted as rude and insulting, unless the students had an especially close relationship with their professor and the utterance were made in jest. A more appropriate response might be the following: "I would love to but I have a prior engagement I can't get out of."

The process of selecting the socioculturally appropriate strategy and the appropriate sociolinguistic forms for that strategy is complex since it is conditioned by the social, cultural, situational, and personal factors indicated above. Strategy selection and selection of forms often depend on the social status of the speaker and the hearer since in most societies deference towards higher status, for instance, is realized via linguistic features (e.g., using vous rather than tu in French) or via modification of the main speech act strategies (e.g., adding intensity to the apology or purposely refraining from cursing). Thus, coming late to a meeting might evoke a more intensified and possibly invective-free apology when the recipient of the apology is the boss, rather than a friend. Other factors such as age and social distance are part of the social set of factors that might play a significant role in strategy selection.

It has been found that situational factors also play an important role in strategy selection. Some situations generalize across cultures and hence will elicit similar strategies in different languages while other situations are more culture-specific and are likely to provoke cross-cultural clashes. One of the situations which was used in the CCSARP project (Blum-Kulka -- House-Edmondson -- Kasper 1989) for apologies had a waiter bring the customer the wrong order. The native respondents in the role of waiter in all the investigated languages avoided the expression of personal responsibility, perhaps because admitting such a mistake might cost them their job. In contrast, the cross-cultural study of complaints showed that noise made by neighbors is perceived in some cultures as a serious offense which deserves a complaint while in other cultures it is viewed as a less significant offense.

2. Empirical data on speech act behavior

While the process of defining and identifying speech acts has been going on since the 1960s, the last fifteen years have marked a shift from an intuitively based anecdotal approach to speech act description to an empirical one. Such empirically based research, encompassing both quantitative and qualitative approaches, has focused on the perception and production of speech acts by learners of a second or foreign language (in most cases, English as a second or foreign language), at varying stages of language proficiency, and in different social interactions. This work has included efforts to establish both cross-language and language-specific norms of speech act behavior, norms without which it would be impossible to understand and evaluate interlanguage behavior.

Early empirical research on speech act sets (e.g., Cohen -- Olshtain 1981) was in part prompted by a realization that although transfer occurs at the sociocultural level, few if any contrastive studies were systematically undertaken in order to characterize such phenomena (Schmidt -- Richards 1981; Riley 1981; Loveday 1982). Research in second language acquisition has helped to provide empirical descriptions of speech acts such as requests, compliments, apologies, complaints, refusals, and expressions of gratitude (see Wolfson -- Judd 1983; Wolfson 1989; Gass -- Neu 1996). Empirical studies concerning the nature of various speech acts in a variety of languages and cultures have been steadily accumulating over the last few years. As a result there is a growing source of empirical data on the strategies for performing these acts (see Cohen 1996, for a review of empirical studies).

The findings from a cross-cultural study by Cohen, Olshtain, and Rosenstein (1986) can serve as an example of empirical data on the similarities and differences between native speakers of American English and advanced nonnatives with respect to apology behavior in English. The 180 respondents for this study included 84 native Hebrew-speaking advanced learners of English studying at one of five Israeli universities and a comparison group of 96 native speakers of American English studying at one of six U.S. universities. The basic finding was that nonnative speakers lacked sensitivity to certain sociolinguistic distinctions that native speakers make, such as between forms for realizing the semantic formula of "expressing an apology", e.g., "excuse me" and "sorry". At least one out of every five times a native speaker offered an expression of apology, it was with "excuse me", while few nonnative speakers used this form. Nonnative speakers were limiting themselves to the use of "sorry" in contexts where "excuse m e" would also be acceptable and possibly preferable.

While native speakers and nonnative speakers did not seem to differ markedly in the use of main strategies for apologizing, striking differences emerged in the various modifications of such apologies, especially in the use of intensifiers such as "very" and "really". Nonnative speakers intensified their expression of apology significantly more in one situation ("forgetting to help a friend buy a bike") than did native speakers. This extra intensity on the part of the nonnative speakers was not necessarily warranted, given the generally low or moderate severity of the offense in that situation.

Not only did nonnative speakers tend to intensify more, but also used a wider and more indiscriminate set of forms. Actually, the nonnative pattern was either to overgeneralize one of the forms ("very" and "sorry") or to use a variety of forms ("terribly", "awfully", "truly"). The nonnative speakers did not use "really" in the way that the native speakers did. They attributed to the intensifier "very" the same semantic properties as to "really", while the native speakers tended to make a distinction whereby "really" expressed a greater depth of apology, regret, and concern, and "very" was used more for matters of social etiquette. For example, in a situation of "scalding a friend with coffee in a cafeteria", the native speakers tended to use "really sorry" while nonnative speakers used "very sorry", which sounded less intensified.

Before the advent of data such as those gathered in the Cohen, Olshtain, and Rosenstein (1986) study just described above, teaching materials dealing with speech acts had to rely for the most part on the curriculum writer's intuition and could best be characterized as reflecting a high level of simplicity and generality. Popular EEL textbooks, for example, only looked at one semantic formula for apologizing, namely, the expression of an apology ("sorry", "I'm sorry", "I'm very sorry"). Brief reference was made to other apology strategies perhaps, but without providing any underlying principles for when to use which strategy (e.g., Blundell -- Higgens -- Middlemass 1982; Berry -- Bailey 1983; Jolly 1984; Swan -- Walter 1985). There were some efforts even in the early years of empirical data collection to move from intuitively-based to empirically-based materials (e.g., Bodman -- Lanzano 1981). In the intervening years textbooks have made more of an effort to reflect genuine language rather than language charac terized at times by linguistic curiosities which were generated in order to teach specific grammatical structures.

Thus, the field of contrastive analysis of speech acts has come a long way in the last twenty years. There are both empirical frameworks for the description of speech act behavior and numerous data sets, replete with data regarding a whole host of speech acts. The question now is what to do with the empirical data -- how best to introduce these findings into the classroom. Since the teaching of second-language words and phrases isolated from appropriate sociocultural contexts (or worse, teaching material that may not be appropriate in any current sociocultural context) may lead to the production of utterances which do not achieve their communicative purposes, the question is how best to create meaningful sociocultural contexts through instruction in a way that does not overwhelm the learners and produces the desired results.

3. The need for selective teaching of speech act behavior

As documented in the empirical research literature, nonnative speakers of a language are prone to produce (and also interpret) some if not many speech acts the way that they would in their native language and culture. Not surprisingly, they may also find that their utterances are not at all appropriate for the target language and cultural situation. Depending on the understanding and generosity of the native interlocutors, the resulting inappropriateness may be ignored, it may be thought humorous, or it may produce friction, annoyance, or even scorn.

Could we expect learners to simply acquire appropriate speech act behavior over time from being in the environment where these forms are produced? According to some researchers, acquisition of native-like production of speech acts by nonnative speakers may take many years (Olshtain -- Blum-Kulka 1984) since the sociocultural strategies and the sociolinguistic forms are not always "picked up" easily. This is clearly evident in speech acts that learners would not even come into contact with except on rare occasions (e.g., what to say to the bereaved at a funeral). But even in the case of more common speech acts such as apologizing, many years could go past before a speaker sorts out the appropriate semantic formulas for given contexts and the sociolinguistic structures that work best with those semantic formulas.

Hence, there seems to be at least some justification for heightening learners' awareness of these distinctions by explicitly teaching some of the finely-tuned speech act behavior that is not simply acquired over time. The fact that speech acts reflect, for the most part, routinized language behavior would ostensibly facilitate classroom learning of these patterns in the sense that much of what is said is predictable. For example, almost half of the time that an adjective is used in a compliment, it is either "nice" or "good" (e.g., "Thai's a nice shirt you're wearing" or "It was a good talk you gave"), with "beautiful", "pretty", and "great" making up another 15% (Wolfson -- Manes 1980). Nonetheless, despite the routinized nature of speech acts, learners may still have to choose from among numerous strategies in order to perform the speech act, depending on the sociocultural context. So, as in the above case, the speaker must recognize that there is a need to apologize after a meal in a Japanese home and must select a strategy for apologizing appropriately. There may also be a need to select from a variety of possible sociolinguistic forms for realizing these various apology strategies (e.g., apologizing for making the disturbance of getting up from the table, apologizing for having inconvenienced the hosts by their having to prepare a meal, and so forth).

4. Research on the teaching of speech acts

Whereas there are numerous studies of speech act behavior, there are as yet few studies on the effects of instruction on speech act performance. One study was conducted with advanced EFL learners in Israel, ten of whom were studying in private language schools and eight in a teachers' college (Olshtain -- Cohen 1990). Native speakers of American English provided baseline data for comparative purposes. The learners were pretested to determine gaps in their apologizing behavior. Then they were taught a set of three 20-minute lessons aimed at filling in the gaps -- information about the strategies within the apology speech act set and about modifications of apologies through the use of intensification and emotionals. Finally, they were posttested to determine what was learned.

The findings suggested that the fine points of speech act behavior, such as (1) types of intensification and downgrading, (2) subtle differences between speech act strategy realizations, and (3) consideration of situational features, can be taught in the foreign-language classroom. Whereas before the instruction, the nonnative speakers' apologies differed noticeably from those of the native speakers, after instruction advanced learners were somewhat more likely to select apology strategies similar to those that native speakers used in that situation. For example, in a situation of "forgetting to buy medicine for a neighbor's sick child", the response of one nonnative before training was a weak expression of responsibility ("Unfortunately not yet ...") and an offer of repair ("but I'll be happy to do it right now"). After training it was an intensified expression of apology ("I'm deeply sorry.") and an offer of repair ("I can do it right now.").

Furthermore, after training, nonnative speakers produced shorter utterances, also more in keeping with native behavior.

Prior to instruction, one learner responded verbosely to a situation of forgetting to meet a friend with, "Did you wait for me? You must forgive me. I could not come because of problems and I tried to warn you by phone but ..." This response was typical of learners at the advanced-intermediate stage of language acquisition who, when uncertain about how to say something, would overcompensate by using too many words (Blum-Kulka -- Olshtain 1986). After training, the utterance was shorter: "Oh, I'm so sorry. It dropped out of my mind." Perhaps the area that met with most success was that of the use of intensifiers. Before training, intensifiers were generally absent in situations like "forgetting to buy medicine for a neighbor's sick child" (only 20% use). After training, intensifiers (e.g., "I'm really sorry I forgot ...") were used in almost all cases (90%).

In another study involving the teaching of speech acts, Billmyer (1990) compared nine female Japanese ESL learners tutored in complimenting and responding to compliments with nine similar learners who were untutored in complimenting. The study looked not just at the speech act but at the reply -- that is, whether the respondent accepted, deflected, or rejected the compliment, and the types of deflecting moves (a comment, a shift of credit, a downgrade, a request for reassurance, or a return). The study found that learners who were tutored in complimenting produced a greater number of norm-appropriate compliments, produced spontaneous compliments (which the untutored group did not), used a more extensive repertoire of semantically positive adjectives, and deflected many more compliments in their reply. The researcher concluded that formal instruction concerning the social rules of language use given in the classroom can assist learners in communicating more appropriately with native speakers outside of the cla ssroom.

Dunham (1992) described an informal study of 45 Southeast Asian high school students, employing the complimenting strategy as outlined by Wolfson (1989). The students in the study were instructed on how to connect -- that is, to maintain or continue the conversation based on the response of the addressee. The author reported that the feedback from the students concerning their use of complimenting and connecting was encouraging, and often resulted in an increased confidence in initiating and maintaining conversations with native speakers.

Finally, a study investigated the effect of teaching 8th-grade French immersion students functional aspects of sociolinguistic competence, specifically sociostylistic variation (Lyster 1994). The treatment was administered by three French immersion classroom teachers during French language arts to 106 students over 5 weeks. Three measures of what was termed sociolinguistic competence (i.e., the ability to recognize and produce socially appropriate language in context) were given as a pretest, immediate posttest, and delayed posttest to the experimental group and to a comparison group. The measures included oral and written production tests and a multiple-choice test.

The oral production test involved viewing slides of people in specific contexts and respondents required to perform different speech acts (requesting or giving directions, requesting help in math, offering to help carry books) in a formal context (unknown adults, school librarian, math teacher) arid an informal one (with peers). The writing production test consisted of two tasks -- writing an informal note (from mother to child about a messy room) and a formal letter (to the landlord asking permission to keep a dog in the apartment). The multiple-choice test had items for which the learners needed to select the utterance that best fit in a given context or give the context which best suited a given utterance. Findings showed improvement in the experimental group in their ability to appropriately and accurately use vous in formal speaking situations, to appropriately use vous in formal letters, and in their ability to recognize contextually appropriate French.

5. Determining what to teach about speech acts

This article has noted that empirical data on speech acts now exist and that some studies have been conducted which demonstrate that instruction can make a difference. The issue still remains as to where a given classroom teacher will find material to teach and how much of it should be taught. Even if the classroom teacher starts with the basic goal to minimize negative transfer of speech act behavior from the native language, there is the question of which speech acts in which situations between which languages -- not to mention the host of other variables such as age, status, gender, and so forth.

The teacher may find that the given textbook has a fair amount of useful information on speech act behavior. If not, the teacher may be able to obtain some information on how natives perform certain important speech acts, such as requesting, complaining, and apologizing from the research literature (such as that cited in this review). If it is not available, then a possible means for obtaining it would be through observing speech acts as they occur naturally. As we have noted above, however, this may not be a very efficient means of obtaining data, especially if fine-tuned distinctions are desired. So, there is a need to turn to more contrived means where data are elicited in a. more or less structured way.

Actually, if the learners have access to native-speaking informants (more likely in a second rather than a foreign language learning experience), they themselves could elicit speech act samples from them. This approach may even enhance the learning process more than if the teacher were simply to lay out the possible alternatives in class. While there are advantages to having learners serve as data gatherers, this mini-research approach may be problematic if they have difficulty obtaining examples of the speech acts in question.

A good example of this problematicity was an effort by a team of students to gather information about apologies in a contrived but natural situation. The students attempted to capture on videotape a series of induced apology situations, but the investigators encountered numerous difficulties (Murillo -- Aguilar -- Meditz 1991). They crouched just outside faculty members' doors; when the professors emerged from their office, they would inadvertently bump into the student and would need to apologize. The method was time-consuming since the targeted faculty members usually did not emerge from the office right away, and too often there was either no audible apology or else a mumbled apology that was not captured on the videotape.

One of the reasons that structured role play and questionnaires are used more than natural data approaches is to control for variables such as age, status, level of imposition, desired level of politeness intended by the speaker, and the like. There could be a range of possibilities depending on the mood of the interlocutor and a host of other variables.

Once descriptions of the speech acts are made available, the next task is to determine the degree of control that learners have over those speech acts through the multiple measures suggested above -- role play, discourse completion tasks, verbal report interviews, and acceptability ratings. Ideally, this information could then be used to prepare a course of instruction that would teach to the gaps in language knowledge, and also give tips as to strategies that may be useful for producing utterances. The role of the learners is to notice similarities and differences between the way that natives perform such speech acts and the way that they do, which is often influenced to some extent by the way they would perform such communicative functions in their native language.

6. Teaching techniques for speech act lessons

Let us look briefly at actual teaching techniques with regard to teaching speech acts. Whatever the goals of teaching, there are various means for presenting and rehearsing the use of speech acts. Whatever approach is used, it is always necessary to specify the situation (e.g., student making request of professor, patron complaining to waiter, etc.) and to indicate the social factors involved (age, sex, social class and occupation, roles in the interaction, status of the participants), and to match these with the most common realizations of the speech act.

The planning and implementation of lessons on speech acts could involve, among other things, the following five steps (adapted from Olshtain -- Cohen 1991):

a. Diagnostic assessment - determining the students' level of awareness of speech acts in general and of the particular speech act to be taught.

b. Model dialogs--presenting students with examples of the speech act in use.

c. The evaluation of a situation--having students decide, in pairs or small groups, whether a speech act realization is appropriate.

d. Role play activities--having the learners provide the details of the speech act and then act out, in role-play fashion, the conversation which is likely to take place between themselves and the interlocutor(s).

e. Feedback and discussion--having students talk about their perceptions, expectations, and awareness of similarities and differences between speech act behavior in the target language and in their first culture.

As we see from this approach, the emphasis is on teaching to the contrasts between the languages in terms of speech acts, rather than just teaching the speech acts of the language and expecting the students to memorize them, as is done in some instructional approaches. In addition, it is possible to furnish other information about the particular speech acts, such as how they are performed, why, when, and with whom. Efforts can be made to contextualize the speech acts in a rich socially situated context that makes then easier to remember. Furthermore, thought needs to be given to the selection of speech acts for instruction. The variables include the importance of the act in terms of its frequency in speaking or written language, the ease involved in Leaching it, the likelihood that it will be acquired naturally without the need to teach it, and the extent of empirical information about the given speech act.

7. Conclusions

Thus, the field of contrastive analysis which in the early years (i.e., the 1950s through the 1970s) focused primarily on grammatical structures and phonology has over the last several decades broadened to encompass crosscultural pragmatics. In recent years, the number of doctoral dissertations focusing on this domain, for example, is increasing. This increase can be explained partly in that contrastive speech act analysis has been a somewhat underdeveloped area for research. Secondly, this is a complex and thus challenging area for research. Thirdly, it is fascinating to observe language transfer at work in areas that go way beyond surface forms to an entire sociocultural outlook.

Clearly, at present there is still more effort being put into investigating speech act behavior than into gathering up the empirical results of these investigations for the purpose of textbook writing. Hopefully, as more descriptions are generated there will be greater effort extended to produce instructional materials and the refinement of techniques for ensuring that learners a working knowledge of speech act behavior appropriate for the situations in which they will need to use the given language.

REFERENCES

Berry, C. -- D. Bailey

1983 Video English. London: MacMillan.

Billmyer, Kristine

1990 "'I really like your lifestyle,' ESL learners learning how to compliment", Penn Working Papers in Educational Linguistics 6: 31-48.

Blum-Kulka, Shoshana -- Elite Olshtain

1986 "Too many words: Length of utterance and pragmatic failure", Studies in Second Language Acquisition 8:165-179.

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

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

Blundell, Jon -- Jonathan Higgens -- Nigel Middlemass

1982 Functions in English. Oxford: OUP.

Bodman, Joan W. -- M. K. Lanzano

1981 Milk and honey: An ESL series for adults. Bellmawr, NJ: HBJ Center for Lifelong Education.

Celce-Murcia, Marianne (ed.)

1991 Teaching English as a second or foreign language. New York: Newbury House/ HarperCollins.

Cohen, Andrew D. -- Elite Olshtain

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

Cohen, Andrew D. -- Elite Olshtain -- David S. Rosenstein

1986 "Advanced EFL apologies: What remains to be learned", International Journal of the Sociology of Language 62: 51-74.

Dunham, Patrick

1992 "Using compliments in the ESL classroom: An analysis of culture and gender", MinneTESOL Journal 10: 75-85.

Fisiak, Jacek (ed.)

1981 Contrastive linguistics and the language teacher. Oxford: Pergamon Press.

Gass, Susan M. -- Joyce Neu (eds.)

1996 Speech acts across cultures: Challenges to communication in a second language. Berlin: Mouton de Gruyter.

Jolly, David

1984 Writing tasks. Cambridge: CUP.

Loveday, Leo 

1982 The sociolinguistics of learning and using a non-native language. Oxford: Pergamon Press.

Lyster, Roy

1994 "The effect of functional-analytic teaching on aspects of French immersion students' sociolinguistic competence", Applied Linguistics 15: 263-287.

Mac Mathuna, Liam -- David Singleton (eds.)

1984 Language across cultures. Dublin: Irish Association for Applied Linguistics.

Murillo, Evelyn A. -- Hellen Aguilar -- Aimee Meditz

1991 "Teaching speech act behavior through video: Apologies", (Paper presented at the Ohio TESOL Fall Conference, Ohio University, Athens, Ohio.)

Olshtain, Elite -- Shoshana Blum-Kulka

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

Olshtain, Elite -- Andrew D. Cohen

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

1991 "Teaching speech act behavior to nonnative speakers", in: Marianne Celce-Murcia (ed.), 154-169.

Riley, Phillip

1981 "Towards a contrastive pragmalinguistics", in Jacek Fisiak (ed.), 121-146.

Schmidt, Richard W. -- Jack C. Richards

1981 "Speech acts and second language learning", Applied Linguistics 1:129-157.

1985 Cambridge English course. Book 2. Cambridge: CUP.

Swan, Michael -- Catherine Walter

Wolfson, Nessa

1989 Perspectives, sociolinguistics and TESOL. Cambridge: Newbury House/HarperCollins.

Wolfson, Nessa -- Elliot Judd

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

Wolfson, Nessa -- Joan Manes

1980 "The compliment as a social strategy", Papers in Linguistics 13: 391-410.