Journal of Language and Social Psychology http://jls.sagepub.com/ How Do Chinese Bilinguals Respond To Variations of Interviewer Language and Ethnicity? Herbert D. Pierson and Michael H. Bond Journal of Language and Social Psychology 1982 1: 123 DOI: 10.1177/0261927X8200100203 The online version of this article can be found at: http://jls.sagepub.com/content/1/2/123 Published by: http://www.sagepublications.com Additional services and information for Journal of Language and Social Psychology can be found at: Email Alerts: http://jls.sagepub.com/cgi/alerts Subscriptions: http://jls.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jls.sagepub.com/content/1/2/123.refs.html >> Version of Record - Jan 1, 1982 What is This? Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 HOW DO CHINESE BILINGUALS RESPOND TO VARIATIONS OF INTERVIEWER LANGUAGE AND ETHNICITY? Herbert D. Pierson and Michael H. Bond The Chinese University of Hong Kong Do Chinese bilinguals change their non-verbal behaviours when they speaking English as opposed to Cantonese and speaking with Americans as opposed to fellow Chinese? In order to answer these questions, 64 female English majors at the Chinese University were video-taped during a standardised interview in either Cantonese or English by an interviewer of American or Chinese ethnicity. Measures of self-perception and perception of the interviewer were also taken to illuminate the meaning of the potential changes in non-verbal behaviour. The language and ethnicity variables had functionally different impacts on the non-verbal behaviours: the higher speed, less frequent use of filled pauses, and increased gazing when using Cantonese are suggestive of its greater redundancy relative to English; the lesser talking combined with Abstract are increased smiles and torso shifts emitted with the American interviewers were with the higher potency and greater informality with which the Americans were perceived by these bilinguals. The American interviewers were relatively less fluent using Cantonese than were interviewers in the other conditions. Perceiving this lesser skill, the interviewees accommodated by downgrading their self-ratings of second language fluency and by increasing their frequency of filled pauses while answering. compatible How Do Chinese Bilinguals Respond To Variations of Interviewer Language and Ethnicity? We American Chinese girls had to whisper to make ourselves Americanfeminine.... We invented an American-feminine speaking personality. Maxine Hong Kingston in The Roman Warrior. Although rigorous comparisons are scarce (LaFrance & Mayo, 1978), it is widely believed that members of different cultures use different non-verbal codes when communicating in their native languages. Studying these codes, many researchers have compared the non-verbal behaviour of bilinguals when speaking their first and second languages. The rationale is made explicit by LaCroix & Rioux (1978): 123 Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 124 If the learning of a tongue implies not only the learning of the language, but also the learning of a cognitive and cultural gestalt of which the language is only one manifestation, then bilinguals should show different non-verbal behaviours in the two languages where they are competent. (p. 131). Studies within this tradition vary considerably in their methodological that it becomes difficult to place much confidence in some of the conclusions. Of the published studies, some have shown differences (Elzinga, 1978; Ervin-Tripp, 1964; Gallois & Markel, 1975; Grujic & Libby, 1978a; von Raffler-Engel, 1976); others have not (LaCroix & Rioux, 1978; Weins et al., 1976). One major problem in understanding these conflicting results arises from the failure of most experimenters to manipulate the ethnicity of the bilingual’s partner or audience while manipulating the language being spoken. It would indeed be surprising if the partner’s ethnicity had no impact on the bilingual’s non-verbal repertoire; interactions between persons of differing ethnicity often introduce the status differences characterising relations between the two ethnic groups which the interactants represent. These status differences will of course have consequences for the non-verbal features of their interaction (see e.g., Bond & Ho, 1978; Thakerar et al., 1982). Obviously, then, one cannot know whether and how the variables of interaction language and interlocutor ethnicity function until both are varied simultaneously and dependent measures taken across a wide spectrum of non-verbal behaviours. A study by Grujic & Libby (1978a) attempted just such a separation. Their subjects were French-Canadian bilinguals from Ontario who conversed in French or English with confederates identified as ethnically French- or EnglishCanadian. The bilinguals were video-taped while they both asked and answered pre-selected questions chosen to stimulate conversation. Measures of interpersonal distance, hand movements, gazing, smiling, and time talking were then taken from the tapes and related to the language/ethnicity variation. Significant effects were due almost entirely to the language of the interaction rather than to the ethnicity of the partner. A number of problems in the presentation of their study make this conclusion premature, however. It is impossible to tell from their report if the confederates’ behaviour was standardised in any way, whether the confederates were in fact ethnically French or English, or the confederates’ degree of fluency in their two languages. Furthermore, the length of the interactions was not controlled, so some of the findings may have been a function of the interaction length rather than a result of the language used. Clearly, a conceptual replication with some methodological tightening is needed. A 2 x 2 design which carefully varies both language and partner ethnicity may also be instructive in exploring the dynamics of speech accommodation (Giles & Smith, 1979; Thakerar et Rl., 1982). Work in this area centres on the assumption that people use variations in speaking patterns and other non-verbal behaviours to alter communication efficiency and to transmit interpersonal attitudes towards their interlocutors (von Raffler-Engel, 1978). This communication can take place through the content of the speech itself or through paralinguistic and other nonsophistication, so Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 125 verbal behaviours. The direction of these changes can be towards the speaking patterns of the interlocutor (convergence) or in the opposite direction (divergence). Changes in evaluation of the interlocutor will often accompany these acts of convergence or divergence (e.g., Giles et al., 1973). Whether convergence or divergence occurs during encounters across ethnic lines will depend on a number of factors including the intentions of the speaker (Bourhis & Giles, 1977), the salience of the intergroup difference (Bourhis et al., 1979), and the importance of the issue discussed for one’s group identity (Bond & Yang, 1982). The relative dominance of one group over the other is undoubtedly important as well (Tajfel, 1978). For, when a member of the higher status group uses the lower status language to a bilingual member of that group, this act will probably be seen as accommodating or at least respectful and should be met with some form of accommodation in return (Simard et al., 1976). Conversely, when a member of the lower status group uses the higher status language to a fellow bilingual, this act may be seen as insulting to the ingroup and be met with divergence or some form of ethnic affirmation (Yang & Bond, 1980). Either of these occurrences would be indicated by elevated scores in the cells of the 2 x 2 design where partner ethnicity and language were not consistent. Hong Kong is an ideal place to explore these status-related issues of convergence and divergence. Ninety-eight per cent of Hong Kong’s population is Chinese and the vast majority speak Cantonese, the main dialect of Kwangtung Province, China. As a British Crown Colony, however, English is the prestige language used almost exclusively in government. Furthermore, Westerners, British and American in particular, are disproportionately represented in the higher echelons of government and business, a reality reflected in the students’ stereotype of the typical English-speaking person as more powerful (Gibbons, 1982; Lyczak et al., 1976; Pierson & Fu, 1982). Nevertheless, the social and political situation is harmonious and the local Chinese have retained a strong sense of their ethnic integrity (Lyczak et al., 1976; Yang & Bond, 1980). This context would thus seem ideal as a for intergroup backdrop exploring changes in non-verbal behaviours arising from variations in language and partner ethnicity. As a first step, a non-threatening interview situation was used because it permits an adequate standardisation of the confederate-interviewer’s behaviour. In summary, the primary aim of the present research is to separate the variables of language and interlocutor ethnicity in order to assess their separate effects on perceptual responses and non-verbal behaviours. By so doing, it was hoped that we might also shed some light on speech accommodation processes in Hong Kong bilinguals. Method Subjects It necessary to select subjects who were sufficiently skilled and fluent in an interview comfortably. It was also important to choose English native speakers of Cantonese who had interacted with native speakers of English, was to handle Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 126 that differences in non-verbal repertoires could have been noticed and mastered. To meet these criteria, 64 female undergraduates from The Chinese University of Hong Kong were recruited from the English Department. All these volunteers were in their second, third or fourth year of a programme where over Z/3’s of the staff were either American or British in origin. Furthermore, over 80% had completed their secondary education in an English-medium secondary school. On the criteria of both fluency and exposure, then, these subjects made suitable interviewees. so Interviewers The experimental design required two groups of interviewers whose ethnicity corresponded to the two languages being compared. Furthermore, we believed that they should be older than their interviewees to reinforce the status distinctions implicit in the interview situation. Finally, they should be sufficiently fluent in their second language to make their use of this language during the interview appear reasonable and effective. This last criterion presented some problem. Cantonese is a difficult language, rarely studied by foreigners before they arrive in Hong Kong. After considerable searching, however, two American males were selected, both in their second year of studying Cantonese. Two Chinese male counterparts, actually fourth year undergraduates, were then chosen to complete the team of four interviewers. All were between the ages of 22 and 25; they were paid an hourly wage for their participation. Sex Composition of the Interview Dyads Males were chosen as interviewers and females as interviewees for two reasons. sex composition reinforces the power differential implicit in the interview situation. As Chinese and American cultures differ in their orientation to power (Hofstede, 1980), the chances of seeing effects on the independent variables are maximised. Secondly,. the male-female mix introduces heterosexual considerations into the interaction. Again American and Chinese cultures differ widely on this dimension (Hsu, 1949), maximising the chances of observing an impact for the ethnicity and language variables. First, this Training the Interviewers It . essential for the interviewers to standardise their non-verbal behaviour the interviews. Otherwise the subjects’ own non-verbal responses would during be influenced by the variation in the non-verbal styles of the different interviewers. This variation could well inflate the experimental error and, if such variation were related to interviewer ethnicity, mask or accentuate ethnicity effects. To this end, the interviewers were given a brief survey of the various classes of was Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 127 non-verbal behaviour. They were then coached on a standardised procedure for their paralinguistic, kinesic, and gazing- behaviour during the interview. Specifically, they were instructed to sit facing the subject with both feet on the floor, keeping all movement to a minimum. They were to look at their interview sheet when asking questions but to fixate the subject’s eyes when she held the floor. Interviewers were asked to give subjects 10 seconds to answer a question, and to wait 5 seconds after the answer, before moving to the next question. Interviewers were advised not to nod, smile, laugh or otherwise reinforce the subject when she was speaking. The intent of these guidelines was to standardise the interviewer’s behaviour while communicating an attentive and moderately serious impression. Finally, the interviewers were asked to spend some time practising this procedure with one another. They were also encouraged to acquaint themselves thoroughly with the interview questions so that they would make as few errors as possible during their questioning of the subjects. Procedure - Contacting the subjects. Due to their varying schedules, both subject and interviewer had to consult by phone in order to arrange a satisfactory time for the interview. The interviewer conducted their conversation in the language to be used in the interview. Pre-interview. Upon arriving, the subject was escorted to the interview room by the interviewer. This room was uncluttered, about 5 x 7 m, and relatively soundproof. The subject was seated in a chair placed in front of that of the interviewer. Slightly to the right of the interviewer was a video-camera, aimed at the subject. To her left was a table with a tape-recorder in one corner. Before the interview began, the subject was asked to fill out a questionnaire providing information about her language training, fluency in English, and overseas experience. This questionnaire took about five minutes to complete during which time it was hoped the subject would adapt to her new surroundings, in particular the video-camera. The Interview. The interviewer collected the questionnaire and turned on the tape- and video-recorder. He then gave a standardised introduction which presented the interview topic, encouraged the subject to take her time, and permitted her to sidestep any question she found uncomfortable. The interviewer then proceeded through the questions in order, terminating the interview after five minutes had elapsed on his stop watch since the beginning of the introduction. Two interview topics were used, viz., friendship and Chinese-American relations on campus. These topics obviously vary in terms of their relevance to the ethnic identities of the participants. Should non-verbal and perceptual responses be robust across these two interview topics, however, some confidence could then be placed in the generalisability of effects. Post-Interview. The interviewer turned off the recorders and gave the subject two rating forms, one for her self-perceptions during the interview, the other for Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 128 her perceptions of the interviewer. These forms were written in the language used during the interview. In order to reinforce the stated confidentiality of these measures, the interviewer left the room while the subject completed the forms. She then put the two sheets into an envelope, sealed it, put it in a box, and left the room. Outside, the interviewer thanked the dismissed her. subject for her co-operation and The Interview Schedule The questions were written in English and translated into Chinese. They were then back-translated to assess linguistic equivalence. Changes were made in the original English version until comparable versions were obtained. The Chinese interviewers then helped teach the American interviewers the correct pronunciation of the Cantonese version of the questions. Overview of the Design There were two languages used, Cantonese and English; four interviewers, two Americans and two Chinese; two interview topics, friendship and ChineseWestern relations. This yielded a 2 x 4 x 2 design, with four subjects in each of the 16 conditions. Dependent Variables Self-perceptions. Twenty, bi-polar scales were selected in an effort to tap the fundamental dimensions of interpersonal attractiveness, power, activity, nervousness, and competence (Bond, 1979). They were respectively by the above dimensions; lovable-unattractive, friendly-unapproachable, deep-shallow; strong-weak, important-insignificant, respectable-lowly, confident-shy; alertsleepy, energetic-tired; free-restrained, spontaneous-conforming, involvedbored, relaxed-tense, forthright-inhibited; fluent-halting, precise-vague, awareinsensitive, responsive-unresponsive to the reactions of the interviewer, insightful-commonplace. The contrast between direct and indirect was also tapped, as it probes what is often regarded as a classic Chinese-Western difference in approach. These contrasting adjective pairs were placed at either end of seven-point Likert scales, preceded by the adverb ’very’. Subjects were asked to rate confidentially how they felt during the interview. To reinforce the stated confidentiality, subjects completed the questionnaires in private, sealed the forms in an envelope, and mixed it into a box of other like envelopes. Perceptions of the interviewer. The same 20 scales and format were provided for the subject’s similarly confidential rating of her interviewer. Paralinguistic measures. The audio-tapes provided a variety of measures considered worthy of examination. Since the interviews varied somewhat in length, only the first four minutes following the interviewer’s first question were Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 129 analysed. The variables examined are listed below: 1. Speech rate the total number of syllables uttered per second of speaking - time. 2. Unfilled pause frequency the number of discernible, unfilled pauses per second of speaking time. 3. Average length of unfilled pause. 4. Average response latency average time between the interviewer’s and the question subject’s first word of response (’ah’, ’um’, etc. excluded). 5. Total speaking time number of seconds during which the subject was Filled and unfilled speaking. pauses were excluded. 6. Code switching the number of interviewees who used any English at all during a Cantonese interview or any Cantonese during an English - - - - interview. 7. Filled pause the number of discernible filled pauses (e.g., frequency second of speaking time. ‘um’, ’ah’, ’geh’, etc.) per Kinesic measures. The following measures were taken from the video-tapes: 1. Torso shifts - - total movements of the torso 2. 3. more than 15 cm. This type of during major posture change. typically Self-manipulation of the hands the number of seconds during which the hand is in moving contact with some part of the body. Gestures of the hands proportion of the time when the subject is speaking that she is making some hand movement whose amplitude is movement a occurs - - greater than 3 Other non-verbal cm. measures. video-tapes: 1. Smiling a frequency 2. Gazing while talking - These additional count of each measures are taken from the separate episode. proportion of time the subject gazes at the interviewer while she is holding the floor (i.e., speaking plus pausing). 3. Gazing while listening the proportion of time the subject gazes at the interviewer when he holds the floor (i.e., while he is asking a question). These 13 non-verbal behaviours were tapped because they cover the major types typically sampled in contemporary research and are amenable to reliable measurement (Duncan & Fiske, 1977; Mehrabian, 1972). - the - Reducing ~xpe~°i~aer~ter Bias Whenever possible, the video-tapes were scored without sound. One of the two experimenters would have to move the tape into starting position, thus learning the subject’s language and interviewer condition. The other experimenter would leave the room at this time, however, effectively blinding him. The scorers alternated this procedure throughout the scoring, so that only one of the two scorers knew the subject’s experimental condition. As no hypotheses had been made before hand, this limitation is not regarded as a problem. The study. transcribed by an assistant, blind to the purposes of the She also scored the pauses and response latencies at this time. Another audio-tapes were Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 130 naive assistant counted the number of syllables uttered by the subject from these transcriptions. Results Reliability of the Non-verbal Behaviours Pearson product-moment correlations reliabilities for the non-verbal measures. Table 1 All Inter-judge Reliability were judged to be were run to establish the inter-judge of Non-verbal Behaviours acceptable for further analysis. Analyses of Variance As there were two languages, four interviewers and two topics, 2 x 4 x 2 ANOVA’S were run on each of the dependent variables.- Individuals were, of course, nested within the ethnicity grouping. The F value reported for the ethnicity variable was obtained by extracting that portion of the interviewer variance contributed by the ethnicity grouping. All reported F values had 1 and 48 degrees of freedom. Interviewer Identity Two interviewers from each ethnic group were included in the design to ensure that results had some claim to generalisability across interviewers from that ethnic group. The means indicated that the impact of ethnicity was the same for each interviewer from the Chinese and American groups for each of the significant effects reported below. For affected dependent variables, an interviewer of a given ethnicity yielded a mean score which was closer to the other interviewer of Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 131 ethnicity his than it to was the mean from the interviewers of the other ethnic group. The Topic Variable There effects for this variable. These are not of immediate the present study, however, except as the topic variable might have interacted with the language or ethnicity variables. In none of the effects reported below was there any evidence for such interactions. were numerous concern to Language Effects The language effects listed below. across the three categories of dependent variable are , Table 2 a’This is a Language Effects for Categories of Dependent Variables chi-square value. Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 132 Ethnicity Effects The Table 3 ethnicity effects found are similarly listed below. Ethnicity Effects for Categories of Dependent Variables interactions between Language and Ethnicity Subjects rated their fluency in speaking English before they were actually interviewed. This rating interacted with the language and ethnicity variables, F = 4.27, p<0.05. The interaction was produced by the lowered self-rating of English fluency when interviewees anticipated an interview in Cantonese with an American. The cell mean in this condition was different from that in all three other conditions at the 0.05 level. Self-perceptions of strength showed an interaction effect, F 5.60, p<0.025, such that the bilingual interviewees felt stronger working in Cantonese with an American interviewer or in English with a Chinese interviewer compared to the other two combinations. Given the subjects’ language skills relative to those of their interviewers, this result is hardly unexpected. There was an interaction for ratings of interviewers’ precision, F 5.33, p<0.025. Not surprisingly, interviewers were seen as more precise when using their first language compared to their second language. As might also have been anticipated, ratings of interviewer fluency showed an interaction, F 7.37, p<0.01. Analysing for simple main effects across the language variable revealed that the American interviewers were perceived as less fluent in Cantonese than in English, F 5.58, p<0.02~. The Chinese interviewers were seen as no less fluent in English than Cantonese, ~<l. On the non-verbal measures, language and ethnicity interacted for filled pause frequency, F 5.34p<0.025. When speaking Cantonese, bilingual Chinese used pause fillers more frequently with Americans than with Chinese interviewers, F = 6.25, p<0.025; the ethnicity variable had no impact when English was being spoken, F’< 1. = = = = = Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 133 Discussion The major purpose of this study was to disentangle the variables of interlocutor ethnicity and language in order to assess their separate effects. By using an extensive battery of non-verbal indicators, we were additionally able to determine if these variables have similar or different functional impacts. As the present study is the first adequately reported research to meet these two conditions, only tentative generalisations can be drawn. First, both the ethnicity and the language variables can have considerable impact on non-verbal and perceptual indicators. Studies which confound these variables therefore present the problems of interpretation suggested in the introduction to this study. Secondly, in .the non-verbal realm these impacts appear to be functionally distinct as there was no overlap in the variables affected. Those attempting replications of this design for other groups and languages are therefore encouraged to retain the multivariate approach to measurement adopted here. At this early stage we do not know what these non-verbal changes communicate to observers of the two ethnic groups or whether they represent convergence towards, or divergence from, the non-verbal style of the out-group. Further, it is too early to be sure which of the non-verbal changes would be found in comparisons between any two languages or between any two ethnic groups. In our discussion below, we speculate about the factors which may have given rise to the effects observed in this particular study. Language Effects When speaking English rather than Cantonese with partners of both Chinese and American ethnicity, our bilingual interviewees interrupted their speech more frequently with both filled and unfilled pauses, gazed at the interviewer a smaller portion of the time when he was speaking, and spoke more slowly. Before analysing the meaning of such differences in non-verbal behaviour, an important conceptual issue must be considered. Our bilinguals were English majors, but not equally fluent in their second language as compared with Cantonese. Could some or all of these non-verbal differences have reflected their relative fluency in Cantonese compared to English? If so, the obtained differences would not be informative about non-verbal changes which arise from a change of language in and of itself. Readers will recall that all subjects rated their fluency in speaking English. For those in the English interviews, this self-rating correlated significantly with the total time they talked, r (30) 0.62, p<0.01, and with the frequency of unfilled pausing, r (30) 0.44, ~<0.0~5. Self-rated fluency, however, did not correlate with gazing at the interviewer when he was talking, with frequency of filled pauses, or with speed of speaking. Such an outcome suggests that these three nonverbal differences across languages did not arise because our bilinguals were relatively less fluent in English. Rather, they appear to be manifestations of different non-verbal styles associated with Cantonese and English. = = ’ Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 134 That English is spoken with less speed and more pausing suggests that it may be characterised as a denser, less repetitive encoding tool than Cantonese. As such, it requires a slower delivery with more frequent pauses to ensure efficient decoding. This hypothesis is consistent with our bilinguals’ looking a lesser portion of the time at an English-speaking interlocutor; looking at others produces additional non-linguistic input which probably interferes with decoding a tightly textured language. Consequently, looking away from a speaker addressing you in your second language may be a strategy developed in the early stages of second language learning to avoid this distraction. Such a non-verbal style might then persist well after the second language has been mastered. If so, then further studies comparing bilinguals using any two languages should show such a difference in the second language. These differences in non-verbal behaviour were, however, uncorrelated with the way bilinguals felt when speaking English or to the way they perceived an English-speaking interviewer. None of the non-verbal variables showing language effects was correlated with any of the feeling or perceptual measures showing language effects. These non-verbal changes when using English would seem to be independent of one’s own reaction to using the language and independent of one’s reaction to a conversational partner using that language. This finding strengthens the argument that the non-verbal differences between the languages reflect codes associated with the languages themselves and not with situational factors surrounding the language use. What do the perceptual data suggest about the use of English by Chinese bilinguals? Self-reports indicated that our speakers felt more restrained, shy, and tense when using English. These feeling states appear to reflect an inhibition factor, hardly surprising when one reflects that English is a second language. In addition, however, our subjects reported feeling more respectable and aware when using English. These feelings suggest greater authority, perhaps reflecting the prestige surrounding the English language in the colonial situation characterising Hong Kong. When speaking English, our interviewers were perceived as more sleepy, tired, and bored. This perception of relative inactivity may reflect the lesser speed with which English is spoken, a finding discussed above in comparing its non-verbal features with Cantonese. The greater perceived respectability of interviewers using English is consistent with the greater status surrounding speakers using English. Regardless of that speaker’s ethnic background, English-speaking persons are likely to be better educated and hold more prestigious jobs in Hong Kong. Ethnicity Effects Before considering the ethnicity effects, we must address the issue of generalisability. Specifically, there were only two American and two Cantonese interviewers, so how much confidence can be placed in the ethnicity effects when applied to a whole range of American or Cantonese male partners? We believe the results are reasonable first approximations because they generalised across Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 135 specific personalities of the interviewers in each ethnic group and because the perceptual results are consistent with previous findings in Hong Kong using the matched guise methodology (Lyczak et al., 1976; Gibbons, 1982). These investigators have found that English speakers are perceived as higher on a the power-competence-status dimension than are Chinese speakers. Using the present interactive methodology, we have found that the American interviewers are seen as stronger, more respectable, and more confident than the Chinese interviewers, a similar clear difference along the potency dimension. What then does the pattern of results for the ethnicity variable reveal? In addition to the perceived higher potency of the Americans, they were also regarded as more spontaneous, relaxed, and free. This grouping would appear to tap a dimension of extroversion, a salient component of perception across cultures (Bond, 1979; Osgood et cal., 1975). This difference in perception has not emerged from the matched guise not tapped. studies because the dimension was apparently The differences in non-verbal behaviour displayed towards Americans compared to Chinese interviewers are consistent with these differences in impressions. Talking a lesser proportion of the time to the American interviewers suggests the greater deference seen in the perceptual measures. Likewise, smiling and shifting the torso more are compatible with the greater naturalness the American interviewers were perceived as communicating. Of course increased smiling and shifting could be construed as reflections of general nervousness. Observing the tapes, however, our clear impression was that our Chinese bilinguals had been ’released’ from the stiffness that characterised their behaviour with the Chinese interviewers. Their self-perception as feeling more direct with the Americans fits this construction of the data. Such interpretations must be regarded as tentative, however, because again there were no significant correlations between these critical non-verbal behaviours and the significant perceptual measures. Support for Speech Accommodation Theory A secondary interest animating the present study was to explore the dynamics speech accommodation. What evidence does the present research provide? A salient goal in an interview situation is that of transmitting information. In this regard the problem condition for our Chinese bilinguals occurred when they were interviewed by Americans in Cantonese. The initial telephone contact established that the Americans wished to use Cantonese, but undoubtedly made their lack of complete fluency clear. This impression of linguistic difficulty was reinforced during the interview. The interviewees frequently requested clarification and, as reported earlier, rated the Americans as significantly less fluent in Cantonese than in English. This state of affairs made salient what Thakerar et al. (1982) have called the function of cognitive organisation the need to adjust one’s speech in light of the listener’s perceived capacity to receive. Such was not the case when the interviewees were working with the Chinese interviewers in English; in most cases the interviewers were as skilled in English of - Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 136 the English majors they interviewed. How then did the subjects respond to the problem of communicating with a less-than-fluent American using Cantonese? To begin with, they converged by estimating their own fluency in English as low. This relative downgrading of their own skills probably prepared the interviewee for coping charitably with the difficulties ahead; a person who regards herself as less skilled is likely to be more sympathetic towards someone struggling for mastery than is someone who regards herself as highly competent. During the interview itself, Cantonese bilinguals broke up their speech units with filled pauses more when working in Cantonese with Americans. They also slowed down their speed of speaking Cantonese by 8.4%, even though this reduction was not statistically significant. Both changes functioned to assist the interviewer in decoding their meaning by giving him more time. Language switching was not used as an accommodating strategy we believe, because the power differential implicit in the interview format allowed the interviewer to dictate which language was to be used. Having accepted the language code during the initial telephone contact, the interviewee was at pains to protect the interviewer’s face by continuing to work in that language. Her accommodation was effected in a less obvious fashion by slowing down the density of meaningful discourse she produced. Despite the obvious effort the American interviewers made in using Cantonese, there were no indications that the interviewees evaluated them more positively for this act of convergence. Nor was there any devaluation of the Chinese interviewers who diverged by using English. A similar lack of evaluative changes was also observed in Grujic & Libby’s (1978b) study. Simard ee al. (1976) have argued persuasively for the importance of considering the interlocutor’s attributions for the second language choice. Unfortunately, we did not assess why the interviewees thought the interviewers used the language they did. When their help was initially solicited, the interview was introduced as part of a research programme, however. We believe the obvious attribution to make in such a situation is that the interviewer has been required to speak the language he is using. Under such conditions, the language selected communicates no interpersonal attitude to the interviewee and therefore elicits no reciprocal change in evaluative response. Not setting out primarily to explore speech accommodation theory, we have undoubtedly overlooked some potentially revealing dependent variables. Measures of speech complexity, clarity of pronunciation, ethnocentric content and so forth might also have shown effects. Likewise other perceptual measures could have been used. In fact the potential breadth of the theory is daunting and suggests that it may be time to begin considering what types of variables will be affected by what types of manipulations. as were Conducting Remarks This research has indicated that both language and ethnicity can be functionally distinct and powerful determinants of non-verbal behaviours and person Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 137 in this type of structured interaction. Whether this outcome will found be always probably depends on the language and ethnicities sampled. Further research using this paradigm is thus encouraged to provide information on the following issues: perception 1. How does the non-verbal code of different languages vary? This issue has received little systematic attention by linguists and is important for teachers of second languages who aspire to train for more than verbal facility alone. 2. How do variations in power differences between ethnic groups influence the process and outcomes of such interactions? The vast majority of work in non-verbal communication has involved only minor and unrealistic variations in status, so that the present approach can be used to make our psychology more social and representative. 3. How do the language and ethnicity variables function in more natural intergroup contexts? Ickes (1982) has argued persuasively for loosening the amount of experimental control exercised over dyadic encounters in the laboratory. Such a relaxation may give us a more accurate picture of typical interaction by allowing objective characteristics of the different ethnicities to exercise fuller impact. 4. This research pushed forward our understanding of how interlocutors manage an interaction with a less-than-fluent bilingual. Such contacts are legion in today’s world, but we know very little about how communication changes with variations in partner fluency nor how it should change to achieve greater efficiency. This practical problem deserves more attention. Acknowledgments The authors wish to express their appreciation for funding from the Institute of Social Studies and the Humanities of the Chinese University. We are also grateful for the ever cheerful co-operation from the statistician of the University’s Social Research Centre, Suet-ming Hsu. Detailed and thoughtful comments on the first version of this report were provided by Miles Hewstone, this Journal’s editor, and two anonymous reviewers. Their professional colleagueship helped immeasurably. in Our thanks to Ng Sau Mei and scoring the tapes. Wong Chau Hing for their conscientious efforts References Bond, M. H. (1979) Dimensions of personality used in perceiving peers: Cross-cultural comparisons of Hong Kong, Japanese, American, and Filipino university students. International Journal of Psychology. 14, 47-56. Bond, M. H. and Ho, H. Y. (1978) The effect of the relative status and sex composition of dyad on cognitive responses and non-verbal behaviour of Japanese interviewees. Psychologia, 21, 128-36. Bond, M. H. and Yang, K. S. (1982) Ethnic affirmation vs. cross-cultural accommodation: a Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 138 The variable impact of questionnaire language on Chinese bilinguals in Hong Kong. Journal of Cross-Cultural Psychology, 13, 169-85. Bourhis, R. Y. and Giles, H. (1977) The language of intergroup distinctiveness. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Bourhis, R. Y., Giles, H. Leyens, J. P. and Tajfel, H. (1979) Psycholinguistic distinctiveness: Language divergence in Belgium. In H. Giles and R. N. St. Clair (eds), Language and Social Psychology. Baltimore: University Park Press. Duncan, S. and Fiske, D. W. (1977) Face-to-Face Interaction: Research, Methods, and Theory. Hillsdale, N. J.: Erlbaum. Elzinga, R. H. (1978) Temporal organization of conversation. Sociolinguistics Newsletter, 9, 29-31. Ervin-Tripp, S. (1964) An analysis of the interaction of language, topic, and listener. American Anthropologist, 66, 86-102. Gallois, C. and Markel, N. N. (1975) Turn taking: Social personality and conversational style. Journal of Personality and Social Psychology, 31, 1134-1140. Gibbons, J. (1982) The issue of the language of instruction in the lower forms of Hong Kong secondary schools. Journal of Multilingual and Multicultural Development, 3, 117-128. (1983) Attitudes towards languages and code-switching in Hong Kong. Journal of Multilingual and Multicultural Development, 4, in press. Giles, H. and Smith, P. (1979) Accommodation theory: Optimal levels of convergence. In H. Giles and R. St. Clair (eds), Language and Social Psychology. Baltimore: University Park Press. Giles, H., Taylor, D. M. and Bourhis, R. Y. (1973) Towards a theory of interpersonal accommodation through language use. Language in Society, 2, 177-192. Grujic, Z. and Libby, W. L. Jr. (1978a) Nonverbal aspects of verbal behaviour in French Canadian French-English bilinguals. Paper presented at the American Psychological Association Convention, Toronto. Grujic, Z. and Libby, W. L. Jr. (1978b) When do French Canadians regard themselves as inferior? Paper presented at the American Psychological Association Convention. Toronto. Hofstede, G. (1980) Culture’s Consequences: International Differences in Work-Related Values. Beverly Hills: Sage. Hsu, F. L. (1949) Suppression versus repression: A limited psychological interpretation of four cultures. Psychiatry, 12, 223-242. Ickes, W. (1982) A basic paradigm for the study of personality, roles and social behavior. In W. Ickes and E. S. Knowles (eds) Personality Roles and Social Behavior. New York: — Springer-Verlag. Y. (1978) La communication non-verbale chez les bilangues. Canadian Journal of Behavioural Science, 10, 130-40. LaFrance, M. and Mayo, C. (1978) Cultural aspects of non-verbal communication. International Journal of Intercultural Relations, 2, 71-89. Lyczak, R., Fu, G. S. and Ho, A. (1976) Attitudes of Hong Kong bilinguals towards English and Chinese speakers. Journal of Cross-Cultural Psychology, 7, 425-438. Mehrabian, A. (1972) Nonverbal Communication. Chicago: Aldine. Osgood, C. E., May, W. H. and Miron, M. S. (1975) Cross-Cultural Universals of Affective Meaning. Urbana, Illinois: University of Illinois Press. Pierson, H. and Fu, G. (1982) Report on the linguistic attitudes project in Hong Kong and its relevance for second language instruction. Language Learning and Communication, LaCroix, J. M. and Rioux, 1,203-216. Simard, L., Taylor, D. M. and Giles, M. (1976) Attribution processes and interpersonal accommodation in a bilingual setting. Language and Speech, 19, 374-87. Tajfel, H. (1978) Social categorization, social identity, and social comparison. In H. Tajfel (ed), Differentiation between Social Groups. London: Academic Press. Thakerar, J. N., Giles, H. and Cheshire, J. (1982) Psychological and linguistic parameters Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014 139 of speech accommodation theory. In C. Fraser and K. R. Scherer (eds), Advances in the Social Psychology of Language. Cambridge: Cambridge University Press. von Raffler-Engel, W. (1976) Linguistic and kinesic correlations in code-switching. In W. G. McCormack and S. Wurm (eds), Language and Man: Anthropological Issues. The Hague: Mouton, 1976. — (1978) The unconscious element in intercultural communication. In R. St. Clair and H. Giles (eds), The Social and Psychological Contexts of Language. Hillsdale, N.J.: Erlbaum. Wiens, A. N., Manaugh, T. S. and Matarazzo, J. D. (1976) Speech and silence behavior of bilinguals conversing in each of two languages. International Journal of Psycholinguistics, 5, 79-94. Yang, K. S. and Bond, M. H. (1980) Ethnic affirmation by Chinese bilinguals. Journal of Cross-Cultural Psychology, 11, 411-25. Downloaded from jls.sagepub.com at Guangdong University of Foreign Studies on March 10, 2014
© Copyright 2026 Paperzz