Chen and Bond 2010 Two Languages Two Personalities - Examining Language Effects on the Expression of...

download Chen and Bond 2010 Two Languages Two Personalities - Examining Language Effects on the Expression of Personality in a Bilingual Context

of 15

description

Chen and Bond

Transcript of Chen and Bond 2010 Two Languages Two Personalities - Examining Language Effects on the Expression of...

  • Article

    Personality and Social Psychology Bulletin36(11) 1514 1528 2010 by the Society for Personality and Social Psychology, IncReprints and permission: sagepub.com/journalsPermissions.navDOI: 10.1177/0146167210385360http://pspb.sagepub.com

    Two Languages, Two Personalities? Examining Language Effects on the Expression of Personality in a Bilingual Context

    Sylvia Xiaohua Chen1 and Michael Harris Bond1

    Abstract

    The issue of whether personality changes as a function of language is controversial. The present research tested the cultural accommodation hypothesis by examining the impact of language use on personality as perceived by the self and by others. In Study 1, Hong Kong Chinese-English bilinguals responded to personality inventories in Chinese or English on perceived traits for themselves, typical native speakers of Chinese, and typical native speakers of English. Study 2 adopted a repeated measures design and collected data at three time points from written measures and actual conversations to examine whether bilinguals exhibited different patterns of personality, each associated with one of their two languages and the ethnicity of their interlocutors. Self-reports and behavioral observations confirmed the effects of perceived cultural norms, language priming, and interlocutor ethnicity on various personality dimensions. It is suggested that use of a second language accesses the per-ceived cultural norms of the group most associated with that language, especially its prototypic trait profiles, thus activating behavioral expressions of personality that are appropriate in the corresponding linguistic-social context.

    Keywords

    bilingual personality, language prime, cultural accommodation, cultural priming, Chinese-English bilinguals

    Received April 19, 2009; revision accepted June 9, 2010

    Do bilinguals have two personalities? This intriguing issue is controversial in the fields of personality and social psychol-ogy, since research testing language effects on personality has been scarce and equivocal. On one hand, personality has long been conceptualized as stable over time and consistent across situations (e.g., Byrne & Kelly, 1981; Pervin, 1980). From this perspective, language is not a contextual factor that could shift a bilinguals personality, but merely a tool that permits the expression of underlying traits. On the other hand, studies on cultural priming have documented language effects on values, self-concept, relationality, and cognition (Oyserman & Lee, 2008), but only a few published studies directly address the issue of whether bilinguals exhibit cross-language differ-ences in personality.

    Ervin (1964) administered the Thematic Apperception Test verbally in French and English to French-English bilinguals using the same pictures across a 6-week interval. The stories were tape-recorded, transcribed, and content-analyzed. Achievement themes were more commonly found in the English stories among female bilinguals, whereas verbal aggression against peers, autonomy, or withdrawal from others were more common in the French stories. Based on these

    findings, the author inferred that bilinguals had two personali-ties, at least in terms of their verbal productions.

    Hull (1996) tested three groups of immigrants, namely, Chinese, Korean, and Mexican Americans, who had learned English after they had immigrated to the United States. Using the California Psychological Inventory (CPI) as a measure of personality, the author adopted a within-subjects, repeated measures design by administering the CPI twice over intervals of 5 to 15 days, in English and in the bilinguals native lan-guages of Chinese, Korean, or Spanish, respectively. Signifi-cant differences were found on most scales across the three cultural groups to the same English-language version. This result could be considered as showing cultural differences in traits. Between-language, within-group differences were also detected and could be considered as demonstrating cross-language differences in traits within the same respondent.

    1The Hong Kong Polytechnic University, Kowloon, Hong Kong

    Corresponding Author:Sylvia Xiaohua Chen, Department of Applied Social Sciences, Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong KongEmail: [email protected]

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1515

    Given that Hull (1990) had earlier found similar results with a different measure of personality, the author concluded that these studies had demonstrated cross-language differences in personality, namely, two languages, two personalities (Hull, 1990).

    Recent research in support of personality shift as a func-tion of language was conducted by Ramrez-Esparza, Gosling, Benet-Martnez, Potter, and Pennebaker (2006), examining cultural frame switching (CFS; Hong, Morris, Chiu, & Benet-Martnez, 2000) in the domain of personality. Using the Big Five Inventory (BFI; John, 1990; John & Srivastava, 1999), they found that in English, Spanish bilinguals responded in a more American manner, reporting themselves to be more extraverted, agreeable, and conscientious, but less neurotic, a personality pattern consistent with the modal personality type of Americans among whom the Spanish bilinguals had learned their English. The researchers thus concluded that language activated CFS for these personality dimensions, not because of translation effects or self-enhancement tendencies, but because of cultural shifts appropriate to the language com-munities involved.

    Bilingual PersonalityButcher and colleagues (e.g., Butcher, 1996, 2004) have exten-sively investigated language effects on personality assessment, particularly on the Minnesota Multiphasic Personality Inven-tory (MMPI) and MMPI-2, and have emphasized the impor-tance of understanding cultural influences on personality. For effective cross-cultural applications and adaptations of per-sonality tests, they suggested using bilingual test-retest studies, in which the English version and the target-language version are administered to a group of bilinguals within an interval of 1 to 2 weeks (Butcher, Mosch, Tsai, & Nezami, 2006). The bilingual test-retest design can remove the variance stemming from individual differences and estimate the variance due to language and all interactions involving language.

    As bilinguals differ in the contexts where they acquire their two languages, what linguistic and social features they respond to in language manipulations may well differ (Ervin, 1961; Weinrich, 1953). Coordinate bilinguals acquire and use their first and second languages in separate and distinct cultural environments. For instance, some immigrants learn their English after arriving in the United States. In contrast, com-pound bilinguals acquire their two languages in the same cultural environment. A widespread example is individuals who learn a second language in local schools while residing in their mother-tongue culture. Previous studies have sampled coordinate bilinguals to investigate personality differences across linguistic contexts (e.g., Ervin, 1964; Hull, 1996). The two languages of these bilinguals activate corresponding cul-tural scripts, and behavioral expressions associated with those cultural systems are thus more readily elicited from coordinate bilinguals (Ervin & Osgood, 1954).

    Conceivably, these observable differences in bilingual per-sonality could also be attributed to an acculturation effect (McCrae, Yik, Trapnell, Bond, & Paulhus, 1998). First, physi-cally relocating from one culture to another may bring about changes in immigrants personality or behavioral expressions associated with different social roles and developmental stages assumed in the second-language culture. Second, language may influence the social judgment of respondents, a term coined by McCrae et al. (1998), referring to processes of person perception, social comparison, and self-presentationin ways that affect responses to personality questionnaires (p. 1050). That is, the implicit standards set by the norms of their host culture and their own ethnic communities may lead immigrants to rate themselves accordingly in their host and native languages, thereby choosing response options to describe themselves in culturally appropriate yet distinguish-able ways. Hence, their cross-language differences in reported traits might reflect different expression of personality associ-ated with the social roles attached to a given language.

    Cultural Accommodation HypothesisIn their meta-analysis on cultural priming, Oyserman and Lee (2008) identified 10 studies using language as a priming method, but the overall effect of language priming was rela-tively weak. It remains unclear what exactly is being evoked by language use, as the meanings and structure of language are highly contextualized by political and historical factors (Oyserman & Lee, 2007).

    We use the cultural accommodation hypothesis to ascertain what processes are being evoked by language use and suggest that language activates bilinguals perception of differences in cultural norms, which then guide their behavior and affect the expression of their personality. As such, one of the mecha-nisms underlying language effects on personality is the per-ceived norms of prototypic personality characteristics in the corresponding cultural group. Because language functions as a communication tool, the use of a given language creates a social situation for receiving and delivering messages by means of either written or spoken channels. In this interactive situation, a bilinguals expression of intrapersonal disposi-tions is contingent on both the salient features of the social situation, that is, the first versus second language being used, and the expected dispositions of their interlocutor, that is, the perception of the prototypic personality characterizing the interlocutors cultural group. The press to make culturally congruent responses in a cooperative interpersonal exchange then motivates the bilingual to accommodate his or her expres-sions to the interlocutors cultural norms, personality, values, and beliefs.

    Earlier studies in line with the cultural accommodation framework have demonstrated that people change their communicative behaviors and converge to or diverge from the outgroup (e.g., Gallois, Ogay, & Giles, 2005; Giles &

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1516 Personality and Social Psychology Bulletin 36(11)

    Ogay, 2006). Mixed findings on cultural accommodation have been documented for the expression of values and self-construals (e.g., Bond, 1983; Bond & Yang, 1982; Kemmelmeier & Cheng, 2004; Ralston, Cunniff, & Gustafson, 1995; Watkins & Gerong, 1999; Yang & Bond, 1980). How-ever, such evidence for personality traits is still scarce. The equivocal pattern of results, that is, two languages eliciting two personalities versus weak language effects in priming studies, signifies the need to make finer grained analyses of language and culture.

    The Present StudyThe objectives of the present study are to adopt multiple meth-ods to disentangle the effects of language and culture on per-sonality in the bilingual context and to uncover the mechanisms underlying such language effects. The limitation of self-reports on ones own personality, as pointed out by McCrae et al. (1998), is that

    the same individual is both target and rater, and we do not yet know whether culture has affected the personality characteristics of the target or the social judgment char-acteristics of the rater, or both. As long as personality assessment is limited to self-reports, these questions cannot be answered, because substantive variance is confounded with method variance. (p. 1050)

    To disentangle target and rater judgment calls for assessment procedures that go beyond self-reports.

    We conducted two studies among Chinese-English bilin-guals in Hong Kong to test the cultural accommodation hypothesis on personality and examine whether compound bilinguals also exhibit cross-language differences in personal-ity. Hong Kongs colonial background and current conditions provide unique samples of compound bilinguals, who identify with both Chinese and Western cultures (Chen, Benet-Martnez, & Bond, 2008). Using written measures, Study 1 adopted a between-subjects design to identify bilinguals perceptions of their own traits and prototypic traits in Chinese- and English-speaking cultures. Study 2 used a within-subjects, repeated measures design with multiple time points to evaluate whether bilinguals exhibit their perceived prototypic traits of the culture primed by spoken language and interlocutor ethnicity.

    At Time 1 and Time 2, we measured bilinguals self-perceived traits, perceived traits of native Chinese speakers, and perceived traits of native English speakers, in their two languages. At Time 3, bilinguals carried out actual conversa-tions in both languages with native Chinese and English speak-ers, separately. These conversations were videotaped and later rated independently by two bilingual observers. By incorporat-ing an intrapersonal variable (perceived norms), an interper-sonal variable (interlocutor ethnicity), and a contextual

    variable (language used) in the same study, we intend to take a social view of language effects on the expression of person-ality and gain insight into bilingual personality, using the Person Situation paradigm.

    Study 1Contextualized personality has usually been studied in role-related contexts, suggesting that personality shift is a function of changes in social roles (e.g., Heller, Watson, Komar, Min, & Perunovic, 2007). In the bilingual context, personality dif-ferences could arise from varying social roles typically enacted across languages. Often, bilinguals adjust their behaviors and are perceived differently because they undertake one role in their first language and another role in their second language, such as immigrants who speak their native language with family members at home and English with colleagues at work. So, if they report more communal traits in their native language than in English but more agentic traits in English than in their native language, it is difficult to tease apart the effects of language, culture, and social role. The present research sam-pled university students in Hong Kong who had not experi-enced a physical relocation from one culture to another. These bilinguals had spent most of their time on campus as students using both Chinese and English, thus reducing switch-over in roles associated with language use.

    In McCrae et al.s (1998) study, Hong Kong Chinese (who responded in Chinese and English) or Canadian-born Chinese (who responded in English) scored lower on Extraversion and Openness to Experience but higher on Agreeableness and three facets of Neuroticism, such as Anxiety, when compared to North Americans (who responded in English); the findings were mixed for the Conscientiousness factor. Yik, Bond, and Paulhus (1998) found that Hong Kong Chinese (who responded in Chinese) perceived themselves to be lower on the personal-ity dimensions of Emotional Stability, Sociability, Helpful-ness, Application, and Restraint (Yik & Bond, 1993) than their peers perceived them to be on those same dimensions. These findings provided a basis for us to predict the direction of effects in the present study.

    Perception effects. We anticipated that Hong Kong bilinguals would perceive native speakers of English as higher on Extra-version and Openness to Experience than native speakers of Chinese, and native speakers of Chinese as higher on Agree-ableness and Neuroticism than native speakers of English. Given the consistently higher academic achievement of Chinese students (Stevenson & Lee, 1996), we expected that native speakers of Chinese would be perceived as higher on Conscientiousness than native speakers of English.

    Language effects. Accordingly, when responding in Chinese, bilinguals would exhibit traits consistent with their own per-ceptions of typical Chinese speakers; in English, they would exhibit traits consistent with their perceptions of typical English speakers.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1517

    Method

    Participants. The sample consisted of 213 (104 males and 109 females) university students, with a mean age of 20.58 (SD = 1.51). Participants were Chinese and English bilinguals (all of Chinese descent) recruited from the Chinese University of Hong Kong and invited to take part in the study on a vol-untary basis. Born and brought up in Chinese culture, partici-pants were all compound bilinguals.

    Measures. The original English version of the instruments was translated into Chinese by bilinguals using the back-translation method (Brislin, 1986) and verified by another bilingual. The equivalence of meaning on all items was ensured through consultations with native Chinese and English speakers.

    Big Five Inventory (BFI; John, 1990; John & Srivastava, 1999). Three versions of the BFI were employed to measure the per-ceived personality of the self, typical native speakers of Chinese, and typical native speakers of English (primarily British and Americans). Responses to each target were anchored on 5-point Likert-type scales ranging from 1 (strongly disagree) to 5 (strongly agree). As a well-established measure of personality, the BFI uses 44 short phrases to assess the most prototypical traits associated with the Big Five factors (see John, 1990), namely, Extraversion, Agreeableness, Consci-entiousness, Neuroticism, and Openness to Experience. The reliability coefficients of the five factors for each language and each target are presented in Table 1.

    Language proficiency and usage (Benet-Martnez & Haritatos, 2005). Participants were asked to report on both their first and second languages in the following domains: (a) language ability (e.g., Rate your overall Chinese language ability), (b) past and present language usage (e.g., How much do you use/have used Chinese to speak with your parents?), and (c) media exposure (e.g., How often do you watch TV shows/movies in Chinese?). The two scales consist of 14 items rated on 6-point Likert-type scales, with those items tapping lan-guage ability ranging from 1 (very little ability) to 6 (very high ability) and the rest from 1 (almost never) to 6 (very often).

    In this study, Cronbachs alphas for Chinese and English pro-ficiency and usage were .68 and .74, respectively, for the Chinese group and were both .78 for the English group.

    Procedure. The questionnaire sets were administered in quiet classrooms to participants in small groups. We randomly assigned about half of the participants (n = 105) to complete the English version (52 males and 53 females) and the other half (n = 108) to complete the Chinese version (52 males and 56 females). They also reported demographic information, including age and gender. We assured participants of their confidentiality and collected anonymous self-reports on the previously described instruments.

    Results and DiscussionDescriptive statistics, including means and standard deviations of the measures, are presented in Table 1. First, we checked the equivalence of the two groups derived from the language manipulation. No significant differences were found in lan-guage proficiency and usage between the two groups assigned to the English and Chinese versions, ps > .05. Thus, any dif-ferences in responding as a function of questionnaire language did not arise because of differential language ability or usage.

    To test for language effects and perception effects on the five personality factors as hypothesized, we conducted five sets of 2 3 ANOVAs (see Table 2). The dependent variables were the five personality factors, as measured by the BFI. The between-subjects factor was language group (English and Chinese); the within-subjects factor was type of target (self-perception, perception of native Chinese speakers, and percep-tion of native English speakers). The target main effects and Language Target interaction effects were tested using the multivariate criterion of Wilkss lambda.

    For the personality factor of Extraversion, the target main effect was significant, F(2, 210) = 204.68, p < .001, whereas the Language Target interaction was not, F(2, 211) = 1.46, p > .05. The univariate test associated with the language main effect was not significant, F(1, 211) = 1.46, p > .05. Simple main effect analyses indicated that across the two

    Table 1. Means, Standard Deviations, and Reliability Coefficients () of Participant Ratings for the Big Five Inventory in Study 1

    Self-perceptionPerception of native Chinese

    speakersPerception of native English

    speakers

    Chinese English Chinese English Chinese English

    Factor M SD M SD M SD M SD M SD M SD

    Extraversion 2.82 0.72 .80 2.92 0.64 .77 3.07 0.62 .74 3.15 0.54 .68 3.96 0.48 .69 3.93 0.42 .60Agreeableness 3.29 0.53 .61 3.49 0.49 .64 2.70 0.50 .62 2.68 0.60 .78 3.27 0.49 .66 3.23 0.50 .69Conscientiousness 3.08 0.65 .78 3.23 0.53 .69 3.30 0.55 .73 3.42 0.52 .72 3.11 0.57 .82 2.99 0.51 .73Neuroticism 3.20 0.82 .86 3.15 0.73 .81 3.22 0.54 .71 3.38 0.50 .62 2.47 0.43 .66 2.68 0.38 .48Openness to

    Experience3.05 0.56 .69 3.17 0.61 .75 2.74 0.56 .72 2.75 0.50 .65 3.77 0.41 .67 3.79 0.42 .65

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1518 Personality and Social Psychology Bulletin 36(11)

    languages, participants perceived themselves as significantly less extraverted than typical native speakers of Chinese, who were in turn less extraverted than typical native speakers of English, ps < .001.

    For Agreeableness, the target main effect was significant, F(2, 210) = 114.05, p < .001, as well as the Language Target interaction effect, F(2, 210) = 4.82, p < .01; the language main effect was not significant, F(1, 211) = 1.15, p > .05. We then conducted t tests to follow up the significant interaction. Using Holms sequential Bonferroni approach to control for fami-lywise error rate across multiple hypothesis tests (Green & Salkind, 2003), two of the three tetrad comparisons were sig-nificant. The comparison with the smallest p value was the mean difference across the two languages between the self and typical native speakers of English, t(211) = 2.64, p = .009, which was less than = .05/3 = .0167. The comparison with the next smallest p value was the mean difference across lan-guages between the self and typical native speakers of Chinese, t(211) = 2.45, p = .015, which was less than = .05/2 = .025. The last comparison was the mean difference across languages between typical native speakers of Chinese and native speakers of English, t(211) = .19, p = .847, which was more than an = .05/1 = .05, and thus nonsignificant. In other words, across languages, participants perceived themselves as significantly more agreeable than typical native speakers of English, who were in turn perceived as significantly more agreeable than typical native speakers of Chinese, ps < .05. This difference was greater between languages in self-perceptions, p < .05, where participants reported themselves to be more agree-able when responding in English than in Chinese.

    For Openness to Experience, only the target main effect was significant, F(2, 210) = 231.70, p < .001; neither the Language Target interaction effect nor the language main effect was significant, F(2, 210) = .60, and F(1, 211) = 1.67, respectively, ps > .05. Across languages, participants perceived typical native speakers of English as significantly more open than native speakers of Chinese, p < .001, whereas self-ratings were significantly lower than those for native speakers of English but higher than those for native speakers of Chinese, ps < .001.

    For Conscientiousness, the target main effect was signifi-cant, F(2, 210) = 17.49, p < .001, as well as the Language Target interaction effect, F(2, 210) = 3.41, p < .05; the main effect of language was not significant, F(1, 211) = 1.31, p > .05. Interaction comparisons showed different patterns for conscientiousness across English and Chinese. In English, participants perceived typical native speakers of Chinese as significantly more conscientious than typical native speakers of English, with themselves scoring between the two groups. In Chinese, they perceived typical native speakers of Chinese to be more conscientious than themselves and typical native speakers of English, with no difference between these latter two targets.

    For Neuroticism, the main effects of target and language were significant, F(2, 210) = 133.54, p < .001, and F(1, 211) = 5.26, p < .05, respectively, whereas the Language Target interaction effect was not significant, F(2, 210) = .98, p > .05. Participants perceived typical native speakers of Chinese as significantly more neurotic than typical native speakers of English, p < .05, with this effect stronger in English than in Chinese, and themselves as closer to typical native speakers of Chinese, p > .05. They perceived themselves as slightly more neurotic in Chinese than in English, but at the trend level, p = .073.

    In summary, target of perception effects were evident for all five factors, consistent with our hypotheses (except for the direction of Agreeableness). Native speakers of English were perceived as higher on Extraversion and Openness to Experience than those of Chinese, whereas native speakers of Chinese were perceived as higher on Neuroticism and Conscientiousness than those of English. Yet, bilinguals per-ceived native speakers of English as higher on Agreeableness than those of Chinese, perhaps because the native speakers of English whom they encountered in Hong Kong are mostly friendly, sociable, and supportive, especially in academic settings (Bond, 1986). Language effects were smaller relative to target effects, and only significant for Neuroticism. The directions of self-rated Extraversion and Openness to Experi-ence were in line with prototypic cultural traits, higher in English than in Chinese at the trend level, but did not reach statistical significance.

    Study 1 demonstrated the effects of perception on dif-ferent cultural groups and yet could not confirm the cultural accommodation hypothesis directly. The stimulus of a

    Table 2. ANOVA for Self-Ratings in Study 1

    Source df F p p2

    Extraversion Target 2 204.68 < .001 .66 Language 1 1.46 > .05 .01 Target Language 2 0.76 > .05 .01Agreeableness Target 2 114.05 < .001 .52 Language 1 1.15 > .05 .01 Target Language 2 4.82 < .01 .04Conscientiousness Target 2 17.49 < .001 .14 Language 1 1.31 > .05 .01 Target Language 2 3.41 < .05 .03Neuroticism Target 2 133.54 < .001 .56 Language 1 5.26 < .05 .02 Target Language 2 0.98 > .05 .02Openness to Experience Target 2 231.70 < .001 .69 Language 1 1.67 > .05 .01 Target Language 2 0.60 > .05 .01

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1519

    second language in self-report personality inventories may not constitute a social situation strong enough to evoke the motive of accommodating to the norms of the corresponding culture. Speaking a second language with a real partner, how-ever, creates a stronger prime for assessing the effects of lan-guage and interactions involving language on the behavioral expression of personality.

    Study 2Study 1 used self-report inventories to examine personality differences in the bilingual context. However, such inventories are limited to assessing self-perceived personality rather than personality as perceived by others. In a social context, per-sonality also presents itself to the eyes of the beholder. Using the matched-guise technique (Lambert, Hodgson, Gardner, & Fillenbaum, 1960) as an unobtrusive method for examining person perception, Bond (1985) presented audiotaped English or Cantonese (the spoken language of Chinese in Hong Kong) passages to Chinese bilinguals, with associated photos of either Chinese or British male speakers. Speakers of Cantonese, regardless of ethnicity, were viewed as humble, honest, and friendly; British speakers, regardless of language, as compe-tent. Thus, both language and ethnicity effects were observed for person perception when these two variables, normally confounded, were disentangled.

    Bilingual competence may reduce cross-language differ-ences, as it makes shifts between languages easier. Bilinguals who are not fluent in their second language may feel inhibited because of their limited ability to express themselves and perhaps thereby distort the expression of their personality. Fluent bilinguals switch languages and communicative behav-iors with ease. In Study 2, we recruited fluent bilinguals and used behavioral observation in addition to self-reports to assess self-perceived personality and personality as perceived by others.

    Based on the findings of Study 1, Chinese-English bilinguals perceived Extraversion and Openness to Experience as dis-tinctively Western traits. Additionally, since McCrae et al. (1998) found that North Americans scored higher on Assertive-ness than did Hong Kong Chinese and Canadian-born Chinese, we focused our investigation on these three Western traits that could demonstrate the direction of accommodation.

    Perception effects. We hypothesized that Hong Kong bilin-guals would perceive native speakers of English as higher on Extraversion, Openness to Experience, and Assertiveness than native speakers of Chinese.

    Language effects. Accordingly, when responding in English, bilinguals would present traits consistent with their own per-ceptions of typical English speakers, that is, higher on Extra-version, Openness to Experience, and Assertiveness, than when responding in Chinese.

    Ethnicity effects. The cultural accommodation hypothesis led us to predict the effects of interlocutor ethnicity derived

    from bilinguals expected personality dispositions of their conversational partners. When communicating with native speakers of English, bilinguals would exhibit typical Western traits, that is, score higher on Extraversion, Openness to Expe-rience, and Assertiveness, than when communicating with native speakers of Chinese.

    MethodParticipants. To control for possible gender effects, all inter-

    viewers were males, and all interviewees were females. The sample consisted of 76 female Chinese-English bilinguals (all of Chinese descent) from the Chinese University of Hong Kong, with a mean age of 20.34 (SD = 1.49). Participants were recruited on the basis of their English and Chinese pro-ficiency so that they were competent in both languages. We selected those whose grades were C or above both in the Hong Kong Certificate of Education Examination and in the Use of English examination of the Advanced Level Examination, a selection criterion used in Hui and Chengs (1987) study. Only those scoring above average on both lan-guage tests were selected.

    Interviewers. Four male interviewers were recruited, two Caucasians and two Hong Kong Chinese, all fluent in both English and Cantonese. We screened and interviewed Chinese and Caucasian candidates not only to evaluate their bilingual competence but also to ensure that their facial features resem-ble prototypic Chinese and Anglo-Saxons. Those with Asian ancestry were excluded as Caucasian interviewers, even if they were born in the United States with English as their first language, because their mixed appearance would undercut the priming of ethnicity.

    All interviewers were between 20 and 30 years old, com-parable in height and physical attractiveness. The reason for selecting young interviewers was that Chinese traditions socialize young people to listen more and speak less in front of seniors, but participants would feel a lesser need to restrain their talking in front of young interviewers. The interviewers and all research assistants were blind to the hypotheses of this study.

    The interviewers were trained to standardize their nonverbal behaviors, such as their paralinguistics, kinesics, and gazing, during the interviews. Meetings were held to clarify the inter-view procedure, and four sets of interview scripts in both English and Chinese, for a total of eight versions, were devel-oped, so that the interviewers could practice beforehand. We then conducted practice sessions to allow interviewers to rehearse with student helpers and to test the equipment used and the flow of the interview process. Comments and sugges-tions were given on interviewers pronunciation, intonation, expressions, and gestures, and so on. They were instructed to wear white shirts and minimize their verbal and nonverbal reactions to control for possible experimenter effects across interviewers.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1520 Personality and Social Psychology Bulletin 36(11)

    ProcedureWave 1. Participants were randomly assigned to two

    groups. Half completed the English version of the written measures; half, the Chinese version. They were asked to assess their own traits, perceived traits of typical Chinese native speakers, and perceived traits of typical native English speakers. All participants were instructed to indicate their age, year and major of study, and length of stay in English-speaking countries.

    Wave 2. After 2 to 3 weeks, members of the two groups completed the other language version of the same measures completed previously. Again, they were asked to assess their own traits, perceived traits of typical native Chinese speakers, and perceived traits of typical native English speakers.

    Wave 3. Three weeks after completing the second set of personality inventories, participants were interviewed by one of the Caucasian interviewers and one of the Chinese interviewers in English and Cantonese separately, for about 10 min with each interviewer in each language. In other words, each participant took part in four conditions, that is, talking with a Caucasian interviewer in English, with a Cau-casian interviewer in Cantonese, with a Chinese interviewer in English, and with a Chinese interviewer in Cantonese. The order of interview language and interviewer ethnicity was counterbalanced.

    The interviews were recorded by camcorders, with each condition saved in a separate video file for later review. Thus, there were four video files for each participant, one for each condition. At the end of the interview, participants were asked to answer questions tapping their self-perceived differences in personality across English and Chinese. After all partici-pants were interviewed, we recruited two Chinese-English bilingual observers to review the video files separately and complete a rating sheet independently for each participant under each condition. To avoid possible interference effects, the observers were not given all the files at the same rating session, so that they did not review the same participant in two or more conditions within one rating session. As the camera was only focused on the participants, the interviewers were not filmed and their ethnicity was not revealed to the observers.

    MeasuresLanguage proficiency and usage (Benet-Martnez & Haritatos,

    2005). This measure was the same as used in Study 1. In the present study, the alphas for Chinese and English proficiency and usage were, respectively, .67 and .71 for the Chinese version and .77 and .72 for the English version.

    Perceived personality. The Sino-American Person Perception Scale (SAPPS; Yik & Bond, 1993) was adopted to assess the perceived personality of the self, typical native speakers of Chinese, and typical native speakers of English. The SAPPS assesses traits on a 7-point scale, based on the Western Five Factor Model (Norman, 1963; see also McCrae & Costa, 1985,

    1987) and indigenous Chinese adjective checklists (Lew, 1985; Yang & Bond, 1990). The scale has both Chinese and English versions consisting of 32 bipolar adjectives on eight orthogonal dimensions of personality: Emotional Stability, Extraversion, Application, Openness to Experience, Assertiveness, Restraint, Helpfulness, and Intellect.

    The reason for adopting this scale in Study 2 is that the SAPPS captures socially relevant traits and has been well validated as a comprehensive measure of personality percep-tion with Chinese populations (e.g., Chen, Bond, & Fung, 2006). In addition to the personality dimensions similar to those of the five-factor model, the SAPPS dimension of Application, for example, refers to being hard-working, diligent, and practi-cal, and that of Restraint denotes being cautious, dignified, and thorough. The alphas of the eight factors for each language and each target are presented in Table 3.

    Interview scripts. To standardize interview questions across the four conditions, four sets of questions were designed in parallel forms and in both languages, such that each participant answered comparable but not identical questions across the four conditions. The content was general information about the participants, their hobbies, and their social activities. After a brief introduction to the procedure, the interviewer asked the participant to talk about sports, movies, songs, and paint-ings, respectively, for each of the four conditions, and asked the participant to describe her favorite sport, movie, song, and painting and why she liked it, as well as her favorite sport star, movie star, singer, and painter, respectively. In addition to hobbies, the interviewer also asked the participant to talk about people around her, such as parents, relatives, friends, and teachers, and her social activities. We used easy and informal topics of conversation so that they resembled informal con-versations rather than formal interviews.

    Observer ratings. To provide a tangible scoring scheme for observer ratings, the authors selected a representative word out of four items defining each personality dimension (e.g., extraverted, open, and assertive). If a representative word was difficult to find from the four items, a key word was used to represent the personality factor (e.g., restrained for the Restraint factor). The observers independently completed a rating sheet after they reviewed one video file of a participant in each of the four conditions. The questions included key items for the eight personality dimensions as measured by the SAPPS. They were asked to judge the participants traits from the video segment (e.g., Did the participant seem extro-verted?) and to circle their answers on a 6-point scale, ranging from 1 (strongly disagree) to 6 (strongly agree). Thus, four sets of ratings were collected for each participant, one set for each condition.1

    Postinterview questions. After the interviews, participants were asked to rate on a 6-point scale the extent to which they perceived themselves feeling, thinking, and behaving differ-ently across languages and to which they perceived their own personality as different, ranging from 0 (no difference),

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1521

    1 (slightly different), 3 (moderately different), and 5 (very different).

    Results and DiscussionLanguage proficiency and usage. We evaluated participants

    subjective ratings of English proficiency in addition to their objective exam grades as a manipulation check. On 6-point scales, the mean of English proficiency was 4.33 (SD = .56) and that of Chinese proficiency was 5.21 (SD = .61), indicating high English and Chinese proficiency as perceived by the participants themselves.

    Ratings by participants. Table 3 presents means and standard deviations of participant ratings on the eight personality dimen-sions in Chinese and English. Using similar analytic strategies as in Study 1 to test for language effects and perception effects on the personality factors as hypothesized, we conducted eight sets of 2 3 repeated measures ANOVA. The dependent variables were the eight personality dimensions of the SAPPS. The within-subjects factors were language (English and Chinese) and target of perception (self-perception, perception of native Chinese speakers, and perception of native English speakers). Instead of presenting an exhaustive report on all eight dimensions, we explicate the results of three factors (Extraversion, Openness to Experience, and Assertiveness) for the purpose of our hypotheses, and then summarize those of all eight dimensions (see Table 4).

    For Extraversion, the target main effect was significant, F(2, 68) = 138.16, p < .001, as well as the Language Target interaction effect, F(2, 68) = 4.47, p < .05. The univariate test associated with the language main effect was marginally sig-nificant, F(1, 69) = 3.89, p = .053. Using Holms sequential Bonferroni approach, interaction comparisons showed that the mean differences in personality ratings across English and Chinese were significantly different between perceived native speakers of English and Chinese, p < .01, and between self-perception and perceived native speakers of English, p < .025.

    Across languages, participants perceived native speakers of English as more extraverted than native speakers of Chinese, whereas self-ratings were significantly higher than those of

    Table 3. Means, Standard Deviations, and Reliability Coefficients () of Participant Ratings for the Sino-American Person Perception Scale in Study 2

    Self-perceptionPerception of native Chinese

    speakers Perception of native English speakers

    Chinese English Chinese English Chinese English

    Factor M SD M SD M SD M SD M SD M SD

    Emotional Stability 3.88 1.00 .71 3.84 0.96 .67 3.95 0.82 .61 3.84 0.94 .62 5.28 0.61 .55 4.58 0.66 .40Extraversion 4.34 1.09 .81 4.34 1.04 .77 3.60 0.81 .66 3.60 0.81 .60 5.73 0.59 .75 5.51 0.72 .74Application 4.54 1.00 .74 4.75 0.87 .65 5.11 0.90 .68 5.11 0.90 .72 3.51 0.71 .68 3.52 0.84 .70Openness 4.26 1.11 .83 4.35 1.02 .80 2.84 0.78 .75 2.84 0.78 .75 5.53 0.69 .75 5.50 0.64 .78Assertiveness 4.03 0.98 .74 4.15 1.04 .62 3.92 0.83 .68 3.92 0.83 .67 5.22 0.63 .72 5.10 0.72 .52Restraint 4.43 1.06 .76 4.52 0.87 .66 4.70 0.81 .69 4.70 0.81 .65 3.53 0.71 .64 3.71 0.91 .52Helpfulness 3.96 0.93 .63 4.78 0.74 .60 3.70 0.81 .64 3.70 0.81 .62 4.44 0.76 .61 4.71 0.64 .55Intellect 4.81 0.79 .71 4.60 0.73 .69 4.54 0.81 .67 4.54 0.81 .66 4.51 0.69 .76 4.68 0.79 .72

    Table 4. ANOVA for Participant Ratings in Study 2

    Source df F p p2

    Emotional Stability Target 2 41.23 < .001 .55 Language 1 45.16 < .001 .40 Target Language 2 23.83 < .001 .41Extraversion Target 2 138.16 < .001 .80 Language 1 3.89 .053 .05 Target Language 2 4.47 < .001 .12Application Target 2 65.57 < .001 .66 Language 1 3.81 .055 .05 Target Language 2 4.01 < .05 .11Openness Target 2 185.85 < .001 .85 Language 1 0.33 > .05 .01 Target Language 2 0.56 > .05 .02Assertiveness Target 2 55.78 < .001 .62 Language 1 0.00 > .05 .00 Target Language 2 2.03 > .05 .06Restraint Target 2 43.39 < .001 .56 Language 1 4.02 < .05 .06 Target Language 2 2.10 > .05 .06Helpfulness Target 2 23.94 < .001 .41 Language 1 69.49 < .001 .50 Target Language 2 39.06 < .001 .54Intellect Target 2 1.11 > .05 .03 Language 1 0.15 > .05 .00 Target Language 2 5.62 < .001 .54

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1522 Personality and Social Psychology Bulletin 36(11)

    Chinese speakers but much lower than those of English speak-ers, ps < .001. This effect was slightly stronger when measured in Chinese than in English.

    For Openness to Experience, only the target main effect was significant, F(2, 68) = 185.85, p < .001; neither the Language Target interaction effect nor the language main effect was significant, F(2, 68) = .56, p > .05, and F(1, 69) = .33, p > .05, respectively. Simple main effect analyses indi-cated that across languages, participants perceived native speakers of English as significantly more open than native speakers of Chinese, p < .001, with the difference between languages nonsignificant, p > .05; self-ratings were signifi-cantly higher than those perceived for Chinese speakers but lower than those perceived for English speakers, ps < .001.

    For Assertiveness, only the target main effect was signifi-cant, F(2, 68) = 55.78, p < .001; neither the Language Target interaction effect nor the language main effect was significant, F(2, 68) = 2.03, p > .05, and F(1, 69) = .00, p > .05, respec-tively. Simple main effect analyses indicated that across lan-guages, participants perceived native speakers of English as significantly more assertive than native speakers of Chinese, p < .001, with the difference between languages nonsignificant, p > .05; self-ratings were significantly lower than those for per-ceived English speakers, p < .001, but not different from those for perceived Chinese speakers, p > .05.

    To summarize participants ratings, target of perception effects were significant for seven out of eight personality factors (except for Intellect), ps < .001, whereas language effects were present only on Emotional Stability, Restraint, and Helpfulness, ps < .05. Native speakers of English were perceived to be more emotionally stable, extraverted, open, assertive, and helpful, but less restrained and lower on application than were native speakers of Chinese. Partici-pants rated themselves in between English speakers and Chinese speakers on these personality dimensions, except for Helpfulness, but closer to Chinese speakers than English speakers. In Chinese, they scored higher on Emotional Sta-bility and lower on Restraint than in English. Their self-ratings of Helpfulness were differentiated by language; that is, participants perceived themselves as not significantly different from English speakers when responding in English but closer to Chinese speakers when responding in Chinese. Native speakers of English and Chinese were not perceived to be different in Intellect, whereas participants viewed themselves to be higher on Intellect than both groups when responding in Chinese, but not different when responding in English.

    Ratings by observers. Likewise, to test language effects and ethnicity effects on the personality factors rated by observers for each condition, we conducted eight sets of 2 2 repeated measures ANOVAs. For the purpose of our hypotheses, we explicate the results for three factors (Extraversion, Openness to Experience, and Assertiveness) in the text and Figure 1, and then summarize those of all eight factors (see Table 5 for

    means and standard deviations and Table 6 for the ANOVA results).

    For Extraversion, the ethnicity main effect and the Language Ethnicity interaction effect were significant, Fs(1, 71) = 34.53 and 20.61, respectively, ps < .001, whereas the language main effect was not, F(1, 71) = 2.79, p > .05. Participants were rated as significantly more extraverted when talking with Caucasian interviewers than with Chinese interviewers, p < .001, without a significant difference between Cantonese and English, p > .05. When talking with Chinese interviewers, participants were perceived to be sig-nificantly more extraverted when speaking in English than in Cantonese, p < .001.

    For Openness to Experience, the language and ethnicity main effects as well as the Language Ethnicity interaction effect were all significant, Fs(1, 71) = 6.34, 18.57, and 19.20, respectively, ps < .05. Participants were rated as significantly more open when talking with Caucasian interviewers than with Chinese interviewers, p < .001, without a significant difference between Cantonese and English, p > .05. When talking with Chinese interviewers, participants were perceived to be significantly more open when speaking in English than in Cantonese, p < .001.

    For assertiveness, the language and ethnicity main effects as well as the Language Ethnicity interaction effect were all significant, Fs(1, 71) = 7.03, 40.23, and 13.14, respectively, ps < .05. Participants were rated as significantly more assertive when talking with Caucasian interviewers than with Chinese interviewers, p < .001, without a significant difference between Cantonese and English, p > .05. When talking with Chinese interviewers, participants were perceived to be significantly more assertive when speaking in English than in Cantonese, p < .001.

    To summarize the results from observers ratings, partici-pants were perceived to be more extraverted, open, assertive, helpful, and higher on application and intellect when talking with Caucasian interviewers than with Chinese interviewers, ps < .05. When talking with Chinese interviewers, participants were rated as more extraverted, open, assertive, helpful, and higher on application and intellect when talking in English than in Cantonese, ps < .05. Such differences were not sig-nificant on emotional stability. When conversing in English, participants were perceived to be more restrained with Chinese interviewers than with Caucasian interviewers, but showed no significant difference when conversing in Cantonese.

    Overall personality differences. Having tested the effects of an interpersonal variable (interlocutor ethnicity) and a con-textual variable (language use), we conducted multiple regres-sion analysis to predict the overall personality differences from intrapersonal variables. In Block 1, we entered version of questionnaire at Time 1 and Time 2, age, major, overseas stay (dichotomous), and length of overseas stay to control for the possible effects of these demographics. In Block 2, we entered English and Chinese language proficiency and usage,

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1523

    to examine the effects of language ability. Block 3 included cross-language differences in feeling, thinking, and behavior, and perceived influence of overseas experiences on language use and on personality, to identify significant predictors.

    The regression model accounted for 39.2% of the total vari-ance in overall personality differences, R2 = .39, F(13, 65),

    p < .01 (see Table 7). None of the demographic variables was significant, ps > .05. Nor were the effects of language profi-ciency and usage significant, ps > .05, perhaps because this sample of participants perceived themselves as high on English proficiency and hence showed less variability on this factor. Among the predictors, only cross-language difference in

    3.543.60

    4.08

    3.65 3.67

    4.12

    3.41

    3.58

    3.87

    3.08

    3.21

    3.47

    3

    4

    Extraversion Openness Assertiveness

    Mea

    n R

    atin

    gs o

    f Obs

    erve

    rs

    Conversing with Caucasian interviewers in English Conversing with Caucasian interviewers in Cantonese

    Conversing with Chinese interviewers in English Conversing with Chinese interviewers in Cantonese

    Figure 1. Observer ratings of bilinguals Extraversion, Openness to Experience, and Assertiveness when conversing with Chinese and Caucasian interviewers in Cantonese and English

    Table 5. Means and Standard Deviations (in Parentheses) of Observer Ratings Under the Four Conditions in Study 2

    Factor

    With Caucasian interviewers

    in English

    With Caucasian interviewers in Cantonese

    With Chinese interviewers

    in English

    With Chinese interviewers in Cantonese

    Emotional Stability 2.65 (0.59) 2.63 (0.63) 2.74 (0.61) 2.60 (0.69)Extraversion 3.54 (0.63) 3.65 (0.77) 3.41 (0.70) 3.08 (0.91)Application 4.06 (0.51) 4.09 (0.48) 4.05 (0.42) 3.59 (0.53)Openness 3.60 (0.60) 3.67 (0.59) 3.58 (0.64) 3.21 (0.72)Assertiveness 4.08 (0.69) 4.12 (0.54) 3.87 (0.74) 3.47 (0.82)Restraint 3.03 (0.66) 3.04 (0.90) 3.29 (0.70) 2.94 (0.83)Helpfulness 4.10 (0.46) 4.01 (0.42) 4.06 (0.42) 3.42 (0.61)Intellect 3.85 (0.55) 3.96 (0.49) 3.72 (0.57) 3.28 (0.68)

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1524 Personality and Social Psychology Bulletin 36(11)

    thinking was significant, p < .05, and positively related to overall personality differences, but overseas experience and differences in feeling and behavior did not make significant contributions, ps > .05.

    General DiscussionThe present research attempted to address the question of whether personality changes as a function of language use, an issue that has both theoretical importance (e.g., what is being evoked by language priming?) and applied significance (e.g., do bilinguals have two personalities?). Language effects have been studied more frequently on values and self-concept (e.g., Earl, 1969; Ross, Xun, & Wilson, 2002; Trafimow, Silverman, Fan, & Law, 1997) but relatively infrequently with personality inventories. Findings from those few studies that examined language effects on bilinguals personality and from cultural priming studies that used language as a prime have been equivocal. To fill in this gap, we adopted different meth-ods to test the cultural accommodation hypothesis.

    Target of Perception Effects

    On the key variables, bilinguals ratings were consistent in Studies 1 and 2 with respect to the perception of culture-related norms for personality traits. Bilinguals perceived native speak-ers of English as higher on typical Western traits, such as Extraversion, Openness to Experience, and Assertiveness, but lower on the typical Eastern trait of Restraint, than did native speakers of Chinese. These perceptions are aligned with the cultural norms for individualism and collectivism (Chen, Bond, Chan, Tang, & Buchtel, 2009). As native speakers of English are mostly from individualistic cultures and native speakers of Chinese mostly from collectivist cultures, they are perceived to embody the characteristics stereotypic of their cultural groups and are more socially appropriate in their respective cultures. These expected dispositions help bilinguals interact with cultural others when they do not have prior, personally developed information about their interac-tion partners.

    Understandably, bilinguals positioned themselves between native speakers of English and Chinese but closer to the Chinese group than the English group. The colonial history of Hong Kong has influenced its residents with Western insti-tutions, for example, political structure, educational system, exposure to English, entertainment industry, and communica-tion styles, in their everyday living (Chen et al., 2008).

    Table 6. ANOVA for Observer Ratings in Study 2

    Source df F p p2

    Emotional Stability Language 1 2.42 > .05 .03 Ethnicity 1 0.45 > .05 .01 Language Ethnicity 1 2.40 > .05 .03Extraversion Language 1 2.79 > .05 .04 Ethnicity 1 34.53 < .001 .33 Language Ethnicity 1 20.61 < .001 .23Application Language 1 19.38 < .001 .21 Ethnicity 1 36.75 < .001 .34 Language Ethnicity 1 33.82 < .001 .32Openness Language 1 6.34 < .05 .08 Ethnicity 1 18.57 < .05 .21 Language Ethnicity 1 19.20 < .05 .21Assertiveness Language 1 7.03 < .05 .09 Ethnicity 1 40.23 < .05 .36 Language Ethnicity 1 13.14 < .05 .16Restraint Language 1 4.52 < .05 .06 Ethnicity 1 1.56 > .05 .02 Language Ethnicity 1 9.58 < .05 .12Helpfulness Language 1 44.22 < .001 .38 Ethnicity 1 55.18 < .001 .44 Language Ethnicity 1 30.99 < .001 .30Intellect Language 1 9.33 < .01 .12 Ethnicity 1 88.30 < .01 .55 Language Ethnicity 1 38.39 < .01 .35

    Table 7. Hierarchical Regression Model for Overall Personality Differences Across Languages in Study 2

    Block 1 Block 2 Block 3

    Variable

    Language version (Time 1) .36 .35 .14Language version (Time 2) .29 .26 .10Age .07 .07 -.00Major -.10 -.11 -.03Overseas experience .09 .08 -.06Length of overseas stay .26 .27 .02English proficiency and usage -.12 -.04Chinese proficiency and usage -.03 -.01Cross-language differences in

    feeling.00

    Cross-language differences in thinking

    .32*

    Cross-language differences in behavior

    .22

    Perceived influence of overseas experience on language use

    -.29

    Perceived influence of overseas experience on personality

    .36

    R2 .11 .13 .39R2 .11 .02 .27F change 1.46 0.66 5.65***df 6/72 2/70 5/65

    *p < .05. ***p < .001.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1525

    University students in Hong Kong are an elite group more exposed to Western influences. Consequently, they character-ize themselves as deviating from typical Chinese speakers toward typical English speakers, but the extent of such devia-tions is only moderate and not to the extent of identifying themselves more with the English group.

    Other than the key variables, participant ratings of helpful-ness were different in Study 2. If we predicted that native speakers of Chinese were perceived to be higher on typical Eastern traits such as Agreeableness, as in Study 1, they should also be higher on a prosocial trait like Helpfulness, but this pattern was not supported in the two studies. Instead, native speakers of English were perceived as more agreeable than native speakers of Chinese, although participants perceived themselves as more agreeable than typical members of both groups. These perceptions may stem from bilinguals social interactions with native speakers of English in Hong Kong, especially in school settings where the Caucasians whom they encountered tend to be warm and supportive. In fact, these results underscore the importance of assessing bilinguals own perceptions about cultural others (perceived norms), as identi-fied in Study 1, instead of the actual characteristics of these other groups (actual norms).

    The results on helpfulness converge with earlier findings showing that Hong Kong Chinese scored lower on the Altru-ism facet of the Agreeableness dimension (McCrae, Costa, & Yik, 1996; McCrae et al., 1998). Based on their study of European and Chinese Canadians, McCrae et al. (1998) sug-gested that exposure to Western culture increased openness, cheerfulness, and prosocial behavior and attitudes. In the pres-ent study, bilinguals shifted their responses on helpfulness as a function of language. In English, they perceived themselves as closer to native speakers of English, but in Chinese as closer to the Chinese group. In this case we observe cultural accom-modation, an assimilation effect.

    Language EffectsCompared to perception effects, language exerted less influ-ence on self-ratings, and cross-language differences were rather small in magnitude but greater for observer ratings than for self-ratings in Study 2. Self-reports tap into targets intrapersonal evaluation of feelings, thoughts, and behaviors, but some personality traits only present themselves under certain circumstances (Funder, 2003). Behavioral observa-tions take into account interpersonal dispositions and situ-ational factors, probably magnifying the expression of personality. Furthermore, observers did not have to perform and respond to interactive situations, so that they had more mental resources to attend to characteristic patterns of their targets behaviors.

    Based on observer ratings, language effects were qualified by the significant interactions between spoken language and interlocutor ethnicity. When conversing with Chinese

    interviewers, participants were perceived as more extraverted, open, and assertive in English than in Chinese. These judge-derived personality profiles match bilinguals self-derived profiles, supporting our contention that cross-language dif-ferences in personality arise from bilinguals perception of prototypic personality in the corresponding culture. When conversing with Caucasian interviewers, participants were perceived as more extraverted, open, and assertive than when conversing with Chinese interviewers. The differences between the two languages were not significant with Cauca-sian interviewers, indicating that the presence of a native English speaker was strong enough to prime these Western traits and elicit accommodating patterns, regardless of the language used.

    Ethnicity EffectsObserver ratings confirmed ethnicity effects, except for intel-lect. These patterns were similar to their response propensities in English, be it self-reports or observer ratings, with both effects demonstrating cultural accommodation. Using English and interacting with native speakers of English engage con-sistent cultural norms as perceived by bilinguals, and accord-ingly they responded in similar ways. These findings reveal the underlying mechanism of bilingual personality, that is, the expectations and goals of making culturally congruent responses, elicited by the language or the interaction partner, motivate bilinguals to realize their perception of cultural norms. As pointed out by Holmes (2002), the influences of social situations and interpersonal expectations determine interaction behaviors.

    The interaction between ethnicity and language on restraint reflects a comfort effect. When communicating with native speakers of Chinese, bilinguals were perceived as more restrained in English than in Chinese but showed no significant difference between languages with native speakers of English. Perhaps speaking a second language with an interlocutor from ones own culture is an unnatural situation, prompting bilin-guals to act with restraint. Because of social comparison, participants might feel inadequate and less competent in front of the Chinese interviewers with better perceived English proficiency, thereby impeding the spontaneous expression of their interpersonal dispositions. Thus, they appeared more restrained compared to their communications in their first language.

    Overall Personality DifferencesThe regression results shed some light on what intrapersonal factors predicted cross-language differences in personality. Among other things, bilinguals believed that cognitive dif-ferences activated across languages made a significant contribution to this process. From the perspective of com-munication accommodation theory, the cognitive function

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1526 Personality and Social Psychology Bulletin 36(11)

    can explain convergent as well as divergent behaviors in the communication process (e.g., Gallois et al., 2005; Giles & Ogay, 2006; Giles, Scherer, & Taylor, 1979). For example, when the speakers goal is to facilitate comprehension, he or she may accommodate to the recipients speech characteristics. It also supports Holmess (2002) emphasis on interpersonal goals and expectations, which are cognitive elements that combine with influences of social situations to bring out behavioral outcomes. It is bilinguals cognitive organization of languages, perceptions, expectations, and goals that leads to the expression of traits during social interactions across languages.

    Conclusion and Future DirectionsTo conclude, the present findings demonstrate that cross-language differences exist on some personality dimensions as perceived by both targets and observers. After controlling for cultural background (by using compound bilinguals), social role, gender, age, and fluency, language effects on personality shift become relatively weak in magnitude. An important mechanism underlying such shifts is cultural accommodation. When bilinguals interacted with interlocutors from different cultures, they showed characteristics corresponding to their perceptions of normative personality in those cultures. Lan-guage (written or spoken) and its associated feature (native speakers of the language) present cultural cues to activate expectations and goals of making culturally congruent responses, and motivate bilinguals to realize their perceptions of cultural norms.

    In closing, we do not claim that language per se does not matter in bilingual personality, a question that could not be directly answered by our study. Language proficiency and cultural frames associated with language use may affect cross-language differences in personality directly. Further studies can adopt processing measures as a means to assess language proficiency and language dominance, for example, using reac-tion time on a verbal categorization task as a performance measure of proficiency. Future research can also examine other compound bilinguals within the same cultural set, such as Spanish-Catalan bilinguals in Barcelona, Spain, or Dutch-Frisian bilinguals in the North of the Netherlands, to test the cultural accommodation hypothesis. The important question is not only whether language really matters but also when and how it matters.

    Acknowledgments

    We would like to express our appreciation to Catherine A. McBride, Helene H. L. Fung, Cecilia Cheng, and Daphna Oyserman for their insightful comments on earlier versions of this article, and to Wai Chan, Chin-Ming Hui, and Ben C. P. Lam for their helpful advice on the statistical analyses. Thanks also go to Wesley C. H. Wu, Brian Willis, Stefan White, Tony T. H. Cheng, Nelson C. Y. Yeung, Angus M. Y. Lok, Queenie K. Y. Lai, and Anita W. Y. Cheung for their assistance in data collecting, interviewing, and coding.

    Declaration of Conflicting Interests

    The authors declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

    Funding

    The authors disclosed receipt of the following financial support for the research and/or authorship of this article: the Direct Grant (#2020905) of the Chinese University of Hong Kong and the General Research Fund (PolyU 5412/08H) by the Research Grants Council, Hong Kong.

    Note

    1. Mean interobserver agreements were .16 and .14 for convers-ing with Chinese native speakers in Cantonese and English, respectively, and .13 and .09 for conversing with English native speakers in Cantonese and English, respectively. These corre-lations were lower than the mean interobserver agreement of .22-.25 found in previous studies on trait ratings (John & Robins, 1993), perhaps because some traits measured in the present study were less observable. Unlike coding specific acts and behaviors, levels of observer agreement are usually lower for judgments of general personality traits, as internal experiences and processes are less available to observers (Gosling, John, Craik, & Robins, 1998). Using two observers is another limita-tion of our study; more judges could provide useful information on the person perception process.

    References

    Benet-Martnez, V., & Haritatos, J. (2005). Bicultural identity inte-gration (BII): Components and psychological antecedents. Jour-nal of Personality, 73, 1015-1050.

    Bond, M. H. (1983). How language variation affects inter-cultural differentiation of values by Hong Kong bilinguals. Journal of Language and Social Psychology, 2, 57-66.

    Bond, M. H. (1985). Language as a carrier of ethnic stereotypes in Hong Kong. Journal of Social Psychology, 125, 53-62.

    Bond, M. H. (1986). Mutual stereotypes and the facilitation of inter-action across cultural lines. International Journal of Intercul-tural Relations, 10, 259-276.

    Bond, M. H., & Yang, K. (1982). Ethnic affirmation versus cross-cultural accommodation: The variable impact of questionnaire language on Chinese bilinguals from Hong Kong. Journal of Cross-Cultural Psychology, 13, 169-185.

    Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner & J. W. Berry (Eds.), Field meth-ods in cross-cultural research (pp. 137-164). Beverly Hills, CA: Sage.

    Butcher, J. N. (1996). International adaptations of the MMPI-2: Research and clinical applications. Minneapolis: University of Minnesota Press.

    Butcher, J. N. (2004). Personality assessment without borders: Adaptation of the MMPI-2 across cultures. Journal of Person-ality Assessment, 83, 90-104.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • Chen and Bond 1527

    Butcher, J. N., Mosch, S. C., Tsai, J., & Nezami, E. (2006). Cross-cultural applications of the MMPI-2. Washington, DC: American Psychological Association.

    Byrne, D., & Kelly, K. (1981). An introduction to personality. Upper Saddle River, NJ: Prentice Hall.

    Chen, S. X., Benet-Martnez, V., & Bond, M. H. (2008). Bicultural identity, bilingualism, and psychological adjustment in multi-cultural societies: Immigration-based and globalization-based acculturation. Journal of Personality, 76, 803-838.

    Chen, S. X., Bond, M. H., Chan, B., Tang, D., & Buchtel, E. E. (2009). Behavioral manifestations of modesty. Journal of Cross-Cultural Psychology, 40, 603-626.

    Chen, S. X., Bond, M. H., & Fung, I. (2006). Perceiving commu-nion in the dyadic relationship of others. Asian Journal of Social Psychology, 9, 103-111.

    Earl, M. (1969). A cross-cultural and cross-language comparison of dogmatism scores. Journal of Social Psychology, 4, 1-20.

    Ervin, S. M. (1961). Semantic shift in bilingualism. American Jour-nal of Psychology, 74, 233-241.

    Ervin, S. M. (1964). Language and TAT content in bilinguals. Jour-nal of Abnormal and Social Psychology, 68, 500-507.

    Ervin, S. M., & Osgood, C. E. (1954). Second language learning and bilingualism. Journal of Abnormal and Social Psychology, 49, 139-146.

    Funder, D. C. (2003). Toward a social psychology of person judg-ments: Implications for person perception accuracy and self-knowledge. In J. P. Forgas, K. D. Williams, & W. von Hippel (Eds.), Social judgments: Implicit and explicit processes(pp. 115-133). New York, NY: Cambridge University Press.

    Gallois, C., Ogay, T., & Giles, H. (2005). Communication accommo-dation theory: A look back and a look ahead. In W. B. Gudykunst (Ed.), Theorizing about intercultural communication (pp. 121-148). Thousand Oaks, CA: Sage.

    Giles, H., & Ogay, T. (2006). Communication accommodation theory. In B. Whalen & W. Samter (Eds.), Explaining commu-nication: Contemporary theories and exemplars (pp. 293-310). Mahwah, NJ: Erlbaum.

    Giles, H., Scherer, K. R., & Taylor, D. M. (1979). Speech markers in social interaction. In K. R. Scherer & H. Giles (Eds.), Social markers in speech (pp. 343-381). Cambridge, UK: Cambridge University Press.

    Gosling, S. D., John, O. P., Craik, K. H., & Robins, R. W. (1998). Do people know how they behave? Self-reported act frequencies compared with on-line codings by observers. Journal of Person-ality and Social Psychology, 74, 1337-1349.

    Green, S. B., & Salkind, N. J. (2003). Using SPSS for Windows and Macintosh: Analyzing and understanding data (3rd ed.). Upper Saddle River, NJ: Prentice Hall.

    Heller, D., Watson, D., Komar, J., Min, J.-A., & Perunovic, W. Q. E. (2007). Contextualized personality: Traditional and new assess-ment procedures. Journal of Personality, 75, 1229-1253.

    Holmes, J. G. (2002). Interpersonal expectations as the building blocks of social cognition: An interdependence theory perspec-tive. Personal Relationships, 9, 1-26.

    Hong, Y. Y., Morris, M. W., Chiu, C. Y., & Benet-Martnez, V. (2000). Multicultural minds: A dynamic constructivist approach to culture and cognition. American Psychologist, 55, 709-720.

    Hui, C. H., & Cheng, I. W. M. (1987). Effects of second language proficiency of speakers and listeners on person perception and behavioral intention: A study of Chinese bilinguals. Interna-tional Journal of Psychology, 22, 421-430.

    Hull, P. V. (1990). Two languages, two personalities? Unpublished doctoral dissertation, University of California, Berkeley.

    Hull, P. V. (1996). Bilingualism: Some personality and cul-tural issues. In D. I. Slobin, J. Gerhardt, A. Kyratzis, & J. Guo (Eds.), Social interaction, social context, and language: Essays in honor of Susan Ervin-Tripp (pp. 419-434). Mahwah, NJ: Erlbaum.

    John, O. P. (1990). The Big Five factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of personality: Theory and research (pp. 66-100). New York, NY: Guilford.

    John, O. P., & Robins, R. W. (1993). Determinants of interjudge agreement on personality traits: The Big Five domains, observ-ability, evaluativeness, and the unique perspective of the self. Journal of Personality, 61, 521-551.

    John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: His-tory, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102-138). New York, NY: Guilford.

    Kemmelmeier, M., & Cheng, B. Y.-M. (2004). Language and self-construal priming: A replication and extension in a Hong Kong sample. Journal of Cross-Cultural Psychology, 35, 705-712.

    Lambert, W., Hodgson, R., Gardner, R., & Fillenbaum, S. (1960). Evaluational reactions to spoken languages. Journal of Abnor-mal and Social Psychology, 60, 41-51.

    Lew, W. J. F. (1985). Traits and dimensions of personality: Chinese intellectuals in Taiwan. CUHK Education Journal, 13, 37-48.

    McCrae, R. R., & Costa, P. T., Jr. (1985). Openness to experience. In R. Hogan & W. H. Jones (Eds.), Perspectives in personality (Vol. 1, pp. 145-172). Greenwich, CT: JAI.

    McCrae, R. R., & Costa, P. T., Jr. (1987). Validation of the five-factor model of personality across instruments and observers. Journal of Personality and Social Psychology, 52, 81-90.

    McCrae, R. R., Costa, P. T., Jr., & Yik, M. S. M. (1996). Uni-versal aspects of Chinese personality structure. In M. H. Bond (Ed.), The handbook of Chinese psychology (pp. 189-207). Hong Kong: Oxford University Press.

    McCrae, R. R., Yik, M. S. M., Trapnell, P. D., Bond, M. H., & Paulhus, D. L. (1998). Interpreting personality profiles across cultures: Bilingual, acculturation, and peer rating studies of Chinese undergraduates. Journal of Personality and Social Psy-chology, 74, 1041-1055.

    Norman, W. T. (1963). Toward an adequate taxonomy of person-ality attributes: Replicated factor structure in peer nomination personality ratings. Journal of Abnormal and Social Psychology, 66, 574-583.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/

  • 1528 Personality and Social Psychology Bulletin 36(11)

    Oyserman, D., & Lee, S. S. W. (2007). Priming culture: Culture as situated cognition. In S. Kitayama & D. Cohen (Eds.), Hand-book of cultural psychology (pp. 255-279). New York, NY: Guilford.

    Oyserman, D., & Lee, S. S. W. (2008). Does culture influence what and how we think? Effects of priming individualism and col-lectivism. Psychological Bulletin, 134, 311-342.

    Pervin, L. (1980). Personality: Theory, assessment, and research (3rd ed.). New York, NY: Wiley.

    Ralston, D. A., Cunniff, M. K., & Gustafson, D. J. (1995). Cul-tural accommodation: The effect of language on the responses of bilingual Hong Kong Chinese managers. Journal of Cross-Cultural Psychology, 26, 714-727.

    Ramrez-Esparza, N., Gosling, S. D., Benet-Martnez, V., Potter, J. P., & Pennebaker, J. W. (2006). Do bilinguals have two person-alities? A special case of cultural frame switching. Journal of Research in Personality, 40, 99-120.

    Ross, M., Xun, W. Q. E., & Wilson, A. E. (2002). Language and the bicultural self. Personality and Social Psychology Bulletin, 28, 1040-1050.

    Stevenson, H. W., & Lee, S.-Y. (1996). The academic achievement of Chinese students. In M. H. Bond (Ed.), The handbook of Chi-nese psychology (pp. 124-142). Hong Kong: Oxford University Press.

    Trafimow, D., Silverman, E. S., Fan, R. M.-T., & Law, J. S. F. (1997). The effects of language and priming on the relative accessibility of the private self and the collective self. Journal of Cross-Cultural Psychology, 28, 107-123.

    Watkins, D., & Gerong, A. (1999). Language of response and the spontaneous self-concept: A test of the cultural accommoda-tion hypothesis. Journal of Cross-Cultural Psychology, 30, 115-121.

    Weinrich, U. (1953). Languages in contact. New York: Linguistic Circle of New York.

    Yang, K. S., & Bond, M. H. (1980). Ethnic identification by Chinese bilinguals. Journal of Cross-Cultural Psychology, 11, 411-425.

    Yang, K. S., & Bond, M. H. (1990). Exploring implicit person-ality theories with indigenous or imported constructs: The Chinese case. Journal of Personality and Social Psychology, 58, 1087-1095.

    Yik, M. S. M., & Bond, M. H. (1993). Exploring the dimensions of Chinese person perception with indigenous and imported constructs: Creating a culturally balanced scale. International Journal of Psychology, 28, 75-95.

    Yik, M. S. M., Bond, M. H., & Paulhus, D. L. (1998). Do Chinese self-enhance or self-efface? Its a matter of domain. Personality and Social Psychology Bulletin, 24, 399-406.

    at Kings College London - ISS on January 30, 2015psp.sagepub.comDownloaded from

    http://psp.sagepub.com/