Internet-Zeitschrift für Kulturwissenschaften	17. Nr.	September 2010
Sektion 1.3.	Re-writing linguistic history – (post)colonial reality on the fringes of linguistic theories Sektionsleiter | Section Chair: Eric A. Anchimbe (University of Bayreuth, Germany) Dokumentation | Documentation | Documentation

Lexical gap, semantic incongruence,
and medium-of-instruction-induced code-switching:
Evidence from Hong Kong and Taiwan(1)

David C. S. Li (City University of Hong Kong)

Email: endavidl@cityu.edu.hk

Abstract

Using a combination of three methodologies – revelation through disruption, reflective diaries, and focus group interviews – this comparative, replication study presents empirical evidence of code-switching being a result of three closely related motivations: lexical gap, semantic incongruence, and the ‘medium-of-learning effect’. 108 university students in Hong Kong and Taiwan were asked to speak only their dominant local community language – Cantonese and Mandarin respectively – for one day, and reflect on the reasons why they wanted to use other languages in specific contexts. Data consist of 108 reflective diaries and 13 focus group interviews. This study draws essentially on diary data, focusing on instances of Chinese-English code-switching involving technical terms and academic jargon learned or introduced in English. The findings show that in situations where little or no negotiation of speaker roles and identity are at work, code-switching may be due to three related factors: (a) there is a lexical gap in the bilingual’s mental lexicon; (b) the putative translation equivalent is perceived as semantically incongruent; and (c) the technical term was taught and learned through English.

1. Introduction

There are two prevailing theoretical frameworks explaining motivations behind code-switching: the Markedness Model (e.g. Myers-Scotton, 1993; Myers-Scotton & Bolonyai, 2001) and conversation analysis (e.g. Auer, 1995; W. Li, 1994, 2002). Both frameworks have greatly enhanced our understanding of some of the typical motivations behind code-switching. For example, Myers-Scotton’s (1993) analysis of code-switching data in East Africa shows convincingly that code-switching from a local vernacular to a prestigious supranational language such as Swahili and English is very commonly found in situations marked by a clear power differential, as in interactions between employer and employee, gate-keeper and visitor. Conversation analysis has shown that sometimes language choice may index the bilingual’s dispreference when responding to a question raised in a different language (W. Li, 1994).

Notwithstanding these insightful contributions, there is one theoretical issue which to my knowledge has not been dealt with satisfactorily: by postulating that some social motivation or discourse-analytic factor is at work during code-switching, it is tacitly assumed that whatever the referential meaning(s) of an embedded language element, there exists a semantically and stylistically congruent counterpart – ‘translation equivalent’ – in the matrix language, such that referential meaning could be regarded as constant. But is this always the case? The validity of this tacit assumption has been called into question (D. Li, 1999, 2001). The literature abounds with examples from different multilingual contexts where lexical borrowing is semantically motivated. On problems of referential equivalence across languages, see for example French-English (Grosjean 1982: 336); German-English (Berns 1992: 158); Japanese-English-Chinese (Honna 1995: 46); and Cantonese-English (D. Li, 1996, 2001; Li & Tse, 2002).

A concern for referential meaning is arguably one of the most salient factors triggering code-switching. Lexical gaps in the matrix language are perhaps the most obvious types of evidence whereby intra-sentential code-switching is so difficult to avoid, for example, unplugged, skyline, present (v.), presentation, and project (n.), for which there is as yet no satisfactory translation equivalent in Chinese (Chen & Carper, 2005). But even where a translation equivalent appears to exist, there is no guarantee that the speaker is aware of it at the moment of speaking or writing. Psycholinguistically, it may be that the speaker is forgetful, tired or nervous. It is also conceivable that despite the existence of a putative equivalent, the speaker finds it unsatisfactory because it carries additional, albeit unwanted, associations. This is one of the findings in Li and Tse’s (2002) experimental “purist” study, in which 12 English majors were asked not to use English for one day, the main objective being to see to what extent English was considered necessary and desirable in context-specific social interaction with others (for more details on methodology, see below).

In one instructive example, a female participant (F3) wanted to invite a friend to play wargames with her in the countryside. The idiomatic translation of wargame in Cantonese is 打野戰 (daa35 je23 zin33, literally ‘fight wild battle’).2 In addition to that meaning, however, it is well-known that this Cantonese expression is commonly used in soft-porn literature (typically written in vernacular Cantonese) alluding to some illicit, erotic activity. It so happened that F3’s invitation was made to a male friend. To abide by the artificial Cantonese-only rule of speaking, she invited him to ‘fight wild battle’ with her, which turned out to be a great embarrassment for both. As she explained at the focus group discussion, she would have used the code-mixed expression打 wargame (daa35 wargame) if no artificial rule of speaking had been in force. Examples such as these suggest that one important motivation behind code-switching is to avoid unwanted semantic loss or gain. In sum, there is considerable evidence in the literature to make a strong case for further scrutinizing semantically motivated code-switching, with a view to examining how it articulates with the prevailing theoretical frameworks.

This paper reports findings of an experimental study showing that referential meaning across languages cannot be assumed to be constant in code-switching research. Where there is (perceived) semantic and/or stylistic discrepancy between the target words in the embedded language and their putative translation equivalents in the matrix language, code-switching may be more adequately explained by the speaker’s attempt to avoid unwanted semantic loss or gain. One special case involves academic jargon and technical terms learned or introduced in English.

2. Data, methodologies and predictions

The research design of this comparative study draws on three methodologies: the ethnomethodological technique of ‘revelation through disruption’ (Garfinkel, 1967), language diaries (Gibbons, 1987), and focus group interviews (see, e.g., Lunt & Livingstone, 1996; Steward & Shamdasani, 1990). Earlier studies in sociolinguistics and anthropology showed that speakers are not always conscious of their language use patterns, and when asked whether, and if so, when they use a particular language variety, their self-report data tend to be inaccurate and thus unreliable (see, e.g., Blom & Gumperz, 1972; Gumperz, 1972; Gumperz & Hymes, 1972). This is why in sociolinguistic research in the last three decades, little attempt has been made to include speakers’ metalinguistic comments about their own language behavior as a source of data.

It is in this regard that Garfinkel’s (1967) research technique of ‘revelation through disruption’ proved to be extremely useful and productive. By disrupting the normal patterns of language use through some artificial rule of speaking, this technique obliges the participants to reflect on what they perceived as actually happening in contextually ‘rich’ situations. This technique has proved to work very well in Li and Tse’s (2002) experimental study. In their experimental study, twelve Cantonese-speaking English majors at City University of Hong Kong were asked (a) not to use any English for one day; (b) to reflect on the reasons why they wanted to use English in context-specific situations with the help of a proforma which helped them to record key contextual information regarding ‘who speaks what to whom and when’ (compare Appendix I); and (c) to participate in a focus group interview two days after the experiment. Very instructive findings were obtained. Practically none of the 12 participants were able to prevent at least some English expressions from cropping up when interacting with friends and peers on the day of the experiment. Among other things, they confirmed that where no negotiation of identity was involved, the typical motivation behind their wish to use English while interacting with others was out of a concern for referential meaning. In particular, they were either unable to find a suitable and satisfactory referential equivalent in Cantonese (i.e. due to a lexical gap in their mental lexicon), or, where an equivalent appeared to exist, they were concerned that the meaning was somewhat different from what they wanted to express (see, e.g., daa35 je23 zin33 vs. daa35 wargame discussed above).

The data reported in this paper were collected for a project designed to replicate the Li and Tse (2002) study. The experiment took place at three universities: two in Taiwan and one in Hong Kong (see http://personal.cityu.edu.hk/~endavidl/index3.htm). 108 students participated in the experiment (65 in Taiwan, 43 in Hong Kong). For one day, they were asked to:

1. speak only their local, dominant community language (Mandarin in Taiwan, Cantonese in Hong Kong);
2. keep a record of speech events specifying ‘who speaks what to whom and when’ with the help of a proforma (soft copy sent to all participants by email before the experiment; see Appendix I);
3. write a reflective diary (up to two pages) in a language of their choice and, when completed, send it to the investigators in the form of an email attachment; and
4. take part in a focus group discussion attended by participants studying the same discipline, sharing their experiences and views on the reasons behind their preferred language choice in context-specific situations.

At each site of investigation (Dong Hwa University, Hualien; Chengchi University, Taipei; and City University of Hong Kong), a briefing was held in the evening before the day of the experiment, where detailed instructions were given and participants’ questions clarified. The rationale behind the study was vaguely disguised as ‘a comparative study of tertiary students’ language use patterns in Taiwan and Hong Kong’. All participants were rewarded with a modest hourly remuneration. An overview of their major disciplines and numbers is presented in Table 1.

Table 1: The number of student participants and their major disciplines at each of the 3 universities

In terms of language choice, there is a clear difference between participants in Taiwan and their counterparts in Hong Kong. Given that the majority of educated Hong Kong Chinese are Cantonese-English bilinguals, that artificial rule of speaking in effect obliged the participants to use only Cantonese (cf. Li & Tse, 2002). In contrast, since a majority of the student participants in Taiwan have Minnan Hua, Hakka or an Aboriginal language as their main home language in addition to Mandarin (the national language) and English which they learned in school, being obliged to use only Mandarin would mean that they should make every effort to prevent elements of their home language and English from cropping up in their conversation.

The experiments proceeded smoothly. Participants’ reflective diaries received by email were analyzed thematically with a view to extracting contextually rich descriptions and reflective commentaries, which were then adapted anonymously into a list of rich points grouped under specific headings serving as stimulus material for discussion by the focus group in question. During the focus group discussion, participants were invited to elaborate on the rich points they documented, while others were encouraged to share their views by citing similar or different experiences. With the participants’ consent, all the focus group discussions were video- and audio-recorded. Data thus consisted of two main sources: 108 participants’ language diaries and the transcriptions of 13 focus group interviews.

In terms of coding, the diary data were first processed carefully and inductively, allowing for recurrent themes or categories to emerge emically. The identified categories were then coded inductively and exhaustively with the help of MAXqda, a Windows-compatible software which is especially user-friendly for the multiple coding of selected text segments. To optimize inter-coding reliability, all the 13 focus group transcriptions were coded independently by a full-time research assistant and me, the principal investigator. Where differences or omissions occurred in the process of coding, they were resolved through clarification and, if necessary, by making slight modifications to the inventory of the coding categories (e.g., change in wording of existing categories, addition of sub-themes to given categories, or creation of new categories). In this way, a total of 63 sub-themes organized under 12 broadly defined categories have been identified.3 In this study our focus is on the participants’ own words to account for the main reasons why they code-switched in context-specific situations. Of relevance to this paper are three recurrent sub-themes under the category “Linguistic motivations of code-switching”:

1. A lack of translation equivalent: participants reported that code-switching was triggered by a lack of translation equivalent, or the improvised translation / circumlocution (in Mandarin in Taiwan, Cantonese in Hong Kong) failed to get across the intended meaning.
2. The putative translation equivalent is semantically infelicitous: participants reported that while a translation equivalent appeared to exist, it was semantically infelicitous or inappropriate in specific contexts.
3. Metalinguistic comments on the ‘first-impression hypothesis’ or ‘medium-of-learning effect’: participants attributed their code-switching to English in part to the cognitive salience of English terms as a result of encountering them first in English, or as a consequence of learning the related concepts through the medium of English.

Owing to space constraints, this study will only draw on the diary data. When analyzing the diary data, care was taken to cross-check the focus group discussion data for consistency and elucidation where appropriate.

Following are three key predictions based on the findings of Li and Tse (2002):

1. All participants will be inconvenienced at least to some extent by being artificially prevented from using any language other than their dominant community language.
2. Despite conscious monitoring, elements from languages other than their respective community languages cannot be entirely suppressed, resulting in code-switching.
3. When the participants want to code-switch but are prevented from doing so due to the artificial rule of speaking, linguistic motivations, especially a lack of semantically and/or stylistically congruent ‘translation equivalents’, account for the majority of the cases.

3. Results

It is important to note that the analysis is based on the participants’ self-report data. In other words, when a participant claims that some specific instance of code-switching was triggered by a lack of translation equivalent, it may or may not be accurate. This is because the participant may not be aware that there exists a dictionary equivalent, or that a fairly widely used translation exists. For example, one Hong Kong science major (HSM2) improvised the translation of FYP (‘final year project’) as最後一年報告 (zeoi³³ hau²² jat⁵⁵ nin²¹ bou³³ gou³³), which sounded very strange to his interlocutors. HSM2 was not the only one who had problems expressing FYP in Chinese. When chatting about school work with peers, many CityU students in Hong Kong – whatever their discipline – found it impossible to avoid using this term, which is a graduation requirement of all undergraduate degree programmes. While the Chinese translation certainly exists somewhere (e.g. in programme documents and the Chinese page of the CityU website), it clearly did not have any currency among CityU students. Another Hong Kong participant (HSM3) reported wanting to say ‘log in’ but could not; he ended up saying 注冊 (sic, zyu³³ caak³³) instead of the standard equivalent in Chinese Windows: 登入(dang⁵⁵ jap²²). However, inaccuracies in the participants’ perception is methodologically not a problem, for what matters for our purpose is their own awareness and perception which guided their code choice in context. Table 2 gives the number of participants who made such a claim in their reflective diary.

Table 2: Number of participants who claimed that code-switching was due to a lack of translation equivalent, or the improvised translation / circumlocution failed to get across the intended meaning

Note: The number within brackets indicates the total number of participants in the group

One domain in which English clearly prevails is electronic communication mediated by personal computers and the internet. Table 3 shows a subset of the English expressions which are frequently used in e-communication or computer applications, for which the participants claimed that there were no usable translation equivalents. Or, the expression in question was suppressed, but then the improvised translation or circumlocution in Chinese resulted in communication problems (those highlighted were mentioned by more than one participant):

Table 3: A subset of English terms in the domain of e-communication which participants wanted to use but could not                                     

It can be seen that the Hong Kong e-communication word list (types rather than tokens) is not only longer than the Taiwanese counterpart; it also consists of nouns (including noun groups) and verbs denoting word processing commands: download, log in, print, save, and send. This is not the case in the Taiwanese word list, which consists only of substantives. After cross-checking the focus group discussion data with Taiwanese participants (n=65), however, it became clear that word-processing commands such as click, delete, highlight and print were also heard and used occasionally, but apparently not as frequently compared with their Chinese equivalents àn (按), shānchú (刪除), fănbái (反白) and yìn (印), respectively. This is in sharp contrast with the language preference of bilingual Hong Kong Chinese learners, who typically refer to word-processing functions directly in English. To what extent such free variation is a community-wide or idiosyncratic practice in Taiwan remains unclear. There is clearly room for further research in this area. What seems certain is that, when the conversation touches upon e-communication, the pressure to code-switch to English is higher among Hong Kong participants than among their Taiwanese counterparts.

A parallel pattern is found with regard to English abbreviations used for e-learning. Since most universities in Taiwan make use of a Bulletin Board System (abbreviated as ‘BBS’ in speech) for e-communication between staff, students and the school administration on the intranet, it is understandable why the acronym ‘BBS’ is so indispensable when reference is made to various aspects of e-communication on campus. Over half of the 65 Taiwanese participants mentioned how inconvenient it was when they were prevented from using ‘BBS’ with friends and peers. Similar remarks of inconvenience were mentioned by most CityU participations when discussing CityU-related topics with their classmates, but the range of English expressions cited is considerably larger (those highlighted were mentioned by more than one participant):

• add-drop (courses on the web)
• AIMS:        ‘Academic Information Management System’
• AWW:        pronounced as ‘A dub dub’; abbreviated title of the course ‘Advanced Writing Workshop’
• BACH:       pronounced as [ba:tS]; abbreviated title of undergraduate programme ‘Bachelor of Arts in Chinese’
• BATI:         pronounced as [ba:ti]; abbreviated title of undergraduate programme ‘Bachelor of Arts in Translation and Interpretation’
• CAPP:       ‘Curriculum, Advising and Programme Planning’
• CSC:          ‘Computer Services Centre’
• CTL:           Department of Chinese, Translation and Linguistics
• e-portal: name of the intranet for CityU staff and students
• FMO:         Facilities Management Office
• FYP:          Final Year Project (graduation requirement of all undergraduate Programmes at CityU)
• GPA:          Grade Point Average

Many of the English words that participants wanted to use but could not are high-frequency vocabulary words, especially nouns and verbs, in English for Academic purposes (EAP). One Hong Kong Chinese major (HCF3), for example, reported how inconvenient it was when she could not use English at a meeting with other fellow students. She cited a fairly long list of English words. According to HCF3, all of these words had no usable, satisfactory equivalent in Chinese, which is why she felt greatly inconvenienced and found it such a pain. She went on to state one instructive example concerning the improvised translation of ‘pair up’, which sounded awkward to herself as well as her interlocutors:3

Terms of address in English constitute another area where some Hong Kong participants felt that the putative Chinese equivalent was less appropriate. Thus one major of translation and interpretation in Hong Kong (HCF8) reported being inconvenienced by not being able to address her lecturers in English using ‘title plus last name’, for example, Dr. Sin, Dr. Cheng. Such a practice appears to be less common in Taiwan.

Apart from a lack of translation equivalent in the bilingual’s mental lexicon, a perceived lack of semantic congruence between the English term and the putative translation equivalent in Chinese is also thematized in many participants’ reflective diaries, including situations where no problem was encountered. Thus HCF5 pointed out that the English terms ‘selling point’ and ‘hard sell’ were suppressed and replaced with賣點 (maai22 dim35) and硬銷 (ngaang22 siu55), respectively, without triggering any communication problem. Similarly, a business major in Taiwan reported having no difficulty replacing ‘calories’ that she usually used with its Chinese counterparts, 卡路里 (kālùlĭ) or 熱量 (rìlìang):

These examples suggest that the Chinese translations of the English terms in question, either through transliteration or translation, or both, have been integrated into the Chinese lexicon, though the extent of integration in the local speech community remains unclear.

However, there were many more participants who reported that, while a translation equivalent of a term in English or Mandarin appeared to exist, very often it was dispreferred because it was perceived as semantically infelicitous or stylistically incongruent. One high-frequency example is the translation equivalent of email: 電子郵件 (diànzi yóujiàn / din²² zi³⁵ jau²¹ gin³⁵), which reportedly has little currency among the participants in Taiwan and Hong Kong alike (cf. Li & Tse, 2002). Where the availability of a translation equivalent was thematized, the participant typically showed awareness of, or thought that he or she knew, what the equivalent was, before explaining why it was dispreferred. For example, one engineering major in Taiwan pointed out why he had wanted to use the term ‘voltage regulator’ because its Chinese translation had hardly any currency among his peers. He also commented that the Chinese equivalent diànyā tiáozhĕng qì was more wordy and ‘not as smooth’:

Table 4 gives an overview of the number of participants who made explicit mention (at least once) that the translation equivalent of a target word in English or other Chinese varieties than Mandarin led to some form of communication problem. In some cases, this awareness was arrived at after the participant failed to use the putative translation equivalent; in other cases, the putative translation was used, but communication was adversely affected in some way.

Table 4: Number of participants who claimed that a translation equivalent might exist but it was semantically infelicitous or inappropriate in specific contexts

Note: The number within brackets indicates the total number of participants in the group

One instructive example regarding a lack of semantic congruence between an English expression and its translation ‘equivalent’ in Chinese concerns a casual remark made by a Hong Kong participant (HCF8) majoring in translation and interpretation on a psychologically ‘heavy’ topic. That remark was uttered in order not to violate the Cantonese-only rule of speaking. In so doing, however, she felt an acute sense of discomfort because the improvised Cantonese ‘equivalent’ made her appear rude, resulting in unwanted semantic loss or gain, and regret:

There are several similar reported instances of unwanted semantic loss or gain in the diary data. Such examples, together with those where participants made explicit reference to a lack of a usable translation equivalent, constitute strong evidence that referential equivalence of a given term in English (or for that matter, in any language or language variety) is not always assured. This lends empirical support to the observation that code-switching is sometimes due to the bilingual’s concern for referential meaning (cf. Li, 1999, 2001; Li & Tse, 2002).

In addition to lexical gap in the bilingual’s mental lexicon and a perceived lack of a semantically congruent translation equivalent, more compelling evidence is constituted by the participants’ metalinguistic comments on the reasons for their own inability to respect the Mandarin-only or Cantonese-only rule of speaking on the day of the experiment. Where English expressions were found difficult to avoid or popped out despite the bilingual’s self-monitoring and control, a few participants postulated that the cognitive salience of English terms might be a natural result of the ‘first impression’ being in English. For example:

Interestingly, according to some other participants, the cognitive salience resulting from the first impression helped account for the reason why the putative Chinese equivalent of an English term was relatively opaque. Thus one English major in Taiwan (CEF1) explained why it never occurred to her to refer to the Chinese equivalent of the word ‘syllabus’ (of a course), because that word was used by the professor from day one of the course:

A very similar point was made by a business major in Hong Kong (HBM4) with regard to the technical terms ‘sample size’ and ‘pilot test’ when talking to a lecturer:

Likewise, a science major in Hong Kong (HSF5) reported that she had never thought about using the Chinese equivalents of ‘mentors’ and ‘mentees’, because the two concepts had always been referred to in English:

One particularly instructive example comes from a non-Cantonese-speaking exchange student from mainland China, who had been in Hong Kong only for four months at the time of the experiment. As she explained in her reflective diary written in Chinese (simplified characters), before coming to Hong Kong she had rarely found it necessary to insert English words into her Mandarin. But after studying at CityU for only four months, she found it difficult to avoid inserting English words of various lengths into her Mandarin, a surprising change in her everyday language use patterns that she became aware of after this experiment. One example of an English term that she gave is an abbreviated course title generally known to CityU staff and students as CCIV (pronounced as C-C-I-V), which stands for ‘Chinese civilization’ – a term that she cited as evidence in support of her ‘first impression hypothesis’:

In effect, what these participants were saying amounts to the same observation made by the participant F4 in Li and Tse’s (2002: 174) earlier study, namely, sin⁵⁵ jap²² wai²¹ zyu³⁵ (先入為主), or ‘the first one who entered is the master’. There is thus prima facie evidence suggesting that the first encounter of a new term may have direct impact on its subsequent cognitive retrieval. This ‘first-impression hypothesis’ may be stated as follows:

When a concept C is first encountered in language X, then C tends to be cognitively mediated through the language X (Cx), even if a direct translation of C is encountered later on in language Y (Cy).

This helps explain why Cx tends to be cognitively more salient than Cy. Now if Y is the matrix language, the insertion of Cx will result in (intra-sentential) code-switching. Of course, more empirical research is needed to ascertain the validity of the ‘first-impression hypothesis’. It is however interesting to note that this hypothesis was generated emically by a number of participants after they had undergone the artificial Cantonese-only or Mandarin-only experiment for one day.

One activity in which the ‘first-impression hypothesis’ reportedly prevails in this study is learning through the medium of English. Quite a few participants suggested that many instances of their code-switching to English were directly a result of learning content subjects in English. This is especially true of participants from Hong Kong, who had undergone English-medium teaching and learning to different extents from secondary school onwards. For example:

This point is nicely illustrated by a number of instructive examples in our data. Some English majors at Cheng Chi University (Taipei) obviously had already come across the English term ‘code-switching’ before taking part in this experiment. This is probably why this term (and in the case of CEF2, the verb ‘switch’ as well) was used in two participants’ reflective diary:

Examples such as these point to a ‘medium-of-learning effect’ (cf. ‘the learning effect’, Gibbons, 1987), which appears to be one important factor leading to Chinese-English code-switching in Hong Kong and Taiwan. The ‘medium-of-learning effect’ helps explain why technical terms in English are so difficult to avoid in the middle of a conversation in Chinese (Cantonese or Mandarin). Table 5 lists the number of participants who made explicit reference to the ‘first-impression hypothesis’ or the ‘medium-of-learning effect’ as a possible explanation of their dependence on English terms and code-switching to English.

Table 5: Number of participants who attributed their code-switching in part to the ‘first-impression hypothesis’ or the ‘medium-of-learning effect’

Note: The number within brackets indicates the total number of participants in the group

It can be seen that Taiwanese participants (n=9, or 13.8%) are clearly outnumbered by Hong Kong participants (n=14, or 33%). This is understandable to the extent that more English is used in universities in Hong Kong – both as a medium of teaching and learning as well as school administration – compared with their Taiwanese counterparts. This pattern is also reflected in the language choice of the participants’ reflective diaries: over 44% of the Hong Kong participants (n=19) opted to write their diary in English, whereas the number of Taiwanese participants who chose English for their diary-writing is negligible (3 out of 65; see Table 6).

Table 6: Number of diaries written in English

Note: The number within brackets indicates the total number of participants in the group

The above analysis shows that the use of English as the medium of teaching and learning is one important factor behind Chinese-English code-switching in Hong Kong, and to a lesser extent in Taiwan. A substantial part of English lexical items that occur in Chinese-English code-switching in Hong Kong and Taiwan are technical terms or academic jargon taught or introduced to students in English, resulting in cognitive salience in terms of the ease with which cognitive retrieval of these terms takes place. This code-switching motivation, which is widely attested in our data, is topic-specific (cf. ‘topical regulation of language choice’, Fishman, 1972: 439) and is clearly a consequence of the medium of learning, hence the ‘medium-of-learning effect’ (cf. ‘the learning effect’, Gibbons, 1987). Fishman’s (1972) insightful observation is worth quoting at length:

The implication of topical regulation of language choice is that certain topics are somehow handled ‘better’ or more appropriately in one language than in another in particular multilingual contexts. However, this greater appropriateness may reflect or may be brought about by several different but mutually reinforcing factors. Thus, some multilingual speakers may ‘acquire the habit’ of speaking about topic x in language X (a) partially because this is the language in which they are trained to deal with this topic (e.g., they received their university training in economics in French), (b) partially because they (and their interlocutors) may lack the specialized terms for a satisfying discussion of x in language Y, (c) partially because language Y itself may currently lack as exact or as many terms for handling topic x as those currently possessed by language X, and (d) partially because it is considered strange or inappropriate to discuss x in language Y. (Fishman, 1972: 439-40; emphasis in original)

In a footnote on the same page, Fishman explains point (b) further as follows:

This effect [i.e. lacking the specialized terms for a satisfying discussion of x in language Y] has been noted even in normally monolingual settings, such as those obtaining among American intellectuals, many of whom feel obliged to use French or German words in conjunction with particular professional topics. English lexical influence on the language of immigrants in the United States has also been explained on topical grounds. (Fishman, 1972: 439)

4. Discussion and Conclusion

Through the use of ‘revelation through disruption’, we are able to tap into the metalinguistic awareness of bi- and multilingual speakers regarding the reasons why they feel the need to code-switch in context-specific situations. It is a valuable source of data to the extent that the question ‘why do bilinguals code-switch?’ cannot be adequately researched without including the code-switchers’ own voices and views in the data for triangulation purposes (cf. Ten Have, 2004: 180-181). In the past, self-report data tended to be dismissed because it was believed that speakers lacked the linguistic awareness needed to describe their own patterns of language use accurately (Blom & Gumperz, 1972). The research design of this study, however, shows that qualitatively reliable self-report data can be obtained provided the subjects’ metalinguistic awareness has been raised in regard to their code choice in context-specific situations. To this end, the complementary methodologies – revelation through disruption, language diary and focus group interview – have been shown to be very productive.

Regarding the motivations of code-switching, existing explanatory frameworks tend to emphasize either social motivations or discourse-analytic factors. In both frameworks, it is tacitly assumed that referential meaning may be held constant when switching between languages. The findings in this comparative study, however, suggest that such an assumption is not always warranted. One special case is switching that involves technical concepts acquired or introduced in a particular language – in this case, English. There is plenty of evidence in our data showing that our participants are either unaware of the Chinese equivalents of field-specific technical terms in question, hence a lexical gap in their mental lexicon. Or, where such translation equivalents appear to exist, they are perceived as ‘not saying the same thing’ owing to a lack of semantic congruence. This is further compounded by a consideration of the (improvised) translation having little or no currency in the local speech community.

This study has also found prima facie evidence of a ‘first-impression hypothesis’, whereby a new concept C encoded and introduced in language X tends to be retrieved as Cx, even though the same concept is encountered later in another language Y (Cy). This helps explain the cognitive salience of Cx vis-à-vis the relative opacity of Cy. Code-switching will result when instances of Cx and other technical concepts in X occur in the middle of a conversation in language Y. According to our data, one activity in which the ‘first-impression hypothesis’ prevails is teaching and learning through the medium of English, resulting in a ‘medium-of-learning effect’. This motivation of code-switching receives considerable support in this study, where many participants provided logically sound first-person accounts of the reasons why they found certain English terms they had learned earlier cognitively more salient – and thus so difficult to avoid – compared with their Chinese counterparts (if they existed).

In sum, this study has furnished strong empirical evidence in support of Gibbons’s (1987) original observation that code-switching may be the direct result of English-medium education (‘the learning effect’). Such a motivation was confirmed and reported by many student participants themselves in their own words (first in diaries, then in focus group discussions) after they had undergone a one-day experiment requiring them to abide by an artificial Cantonese-only or Mandarin-only rule of speaking. The medium-of-learning effect is particularly revealing in the analysis of code-switching that involves academic topics in the social interaction between ESL/EFL students at the university level in Hong Kong and Taiwan.

It should be noted that the motivations discussed in this study – lexical gap in the bilingual’s mental lexicon, perceived semantic incongruence between an original English term and its putative translation equivalent in Chinese, and the medium-of-learning effect – tend to prevail in informal interaction between educated Chinese friends and peers where little or no negotiation of identity is at stake. They constitute three more or less discrete reasons why code-switching is perceived as the unmarked code choice (cf. Myers-Scotton, 1993). In short, the three predictions based on Li and Tse’s (2002) one-day experiment are all supported in this replication study, especially linguistic motivations of code-switching, suggesting that referential meaning cannot be bracketed off as constant when investigating code-switching motivations in bilingual interaction.

References

• Auer, Peter. (1995) The pragmatics of code-switching: A sequential approach. In L. Milroy and P. Muysken (eds) One speaker, two languages: Cross-disciplinary perspectives on code-switching (pp. 115-135). Cambridge: Cambridge University Press.
• Berns, Margie. (1992) Bilingualism with English as the mother tongue: English in the German legal domain. World Englishes 11, 155-161.
• Blom, Jan-Petter and Gumperz, John J. (1972) Social meaning in linguistic structure: Code-switching in Norway. In J.J. Gumperz and D. Hymes (eds), 407-434.
• Chen, Katherine and Carper, Gray. (2005). Multilingual Hong Kong: Present一個project. www.foryue.org.
• Fishman, Joshua. (1972) Domains and the relationship between micro- and macrosociolinguistics. In J.J. Gumperz and D. Hymes (eds), 435-453.
• Garfinkel, Harold. (1967) Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
• Gibbons, John. (1987) Code-mixing and code choice: a Hong Kong case study. Clevedon: Multilingual Matters.
• Grosjean, François. (1982) Life with two languages: An introduction to bilingualism. Cambridge, Mass.: Harvard University Press.
• Gumperz, John J. (1972) Introduction. In J.J. Gumperz and D. Hymes (eds), 1-25.
• Gumperz, John J. and Hymes, Dell. (eds) (1972) Directions in sociolinguistics. The ethnography of communication. New York: Holt, Rinehart and Winston, Inc.
• Honna, Nobuyuki. (1995) English in Japanese society: Language within language. Journal of Multilingual & Multicultural Development 16(1&2), 45-62.
• Li, David C.S. (1996) Issues in bilingualism and biculturalism: A Hong Kong case study. New York: Lang.
• Li, David C.S. (1999) Linguistic convergence: Impact of English on Hong Kong Cantonese. Asian Englishes 2(1), 5-36.
• Li, David C.S. (2001) L2 lexis in L1: Reluctance to translate out of concern for referential meaning. Multilingua 20(1), 1-26.
• Li, David C.S. and Tse, Elly C.Y. (2002) One day in the life of a ‘purist’. International Journal of Bilingualism 6(2), 147-202.
• Li, Wei. (1994) Three generations, two languages, one family: Language choice and language shift in a Chinese community in Britain. Clevedon and Philadelphia: Multilingual Matters.
• Li, Wei. (2002) ‘What do you want me to say?’ On the Conversation Analysis approach to bilingual interaction. Language in Society 31(2), 159-180.
• Lunt, Peter and Livingstone, Sonia. (1996) Rethinking the focus group in media and communications research. Journal of Communication 46(2), 79-98.
• Myers-Scotton, Carol. (1993) Social motivations for codeswitching: Evidence from Africa. Oxford: Clarendon Press.
• Myers-Scotton, C. and Bolonyai, A. (2001) Calculating speakers: Codeswitching in a rational choice model. Language in Society 30(1), 1-28.
• Stewart David W. and Shamdasani, Prem N. (1990) Focus groups: Theory and practice. Newbury Park: Sage.
• Ten Have, Paul. (2004) Understanding qualitative research and ethnomethodology. London: Sage.

Notes

1.3. Re-writing linguistic history – (post)colonial reality on the fringes of linguistic theories

Sektionsgruppen | Section Groups| Groupes de sections

Inhalt | Table of Contents | Contenu  17 Nr.

Webmeister: Gerald Mach     last change: 2010-09-11