English Publications

Research Papers

                 CODE-SELECTION IN CHINESE/ENGLISH BILINGUAL COMMUNICATION

                                                                      INTRODUCTION

In a society where many members can speak more than one language, communication in which two or more languages are alternatively used is a common phenomenon. This language situation has been the subject of considerable research in the past few decades, and studies using different approaches have revealed various kinds of factors affecting language alternation in bilingual communication, including linguistic factors (Gumperz 1982; Pfaff 1976; Poplack 1980; Sankoffand Poplack 1981; Woolford 1983; etc.), sociological factors (Blom and Gumperz 1972; Ferguson 1959; Fishman 1972 1978; Fishman, Cooper and Ma 197 1; Scotten 1983; Scotten and Ury 1977; etc.), sociocultural factors (Blom and Gumperz 1972; Ervin-Tripp 1964; Giles 1979; Zentella 1981; etc.), pragmatic factors (Auer 1984 1988; Gumperz: 1982; Scotten 1988; Vald6s 198 1; etc.), and environmental and psychological factors (Bentahila 1983; Brown and Fraser 1979; Pyrne1969; Genesee and Bourhis 1981; Giles 1979; Lambert 1967;,Ryan and Giles 1982;.etc.). Although these studies have made significant achievements, language alternation is still a subject of much debate among researchers.

As a departure from the above work, this study takes a communication approach that treats language alternation basically as a communication phenomenon, functioning to establish relationships and transmit messages. It uses a relational system to analyze Chinese/English bilingual communication and provides its general modes of codeselection. It deals with code-selection only in the initial conversation and not with code-shifting or code-switching during the conversation.

                                                                   DATA AND RESULTS

Data were collected during the one-year fieldwork (May 1991 - May 1992) in and out of Chinese communities in New York City, in which lived a Chinese population of "about 30,000, with halfofthem Bring in Chinatowns" (Xia 1992:202). They speak a variety of Chinese dialects, I such as Mandarin, Cantonese, Min-nan, Wu, etc. Although the exact number of the people of each dialect is unknown, it is estimated that "71.5% of Chinese Americans speak Mandarin, 5% speak Cantonese, 4% speak Min-nan. Although many people speak other dialects than Mandarin at home, most of them can speak Mandarin" (World Journal May 15th 1992: 20). In Chinatown, Cantonese and Min-nan are the dominant dialects. In other Chinese communities, Mandarin is dominant over other dialects.

897 conversations were recorded by observation of and personal participation in daily natural social interaction in a variety of societal domains, including communication in the streets, stores, schools, offices, public services, companies, parties, and homes, between strangers, friends, colleagues and fellow workers, students and teachers, doctors and patients, servicemen and customers, family members, etc. These conversations ranged from the shortest consisting of only a few words to the longest made of dozens of utterances.

Of the 897 conversations, there were 394 initial selections of English and 503 initial selections of Chinese, which demonstrates a high frequency in language alternation. The data show that these codeselections were made in a close relation with the relationships of the communicators. Table I describes this relation. It shows that English was selected more frequently between non-intimates than Chinese, which was more often used between intimates, and that the more intimate the relationships were, the more frequently Chinese was selected.

                               METHODOLOGY: AN ANALYTIC SYSTEM OF RELATIONSHIP

There are certainly many ways to view the possible relationships that can bring people into communicative interaction, and the nature of these relationships can be defined within a variety of different perspecfives and hence within a different system of interaction categories. As some aspects of information processing are unique to bilingual communication, I will, for my analytic purpose, define relationship in terms of three dimensions: (1) ethnic backgrounds, (2) levels of intimacy, and (3) roles. Each of these dimensions contains a dichotomy of relational categories. The dimension of ethnic backgrounds has categories of interethnic and intraethnic relationships; the dimension of levels of intimacy has categories of intimate and non-intimate relationships; and the dimension of roles has categories of long-term (L-Term) and short-term (S-Term) relationships. These three dimensions together with their six categories form a complete analytic system ofrelationship (Diagram 1).

Table 1 Relationships and Code-Selections in Chinese/English Bilingual Communication

Diagram1. An Analytic System of Relationship in Bilingual Communication

Diagram  l illustrates a descending order of level among these three dimensions. "Intraethnic" and "interethnic" are the first level of bilingual communication, under which different kinds of relationships are categorized: (1) long-term intimate intraethnic, (2) short-term intimate intraethnic, (3) long-term non-intimate intraethnic, (4) short-term non-intimate intraethnic, (5) long-term intimate interethnic, (6) short-term intimate interethnic, (7) long-term non-intimate interethnic, and (8) short-term non-intimate interethnic. These kinds of relationships are finally defined by the roles communicators play in their social interaction. There are two kinds of roles in bilingual communication: social and cultural roles. Social roles are those such as father, son, supervisor, employee, doctor, patient, etc. Cultural or ethnic roles are those such as Chinese, Korean, Japanese, Hispanic, American, etc. Among the social roles, some relationships are longer-term and more intimate than other relationships. For example, a father-son relationship is longer-term and more intimate than adoctor-patient relationship. For the cultural/ethnic roles, an intraethnic relationship is generally more intimate than an interethnic relationship. The structural relations among the three dimensions of relationship in bilingual communication is, in this way, best represented by the interrelationship between levels of intimacy and the other two dimensions, represented by the two kinds of role-relationships respectively: the social and cultural/ethnic role-relationships.

Four levels of relationships are formed, based upon frequency of interaction and levels of mutual knowledge.

(a) short-term non-intimate (SN): the lowest level of interactive frequency and mutual knowledge;

(b) long-term non-intimate (LN): the lower level of interactive frequency and mutual knowledge;

(c) short-term intimate (Sl): the higher level of interactive frequency and mutual knowledge; and,

(d) long-term intimate (Ll): the highest level of interactive frequency and mutual knowledge.

SN refers to a one-time-or-never communication relationship between people with no prior communication. Communicators with this kind of relationship are strangers, who meet for the first time and may or may not meet again. This kind of communication usually takes place in public domains, such as asking directions, shopping, first visits to doctors and lawyers, first meetings at parties, etc. LN refers to occasional communication relationships between people who have a very short history of communication and know each other, but not very well. Communication of this kind usually takes place between people who have lived, studied, or worked together for a very short period of time, or between people who have met several times at stores, parties, doctors' and lawyers' offices, etc. SI refers to frequent communication relationships between people who have a long history of communication. Communication of this kind is usually found among people who have lived, studied, or worked together for a rather long period of time, or had frequent meetings at public services, and know each other rather well. LI refers to extensive communication relationships between people who have a very long history of communication and know each other very well. Communication of this kind takes place among family members, relatives, very good friends, colleagues of long standing, etc.

lf we group (b) and (c) into the level of serni-intimacy, we get three levels of intimacy on the following relationship continuum.

Graph 1. The Three Levels ofIntimacy ofthe Relationship Continuum.

                      Intimate                     Semi-intimate              Non-infuriate

                      (LI)                              (SI and LN)                (SN)

These three levels of intimacy represent three developmental stages of relationship in the process of communication: (1) non-intimacy is the establishing stage, (2) semi-intimacy is the developing stage, and (3) intimacy is the developed stage. Relationships develop from the short-term non-intimate relationship (SN) through the long-term nonintimate (LN) and the short-term intimate (SI) relationship to the longterm intimate relationship (LI); or they may change the other way round.

                                                                             ANALYSIS

Non-Intimate Communication

The data indicate that English is the most frequently selected in non-intimate Chinese/English bilingual communication. Two characteristics of non-intimate communication are important in determining this selection: (1) there is no prior social interaction between communicators; and (2) they have a formal relationship.

People who have no prior social interaction and engage in communication for the first time know (almost) nothing about each other, including their ethnic backgrounds and language proficiency. This implies that their communication can be either interethnic or intraethnic, and that the language used in the communication can be in-group or outgroup. In such an ambiguous situation, in order to make communicabon possible, people will select a language which can cover both possibilities, a language that can be used in both intraethnic and interethnic communications. In the United States, this language is English, a majority language used in most social and cultural domains and between people with the same and different ethnic groups.

People who have an initial social interaction also have, in most cases, a formal relationship. In an environment where two languages are used, people will select a language that can better establish this formal relationship. In a multilingual society, different languages play different social functions. Ferguson (1959) and later Fishman (1967; 1972) describe a language situation called diglossia, a situation in which two languages (or two varieties of a language) have very precise and distinct social functions. One language is learned largely through formal education and used in formal situations like church sermons, political speeches, university lectures, and news broadcasts, and the other language is largely learned through informal channels and used in informal situations, like instructions to servants, and conversations with family members, friends and colleagues (see Ferguson 1972: 236). The former language is called H language and the latter, L language. Situations in which H language is used usually involve communicators with formal relationships, while situations in which L language is used usually involve communicators with informal relationships. In the United States, English as H language is used in most formal communications, and other minority languages (L languages) are mostly used in informal communications. Studies of language use in the United States have shown that English is usually used in bilingual communication involving non-intimate or less intimate formal relationships, while other languages are more often used in bilingual communication involving intimate or more intimateiformal relationships (see, among others, Ervin-Tripp 1964; Fishman 1966; Flores and Hopper 1975; Haugen 1953; Lance 1975; McClure 1977). The Chinese/English communication data have the same results (see Table 1).

As Table 1 shows, Chinese is also used in non-intimate Chinese/ English bilingual communication. Several reasons were found for this selection. First, a low language proficiency in English will limit the possibility to select it in communication. Of the 42 cases in which Chinese was selected, there were 18 in which communicators had difficulty in using English in communication. They spoke very little, broken English.

Second, kinds of communities in which communication takes place will affect code-selection. Table 2 shows that Chinese is used more frequently in the Chinese communities than in the non-Chinese communities. This is because of the function of the Chinese communities in defining the ethnic and linguistic contexts of communication. In the Chinese communities, most communicators are Chinese and Chinese is the majority language, even though English is the formal language. There is greater communicability by speaking Chinese than by speaking English. By using Chinese one will meet less difficulty in communication in a Chinese community than in a non-Chinese community.

Table2. Code-Selection in Non-Intimate Chinese Americans' Bilingual Commufffication

Third, informal situations can redefine the formal relationships between non-intimate communicators and reduce to some degree the formality in initial social interaction, giving communicators more freedom in selecting codes. Table 2 indicates that Chinese is used more often in informal situations than in formal situations. This is because in informal situations, communicators, though non-intimate, are less controlled by situations and are thus less restricted by the mode of code-selection required by non-intimate communication. In brief, non-intimate communicators are freer in selecting a code in their own communities and in informal situations than in other groups' communities and in formal situations.

Fourth, the involvement of an intimate in the communication between non-intimates can modify the non-intimate relationship between the latter, leading to their selection of a more intimate code, Chinese. This phenomenon is found common in introductory communication, in which the code selected by the introducer often determines the code selected by the introduced. The reasons for this are that (1) the code selected by the introducer reveals information about the language backgrounds of the introduced: it indicates that the introduced can understand or may speak that code; (2) as the introduced are the intimates ofthe introducer, they may know something about each other from the latter; (3) as the communication between the introduced is initiated by their intimate, the introducer, the mode ofcommunication is not as formal as that between two complete non-intimates; and (4) as the introduced are the introducer's intimates, they are psychologically attached with a little, if not great, intimacy. In this way, the more intimate code, Chinese, will be selected by the non-intimate introduced.

Semi-Intimate Communication

The two kinds of semi-intimate communicators - long-term nonintimate (LN) and short-term intimate (Sl) - represent two phases of relational development with the latter being more intimate than the former, and as the Chinese/English communication data show (Table 3), their bilingual communication indicates a general tendency towards an increased use of the native language in communication.

Communicators in semi-intimate communication are in a stage of developing relationships from less intimacy to more intimacy. To develop their relationships, they need to use a language that marks more intimacy than non-intimacy. As the native language of an ethnic group marks an ingroup relationship, while the non-native language marks an outgroup relationship, in ingroup communication the native language will mark a more intimate relationship than the non-native language. Therefore, the native language of an ethnic group plays a more important role in developing relationships in intraethnic communication. By using the native language in intraethnic communication, members of an ethnic group can develop a less intimate relationship into a more intimate one more effectively than by using a non-native language, that marks a less intimate, intereffinic relationship.

Table3. Code-Selection in Chinese/English Bilingual Communication between Serni-Intimate Chinese Americans Whose First Language was Chinese.

The following factors were found important in developing relationships and increasing the use of Chinese. First, people who were socially equal would start to use Chinese earlier than people who were socially unequal. This is because the more equal the social relationship is, the shorter the social distance will be, and the more social interaction results. Second, people would start earlier to use more Chinese as a result of a faster relational development in informal social interaction- at a party, afamily gathering, atour, etc.- than in formal social interaction, such as at a public meeting, lecture, and work. This is because informal communication leads to more exchange ofpersonal information, speeding up mutual understanding and knowledge. Finally, people would develop a more intimate relationship faster with people who had the same geographic backgrounds and would start to use Chinese earlier with them than with the people who came from different places. Mainland Chinese have a greater ease in developing an intimate relationship among themselves than with Taiwanese. Both Mainland Chinese and Taiwanese have greater difficulty in developing intimate relationships with Hong Kong Chinese and American-bom Chinese. This is probably because Mainlanders share more social and cultural traditions with Taiwanese than with Hong Kong and American-born Chinese. Among Mainlanders, people from the same locale would feel closer to each other than to people from different locales. They would prefer to use local dialects in their communication because of its more intimate nature.

English, however, was still used rather frequently in semi-intimate communication, especially in LN communication. This is because semi-intimate communication is still in a stage of developing relationships. Therefore, communication, although not primarily, still marks the other "half' of the semi-intimate relationship, i.e., the semi-non-intimate relationship. To mark the semi-non-intimate relationship, communicators will select a code that marks a more non-intimate relationship than an intimate relationship. As discussed above, the non-native or the outgroup language usually marks the former relationship and the native or the ingroup language usually marks the latter relationship. Therefore, while usingthe native language to develop their relationships, communicators still use the non-native language to mark their not well-developed semiintimate relationships, although not morethan that used in non-intimate communication.

Table 3 shows that English is more used in LN communication than in SI communication. This is because the relationship involved in the former is less developed and less intimate than that in the latter. In LN communication, the relationship is rather formal and sometimes socioculturally unequal, while in SI communication, the relationship has been developed into a rather informal or socioculturally equal one. With this development, English is used less and less than Chinese, as the relationship develops more and more intimate.

Intimate Communication

In intimate communication, as relationships of communicators have been well developed, communication functions primarily to transmit messages. Code-selection will be mainly based upon the language proficiency of communicators to transmit messages more efficiently.

Of the 454 conversations between intimate communicators, eleven patterns are isolated, that indicate that language proficiency is important in code-selection (Table 4).

Table 4 Patterns of Code-Selection in Intimate Chinese/English Bilingual Communication

Notes.

* E=C: almost balanced English-Chinese bilinguals (the order ofthe languages indicates the order of bilingual proficiency: the first is the first language and the second, the second language). Bilinguals of this kind have English as the first language and Chinese as the second language but can speak and understand Chinese almost as well as English.

C=E: almost balanced Chinese-English bilinguals. Bilinguals of this kind can speak and understand English almost as well as Chinese.

I>C: imbalanced English-Chinese bilinguals. Bilinguals of this kind have some difficulty in speaking or understanding Chinese.

C>E: imbalanced Chinese-English bilinguals. Bilinguals of this kind have some difficulty in speaking or understanding English.

E>>C: extremely imbalancedEnglish-Chinese bilinguals. Bilinguals of this kind have great difficulty in both speaking and understanding Chinese.

C>>E: extremely imbalanced Chinese-English bilinguals. Bilinguals of this kind have great difficulty in both speaking and understanding English.

+ E>C bilinguals speak English to C>E bilinguals, who speak Chinese to the former.

++ E=C bilinguals speak English to C=E bilinguals, who speak Chinese to the former.

Table 4 shows two general situations in intimate Chinese/English bilingual communication: (1) if the first language of the communicators is the same, that language is used (a-f); and (2) if the first languages of the communicators are different, the mode of code-selection is rather complicated. There are two possible ways: (a) each of the communicators uses his/her own first language (i and j); and (b) one of the communicators accommodates to the language of the other communicator(s) (g and h; k involves both of the situations). These two possible ways are often determined by the levels of language proficiency ofcommunicators.

From Pattern (a) through Pattern (d), the speaker and the listener are identical in their language proficiency. In Patterns (a) and (c), English is the first language for both the speaker and the listener; and in Patterns (b) and (c), Chinese is the first language for both the speaker and the listener. In Patterns (e) and (f), although the speaker is different from the listener in his/her second language proficiency, they share the same first language; so they select their first language in communication. In Patterns (g) and (h), the speaker and the listener differ in both the first and the second language; they further differ in their second language proficiency: one is an imbalanced bilingual and the other is an almost balanced bilingual. Communication between bilinguals ofthis category will require that the almost balanced bilingual accommodates to the language ofthe imbalanced bilingual. Therefore, in (g), the C=E bilingual accommodate to the E>C bilingual: English is used; and in (h), the situation is just the other way round. In Patterns (i) and (j), the speaker and the listener are different from each other in the first and the second language, butthey share one thing, that is, either they are both imbalanced bilinguals, or they are both almost balanced bilinguals. Therefore, in terms oflanguage proficiency, they are equal. In (i), if the E>C bilingual selects Chinese, the C>E bilingual's first language, it will be easier for the latte rto decode messages, but it will be difficult for the former to encode messages. If the E>C bilingual selects English, it will be easier for him/her to encode messages but it will be difficult for the C>E bilingual to decode messages. It is the same with the C>E bilingual. So no matter which language is selected, there will always be some difficulty for one or the other of them. In this case, there will be almost no reason for them to accommodate each other. Each selects his/her own first language, which is at least faster and more accurate for the speaker to encode messages. Encoding goes before decoding and the speaker supposes that the listener can decode the messages. If the former finds that the latter cannot decode the messages, he/she will, with some difficulty, switch to the other language, the listener's first language; that is why there is more frequent code-switching in (i) communication. In Pattern (j), the reason for communicators to select their first language is almost the same as that in Pattern (i): no matter which language is selected, there will always be some difficulty for one of them, owing to the difference between the first and the second languages. Moreover, since both the speaker and the listener are almost balanced-bilinguals, they are more proficient in using the second language in communication: the listener has less difficulty in decoding messages in his/her second language, so communication is more successful than communication of Pattern (i) this is why there is less code-switching in (j) than in (i). In Pattern (k), as one of the communicators has much difficulty in his/her second language, s/he and other communicators have to select his/her first language in communication.

To illustrate these patterns of code-selection, an examination of communication in the family domain is necessary. Two situations of code-selection were found in family communication: (1) all family members use either Chinese or English in their communication; and (2) some family members use Chinese and some use English in their communication. For the first case, either Chinese or English is the first language of all the family members. The situation in which Chinese is the first language of all the family members and is used as their medium of communication is found in most newly immigrant Chinese families or families residing and working in Chinatowns, where daily communication is conducted with Chinese as the main medium. The situation in which English is the first language of all the familymembers and is used as their medium ofcommunication is found mostly in second- or third-generation Chinese American families, whose members were all born in America, or in families in which childrenwere either born or raised in America and in which first-generation Chinese-American parents have received higher American education or work in places where English is the main or the only communication medium. All the members of this kind of family actually have acquired and use English as their first language.

The second case in which some family members use Chinese and some use English in their communication is found common in families where Chinese is the first language for some family members, and English, for the others. The most common situationis as follows: parents are first-generation Chinese Americans with Chinese as their first language, and their American-born or American-raised children speak English as their first language. Communication is conducted between parents in Chinese, among children in English, and between parents and children in their first languages respectively: parents speak Chinese to the children, who respond in English.

This kind of language-crossing phenomenon is found most commonly in families in which the parents' language proficiency in both Chinese and English is almost balanced or their English is proficient enough for them to understand it, though not good enough for them to speak it. The reasons why parents with language proficiency almost balanced in both Chinese and English do not speak English to their English-speaking children are the following: (1) most of their daily communication is conducted in Chinese; (2) the communicative habit of using Chinese formed through past years of communication with their children makes them feel very uncomfortable in changing to the use of English; and (3) it is easier for them to encode in Chinese though they decode well in English. Such language-crossing communication is conducted often between very intimate communicators. It could arouse embarrassment or annoyance when it takes place between less intimate communicators. Some people complained that some Chinese children spoke English to them even though they insisted on speaking Chinese to the children. However, these same people found that itwas natural when their own children spoke English to them while they replied in Chinese.

A kind of code-shifting phenomenon is found in Chinese family communication, that results in a shifting in the above two patterns of family communication. In intimate communication, the relationships of communicators are well-developed and unmarked in communication, and the principal function of communication is to transmit linguistic information. The first language is the fastest and most accurate language to transmit linguistic information. A speaker's first language can change during his/her life. In a bilingual community where a mother tongue (L language) is different from the language used in school and in most of other societal domains (H language), the mother tongue is a child's first language. However, when the child grows up and starts school, the school language gradually becomes his/her first language, as it is used most often and in most societal domains. This gradual change of one's first language often results in a change in communicative patterns of family communication. My observation shows that this change occurs more frequently when children have studied in school for three or four years. After three or fouryears' schooling, a child's ability in English has become sufficient for daily communication and they start to use it more than Chinese. This results in a shift in the patterns of family communication. The former pattern with Chinese as the main communicative medium for all the family members gives way to a new pattern: children speak English (their first language) to their parents who in turn speak Chinese (their first language) to the children. This kind of language shiffing is possible when communication takes place between intimates, such as family members, whose relationships have been well-developed, and communication functions mainly to exchange linguistic information. In this kind of linguistic-information-oriented communication, communicators would use their first languages to exchange messages in a more efficient way.

So far the parents we have looked at in the second case are either almost balanced bilinguals or are good enough at understanding English. In cases where their English language proficiency is not sufficient for them to understand normal English communication, children have to give up their own first language and accommodate their parents' first language, Chinese. This phenomenon is found in almost all studied families where parents are first-generation immigrants who speak little English and where children are either born or raised in the United States. If the children are proficient in their parents' first language, communication can be conducted without much effort. If the children are not proficient in the first language of their parents, communication could be conducted in a difficult manner for both the children and the parents. The former have to use a less proficient second language to communicate, resulting in much delay and inaccuracy in both encoding and decoding messages. The latter have difficulty too in decoding the messages they are receiving and in getting across the messages they are sending.

Cases are also found in which parents accommodate their children's first language. This happens when the children are English-Chinese bilinguals with low proficiency in Chinese. These children have acquired English before they have had a good command of Chinese. In many cases when communicating with their children, Chinese-English bilingual parents find it very difficult to use Chinese to get across to their children many things that the latter have never learned in Chinese. In this case, the parents have to give up their first language, Chinese, and accommodate their children's first language, English.

In sum, the patterns of code-selection in intimate communication are generalized as follows: (1) When the first language of communicators is identical, that language is readily used; (2) When the first languages of communicators are not identical, each communicator may (a) use his/her own first language or (b) accommodate the other's first language, depending upon the language proficiency ofthe communicators. If the communicators are almost balanced in both languages or their language proficiency is sufficient for them to understand the second language, type (a) happens. If a speaker is not an almost balanced bilingual and his/her second language proficiency is not good enough for hinvber to conduct normal communication, the other communicator(s) will accommodate his/her first language.

                                                       SUMMARY AND DISCUSSION

The Chinese/English bilingual communication has demonstrated that code-selection in bilingual communication is essentially based upon relationships between communicators and their language proficiency. The relationship between communicators is dynamic: it develops from its initial defining stage through the intermediate developing stage to its final developed stage. This dynamic nature of relationship ascribes a dynamic nature of code-selection (Graph 2). In communication between non-intimates, communicators know almost nothing about each other, including ethnic backgrounds, language proficiency, etc., communication is usually conducted in formal situations with an interethnic relationship, and topics are most often non-intimate. Communication at this stage is marked by its two fundamental functions: to transmit messages and to establish relationships. At this stage, the majority or H language is used, as it can best perform these two functions and mark an interethnic relationship. As communication continues, communicators get to know each other more and more, and their relationships develop from the nonintimate stage to the semi-intimate stage. At this point, communication is still marked by its two fundamental functions. At its early stage, communication begins to involve some personal information, including ethnic backgrounds, language proficiency, personal matters, etc., and the relationship gradually develops from an interethnic one into an intraethnic one. To develop amore intimate intraethnic, socially equal and informal relationship, communicators use a native language more often in communication. As communicators enter the final stage of relational development that is, the stage of the intimate relationship, relationships between cominunicators have been well developed and communication functions primarily to transmit messages. To transmit messages in a more efficient way, communicators, depending upon their language proficiency, will select a code that is better for encoding and decoding messages.

Graph 2. Relationships and Code-Selection in Chinese Americans' Bilingual Communication.

Language proficiency will make some adjustments to the above general modes ofcode-selection. Lower language proficiency will limit the selection of the codes specified in the above modes. Kinds of communities, native communities ornon-native communities, situations, formal or informal, and involvements ofdifferent relationships will also modify the above general modes of code-selection.

NOTES

1. It is disputed whether all these local speeches are dialects or languages. I use "dialect" as it is traditionally used by Chinese from political, historical, and linguistic points of view.

REFERENCES

Auer, J. C. P. 1984. Bilingual conversation. Amsterdam: John Benjamins.

--- 1988. A conversational analytic approach to code-switching and transfer. In M. Heller, ed., Codeswitching: anthropological and sociolinguistic perspectives Berlin: Mouton de Gruyter. 187-214.

Bentahila, A. 1983. Motivations for code-switching among Arabic-French bilinguals in Morocco. Language and Communication, 3: 233-243.

Blom, J-P. and Gumperz, J. J. 1972. Social meaning in linguistic structures: codeswitching in Norway. In J. J. Gumperz and D. Hymes, eds., Directions in sociolinguistics. New York: Holt, Rinehart and Winston. 407-34

Brown, P., and Fraser, C. 1979. Speech as a marker of situation. In K R. Scherer and H. Giles, eds., Social markers in speech. Cambridge: Cambridge UP. 33-108.

Byrne, D. 1969. Attitudes and Attraction. Advances in Experimental Social Psycholo~y,4,35-89.

Ervin-Tripp, S. M. 1964. An analysis of the interaction of language, topic, and listener. American Anthropologist 66, 86-102.

Ferguson, C. A. 1959. Diglossia. Word, 15,325-40. AlsoinP.P. Gigholied. 1972. Language and social context. Harmondsworth: Penguin. 232-5 1.

Ferguson, C. A. and Heath, S. B., eds. 198 1. Language in the USA. Cambridge: CambridgeUP.

Fisher, B. A. andEllis, D. G. 1990. Smallgroup decision making: Communication and the group process. 3rd ed. New York: McGraw Hill.

Fishman, J. A. 1966. Language loyalty in the United States. The Hague: Mouton.

---- 1967 Bilingualism with and without diglossia: diglossia with and without bilingualism. Journal ofSocial Issues 23, 29-38.

---- 1972. Societal bilingualism: stable and transitional. The Sociology of Language 4,91-106.

Fishman, J., Cooper, R. L. and Ma, R. 1971. Bilingualism in the Barrio. Bloomington: Indiana UP.

Flores, N. and Hooper, R. 1975. Mexican Americans' evaluation of spoken Spanish and English. Speech Monographs 42, 91-8.

Genesee, R., and Bourhis, R. 198 1. Language variation in social interaction: the importance of situational and interactional context. Unpublished manuscript, Department of Psychology, McGill University.

Giles, H. 1979. Ethnicity markers in speech. In K.R. Scherer and H. Giles, eds., Social markers in speech. Cambridge: Cambridge UP. 251-290.

Greenfield, L. 1970. Situational measures of normative language views. Anthropos 65,602-18.

Gumperz, J. J. 1982. Discourse strategies. Cambridge: Cambridge UP.

Haugen, E. 1953. The Norwegian language in American: a study in bilingual behavior. Philadelphia: U of Pennsylvania P.

Heller, M., ed. 1988. Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Mouton de Gruyter.

Lambert, W. E. 1967. The social psychology of bilingualism. Journal of Social Issues, 23,91-109.

Lance, D. 1975. Spanish/English code-switching. In C. E. Hernandez-Chaves and A. Beltramo, eds., El lenguaje de los Chicanos. Arlington, VA: Center for Applied Linguistics. 138-53.

McClure, E. 1977. Aspects of code-switching in the discourse of bilingual Mexican-American children. In M. Saville-Troike, ed, Linguistics and anthropology. GURT. Washington, DC: Georgetown UP.

Pfaff, C. 1976. Functional and structural constraints on syntactic variation in code-switching. In S. B. Steever, C. A. Walker and S. S. Mufwene, eds., Papers from the parasession on diachronic syntax. Chicago: Chicago Linguistic Society. 248-59.

Poplack,S. 1980. Sometimes I'll start a sentence inEnglishytermino enespanol: Toward a typology of code-switching. Linguistics, 18, 361-79.

Ryan, E. and Giles, H., eds. 1982. Attitudes towards language variation: social and applied contexts. London: Edward Arnold.

Sankoff, D., and Poplack, S. 1981. A formal grammar for code-switching. Papers in Linguistics 14, 3-46.

Scotten, C. M. 1983. Negotiation of identities in conversation: A theory of markedness and code choice. International Journal of Sociology of Language 44,115-35.

1988. Code switching as indexical of social negotiations. In Heller, ed,   Codeswitching: Anthropological and sociolinguistic perspectives. Berlin: Mouton de Gruyter. 151-86.

Scotten, C. M., and Ury, W. 1977. Bilingual strategies: the social flinctions of code-switching. Linguistics 193, 5-20.

Valdds-Fallis,G. 1981. Codeswitchingasdeliberateverbal strategy: amicroanalysis of direct and indirect requests among bilingual Chicano speakers. In R. P. Duran, ed, Latino language andCommunicative behavior. Norwood, NJ: ABIEX

Watzlawick, P., Beavin, J., and Jackson, D. 1967. Pragmatics of human communication. New York: WW Norton and Company.

Weinreich,U. 1968. Languages in contact: Findings and problems. The Hague: Mouton.

Woolford, E. 1983. Bilingual code-switching and syntactic theory. Linguistic Inquiry, 14 520-36.

Xia, N. 1992. Maintenance of the Chinese language in the United States. Bilingual Review, 17 (3):195-209.

Zentella, A. C. 198 1. Hablamos los dos. we speak both: Growing up bilingual in El Barrio. PhD thesis: University of Pennsylvania.

- End -

About the author: (Published in Language Quarterly Vol. 31:304: 177-195)

Contact the author

Email your comments to chinatown-online.com

| Home | - | About Us| - | Contact Us| - | Advertise with Us | - | Help |  - | Pay |