To -Er Is to Err:
A Case of Code-switching in Standard Mandarin

Jennifer M.Y. Wei
Soochow University, Taipei

The purpose of this paper is to propose a new method for the analysis of code-switching.⁽¹⁾ Code-switching, one of the most prominent language phenomena in the study of language in contact, is defined in this paper as the change from one language to another in an on-going discourse. By treating the suffixual -/r/ in Standard Mandarin as a socio-culturally constructed symbol rather than as an invariant and well-bounded linguistic code, this paper probes into the many forces--economic, cultural and political--acting upon and within a multi-lingual speaker in performing the act of code-switching.

As linguists of Chinese have noted, the northern speech variety of Chinese often contains word final -erization, as in fanguan-r (restaurant) and chaguan-r (tea house). Dayle Barnes (1974, 1977) defines this language phenomenon from a socio-structural perspective. Like Chao Yuan Ren (1968), Barnes treats the word final -erization as a well-defined linguistic feature most prominently displayed in Beijing speech. To scholars such as these, it is a linguistic feature associated with prestige and knowledge that was adopted as the standard for both Putonghua (the common language) in Beijing and Guoyu (the national language) in Taiwan in the 1940s.

Through my own experience as a linguist trained in the 1980s in the United States and as a Chinese citizen born and raised in Taipei in the 1960s, I have come to notice some of the blind spots created by such a normative approach in treating word final -erization in Chinese as governed solely by internal lexical rules. This approach neglects a multilingual speaker's concerns over a language choice, a choice which often comes with a great deal of cultural and political baggage. The tumultuous history of modern China and the complexities associated with the Chinese diaspora--as well as issues of power and identity within Chinese communities in places such as Hong Kong, Beijing and Taipei--have implications and consequences even for the most seemingly objective linguistic choice as "to -er or not to -er."

I propose an interactional approach rather than a structural or correlational approach to the study of this issue. I suggest that it is best to probe into the translinguistic context of creating a standard language in Chinese in order to understand this linguistic phenomenon. Following Bakhtin (1981), Hill and Hill (1986) and Heath (1983), I argue for a dialogic model in order to deconstruct this issue from a viewpoint of "epistemological relativism." The meaning of suffixual -/r/--comparable to that of the postvocalic -/r/ in English in the United States or the "received" uvular /r/ as opposed to the rolled /r/ in French--is local and contingent. I conclude that what a multilingual speaker faces in a code-switching situation is more characteristic of an ideological contest rather than of a well-defined linguistic exercise where, presumably, an absolute line can be demarcated within a static system of forms and meanings.

This perspective came home to me when I was working on a dissertation project on Chinese-language choices and adaptations in New York City. I had long been fascinated by how speakers of Chinese adopt from an array of language choices to act out their social identities, and a range of questions puzzled me. What's on their mind when they switch from one language to another?⁽²⁾ How do they react to varieties of spoken or written Chinese, especially those different from the ones they are using? What kind of symbolic capital⁽³⁾ do they have in defining or defending one set of varieties but not the other? How would I know?

My mind was constantly occupied by these interrogations even when I was engaged in conversations with an American student of Chinese at Columbia University. The conversation warmed up with our discovery of similar interests: Chinese languages, social discourse, language and power. I was drawn to him by the way he talked, assuming that he was well-educated, good mannered, and knowledgeable. I thought that he could be more than just a casual acquaintance; he could be a friend, a companion, an ally. After a short time, our conversation shifted to our plans for the summer. "Wo yao zai fanguan-r dagong" (I will work in a restaurant), he said. Hearing that -erized⁽⁴⁾ word, which is characteristic of Beijing speech, I immediately felt a gulf between us. Our potential friendship instantly ended in perplexity and silence, because there was no way, given my background, that I could achieve friendship with a man who talked in this fashion.

To say I suspended this potential friendship because I did not or could not understand his Mandarin would be false. I understood it only too well. I am keenly aware of the sensitivity of people who refrain from demonstrating -erization because they recognize this linguistic feature as an ideologically charged sign. Being born and educated in Taiwan, my impulse was to identify with those who, like myself, have been labeled as "non-standard speakers of Mandarin." This has everything to do with my anxiety about Taiwan's future and my grievances towards China's past. "To -er or not to -er" is hardly a question in my mind: to -er is to err.

My point is that the contested social meanings embedded in this linguistic sign r have taken on specific social valuations and are perceived at a very conscious level (Labov 1972). It is more than an issue of correctness or a demonstration of linguistic competence. It is an act of identity (Le Page and Tabouret-Keller 1985). In other words, my cultural self⁽⁵⁾ as a Taiwanese female and my language self as a "non-standard speaker of Mandarin" were in conflict with my linguistic self as a "competent and objective professional." It was more than a choice of a code, which still implies a set of laws which doom a speaker to obey; thus, his or her free will or language adaptations can only be surrendered under a set of prescribed patterns. I do not mean to challenge the laws; I want to point out the duel of socio-political forces that act upon and within a speaker in his or her making of a linguistic choice. Mine was an ideological calculation of "how I should behave against what I want to be." It was a consciousness, as Susan Gal proposes, of "how speakers respond symbolically to relations of domination between groups within the state, and how they understand their historic position and identity within a world structured around dependency and unequal development" (1988, 248).

My silence at that time to the Standard Mandarin of a linguistic student only reminded me of the many silences from other people, like myself, well-educated, sensitive, and sympathetic, yet labeled as non-standard speakers of Mandarin. I chose silence in order to identify with the linguistic others. This act of code-switching cannot be understood apart from the socio-political evolution of the linguistic symbol -/r/ and its representations in various Chinese communities.

Word final retroflexion, termed -erization by Barnes (1977), occurs in the northern variety of Chinese, which was adopted by the authorities in Beijing and Taiwan in the 1940s as the linguistic norm for Putonghua and Guoyu, respectively. However, -erization appears on the whole to be absent from the speech of many other Mandarin-speaking areas (such as Singapore) and completely absent from the distinct linguistic varieties of Chinese indigenous to the provinces of southeast China, such as Guangdong, Hong Kong and Taiwan where the Yue and Min varieties of Chinese are practiced. The selection of a linguistic norm as a national language, as is the case in Taiwan, or as a common language, as is the case in China, is not a straightforward linguistic exercise. The choice of a standard variety against which other varieties are measured and valued is a prescribed linguistic phenomenon. Its codification and imposition have to do with the socio-political status of a linguistic variety at a certain time and space.

Linguistically, Yue variety (Cantonese) and Min variety (Taiwanese) of Chinese preserve more distinct linguistic traces of ancient Chinese than those of the northern varieties such as Beijing Mandarin. For example, Cantonese and Taiwanese have more elaborate tonal repertoires than their Beijing counterpart. The former also preserves word final constants from Old Chinese which are absent in the Beijing variety. In the case of Cantonese, the indirect object-direct object sequence as in the sentence of neih bei min ngoh⁽⁶⁾ (you give face [to] me) further distinguishes itself from Standard Mandarin which retains the direct object-indirect object sequence as in ni gei wo mianzi (you give me face).

According to Robert Cheng, Taiwanese (a variety of Min) differs from Standard Mandarin in the following aspects: contrast between habitual and future action, contrast between present and past tense, use of preverbal auxiliary verbs, and the obligatory use of auxiliary verbs as operators (Cheng 1985, 353-68). In addition to the grammatical differences, the dental retroflexations and suffixual -erization are also absent from Taiwanese (Kubler 1981).

Yue and Min have been widely used not only by Chinese in the southeast provinces of the mainland, such as Guangzhou or Fujian, but also in the Chinese diaspora since the end of the 18th century. In North America, Australia, and Southeast Asia, one can hear these two varieties along with other varieties of Chinese practiced daily. Cantonese has been the native language for the Chinese in Hong Kong, though it has never had an official status such as that held by British English during the colonization or that will be granted to Putonghua after 1997 (when Hong Kong returns to the People's Republic of China).

A similar sociolinguistic situation occurred in Taiwan. According to Kubler, on two occasions in Taiwan's recent history, non-native languages were widely promoted by central governments originally located outside the island (1985, 156). Japanese was declared the official language during Japan's colonization of Taiwan, and Standard Mandarin replaced the official and national linguistic status of Japanese after World War II. DeFrancis provides similar observations:

The native speakers of Mandarin who took over Taiwan after 1945 comprised only 2 or 3 million people as against the 5 or 6 million inhabitants already there. Most of the latter are native speakers of what is variously called Taiwanese or Fukienese or Min, spoken in the adjacent mainland province of Fujian, from which their ancestors migrated some three centuries ago. Guoyu (the national language) was imposed on to this non-Mandarin majority as the only language of education (DeFrancis 1984, 59).

On the Chinese mainland, the standardization of Chinese, based on the northern speech of Beijing, was promoted with the circulation of the official language policy in 1956. By adopting the term Putonghua, the Chinese Communist leaders meant to play down the inherent exclusiveness of a standard language and bring the most "viable and potent" elements from the working class and peasants (Ramsey 1987, 15). However, the linguistic difference between standard Chinese, Guoyu and Putonghua appears more like linguistic make-up in appearance, marked by subtle phonetic variations such as "to -er or not to -er." It is more of a felicitous socio-political ideology termed and imposed by the authorities than a distinct linguistic exercise practiced by those who speak "Standard Mandarin."

It is notable that authorities in Mainland China, Hong Kong, and Taiwan have opposing opinions as to what is authentic Chinese and what is merely a dialect. The definition of which variety of Chinese qualifies as the language, and which is, by default, an abbreviation, a dialect, or "non-standard" is not a straightforward linguistic exercise. Standard Mandarin, alias Guoyu in Taiwan and Putonghua in China, is a prescribed language and is based on a Beijing variety. In the general linguistic paradigm, the contesting nature of language or dialect is not an interesting issue. As Milroy and Milroy point out: "One consequence of the doctrine of arbitrariness is the linguist's working assumption that no language or dialect can be shown to be better or worse than another on linguistic grounds alone" (Milroy and Milroy 1985, 19). In reality, it is very difficult to test this scientific assumption empirically since the antithesis to this assumption, that one linguistic variety is superior to the other, is like two sides of a coin. Generally, in a linguist's mind, the value as opposed to the signification of a linguistic variety is purely ideological, not scientifically interesting.

Thus, all things being equal, varieties of Chinese--like those of English, Latin, or Arabic--are entitled to be qualified as standard languages, regardless of the evolution of economics, politics or history. If treating one language as superior to the other is ideological but not scientific, then this notion of Standard English, Standard Latin, Standard Arabic, or Standard Chinese is certainly ideological too. As Milroy and Milroy write:

The spread of English is due, not to its superiority as a system of language, but to the greater economic and political success of its speakers in recent centuries. In a similar way, classical Latin became the official language of a great empire; yet, its great prestige did not ensure its ultimate survival in the face of political and economic change (Milroy and Milroy 1985, 19).

Norman compared this linguistic ambiguity of language and dialect in Chinese to that of Romance languages:

Both have their roots in a large-scale imperial expansion. In both instances the empirical languages are carried by armies and settlers to areas previously occupied by speakers of different languages; in both cases, the newly developing vernaculars existed alongside an antiquated written language and were profoundly influenced by it. In view of these parallels, it would not be surprising if we found about the same degree of diversity among the Chinese dialects as we do among the Romance languages, and in fact I believe this to be the case (Norman 1988, 187).

To call Cantonese and Mandarin major Chinese dialects is like calling French, Italian and Spanish major Latin dialects, or, worse yet, major French dialects, major Italian dialects, or major Spanish dialects (depending on which end of the spectrum one is viewing). The socio-cultural evolution of varieties of Chinese parallels that of the varieties of Latin. A change of name from major languages to major dialects lowers the "dialectal" varieties in the linguistic hierarchy, implying as well the subordination of their speakers to the speakers of "the language" in a socio-psychological ladder. This kind of labeling goes on every day in New York City's Chinatown, and it has profound consequences for the ways people view themselves and others. It colors their interactions in particular ways. Their daily linguistic transactions fluctuate as the socio-economic forces change; they are contingent on the ever-changing socio-cultural predicament. If Rome were still the political center of European power, speakers of the Italianate varieties of the Romance language might be doing just what the speakers of Chinese are doing in a multilingual community such as New York City's Chinatown. The ambiguities of a well-bounded and invariant linguistic boundary pose challenges to the structuralist paradigm.

This paper follows Barnes (1977) in focusing only on nouns -erized in word final position. This linguistic phenomenon is characteristic in northern Chinese speech and notably in Beijing, the capital of People's Republic of China. Not all word final nouns are subject to -erization in Standard Mandarin. Chao provides extended examples for lexical distributions of -erization (1976, 216-49). According to some of my informants from Beijing, both Beijing patois spoken by the uneducated workers and Beijing language (Standard Mandarin) demonstrate -erization. For example, when I -erized dashiguan-r (embassy), I was laughed at by a Beijing speaker who later told me that dashiguan, as opposed to fanguan-r (restaurant) or chaguan-r (tea house), is a grand place and people (speakers of Beijing language as opposed to Beijing patois) do not -erize this word. When I -erized Shifu-r (master), a term which has gained currency among people in China, she commented on the improper usage of such a fashion as "the uneducated trying to imitate the speech of the learned." I later asked this speaker what came to her mind when she heard this "mistake" from other people (those who -erize in the "wrong" manner, and therefore, by her definition, speak the patois of Beijing). "You are either not from Beijing or you are from a low class in Beijing." "To -er or not to -er" is neither a dilemma nor a mistake in her mind. According to her, only those who do not speak the Beijing language have to worry about this issue. Thus, word-final -erization for a Beijing speaker is a linguistic marker not only to differentiate who speaks the Beijing language but also to differentiate different classes. (You are either not from Beijing or you are a speaker of Beijing patois.)

As mentioned above, suffixual -erization was absent in Min dialect and has been a highly marked linguistic feature, even among those who were born in the sixties and learned to speak Standard Mandarin at school. Standard Mandarin was imposed in Taiwan at the end of the 1940s, after the independence of Taiwan from Japanese colonization. As a result of the colonization (1894-1945), most of the people over fifty years of age learned Japanese in school prior to their learning of Standard Mandarin, the language promoted by the central government in Taiwan which came from mainland China in 1949. The distinction between the Taiwanese, those who were in Taiwan prior to 1949, and the Chinese mainlanders, those who came to Taiwan after 1949, was a very sensitive socio-political issue in the fifties. The February 28th Event in 1947, when the Nationalist Party used force to repress the discontented Taiwanese, highlighted the ethnic tensions among the Chinese in Taiwan. As a result, Standard Mandarin, characterized by word-final -erization and a set of dental retroflexations, was regarded as highly marked, not only linguistically but also socio-culturally. As one of my informants from Taiwan commented: "When I was young, if I heard people speak with -erization, I knew that person was a mainlander. None of my friends would speak in such a fashion."

Demonstration of "correct" pronunciation of Chinese is often compounded with pride and prejudice regarding one's native place: Ni(n) shi nali ren? in Taiwanese Mandarin as opposed to Ni(n) shi na-r ren? in Standard Mandarin. Suffixual -erization in this context receives a social referent in differentiating the mainlanders who came to Taiwan in the forties and the Taiwanese who came to Taiwan before that period. Over the years, interethnic marriage and the rise of the second or third generations of Taiwanese (those who were born and raised in Taiwan after the 1949 emigration from the Chinese mainland) have made the distinction between mainlander and Taiwanese less socio-culturally sensitive. The increased wealth enjoyed by the Chinese in Taiwan since the 1980s and their search for a legitimate identity on the international stage have consolidated and revised the status of "Mandarin in Taiwan" and have added a new referent to Standard Mandarin. Suffixual -erization in this context accords this sound with a group of Beijing political authorities who are distrusted and resented for their social and political actions since the 1960s.

What does suffixual -erization mean to an American student of Chinese? Is it a put-on speech act which native speakers of Chinese love to perform in order to have an upper-hand over their non-Chinese counterparts, a deja-vu sound (the post-vocalic /r/ in English or the uvular /r/ in Parisian French?) whose social referents and meanings are as intriguing as its Chinese counterpart, or is it a peculiar linguistic marker identified with a particular place, Beijing, which has great symbolic power in modern civilization and shows a Northern Chinese bias in its political history which has been reinforced by the actions of the Chinese Communist Party?⁽⁷⁾

When I first learned Chinese in Taiwan in the 70s, I always detested this /r/ sound. It reminded me of the way my mother spoke English. She and my dad were both from a small city between Rhode Island and Massachusetts. She tended to add an /r/ to those words ended with an open vowel, like the word, idea (/aidir/), and dropped the /r/ when the word is ended with an /r/ sound as in car (/kaa/). I always felt embarrassed to hear her speak like that. One of my Chinese teachers from Beijing didn't really use this -erization style of speaking with other people. She only wanted us, the foreign students of Chinese to learn. I felt she sounded so unnatural. Plus, nobody speaks in this fashion in the street. However, I could accept those teachers who -erized their Chinese all the time, not just in front of the foreign students. Once when I was walking in Hong Kong, I heard a person behind me -erizing almost every word. I turned around and found out that it was a foreigner speaking Chinese. He looked like us and he spoke like someone from Beijing. Someone told me later that this person's speech sounded like Beijing patois. But my attitude has changed over the years as I met more speakers from Beijing and become friends with them. I remember that the first time I fell in love with the sound, which I think is acceptable to me now, was when I heard the six-year old son of my friends from Beijing talking. -Erized words just came out his mouth naturally. I said to myself, "This is the very thing that we are killing ourselves for and this little boy just produces it without any difficulty or confusion.⁽⁸⁾

In recent years, there has been rising interest in the Russian literary scholar and linguist M.M. Bakhtin's translinguistic approach to the study of language (Bakhtin 1981). The voice of literary prose is a central structural element in his theory, which is not a set of terms for the description of language but is instead a study of language in domains beyond the bounds of form and content (Hill 1987, 92). Such a notion, therefore, is different from the notion of meta-messages (Bateson 1972) which still implies a grammatical structure determining the shape of messages.

Central to a translinguistic approach is the identification and description of voice or, better yet, voices. The translinguistic approach regards speakers as in control of a range of ways of speaking and writing and, consequently, seeks an account of the possibilities open to them for the juxtaposition of complementary and conflicting voices (Hill and Hill 1986). The notions of heteroglossia and polyphony are designed to capture the multiplicity of possibilities that reside within communities and within individuals.

Polyphonic discourse is characteristic of all speech communities, including the academic one. Within the discipline of linguistics, structuralists organize their metaphors so as to talk about language as a system of rules which describe ideal and invariant objects. Functionalists organize their metaphors to predict the variation in language as determined by well-defined and unchanging variables (such as gender and class) in a society where everyone defines gender and class the same way. The modes of linguistic discourse treating languages in one way or the other may also be seen as on-going dialogues where different voices come into play in shaping a contested paradigm. Several of Bakhtin's terms and metaphors are useful in understanding the dynamics of the act of code-switching. They are dialogue, discourse, and dialogic relation.

Dialogue, apart from its normal meaning of an exchange of utterances, does not necessarily involve the speaker and the listener physically; rather, it can also refer to inner speech communication (Vygotsky 1986). For example, when I was making a choice of "to -er or not to -er," I was engaging in my mind in a dialogue with what I had internalized as "Chinese speaking" learned in years of schooling in Taiwan. In other words, I was engaged in a conversation or dialogue with my family, friends and teachers from whom I constructed a cultural identity as a middle-class Taiwanese female, though none of them were physically present. It was this dialogue that guided my language choice rather than the dialogue I had with the American student of Chinese, even though the latter was physically present. My prior social experiences organized my linguistic transactions with others according to internal voices and different social consciousnesses.

In other words, though I understand my actual interlocutor and his Standard Mandarin, I comprehend the form and content of what it represents at a given moment against previous dialogues. The prior dialogues I had during school days in Taiwan and the dialogue with a well-educated, knowledgeable student of Chinese guide my perception of the situation at hand and constitute my expectations for future interactions. Each dialogue then leads to a more complex one and is further internalized and constantly recalled and reinterpreted to enrich the dynamics of our relationship. Each dialogue thickens understandings and complicates strategies for interactions.

Discourse in a Bakhtian sense means a specific point of view constituted by a particular social consciousness (Voloshinov 1986). For example, my graduate training in various formal linguistic discourses equipped me with theoretical insights as to what a language is and what dialects are. My Taiwanese middle-class upbringing imparted a very strong sense of what it means to be a genuine Chinese. Enduring relationships with family, friends, and teachers in a community outside Mainland China further inculcated in me aesthetic and emotional ties to various Chinese usages. These intellectual and emotional investments conditioned my language behavior, and I further internalized them as a persuasive frame of reference to interpret those who behaved differently from me, either linguistically or culturally. And this set of implicit socio-cultural ideologies underlies my social identities and guides my language choices. The notion of dialogic relation is prominent in a heterglossic situation. That is, the meaning of word-final /r/ in Standard Mandarin is not as static or well-defined as an objective linguist would assume. Historical contexts give -/r/ life and substance. As Voloshinov points out: "In actual fact, . . . contexts of usage for one and the same word often contrast with one another. . . . Contexts do not stand side by side in a row, as if unaware of one another, but are in a state of constant tension, or incessant interaction and conflict" (Voloshinov 1986, 80).

The objective linguist throws the issue of historical context overboard along with the notion of parole, the non-standard, non-systematic speech. What is left in this analysis is a monologizing discourse where each repetition of the same word aims at the same direction. This, of course, is far from the truth. The polemic nature of a symbol profiles the dialogic relations to the interest institutions and/or individuals. Standard Mandarin typified by suffixual -erization has gained several referents over the years. The dramatic historical events that have taken place since 1949 have shaped and are still shaping the meaning of suffixual -erization in Standard Mandarin. It can signify authority, prestige, prejudice, bureaucracy, or hypocrisy in instances such as the designation of Standard Mandarin as the national language by the Chinese government in Taiwan, the elevation in social status of the working class and the peasants during China's communization, and the designation of Putonghua as the common language for all the people in China by the Chinese Communist Party. -Erization further signifies these attitudes as when the Nationalist Party used force to repress the discontented Chinese in Taiwan, the radicals during the Cultural Revolution denounced old values and abolished the use of native place as a reference or source for social identity, Beijing was recognized by the U.N. in 1979 as the sole political representative for all Chinese, and when the massacre took place in Tiananmen in 1989. These attitudes will take on further nuances as the 1997 deadline of capitalist Hong Kong's return to Beijing approaches and as the tensions and contentions between Beijing and Taipei take on more dramatic forms. Only by relocating -/r/ in a given socio-cultural discourse can we begin to understand its multivocality.

This paper points out how speakers of Chinese can create and manipulate meanings of a linguistic feature to abolish or establish a relationship with their interlocutors. By situating the suffixual -erization in Standard Mandarin in different times and spaces, we find several social referents to this linguistic phenomenon. For a speaker from Beijing, for whom suffixual -erization is an unmarked feature, it can be used to distinguish different classes (working class and the educated) or native place (from Beijing or not). For a speaker from Taiwan, for whom suffixual -erization is a highly marked feature, it can be used to distinguish those who came from Taiwan before 1949 (the Taiwanese) from those who came after 1949 (the mainlanders) or it can be used to distinguish we (the Chinese in Taiwan) and they (the Chinese bureaucrats in Beijing). For an American speaker of Chinese, who has to relativize sets of communicative rules against his/her social class background and intentions, what really marks the distinction in "to -er or not to -er" is not the formal difference in phonetics but the dynamic socio-political dimension which speaks through it.

By deconstructing the ambiguities in Standard Mandarin and presenting the indeterminacy in its practical implications, this paper tries to provide an alternative analysis to the study of code-switching, and of acts of identity. It aims to show that in the case of "Standard Mandarin" "to -er or not to er" is analogous to "to be or not to be"; what you -er is what you are as a marker of class or social status. Recognizing the inequality in access to and command of a standard language, this paper advocates locating the study of code-switching beyond bounded forms and meanings and situating it in a context of economic and political power relations. Standard Mandarin is going through constant deconstruction and reconstruction in conjunction with the changing socio-economical transformation taking place in China, Hong Kong and Taiwan. Adopting the study of code-switching as a symbolic creation concerned with construction of self and other within a broader political, economic and historical context will give significance not only to the theoretical paradigm but also to the speakers of Chinese.

Bakhtin, M.M. 1981. The dialogic imagination. Austin: University of Texas Press.

_____. 1986. Speech genres and other late essays. Austin: University of Texas Press.

Barnes, Dayle. 1974. Language planning in Mainland China: A sociolinguistic study of Putonghua and Pinyin. Ph.D. dissertation. Georgetown University.

_____. 1977. To er or not to er. Journal of Chinese Linguistics 5:211-36.

Bateson, Gregory. 1972. Steps to an ecology of mind. San Francisco: Chandler.

Blom, J. and Gumperz, John J. 1972. Social meaning in linguistic structure: Code-switching in Norway. In Directions in sociolinguistics, ed. D. Hymes and J. Gumperz, 407-35. New York: Holt, Rinehart and Winston.

Bourdieu, Pierre. 1977. Outline of a theory of practice. Trans. Richard Nice. Cambridge: Cambridge University Press.

Chao, Yuen Ren. 1968. Grammar of spoken Chinese. Berkeley: University of California Press.

_____. 1976. Aspects of Chinese sociolinguistics. Stanford: Stanford University Press.

Cheng, Robert. 1985. A comparison of Taiwanese, Taiwan Mandarin, and Peking Mandarin. Language 61 (2):352-77.

Culler, Jonathan. 1976. Saussure. Hassocks: Harvester Press.

DeFrancis, John. 1984. The Chinese language: Fact and fantasy. Honolulu: University of Hawaii Press.

Fairclough, Norman. 1989. Language and power. New York: Longman.

_____. 1992. Discourse and social change. Cambridge: Polity Press.

Gal, Susan. 1988. The political economy of code choice. In Codeswitching: Anthropological and sociolinguistic perspectives, ed. M. Heller, 245-64. New York: Mouton de Gruyter.

Heath, Shirley Brice. 1983. Ways with words. Cambridge: Cambridge University Press.

Heller, Monica, ed. 1988. Codeswitching: Anthropological and sociolinguistic perspectives. New York: Mouton de Gruyter.

Hill, Jane. 1987. The refiguration of the anthropology of language. Cultural Anthropology 2:89-103.

Hill, Kenneth C. and Jane H. Hill. 1986. Speaking Mexicano: Dynamics of syncretic language in Central Mexico. Tucson: University of Arizona Press.

Kubler, Cornelieus. 1981. The development of Mandarin in Taiwan: A case study of language in contact. Ph.D. dissertation, Cornell University.

_____. 1985. The influence of Southern Min on the Mandarin of Taiwan. Anthropological Linguistics 27 (2):156-77.

Labov, William. 1966. The social stratification of English in New York City. Washington, D.C.: Center For Applied Linguistics.

_____. 1970. The study of nonstandard English. The National Council of Teachers of English by special arrangement with the Center for Applied Linguistics.

_____. 1972. Language in the inner city. Philadelphia: University of Pennsylvania Press.

Le Page, R.B. and Andree Tabouret-Keller. 1985. Acts of identity: Creole-based approaches to language and ethnicity. Cambridge: Cambridge University Press.

Link, Perry, Richard Madsen, and Paul H. Pickowicz. 1989. Unofficial China: Popular thought and culture in the People's Republic. Boulder, CO: Westview Press.

Milroy, James and Milroy, Lesley. 1985. Authority in language: Investigating language prescription and standardization. London and Boston: Routledge & Kegan Paul.

Norman, Jerry. 1988. Chinese. Cambridge and New York: Cambridge University Press.

Rabinow, Paul. 1977. Reflections on fieldwork in Morocco. Berkeley: University of California Press.

Ramsey, S. Robert. 1987. The languages of China. Princeton: Princeton University Press.

Saussure, de Ferdinand. 1966. Course in general linguistics. New York: Mcgraw-Hill.

Tennessen, Carol. 1985. Authority and resistance in language: From Michel Foucault to Compere Lapin. Ann Arbor: University Microfilms International.

Urciuoli, Bonnie. 1985. Bilingualism as code and bilingualism as practice. Anthropological Linguistics 27 (4):363-86.

Voloshinov, V.N. 1986. Marxism and the philosophy of language. Cambridge: Harvard University Press.

Vygotsky, L.S. 1978. Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.

_____. 1986. Thought and language. Cambridge: MIT Press.

1. Thanks are due to the many people who have commented on the previous drafts of this paper: Charles Bird, Martha Kendall (without whose encouragement this piece would be nothing but passing paragraphs in a lost diary), Frederic Blake, John DeFrancis, Jeffrey Wasserstrom, Cornelieus Kubler, Robert Sanders, Bonnie Urciouli, Bill Graves, Elise Breene, Anthea Gupta, Kris Hongladarom, Zhang and Lo.

2. The change from one language to another in a discourse is better understood in the linguistic literature as code-switching. According to Monica Heller, this language phenomenon is so prominent that even those who code-switch can be unaware of their behavior and vigorously deny doing anything of the kind (Heller 1988, 1).

3. The term symbolic capital is from Pierre Bourdieu (1977) who uses it to refer to non-material resources which are socially valued within a community, such as honor, reputation, good breeding, or talent.

4. This paper only focuses on the word final retroflexation, termed as suffixual erization by Barnes (1977), which occurs frequently in Northern Chinese speech.

5. Following Rabinow, the notion of self described here is defined as " . . . neither the purely cerebral cogito of the Cartesians, nor the deep psychological self of the Freudians. Rather, it is the culturally mediated and historically situated self which finds itself in a continuously changing world of meaning (1977, 10).

6. The romanization for Cantonese in this paper follows Professor Parker Pok-Fei Huang's system.

7. I am grateful to Professor Fred C. Blake for sharing this insight with me (personal communication).

8. I am grateful to Melissa Richardson for discussing this issue with me.