Introduction: Class and ideology in code-switching research

Session on Code-switching, class, and ideology
II University of Vigo International Symposium on Bilingualism, Vigo, 22-26 October 2002

Celso Alvarez Cáccamo
Departamento de Galego-Português, Francês e Linguística
Faculdade de Filologia
Universidade da Corunha
15071 A Corunha, Galiza


     This session deals with the relationships between code-switching, class, and ideology. It is an ambitious session topic, and I apologize for having chosen it. The reason why it is ambitious is three-fold, as the topic itself: One, we still don’t understand what exactly is subsumed under the malleable label of "code-switching" (despite its multiple technifications). Two, there are various interpretations of what "social class" means. And, three, we still don’t know what ideology is, or where it is. Therefore, this session runs the risk of beating around the bush. The fault will not lie in the individual contributions, but in the adventure of putting them together. Fortunately, I trust that our discussant, Monica Heller, will help us make some sense of our difficult triangle: code-switching, class, and ideology.

     I would just like to lay out some reflections on the topic. One way to handle the session topic is to look at how heteroglossic code-switching practices index or are produced by other constructs, namely "class", and "ideology". I suggest that we read the session papers in this way. Another approach (my own) is in a way metadiscursive: I would like to highlight some aspects of the very discursive representations of code-switching in the research done so far.

     The first thing that has intrigued me for some time is how, in the 50 or so years of history of the term "code-switching", the literature has come to encircle such a number of varied communicative phenomena into a single label, and how this label has been hypertechnified to the point of producing offshoots such as "code-mixing", "code-shifting", and so on. It is also interesting that very little of this technical discourse on code-switching, constraints, matrixes, contextualizations,etc., circulates outside of the academic field. This might be said of many scientific notions, but by no means all. We all know (or believe we know) what an "atom" is, or what a "black hole" is. At least we talk about them. As for the circulation of sociolinguistic terms themselves, "bilingualism" is almost a daily word, and its technical sense approaches common semantics quite closely: the use of two languages. Another central term, "diglossia", has reached immense popularity by way of the educational system, political activism, and other channels, at least in bilingual communities where issues of sociolinguistic conflict and national identity are prominent, such as here in Galiza and in other countries of the Spanish State.

     However, "code-switching" as such barely circulates outside the academic circles. The everyday correlates that "code-switching" has in common speech are expressions such as "people mix languages", or "we constantly change languages", or expressions of that sort. No doubt, labelings of both the technical and popular kinds may reflect different language ideologies. But technical discourse has had very little impact, I believe, on people’s speech practices, that is, it has generated very little reflexivity, except perhaps in those groups of young students who approach the study of code-switching for the first time, and who, in one way or another, become initially fascinated with the topic. This fascination is good, but it is not good enough to really understand code-switching. The truth is that the types of questions being asked about code-switching by younger researchers are quite similar to those posed ten or twenty years ago. The usual questions are, for example: "I am studying a bilingual immigrant community. I’d like to know what triggers speakers to switch languages". Or: "I live in a nation without a state. The state language dominates the traditional language. I’d like to know how each language in my community works to signal the speaker’s ethnic or national identity". Or, "I’d like to find out the grammatical rules that permit the combination of elements from my two languages, A and B. When do people code-switch and when do they transfer?". Of course, all these questions are being asked and answered in a monolingual language, in monolingual discourse.

     Studying the particularities of code-switching between various pairs of languages or dialects in different communities and contexts is no doubt interesting -- it is in this way that generalizations can be obtained, although (I am afraid) very few of them exist nowadays nevertheless. But "interesting" is not the same as "essential". And I believe we need to focus with increasing urgency on, precisely, the essential aspects of code-switching as communicative behavior. We must re-situate what is called code-switching within the universal human ability to communicate, that is, to make use of whatever signaling resources humans differentially have at their disposal in order to convey intentions and meanings. We must articulate the study of code-switching, then, with the study of the material and the ideational, the symbolic.

     The triangle we are dealing with today (language, class, ideology) contains the main stuff that makes up social reality: the material and the ideational. On the one hand, "class" represents the material basis of societal organization. We may take "class" as structurally anchored in the system of material and symbolic production, as economically based, as occupationally based, as culturally based, etc.; but in this world it is hard to conceive of any respectable perspective on "class" that does not deal with acute economic inequality, not only difference. Second, we have "ideology", which refers to contents. We know, of course, that "ideology" is also a construction, as the content of the mind is irretrievable for study. So, we may understand "ideology" (at least I do) as a meta-term: as something like a coherent set of statable (but not necessarily stated) propositions which predicate about some dimension of social reality. An ideology is coherent by definition: coherence may result from logical contradictions between propositions, true, but it is the analyst’s role to construct such coherence, not the role of statements themselves to be magically organized into coherence. Ideologies may take the shape of ‘beliefs’, ‘ideas’ or ‘opinions’, and they may manifest themselves in discourse or in other configurations of behavioral signs.

     Finally, we’ve got language and discourse at the top of this triangle (at the top because we are linguists, not sociologists), linking both social classes and social ideologies. Language, discourse, and therefore code-switching each participate in both the material and the ideational dimensions. We also know that discourse is viewed by some as a set of sequential statements, moves or utterances, and by others as a constellation of meanings. Discourse is socially (economically) generated and circulated, while it carries ideological contents.

     Research on code-switching has also selectively focused on one of these two dimensions (material and ideational) that constitute communication. I’m not trying to be binarist here: the ideational (the mind) is an abstract property of physical processes and a result of social ones. But the history of the "code-switching" notion itself reflects the flowing interests that its constructors have had on either dimension of social reality.

     It should be known by now that it all started in information theory and acoustics. From the notion of "speech code" as a very graspable, material object anchored in the acoustic sciences in the late 1940's, functional structuralism (primarily under Jakobson’s work) has abstracted the notion of "code" to represent sets of unobservable grammatical rules, including the semantic. In a 1950 article entitled "The information theory point of view in speech communication," R. M. Fano referred to "switching codes" as the mechanism by which speakers alternated mentally between different phonetic representations of phonemes in order to recognize "accents". So, for example, what native speakers of English in the audience are doing now in order to understand my words would be an instance of their mental "switching of codes" to match my non-native phonetics with their native phonologies. In the 1960's, the discourse on codes and switching veered toward "code-switching" (a term coined by Hans Vogt in 1954) as a form of bilingual behavior, and, in the 1970's, toward a discursive practice that conveys pragmatic meanings and signals social identities. Therefore, we can see how the field of code-switching itself has fluctuated between the material and the symbolic. It is in this indeterminacy where it remains today, as the study of a sort of particular phenomenon of a particular group of humans who would possess (so they say) two grammars within one competence (or was it one grammar within two competences?).

     Further, while practically any student of code-switching would defend Gumperz’ and Hernández-Chavez’ pivotal assertion that the alternation of languages may be equivalent to that of dialects or accents in monolingual speech, and that therefore "code-switching" is broader as speech variability, the fact is that the vast majority of work on code-switching has attended to alternation between two supposedly distinct languages. In an equation that parallels academic research on monolingual practices, much code-switching research has started from the assumption that, if two languages were detectable, what should be looked for is the relationships between these two languages with two forms of social identity, primarily, two ethnicities or two national groups. Some current research (but by no means all) has fortunately overcome the reductions of assigning one-to-one social indexical meanings to languages, so that "hybridity", "bivalence", "neutrality" and related notions underscore the indexical complexity of language practices. But what remains even in these contributions is a view of the indexicality of code-switching as a subsidiary by-product of seemingly preexisting unitary languages and monolithic identities: only two (or more) things can "hybridize", and a given socio-indexical speech practice can be "neutral" only in regards to at least two previous identities. When code-switching is claimed to be related to "new" identities and ethnicities, this "new", again, presupposes the precedence of the "old".

     So, what we are dealing with is, evidently, a matter of discursive representation within our own field. In summary, the term "code-switching" was initially (as with many scientific notions) a metaphor to understand the workings of speech pattern alternation in its material sense: speech frequencies. Coincident in historical (and political) time with a growing focus in western applied communication research on the notions of code, encoding, encrypting, decrypting, ciphering, deciphering, and so on (we are talking about the Cold War period), "code-switching" became strongly solidified around the Jakobsonian core tenet of language as a code, an equation that raises more questions than it offers answers. The term "code-switching" took on a life of its own, and the psychological "switching" part became backgrounded, as in today’s frequent lexicalization of "codeswitching" without the hyphen, and with only one main lexical stress over "code", "códeswitching", which somehow reifies even further the notion and obscures its origins. Further, nowadays the English term is widely borrowed into research in other languages, without translation.

     One of the outcomes of this reification is that the active, "switching" part (the one that involves agency) has been neglected, and hence comes the somewhat conceptual abhorration of "unmarked" or "meaningless code-switching", to refer to what Gumperz’s first called simply "code-switching style". For, if it is the mind that does the switching between "codes", and if this switching is "meaningless", which codes is then the mind switching between? Two meaningless languages-as-codes? Or two languages that are only one code?

     The easiest answer to this puzzle is, well, that if nothing substantial in terms of interaction is being switched in "meaningless code-switching" when languages alternate, then there probably is no switching at all. That is, we have two languages in sequential contact. But, what types of codes might these languages be if they do not contrast in discourse? Semantic, interpretive codes? Or "syntactic", surface codes, in Eco’s sense? The effect of two juxtaposed languages is then comparable to that of two juxtaposed sounds, not to that of two opposing terms.

     On the other hand, when alternation between a pair of languages does carry interactional meaning (that is, when there is mental "switching"), the question is, does alternation carry such meaning by virtue of the two languages being used? How is this possible, if in other cases the same two languages do not carry opposing interactional meanings? What is, then, actually being "switched": the surface form of languages, or two previous, relatively independent devices or principles that organize interaction, and that we may call codes?

     These issues are as yet unsolved. However, because of the strong primacy of a monoglossic ideology centered around "language" and Language (with a capital "L") in the academic field, the fact is that most research continues to view language alternation always as "code-switching" even when there is no switching of meaningful interactional codes involved. This is, in a way, a betrayal of the founding principles of code-switching as contextualization.

     One reason for this inertia and lack of reflexivity on the discourse about code-switching can be looked for in the articulation of the academic field with other social domains. Bilingual studies articulate, for example, with applied fields where issues of ethnic and national identity are manifest. In the educational field, practices of language choice and alternation amongst students in the classroom affect issues such as the differential access to the curriculum, educational success and failure, biased social categorization, or resistance to the educational institutions. In the field of language planning, social hybrid practices such as daily language alternation defy totalizing notions about national and ethnic languages being essential expressions of natural identities, as elite ideologies of linguistic unity and purity are problematized. In the growing field of computational linguistics and its accompanying "language industries", language alternation in writing or speech poses enormous challenges for automated language recognition, not to mention speech synthesis! In sum, in what concerns dominant monoglot ideologies, frequent language alternation as a daily social practice is not a very friendly reality for some important practical goals of western class society, which, in terms of language, consist of providing fully standardized school curricula as tools for social reproduction, fully normalized national languages as emblems and instruments of social classification, and fully standardized language technologies as differentially distributed commodities.

     However, bilingual studies are oftentimes subject as well to dominant monoglot ideologies. In code-switching research, too, the objects are parcelled out, territorialized before they are effectively found. By starting from predefined notions about languages and by focusing on discreteness rather than on fluidity, the Language-Nation and Language-Ethnicity binomials (misrepresented forms of the powerful equation Language-State) are reproduced. Importantly, this equation foregrounds totalizing identities such as nation and ethnicity, while it concurrently backgrounds other identities, such as gender (indeed) and social class. And it does so in various ways. First, much of the discourse on code-switching and bilingualism in general classifies individuals as "bilinguals" or "monolinguals" while it in general de-classifies them as belonging to materially dispossessed or privileged groups (bilingualism is by definition inter-classist, just as is the Nation, although we should know with Heller that bilingualism is not only a matter of degree of competence, but of social commodification). In a way, a slippery social distinction between "code-switchers" and "non code-switchers" -- between those speakers who "have" code-switching, and those who don’t -- is drawn, which puts into question the universality of communicative principles. Secondly, (as it could not be otherwise), research on code-switching backgrounds the very social conditions that allow for the representation of these speakers as "bilinguals" or as "code-switchers", and for the representation of an array of heteroglossic discursive practices under the common label of "code-switching". These conditions refer to differential class and epistemological positions. And finally, and importantly, this research and classification generates the very discursive conditions by which the class position of the researcher is actualized and articulated with and within the institutions of knowledge and power: in other words, the field reproduces itself.

     Perhaps these arguments are not extraordinarily new to all of us who reflect on how Science works. But it is this topic that occupies me today. Now, if we recall some of the foundational work on code-switching, namely John Gumperz’s, the material dimension was ever present, though not always explicitly articulated. The applied vocation of Gumperz’s work was patent in his analyses of the implications for social advancement of diverse communicative practices by minority groups in England or the US. The point was not only that people alternate languages or registers, but that by so doing they manage their identity capital, which depends on their structural position vis a vis other groups. This materialist view is present also in work by Gal, Heller and other Gumperz’s disciples. But it is absent from a large part of the literature that has selectively capitalized (at times without due credit) on Gumperz’s and others’ contributions.

     In this sense, I believe research on code-switching could benefit (as some research has already done) from reflexively addressing the two issues that concern us today: the material bases for the differential distribution and circulation of discursive resources, including languages, varieties, and contextualization devices, and the differential distribution and circulation of ideologies about those languages and switching practices.

     When speakers are categorically territorialized as "ethnic" or "national" groups over class groups and gender groups, the constitutive gap between researcher and subjects is masked under totalizing identity alliances. When, on the contrary, class identities and positions are emphasized, then this unbridgeable gap between researcher and subjects emerges clearly, and in this sense the class exercise becomes apparent and effective, as it should be in Science. It is by virtue of class position that the researcher is able to "find", "discover" or "construct" which "languages" the speakers are using or switching, as it is only the researcher, by virtue of his or her class position, that has the key to a technical inspection of the material bases of discourse practices. I suppose that something should be said here to exculpate the ethnomethodological programme in its attempt not to impose "expert" interpretations on the data, including what "language" the speaker is using. But the fact is that ethnomethodologists are experts, too, experts in ethnomethodology, which is also a privileged resource that the speaking subjects do not possess.

     In brief, when material conditions surrounding the production of so-called bilingual speech are highlighted over other conditions and subjects’ identities, the researcher’s own class position is effectively and transparently articulated through his or her practices of representation. The semiotic contrast between the political transcription of the speaker’s "bilingual", heteroglossic practices, on the one hand, and the researcher’s monoglossic, monolingual discourse on those practices, on the other, becomes foregrounded. And this, uncomfortably, leaves us very little room for avoiding reflexivity, that is, for avoiding asking ourselves an essential question: How much do we know about language and communication which legitimates us for representing in specific ways our subjects’ languages and communicative practices?

     And I believe we still know very little about what language is, what communication is, where codes and ideologies reside. In our ignorance (at least, my own), sometimes we push forward to try to understand the topographic details of language boundaries, to detect the often undetectable in language alternations, while we leave behind the necessary quest for, simultaneously, the universal principles of communication and symbolization, the historically contingent bases of inequality through talk, and the relationships between the two.

Última alteração: 25-03-2005