Language choice as a significant aspect of talk organization: The orderliness of language alternation

A common assumption in research on talk, especially in research on talk among bilingual speakers, is that people speak languages. This assumption leads to a serious issue of the orderliness of language alternation among bilingual speakers, of the possibility of conducting a conversational activity in two languages. To account for the orderliness of this type of talk, most researchers (e.g., Gumperz 1982, Myers-Scotton 1993b) have tended to focus on the symbolic dimension of language and argued that language alternation is not random for it is socioculturally significant. While language alternation does indeed have a symbolic dimension, it is also an instance of practical action and it is on this level that issues of its feasibility arise. As a consequence, another tradition of research (e.g., Auer 1984, 1988, 1995), which investigates talk among bilingual speakers as an instance of practical action, is fast developing. The problem with this tradition is that, drawing on the concept of language as an analytical tool, it can account only for some instances of language alternation. Therefore, this article comes in the thrust of this second research perspective and suggests ways of overcoming that difficulty. More specifically, the article argues for the need to suspend the concept of language, to see talk as an orderly activity and language choice as a significant aspect of that order. it argues that the concept of medium of an interactional episode, rather than that of language, can allow the analyst to account for the orderliness of language alternation, for the possibility of conducting a conversation in two languages.