ABSTRACT
Debates over the nature of gender identity and its social construction, originating in feminist work of the 1990s, have in recent years informed research in sociolinguistics generally and feminist linguistics more speci®cally. In particular, conceptions of gender as categorical, ®xed and static have increasingly been abandoned in favour of more constructivist and dynamic ones. Cameron (1990: 86), for example, makes the point (paraphrasing Harold Gar®nkel) that `social actors are not sociolinguistic ``dopes'' ', mindlessly and passively producing linguistic forms that are de®nitively determined by social class membership, ethnicity or gender. Rather, Cameron argues for an understanding of gender that reverses the relationship between linguistic practices and social identities traditionally posited within the quantitative sociolinguistics or variationist paradigm. Work in this tradition has typically focused on establishing correlations between linguistic variables and social factors such as age, race, ethnicity and sex, implicitly assuming that these aspects of social identity exist prior to and are determinate of linguistic behaviour (and other social behaviour). Indeed, early research in language and gender (in the 1970s and 1980s) was largely conducted within this research paradigm, focusing speci®cally on the correlation of linguistic variables with the independent variable of sex. By contrast, more recent formulations of the relationship between language and gender, following Butler (1990), emphasize the performative aspect of gender: linguistic practices, among other kinds of practices, continually bring into being individuals' social identities. Under this account, language is one important means by which gender ± an ongoing social process ± is enacted or constituted; gender is something individuals do ± in part through linguistic choices ± as opposed to something individuals are or have (West and Zimmerman 1987). Cameron's comments are illustrative:
Whereas sociolinguistics would say that the way I use language re¯ects or marks my identity as a particular kind of social subject ± I talk like a white middle-class woman because I am (already) a white middle-class woman ± the critical account suggests language is one of the things that constitutes my identity as a particular kind of subject. Sociolinguistics says that how you act depends on who you are; critical theory says that who you are (and are taken to be) depends on how you act. (emphasis in original)
(Cameron 1995: 15±16)
The idea that an individual's linguistic behaviour does not simply arise from a set of permanent and invariant social attributes is also suggestive of the contextually-variable nature of social identities. If identities are not ®xed and static, then their `performance' can vary across social, situational, and interactional contexts. It is in this regard that Schiffrin (1996) is critical of
variationist studies within sociolinguistics, in particular, the practice of coding aspects of social identity as categorical and invariant across contexts. Schiffrin argues for a different view, one in which social identities are locally situated and constructed: `we may act more or less middle-class, more or less female, and so on, depending on what we are doing and with whom. This view forces us to attend to speech activities, and to the interactions in which they are situated' (Schiffrin 1996: 199). Likewise, Goodwin in an ethnographic study of urban African-American children in Philadelphia suggests that stereotypes about women's speech collapse when talk in a whole range of activities is examined:
In order to construct social personae appropriate to the events of the moment, the same individuals articulate talk and gender differently as they move from one activity to another. The relevant unit for the analysis of cultural phenomena, including gender, is thus not the group as a whole, or the individual, but rather situated activities. (emphasis mine)
(Goodwin 1990: 9)
Goodwin's comments not only argue for a dynamic and variable conception of gender identity, they also point to the variable linguistic resources drawn upon in performances of gender from one activity to another.
