* An important concern of critical approaches to language and language learning is to go beyond simply describing conventions of language form and use to show the ways in which such conventions are tied to social relations of power. "Critical approaches differ from non-critical approaches in not just describing discursive practices, but in showing how discourse is shaped by relations of power and ideologies, and the constructive effects discourse has upon social identities, social relations and systems of knowledge and belief" (Fairclough, 1992b, p. 12). This awareness of the ideological effects of discourse that critical analysis can bring about can then, it is argued, lead to changes in discourse practices that will result in greater social equality and justice (Fairclough, 1992a, p. 10). As Widdowson (1998) points out, however, if all discourse is ideological, then "ideological significance can never be discovered, for it is always a function of a particular ideological partiality" (p. 149). Critical approaches, therefore, present a fundamental dilemma: On the one hand, subjects (language users) are positioned by discourse or ideology (Fairclough's "constructive effects discourse has upon social identities," 1992b, p. 12); on the other hand, subjects (language users) are supposed to be able to use discourse to create a position of their choosing (an awareness leading to greater social justice). That is, on the one hand, ideology is seen as coercing subjects into certain social practices through discourse; on the other hand, subjects are seen as capable of manipulating the code for their own interests. But if the subjects (or the possible identities of the language user) are formed in discourse, where do awareness and the possibility of interests independent of discourse come from?