D Vinod Pavarala the foundation for David Morley's influential ethnography of television audiences. One of the perspectives on audiences that is textualist is based on psychoanalytic theory and concerns itself with how various media texts constitute subjectivities. This perspective seeks to infer audience responses from the structure of the text itself. Based as it is on universal, primary, psychoanalytic processes, this approach does not provide for any differential 97 readings of the text or the investigation of such multiple interpretations by empirical audience study. Hall (1978) argues that "without further work, further specification, the mechanisms of the Oedipu$ complex in the discourse of Freud and Lacan are universalist, trans-historical and therefore essentialist." Jane Feuer (1986) suggests that the individual subject constituted by cinema is pre-Oedipal and thus may be understood by a Lacanian model, but television viewers located in a family/household context are post-Oedipal, socialized, and thus require a theory of subjectivity that is more socially derived. The question is whether audience members who have assumed the positions inscribed in the text will also subscribe to the ideological framework of the text. Morley (1992:65) recommends making a distinction between "inhabiting inscribed subject positions, adopting an ideological problematic and making a dominant reading of a text." Making an important intervention in Screen, Willemen (1978) is critical of what he sees as the conflation of the textual subject and the social subject in psychoanalytic theories. For him, the 'real' readers (not 'inscribed' ones) are subjects located in specific historical and social contexts, rather than mere subjects of a single text. Since the 1970s, many scholars within British cultural studies started expressing their concern over 'overwhelming textualization' of the field, taking cultural phenomena away from their social and material foundations. Calls were given to shift from a text-centric semiology of media toward a sociology of television which would study text-audience interactions within a matrix of classes, subcultural groups, institutions, nations, races and genders.1 This shift required an interrogation of the taken-for-granted power of the media over their consumers. This 'new audience research' (distinguished from the older, effects or gratifications school) assumed that the audiences are always active and that media texts aye polysemic in nature (Evans, 1990). The theoretical and empirical consequences of how these assumptions are understood make audience studies a fascinating yet contentious field. As len Ang (1989) points out, we need to recognize that there is a certain politics of empirical audience studies which is at play. In the name of conducting 'demand-side' research (Corner, 1991), audience research often becomes captive to commercial or political interests. The urge to do research "on the side" of the audience should not descend into offering formulations such as "the audiences are free to derive any meaning they want from media texts," which would only end up as populist, neo-liberal apologia for media producers and industries (Modleski, 1986). The superficial assertion of the polysemic character of television texts and the presumption of widespread interpretative resistance present an overly romantic image of the audience. John Fiske's (1987a, 1987b) celebratory work on a "semiotic democracy" or a "readers' liberation movement" in which people from diverse backgrounds have the right to construct their own meanings within an autonomous cultural economy (posited against the Numbers 32-33