produce reality. In the absence of fixed and essential meaning, signifier and signified can only be negatively defined as relational or differential entities. Consequently, for Saussure, the study of language is rooted in form and structure, not in content and reference. It is langue (abstract system), not parole i (individual speech act) which is the proper object of a synchronic analysis. Paradoxically, while the contingent nature of the sign clearly subjects it to historical change, language as Saussure conceives it, is only open to ahistorical analysis.4 Further, though langue is defined as a social product, and always implies parole, the preoccupation with langue at the expense of parole empties langue of its sociality. Language emerges as an autonomous (self-defining, self-contained, self-regulating, self-generating) relational structure, looking inward to its own mechanisms, and allowing no appeal to a reality beyond itself. Some of the difficulties of the application of this model emerge when language is privileged as the characteristic human structure, that which establishes and perpetuates all social life, as it is by Levi-Strauss in Structural Anthropology5. By taking language as the primary constituent of humanness the structuralist can authoritatively use the linguistic model for the analysis of all culture. Yet since linguistic structures always precede and limit understanding, man has little control over that which defines his humanness. His being is encoded in, contingent on language. Despite the claims that structuralism is a method or an activity rather than an absolute truth6, despite the obvious and necessary assault on the solaces of Western idealism, of universal hunran nature, and of the notion of the individual as 'independent creator, structuralism ends up with other versions of idealism. Other universals are smuggled in through the backdoor. Language no longer represents, reflects or gestures to an underlying ethical system, divine or human, but is itself a regulated mechanism. The internal construction of the sign may be irrational, but its systematic deployment within structures of signification is not. The mechanisms of language are related to certain fundamental operations of the brain and transformed into universal structures characteristic of human nature. Thus Levi-Strauss argues for a universal grammer of myth, and Todorov assumes a universal grammar underlying all languages and sign systems which 'coincides with the structure of the universe itself7. Subject and object, myth and mind become identical. It is precisely this circular and totalizing aspect of structuralism that is challenged by the poststructuralism of the later Barthes and Derrida. The structuralist package has been difficult to reconcile with either liberal humanism or with marxism. The liberal humanist strategy has been to undercut its claim to put the study of literature on a scientific basis. It is either taken to task for being too scientific—too crude and reductive to have any truck with literature, or for not being scientific enough—the linguistic model is seen as an archetype or a special kind of metaphor and structuralism as a skill or an art8. The undermining of the scientific component is one way of reconciling structuralism with traditional humanities, of accorftodating it as one supposedly 'neutral' method amongst many others. Another way has been to moderate its Journal of Arts and Ideas 63