This chapter continues the previous chapter's examination of alternatives to attribution theory and other orthodox approaches to social cognition. A central issue here is the question of whether explanations are predominantly inferential and logical, as many theories of social cognition suggest, or more functional, in the sense of serving to present an image of self to particular audiences. The chapter begins with arguments for the self-presentational analogy, considers disputes about the ideological variables affecting models, and proceeds to discourse analysis, which examines the discursive quality of people's communications in naturalistic contexts.

Which analogy: scientist or actor?

The critique of inferential models

Harré (1981a) claims that attribution theories and related models suppose that lay explanations are inferential or propositional, a supposition that results in a ‘rhetoric of scientism’. This rhetoric is most definitively expressed in Kelley's (1972a) covariation model, which proposes that lay attributors establish the causes of behaviour by employing strategies analogous to statistical procedures used by scientists. Harré claims that in the ‘real’ world, explanations are functional in a self-presentational sense. They function as a ritual display, as rhetoric, as show; they are designed to impress, and affect others' impressions. Harré proposes that a dramaturgical analogy is more appropriate than an analogy with the scientist or than assumptions that people normally explain actions on a propositional—inferential basis.