Abstract

In this paper I present arguments against the epistemological ideal of ‘correspondence’, namely the deeply entrenched notion that empirical truth consists in the match between our theories and the world. The correspondence ideal of knowledge is not something we can actually pursue, for two reasons: it is difficult to discern a coherent sense in which statements correspond to language-independent facts, and we do not have the kind of independent access to the ‘external world’ that would allow us to check the alleged statement–world correspondence. The widespread intuition that correspondence is a pursuable ideal is based on an indefensible kind of externalist referential semantics. The idea that a scientific theory ‘represents’ or ‘corresponds to’ the external world is a metaphor grounded in other human epistemic activities that are actually representational. This metaphor constitutes a serious and well-entrenched obstacle in our attempt to understand scientific practices, and overcoming it will require some disciplined thinking and hard work. In real practices of representation correspondence holds between one conceptualised structure and another, not between theory and ‘world’. Any real objects we can speak about are already conceptualised entities, not mind-controlled but mind-framed.