Computability as a test on linguistics theories
Tony Veale · 2006
Abstract. Scientific theories are more than purely formal constructs, but linguistic artefacts that often rely on the rhetorical qualities of language to give their claims additional resonance and argumentative force. This reliance of theory upon language is even greater in those theoretical domains whose main concern is language itself, leading to a sometimes convenient blurring of content and form. Though Cognitive Linguistics has consistently revealed metaphor to be a fundamental building block in the development of complex conceptualizations, Cognitive Linguistic theories often exploit metaphor as an allusive place-holder when more formal clarity is demanded. Nonetheless, in this paper we argue that one such metaphor – the MIND-AS-COMPUTER metaphor that underlies the enterprises of Artificial Intelligence and Cognitive Science – can yield precisely the kind of formal clarity that is required by the most suggestive or radical of theories in Cognitive Linguistics. An exploration of how this metaphor informs the computational realization of cognitive theories will allow us to illuminate the often wide gap that exists between the descriptive suggestiveness of a linguistic theory and its actual computational sufficiency.