| description |
The thesis aims to provide a heuristic critique of the meta-theoretical foundations of Chomsky's project for an explanatory linguistics. The critique is 'heuristic' in that it attempts to take the considerations adduced to indicate how those conceptual foundations are to be re-designed on lines parallel to constructivism in the philosophy of mathematics. The net result is the provision of an outline of a meta-theoretic rationale for a process orientated linguistic theory (e.g. Kempson et al.'s LDSNL framework). The thesis investigates, and is organized around, three central strands of the Chomskyan paradigm: 1) The mathematization of linguistics: the use of formal/mathematical systems as theory constitutive metaphors. 2) A scientific realist (as opposed to instrumentalist) construal of linguistic theories. 3) A conceptualist/psychologist ontology for linguistic objects with a concomitant explanation for the nature of the linguistic in terms of properties of the modularized human "mind/brain" articulated through a system of mental representations. The central conclusions drawn are: 1) There is a failure to achieve adequate warrant for a scientific realist construal of Chomskyan linguistic theories. 2) The object(s) of study that is (are) posited in the Chomskyan paradigm require a Platonist or autonomist ontological status. A corollary of this is the inability to achieve an adequate explanation for the nature of linguistic phenomena. These conclusions, together with the observation of certain conceptual tensions and antimonies in generativist thinking (e.g. the relation between types and tokens), are taken to be sufficient to prompt a re-examination of the (metaphysical realist) assumptions that underlie that thinking. The solution that is canvassed, and which promises to resolve these tensions, is by way of a linguistic version of mathematical constructivism in which the emphasis lies in linguistic phenomena being construed as primarily cognitive events in which the constructive procedures are crucially constitutive of then linguistically individuating properties.
|
