Aspects of a psychologically informed theory of agreement

Abstract

AbstractIn this paper I try to establish bidirectional links between the grammar and the processing (especially production) of agreement in order to provide the broad strokes of a psychologically viable theory of agreement. I start by arguing that full encapsulation and full interactivity in agreement operations are not realistic options. The question therefore becomes how much of each should be posited on principled grounds. It is further argued that in language production agreement ‘leaks’, in the sense that conceptual structure is ready to interfere in the establishment of agreement ties, and that that interference is neatly modulated by morphological strength, in the sense that morphology acts a barrier to it. I suggest a series of components that a theory of agreement must contain if it is to be psyscholinguistically realistic. Among these: a. constant conceptual pressure (leaking) and varying degrees of morphologisation both inter- and intra-linguistically (blocking); b. constructional ‘listing’, which adds to the division of labour between direct semantic influence and encapsulated feature transmission; c. ‘Avalanching’ (a chain reaction of chunking), which results in near encapsulation in practice in the minds of speakers of languages with a rich morphology; d. a process of ‘Match and Check’, which ensures automatic, non-strategic computations of massive feature redundancy (Match) without the need for accompanying access to conceptual structure (Check).