Going beyond the input: Three Factors and language acquisition, variation and change

Going beyond the input: Three Factors and language acquisition, variation and change

Invited talk by Theresa Biberauer, University of Cambridge

My objective is to outline a new generative model of language acquisition that accommodates the fact that the input acquirers receive does not reflect a single, invariant ‘parent grammar’, while still allowing us to understand (i) why transmission across speaker generations looks mostly successful, and (ii) why and how acquirers are able to innovate in distinctive ways in the absence of (unambiguous) input. I assume a “three factors” model (Chomsky 2005), in which Universal Grammar (Factor 1) is minimally specified; this interacts, in particular, with specific aspects of the input (Factor 2) – what I call systematic departures from Saussurean arbitrariness – and a general cognitive (Factor 3) bias, Maximise Minimal Means (MMM), to produce successively more adult-like I-language grammars. Importantly, the interaction of these three factors creates what I call NONE>ALL>SOME learning paths, which have also been observed outside syntactic learning (e.g. in phonology, and in lexico-cognitive concept formation), and also in concept formation more generally. I discuss experimental and real-life contexts where NONE>ALL>SOME patterns seem to have arisen where children are confronted with ambiguous or missing input (poverty of the stimulus scenarios), and also consider whether MMM might play a role in helping us to understand how adults learn and reshape language in L2 and contact scenarios.