Colloquia Green

– Culture as Computation –

The generative enterprise in linguistics is associated with two major propositions about language and its place in nature. The first has been to argue that language can be understood as a computational innovation. The second has been to argue that the primary explanandum in the evolution of language is the emergence of a capacity for individual human brains to represent a specific formal class of grammatical structures. I’d like to argue that the first of these propositions is broadly right, but the second is too restrictive. Humans have achieved something computationally remarkable by building languages together, but the unique computational system responsible for this achievement is not (just) Merge: it is our species-unique platform for storing and re-using computations performed by earlier individuals — cumulative cultural evolution. I’ll present some of the work I have been doing to try to move towards a computational understanding of language at this level of analysis, using probabilistic inference as a unifying model for acquisition, transmission, and social interaction.

23 MAR 2017 | MPI Room 163 | 15.45 – 17.00