CNL Lunch Talks

Andrew Ira Nevins

Hindi ERP, and a biolinguistic perspective on morphosyntactic licensing

Thursday December 18th 2003, 12:30 PM, 3416 Marie Mount Hall

Event-related potentials afford us the possibility to explore two biolinguistically relevant questions:

1) Is the mechanism for licensing inflectional morphology the same? What about when there are syntactic and semantic factors. Could connectionists make a distinction between types of grammatical licensing relation? Let's look on the scalp. Maybe there is something to * vs. # as used in linguistic judgements.

2) What is the nature of the agreement space charted by phi features? Is incorrect agreement in two features worse than dis-agreement in only one feature, and is this part of the linguistic computation, or a posthoc introspection on the nature of a violation? Let's look on the scalp. Maybe there is something to ** vs * as gradient degrees of total ungrammaticality.

We look at Hindi, because it has the right type of morphology to ask these questions. I don't promise definitive answers, but I will share the results of hundreds of hours of research.