Workshop description

The notion of simplicity played a central role in early generative linguistics as an evaluation metric for both for the linguist and for the child acquiring the grammar. A simplicity-based metric was appealing for a variety of reasons: it is a very general criterion for comparing hypotheses, and it works directly with the representations provided by linguistic theory, which in turn opens the way for the linguist to compare theories of UG based on their predictions regarding learning.

This workshop will bring together members of MIT’s Linguistics and Brain & Cognitive Sciences departments to share current work that revisits the relevance of simplicity criteria in the study of language. Our goal is to build a dialogue between the two departments around questions of common interest about simplicity such as: Can a simplicity-based evaluation metric work? If so, how should simplicity be measured? Can current work on learning in linguistics be reconciled with a simplicity-based metric? And can the idea of theory comparison using a general learning criterion be made to work?

Presenters and commentators will be from the local MIT community but attendees and poster submissions from elsewhere are welcome. Our hope is that this small forum will generate interest in the topic and may form the foundation for a larger workshop in the future.

Abstracts for posters can be submitted to by August 3, 2017.