W O R K S H O P D E S C R I P T I O N
Joan Bresnan and Chris Manning
Our main theme is that the same categorical phenomena which are attributed to hard grammatical constraints in some languages continue to show up as statistical preferences in other languages, motivating a grammatical model that can account for soft constraints.
In this workshop we introduce Stochastic Optimality Theory (Boersma and Hayes 2001), a way of modelling variation which extends the Optimality Theory framework, and show how this theme can be captured in this framework in the domain of syntax.
II.
We will work through a detailed example where Stocastic OT can provide a unified analysis: crosslinguistic variation in passivization (Bresnan, Dingare, and Manning 2001; Dingare 2001; Manning 2002). Time permitting, we will give a brief overview of related recent work in other areas of syntax.
III.
We will provide an introduction to software for doing Stochastic OT, focussing particularly on a tutorial on Boersma's Praat system, an open source system (usable on Windows, Macintosh, Linux and other Unices) which among many other tasks such as phonetic analysis can do Stocastic OT analyses.
IV.
Finally, we will make some comparisons between Stochastic OT and other approaches in current sociolinguistics, in particular with the logistic regression model found in Varbrul.
[Workshop description revised: 9/19/02]