El grupo de lingüística computacional del Instituto de Filología y Literaturas Hispánicas "Dr. Amado Alonso" invita a la charla abierta que dará la doctoranda Marina Ermolaeva (University of Chicago) el sábado 26 de septiembre a las 11 con el título "Learning minimalist grammars via lexical item decomposition". La charla tendrá una duración aproximada de una hora y media y será dictada en inglés a través de Skype. Para inscribirse, solicitamos completar el siguiente formulario.
Learning minimalist grammars via lexical item decomposition
Within the currently dominant framework for syntax, based around Chomsky's (1995, 2000) Minimalist Program, it is not uncommon to encounter multiple analyses of the same natural language pattern in the literature. A natural question, then, is whether one could evaluate and compare syntactic proposals from a quantitative point of view, approaching the issue as a learning problem. Minimalist grammars (Stabler 1997) are a natural choice for this task. As a formalization of the Minimalist Program, they are well-suited for implementing analyses of syntax phenomena, yet at the same time explicit in spelling out assumptions about syntactic units and operations.
Taking a step towards the goal outlined above, this presentation describes a principled way of making linguistic generalizations by detecting and eliminating syntactic and phonological redundancies in the data. As proof of concept, I will first introduce a toolkit of basic operations over lexical items and provide a small step-by-step example transforming a "naive" minimalist grammar over unsegmented words into an "optimized", linguistically motivated grammar over morphemes. I will then demonstrate and discuss a description of the English auxiliary system, passives, and raising verbs produced by a prototype implementation of the learning procedure based on these operations.