Extended Abstract
Author: Marina Ermolaeva (University of Chicago)
Syntactic literature tends towards a big-picture outlook, abstracting away from details such as full specifications of lexical items or features involved in derivations. However, in order to identify quantifiable differences between competing analyses, a lower-level description is required. We frame this task as a learning problem, using the formalism of Minimalist Grammars. The learner makes morphosyntactic generalizations over a corpus of dependency structures and expresses them as standalone lexical items. The present work-in-progress explores how this approach can be used to induce linguistically plausible grammars and, ultimately, to evaluate proposals of theoretical syntax in quantitative terms.
Keywords: minimalist grammars, grammar induction, lexical item decomposition, dependency structures
How to Cite: Ermolaeva, M. (2020) “Induction of Minimalist Grammars over Morphemes”, Society for Computation in Linguistics. 3(1). doi: https://doi.org/10.7275/jed1-5j57