Grounding the Lexical Sets of Causative-Inchoative Verbs with Word Embedding

摘要

Lexical sets contain the words filling the argument positions of a verb inone of its senses. They can be grounded empirically through their automaticextraction from corpora. The purpose of this paper is demonstrating that theirvector representation based on word embedding provides insights onto manylinguistic phenomena, and in particular about verbs undergoing thecausative-inchoative alternation. A first experiment aims at investigating theinternal structure of the sets, which are known to be radial and continuouscategories cognitively. A second experiment shows that the distance between thesubject set and object set is correlated with a semantic factor, namely thespontaneity of the verb.