Papers due: June 30, 2013
Acceptance notice: July 15, 2013
Camera-ready version due:August 15, 2013
The goal of the GL conferences is to bring together diverse contributions from theoretical and computational linguistics, computer science, cognitive science, and lexicography, which explore compositionality from the point of view of generative approaches to the lexicon. Historically, contributions have assumed, as a starting point, the view outlined in Generative Lexicon theory (Pustejovsky, 1995, Pustejovsky et al, 2012).
Traditional topic of the GL conferences are:
- Polysemy and sense shifting
- Co-compositionality and creation of new word senses
- Type coercion and argument selection phenomena
- Argument realization: mapping from lexicon to syntax
- Cognitive foundations for semantic categories
- The trade-off between pragmatics and lexical knowledge
- Presupposition and commonsense knowledge
- Underspecification and word sense disambiguation
These topics can be approached from either a theoretical or computational perspective.
This year, GL2013 has a special focus on the relationship between generative approaches to the lexicon and distributional semantics. Distributional semantic models represent the meaning of lexical items in terms of vectors recording their pattern of distribution in linguistic contexts. They share with GL the goals of overcoming the limitations of classical models of the lexicon based on context-independent sense distinctions, and promoting a different view of lexical content generated in contexts and with contexts. The conference aims at exploring the potential synergies coming from the similarity as well as from the the complementarity of GL and distributional semantics. Possible topics include:
- classical semantic models that distributional semantics claims to be able to solve;
- current solutions and limits of distributional semantics theories to account for linguistic compositionality;
- prospects to enrich distributional semantics with robust first-order models of inference;
- the integration of distributional semantic principles and techniques into a broader dynamic model theoretic framework.