The past two decades have witnessed the exploration of a range of grammatical formalisms by computational, theoretical, and psycho linguists, for their utility in building natural language interfaces and machine translation systems, characterizing the nature of human linguistic knowledge, and constructing models of language processing. Because many of these formalisms share important formal and linguistic properties (most prominently, mild context-sensitivity and lexicalization), there are many potential synergies, computational, theoretical and psychological, that can be gotten by considering ideas that stem from work outside of particular formalism. Moreover, though theoreticians, computationalists, and psychologists are concerned with solving different problems, ideas that derive from one community often turn out to have a significant consequences for the others.
The goal of this NSF-sponsored workshop, which will take place at Yale University on June 10-12, 2010, is to foster both of these types of connections: across the formalism divide and the theoretical-computational-psycho divide. The Tree-Adjoining Grammar community has a history of exploring these connections, and this workshop aims to expand the community of researchers involved in such cross-pollination even further. The workshop will bring together researchers from the Tree-Adjoining Grammar, Minimalism, Categorial Grammar, Dependency Grammar, HPSG, and LFG communities to look at the similarities and differences of the formalisms, with the goals of developing shared, broad-coverage grammars, transferring parsing and machine learning algorithms from one formalism to another. and gaining new insights into the properties of different formalisms and their capacity for linguistic and psycholinguistic explanation.
This award provides support crucial to attract to the workshop not only prominent researchers, who will give invited presentations and tutorial lectures, but also, and perhaps even more importantly, PhD students. By introducing junior researchers to the fruitfulness of cross-framework and cross-disciplinary interactions at an early stage in their careers, our hope is that the award will have a transformative effect on the kind of work they will engage in during their entire careers, potentially leading to a broader, more integrated perspective in the field at large.
The past two decades have witnessed the exploration of a range of mathematical systems (or grammars) for representing linguistic information. The systems have been used for building natural language interfaces and machine translation systems, characterizing human linguistic knowledge, and constructing models of human language processing. Because many of these systems share important mathematical and linguistic properties, there are many potential synergies, computational, theoretical and psychological, that can be gotten by considering ideas that stem from work outside of any one of them. Moreover, though theoreticians, computationalists, and psychologists are concerned with solving different problems, ideas that arise in one community often turn out to have a significant consequences for the others. This NSF-sponsored workshop, which took place at Yale University on June 10-12, 2010, fostered these types of connections: across the divide of different grammatical systems and across the divide of different disciplines (theoretical linguistics, computer science, and psychology). The workshop brought together researchers from the Tree-Adjoining Grammar, Minimalism, Categorial Grammar, and Dependency Grammar communities to look at the similarities and differences of the formalisms, with the goals of developing shared, broad-coverage grammars, transferring parsing and machine learning algorithms from one formalism to another and gaining new insights into the properties of different grammatical systems and their capacity for linguistic and psycholinguistic explanation. Among other topics, attendees explored novel approaches to the problem of encoding of the mapping between sentences and their meanings, and to the problem of learning language from the kind of input to which children are exposed. This award provided support crucial to allowing not only prominent researchers to attend this interdisciplinary workshop, but also, and perhaps even more importantly, a sizable group of (undergraduate and graduate) students in the fields of linguistics and computer science. By introducing these junior researchers to the fruitfulness of cross-framework and cross-disciplinary interactions at an early stage in their careers, our hope is that the award will have had a transformative effect on the kind of work they will engage in during their entire careers, potentially leading to a broader, more integrated perspective in the field at large.
