ELSNET Summer School 1997
14-25 July 1997

Gerald Gazdar

Cognitive and Computing Sciences
University of Sussex
Brighton, BN1 9QH, UK

Email: geraldg@cogs.susx.ac.uk
Phone: +44 1273 678029
URL: http://www.cogs.susx.ac.uk/lab/nlp/gazdar/gazdar.html

Recent papers

1. Lynne J. Cahill & Gerald Gazdar (1997) The inflectional phonology of German adjectives, determiners and pronouns. Linguistics, 35.2, 211-245. Postscript version available

2. Roger Evans & Gerald Gazdar (1996) DATR : A language for lexical knowledge representation. Computational Linguistics, 22.2, 167-216. Postscript version available

3. Gerald Gazdar (1996) Paradigm merger in natural language processing. In Robin Milner & Ian Wand, eds., Computing Tomorrow: Future Research Directions in Computer Science, Cambridge: Cambridge University Press, 88-109. Postscript version available

4. Lynne J. Cahill & Gerald Gazdar (1996) A lexical analysis of numeral expressions in three related languages. Proceedings of the AISB-96 Workshop on Multilinguality in the Lexicon, 69-75. Postscript version available

5. Lynne J. Cahill & Gerald Gazdar (1995) Multilingual lexicons for related languages. In Proceedings of the 2nd DTI Language Engineering Conference, 169-176. Postscript version available

6. Adam Kilgarriff & Gerald Gazdar (1995) Polysemous relations. In F.R. Palmer, ed. Grammar and Meaning: Essays in Honour of Sir John Lyons, Cambridge: Cambridge University Press, 1-25. Postscript version available

7. Roger Evans, Gerald Gazdar & David Weir (1995) Encoding lexicalized tree adjoining grammars with a nonmonotonic inheritance hierarchy. In Proceedings of the 33rd Annual Meeting of the Association for Computational Linguistics, 77-84. Postscript version available

8. Evans, Roger Gazdar, Gerald and Weir, David (1994) Using default inheritance to describe LTAG, in Proceedings of 3rd International TAG+ Workshop. Postscript version available

Research interests

1. Lynne Cahill and I are developing a trilingual computer lexicon for the core vocabulary of Dutch, English and German. From a linguistic perspective, we are ascertaining the extent to which these Germanic languages can be lexically related, examining formal ways of expressing linguistic generalizations that hold across two or more languages, and assessing the degree to which the historical links between languages can be exploited in descriptions of the languages as they are now. From a computational perspective, we are evaluating how well existing techniques for representing monolingual lexicons generalize to the multilingual case and investigating the extent to which multilanguage lexical representation techniques can be applied within monolingual lexicons.

   In addition to the trilingual lexicon itself, the project will also deliver automatically generated and consistently structured monolingual lexicons for Dutch, English and German. See Cahill & Gazdar, 1995; Cahill & Gazdar, 1996. Cahill & Gazdar, in press.

2. Roger Evans, Bill Keller and I have been responsible for the design of a formal language for lexical knowledge representation: DATR is a declarative language for representing a restricted class of inheritance networks, permitting both multiple and default inheritance. The principal intended area of application is the representation of lexical entries for natural language processing. The goal of the DATR enterprise is the design of a simple language that (a) has the necessary expressive power to encode the lexical entries presupposed by contemporary work in the unification grammar tradition, (b) can express all the evident generalizations about such entries, (c) has an explicit theory of inference, (d) is readily implementable, and (e) has an explicit declarative semantics. See:

   Roger Evans & Gerald Gazdar (1996) DATR : A language for lexical knowledge representation. Computational Linguistics, 22.2, 167-216. Postscript version available

   Bill Keller (1996) An evaluation semantics for DATR theories, in Proceedings of COLING-96 , 646-651. postscript version available

   Bill Keller (1995) DATR theories and DATR models, in Proceedings of the 33rd Annual Meeting of the Association for Computational Linguistics, 55-62. postscript version available

3. In work with Roger Evans and David Weir, I have been investigating the use of DATR to encode a Lexicalized Tree Adjoining Grammar (LTAG) lexicon as an inheritance hierarchy with internal lexical rules. Such an encoding eliminates the considerable redundancy otherwise associated with an LTAG lexicon. See Evans, Gazdar and Weir, 1995; Evans, Gazdar and Weir, 1994.

Former and current research students

• Roger Evans (1987) "Theoretical and computational interpretations of GPSG";
• Graham Russell (1988) "Verbal ellipsis in English coordinate constructions";
• Lynne Cahill (1990) "Syllable-based morphology for natural language processing";
• Bill Keller (1991) "Feature logics, infinitary descriptions and the logical treatment of grammar";
• Adam Kilgarriff (1992) "Polysemy";
• Ann Copestake (1992) "The representation of lexical semantic information";
• Alan Knowles (1993) "The syntax and semantics of 'each'";
• Jasper Taylor (1994) "A multi-agent planner for modelling dialogue";
• Anja Belz (current) on automatic acquisition of morphological lexical knowledge;
• Gabriel Illouz (current) on information extraction from chronological corpora;
• Diana McCarthy (current) on automatic acquisition of syntactic lexical knowledge;
• Makoto Shimizu (current) on a corpus-based comparison of coreference in English and Japanese.

Curriculum Vitae
NLP at the University of Sussex page
COGS home page
Email: geraldg@cogs.susx.ac.uk