WORKSHOP ON COLLOCATION

Computational Extraction, Analysis and Exploitation

ACL'2001 Conference
Toulouse, France
July 7th, 2001

The term "collocation" was introduced in the nineteen thirties by J. R. Firth, founder member of the British Contextualist school, to characterise certain linguistic phenomena of cooccurrence that stem principally from the linguistic competence of native speakers (Firth 1957). By its very nature collocation remains a relatively fuzzy concept, the consequence of which being that traditional grammarians and semanticists have tended to ignore it, the exception being some lexical semanticists as Cruse (1986). The study of collocation is above all a practical one aimed at assisting language learners and translators in their tasks.

Essentially idiomatic in nature, collocation defies rigid formalisation which explains the existence of different schools of thought between those seeking a descriptive contextualised view of linguistic phenomena and those who seeks formalised applications for translation, lexicography or computational purposes. This has led to a variety of approaches based around a general core meaning for the phenomenon.

For several years, NLP has been concerned with collocation largely through the following fields:

• Formalisation through specialised formalisms for different NLP tasks: dictionary formalism such as lexical function; HPSG, LFG, TAG, ... formalisms for analysis or generation.
• Extraction from monolingual or bilingual texts or dictionairies using either raw statistics or statistics combined with linguistic information such as part-of-speech or grammar dependancy.
• Exploitation through specific NLP systems dedicated to second language learning or translation, or for such NLP tasks as information retrieval or thematic structuration.

• Formal description of collocation through existing or  dedicated specialised formalisms
• New methods adopted for the identification of collocations. This would  include statistics and also more linguistic oriented methods.
• NLP systems dedicated to collocation.
• Exploitation of collocations for other NLP tasks through monolingual or multilingual environments.

ORGANIZERS

• Béatrice Daille     IRIN - University of Nantes, France - daille@irin.univ-nantes.fr
• Geoffrey Williams     CRELLIC - University of Bretagne-Sud, France - Geoffrey.Williams@univ-ubs.fr

PROGRAM COMMITTEE

• Jeremy Clear, Honorary Research Fellow, University of Birmingham
• Pernilla Danielsson, TELRI
• Chris Gledhill, University of St Andrews
• Syvain Kahane,  LaTTiCe/TALaNa
• Marie-Claude L'Homme, University of Montreal
• Julia Pajzs, Hungarian Academy of Science
• Antoinette Renouf, University of Liverpool
• Alain Polguère, OLST - University of Montreal
• Laurent Romary, LORIA
• Dan Tufis, Romanian Academy - RACAI
• Jean Véronis, University of Provence
• Leo Wanner, University of Stuttgart

 

SCHEDULE

Workshop paper submissions

April 8, 2001

Notification of acceptance

April 30, 2001

Deadline for camera-ready papers

May 13, 2001

WORKSHOP DATE

SUBMISSION FORMAT AND INSTRUCTIONS