Advertisement

Learnability of Pregroup Grammars

  • Denis Béchet
  • Annie Foret
  • Isabelle Tellier
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3264)

Abstract

This paper investigates the learnability of Pregroup Grammars, a context-free grammar formalism recently defined in the field of computational linguistics. In a first theoretical approach, we provide learnability and non-learnability results in the sense of Gold for subclasses of Pregroup Grammars. In a second more practical approach, we propose an acquisition algorithm from a special kind of input called Feature-tagged Examples, that is based on sets of constraints.

Keywords

Learning from positive examples Pregroup grammars Computational linguistics Categorial Grammars Context-Free grammars 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Lambek, J.: Type grammars revisited. In: Lecomte, A., Perrier, G., Lamarche, F. (eds.) LACL 1997. LNCS (LNAI), vol. 1582, p. 1. Springer, Heidelberg (1999)CrossRefGoogle Scholar
  2. 2.
    Bargelli, D., Lambek, J.: An algebraic approach to french sentence structure. In: [17]Google Scholar
  3. 3.
    Casadio, C., Lambek, J.: An algebraic analysis of clitic pronouns in italian. In: [17] Google Scholar
  4. 4.
    Kanazawa, M.: Learnable classes of categorial grammars. Studies in Logic, Language and Information. In: FoLLI & CSLI, Cambridge University Press, Cambridge (1998)Google Scholar
  5. 5.
    Gold, E.: Language identification in the limit. Information and control 10, 447–474 (1967)MATHCrossRefGoogle Scholar
  6. 6.
    Béchet, D., Foret, A.: Remarques et perspectives sur les langages de prégroupe d’ordre 1/2. In: Atala (ed.) Actes de la conférence internationale Traitement Automatique des Langues Naturelles, TALN 2003 (2003) (poster in French)Google Scholar
  7. 7.
    Dudau-Sofronie, D., Tellier, I., Tommasi, M.: Learning categorial grammars from semantic types. In: Proccedings of the 13rd Amsterdam Colloquium., pp. 79–84 (2001)Google Scholar
  8. 8.
    Dudau-Sofronie, D., Tellier, I., Tommasi, M.: A learnable class of ccg from typed examples. In: Proccedings of the 8th conference on Formal Grammar (FGVienna), pp. 77–88 (2003)Google Scholar
  9. 9.
    Dudau-Sofronie, D., Tellier, I.: A study of learnability of lambek grammars from typed examples. In: Categorial Grammars 2004, pp. 133–147 (2004)Google Scholar
  10. 10.
    Lambek, J.: Mathematics and the mind. In: Abrusci, V., Casadio, C. (eds.) New Perspectives in Logic and Formal Linguisitics, Proceedings Vth ROMA Workshop, Bulzoni Editore (2001)Google Scholar
  11. 11.
    Foret, A.: Some unlearnability results for lambek categorial and pregroup grammars (oral communication). In: ESSLII, Trento, Italy (2002)Google Scholar
  12. 12.
    Foret, A., Le Nir, Y.: Lambek rigid grammars are not learnable from strings. In: COLING 2002, 19th International Conference on Computational Linguistics, Taipei, Taiwan (2002)Google Scholar
  13. 13.
    Buszkowski, W.: Mathematical linguistics and proof theory.In: [18] , ch.12 , pp.683– 736 Google Scholar
  14. 14.
    Moortgat, M.: Categorial type logic.In: [18] ch. 2,pp. 93–177 Google Scholar
  15. 15.
    Buszkowski, W.: Lambek grammars based on pregroups. In:[17] Google Scholar
  16. 16.
    Nicolas, J.: Grammatical inference as unification. Technical Report 3632, IRISA, Unité de Recherche INRIA Rennes (1999)Google Scholar
  17. 17.
    de Groote, P., Morill, G., Retoré, C. (eds.): LACL 2001. LNCS (LNAI), vol. 2099. Springer, Heidelberg (2001)MATHGoogle Scholar
  18. 18.
    van Benthem, J., ter Meulen, A. (eds.): Handbook of Logic and Language. North- Holland Elsevier, Amsterdam (1997)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Denis Béchet
    • 1
  • Annie Foret
    • 2
  • Isabelle Tellier
    • 3
  1. 1.LIPN – UMR CNRSVilletaneuse
  2. 2.IRISA – IFSICRennes
  3. 3.MOSTRARE project RU Futurs INRIAGRAppA & Inria FutursLille

Personalised recommendations