Polymnie Fourth Meeting

Subject to changes!

/doku.php?id=projects:polymnie:events:fourth-meeting

Le projet du French Discourse Treebank (FDTB) consiste à ajouter une couche d’annotation discursive sur un corpus (journalistique) déjà annoté pour la morpho-syntaxe, le French Treebank [1]. Ce projet s’inspire de celui mené à l’université de Pennsylvanie, le Penn Discourse TreeBank [2]. L’exposé propose de présenter les avancées et les futures étapes de ce projet, qui se divise en différentes phases. La première phase (FDTB1) a consisté à repérer de façon semi-automatique toutes les occurrences de connecteurs de discours dans le corpus sur la base d’une liste aussi complète que possible de 350 connecteurs de discours du français, celle de LexConn [3]. La seconde phase (FDTB2), en cours, consiste à annoter pour chaque connecteur repéré les arguments qu’il lie et la relation de discours qu’il exprime, et ce à partir d’une nomenclature de relations de discours inspirée surtout de celle proposée pour le PDTB, mais aussi de celle proposée pour le corpus ANNODIS [4] (elle-même inspirée des travaux en RST et SDRT). Enfin, on proposera, dans un troisième temps, de repérer dans le corpus et de dresser la liste de ce que les auteurs du PDTB nomment AltLex (Alternative Lexicalisations), c’est-à-dire des expressions compositionnelles qui expriment une relation de discours qui pourrait également être exprimée par un connecteur (par exemple, “la raison en est que” qui exprime la relation que “parce que”).

We present a method to interface a sentential grammar and a discourse grammar without resorting to an intermediate processing step. The method is general enough to build discourse structures that are direct acyclic graphs (DAG) and not only trees. Our analysis is based on Discourse Synchronous TAG (D-STAG), a Tree-Adjoining Grammar (TAG)-based approach to discourse. We also use an encoding of TAG into Abstract Categorial Grammar (ACG). This encoding allows us to express a higher-order semantic interpretation that enables building DAG discourse structures on the one hand, and to smoothly integrate the sentential and the discourse grammar thanks to the modular capability of ACG.

Chung-chieh Shan and Simon Charlow have both observed that the type constructions used by formal semanticists tend to share the same algebraic structure: that of a monad. This structure is very well known to computer scientists who use it to study the semantics of programming languages. I will stress the point that parallels between the study of side effects in programming languages and pragmatics in natural languages are not accidental and hint at some of the benefits that can arise from adopting a theory of side effects in a natural language. I will close by evaluating the suitability of such a framework wrt. to more recent type constructions of natural language semantics, such as the treatment of double negation proposed by Sai Qian.

[1] A. Abeillé, L. Clément, and F. Toussenel, “Building a treebank for french,” in Treebanks. building and using parsed corpora, A. Abeillé, Ed., Springer, 2003, p. 165–187.
[Bibtex]

@Inbook{abeille-et-al:2003,
author="Abeill{\'e}, Anne and Cl{\'e}ment, Lionel and Toussenel, Fran{\c{c}}ois",
editor="Abeill{\'e}, Anne",
title="Building a Treebank for French",
bookTitle="Treebanks. Building and Using Parsed Corpora",
year="2003",
publisher="Springer",
pages="165--187",
crossref="abeille2003",
doi="10.1007/978-94-010-0201-1_10",
}

[2] R. Prasad, N. Dinesh, A. Lee, E. Miltsakaki, L. Robaldo, A. Joshi, and B. Webber, “The penn discourse treebank 2.0.,” in Proceedings of the sixth international conference on language resources and evaluation (lrec’08), Marrakech, Morocco, 2008-05-28/2008-05-30 2008.
[Bibtex]

@InProceedings{prasad-et-al:2008,
author = {Rashmi Prasad and Nikhil Dinesh and Alan Lee and Eleni Miltsakaki and Livio Robaldo and Aravind Joshi and Bonnie Webber},
title = {The Penn Discourse TreeBank 2.0.},
booktitle = {Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)},
year = {2008},
month = {may},
date = {2008-05-28/2008-05-30},
address = {Marrakech, Morocco},
editor = {Nicoletta Calzolari and Khalid Choukri and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis and Daniel Tapias},
publisher = {European Language Resources Association (ELRA)},
isbn = {2-9517408-4-0},
pdf = {http://www.lrec-conf.org/proceedings/lrec2008/pdf/754_paper.pdf}
}

[3] C. Roze, L. Danlos, and P. Muller, “LEXCONN: a french lexicon of discourse connectives,” Discours, vol. 10, 2012.
[Bibtex]

@Article{roze-et-al:12,
author = {Charlotte Roze and Laurence Danlos and Philippe Muller},
title = {{LEXCONN}: A French Lexicon of Discourse Connectives},
journal = {Discours},
year = 2012,
volume = 10,
doi = {10.4000/discours.8645}
}

[4] S. Afantenos, N. Asher, F. Benamara, M. Bras, C. Fabre, M. Ho-Dac, A. L. Draoulec, P. Muller, M. Pery-Woodley, L. Prevot, J. Rebeyrolles, L. Tanguy, M. Vergez-Couret, and L. Vieu, “An empirical resource for discovering cognitive principles of discourse organisation: the ANNODIS corpus,” in Proceedings of the eight international conference on language resources and evaluation (lrec’12), Istanbul, Turkey, 2012-05-23/2012-05-25 2012.
[Bibtex]

@InProceedings{afantenos-et-al:2012,
author = {Stergos Afantenos and Nicholas Asher and Farah Benamara and Myriam Bras and Cecile Fabre and Mai Ho-Dac and Anne Le Draoulec and Philippe Muller and Marie-Paul Pery-Woodley and Laurent Prevot and Josette Rebeyrolles and Ludovic Tanguy and Marianne Vergez-Couret and Laure Vieu},
title = {An empirical resource for discovering cognitive principles of discourse organisation: the {ANNODIS} corpus},
booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},
year = 2012,
month = {may},
date = {2012-05-23/2012-05-25},
address = {Istanbul, Turkey},
editor = {Nicoletta Calzolari and Khalid Choukri and Thierry Declerck and Mehmet Uğur Doğan and Bente Maegaard and Joseph Mariani and Asuncion Moreno and Jan Odijk and Stelios Piperidis},
publisher = {European Language Resources Association (ELRA)},
isbn = {978-2-9517408-7-7},
pdf = {http://www.lrec-conf.org/proceedings/lrec2012/pdf/836_Paper.pdf},
url = {http://www.lrec-conf.org/proceedings/lrec2012/summaries/836.html}
}