CAuLD

CAuLD

Construction Automatique de représentations Logiques du Discours

Action de recherche collaborative (ARC) INRIA

Logical Methods for Discourse

Logical Methods for Discourse

December 13-14 2010, Nancy (France)

Room B013, LORIA

Presentation

As part of the CAuLD project, this workshop aims at sharing related views on logical and type-theoretical approachs to discourse modelling. This includes aspects related to the underlying formalisms or to the formal modelling of various phenomena.

Participants


(Tentative) Program

(Note that the schedule is still subject to changes.)

Monday Dec. 13

10:15-10:30
 Welcome (and coffee...)
10:30-11:30
 Categorial Grammar Inference (Slides)
 Noémie-Fleur Sandillon-Rezer
11:30-12:30
 Towards Wide-Coverage Semantics for French (Slides) Richard Moot
12:30-14:30
 Lunch
14:30-15:30
 Presupposition Accommodation as Exception Handling (Slides)
 Ekaterina Lebedeva
15:30-16:30
 L'adverbe "bientôt" : portée et repérage (Slides) Anaïs Lefeuvre
16:30-16:50
 Break
16:50-17:50
 Modeling Event Implication for Compositional Semantics (Slides) Sai Qian

Dinner at 20:00 at the Grand Café Foy

Tuesday Dec 14.

9:00-10:00
 Modal Subordination and Continuation Semantics (Slides)
 Sylvain Pogodalla and Nicholas Asher
10:00-11:00
 SDRT and Continuation Semantics (Slides)
 Nicholas Asher and Sylvain Pogodalla
11:00-11:20
 Break

11:20-12:20
 Generation with Grammars enriched with Lexical Semantics Information (Slides)
 Pierre Bourreau
12:30-14:30
 Lunch
14:30-15:30
 Constraints on the Simultaneous Resolution of Multiple presuppositions (Slides)
 Hans Kamp
15:30-16:30
 An SDRT based analysis of Pathological Dialogues (Slides)
 Maxime Amblard

Abstracts

Towards Wide-Coverage Semantics for French

Richard Moot

I will talk about the development of a wide-coverage categorial grammar for French and an associated semantic lexicon which, in combination with the Grail parser, form a first step towards wide-coverage semantic analysis for the French language. Two semantic lexicons are being developed, the first one producing discourse representation structures and the second one producing lambda-term semantics in Philippe de Groote's Montegovian Dynamics. I will discuss both syntactic and semantic choices made in the grammar and give a short demonstration of the current prototype system.

Presupposition Accommodation as Exception Handling

Ekaterina Lebedeva and Philippe de Groote

Van der Sandt's algorithm for handling presupposition is based on a ``presupposition as anaphora'' paradigm and is expressed in the realm of Kamp's DRT. In recent years, we have proposed a type-theoretic rebuilding of DRT that allows Montague's semantics to be combined with discourse dynamics. Here we explore van der Sandt's theory along the line of this formal framework.
It then results that presupposition handling may be expressed in a purely Montagovian setting, and that presupposition accommodation amounts to exception handling.

L'adverbe "bientôt" : portée et repérage

Anaïs Lefeuvre and Natalia Vinogradova

Afin de traiter les itinéraires dans le cadre du projet ITIPY, il nous faut ordonner les éventualités du discours en fonction, entre autres, des adverbiaux de localisation temporelle. Bientôt fait partie de la classe syntaxique des adverbes et observe un emploi typique des adverbiaux. Après étude des différents contextes où apparait cet adverbe au sein de notre corpus (récits de voyage du XIXème siècle en région pyrénéenne), nous proposons d'en préciser la portée et le repérage temporel en fonction d'autres marques. Nous avons aussi observé quelques cas qui nécessitent un traitement particulier. La notion de contexte dans le cadre de la DRT et d'anaphore temporelle telle qu'elle a été étudiée par Partee (1984) constitueront nos principaux outils. Nous tenterons alors de formuler les contraintes pour éclaircir la temporalité de l'éventualité à laquelle se rattache l'adverbe. Nous accorderons une attention particulière aux cas atypiques mentionnés.

Modal Subordination and Continuation Semantics

Sylvain Pogodalla and Nicholas Asher

Following [1]'s proposal of using continuations to model the dynamics of discourse, we propose a new way to build semantic representation of Modal Subordination. We show how different proposals (embedded and not embedded modalities, respective scope of quantifier and modalities, etc.) can be implemented compositionally without any change to the formalism, type theory, using only lexical semantics.

[1]  Philippe de Groote, 2006. "Towards a Montagovian account of dynamics". In proceedings of SALT 16, Tokyo.

SDRT and Continuation Semantics

Nicholas Asher and Sylvain Pogodalla

Segmented Discourse Representation Theory (SDRT) [1,2] provides a dynamic semantics for discourse that exploits a rich notion of discourse structure.  According to SDRT, a text is segmented into constituents related to each other by means of rhetorical relations; the resulting structure, known as a {\em segmented discourse representation structure} or SDRS has various semantic effects. This theory has shown how discourse structure makes contributions to the interpretation of a variety of linguistic phenomena, including tense, modality, presupposition, the interpretation of anaphoric pronouns and ellipsis.  SDRT exploits dynamic semantics [5,4] to interpret SDRSs.  We investigate here the advantages of integrating SDRT within continuation style semantics of the sort developed in [3].

[1] Asher, N.: Reference to Abstract Objects in Discourse. No. 50 in Studies in Linguistics and Philosophy, Kluwer, Dordrecht (1993)
[2] Asher, N., Lascarides, A.: Logics of Conversation. Cambridge University Press (2003)
[3]  Philippe de Groote, 2006. "Towards a Montagovian account of dynamics". In proceedings of SALT 16, Tokyo.
[4] Groenendijk, J., Stokhof, M.: Dynamic predicate logic. Linguistics and Philosophy 14(1), 39­100 (1991)
[5] Kamp, H.: A theory of truth and semantic representation. In: Groenendijk, J.A.G., Janssen, T.M.V., Stokhof, M.B.J. (eds.) Formal Methods in Study of Languages. Mathematical Centre Tracts, Amsterdam (1981)

Generation with Grammars enriched with Lexical Semantics Information

Pierre Bourreau

The recent emergence of linguistic formalisms based on typed lambda calculus led to a similar treatment of both the tasks of parsing and generation of natural language. Thanks to strong typing properties, Kanazawa showed that those tasks can be done in polynomial time in the case of linear terms (which leads to grammars where copies and deletions are not allowed) and almost linear terms (i.e. grammars on which a restricted copy is allowed), while Salvati proved decidability in the general case. In this talk, we extend Kanazawa's technique to allow deletion on the considered grammars and show the problem remains polynomial. We also show that deletion can be used in the design of our grammars to introduce lexical semantics information for instance.

Constraints on the Simultaneous Resolution of Multiple Presuppositions

Hans Kamp

Many natural language expressions and constructions trigger presuppositions. As a consequence it is common for sentences to generate not just one bur several presuppositions. Utterances of such sentences require verification that all the presuppositions generated by the sentence are justified in the context in which the utterance is made. Often this requires accommodation of some sort, usually to complement information that is already present in the context as it is given to the interpreter. In cases where some accommodation is needed, the accommodations which the interpreter is led to make often seem fully determined by the sentence and its discourse context - so much so that the accommodation comes across as an entailment of the discourse. To my knowledge the principles that determine what accommodations are to be chosen in such cases are still poorly understood.
Before questions concerning these principles can even be addressed we need a precise way of identifying the presuppositions that the interpreter must justify. That is, we need some way of identifying their logical forms or semantic representations, as part of a logical form or semantic representation for the sentence that gives rise to them (in DRT terms: as part of a “provisional” sentence representation in the sense of (Van Der Sandt, 1992)). To this end we need (a) rules for constructing provisional representations and (b) for the lexical presupposition triggers lexical entries that contain the information needed to construct the presuppositions generated by these lexical triggers.
In the talk I will discuss some of the issues arising in connection with the construction of presupposition representations and with the form and content of the lexical entries of (lexical) presupposition triggers. But the emphasis will be on the principles that guide accommodation as part of multiple presupposition resolution.

An SDRT based analysis of Pathological Dialogues

Maxime Amblard

The aim of this work is to combine pragmatic of interactions from psychology researches and formal discourse representations. We focus on a corpus of drive dialogues between schizophrenic and psychologist. In the first part, we propose SDRT representations of theses dialogues. To build these representations, we need to extend usual SDRT relations to specific ones for this corpus. We assume that schizophrenics are logically consistent and troubles appear in the discourse representation. Then, we introduce thematic structure which mark the theme coherence in the SDRT tree. We discuss pathological rules that schizophrenics seems to be able to use, especially the break of the right frontier and the rise through the tree without completeness of the thematic structure, and note that targets of theses rules seems to come from ambiguities (lexical, referential, syntactic, semantic, ...).

Practical informations

The workshop will be held at  the LORIA, room B013. There is a map of the campus describing the (pedestrian) way (about 500m)  from the Callot tram stop to the LORIA entrance (see the red arrow).  General informations to access the LORIA are available here. At the reception desk, you should ask for a visitor badge and, if you need one, for a WiFi access.

General informations on public transportation are available here. The easiest way from Nancy  downtown is to take the TRAM (line n°1 to "Vandoeuvre CHU Brabois").

The schedule of the shuttle for the Nancy downtown <-> Lorraine TGV station (XZI) or Nancy-Metz airport link (about 35 min. You have to buy your ticket in board: 4.20 euros single, 7.40 euros return) is here (note that the schedule will change on Dec 12th. So there might be some small differences. However, synchronisation with TGV will remain). Note that Lorraine TGV station is not the same as Nancy station (which is downtown).

The workshop dinner will take place Monday evening at the Grand Café Foy, which is on the "place stanislas" (see the access map).

On Sunday evening, there are very few restaurants that are open. In addition to l'Excelsior (which is on the opposite side of the "place Thiers" from the Nancy station, see the access map) or le Café de Foy, you may also try La Taverne Karlsbrau which is also close to the station, or try your luck around the "place stanislas".

Weather forecast

météo Nancy

Contact