Publications by F. Lareau

Le temps verbal dans l'interface sémantique-syntaxe du français

16 Jun 2009 12:00:00 EST

English title: Verbal tense in the semantics-syntax interface of French
F. Lareau (2009)
In Proceedings of MTT, Montreal: 223-232.

We briefly introduce a new model of French verbal tenses, where two complementary inflectional categories, tense and shifting, work together to locate facts in time. The category of shifting poses a reference point in the past or in the non-past. The category of tense locates facts in relation with this reference point. The core of this article is dedicated to the modeling of tense and shifting grammemes in the semantics-syntax interface of the Meaning-Text Unification Grammar of French. We also discuss the question of grammatical idioms and collocations.

Download the PDF.

Applying the Meaning-Text Theory Model to Text Synthesis with Low- and Middle-Density Languages in Mind

15 Jan 2009 12:00:00 EST

L. Wanner, F. Lareau (2009)
In S. Nirenburg (ed), Language Engineering for Lesser-Studied Languages, IOS Press, Amsterdam.

Large scale text generation requires substantial resources, namely grammars and lexica. While these resources are increasingly compiled for high density languages, for low and middle density languages often no generation resources are available. The question on how to obtain them in a most efficient way becomes thus imminent. In this article, we address this question for MTT-oriented text generation resources.

This publication is not available in PDF. You can get the book from your local Amazon store: ca / us / uk / de / fr.

Vers une grammaire d'unification Sens-Texte du français : le temps verbal dans l'interface sémantique-syntaxe

11 Dec 2008 12:00:00 EST

English title: Towards a Meaning-Text Unification Grammar of French: Verbal Tense in the Semantics-Syntax Interface
F. Lareau (2008)
Ph.D. thesis, University of Montreal / University Paris Diderot (Paris 7).

This thesis has two main objectives. First, we aim at describing French verbal inflection, more particularly tense in the semantics-syntax interface. Second, we want to develop the Meaning-Text unification grammar (MTUG) formalism and test its adequacy for language description.

To achieve our first goal, we propose a lexicography-inspired methodology that is centered on the linguistic sign as a whole. At the descriptive level, our main thesis is that French has not only one, but two inflectional categories of tense that complement each other. The category of shifting comprises the grammemes SHIFTED and NON-SHIFTED, which locate a temporal reference point in relation with the time of speech. The category of tense proper comprises the grammemes ANTERIOR, SIMULTANEOUS and POSTERIOR, which situate facts in relation with this reference point.

At the formal level, our main contributions to the MTUG formalism are the representation of structures in terms of objects and functions on these objects, as well as the modeling of semantic decompositions. The formalization of our French inflection model turns out to be an interesting way of testing MTUG’s adequacy, since this phenomon involves signs of different natures. It also provides a good opportunity to observe the articulation of the grammar’s modules.

Download the PDF.

Automatic Production of Multilingual Environmental Information

14 Sep 2007 12:00:00 EST

B. Bohnet, F. Lareau, L. Wanner (2007)
In Proceedings of EnviroInfo, vol. 2, Warsaw: 59-66.

Multilingual environmental information is communicated via different media. These media can be newspapers, TV, internet, WAP, SMS, etc. Each medium has a presentation mode which fits best. Thus, for newspapers, pictograms, maps, and/or short text are to be preferred. In contrast, the information provided on the internet can be very detailed and personalized and contain the latest data available. Furthermore, on the internet, the user can interactively select more details or change the presentation mode. For all media, the most challenging information mode is text. Since a template-based method where predefined sentences with empty slots filled at the time of generation cannot ensure coherent and cohesive text for all contextual settings, full-fledged generation techniques are needed. In this paper, we present the generation techniques as used for the production of multilingual air quality information in the framework of the MARQUIS project.

Download the PDF.

Towards a Generic Multilingual Dependency Grammar for Text Generation

15 Jul 2007 12:00:00 EST

F. Lareau, L. Wanner (2007)
In Proceedings of GEAF, Stanford: 203-223.

For practical multilingual text generation, efficient development and representation of large scale grammatical and lexical resources are crucial. One way to ensure efficiency is to share as much resources as possible between languages. We present some preliminary work on shared grammatical resource development within the framework of the Meaning-Text Theory, emphasizing the lexicalist point of view. We show that rich dictionaries allow for more generic grammar rules which can be used for several languages, so that the number of language-specific rules is kept low. We also discuss the benefits of shifting the workload to the dictionaries from the viewpoint of extension and consistency control as well as the impact it has on the organization of work. Evaluation of the developed shared grammatical resources constitutes another topic of the article.

Download the PDF.

Vers une formalisation des décompositions sémantiques dans la Grammaire d'Unification Sens-Texte

08 Jun 2007 12:00:00 EST

English title: Towards a formalization of semantic decompositions in the Meaning-Text Unification Grammar
F. Lareau (2007)
In Actes de TALN, Toulouse.

We propose a formal representation of meaning decomposition in the framework of the Meaning-Text Unification Grammar. The proposed technique aims at offering a better integration of such semantic decompositions into a global model of the language. It relies on the saturation of polarities to control the construction of decomposed respresentations as well as their mapping to the syntactic trees that express them. The proposed formalism is discussed from the viewpoint of generation, but it applies to analysis as well.

Download the PDF.

From Measurement Data to Environmental Information: MARQUIS - A multimodal air quality information service for the general public

22 May 2007 12:00:00 EST

L. Wanner, D. Nicklaß, B. Bohnet, N. Bouayad-Agha, J. Bronder, F. Ferreira, F. Friedrich, A. Karppinen, F. Lareau, A. Lohmeyer, A. Panighi, S. Parisio, H. Scheu-Hachtel, J. Serpa (2007)
In Proceedings of ISESS, Prague.

This article introduces the Multimodal AiR QUality Information Service for General Public MARQUIS. The service, which is based on air quality measurements, is able to provide information on air quality in Europe in a way that is easily understandable for the general public. The goal of the project is presented below with a special focus on the role of the users. Furthermore the architecture of the system and the components are described.

Download the PDF.

On the Challenge of Creating and Communicating Air Quality Information

22 May 2007 12:00:00 EST

L. Wanner, B. Bohnet, N. Bouayad-Agha, F. Lareau, A. Lohmeyer, D. Nicklaß (2007)
In Proceedings of ISESS, Prague.

Building environmental information systems is a challenging task. It involves a detailed analysis of the addressee’s needs, development of air pollutant concentration assessment and interpretation techniques, which must be tailored to the addressee, as well as the development of information planning and production strategies. Environmental engineers are increasingly often in charge of setting up environmental services. They must be thus knowledgeable in a number of issues they did not face before. We present an air quality information system MARQUIS, which is usable, on the one hand, as a base service that can be extended and adapted to a new environmental information scenario, and, on the other hand, as a source of teaching and training material for environmental engineers.

Download the PDF.

Meaning-Text Unification Grammar: modularity and polarization

25 Jun 2005 12:00:00 EST

S. Kahane, F. Lareau (2005)
In Proceedings of MTT, Moscow.

This article presents the Meaning-Text Unification Grammar’s current state, now that its formal foundations have been clarified with the development of Polarized Unification Grammar. Emphasis is put on the model’s architecture and the role of polarization in linking its modules — semantics-syntax interface and the well-formedness grammars of each representation level.

Download the PDF.

Grammaire d'Unification Sens-Texte : modularité et polarisation

10 Jun 2005 12:00:00 EST

English title: Meaning-Text Unification Grammar: modularity and polarization
F. Lareau (2005)
In Actes de TALN, Dourdan: 23-32.

This article presents the Meaning-Text Unification Grammar’s current state, now that its formal foundations have been clarified with the development of Polarized Unification Grammars. Emphasis is put on the model’s architecture and the role of polarization in linking the various modules — semantic-syntax interface, syntax-morphotopology interface and the well-formedness grammars of each representation level. We discuss how various polarity neutralization strategies can control different analysis and generation procedures.

Download the PDF.

Towards an explanatory-combinatorial dictionary of Japanese

29 Aug 2003 12:00:00 EST

F. Lareau (2003)
In Proceedings of Asialex, Tokyo.

Within the Meaning-Text Theory, the vocabulary of a language is described by an Explanatory-Combinatorial Dictionary (ECD). An ECD of a language aims at describing exhaustively all three components of its lexical units: their meaning, their form and their combinatorics. It describes not only the syntactic and morphological behavior of words, but also their restricted lexical cooccurrence, via lexical functions. Until now, most of the work in explanatory-combinatorial lexicography has been done on French and Russian. The main goal of this paper is to show that it would be possible and useful to develop an ECD of Japanese, and to discuss potential problems inherent to this task and propose some leads towards a solution.

Download the PDF.

La synthèse automatique de paraphrases comme outil de vérification des dictionnaires et grammaires de type Sens-Texte

15 Nov 2002 12:00:00 EST

English title: The automatic synthesis of paraphrases as a tool for the verification of Meaning-Text dictionaries and grammars
F. Lareau (2002)
M.A. dissertation, University of Montreal.

This dissertation presents an automatic text synthesis prototype system, Sentence Garden, which is intended to be used as a “test lab” for the validation of Meaning-Text linguistic models. The originality of this system lies mainly in the fact that it aims at producing not only one sentence, but all possible sentences that express a given meaning. Meaning-Text models are introduced first from a theoretical point of view, then from the viewpoint of their encoding in Sentence Garden.

Download the PDF.

Moyo comme un petit peuple tranquille

01 Mar 2002 12:00:00 EST

English title: Moyo as a small, quiet nation
F. Lareau (2002)
In Hiatus, vol. 3: 6-9.

This is an informal article that was published in the Linguistics Student Association's journal at the University of Montreal. It is an editorial view of French Canadian's opinion of their own dialect.

This publication is not available in PDF.

A practical guide to writing DiCo entries

15 Jul 2001 12:00:00 EST

F. Lareau (2001)
In Proceedings of Papillon, Grenoble.

This paper is intended for people who are already somewhat familiar with the DiCo project. Its purpose is to offer a practical, step-by-step guide to writing DiCo entries.

Download the PDF.