This contribution aims at giving an overview of automatic speech recognition research, highlighting the needs for corpora development. As recognition systems largely rely on statistical approaches, large amounts of both spoken and written corpora are required. In order to fill the gap between written and spoken language, speech transcripts need to be produced manually using appropriate tools. Methods and resources accumulated over the years now allow, not only to tackle genuine oral genres, but also to envision large-scale corpus studies to increase our knowledge of spoken language, as well as to improve automatic processing.

Corpus Taurinense (CT) is the POS tagged version of ItalAnt Corpus, an electronic corpus of Old Italian texts (between 1251 and 1300). In this article we aim to describe the approach followed in CT for the annotation of multiword units (MWU). MWU in our work is a set of two or more graphic words which receive (also) an overall POS tagging because this set of words is in paradigmatic relation with one word lexical unit with the same POS.Our POS tagging confirms that most of the Modern Italian compound conjunctions at that time were not lexicalised. The order of the components is already the Modern Italian order but they can still be interrupted by occasional elements.

Metrics and rythmics aspects are examined on a 80 000-verse corpus, analysed with computational linguistics tools. We propose a cumulative experimental approach consisting in building a verse pattern with a series of features (morpho-syntactic, stress, rhyme.). Features may characterize units of different levels (syllables, hemi-verse, verse, etc.) and are evidenced by different tools, but all are integrated in a single database. Thus we can verify classic metric rules and hypotheses. We also document new regularities, for example on stress patterns, and we test some new hypotheses about links between traits and patterns. This empirical approach on a large corpus, beyond verification of hypotheses, may lead to the construction of grounded theories.

This article aims to demonstrate how a corpus of spoken French, that was started in Aix-en-Provence around 1975, has developed over time in connexion with the development of what has been called since 'corpus linguistics'. The story of that corpus is related here, whilst at the same time the possibilities for exploitation that it offers today are traced briefly.

It is important to realise that human translation is difficult and diverse, and that automation is needed not only by end users, but also by translators and Interpreters. Also, automation itself comes in many forms. After briefly describing computer tools for translators, we will concentrate on the linguistic and computer approaches to the automation of translation proper. This survey will yield an array of criteria for categorizing existing CAT systems, with brief examples of the state of the art. Finally, we present perspectives of future research, development, and dissemination.

The aim of this paper is to focus on the necessity of a double marking-out when doing the semantic analysis of the data of a corpus. The first mark lies in the situation in which texts are produced and the second one lies in the interpretation of the texts. In both cases, the author suggests to use the notion of 'genre' (textual 'genre' and interpretative 'genre') in order to classify and categorize situations. The issue is exemplified by the problem of the building of terminologies according to a particular interpretative 'genre'. The paper shows how textual 'genre' influences the functioning of conceptual relations patterns (e.g. the preposition avec is used to spot a meronymic relation). It demonstrates that this kind of analysis may help to refine the descriptions initially made by introspection.

In this article, the authors describe first how the Corpus of Medieval Portuguese has been constitued, in particular how it has been coded ; secondly, an attempt will be made at demonstrating how it can be used for the construction of a dictionary of medieval Portuguese, more specifically of its verbs, proper and common nouns.

The first part of this article explains which methodological issues need to be examined in order to establish and transcribe a large corpus of texts written by pupils, along with their school context. The second part of the article states the various lines of epistemological questioning which led to a second research project, i.e. questions about how to define types of school writing as well as a corpus and context, and about the necessary links between those three elements. A variety of software programs was used to analyse corpora which were not in conformity with orthographical and stylistical standards. Such a use seems possible, joined with qualitative analysis.

This paper is part of the PROLEX project of automatic processing of proper nouns. Our objective consists, with the help of the computer, not only in identifying the various occurrences of the definite proper noun (modified or not modified), but also in tagging them with a relevant type of interpretation: referential, denominative, model, metaphoric or split. After the elaboration of a typology of the various uses of the definite proper noun, we try to extract the formal and lexical indications allowing for the deambiguation of the referential and semantic functioning of the proper noun. After isolating these differential units (e.g. determiners, adjectives, predicates of existence), we build local grammars intended for automatic recognition.

In this article, we present a novel approach to syntactic parsing. As opposed to usual methods that consider syntactic parsing as a series of processes, we suggest here an architecture based on hybrid agents which task is robust deep syntactic parsing. After a short summary of the research on which we have based our work we present by means of an example the theoretical functioning of the architecture. This then allows us to describe the APA architecture, developped by Girault, that we have used in this joint research. Finally, in conclusion, we present some perspectives for implementation.

Online tourist guides are information tools communicating destination image and specialised knowledge at the same time. They feature a large variety of lexicographic structures including word lists, articles, conceptual schemes, indexes and registers, search options on keywords, internal and external cross references etc. This is by no means surprising in so far as what is needed is effective data access in order to extract information – precisely in the same way as in lexicography. The functional thesis we defend in this article is that lexicographisation in a user perspective can improve the access process. Taking œnotouristic online guides as a case in point, we will examine different user situations leading to consultation, in particular the need for experiential information, in which users simply wish to improve the conditions of their œnotouristic experience. We will then formulate theoretical proposals aimed at ensuring better interaction of lexicographic functions, data presentation and access possibilities.

After having reminded what distinguishes a database from a corpus, the author of the present article gives a rapid overview of the most important documentary sources that exist today in the domain of French diachrony, after which she demonstrates first, what use can be made of them thanks to a series of tools available today, and second, and more important, how the access to large corpora allow us to review our analysis of linguistic facts, and invites us to a qualitative change in our linguistic reasoning.

This paper presents concepts and systems for multilingual terminological and textual knowledge codification, representation, validation, management and sharing structured around the notion of genre. These systems operationalize the different stages of the ‘virtuous knowledge cycle’ within a dynamic, multilingual specialized web-dictionary and a multilingual, genre-based corpus of medical texts genre hierarchies or systems. The knowledge cycle approach mirrors ‘real life’ working processes and allows for repeated conversions of knowledge between its tacit and explicit forms, allowing knowledge to codify and spiral up from the individual to the collective level at corporate, ‘community of practice’. The paper reports on the results of the implementation of these concepts and systems in general and the web-dictionary in particular within the Danish health care, pharmaceutical, medical device and translation sectors which technologically have been fused into one collective ‘knowledge cluster’ and it discusses the opportunities for research and business that spring from fusion of language and health technologies.