Guy Denhiere | Centre National de la Recherche Scientifique / French National Centre for Scientific Research (original) (raw)

Uploads

Papers by Guy Denhiere

Le Langage Et L Homme, 2004

EMOSEM is a psychological model and an automatic analysis system of emotions in texts. Element of... more EMOSEM is a psychological model and an automatic analysis system of emotions in texts. Element of the text analysis platform SEMOTEX, EMOSEM links to the analysis of latent semantics (LSA), lexical resources of the emotional categories offered in literature. The results of the analysis of twelve test-texts representing six emotions, positive (" Joy " , " Trust " , " Desire ") and negatives (" Sadness " , " Fear " , " Anger ") are presented and discussed.

Understanding texts : how to help 7 year olds construct the coherence in the overall meaning of a... more Understanding texts : how to help 7 year olds construct the coherence in the overall meaning of a text ? Résumé L'automatisation par l'enfant des processus d'identification et de reconnaissance des mots est une étape nécessaire à l'apprentissage de la lecture à la fin du cycle 2. Cependant, dès cette période, il semble important de contribuer à la mise en place ou au renforcement de sa capacité à construire la cohérence de la signification locale et globale d'un texte. Le projet Diagnos-Remédiation dans lequel s'inscrit ce travail a pour but de concevoir et de valider des séquences d'apprentissage et de remédiation adaptées aux dysfonctionnements repérés au cours d'épreuves de la batterie Diagnos-Lecture™ (2). L'objectif de notre recherche est d'évaluer, chez des enfants de C.E. 1 , l'effet de séquences d'aide à la construction de la cohérence de la signification globale de ce qui est dit dans un texte.

EMOVAL est un modèle d'analyse dimensionnelle de la valence affective des textes de la plateforme... more EMOVAL est un modèle d'analyse dimensionnelle de la valence affective des textes de la plateforme d'analyse SEMOTEX. Il utilise un ensemble de 6 normes représentant 5582 mots différents caractérisés en valence et en activation et dont les valeurs sont significativement corrélées entre elles. Les tests réalisés confortent la pertinence du modèle. Nous présentons également les principales fonctionnalités de la plateforme.

The aim of this study was to investigate the construction of a retrieval structure during reading... more The aim of this study was to investigate the construction of a retrieval structure during reading, according to the hypothesis that text macrostructure is used in Long-term working memory to maintain encoded information in an accessible format. We first designed an experiment for testing the hypothesis that retrieval structure is a macrostructure of the text. Then, we conceived and run a model inspired by CI-LSA Framework in which a generalization process create macropropositions. Results are that simulation data were found to be highly correlated with participants' data.

Behavior Research Methods, 2006

In the present article, we outline the architecture of a computer program for simulating the proc... more In the present article, we outline the architecture of a computer program for simulating the process by which humans comprehend texts. The program is based on psycholinguistic theories about human memory and text comprehension processes, such as the construction-integration model (Kintsch, 1998), the latent semantic analysis theory of knowledge representation (Landauer & Dumais, 1997), and the predication algorithms (Kintsch, 2001; Lemaire & Bianco, 2003), and it is intended to help psycholinguists investigate the way humans comprehend texts.

Computing Research Repository, 2008

In contrast to many LSA semantic spaces in the literature which are based on domain -specific cor... more In contrast to many LSA semantic spaces in the literature which are based on domain -specific corpora, we chose to build a general child corpus intended to model children's semantic memory (Denhière, Lemaire, Bellissens & Jhean -Larose, 2004) and to offer a layer of basic semantic associations on top of which computational models of children cognitive processes can be designed and simulated (Lemaire, Denhière, Bellissens & Jhean -Larose, to appear).

The hypothesis of study-time regulation states that Ss studying a material, spend on each item a ... more The hypothesis of study-time regulation states that Ss studying a material, spend on each item a time related to its objective difficulty. In this experiment, difficulty is defined according to interstimulus similarity (high, medium, or low) and the items are given a different incentive value (10 points, or 1 point). Thirty-six Ss learned 8 paired-associates with self-paced presentation. The Ss interrupted their study when they judged they had achieved a perfect mastery of the material; they were submitted to a double retention test (recall and recognition). The results showed that the higher the interstimulus similarity of the items, the longer their study-time and the lower the number of correct responses in the tests. Study-time generally decreased from cycle to cycle during the periods of study; but between these periods it rose after the test for the Ss who resumed study. This evolution of study-time interacted with interstimulus similarity. No effect of incentive value was evidenced.

L'année psychologique, 1982

Reeducation Orthophonique, 1983

RefDoc Bienvenue - Welcome. Refdoc est un service / is powered by. ...

The resonance model framework (Myers & O’Brien, 1998) assumes a passive, fast-acting resonance pr... more The resonance model framework (Myers & O’Brien, 1998) assumes a passive, fast-acting resonance process as the basic mechanism for retrieving prior information. This process is triggered as soon as a semantic and contextual overlap is present between information currently processed and information stored in long term memory. We investigated the hypothesis that a specific overlap, related to the temporo-causal structure of the text, may influence the resonance process. Results stress the importance of contextual overlap for the resonance process and highlight the role of text structure in the manipulation of the overlap.

Enfance, 1994

L'identification d'items mono-et bisyllabiques par des enfants apprentis lecteurs. Cett... more L'identification d'items mono-et bisyllabiques par des enfants apprentis lecteurs. Cette recherche a pour but d'étudier l'influence de facteurs relevant de la structure de la langue (structure syllabique, écart de sonorité, distance ...

Revue française de pédagogie, 1983

Acta Psychologica, 1972

Le Langage Et L Homme, 2004

Behavior Research Methods, 2006

Computing Research Repository, 2008

L'année psychologique, 1982

Reeducation Orthophonique, 1983

RefDoc Bienvenue - Welcome. Refdoc est un service / is powered by. ...

Enfance, 1994

Revue française de pédagogie, 1983

Acta Psychologica, 1972

Four lexical decision tasks with varying presentation duration of idioms (240 vs 340 ms) and inte... more Four lexical decision tasks with varying presentation duration of idioms (240 vs 340 ms) and interval between Idiom and Target (50, 300 and 450 ms) were done to test the "Graded Salience" Hypothesis (Giora, 1997, 1999) vs "Configurational" hypothesis proposed by Cacciari & Tabossi (1988). Idioms extracted from Idimatic™ data-base (G. Denhière & J.C. Verstiggel, 1998) were controlled for frequency, familiarity, conventionality and typicality and varied for salience. Results are in good agreement with G.S. hypothesis.

In this article, we study the organization of emotions in memory supported by strategical process... more In this article, we study the organization of emotions in memory supported by strategical processes. Three hypotheses are tested. If emotions are structured in a n-dimension spatial, to which emotional characteristics return these dimensions. If there is a discrete organization of emotions, is it possible to give an account of it through the organization of emotions in memory. Finally, can semantic relations between terms give an account of their emotional relations. Two word sorting experiments are performed. In the first experiment, 128 pleasant and unpleasant words are divided into 8 lists. In the second experiment, two lists of 128 words, one positive and the other negative are each divided into 8 groups of 16 words. Results support the hypothesis of an organization around a core affect (Russell, 2003), reinforce the existence of discrete emotions, and confirm the predictive function of semantic relations on the organization of emotions.

EMOSEM est un modèle psychologique et un système d’analyse automatique de la coloration émotionne... more EMOSEM est un modèle psychologique et un système d’analyse automatique de la coloration émotionnelle de textes. Composante de la plateforme d’analyse de textes SEMOTEX, EMOSEM associe à l’analyse de la sémantique latente (LSA), les ressources lexicales des catégories émotionnelles proposées dans la littérature. Les résultats de l’analyse
de six textes-tests décrivant 12 émotions positives et négatives (« Joie », « Confiance »,« Désir », « Tristesse », « Peur », « Colère ») sont présentés.

Une adaptation française de l'échelle de vocabulaire de Biemiller (2001, 2008) a été administrée ... more Une adaptation française de l'échelle de vocabulaire de Biemiller (2001, 2008) a été administrée à des élèves de Cours Préparatoire (CP) et de Cours Élémentaire (CE1). Les résultats obtenus sont comparés à ceux obtenus par Biemiller et simulés à l'aide de l'Analyse de la Sémantique Latente (LSA).