Archive Electronique - Institut Jean Nicod
Not a member yet
1997 research outputs found
Sort by
Les relations spatiales dans le raisonnement : un modèle computationnel
No full textStudies in the psychology of spatial reasoning often focus on positional problems, such as "A is to the left of B, C is to the right of B, where is A compared to C?". The computational model PRISM developed within the model-based approach, simulates in detail the psychological processes that enable us to solve this type of problem, especially by including relevant cognitive elements such as attentional limits. However, no computational model seems to exist for our reasoning about other types of relations, such as circle inclusion, or those used in geometry. This dissertation aims to fill this gap, in particular by combining the psychological principles already present in PRISM with the possibilities offered by the recent computational tool Answer Set Programming Modulo Theories for the Qualitative Spatial domain ASPMT(QS). Moreover, some theories of cognitive linguistics allow us to use the model on a larger part of language. This offers the possibility to shed new light on certain cognitive science topics, such as the study of logical connectives.Les études en psychologie du raisonnement spatial se concentrent généralement sur des problèmes de positionnement d'objets, comme par exemple "A est à gauche de B, C est à droite de B, où est A par rapport à C ?". Le modèle computationnel PRISM simule en détail les processus psychologiques qui nous permettent de résoudre ce genre de problèmes, notamment en intégrant des éléments cognitifs pertinents comme les limites attentionnelles. Toutefois, aucun modèle computationnel ne semble exister pour modéliser nos raisonnements sur d'autres types de relations, comme l'inclusion de cercles, ou celles utilisées en géométrie. Cette thèse propose un tel modèle, notamment en combinant les principes psychologiques déjà présents dans PRISM avec les possibilités offertes par le récent outil computationnel Answer Set Programming Modulo Theories for the Qualitative Spatial domain ASPMT(QS). En outre, des théories de linguistique cognitive nous autorisent à utiliser ce modèle sur une plus large partie du langage. Cela permet d'apporter un éclairage nouveau sur certains sujets en sciences cognitives, comme par exemple l'étude des connecteurs logiques
Caccia alla verità: Persuasione e propaganda ai tempi del virus e della guerra
No full textInternational audienc
Choice seeking is motivated by the intrinsic need for personal control
No full textWhen deciding between options that do or do not lead to future choices, humans often choose to choose. We studied choice seeking by asking subjects to decide between a choice opportunity or performing a computer-selected action. Subjects preferred choice when these options were equally rewarded, even deterministically, and were willing to trade extrinsic rewards for the opportunity to choose. We explained individual variability in choice seeking using reinforcement learning models incorporating risk sensitivity and overvaluation of rewards obtained through choice. Degrading perceived controllability diminished choice preference, although willingness to repeat selection of choice opportunities remained unchanged. In choices following these repeats, subjects were sensitive to rewards following freely chosen actions, but ignored environmental information in a manner consistent with a desire to maintain personal control. Choice seeking appears to reflect the intrinsic need for personal control, which competes with extrinsic reward properties and external information to motivate behavior. Author summary Human decisions can often be explained by the balancing of potential rewards and punishments. However, some research suggests that humans also prefer opportunities to choose, even when these have no impact on future rewards or punishments. Thus, opportunities to choose may be intrinsically motivating, although this has never been experimentally tested against alternative explanations such as cognitive dissonance or exploration. We conducted behavioral experiments and used computational modelling to provide compelling evidence that choice opportunities are indeed intrinsically rewarding. Moreover, we found that human choice preference varied according to individual risk attitudes, and expressed a need for personal control that competes with maximizing reward intake
Indirect illusory inferences from disjunction: a new bridge between deductive inference and representativeness
No full textInternational audienceWe provide a new link between deductive and probabilistic reasoning fallacies. Illusory inferences from disjunction are a broad class of deductive fallacies traditionally explained by recourse to a matching procedure that looks for content overlap between premises. In two behavioral experiments, we show that this phenomenon is instead sensitive to real-world causal dependencies and not to exact content overlap. A group of participants rated the strength of the causal dependence between pairs of sentences. This measure is a near perfect predictor of fallacious reasoning by an independent group of participants in illusory inference tasks with the same materials. In light of these results, we argue that all extant accounts of these deductive fallacies require non-trivial adjustments. Crucially, these novel indirect illusory inferences from disjunction bear a structural similarity to seemingly unrelated probabilistic reasoning problems, in particular the conjunction fallacy from the heuristics and biases literature. This structural connection was entirely obscure in previous work on these deductive problems, due to the theoretical and empirical focus on content overlap. We argue that this structural parallelism provides arguments against the need for rich descriptions and individuating information in the conjunction fallacy, and we outline a unified theory of deductive illusory inferences from disjunction and the conjunction fallacy, in terms of Bayesian confirmation theory
Contraste phonologique et inventaires des traits dans les langues des signes : étude de la Langue des Signes Française (LSF)
No full textIn this dissertation, we aim at providing a comprehensive analysis of contrast and feature inventories in sign language (SL) from a theoretical, empirical, and experimental point of view. We use data from French Sign Language (LSF), a language still underrepresented and for which little is known about its phonology. First, we dig into feature-based theoretical models to better understand the nature of SL inventories. An investigation of one particular location on the hand, the webbing part of the fingers, provides us with a better sense of how to determine the content of features inventories. We show that because of the difficulties in finding genuine minimal pairs, other forms of contrast must be used to compose phonological inventories, among which near-minimal pairs, productivity, as well as neutralization. Contrary to previous analyses based on American Sign Language (ASL), we provide evidence from LSF that the feature [web] must be included among the features that signers can use to create phonological contrast. Its status then led us to reanalyze (sub)location features as orientation features in some phonological models for SL. Our second study on features then allowed us to revisit the status of orientation in the composition of signs. In many phonological models, orientation is derived from a specification in handshape and a specification in location, while in others it is considered as an absolute property of the hand. Our analysis of two-handed signs produced on the signer's body in LSF provides evidence that both are necessary for the representation of this category of signs, and the implementation of secondary planes in a formal framework allows us to do so. In the third study, we capitalize at a more global level on the notion of feature and feature geometry models to capture complexity in sign articulation. We evaluate the ability of a theoretical model to reflect the complexity of signs in LSF. We compare an error-driven measure based on a repetition task, in which hearing non-signers had to copy existing signs, to a model-driven measure derived from the phonological structure of signs in the framework of the Prosodic Model (Brentari, 1998). Our results show that phonological models manage to capture overall articulatory complexity, as well as the one in handshape and location, but not in movement, the only dynamic phonological class. Altogether, these results indicate that an accurate theoretical model of sign phonology/phonetics reflects the degree of complexity as a result of the perceptual and articulatory properties of signs. In the last study, we focus on the psycholinguistic mechanisms behind the perception of contrast. We explore the perception of contrast at the experimental level by conducting a study on categorical perception (CP) in LSF, in both handshape and location. We show that a CP effect is observed in handshape in deaf signers, but not in hearing non-signers, a population never exposed to the phonological system of LSF. In location, no CP effect was observed which is in accordance with the literature. The unexpected significant effect observed in our allophonic pair in handshape puts into perspective the perception of (non-)contrastive units in SL and allows us to highlight the finesse of contrastiveness in the visual-gestural modality. From the various dimensions we have addressed in this thesis, we demonstrate that the study of LSF provides useful insights in the understanding of how distinctive features capture contrast in SL in general, as well as their role in determining sign complexity and perceiving contrast.Dans cette thèse, notre objectif est de fournir une analyse complète du contraste et des inventaires de traits dans les langues des signes (LS) d'un point de vue théorique, empirique et expérimental. Nous utilisons les données de la Langue des Signes Française (LSF), une langue encore sous-représentée et dont on connaît peu la phonologie. Nous explorons les modèles théoriques basés sur les traits afin de cerner la nature des inventaires des LS. L'étude d'un emplacement particulier sur la main, la partie palmée des doigts (webbing), permet de mieux comprendre comment déterminer le contenu des inventaires des traits. Nous montrons qu'en raison des difficultés à trouver de véritables paires minimales, d'autres formes de contraste doivent être utilisées pour composer les inventaires phonologiques, telles que les paires minimales approximatives, la productivité, ainsi que la neutralisation. Contrairement aux études sur la Langue des Signes Américaine (ASL), nous apportons la preuve à partir de la LSF que le trait [web] doit être inclus parmi les traits que les signants peuvent utiliser pour créer un contraste phonologique. Son statut nous a alors conduit à revoir les traits de (sous-)emplacement comme traits d'orientation dans certains modèles phonologiques. Notre deuxième étude sur les traits nous a ensuite permis de revoir le statut de l'orientation dans la composition des signes. Dans de nombreux modèles phonologiques, l'orientation est dérivée d'une spécification de la configuration manuelle et d'une spécification de l'emplacement, tandis que d'autres la considère comme une propriété absolue de la main. Notre analyse des signes à deux mains produits sur le corps du signant en LSF apporte la preuve que les deux sont nécessaires à leur représentation, et l'implémentation de plans secondaires dans un cadre formel nous permet de le faire. Dans la troisième étude, nous capitalisons à un niveau plus global sur la notion de traits et les modèles géométriques pour appréhender la complexité articulatoire des signes. Nous évaluons la capacité d'un modèle théorique à refléter la complexité des signes en LSF : nous comparons une mesure basée sur les erreurs de répétition de signes existants produites par des entendants non-signants à une mesure basée sur la structure phonologique des signes dérivée dans le cadre du Prosodic Model (Brentari, 1998). Nos résultats montrent que les modèles phonologiques saisissent la complexité articulatoire globale, ainsi que celle de la configuration manuelle et de l'emplacement, mais pas celle du mouvement, la seule classe phonologique dynamique. En somme, ces résultats indiquent qu'un modèle théorique précis de la phonologie/phonétique des signes reflète le degré de complexité associé aux propriétés perceptives et articulatoires des signes. Dans la dernière étude, nous examinons les mécanismes psycholinguistiques qui sous-tendent la perception du contraste. Nous explorons la perception du contraste à partir d'une étude sur la perception catégorielle (PC) de la configuration manuelle et de l'emplacement en LSF. Nous montrons qu'un effet de PC est observé dans la configuration manuelle chez les sourds signants, mais pas chez les entendants non-signants, une population jamais exposée à la phonologie de la LSF. Dans l'emplacement, aucun effet de PC n'a été observé, conformément à la littérature. L'effet significatif inattendu observé dans notre paire allophonique de la configuration manuelle remet en perspective la perception des unités (non-)contrastives dans les LS, et nous permet de mettre en évidence la finesse de la contrastivité dans la modalité visuo-gestuelle. A partir des différentes dimensions abordées dans cette thèse, nous démontrons que l'étude de la LSF apporte des éléments nécessaires à la compréhension de la manière dont les traits distinctifs rendent compte du contraste des LS en général, ainsi que leur rôle dans la détermination de la complexité des signes et dans la perception du contraste
Modulation des représentations corporelles par l’hypnose
No full textAlmost a century of modern scientific inquiry into hypnosis has revealed that it can have dramatic effects on perception, cognition, and behaviour. After a brief historical review, we argue against hypnosis being a unitary phenomenon. Therefore, we restricted our scope to a single hypnotic suggestion, namely modulating body-size representations. Our experimental work revealed that hypnotic suggestion is a reliable and robust tool for modulating these representations, both according to self-reports and implicit measurements. We interpret these effects as the result of classic situational and trait factors, and we highlight the important role of imagery and source monitoring. In this view, imagery taken for perception supports contextually cued expectancies, shaping primitive bodily representations.Près d'un siècle de recherche scientifique sur l'hypnose a révélé que celle-ci peut avoir des effets spectaculaires sur la perception, la cognition et le comportement. Après une brève revue historique, nous argumentons contre une vision unitaire de l’hypnose. Ainsi, nous avons restreint notre investigation à une unique suggestion hypnotique : la modulation des représentations de la taille du corps. Notre travail expérimental révèle que la suggestion hypnotique est un outil fiable et robuste pour moduler ces représentations, comme le révèlent les mesures perceptuelles et implicites. Au-delà du rôle des facteurs contextuels et des traits individuels, nous soulignons l’importance de l'imagerie mentale et du monitoring de source. Dans cette optique, l'imagerie confondue avec la perception appuie les attentes contextuelles, façonnant ainsi les représentations corporelles de bas niveau
Misinformation on Misinformation: Conceptual and Methodological Challenges
No full textAlarmist narratives about online misinformation continue to gain traction despite evidence that its prevalence and impact are overstated. Drawing on research examining the use of big data in social science and reception studies, we identify six misconceptions about misinformation and highlight the conceptual and methodological challenges they raise. The first set of misconceptions concerns the prevalence and circulation of misinformation. First, scientists focus on social media because it is methodologically convenient, but misinformation is not just a social media problem. Second, the internet is not rife with misinformation or news, but with memes and entertaining content. Third, falsehoods do not spread faster than the truth; how we define (mis)information influences our results and their practical implications. The second set of misconceptions concerns the impact and the reception of misinformation. Fourth, people do not believe everything they see on the internet: sheer volume of engagement should not be conflated with belief. Fifth, people are more likely to be uninformed than misinformed; surveys overestimate misperceptions and say little about the causal influence of misinformation. Sixth, the influence of misinformation on people's behavior is overblown as misinformation often 'preaches to the choir'. To appropriately understand and fight misinformation, future research needs to address these challenges
Cultural Evolution of Precise and Agreed‐Upon Semantic Conventions in a Multiplayer Gaming App
No full textInternational audienceThe amount of information conveyed by linguistic conventions depends on their precision, yet the codes that humans and other animals use to communicate are quite ambiguous: they may map several vague meanings to the same symbol. How does semantic precision evolve, and what are the constraints that limit it? We address this question using a multiplayer gaming app, where individuals communicate with one another in a scaled-up referential game. Here, the goal is for a sender to use black and white symbols to communicate colors. We expected that the players' mappings between symbols and colors would grow more specific over time, through a selection process whereby precise mappings are preferentially copied. We found that players become increasingly more precise in their use of symbols over the course of their interactions. This trend did not, however, result from selective copying of precise mappings. We explore the implications of this result for the study of lexical ambiguity, Zipf's Law of Meaning, and disagreements over semantic conventions
Fighting misinformation or fighting for information?
No full textInternational audienceA wealth of interventions have been devised to reduce belief in fake news or the tendency to share such news. By contrast, interventions aimed at increasing trust in reliable news sources have received less attention. In this article, we show that, given the very limited prevalence of misinformation (including fake news), interventions aimed at reducing acceptance or spread of such news are bound to have very small effects on the overall quality of the information environment, especially compared to interventions aimed at increasing trust in reliable news sources. To make this argument, we simulate the effect that such interventions have on a global information score, which increases when people accept reliable information and decreases when people accept misinformation
Give me your Intentions, I'll Predict Our Actions: A Two-level Classification of Speech Acts for Crisis Management in Social Media
No full textInternational audienceDiscovered by (Austin,1962) and extensively promoted by (Searle, 1975), speech acts (SA) have been the object of extensive discussion in the philosophical and the linguistic literature, as well as in computational linguistics where the detection of SA have shown to be an important step in many down stream NLP applications. In this paper, we attempt to measure for the first time the role of SA on urgency detection in tweets, focusing on natural disasters. Indeed, SA are particularly relevant to identify intentions, desires, plans and preferences towards action, providing therefore actionable information that will help to set priorities for the human teams and decide appropriate rescue actions. To this end, we come up here with four main contributions: (1) A two-layer annotation scheme of SA both at the tweet and subtweet levels, (2) A new French dataset of 6,669 tweets annotated for both urgency and SA, (3) An in-depth analysis of the annotation campaign, highlighting the correlation between SA and urgency categories, and (4) A set of deep learning experiments to detect SA in a crisis corpus. Our results show that SA are correlated with urgency which is a first important step towards SA-aware NLP-based crisis management on social media