Archive Electronique - Institut Jean Nicod
Not a member yet
1997 research outputs found
Sort by
The spectrum of perspective shift: protagonist projection versus free indirect discourse
No full textInternational audienc
When Is It OK to Call Someone a Jerk? An Experimental Investigation of Expressives
No full textInternational audienceWe present two experimental studies on the Italian expressive 'stronzo' (English 'jerk'). The first study tests whether, and to which extent, the acceptability of using an expressive is sensitive to the information available in the context. The study looks both at referential uses of expressives (as in the complex demonstrative 'that jerk Marco') and predicative uses of expressives (as in 'Marco is a jerk'). The results show that expressives are sensitive to contextual information to a much higher degree than the non-expressive control items (such as 'Piedmontese') in their referential use, but also, albeit to a lesser degree, in their predicative use. The second study tests whether the lower acceptability of expressives in their predicative use may be simply due to saying something negative about someone. A comparison between expressives, such as 'jerk', and non-expressive negative terms, such as 'nasty' or 'unbearable', suggests that it is the expressive nature of these terms, rather than the mere negative valence, that affects acceptability. Our studies present a major challenge to the existing accounts of expressives, and raise several theoretical issues that still call for an answer
Social deliberation systematically shifts resource allocation decisions by focusing on the fate of the least well-off
No full textInternational audienceHow much inequality should be tolerated? How should the poorest be treated? Though sometimes conflated, concerns about inequality and the fate of the poorest involve different allocation principles with different sociopolitical implications. We tested whether deliberation—the core of democracy—influences reasoning about distributive principles. 322 participants faced allocation decisions for others between egalitarian (low variance in allocation), utilitarian (high total amount), and maximin (maximizing the welfare of the poorest) options. After their initial decisions, participants either reflected upon similar decisions solely or discussed them in pairs before facing the same choices again individually. Social, but not solitary, deliberation led to more maximin and fewer egalitarian choices, and this change lasted at least 5 months after the experiment. Conversation analyses of approximately 7500 utterances suggest that some participants initially made egalitarian choices heuristically, when in fact they mostly cared about the poorest, and dialogue promoted more internally coherent maximin preferences
Comprendre la désinformation et promouvoir l'information fiable
No full textAmericans are more worried about misinformation than about sexism, racism, terrorism, and climate change. Fears over misinformation on social media are overblown. Misinformation represents a minute proportion of the news that people consume online (~ 1%), and a small minority of people account for most of the misinformation consumed and shared online. People, on average, are good at detecting fake news and identifying reliable sources of information. People do not believe everything they see and read on the internet. Instead, they are active consumers of information who domesticate technologies in unexcepted ways. It’s very unlikely that social media exacerbates the misinformation problem, that fake news contributes to important political events or that falsehoods spread faster than the truth. Yet, some fake news stories do go viral, and understanding why, despite their inaccuracy, they go viral is important. In a series of experiments, we identified a factor that, alongside accuracy, drives the sharing of true and fake news: the ‘interestingness-if-true’ of a piece of news, e.g. if alcohol was a cure against COVID-19, the pandemic would end in an unprecedented international booze-up. In three experiments (N = 904), participants were more willing to share news they found more interesting-if-true, as well as news they deemed more accurate. They rated fake news less accurate but more interesting-if-true than true news. People may not share news of questionable accuracy by mistake, but instead because the news has qualities that compensate for its potential inaccuracy, such as being interesting-if-true. Despite these qualities, why are most people are reluctant to share fake news? To benefit from communication, receivers should trust less people sharing fake news. And the costs of sharing fake news should be higher than the reputational benefits of sharing true news. Otherwise we would end up trusting people misleading us half of the time. Four experiments (N = 3,656) support this hypothesis: sharing fake news hurts one’s reputation in a way that is difficult to fix, even for politically congruent fake news. Most participants asked to be paid to share fake news (even when politically congruent), and asked for more when their reputation was at stake. During the second part of my PhD, I tested solutions to inform people efficiently. I found that discussing in small groups the scientific evidence on Genetically Modified (GM) food safety and the usefulness of vaccines changed people’s minds in the direction of the scientific consensus. To scale up the power of discussion, we created a chatbot that emulated the most important traits of discussion. We found that rebutting the most common counterarguments against GMOs with a chatbot led to more positive attitudes towards GMOs than a non-persuasive control text and a paragraph highlighting the scientific consensus. However, the dialogical structure of the chatbot seemed to have mattered more than its interactivity. During the pandemic, we deployed a chatbot to inform the French population about COVID- 19 vaccines. Interacting a few minutes with this chatbot, which answered the most common questions about COVID-19 vaccines, increased people’s intention to get vaccinated and had a positive impact on their attitudes towards the vaccines. In the end, people are not stupid. When provided with good arguments, they change their mind in the direction of good arguments. Most people avoid sharing misinformation because they care about their reputation. We do not live in a post-truth society in which people disregard the truth. Overall, we should probably be more concerned about the large portion of people who do not trust reliable sources and are uninformed because they do not follow the news, rather than the minority of people who trust unreliable sources and are misinformed.Les fausses nouvelles affolent. Les Américains sont plus préoccupés par la désinformation que par le sexisme, le racisme, et le changement climatique. Ces craintes sont très largement exagérées. La désinformation ne représente qu'une infime portion des nouvelles consommées en ligne (~ 1 %) et une petite minorité de gens est à l'origine de la majorité des fausses informations consommées et partagées en ligne. En moyenne, les gens sont capables de reconnaître les fausses nouvelles et d'identifier les sources d'information fiables. Les gens ne croient pas tout ce qu'ils voient et lisent sur l'internet. Il est peu probable que les réseaux sociaux exacerbent le problème de la désinformation, que les fausses nouvelles aient contribué à des événements politiques importants ou que les fausses nouvelles se répandent plus vite que la vérité. Cependant, certaines fausses nouvelles sont virales, et il est intéressant de comprendre pourquoi, malgré leur manque de fiabilité, ces fausses nouvelles deviennent virales. Au cours d'une série d'expériences, nous avons identifié un facteur qui motive le partage des vraies et des fausses nouvelles : "l'intérêt-si-vrai" d'une nouvelle, e.g. si l’alcool était un remède contre la COVID-19 il suffirait de faire la fête pour se protéger du virus. Au cours de trois expériences en ligne (N = 904), les participants étaient plus disposés à partager des nouvelles qu'ils trouvaient plus intéressantes-si-vraies, ainsi que des nouvelles qu'ils jugeaient plus fiables. Ils considéraient les fausses nouvelles comme moins fiables mais plus intéressantes-si-vraies que les vraies nouvelles. Les gens pourraient partager des fausses nouvelles non pas par erreur, mais plutôt parce que ces nouvelles possèdent des qualités qui compensent pour leur manque de fiabilité, comme le fait d'être intéressantessi- vraies. Malgré ces qualités, pourquoi la plupart des gens sont-ils réticents à partager des fausses nouvelles ? Quatre expériences (N = 3 656) montrent que le partage de fausse nouvelle nuit à la réputation de son transmetteur d'une manière difficile à compenser par le partage de vraies nouvelles. La plupart des participants demandèrent à être payés pour partager des fausses nouvelles, et ce montant était d’autant plus important que leur réputation était en jeu. Durant le deuxième parti de mon doctorat j’ai mesuré l’efficacité d’interventions pour informer efficacement les gens. J’ai montré que discuter en petits groupes des preuves scientifiques portant sur la sûreté des organismes génétiquement modifiés (OGM) et de l'utilité des vaccins, influençait l’opinion des gens en direction du consensus scientifique. Pour étendre le pouvoir persuasif de la discussion, nous avons développé un chatbot simulant les caractéristiques les plus importantes d’une discussion. Interagir avec ce chatbot réfutant les contre-arguments les plus courants contre les OGMs entraîna des attitudes plus positives à l'égard des OGMs que plusieurs conditions de contrôle (N = 1306). Pendant la pandémie, nous avons déployé un chatbot répondant aux questions les plus courantes sur les vaccins COVID-19. Interagir quelques minutes avec ce chatbot augmenta l'intention des gens de se faire vacciner et eu un impact positif sur leurs attitudes envers les vaccins. Au final, les gens ne sont pas stupides. Lorsqu'on leur présente de bons arguments, ils changent d'avis en direction de ces bons arguments. La plupart des gens évitent de partager des fausses nouvelles par souci pour leur réputation. L’ère de la « post-vérité » n’existe pas, la fiabilité de l’information est aussi importante aujourd’hui que par le passé. Dans l'ensemble, il est probablement plus important de se préoccuper du grand nombre de gens qui ne font pas confiance aux sources fiables et ne sont pas informés parce qu'ils ne suivent pas l'actualité, plutôt que de la minorité de gens qui font trop confiance aux sources douteuses et sont mal informées
Ground Zero. The Triangle of Humiliation
No full textInternational audienceThe eponymous heroine of The Lost Honour of Katharina Blum, Heinrich Böll's 1974 novel [made into a film in 1975 by Volker Schlöndorff and Margarethe von Trotta] is a candid, honest and reliable housekeeper who finds herself killing a journalist in cold blood-out of the humiliation she feels at having her self-image sullied in public after having been accidentally involved in a scandal. She had been innocent of any wrongdoing: her only fault was to have danced one night with a young man who turned out to be a dangerous criminal. But she is nonetheless eviscerated in the papers; her mother, employers and ex-husband are interrogated; and she is so ashamed of this public image in which she does not recognize herself that, blinded indeed by humiliation, she commits the crime for which she is then condemned
Combining Vagueness Detection with Deep Learning to Identify Fake News
No full textPaper to appear in the Proceedings of the 24th International Conference on Information Fusion. Johannesburg. (2nd version: Typo corrected in metadata in one of the authors' names)International audienceIn this paper, we combine two independent detection methods for identifying fake news: the algorithm VAGO uses semantic rules combined with NLP techniques to measure vagueness and subjectivity in texts, while the classifier FAKE-CLF relies on Convolutional Neural Network classification and supervised deep learning to classify texts as biased or legitimate. We compare the results of the two methods on four corpora. We find a positive correlation between the vagueness and subjectivity measures obtained by VAGO, and the classification of text as biased by FAKE-CLF. The comparison yields mutual benefits: VAGO helps explain the results of FAKE-CLF. Conversely FAKE-CLF helps us corroborate and expand VAGO's database. The use of two complementary techniques (rule-based vs data-driven) proves a fruitful approach for the challenging problem of identifying fake news
Social information use and social information waste
No full textInternational audienceSocial information is immensely valuable. Yet we waste it. The information we get from observing other humans and from communicating with them is a cheap and reliable informational resource. It is considered the backbone of human cultural evolution. Theories and models focused on the evolution of social learning show the great adaptive benefits of evolving cognitive tools to process it. In spite of this, human adults in the experimental literature use social information quite inefficiently: they do not take it sufficiently into account. A comprehensive review of the literature on five experimental tasks documented 45 studies showing social information waste, and four studies showing social information being over-used. These studies cover ‘egocentric discounting’ phenomena as studied by social psychology, but also include experimental social learning studies. Social information waste means that human adults fail to give social information its optimal weight. Both proximal explanations and accounts derived from evolutionary theory leave crucial aspects of the phenomenon unaccounted for: egocentric discounting is a pervasive effect that no single unifying explanation fully captures. Cultural evolutionary theory's insistence on the power and benefits of social influence is to be balanced against this phenomenon. This article is part of the theme issue ‘Foundations of cultural evolution’
Quantity–quality trade-off in the acquisition of token preference by capuchin monkeys (Sapajus spp.)
No full textNational audienceMoney represents a cornerstone of human modern economies and how money emerged as a medium of exchange is a crucial question for social sciences. Although non-human primates have not developed monetary systems, they can estimate, combine and exchange tokens. Here, we evaluated quantity–quality trade-offs in token choices in tufted capuchin monkeys as a first step in the investigation of the generalizability of tokens as reinforcers, which is a potentially relevant factor underlying the emergence of money in humans. We measured capuchins' exchange preferences when they were repeatedly provided with 10 units of three token types yielding food combinations varying in quantity and quality. Overall, capuchins maximized their quantitative payoff, preferring tokens associated with a higher food amount, rather than showing violations of rationality. However, some individuals did not maximize their qualitative payoff, possibly because of conditional valuation effects or owing to the choice overload phenomenon, according to which too many options reduce the accuracy of choice. Our study supports the importance of comparative research to finely analyse the multiple components shaping the economic behaviours of other species, possibly to achieve a more comprehensive, evolutionary- and ecologically based understanding of human economic behaviour
Fifty Shades of Affective Colouring of Perception
No full textInternational audienceRecent evidence in cognitive neuroscience indicates that the visual system is influenced by the outcome of an early appraisal mechanism that automatically evaluates what is seen as being harmful or beneficial for the organism. This indicates that there could be valence in perception. But what could it mean for one to see something positively or negatively? Although most theories of emotions accept that valence involves being related to values, the nature of this relation remains highly debated. Some explain valence in terms of evaluative content, others in terms of evaluative attitude. Here I shall argue that an account of affective perception in terms of attitude has more chance to succeed. To do so, I will first highlight the difficulties that a content-based approach faces, considering the many forms it might take. I will conclude that seeing the world positively or negatively involves more than a positive or negative content; it involves a distinctive attitude, but which one? Should it be conceived in imperative or evaluative terms? And what makes this attitude distinct from a proper emotion