Cet enseignant est référent pour cette UE
S'il s'agit de l'enseignement principal d'un enseignant, le nom de celui-ci est indiqué en gras.
Mardi de 14 h à 16 h (EHESS-Marseille, Centre de la Vieille-Charité, 2 rue de la Charité 13002 Marseille), du 8 novembre 2016 à juin 2017, calendrier ci-dessous
Experiments have become part of the scientific apparatus in social science quite recently, if one considers the status they acquired as a superior proof practice in the late seventeenth century in natural science. For the second year, the seminar invites participants from different disciplines, such as animal studies, cognitive sciences, economics, linguistics, neurosciences, philosophy, psychology, etc., to analyse the experimental paradigm, its historical developement and its epistemological constraints and open, as practitioners, the blackbox of experimental settings . This seminar aims at reflecting on the status of experiments in the social sciences comparatively, the part they play in providing insights and proofs, the methodological and rhetorical differences in experiments and experimental settings, the necessity and role of « incentives » in eliciting « truthful » answers, the transfer from results in the lab or in the « field » to the « world », the status of reference-models to which experimental results are compared, etc.
This seminar will be jointly organized by researchers at different sites of the EHESS and other universities by using a video-conferencing system.
Tuesday November 8, 2017, 2pm - 4pm : Paolo Marocco (Università di Genova), « Organizing lab experiments using platform Amazon Mechanical Turk »
Tuesday December 13, 2017, 2pm - 4pm : Carole Treibich (SESSTIM-Greqam, Marseille), « Behavioral economics in the field : Estimating misreporting of condom use of female sex workers in Senegal »
Tuesday January 10, 2017, 2pm - 4pm : Jean-François Bonnefon (Toulouse School of Economics, Institute for Advanced Study in Toulouse, Center for Research in Management), « Ethical challenges for self-driving cars »
Even if self-driving cars prove safer than conventional cars, their widespread adoption is imperiled by a moral challenge. If a crash is unavoidable and the car must decide whether to save its passengers or to save pedestrians, what should it do? Minimize casualties (even if it means harming its passengers), or protect its passengers at all costs (even if it means harming a greater number of pedestrians)? Studying problems of this sort, we identified a no-win situation for manufacturers of self-driving cars. Either they offer cars that protect their passengers at all costs, and they invite moral outrage (because people have a strong feelings that self-driving cars should minimize casualties impartially); or they offer cars that minimize casualties impartially, and they scare consumers away (because people want their own car to protect them first). I will describe these findings, discuss the reaction of car manufacturers, and talk of our ongoing work, including the Moral Machine website which already attracted several millions of users.
Tuesday February 7, 2017, 2pm - 4pm : Antoinette Baujard (GATE Lyon Saint-Étienne), « The sense of the act of voting depends on rules. A teaching from in situ voting experiments »
We have conducted voting experiments during the first round of the French Presidential elections in 2007, 2012 and so shall we in 2017. These experiments allow to compare the current 2-round plurality rule, i.e. a mononominal rule, with variants of multinominal rules, i.e. where voters may vote for each candidate, among which approval voting and evaluative voting. This communication presents the threefold teaching to be derived from these data.
Firstly, when voting rules vary, results show significant discrepancies in the expression of electoral preferences, rankings of candidates, relative importance of candidates and, last but not least, in the winning candidates (Baujard et al. 2014).
Secondly, beyond the straigthforward teaching on the distinct properties of distinct rules, we learn that individual behaviors adapt to contexts. Voters may consider these different rules as different exercices (e.g. Lebon et al. 2017).
Thirdly, from one exercice to another, the sense of the act of voting for each voter evolves. This communication shall make the bold conjecture that this discrepancies may be due to the distinct domains of definition of voting rules. In mononominal rules, voting implies an individual thinking of the final collective decision: the expression of one voter are given the same aim and properties than the collective decision. Alternatively in multinominal rules including information on intensity of preferences, voting suppose that individuals take part in a collective endeavour: different properties are required whether for each individual vote or for the collective decision.
Tuesday, March 14, 2017
Tuesday, April 11, 2017 (2pm-4pm) : Stéphanie Dupouy (philosophie, Université de Strasbourg) & Mark Wexler (sciences cognitives, CNRS, Université de Paris-5 Descartes),
Tuesday, May 30 (2pm-4pm) : John Tresch (University of Pennsylvania), « Experimental Scenographies in the Science of the Meditating Brain »
Tuesday, June 6 (2pm-4pm) : Colin Tredoux (University of Cape Town et Visiting Professor of University of Toulouse I): "The Ecology of Experiments in Forensic Psychology"
Mots-clés : Épistémologie,
Aires culturelles : Europe,
Intitulés généraux :
pour tous renseignements, veuillez contacter Miriam Teschl par courriel.
Adresse(s) électronique(s) de contact : miriam.teschl(at)ehess.fr
Dernière modification de cette fiche par le service des enseignements (firstname.lastname@example.org) : 24 mars 2017.