Institut Jean Nicod

Accueil > Séminaires/Colloques > Archives > Colloques > 2010-2011 > Coordination and cooperation : game-theoretical and cognitive (...)



Coordination and cooperation : game-theoretical and cognitive perspective

EHESS, 190-198 av de France 75013 Paris, salle Jean-Pierre Vernant (8e étage) - May 26-27, 2011

Abstracts: 

Guillaume Dezecache (IJN, LNC), Laurence Conty (LNC), Michele Chadwick (LNC), Dan Sperber (IJN), Julie Grezes (LNC)
« I fear your fear of his fear »: an experiment of transitive emotional contagion

Crowds are often defined by their extraordinary homogeneity. This work aims at identifying one of the mechanisms by which such homogeneity might emerge. We assume that it may occur through local emotional transmission: to perceive the fear of an individual (through facial expressions, body expressions and vocalizations) is sufficient to affect the emotional experience of an observer. In turn, this observer can affect the emotional experience of another observer, and so on, till the emotional homogeneity of the crowd is achieved. To test this hypothesis, we recorded the physiological cues of participants placed in a local situation of transitive emotional contagion (which we assume occurs in crowds on a much larger scale). Within this paradigm, we recorded the physiological cues of an individual C observing an individual B watching an actor (individual A) displaying either fearful or joyful expressions. We also tested the hypothesis that the transitive emotional contagion would be more intense when B and C are engaged in an affective relationship with one another. Our preliminary results suggest that the information contained in facial expressions is sufficient to produce a transitive emotional contagion, confirming that a local emotional transmission model can be a good candidate to help account for emotional transmission in crowds. 

Natalie Sebanz (Radboud University Nijmegen & Central European University, Budapest)
Is there a ‘we-mode’ in joint action planning?

Psychological research has shown that individuals performing joint actions effortlessly accommodate to the perspective, tasks, and actions of their co-actors. However, there is little evidence to suggest that a switch to a ‘we-mode’ occurs such that actions are not just planned from an individual perspective but rather from a joint perspective. I will present new results from experiments on inter-group mimicry that provide evidence for a ‘we-mode’ in action planning. In particular, the findings suggest that individuals form action plans that specify the actions to be performed jointly. I will discuss how these findings link up with accounts of team reasoning that postulate the existence of a ‘we-mode’.


Cédric Paternotte (University of Bristol)
Some more reasoning about we-reasoning

Hakli et al. (2010) recently argued that collective reasoning is irreducible to individual reasoning by highlighting the similarities between two accounts of we-reasoning, a philosophical one (Tuomela's we-mode) and a game-theoretic one (Bacharach's unreliable team interactions). This paper argues that although their conclusion is probably correct, it is not for the reasons they provide. Indeed, the similarities between the accounts have been overestimated, the difference between collective and individual reasoning in Bacharach's model exaggerated and the major role that probabilities play in it underestimated. The paper also shows that Bacharach's account is neither necessary nor sufficient for Tuomela's; that contrary to HMT's claims, 1/ it is doubtful whether we-reasoning is more rewarding or creates more collective order than individual reasoning, and 2/ the two accounts do not make the same action recommendations. They are distinct theories which share a basic common we-reasoning structure and refine it in idiosyncratic ways.

Günther Knoblich (Radboud University Nijmegen & Central European University, Budapest)
What is 'rational' in joint action?

In a review of recent empirical research on joint action I will ask in which sense joint action may be considered as 'rational'. I will address low level interpersonal processes such as temporal entrainment and mimicry, basic planning processes for performing coordinated joint actions in real time, and communicative signals in the service of joint actions and decisions.
Elisabeth Pacherie (Institut Jean Nicod, Paris)
Framing Joint Action
Many philosophers have offered accounts of shared actions aimed at capturing what makes joint actions intentionally joint, what makes it rational for agents to engage in joint actions, and how the intentions of co-agents should be coordinated to yield joint actions. I first discuss two leading accounts of shared intentions, proposed by Michael Bratman and Margaret Gilbert. I argue that Gilbert's account imposes more normativity on shared intentions than is strictly needed and that Bratman's account requires too much cognitive sophistication on the part of agents. I then turn to the team-agency theory developed by economists that I see as offering an alternative route to shared intention. I concentrate on Michael Bacharach's version of team-agency theory, according to which shared agency is a matter of team-reasoning, team-reasoning depends on group identification and group identification is the result of processes of self-framing. I argue that this approach can yield an account of shared intention that is less normatively loaded and less cognitively demanding. I further argue, however, that sophisticated mindreading and/or social normativity may be needed to sustain cooperation in situations that do not by themselves reliably induce group identification.

Clare M. Eddy, Ian J. Mitchell, Sarah R. Beck, Andrea E. Cavanna & Hugh E. Rickards (University of Birmingham)
Something is Not Always Better Than Nothing: The 
Ultimatum Game in Tourette Syndrome and Huntington’s Disease

In Tourette syndrome (TS) and Huntington’s disease (HD) striatal dysfunction can affect the functioning of the frontal cortex, through alterations within frontostriatal pathways. Patients can show impairments in executive function and Theory of Mind (ToM): the ability to reason about mental states. We investigated patients’ performance on the Ultimatum Game (UG), an economic decision making task thought to involve social reasoning.
In two experiments, patients with TS and patients with HD completed a version of the Ultimatum Game, whereby they had to choose to accept or reject offers of money which reflected an unfair split between themselves and a proposer. A rejection meant that the participant and proposer received nothing. Participants also completed ToM tasks and executive measures. 
Both patient groups rejected significantly more UG offers than controls. Some patients with TS and HD rejected higher offers that were always accepted by controls. Patients showed evidence of executive dysfunction and performed more poorly than controls on ToM tasks, but their deficits were not correlated with UG rejections.
Increased UG rejections in TS and HD could be related to behavioural symptoms involving poor emotion control and problems with social interaction. However, patients may reject offers for different reasons. Explanations for rejections framing the proposer’s behaviour in a larger social context in TS could indicate ‘altruistic punishment’ rejections. 
These findings implicate a role for the striatum in decision making with social contexts, and suggest that striatal dysfunction could impair decision making through poor emotion regulation. 

Sacha Bourgeois-Gironde (Aix-Marseille University, Ceperc, Institut Jean-Nicod) & Hadrien Orvoën (EHESS, CAMS, Institut Jean-Nicod)
How emotional social cues vs. economic factors affect trust

In a recent paper (Bourgeois-Gironde & Corcos 2011) we proposed a method to discriminate between the influence of strategic reciprocation and personal trust acquired through exchanges in the trust-game. We further showed that acquired trust crowds out background prosocial dispositions upon entering the trust game, and that the variance of monetary returns may be a have a stronger impact on trust than their level. This underlines the importance of a dynamic treatment of social cues (of different types) both by contrast with alleged static levels of pure altruism (Fehr et al. 2002) and with a purely calculative processing of monetary returns through repeated interactions in the game. The interaction of moral social cues with the investor’s behavior in the trust game has been investigated (Delgado et al. 2005). The main result is that “good” moral profiles of trustees freeze reciprocating adaptation by the investors when it happened that these trustees are not cooperative. The main limitation of that study is that moral profiles are static while we would like to observe dynamic interactions of social and economic cues in the trust game. In order to do so we have adapted Sanfey's et al. (2010) so that each 53 subjects that participated in our experience received after each round a monetary return by the trustee and a social emotional feedback in terms of a smiling, neutral or frowning face. We studied how these two parameters affected the dynamics of the game, especially in cases monetary and social returns by the trustees were incongruent. We observed a main effect of partners' cooperativeness on partner-averaged investments, but also a significant influence of congruence in the sense that incongruent partners are less easily trusted than congruent ones. The effect of congruence does not vanish throughout the game. In our experiment players thus seem to be sensitive not to social emotional or economic cues per se but to the dynamic coherence of both types of cues. Our results then suggest a distinct account of how social emotional and purely monetary cues interact in experimental cooperative games and prompt further investigations among clinical populations that would present particular dysfunction in integrating social and emotional cues in strategic behavior.

Mattia Gallotti (University of Exeter) 
From conceptual to biological irreducibility: collective intentionality naturalized 

Over the last twenty years the theory of collective intentionality has contributed important insights into many forms of human sociality, like cooperation and communication. The root idea of collective intentionality is that whenever people intend to do something together, they see their actions as being directed at the same goal. And for a goal to be shared, people must be able to think of individuals, both themselves and the others, as a group. 
In this paper I shall address the long-standing issue of the naturalization of collective intentionality as a question of experimental philosophy. The tacit premise is that it is not sufficient that philosophers embrace some form of self-proclaimed naturalism in principle, if the goal is to let philosophical investigation be informed by scientists’ research agenda and methods. Collective intentionality theory is particularly interesting in this respect: it grew out of philosophers’ intuitions about the role of group thinking in grounding phenomena of sociality, thus exemplifying the approach of paradigmatic research programs in the philosophy of social science that address philosophical questions on aprioristic grounds. Yet, more recently, it has become a topic of investigation in field and laboratory settings where mind and brain scientists have provided us with a rich experimental literature on the proximal and ultimate bases of collective intentional behavior, thus challenging the suspicion that group thinking is not scientifically observable. 
I shall start by distinguishing between distinct interpretations of the central problem of collective intentionality - whether collective intentionality can be decomposed into the concepts that we already deploy in understanding individual intentional behaviour. I then outline Searle’s claim that collective intentionality might play a role in causing people to achieve the kind of ‘mutual understanding’ necessary for joint action. While this is clearly an empirical question, it has so far been tackled by means of ordinary language analysis, thought experiments and counterfactual thinking. I shall counter this approach with an empirically-informed argument that bridges the gap between the scientific and the philosophical literature, based on Michael Tomasello’s research in the development and evolution of proto-social forms of behaviour such as ‘joint attention’. 
I conclude by spelling out the implications of this theory in more depth. Based on Tomasello’s findings, it can be argued that collective intentionality is developmentally prior and causally necessary for individuals to comprehend the goal-directed structure of joint action. It also suggests that collective intentionality might be a dedicated mechanism along the lines suggested by Searle, and that philosophical and scientific insights can only be integrated in more complex ways than previously supposed.

Christophe Heintz (Central European University, Budapest)
A preference for fulfilling other's expectations lead to the cultural diversity of prosocial behaviour in experimental settings

Experiments in behavioural economics have shown that people participating in experimental games behave more or less altruistically depending on their culture: there is cross-cultural variation of prosocial behaviour in these experimental setting. The main explanation that has been put forward in the recent works of Fehr, Gintis, Henrich and others, appeals to variations in social preferences: if prosocial behaviour varies, it is because the underlying preferences vary. Against this view, I will defend the hypothesis that social preferences are stable across cultures and that other factors are responsible for the observed cultural variations. The underlying universal and stable prosocial preference at work in these experimental settings is a preference for fulfilling the expectations that others have towards oneself. I will argue that humans monitor others' expectations towards themselves and that they have a tendency to fulfil these expectations. Of course, the legitimacy of other's expectations and the cost of fulfilling them are most often evaluated before they are actually fulfilled. I will first provide evidence from the literature in experimental economics that can be interpreted in favour of this process: monitor, evaluate, then fulfil others' expectations towards oneself. Then I will show how it can account for the cross-cultural variations of prosocial behavior in experimental settings.

Francesca Giardini (Central European University, Budapest)
Reputation for cooperation and social control in artificial societies

In human societies, exchanging social information is fundamental for partner selection, cooperation, social control and coalition formation, but it also plays a role in social comparison and group cohesion, just to name some of its main functions. Reputation spreading is a social activity that plays a fundamental role in linking nodes and providing individuals with new knowledge about unknown peers. Using language, humans are able to manipulate social information in different ways, ranging from reporting completely false information, to modifying the source of information. Aim of this work is to explore how reputation enters individual choices about cooperation, exploring whether and to what extent the way in which information is presented may affect these choices. I will try to show the complex interplay between the micro-level of agents' motivations and the macro-level of collective behaviors by presenting some results from experimental studies within the framework of Agent-Based Social Simulation (ABSS). In this computational approach, social phenomena may emerge as a result of interactions among heterogeneous artificial agents endowed with internal representations of themselves, their peers and their environment.exmple d'évènement lié.

Thursday May 26th, 2011

Morning session

Chair: Elisabeth Pacherie (IJN)

9.30 – 10.45
Guillaume Dezecache (Institut Jean-Nicod, Laboratoire de Neurosciences Cognitives, Paris) & Julie Grezes (Laboratoire de Neurosciences Cognitives, Paris)
« I fear your fear of his fear »: an experiment of transitive emotional contagion

10.45 - 11.00
Coffee Break

11.00 -12.15
Natalie Sebanz (Radboud University Nijmegen & Central European University, Budapest)
Is there a ‘we-mode’ in joint action planning?

12.15 – 14.00 
Lunch Break

Afternoon session

Chair: Christopher Kutz (UC Berkeley)

14.00 – 15.15
Cédric Paternotte (University of Bristol)
Some more reasoning about we-reasoning

15.15 – 16.30
Günther Knoblich (Radboud University Nijmegen & Central European University, Budapest)
What is 'rational' in joint action?

16.30 – 16.45
Coffee Break

16.45 – 18.00
Elisabeth Pacherie (Institut Jean Nicod, Paris)
Framing Joint Action

Friday May 27th, 2011

Morning session

Chair: Paul Egré (IJN)

9.30 – 10.45
Clare M. Eddy , Ian J. Mitchell, Sarah R. Beck, Andrea E. Cavanna & Hugh E. Rickards (University of Birmingham)
Something is Not Always Better than Nothing: The Ultimatum Game in Tourette Syndrome and Huntington’s Disease

10.45 - 11.00
Coffee Break

11.00 -12.15
Sacha Bourgeois-Gironde (Aix-Marseille University, Ceperc, Institut Jean-Nicod) & Hadrien Orvoën (EHESS, CAMS, Institut Jean-Nicod)
How emotional social cues vs. economic factors affect trust

12.15 – 14.00 
Lunch Break

Afternoon session

Chair: Sacha Bourgeois IJN

14.00 – 15.15
Mattia Gallotti (University of Exeter) 
From conceptual to biological irreducibility: collective intentionality naturalized 

15.15 – 16.30
Christophe Heintz (Central European University, Budapest)
A preference for fulfilling other's expectations lead to the cultural diversity of prosocial behaviour in experimental settings

16.30 – 16.45
Coffee Break

16.45 – 18.00
Francesca Giardini (Central European University, Budapest)
Reputation for cooperation and social control in artificial societies


CNRS EHESS ENS ENS