WORKSHOP “From Minimal to Complex Collective Actions”

September, 4 2017 at SALA ENZO PACI, CSSA, Philosophy Department, University of Milan, Via Festa del Perdono, 7, 20122 Milano


9.00-10.00      Kirk Ludwig (Indiana University, Bloomington): Scale in Collective Action: From Meetings to Mergers

10.00-10.40    Michael Schmitz (University of Vienna): From Walking Together to the March on Washington: A Layered Account of Joint Action

10.40-11.10    Break

11.10-11.50    Helen Lauer (University of Dar Es Salaam): The anatomy of complexity in scientific collaboration driven by non-evidential criteria for consensus

11.50-12.30    Matti Heinonen (University of Helsinki): Mechanistic Explanation as an Interdisciplinary Endeavor: The Case of Shared Task Representations

12.30-13.10    Gunnar Schumann (University of Hagen): Why There Are No “We-Intentions” (And Not Even A Puzzle About Collective Intentions)

13.10-14.40    Lunch Break

14.40-15.40    Sara Rachel Chant (Tulane University): From Minimal to Complex Collective Actions

15.40-16.20    Guglielmo Feis (University of Milan) and Alice Borghi (University of Milan): Groups and Group Agency as a Democratic Reply to Popularism

16.20-16.50    Break

16.50-17.30    Sandra Larsy (Institut Jean Nicod Paris): Shared Representations in Joint Actions: Developmental Issues

17.30-18.30    Stephen Butterfill (University of Warwick): Joint Action and the Emergence of Referential Communication


Kirk Ludwig (Indiana University): Scale in Collective Action: From Meetings to Mergers

Many accounts of joint intentional action focus on small-scale interactions in which the participants know each other and the task is one that they all work on at the same time and in proximity to one another.  A paradigm of this is Michael Bratman’s account of modest sociality in which mutual responsiveness plays a central role and requires all the agents to be aware of each other in order to respond in ways that are mutually supportive.   Margaret Gilbert’s account, which is grounded in joint commitment, likewise starts with the case of groups small enough for all members to express in their behavior to each other their readiness to jointly commitment to their doing something, and even extensions to non-basic cases presuppose a prior arrangement in which everyone expresses their readiness to everyone else. Raimo Tuomela’s basic account requires mutual knowledge of the preconditions for success and so appears to place significant constraints on the size of the group.   But while small-scale joint action is a pervasive feature of social life, we live in a civilization whose basic structures are dominated by large-scale collective and institutional action.  How do we bridge the gap between, on the one hand, accounts of small-scale joint action, like committee meetings, where all the participants are aware of all the others, and, on the other, multi-generational projects like building the Great Wall of China, and mergers where corporations themselves appear to be the agents, and most of their operators and shareholders never meet and are ignorant of much that the corporations do?  In this talk, I sketch a general account of plural action and joint intention that scales up from two people shaking hands or meeting in the park, to generations-long construction projects like the Great Wall of China, and apparently irreducible organizational actions like corporate mergers.

Michael Schmitz (University of Vienna) From walking together to the march on Washington: a layered account of joint action

I will explore continuities and discontinuities between minimal and complex joint actions as well as their relation from the point of view of a theory of the layered intentionality of groups. On all levels of collective intentionality subjects act in certain modes in which they experience and represent each other as co-Subjects of actional and perceptual, practical and theoretical positions towards the world. The levels differ in terms of the representational format of the relevant representational states, acts and artifacts. For purposes of orientation, I distinguish three broad levels/modes. The mode of joint action and attention has a nonconceptual representational format, the level of We-mode beliefs, intentions and group speech acts a conceptual and propositional one. The level of institutional reality can be characterized in terms of the essential role of writing and other forms of documentation and the fact that individuals and groups take positions in terms of their roles in institutional settings. For example, an individual might give an order or propose a plan for action in their role as prime minister of a country. A more precise characterization of these layers and their relation is proposed in terms of such parameters as their degree of context dependence, the degree of the externalization and of the durability of the relevant representations and their degree of abstraction, the degree of differentiation of representational role, the size of the relevant collectives and more.

Helen Lauer (University of Dar Es Salaam): The anatomy of complexity in scientific collaboration driven by non-evidential criteria for consensus

Deference to relevant expertise has become a central tenet within complex scientific collectives that straddle more than one discipline and respond to conflicting criteria of success. But with increasing epistemic dependence (Hardwig 1985), there emerges a prevailing climate of complicity in censorship of well-informed, evidence-based dissent from mainstream orthodoxies (Bauer 2017). Paradoxically, carefully orchestrated suppression of critical exchange appears to be affected by collective adherence to the very regulations and protocols designed to ensure that scientific group endeavours remain open to the heuristic rigours of conjecture and refutation among rival perspectives. The point of this analysis is to demystify and account for this paradox. First I will discuss how salient models of “structured social groups” (Bratman 1993: 98) fail to capture the conditions constituting non-evidence based scientific consensus by oversimplifying: either the focus of “group commitment” (Tollefsen 2002), or the content of members’ attitudes towards the norms that “create and maintain those institutions” (Tuomela 2003), or the nature of the “plural” subjectivity constituting professional knowledge-producing networks (Gilbert 2014), or the essential role of group members’ awareness of each other’s “sub-plans” and intentions (Bratman 1993: 103, 109; 1999; Chant and Ernst 2008; Kutz 2000: 20). Two innovations will help illuminate the professional dynamics that fortify the increasing tolerance for disinformation and the growing resilience of partial knowledge endorsed by scientific consensus in high profile collectives. (i) I replace the restrictive trait of ‘rule following’ as a central characteristic of rational collaboration, with Conte’s (2012) heuristic notion of “nomotropic” behaviour – whereby an intentionally collaborative norm may result from awareness of a regulatory principle without conforming to it; and (ii) I explore the non-epistemic values of espousing beliefs accredited not for their factive content but rather objectually (Brogaard 2016)—i.e. not in virtue of the available evidence, but because the agent thereby intends to convey a stance, or to exhibit allegiance to a professional group.

Gunnar Schumann (University of Hagen): Why there are no “We-Intentions” (and not even a puzzle about collective intentions)

Some philosophers hold that the minimal criterion for there to be a collective action is that the actions of the individual agents are caused by so-called “we-intentions”. But, as I want to argue (from a Late-Wittgensteinian point of view), intentions cannot be thought of as mental entities of any kind that can be in some sense possessed by individuals and which causes an action. The point of the concept of intention is that we would not ascribe an intention to an agent when she does not perform the relevant action – although she has the opportunity to do. An intention must therefore not be thought of as an entity, but as a way of behavior that can be understood in a context as being bound by a self-prescription by an agent. Accordingly, a collective intention is the behavior of the individual members of the group that can be understood in a context as being governed by one self-prescription. When two people intend to X together, then they have one and the same intention, not two very similar, but numerically distinct, mental states. It is the context of the performances of the actions that ensures that the partial actions are parts of the performance of one collectively intended action. (The context of an item of individual or collective behavior is the conceptual criterion for there to be a collective action.) To that context must belong that the shared intention was expressed either verbally (by agreement, promise or contract) or non-verbally (by the way the joint action or attempts to it were performed). As I will argue further, this also holds for more complex collective actions of larger groups of agents, so that there is no difference between basic and complex collective actions in principle.

Matti Heinonen (University of Helsinki): Mechanistic Explanation as an interdisciplinary endeavor – the case of shared task representations

Joint action has been the topic of intensive study in the behavioral and cognitive sciences during recent years. Researchers in developmental psychology and evolutionary anthropology have sought to understand the ontogenetic and phylogenetic development of our capacity for joint action, and the disposition to act together has even been recruited as a key explanatory variable for the emergence of natural language and cumulative cultural evolution. Researchers in cognitive psychology and neuroscience have been concerned with disentangling the cognitive mechanisms underlying our phylogenetically distinctive capacity for joint action in order to determine whether joint action is produced by dedicated cognitive mechanisms, or whether joint action is the outcome of general-purpose mechanisms for action coordination and control. Much of this research has converged today on whether individuals are disposed to form shared task representations of their common task, and whether this hypothesis best accounts for much discussed interference effects in new experimental paradigms, such as the Joint Simon Task. I will survey this on-going experimental and theoretical research in cognitive science from the point of view of mechanistic philosophy of science so as to answer the following two questions: 1) Could shared task representations function as a mechanism for minimal forms of joint action? 2) Could minimal forms of joint action function as a mechanism for the more advanced forms of joint action that are studied by the behavioral scientists?

Sara Rachel Chant (Tulane University): From Minimal to Complex Collective Actions

Collective action theory models its questions and answers for collectives after those of individual action theory. Because the first problem for individual action theory concerns the conditions for something’s being an individual action, the first problem for collective action theory concerns the conditions for something’s being a collective action. Likewise, since standard answers to the first problem for individual action theory concern the mental states of an agent (e.g. their beliefs, desires, plans, and intentions), standard answers to the first problem for collective action theory concern the mental states of the group. Since this approach is straightforwardly applied from individual action theory, and individuals are saliently different from groups, providing a unified analysis across the wide range of actions taken by groups proves far more challenging than its individual counterpart. As a result, most theories of collective action focus on ‘small-scale’ actions (e.g. walking together or lifting a heavy table) when they discuss the minimal conditions for something’s being a collective action and hope that they can extend that minimalist account to cover the very wide array of complex actions of groups. In this talk, I suggest that the standard minimalist proposals are not minimal enough and that that mistake arises from the general approach taken in collective action theory. Once we recognize that mistake and consider truly minimal collective actions, we may begin to see the way toward a unified theory from minimal to complex collective actions.

Guglielmo Feis (University of Milan) and Alice Borghi (University of Milan): Groups and Group Agency as a Democratic Reply to Populism

The theory of group agency (List and Pettit 2011) provides a minimal set of functions to describe groups as rational decision makers. According to the organizational structure, List and Pettit distinguish between two kinds of group agents. Group agents could be functionally explicit, when they clearly use an aggregation function (e.g. majority voting), or functionally inexplicit, when they determine on case base which (aggregation) function works better. Given that functionally inexplicit groups embody feedbacks from group-level attitudes about propositions to individual attitudes, they could be reasonable. On the contrary, it is more difficult for a functionally explicit group to be reasonable, given its mechanical structure. We argue that a democratic way of taking decision requires a functionally inexplicit group structure and, given that, tends to be reasonable, not only rational. This characterization is relevant in order to reply to populist appeal to people, if we manage to characterize populist way of taking decision as requiring a functionally explicit structure, that is non-reasoning.

Sandra Larsy (Institut Jean Nicod): Shared Representations in Joint Action: Developmental Issues

Many accounts of joint action argue that to coordinate successfully, agents must be able to form shared representations of their environment, their goals and actions. Since this topic has only recently become a topic of inquiry in philosophy and psychology, the nature and the emergence of shared representations are still being discussed. A first debate, mainly in philosophy of action, concerns what must be shared to support joint action: intentions or goals; raising the question of whether robust theory of mind abilities (absent before four years of age) are required in joint action. A second debate, in developmental psychology, investigates the nature of the theory of mind abilities young children have and deploy when they engage in joint action. This body of research is notably interested in understanding the involvement of other cognitive processes such as joint attention, joint commitments, action prediction, inhibitory control or perspective-taking. After reviewing this research, I will introduce a typology of joint action that incorporates these conceptual distinctions and combines two dimensions: (i) the type of understanding of others' actions, goals and intentions needed to meet these demands; (ii) the forms of joint-ness the action demands which refers to the identical or complementary agents’ roles during joint action. This typology will distinguish three categories of joint action: the first two categories will not require any theory of mind abilities but rather respectively the ability to take the other’s perspective and to maintain or confront their own and the other’ perspectives.

Stephen Butterfill (University of Warwick): Joint Action and the Emergence of Referential Communication

Consider the conjecture that abilities to perform joint actions play a role in explaining the developmental emergence of referential, purposively communicative actions.  If this conjecture is correct, how should we characterise joint action?  Leading philosophical and developmental theories propose to do so by invoking abilities to coordinate planning. But a body of research indicates that abilities to coordinate plans have a protracted development, with even the simplest and earliest manifestations typically not occurring until years after a child’s first words. After reviewing this research, an alternative approach to characterising joint action involving expectations about collective goals will be introduced. But how could abilities to perform joint actions enable production and comprehension of referential, purposively communicative actions? By constructing a model, this talk aims to show that a surprisingly rich and flexible range of communicative actions can be characterised by appeal to simple forms of joint action, without any need for postulating either coordinated planning or communicative intentions.  The route from minimal to complex joint actions may therefore go via communication.