GROUP AGENTS AND COLLECTIVE INTENTIONS Raimo Tuomela University of Helsinki University of Munich
Individualism in Social Science (a) Conceptual understanding of an individual’s action must be based on either the individual’s own attitudes or reasons or some other agent’s (individual’s or reducible group agent’s) attitudes and reasons as its ground. ( Meaning ) (b) Explanation of an individual’s action must have either the individual’s own attitudes and reasons or some other agent’s (individual or reducible group agent’s) attitudes and reasons as its explanatory basis. ( Explanation ) (c) The basic ontology of the best explaining social scientific theory must consist solely of the activities and properties and interactions of either individuals or groups reducible to the individualistic basis referred to in (a) and (b). ( Ontology )
Why we-thinking? We-thinking in the full we-mode sense is in some cases conceptually necessary , e.g. in contexts where the we-mode constitutes (and thus constructs) full-blown group notions — collective artifacts — such as group beliefs or social institutions; functionally required in many contexts, especially in cases of joint action requiring synergy effects for collectively (and individually) beneficial results. theoretically sufficient for (dis)solving central collective action dilemmas (e.g. the PD) and thus for creating collective order . needed for group-based cultural evolution in “developed” cases. is capable of handling large groups better than the I-mode (primary dependence concerns the group and not individuals).
We-mode group A person functions in the we-mode if she functions as a group member (in a strong sense) and in the I-mode if she functions as a private person, possibly in a group context. A group is in most contexts below assumed to be a we- mode group ( a species of corporatio) . It is assumed to commit itself to a group ethos (certain constitutive goals, beliefs, standards, norms, etc.) and to relevant we- reasoning and we-acting. A we-mode group constructs itself as a group in a quasi-entifying sense and can be viewed as a functional group agent . Because of group membership the members of a we- mode group ought to act as group members and thus to “identify with the group”.
History 1: Group minds and group agents Group minds and agents have been discussed for thousands of years. The ancient Romans built into their law the idea of corporate responsibility, speaking of organized collectives that were also referred to by the terms “ universitas ”, “ corporatio ,” and “ collegium ”. A corporation ( corporatio ) contrasts with “ societas, ” a collective based on interaction of individuals who do not form an intentional group agent or a “group person” (where the term “person” indicates a theatrical mask). A corporatio is a persona capable of action and of making promises and fulfilling them (Hobbes).
History 2: Accounts of group minds and group agents Intrinsically intentional group agents: groups are intrinsically intentional analogously with the intentionality of individual agents. Extrinsically intentional group agents : group members form the group mind (collective attitudes, etc.) e.g. by their relevant kind of collective acceptance or some related group-internal process or mechanism.
Group Agents 1 An organized group (e.g. a we-mode group) can functionalistically be taken to be an agent (person) if it is goal-directed and if it can also be taken to reason and even reflect upon its activities. Its mental states are functionalistically construed as group-level states . In simple cases, a group agent is (possibly “emergently”) constructed on the basis of individuals’ properties and relations. In the case of corporations and states and other group agents with normatively characterized positions a more top-down kind of construction is typically used. We thus have a group-level description of collective intentions and beliefs, etc. and a member-level description of them.
Group Agents 2 A group agent exists as a functional social system capable of producing uniform action, not as an intentional agent with phenomenal features. It can only function via its members’ functioning appropriately. In many cases a group agent involves some fictitious and irreducible constructive elements and cannot be fully accounted by individualism.
Three central we-mode criteria The we-mode involves the following three central ideas on the member level (Tuomela, The Philosophy of Sociality , Oxford UP, 2007 and Tuomela (forthcoming)): ( authoritative) group reason , collectivity condition , and collective commitment . We-mode mental attitudes are had in the way or mode satisfying the above requirements for thinking and acting with the full we-perspective with a “togetherness - we”. People can e.g. carry a table either in the I-mode or in the we-mode. A content can in this sense be intended in various modes.
I thinking and we-thinking in the I-mode versus the we-mode 1 In general, there can be we-thinking and action, etc. in the we-mode and we-thinking or action in the I-mode (even for the benefit of the group’s goals and interests). On the other hand, there can be I-thinking and action in the we-mode (conceptually group-dependent thinking, e.g. I we-intend to participate in joint action) and in the I-mode (“private” I -thinking without conceptual dependence on the group).
I thinking and we-thinking in the I-mode versus the we-mode 2 The above classification can be presented concisely in terms of symbols as follows, where e.g. J = carrying a table jointly: (1) IM(I, PJ) (2) WM(I, PJ) (3) IM (W, J) (4) WM(W, J) Here IM and WM are operators covering the contents within the parentheses. I and W respectively mean I-intention and we- intention where the mode of the we-intention determines in which way or manner the content is held and in the case of WM the “strength” of the “we” involved in it. J means joint action as the content of a we- intention and PJ an individual member’s part action in the members’ joint performance of J.
An argument for the central we-mode criteria The idea of a group agent capable of acting as a group can in part be based on an intuitive analogy of intentional action (as action for a reason) in both the individual and the group case. Analogously to typical intentional action by an individual agent, intentional action by a group agent (and its parts, the members) is normally based on reasons for actions. Analogously to an individual’s having to coordinate the movements of her body parts, the members of a (we-mode) group coordinate their activities (including mental ones) in order to achieve group goals. Analogously to an individual agent committed to her intended actions, the group members are committed as a group, i.e. collectively committed, to the group's actions.
We-mode versus I-mode: functional differences One of Bacharach’s (1999) results can be applied to show that the (pro-group) I-mode and the we-mode , probabilistically construed concerning mode adoption, do not entail the same equilibrium behaviors. This holds also for cases (even) where the choices, utilities and the probabilities of the players acting for their own benefit instead of the group’s benefit are the same: The pro -group I- mode admits Pareto-suboptimal equilibria (e.g. DD in Hi-Lo) that in many cases will not be equilibria in the we-mode case. The above applies especially to common interest (Paretian) game situations with strong interdependence (such as, the Hi-Lo game and the PD). In Hi-Lo a full- blown group “framing” obviously makes joint outcome Hi-Hi (rather than Lo- Lo) a rational group’s choice.
II COLLECTIVE INTENTIONS
Joint intentions, we-intentions, and group agent’s intentions Collective intentions in this talk include joint intentions , we-intentions , and group agents’ intentions . One can act jointly either on an I-mode joint intention or, in a conceptually and functionally stronger sense, on a we-mode joint intention. We-intentions are components of joint intentions. In addition also intentions attributed to groups qua collective intentions will be discussed. All these three kinds of collective intentions are closely interconnected. In some agents’ we -mode joint action the agents form a we-mode group even in fleeting cases.
I-mode intention (IMI) Agent A has the intention that P in the I-mode if and only if A is privately committed to satisfying P (or participating in satisfying P) and he intends to do it at least in part for himself qua private person (rather than qua group member). (PLIMI) Agent A has the intention that P in the plain I-mode if and only if A is privately committed to satisfying P (or participating in satisfying P) and he intends to do it only for himself qua private person.
Pro-group I-mode intention (PROGIMI) Agent A has the intention that P in the pro-group I-mode in group g if and only if A is functioning qua member of g (in a weak sense), A is privately committed to participating in satisfying P and he intends to do it at least in part for the members of group g but in part for himself qua private person.
Recommend
More recommend
