The Agency-Based Collectivist Argument from the Discursive Dilemma

The agency-based argument from the discursive dilemma is typically combined with function­alism (or sometimes with the position known as interpretationism).³¹ According to functionalists, any system can be classified as an agent provided (i) it generally forms consistent and evidence­based belief-like cognitive states about its environment, (ii) generally forms not incompatible or

impossible desire-like conative states, and (iii) is generally capable of acting on these cognitive and conative states.

In general, if collectives can be autonomous agents, then it seems intuitively more plausible to eliminate potential “deficits in the accounting books” by ascribing responsibility to them. Agents can act wrongly and do bad things. If a collective as a whole is an agent in its own right, then it, sooner or later, will also end up doing wrong, and if so, then it will be responsible for doing so. The intuitive idea is clear enough, but of course the crux of the matter is who is responsible for how, morally speaking, the actions of collective agents turn out to be.³³

We have already seen in section 21.5.3 that group rationality and individual rationality can come apart, perhaps because the premise-based procedure is in place. When this happens, individual members will embrace a view which as individuals they reject.

the group will count as a reasoning subject in the image of the reasoning subjects that we individuals constitute. It will exercise a sort of control over its own processes of judgment-formation that resembles the personal control associated with individual reasoning. The members will act together in implementing an intentional exercise of group control. And they will do this in respect of themselves as a unified center of attitude formation and enactment.

(Pettit 2007b: 513; see also List & Pettit 2011: 104; Rovane 2004)

A considerable advantage of this approach is that it seems to be able to avoid the specter of metaphysical dualism, and at the same time preserve the autonomy of group rationality.³⁴ This is an advantage because all responsibility collectivists, and certainly those discussed in this chapter,³⁵ embrace ontological individualism (Zahle 2007; Epstein 2009). Ontological individu­alism is the view that group-level properties supervene on properties of individuals (List & Pettit 2011: 89). If so, then group rationality too, while autonomous, must supervene on individual reasoning and deliberation, and the actions of the collective agent must be carried out by indi­viduals. But as we have said, this stricture of ontological individualism is fulfilled because even when individuals are guided by group reasons it is still individuals who act, i.e., on this collect­ivist account it remains true that, strictly speaking, it is always the manager(s) who sign(s) the contract or the soldier(s) who pull(s) the trigger, and so on. On the other hand, group rationality remains autonomous too — supervenient on but irreducible to individual rationality — because when these individuals undertake various actions, their actions are guided by the interests and purposes of the group.

The individualist will insist that this account still does not establish the responsibility of col­lective agents, which are said to emerge through the collectivization of reason. If we grant (as responsibility-collectivists who are ontological individualists also do) that it is indeed the job of individual agents to monitor the epistemic states of the collectives, then there seems to be no ground for attributing (non-distributive) responsibility to the collective — at least not in situ­ations which involve the discursive dilemma. It may of course frequently happen that individual members indeed “hear the call of the group” (Pettit 2007b: 516), and enable the constitution of the collective agent, for example, by not insisting on their individual beliefs allowing the group to endorse (p&q&r) and act accordingly (e.g., release the felon or recommend awarding tenure to Dr. Borderline). If the relevant individuals are so disposed, then (let us grant for the sake of the argument) a collective agent may have indeed come into existence. However, once the story is spelled out in a piecemeal way as we have just done here, it becomes clear that any morally objectionable action by the collective will be unequivocally traceable to the indi­vidual members’ actions. Notice that this objection remains forceful even if we set aside the individualist worries discussed in section 21.5 that collectives do not seem to be at fault in the relevant sense in situations involving the discursive dilemma. For example, the recom­mendation to award tenure to Dr. Borderline will be made because one or more individual members have chosen to allow the group to go with the premise-based procedure and endorse (p&q&r). Therefore, those individual members will be clearly identifiable as responsible for the decision who have allowed the group to take the decision that (p&q&r).

In sum, responsibility-collectivists are caught between a rock and a hard place because, on the one hand, ontological individualism and responsibility-collectivism do not seem to go together, but the metaphysics of ontological dualism is found so unappealing by most, on the other. On an ontologically plausible account of collective agency, failures by the collective will be clearly traceable to individual actions, and so even if the collective is an autonomous agent in some sense, individual members will be responsible for whatever is done by the collective. Conversely, if collectivists deny that failures of collectives are traceable to individual actions, then they will have no plausible, metaphysically unmysterious account to offer about how autonomous col­lective agency is supposed to emerge.

21.7