Review: Bratman, Michael. Shared Agency

Review: Bratman, Michael. Shared Agency

New York: Oxford University Press, 2014. Pp. 219. $105.00 (cloth), $31.95 (paper).

Shared Agency brings together and further refines ideas in the theory of collective agency and action that Michael Bratman has been developing over the past twenty years or so. In it, he presents his account of shared intention, and its role in small-scale collective action, in its most comprehensive and systematic form. The theory presented in Shared Agency is developed in impressive detail; the arguments are careful and rigorous; and the overall view is both illuminating and challenging. It is a major achievement and a must-read for anyone interested in issues of collective action and intentionality.

FROM INDIVIDUAL TO SHARED AGENCY

The target phenomenon Bratman sets out to theorize is the form of agency exhibited by relatively small, non-hierarchical groups of people when they intentionally do something together—for example, singing an impromptu duet, or painting a house together. The central project of the book is to offer a model of this type of “modest sociality” that relies only on the conceptual and normative resources provided by Bratman’s earlier, well-known theory of individual planning agency. Doing so, Bratman thinks, will provide a strong case for the continuity thesis—the claim that we do not need any new basic elements in our metaphysics or philosophy of mind, beyond what is already given in the structure of individual planning agency, in order to account for robust forms of collective agency. Nor do we need to look for any basic norms governing shared intentional activity that cannot be accounted for in terms of norms of individual practical rationality—norms of means-end coherence, consistency, and the like.

Bratman situates his view between two alternative interpretations of the relevant phenomena. On the one hand, he argues that genuinely shared agency is a richer phenomenon than can be captured by the idea of a strategic coordination of individuals’ actions in a context of common knowledge. On the other hand, he does not think the relevant forms of modest sociality are necessarily associated with reciprocal obligations or commitments the participants have to one another.

The first contrast is with strategic equilibrium theories. On these views, roughly, collective action occurs when it is common knowledge that there is a stable solution to the problem of coordinating separate parties’ strategies for achieving their goals, in light of what each expects the others to do. Bratman claims that this is not sufficient for the sort of shared intentional activity he wants to capture. We might imagine, for example, that you and I are walking next to each other, each of us adjusting our movements to the other’s so as to avoid a collision, both aware that we are doing this. Bratman argues, following Margaret Gilbert and others, that this does not entail that we are walking together.

If you and I are walking together in the relevant sense, what is missing from the description of our activity as the strategic coordination of our individual pursuits in a context of common knowledge? One answer, defended, for example, by Gilbert, is that what’s missing is an account of the entitlements and obligations we have toward one another in virtue of our joint endeavor. For example, we might think I have some obligation to you not to suddenly change course without warning or explanation. But Bratman rejects this characterization of the phenomena as well. He does allow that mutual obligations will frequently apply in cases of shared activity—triggered, for instance, by promises and assurances people often offer one another in these contexts. But he doubts they are necessarily present.

In Bratman’s view, what distinguishes a group’s doing something together from mere strategic coordination of individual actions is the group’s shared intention to engage in the joint activity as such. And on his account, shared intentions do not essentially engage any basic norms beyond those of individual planning agency and rationality. This is because a shared intention is just a complex of interdependent intentions and beliefs with distinctive contents. Not everyone will be convinced that the account does full justice to the distinctive interpersonal commitments and expectations characteristic of intentional joint activity. But even those who have doubts will appreciate that Bratman has done a considerable service in presenting such a thoroughly worked-out and challenging alternative.

The heart of Bratman’s theory is given in what he refers to as the basic thesis. The basic thesis in (if you can believe it) its compressed form is as follows:

First, a group of us shares an intention to perform a joint activity J if the following conditions A-D are met:

1. Intention condition: We each have intentions that we J; and we each intend that we J by way of each of our intentions that we J and by way of relevant mutual responsiveness in sub-plan and action, and so by way of sub-plans that mesh.
2. Belief condition: We each believe that if the intentions of each in favor of our J-ing persist, we will J by way of those intentions and relevant mutual responsiveness in sub-plan and action; and we each believe that there is interdependence in persistence of those intentions of each in favor of our J-ing.
3. Interdependence condition: There is interdependence in persistence of the intentions of each in favor of our J-ing.
4. Common knowledge condition: It is common knowledge that A-D (103).

And second, we actually engage in a shared intentional activity if, in addition, the following condition is met:

1. Mutual responsiveness condition: Our shared intention to J leads to our J-ing by way of public mutual responsiveness in sub-intention and action that tracks the end intended by each of the joint activity by way of the intentions of each in favor of that joint activity (103).

There is a complex set of ideas here, which Bratman carefully lays out and defends over the course of the first four chapters of the book. He goes on to provide insightful discussions of the relation between modest sociality and moral obligation, the prospects for thinking of groups as agents or subjects in their own right, and shared deliberation in the context of joint activities. But I will only be able focus here on some key elements of the basic thesis itself.

PERSISTENCE INTERDEPENDENCE AND THE SETTLE CONDITION

Perhaps the most distinctive and controversial element of Bratman’s thesis is his claim that in central cases of modest sociality, every participant intends the joint activity itself. And this requires that each not only intends to do his or her part, but intends also that the others do their parts (and indeed, that they all do their parts by way of everyone’s intentions that they do their parts). As David Velleman once pointed out, however, on the face of it this seems to be in tension with an important fact about the nature of intention, namely, that one can rationally form an intention on some matter only if it is up to one to settle what will be the case. This “settle condition,” as Bratman calls it, presents a problem since it implies that, on his account, it must be rational for more than one member of a group to regard it as up her whether the group as a whole does something.

Bratman’s account requires that, if you and I are to share an intention to J (at least assuming we both know we are rational) it must be both up to me and up to you to settle whether we will J. But how can I take it to be up to me whether we will, say, go to New York together—where this implies that it’s up to me not only whether I go, but whether you go as well—while at the same time regarding it as up to you whether you will go—and indeed, whether I myself will go?

Bratman’s solution to this puzzle is to appeal to a condition of interdependence in the persistence of our individual intentions that, e.g., we go to New York. The idea is that if we will each continue to intend that we go to NY just in case the other does, and if we will actually go so long as we both intend that we go, then we each do have the power to settle whether we will go by way of our intentions to do so. If I know that your intention depends in this way on mine, then in forming and maintaining my intention that we go to NY, I can then view myself as thereby settling the matter. And the same goes for you.

This is an ingenious response to the objection. And Bratman goes on to provide a very interesting discussion of different possible sources of the relevant persistence interdependence, and of the various forms it might take. Nevertheless, I have some doubts about whether the appeal to persistence interdependence will ultimately prove to be a satisfactory solution to the problem.

The appeal to persistence interdependence does not seem to entirely eliminate the tension between my regarding it as both up to me, and up to you, to determine whether or not we go to NY. Let’s assume I take it to be up to me whether we go. If I’m rational, I will intend for us to go only if I judge that there is sufficient reason for us to do this. But suppose that I nevertheless regard it as equally up to you whether we go. In order to think this, I must believe that I will intend that we go to NY as long as you do—that I won’t change my mind unless you change yours. This is just the requirement of persistence interdependence. But this suggests that I regard your assessment of the continuing force and relevance of the reasons supporting our trip to New York as trumping my own. And this seems incompatible with regarding the matter as being within my discretion to settle.

The difficulty is particularly acute in cases where our reasons for wanting to engage in the joint activity differ. You and I intend to play tennis. But whereas you just want to play with whomever is available—which happens to be me—my interest is in playing a match with you in particular. Here it’s not clear that I am even justified in assuming that you will persist in your intention that we play so long as I do. Imagine that some third person comes along who is willing to play. This would not affect the grounds of my commitment to our game, but it would affect yours. You will no longer need me in order to play. In this case, if it were up to me, we would play our game. But the possibility that you will give up your intention prevents me from taking my intention to settle whether we will play the match. Assuming we are both rational, I cannot take the persistence of my intention that we play to be sufficient for the persistence of yours, since whether or not you have reason to change your mind does not depend on whether I have reason to change mine.

It would be one thing if the implication here were merely that, in this sort of case, our shared intention will lack a certain stability. That would be plausible enough. But the implication of Bratman’s view appears to be that, if our reasons for engaging in the joint activity are different, then at least one of our intentions toward our joint activity will not meet the settle condition. And this means that (barring irrationality or error) shared intention in this sort of case is just not possible. This is both a counterintuitive result and also seems to be at odds with Bratman’s claim that his view is well suited to account for the possibility that participants in a joint activity may have different reasons for engaging in it.

THE SCOPE OF THE BASIC THESIS

The requirement that the parties involved in a shared activity have common knowledge of the persistence interdependence of their intentions is limiting in another way. For it seems to significantly restrict the range of cases the basic thesis applies to. One can think of examples of small-scale, shared intentional activity that appear not to meet the conditions of the basic thesis.

Consider, for example, an informal reading group with several members who gather once a month for discussion. It’s easy to imagine a case where the contents of the participants’ intentions refer to the group as such, and not to specific members individually, and where no one participant’s intention to meet with the others is necessary or sufficient for the group to get together. If a member decides not to participate one month, the discussion will take place without him. Thus, none of the individual members is in a position to determine whether the group does or does not meet on a given day. Yet it would still be natural to say that the group intends to discuss a certain book on a certain day of the month. And if their plan leads in the right way to their meeting on that day, it seems we have a familiar type of shared intentional activity. But the group’s intention and the resulting shared activity apparently demand a different sort of account than the one Bratman offers.

Now, this is not, in itself, an objection to Bratman’s basic thesis. His official aim is to provide sufficient conditions for a central form of modest sociality; he does not argue against the possibility that there might be other forms that require separate analysis. Still, there are two relevant points to be made about this.

The first is that the limited range of the basic thesis potentially reduces the interest of Bratman’s continuity thesis, mentioned above: the idea, as Bratman puts at one point, that “once God created individual planning agents and placed those agents in a world in which they have relevant knowledge of each other’s minds, nothing fundamentally new… needs to be added for there to be modest sociality” (8). Perhaps Bratman has captured a form of shared agency that is continuous with individual agency, and does not require the introduction of fundamentally new conceptual, metaphysical, or normative elements. But since there are evidently other forms of modest sociality (not to mention larger-scale forms of collective agency), Bratman’s model does not in itself do much to justify our faith that, after creating individual agents, God rested.

The second, related, point is that the fact that there appear to be forms of modest sociality not analyzable in terms of the basic thesis would seem to undercut Bratman’s occasional appeals to Ockham’s Razor in arguing against non-reductive views like Gilbert’s and Searle’s. Bratman argues that, since we do not need to posit new primitive elements to account for the transition from individual to social agency, we should not incorporate these elements into our theories. But it remains to be seen whether we will ultimately need to appeal to such things as joint commitments (Gilbert) or irreducible we-intentions (Searle) in order to account for the full range of phenomena. And if such an appeal promises to capture a wider class of cases, then it’s less clear that the reducibility of Bratmanian shared intentions to individual intentions is a point in his favor.

Ultimately, however, whatever the prospects for extending Bratman’s theory beyond the central cases he focuses on, there is no doubt that Shared Agency is an enormous contribution to our understanding of social interaction. The view it presents is powerful and illuminating and will serve as a touchstone for future work in this area.

Stephen J. White

Northwestern University

1