Introduction
The notion of collective intentionality – which refers to instances in which a plurality of actors holds an intentional state together, such as when they jointly believe something, intend to act, or pursue a shared goal – underpins the ways in which many scholars try to make sense of the world of international politics.Footnote 1 Most conspicuously, international relations (IR) scholars routinely treat states as intentional collectives or even agents. They portray states as ‘intentional – purposive or goal-directed – systems’ made up of numerous individual actors and invoke states’ collective intentionality to justify featuring them as key actors in causal explanations.Footnote 2 Others claim that groups of states sometimes form and act upon collective intentions in the course of pursuing shared goals together. For instance, explicitly ‘drawing on the concept of collective intentionality,’ Jennifer Mitzen argues that one can think about global governance arrangements as ‘states concerting their power for common public interests,’ noting that ‘[w]hen states commit to address particular problems together they can constitute a public power beyond the state with the capacity to steer international political outcomes.’Footnote 3 Similarly, Toni Erskine posits that states can form intentional collectives – such as coalitions of the willing – to engage in what she labels ‘joint purposive action,’ an involved, sophisticated mode of collaboration.Footnote 4 In what they term ‘“thick” cooperation,’ Moritz Graefrath and Marcel Jahn ‘identify a distinct sense of [interstate] cooperation that does not merely refer to actors … coordinating their policies while pursuing their individual goals, but instead involves actors who jointly work toward the attainment of a shared goal.’Footnote 5 And finally, Jelena Cupać makes the case that some international organizations, including the United Nations Security Council, can be understood as involving collective intentionality.Footnote 6
Over time, however, questions have arisen about the ontological underpinnings of this practice. On the one hand, many of the scholars who invoke collective intentionality in their empirical and normative analyses of international politics intuit a reductionist understanding of the phenomenon, that is, they suggest collective intentions are reducible to the intentional states of the relevant individual actors. In stark contrast, those who have studied collective intentionality and its application in IR have largely rejected reductionism as theoretically unconvincing and instead embraced various alternative, explicitly non-reductionist, ontological frameworks.
In this piece, we argue that this rejection rested on a foundational misreading of what reductionism entails, that a sophisticated reductionist account avoiding prior theoretical charges is possible, and that such an account promises to significantly advance the theoretical study of IR. To make this case, we first establish that the IR literature on collective intentions has based its rejection of reductionism on what amounts to a caricature account that bears little resemblance to the sophisticated accounts philosophers have developed in their works. The central issue is that, contrary to IR scholars’ common assertion, reductionist accounts need not be summative, that is, treat collective intentions as mere aggregates of individual ones. In fact, what arguably constitutes the most influential reductionist view – the philosopher Michael Bratman’s account of shared agency – is decidedly non-summative: while retaining the idea that collective intentions are ultimately reducible to individual intentions at its core, Bratman argues that it is only when these individual intentions have the right social contents, are appropriately interconnected, and are accompanied by dispositions to respond to others’ intentions and actions that they give rise to collective intentionality.
Based on Bratman’s philosophical groundwork, we develop reductionist accounts of the two main phenomena involving collective intentionality in IR: state intentions and joint interstate action. We argue that the members of a state’s foreign policy executive – the small group of individuals that together debate and decide upon what course of action to adopt – constitute the ultimate wellspring of a state’s intentions. In cases where states come together and jointly pursue a shared goal, these state intentions, then, become the base for the collective intention that underlies such sophisticated forms of cooperation.
This first sophisticated reductionist view explicitly developed for the purposes of IR enables innovative theorizing on both empirical and normative issues in international politics. To some, debates about how to conceive of the ontological foundations of collective intentionality might appear pointless. After all, there is no meaningful way to identify what one might call the ‘true’ ontological account of collective intentions. In Nuno Monteiro and Keven Ruby’s words, ‘there is no Archimedean point capable of settling, once and for all, the debate among foundational positions’ on the matter.Footnote 7 What, then, is the value added of developing yet another alternative ontological account?
To put it simply, we are driven by the belief that greater ontological pluralism generally advances international theory. New ontological frameworks that are based on different theoretical priors than existing approaches might allow scholars to productively employ concepts that previously appeared incompatible with their worldviews. To this day, the concept of collective intentionality has largely featured in social constructivist research, in part because the commonly presumed ontology underlying the phenomenon is incompatible with the foundational premises of alternative theoretical frameworks, such as political realism. Our reductionist view, in contrast, conceives of the phenomenon in terms compatible with such theoretical traditions, thus paving the way for new and exciting theoretical insights that a debate located almost exclusively within one school of thought would be unlikely to generate.
To demonstrate that our reductionist account of collective intentionality is well-suited to enable precisely this kind of theoretical advance, we discuss several ways in which it challenges existing views on both the empirical and normative dimensions of state intentions and joint interstate action. For the empirical study of international politics, our account’s focus on states’ foreign policy executives and the particular character of their deliberations, voting, and acting entails concrete guidelines for how to study the origins of foreign policies and grand strategies. In addition, by entailing a novel theoretical perspective on institutions such as coalitions of the willing or security concerts – one that acknowledges them as involving a form of collaboration that goes over and above mere policy coordination, but nevertheless considers them inherently fragile – reductionism opens productive avenues for future research on the origins and character of various types of interstate cooperation and global governance. Finally, reductionism also aids the normative study of IR. In particular, it is well-positioned to serve as a framework for assessing collective moral responsibility and its relation to individual responsibility, which is an issue of pressing concern for scholars and the general public alike.
The remainder of this piece proceeds as follows. We begin by reviewing the literature on collective intentions in IR, distinguishing the concept from others related to it, and tracing how the field came to reject reductionism. Thereafter, we outline Bratman’s account of shared agency and, based on it, develop our reductionist accounts of state intentions and joint interstate action. The subsequent section outlines the various contributions reductionism can make to empirical and normative theorizing in IR. Finally, the conclusion recapitulates our main findings and discusses both the promise and limits of a reductionist view for IR.
The state of the art: collective intentions in IR
This section first clarifies the article’s concept of interest – collective intentionality – by contrasting it with the related notions of collective agency and personhood. It then surveys the IR literature on the ontology of collective intentions to show that while scholars have developed a series of competing accounts, to date, the literature has not proposed a sophisticated reductionist view. Specifically, we demonstrate that although reductionist intuitions have long permeated large parts of IR, the long-standing belief that reductionism necessarily entails a summative view of collective intentionality has led the field to largely reject it as a possible contender.
Conceptual preliminaries
IR debates on the ontological status of states and groups of states often refer to three notions: collective intentionality, collective agency, and collective personhood. While the three frequently appear together, they are conceptually distinct, and distinguishing them matters because scholars typically use them to make claims of different strength about a collective.
Collective intentionality is the least demanding concept of the three. It obtains when a plurality of actors holds an intentional state together – for example, when they jointly believe something, intend to act, or pursue a shared goal.Footnote 8 Collective intentionality thus marks a shift from the ‘I-mode’ of separate individual intentionality to a ‘we-mode’: the relevant intentional states are attributable to the group as a whole rather than consisting merely in a collection of separate individual attitudes.Footnote 9 Our main concern in this article is one central form of collective intentionality: joint action, where, among a collective such as a state or a group of states, there exist joint intentions to act together in pursuit of shared ends.
Collective agency is usually understood as a more demanding phenomenon.Footnote 10 A collective agent is not simply a plurality whose members share intentional states or act together, but a collective organized so that it can function as a sufficiently unified center of attitude and action.Footnote 11 There are several different ways to specify this requirement. For instance, Christian List and Philip Pettit suggest, roughly, that group agency requires a standing structure capable of generating rationally coherent group-level beliefs and desires, sustaining and revising them over time, and connecting them to mechanisms of collective action.Footnote 12 Alternatively, Thomas Gehring and Kevin Urbanski argue that corporate agency requires corporeality, action capability, and autonomy.Footnote 13 What these accounts share is the idea that agency requires a degree of organization, stability, and action-capacity that need not be present in ordinary cases of collective intentionality.Footnote 14
Finally, IR scholars sometimes invoke the concept of personhood, particularly in relation to the state, although there is little consensus on what exactly attributing personhood to states involves. While a few accounts treat personhood as analyzable in terms of collective intentionality, most hold that it implies more.Footnote 15 For example, on Alexander Wendt’s account, personhood involves collective intentionality plus further features associated with rational intentional systems. Other accounts are more demanding still, arguing that full personhood requires consciousness or even state-level phenomenal experience.Footnote 16 For present purposes, what matters is that collective personhood, like collective agency, is usually treated as more demanding than collective intentionality.
This conceptual separation is important. It locates our object of study at the least demanding end of the spectrum that, somewhat surprisingly, has received little theoretical attention. Likely because the contributions that first brought debates on the ontological status of collectives to the forefront of theoretical IR immediately concerned themselves with the more demanding notions of agency and personhood, the literature today still focuses much of its attention on them as well. In doing so, however, it passes over potentially productive conceptual space. Indeed, it appears prudent to begin at the most foundational level – that of collective intentionality – before engaging in debates on more sophisticated concepts. After all, errors and oversights at this lower conceptual level are likely to have important implications for downstream debates that build on it as a foundation.
Prior views on collective intentionality and the rejection of reductionism
The ontology of collectives is now the subject of a rich debate in IR. Reductionism, however, is not a part of it. Where it appears at all – before quickly being rejected as unconvincing – it takes the form of a crude summative view, on which collective intentions are mere aggregates of individual intentions.
Collective intentionality and the state. In their thinking about the question of whether states can have intentions or not, many scholars intuitively embrace some form of reductionism. Robert Gilpin offers the locus classicus, positing that ‘only individuals act, even though they may act on behalf of one of these collective social entities.’ To speak of states as agents is, for him and many others espousing this view, simply ‘a matter of convenience and economy’ – a way of avoiding the tedium of ‘listing the individual[s] … who in reality did the acting.’Footnote 17 Similar reductionist intuitions recur in the work of Peter Lomas, who insists that ‘it is never right to speak of states acting,’ since ‘[a]n act is a literal thing’ and ‘[o]nly people can carry one out.’Footnote 18 Michael Marks likewise observes that ‘states are not individuals’ and that writing ‘France’ rather than ‘French foreign policymakers’ is merely a matter of convenience.Footnote 19
These broadly reductionist intuitions, however, never found expression in a fully theorized reductionist account of state intentionality.Footnote 20 If anything, they imply a simple summative view, according to which a state’s intentions are simply the sum of same-content intentions held by the individuals who compose or represent it.Footnote 21 But this is less a theory of collective intentionality than a way of eliminating it: claims about what states intend or do become dispensable shorthand for claims about particular human beings.Footnote 22
On substantive grounds, such a summative-eliminativist view is problematic for two main reasons. The first is meta-theoretical. As Deborah Tollefsen notes, ‘a great deal of explanatory work in the social sciences involves the attribution of intentional notions to groups,’ a point that is especially salient in IR, where, as Christian List and Kai Spiekermann observe, ‘a common unit of explanation is the state.’Footnote 23 If state-level intentional attributions do genuine explanatory work, we should not eliminate the collective intentionality they invoke.Footnote 24 The second is that the elimination cannot be carried through in any case. The chief obstacle, pressed throughout the literature, is the problem of multiple realizability: the same claim about what a state did can remain true regardless of which individuals realized the relevant state action.Footnote 25 ‘France conducted airstrikes,’ for instance, may still be true even if different pilots had flown the mission.Footnote 26 A summative view is too thin to capture this. Lacking any structure beyond the aggregation of individual intentions, it cannot explain why the same kind of state action can be realized by different individuals.
Wendt’s landmark article ‘The State as Person in International Theory’ proved, in many respects, the death knell of reductionism in the IR-ontology debate by categorically rejecting reductionist views as untenable and arguing instead for the superiority of non-reductionism.Footnote 27 Drawing on the notions of supervenience and emergence, and on philosopher Margaret Gilbert’s non-reductionist account of collective intentionality, Wendt argued that states can possess collective intentions as higher-level social phenomena that are not reducible to the intentions of the individuals who make up the state. In working out this account, Wendt explicitly identified the alternative he sought to reject – reductionism – with a summative view of state intentions. Reductionism, in his telling, ‘takes a “summative” or “aggregative” view of groups.’ In further exchanges he doubled down on this point, labeling reductionism a ‘“heap” view of collective intentionality’ and insisting that ‘no heaping of individual intentions will ever produce’ a genuinely collective one – which, he concludes, makes ‘a reductionist account impossible.’Footnote 28 The real task, he claimed, was ‘to show that state action is anything more than the sum of … individual governmental actions.’Footnote 29
Wendt’s article thus entrenched a key assumption about reductionism on collective intention: that it is necessarily summative and, for that reason, theoretically untenable. Along these lines, Gehring and Urbanski, for example, describe the individualist-reductionist position as one on which scholars ‘refer to corporate agency merely as a metaphorical shorthand for the actions of involved individuals.’Footnote 30 Cupać similarly characterizes reductionists as holding that ‘only individual humans possess the capacity for action’ and that state personhood is therefore merely ‘a pragmatic tool’ without ‘real ontological presence.’ As in these cases, the field as a whole largely equates reductionism with a thin, summative picture, on which collective intentions amount to nothing more than ‘personal interests and preferences … simply aggregated’ – and rejects it on that basis.Footnote 31
To be clear, while non-reductionism has become increasingly popular in the debate on how to think of the state and its attributes, there exists a host of alternative approaches, none of which, however, seek to rescue reductionism from the exile Wendt inadvertently banished it to. For example, a prominent strand of constructivist and post-positivist scholarship – most notably, the practice- and narrative-oriented work of scholars such as Patrick Jackson, Iver Neumann, and Erik Ringmar – approaches the state from a different angle altogether. Rather than asking how state intentions relate to the intentions of individuals, these scholars emphasize the ways in which states are constituted through discourse, practice, and representation.Footnote 32
If any existing position comes close to a genuinely reductionist ontology, it is Fleming’s account of states as owners rather than agents of actions.Footnote 33 Fleming occupies a middle ground between the two poles canvassed so far. Against the eliminative view, he holds that states really can be said to act; against Wendt, he denies that this requires positing an irreducible group agent. When we say ‘North Korea conducted a nuclear test,’ he argues, we do not mean that North Korea is literally an acting agent; we mean that authorized individuals acted in North Korea’s name, so that the action is attributable to the state, which owns it and bears responsibility for it.Footnote 34 The account thus keeps the locus of action at the individual level while explaining how states can nonetheless be said to act.
Yet, even this approach leaves the central ontological question unresolved. Attribution is meant to explain how the actions of authorized individuals come to count as actions of the state, not how several such individuals, acting together in the state’s name, jointly form a collective intention in the first place. State officials do not merely perform separate actions that are subsequently attributed to the state; they deliberate, plan, and decide together. It is precisely this joint intentional structure that an ontology of collective intentionality must explain. Fleming’s account thus addresses how state action is attributed, but not how state intention is constituted.
Collective intentionality and collectives of states. A similar neglect of reductionism reappears in the contemporary IR literature on collective intentionality as applied to groups of states, including international organizations (IOs) such as the United Nations (UN) as well as initiatives like security consortia or coalitions of the willing among other forms of interstate cooperation.
Recent work on collective intentionality among groups of states has focused primarily on IOs. This literature begins from the observation that mainstream IR theory has long accorded IOs little ontological significance.Footnote 35 In neorealist and neoliberal-institutionalist accounts, IOs usually figure either as arenas in which states pursue their separate interests or as instruments for lowering the transaction costs of cooperation; in neither case are they treated as intenders in their own right.Footnote 36 Principal-agent theory, meanwhile, gives IOs a more active role by explaining how delegated agents may acquire autonomy through agency slack, but its focus remains on delegation and control rather than on joint intention formation.Footnote 37 Constructivist work goes further still. Michael Barnett and Martha Finnemore’s account of IOs as bureaucracies treats them as autonomous actors whose authority rests on rational-legal legitimacy and control of expertise, and who exercise power by making rules that classify and fix meanings in world politics.Footnote 38 Yet, even this literature locates IO agency in bureaucratic autonomy without explaining how that autonomy relates to the member-dependent structures through which IOs are constituted and authorized.Footnote 39
Working toward a more systematic understanding of how IO agency relates to IO members, Gehring and Urbanski seek to explain how even member-dominated IOs, such as the UN Security Council, can become actors in their own right.Footnote 40 In their view, IO agency arises when member states pool governance resources, authorize the organization to deploy them, and establish decision-making procedures through which organizational decisions are made. Although members thus play a crucial role in generating IO agency, in the end, the account remains non-reductionist. As they explain, ‘although IO agency arises entirely from activities of lower-level actors, it reflects an organizational quality that cannot be reduced to a particular constellation of constituent actors and their preferences.’Footnote 41
Cupać presses the non-reductionist argument further, and on more explicitly ontological grounds.Footnote 42 Like Gehring and Urbanski, she focuses on member-dominated IOs such as the UN Security Council and rejects any ‘purely summative’ picture of collective action.Footnote 43 Yet, whereas Gehring and Urbanski’s account rests on organizational features such as resource pooling, Cupać begins from the idea that IO agency is inseparable from the person-like traits – intentions, beliefs, and even emotions – that are routinely ascribed to IOs. To make sense of these attributions, Cupać draws on Gilbert’s account of plural subjects and adapts it to the case of IOs: IOs become unified subjects when their members ‘jointly commit to act as a single body.’Footnote 44 Once established, this joint commitment structures members’ intentions and actions from the top down, endowing IOs with agency and person-like traits that cannot be reduced to the intentions, preferences, or actions of individual member states.
The same Gilbertian picture underlies Mitzen’s account of security concerts. On her telling, states acting in concert form a plural subject bound by a joint commitment to pursue shared goals, such that their collective intentionality ‘cannot be reduced or disaggregated to [individual actors’] preferences and/or interactions.’Footnote 45
The most systematic and ontologically grounded treatments of collective intentionality in the groups-of-states literature have thus converged on the same non-reductionist, broadly Gilbertian ontology that also came to dominate the debate over the state. There exists, however, an important exception that suggests it might be possible to conceive of collective intentionality in such cases in a different way: Erskine’s work on the moral agency of collectives, and in particular her treatment of coalitions of the willing.
On most institutions, Erskine is not a reductionist: drawing on Peter French, she treats many collectives as genuine corporate moral agents, irreducible to their members.Footnote 46 Coalitions of the willing, however, are different. As ‘informal associations’ that ‘lack formal organizational structures and decision-making procedures,’ they do not, on her account, qualify as corporate agents. Crucially, for Erskine, this does not mean that no collective intentionality is involved in these cases. Even absent corporate agency, she argues, the members of such coalitions can engage in ‘joint purposive action’ and thereby exhibit a form of non-eliminable collective intentionality.Footnote 47
In making this move, Erskine sketches a theoretical structure that noticeably departs from the non-reductionist approaches summarized previously. ‘When it comes to this category of collectivity,’ she writes, ‘descriptions of actions and intentions … are reducible to their individual constituents,’ yet remain ‘different in important ways from accounts of individuals acting independently or in isolation.’ What makes the difference is that ‘joint action’ of this kind cannot be analyzed as the ‘mere summation of individual acts.’ It requires instead that ‘the members of a collectivity’ have ‘compatible interests,’ a ‘concomitant willingness to cooperate,’ and ‘the capacity to deliberate … in order to coordinate their actions.’Footnote 48
These remarks are striking because they intuit a ‘reductive account of shared activity and shared intention’ that is nonetheless non-summative.Footnote 49 Erskine, however, never provides the ontological microstructure needed to underwrite such a view, leaving the question of how it may be possible to reduce collective intentionality to individuals without it collapsing into mere summation unaddressed. In fact, she concedes that her ‘reductive account … requires further explanation and analysis.’Footnote 50 In a sense, she describes a vision without developing the theoretical resources needed to realize it. The account we develop below does exactly that.
A reductionist account of collective intentionality for IR
In what follows, we argue that in rejecting reductionism as a viable view on collective intentionality for the purposes of IR, the field has prematurely discarded what, in fact, constitutes a productive theoretical perspective. Drawing on the plan-theoretic account of philosopher Michael Bratman, we demonstrate that IR scholars have been wrong to equate reductionism with summative views on collective intentions. Based on Bratman’s groundwork, we then develop reductionist views on the two central issues involving collective intentionality in IR: state intentions and joint interstate action.
Michael Bratman on collective intentionality
Within contemporary social ontology, Bratman’s planning theory of shared agency offers one of the most influential and fully worked-out reductionist accounts of collective intentionality. It is designed, first and foremost, to explain relatively small-scale cases of acting together – what Bratman calls ‘shared intentional activity’ – that do not involve the hierarchical and role-differentiated structures characteristic of complex institutions.Footnote 51
The account builds on Bratman’s planning theory of individual agency, which treats intentions as plan states that organize action over time.Footnote 52 Bratman’s guiding thought is that the very planning capacities that make temporally extended individual agency possible can also explain basic forms of shared agency. The transition from individual to shared agency is significant, but it does not require ‘any new conceptual or metaphysical resources’ such as a primitive we-attitude, a group mind, or an irreducible plural subject.Footnote 53
What sets cases of joint action apart is that they involve intentions of a distinctive kind: ‘intentions of individuals whose contents appeal to the group activity.’Footnote 54 Such intentions are not simply intentions to do one’s part, nor intentions to act in a way that, together with the others’ actions, will produce some outcome. Rather, they have the form ‘I intend that we J.’ For instance, if you and I intend to paint a house together, then, in Bratman’s view, our shared intention involves my intention that we paint and your intention that we paint.Footnote 55 Each participant’s intention is thus directed at the joint activity itself.
Intentions of the form ‘I intend that we J,’ however, are not yet sufficient for joint intention and action. You and I also each need to intend that we J ‘in accordance with and because of meshing subplans.’Footnote 56 This is to say that our subplans for how to J must be co-realizable. Suppose, for instance, that ‘I intend that we paint [the house] red all over, and you intend that we paint it blue all over.’ These subplans do not mesh, and even if we still ‘end[ed] up painting the house together (some combination of red and blue),’ this would not constitute a joint action based on collective intentionality. If, however, ‘I intend that we paint the house with an inexpensive paint, and you intend that we paint it with a paint purchased at Cambridge Hardware,’ no such problem arises. After all, ‘I don’t care where we buy the paint, and you don’t care about the expense,’ so ‘we could proceed to paint the house with an inexpensive paint from Cambridge Hardware.’ In this case, our subplans mesh.Footnote 57
Besides meshing subplans, in Bratman’s analysis the intentional structure involved in joint action exhibits further features, among them interlockingness and reflexivity. Interlockingness means that each participant’s intention is directed not only at the group’s J-ing, but also at the group’s J-ing proceeding partly by way of the other participants’ relevant intentions. This rules out cases in which one actor intends an activity involving another while bypassing that other actor’s intentional agency. If I intend that we go to New York by forcing you into the car, I may intend an outcome involving both of us, but not one that proceeds through your intentional participation. Reflexivity, in turn, means that each participant’s intention is also directed at the activity’s proceeding partly by way of their own relevant intention.Footnote 58
Another central element of Bratman’s account is that the actors engaged in joint action exhibit mutual responsiveness of both intention and action. For instance, imagine you and I are singing a song together: ‘[o]ur intentions lead each of us to be appropriately responsive to the actions of the other: I listen closely to when and how you come in, and this helps guide my own singing; and you are similarly responsive.’Footnote 59 The intentional structure of joint action thus involves not only that we intend to J, but also that the joint activity proceeds ‘by way of [J]-tracking mutual responsiveness in sub-intention and action.’Footnote 60
Overall, what emerges from these conditions is a sophisticated reductionist account: shared intention is constructed out of the intentional states of individual planning agents. But only when those intentional states have the right social contents, are appropriately interconnected, and are accompanied by dispositions to respond to others’ intentions and actions do they suffice to give rise to collective intentionality.Footnote 61 This is why the social ontology literature consistently classifies Bratman’s account as reductionist (collective intentionality is fully constituted by facts about individual participants without requiring an irreducible group-level intentional state) but non-summative (it does not analyze group attitudes as the mere sum of individual attitudes with the same content but as a structured configuration of individual-level attitudes).Footnote 62
Bratman’s reductionism and prior theoretical challenges
In offering this sophisticated reductionist account, and specifically one that goes beyond a summative view, Bratman’s account avoids the problems that have driven prior skepticism toward reductionism among IR scholars. First, it does not collapse into eliminativism: rather than treating talk of collective intentions as mere shorthand for parallel individual intentions, it preserves collective intentionality as a distinctive phenomenon while explaining it in individualistic terms. In this vein, Bratman calls this approach an ‘augmented individualism’ that is ‘not dismissive or debunking’ of collective intentionality but provides ‘a theoretically rich structure of conceptual, metaphysical, and normative resources in terms of which we can more deeply understand these phenomena.’Footnote 63 Indeed, shared intentional activity can, on his view, render an internally structured group a ‘single agent’ in a weak sense: not as a ‘novel center of intentional attitude,’ but as a locus ‘causally responsible for … action.’Footnote 64 This is also why Bratman’s account does not deprive collective intentionality of its explanatory significance. It identifies a genuine site of collective action that can figure in causal explanations without positing an irreducible group subject.Footnote 65
Second, Bratman’s account effectively answers the challenge of multiple realizability. The worry, as noted above, is that joint action cannot be identified with any fixed set of individuals, since the same kind of joint action can be carried out by different people. Bratman’s reductionism sidesteps this issue because it does not make the identities of the participants part of the conditions for joint action. What it requires instead is that their planning attitudes exhibit the right structure – reflexive, interlocking intentions that track meshing subplans and are sustained by mutual responsiveness. Because different individuals can instantiate this structure, the same kind of joint action can be realized by different configurations of participants.Footnote 66
Third, Bratman’s account also addresses another worry some IR scholars have pressed against reductionism, namely, that it is overly demanding by requiring every member of a group to ‘personally share its intention.’Footnote 67 This objection needs to be handled carefully. In the basic model of shared intention, the members do indeed share an intention to act together. But this does not mean that they must also share the same evaluative judgments about the merits or desirability of the joint action. A group of decision-makers may settle on a policy that some of them regard as unwise or even mistaken; what matters for shared intention is not evaluative agreement, but the relevant structure of interrelated intentions.Footnote 68
Bratman’s later work strengthens this response by addressing the harder case in which some participants do not share the collective intention at all.Footnote 69 To do so, he supplements the basic plan-theoretic framework with a distinction between group intentions and shared intentions. Groups, in this view, can have intentions ‘that are not themselves shared intentions,’ that is, intentions not personally held by every member.Footnote 70 This can happen when members operate under ‘shared procedural policies,’ understood as rules specifying how certain questions are to be settled, for example by majority vote.Footnote 71 When such a procedure issues an outcome poised to guide the group’s subsequent deliberation and action, that outcome can constitute a ‘procedure-based intention of the group,’ even if some members do not personally intend what the group has settled on.Footnote 72
Bratman illustrates the point with a group ‘engaged in a shared intentional activity of providing medical aid in crises around the world.’Footnote 73 The group has procedures for allocating its limited resources, and when one such procedure issues an outcome set to guide the group’s further activity, the group thereby holds a procedure-based intention to act. This remains true even where some members form no corresponding intention of their own: they ‘might simply be unaware’ of the procedure’s result, or consider the decision a ‘serious mistake’ and even ‘resist intending that the group act accordingly.’Footnote 74 What makes the outcome the group’s intention is that members continue to participate in the shared procedural policies through which their course of action is settled.Footnote 75 This extension preserves Bratman’s reductionist orientation. It shows that group intentions and shared intentions may be analytically separable, but not that group intentions float free of individual-level intentional structures. On the contrary, procedure-based group intentions count as intentions of the group only because they are embedded in shared procedural policies, so that the ‘infrastructure of procedure-based group intention systematically involves shared intention.’Footnote 76
Overall, Bratman’s reductionism thus appears to constitute a promising ontological account for the purposes of IR: strictly non-summative in nature, it solves many of the challenges IR scholars have leveled at reductionist views on collective intentionality in the past. In what follows, we combine Bratman’s basic framework with original arguments to develop from it a new perspective on the two key issues involving collective intentions in IR: state intentions and joint interstate action.
Reductionism and state intentions
The first question that arises when seeking to apply reductionism to the case of state intentions is whether there exists a group of individuals whose shared intention could plausibly serve as the foundation of the state’s intentions. Bratman’s account suggests such a group should be relatively small and remain below the level of organizational complexity characteristic of large-scale institutions: after all, his plan-theoretic account is primarily targeted at explaining the sociality of small groups, although in later works he leaves open the possibility that it could be extended to larger and more complex groups as well.Footnote 77
A survey of the IR theory literature suggests that, at least in the context of foreign economic and military policy, many scholars would argue that there is, indeed, such a group: the ‘foreign policy executive’ (FPE). In their view, the FPE – consisting of ‘the president, prime minister, or dictator, and key cabinet members, ministers, and advisors charged with the conduct of foreign and defense policies’ – constitutes ‘the essence of the state’ because it ‘is the sole authoritative maker of foreign policy.’Footnote 78 In extreme cases, where the power to decide on foreign policy is highly concentrated, the FPE can be very small. Think, for instance, of Soviet foreign policy decision-making during the early Cold War. In most cases, however, the group will involve more individuals while still remaining relatively small. Paradigmatic examples include U.S. President John F. Kennedy’s Executive Committee of the National Security Council (ExComm) during the Cuban Missile Crisis or Israeli Prime Minister Menachem Begin’s Ministerial Committee on Security Affairs.Footnote 79
In our view, at least in those issue areas in which the FPE plays a decisive and largely independent planning role, it is the collective intention based on the appropriately structured individual intentions of the FPE members that scholars refer to when they speak of the state’s intentions. They serve as authorized proxies, representatives, or delegates of citizens and, embedded in relevant institutional practices, deliberate amongst themselves and ultimately decide on a plan to address the foreign policy challenges they confront.Footnote 80 More concretely, they decide how to utilize the means at their disposal, which include natural resources, the industrial base, as well as military equipment and personnel, in pursuit of shared goals associated with their official roles, such as safeguarding territorial integrity, political autonomy, wealth, and the general well-being of the citizenry.
This understanding broadly echoes strands of IR scholarship that invoke state intentions as central to studying international politics while conceiving of them, at least intuitively, in reductionist terms. For instance, Sebastian Rosato notes that ‘it is a state’s executive … made up of … the head of government plus a handful of ministers and advisers … that formulates [the state’s] intentions.’ In his eyes, therefore, ‘a state’s intentions are the product of bargaining among the individuals that make up its executive.’Footnote 81 Along the same lines, Norrin Ripsman et al. argue that the FPE ‘is the most important actor to focus on when seeking to explain foreign policy and grand strategic adjustments.’ As they note, it is undoubtedly true that ‘many individuals inside and outside of the government have an interest in foreign policy and make statements about policy, yet not all of these actors have meaningful input into policymaking.’ Accordingly, they recommend a focus on those that actually ‘weigh in on the final decision.’Footnote 82
Note that understanding state intentions along Bratmanian lines goes decidedly beyond a simple summative-reductionist view, which would analyze a state’s intention as the mere aggregation of individual intentions with the same content. Such a view would reduce state intention to the individual intentions of each FPE member, all of which have the same content, say, to pursue policy X. Our account also requires more than that the members of the FPE come together, deliberate, and vote to pursue policy X by whatever agreed-upon procedure.
Rather, it requires the individual decision-makers’ intentions to be appropriately structured and interrelated. Once the FPE has decided, through an agreed-upon procedure, to pursue policy X, most (but, as noted above, not necessarily all) of its members – Prime Minister, Foreign Minister, and so on – hold the intention ‘I intend that we X.’ In addition, their intentions must be reflexive (in that each FPE member’s intention is directed at the policy’s implementation proceeding at least partly by way of their own intention) and interlocking (in that the FPE members intend not just to somehow achieve X but by way of the others’ participation), track meshing subplans (in that the FPE members ensure their plans for how to implement X are co-realizable), and be sustained by mutual responsiveness (in that the various officials support each other in fulfilling their parts needed for the joint activity to succeed). The resulting picture is one of individual government officials who jointly settle on a policy and then act in concert to implement it.
This account of how states form intentions, once again, closely mirrors how various scholars – specifically those working in theoretical traditions embracing the notion of the state as a rational unitary actor – conceive of the planning processes involved at the highest level of government. For instance, many contemporary political realists posit that states are typically rational precisely because their policy decisions result from an earnest deliberative process after which FPE officials, even if they initially disagreed, work together to implement the decision taken. Consider, for instance, how John Mearsheimer and Sebastian Rosato describe the process leading to the U.S. decision to escalate during the Korean War. In their telling, ‘the relevant leaders,’ who were each committed to determining the best course of action for the country, ‘met frequently throughout the crisis to discuss the appropriate policy,’ and in the end decided to move U.S. forces into North Korea. Although the authors do not couch the issue in Bratmanian terms, what they describe is effectively the kind of shared intention formation among a small group of decision-makers posited by our reductionist view.Footnote 83
More direct support for the claim that Bratmanian shared intentions arise within FPEs can be found in the historical record itself, which contains ample evidence of foreign-policy officials describing their own attitudes and thoughts in language reminiscent of our account. To name just one example, during a 1988 meeting on the U.S. position in the Strategic Arms Reduction Talks (START), Secretary of State George Shultz commented that ‘we should all pledge to make an all-out effort.’ After further deliberation, National Security Adviser Colin Powell concluded that ‘we have, therefore, Mr. President, a decision, and the decision is that we’ll go for the gold, and we will drive towards that end,’ while adding that, to succeed, ‘[w]e will need high-level involvement now.’ Later, when Shultz proposed complementary diplomatic and defense-level efforts as part of this strategy, Defense Secretary Frank Carlucci explicitly took up his assigned contribution – ‘I’ll be happy to do this’ – while discussing how his contribution should be coordinated with the broader negotiating effort.Footnote 84 This exchange illustrates that FPE members frequently understand themselves as engaged in the joint pursuit of shared goals, with interdependent attitudes, meshing contributions, and mutual support among their respective actions.
Reductionism and joint interstate action
The second issue area in which collective intentionality features prominently in the IR literature concerns sophisticated forms of interstate cooperation. For this phenomenon, too, we can develop a reductionist view based on Bratman’s framework.
Simply put, a reductionist account constructs the phenomenon as involving a group of states with an intentional structure fulfilling all the requirements needed to give rise to a collective intention in pursuit of shared goals. To be more specific, this means that the intentions of the states involved in a cooperative endeavor need to be reflexive and interlocking track-meshing subplans and be sustained by mutual responsiveness. If these conditions are fulfilled, our reductionist view holds that the respective group of states forms a collective intention in the pursuit of shared goals, giving rise to a phenomenon that is qualitatively different from other forms of cooperation where states merely coordinate their policies, either explicitly or implicitly, in the pursuit of their individual goals.Footnote 85
Before moving on, note that elucidating joint interstate action in these terms presupposes that there exists such a thing as the intentions of states. Reductionism, of course, provides the conceptual resources to justify this move. It offers a nested view in which the collective intention of a group of states is reducible to the individual intentions of the participating states, while each state’s intention is in turn reducible to the individual intentions held by its FPE members. This view treats each participating state as an agent in its own right, but only in the weak sense outlined above: as an internally structured collective capable of coordinated intentional activity. Each state can therefore figure as an intentional participant in the formation of the collective intention of the group of states, while its own intention remains reducible to the intentions held by its FPE members. Importantly, however, scholars need not accept our account of state intentions in order to accept our account of joint interstate action. Even those who reject the view that state intentions are grounded in the individual intentions of FPE members can still endorse a reductionist account at the interstate level. As long as states are taken to form intentions on some ontological basis, the collective intentions of a group of states can be analyzed as reducible to the individual intentions of the participating states.
As highlighted previously, Bratman’s basic ontological framework is primarily targeted at illuminating instances of joint action involving a relatively small number of participants and lacking the complexity of larger institutions. What kind of real-world constellations of states coming together are likely to meet these requirements?
In our view, coalitions of the willing and other such ad hoc coalitions are a paradigmatic example: they typically involve a limited number of states that ‘share a common purpose’ and possess the ‘deliberative capacity in order to pursue it,’ coming together in a loosely organized format typically characterized by ‘task-specific mandates,’ ‘speedy set up,’ and ‘short timeframes.’Footnote 86 An illustrative example of such a grouping is the International Force for East Timor (INTERFET). This ad hoc initiative was primarily championed by a small number of core states, including Australia and New Zealand, and even by the end came to include no more than 22 nations. Formed in response to UN Security Council Resolution 1264, it had the purpose of restoring peace and security in conflict-ridden East Timor. In her discussion of joint purposive activities, Erskine points to INTERFET as a prime example of states that, as ‘members of a coalition of the willing,’ are ‘temporarily united in pursuit of a common purpose’ and thus exhibit a particularly sophisticated form of cooperation.
Rather than simply coordinating their policies in order to realize their individual goals, the states involved in INTERFET formed a collective intention to pursue their shared goal of bettering the situation in East Timor together. Where Erskine herself merely hinted at conceiving of cases like this in reductionist yet non-summative terms, our reductionist view on collective intentionality provides the required theoretical scaffolding to support it: the FPEs of Australia, New Zealand, and the other core nations of INTERFET each formed an intention for their state to participate in the joint endeavor, and the resulting set of state intentions can plausibly be reconstructed as exhibiting all the features required for a shared activity in Bratman’s sense. The collective intention that undergirded their joint goal pursuit, per reductionism, is thus reducible to the states’ individual intentions, which in turn reduce to the individual intentions of their respective FPE members.
The account, however, is not limited to explaining such groupings with narrowly defined tasks, but can also elucidate instances in which a small group of states comes together with the intent of pursuing broadly defined shared goals for an indefinite period. These include, for example, security concerts, that is, condominia in which the great powers within a system come together to jointly manage their security environment.Footnote 87 The Concert of Europe, which the European great powers of Austria, France, Prussia, Russia, and the United Kingdom formed following the Napoleonic Wars, serves as an instructive example. A reductionist view explains the collective intentionality that the great powers displayed in the Concert in broadly individualistic terms. Within each of the great powers, the respective FPE members formed the shared intention to work with the other great powers to keep the balance of power in Europe intact and prevent revisions to territorial boundaries as previously established at the Congress of Vienna. These state-level intentions, then, provided the basis for the great powers’ joint pursuit of this shared goal. The Concert therefore differed from more ordinary forms of great-power cooperation because it involved not merely parallel or coordinated policies, but interrelated intentions toward a joint political project.
Reductionism and the study of international politics
The point of this piece is not just to show that developing a sophisticated reductionist account of state intentions and joint interstate action is possible, but also to demonstrate that adding such an account to the ongoing debate on the ontology of collective intentionality promises to enable productive theorizing on issues of concern to IR scholars. To this end, we generate novel perspectives on current issues in both the empirical and normative study of IR, thus illustrating how greater ontological pluralism can be an important catalyst of theoretical advances.Footnote 88
Empirical theorizing
When it comes to the empirical study of international politics, that is, attempts to produce knowledge of how the world of international politics works, reductionism promises to aid scholars’ efforts in several ways.
First, recall that our reductionist account of collective intentionality ultimately centers individual actors – highly influential government officials who are members of their state’s FPE – and their interactions as the base of state intentions. Such a view on intention formation within the state implies clear guidelines for how to empirically study why states decide to adopt or change their foreign policies and grand strategies. In particular, it suggests that scholars should begin by ‘identifying the members and composition of the FPE; pinpointing those policymakers “who matter” in the formulation of foreign policy and where decisions are made; and determining the relative power and influence of actors within the FPE.’Footnote 89
By directing attention to individual actors and their interactions, the account dovetails nicely with IR scholars’ resurgent emphasis on how causal factors located at the individual level – such as psychological biases and heuristics, personality, or age – are critical for explaining state behavior and international outcomes. At the same time, however, it remains compatible with approaches that give causal primacy to state- or system-level variables, since even these typically treat decision-makers as an important link in the causal chain leading from strategic incentives and constraints to foreign policy. Our reductionist account is agnostic about where the FPE members’ individual intentions come from – whether they are shaped by personal idiosyncrasies, domestic political incentives, material threats and opportunities, social structures, or other factors. What it does suggest is that scholars seeking to understand the determinants of a particular policy should begin their inquiries by examining the thought processes of and deliberations among the few government officials present in the metaphorical smoke-filled rooms where foreign policies are typically decided.
The account, however, provides tools for understanding foreign policy decision-making that go beyond these general guidelines, which in many ways echo traditional foreign policy analysis. In particular, by highlighting that state intentions are constituted by the FPE members’ appropriately structured intentions, it allows scholars to identify deviations from this ‘normal’ mode of foreign policymaking and describe what exactly distinguishes such deviant cases from how states usually form intentions.
Consider, for instance, the case of the second Trump administration. Many observers worldwide describe the way in which the United States under President Donald J. Trump decides on and implements its foreign policies as a drastic departure from standard practices. Among other things, Trump’s personal instincts appear to drive much of the process; there are doubts as to whether meaningful, uninhibited deliberations take place before decisions are taken, and even within his inner ring of officials and advisors there often seem to be no unifying beliefs about what policy to pursue for what purpose. Our account provides the tools to describe this set of circumstances and their far-reaching implications succinctly. Specifically, a reductionist view on the matter suggests that U.S. officials today may not, in fact, hold the kind of shared intention typically characteristic of foreign policy decision-making. To put the matter more concretely, upon analysis of the archival record when it becomes available, scholars may find that the United States, in no meaningful sense, had an intention to act in a particular way during Trump’s second time in office. Instead, the record may suggest that whatever policies were implemented resulted from a single individual who, driven by personalistic instincts and interests, imposed his plans for how to behave on others around him.
Second, reductionism’s unique view on sophisticated forms of cooperation yields a new understanding of coalitions of the willing, concerts, and similar small group formats that differs markedly from the perspectives generated by alternative theoretical frameworks, such as many IR theorists’ favored variety of non-reductionism. In particular, it suggests that such cooperative endeavors are much more fragile than others who have written on them appear to claim.
On Gilbert’s plural subject account, states engaging in joint action are ‘jointly committed to intend as a body’ to perform the joint action.Footnote 90 The joint commitment binds states to one another in a normative ‘relationship of owing,’ leading them to align their individual behavior with that commitment. Through the formation of such a collective intention, a ‘plural subject’ emerges, and some even go as far as claiming that this creates ‘a new agency in which authority and responsibility over action is shared.’Footnote 91 Endorsing such a view paints a picture of joint interstate action as strongly constrained by the ‘collective reason’ emerging from the joint commitment.Footnote 92 It implies a high degree of stability because the obligation constituted by the joint commitment exerts normative pressure on each state to act in support of the joint activity.Footnote 93 A state may, of course, choose to violate the commitment unilaterally, but in doing so would incur high normative costs – including giving others standing to rebuke it – and would still not thereby rescind the commitment on Gilbert’s view.
In contrast, a reductionist account leads to a very different depiction of joint interstate action. As reductionism denies the presence of an irreducible collective reason or joint commitment that effectively aligns the behavior of the participating states, the continuation of joint action depends on whether the participating states – in light of their evolving interests, plans, and priorities – continue to maintain their respective intentions to act together. A reductionist account therefore portrays joint interstate action as comparatively fragile: because the state-level attitudes that constitute it may change over time, the joint activity remains more susceptible to disruption or termination than it would be under a Gilbertian account, where a joint commitment generates obligations that exert continuing normative pressure on the participating states.Footnote 94
Against this backdrop, reductionism suggests that even cooperative endeavors that stand out for their relative longevity – such as the Concert of Europe discussed previously – may have been much less stable than proponents of non-reductionism have presented them to be. For example, previous non-reductionist explanations of cooperation in the Concert have posited that the great powers’ joint commitment to manage their security environment together served as the basis of their collective intentionality and led to the emergence of a new plural subject in the sense of Gilbert. This joint commitment, in turn, imposed important restrictions on the conduct of each participating power. As Mitzen explains, each member of the Concert came to ‘stand […] in a relationship of obligation to the others and ha[d] the standing to hold them accountable for actions relevant to the commitment.’Footnote 95 This lent significant stability to their joint endeavor.
Reductionism, in contrast, explains the collective intentionality the great powers displayed in the Concert in broadly individualistic terms, reaches different conclusions about its stability, and suggests an alternative explanation for why the arrangement persisted for as long as it did. Since, per reductionism, the individual actors did not face an overarching joint commitment that exerted a distinctly normative pressure on them to fulfill their role in the joint activity, reductionism suggests the cooperation that took place in the Concert to have been much more fragile. The great powers’ intention to pursue their shared goal together critically depended on the attitudes and behavior of each individual group member, meaning that a decision by any FPE to rescind its participation in the joint endeavor – for instance, because it no longer served the state’s strategic interests – would have been enough to lead the system to collapse. In this view, it was only possible for the Concert to have lasted as long as it did because of a unique constellation of compatible interests that persisted for much longer than is typical for the often rapidly changing world of great power politics. As soon as interests started to diverge, reductionism would expect the collective intention to work together to erode.
Indeed, such a depiction of the Concert aligns closely with how some IR theorists have presented its underlying dynamics. Mearsheimer, for instance, noting that ‘[c]oncerts usually last only a few years,’ suggests that the Concert of Europe worked well for several years because the great powers did ‘not have much to gain militarily by attacking each other, given the rough balance of power between them,’ and that each of them had a ‘significant interest in maintaining the status quo’ during this period. As the balance of power began to change and more and more disputes on how to deal with other states arose among the great powers, however, Mearsheimer argues the Concert became increasingly ineffective and, while no change was apparent from the outside, it essentially continued to exist in little more than ‘an abortive form’ prior to the outbreak of the Crimean War.Footnote 96
At the same time, reductionism avoids the perhaps tempting pitfall of concluding from this that the Concert was simply yet another case of ordinary cooperation of the sort frequently observed between great powers. In stressing that the Concert involved a form of collective intentionality that made possible the kind of joint pursuit of a shared goal characteristic of sophisticated forms of cooperation, our account serves to distinguish it from other, less ambitious cases of cooperation. In effect, reductionism makes possible a nuanced middle-ground position that acknowledges such cases as involving a qualitatively different form of state interaction while simultaneously maintaining the primacy of states’ individual interests and the fragility that it introduces into all cooperative endeavors under international anarchy.
Normative theorizing
Collective actors in international politics often make decisions with significant moral consequences. By analyzing collective action in terms of the structured attitudes of individual participants, reductionism provides a fruitful framework for assessing collective moral responsibility and its relation to individual responsibility. More specifically, it directs attention to the individuals through whose attitudes and actions collective intentions are formed and carried out, and thus – more directly than non-reductionist views – suggests that moral responsibility lies primarily with those individuals.Footnote 97
Erskine’s normative analysis of coalitions of the willing reinforces the thought that, in many IR contexts, individual participants often remain the central bearers of moral responsibility. As noted above, Erskine treats such informal associations in broadly reductionist terms: they may engage in ‘joint purposive action,’ but they do not qualify as institutional moral agents in their own right. Responsibility therefore remains, in her view, with the individual states and other actors that constitute, or could constitute, them. In her discussion of the 2011 intervention in Libya, for example, she locates the relevant responsibilities in the states that formed the coalition in response to UN Security Council Resolution 1973.Footnote 98
If, as we propose, state intentionality is understood in similarly reductionist terms, Erskine’s analysis can be extended to the moral evaluation of many state decisions and policies. In cases where a small group of influential decision-makers deliberates, plans, and authorizes a policy while being aware of its likely consequences, it seems appropriate to assign primary moral responsibility to those who shaped and approved the decision. Think, for instance, of the United States’ decision to authorize the Central Intelligence Agency’s enhanced-interrogation program, which has been widely criticized as involving torture and grave human rights violations. President George W. Bush’s approval of the program followed recommendations from the National Security Council’s Principals Committee, which comprised a small number of high-ranking officials such as Dick Cheney and Condoleezza Rice. Many today believe that Bush and the advisors who directly shaped the decision-making process are morally responsible for what has been described as constituting nothing less than ‘war crimes.’Footnote 99
Reductionism therefore undergirds an individual-centered account of collective moral responsibility that seems appropriate in many IR contexts. Beyond suggesting that individuals are often the primary bearers of responsibility, however, reductionism also helps specify the scope of their responsibility, and the degree of their blameworthiness, in cases of collective wrongdoing.
Recent philosophical work, in fact, makes exactly this point. Olle Blomberg and Frank Hindriks argue that, in cases of shared intentional wrongdoing understood along Bratmanian lines, participants are normally more blameworthy than actors who produce the same outcome through parallel or strategic interaction. The reason is that, in shared intentional wrongdoing, each participant intends not only the wrongful outcome, but also that the others’ intentions help bring it about. Their wills are therefore mutually implicated: each is responsible not merely for their own discrete contribution, but for the shared activity realized through their interrelated intentions.Footnote 100 In the IR context, this helps explain why officials or states involved in jointly authorizing wrongful policies may bear responsibility not only for their individual contributions, but also for the collective policy they helped realize.
The broader U.S. detainee-abuse controversy illustrates the point. The Senate Armed Services Committee explicitly rejected the view that detainee abuse was merely the work of ‘a few bad apples,’ concluding instead that senior officials, through a shared policymaking process, ‘solicited information on how to use aggressive techniques, redefined the law to create the appearance of their legality, and authorized their use against detainees.’Footnote 101 In line with Blomberg and Hindriks’ argument, the officials’ joint role in formulating and approving this policy framework made them individually responsible for the shared wrongdoing, not merely for their own contributions, and rendered them more blameworthy than they would have been had they acted independently.
This straightforward case of joint authorization is not the only one reductionism helps illuminate. As noted above, Bratman’s framework also allows for the possibility that a group has an intention that is not shared by every member of the collective, namely when that intention is generated through a shared procedural policy. This is perhaps one of the most interesting cases of moral evaluation that the framework brings into view. It raises the question of how individual responsibility and blameworthiness should be assessed when a member does not personally share the group intention, but continues to endorse the procedure through which it was formed. Relevant differences may arise between cases in which a member is simply unaware of the procedure’s outcome and cases in which a member knows about the outcome, rejects it, but nevertheless continues to accept the procedure that authorizes it.
To name just one example of this latter possibility, consider George Ball’s dissent within President Lyndon B. Johnson’s Vietnam decision-making circle in July 1965, when the administration was considering a major expansion of the U.S. military commitment. As Under Secretary of State, Ball was part of the senior foreign-policy elite advising Johnson to escalate the U.S. involvement. Yet, he himself was a prominent dissenter from the course favored by most of the President’s senior advisors. In a key meeting, he warned, ‘I can foresee a perilous voyage,’ while adding, ‘[b]ut, let me be clear, if the decision is to go ahead, I’m committed.’Footnote 102 The administration subsequently settled on escalation, and Ball remained within the FPE for some time thereafter.Footnote 103
Although a detailed normative assessment of the case lies beyond the scope of this article, it illustrates how the reductionist view we propose can both illuminate underexplored questions of moral responsibility in international politics and provide novel conceptual resources for approaching them. Most immediately, it raises the question of whether a dissenter such as Ball, who opposed the policy but ultimately remained within and acted through the decision structure that produced it, nevertheless bears moral responsibility for the resultant wrongdoing. If so, further questions arise about how far that responsibility extends to earlier decisions and forms of participation that helped make the later policy possible, and whether resignation from the FPE could be sufficient to sever that responsibility. These are complex moral questions that require theorizing beyond Bratman’s account, but a reductionist analysis helps identify and structure them.
Conclusion: the promise and limits of reductionism
This article has argued that IR theorists have been too quick to dismiss reductionism as a viable ontology of collective intentionality. Much of the existing literature has treated reductionism as if it necessarily entailed a crude summative view, according to which collective intentions are nothing more than aggregates of parallel individual intentions. Against this assumption, this piece has developed a more sophisticated, non-summative view. Drawing on Bratman’s account of shared agency, this view holds that collective intentionality is reducible not to the mere aggregation of individual intentions, but to a structured configuration of individuals’ intentional and dispositional states. We have argued that this view overcomes many of the concerns that have motivated skepticism toward reductionism in IR, yields distinctive accounts of both state intentions and the intentions of groups of states, and helps advance empirical and normative theorizing in the field.
At the same time, it is important to be aware of the potential limitations of the reductionist view developed here. As noted previously, the account most straightforwardly applies to cases in which the relevant collective is sufficiently small, and organizational complexity and hierarchical differentiation remain sufficiently manageable. These include many of the phenomena on which this article, like much of the literature, has focused, such as foreign-policy executives, coalitions of the willing, security concerts, and other relatively small groups of states engaged in the joint pursuit of shared goals.
Not all collectives of potential interest to IR scholars, however, fit this model equally well. Complex intergovernmental organizations such as the UN are larger, more internally differentiated, and more institutionally elaborate than the small-scale cases for which Bratman’s basic account was originally developed. The UN, for instance, possesses international legal personality, operates through codified rules and institutionalized voting procedures, maintains institutionalized budgeting and financing arrangements, and has developed an extensive specialized bureaucracy, including an international civil service formally responsible only to the organization. These features make much of its conduct difficult to capture solely through the small-scale intentional structures emphasized in Bratman’s basic framework.
This does not mean, however, that reductionism cannot be extended to account for such cases as well. Bratman himself, in fact, has recently argued that his planning theory can be developed into an account of institutional agency without positing an irreducible group mind or holistic institutional subject.Footnote 104 Reductionism may therefore reach further in an IR context than its basic small-scale formulation suggests. Still, such an extension is more contentious than the applications defended in this article: once one moves from small-scale joint action to highly institutionalized organizations, it becomes less clear whether collective intentionality, understood as the least demanding concept in the ontology of collectives, is sufficient to capture what is at stake.
In particular, for highly institutionalized organizations, the central question may no longer be merely whether a plurality holds an intentional state together, but whether the collective has acquired more robust forms of group-level agency or personhood. Does it function as a relatively unified center of belief, decision, and action? Can it sustain and revise attitudes over time through institutional mechanisms? Does it possess forms of autonomy that make it more than a site of member-state coordination? These questions seem to take us beyond collective intentionality per se and toward more demanding theoretical notions.
In many ways, however, this only further underscores the methodological value of beginning with investigating collective intentionality before engaging in extensive debates about agency, personhood, and other such notions. Starting with the least demanding concept allows us to show that reductionism explains considerably more than has been recognized, while also identifying the point at which more demanding questions about institutional agency and autonomy begin to arise. Whether one judges this to mark a serious limitation ultimately depends on how one understands the role of complex international institutions in world politics. If one does not regard such institutions as causally relevant actors in their own right, but rather as arenas, instruments, or products of member-state interactions, then the reductionist account developed here may take us a considerable distance. If, however, one’s theoretical priors are that these organizations possess forms of agency and autonomy that go beyond this, then additional theoretical resources may be needed.Footnote 105
Overall, then, future research should ask whether, and how, the reductionist view developed here can be extended to complex international organizations. Irrespective of what scholars will find, however, the central conclusion of this article stands: reductionism deserves to take its place as a serious contender for how to understand collective intentionality in IR, and applying it promises to produce exciting new perspectives on issues of perennial interest in the empirical and normative study of international politics.
Competing interests
The authors declare none.
Open access
