Introduction
We might argue that personal identity is both a relational fact and a normative relationship; that is that personal identity is identified by mutual obligations and the authority of an intention or principle over temporal stages of a person. At the same time it becomes apparent that integrity is not something we just have or not have, but also a virtue and an ideal. We at least need to recognise integrity as a normative standard of conduct. Therefore, the person may still form a fundamental unit in normative analysis, in spite of the reductionist truth about personal identity, as Parfit sees it.Footnote 2
With these arguments in hand, I now want to develop the idea of the corporate person. The standard case against corporate personality is articulated by claims that mirror Parfit's objection to ethical individualism.
(Q1) Reductionism about groups is the familiar claim that the actions and intentions of collectives like states, nations and corporations are ontologically derivative from the actions and intentions of their individual membership. States are not grounded in some substratum, existing underneath its individual members, as some corporate realists may have mistakenly thought, but merely emerge from individual intentions and acts and the relatedness between them.
(Q2) Ethical individualism asserts that collectives, conceived of as a relationship between individuals, only to have derivative significance in morality and rationality. In contrast to the temporal stages of a person, the fact that two individuals are in some sense connected contains no moral and prudential significance as such. Instead, the significance and value of relational acts should ultimately be derived from their significance and value to the individual good.
Corporate personality is conceivable only if at least one of these claims could be rejected. Traditionally, perhaps, this has led us to deny Q1. But what I want to suggest is that we reject Q2 in favour of its counterpart:
(Q3) Associationism argues that even though the actions and intentions of collectives are reducible to a relationship between individuals, it could still form a non-derivatively significant unit of normative theory. The fact that two individuals are in some sense connected could be just as significant for rationality and morality as the fact that two person-stages are so related. So associationism argues that interpersonal relatedness is inherently reason-giving, that is, its value does not always have to be reduced to the interests and desires of separate individuals.
Why accept associationism? We need to show that the same pivotal arguments could be made in the case of interpersonal relations. To begin with it is wise to invoke practice theory. That is, we need to start by assuming that certain practices are an inescapable feature of human life or of our form of life, and from this reason towards a theory of personal identity. We must demonstrate that there are normative grounds for accepting Q3. And individualists will probably want to reject this claim. They will want to support the moderate claim and reject associationism at the same time. Consequently, we need to provide an independent argument for the necessity of corporate personality over and above individual personality. Such an argument must illustrate that a purely individualist interpretation of the idea of the person leaves an important gap in normative theory.
Second, we need to claim that Q3 refers at a more fundamental and normative level to what might be termed integrity. The challenge is to show that integrity is not necessarily confined to associative links between ‘person-stages’ or ‘person-events’ contained in one human life span. Integrity, it will have to be argued, is also capable of covering a whole range of associative links between individuals. In this article, I shall argue that the idea of integrity as a virtue is as applicable to interpersonal association as it is to the association between ‘person-stages’ or ‘person-events’. The argument for corporate personality needs to remain within four critical side-constraints.
First, it should be able to explain corporate personality within the limitations of reductionist metaphysics. I call this the reductionist constraint. This notion should therefore be framed in reductionist terms of ‘shared’ and ‘collective’ intentions and preferences.
Second, it should nevertheless give us an account of corporate personality that is independent from a normative point of view. I call this the sovereignty constraint. Corporate interests, preferences and intentions, I want to argue, are not merely the summation or function of interests, preferences and intentions. The concept of corporate personality needs to be defined in such a way that it can add something distinctive to the normative domain. Both constraints are largely conceptual and are addressed in Parts I and II below.
Third, if the idea of corporate personality is to be applied to associations like the state, it should be able to avoid romantic and nostalgic conceptions of a monolithic community. I call this the empirical constraint. I propose an interpretative and institutional conception of integrity that accommodates the concern that our present-day world is characterised by too much complexity and a diversity of identities. Ronald Dworkin in his Law's Empire has told us to think of intention in interpretive rather than psychological terms.Footnote 3
Finally, corporate personality must have normative purchase. We could refer to this as the practicality constraint. As my argument plainly rests on the so-called practice theory, corporate personality must be able to do normative work. It must explain something that ethical individualism cannot adequately address.Footnote 4
The main thrust of this article will be to introduce a general model of corporate personality using a reductionist account of ‘shared’ and ‘collective’ intention and I do this in Part I. Couching corporate personality in these terms, I argue, takes care of the first constraint. However, if we follow this model, the sovereignty constraint turns out to be the greater challenge. Part II addresses no less than three different understandings of corporate personality in which we collapse back into a mode of ethical individualism: ‘conventionalism’, ‘contractualism’ and ‘voluntarism’. For each of these understandings, therefore, I advance an opposing interpretation.
I. The reductionist constraint
I begin with the issue of how the idea of corporate personality could be made consistent with ontological reductionism about collectives. In order to stay within this constraint, I want to introduce the notions of ‘shared’ and ‘collective’ intention and preferences. These notions avoid anthropomorphism by arguing that the intentions of collectives emerge from a rational bundling of the intentions of a number of individuals, just as the intentions of individual persons consist out of the rational bundling of intentional episodes or person-stages. Essentially, the intentions of a group are viewed as the intentions shared by its members, and here I am indebted to Bratman's work.Footnote 5 I argue that the notion of a shared intention or preference meets the first challenge.
It is important to note immediately that there is something strange about the form of the intentions that are to be shared. While occurrences in the brains of individual people fluctuate, the key is that they need to be designed not for individual action but for group action. Another way of putting it is to say that while individuals always do the thinking and the intending, they can think and intend in different modes. More precisely, people can think in an individual and group mode. When individuals think in the former mode, so-called ‘I-intentions’ are put in charge over their decisions and actions. When they think in the latter mode then so called ‘we-intentions’ are put in charge over their decisions and actions.
The notion of a we-intention (or collective intention) needs to be unpacked a bit further. What does it mean for action to be guided by collective intention? On Bratman's account of shared intention, a we-intention is an intention to undertake something together. I can intend that I do this or that – in this case I think in the mode of an individual – but I can also intend that we do this or that. In the latter case I have a collective intention. The hallmark of such intentions, then, is that they are defined over the actions of a group of people. For example John, Paul, George and Ringo intend to form a band and compose and play music together. Here John does not just intend to play himself; he intends them to play together. And Paul, George and Ringo we-intend the same.
John Searle advances a similar notion of collective intention. As he puts the idea: ‘We simply have to recognise that there are intentions whose form is: We intend that we perform an act A; and as such an intention can exist in the mind of each individual agent who is acting as part of the collective.’Footnote 6 The point Searle emphasises is that such collective intentions are as primitive or as basic as so-called I-Intentions.
Robert Sugden focuses on collective preferences rather than collective intentions. But the former have precisely the same features as the latter. According to Sugden, I can define a preference in order to guide my actions. If the former sort of preference directs my reasoning, I think as an individual. When the latter sort of preference directs my reasoning, I ‘think as a team’. He notes:
‘… although team-directed reasoning is carried out by individuals, it is not an instance of individual reasoning as that is represented in the standard theory of rational choice. The two kinds of reasoning are different in structure. In the standard theory, the individual appraises alternative actions by her in relation to some objective (her preferences), given her beliefs about the actions that other individuals will choose. An individual who engages in team-directed reasoning appraises alternative arrays of actions by members of the team in relation to some objective.Footnote 7
Now, since we engage in collective action on a daily basis, it is not implausible to think that collective intentions and preferences are everyday phenomena. However, the idea of collective intentions and preferences invites an important logical obstacle, which needs to be removed.Footnote 8 Let us concentrate on we-intentions. Evidently at a physical level John, Paul, George and Ringo can only act individually. There is, of course, no separately existing band that could generate music by itself. And this fact points to a discontinuity between intention and action. Collective intentions can only result in individual actions. But is it really intelligible to say John we-intends that he and Paul do something? After all, we have to assume Paul, not John, is ultimately in control of his actions and will have to make up his own mind whether or not to participate. How, then, could John intend for Paul? And to what use?
Bratman shows, in my view persuasively, how the discontinuity problem could be overcome.Footnote 9 John can we-intend, he argues, on the basis of the further belief, assumption or prediction that Paul will have the same intention. Given that John thinks that Paul intends to do his part, he can form the intention that they undertake something together, without assuming that he has direct discretion over Paul's action. Whether they play together is under his control. It might be objected that the need for such beliefs and assumptions detracts from this view as compared to individual agency. But the obvious response is here that future-directed intention operates by the same logic. Similar assumptions and predictions are required in order to understand how I could logically intend in the morning to go to the cinema tonight. After all, I have currently no more discretion over my future selves than John over Paul's actions. Hence, I can only intend something in the future, on the basis of the assumption that my future selves will co-operate. I need to say something about personal identity here, for it may be argued that our future selves are related to our present selves, in the sense that they are the same self but that we are not related to others in this same direct way. Here I think we need to augment my remarks on personal identity with reference to what we might call practice theory. That is, instead of grounding normative theory in a conception of the person and personal identity alone we have also to begin by assuming that certain practices are a basic feature of human life, as we know it, and from these practices we may reason towards a certain general and normative understanding of the person and personal identity. Otherwise certain basic intuitions, judgments and would not be intelligible. Rawls argues for something along these lines in his essay ‘The Independence of Moral Theory’.Footnote 10 Moreover, Rawls argues in Political Liberalism that any workable notion of the person has also to assume that we all have a basic sense of justice and a capacity for a conception of the good.Footnote 11 There is no place in Rawls's account for the extreme claim given by Parfit, in Chapter 14 of Reasons and Persons (1984), since that is so radically at odds with our political and democratic processes that it cannot be entertained seriously in terms of any reasonable model that involves process considerations in the real world.
I return to this analogy between collective and future-directed intention below. There is, however, a further point that is helpful in allaying the concern about discontinuity. It is certainly true that John's we-intention has no discretion over Paul's actions. But the point is rather that John's we-intention guides and constrains his own intentions and actions in a particular way.Footnote 12 And if Paul has the same we-intention, this will serve to frame and constrain his own individual options in a particular way. This point is captured by Bratman's important claim that the structure of collective agency is two-tiered. A collective intention to do something together necessitates a degree of planning and co-ordination so that each individual member knows what he must ‘I-intend’ to do himself in order to achieve the intended result at the level of collective action. In the above example, apart from having a we-intention, John will have to I-intend that he sings a certain tune, Paul will have the I-intention that he play certain chords on the guitar, and so forth. A we-intention then, involves having a plan delegating specific tasks in such a way that subsidiary intentions are defined over individual actions and their outcomes.Footnote 13 Collective agency, among other things, entails that lower-tier I-intentions are framed, constrained and guided by a higher-tier we-intention. To put it differently still, collective agency could only emerge if the collective intention or preference is ultimately in charge of an individual's decisions. Thinking as a team, an individual will ultimately have to understand the rationality of his choices and actions in terms of this higher-order, collective intention.Footnote 14
Three further features of this model of we-intentions need to be mentioned here. First, a collective intention or preference could still leave ample discretion for individual choice and interpretation. Rather than having to describe to the last detail what an individual member will have to do, the latter may just constrain our options in a more general sense.Footnote 15
Second, since associations often have specific purposes, we need to demand only that individual members put we-intentions in charge over those decisions and actions that are relevant to those purposes. So group members do not altogether lose their sense of individual agency. The point is merely that we think in the mode of the group where appropriate. While we think as citizens, members of staff, and so forth, on particular occasions, in numerous other contexts we can continue to think as individual agents. What we have reason to do, then, depends on the standpoint that we take. We can either conceive of ourselves as members of a particular group or as individuals; and Dworkin has detailed this.Footnote 16
Third, we do not need to impose external restrictions on the content of collective intentions and preferences. In rational choice theory, individual preferences are generally taken as given. They provide a ranking of action-outcomes, capable of guiding the agent towards one action or another. As has often been noted before, the ranking is not itself the subject of further normative evaluation. Collective preferences are similar in the sense that they provide the agent merely with instrumental reasons.Footnote 17
Once this two-tiered structure is in place, the question is not whether John can logically intend to play while he has no discretion over the actions of Paul, George and Ringo, but whether or not they all share this collective intention and are prepared to accept its authority. Here I use authority with reference to my earlier remarks concerning practice theory. Korsgaard makes the case from a Kantian stance that action needs to rest on a justification that has the form of a law. That is to say that we are all persons, and thus capable of action, only if our higher-order volitions have the property that they could be willed as a universal law, not just governing our own action but those of all agents. Moreover, that ‘… from a practical point of view our relationship to our actions and choices is essentially authorial’.Footnote 18 It is important at this point to emphasise the distinction between a shared intention and a collective intention. It is perfectly possible for one person to have a shared intention. For example, John might have a we-intention to form a band with Paul, George and Ringo, while the latter do not. In this case, John believes or predicts, mistakenly, as it turns out, that they are all agreed on forming a band and that they are all agreed on a certain plan, delegating more specific tasks to each of them. Until he finds out about this mistake, John will think he had a reason to proceed to carry out what the plan instructs him to do.Footnote 19 But it is still wrong to suggest that in this example there is a group agent. An association will exist only if Paul, George and Ringo share this intention. And an intention is shared if two or more people both intend to undertake a certain activity together. As Bratman puts it: ‘It takes at least two not only to tango but even for there to be a shared intention to tango.’Footnote 20 So the concept of shared intention refers not only to the intentions or preferences of a single individual, but to a state of affairs in which the intentions and attitudes of several individuals interlock in a purposive way.
In the literature, this relation is conventionally described by a number of jointly necessary and sufficient conditions, which are the subject of ongoing debate and many subtle differences. To keep things simple, I want to introduce Bratman's account as a working hypothesis.Footnote 21 Roughly, his model consists of the following conditions. We share an intention if:
(1) You and I intend that we J;
(2) You and I intend that we J in accordance with sufficiently meshing plans;
(3) I intend that we J partly because you and I intend so, you intend that we J partly because you and I intend so;
(4) It is common knowledge between us that the above conditions apply.Footnote 22
The first condition requires that John must generally intend to play and Paul must intend to play, and so forth. By the same token, what Sugden calls a ‘team preference’ requires that two or more persons have the same collective preference, or what he refers to as a ‘team-directed preference’.Footnote 23
The second condition specifies that the parties also need sufficiently compatible plans. While it is not necessary that co-ordinating plans are fitted out to the smallest details, they should overlap at a more abstract level such that outstanding issues could be resolved in further rounds of bargaining or deliberation.
More controversially, the third condition demands a degree of respect for the individual agency of the parties involved. Using one of Bratman's examples, if Sonny handcuffs Donny and transports him to New York in the boot of his car, then they do not share an intention to go to New York together. Even though their actions are co-ordinated, it is not at all evident that Donny intends that they go to New York together. As Bratman puts it, this is ‘the Mafia sense’ of having a sharing an intention. I shall refer to this condition as the agency requirement.
The fourth condition stipulates that it is common knowledge or, weaker, a common belief between the parties that the above conditions apply. This means, as Sugden puts it, that each member must have ‘full team confidence’.Footnote 24 That is to say, each individual must be confident that the other members intend to J, and each individual is confident that the other members are similarly confident, and so on indefinitely.Footnote 25
Now, something like this forms the conceptual core of corporate personality. We can begin to understand how small groups and associations, corporations and states could intend and act collectively without abandoning reductionism. To see why corporate personality is no less credible than individual personality, it may be helpful to compare the following two cases. First, let us say that you and I intend to write a book together. You and I define an intention and a preference over something we want to do together. To achieve this purpose, we need a plan for the co-ordination of our individual efforts. It is my job, say, to write the first three chapters, while it is your job to write the last three. I take care of finding a publisher; you take care of certain other things. We also need to think in the mode of a group. That is, we need to evaluate our choices and actions in terms of our joint purpose. If our subsidiary intentions interlock in a coherent manner and we have ‘team-confidence’, then you and I share an intention in the above sense.
Compare this to the case where I intend to write this book on my own. In order to succeed, it is obviously not sufficient to just form an intention to start with the first section starting at t. Starting the whole project also depends on my having an intention to write the next sections of the book at t1, t2 … tn. What is required, then, is not just an intention to start writing now but another sort of intention covering my future actions over that entire period of time.Footnote 26 So I will have to intend at t to write further sections at t1, and for however long I think it is going to take to finish.
Now, the crucial point is that having a future-directed intention like this is the equivalent of having a collective intention. To begin with, it is bogged down by the same discontinuity between intention and action. I presently need to define an intention over what I will do in the future. Yet I know that presently I have no discretion over my future actions. My intention at t has just as little discretion over what I do at t1 as my intention that you write Chapter 4. In other words, my intention to write a book myself is defined over what we, my future person-stages and I, are up to together.
Furthermore, and also similar to collective intention, the success of my project therefore depends on the question whether or not my future person-stages will take this intention as authoritative. At any rate, having a future-directed intention necessarily involves a co-ordinating scheme allocating specific tasks over the weeks, months and years. If I begin with Chapter 1 this week, I will move on to the next within two months, and so on. To put it differently, my future selves and I have to think in a style analogous to the group mode of agency. Thinking in this mode, I ultimately take my future-directed intention to be in charge over my present decisions. So, my person-stages are guided by a purpose or a preference not just defined over what I could presently do, but over the outcomes of a temporally sustained sequence of actions, over what I could achieve over a period of time. Again, it is this two-tiered model that explains my agency.
Notice also that my project will succeed not only if my future person-stages intend to write this book together, but also on the condition that our subsidiary intentions are sufficiently adjusted to each other. There is, of course, some discretionary space for each person-stage. And my person-stages will probably also entertain other future-directed projects. But there must be a minimum of agreement on the co-ordinating scheme that distributes tasks over the coming period. Thus, my intention to write the book myself has all the hallmarks of shared intention, reductively understood. That intention is, in fact, a temporal bundle of interlocking intentions of several person-stages.Footnote 27
Recall that I claimed that the personal identity of individuals should be understood entirely in terms of a rational bundling of purposive mind-states which I also referred to as integrity. Integrity essentially means that we guide our actions according to a general policy. This policy must be general in the sense that it is guiding not only in the context of one concrete and temporally discrete decision, but also in a whole range of decision moments. In addition, it should be stressed that integrity is not just a kind of consistency or stability between temporal stages that, understood in terms of a system of rules or laws, allows us to predict behaviour. Integrity also involves a mode of thinking according to which this system of rules and laws is a normative standard for the appraisal of our attitudes and behaviour. So this system of rules is authoritative in the sense that it provides us with reasons for action.
The above notion of shared intention precisely to such a bundling, the only difference being that this bundle does not consist of the temporally discrete mind-states of one individual, but of those of several individuals. As an individual but temporally extended agent, future-directed precepts like plans and intentions are put in charge of my choices and actions. In the context of corporate agency, integrity assumes the shape of collective intentions and preferences, that is, policies, plans and principles defined over the actions of a number of people and the outcomes thereof.Footnote 28
II. Sovereignty constraints
The notion of shared intention, I want to claim, takes care of the reductionist constraint. We do not, at least, slip in metaphysical theories postulating a spooky mind-substance of the group. Yet a number of serious challenges still remain. In this section, I argue that Bratman's model – his agency condition in particular – needs to be read in a specific way to meet what I referred to as the sovereignty constraint. Bratman's model of a shared intention is open to attack from the opposite end. That is, it may collapse back into a form of ethical individualism, thus rendering the idea of corporate personality ultimately incapable of doing any real and independent work. To prevent such a collapse, the model needs to be read and fleshed out in a particular way.
2.1. Conventionalism
To begin with, shared intention must be carefully distinguished from what could be described as conventionalism.Footnote 29 This view holds that the notion of we-intention could ultimately be broken down in standard I-intentions. The model discussed above argues that collective intentions are basic. However, a collective intention becomes a shared intention only if a we-intention is jointly held, and certain supporting attitudes are interlocking in the right way.Footnote 30 In contrast, conventionalism argues that the idea of a collective intention itself consists in an interlocking network of I-intentions. Therefore, it rejects the entire idea that we first have to think in the mode of a group. It explains collective action on the basis of the normativity of individual intentions and preferences alone.
Consider again the above example. On Bratman's model, John and Paul have a we-intention to undertake something together. In addition, their attitudes converge according to a plan that co-ordinates their individual efforts, they respect each other's agency, and so on. Instead, the conventionalist picture suggests that John I-intends to sing alone, he further believes that Paul I-intends to play guitar, and that all this is common knowledge between them. What is more, Paul does in fact I-intend to play guitar, and he also believes that John intends to sing, and that this is common knowledge between them. In other words, it suggests that there is here a kind of convergence between their I-intentions and certain supporting beliefs so that they could successfully co-ordinate their efforts. As I understand it, conventionalism provides a model for social co-ordination and co-operation without having to resort to the idea of collective intentions or preferences.
Is conventionalism plausible? This is of course a key question. Fortunately, we do not have to settle it right now. Here, it is important just to emphasise the difference between the two models. The idea that John and Paul share an intention remains fundamentally at odds with the conventionalist picture.Footnote 31 Essentially, the notion of collective intention as a basic entails taking a different kind of standpoint. It implies that individuals think as a group and guide their deliberations on the basis of collective intentions, preferences and purposes. The conventionalist account of we-intention cannot in the end be differentiated from ethical individualism. It supports the standard view that we guide our deliberations on the basis of individual intentions, preferences and purposes.
Nothing I have said so far actually implies that this notion of corporate personality also provides a more plausible explanation than the conventionalist model for a certain range of events. As far as I can see, the position taken here is still neutral to that question. But if we are serious about the idea of corporate personality, I suggest, we will have to commit to the view that we-intentions are primitive, even though collective agency requires an interlocking of further attitudes. Of course, it must ultimately be demonstrable that this difference has implications; and this has been attempted.Footnote 32 However, for now the point is to show that collective intentions are sui generis. In concluding this section I want to underscore the necessity, for my purposes, of rejecting conventionalism. My approach argues for a model of corporate personality. My model of corporate personality holds that collective intentions are basic and that whether or not intentions or preferences are shared is an empirical matter; and that while voluntary consent may be a sufficient condition, it is not a necessary condition. Whereas conventionalism argues that collective intentions and preferences could ultimately be broken down to individual intentions and preferences.
2.2. Contractualism
I return to the concept of shared intention. As set out above, this concept is vulnerable to a further problem. The objection results from the interpretative discretion that the agency requirement, as set out above, appears to leave. How we understand this condition turns out to have pivotal significance, and I shall be occupied with it in the rest of this article.
One possible and strongly normative reading of the requirement demands that shared intentions have a contractualist basis. Essentially, this reading attempts to justify or rationalise collective intentions or preferences in terms of individual intentions or preferences. In other words, a shared intention must fulfil the normative requirement that for all members it be ‘reasonable’ to join the collective endeavour. And ‘reasonable’ is likely to mean here that it must pay off individually to start reasoning in the group mode. Natural candidates for these normative requirements are the Pareto principle and, more restrictively, Rawlsian or other egalitarian principles of distributive justice.
The problem with the contractual view, I want to argue, is that it rules out too much.Footnote 33 Of course, I see that when such conditions are not met, suspicions may be raised as to whether or not the participant's agency is sufficiently respected. But it goes too far, in my view, to rule out the possibility that people can willingly have a we-intention even if the shared activity does not pay off in terms of their own individual preferences and interests. While it may well be that such conditions set out an external normative parameter for co-operative action, it is not persuasive to claim that individuals could not think in the mode of a group if there were good reason to do so from the individual point of view. To put it differently, I want to suggest that collective intentions are not normatively entrenched I-intentions. Whether or not an agent has a we-intention is in my view ultimately an empirical question, no less an empirical question than whether or not agents have a certain individual intention or preference. Thus, what we need is the presence of an intention to undertake something together. This presence is a reason for action in and of itself, if we follow Sugden. Sugden argues we do not need a further normative test to the effect that by the lights of some criterion it must also be reasonable to adopt a collective intention or preference.Footnote 34 Moreover, we might follow Gold and Sugden who afford us a theory of personal action which takes into account collective preferences. In the Gold and Sugden model individuals rationalise their actions in accord with collective preferences (i.e. team preferences) as opposed to their own preferences per se. Michael Bacharach has argued something similar.Footnote 35
However, let me try to make the point against the contractual view in a different way. Contractualism, I have just said, is grounded in the idea that in order to respect the agency of the parties involved, collective intentions must be reasonably acceptable. And ‘reasonably acceptable’ is hammered out in terms of individual intentions and preferences or rather in terms of individual gains and payoffs. But why should this be true? Perhaps the most promising argument trades on a perceived distinction between shared and collective intentions on the one hand and I-intentions on the other. The idea, it seems, is that the former intentions do not count as a form of agency, while the latter do, because we-intentions require us to think in a modus operandi fundamentally estranged from ourselves.
I believe this argument to be flawed. Take again the case where I intend to write a book on my own. As we saw, I do not just form an intention to start with the first few words. Instead I also intend that my future selves write later paragraphs and chapters. Therefore, my intention is defined, not over immediate action, but over temporally sustained action. Notice that my future selves have to think in a mode of thought in which this future-directed intention is taken to be reason-giving. This explains why we think of our lives as continuous wholes rather than unconnected sums of infinitesimally brief moments. To put it differently, we have preferences and goals pertaining to how our whole life goes, or at least large chunks of it. And in considering what we ought to do, we abstract from our present interests in order to also look at the good of our life as a whole.
The problem, then, is that we could also apply the contractualist thesis to individual personality.Footnote 36 Contractualism would demand that our future-directed intentions be reasonably acceptable from the point of view of each person-stage. My future selves would all have to be able to share in the gain or receive their fair share under some criterion of justice. But this is absurd. We do not subject ourselves to such a restricting regime. If I presently decide against going to the theatre, while I would desperately like to, and miserably continue to work with the ulterior purpose of completing the task at hand, there is no one who would seriously doubt my free agency. We do not say that this could not reasonably be an intention or an end of mine because my present self is worse off than he would have been without the intention. On the contrary, we often take this to be an exemplary case of rational agency that needs to be singled out for praise. There is, of course, no insinuation of unfairness. I just happen to have this long-term goal, and if I endorse that goal, that is pretty much the final word. That goal provides me with a reason not to go to the theatre.
It is important to note that I am not arguing against some distributive policy according to which, say, future-directed action must not disturb a certain allocation of wellbeing over my entire life span as a measure of the good life. I may well have good reasons to prefer to live the egalitarian life, a life without having to suffer such miserable moments, but perhaps also a life without having written a book. Distributive considerations of this kind, I suspect, do play a role in many of our deliberations about the good life. We set limits to how much sacrifice we presently want to endure in order to realise long-term goals, even though these limits seem a good deal less restrictive than Rawls's difference principle or even the Pareto principle.Footnote 37 But here my argument is not with these principles. It may not be absurd to impose distributive limits on how we divide wellbeing over time. What is absurd, however, is the suggestion that these limits are normatively powerful enough to rule out the conceptual possibility of my reasonably and willingly adopting a future-directed intention that requires a present sacrifice for a future good. This suggestion is absurd because such normative principles are external to agency. Whatever use they have, they are not a prerequisite for agency to emerge.
In resisting the conceptual reduction of we-intentions to I-intentions, I am pressing the exact same claim. I am not suggesting that there are no external grounds for adopting some distributive criterion such that states or other co-operative schemes can be ethically appraised. Perhaps it is true in a particular context that some egalitarian distribution is what justice demands. I have, at any rate, no real quarrel with such a norm here. But I do have a problem with its specifically contractual grounding in agency. It is simply implausible to claim that it is impossible for someone to reasonably and willingly agree to participate in a joint venture that violates this criterion and, say, leaves him worse off than before. Contractualism effectively argues that such a joint venture cannot be grounded in a person's will and, therefore, that there cannot be a shared intention in these cases. Now, it may not be mistaken to embed shared intentions in an individual's volition and agency. What is mistaken, however, is the presumption that only individual intentions could count as constitutive of people's volition and agency. This presumption takes the distributive requirement for an internal requirement constitutive of agency itself.
On the account I am suggesting here, we-intentions and collective preferences are no less expressive of a person's will, no less normative than I-intentions. If John and Paul come to hold the collective preference that they play music together, then they rationally ought to do so, even if this means that they are worse off individually as a result. In this respect, I can really see no reason why we ought to treat we-intention or shared intention differently from future-directed I-intention. We do not generally think that future-directed I-intentions, like the one in my example above, deflect us from our agency simply because they require us to think in a particular mode of thought. We do not think that this kind of intention must be justifiable from the perspective of present-directed volition in order to count as a true expression of our agency. In fact, when our future-directed intentions are governed by the present-directed viewpoint we are vulnerable to what we commonly describe as weakness of will. Why should we think differently in the case of we-intentions?
I should add an important caveat. The very notion of the weakness of will clearly implies that our temporally discrete desires, far from being expressive of our free agency, actually threaten to undermine it. Our agency, according to this view, is always located in future-directed intention. In other words, by overriding the desire of a person-stage, we fulfil rather than impede our freedom.Footnote 38 That said we should be careful not to press the analogy with collective intention to its ultimate conclusion. If we were to claim that collective intention is always more expressive of our agency than individual intention, then we would end up in the position that Rousseau may have set out in The Social Contract, and which has provoked the wrath of classical liberalism ever since, namely that elements resisting the General Will could be forced to be free.Footnote 39 Instead, we should emphasise the pluralist consequences of the argument. The main point I am trying to make is not that collective intention is ‘more real’ than individual intention; it is merely that it is conceivable that a collective intention is expressive of our agency. Individualists assume that future-directed intention is expressive of our agency. Individualists assume future-directed intentions are somehow privileged, that it is somehow self-evidently true that these intentions correspond to our agency, and that our temporal selves can thus indeed be forced to be free. But there is in my view no ironclad argument to back up this assumption. As Parfit demonstrates, there is no logical reason why we could not argue for the autonomy or the rights of temporal selves, and deplore the repression that ethical individualism necessarily involves.Footnote 40 More to the point, there is no logical reason to presuppose at the outset that collective intentions could never be truly expressive of the will, and should always be subjected to the contractualist test. What this leaves us with is not the Rousseaunian position, which may indeed lead to the Gulag, but rather the view that the unit of human agency is more flexible than individualists normally assume, and could in principle take on several forms. We do not rub out the practical importance of the individual and conflicts between individual interests in favour of the interests of a collective person. But what we have to affirm is that individual and corporate personality may both have a role to play in practical reason. Moreover, the question which particular form our agency should assume is in itself the subject of normative inquiry.Footnote 41
To sum up, we must be careful, at this stage, not to place conditions on corporate agency that prove to be too strict and are, also, not properly applied to individual agency. Certainly, some conditions of agency need to be met if intentions are to be really ‘shared’. But contractualism goes wrong in fixating on a norm that insofar as I can see is external to corporate personality.
2.3. Voluntarism and obligations of shared intention
While I reject the contractualist interpretation of shared intention, I have not rejected the agency requirement as such, namely that such intentions must respect the volition and agency of the individuals involved. I merely want to keep it separate from distributive norms testing the reasonableness of collective intentions and preferences.
If the contractualist view is rejected, it might be thought alternatively that shared intention requires something like an explicit agreement, a promise or an actual contract between the parties involved. This thought is appealing because an actual contract would seem to satisfy all the conditions of shared intention set out above. The problem, however, with this view is that it involves the danger of excluding associations that do not obviously rest on such an agreement, yet should still be counted as corporate persons. The state is an obvious case in point. It is widely agreed that membership does not rest on explicit consent or an actual contract.Footnote 42 Hence, if we want to rule in non-voluntary associations like states, the actual contract view must be dismissed as too restrictive.
I suggest the following argument. For a shared intention we need two or more persons possessing a collective intention, sufficiently meshing subplans, and so forth, as specified above. Now express consent, or a contract or a promise, are merely manifestations of the fact that such intentions are present. To put it differently, agreement is not itself part of the criterion for shared intentions, but merely one indicator among a host of possible indicators revealing the prevalence of the we-attitudes in question. If John and Paul come to an agreement, sign the relevant papers, and so forth, this may be ample to satisfy the conditions of sharing an intention. However, while it is one thing to say that the presence of an explicit agreement is sufficient evidence for the presence of a shared intention, it is something else to claim that this is also a necessary condition for shared intention to emerge.
The latter half of this claim is false because it is possible for people to share an intention without explicit agreement or without making a promise. Bratman approvingly refers to Hume's conventionalist conception of justice: ‘Two men, who pull the oars of a boat, do it by an agreement or convention, tho’ they have never given promises to each other.'Footnote 43 Hume's rowers have the collective intention to row the boat together. They agree on where they are going, at what speed, and so on. In other words, they have appropriately interlocking attitudes. They are in fact prepared to row the boat. It might be claimed that their express consent is required as a means to signal to each other that they are prepared to row the boat. To be sure, consent and promises may help to facilitate mutual knowledge. But, again, it is a mistake to claim that they form a necessary means. What is necessary is that both men are confident that the other has the appropriate collective intention and supporting attitudes. However, this confidence may also rest on a different kind of evidence. Since both men are in fact rowing the boat, their collective preference is revealed by their continued action. The rowers may simply base their expectations on stable patterns of behaviour. Shared intentions, then, may be different from conventions in that they could emerge in equally spontaneous and informal ways.Footnote 44
The express consent interpretation, in short, is unnecessarily stringent. What is more important, it does not help to alleviate the concern with agency of the parties involved. Essentially, the problem is that shared intentions could be elicited under duress. Recall the ‘Mafia sense of going to New York together’. Here, it is important to distinguish two versions of this general case. Let ‘Case A’ refer to the scenario where Sonny wrestles Donny into the boot of his car and drives to New York. Clearly, Donny is brought to New York entirely against his will. Since the latter does not intend to go to New York at all, it is evident that there cannot be a shared intention either.
However, consider ‘Case B’. This time Sonny does not employ his brute strength, but threatens Donny with a gun. So Donny faces a choice. It may be a harsh choice, perhaps, but he can still decide to accompany Sonny to New York or get shot in Miami. Now, if he decides to join Sonny, it may be plausibly claimed that Donny does indeed form a collective intention to go to New York together, their intentions may also be sufficiently meshing, and it is not unlikely that this is common knowledge. Raz has offered a discussion of this sort of issue.Footnote 45
The key question, however, is whether or not Donny's agency is respected such that this collective intention could be shared. According to one view, it matters that Donny has a choice, makes up his mind and intends that they go to New York together. Albeit under duress, there is a shared intention. Call this the harsh view. According to another view, Donny's agency has not been respected. His collective intention is declared void precisely because it has been obtained under duress. Call this the no-real-choice view.
Notice first that the demand for express consent or a promise does not coincide with this latter view. After all, Sonny might even make Donny sign a contractual statement to the effect that he obligates himself to go to New York. In order to share an intention on the no-real-choice view we need voluntary consent. Whether consent is explicit or implicit is less important. What matters is that collective intentions are arrived at without any threats.
The real question, then, is whether or not we have to subscribe to this interpretation. In providing an answer, much seems to depend on the further question whether or not shared intentions involve mutual obligations (and entitlements). Bratman and Sugden propose a model on which shared intentions and preferences do not obligate the parties involved.Footnote 46 For example, it may create mutual expectations that we are obligated to meet. However, these obligations are not internal to the idea of a shared intention itself. You and I could share an intention without acquiring any obligation whatsoever. If this claim were correct, then the harsh view would be a good deal less harsh. As I understand them, Bratman and Sugden do indeed hold that in Case B there is a shared intention. But this result does not imply that Sonny has an obligation to do his share, at least not in virtue of sharing an intention.
Gilbert, on the other hand, advances a model on which such mutual obligations and entitlements are conceptually constitutive of sharing an intention – what she prefers to call a ‘joint commitment’ or ‘joint decision’.Footnote 47 This idea is to operate in the spirit of the obligations of actual contracts and promises. Yet she also supports the harsh view. According to Gilbert, if Donny decides to go to New York in Case B, this becomes a ‘joint decision’ of theirs. Typically, joint decisions cannot be rescinded unilaterally. Therefore, if Sonny and Donny share an intention, this also generates an obligation on Donny's part to see their joint commitment through.Footnote 48
Now, Gilbert's view that obligations are in part constitutive of shared intention seems correct to me. Elster has argued that the personal identity of individual agents over time should be defined in terms of a normative and obligation-creating connectedness or integrity.Footnote 49 Since I now invoke the same notion of integrity in order to advance the possibility of corporate personality, shared intention has to be given a similarly normative foundation. However, intuitively it seems that something goes horribly wrong on this view. Surely, we would not seriously want to contend that Donny has a genuine obligation to accompany Sonny to New York. Either they do not share an intention at all, or they do not have Gilbert's obligations and entitlements. Thus, we are faced with a dilemma. If we want to retain the view that obligations are constitutive of shared intention, there seems to be a strong case for the no-real-choice view. If we want to maintain the harsh view, then we have to drop the inherent obligations and entitlements. Something has to give.
Gilbert suggests an interesting way to slip between the horns of this dilemma. On her account, Case B contains both a shared intention as well as mutual obligations. Her solution is to distinguish between having an obligation and having a morally binding obligation. The key assumption of her argument is that practical reason is pluralist in nature. On this particular view, Donny's obligation constitutes just one reason for action. Apart from his obligation of joint commitment to Sonny, he may have obligations of fidelity or reciprocity to other people. He may have certain general obligations of justice or benevolence. And apart from other-regarding concerns, he may also have prudential reasons. Gilbert's point, I think, is that it is perfectly possible to say that Donny is under an obligation to go to New York, all things considered. Whether he ought to go in the final analysis depends on the relative weight of this obligation vis-à-vis other reasons. To put it differently, Donny has a genuine obligation, but we could claim that this obligation is trumped by more significant reasons.
It sometimes looks as if Gilbert regards the collective weight or the strength of obligations of joint commitment as a constant factor. Whether or not they are outweighed by other reasons does not appear to depend on the nature of the circumstances under which joint commitments are made. But I think it would be more accurate to suggest that this relative weight is a function, in part, of precisely these circumstances. Viewing the relative weights as such a function, something like the following picture results. The analysis of Case A remains unchanged. There clearly is no shared intention – Donny's agency is simply not effective – and consequently there are no obligations either. In Case B, Donny has a genuine obligation to Sonny in virtue of the fact that they share an intention. However, the relative weight of his obligation is significantly reduced. This is a genuine obligation, even in this case, but the weight that this obligation can put in the balance in comparison with competing obligations is relatively weak.
When I say relatively weak, I mean weaker than in ‘Case C’. In this case Donny proposes to go to New York, with Sonny, of his own accord. It is more likely, on the balance of things, that Donny ought to go to New York in Case B than in Case C, or so I argue. By the same token, it is also less likely, on the balance of things, that Sonny has reason to feel aggrieved if Donny ends up not going to New York in Case B than in Case C. In the latter case Donny may have stronger reasons to stay in Miami, say, because of a previous commitment to his pregnant wife. But it seems altogether more plausible that Sonny can rightfully censure Donny for breaking his commitment than in Case B. If Donny goes back on his commitment to Sonny he needs a stronger reason to trump his obligation to Sonny.
In this way we can do justice to the central intuition that something goes wrong in this view of associative obligations. Our intuitions, of course, are that such obligations and entitlements must at least be sensitive to coercion, pressure and threats. If I am right about Gilbert's view, her solution is not really responsive to these circumstances. She merely argues that such obligations could be trumped by other reasons, all things taken into consideration. But presumably there can be such trumping even in cases where there is no duress. No amount of trumping does justice to the central intuition, unless we can also show that trumping is in fact more likely to occur in cases where there is duress than in cases where there is not. In contrast to this, I argue for the view that while threats may not snuff out our choice completely, they do constitute a more or less serious impairment to our agency. On this approach, then, there is scope for the view that obligations are inherent in shared intention only insofar as our agency remains intact. (In parenthesis, it might be asked why, on this line of reasoning, the state, a non-voluntary association, has any authority over us. The overall argument I am propounding relates to corporate personality. First, I hold that by framing corporate personality in terms of shared and collective intentions we remain faithful to reductionism about persons and personal identity. Second, by pressing an interpretation on which shared intention remains independent from ethical individualism, we have opened up a conceptual space for corporate personality to do normative work. Third, by injecting interpretative and institutional elements into the basic model, we can now see how corporate personality could be applied to larger and more complex associations, including the state.)
Alternatively, the no-real-choice view aims to capture the same intuition that shared intentions, like agreements and promises, cannot be extracted under duress. If some credible threat is issued, or so the view goes, the shared intention is void and, consequently, there is no scope for obligations of joint commitment either. What commends my approach over this view? One reason, I think, is that it does not seem very plausible to argue that joint commitment, and therefore obligation, becomes entirely void as soon as a threat is issued or some degree of pressure exacted. Clearly, agreements and joint commitments are reached under the sway of threats and pressures all the time. Furthermore, what we have to bear in mind is that there are many kinds of threats and pressures, differing in degree and intensity. For example, instead of reaching for his gun, Sonny may threaten to blow the whistle on Donny's illegal nightclub, he may threaten to pull out of some potentially lucrative business deal, in fact, he may just tell Donny he will no longer be welcome at his next birthday party. There is, I want to suggest, a fairly elastic difference between the issuing of threats on the one hand and what we would normally refer to as ‘bargaining power’ or ‘cajoling’ on the other. It is not entirely vacuous to suggest that these are genuine commitments that also generate genuine obligations.
The no-real-choice view could try to remedy this by capturing the complexity involved in what it means by real choice. But it would still need to propose a problematic cut-off point between joint commitment reached under the sway of legitimate cajoling and joint commitment under illicit duress. The refined view would declare a joint commitment void once agency is not sufficiently respected. At that point, concomitant obligations and entitlements simply cease to exist. There may or may not be such a cut-off point, but the point is that it is difficult to see how there could be degrees in the ‘sharedness’ of an intention. We either share an intention or not. Similarly, on the no-real-choice view, either these obligations exist in their full glory or they do not exist at all. In contrast, my suggestion is that shared intention persists in the face of any form of pressure (with the exception of Case A). But the restriction placed on agency by this coercion is discounted in the relative weight of the resultant obligations.
However, the difference between the no-real-choice view and the position that I just set out may seem minor. Ultimately, both views make obligations of shared intention sensitive to coercion, pressure and threats. Therefore, we cannot claim that this argument answers any outstanding issues, for example, the question of political obligation, with any finality. Nevertheless, it still produces more than a merely semantic result. Two points stick out. To begin with, if we take the no-real-choice view, then we have to accept that non-voluntary associations like states may not qualify as corporate persons, implying that corresponding obligations of shared intention could not exist either. However, there is a difference between saying that associative obligations could not exist and saying that they may be outweighed by other reasons.
Second, and more important to my purposes here, this way of approaching the issue allows us to enjoy two benefits. On the one hand, we can treat the issue as to whether a corporate person exists separately from the question whether or not interlocking of attitudes has emerged under duress. This is the key advantage of the harsh view that Bratman defends. On the other hand, it still advances Gilbert's view that obligations are constitutive of shared intention, without necessarily embracing counter-intuitive results. To put it differently, we can still claim that shared intention is a normative relationship comparable to personal integrity. Of course, once the existence issue has been settled, we must still decide on the weight of associative obligations of shared intention. The presence or absence of duress is clearly one variable that has to be taken into consideration. And in terms of political obligation, this variable may have a rather limited justificatory power.
Conclusion
It can be argued that the identity of persons over time should be understood entirely in terms of a bundling of mind-states without reference to a physical or spiritual substratum, that this bundling should be in part defined in normative terms of integrity. This conception of personal is itself irreducibly practical.Footnote 50 If such an argument is correct, then we face the following two challenges to get the idea of a corporate person up and running. First, we have to present an understanding of human association that corresponds to the idea of integrity. Second, because the model depends on its normative impact, we also need to show that this understanding of human association has practical benefits.
In this article, I have tackled the first challenge. I wanted to show that the notion of integrity could be applied, not just to the temporal stages of a life, but also to the relatedness between individuals. To this purpose, I introduced the model of collective and shared intention. Essentially, this model explains how we could say that an association could have intentions without resorting to implausible versions of corporate realism, which postulate a group being separate from its membership. The model of shared intention relies on the view of intention as an interlocking set of collective intentions and preferences, and other supporting attitudes.
In order to prevent shared intention from collapsing into a mode of ethical individualism, I rejected a number of interpretations. Contrary to conventionalism, corporate personality argues that collective intentions are primitive. That is, collective intentions are not to be broken down and analysed in terms of individual intentions. Against contractualism, corporate personality holds that collective intentions are not subjected to a contractualist test. Collective intentions do not have to be reasonably acceptable to an association's members. Finally, in opposition to voluntarism, corporate personality maintains that shared intentions could exist under duress, and that obligations and other normative concerns are constitutive of shared intentions.
