Corporate Responsibility Revisited

It is a privilege to have my work taken so seriously, and addressed to such a challenging effect, in these four commentaries on “Responsibility Incorporated” and related material.¹xPhilip Pettit, “Responsibility Incorporated”, in this issue. I am most grateful to the authors of those pieces for the time and attention they have given me and to the editors, Hans Lindahl and Erik Claes, for organizing the exchange. In this response I set out my views on the topics addressed in a more or less systematic way, restating, revising and expanding on existing work, and I comment as appropriate on the multiple contributions and criticisms made in the papers. I hope that this will make for a more coherent, free-standing discussion than an author-by-author set of replies.
The paper is in three sections. In a first, background section, I consider a set of issues related primarily to responsibility at the individual level. In section 2, I look again at responsibility at the corporate level. And in section 3 I consider the connection of this discussion to issues of responsibility in law and politics. The paper by Bert van den Brink and the paper by Govert den Hartogh figure mainly in section 2, the paper by Bert van Roermund and Jan Vranken and the paper by Ronald Tinnevelt in section 3.

• 1 Responsibility at the Individual Level

  In order to be fit to be held morally responsible an agent has to face a choice between options over which it makes sense to deliberate. This means that the person has to be able to think of each option, taken under an appropriate description: I can do that; whether or not I do it is up to me. And if the person is right to deliberate, of course, then that thought has to be correct. The agent must have a can-do ability in respect of the different options and must be aware of having that ability.
  But we will not always hold a person responsible for a choice made when this condition is fulfilled. We won’t do so, of course, if the choice was between options that did not differ in evaluative significance; the decision may have been utterly trivial, as in picking one box of soap powder at the store rather than another identical box. More importantly, we won’t do so if the agent, as we think, was not in a position to know of the rival merits of the options available; the person, as we say, was blamelessly ignorant, not being at epistemic fault for the failure to register those merits.²xThere is no contrast on this matter with the law which, as Van Roermund and Vranken note, “takes a normative view (…) an agent ought to be aware of certain risks”. But even if the agent was in a position to know of those merits, or perhaps actually registered them, we may still abstain from ascribing responsibility. We may think that the agent did not have the capacity, or at least not the full capacity, to let a recognition of those merits impact on the decision made.
  These comments reiterate the three conditions for being fit to be held responsible that I distinguish in “Responsibility Incorporated”: the choice must be value-relevant, the agent must be able to make the appropriate value judgment, and the agent must be value-sensitive, having the ability to act on the judgment made. The value-judgment or the value-sensitivity condition may fail utterly, in which case the agent will be exempted from blame for doing something bad, or either may fail in some partial measure, in which case the agent can appeal to excuses in mitigation of the act.³xSee John Gardner, Offences and Defences: Selected Essays in the Philosophy of Criminal Law (Oxford: Oxford University Press, 2007). Excuses may show that it would have been hard, though not impossible, for the agent to have known about the relative merits of the options. Or they may show, going to the value-sensitivity condition, that it would have been hard, if not impossible, for the agent to have resisted certain impulses or temptations, or to have deliberated properly under such and such conditions of duress or coercion or whatever.
  These observations indicate that the can-do ability that any agent must display and register in a choice is not sufficient for fitness to be held responsible.⁴xI am grateful on this issue for points made by Gideon Rosen in discussion. Such fitness presupposes the richer capacity, as we might put it, to track the evaluative considerations relevant to the choice: to recognize the considerations that apply there and to respond as they require.⁵xSee Philip Pettit, “The Elements of Responsibility”, Philosophical Books 46 (2005): 210-19; Philip Pettit and Michael Smith, “Freedom in Belief and Desire”, Journal of Philosophy 93 (1996): 429-49. This capacity to track the reasons – this “discursive control”⁶xI use this phrase in particular, as Van Roermund and Vranken point out, in Philip Pettit, A Theory of Freedom: From the Psychology to the Politics of Agency (Cambridge and New York: Polity and Oxford University Press, 2001). – is at the core of what it means to be fit to be held responsible and, clearly, it is a capacity that comes in degrees. Various departures from full satisfaction of the value-judgment and value-sensitivity conditions exemplify ways in which it can be graduated in this way.
  The capacity that fitness to be held responsible for any action presupposes, then, is not the capacity to have done otherwise in that particular choice, in the contra-causal sense of that phrase.⁷xFor analysis of an unproblematic sense in which the sentence can be true, see Philip Pettit, “The Capacity to Have Done Otherwise”, in Peter Cane and John Gardner (eds.), Relating to Responsibility: Essays in Honour of Tony Honore on his 80th Birthday (Oxford: Hart, 2001), 21-35; reprinted in Philip Pettit, Rules, Reasons, and Norms (Oxford: Oxford University Press, 2002). It is not the capacity, back at the exact moment of choice, and notwithstanding the causal pressures that came on stream at that point, to have made a different choice.⁸xRonald Tinnevelt rightly says that “one can’t hold actors responsible for things that lie beyond their control” but it is important to understand how control is understood here. See Ronald Tinnevelt, “Collective Responsibility, National Peoples, and the International Order”, in this issue. You need not have had that capacity; indeed it does not make sense under naturalistic assumptions – deterministic or even indeterministic – to think that you could have had it. What must be true is only something that is not so naturalistically problematic: that you acted in the presence of a capacity – ideally, a full capacity – to track the relevant reasons in that sort of choice. It is possible to act in the presence of such a capacity, without actually manifesting it, and that is what happens when you do something bad and are fit to be held responsible for it.
  Why do I stress this? Mainly to emphasize a point mentioned in my original paper but perhaps insufficiently stressed. This is that there is bound to be a grey area between those choices in which you are fully, even partially, fit to be held responsible and those choices in which you are not. The ascription of capacities is not subject to a tight discipline in any area, even the most mundane. Does the horse that failed to run a mile in a minute have the capacity to do so and just fail to manifest the capacity? Does the horse that actually runs the mile in a minute really have the capacity apparently manifested – or is its performance just a fluke? Such issues arise in every area where we ascribe capacity and they do so in super-abundance when it comes to ascribing the capacity to track reasons that is presupposed to fitness to be held responsible.
  If there is no bright line to mark off cases where the presupposed capacity is fully or partially present and cases where it is not, then how do we reliably manage to sustain the practice of holding one another responsible in ordinary life? My own view is that a convention of interpersonal exchange – a convention that may be sourced deep in our nature – forces us to hold each other, and indeed ourselves, to standards that we may routinely fail to satisfy. It is as if we conspire in the flattering fiction that there is rarely reason in the course of ordinary life to give up on one another’s capacity to track relevant reasons; there is rarely reason just to resign ourselves, free of hope, to whatever level of underperformance we encounter.⁹xFor a discussion of this “conspiracy in conversability” see Philip Pettit, “Joining the Dots”, in Common Minds: Themes from the Philosophy of Philip Pettit eds. Geoffrey Brennan, Robert Goodin, Frank Jackson and Micheal Smith (Oxford: Oxford University Press, 2007), 215-344. Peter Strawson underscores the point when he argues, famously, that to stop feeling resentment or gratitude towards one other – to opt for resignation rather than a continuing presumption that the person is able to track the reasons – would be to take a wholly objective stance that is inconsistent with ordinary life and interaction; it would be to take a stance that we reserve for those we deem to be out of their minds: subjects of treatment rather than partners in exchange.¹⁰xSee Peter Strawson, “Freedom and Resentment”, in Gary Watson (ed.) Free Will (2nd ed) (Oxford: Oxford University Press, 2003).
  This way of thinking naturalizes the notion of fitness to be held responsible. Daniel Dennett has worked over many years to dislodge the illusion that we human beings achieve insight and consciousness in virtue of an inner Cartesian theater in which we scan the contents of our every proper experience and thought and intention.¹¹xDavid Dennett, Consciousness Explained (New York: Penguin, 1992). I think it is equally important to get rid of the illusion that we human beings achieve freedom and control in virtue of an inner Cartesian HQ in which we select the options we choose with the discretionary independence of a god.
  I am individually assembled out of many modules and systems, each with its own domain of operation, and there is no control center at which I, an indivisible sovereign self, dictate my actions from on high.¹²xSee Philip Pettit, “Neuroscience and Agent-Control”, in Distributed Cognition and the Will. Individual Volition and Social Context eds. Don Ross, David Spurrett, Harold Kincaid and G. Lynn Stephens (Cambridge, Mass: The MIT Press, 2007), 77-92. I may enjoy a certain control in my behavior but that control has to emerge from a constant struggle among contradictory impulses and insights to let a consistent, persuasive set of values shape my responses. If in the wake of that struggle I assume a unified profile, that unity does not reflect any unity at the level of production. Whatever capacity I display to track the reasons is hard won, reflecting a victory for agonistic skill, not any unworldly autonomy; it is manufactured out of a mix of tactics and tricks that I have learned to deploy in the cause of self-control.¹³xSee Victoria McGeer, “The Moral Development of First-Person Authority”, European Journal of Philosophy 16 (2008): 81-108.
  The naturalistic, even pedestrian character of our capacity to track reasons and our fitness to be held responsible teaches us, as emphasized, that there is no bright line dividing cases where we are responsible and cases where we are not. But it also supports three other lessons that I want to emphasize.
  A first is that there is absolutely nothing surprising about the fact that a principal strategy for inducing fitness to be held responsible, deployed canonically in the case of children, is to treat agents as if they were responsible. By doing this we provide each other with a motive for taking steps that may help us to become properly fit to be held responsible; we responsibilize one another, as I put it in my paper. When I treat you as if you were fit to be held responsible, I credit you with a capacity that you will be happy to be thought to have: the capacity to perform as a proper person amongst persons. And at the same time I hold out the prospect of your being subject to sanction, and perhaps even losing a title to the status of being fit to be held responsible, should you fail to behave as relevant reasons require. Thus I give you an incentive to make the extra effort to prove yourself worthy of the tribute I pay. The effort may involve giving sustained attention to the considerations that keep you going in the right direction, or guarding against the lapses or impulses that might lead you astray. If repeated, this is the sort of effort that ought over the longer haul to generate the skills that will enable you to track the reasons reliably and to count as generally fit to be held responsible.
  A second lesson that the naturalistic character of our fitness to be held responsible supports is that there is good social sense in the practice of holding one another responsible – and indirectly of holding ourselves responsible. That practice tends to elicit and reinforce the very capacity it presupposes; it makes us into creatures who can be brought to a higher level of performance than might have been achieved under a brute regimen of inclination and impulse. The use of the practice in individual cases is backward-looking, focusing as it does on particular actions and on the responsibility – or perhaps the responsibilizability – of particular agents. But the rationale of the practice, as we might put it, is inherently forward-looking. Holding one another responsible for what we do is a practice whereby we raise ourselves by the bootstraps and make one another capable of being held to socially well-coordinated expectations. It amounts to a strategy for collective, socialized self-discipline.
  The third lesson of naturalism follows naturally on this. When it is within our discretion to extend or restrict the practice, we should make the decision on a forward-looking, consequentialist basis. It may not be the case that wherever there is a capacity to hold someone responsible for certain actions we ought therefore to extend the practice to that domain. There might be no point in imposing the practice in that domain, as there is no point in holding people responsible in choices that do not engage with matters of good and bad, right and wrong. This lesson is that which Rawls upheld when he argued in an early article that while we ought to punish offenders within a penal practice on the basis of how severely and culpably they offended – that is, on a backward-looking basis – we should look to its forward-looking consequences, in determining the shape and the scope that we ought to give to the practice.¹⁴xJohn Rawls, “Two Concepts of Rules”, Philosophical Review 64 (1995): 3-32.

• 2 Responsibility at the Corporate Level

  Corporate agents

  These remarks set the appropriate background, as I see things, for thinking about corporate responsibility. The issue of corporate responsibility gets off the ground in virtue of the argument that corporate entities do indeed count as agents in their own right, an argument that I develop more properly in a forthcoming book with Christian List.¹⁵xChristian List and Philip Pettit, Group Agency: The Possibility, Design and Status of Corporate Agents (London: London School of Economics, 2009).
  A group of people will constitute a corporate agent, in my terminology, insofar as they simulate an individual human person, being capable of advancing considered goals in light of considered representations and, in particular, being capable of registering and responding to desiderata of rationality as well as to the values that members of the group personally countenance. The members may not collectively achieve perfect consistency, perfect sensitivity to evidence, or any such ideal of rationality, but as a group they must recognize that when a shortfall is pointed out on those fronts, then they cannot be indifferent. They must aspire to perform as an individual person might perform in these respects. Whether they constitute a town meeting, a political party, a commission of inquiry, a private association, or a commercial enterprise, they must recognize that in their collective identity, as well as in their individual identities, they have to be an agency with whom others can do business; they must constitute a conversable interlocutor.
  We know from the individual case that this status does not come into being in virtue of the rule of a Cartesian HQ. Building on a multiplicity of subsystems, and on a toolkit of contingent resources, individual persons have to strive agonistically for the internal coherence and interactive sensitivity that will enable them to perform as conversable interlocutors. What is true of individuals, taken severally, is equally true of the corporate agents that they form together. The groups that incorporate in the manner envisaged have to coordinate around an implicit or explicit constitution, establish roles and protocols of interaction, and put safeguards in place against a variety of disruptive contingencies, in order to ensure their satisfactory agential performance. They have to work hard at achieving the integrity and the autonomy that we look for in any human-like agency. Individual persons are made, not born, and neither are their incorporated counterparts.
  In the context of this observation, I welcome the comment by Bert van Roermund and Jan Vranken: “Lawyers were well head of philosophers in realizing that ‘mind’ is the concept of a capacity supervenient on complex organizations of biological hardware (neuronic synapses, for instance). Equally, they saw no problem in extending supervenience to other complex organizations of hardware (social relationships, for instance).”¹⁶xBert van Roermund and Jan Vranken, “Morality Incorporated? Some Peculiarities of Legal Thinking”, in this issue. Corporate agents only get to be recognizable as such when we banish the specter of the Cartesian HQ from consideration and acknowledge that agency can be achieved on a distributed basis, in virtue of a suitable coordination of parts; it does not require the subordination of other parts to a single center, as in an analogue of the Cartesian subject. This is the point that was recognized by the medieval legal commentators, when they argued strenuously for the legal personhood of the guild or town or church. And it is the point that was lost to philosophers, including even Thomas Aquinas, as they followed the lead of the pope, Innocent IV, in arguing that the lack of a soul – the medieval counterpart of the Cartesian HQ – meant that a group could only provide a simulacrum of agency, not a proper instance.¹⁷xFor the references to Innocent IV, Aquinas and the commentators see “Responsibility Incorporated”.

  Networked, hierarchical corporate agents

  Bert van den Brink suggests that in my original paper I provide a persuasive argument for thinking only that more or less egalitarian, participatory groups, not hierarchical bodies, count as agents that are fit to be held responsible in their own right. “Pettit assumes that, on principle, all members of the corporation should be seen as having played an equally influential role in the generation of the constitution and the beliefs, desires and intentions that made the harmful actions that resulted likely or inevitable.”¹⁸xBert van den Brink, “On the Enactment of Corporate Arrangements”, in this issue. The reason he thinks this is that it is only such groups that put in place the arrangements whereby different policies are formed and implemented. With hierarchical group agents, it is those in commanding positions, not the foot soldiers – and so not the group as a whole – that take charge of constitutional matters; and so it is only these parties, therefore, not the group, that ought to be held responsible. Take the company whose sloppiness, according to the government report, led to the disaster with the Herald of Free Enterprise. According to Van den Brink, the responsibility does not lie with the deckhands but “only or mainly with those who had [suitable] roles and powers within the corporation”: in effect, “the owners, board, and management”.¹⁹xIbid.
  I am happy to concede that with some groups that behave as agents, it is really only a minority of those in charge who constitute the true group agent; others are like hired hands that that group agent recruits to its purposes: they have little or no role to play in constitutional matters, i.e. in making the arrangements under which policies will be formed and implemented, and they will not count as members proper. But I do not agree that this means that corporate agents proper – in particular, agents that can be held responsible in their own right – have to be structured in an egalitarian, participatory way. If I concentrated on such groups in my paper, that was for convenience only.
  While an incorporated group counts an agent in virtue of its members being coordinated so as to maintain a collective agential profile, there is no reason why that coordination should not involve specialization and hierarchy. Think of the condominium that operates as a single agent, and generally does so in a way that serves the common interest of members. The decisions reached by such a body, and the actions taken, will have to pass through a series of procedural tests in order to be attributable to the body as a whole. They will have to emanate from an elected committee, they will have to survive the possibility of objection in a period when they are subject to the review of members, they will have to pass muster with the lawyers and auditors that the body employs, and so on. And it is only if they surmount all those hurdles that they will count as decisions or actions of the condominium as a whole; their attributability to the corporate body will depend, as a matter of knowledge shared amongst members, on their successfully running that gauntlet of tests. In this sort of corporate agent, there is a great deal of specialization and hierarchy. Yet a large number of members, not just those in more central roles – not just those on the committee, for example – will have a role in letting the decisions form. The role may be just that of the ordinary member who decides not to contest a committee initiative or indeed that of a member who fails even to open the letter in which objections are solicited.
  These observations apply, not just to particular decisions that such a group takes, but also to those larger decisions about whether the very arrangements for taking decisions – in effect, the constitution of the group – should be changed or not. Where the constitution of a group leaves a role for every member to play a role in amending the constitution itself, however minor that role may be, the success of a proposed amendment will always be hostage to the possibility that this or that member does not go along. Members may not have a veto power but they will all have a power that is capable in principle of frustrating the passage of the amendment. Thus the passage of an amendment, or indeed the rejection of an amendment, should be taken to be the work of the body as a whole, not just the work of more influential officials. The officials cannot act in promoting those amendments with indifference to other members. And so, under standard rules of attribution, what is done will count as having been done, not just by those officials, but by the group in whose name they act.
  As it is with a condominium, so it may be with a commercial organization. The deckhands in the company that owned the Herald of Free Enterprise may be implicated in the sloppiness just as much as the managers, even though their role as enactors of company decisions did not generally amount to much. Perhaps it fell to their part, explicitly or as a matter of presumption, to send warning signals to management about any dangers created by the existing way of doing things. Or perhaps they could always have brought up the matter of safety at a union meeting – and the general matter of the arrangements under which they were forced to operate – with a view to having their representatives pass on the concerns. In any such case they would have been a connected part of the total organization, not just hands hired by the board, as by a distinct group agency.
  I regret the focus in my paper on egalitarian, participatory agents, although it served the purposes of convenience, making for a ready connection with the discursive dilemma. Most group agents, as these commentaries all recognize, are not of this kind; they are constituted by a networking of individuals in different roles, and with different degrees of influence. The medieval commentators whom I mentioned in the paper, such as Bartolus of Sassoferrato, were well aware of this. Thus they were happy to describe a city like Perugia – a civitas or state – as a group agent – an universitas or collegium – though it operated under a mixed constitution and enforced a separation and sharing of decision-making powers; they saw it as a persona ficta: an artificial person.²⁰xSee Cecil N. Sidney Woolf, Bartolus of Sassoferrato (Cambridge: Cambridge University Press, 1913).
  It was only with the work of Jean Bodin and Thomas Hobbes that this possibility of a networked agent got to be elided. Following the lead of Bodin, Hobbes argued in the central case of politics that the people with whom he identified the civitas or state can get to be incorporated only in one of two distinct way. First, by being represented by a monarch or aristocracy, in which case ordinary individuals are indeed like hired hands, not proper members.²¹xHobbes does require that people authorize these representatives but authorization comes cheap in his books; it need involve nothing more than submission to a conqueror that is motivated by fear for your life. On these aspects of Hobbes see Philip Pettit, Made with Words: Hobbes on Language, Mind and Politics (Princeton: Princeton University Press, 2008). Or second, by being members of the sort of egalitarian, self-determining assembly that both Bodin and Hobbes described, in a novel turn of terminology – designed strategically, I suspect, to give the word connotations of infeasibility – as a democracy.

  The fairness issue

  But will it be fair to assign corporate responsibility for some ill to a networked group, allowing that responsibility to be equally shared amongst those with minor and those with major roles, indeed amongst those who dutifully enact their roles and those who don’t, and even amongst those who go along with what is done and those who raise a protest?²²xIt was mistaken on my part – certainly it was misleading – to introduce the word “nonprotesting” in this remark from “Responsibility Incorporated”, about the members of responsible group: “As continuing, nonprotesting members, they will be complicit in the group performance”. Bert van den Brink and Govert den Hartogh each argue that this would be unfair. It will be unfair, as Den Hartogh puts the complaint, “if these people (…) have done nothing for which they personally deserve to be blamed”.²³xGovert den Hartogh, “Collective Criminal Responsibility: Unfair or Redundant”, in this issue.
  Those who personally deserve to be blamed for a corporate offense will have “enactor responsibility”, in my terminology, for what was done. They will be exposed to blame for the part they played – assuming it was a part they might have refused to play – in implementing the corporate act. This part may be positive or negative; they may have actively contributed to the act or they may just have failed to protest against it. The fact that everyone in the networked group agent shares “member responsibility” equally – this is the responsibility that goes with being a member of a group agent that is held corporately responsible – is quite consistent, then, with a great inequality in the distribution of enactor responsibility. And so the distribution of responsibility won’t be insensitive to the differences in the roles played by different people.
  But is it fair to impose equal member responsibility on those in a networked group, even when some may have had little or nothing to do with something reprehensible that was done by the group? I respond with a counter-question. There are many cases, as I tried to show in “Responsibility Incorporated”, where members of a corporate offense may be able to appeal to credible exemptions or at least partial excuses for their individual, enactor responsibilities. Would it be fair to allow incorporated collections of individuals to get away with imposing various harms, then, when none of the individuals involved may be subject to blame, or at least a suitable degree of blame, as an enactor – when each has a good excuse, perhaps even a claim to exemption, in respect of his or her behavior? Would this be fair when the incorporated group does indeed count as an agency that is fit to be held responsible?²⁴xBoth Den Hartog, “Collective Criminal Responsibility”, and Van Roermund and Vranken, “Morality Incorporated?”, raise the possibility that having a regime of holding corporate entities responsible as such may itself create an opportunity for individuals to hide behind the corporation. But they do not expand on this claim. For my own part, I do not see why the regime will create any more opportunities for evading responsibility than would exist in its absence.
  Suppose it is granted in response to my counter-question that group agents should be held responsible in their own right, in order to guard against possible lapses in the regime of responsibility – possible failures in our practice of keeping the book on one’s another’s actions. It will then follow that the members of culpable groups will have to share in the corporate responsibility just in virtue of being members and so that they will have to bear member responsibility equally. The equal responsibility they will have to bear as members will just be a reflection of the fact that the incorporated collection in which they are members is a responsible agent: they, together with other members, have done something blameworthy and reprehensible.
  Given that enactor responsibility is divided in proportion to the importance and influence of members in a particular corporate act, and given that there is good reason why we should hold corporate agents responsible in their own right as well, it follows that we should have a category of member responsibility that everyone can share equally. Extending the practice of holding agents responsible to corporate bodies – and so ascribing member responsibility equally – is justified on precisely the consequentialist sorts of grounds that make sense of why we maintain the practice with one another in our individual capacity.²⁵xPhilip Pettit, “The Inescapability of Consequentialism”, University Center for Human Values, Princeton, 2009, provides an argument that non-consequentialist considerations cannot serve in this role. Indeed those same grounds may argue, as I emphasized in the original paper, for holding corporate bodies responsible even when they have not yet put in place the safeguards required to protect against ill-doing. Provided such bodies are at least responsibilizable, it will make developmental sense to include them in the domain of the practice; it will provide an incentive for them to put appropriate safeguards in place.
  It would be objectionable to hold all members equally responsible, as holding a corporate entity responsible entails, if members really were like the hired hands discussed earlier, and had no possible role in affecting corporate decisions. But it is not unfair or objectionable to do so, given that membership entails some possibility of a decision-making influence, however marginal and however unequal, on group initiatives. This fact means that the will of each member is implicated in what the group does and that provides an intuitive ground for including all members within the ambit of a complaint against the group as a whole.²⁶xSee Christopher Kutz, Complicity: Ethics and Law for a Collective Age (Cambridge: Cambridge University Press, 2001).
  Despite the charge of unfairness, then, my view remains that people should have to answer, not just for things they individually do, but also for anything voluntarily done by a group agent of which they are members in the sense invoked here. This is particularly salient with voluntary associations. But it will also hold with states, or at least with states that are sufficiently democratic for the particular profile assumed by a state – in effect, its government – to be responsive to the inputs of citizens.²⁷xAnna Stilz, “Collective Responsibility and the State”, Dept of Politics, Princeton, 2009. The members of such a state will not have the right of exit, as with a voluntary association, but they will still have the right of voice: the right to contest what is done, with some hope of influence; they will not be condemned to mindless loyalty.²⁸xSee Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Cambridge, Mass: Harvard University Press, 1970).

  The redundancy issue

  But Govert den Hartogh offers a distinct argument against having resort to corporate responsibility. He argues that where I posit corporate responsibility, and so the responsibility of individuals as members of a group agent, we ought really to make do with the responsibility of agents as the makers and/or maintainers of a group agent. We ought to make do, in a phrase, with constituter responsibility, as I shall call it; ‘constituter’ is meant to refer either to a person who plays a part in manufacturing a group agent – a ‘designer’, as I said in the original paper – or to a person who plays a part in maintaining it in existence. In this argument Den Hartogh focuses on the legal practice of holding agents criminally responsible and imposing punishment. But he thinks that the argument, if sound, would apply in the moral area too: “In holding someone responsible for wrongful harms we at least express blame, and all the points (…) about the allocation of punishment can also be made about the allocation of blame.”²⁹xIbid., 1.
  Den Hartogh starts from the following premise, which I am happy to endorse. “All properly collective intentions of incorporated agents presuppose the merely joint intentions of individual agents to participate in the incorporation, each acting for his own reasons.”³⁰xIbid., 4. My own understanding of what a joint intention involves fits with his, building on Michael Bratman’s work.³¹xMichael Bratman, Faces of Intention: Selected Essays on Intention and Agency (Cambridge: Cambridge University Press, 1999). Two or more people jointly intend to do something together, X – or at least jointly acquiesce in their Xing – just in case conditions like the following obtain. There is a salient scheme of cooperation, however embryonic, under which they can together X; they each intend that they together X by following that scheme: that is, they each intend to play their role in enacting that scheme as a result, at least in part, of expecting others to play their roles; and all of this is a matter of shared awareness: they each believe that the conditions obtain, they each believe that each believes that, and so on.³²xThis formula involves some inessential variation on Philip Pettit and David Schweikard, “Joint Action and Group Agency”, Philosophy of the Social Sciences 36 (2006): 18-39. Like Den Hartogh I am prepared to grant, at least for all the cases that concern us here, ³³xList and Pettit, Group Agency, Ch 1, argues for cases where corporate agents exist without such an intention on the part of members. But they are quite marginal. that individuals constitute a corporate agent together in virtue of jointly intending to do so or, to sound a somewhat deflationary note, in virtue of jointly acquiescing in doing so.³⁴xSee ibid. But of course it is not the case, as Den Hartogh also recognizes, that people who jointly intend to do something together also thereby constitute a group agent. This may be denied by Margaret Gilbert, On Social Facts (Princeton: Princeton University Press, 1989). There is a salient scheme of cooperation whereby they can set up such an agent, or maintain it in existence and operation, and they each intend to play their part in that scheme, believing that others will play their parts too. And they do this, moreoever, as a matter of shared awareness.
  Where does this claim take Den Hartogh? At first, it seems that he wants to go straight to a radical conclusion: that the members of any corporation “jointly intend everything which is done in their name”.³⁵xDen Hartogh, “Collective Criminal Responsibility”. Did that conclusion hold, he suggests, then there would be no need to go to the notion of corporate responsibility. We could say that all the individuals involved are responsible as individuals for the joint action of setting up and/or maintaining the corporate body and for the consequences of that joint action: the actions that the corporate body takes.
  But of course that conclusion does not follow, as Den Hartogh immediately observes, for the constituters of a group cannot be expected to keep track of all that the corporate body does. Thus, without being particularly negligent, they will have to shift or offload responsibility for those actions to appropriate enactors. We may want to guard against negligence and specify in advance – presumably in morality as well as law – the duties that members will have to bear for what is done by enactors in the group’s name. But “there is an upper limit of fairness to the extent to which their responsibility can be increased”,³⁶xIbid.and we should look for principles under which it may be reasonably assigned.
  Without going into his views in any more detail, we can summarize Den Hartogh’s claim in the following terms. Apart from the responsibility of enactors that he and I can agree in recognizing, we need not acknowledge member responsibility in the sense I have in mind. We need only recognize the joint responsibility that makers and maintainers of a corporate group will have, though this constituter responsibility will not be distributed equally, only under a fair allocation of duties.
  I agree that apart from the disjoint responsibility that individuals may have for the things they do as individuals, they have joint responsibility for the things they do together, acting on joint intentions. And I agree that this sort of responsibility is different from the corporate responsibility that group agents should be ascribed and from the member responsibility that goes with it. But I still think that we need to recognize this third sort of responsibility as well.
  When members set up or sustain a group, intending that they act suitably together, they may do so reluctantly; that is why I suggested that ‘acquiesce’ might figure in the definition of joint intention. Thus it may only be for want of a better alternative, or because the costs of exit are high, or because they are pressured or unimaginative, that some individuals continue to intend that they together with others sustain a certain association: perhaps a voluntary group, perhaps a commercial corporation, perhaps even a state. But once we see that some individuals may only acquiesce reluctantly in making or at least maintaining a group agent, we can recognize that they might each have a good excuse for sticking by such a body, even when it is liable to generate objectionable acts. As those who enact the harm imposed by a corporate entity may be exempted from blame, on grounds of ignorance or fear for their lives, or may be partially excused on grounds of the difficulties or duress they face, so those who act jointly to support a group may, as such, be exempted or excused on similar grounds.
  Thus, as Den Hartogh himself recognizes, adding constituter responsibility to enactor responsibility does not get over the problems discussed earlier, and in the original paper. If the only responsibility that can be ascribed in the case of corporate wrong-doing is a responsibility that attaches to individuals as enactors or constituters, then there will be gaps in the books that we can keep on individuals. There will be corporate opportunities for individuals to combine in wrong-doing, without anyone’s being exposed to rebuke; there may even be opportunities for individuals to exploit this loophole intentionally. The consequentialist argument in favor of recognizing member responsibility – a sort of responsibility that no member can duck – remains still in place. Unless we hold corporate agents responsible as such, thereby holding members responsible just as members, there will be a shortfall in what we should expect the practice of holding agents responsible to deliver.
  These are difficult matters and I can quite see why others might take a different view from mine. So let me add three further observations in the hope of making my own view more plausible. The first thing that I want to emphasize is that there is nothing mysterious about holding a corporate entity responsible as such; it is not as if doing so gives responsibility to a weird or even fictional entity. All it means is that the individuals, qua group agent, are held responsible; or, equivalently, that the individuals, qua members, are held responsible. The aspect under which people count as members is distinct from their aspect as enactors and even, as in Den Hartogh’s proposal, from their aspect as joint constituters of the group.
  The second thing I want to add is that there is no double counting involved in holding people responsible qua enactors, or indeed constituters, and holding the group agent responsible at the same time: that is, holding the people responsible qua members as well as qua occupants of those other roles. This is because people are held responsible in each case for a different act, as I stressed in my paper. The people qua members – that is, the group agent itself – are responsible for allowing, even programming for a harmful or otherwise evil act. The people qua enactors are responsible for performing the act allowed, implementing the group’s program. The people qua constituters are responsible for becoming members in the first place and for continuing to remain members.³⁷xI allowed for something akin to this third role in speaking of the responsibility of people as designers of a group agent. What Den Hartogh nicely shows is that the notion of designer should be expanded to encompass his constituters and that it should be interpreted in the light of work on joint intention and action.
  The third observation is that there is no inconsistency in allowing that people may be excused or even exempted from responsibility for helping to constitute a group agent – that is, in effect, for joining or staying in the ranks – and holding that they can be equally held responsible as members. Their equal responsibility as members derives from the responsibility of their group. Whatever the motives that lead them to become or remain members – and whatever the excuses they may be able to offer, therefore, for playing such a constituter role – the cost of being members is that they share in any responsibility that we assign to the group as such. And that cost, as I argued, is fairly imposed insofar as membership gives individuals some degree of influence or control and thereby implicates their wills; it means that they are not just hired hands.

• 3 The Connection to Law and Politics

  Responsibility and legal liability

  Nothing in what I have said so far, and nothing in my original paper, addresses issues of legal liability, criminal or civil. Where a regime of moral responsibility dictates when people are morally blameworthy for what they do, a regime of legal responsibility will dictate when people are exposed to legal sanction, criminal or civil, for their actions. Given the difference in the point of each institution, it is not surprising that they should come apart in various ways and Van Roermund and Vranken give a helpful account of various points of divergence.
  Focusing mainly but not exclusively on private or civil law, they argue that the crucial legal issues are: one, whether an action or other consequence should be imputed to an agent, including a corporate agent; and two, how far the agent should be called to account before an adjudicator and subjected to costs or penalties in the event of being found liable. They then point out that “responsibility and imputability are not congruent”, since agents may be held legally imputable, accountable and liable in a way that diverges from how they may be held morally or socially responsible.³⁸xVan Roermund and Vranken, “Morality Incorporated?”
  This divergence is not surprising, nor do Van Roermund and Vranken suggest it is. An agent, individual or corporate, may be exposed to the law under a regime of strict liability or vicarious liability, for example, without being exposed to blame. And an agent may be exposed to blame and be held morally responsible without being taken to be legally accountable or liable; thus it will often be ineffective, counterproductive or just culturally inappropriate to subject certain actions that we take to be morally blameworthy to legal sanction. The only qualification to add to this picture is that the observation about responsibilizability suggests that there may be a developmental rationale for having an analogue of strict liability within the regime of moral responsibility. There may be a point in exposing to blame individual or corporate agents that are not strictly fit to be held responsible; this may be the best way of making them fit to be held responsible, and it may be desirable that they achieve such fitness.
  Van Roermund and Vranken appear to be happy to ascribe a mind or mens to such an agent, as I would be happy to do, and so to think that it may satisfy a mens rea or guilty-mind condition. But they say that this ascription or “imputation is governed by liability right from the start”, suggesting that imputability is a shadow cast by the coercive, state imposition of liability.³⁹xIbid. The suggestion carries them dangerously close to a radical conventionalism about imputability and liability – and perhaps, by extension, about moral responsibility – as when they comment that “the whole idea of the legal person, as distinct from the natural person, was invented to organize and facilitate social functions”. ⁴⁰xIbid.
  While I agree that corporate entities have minds and count as persons – artificial persons – I should emphasize that I do this on the basis of an analysis of what it is to have a mind and to be a person, not out of a belief in the sort of radical conventionalism signaled by this remark. Van Roermund and Vranken may actually agree with me but the point is still worth making for the sake of clarity.
  On a Cartesian view of what it is to have a mind or be a person, there is a deep metaphysical fact involved: the minded, personal status of an agent turns on whether there is a non-physical res cogitans present. On this view, nothing artificial like a corporate group could have such a status. But I reject that Cartesian view in favor of a functionalist account, according to which it is sufficient for being a minded person that a system, natural or artificial, function appropriately. Such a system will have the required status if it is organized, no matter out of what materials, to pursue considered goals on the basis of considered representations, and to be able to acknowledge and respond to desiderata of rationality – and values more generally – in the formulation of its goals, representations and actions. It is in virtue of naturalizing and demystifying what it is to have a mind and be a person in this way that I am led to acknowledge that corporate entities are agents that are often going to be fit to be held responsible.
  What is the difference between this sort of functionalism and the conventionalism that I reject? At the risk of striking a risible note, we might say this. According to functionalism, something that walks like a duck, talks like a duck and otherwise behaves like a duck is a duck. According to the conventionalism I would reject, anything that is called a duck is a duck; words rule but without being subject to the discipline of consistent definition.
  Before leaving the issue of legal liability I should comment on the legal sanctions – in the criminal case, the punishments – that are suitable with group agents. Corporate entities cannot be put in jail, as everyone recognizes, but how then are they to be punished, whether in law or in ordinary morality? Moral condemnation, especially when widely promulgated, may be punishment enough for a corporation, as common sense dictates.⁴¹xSee Geoffrey Brennan and Philip Pettit, The Economy of Esteem: An Essay on Civil and Political Society (Oxford: Oxford University Press, 2004). But fines and other restrictions will often also be appropriate. Govert den Hartogh protests that in a commercial case the fine may be cast as “a debit entry in the company’s book-keeping which in no way is understood to express blame”.⁴²xVan Roermund and Vranken, “Collective Criminal Responsibility”. But this is no objection. An individual might choose to see a fine in the same way, defying the social intention with which it is imposed. It is up to society to make the intent of the sanction clear; there is nothing about the sanction in itself that resists its being given a reprobative significance.
  Both Van Roermund and Vranken, and indeed also Den Hartogh, suggest that the law rightly identifies those in suitable positions within an incorporated group as the functionaries who should be held responsible for one another corporate offense.⁴³xIbid., and Den Hartogh, “Collective Criminal Responsibility”. I have no objection to this idea, nor to its counterpart in the broader enterprise of holding a group morally responsible. But I take that sort of prior setting to apply, of course, in the realm of enactor responsibility. It need not in any way compromise the idea of corporate responsibility, and corporate imputability, that I favor.

  Responsibility and politics

  Ronald Tinnevelt does a service in challenging me to connect my views on corporate and collective responsibility with my view of the role of states and peoples in the international order and raises some good questions.⁴⁴xTinnevelt, “Collective Responsibility, National Peoples, and the International Order”. I am happy to take up that challenge in this concluding discussion, though inevitably I have to be fairly brief. I am also hampered, in truth, by the fact that my views on international issues are not as fully developed as I would wish.
  In “Responsibility Incorporated” I suggested that even where peoples do not fully control the governments – more broadly, the states – that represent them, it may make some sense to hold those peoples, and not just the political authorities, responsible for things that are done in their name. It is unfortunate that in some discussions that Tinnevelt addresses I speak of “national groupings” rather than speaking, as I usually do, of “peoples”. Although I generally avoid the word “nation”, it leads him to raise the question as to whether I conceive of such groupings in the way that liberal nationalists like David Miller think of nations.⁴⁵xIbid. The answer is that I do not. I prefer to abstract from the issues of cultural identification raised by talk of nationhood. What I mean by a people, for purposes of this discussion, is the group that a government and state claims, presumptively with some authority – that is, some authorization by members of the group – to represent. That group may be a single nation, by whatever criteria are thought important, or an amalgam of many nations; I simply abstract from that issue.
  I suggest towards the end of “Responsibility Incorporated” that it may be useful to have a convention – a general, international convention – of holding a people responsible for the things their political representatives do in their name, even when the individuals involved do not have a degree of control, individual or collective, over those representatives. When ordinary individuals do not have such control, the larger group that is composed of the represented and their representatives – ordinary folk and political authorities – does not constitute a group agent that is fit to be held responsible; the only group agent that is properly present will be the smaller group that is composed of the political authorities. So why do I say that there may be point in holding the larger group responsible, as if it were a corporate entity? Because such a group may be responsibilizable. By being exposed to the discipline of corporate responsibility, ordinary folk may be given an incentive to assume more control over the authorities – over those they authorize as representatives – and to incorporate properly rather than remaining in the position of passive onlookers.
  But why, in turn, might it be desirable to make a people into an entity that is properly fit to be held responsible? Two considerations strike me as pertinent. First of all, as this group assumes more control over the representatives it authorizes, it will approximate more closely to a democratic ideal under which the demos achieves kratos or control over government. And, second, the more inclusively incorporated state is likely to behave better on the international front, as the actions of the government come under broader scrutiny and discussion; this might be thought to follow, for example, from the democratic-peace hypothesis under which democratic regimes are less likely to resort to war with one another than non-democratic regimes.⁴⁶xMichael W. Doyle, Ways of War and Peace: Realism, Liberalism and Socialism (New York: Norton, 1997).
  Tinnevelt is right to surmise that this developmental rationale for treating peoples as responsible is “related to the editorial dimension of democracy”.⁴⁷xTinnevelt, “Collective Responsibility, National Peoples, and the International Order”. He refers at this point to my “neorepublican” argument that in a democracy there should not only be the electoral control whereby the people indirectly author the laws that representatives enact and implement; there should also be a range of formal and informal contestatory powers – over and beyond the contestatory powers ensured by periodic elections – whereby the people can edit the actions of government, whether individually or in groups.⁴⁸xThe papers that he cites in particular are: Philip Pettit, “A Republican Law of Peoples”, European Journal of Political Theory, Special Issue on ‘Republicanism and International Relations’ 9 (2010); Philip Pettit, “Legitimate International Institutions: A Neorepublican Perspective”, in The Philosophy of International Law eds. John Tasioulas and Samantha Besson (Oxford: Oxford University Press, 2010). Tinnevelt goes on to say that this developmental rationale will only apply to countries where citizens truly have “room to keep their governments in check”.⁴⁹xTinnevelt, “Collective Responsibility, National Peoples, and the International Order”. I, of course, agree. He insists that citizens will have such room only where “they are able to initiate public deliberation and set items on the political agenda”, and indeed that they must have this sort of power if the government is to meet republican ideals and exemplify the ideal of non-domination in how it relates to citizens.⁵⁰xIbid. And again I agree.
  In making these points, Tinnevelt comments that he finds it surprising that I think of non-domination – or freedom, understood as non-domination – as an unmoralized ideal, declaring myself loath to accept “a more normative and political interpretation”.⁵¹xIbid. The remark is followed up in the conclusion, when he criticizes me for the lack of “a strong political interpretation of the ideal of non-domination”. I am puzzled about what he has in mind here. To hold that the ideal is unmoralized – my preferred term – is only to hold that it is definable in such a descriptive way that people with different evaluative commitments can still agree on whether some individual or body is dominated in a given context; it is to insist that the recognition of non-domination is not value-dependent. And it is not clear why that issue is relevant. Certainly it is not clear why a commitment to the point of view that Tinnevelt defends – and, as it happens, that I share – is threatened by keeping the ideal of freedom as non-domination unmoralized.
  On my own definition, one person or body, X, dominates another, Y, just in case X has enough actual or apparent power to impose his or her or its will on Y.⁵²xThis summarizes the main thrust of the variant formulations that I have used, for example in Philip Pettit, Republicanism: A Theory of Freedom and Government (Oxford: Oxford University Press, 1997); Philip Pettit, A Theory of Freedom: From the Psychology to the Politics of Agency (Cambridge and New York: Polity and Oxford University Press, 2001); and Philip Pettit, “Republican Liberty: Three Axioms, Four Theorems”, in Republicanism and Political Theory eds. Cécile Laborde and John Manor (Oxford: Blackwells, 2008). The power involved may be physical, cultural, or informational; the imposition may be total or partial and the means whereby the will is imposed may be by X’s active interference, by X’s invigilating or monitoring Y with a view to interfering if necessary, or by X’s intimidating Y. It follows from this account that if X interferes on Y’s own instructions and terms – say, as his sailors interfere with Ulysses – then the interference does not dominate Y and reduce Y’s freedom; it is non-arbitrary, as I say. Those who wish to moralize the notion of freedom as non-domination agree that non-arbitrary interference does not reduce freedom but take ‘non-arbitrary’ to have an evaluative meaning: to mean something like ‘legitimate’.
  I see no problem in refusing to moralize the definition of freedom as non-domination, while still agreeing with Tinnevelt on the importance of a contestatory, deliberatively regulated democracy. On the contrary, indeed, I think that a problem arises for those who moralize the ideal. There will be a point in invoking freedom as non-domination in deliberative exchange and contestation, only if the different sides can at least agree on when domination is present and when it is not. A deliberative regime will work well only if the basic terms of reference used by different sides are unmoralized and ensure a common currency of argument.
  The final issues raised by Tinnevelt relate to my views of the international order and how it should best be organized. I explore the question of how best to organize this domain, on the assumption that states are not going to wither away in the foreseeable future. Even someone who thinks they might wither away should be able to see value in that exploration, as I see value in the exploration of how it might be best to organize the international domain, allowing that states might prove dispensable. The only issue between the approaches will be, which is more likely to provide a usable public philosophy by which to assess and evaluate actual performance.⁵³xOn the requirements for such a usable public philosophy see Ch 5 of J.L. Marti and Philip Pettit, A Political Philosophy in Public Life: Civic Republicanism in Zapatero’s Spain (Princeton: Princeton University Press, 2010).
  Ronald Tinnevelt is impatient with my assumption that states in broadly the form we know them are likely to be a permanent feature of the international landscape – and, in particular, that a federal world republic is infeasible. It is difficult, however, to have a fruitful argument around this point; the empirics are just too opaque. But Tinnevelt has another complaint that strikes a chord. He takes me – probably fairly, on the basis of the texts he cites – to be insufficiently sensitive to the possibility that even if states play a crucial role in the international order, there may be room for individuals to enjoy an independent role, as under the proposal for a second chamber, representative of individuals, in the United Nations. There may be a place, as he puts it, for “a two-track model of legitimation (individuals and states)”.⁵⁴xTinnevelt, “Collective responsibility, national peoples, and the international order”. Here I readily admit that I may have been led by the assumption that states are likely to remain a part of any international order to a neglect of the possibility and utility of such arrangements. I am happy to consider the initiatives in institutional design that he recommends. Future work.

Noten

• 1 Philip Pettit, “Responsibility Incorporated”, in this issue.

• 2 There is no contrast on this matter with the law which, as Van Roermund and Vranken note, “takes a normative view (…) an agent ought to be aware of certain risks”.

• 3 See John Gardner, Offences and Defences: Selected Essays in the Philosophy of Criminal Law (Oxford: Oxford University Press, 2007).

• 4 I am grateful on this issue for points made by Gideon Rosen in discussion.

• 5 See Philip Pettit, “The Elements of Responsibility”, Philosophical Books 46 (2005): 210-19; Philip Pettit and Michael Smith, “Freedom in Belief and Desire”, Journal of Philosophy 93 (1996): 429-49.

• 6 I use this phrase in particular, as Van Roermund and Vranken point out, in Philip Pettit, A Theory of Freedom: From the Psychology to the Politics of Agency (Cambridge and New York: Polity and Oxford University Press, 2001).

• 7 For analysis of an unproblematic sense in which the sentence can be true, see Philip Pettit, “The Capacity to Have Done Otherwise”, in Peter Cane and John Gardner (eds.), Relating to Responsibility: Essays in Honour of Tony Honore on his 80th Birthday (Oxford: Hart, 2001), 21-35; reprinted in Philip Pettit, Rules, Reasons, and Norms (Oxford: Oxford University Press, 2002).

• 8 Ronald Tinnevelt rightly says that “one can’t hold actors responsible for things that lie beyond their control” but it is important to understand how control is understood here. See Ronald Tinnevelt, “Collective Responsibility, National Peoples, and the International Order”, in this issue.

• 9 For a discussion of this “conspiracy in conversability” see Philip Pettit, “Joining the Dots”, in Common Minds: Themes from the Philosophy of Philip Pettit eds. Geoffrey Brennan, Robert Goodin, Frank Jackson and Micheal Smith (Oxford: Oxford University Press, 2007), 215-344.

• 10 See Peter Strawson, “Freedom and Resentment”, in Gary Watson (ed.) Free Will (2nd ed) (Oxford: Oxford University Press, 2003).

• 11 David Dennett, Consciousness Explained (New York: Penguin, 1992).

• 12 See Philip Pettit, “Neuroscience and Agent-Control”, in Distributed Cognition and the Will. Individual Volition and Social Context eds. Don Ross, David Spurrett, Harold Kincaid and G. Lynn Stephens (Cambridge, Mass: The MIT Press, 2007), 77-92.

• 13 See Victoria McGeer, “The Moral Development of First-Person Authority”, European Journal of Philosophy 16 (2008): 81-108.

• 14 John Rawls, “Two Concepts of Rules”, Philosophical Review 64 (1995): 3-32.

• 15 Christian List and Philip Pettit, Group Agency: The Possibility, Design and Status of Corporate Agents (London: London School of Economics, 2009).

• 16 Bert van Roermund and Jan Vranken, “Morality Incorporated? Some Peculiarities of Legal Thinking”, in this issue.

• 17 For the references to Innocent IV, Aquinas and the commentators see “Responsibility Incorporated”.

• 18 Bert van den Brink, “On the Enactment of Corporate Arrangements”, in this issue.

• 19 Ibid.

• 20 See Cecil N. Sidney Woolf, Bartolus of Sassoferrato (Cambridge: Cambridge University Press, 1913).

• 21 Hobbes does require that people authorize these representatives but authorization comes cheap in his books; it need involve nothing more than submission to a conqueror that is motivated by fear for your life. On these aspects of Hobbes see Philip Pettit, Made with Words: Hobbes on Language, Mind and Politics (Princeton: Princeton University Press, 2008).

• 22 It was mistaken on my part – certainly it was misleading – to introduce the word “nonprotesting” in this remark from “Responsibility Incorporated”, about the members of responsible group: “As continuing, nonprotesting members, they will be complicit in the group performance”.

• 23 Govert den Hartogh, “Collective Criminal Responsibility: Unfair or Redundant”, in this issue.

• 24 Both Den Hartog, “Collective Criminal Responsibility”, and Van Roermund and Vranken, “Morality Incorporated?”, raise the possibility that having a regime of holding corporate entities responsible as such may itself create an opportunity for individuals to hide behind the corporation. But they do not expand on this claim. For my own part, I do not see why the regime will create any more opportunities for evading responsibility than would exist in its absence.

• 25 Philip Pettit, “The Inescapability of Consequentialism”, University Center for Human Values, Princeton, 2009, provides an argument that non-consequentialist considerations cannot serve in this role.

• 26 See Christopher Kutz, Complicity: Ethics and Law for a Collective Age (Cambridge: Cambridge University Press, 2001).

• 27 Anna Stilz, “Collective Responsibility and the State”, Dept of Politics, Princeton, 2009.

• 28 See Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Cambridge, Mass: Harvard University Press, 1970).

• 29 Ibid., 1.

• 30 Ibid., 4.

• 31 Michael Bratman, Faces of Intention: Selected Essays on Intention and Agency (Cambridge: Cambridge University Press, 1999).

• 32 This formula involves some inessential variation on Philip Pettit and David Schweikard, “Joint Action and Group Agency”, Philosophy of the Social Sciences 36 (2006): 18-39.

• 33 List and Pettit, Group Agency, Ch 1, argues for cases where corporate agents exist without such an intention on the part of members. But they are quite marginal.

• 34 See ibid. But of course it is not the case, as Den Hartogh also recognizes, that people who jointly intend to do something together also thereby constitute a group agent. This may be denied by Margaret Gilbert, On Social Facts (Princeton: Princeton University Press, 1989).

• 35 Den Hartogh, “Collective Criminal Responsibility”.

• 36 Ibid.

• 37 I allowed for something akin to this third role in speaking of the responsibility of people as designers of a group agent. What Den Hartogh nicely shows is that the notion of designer should be expanded to encompass his constituters and that it should be interpreted in the light of work on joint intention and action.

• 38 Van Roermund and Vranken, “Morality Incorporated?”

• 39 Ibid.

• 40 Ibid.

• 41 See Geoffrey Brennan and Philip Pettit, The Economy of Esteem: An Essay on Civil and Political Society (Oxford: Oxford University Press, 2004).

• 42 Van Roermund and Vranken, “Collective Criminal Responsibility”.

• 43 Ibid., and Den Hartogh, “Collective Criminal Responsibility”.

• 44 Tinnevelt, “Collective Responsibility, National Peoples, and the International Order”.

• 45 Ibid.

• 46 Michael W. Doyle, Ways of War and Peace: Realism, Liberalism and Socialism (New York: Norton, 1997).

• 47 Tinnevelt, “Collective Responsibility, National Peoples, and the International Order”.

• 48 The papers that he cites in particular are: Philip Pettit, “A Republican Law of Peoples”, European Journal of Political Theory, Special Issue on ‘Republicanism and International Relations’ 9 (2010); Philip Pettit, “Legitimate International Institutions: A Neorepublican Perspective”, in The Philosophy of International Law eds. John Tasioulas and Samantha Besson (Oxford: Oxford University Press, 2010).

• 49 Tinnevelt, “Collective Responsibility, National Peoples, and the International Order”.

• 50 Ibid.

• 51 Ibid. The remark is followed up in the conclusion, when he criticizes me for the lack of “a strong political interpretation of the ideal of non-domination”.

• 52 This summarizes the main thrust of the variant formulations that I have used, for example in Philip Pettit, Republicanism: A Theory of Freedom and Government (Oxford: Oxford University Press, 1997); Philip Pettit, A Theory of Freedom: From the Psychology to the Politics of Agency (Cambridge and New York: Polity and Oxford University Press, 2001); and Philip Pettit, “Republican Liberty: Three Axioms, Four Theorems”, in Republicanism and Political Theory eds. Cécile Laborde and John Manor (Oxford: Blackwells, 2008).

• 53 On the requirements for such a usable public philosophy see Ch 5 of J.L. Marti and Philip Pettit, A Political Philosophy in Public Life: Civic Republicanism in Zapatero’s Spain (Princeton: Princeton University Press, 2010).

• 54 Tinnevelt, “Collective responsibility, national peoples, and the international order”.