1. Retributivism, deterrence and the corporate agent

In this section, I make a preliminary argument that while deterrence is the theory that is most often invoked to justify our practice of imposing corporate criminal liability, desert should be given more prominence, in light of the corporate entity’s ability to make moral choices and therefore to be morally culpable. I describe this argument as preliminary because I will have reason to add to it and complicate it in §3 below. To illustrate some of the points in this argument, I will refer to the facts of the well-known Ford Pinto case of the 1970s: according to standard accounts of the case, employees of the Ford Motor Company were under pressure to produce an inexpensive and compact vehicle. In order to do this, they put the car’s fuel tank in the rear. They became aware that the car’s design, including its fuel tank placement, meant that some number of car passengers would likely suffer extreme burns or death. The company created an internal report that calculated that paying out civil costs to burn victims would be less expensive than recalling and fixing all the extant Pintos. And so the car was left unsafe, and more deaths resulted before the company was forced to fix it.^{Footnote 3} According to a contemporaneous news exposé, at least 500 people died as a result of the exploding gas tank flaw (who would not have died had they been driving a different model of car). Ford paid out millions of dollars in out-of-court settlements, and did not fix the car until new government regulations forced it to do so.^{Footnote 4} The cost-benefit analysis valued human life at $200,000 per human. Although many people died, it was the case of three teenage girls whose Pinto exploded, killing all of them, that brought the car’s defects to widespread public attention. Let us assume, hypothetically, that the company had been found guilty of manslaughter due to its recklessness in not recalling the car after it failed safety tests (in fact, there was a criminal trial but Ford was not convicted). Under what theory would we want to set a punishment for the company’s manslaughter of three young women? That is, assuming we believe that the decision to continue manufacturing and marketing a car that Ford knew was unsafe, and knew could be fixed with an inexpensive improvement (apparently a $5-dollar piece of rubber around the gas tank could have solved the problem) but chose not to fix, is indeed a criminal wrong and not a civil or regulatory issue.

There are several rival theories that seek to justify the state imposition of punishments on individuals. Deterrence is the theory that justifies punishment on the basis that it creates disincentives for those contemplating committing crimes in the future. Indeed, those who defend corporate criminal liability most frequently do so on the basis that it will achieve some deterrent aim.^{Footnote 5} Retributivism holds, by contrast, that we punish not because we want to create a deterrent for future behaviour, but because the offender deserves the punishment, in direct proportion to her moral blameworthiness in committing the offence. The idea that people deserve to be sentenced for wrongdoing is one that is extremely influential in philosophical justifications of the punishment of individuals, having surpassed deterrence in its influence as the leading sentencing theory in the 1970s.^{Footnote 6} One reason this happened was that the best evidence began to point to the fact that deterrence was not as effective as earlier theories had presumed.^{Footnote 7} Some philosophers seeking to find a justification for state punishment turned away from the instrumental uses of punishment, and considered issues of justice (including the state’s obligations of justice to offenders) more squarely. The focus on justice has always existed in our ideas about punishment, but in the last few decades it has been expanded by academics in a way that brings a sophistication to our understanding of what value a criminal sanction might have as a condemnation or rectification of wrongful acts.

Despite its being widely adopted as at least a significant part of the justification for punishment in the individual context, corporate sentencing has not been subject to the same analysis. Punishment theorists writing about retributivism rarely, if ever, take up arguments about corporate sentencing. Likewise, though with some notable exceptions, retributivism is rarely mentioned as a justification for corporate sentencing.^{Footnote 8} On the other hand desert theory is often raised when theorists want to argue against the criminalization of corporate misconduct, on the basis that the absence of moral blameworthiness in corporate acts shows why criminal punishment is unjustified or should be limited.^{Footnote 9} Retributivism might also be cited as a kind of folk explanation for corporate punishment: retributivism is painted as a palliative for uneducated people who are angry and need to see that an entity is being blamed.^{Footnote 10} The impression here is that observers see an unperceptive lay population that needs to be placated with a criminal conviction, and not a lay audience that might have a valid moral intuition about the possibility of moral responsibility residing in corporate entities.

At the same time, the ideas of retributivism and moral blameworthiness are often brought up by critics of corporate liability: the argument is made that it is foolish to punish corporations criminally when criminal punishment relies on ideas of moral blameworthiness and moral accountability that invest subjects with faculties that corporations do not possess.^{Footnote 11} So, we have the situation that in individual sentencing theory, moral blameworthiness is a feature that justifies punishment, while in corporate sentencing theory it has been presented as an obstacle to the genuine deterrent aim of corporate criminal liability.

Yet the response to corporate crime is most effective if it has a strong desert justification. This is true for two reasons, one we could call a ‘positive’ reason, a justification for criminal punishment, and one we could call a ‘negative’ reason, or a way of analyzing when criminal law is being used wrongly, and where a regulatory response would serve better. First, the positive reason: desert theory allows us to use the criminal law to combat the cultural idea that corporations are merely profit calculators, incapable of acting for moral reasons. If corporations are amoral, in the sense of being incapable of taking ethical considerations into account when choosing their activities, then the best we can hope for is a legal system that encourages them, through a system of incentives and disincentives, to conform to what society deems to be better behaviour (for instance, operating mines with safety standards rather than without). One of the things that went wrong in the Ford Pinto case may have been that members of the group believed that the only way to serve as agents of a corporation was to take decisions as amoral calculators, following the logic of their cost-benefit analysis as though there were no other option. But as group agents, as I will argue more fully below, corporations are in fact just as capable of making morally sensitive decisions as individuals are (as are their agents). It is to the person’s ability to be a moral agent that retributivism speaks.

As Andrew von Hirsch and Andrew Ashworth explain, the disapprobation of criminal law’s censuring response conveys, under a retributive (or proportionalism) rationale, the message that the offense was reprehensible and that, as a moral agent, the offender ought to have done better.^{Footnote 12} Deterrence-motivated punishment does not convey this message. Even if, as in Alan Brudner’s variant of retributivism, retributivism does not punish moral wickedness but rather vindicates the legal relation of autonomy between individuals,^{Footnote 13} a retributivist theory acknowledges the choice that an agent makes to either contravene or respect the agency (and the autonomy, the choice of personal safety, and all that follows) of others. If we are hopeful that the criminal law can be part of a process of leading corporate agents to understanding the fullness of their responsibility to others, rather than viewing norms as rules to be followed only when necessary, we ought to care about applying desert theory to corporations.

The negative reason retributivism is important as part of our understanding of punishment can also be seen in the work of von Hirsch and Ashworth. They point out that the censuring—blaming—part of criminal sanctioning creates a rationale for what should and should not be criminalized—where conduct ‘cannot reasonably be accounted for as involving wrongdoing’ it ought not to be treated under the criminal law.^{Footnote 14} This does not mean that such conduct cannot be regulated by the state, but it provides a way of distinguishing between conduct that infringes improperly on the dignity or autonomy of others and conduct that is merely undesirable in a specific social context. A strict focus on deterrence provides no way of keeping the criminal/regulatory distinction, and creates a tendency towards overcriminalization, which further allows business-minded individuals to disdain the criminal law as state overreaching.^{Footnote 15} Retributivism is often on the side of creating less criminalization, and so bolstering the legitimacy of the criminal liability that the proportional state does pursue.

There are three apparent reasons why theorists of corporate crime and punishment may have restricted themselves to deterrence and rehabilitation theories: the first is that the offender (a corporate entity) is the wrong kind of criminal offender—that because the corporation is not a moral agent, it does not make sense to try and analyse it through the lens of moral accountability. The second is that the offence is the wrong kind of offence: that regardless of whether corporations are moral agents, the types of offences that they are going to be tried for are not really wrongful, they are merely a type of conduct that the state seeks to disincentivize. The third is that people place more reliance on the effectiveness of deterrence in corporate cases. This third, however, is not a bar to applying retributive theory, but merely a reason why thinkers have not stopped to consider the issue: even if deterrence were more effective for corporations (and there is some reason to suppose that it is not)^{Footnote 16} it would still be an independent good for sentencing theory to address the moral capacity of corporate entities (if such exists) and their moral blameworthiness when they knowingly commit bad acts, for the reasons given above. I will now address the two objections to retributivism for corporations—lack of moral agency and lack of blameworthy action—in the hopes of convincing the reader that neither is ultimately persuasive.

a. Corporations as moral agents

If we are to discuss applying criminal and penal norms to corporations, then we must at the very least be able to identify these entities as rational agents, and not simply as loose associations of individual agents who would then be the sole proper loci of responsibility. This in itself is a controversial idea. And even if one accepts that corporations are rational agents, one might question whether corporations are also moral agents. I defend the position that corporations can indeed have full moral capacity and agency, even if we consider moral agency to be separate from rational agency. Bringing these points together, I conclude that the corporation is the kind of entity that can properly be brought within the ambit of a retributive punishment framework.

While this article does not intend to put forth any novel theory of corporate rational agency, it is necessary to understand a bit about what we mean by rational agency in order to move on to moral agency, which requires rational agency in order to exist. What does it mean to say that a corporation is a rational agent? As a first step, it certainly means that the corporation is more than a simple object, and that imbuing it with attitudes and judgments is not the same as imbuing a machine with the same.^{Footnote 17} It means that the agent responds to reasons, but it is more than that. It also means that the agent has some self-awareness, and that it takes responsibility for its attitudes and actions over time. While corporations’ actions are constituted of the decisions and actions of multiple people, the group becomes distinct from the individuals that constitute it, with decisions that can sometimes diverge from the decision of a majority of group members if they decided alone for the corporation.

Philip Pettit demonstrates the autonomy of corporate intentionality using what he calls the discursive dilemma, which this example illustrates: the three members of a board of directors must vote on three propositions that together will determine a certain course of action. For instance, a member has proposed that the corporation buy a new fishing boat (see table 1). A second member proposes that the corporation install expensive safety provisions if they do buy the boat. This same member proposes that the corporation take the boat close to a popular beach where there is lucrative fishing to be exploited. Member A votes for propositions 1, but against 2 and 3. Member B votes against 1 and 2, but for 3 (being of the opinion that they should use an existing boat to exploit the fish near the crowded beach), while Member C votes for 1, 2 and 3.

Table 1

The group passes the composite decision 1 + not-2 + 3—buy the boat, don’t install the safety provisions, and still fish near the heavily-attended beach. Note that no member individually voted for this conclusion, and each member may depart with the understanding that the corporate position is not really a reflection of her own views. If the boat kills a sunbather, and would not have done this had it had adequate safety provisions, arguably each single board member could say, I am not responsible because I voted against a crucial element of this plan. The adoption of the composite view casts the corporation as ‘an autonomous subject.’^{Footnote 18} The opinion of the corporation cannot be traced back to any individual member’s opinion, and so there is no way that it can simply be an aggregate of the three members’ minds. One cannot draw a line from this to an idea of non-responsibility for the individuals, of course, but it is still interesting that a corporation could be constructed so as to behave according to intentions that no individual member would approve.

What Pettit has drawn from this account is that not all aggregates are rational agents, but that there will be pressure on an aggregate to become rational, and that if the aggregate of individuals enacts procedures to turn it into a rational agent (for instance, by having a procedure to resolve internal contradictions like the one that A, B and C encountered above), and that when it does find procedures to become internally consistent, then it becomes a collective agent,^{Footnote 19} ‘an intentional subject in its own right’.^{Footnote 20} This is not to say that the corporation is in some mystical way separate from its members—the member individuals are what constitutes the collectivity. But it is, in Pettit’s structure, ‘distinct in the sense of being a center for the formation of attitudes that are capable of being quite discontinuous from the attitudes of the members.’^{Footnote 21} This comes from the fact that the group might come to different conclusions on an issue than its individual members (if for instance that issue is resolved through a combination of votes on sub-issues) but could also organize itself to be aware of its views and to hold itself to being consistent and logical in its attitudes and actions.

This is a simple sketch of rational agency. It is not within the purposes of this paper to mount a defence of rational agency, but merely to show that it exists and to adopt it as a framework for the understanding of what it means for a corporation to be a subject of criminal norms. We can now move on to consider whether or not it is also a moral agent. On one theory, rationality immediately implies moral agency as well. To respond to reasons is also to respond to normative reasons. Or, similarly but without taking a stance on what the content of moral reasoning is, one could hold that simply if corporate decisions are always a result of individual votes, then the moral dimensions of a choice necessarily get weighed in that way. This is the view that Pettit advances in ‘Responsibility Incorporated.’^{Footnote 22} It takes no stand on whether moral agency is a direct consequence of (human) rational agency. In the case of corporations, it seems to me that either way it is necessary to consider the issue of moral responsibility separately, since so much doubt has been cast on corporate moral agency specifically, and separately from rational agency.

A central way in which writers choose to cast doubt on the existence of corporate moral agency is to tie the ability to make moral judgments to the capacity to experience certain emotional states, and deny that corporations can have sufficient access to those emotional states. Corporations cannot fall in love, experience fear, and so on. Luban, Strudler and Wasserman view guilt and shame as ‘suitable moral emotions’ without which an entity could not ‘appreciate the significance of moral criticism.’^{Footnote 23} This account places emotions at the centre of moral understanding, going so far as to assign moral worth to emotions themselves. Similarly Michael McKenna asks whether collectives are able to feel resentment, gratitude, indignation or guilt, and argues that these states ‘involve attendant emotions that have their own phenomenal peculiarities.’^{Footnote 24} Wolgast draws attention to the fact that corporations have no friends and family, among other features.^{Footnote 25} The human-centered language that is used to address moral agency is often frustratingly misleading.

My response to this is that even if moral agency does not flow directly from rational agency, their existence sits side by side within the corporate agent. A rational agent has beliefs, attitudes and desires.^{Footnote 26} The corporate ability to possess financial knowledge comes from contributions of various corporate agents who intend that their knowledge should become part of the corporate entity’s knowledge. Corporations do not have distinct capacities of perception or memory, yet they are, as Pettit points out, able to ‘register and endorse’ information that their member agents perceive or remember. Mental acts of individuals can clearly be imputed to the corporation.^{Footnote 27} In the same way that not all physical acts of an employee are imputed to the corporation, not all mental acts become corporate acts. Board members or other decision-makers are able to transfer to the corporation such business-related mental states as a wish to protect the environment, to minimize the risk of harm to consumers, and to act out of certain fears: not the mortal human’s fear of dying alone, but the fear of becoming bankrupt. If the corporation can have knowledge, then through that same mechanism it can hold at least some emotional states.

The range of topics on which corporations think and make judgments is narrower than that of human individuals, since it is restricted to thoughts and judgments of members when they are acting in pursuit of corporate goals.^{Footnote 28} For this reason, a corporation may not be susceptible to certain emotions. In the case of romantic love, for example, it is hard to imagine that even if all the members of a corporation were individually in love with one person, they would form a joint intention to be in love as a group. On the other hand, we could imagine a group of individuals (say, a football team) suffering as a result of a group experience (say, a humiliating defeat at the hands of another team) and acting out of this jointly-held suffering (perhaps by being rude to the other team at the after-game party). At least when members act from their emotions in ways that make up group action, their emotional states become the group’s emotional states.

Larry May argues that corporate structure renders corporate entities incapable of assimilating emotional knowledge into their store of knowledge. He begins by noting that because corporations are designed to look after business interests, the relevant corporate individuals judge corporate actions on their business efficacy rather than for their moral worth. As a result of this, though moral concerns may be considered and even subjected to a vote, these concerns are nevertheless only the concerns of ‘each conscientious member of the board’ rather than of the corporation.^{Footnote 29} He argues that there cannot be a ‘melding of consciences into an institutionalized superconscience’ in a manner analogous to how other types of individual decisions become institutional decisions because only thinking which is ‘goal- or task-oriented’ could possibly be incorporated into a corporate mind.^{Footnote 30} This is despite the fact that May has allowed that members of the board of a corporation might be taking a vote on issues of conscience: so do those issues of conscience and the reasoning that goes into the vote become issues of business because they are incorporated into corporate structure? If one allows, as May does, that a corporate board may make a decision on explicitly ethical grounds,^{Footnote 31} then it seems very odd to claim that the decision, qua corporate decision, has no ethical dimension. The votes engendering the decision are created in minds that contain moral judgment, and this judgment is part of the reflection that goes into the vote, and through this mechanism it enters the corporation’s structure.

While there is much more to say on the intricacies of the moral agency of corporations, for the present purposes it is enough to note that there is a mechanism whereby the corporation can take note of and even experience, though this language may seem odd applied to a group agent, the necessary ingredients for a fully informed moral judgment. Thus, whether or not moral agency requires emotional understanding, corporations are moral agents. Since they are moral agents, they may be blamed for their wrongful actions, and their particular level of culpability can be assessed.

2. Punishment, suffering and the corporate entity

We turn now to consider the nature of punishment, and whether it is possible to criminally punish a corporation. In this section, I come to the rather pessimistic conclusion that it is in fact impossible for the state to punish corporations, despite there being good reasons (outlined in the previous section) for it to do so, if it could. The argument addressed in this section is in three parts: 1) suffering is a component of successfully imposed punishment; 2) corporations cannot suffer; 3) ergo, corporations cannot be punished. Corporations can experience emotions as collective entities, even those that we associate with wrongdoing and punishment, for instance an emotion such as remorse. They may even be able to suffer under some circumstances. But they cannot suffer as a direct result of the state imposing punishment on them. If suffering is a constituent of successful punishment, no corporation can ever successfully be punished.

There is a related argument that corporations cannot be punished because the punishments will never be significant enough to hurt the corporation. That position may rest on an intuition that it is hard to make the corporation ‘care’, as it were, about the punishment. But it seems to me that we can say, at least, that a corporation can be harmed by punishment. If you take away all of its assets, the corporation is harmed. This is because the corporation is an agent with goals and interests.^{Footnote 42} This is in contrast to a car or other object, which can be damaged but not harmed. But harm and punishment are not equivalent. For instance, a plant or tree may be distinguished from a car in that it can be harmed, rather than just damaged.^{Footnote 43} But it does not suffer. If it is true that a plant can be harmed but cannot suffer as a result of the harm, then harm and suffering are not coextensive. It is therefore also plausible that punishment harms the corporation without directly causing it to suffer.

There is a strong case to be made that punishment must involve at least some suffering on the part of the person punished.^{Footnote 44} The opposite view is that punishment need only involve a deprivation, and that suffering is merely a frequent side effect of the deprivations imposed. But if not to make the offender feel the sting of her criminal conviction, it is very hard to see why hard treatment would be imposed at all (I refer to the motivation to impose hard treatment rather than the ultimate justification for it). This does not mean that every time the state metes out a sentence, the person sentenced suffers. It means that where punishment does not make someone suffer, but rather causes them enjoyment or indifference, then the state has failed in its attempt to punish them. The fact that sometimes punishments fail to cause suffering is not relevant to the analysis. If some people do not suffer it only shows that sometimes the state’s attempt to punish is unsuccessful, which is an acceptable result for a state that must make rules of general application.

There are several opposing views to the position that the intention to impose suffering is a part of punishment. One is that deprivations constitute successful punishments, and that suffering is not involved in the deprivations that modern punishment imposes.^{Footnote 45} But if the deprivation is without suffering, in what way is it hard treatment at all? It begins to more closely resemble a tax or a penalty. Certainly criminal punishment is supposed to affect the offender in some way that is qualitatively different from taxation. Joel Feinberg famously opined in his essay ‘The Expressive Function of Punishment’ that what makes punishment distinctive among other kinds of penalties is its ‘symbolic significance.’^{Footnote 46} Is it enough then to combine the stigma of a criminal conviction with the deprivation of the sort that a tax imposes, to arrive at a punishment? No, for the symbolic significance would then be divorced from the penalty itself, and conviction and sentence would not be related to each other anymore. A conviction followed by a deprivation that everyone acknowledged would be meaningless to the person on whom it was imposed might be beneficial to society as a whole but this would not be a punishment.

Another alternative to the view that suffering is a constituent of punishment is that punishment instead requires a response that is objectively the kind of response that would be burdensome to most people, without considering whether it is burdensome to the individual being sentenced. Bill Wringe distinguishes between an intention to cause suffering and treatment that is normally found burdensome by individuals of the type being targeted.^{Footnote 47} David Gray sets up a distinction between subjective punishment theory, in which the state must consider such things as the offender’s victimization in prison to be part of his prison sentence, and objectivism, in which punishment is justified without considering the subjective experience of distinct individuals.^{Footnote 48} Gray seems to take it that this move requires him to adopt a theory in which suffering is an incidental side-effect of punishment. Wringe, similarly, argues in a way that seems to turn on the idea that the state must not be seeking to impose hard treatment in a way that intends to cause subjective suffering in the person being sentenced: if the goal were to cause suffering and suffering were not achieved, perhaps the state would need to try again, or would have a problematic failure on its hands.^{Footnote 49} But while we might want to frame the state’s approach as objective rather than subjective, there is, I submit, no other purpose for the state’s actions other than to cause suffering.

The fact that we do not tailor sentences to ensure that each individual will suffer by scrutinizing their inner desires and fears is not evidence that successful punishment need not include the imposition of some amount of suffering. The reasons we do not tailor sentences too closely to an offender’s inner life are unconnected with issues of the effectiveness of punishment. It is, as Wringe himself points out, because ‘there seems something worryingly illiberal about the idea that the state should be concerned with the state of my individual soul.’^{Footnote 50} And it is for that reason by itself that we do not ask judges to consider the level of suffering that hard treatment will impose upon each individual that they sentence. But consider the story of the old destitute man who commits a minor crime so that he can be housed and fed for some period of time at the state’s expense. When the judge orders a period of incarceration (and thus a bed and three meals a day), the judge is not properly punishing the offender. If, as an audience, we had less sympathy for the convict (say that his choice of offence was punching a small child viciously in the face), we would feel that giving him exactly what he wants is a travesty. This is a failed punishment. The failure of course does not mean that the state needs to radically rethink its practices; occasional failures are acceptable, as they are at the stage of detection and prosecution of crimes.

Consider another case of punishment: that of parents punishing children. Suppose Albert punishes his daughter Beatrice for hitting another child by making her mop the floors. The next day, Albert’s son Carl steals candy from a friend. But Carl absolutely loves mopping. Albert ought not say, ‘I will make Carl mop the floors as I did Beatrice, because it is generally considered burdensome treatment.’ Here, because Albert is not the state but a parent, there is nothing illiberal in his looking into his son’s personal tastes and using that to tailor his punishment, say by having Carl wash the dishes instead. Because Albert is not an agent of the state, Albert can ensure that he causes the correct amount of suffering (more or less) to address the wrong done by his child.^{Footnote 51} Albert could not correctly claim to have punished his child by imposing an activity his child enjoyed. In both the state punishment and parental punishment cases, we are willing to allow for some successes and some failures, since the goal of imposing justified punishment does not override all of our other interests. As a further reply to Wringe, it is not true that we do not tailor punishments to people’s personal circumstances: we might for instance expect a judge to consider the size of an offender’s personal wealth when setting a monetary penalty as a criminal punishment, and hope that the judge considers a mother’s young children when deciding whether to set a custodial sentence.

It is clear that the purpose of hard treatment is its suffering aspect because the opposite view, that the state imposes hard treatment without caring whether the sentence causes the offender to suffer, leaves no rationale for imposing that hard treatment at all. Wringe argues, against this view, that on the expressivist view of punishment (a variant of desert theory in which the purpose of sentencing is to express the community’s disapproval of the crime), the fact that punishment is objectively harsh is sufficient, even if it happens not to be harsh for the person sentenced, and that this demonstrates that the state does not intend to impose suffering.^{Footnote 52} I propose that we would do better to frankly acknowledge that part of the purpose of punishment is the infliction of suffering, and that there are countervailing reasons why the state does not look too closely at the personal circumstances of the offender in tailoring a sentence. Without its attendant suffering, there is no particular justification for hard treatment on either the deterrence view (at least in the case of individuals) or on the retributive view.

I have said that punishments may fail in individual cases and that this is perfectly acceptable, but things start to look different if we know, a priori, that all punishments imposed on a particular group will be unsuccessful. Consider a state that is so weak that the only punishments it is capable of imposing on wrongdoers are punishments that no wrongdoer would even notice, say the state threw a grain of sand at them. We would think it best for the state to acknowledge that it is not punishing these individuals. It would in fact be detrimental to continue in the attempt, as it would strengthen the offenders’ sense of impunity. The knowledge of systematic inability to punish would be a reason to abolish the punishment entirely for that class of persons. Similarly, if it is impossible to punish corporations we might have to rethink the practice of imposing corporate criminal liability on them. If suffering is necessary for punishment to be successful, and punishment must have the possibility of success to be justified, then in order to justify punishing corporations one would have to show corporations are capable of suffering.

A corporation can have emotional states, and suffering is an emotional state, so one might think that there is no further problem to be solved. But the method of imputing emotional states to corporations will not work here. The key difference lies in the direction of imputation. We attribute mental states—knowledge of financial records, moral judgment, repugnance, competitiveness—to the corporation from the mental and emotional states that individual members of the corporation turned into corporate mental states through their actions, such as writing reports, tabling motions, voting, and so on. But here we have a situation in which we contemplate punishing the corporate entity itself, punishing it directly, not punishing its members. The corporation can be deprived of assets and that deprivation might result in suffering for the individual members, but the suffering of individuals who are not named in the punishment cannot be used as evidence that the punishment worked.^{Footnote 53} A corporate entity cannot suffer directly, prior to the involvement of individual agents. Obviously it has no separate ability to feel or know, just as it has no separate ability to act. As with its actions, the corporation’s mental states must be performed by an agent.

Let us push a little further on this idea that individual agents of the group might suffer in such a way as to make it the case that the corporation suffered as a result of the punishment. Larry May brings up the possibility of assigning the role of feeling shame for corporate past acts to a manager. He dismisses this idea on the basis that corporate decision-making structure ‘is not at all set up for that type of backward-looking personal reflection.’^{Footnote 54} But May is too quick to dismiss the possibility that a corporation, as a group agent, could feel shame or remorse. It can act on the basis of a shared remorse, and can hold this feeling in the same way that it holds knowledge or other intentional states. Margaret Gilbert argues that group agents are indeed capable of remorse.^{Footnote 55} This occurs when individual members feel membership remorse for the collective’s actions. It is distinct from remorse for members’ own roles in those actions. A member of a church group could feel remorse for the part she played in discriminating against same-sex marriage. But she could also feel remorse as a member of the group for negative effects of an anti-same-sex-marriage position that the church held only before she became a member, because her ties to the group give her a sense of identification that extends to feeling remorse for the group’s prior actions. It is possible, on this account, for a corporation (as a group agent) to feel remorse for a wrongful corporate action and its resulting harms, through the interlocking remorse-intentions of its members.

So does the line of thinking from remorse work equally for suffering from punishment at the group level? It does, but not in the relevant sense for criminal punishment. This is because the punishment must be experienced directly, and not follow from the suffering of others. It would be impermissible for the state to directly punish individuals for their innocent membership in a corporation that committed wrongful acts. Perhaps this is the unspoken insight that causes J. Angelo Corlett to argue that corporate retributive punishment would be permissible but only if corporations were reorganized so that every member, every secretary and janitor, had meaningful decision-making authority within the corporation.^{Footnote 56} And yet, while this would make the flow-through of suffering more ethically permissible, which might then create a loop back to being corporate suffering itself through the institutionalized suffering of the members, requiring every single individual to be guilty in order for the corporation to be guilty has the effect of collapsing the corporation entirely, and negating its existence as separate from its individual members. Corlett’s position really amounts to a denial of corporate-entity blameworthiness, though he does not identify it as such. Perhaps a similar effect could be achieved through bringing suffering to the corporation by imposing suffering only on its individually implicated members. Typically, when an individual agent of the corporation is found guilty of a crime, that agent ceases to be a member of the corporation, and so her suffering cannot be looped back into the corporation. But what if a particularly philosophically-inclined judge chose to punish the responsible individuals but also sentence the corporation itself to keep those individuals in their current jobs so that their suffering could be fed back into corporate intentions and decisions? Would this constitute successful corporate retributive punishment, complete with a requirement of suffering? The judge-imposed individual suffering would still not constitute what we mean by suffering. For one thing, remember that not all member emotions become group emotions, and it is likely that the individuals suffering from hard treatment would not use that suffering in a way that caused it to become group suffering. Additionally, it is not a permissible move by the state to create a punishment by causing suffering to those surrounding the offender.

For all these reasons, and even though it would make corporate sentencing theory simpler, there is no way to understand suffering as something imposed directly on the corporation itself. And even if hard treatment has the indirect effect of making members suffer, that is not in any way connected to the principle of punishing the corporation itself for its wrongdoing. Corporate punishment fails. This may mean that corporate criminal liability fails as well, and that its pursuit ought to be abandoned. If we cannot punish, then why prosecute? But I am going to argue, in the remainder of this article, that there are still meaningful ways to sentence corporations for criminal wrongdoing, though it weakens the specific arguments made above about the applicability of retributivism to corporations. I turn now to investigate the possibility that the criminal law may have other reasons to seek corporate criminal convictions than punishment per se.

3. Ramifications for corporations as subjects of criminal law

The criminal law cannot communicate its disapprobation of wrongful conduct to the corporation in the same way that it can to individuals because of the absence of the possibility of true punishment. But can it still accomplish some of the goals of desert theory? An argument of Bill Wringe’s may be able to be adapted to fulfill this need. Though Wringe’s idea is not based on the problem of suffering, he maintains that Duff’s communicative theory of punishment does not work for corporations and proposes that instead of communicating to the offending corporation, one might save the communicative theory by admitting that instead we are communicating to the community at large.^{Footnote 59} While this would impermissibly treat an individual offender as a means rather than an end, Wringe argues that treating a corporation as a mere means is indeed permissible.^{Footnote 60} One need not get caught up in the distinction between ends and means to follow Wringe’s thinking, but simply acknowledge that although groups may have intrinsic social value both for their members and for the surrounding community, they do not need to be treated with the same respect as individuals, since groups can form and reform, but individuals are indivisible units.

This gives us good grounding to reach the conclusion that is consonant with many peoples’ intuitions: groups do not require the respect that individuals do, even if they are often important to our social existence. This seems consonant with the nature of a group as distinct from an individual agent and may also follow from the fact that they are immune to direct suffering. Treating corporations as a mere means would allow us to keep many of the communicative aspects of moral condemnation in desert theory, and use that as an alternate value for desert-based sentencing, while allowing that punishment is impossible. We might instead accept as a form of punishment a deprivation that is serious enough to demonstrate to the surrounding community that the system strongly denounces the conduct that is being punished. Considering that the community also includes the members who contribute their individual moral capacities to make up the moral agency of the corporation, expressing condemnation to the community through proportionate sentences would also have the effect of enhancing the moral abilities of the corporate entity itself.

Retaining a modified retributivism while acknowledging that the foundation of punishment on which it rests is doomed to failure in the corporate case allows us to retain the feature of criminal law that it ascribes moral responsibility, while remaining honest about its limitations in regard to corporations. Criminal punishment expresses the public’s condemnation to observers and to the members of the corporate entity qua group members by imposing hard treatment. While the criminal law does not apply to the corporation in the same way that it applies to individuals, there is still a benefit to having corporate criminal liability, in terms of communication and enhancement of moral values for the corporate entity and a robust expression to surrounding individuals that the state denounces corporate offending.

This analysis also has at least one practical consequence: someone who is setting corporate sentences (whether legislatively or in a particular case) may have more upward flexibility if retributivism is attenuated due to the impossibility of punishment. While desert principles require that an upper limit be strictly adhered to (a person must be punished in direct proportion to her culpability, and no more), showing that punishment is not possible in the case of corporations means that deterrent aims can be allowed to have more sway, with the result that upper limits to sentences will be higher (generally, a deterrence theory favours higher sentences, as these are supposed to create a greater disincentive to those considering engaging in the conduct). Because we are using the criminal law to enhance the moral agency relations of the corporation, it would not be desirable to have completely deterrence-based sentences that took no account of the culpability of the corporation. But, in keeping with the modified duty of respect that is owed to corporations based on the analysis of their personhood, somewhat higher limits on sentences could be countenanced as being in line with the proper principles that govern corporate punishment for wrongdoing. It also explains why punishments such as the revocation of a corporation’s charter (a corporate ‘death penalty’) could be an acceptable response to certain types of wrongdoing.