On April 9, 2017, the sixty-nine-year-old physician David Dao was seriously injured when he was forcibly removed from United Airlines flight 3411. United immediately faced widespread criticism.Footnote 1 The airline’s initial attempts to assign blame for the incident to Dr. Dao prompted a surge in pejorative mentions of the airline on social media and threats of customer boycotts.Footnote 2 United’s subsequent apology to Dr. Dao was characterized by a corporate communications specialist as being “back-against-the-wall, through-gritted-teeth.”Footnote 3
The most striking feature of the United case is not that the firm perpetrated an apparent moral wrong but, rather, that its attempt to deny responsibility for its actions triggered such strong and intuitive moral outrage. This is a conceptually difficult observation. Talk of United’s moral responsibility makes sense only insofar as we are prepared to allow that United is a moral agent. People commonly express indignation toward corporations on moral issues; when they do so, they assume, possibly unconsciously, that corporations are moral agents independently of their members. As the case of flight 3411 illustrates, if a corporation responds inappropriately to expressions of indignation, then, in turn, it attracts additional indignation; morality in the corporate sphere therefore appears to be ineluctably intertwined with the ways that corporate actors receive, process, and respond to addresses from outsiders.
We take a relational approach to understanding corporate moral agency. We are not concerned merely with how a corporation affects external actors but also with the ways in which it engages with them. This perspective is consistent with a philosophical tradition that contends that holding an agent morally responsible is synonymous with the formation of reactive attitudes, such as indignation, resentment, and guilt, in response to its actions (Strawson Reference Strawson[1962] 1974). The relational approach thus identifies corporate moral agency by attempting to understand the ways in which noncorporate actors hold the corporation morally responsible by expressing reactive attitudes. We rely in our analysis on the work of the philosopher Stephen Darwall (Reference Darwall2006). Darwall is concerned with the perspective that agents take when they make claims on one another and with the moral authority that underpins those claims. When an agent expresses a reactive attitude to another, it issues what Darwall calls a second-personal address; Darwall refers to the ability to respond appropriately to these addresses as second-personal competence, and that competence is central to his understanding of moral agency.
We construct a functional characterization of second-personal competence and use it to identify three conditions that a corporation must satisfy to be a moral agent. We therefore introduce a particular kind of relational approach to the debate of corporate moral agency: a second-personal one. A novel element of our analysis is that it starts from the emotions that people experience in response to corporate actors and from the pragmatics of expressing those emotions to the corporation. This approach is in contrast to other work on corporations and affect, which is concerned with the extent to which corporations can themselves be said to experience emotions. That work has a bearing on our analysis insofar as a capacity for emotion is necessary in the putative targets for emotional reactions; we discuss this point in detail in these pages.
Our analysis is meta-ethical: we present criteria for a corporation to be a moral agent, but we do not stake out a position on whether a corporation satisfies those conditions. Instead, we identify the critical points on which, on a second-personal account, a claim of corporate moral agency succeeds or fails. Those points relate to whether it is possible to experience emotional responses toward a corporation, to the ability to trigger dialogue with a corporation by expressing those responses in addresses to the corporation, to the status that those addresses are accorded in corporate deliberation, to the nature of corporate communication, and to the corporation’s ability to bind itself in response to its deliberations.
The second-personal approach to corporate moral agency stands in contrast to the dominant approach in the existing literature, which emphasizes the corporation’s capacity to act in pursuit of its own goals and to reason about what goals it ought to pursue. We characterize that type of analysis as first-personal: it is concerned with the corporation’s internal capacity to reason about the outside world and so to affect it; in contrast, the central pillar of the second-personal approach is the way that external actors experience the corporation as a social actor and the way that the corporation views those actors in a social context. The second-personal approach offers two ways of making progress on the debate about corporate moral agency. First, because it grounds corporate moral agency on outsiders’ capacity meaningfully to view the corporation as a social actor simpliciter, the second-personal approach sidesteps complex first-personal analyses of the ways in which corporate agency and responsibility might be constructed by aggregating the moral agency of the individuals who compose a corporation. Second, we argue that a second-personal analysis of the practice of holding responsible yields a new way to specify the first-personal criteria for corporate moral agency. A second-personal account of corporate moral agency therefore makes it easier to judge whether a corporate qualifies as a moral agent even under first-personal criteria.
The remainder of this article proceeds as follows. In section 2, we review the philosophical literature that argues that moral responsibility is associated with the social practice of holding, and being held, responsible. We identify three conditions that scholars working in this tradition view as essential for moral agency. The second-personal account of corporate moral agency is presented in section 3, where we describe the hurdles that must be cleared if the three conditions are to be satisfied for a corporate actor. Section 4 analyzes the relationship between first- and second-personal accounts of corporate moral agency, discusses the implications that the second-personal approach has for corporate policy, and discusses corporate punishment from a second-personal perspective. Section 5 concludes.
1. TWO APPROACHES TO CORPORATE MORAL AGENCY
One of the most important questions in business ethics is whether it is reasonable to view corporate bodies as being morally responsible for their actions, that is, as possessing moral agency. In this section, we contrast two ways of thinking about this question: a first-personal approach, which relies on methodological individualism, and a relational approach, which analyzes the way that corporations are experienced as social actors.
It appears obvious that a fully developed adult human being is a moral agent, but an attempt to argue that corporations are moral agents because they have all of the characteristics of a fully developed adult human is doomed to fail (Hess Reference Hess, Orts and Smith2017, 179): corporations cannot experience love and fear as humans do, for example. Proponents of corporate moral agency do not therefore start from an assumption of humanity (see, e.g., Donaldson Reference Donaldson1982; Manning Reference Manning1984; Ozar Reference Ozar1985; Goodpaster Reference Goodpaster1983; Moore Reference Moore1999; Phillips Reference Phillips1992). Instead, they attempt to identify formal characteristics that any moral agent must exhibit, and then they argue that those characteristics can meaningfully be found in corporate bodies. Scholarly discussion of corporate moral agency then asks what formal characteristics a moral agent must possess and whether any nonhuman agent can exhibit those characteristics.
The academic discourse that underpins standard accounts of corporate moral agency deploys a first-personal characterization of moral agency. First-personal accounts of moral agency are concerned with an actor’s capacities to deliberate for itself, to originate its own moral reasons, and to legislate for itself. Proponents of corporate moral agency argue that formal aggregation of human decisions and actions creates these first-personal capacities in an organization (French Reference French1979; Pettit Reference Pettit2007; List and Pettit Reference List and Pettit2011; Hess Reference Hess2010); skeptics claim that it cannot, because the individual decisions underlying both the design of the aggregation system and members’ acceptance of its outcomes are the only morally relevant choices (Keeley Reference Keeley1981; Velasquez Reference Velasquez1983, Reference Velasquez2003; Rönnegard and Velasquez Reference Rönnegard, Velasquez, Orts and Smith2017; Rönnegard Reference Rönnegard2015). In short, proponents and opponents of first-personal accounts of corporate moral agency hold such apparently irreconcilable metaphysical positions regarding the outcome of aggregation that it is hard to see how they could ever reach a consensus. If we are to make further progress, then it may be necessary to take a different perspective.
This article presents an analysis of corporate moral agency that builds on the social practice of holding corporations morally responsible. Following Strawson’s (Reference Strawson[1962] 1974) account of moral responsibility in individual human agents, we argue that, if it is appropriate to hold an organization responsible, then it must be morally responsible. That is, the practice of holding corporations morally responsible may provide its own justification. By this relational account, metaphysical freedom is not the foundation of corporate moral agency, although it may be a reasonable implication.
We begin our analysis in the next section by reviewing an important body of literature that studies the social practice of holding natural, human, persons responsible. This provides the foundation for the relational analysis of corporate moral agency in section 3; in section 4.1, we return to the first-personal account and examine its relationship to our approach.
2. THE RELATIONAL APPROACH TO INDIVIDUAL MORAL AGENCY
In this section, we discuss a functional characterization of individual moral agency that is grounded in social relations; our goal is to establish a framework that we can use to reason about corporate actors. The fundamental concept in our analysis is the practice of holding an agent responsible. We discuss the relevance of this practice, and we define a moral agent or, equivalently, a member of the moral community to be the sort of thing that can be held responsible (this is, of course, a distinct question from the quality of the conduct for which the agent should be held responsible). We conclude that an agent can be held responsible if and only if that agent has the competence to participate in a particular type of conversation. That competence is second-personal, and it requires a moral agent to satisfy three conditions: the Attitude, Address, and Authority Conditions. We discuss each condition in turn.
2.1 The Attitude Condition
We derive the Attitude Condition from the claim that moral responsibility can be defined in terms of the practice of holding responsible. This claim was first advanced in P. F. Strawson’s (Reference Strawson[1962] 1974) essay Freedom and Resentment. Strawson argues that we expect others to exhibit goodwill toward us in our personal relationships. In response to others’ goodwill or lack of it, we naturally experience reactive attitudes, such as love, resentment, anger, or indignation. It is possible to form relationships in which the reactive attitudes are suspended, for example, when dealing with a very young child or in situations where our approach to the relationship is instrumental rather than properly involved, but Strawson argues that this type of objective attitude is hard to sustain in relationships between capable and mature people. The reason is that reactive attitudes are such natural reactions to good and ill will that it is profoundly difficult to suppress them for long.
Strawson uses reactive attitudes to reconcile the practice of holding people responsible for their actions with the possibility that those actions are governed by deterministic laws, and his analysis sheds light on the nature of moral agency.Footnote 4 The act of holding someone responsible is associated with specific reactive attitudes like resentment, indignation, and guilt. Those attitudes are fundamental to our lived experience: even if we were intellectually convinced of the truth of determinism, we probably could not abandon them. Strawson argues from the fundamental character of reactive attitudes that any attempt to understand the metaphysical meaning of moral responsibility is “to over-intellectualize the facts”: the general framework of reactive attitudes is a fact of human society that “neither calls for, nor permits, an external ‘rational’ justification” (Strawson Reference Strawson[1962] 1974, 25). Strawson’s essay therefore points toward an approach to moral agency that is grounded in our interactions with one another, rather than directly in our individual freedom. We will refer to approaches derived from Strawson’s insights as relational.
The foundational nature of the reactive attitudes generates a first necessary condition for moral agency. Strawson argues that it makes no sense to form a reactive attitude toward an agent if that agent is not capable of bearing moral responsibility. Conversely, if the agent is capable of bearing moral responsibility, then it must be possible to form a reactive attitude toward the agent. This reasoning yields the following condition:
Attitude Condition: If an agent can be held morally responsible, then it must be possible for other agents to form reactive attitudes toward that agent.
2.2 The Address Condition
We derive the Address Condition from the claim that an agent can be viewed as morally responsible only when it is appropriate to address reactive attitudes toward that agent, that is, only when the agent has no excuse or exemption for its conduct. We therefore start our discussion with an examination of the circumstances that would either inhibit the formation of reactive attitudes or mollify or entirely dissipate an already-formed attitude (Strawson Reference Strawson[1962] 1974, 7–8). These circumstances can be classified as excusing or exempting conditions, depending on whether they demonstrate that the moral demand at stake was not in fact violated or that the agent is appropriately discharged from the moral demand (Watson Reference Watson and Schoeman1987, 259–61). Excusing conditions refer to temporary conditions under which a morally competent agent is not held responsible and so do not help us to determine what it means to be a moral agent. In this section, we therefore present an analysis of exempting conditions; exemptions are important because, as we argue in section 3.2, they have a bearing on the validity of the reactions that underpin a relational account for corporate moral agency. In particular, we ask in that section whether the corporation’s inability to experience emotions constitutes an exempting condition and hence violates the internal logic of reactive attitudes.
Relational approaches to moral agency start from the position that to hold an agent morally responsible is to believe that it would be appropriate to experience reactive attitudes when the agent violates a moral standard (see, e.g., Wallace Reference Wallace1994, 21), and it is therefore important that we can identify what reactive attitudes are “appropriate.” We have already noted that, because the relational approach uses reactive attitudes as the basic building blocks from which moral responsibility is constructed, we cannot define appropriate reactive attitudes in terms of a deeper metaphysical idea (autonomy or transcendental freedom, for example). Instead, Strawson (Reference Strawson[1962] 1974, 24–25) claims that reactive attitudes have an internal logic that dictates when they are appropriate. A reactive attitude formed when either an excusing or an exempting condition obtained would violate its internal logic. For example, a reasonable person cannot logically sustain moral reactive attitudes toward a severely cognitively impaired person.
This discussion sheds light on the relationship between the felt experience of an emotion and its justification. Suppose that John feels that other people would respond in a useful way if he were resentful toward Jane, who is severely cognitively impaired. John knows that Jane is exempted: can the emotion that he experiences toward her really be resentment? By Darwall’s (Reference Darwall2006, 15–17) account, John’s emotion cannot be resentment, even if the two are phenomenologically indistinguishable; this is because John’s emotion is motivated by a “reason of the wrong kind.” Darwall notes that it is impossible to come to believe a proposition p (say, “I am resentful” or “there is a God”) by reflecting on the desirability of a belief in p: this impossibility renders the desirability of a belief in p a reason of the wrong kind for believing p. John’s emotion is pragmatically useful, and so can be viewed as an objective attitude, but it cannot be a reactive attitude: John may behave as if he holds Jane morally responsible for reasons of social utility, but this is not the same as actually holding her so responsible (for a discussion, see Wallace Reference Wallace1994, 74–83).
We devote the remainder of this section to a conceptual analysis of the characteristics of an agent that is not exempted. Our first observation is that a moral reactive attitude does not simply record a judgment on past behavior; it also comes with the expectation that the target of the attitude will recognize the judgment and respond appropriately (Stern Reference Stern1974, 78). That is, reactive attitudes are “incipiently forms of communication” (Watson Reference Watson and Schoeman1987, 264). Meaningful communication requires an interlocutor who understands and is capable of applying the principles that underpin the communicator’s practical reasoning (Stern Reference Stern1974, 77–78; Watson Reference Watson and Schoeman1987, 267–68); in other words, the interlocutor must possess “normative competence” (Doris Reference Doris2002, 136; Shoemaker Reference Shoemaker2007, 71). We tease out the content of that competence later, but it is clear at this stage that it must at least involve an ability to engage in dialogue (Stern Reference Stern1974, 79). Expressions of anger against entities that cannot engage in dialogue is common (for discussions, see Sepinwall Reference Sepinwall2011, 414; Reference Sepinwall2016, 6; Wolf Reference Wolf1985), but, by a relational account, that type of anger is not a reactive attitude. Hence, for example, when a person’s house is destroyed by a hurricane, the resultant anger cannot presuppose the hurricane’s moral agency, because the hurricane cannot engage in dialogue with the person. We extract the following condition for moral agency from this argument:
Address Condition: If an agent can be held morally responsible, then it must be possible to trigger a dialogue by addressing reactive attitudes to it.
The Attitude and Address Conditions are not (jointly) sufficient for a relational account of moral agency, because some people satisfy both of them and yet sit outside the moral community. A standard example is the psychopath, who is the subject of particularly strong emotional responses, in part because psychopaths frequently refuse to deploy their capacity for dialogue (Watson Reference Watson and Schoeman1987, 267–75). Psychopaths therefore appear to satisfy the Attitude and Address Conditions. But some authors have nevertheless argued that psychopaths are not members of the moral community. Although they are capable of engaging in dialogue, psychopaths are not moved by moral reasoning; indeed, they may be moved deliberately to violate moral standards by a desire to harm others (Shoemaker Reference Shoemaker2007, 97). Some commentators have attempted to explain what missing quality excludes psychopaths from the moral community by arguing that they are incapable of understanding moral reasons, but Shoemaker (Reference Shoemaker2007, 77–85) argues that, on the contrary, they can understand those reasons; they are, however, incapable of being moved by them because they cannot care about other people. He distinguishes caring from empathizing, because he claims that high-functioning autistics can be moral agents despite an inability to empathize with other people in the usual way (Shoemaker Reference Shoemaker2007, 93): he claims that autistic people can figure out moral demands “by alternate routes” and then be motivated to act accordingly (Shoemaker Reference Shoemaker2007, 96–97). In later work, Shoemaker (Reference Shoemaker2011) presents three different conceptions of responsibility and uses them to distinguish between the responsibilities of different agents at the boundaries of the moral community. We draw on the work of the moral philosopher Stephen Darwall to identify an alternative unifying account of moral responsibility and moral agency.
2.3 The Authority Condition
We derive the Authority Condition from the conception of normative competence presented in Stephen Darwall’s (Reference Darwall2006) book The Second-Person Standpoint. Darwall’s central point is that moral agents acknowledge that they stand in relations of authority to one another and that those relations generate practical reasons for action. A moral agent must therefore be capable of this type of acknowledgment. Darwall refers to this capability as second-personal competence. Our functional characterization identifies three elements that are required for an agent’s second-personal competence: valid sources, second-personal deliberation, and appropriate response. We discuss each element in turn, and we conclude with the formal statement of the Authority Condition.
2.3.1 Valid Sources
If John is to recognize Jane’s moral authority, then he must acknowledge that she is entitled to make certain claims of him. When Jane makes a demand of John, he should therefore view her as creating a new moral reason, rather than simply pointing out facts that would have existed without her intervention (Darwall Reference Darwall2006, 57–60): Jane is a source of authority, rather than a medium between John and morality (Shoemaker Reference Shoemaker2007, 89).
To understand better what it means for one person to view another as a valid source of moral reasons, suppose that John’s behavior toward Jane is interpreted by Jane as harassment and that she informs John of this fact. John could decide that it is undesirable in general that anyone should experience harassment and, for this reason, cease his harassment of Jane. In drawing this conclusion, John would be exercising a first-personal capacity for independent reasoning. But his reasoning would be agent-neutral, and would not really acknowledge that Jane qua Jane has a particular moral standing with regard to John. In contrast, John could recognize Jane’s moral authority and correlatively appreciate, as a free moral agent himself, that Jane has a claim on his will: this is what it means to view Jane as a source of moral reasons (see Darwall Reference Darwall2006, 5–10).Footnote 5 John’s reasons for acting in this case are agent-relative and therefore inherently relational; they can be contrasted to those deriving from advice, logic, or solely from science, which do not depend on Jane’s possession of any particular authority with regard to John (Darwall Reference Darwall2006, 59).
John’s acknowledgment of Jane as an independent source of valid moral reasons has implications for the way that Jane communicates with John: her communication is addressed to John, rather than at or about him (Darwall Reference Darwall2006, 40), and she expects John to acknowledge her authority to address him. When she communicates in this way, Jane is making a second-personal address; she could do so through an explicit speech act (“you treated me with disrespect then; why did you do it?”) or implicitly, for example, by assuming a resentful attitude (Darwall Reference Darwall2006, 3). This type of dialogue requires the interlocutors to take up the second-person standpoint, which is “the perspective [people] take up when [they] make and acknowledge claims on one another’s conduct and will” (Darwall Reference Darwall2006, 3). When Jane and John take up this perspective, they attribute moral agency to one another, acknowledge each other’s moral freedom, and presuppose that they have the authority to engage in moral discourse.Footnote 6
2.3.2 Second-Personal Deliberation
The fact that Jane makes a second-personal address to John does not automatically obligate John to do precisely what Jane would like him to do, but it does obligate him to give Jane’s reasons due consideration in his deliberation. How, then, should John deliberate? Darwall argues that John should view himself independently of his own beliefs and desires as another person would; that is, he should adopt the second-person standpoint toward himself. The resultant second-personal deliberation therefore gives weight to Jane’s reasons because they came from Jane, and not because of any good or bad consequences that John might expect to experience if he were to ignore them (Darwall Reference Darwall2006, 225–28, 249–52). In this way, John comes to a conclusion that he views as just and so is able to hold himself responsible for the actions that caused Jane’s second-personal address, for example, by appreciating that it would be appropriate for him to experience guilt (Wallace Reference Wallace1994, 23; Darwall Reference Darwall2006, 111–12). Engaging in second-personal deliberation also enables John to hold himself responsible for his future actions by electing to bind himself in accordance with his deliberation (Darwall Reference Darwall2007, 55–56).
2.3.3 Appropriate Response
When John recognizes Jane’s moral authority, he acknowledges that she is owed a response to her address. John may conclude from his second-personal deliberation that, notwithstanding Jane’s claim, his actions were justified. In this case, he acknowledges her moral authority by providing a valid justification for his actions. If Jane accepts John’s justification, her initial reactive attitude will be modified or may disappear altogether; if she does not, she may continue her dialogue with John with a fresh expression of a reactive attitude.
If John concludes that he cannot justify his action and hence that it would be appropriate for him to experience guilt, then he acknowledges Jane’s moral authority by admitting his fault (“I understand the harm that I did you; I was wrong, and I am sorry”). John should make amends to Jane and change his future behavior (Darwall Reference Darwall2006, 70–79).
2.3.4 Formal Statement of the Authority Condition
We are now in a position to present the last of the three conditions for moral agency under a second-personal account:
Authority Condition: If an agent can be held morally responsible, then it must be able to acknowledge the moral authority of other moral agents. That is,
-
(i) the agent must be able to recognize others as valid sources of moral reasons;
-
(ii) the agent must be able to perform second-personal deliberation;
-
(iii) the agent must be able to respond appropriately to second-personal addresses.
As presaged earlier, we can use the Authority Condition to distinguish between the difficult borderline cases of psychopathy and high-functioning autism without relying, as Shoemaker’s (Reference Shoemaker2007, Reference Shoemaker2011) analysis does, upon either emotions or a distinction between different conceptions of responsibility. By our account, psychopaths are not moral agents; they cannot be moved by moral reasons, because they do not acknowledge the moral authority of other people. In contrast, despite their inability to experience standard emotional responses, high-functioning autistic people are able to acknowledge others’ moral authority and hence are moral agents.
This section presented a relational view of moral agency. Under the relational view, an agent is held responsible whenever it is appropriate to form a reactive attitude toward that agent. Reactive attitudes are appropriate if there are no excusing or exempting conditions; we focus on the latter, because they identify actors who cannot ever be held morally responsible for their actions and hence are not members of the moral community. We argue that Darwall’s (Reference Darwall2006) second-person standpoint provides the framework required to delineate the contours of moral agency. By Darwall’s account, a person is a valid target of reactive attitudes if and only if he or she is second-person competent, that is, if and only if the person is able to acknowledge the moral authority of other agents and to respond appropriately to their addresses of reactive attitudes (Darwall Reference Darwall2006, 70–79). People who are not second-person competent are therefore exempted from moral responsibility, and our emotional responses to those people should reflect an objective attitude; people who possess that competence can legitimately be held responsible for their actions. The second-person standpoint underpins a theory of moral agency in human persons. We will deploy it to understand how, in a relational context, corporations can possess moral agency.
3. A SECOND-PERSONAL ACCOUNT OF CORPORATE MORAL AGENCY
Section 1 of this article contrasted first-personal and relational approaches to corporate moral agency. Arguments conducted on first-personal grounds are concerned with the corporation’s capacity to reason for itself and to act in response to its reasoning. Under this approach, holding responsible is a social practice that is invoked when an agent fails to honor an already identified moral obligation. In contrast, the relational approach grounds moral obligations in the practice of holding responsible. Section 2 discussed the relational approach to individual moral agency and combined theories based on the reactive attitudes with the second-person standpoint. We presented necessary and sufficient conditions for second-personal competence, which, we claimed, is essential for moral agency. We now ask how these conditions could be applied to a corporation and hence what it would mean for a corporation to be a moral agent under a second-personal approach. We start our analysis by presenting formal conditions for a corporation to possess second-personal competence; we then discuss the meaning of each condition and present the main obstacles to their satisfaction.
3.1 Corporate Second-Personal Competence Conditions
In line with the analysis of section 2, we say that a corporation has second-personal competence if and only if the following corporate second-personal (CSP) conditions are satisfied:
-
C-1. Attitude Condition: If a corporation can be held morally responsible, then it must be possible for other agents to form reactive attitudes toward that corporation.
-
C-2. Address Condition: If a corporation can be held morally responsible, then it must be possible to trigger a dialogue by addressing reactive attitudes to it.
-
C-3. Authority Condition: If a corporation can be held morally responsible, then it must be able to acknowledge the moral authority of other moral agents. That is,
-
(i) it must be able to recognize others as valid sources of moral reasons;
-
(ii) it must be able to perform second-personal deliberation;
-
(iii) it must be able to respond appropriately to second-personal addresses.
By analogy with the second-personal theory of individual moral agency presented in the previous section, we claim that the CSP conditions are necessary and sufficient for a corporation to be deemed a moral agent. We do not attempt to argue in this article that these conditions are actually satisfied for any corporation; instead, we ask what it would take for each of the conditions to hold.
3.2 The Attitude Condition
The Attitude Condition is satisfied when outsiders can form reactive attitudes toward the corporation. It is certainly the case that people speak as if they form reactive attitudes toward corporations. For example, a year after the events described in our opening paragraph, social media commentators spoke as if they held United responsible for the treatment of Dr. Dao. A March 2018 tweet stated that “I know the internet quickly and quietly forgot about what happened to David Dao, but I haven’t. And I’ve flown several times wherein I’ve paid extra money by siphoning @United out of my price search because I refuse to spend a penny on their toxic [expletive] company.”Footnote 7 A skeptical position holds that the use here of “their” is evidence that this type of statement deploys a linguistic convention that references United as a convenient shorthand for detailed statements about multiple individuals working for United, but exercising their own agency (Maitland Reference Maitland, Orts and Smith2017; Rönnegard and Velasquez Reference Rönnegard, Velasquez, Orts and Smith2017; Sepinwall Reference Sepinwall, Orts and Smith2017; Velasquez Reference Velasquez1983). But, to satisfy the Attitude Condition, we need not establish that this linguistic convention is not employed, nor even that it is not commonly employed; we need only establish that it is possible and meaningful to form a reactive attitude toward a corporate body.
Some authors argue that it is impossible to form a meaningful reactive attitude toward a corporation because, first, a reactive attitude requires an appropriate emotional response and, second, a corporation cannot experience that type of response. That is, following section 2.2, they hold that an inability to experience emotions qualifies as an exemption that rules out corporate moral responsibility. For example, Sepinwall (Reference Sepinwall, Orts and Smith2017) claims that we form reactive attitudes to engender feelings of guilt, because the discomfort experienced by a guilty agent leads the agent to change its future behavior. Shoemaker (Reference Shoemaker2019) argues that we blame in order to induce remorse, which he defines to be “the painful acknowledgement of what I did to you from your perspective” (907, emphasis original); this pain is necessary because it allows the wronged party to withdraw its reactive attitude and to forgive its interlocutor. Shoemaker (Reference Shoemaker2019, 970–80) argues that remorse, conceived as an internalization of the other party’s perspective, is necessarily phenomenological and that it cannot, if sincere, have an instrumental purpose.
Sepinwall and Shoemaker advance strong arguments in favor of what we will refer to as the affect position: that the targets for reactive attitudes must have a capacity for affect (affect is referred to throughout Sepinwall’s account of the moral attitudes). If we accept the affect position, then, if it turns out that corporations cannot experience affect, we must also reject the claim that corporations can be moral agents. The affect position is therefore a central point of contention in the study of corporate moral agency.
The claim that corporations cannot experience discomfort as people do seems compelling; Sepinwall (Reference Sepinwall, Orts and Smith2017, 146) therefore concludes that “even if the corporation is a moral agent, it is not an appropriate target for blame,” and similarly, Shoemaker (Reference Shoemaker2019, 898) argues that corporations can be blameworthy but that they cannot be the subject of angry blame. But a convincing argument that corporations cannot experience affect is not enough to show that a corporation cannot be the target of an appropriate reactive attitude; that is, it is not enough to establish the affect position. The reason it is not enough is that, in a relational account, blameworthiness derives from the affective experience of the agent who forms an appropriate reactive attitude and not necessarily from the feelings of the target of those attitudes; we discuss this point with regard to high-functioning autistic people in section 2.3.4. This is a distinctively second-personal point, because it is concerned with the relationship between an actor and the corporation being held to account and not directly with the internal, first-personal characteristics of the corporation.
We suggest that arguments regarding the affect position should focus on the internal logic of the reactive attitudes. Recall from section 2.2 that the internal logic of a reactive attitude is violated whenever an excusing or exempting conditions obtains, in which case the attitude is inappropriate. Silver (Reference Silver2005) uses an argument about internal logic to identify the situations in which a reactive attitude toward a corporation is rationally warranted. Silver argues, in line with our earlier discussion of Strawson’s essay, that corporations do not need free will to be morally responsible: he analyzes his own reactive attitudes toward corporations and concludes that they require the corporations to possess neither corporeality nor consciousness. We discuss the internal logic of reactive attitudes with regard to Sepinwall’s and Shoemaker’s positions in turn.
Sepinwall (Reference Sepinwall, Orts and Smith2017, 153–54) emphasizes the punitive element of blame, which, she argues, derives from the negative appraisal inherent in blame and in the agonizing self-reflection that it engenders (see also Sepinwall Reference Sepinwall2011, 428; Reference Sepinwall2016, 8). Recall that Strawson (Reference Strawson1968, 4) claims that the socially desirable consequences of the practice of holding responsible are not an appropriate basis for that practice. According to Strawson, desirability and efficacy are external to the logic of the reactive attitudes and hence can only be the basis for objective attitudes. An important question with regard to Sepinwall’s analysis is therefore whether a justification for blame in terms of its punitive effect is a part of the internal logic of the reactive attitude underpinning blame or is external to that logic. If the punitive element of blame is an external justification, then Sepinwall’s argument has implications for the type of objective attitudes corporations can attract, but not for moral agency. If punishment is intrinsic to blame, then she has a strong argument against the Attitude Condition and hence against the possibility that corporations can be moral agents under the second-personal account.
Shoemaker (Reference Shoemaker2019) presents an alternative account of the self-reactive attitudes triggered by blame: he argues that angry blame demands a response that can lead to forgiveness and hence to the withdrawal of the initial reactive attitude. For Shoemaker, forgiveness requires the blamed party to understand its behavior from the perspective of the wronged party and to experience the associated pain of remorse. This type of understanding is, Shoemaker argues, dependent on the blamed party’s ability to interpret the content of the wronged party’s emotional communications (on this point, see McKenna Reference McKenna2006, 26–29); both this and the ability to experience remorse depend on a capacity for affect. Shoemaker concludes that it is impossible to direct anger toward corporate bodies, and if he is correct, our Attitude Condition for moral agency is violated.
In contrast to Sepinwall, Shoemaker (Reference Shoemaker2019, 904n37) believes that functional analogues to guilt exist, but he argues that there is no such analogue for remorse, which, he claims, is inseparable from affect. Moreover, Shoemaker (Reference Shoemaker2019, 912) appears to view the affective component of remorse as an acknowledgment of moral authority. Such an acknowledgment must be sincere if it is to hit its mark, and Shoemaker’s (Reference Shoemaker2019, 916) key contention is that, if an expression of remorse has no affective component, the angry party can never be sure of its sincerity. If his argument is correct, then Shoemaker has established that the corporation cannot be a target of at least one important reactive attitude and hence that the Attitude Condition is violated. But Shoemaker’s argument rests on the impossibility of establishing sincerity absent an emotional response, which, he claims, serves as a commitment device. A natural response is to note that certainty of another agent’s sincerity requires privileged first-personal epistemic access to that agent, which, as a matter of fact, we have neither to corporate nor to human agents. The best we can hope for is a credible signal; indeed, Shoemaker acknowledges at the end of his article that such a signal is at least plausible for corporate agents, and hence he implicitly allows that the Attitude Condition could be satisfied for those agents.
In summary, our response to the affect position is to present an account of reactive attitudes toward corporations whose internal logic allows blamed parties to achieve an adequate response in ways that need not entail affect. This type of account holds that the lack of any capacity for affect does not qualify as an exempting condition and hence that reactive attitudes toward corporations may be appropriate. Björnsson and Hess (Reference Björnsson and Hess2017) argue that corporations can adequately respond to a reactive address if they are able to find ways to process its epistemic content and to respond noninstrumentally; although each expresses skepticism that this happens in practice, both Shoemaker (Reference Shoemaker2019, 916–17) and McKenna (Reference McKenna2006, 32) allow that it may conceivably be possible. The internal logic on which we rely in our response is derived from the second-personal competence required by the Authority Condition.
If, for the sake of argument, we allow that the Attitude Condition is satisfied, then it is interesting to ask under what circumstances reactive attitudes expressed toward a corporation are most likely to refer to the corporation, rather than to individuals within the corporation. The first and most obvious requirement is that the corporation be visible: a firm whose existence is hidden from its customers cannot be the target of a reactive attitude. The second requirement is that the corporation have some sort of formal governance procedures so that outsiders can believe that the firm’s activities are unlikely to change dramatically when a person leaves. The final requirement is one of scale: for reactive attitudes in very small firms, it is easy to attribute the collective’s choices to specific individuals (McKenna Reference McKenna2006; Sepinwall Reference Sepinwall2011, 424). This point is illustrated by the Colorado bakery Masterpiece Cakeshop, which is an LLC that precipitated a Supreme Court case when it refused to sell wedding cakes to same-sex couples. Masterpiece is a small company, and the reactive attitudes that this case inspired appear largely to have been directed toward the firm’s owner, Jack Philips (see, e.g., Delfin Reference Delfin2018). In contrast, actions appear more likely to be identified with corporations when they are very large and have strong brand identities; it is toward these corporations that reactive attitudes seem most likely to form, as they did in the United Airlines case discussed in the introduction. Such corporations have procedures and cultures that are robust to staff turnover; Silver (Reference Silver2005) argues that, in such an organization, corporate culture is the relevant disposition that renders reactive attitudes warranted (for a closely related economic argument, see Tadelis Reference Tadelis1999).
3.3 The Address Condition
The Address Condition is satisfied when it is possible to trigger a dialogue by addressing reactive attitudes to a corporation. To understand the ways in which actors express reactive attitudes to corporations, we consider the different ways that organizations and people receive messages. We argue that, in contrast to a human person, one can only address a reactive attitude to a corporation if an appropriate forum for that address exists. We conclude that corporate moral agency requires that each person, or the person’s representative, have access to a forum and that a corporation’s forums be wired into its formal decision-making procedures. Finally, we discuss the roles of the state, of third parties, and of the organization in creating and managing access to forums.
Talking to a biological person is different from talking to a corporate actor. It is possible to express moral indignation to natural persons by speaking to them, but it is only possible to address a corporation if a suitable forum exists. A forum is a medium for expressing opinions about a subject corporation. Forums are created by someone. We call a forum corporate if it is created by the subject corporation, it is third party if it is created by any nonstate actor other than the subject corporation, and it is a state forum if it is an organ of the state. For example, a contact form on a company website is a corporate forum, Twitter and other social media platforms are third-party forums, and the courts are state forums.
If a subject corporation’s forums are to enable dialogue, then they must satisfy two validity requirements. First, the forums must collectively satisfy an access requirement. To understand this point, recall that Darwall’s account of moral agency is grounded in the capacity of human agents to engage in a particular form of dialogue. Darwall assumes that it is always possible to trigger this type of dialogue, so that, if John is harmed by Jane, either Jane or her representative can call John out and so induce him to assume the second-personal perspective. But it may not be possible for some types of actors to trigger an appropriate dialogue with a nonhuman agent. Accordingly, the access requirement is that at least one of the subject corporation’s forums is available to every agent or to representatives of those agents. Second, each forum must satisfy a procedural requirement that the corporation has some formal process for receiving and considering the reactive attitudes expressed on the forum. This process is necessary if the forum’s participants are to address their attitudes to the corporation, rather than simply to speak to one another about it.
The access requirement implies that, under the relational theory, a corporation cannot be viewed as a moral agent if it can only receive addresses via shareholder meetings and analyst calls. Access to these forums is restricted to a limited set of people, and hence they exclude many actors who have a moral right to address the corporation. In most developed countries, state forums like the courts are open to everyone. However, accessing those forums can be expensive, may require professional representation, and is frequently intimidating.Footnote 8 State forums are complemented by third-party forums: these can be as diverse as a free press, nongovernmental organizations (NGOs), and social media. Better access to third-party forums increases the likelihood that the access requirement will be satisfied and so strengthens the case for corporate moral agency. An immediate consequence of this is that states that restrict third-party forums, for example, by undermining NGOs or by blocking access to social media platforms, weaken corporate claims to moral agency. This observation, which predicates corporate moral agency on the combined behavior of multiple actors in their social context, is explicitly relational and does not emerge from the traditional requirements of first-personal accounts.
The procedural requirement ensures that participants in a forum are speaking to its subject corporation, rather than speaking to one another about it. The subject corporation must capture the content of the forum and must ensure that the content is considered in line with the company’s formal decision-making procedures. For example, a company that ignores Twitter comments renders Twitter an invalid forum for itself. And hiring a person to read Twitter comments is on its own insufficient: Twitter is an invalid forum for the company if that reader is not required to funnel comments into the company’s governance organs.
3.4 The Authority Condition
The Authority Condition is satisfied for a corporation when it is able to acknowledge the moral authority of other moral agents. In turn, this implies that the corporation be able to recognize others as valid sources of moral reasons, that it can perform second-personal deliberation, and that it can respond appropriately to second-personal addresses. We discuss each of these requirements in turn.
3.4.1 Valid Sources
A corporation is capable of viewing other moral agents as valid sources of moral reasons to the extent that it is able to weigh addresses from those agents in its deliberation because they come from those agents. This is different from accounting for the opinions of others insofar as they affect its bottom line, for example, by impairing its reputation in a way that reduces sales. A corporation whose governance systems, policies, and culture force it to view the opinions of another agent as important not because they come from that agent but because they have a potential instrumental effect fails to view the agent as a valid source of moral reasons and hence violates the Authority Condition.
Precisely because a corporation’s ability to treat others as valid sources of moral reasons depends on its governance systems, policies, and culture, it is extremely hard to observe this capability directly. As a practical matter, one way to form a view on this ability is to examine the reasons that the corporation gives for its actions. For example, a corporation might claim that a decision was taken solely because it was requested by its host community; that is, the corporation might advance the type of reason characterized above as agent-relative. Of course, corporate explanations can be insincere, as can those of human agents, but if the claim is credible, then there is a strong case that this corporation is capable of acknowledging other agents’ moral authority. If, on the other hand, the corporation always justifies itself with reference to profit maximization, then its espoused reasons are agent-neutral. Two conclusions are possible in this case: either the corporation is incapable of recognizing others’ moral authority and hence, like the psychopath of section 2.2, is not itself a moral agent or the corporation has this capability but does not exercise it, in which case it is simply acting immorally. A more reliable way to form a view on the way that a corporation views others’ addresses is to observe its day-to-day operations and so to draw a conclusion on the way that its procedures and cultural values inform its practical deliberations.
3.4.2 Second-Personal Deliberation
A corporation is capable of second-personal deliberation to the extent that it can adopt the second-person standpoint toward itself. There are good reasons to believe that a human person can view herself from a second-personal perspective and form self-reactive attitudes (Darwall Reference Darwall2006, 23). But a corporate entity does not have human mental states and it cannot experience emotions; it follows, as in our discussion of the Attitude Condition, that this element of the Authority Condition can be satisfied only if the corporation has the functional capacity to view itself independently of its own (functional) beliefs and desires.
We suggest that the best argument that corporations can possess the functional capacity for second-personal deliberation is, first, that the members of corporations have that capacity and, second, that corporations facilitate a type of distributed cognition that synthesizes the second-personal deliberations of their members. This is analogous to an approach that Tollefsen (Reference Tollefsen2003, 231–32) and Björnsson and Hess (Reference Björnsson and Hess2017, 289–92) use to explain the amalgamation of emotional judgments in corporations and that is characterized by Sepinwall (Reference Sepinwall, Orts and Smith2017, 157) as a “recruiting account.” We accept that it would make no sense to claim that, because its members have emotions, a corporation can be treated as capable of affect (see Sepinwall Reference Sepinwall, Orts and Smith2017, 146; for a contrasting opinion, see Gilbert Reference Gilbert2000). But we claim that this element of the Authority Condition stands or falls on the corporation’s ability rationally to combine the second-personal judgments of its members; it may conclude that, were it a human agent, the emotions associated with guilt would be appropriate, but to satisfy the functional requirements of second-personal deliberation, it need not actually experience that guilt. Indeed, the fact that the corporation relies on the collective deliberation of multiple agents suggests that its attempts to adopt the second-personal perspective may actually be more successful than those of any individual person: multiple insights are aired in group deliberation, which frequently also serves to neutralize individual subjective biases (for related arguments, see Shoemaker Reference Shoemaker2019, 903–4, and Tollefsen Reference Tollefsen2003, 228; Reference Tollefsen2015, 129). For example, a diverse board may be better able to adopt the second-personal perspective when considering addresses from underrepresented groups than any single person could, and individual biases are less likely to survive the deliberations of such a board.
We have argued that the second-personal deliberation element of the Authority Condition can be satisfied if and only if group deliberation produces a functional equivalent of individual second-personal deliberation, and we have advanced reasons to believe that it might. A related question is how a particular corporation’s capacity for second-personal deliberation could be assessed. Just as a corporate capacity to view others as valid sources of moral reasons is hard to observe, so, too, is the capacity for second-personal deliberation. In both cases, the most direct assessment is achieved by observing the corporation’s day-to-day operations, although this type of observation is costly and time consuming. A corporation that provides credible agent-relative reasons for actions provides prima facie evidence that it has performed second-personal deliberations (as it does of its ability to view others as valid sources of reasons). This type of deliberation appears to have underpinned Rio Tinto’s approach to local communities in the 1990s: Cochrane (Reference Cochrane2017, 4) reports that Rio Tinto staff “seemed to respect local communities rather than viewing them as problems to be dealt with.” Our earlier observation that an absence of evidence does not imply an absence of moral agency also applies here.
3.4.3 Appropriate Response
A corporation can respond appropriately to second-personal addresses to the extent that it can explain itself, make amends for its actions if necessary, and change its future behavior. We discuss each of these competences in turn.
If a corporation concludes after second-personal deliberation that its actions were justified, then moral agency requires that it be able to explain itself. Corporations are unable to speak in the way that a human person does, and, as earlier, we therefore require a functional equivalent to a human speech act. We follow Pettit (Reference Pettit, Orts and Smith2017) and argue that this is achieved when corporate procedures identify a person who has the formal authority to speak for the corporation, so that the corporation can understand and explain itself in words: formally, the corporation is a conversable agent (Pettit and Smith Reference Pettit and Smith1996; Pettit Reference Pettit2002, Reference Pettit2007). That is, appropriate response requires conversability, and hence so does corporate moral agency.
If, on the other hand, a corporation concludes that its actions were not justified, then moral agency requires that it be able sincerely to acknowledge its fault. This is arguably what happened when, in 2015, the Coca-Cola chief executive Muhtar Kent responded to negative comments about the firm’s approach to scientific research into obesity by acknowledging fault and committing to improved transparency and to overhauling the firm’s governance (Kent Reference Kent2015). As discussed, sincerity here cannot derive from a capacity for affect and must be identified functionally. An organization cannot be viewed as second-person competent if its procedures and policies require it always to deny liability or to suppress uncomfortable information. Casual empiricism suggests that individuals understand this: the public responds far less positively to an organization that appears insincere than to one that appears genuinely to engage with feedback.
After acknowledging fault, a corporate moral agent must be able to make amends and to commit to changing its future behavior. This statement derives from Darwall’s (Reference Darwall2006) observation that, when one person (John) acknowledges the moral authority of another (Jane), he accepts her reasons as his own and then binds himself accordingly. We discuss the requirement that a corporate moral agent be able to accept another’s reasons earlier; appropriate response therefore reduces to the requirement that the corporation be able to commit itself by adopting the appropriate policies, procedures, or cultural mores. Whether it can do so is an empirical question upon which this element of the Authority Condition hinges. Consistent with Shoemaker’s (Reference Shoemaker2019, 916–17) analysis, the corporation demonstrates its sincerity when it makes amends and commits itself to change. In related work on responsive adjustment, French (Reference French1984) argues that if a corporation does not change, others may reasonably view its past actions as intentional and legitimately impose sanctions; he thus implicitly assumes that corporations can acknowledge their wrongs and attempt to fix them.Footnote 9
4. DISCUSSION AND IMPLICATIONS
Section 3 provides a second-personal account of what it would take for a corporation to be a moral agent. We do not claim that every, or indeed, any, corporation necessarily satisfies the CSP conditions that we present, but we believe that our approach can generate new insights into first-personal approaches to corporate moral agency, into the ways that corporate policy affects moral responsibility, and into the differences between the legal liability and moral responsibility of corporations. This section considers each of these points in turn.
4.1 Relationship between the First- and Second-Personal Approaches to Corporate Moral Agency
We have presented a second-personal account of corporate moral agency that is founded on the corporation’s ability to engage appropriately with other moral agents. We now discuss the relationship between the second-personal approach of section 3 and the first-personal reasoning that dominates the existing literature on corporate moral agency. We make two points. First, we argue that, because the second-personal account identifies the way that moral reasoning is deployed, it generates fresh insights into why the first-personal requirements for moral agency are important. Second, we demonstrate that it is possible for a firm to satisfy the first-personal requirements for moral agency without satisfying the CSP conditions; we do this by exploring the role of corporate visibility and the importance of cultural norms.
4.1.1 Implications of Corporate Second-Personal Competence for First-Personal Requirements
The second-personal approach yields a fresh way to develop the first-personal requirements for moral agency. A corporate actor that satisfies the CSP conditions is a suitable target for reactive attitudes precisely because it is capable of recognizing the moral authority of other actors and of deliberating appropriately in response to their second-personal addresses. This type of deliberation demands that the corporation be capable of the autonomous reasoning required to identify the morally correct response to second-personal addresses and of acting appropriately in response to that reasoning. These are precisely the capacities that are stressed in French’s (Reference French1979) classic treatment of corporate moral agency and in recent important work on corporate autonomy (Pettit Reference Pettit2007; List and Pettit Reference List and Pettit2011). Hence the first-personal conditions for moral agency are implied by the CSP conditions. Cashing out first-personal competence from the CSP conditions yields additional insights, because it indicates the role that first-personal competence plays in the social interactions upon which moral agency rests. Our analysis in section 3.2 of the internal logic of the reactive attitudes is an example of this type of cashing out, because it provides insights into the relationship between affect and corporate moral agency that cannot be achieved using a purely first-personal approach.
4.1.2 Satisfaction of First-Personal Requirements Do Not Imply CSP Conditions
We now show that an organization that satisfies the first-personal requirements need not satisfy the CSP conditions. We present two examples of organizations that could be viewed as moral agents under a first-personal approach but that lack second-personal competence. The first organization does not satisfy the Attitude or the Address Condition; the second violates the Authority Condition.
First, consider a secret society that has a particular purpose, a constitution, and a mechanism for deliberating over its actions and for incentivizing its members. Such a society possesses the formal systems for generating and recognizing organizational choices that, by French’s (Reference French1979) account, create corporate moral agency. That is, the secret society satisfies first-personal criteria for moral agency. But the society fails to satisfy the CSP conditions. The reason is that, precisely because the society is secret, it cannot be the target of reactive attitudes and, again because it is secret, it is impossible for other moral agents to address the society. The society therefore violates the Attitude and Address Conditions, and hence, from a second-personal perspective, it is not a moral agent.
Second, consider the imagined charity ImpactCo. ImpactCo runs schools in the developing world, and their graduates have better-than-average life chances. All decision-making at ImpactCo is driven by a strong cultural belief that it should protect its ability to continue to operate at all costs; indeed, it is impossible for most members of ImpactCo to imagine another way of doing business.
Now imagine that ImpactCo learns that children at one of its schools have been sexually abused by a member of staff. ImpactCo appreciates that a terrible crime has been committed; if it were publicized, its long-term educational mission would be compromised. It weighs up the desirability of achieving justice for the abused children against the costs of reduced access to funds should the abuse come to light. It concludes that revelation is unlikely and is guided by its cultural beliefs to cover up the abuse in order to protect its image.
ImpactCo satisfies the first-personal requirements for corporate moral agency: it can make choices and stick to them, and, as witnessed by its consideration of the trade-off that it perceives between justice and its long-term mission, it is able to use reason to choose and to revise its goals. But ImpactCo’s strong culture means that it can only acknowledge the victims of abuse insofar as they have an instrumental effect upon its long-term mission. Consequently, it fails to acknowledge their moral authority and so fails to satisfy the Authority Condition.
We have presented two organizations that satisfy first-personal criteria for moral agency but that lack second-personal competence. Each of those organizations has the capacity to reason freely, to commit itself, and to act in pursuit of its goals. But neither is able to engage in the social interactions that, by a second-personal account, underpin moral agency. In both cases, this lacuna could be addressed by policy. We therefore turn now to a discussion of the relationship between corporate moral agency and corporate policy.
4.2 Corporate Policy and Moral Responsibility
Our analysis identifies the key points on which the satisfaction of the CSP conditions for corporate moral agency hinges and indicates that the satisfaction of several of the conditions depends on policy choices, either by the corporation or by another actor, such as the state. We now discuss possible reasons to believe that policy ought to encourage corporate moral agency, and we outline some policies that could facilitate this.
The reason to believe that corporations should be moral agents is that, if they were, then they could be more likely to take the sorts of actions that we would like them to take, even if those actions are not legally mandated. A corporation that is second-person competent appreciates the effect that it has on other agents and can reason about that effect from their perspective; it is better able to engage with other moral agents and, through its dialogue with those agents, is more likely to find a morally satisfactory solution to complex problems. That is, moral agents have less need of the law to identify right actions. This observation is in contrast to the claim that moral agency is required in order that (at least some) criminal sanctions can justly be applied to corporations; we argue in the following pages that such sanctions do not require moral agency.
We now outline the ways that a corporation can render itself morally responsible according to the second-personal account; the appropriate policies follow from the CSP conditions. A corporation is more likely to satisfy the Attitude Condition if it enables individuals to form reactive attitudes toward itself by creating a strong brand identity and ensuring that it acknowledges authorship of its actions. To ensure that it satisfies the Address Condition, the corporation can ensure that it creates, or participates in, appropriate forums for the expression of reactive attitudes. Those include the reactive attitudes of its employees, so that the corporation should also facilitate whistleblowing. Participation in forums entails the creation of processes that receive attitudes transmitted in the forums and that ensure that such attitudes are adequately considered by the corporation’s decision-making bodies. As noted, the Authority Condition is complex, because at least valid sources and second-personal deliberation depend on the organization’s culture, and hence efforts to induce moral agency must attempt the difficult task of culture management. But the increasing regulatory emphasis on “tone at the top” in the financial sector at least indicates that this type of management is viewed as a valid target of regulatory intervention (see, e.g., Patelli and Pedrini Reference Patelli and Pedrini2015; Lail et al. Reference Lail, MacGregor, Stuebs and Thomasson2015). Appropriate response has two components. First, it requires the corporation to address others in the right way, and the corporation could attempt to do this, or could be mandated to do so, by designing an appropriate corporate communications policy. Second, it requires the corporation to act in accordance with its second-personal deliberations, possibly by making amends and possibly by changing its future behavior. This requirement is partly a cultural one, and it is partly concerned with the formal procedures by which decisions are recognized and implemented in organizations; appropriate changes to the latter may be constrained by a legal system that gives complete primacy to shareholders.
The CSP conditions make it possible to assess whether a corporation is a moral agent and, if it is not, how much work is needed to achieve that status. The scale of the task facing a corporation that wishes to achieve moral agency could vary enormously; we have presented a very brief outline of the policies that could be adopted in pursuit of this goal.
4.3 Corporate Punishment and the Objective Attitude
The relationship between corporate moral agency and corporate punishment is a complex one. Several authors argue that corporations can reasonably be punished, for example, as a result of criminal proceedings, only if the corporation is a moral agent (Hasnas Reference Hasnas2012; Laufer Reference Laufer1996; Moore Reference Moore1999; Rönnegard Reference Rönnegard2013; Velasquez Reference Velasquez1983). We now examine corporate punishment from the second-personal perspective. We conclude that, because it is useful to punish corporations for wrongdoing, it can be understood as an expression of an objective, as opposed to a reactive, attitude. When punishment is levied for this reason, it is unrelated to questions of moral agency, so that corporations cannot avoid punishment by deliberately avoiding moral agency, nor can corporate officers evade responsibility by invoking corporate moral agency (Maitland Reference Maitland, Orts and Smith2017).
Our argument falls back on Strawson’s distinction between reactive and objective attitudes. To illustrate our point, consider the practice of training a dog. Dogs are not second-person competent and so, by our account, are not moral agents. Our attitudes toward them must therefore be objective, rather than reactive: when we train a dog, we do not expect it to understand or give reasons. We nevertheless rebuke dogs for inappropriate actions; we do so to manipulate their behavior because this enables our harmonious coexistence. It would be mistaken to claim that it would be unreasonable to train a dog in this way simply because it is not a moral agent.
We contend that, as a practical matter, corporate punishment is usually imposed from the objective stance as a useful way to train, treat, and manipulate corporate actors. The objective attitude can be adopted toward a moral agent like a person, or toward a nonmoral agent like a dog, and hence punishment can be justified without reference to corporate moral agency (for a related argument, see Hussain and Sandberg Reference Hussain, Sandberg, Orts and Smith2017). Corporate punishment succeeds because it influences the processes that govern corporate life and alters the incentives that corporate officers face. It forces those officers to recognize shared liability and fiduciary obligations that we impose because they are socially desirable (Sepinwall Reference Sepinwall2011, Reference Sepinwall2015). This recognition is a moral act and therefore relies on the moral agency of corporate officers. We cannot in this brief discussion address important questions of justice that are raised by this practice (Hasnas Reference Hasnas, Orts and Smith2017).
5. CONCLUSION
We have presented a second-personal account of corporate moral agency that views an agent’s ability to recognize others’ authority as central to its moral agency. Building on the work of Stephen Darwall (Reference Darwall2006), we identify three functional characteristics that are necessary and sufficient for a second-personal account of moral agency, and we use these attributes to analyze the case for corporate moral agency. Our approach is an alternative to the dominant “first-personal” approach of the existing literature, which is concerned with an agent’s capacity to reason for itself and to act on its reasoning. One can reason about organizations first personally without necessarily considering the way that the corporation is experienced as a social actor by outsiders and, in particular, without considering the conversations that the organization has with outsiders. In contrast, the second-personal account holds that corporate moral agency is inseparable from corporate communications: a corporate body that cannot receive, process, or respond to the appropriately expressed reactive attitudes of outsiders is not, by our account, a moral agent. Our emphasis on the ability to acknowledge others’ moral authority distinguishes the second-personal account of corporate moral agency from prior work that relies on reactive attitudes; it also yields a straightforward functional characterization of the features that render a corporate agent capable of being the target of reactive attitudes.
Our analysis in this article identifies and justifies necessary and sufficient conditions for a group to be a moral agent under the second-personal theory of corporate moral agency. We believe that our approach to moral agency may lead to a better understanding of the ways in which branding and reputation contribute to moral agency; that it could underpin a serious analysis of the whistleblower as second-personal addresser; that it could help us better to understand the design of formal and informal forums; and that it may help us to achieve a formal understanding of moral corporate communications functions. We leave these avenues of inquiry for future work.
Acknowledgments
We are especially grateful for the hard work of two anonymous referees and the editor. We also thank participants in brown bag seminars at Oxford and the members of Write Club for their comments and suggestions. Thanks in particular to Heli Helanummi-Cole, Rohini Jalan, Tom Lawrence, Grant Rozeboom, and Eva Schlindwein for carefully reading and commenting on earlier drafts.
Alan D. Morrison (alan.morrison@sbs.ox.ac.uk, corresponding author) is professor of law and finance at the Said Business School and a fellow of Merton College at the University of Oxford. Morrison teaches ethics and governance at Oxford. His research is concerned with the regulation of financial institutions, with corporate intentionality, and with moral learning.
Rita Mota is the Intesa Sanpaolo Research Fellow at the Centre for Corporate Reputation at the Said Business School and is a junior research fellow at Lady Margaret Hall at the University of Oxford. Mota teaches ethics at Oxford. She performs research in business and human rights, corporate intentionality, and moral learning.
William J. Wilhelm Jr. is the William G. Shenkir Eminent Scholar Professor at the McIntire School of Commerce at the University of Virginia. He is an investment banking specialist and teaches a wide range of related courses. His current work is concerned with professional standards in banking and with the effects of technological change on moral standards in investment banks.
