Balancing, Shielding, Filtering: Three Models of Role Morality

Abstract

How does a role—whether in business, law, government, or some other institution—change what is morally permissible or obligatory? Here I present three options and argue for the third. On the balancing model, a role simply gives its occupant additional normative reasons, to be weighed against all other normative reasons. On the shielding model, a role comes with its own moral code, blocking the force of all role-external reasons. On the filtering model, a role selectively filters its occupant’s reasons for action, creating obligations or permissions to act on a narrowed range of considerations. I argue that the filtering model offers a superior analysis of the ethics of roles, including the concepts of professional integrity and discretion. I focus on three difficult cases: a nuclear safety regulator, a criminal defense lawyer, and a corporate lobbyist. I conclude by discussing the implications of the filtering model for business ethics.

Business ethics, legal ethics, and the ethics of public service all engage with questions of role ethics. This branch of ethical theory asks how agents are morally required or permitted to act, given their role within a social institution: a for-profit corporation; a criminal trial; the state. Yet contributions to normative business ethics do not usually proceed from a clear position on the kind of difference that a role makes to an agent’s moral situation. Clarity on this matter is important because the manner and extent to which a role reshapes its occupant’s moral responsibilities also shapes the kinds of criticisms that are appropriate. If a role majorly transforms what is morally permissible or obligatory, it might be entirely misguided to criticize people in business roles—management, sales, marketing, finance—for failing to comply with our ordinary moral standards—fairness, compassion, honesty, charity. In contrast, if roles make a more marginal difference to an agent’s moral responsibilities, a much broader range of criticisms of people in business (and law, public service, etc.) shall be on the table: we should expect these role actors to do whatever a good person ought to do. ¹

In this article, I seek to advance our understanding of these matters by comparing three theories or models of the moral structure of roles. These models can be introduced as follows:

• • The balancing model: A role simply provides normative reasons for action that its occupants would not otherwise have, reasons that must be weighed against all the reasons that agents have as persons, independently of the role. Role actors are criticizable for failing to do what is most justified on balance, all things considered.

• • The shielding model: A role comes with its own moral code, and role actors must follow that code. Role actors are shielded from criticism in terms of the normative reasons or standards that apply to persons outside of the role.

• • The filtering model: Roles filter out some, but not all, of the normative reasons that apply to persons independently of their roles. Because role actors are required or permitted to act on a narrower range of reasons, a narrower range of moral criticisms are appropriately directed at their conduct.

The filtering model has been the least explored of these options, but I believe it is the most attractive. It best captures the moral experience of acting in a role, and it offers a more expressive vocabulary for evaluating hard cases in professional ethics.

I motivate the filtering model with an argument by cases. In the first section, I present three cases—two real, one satirical—that illustrate the moral quandaries surrounding the concept of role. These cases span the institutions of government regulation, criminal law, and corporate lobbying. In the second section, I present and compare the balancing, shielding, and filtering models of role morality, using the three cases to illustrate the advantages of the filtering model. I conclude by discussing the implications of the filtering model for business ethics.

THREE CASES

1. The Chalk River Nuclear Reactor

My first case concerns the responsibilities of a regulatory role. On November 18, 2007, the nuclear reactor at Chalk River, Ontario, was shut down for routine maintenance. At this time, inspectors from the Canadian Nuclear Safety Commission (CNSC), a federal regulator, became aware that the facility was not in compliance with national and international safety standards: its cooling pumps lacked earthquake-safe emergency generators. CNSC regulators decided to keep Chalk River closed until the necessary renovations were completed. However, the Chalk River reactor was no mere power generator: it was the National Research Universal (NRU) reactor, one of only a handful of facilities in the world to produce medical isotopes, used in tens of millions of diagnostic scans and cancer treatments every year. Closing the NRU, even for only a matter of months, would cause an abrupt shock to the global supply of medical isotopes. Hospitals would begin experiencing shortages, disrupting critical procedures. Within a few weeks of the reactor’s shutdown, the shortages were proving worse than anticipated, and heavy public scrutiny fell on the CNSC’s decision. ²

What followed was a showdown between Linda Keen, president of the CNSC, and leaders of the federal government, including Prime Minister Steven Harper and Minister of Natural Resources Gary Lunn. Treating the isotope supply crisis as the more pressing issue, Harper and Lunn tried to pressure Keen to approve the immediate reopening of the Chalk River reactor. Keen, however, stood firm in her position that the backup power generators be installed first. After this stalemate persisted for a few weeks, Harper and Lunn decided to go over Keen’s head, passing emergency legislation to overrule the CNSC’s decision and restart the reactor as soon as December 16, while the renovations remained ongoing. Happily, that work was completed without incident within a few months.

That might have been the end of the matter, but events escalated into a different kind of crisis when a letter from Lunn to Keen was leaked to the press. In the letter, Lunn told Keen that he doubted she possessed the “fundamental good judgment” necessary for her job and said he was thinking about recommending her removal. In response, Keen accused Lunn of “improper interference” with the CNSC and reasserted her authority and responsibility as an independent regulator (CTV News 2008). Soon after, Keen was abruptly removed from her post at the head of the CNSC. Defending her dismissal, Lunn said that Keen had displayed a “lack of leadership” and that she “did not execute the responsibilities of her role” (CBC News 2008).

Was Lunn right? In a surprising moment of political consensus, the emergency legislation to restart Chalk River passed unanimously in the House of Commons. The general will seemed clear: in this case, the risk associated with restarting the reactor before the backup generators were installed was outweighed by the radioisotope shortage resulting from its closure. The counterarguments were nothing to shrug at. Until the backup generators were installed, Keen estimated the risk of a major reactor failure at Chalk River as one in a thousand—a seemingly low risk, but a thousand times more risky than international standards tolerate (Standing Committee on Natural Resources 2008, 10). Still, let us assume that Parliament weighed the reasons correctly: it was better to reopen the reactor before the backup generators were installed, accepting the greater risk of a nuclear disaster in order to end the isotope supply crisis.

But from Keen’s perspective, the real question was not which issue was more pressing but who had the authority to make such a judgment. Not her, she said. As she interpreted her office as nuclear safety regulator, she could only approve the operation of a nuclear reactor once it satisfied a thorough safety assessment. She could not, ought not, recognize the isotope supply crisis as grounds for an exemption. Testifying to Parliament some months after her ousting, Keen stressed that “[t]he role of the CNSC is to ensure the safety of Canadians by regulating Canada’s nuclear facilities. That’s our job, that’s our responsibility, and that’s what the law prescribes.” Given the restricted mandate of this regulatory role, “the commission did not have the authority to take the issue of isotopes into consideration” (Standing Committee on Natural Resources 2008, 10). Keen made clear that she thought it entirely appropriate for Parliament to overrule the CNSC’s position on Chalk River; her objection was to the informal pressure she encountered from Lunn and Harper to reverse her regulatory ruling. As she put it:

Parliament is supreme, period. Neither I nor any Canadian would ever question the right of Parliament to act as it did. Parliament was faced with two competing interests: nuclear safety, on the one hand, and the need for medical isotopes, on the other—not an easy decision, and one appropriately made by the elected representatives of Canadians… (Standing Committee on Natural Resources 2008, 11).

Keen glossed the situation as one in which “[d]ifferent actors had different roles to play and different responsibilities to fill” (Standing Committee on Natural Resources 2008, 11).

Lunn’s and Keen’s differing interpretations of the role responsibilities of the president of the Canadian Nuclear Safety Commission reflect importantly different conceptions of role morality generally. Lunn’s view, interpreted charitably, was that the medical isotope crisis was of such demonstrative importance that it was irresponsible for Keen to not recognize an exemption to the ordinary safety protocols. ³ This position makes sense if we deny that a role supplies grounds for acting on anything less than an all-things-considered judgment, with the formal rules or mandate of the role serving as no more than general guidelines, to be set aside when they no longer prescribe the best course of action. Keen’s view, in contrast, makes sense if we think that the actors in a role must sometimes make judgments based on a narrower range of reasons, with the appropriate range dictated by the role’s mandate and rules. On Keen’s interpretation of her role responsibilities, she was legally and morally prohibited from allowing any considerations beyond nuclear safety to influence her official decision on Chalk River. Elected legislators, not CNSC regulators, were properly tasked with balancing all the factors in play.

I think that Keen was right. A system of arm’s-length regulation loses its point and value if those regulators are to ultimately base their decisions on an all-things-considered judgment rather than their official mandate. That is especially true when a regulator is tasked with safeguarding considerations that are liable to be overlooked or underweighted by other decision-makers, as seems true of nuclear safety. If lawmakers were for some reason unable to react quickly to the issue, we might judge the case differently, for the intended division of moral labour would then have been dysfunctional. But that was not the situation: the balancing of considerations was competently and appropriately performed by Canada’s elected representatives. Not every moral theory of roles is able to make sense of Keen’s judgment in this case, however. As we will see, a natural and pervasive way of thinking about the ethics of roles is committed to Lunn’s view—and worse for it, I shall argue.

2. The Buried Bodies Case

My second case is a staple of legal ethics education. In 1973, a group of friends were camping in upstate New York when they were attacked by Robert Garrow, who murdered one while the others escaped. After a twelve-day manhunt, Garrow was apprehended. For his criminal trial, Garrow insisted on being defended by a local attorney, Frank Armani, who had represented Garrow on some prior charges but had never handled a murder case. Armani was reluctant, but agreed, believing he had an obligation to provide the representation that Garrow requested. He enlisted the support of another attorney, Francis Belge.

In confidence with his counsel, Garrow revealed that he had kidnapped, raped, and murdered two other teenagers on separate occasions. Garrow even told his counsel where he had hid the bodies, and the lawyers—doubting Garrow’s soundness of mind—trekked into the woods to confirm his gruesome story. Armani and Belge did not share this knowledge with anyone, even as police continued to explore Garrow’s involvement in the disappearance of the other teenagers, and even as those victims’ parents, some of whom knew Armani personally, appealed to him directly for information. Instead, the lawyers tried to use the secret as leverage in plea bargaining, suggesting they were in a position to provide information concerning two unsolved murders in exchange for a more lenient sentence for their client. When the prosecution refused this deal, the lawyers advised Garrow to plead not guilty by reason of insanity, and to substantiate this plea Garrow admitted to all his murders under cross-examination. The ploy failed, and Garrow was found guilty of murder, at which point the attorneys publicly revealed that they had known about his other crimes for months. ⁴

The Buried Bodies Case (or the Lake Pleasant Bodies Case, as it is also known) is an especially vivid example of “ordinary morality” coming apart from “role morality.” Armani and Belge became the target of widespread outrage, ranging from disgusted editorials to hate mail and death threats. Yet the legal community was largely supportive, with many in the profession celebrating their commitment to attorney-client privilege and partisan advocacy under extraordinary circumstances (Hansen 2007; Lerman et al. 2007). Belge was charged with failing to report a dead body, and an ethics complaint was brought against him by the parents of a victim, but in neither case did the charges stick; instead, both the trial court judge that considered the criminal charge and the professional committee that heard the ethics complaint praised Belge for his commitment to his obligations as an attorney.

This case is in several ways more difficult than the first one. In the Chalk River crisis, lawmakers were fully apprised of the information they needed to form an all-things-considered judgment when Linda Keen insisted on her limited mandate. In the Buried Bodies Case, Armani and Belge were privy to a secret that, if shared, would give direction to open police investigations, provide closure to the victims’ families, and bring the justice system to bear on all of Garrow’s crimes. It was thus a live question whether the fundamental goods or purposes of the legal profession were actually being served by maintaining attorney-client privilege in this case. If the partiality and adversarialism expected of criminal defense attorneys is merely an imperfect means to some end—impartial justice, say—then that end might have been better advanced by betraying Garrow’s confidence. But if legal adversarialism is instead an internal component of an intrinsically valuable right to a partial advocate, then the lawyers were quite plausibly obligated to play their role in securing that right. Armani appears to have held the latter interpretation, arguing that “if [a client] can’t speak freely to his attorney, then you have no real justice system” (Kielty and Farrell 2016).

Then there is the larger, ethical question—the one that interests me here—of whether the rules of the role settle what Armani and Belge morally ought to have done. On one answer to this question, the lawyers’ professional obligations are the end of the matter: all other considerations are categorically out of place. A canon of the then-current Model Code of Professional Responsibility required a lawyer to “preserve the confidence and secrets of a client” (American Bar Association 1969, Canon 4). On the most straightforward interpretation of this canon, Armani and Belge had a role obligation to keep their client’s secret, revealing it only if this might serve his interests. Yet the public outrage against the lawyers’ actions indicates that many are unpersuaded by this way of thinking about role ethics. The rules of the role are one thing, but didn’t the lawyers have an overwhelming reason to break the rules in this case? Indeed, shouldn’t the rules have admitted an exception for the exceptional situation in which the lawyers found themselves? In an interview many years after the events, a parent of one of Garrow’s victims expressed dismay at learning that Armani and Belge’s actions are often treated as exemplars of professional integrity in legal ethics education:

I’m pretty horrified to think that this is what is considered to be correct. Because I don’t think it’s ethical at all, and to think it’s being taught as the right way to do things in an ethical class is totally incomprehensible to me … Maybe they ought to think not only about the criminal who they’re trying to defend, but what about the victims? And I think that that should at least be an equal thought in their mind, if not a greater consideration (Kielty and Farrell 2016).

Armani, for his part, does seem to have given thought to the victims and their parents. As he puts it in one interview, he was torn between “[t]he question of the Constitution, the question of even a bastard like [Garrow] having a proper defense” and “the fact that I have a dead girl here … the breaking hearts of her parents” (WETA-TV 1986). But Armani believed that the “proper, ethical, legal, moral … thing to do” was put his client’s interests first (Kielty and Farrell 2016).

Did he make the right choice? It depends, as with my first example, on how we think about the structure of role morality in general. It depends in particular on whether a role like defense attorney can require weighty considerations—“the breaking hearts of [the] parents”—to be set aside. I am inclined to say that the lawyers were obligated to act as they did, as disquieting as this may seem. Indeed, I think they deserve admiration for finding a way to disclose Garrow’s crimes without breaking their obligations to him as a client, by revealing the secret in a strategic manner that aimed at Garrow’s advantage. Yet I think it is a hard case, and the best analysis should draw out what makes it hard. In a dilemma such as this, even an admirable commitment to role is appropriately accompanied by a dreadful feeling of unease—unease that, by Armani’s own account, plagued him throughout the ordeal.

3. The Tobacco Lobbyist

My third case brings us to business ethics. In the film Thank You for Smoking (Reitman 2005), Nick Naylor is the chief lobbyist for Big Tobacco. ⁵ He describes his job as “spin control,” which in practice means using charm and creativity to keep people smoking cigarettes. Naylor jets across the United States, variously making media appearances as the face of the tobacco industry, hobnobbing with Hollywood producers to secure product placements in blockbusters, and dissuading a state congress from affixing poison labels to cigarette boxes. His value to the tobacco lobby is his charisma, quick wit, and a knack for turning the tables on a conversation, making those who seek to criticize or regulate his employers look like enemies of honest American business. In the opening scene, Naylor is on a TV talk show with a panel of anti-smoking advocates and a teenage lung cancer patient. He faces a hostile audience, yet manages to disarm that hostility with a clever misdirection: his employers have no interest in pushing products that cause cancer, Naylor argues, because that would only lead to the premature deaths of valued customers. His critics in the press call him a “blood sucker” and “pimp profiteer.” While Naylor seems unphased by these attacks, he is not in denial about the realities of his profession. He regularly dines with his counterparts in the alcohol and firearms industries—the “MOD squad,” they joke, or “merchants of death.”

Thank You for Smoking is a satire, but it parodies real-world occupations that blur the lines between advocacy and profiteering. The character of Nick Naylor also warrants our attention for his philosophy of roles, one that he imparts to his twelve-year-old son, Joey.

JOEY: You go to an office. Then you go on TV and talk about cigarettes. Then you fly out to LA to talk to some guy who works with movie stars. What is that?

NICK: It’s my job. I’m a lobbyist.

JOEY: I know, but did you study to do that?

NICK: No, I just kind of figured it out.

JOEY: Then, can’t anyone just do that?

NICK: No, I think it requires a moral flexibility that kind of goes beyond most people.

JOEY: Do I have flexible morals?

NICK: Well, let’s say you become a lawyer, right? And you were asked to defend a murderer. Worse than that, a child murderer. The law states that every person deserves a fair trial. Would you defend him?

JOEY: I don’t know. I guess every person deserves a fair defense.

NICK: Yeah, well … So do multi-national corporations.

Naylor’s analogy between corporate lobbying and criminal law is on point, given our own discussion of the Buried Bodies Case. That case saw us confront the idea that in representing a murderer—a child murderer, even—a lawyer is expected to respect the rules of their professional ethic even where this conflicts with weighty moral considerations. Naylor’s claim is that his role requires a similar kind of “moral flexibility,” offering a broader range of moral license to those who can bend to its demands.

But Naylor is not a lawyer, and he is not ensuring a person’s right to fair treatment in a criminal trial—the scenario in which the case for loyal, partisan representation is strongest. Instead, Naylor advocates for the business interests of multinationals: they employ him to improve their public image, to argue against expensive regulation, and thus to make them more money. While there is plausibly an intrinsic value in ensuring that even a guilty client has a competent advocate when facing the coercive and complex structure of the criminal law, there is far more dubious value in advocating for the financial interests of large corporations, particularly corporations that seek to downplay the health risks of their products.

I do not wish to deny, however, that the role of lobbyist exists within the legal framework of business and public policy. Naylor’s role is, I will say, just as legitimate as the role of criminal attorney, even if it is in other ways less good or less justified. Let me briefly elaborate on this notion of legitimacy, for I think it is of central importance. Institutions, along with their roles and rules, can be more or less justified along a range of normative dimensions: fairness, welfare-promotion, respect for autonomy, and so on. Legitimacy, in contrast, does not come in degrees: it is a bar that, once cleared, generates an important moral claim on the agents under the institution’s jurisdiction. In simple (but contestable) terms, we are under an obligation to obey the legitimate laws of the legitimate states that we reside in, and we are under an obligation to obey the legitimate rules of the legitimate roles that we occupy. ⁶ Those obligations might be defeasible; understanding the way in which they bind is part of my subject here. Importantly, role obligations can, like laws, be in many ways defective without compromising their claim to legitimacy. I here ask us to treat Naylor’s role as legitimate insofar as it piggybacks on the legitimacy of liberal market institutions: he promotes the interests of a legal enterprise via the rules and procedures of a democratic polity. ⁷ This assumption is consistent with the view that corporate lobbyists exploit unjust inequalities in the political process, and it is consistent with there being little value or worth to the specific interests that Naylor advances. The question is: where do these various assumptions leave Naylor? Is his lobbying, because legitimate, morally unimpeachable? Or is his “spin control,” because it is in the service of objectionable ends, morally indefensible? If the best answer is a complex in-between, we need a model of role morality capable of drawing the important distinctions.

In view of this challenge, Naylor’s flexibility metaphor might actually be a helpful way of thinking about the morality of roles. It suggests that the constraints of ordinary morality are permissibly stretched to allow for the maneuvers one must make in performing a role. This image bucks the idea that a role comes with a wholly detached moral code, with no reference to the ethical rules that bind role actors simply as persons. The image also implies that the flexibility of a role has a limit, meaning some conduct is still off the table. And indeed, Naylor refrains from bald-faced lies, opting instead for misdirections and equivocations; he seems to delight in finding ways to advance his employers’ interests without stretching the prerogatives of his role to their breaking point.

Yet the flexibility image is rather too convenient for Naylor’s purposes. Gymnastic ability might be a flattering simile for the craft of argument, but it is a troubling way of thinking about exemptions to moral constraints, as if the nimble should be able to get away with more than others. That may be the mindset with which Naylor approaches his work, but it is not a suitable basis for a moral theory aimed also at describing the limits of the latitude afforded to a role. We need a different metaphor.

THREE MODELS

1. The Balancing Model

The simplest way for a role to change an agent’s moral situation is for it to supply reasons for action that they would not otherwise have, reasons to be weighed against competing reasons. The balancing model of role morality claims that this is all roles do: they are one source of normative reasons among many, and they supply no grounds for ignoring reasons that bear on role actors simply as persons.

The language of reasons has, in this context, the benefit of neutrality between competing ethical systems. ⁸ Take whatever you think is at the root of morality, whether that is the value of certain ends, or the imperative to follow rational rules, or the self-perfection in cultivating virtue, or something else besides or in addition: these are sources of “role-external” reasons for action. They “count in favor” of how one ought to act—count so in virtue of one’s status as a moral agent or person. Then take whatever you think gives role actors reason to follow the rules or expectations of their office, be that the value of the ends they promote, or the importance of a rule-governed procedure, or the virtues inherent in the professional practice, or something else: these are sources of “role-internal” reasons for action, or reasons for acting as the role prescribes. The central idea of the balancing model is that these role-internal and role-external reasons must be compared—weighed, as if on a fulcrum balance. When there is a conflict between these sets of reasons, the role actor should be guided by whichever set has, in the circumstances, the greater “weight” (or “strength,” “force”). Role actors must therefore be prepared to break with the rules of the role when the balance of reasons demands this. ⁹

Figure 1 represents the balancing model diagrammatically. On this model, the reasons of the role are themselves justified by role-external reasons. This is to capture what David Luban—who, in the context of legal ethics, has presented the balancing model most explicitly—calls “the rather obvious idea that an act absolutely essential for an institution of merely marginal importance itself has only marginal importance” (Luban 1988, 134). The thought is that a role and its rules are not free-floating: they must have some external justification in more fundamental reasons (or values, principles, etc.). The weight of a role’s internal reasons is a function of the overall strength of the justification for the role and the extent to which that justification is furthered in a given scenario. As Luban puts it: “we determine whether a role obligation should be overridden in a particular case by asking whether the damage done to the institution’s mission by defaulting from the role obligation in that case outweighs the good accomplished by breaking the role” (1990, 430). Role obligations, on this view, are merely rules of thumb: they represent how a role actor is generally expected to act, but they can always be overridden by strong role-external reasons, particularly when the role-internal reasons are weak.

Figure 1: The Balancing Model

Such a view has appealed to moral philosophers who seek to deflate or debunk appeals to roles as defenses against charges of immorality. This project has been carried out most thoroughly by Arthur Applbaum, who concludes:

institutions and the roles they create ordinarily cannot mint moral permissions to do what otherwise would be prohibited. They cannot because the filtering of reasons and the redescription of actions and actors that roles demand do not properly have a grip on the judgments of role players who are also persons simply, and so who face all the reasons for action that apply to persons simply (Applbaum 1999, 257–58). ¹⁰

On this view, a role can give its occupants reasons for action they would not otherwise have, but it cannot silence the force of the reasons that apply to agents as persons, independently of their roles. If a role changes what agents are morally permitted to do, it does so only by providing role-internal reasons for action that are strong enough to outweigh all competing, role-external reasons—and Applbaum doubts that the reasons of a role will ordinarily be so strong. ¹¹

The deception, manipulation, and ruthlessness so common in public and professional life certainly call for an ethical reckoning (cf. Nagel 1979). But the balancing model achieves these debunking aims at the cost of undermining the moral significance of all roles, including the more reputable ones. This becomes clear in returning to Linda Keen’s decision in the Chalk River crisis. As president of the Canadian Nuclear Safety Commission, Keen insisted that it was not for her to balance the conflicting considerations in the case: her role was to enforce nuclear safety standards, and she could not exempt Chalk River from those standards on the grounds of the isotope supply crisis without betraying the obligations of her role. I think that Keen acted admirably: she held firm to the mandate of her office, resisting improper pressure from elected officials to disregard that mandate. But the balancing model seems committed to the conclusion that Keen acted wrongly. Role-internal reasons must always be balanced against role-external reasons, and in this case the reactor’s safety deficiencies were less pressing than the shortage of medical isotopes.

Perhaps there is a way to finesse the example so that the balance works out differently for Keen—to show that for her, though not for others, the nuclear safety issue had the greater weight. I find it hard to see how such a story can be told without giving up on the central idea of the balancing model, which is that the rules of a role should not be followed simply as such, but only insofar as the reasons for following them are strongest on balance, taking all factors into consideration. In any case, the deeper problem is that the model is committed to saying that Keen thought about her decision incorrectly: she was overly officious, wrongly insisting on her official mandate when she should have weighed that mandate against the isotope supply issue. This is one way of taking Gary Lunn’s censure of Keen for displaying a “lack of leadership” in the crisis (CBC News 2008). I do not think that is the right appraisal of Keen, and I hope, in presenting my preferred model of role morality, to show that there was integrity in her commitment to her limited mandate.

The problem for the balancing model carries through to my second example, the Buried Bodies Case. This case is especially pertinent because Luban himself hoped to use the balancing model to explain why Armani and Belge’s actions were morally justified (1988, 186–89; 1990). He believes there is good justification for the adversary legal system in criminal cases, and he argues that this institutional justification gave Armani and Belge strong role-internal reasons for loyalty to Garrow, their client. While the lawyers did have many role-external reasons to abandon that loyalty, in this case their role within the criminal justice system supplied the reasons with the greater weight. But as Luban’s critics have pointed out (Wasserman 1990; Wueste 1991; Dare 2009), it is far from clear that the balance of reasons worked out that way. On Luban’s view, Armani and Belge would have sufficiently strong reasons to keep their client’s secret in confidence if, and only if, betraying his confidence would do enough damage to the larger “mission” of adversary legal advocacy to outweigh their role-external reasons (Luban 1990, 430). But couldn’t Armani and Belge have alerted the police discreetly, with an anonymous tip? Even if Garrow suspected that his confidence had been betrayed, would that really have undermined the general faith in the institution of attorney-client privilege? And even if faith in that institution were undermined, would the damage have been enough to outweigh the costs of the ongoing police investigations and the lack of closure for the victims’ families? The balancing model seems to require the lawyers to follow this line of reasoning and thus to realize that their role-based reasons were quite weak indeed, for “the damage done to the institution’s mission by defaulting from the role obligation” would likely be negligible.

Again, perhaps there is a way to finesse the analysis of the case so that the balance of reasons works out differently. ¹² And of course, one might think that the balancing model is right to condemn the lawyers’ conduct—recall the parent who said that defense attorneys “ought to think not only about the criminal who they’re trying to defend, but [also] about the victims” (Kielty and Farrell 2016). But the more fundamental difficulty for the balancing model is once again the way in which it examines the problem. The model directs us to see a conflict between two equally applicable sets of reasons. It neglects the possibility that Armani and Belge’s situation was like Linda Keen’s in that their role required them to set aside some considerations that would otherwise be decisive. Indeed, as Luban’s critics have pointed out, the balancing model seems to render roles otiose: in the end, the lawyers were not to think as lawyers but as persons, balancing all reasons that apply to them as persons (Dare 2009, 38; Wueste 1991). Perhaps this is ultimately the right conclusion to draw, but we should recognize it as a revisionist conclusion, undercutting the moral significance that many professionals and public servants take their roles to have.

The balancing model’s appraisal of Nick Naylor, the lobbyist of Thank You for Smoking, might appear to be its saving grace. Here, the debunking impulse behind the model seems on point: the ends that Naylor serves lack value and so his reasons for advancing those ends are weak and easily outweighed. Indeed, proponents of the balancing model seem bound to conclude that Naylor ought to be routinely disobeying the obligations of his role—perhaps by resigning, or perhaps by internal sabotage, making his advocacy lackluster and incompetent rather than enthusiastic and skillful. I think it is right to challenge the “moral flexibility” that Naylor claims for his role. But even here, the balancing model’s analysis seems too shallow. Corporate lobbyists like Naylor maneuver the levers and pulleys of a liberal market democracy. If these institutions, and Naylor’s role within them, are at least legitimate, then I think this legitimacy should count for something, should change how we evaluate his “spin control.” To settle for an easy condemnation of Naylor, based simply on the balance of reasons he faces, is to be rather too unwilling to explore the moral implications of the institutions we have built.

To be sure, the simplicity of the balancing model remains seductive. It ought to be our default: any departure from its terms must be motivated, and some roles may not satisfy the proposed conditions for this departure. Still, I hope to have offered enough reservations about its picture of role morality to motivate the search for an alternative.

2. The Shielding Model

If we wish to take roles more seriously, how might we think about their obligations? One option is to say that a role comes with its own moral code, distinct from the code of ordinary life. I shall call this the shielding model, for the main idea is that role occupants are shielded from the normative force of reasons that are external to the role.

Figure 2 depicts the structure of the shielding model. The central idea is that the force of role-external reasons is entirely blocked while acting in a role. The morality of roles is solely a matter of following their internal rules. With the balancing model, the shielding model acknowledges that a role must be ultimately grounded in role-external considerations; it cannot be free standing or self-justifying. But the type of grounding involved is different. On the balancing model, what matters is the justification for the role: the degree to which it promotes some valuable end or otherwise has merit. On the shielding model, what matters is the legitimacy of the role: an all-or-nothing condition that, once satisfied, turns the internal norms or rules of the role into their own moral code, uncontested by other, role-external considerations. Importantly, the claim is not that the role-external reasons are outweighed, but that role-external reasons are no longer applicable—no longer bear on how the role occupant ought to act. In contrast, illegitimate roles do not shield their occupants against external criticism. When a role is illegitimate, it offers (at most) some additional reasons for action, to be balanced against all role-external reasons.

Figure 2: The Shielding Model

Importantly, a role and its rules can be unwise, unfair, or otherwise defective without losing the status of legitimacy. On the shielding model, the authoritative rules are those that characterize the role as it is, rather than the rules as we want them to be. Loyalty is owed to actual institutions and their divisions of role responsibilities, not to our ideal imaginings of those institutions. Determining which rules are legitimate is part of what is involved in interpreting the role, and the best interpretation of a role could potentially reveal that its requirements are widely misunderstood. ¹³ Still, role actors are not licensed, as they were on the balancing model, to decline to follow a rule of their role on the grounds that the point of the rule would be better served by breaking it. (I indicate this, in Figure 2, with a solid circle around the internal reasons of the role, as opposed to the porous circle in Figure 1’s representation of the balancing model.) While a role actor may add their voice to calls for reform, when acting in role they are bound by its current rules.

Tim Dare (2009; 2016), who develops an account of role morality along these lines, calls it “the indirect approach.” The idea is that ordinary or role-external reasons have an indirect influence over the morality of conduct in a role: they legitimate its rules, but the rules dictate how role occupants ought to act. ¹⁴ For example, role-external reasons supply the normative grounds for the system of adversary law, but it is the rules of that system, not the reasons behind it, that dictate how a lawyer ought to act. This is the sort of view I intend to capture here, but I prefer the shielding metaphor. The balancing image has a hold on much of our ethical thinking, so we need an equally compelling image with which to contest it. The notion of a shield fits into the physical imagery already introduced by the language of reasons, inviting us to imagine the “strength” or “weight” or “force” of a reason being blocked while acting in a role.

Whereas the balancing model appeals to theorists who seek to undercut the moral significance of roles, the shielding model finds its strongest defenders in practitioners who defend the dignity of loyal adherence to a station and its duties. ¹⁵ Consider the following example from Michael Quinlan, who led a distinguished career in the British civil service. When a cabinet minister, an opposition front-bencher, a civil servant, and a media commentator are each asked whether the prime minister is doing a good job, their different roles dictate different responses: the cabinet minister must say “yes,” the opposition front-bencher must say “no,” and the civil servant must refuse to answer.

Only the media commentator is free (if his editor lets him) to express what he really thinks, but it would be superficial to the point of silliness to suppose that this latitude somehow makes him or his profession more moral than the others. Anyone who seeks to sit in judgment upon the behavior of people in working roles is simply being foolish if he attempts to opine without reference to the necessities—the proper necessities—of their roles and responsibilities (Quinlan 1993a, 539).

On this picture, the morality of conduct is relative to the role, and it is a mistake to judge role occupants by role-external standards. A role has its own code, and that code—not the more general ethical standards of ordinary life or philosophical theory—is the appropriate standard for evaluating the role actor’s conduct. Contra Applbaum (1999, 257–58), a role actor is not a person simply; he is not accountable to all of the reasons that apply to persons simply. ¹⁶

If the shielding model is our guide to role morality, then the protagonists of our three cases—Keen, Armani and Belge, and Naylor—all act in morally defensible or even obligatory ways, given their respective roles. Keen correctly identified her limited mandate, and thus correctly resisted improper pressure to act for considerations external to that mandate. Armani and Belge displayed moral fortitude in continuing to respect the rules of attorney-client privilege in extremely trying circumstances. Naylor simply does his job and does it well: although the “flexibility” afforded to his role “goes beyond most people,” this does not make him or his profession any less moral than the others. As Quinlan (1993a, 540) notes, some roles are not for everyone: “Some temperaments, some mixes of gifts, will find themselves ill-suited by the constrictions of this or that role. Where that mismatch arises, it is not necessarily a discredit to either the individual or the role; but they had better not combine.” Not everyone can stomach what Naylor does, but that does not, on the shielding model, mean his actions are impermissible: we must judge him in light of his role and its responsibilities.

The chief shortcoming of the shielding model is the uniformity with which it evaluates the three cases: it is insufficiently nuanced. Perhaps the bare permissibility of all the protagonists’ actions is secure, given the legitimacy of their roles, but there is more to say about the cases. My own judgment is that their actions become increasingly questionable as we move through the sequence: I find Keen plainly admirable, though I think legislators were right to overrule her; I have conflicted feelings about Armani and Belge, finding their commitment to their client hard to stomach even as I accept the constraints of their professional obligations; and Naylor’s conduct wholly disagrees with me, yet I am not convinced it is wholly wrongful, given that our system of institutions allows for corporate lobbying as a form of free speech and enterprise. If all role-external reasons are categorically out of place when evaluating the conduct of people in working roles, as they are on the shielding model, we are left with no basis on which to draw these more subtle distinctions between the cases. The most we seem able to do is think about what would be acceptable to us, given our personal values. But roles are not immoral just for being ill-suited to some temperaments.

A shielding theorist might try to draw a sharp line somewhere in the sequence, perhaps insisting that Keen and Armani act with legitimacy while Naylor does not. ¹⁷ But, to return to a point stressed above, this does not seem to me to be the right way to think about the tobacco lobbyist. If we have for-profit businesses in a liberal polity, then it is legitimate for these businesses to hire advocates for their interests. Naylor’s role is morally ambivalent: he serves ethically questionable ends via ethically questionable means, yet he operates in the legitimate terrain of liberal market institutions. At least, this is the characterization under which his case raises the most difficult problems. We need a moral framework capable of discussing such cases with nuance, and the shielding model is not that framework.

A different objection to the shielding model is that it is wrong to treat a role as silencing the force of all role-external reasons. The rules of role should be taken seriously, but they surely do not block every other consideration. And indeed, theorists who initially appear to be presenting a view with the shape of the shielding model are often quick to qualify it to allow some external considerations to carry weight—perhaps decisive weight—over how a role occupant ought to act. Quinlan, just before his insistence on the “proper necessities” of working roles, claims the rules of a role cannot override the “basic moral requirements of any human behavior,” including “truth-telling and honesty, and avoiding gratuitous hurt to other people” (1993a, 538; cf. Bowie 1982 for similar claims about the physician’s role). On the shielding model, it is hard to understand why these role-external reasons, but not others, should be taken into account. It cannot be that these concerns are somehow “built in” to the role’s internal code. Why expect that to be universally true, if roles are social constructs, and when what matters is the rules as they are, not as we want them to be? In any case, Quinlan does not treat these “basic” moral requirements as part of the role but rather as exceptions to its rules, stemming from an external source.

The balancing theorist has a ready explanation for why role occupants must continue to respect some moral requirements but not others: some are too forceful to be outweighed by the reasons of the role. But the balancing model has different limitations, discussed already. I next present a third alternative, one that sticks with the shielding model’s idea that some considerations are categorically excluded while acting in a role, but treats those exclusions as more selective—a filtering of reasons rather than a blanket shielding.

3. The Filtering Model

On the filtering model, a role actor is to consider their reasons for action as if through a filter, excluding or discounting some classes of reasons while taking others into account. To act with integrity in a role is to refuse to act on reasons that are properly excluded by the role’s rules and requirements, while continuing to give weight to reasons that survive these filters. Consider an analogy to biological structure: a selectively permeable membrane allows some types of molecules to pass through while blocking the passage of other molecules. Normative filters can be thought of as a similar kind of structure, one imposed by the institution on the role occupant.

I will unpack the filtering metaphor by expanding on the language of normative reasons. We have so far considered first-order reasons: reasons that count directly in favor of actions, with a weight that can be compared against the weight of other first-order reasons. I now introduce second-order reasons: reasons to act, or refrain from acting, for other reasons. The simplest and most powerful kind of second-order reason is an exclusionary reason, or a reason to entirely exclude a range of first-order reasons in determining what to do. ¹⁸ Whereas first-order reasons have weight or force, second-order reasons function as filters, narrowing the range of reasons that the agent is allowed to give weight to.

Joseph Raz (1977; 1979; 1990) developed the concept of second-order reasons to explain the normative character of law, authority, promises, and other rule-governed practices. Although Raz does not discuss role ethics in detail, I think it is no coincidence that he motivates the idea of a second-order reason with an example that features a role obligation.

While serving in the army Jeremy is ordered by his commanding officer to appropriate and use a van belonging to a certain tradesman … His friend urges him to disobey the order pointing to weighty reasons for doing so. Jeremy does not deny that his friend might have a case. But, he claims, it does not matter whether he is right or not. Orders are orders and should be obeyed even if wrong, even if no harm will come from disobeying them. That is what it means to be a subordinate. It means it is not for you to decide what is best (Raz 1990, 38).

As Raz presents Jeremy’s situation, his commanding officer’s order is not merely a first-order reason, to be followed only if it outweighs his other first-order reasons. It is also a second-order reason—an exclusionary reason—to obey the order regardless of how the balance of reasons falls. There are limits to this deference, however. If Jeremy’s commanding officer ordered him to torture the van owner, he should refuse (Raz 1990, 38). On the Razian analysis, this is explained by the scope of exclusionary reasons: they exclude a range of first-order reasons, but not reasons outside of that range. Jeremy’s reason to follow orders does not exclude his reason not to torture; faced with such an order, Jeremy should indeed weigh his role-internal and his (non-excluded) role-external reasons and find the latter to be decisive. The filtering model takes this analysis and expands it to roles with obligations more complicated than “obey orders.” ¹⁹

Figure 3 illustrates how the filtering model departs from the balancing and shielding alternatives. The internal reasons of a role consist, on this view, of both first- and second-order reasons, which together shape the obligations of the role. First-order reasons must be balanced, as on the balancing model, while second-order reasons limit the range of first-order reasons that are added to the balance, partially shielding the role occupant from external criticism. On this view, keeping track of the scope of second-order reasons is essential to acting ethically in a role, for the scope determines the range of role-external reasons that must continue to be given weight in deciding what to do. (One extension to the model, involving second-order permissions, is considered later.)

Figure 3: The Filtering Model

The balancing model focused on the justification for the rules of a role, whereas the shielding model focused on their legitimacy. The filtering model sees both as important. The justification for a role obligation tracks its first-order weight. This reflects the idea that the weight of first-order reasons comes in degrees, like the justification for a role and its rules. The legitimacy of a role obligation bears on whether the role can also create second-order reasons to exclude some first-order reasons. This reflects the idea that legitimacy is an all-or-nothing matter. If the hierarchical structure of the military is legitimate, then Jeremy’s commanding officer has the power to issue orders that count, for Jeremy, as second-order reasons, even if some of those orders lack a good justification. Interpreting the legitimate rules of a role involves interpreting the second-order scope of its obligations: the reasons that are excluded versus the reasons that remain on the normative scales. In the absence of second-order reasons, the filtering model collapses into the balancing model. It is second-order reasons that give shape to the obligations of a role, such that entering a role is a different, more structural moral change than simply entering a different circumstance with different first-order reasons in play.

Some ethicists, particularly those who are drawn to the balancing model of role morality, worry that treating role obligations as exclusionary reasons amounts to an unacceptable narrowing of moral judgment. Recall Applbaum’s insistence that “the filtering of reasons … does not properly have a grip on the judgments of role players who are also persons simply, and so who face all the reasons for action that apply to persons simply” (1999, 257–58; cf. 198–99). In a similar spirit, Luban rejects a Razian analysis of role obligations on the grounds that it implies that role actors “need not think about moral issues at all—for moral questions about acts are eliminated by the exclusionary reasons contained in the role” (1988, 120). I think these authors are wrong to associate the filtering of reasons with the silencing or abdication of moral judgment. Keeping track of the scope of second-order reasons is an exercise of moral judgment, particularly when that scope is unclear or contested. Moreover, it is sometimes right to stand on one’s second-order reasons in the face of strong but legitimately excluded first-order reasons. This is right not just because one is acting in a role, but because in entering a role one takes on a commitment, as a person, to act as the role requires.

I can here say little about what it takes to defend or vindicate the second-order reasons or filters of a role. Defenses will typically appeal to the institutional structures that invoke such filters, including hierarchies, competitions, and agent-principal relationships. When these institutions meet the conditions for legitimacy—however high or low that bar is set—their roles are bound by second-order constraints. This can be true, as I emphasized above, even if the same institutions are unjust or otherwise flawed: legitimacy does not require perfection. With that said, we should remember that the filters of the filtering model are, like the shields of the shielding model, subject to interpretation, and popular interpretations might prove, on critical inspection, to be misguided or exaggerated—implausibly broad in scope. The details depend on the specific role; I do not think that we can get far in attempting general pronouncements. We might try to say that the “basic moral requirements” referenced by Quinlan (1993a, 538)—his examples are honesty and non-maleficence—escape the filters of any legitimate role, but even there I am wary of overgeneralization. So let us return to the specifics: how might we reassess our three cases in the filtering model’s terms?

Linda Keen’s stand on her official mandate exemplifies how moral judgment in a role sometimes demands the filtering of reasons. Keen was attentive to the distinction between the reasons she was obligated to act on (those pertaining to the safety of the nuclear reactor) and those she was obligated to give no weight to in her role as regulator (those pertaining to the supply of medical isotopes). Had she buckled to pressure to green light the immediate reopening of the reactor, she would have been acting with good justification, in the sense that she would have been acting in line with the balance of first-order reasons. But she would have been acting without legitimacy, for she would have been giving weight to reasons she was role-bound to exclude. The reasoning Keen employed in the Chalk River crisis is precisely the sort of reasoning that is called for if role obligations are second-order reasons. It was, as she put it, a situation in which “[d]ifferent actors had different roles to play and different responsibilities to fill,” and where the all-things-considered balancing of reasons was “appropriately made by the elected representatives of Canadians” (Standing Committee on Natural Resources 2008, 11).

Keen’s commitment to her role obligations exhibits, I think, a distinctive form of integrity, one that the filtering model captures especially well. In an insightful discussion, Cheshire Calhoun describes integrity as “standing for something”: not trading one’s own views “too cheaply for gain, status, reward, approval, or for escape from penalties, loss of status, disapproval” (1995, 250). Calhoun focuses on the personal integrity of standing firmly on one’s values in the face of temptation or pressure to betray those values. Keen exemplifies the neglected virtue of professional integrity: standing firm on one’s role obligations in the face of temptation or pressure to do something that the role prohibits—something that might even be the right thing to do, were it not for the constraints of the role. Those with professional integrity recognize a requirement to act on a limited range of reasons, and they stand by that requirement even in the face of “penalties, loss of status, disapproval.” Such integrity is, I think, at the heart of role morality, understood as a domain in which practical judgment is often constrained by second-order reasons. (Keen may also have acted with personal integrity in risking her job for her sense of her role responsibilities, but that is because professional integrity ranked high among her personal values. ²⁰)

I think we can also credit Armani and Belge with professional integrity: they too stand firm in their sense of their role obligations, despite significant pressure and strong first-order reasons to break those obligations. But the case is more difficult. Unlike Keen, whose role obligations were comparatively clear, Armani and Belge faced a live question about the scope of their obligations to their client. They had unearthed the dead bodies of teenagers whom their client confessed to assaulting and murdering: were they really required to hold that secret in confidence? It was not implausible to think that this exceptional situation qualified as a legitimate exception to the general duty of client confidentiality, in the same way that hard legal cases sometimes reveal limits to law that were not anticipated by its crafters (Dworkin 1978). It did not help that the then-current Model Code of Professional Responsibility was a confusing document; interpreting its requirements was, in the words of its reformers, “a three-dimensional chess game that lawyers played at their own peril” (Center for Professional Responsibility 2006, xiv). The current Model Rules of Professional Conduct add clarity: a client’s secrets may be shared for only a limited range of reasons, none of which applied in Armani and Belge’s situation. ²¹ But the lawyers did not have these updated rules, so they were in the unenviable position of having to decide whether their situation was exceptional—whether it was one of the limited set of cases in which they could betray their client’s confidence for reasons not excluded by their role.

Given that it was such a hard case, it is unsurprising that the lawyers experienced such conflicted feelings. As Armani put it, he was torn between his client’s right to “a proper defense” and “the breaking hearts of [the] parents” who still held out hope for recovering their missing children. Luban, with his balancing model, attempted to understand Armani’s predicament as a conflict between two sets of first-order reasons. On my filtering model, the conflict is instead between the all-things-considered balance of first-order reasons and the filtered balance of first-order reasons, once the second-order exclusions of the criminal lawyer’s role are applied. The excluded reasons remain strong reasons: they are reasons that would, were the scope of the filter slightly more narrow, tip the balance decisively in favor of betraying the client’s confidence. A special kind of moral fortitude is therefore needed to resist their genuine pull. As Raz observes, such a choice will often be accompanied by “a peculiar feeling of unease”—the typical phenomenology, he suggests, of a conflict between a strong first-order reason and an exclusionary reason (1990, 41). Indeed, uneasy feelings will likely plague the morally conscious agent no matter what they choose to do in such a situation. Armani may have displayed professional integrity in standing by the obligations of his role, but had he instead abandoned those obligations, we might have credited him with the integrity of conscientious disobedience. In an intractably hard case, a role actor might act in one way wrongfully, and in another way admirably, no matter what they choose to do. ²²

We now return to the most troubling case: Nick Naylor, the tobacco lobbyist. I do not wish to credit Naylor with any kind of integrity. Yet, as stressed above, I also want to treat his role as legitimate and to treat that legitimacy as counting for something. How, then, might we evaluate the case with more nuance? One option is to insist on limited scope of the second-order reasons that his role affords, particularly when it comes to his role-external reasons of honesty. But Naylor is already, in his crafty way, attentive to such scope restrictions. In particular, he doesn’t lie about the harmful effects of his clients’ products; he equivocates, tells half-truths, or redirects the conversation with his ample charm. While he toys with the line set by the scope of his filters, he does not clearly cross it. We might instead argue that the ends Naylor serves are considerably less worthy than those served by Keen or Armani, so there is less dignity in his pursuit of those interests in the face of strong competing reasons—even if those reasons are legitimately excluded. But it is not clear how Naylor can be fairly criticized in terms of reasons that he must, on pain of role immorality, set aside.

There is, however, a more structural distinction we might draw between the tobacco lobbyist and the other two cases. I have so far presented the filters of a role as second-order reasons. These reasons ground role obligations, or requirements to set aside first-order reasons. But the filters of a role might also be discretionary, granting role actors the right to decide, from context to context, whether or not to screen out the first-order reasons in their scope. In Raz’s terminology, these discretionary filters are second-order (or exclusionary) permissions—prerogatives to exclude reasons, rather than requirements to do so. Raz (1990) uses second-order permissions to explain legal rights, such as the right to free expression, even to express foolish or obnoxious things. Second-order permissions helpfully characterize the moral discretion that is often a feature of role ethics: a role actor might have the option, rather than the strict obligation, to set aside a class of first-order reasons when those reasons conflict with how they wish to perform the role.

I think that Naylor’s role is characterized, at least in part, by second-order permissions. He has a basic mandate to lobby for the tobacco companies he represents, but he has the professional discretion to decide how he pursues that mandate, and one aspect of that discretion is the permission to set aside a range of role-external reasons—reasons of truth-telling or harm reduction, for example. When he appears on a talk show and distracts the audience with wit and sophistry, he exercises discretion; when he successfully persuades a state congress to keep poison labels off cigarette boxes, he exercises discretion once again. His role gives him, in his words, the “moral flexibility” to set a broad spectrum of ethical qualms aside, as he judges convenient from case to case. Like second-order reasons, second-order permissions have restricted scopes: perhaps Naylor is not permitted to outright lie, and indeed he refrains from doing so. Still, Naylor seeks to stretch his permissions as far as he can, delighting in the challenge this presents.

I am conceding that Naylor has the prerogative to act as he does, in virtue of his legitimate role. He is within his moral rights. But rights are only partial shields against moral criticism. If someone uses their right to free expression to say obnoxious things, they are still being obnoxious; to respond to this criticism by reasserting one’s rights is to misunderstand the critique. ²³ So too with Naylor and the rights of his role. When Naylor acts discretionarily, rather than with strict obligation, he is exposed to criticism on the grounds that he could perform the role differently. This is why Naylor cannot be credited with the same integrity as Armani or Keen, even though each can appeal to their role as grounds for not acting on strong first-order reasons. When it comes to second-order permissions, the path of integrity is to refrain from milking those permissions for all they are worth. While Naylor’s conduct is permissible, in the thin sense of being within his role rights, he can still be justly criticized in terms of all the first-order reasons that he could be giving weight to. He may still be, in the colorful language of the film, a “blood sucker” or a “pimp profiteer.”

But we know that Naylor is unphased by these accusations. He said that his moral flexibility goes beyond most people.

IMPLICATIONS FOR BUSINESS ETHICS

These matters have not received much attention in the business ethics literature. ²⁴ This is unfortunate, for business ethics directly engages with questions of role ethics, interrogating how specific roles in specific institutions—competitive markets and for-profit firms—change the moral obligations and permissions of their occupants. Depending on the moral theory of roles we adopt, it might be entirely misguided to criticize business actors for failing to conform with our ordinary moral standards: to critique managers for being unfair, or insurance underwriters for lacking compassion, or marketers for engaging in forms of deception. By way of conclusion, I would like to begin—I can only begin—to consider how business ethics might be approached via each of the models discussed here.

The balancing model exposes business actors to the widest range of criticism and condemnation. On this view, managers, marketers, and so on are to be evaluated essentially as persons, not role actors: although their roles give them reasons for action they would not otherwise have, they still “face all the reasons for action that apply to persons simply” (Applbaum 1999, 257–58). Many popular approaches to business ethics seem to presuppose, or at least fit most naturally with, this model of role ethics. Take multi-fiduciary stakeholder theory (e.g. Freeman 1998): the view that managers must balance the interests of all the constituencies with a “stake” in the corporation is very close to the view that managers must balance all the reasons that bear on their decisions. ²⁵ Or take the evaluation of business roles by the standards of a general ethical theory, be it Kantian, Aristotelian, Buddhist, existentialist, or what have you: such an approach is likely to be implicitly assuming the balancing model, for on other models the appeal to role-external considerations is immediately suspect. ²⁶ It is not surprising for the balancing model to have such broad (if tacit and unexamined) appeal, for it comes with the lowest barriers to moral critique: any normative consideration we can identify is a consideration we can criticize role actors for neglecting. My central misgiving with the model is that it does not take roles sufficiently seriously. As I tried to illustrate with my three cases, a legitimate role calls for forms of moral reasoning not captured by the simple image of weights on a balance. I therefore think that a business ethics that assumes the balancing model, at least without further argument, is not really engaging with the questions that make the subject interesting and difficult: questions about how the roles of economic life bind the persons who occupy them.

At the other extreme is the shielding model, which contends that all role-external reasons—and thus all critiques that draw on such reasons—are blocked while in a legitimate role. This model will most likely appeal to theorists who seek to defend the status quo in business, rebutting the critics who “sit in judgment” without attending to the “proper necessities” of business roles and responsibilities (Quinlan 1993a, 539). Milton Friedman’s (1970) famous insistence that a corporate manager’s responsibility is to increase the company’s profits is very much in the spirit of the shielding model: the heart of his claim is that the ethics of management must not be conflated with, or made subservient to, the ethics of ordinary life. To be sure, the economic institutions that define the managerial role must be defended, and that defense must counter a broader array of moral objections; but if these institutions are shown to be at least legitimate, then their internal rules and requirements constitute a self-contained moral code. The chief difficulty with the shielding model is that it is hard to accept that all role-external reasons are blocked from consideration, with no exceptions. Quinlan, who seems at first to defend a shielding theory of public service ethics, qualifies his view by saying that public servants must still respect “basic moral requirements”; Friedman himself admits that managers can only pursue profits within the constraints of “ethical custom.” The comment is unelaborated, but it is the first step to conceding that roles are not impenetrable shields against external reasons, merely selective filters.

The best example of a filtering view in the contemporary business ethics literature is Joseph Heath’s market failures approach (2006; 2014; 2023). The basic idea of the approach is that people in business roles are entitled to set aside a broad range of ordinary moral considerations while doing their jobs, but they remain accountable to certain welfarist reasons, specifically reasons related to Pareto efficiency. Thus, managers are permitted to undercut competitors by slashing prices, even where this causes great hardship to family-run outlets, but they are not permitted to collude with rivals to keep prices high, for that creates a deadweight loss. It is a picture of selective filters, narrowing the range of reasons that business actors are accountable to. For Heath, business actors’ filters do not exclude efficiency-related reasons because efficiency is the underlying purpose of a market economy. Those same actors may, however, disregard reasons of compassion, fairness, or even honesty (within limits: see Heath 2023, 107–17), so that they may more freely pursue the competitive strategies that contribute to market efficiency.

Many balk at the market failures approach. To follow just one line of criticism: on Heath’s view, managers can adopt purely contractual relations with employees, offering the lowest wages they can get away with, rather than a higher wage that better compensates employees for their contributions to the firm (Heath 2018). But surely, the objection runs, this is unconscionable: businesses ought to observe a higher standard of fairness in their hiring practices and labor relations. I’m sympathetic to this objection, but we should distinguish between three ways of formulating it, for these formulations carry very different burdens of proof. The first option is to revert to the balancing model. Corporate managers, like all role actors on this model, remain accountable to all the reasons that apply to persons simply and thus can only act against reasons of fairness for the sake of weightier reasons. The second option is to deny that business roles are legitimate, in the sense necessary for grounding second-order reasons. There is a line to be drawn, perhaps somewhere between criminal attorney and corporate executive. The third option is to challenge Heath from within the terms of the filtering model, arguing that a broader—but still limited—range of first-order reasons escape the second-order scope of the filters of the managerial role. It is not just efficiency-based reasons that business managers must continue to act on, but also some fairness-based reasons, particularly in labor relations. I think that the third strategy is the more promising direction for future scholarship; the challenge is to pursue it in a way that is systematic rather than ad hoc. ²⁷

The filtering account also introduces another distinction: some of the filters of a role might be discretionary, granting a permission or right to exclude some range of first-order reasons, rather than a strict obligation or requirement to exclude those reasons. Heath often presents the ethics of market exchange as permissive: beneficial transactions and innovations come from allowing market actors to compete in an undirected way, where profit maximization is one among many ends that a business and its agents may pursue (Heath 2023, 67–68, 141, and elsewhere). These second-order permissions grant business role actors the license to exclude a range of normative reasons, including reasons of fairness, compassion, and honesty. Business actors are often still allowed to act on these ethical considerations—they violate no role obligation in giving them their due weight—but they also are within their role rights in paying no mind to such matters. ²⁸

If business ethics is indeed a structure of second-order permissions, then many cases of alleged immorality will be like the case of the tobacco lobbyist. In my final evaluation of that case, I suggested that Nick Naylor exercises a permission to exclude first-order reasons, rather than an obligation to do so. Naylor has, as I put it, discretion—or as he puts it, “moral flexibility.” This means that our judgments of his conduct must be qualified: though he could give weight to a broader range of reasons, he nonetheless acts within the prerogatives of his role when he chooses not to. If managers are permitted but not required to exclude considerations of fairness in labor relations, our judgments of their conduct should be similarly ambivalent. Some degree of immorality will persist: just as free speech can still be foolish or obnoxious, the permissible squeezing of wages can still be palpably unfair. That moral vocabulary still applies, because the role does not prevent managers from paying their employees more: they are just permitted not to. We do not miss the mark—do not overlook the “proper necessities” of the role—when we call these wage-squeezing managers greedy or heartless. But like Nick Naylor, many leaders of the business world will be unphased by such accusations. It will be enough for them that they act permissibly, given the discretionary filters of their role.