Institutionally Driven Moral Conflicts and Managerial Action: Dirty Hands or Permissible Complicity?

In the philosophical literature on the subject, the terms ‘irresolvable moral conflict’, ‘moral conflict’ and ‘moral dilemmas’ are often used interchangeably. Either term is used to describe a situation in which a moral actor is said to be under an obligation to carry out two (or more) incompatible courses of action, such that he cannot do both. In such a situation, the agent is doomed to moral failure regardless of what he does.

It is helpful to distinguish between irresolvable and apparent moral conflicts. Opponents of moral dilemmas use the modifier ‘apparent’ so as not to grant that these choices are irresolvable. In other words, to opponents of dilemmas, these situations only appear to be dilemmatic, but in reality there is a right course of action that does not generate a moral residue. If one wishes to be agnostic about the existence of genuine dilemmas, one might refer to such choices as potential conflicts.

In January 2010, Google again surprised the world by announcing that it was considering pulling its search engine operations out of China after it had been targeted by a highly sophisticated cyber-attack thought to have originated from within the country. In the same announcement, David Drummond, the company’s Chief Legal Officer, also mentioned Google’s growing concerns with the Chinese government’s increasing restrictions on freedom of speech and information (Drummond 2010). According to news reports, at least one of Google’s founders, Sergey Brin, argued for ceasing Google’s operations in China due to what he described as the ‘earmarks of totalitarianism’ in the government’s censorship and surveillance of human rights activists. Brin claimed that increased government censorship of the Internet and interference with the company’s operations reportedly made many Google executives increasingly uneasy with the prospect with Google’s policy of self-censorship (Vascellaro 2010b). However, CEO Eric Schmidt had apparently pushed to remain in China, in order to try to bring change over the long-term (Vascellaro 2010a).

In the literature on dilemmas, these choices are interchangeably referred to as ‘moral conflicts’, ‘irresolvable conflicts’, ‘unresolvable conflicts’, ‘tragic conflicts’, ‘tragic choices’ and ‘(genuine) moral dilemmas’. See, for instance, Hill’s (1996) discussion defence of Kantian moral theory against the possibility of moral dilemmas. In it, he refers to moral dilemmas, tragic conflicts, tragic choices and unresolvable conflicts interchangeably. In the same collection of essays, Mason (1996) refers to moral dilemmas and tragic conflicts interchangeably, as do Dahl (1996), Gowans (1996) and McConnell (1996). For the sake of clarity, henceforth, I will primarily use the term ‘moral dilemmas’ or ‘dilemmas’ to refer to such choices.

McConnell words his definition in a way that allows for more than two required courses of action: ‘the agent is required to do each of two (or more) actions; the agent can do each of the actions; but the agent cannot do both (or all) of the actions’ (McConnell 2010). Sinnott-Armstrong points out that this simple definition allows for dilemmas in which the agent ought not to do something, because we can say that ‘the agent ought to adopt the alternative of not doing that thing’ (1988, p. 5). For the purposes of this paper, I will abide by these thinkers’ definition of moral dilemma.

In the literature on moral dilemmas, ‘moral residue’ and ‘moral remainder’ are terms that are used interchangeably to refer to the emotions experienced in the face of dilemmas. I will refer primarily to ‘moral residue’ or ‘residues’.

Brenkert distinguishes between two ways in which we can use the word ‘moral compromise’. The first sense of has to do with compromises that are the result of political negotiations on contentious issues. While he does not rule out that Google and the Chinese government engaged in this kind of moral compromise, this is not the sense of compromise that interests him; rather, the sense of moral compromise that he has in mind is the result of idea of irresolvable moral conflicts (Brenkert 2009, pp. 462–463).

Brenkert argues that utilitarianism, Donaldson’s (1992) ‘Compatibility Proviso’ and what Martin Benjamin calls the ‘Doctrine of Moral Harmony’ (1990, p. 82) are deficient because it cannot account for the moral residues that result from moral conflict (2009, pp. 463–464).

We do not even need the terminology of ‘moral compromise’ in order to understand Brenkert’s diagnosis of the problem. We might instead say that Google’s managers faced a genuine moral dilemma, forcing them to make a choice that involves inevitable wrongdoing.

Brenkert cites reports of police harassment, fines and imprisonment of Chinese portal companies’ employees (2009, p. 467). I would point out that this last duty is conditional upon the decision to set up operations within China, so that this responsibility would not be incumbent upon Google were it to decide not to operate in China in the first place.

Also cited in (Brenkert 2009, p. 467).

Brenkert also speculates that Google is likely an important life project for founders Sergey Brin and Larry Page, as well as for other Google employees. However, a life project might be conceived differently depending on the person, such that it may weigh in favour of self-censorship in one person’s mind while weighing against it in another one. For this reason, I do not include this consideration in this analysis.

Brenkert uses the term self-censorship to refer to what search engines in China must do with regard to their search results, i.e. filter results they believe the government would consider too dangerous or controversial to be made available to the public. Amongst others, George Brenkert and Thompson (2006) use self-censorship to describe what Google and other search engine providers do to comply with the censorship regime. A search for ‘self-censorship in China’ turns up many stories of how the media in China and other parties censor themselves so as not to run into trouble with the government. I also use the term ‘self-censorship’ in this way.

The use of the term ‘self-censorship’ may be controversial, because we generally tend to think of self-censorship as refraining ourselves in a purely voluntary manner without reference to an external authority, e.g. deciding against posting something on Twitter or Facebook for fear of offending others. Nevertheless, the use of the term self-censorship seems reasonable in this case, given the way in which the censorship requirements are enacted by the Chinese government. As Thompson points out, the government itself does not do the censoring of the search results, nor do they provide a comprehensive ‘blacklist’ of terms. Instead, they issue intentionally vague guidelines and rely on the discretion of the search providers to determine what ought to be censored. Thompson raises the possibility that such a strategy likely results in more comprehensive censorship than relying exclusively on direct censorship by the government.

‘Don’t be evil!’ appears at the beginning and the end of the company’s code of conduct (Google 2012a). An alternative version, “You can make money without doing evil,” appears in the ten principles outlining their corporate philosophy (Google 2012b).

In making this point, Brenkert cites the following from Bernard Williams: ‘[M]any people can recognize the thought that a certain course of action is, indeed, the best thing to do on the whole in the circumstances, but that doing it involves something wrong’ (Brenkert 2009, p. 19; Williams 1972, p. 93).

For instance, in describing what constitutes complicity, Gardner (2007, p. 132) argues: ‘There is what I should do simpliciter, and then there is what I should do by way of contribution to what you do. If I fail in the first I am a principal. If I fail in the second I am an accomplice’. Lepora and Goodin (2013, p. 6) make a similar argument about complicity constituting causal contributions to another agent’s wrongdoing: ‘Complicity […] necessarily involves committing an act that potentially contributes to the wrongdoing of others in some causal way’. Wettstein (2012) goes further in writing about corporate silence in the face of human rights violations constituting complicity, even though the causal connection is not immediately obvious.

For these reasons, along with Lepora and Goodin (2013), I would argue that there are more and less morally acceptable ways of being complicit. This allows me to argue that the Google case may constitute a case of permissible complicity.

To make this point, Williams uses the example of Agamemnon’s tragic choice between his duties as commander-in-chief of the Greeks and his role of father in the face of the goddess Artemis’ wrath against his army (Williams 1965, p. 111).

I use the phrase ‘to regret or feel bad about’, because there is some disagreement as to whether regret would even be appropriate in this case. Foot herself argues that regret ought to be reserved for cases in which we choose or consent to a course of action and subsequently wish we had not. Thus, in cases of moral dilemma, so long as one is convinced that one has done the right thing, regret is not the appropriate term for one’s emotional reaction after the fact. See Foot, ‘Moral Dilemmas Revisited’, 185.

Of course, it is possible to take such a position. For instance, see (Hirshman 2006). My objection to such a view is that it does not pay due respect to women’s freedom to pursue what they deem to be a meaningful life and denies that such a life can be found inside the home. However, this is a separate debate that cannot be suitably addressed in this context. For the time being, I can only appeal to the intuition that the moral requirement that all women ought to work outside the home is overly demanding and does not permit women enough freedom to be masters (mistresses) of their own destinies.

For instance, in support of this view, Brenkert (2009, p. 464) cites Isaiah Berlin. In ‘Political Ideas in the Twentieth Century’, Berlin (1950) speaks specifically of the necessary trade-off between individual freedom and the guarantee of a basic minimum in the political realm, and, as is clear from ‘Two Concepts of Liberty’, he believes that the sacrifice of one value to another can also happen in the personal sphere: ‘If, as I believe, the ends of men are many, and not all of them are in principle compatible with each other, then the possibility of conflict—and of tragedy—can never wholly be eliminated from human life, either personal or social. The necessity of choosing between absolute claims is then an inescapable characteristic of the human condition’ (1969, p. 169).

The term ‘agglomeration principle’ first appears Williams (1965, p. 118).

Vranas (2007) has offered a thorough defence of ‘ought implies can’, and Donagan (1984) has offered a way of understanding rationalist systems of morality in a way that makes it possible for opponents of moral dilemmas retain the agglomeration principle.

Although he does not invoke Friedman, Brenkert does cite the fiduciary duties of managers to shareholders (2009, p. 467) when noting the competitive pressures facing Google (2009, p. 466). He also points out that, given these competitive pressures, the failure to make the investment in mainland China could be quite detrimental to shareholders in the long run, and that Google has duties stemming from the law, morality and economic theory to continue further its business interests (2009, p. 466). However, for a competitive context to generate the conditions necessary for moral compromise, managerial responsibilities would have to be highly demanding, binding managers to commit wrongs in order to fulfill them. Other than citing competitive pressures, Brenkert does not propose an argument for thinking that managerial responsibility is as demanding his account of moral compromise requires it to be. However, one possible way to buttress his argument would be to invoke a highly demanding account of property rights, such as Milton Friedman’s. I am grateful to an anonymous reviewer for pointing out this possibility.

Specifically, Stout invokes Fischer Black and Myron Scholes work on pricing options, which indicates that it does not make anymore sense to ‘say the debtholder has purchased the right to the corporation’s future profits from the corporation while also selling a call option […] to the shareholders, as say the shareholders purchased the right to the corporation’s profits from the company but have also bought a put option […] from the debtholders’ (Stout 2012, p. 38).

She also argues against the idea that shareholders are residual claimants, pointing out that there are fundamental differences between what she calls living and dead corporations, the latter being corporations going through the bankruptcy proceedings in which shareholders are treated as residual claimants. Under normal circumstances, i.e. in living companies, shareholders are not residual claimants and are, therefore, not entitled to getting funds out of a company unless the board of directors decides to pay out a dividend (Stout 2012, pp. 39–40).

A good example of this would be (Wettstein 2009). Additionally, see Scherer and Palazzo (2011) for a review of the literature regarding political conceptions of corporate responsibility.

Leaving aside the motivations of those who would politicize the corporation, it is possible that politicization of the corporation opens the door to managerial dilemmas simply by making managers answerable to the demands of public morality. However, this is true only if the corporation is the right kind of political institution. Recall the two key differences between public and private morality. The first is that public officials are expected to act with more impartiality than private actors. Even if we do not go as far as Milton Friedman in eschewing social responsibilities for managers, we certainly do not see managers responsible first and foremost for the public good. When we do ascribe responsibility for the public good to managers, it is not clear that we do so in the same way as for politicians. For example, generally we do not see managers as responsible for promoting our collective interests with the same emphasis on impartiality. The fact that we talk about the importance of respecting, advancing and balancing stakeholders’ interests is itself an implicit rejection of the impartiality we expect from politicians or public officials more generally. After all, political impartiality in pursuit of the public good would manifestly disallow the stakeholder approach to management—to say nothing of shareholder primacy—because it is a species of favouritism, giving greater weight to certain parties’ interests over others. The second key difference between public and private morality is that public morality emphasizes the importance of consequences more than private morality. In contrast, the same is not typically said about the corporation or business managers. Managers are to make decisions with the interests of those who are designated as stakeholders in mind, and not in terms of some kind of far-reaching utilitarian or consequentialist calculus.

This is one of the two ways of defining ‘direct complicity’. The other is that direct complicity occurs ‘when a company provides goods or services that it knows will be used to carry out the abuse’ (UN Global Compact 2011). However, even under this definition, the distinction with Brenkert’s obedient complicity is valid, because filtering search results in itself would not be a human rights violation in itself as forced relocation would be. The UN Global Compact proposes two features that define ‘corporate complicity’. The first is that the company or an individual acting on its behalf acts or omits to act in such a way ‘that ‘‘helps’’ (facilities, legitimizes, assists, encourages, etc.) another […] to carry out a human rights abuse’. The second is that complicit agent does so with the knowledge that the contributory ‘act or omission could provide such help’.

Due diligence entails: ‘assessing actual and potential human rights impacts, integrating and acting upon the findings, tracking responses, and communicating how impacts are addressed’ (Ruggie 2011, p. 16).

One might point out that the Chinese government could claim that they do not harbour the intention to directly infringe or violate their citizens’ human rights, but that such infringements are necessary so as to achieve other laudable goals, such as economic growth and political stability. However, such an objection cannot be taken seriously if one holds the view of human rights as something inviolable. The claim of human rights entails a claim on behalf of the rights-bearer to be treated in certain ways by the bearer of the correlated duty, regardless of what other goal the duty bearer is trying to achieve. Therefore, there are certain means that the duty bearer cannot pursue, even for the purpose of realizing something praiseworthy. In this case, it is the state that bears the responsibility to respect and secure certain fundamental rights and freedoms. Therefore, in enforcing a censorship regime, the state is failing to live up to its responsibilities to act appropriately towards its citizens. In other words, even if we take the claim that rights are being violated in the pursuit of other laudable goals, such as political stability or economic wealth, the claim of human rights entails that there are certain means of pursuing those goals that are impermissible; namely, human rights violations.

On a related note, when it became obvious to Google’s executives that the censorship regime was becoming more restrictive, it was Brin who pushed for the company to exit China’s search engine business (Vascellaro 2010a, b).

This condition shares some similarities with Wettstein’s (2012) argument corporations have a positive duty to speak out against human rights violations. However, Wettstein’s framework relies on the politicization of the firm, whereas mine does not.

More recently, in discussions of just war theory, Walzer (2006) and Luban (2011) have argued for a revised, stricter version of the DDE that requires soldiers act with the intention not to harm civilians, as opposed to merely not having the intention to harm them. This more recent version of DDE goes beyond requiring an absence of the intention to harm. Instead, it requires an intention not to harm, even if this means that the agent must take on additional risk to himself. This account of DDE would seem to yield the conclusion that Internet search providers are not permitted to establish operations in China at all.

With respect to this point, I would argue that, even if my account tracks DDE in this way for this case or any other similarly relevant cases, the two differ as to the explanations that they each offer for why a given case of complicity is or is not permissible. My account would look to see if the manager was acting out of an intention to minimize complicity and was expressing this intention to relevant constituents. Someone applying DDE would look to verify that the manager was not intending to harm and would balance the manager’s good end with the harm brought about from the foreseen, but not intended, side effect. In one sense, this point of difference between the account proposed here and DDE responds to one of the criticisms that has been levied against DDE by its critics; namely, that it may pick out cases of permissible exceptions to certain moral rules, but it does not capture or adequately explain what makes them exceptions (Scanlon 2008). If one finds this criticism of DDE compelling, then the advantage of my account over DDE is that it seeks to explain what makes certain acts of complicity permissible in a way that is more specific than DDE.

I would like to thank Alan Strudler for raising this objection using this particular example. I would also like to thank an anonymous reviewer for raising this objection more generally.

This article has been adapted from the dissertation, “Managerial Action in the Face of Moral Conflict: Role Responsibilities and Moral Dilemmas” (2013). Available from ProQuest. Paper AAI3568064. http://​repository.​upenn.​edu/​dissertations/​AAI3568064.