¹ Stocker, Michael, ‘The Schizophrenia of Modern Ethical Theories’, The Journal of Philosophy 73 (1976), pp. 453–6CrossRefGoogle Scholar.

² Noteworthy examples include: Barbara Herman, ‘On the Value of Acting from the Motive of Duty’, The Philosophical Review 90 (1981), pp. 359–82; Baron, Marcia, ‘The Alleged Moral Repugnance of Acting from Duty’, The Journal of Philosophy 81 (1984), pp. 197–220CrossRefGoogle Scholar; Adams, Robert, ‘Saints’, The Journal of Philosophy 81 (1984), pp. 392–401CrossRefGoogle Scholar; Railton, Peter, ‘Alienation, Consequentialism, and the Demands of Morality’, Philosophy and Public Affairs 13 (1984), pp. 134–71Google Scholar; Louden, Robert, ‘Can We Be Too Moral?’, Ethics 98 (1988), pp. 361–78CrossRefGoogle Scholar; Kagan, Shelley, The Limits of Morality (Oxford, 1989), esp. pp. 357–69Google Scholar; Wilson, Catherine, ‘On Some Alleged Limitations to Moral Endeavor’, The Journal of Philosophy 90 (1993), pp. 275–89Google Scholar; Unger, Peter, Living High and Letting Die (Oxford, 1996), esp. ch. 6CrossRefGoogle Scholar; and Jeske, D., ‘A Defense of Acting from Duty’, The Journal of Value Inquiry 32 (1998), pp. 61–74CrossRefGoogle Scholar.

³ An important exception here is Pettit, Philip, ‘Consequentialism and Moral Psychology’, International Journal of Philosophical Studies 2 (1994), pp. 1–17CrossRefGoogle Scholar. I will return to Pettit, for my conclusion implies that his work on the implementation of consequentialism is indicative of what ought to be a general strategy for ethical theories to meet the challenge of moral schizophrenia.

⁴ Note that my aim is more specific than addressing the overall problem of reconciling the particularity of agents’ personal relationships or integrity with their ethical obligations. Moral schizophrenia is related to this more general issue, but Stocker's challenge is often lumped in with the general problem of capturing goods like friendship and integrity without due attention being paid to the moral schizophrenia that results if ethical theories succeed at being able to capture these goods. Thus, Stocker's challenge is closely related to, but distinct from, other influential challenges to the demands of ethical theory, e.g.: Williams, Bernard, ‘Persons, Character and Morality’, Moral Luck (Cambridge, 1981)CrossRefGoogle Scholar; and Wolf, Susan, ‘Moral Saints’, The Journal of Philosophy 79 (1982), pp. 419–39CrossRefGoogle Scholar.

⁵ The most succinct (contemporary) articulation of this division is found in Bales, R. E., ‘Act-Utilitarianism: Account of Right-Making Characteristics or Decision-Making Procedures?’, American Philosophical Quarterly 8 (1971), pp. 257–65Google Scholar. Other prominent examples of it include: Adams, R. M., ‘Motive Utilitarianism’, The Journal of Philosophy 73 (1976), pp. 467–81CrossRefGoogle Scholar; Hare, R. M., Moral Thinking (Oxford, 1981)CrossRefGoogle Scholar; Parfit, Derek, Reasons and Persons (Oxford, 1984)Google Scholar; Peter Railton, ‘Alienation’; Sumner, L. W., The Moral Foundation of Rights (Oxford, 1987)Google Scholar; Jackson, Frank, ‘Decision-Theoretic Consequentialism and the Nearest and Dearest Objection’, Ethics 101 (1991), pp. 461–82CrossRefGoogle Scholar and Crisp, Roger, ‘Utilitarianism and the Life of Virtue’, Philosophical Quarterly 42 (1992), pp. 139–60CrossRefGoogle Scholar.

⁶ Stocker, ‘Schizophrenia’, p. 454.

⁷ Stocker, ‘Schizophrenia’, p. 462.

⁸ Stocker claims that we want our relations to be ‘infused with emotions’ in the sense that our friends and lovers care for us and not merely recognize our value; see ‘How Emotions Reveal Value and Help Cure the Schizophrenia of Modern Ethical Theories’, How Should One Live?, ed. Roger Crisp (Oxford, 1996), pp. 180–2. However, this emotional engagement need not take the form of overt warmth or passion, since he also claims that warmth is not a necessary feature of acting for the sake of a friend: ‘Values and Purposes: The Limits of Teleology and the Ends of Friendship’, The Journal of Philosophy 78 (1981), pp. 747–65 (755n.).

⁹ Williams, ‘Persons’, pp. 17–18.

¹⁰ Incidentally, this variant of Smith also reveals that Williams's presentation of the ‘one thought too many’ objection is misleading, since the objection is ultimately based on a specific kind of thought that, he claims, should not be in mind during our interactions with loved ones. In other words, the problem is not simply a matter of the quantity of an agent's thoughts, despite what the urgency of his chosen thought-experiment (a rescue situation where one can either save a stranger or one's wife) seems to suggest.

¹¹ Furthermore, Stocker explicitly notes that spontaneity cannot be used to differentiate between acts that are performed for the sake of a friend and acts that promote the friendship without the appropriate concern for the friend herself. See ‘Values and Purposes’, p. 755n.

¹² In fact, Stocker is careful not to use the term ‘impartiality’ in his diagnosis that modern ethical theories are unable to achieve goods like friendship. He instead uses terms like ‘externality’ and ‘impersonality’ when describing what is disconcerting about Smith's confessed reason for his visit.

¹³ Stocker is elsewhere more lucid about the fact that this is the problem exemplified by the Smith example in ‘Schizophrenia’. See ‘Values and Purposes’ and ‘Morally Good Intentions’, The Monist 54 (1970), pp. 124–41.

¹⁴ Stocker, ‘Schizophrenia’, p. 461.

¹⁵ See Williams, Bernard, Ethics and the Limits of Philosophy (Cambridge, Mass., 1985), pp. 9–11Google Scholar; Hursthouse, Rosalind, On Virtue Ethics (Oxford, 1999), pp. 131–2Google Scholar; Driver, Julia, Uneasy Virtue (Cambridge, 2001)CrossRefGoogle Scholar.

¹⁶ In fact, ethical theories beyond those considered ‘modern’ are also capable of endorsing agents with this separation between their motives and reasons, although proponents of virtue ethics rarely acknowledge this fact. This is unfortunate, since virtue ethics should also want to endorse agents like Smithers rather than Smith in order to capture the elusive moral value of virtues like modesty, courage and benevolence. As Thomas Hurka points out, there is considerable irony in the fact that proponents of virtue ethics sometimes object to the fact that modern theories appeal to a divided moral psychology, when virtue ethics is equally (if not more) in need of separating its agents’ immediate motives from the reasons it provides to explain why the virtues are valuable. See Virtue, Vice, and Value (Oxford, 2001), pp. 246–9. The fact that virtue ethics benefits from a separation between principles and motives is also recognized in: Swanton, Christine, ‘Virtue Ethics and the Problem of Indirection: A Pluralistic Value-Centred Approach’, Utilitas 9 (1997), pp. 167–81CrossRefGoogle Scholar; McKerlie, Dennis, ‘Aristotle and Egoism’, The Southern Journal of Philosophy 36 (1998), pp. 531–55CrossRefGoogle Scholar; and Slote, Michael, Morals from Motives (Oxford, 2003), pp. 42–7Google Scholar.

¹⁷ Stocker, ‘Schizophrenia’, p. 458.

¹⁸ This interpretation is probably reinforced by the misconception that schizophrenia is a diagnosis that is synonymous with multiple personality disorder. The two conditions are distinct, but the distinction holds for Stocker's analogy as well. He claims that modern ethical theories precipitate a form of psychological disharmony; he does not claim that modern ethical theories precipitate a psychology in which a struggle for control exists between multiple sources of normative authority.

¹⁹ See, for example, ‘Values and Purposes’, p. 757, and ‘Friendship and Duty: Some Difficult Relations’, Identity, Character and Morality: Essays in Moral Psychology, ed. Owen Flanagan and Amélie Oksenberg Rorty (Cambridge, Mass., 1990), p. 221.

²⁰ Stocker, ‘Morally Good Intentions’, p. 127.

²¹ Stocker, ‘Values and Purposes’, p. 754.

²² Stocker, ‘Friendship and Duty’, especially pp. 219–23.

²³ Dean Cocking and Jeannette Kennett seem perilously close to being committed to this claim when they say it is plausible to think that it is admirable for a character from Death in Brunswick (1990) to be loyal enough to be moved to help a friend cover up an accidental killing of an innocent person (‘Friendship and Moral Danger’, The Journal of Philosophy 97 (2000), pp. 278–96). However, this seemingly counterintuitive claim is balanced by their conclusion that the character fails as a moral agent to do what he ought to. As Railton correctly notes in ‘Alienation’, even an agent's love for her spouse should not be ‘a romantic submersion in the other to the exclusion of worldly responsibilities’ (p. 151). This is not to say that one should not feel conflicted about breaking the bonds of friendship; the point is only that there are at least some clear cases where these bonds must give way to considerations of duty. This fact, and the sense of conflict that one experiences in such situations, is beautifully captured in Carol Reed's film The Third Man (1949).

²⁴ Stocker, ‘Values and Purposes’, p. 754. In this article the friend in the hospital is named Harry instead of Smith, but the situation is otherwise the same. It is striking to notice the similarity between Stocker's sine qua non conditions and the parallel mechanisms presented by proponents of ethical theory. See Railton's counterfactual condition (‘Alienation’, p. 151), Herman's ‘limiting conditions’ (‘Value of Acting’, pp. 372–6) and Baron's ‘secondary motivations’ (‘Repugnance’, pp. 207–9). These mechanisms serve the same function of limiting the actions of agents without necessarily becoming conscious motives during decision-making. It is ironic, then, that they are often understood as cures for moral schizophrenia.

²⁵ See Stocker, ‘Friendship and Duty’, p. 222 and ‘Values and Purposes’, p. 761.

²⁶ Stocker, ‘Friendship and Duty’, esp. pp. 224–9. The point at stake is articulated nicely by Thomas Hurka who points out that Casablanca (1942) and The English Patient (1996) present precisely asymmetric views about the priority of personal relationships and higher political causes. He appropriately concludes that Casablanca is the morally superior film. See Hurka, ‘The Moral Superiority of Casablanca over The English Patient’, The Globe and Mail, January 25, 1997.

²⁷ ‘Friendship and Duty’, esp. pp. 228–30.

²⁸ This view is given its fullest defense in Stocker's Plural and Conflicting Values (Oxford, 1992).

²⁹ See ‘Duty and Friendship’, p. 66. Stocker's view here is strikingly similar to Bernard Williams's claim that no solution is available to reconcile the inevitable tension between critical reflection and the integrity of the moral agent: ‘How truthfulness to an existing self or society is to be combined with reflection, self-understanding, and criticism is a question that philosophy, itself, cannot answer. It is the kind of question that has to be answered through reflective living. The answer has to be discovered, or established, as the result of a process, personal and social, which essentially cannot formulate the answer in advance, except in an unspecific way’ (Ethics and the Limits of Philosophy, p. 200).

³⁰ ‘Friendship and Duty’, p. 223.

³¹ For example, when Bernard Williams objects to R. M. Hare's two-level utilitarianism because it is not stable under reflection – an objection closely related to Stocker's moral schizophrenia diagnosis – Hare is simply baffled: ‘I cannot understand why Williams makes such heavy weather . . . of the combination of critical with intuitive thinking.’ See Williams, ‘The Structure of Hare's Theory’ and Hare, ‘Comments’, Hare and his Critics, ed. Douglas Seanor and N. Fotion (Oxford, 1988).

³² ‘Schizophrenia’, p. 458.

³³ Stocker often seems to assume that there is an in-principle, as opposed to empirical, difference between the burdens of divided versus unified moral psychologies, but this begs the question against ethical theories that have no independent aversion to a divided moral psychology. David Brink articulates this point in the context of defending utilitarianism against Williams's objection that it fails to meet a condition of publicity. See ‘Utilitarian Morality and the Personal Point of View’, The Journal of Philosophy 83 (1986), p. 428.

³⁴ Whether Smithers is balancing the relative strengths of utility, imperfect duties, virtues or obligations to other members of a social contract, it seems a safe assumption that his ethical theory will rule in favor of helping at the children's hospital if the alternatives are presented in the context of a single, isolated choice.

³⁵ Note that the assumption of non-ideal circumstances is significant here, because under circumstances of radical equality the problem at stake will disappear. In other words, if an agent's ethical theory judges that her friends and family are just as deserving of aid than strangers, then the problem described above will not apply. Conversely, it is the assumption of ideal conditions that allows Elinor Mason to successfully give a negative answer to the question, ‘Do Consequentialists Have One Thought Too Many?’, Ethical Theory and Moral Practice 2 (1999), pp. 243–61. Mason is correct to think that Williams intends his objection to hold even in ideal conditions, and she is therefore justified in concluding that his objection fails in these terms. However, a revised version of his objection, presented as a practical dilemma rather than a point of principle, reveals a legitimate problem for ethical theories in non-ideal circumstances.

³⁶ I am assuming, though I think not implausibly, that special relationships like friendship are indispensable for human agents (as opposed to honeybees) because of the way these relationships serve a vital socializing function, e.g. helping us learn to care about the well-being of others. As Elizabeth Ashford notes, ‘It is only in the context of a loving personal relationship that our concern for a person is sufficiently intense to mirror the moral importance of how another person's life goes.’ See ‘Utilitarianism, Integrity, and Partiality’, The Journal of Philosophy 97 (2000), pp. 436–7. The importance of this observation is underrated and, I think, greatly amplified when considered in the context of human moral education.

³⁷ Railton, ‘Alienation’.

³⁸ Strictly speaking, the paradox should be titled the paradox of justified friendship, but I will use the short version for the sake of brevity. Note also that the paradox of friendship is related, but not equivalent, to the more general ‘paradox of loyalty’ as it is described by Pettit, Philip in ‘The Paradox of Loyalty’, American Philosophical Quarterly 25 (1988), pp. 163–71Google Scholar.

³⁹ To draw on the terminology provided by Philip Pettit and Geoffrey Brennan, the benefit of participating in authentic friendships is both ‘calculatively elusive’ and ‘calculatively vulnerable’, because (a) it must be produced by an unreflective disposition rather than by direct calculation and (b) it is unavailable when the relevant disposition is calculatively monitored by a supervisory level of thinking. See their ‘Restrictive Consequentialism’, Australasian Journal of Philosophy 64 (1986), pp. 438–55. It is important to distinguish between these two properties, because only calculative vulnerability generates unique psychological problems. The move to a divided psychology in which an agent's ethical limiting conditions (i.e. sine qua non conditions, counterfactual condition, etc.) do not serve as direct motives to act for the sake of her friends is sufficient to overcome calculative elusiveness without precipitating any special psychological burdens. It is calculative vulnerability that creates the practical dilemma of the paradox of friendship.

⁴⁰ See ‘Alienation’, pp. 159–60.

⁴¹ Note that this point holds whether or not one specifies the criterion of rightness, as Railton does, so that Smithers and Juan perform blameless morally wrong acts when following an overall plan that includes a commitment to friends and loved ones. I am grateful to Roger Crisp for encouraging me to note that the coherence of blameless wrongdoing can be set aside here despite its significance in other contexts.

⁴² A widely read articulation of the fact that Kantianism is at least moderately demanding in this respect can be found in Onora O'Neill, ‘Kantian Approaches to Some Famine Problems’, Matters of Life and Death, ed. Tom Regan (Philadelphia, 1980).

⁴³ See Scheffler, Samuel, The Rejection of Consequentialism (Oxford, 1982)Google Scholar.

⁴⁴ One way of recognizing this is to note that the paradox does not disappear if we assume that all agents subscribe to the same ethical theory. Everyone can accept the same set of limits on friendship imposed by broader ethical obligations, yet the tension between moment-to-moment decision-making and the overall value of friendship will remain in the same way as the paradox of hedonism persists for an individual with a single, hedonistic point of view.

⁴⁵ The persuasiveness of Railton's argument in ‘Alienation’ benefits from: Elinor Mason, ‘Can an Indirect Consequentialist Be a Real Friend?’, Ethics 108 (1998), pp. 386–93. Mason elucidates some of the important empirical assumptions regarding the ethical benefits of friendship that are only suggested in Railton's key example of Juan (‘Alienation’, pp. 150–3 and 159–60), and she rightly emphasizes the fact that the ethical benefits of friendship must be evaluated in a broader context than single instances of decision-making. Unfortunately, Mason invites counterexamples by setting the relevant limiting conditions on friendships too broadly by having agents endorse a general ‘pro-friendship disposition’ that would not be flexible enough to allow for exceptions in emergency cases like, say, Casablanca and The English Patient. Robert F. Card points out this problem for Mason's position in ‘Consequentialism, Teleology, and the New Friendship Critique’, Pacific Philosophical Quarterly 85 (2004), pp. 149–72. Given the overall benefits of friendship, it might be sensible to draw a line in the sand, as Mason does, and set the limiting conditions for agents at the level of a general pro-friendship disposition if one assumes that agents are precluded from the flexibility of a divided moral psychology that allows agents to sometimes make situation-specific evaluations of the value of acting for the sake of their friends. My aim, however, in dispelling the stigma of pathology associated with moral schizophrenia is to cast doubt on precisely this assumption.

⁴⁶ Stocker, ‘Schizophrenia’, p. 458.

⁴⁷ These triggers are part of what Pettit describes as ‘virtual consequentialism’, where the commitment to consequentialist justification is only virtually present in moment-to-moment decision-making in the sense that consequentialist reasoning lies dormant until certain thresholds are met that trigger critical reflection within the agent's conscious deliberations. See ‘Moral Psychology’, pp. 14–16.

⁴⁸ Earlier drafts of this article were presented at meetings of the Canadian Philosophical Association and the British Society for Ethical Theory. I would like to thank Sam Black, Elizabeth Brake, Todd Calder, Roger Crisp, Brad Hooker, Thomas Hurka, Violetta Igneski, Brian Lawson, Ann Levey, Colin Macleod, Dennis McKerlie, Mark Migotti, Emer O'Hagan, Philip Pettit, Peter Railton, Anthony Skelton, L. W. Sumner and Evan Tiffany for their helpful comments.