1. Introduction
In his well-known paper,Footnote 1 John Doris argues that situationist psychology impugns Aristotelian descriptive psychology, which posits the existence of situation-independent virtues (although it should be acknowledged that he is silent on other virtue ethical frameworks). In retaliation, virtue ethicists have tended to argue that the situationist challenge is a strawperson, which erects a caricature of the virtuous person.Footnote 2 In fairness, it must be admitted that the general situationist position lacks nuances, and is therefore susceptible to misconstrual and misapplication. In an insightful response to Doris, Maria Merritt reframes and clarifies the debate by suggesting that situationist psychology is lethal against an Aristotelian form of virtue ethics, but not a Humean conception which respects the importance of social relationships and settings for securing robust dispositional traits. Specifically, she contends that Aristotelianism is distinctly susceptible to the situationist challenge because it posits “Motivational Self-Sufficiency of Character” (MSC)—that is, motivational traits that are “independent of factors outside oneself”Footnote 3 as requisite for virtue, whereas Humeanism does not.
Because Merritt’s analysis deflects the pervasive strawperson objection by clarifying precisely what type of virtue psychology is at issue, I will take this position as my starting point for evaluating the tenability of virtue ethics on the whole. Since Merritt’s paper was published in 2000, it has gained considerable traction with character theorists such as Doris (in his 2010 publication), Christine Swanton, and Peter Vranas. Nonetheless, I believe that Merritt’s argument is incomplete in two significant ways, because it fails to answer some of Doris’ more trenchant objections. The main issue is that Merritt’s response to Doris does not directly assuage his concerns about virtue ethics’ descriptive moral psychology, and some of these objections apply as much to Humeanism as to Aristotelianism. Perhaps Merritt fails to recognize the scope of these objections, or sees them as orthogonal to the more urgent project of discrediting a strong ideal of MSC as embodied in Aristotelianism; but whatever the reason, she does not directly answer two of Doris’ strongest objections. These are presented as two horns of a dilemma, as follows:
1. If virtue ethics requires reflection on the regulative ideal of a virtuous exemplar, then it is subject to the problem of excessive ‘theoretical mediation’ promulgated by Stocker and Williams. According to Stocker, virtuous people act directly on virtuous motives, such as love, friendship, fellow-feeling, and community, without the mediation of theoretical heuristics or abstract calculations (such as a utility-maximizing formula or a categorical imperative). This means, for instance, that when one has an authentic friendship, one acts directly for the friend’s sake, not for any higher-order aim. If one allows regulative ideals to mediate one’s friendship-oriented actions, one incurs ‘moral schizophrenia,’ i.e., a divorce between one’s personal motives and one’s theoretical reasons. This psychological condition is not merely “uncomfortable, difficult, or even untenable,” but actually self-annihilating, for it renders “our lives essentially fragmented and incoherent.”Footnote 4 On Williams’ account,Footnote 5 utilitarians lack moral integrity because they must view their moral motives as mere sensations instead of constitutive elements of their moral personalities, and must regard their life projects as dependent upon the projects of others. Thus, they are estranged from their moral values and prevented from forming authentic personal projects. In light of these objections, Doris articulates the following problem. On an Aristotelian view, cultivating the virtues is difficult, and thus virtuous exemplars are rare. As a result, most people must rely on virtue-based mental representations as mediating ideals, and attempt to instantiate them in their actions. However, if this is the correct psychological model of virtue-acquisition, then Aristotelians must invoke regulative ideals to explain how the majority of humans act virtuously. But doing so goes against the Stocker-Williams objection to excessive theoretical mediation.
2. On the other hand, if virtue ethics is not supposed to rely on mediating psychological heuristics, it requires something impossible: that we achieve the status of moral saints. According to Aristotle’s Nicomachean Ethics, cultivating a virtuous disposition is “not for everyone, nor is it easy,” and so “goodness is both rare and laudable and noble.”Footnote 6 Advocating this view, Rachana Kamtekar argues that situationism does not discredit virtue ethics, because a small percentage of Milgram’s experimental subjects acted well, and “if virtue requires practical wisdom, one would expect virtuous persons to be rare.”Footnote 7 That is, virtue ethics is tenable because exemplars are supposed to be few and far between, which explains why most subjects acted invidiously in the experiment. This captures a common (though not unexceptionable) interpretation of Aristotle, as well as represents a common response to the situationist challenge—i.e., that the few defectors were genuinely virtuous, as one would expect, while the majority were vicious. The problem with this explanation is that it casts us onto the other horn of the dilemma, for if virtuous exemplars are rare, then virtue ethics is psychologically untenable for normal humans, unless we use heuristic devices. The result is that unless we have fairly decisive reason to think that we are exceptional (and we should be wary of thinking this in light of our well-documented susceptibility to self-serving biasFootnote 8 and perceptual distortionsFootnote 9), we are unwarranted in subscribing to virtue ethics as a practical program. Hence, we should endorse a less demanding, less perfectionistic ethical theory, such as utilitarianism or Kantianism.
In sum, virtue ethicists must grapple with two problems. On the one hand, an Aristotelian conception of virtue ethics that eschews regulative ideals appears to be psychologically untenable for ordinary humans; but on the other hand, a theoretically-mediated virtue ethics is unacceptable because it engenders excessive theoretical mediation. One option for the virtue ethicist is to ignore the Stocker-Williams objection and embrace a reflective conception of virtue psychology, but Doris points out that, if we do this, we encounter a third problem: that, absent further argument, we have no reason to prefer the regulative ideals of virtue ethics over and above those of utilitarianism and Kantianism. All things considered, it appears that we should favour the latter theories, as they appear, prima facie, to require less effort. On scrutiny, this objection involves a pluralistic metaethical presumption, which Merritt describes as follows:
For purposes of his argument, Doris takes on board a widely shared understanding of virtue ethics as competing directly, in the same line of business, with dominant ethical theories such as consequentialism and Kantianism. He concludes that the main impact of situationism on virtue ethics is to wipe out whatever competitive advantage it might have had in this contest.Footnote 10
Merritt proposes to defeat this impasse by describing virtue ethics as agent-centered as opposed to act-centered, and dispositional as opposed to rationalistic. This is supposed to defuse the competitive picture by “shifting the terms of the discussion to a different type of practical problem, as opposed to merely offering a different way to deal with the same type of practical problem.”Footnote 11 Her argument in a nutshell follows these contours: in the first place, Humeanism differs from Aristotelianism in that it does not require strong MSC; rather, it permits us to cultivate character traits by embracing social relationships and settings that are likely to last. Second, Humeanism does not engender excessive theoretical mediation because it is not an act-centered theory, and thus does not over-emphasize critical reflection. Rather, it is a dispositional, agent-centered theory, and thus hardly needs regulative ideals at all. This view is encapsulated in the familiar dictum that virtue ethics is concerned with how to live, not how to act. Therefore, Doris’ criticism is misguided on two scores: First, it falsely depicts virtue ethics as act-centered, and secondly, it depicts virtuous behaviour as primarily rationalistic, as opposed to primarily dispositional.
Using this argument, Merritt purports to pass between the two horns of the dilemma. However, I do not think that Doris’ objections can be so easily bypassed. The first problem is that while Merritt may make some headway in the debate by pointing out that virtue ethics is agent-centered and dispositional as opposed to act-centered and rational, she does not give us any reason to prefer the former picture of moral psychology to the latter. Perhaps utilitarianism and Kantianism are ultimately more compelling from an empirical perspective. We cannot decide until we assess their respective psychological implications. Thus, we need a competitive metaethical defense of virtue psychology. This recapitulates Doris’ concern that the virtue ethicist must provide an argument “to the effect that her favored regulative ideals (however, exactly, they are to be construed), are better suited to effecting morally desirable conduct than alternatives offered by her competitors.”Footnote 12 That is, we need to know why, when we reflectively deliberate, we should prefer a virtue ethical model. I aim to answer this question, or at least provide the basis of such an answer, in what follows. More research into moral psychology will help to fill in the details.
To be more specific about my intentions, in the next two sections, I aim to substantiate the following two arguments:
1. Excessive theoretical mediation is a problem not just because of Williams and Stocker’s very valid reasons, but also because rational deliberation tends to be disabled, overwhelmed, or swamped in morally challenging circumstances, rendering these processes useless. That is, in Milgram-type situations, these psychological mechanisms are inert. This means that rationalistic theories that prioritize such mechanisms are flawed. In light of contemporary research into the psychology of expert performance, we should embrace a dispositional account that privileges unconscious cognitive mechanisms.
2. This does not mean that there is no place for reflective deliberation in moral psychology, but simply that such processes are secondary in the execution of ethical thought and action. This view is supported by Joseph Rouse’s psychological model of expertise, which holds that reflective deliberation is at best infrequently useful and at worst counterproductive:
Conceptual understanding (or reflective deliberation) is superfluous wherever we have become skillfully responsive to circumstances, since we can be flexibly and appropriately responsive without any intervening conceptualized representations. Reflective conceptual articulation is also antagonistic to skilled engagement with the world: in stopping to think, we would dissolve any smooth-flowing skilled bodily attunement to what is taking place.Footnote 13
The same considerations extend to ethical expertise and excellence—that is, reflective deliberation can be superfluous and antithetical to virtuous behaviour when we have achieved excellent dispositions. Nonetheless, it may be needed at certain junctures. Because virtue ethics mirrors this distribution of cognitive labour, depicting deliberation as a secondary function, it provides the most accurate model of moral psychology. Moreover, because virtue ethics is supposed to be a non-codifiable, non-rule-based theory, whereas rationalistic theories posit a single action-guiding principle (the utility-maximizing formula or the categorical imperative), virtue ethics is more compatible with what Doris calls “variantism,” the view that different principles are apposite to different situations. Thus, virtue ethics is more empirically adequate than the invariantist alternatives.
2. Dispositionalism Versus Rationalism
As we saw, Merritt tries to deflect Doris’ first objection by emphasizing that virtue ethics is an agent-centered theory as opposed to an act-centered one, but she does not give us decisive reason to prefer the former view. In response to the second objection, she admits that reflective deliberation plays a secondary role in virtue ethics, but does not elucidate or defend this psychological picture. In this section, I address the first problem, and turn to the second problem in section 3.
The consensus amongst philosophers and social scientists is that subjects of situationist experiments act against their better judgment (or deep-seated values or will) because their faculties are, in some sense, “overwhelmed,”Footnote 14 “overridden,”Footnote 15 “swamped” or “short-circuited”Footnote 16 by situational factors. Initially, it is unclear what faculties, or parts of the agent, are supposed to be ‘overwhelmed’ in this manner. If we survey the literature on the Milgram experiments—easily the most famous of all situationist psychology experiments—we find a number of vague and speculative explanations. However, there is one interpretation that has been particularly influential and may be taken as the standard view. Ross and Nisbett explain the Milgram subjects’ behaviour in terms of the following “relevant features of the situation”:
“the stepwise character of the shift from relatively unobjectionable behaviour to complicity in a pointless, cruel, and dangerous ordeal,” making it difficult to find a rationale to stop at one point rather than another. Second, “the difficulty in moving from the intention to discontinue to the actual termination of their participation,” given the experimenter's refusal to accept a simple announcement that the subject is quitting—“The experiment requires that you continue.” Third, as the experiment went on, “the events that unfolded did not ‘make sense’ or ‘add up’ …. The subjects' task was that of administering severe electric shocks to a learner who was no longer attempting to learn anything …. [T]here was simply no way for [subjects] to arrive at a stable ‘definition of the situation.’”Footnote 17
This explanation, which is reiterated and reinforced by philosophers as diverse as Harman, Flanagan, Miller, Zimbardo, and Sosa (who hold widely discrepant views on virtue ethics), suggests that the subjects’ rational faculties were temporarily short-circuited by situational factors. The subjects could not ‘find a rationale’ to stop, ‘make sense of’ the experimenter’s instructions, ‘define the situation,’ and so on. These descriptions imply that the relevant situational factors prevented the subjects from formulating and executing a rational plan of action. In like fashion, Flanagan reiterates Ross and Nisbett’s stepwise explanation of the experimenter’s instructions, remarking that “the gradual stepwise character of the situation creates a justification problem for the subject. He has to ‘find a justification … that would explain why he had to desist now, when he hadn't desisted earlier; how it could be illegitimate to deliver the next shock but legitimate to have delivered one of only slightly less magnitude moments before.’ A sort of moral sorites problem exists here.”Footnote 18 This reference to justifications and the sorites paradox suggests that there was a defect of rational calculation. If this is the correct interpretation, then rationalistic theories such as utilitarianism and Kantianism, which, according to Doris, require agents to perform rational decision-making procedures during or just prior to acting, cannot be action-guiding, since they rely on cognitive processes that are ineffectual precisely when, according to the standard reading, these processes are most crucial. Now, while it must be admitted that this rationalistic (cognitivist) view is not the only possible psychological interpretation of these two theories, it a very common one, endorsed not only by Doris, but also Flanagan, Williams, McDowell, Stocker, and Hursthouse, to name but a few. Moreover, it is generally taken to be the distinguishing feature of these theories in contrast to virtue ethics. Insofar as this is right, utilitarianism and Kantianism appear to be useless in Milgram-type situations.
Although Ross and Nisbett’s explanation is the gold standard, Doris gives an alternative account, which depicts the subjects as unable to express their character. On this view, “the situationist data can … be read in two ways, corresponding to our two readings of Aristotelianism.”Footnote 19 The first reading is a ‘dispositionalist’ one which defines virtue in terms of behavioural regularity, and the second is an ‘intellectualist’ one which defines virtue as perceptual acuity. Doris disputes both of these readings by citing Darley and Batson's Good Samaritan Experiment, in which Princeton seminary students were instructed to give a talk on the other side of campus, and encountered a ‘victim’ slumped and moaning in a doorway on their route. Out of three differential variables, the only one that made a difference was time: ten percent of those in a hurry stopped to help, compared to 45 percent of those in a moderate hurry and 63 percent of those in no hurry. Doris infers that “time pressures swamped subjects’ dispositions to help someone they perceived to be in need of assistance.”Footnote 20 This refutes the dispositionalist reading, since the students’ prosocial dispositions did not manifest themselves in the situation.Footnote 21 Alternatively, it is possible that “some ‘hurried’ non-helpers failed to help … because their haste … dampened the awareness required to notice that someone was in need of their assistance,”Footnote 22 which undermines the intellectualist reading. In addition, Doris cites S.J. Ceci’s ‘contextualist’ theory of intelligence,Footnote 23 according to which “cognitive ability in a particular area [often fails] to ‘transfer’ to closely related areas: mathematical acuity in the classroom, for example, may not strongly correlate with ability to apply similar computations in work-related settings,”Footnote 24 which reinforces the second objection. Thus, Doris’ reading, as opposed to Ross and Nisbett’s, emphasizes the subjects’ dispositional failings.
These arguments are leveled against Aristotelianism. However, it is notable that they can also be lodged, with one minor modification, against rationalistic ethical theories. This is because these theories are amenable to the same two psychological readings available to the Aristotelian, provided that we include rational calculating as a type of ‘intellectualist’ capacity. Specifically, utilitarians and Kantians have two options for understanding moral agency: Either they see the agent as acting dispositionally, in which case they are vulnerable to Doris’ first objection, or they see the moral agent as acting on perceptual and/or rational information (which is the more common reading), in which case they are susceptible to Doris’ second objection and the Ceci example. The central problem is that utilitarians and Kantians, just like Aristotelians, must identify and respond to morally-relevant factors, which requires recognitional and decision-making capacities. However, if Doris is right, these capacities are compromised by situational variability. Taking this into account, it appears that no ethical theory is able to withstand the situationist challenge.
Now in reading Doris’ paper, it may appear at first blush that he is intent on defending utilitarianism and Kantianism against Aristotelian virtue ethics, but this is not so. Rather, Doris is only interested in showing that Aristotelianism is just as bad as the other two, if not worse. Thus, all three are seriously flawed. If this were the final word, it would leave us with no positive ethical theory—in other words, ethical nihilism—which is an unacceptable conclusion, since we need ethics to help us navigate the problems of daily life. But on closer inspection, Doris can be seen as offering a completely new theory based on situationist psychology, which we might call ‘situational ethics.’ This view holds that people should not trust in their (perceived) character, but should aim to avoid “morally dangerous circumstances,”Footnote 25 i.e., circumstances that may tempt us to act against our better judgments, wherever they occur. To illustrate this approach, Doris asks us to imagine that we are invited to a romantic dinner by a flirtatious colleague to whom we are strongly attracted, while our beloved spouse is out of town. Doris says, if we want to remain faithful, we should avoid the dinner like “the plague,”Footnote 26 since it is a known fact that humans are susceptible to cheating. This seems to imply that, as a rule, we should try to avoid all morally dangerous situations.
The obvious objection to this approach is that, if we consider the psychological possibilities for putting it into practice, it is just as susceptible to Doris’ anti-Aristotelian objections as the other rationalistic theories. That is, if we want to avoid ‘morally dangerous circumstances,’ we must be able to identify and respond to them, and these recognitional and reason-responsiveness capacities invoke the two contested readings of Aristotelianism: either we recognize and respond to morally-dangerous situations dispositionally, or we do so intellectually, using perceptual and/or rational capacities. Since Doris has rejected both of these options, we are left with no psychological basis for situational ethics. This once against leaves us at an impasse.
To move the discussion forward, I propose a different approach to the situationist challenge, which resurrects virtue ethics in a psychologically viable form. Specifically, I want to defend a primarily-dispositional account of virtue ethics, which acknowledges reflective deliberation as an auxiliary component in moral psychology, drawing on Rouse’s theory of expertise, together with cognitive science examples showing that habituation, operant conditioning, and subconscious processes are highly effective methods for acting and living well. At least, these mechanisms are likely to be highly effective in stable political climates such as ours, in accordance with Merritt’s Humean defense. Provided that we can first secure a degree of situational regularity by, for instance, building strong interpersonal relationships, we can cultivate stable virtues.
To see this, first consider Rouse’s psychological theory. In his essay on expertise and human psychology, Rouse reiterates Dreyfus’ view that “our most basic perceptual responsiveness to our surroundings consists of a kind of ‘mindless coping’ with situations that are meaningful to us as bodily agents.”Footnote 27 This responsiveness does “not require the use of implicit or explicit representations, reflective deliberation, or even self-consciousness (in the form of a Kantian ‘I think’ or even a Merleau-Pontyan ‘I can’)”: rather, “such bodily skillfulness responds to our surroundings as a field of attractions and repulsions, experienced as tension and its release, or a smooth, coordinated bodily ‘flow.’”Footnote 28 This description clearly fits well with situationism’s view of human behaviour as highly dependent upon and reactive to situational variables. But it also accords with a Humean dispositional account of ethical behaviour, which privileges the role of emotional responsiveness to the world, bringing these two theories into harmony.
Rouse then cites Dreyfus’ view that skilled activity bypasses “rules and reflective considerations … by developing a direct bodily responsiveness to the overall configuration of the situation”Footnote 29:
In acquiring a skill … [there] comes a moment when we finally can perform automatically. At this point we do not seem to be simply dropping these [previously deployed] rules into unconsciousness; rather we seem to have picked up the muscular gestalt which gives our behavior a new flexibility and smoothness. The same holds for acquiring the skill of perception.Footnote 30
Dreyfus supports this account using the examples of blitz chess (which is played under time constraints) and professional baseball. He says that experts in these activities “do not have a concept in mind, but instead respond directly to the affordance or solicitations of a situation on the board or field.”Footnote 31 This explains why professional chess players can move their piece in under two seconds per turn, and professional ballplayers can reliably hit a fast-moving pitch. There is no time, and no need, for the intervention of conscious deliberation. That said, Rouse admits that there is a need for reflective deliberation at certain junctures of skillful activity. Conceptual (explicitly rational) deliberation, he says, plays an important role in acquiring skills and correcting mistakes. In the blitz chess example, “the grandmaster’s ability to recognize and respond almost instantaneously to complex patterns on the chess board is the outcome of an extended ‘selective’ regime (artificial selection involving study of past games leading to recognition of significant configurations) …”Footnote 32 That is, grandmasters must study rules and strategies and critically reflect on their mistakes before and after gameplay. However, this does not substantiate a rationalistic moral psychology, because virtue ethics allows for the intermittent operation of rational deliberation as a means of establishing checks and balances. Rationalistic theories, by contrast, give this mechanism pride of place compared to reflexive responses. The difference between the two accounts is one of degree: while both recognize the importance of two distinct processes, they assign them differential significance. On Rouse’s model, virtue ethics is the more accurate picture of how human beings develop robust competences.
Now one may worry that neither view can secure a positive outcome in Milgram-type situations, which are completely unpredictable. This may unfortunately be true; perhaps in radically strange situations, one cannot act well either rationalistically or dispositionally. However, we can find comfort in the fact that most situations in modern western societies are relatively predictable—in fact, so predictable that many people lament the boredom of modern life. If we combine Merritt’s prescription to cultivate stable, positive relationships and settings as much as possible, with Rouse’s model of psychological expertise, we obtain an endorsement of Humean virtue psychology. If in life, as in sports, there is a degree of predictability, then cultivating dispositional traits should be possible. By contrast, rationalistic theories are psychologically untenable even in relatively stable conditions, since they place too much emphasis on rational processes—processes that are both psychologically secondary, according to Rouse, and normatively inappropriate, insofar as they alienate use from personal values such as friendship, community, and integrity.
My second argument in support of virtue ethics adapts an example from Timothy Shroeder’s analysis of moral responsibility. His account of driving, I believe, offers a striking illustration of how “automatic, operantly conditioned action”Footnote 33 can guide us effectively through complicated situations, in which our rational faculties are off-line. Shroeder asks us to imagine a driver who has developed a reflective impulse to turn at a familiar intersection:
Think of a person who, because of the rewards and penalties attached, has formed the habit of turning left at a certain intersection say, towards a place of work. The person turns left, out of habit, for years, but then gets a new place of work. The old workplace is torn down, and now there is nothing but industrial desolation in that street. Even knowing all this, the person may still find himself with a tendency to turn left at the familiar intersection: a meaningless behavioural tendency. The tendency will not be exercised when the person is paying attention, but when distracted at a crucial moment, the old operant conditioning may take over and produce a left turn in response to the familiar stimulus. This does not appear to be an example of an action produced by a desire. Our imaginary driver does not have an intrinsic desire to make left turns, or to go to a workplace he does not even believe exists any longer. All he has is a lingering tendency to turn left at a certain intersection.Footnote 34
Shroeder explains that impulses due to operant conditioning are not intentional states, but rather, “activity produced entirely through the action of perception, the basal ganglia, and the motor cortex,” and hence “such activity would reflect no present desire.”Footnote 35 He offers this example to show that moral responsibility is compromised in people with Tourette Syndrome, who exhibit tics due to a defect of the basal ganglia and motor cortex. However, I reproduce it here to illustrate that automatic, operantly conditioned impulses can reliably produce characteristic behaviour (i.e., consistent with our evaluative judgments) in situations of moral stress, when rationality is disabled. If the Ross-Nisbett thesis is correct and our rational processes are impaired in the face of overwhelming situational pressures, then we cannot rely on a rationalist picture of moral psychology as a program for navigating difficult situations. Although ethical dispositions may not be efficacious in every situation, the driving example suggests that they may remain operational in rationality-disabling circumstances. Thus, it is reasonable to surmise that with a little apposite training and experience, a person can (conceivably) develop robust ethical habits, just as a motorist can develop robust driving habits, which continue to function in the face of distractions, obstacles, and surprises. If we consider the Good Samaritan experiment in light of this example, it is a reasonable conjecture that if one makes an initial resolution to help suffering victims, then one may, over time, develop an effective disposition to identify and help such victims, even when rational processes are preoccupied with computational or prudential tasks (such as calculating the time required to cross campus, preparing for an urgent assignment, and so on), and therefore unavailable for other purposes. Thus, one could possibly develop a penchant to be a Good Samaritan using dispositional methods. This, of course, does not mean that dispositional training is a silver bullet, or even a particularly effective method; but if the Ross-Nisbett interpretation is correct, then it is our best and last hope, since these mechanisms are our only available resource.
I came across another impressive example of operantly conditioned behaviour in a recent news broadcast about a Newfoundland woman who unwittingly struck and killed a moose with her car on the way to work, and was hit in the head by the Moose’s hoof. Surprisingly, she continued driving to work along her normal route perfectly safely, apparently in an unconscious state, and finally arrived at her destination. She later told reporters, “I can remember pulling into the driveway and my co-workers came out and asked me if I was okay and I said ‘Why?’ and they said ‘Well, Michelle, you're bleeding and look at your car.’ I looked at my car and I was devastated.”Footnote 36 This woman had no recollection of hitting the moose or driving to work, although she navigated the road perfectly. This, again, demonstrates the efficacy of unconscious processes in the absence of rational competence, which lends credence to a dispositionalist ethics.
In fact, this psychological hypothesis is corroborated by a plethora of cognitive science research into morality and the unconscious mind—indeed, too much to cite here. This body of research finds that unconscious cognitive processes are more reliable than conscious ones for performing morally relevant tasks such as judging the world, making decisions, predicting outcomes, achieving success, and the like. The literature is prodigious, but a few examples should suffice for an introduction. In “Strangers to Ourselves: Discovering the Adaptive Unconscious,” Timothy D. Wilson argues that “the mind operates most efficiently by relegating a good deal of high-level, sophisticated thinking to the unconscious, just as a modern jetliner is able to fly on automatic pilot with little or no input from the human, ‘conscious’ pilot. The adaptive unconscious does an excellent job of sizing up the world, warning people of danger, setting goals, and initiating action in an efficient and sophisticated manner.”Footnote 37 These skills are requisite for behaving ethically, so unconscious processes are essential for this purpose. This cognitive picture is also endorsed by Naline Ambady and Robert Rosenthal, who found that students could accurately assess a teacher’s professional competence in fewer than 30 seconds, and by John Gotman, who found that it was possible to predict divorce outcomes by analyzing the first few minutes of a newlywed couple’s conversation. This implies that unconscious cognition produces accurate impressions of the world and helps us make reliable predictions. Similarly, in “Blink: The Power of Thinking Without Thinking,” Malcolm Gladwell argues that intuitions and snap judgments “can be every bit as good as decisions made cautiously and deliberately.”Footnote 38 In one example, he describes a veteran military strategist named Paul Van Riper, who subscribes to the belief that “war is inherently unpredictable and messy and non-linear,” and “analytic, systematic decision making” is inferior to experientially-informed intuitions.Footnote 39 In 2002, Van Riper was commissioned by the U.S. military to lead a traditional military regiment (without advanced technology) in a simulation exercise against a team of soldiers equipped with “the Operational Net Assessment, a formal decision-making tool that broke the enemy down into a series of systems—military, economic, social, political,” and used this data to predict the opponent’s likely course of action.Footnote 40 The U.S. military expected the technology to carry the day, but Van Riper’s team easily trounced its opponent. This corroborates the view that intuitions and experience often trump rational decision procedures in navigating difficult situations.
This evidence is not dispositive (as no social psychological data can be), but it gives ballast to two related arguments: First, that unconscious dispositional processes, such as operantly conditioned impulses, instincts, and intuitions, are more effective than explicit rational decision-making for ensuring ethical behaviour, and secondly, that these processes are the only available resources in startling, rationality-disabling situations, which are the focal point of situational analysis. These two arguments mutually support a virtue ethical model of moral psychology.
In the next section, I turn to the question of whether virtue ethical regulative ideals are preferable to utilitarian and Kantian ones, embodied in the utility-maximizing formula and the categorical imperative, respectively.
3. Virtue Psychology and Variantism
I begin this section by describing an account of moral responsibility defended by Doris in “The Moral Psychology Handbook” (MPH). This is the view that when people make moral responsibility judgments, they apply discrepant criteria to different cases—to judgments about “abstract questions and concrete questions, about morally good behaviors and morally bad behaviors, about close friends and complete strangers.”Footnote 41 Doris calls this theory ‘variantism’. Although it is presented as a descriptive account in the first instance, it is supposed to have (tentative) moral implications. At the end of the chapter, Doris considers whether variantism is normatively the right theory to subscribe to, and begins by stating that an adequate ethical theory “will lie somewhere on the continuum between extreme conservatism and extreme revisionism.”Footnote 42 Thus, ethical theories should be constrained by what philosophers call “ordinary practice.”Footnote 43 This view is reflected in Doris’ earlier book, “Lack of Character,” where he discounts consequentialism because it is “radically revisionary,” whereas an adequate ethical theory must be at most “conservatively revisionary.”Footnote 44 So, since variantism accurately describes our moral practice, it follows that it must also have strong normative appeal. Making a stronger pronouncement in MPH, Doris remarks that, while the data is not yet conclusive, “one cannot help but be struck by how difficult it has been to develop an invariantist theory that does not fall prey to counter-examples. Perhaps it is time to give variantism a try.”Footnote 45 This means that if Doris had to bet, he would place his chips on variantism.
One of Doris’ reasons for favouring variantism is that it captures our attitude toward personal relationships. He cites a study by Arriaga and Russbult which found that married subjects tended to assign less blame to spouses with whom they had a high perspective-taking relationship than to others: “in other words, people’s attributions of blame to an agent seemed to depend on their relationship to that particular agent, with people assigning lower amounts of blame to those agents whose perspectives they were especially likely to take.”Footnote 46 To the extent that Doris is congenial to normative variantism, he must endorse the idea that perspective-taking relationships should positively influence our moral judgments. Indeed, apart from the empirical research, there is support for this view in moral philosophy. Flanagan, and later Hurley and Darwall (respectively), argue that an adequate ethical theory must be minimally psychologically realizable for creatures like us, since only such a theory can be minimally action-guiding, and thus fulfill the practical aims of ethics. Therefore, insofar as impersonal/agent-neutral theories (such as utilitarianism) alienate us from our moral motives, they are inadequate, as they fail to give special status to our personal/agent-relative reasons such as friendship.Footnote 47 This means that, ceteris paribus, we ought to be especially forgiving to our loved ones. In this respect, Flanagan’s ‘principle of minimal psychological realism’ (PMPR) corresponds with normative varianism. Interestingly, Flanagan also takes this principle to corroborate virtue ethics, since this theory explicitly recognizes the importance of personal motives.
My reason for beginning this section with variantism is that, if this view is correct on both a descriptive and a normative level, then we must select an ethical theory that conceives of moral judgments as varinatist in nature, in the sense of applying discrepant criteria to different situations. Utilitarianism and Kantianism do not fit this description, inasmuch as they are single-principle theories that apply an invariant moral criterion to every possible situation. Situational ethics also misses the mark, inasmuch as it requires that we act on an invariant principle stating that we should categorically avoid morally dangerous situations. If this is not the correct interpretation of this theory, then Doris does a poor job of illuminating its proper application. As I see it, his theory does not advise that we should avoid some morally-dangerous situations and approach others with caution, which is arguably the more intuitive prescription. (How are we to improve if we never challenge ourselves?) Rather, it issues a categorical directive to avoid danger. Virtue ethics, however, is different from these single-principle rationalist theories in that it does not purport to give a single, or even a finite set of, general-purpose rule(s) or decision-procedure(s) for how to act. It is for this reason that Merritt describes it as non-action-based and non-reflective, as well as why G.E.M. Anscombe, Hursthouse, and McDowell characterize it as non-codifiable and non-juridical. In Anscombe’s seminal paper on modern moral philosophy,Footnote 48 for instance, she emphasizes that the defining feature of virtue ethics compared to alternatives is that the former is not a ‘juridical’ view, but rather a context-sensitive, non-codifiable one that directs us to use practical wisdom to decide how to act in each unique situation. She says that justice, for instance, cannot be coherently defined in categorical terms, since it differs according to the particularities of the case. For example, letting someone die might be unjust under most circumstances, but justified if the person is suffering from an agonizing terminal illness and has requested passive euthanasia or assisted suicide. Most importantly, determining whether an action is just or unjust is not simply a matter of applying pre-conceived decision-procedures, but must be done by considering the particularities of each case.
McDowell and Hursthouse both reject codifiability, though to different extents. McDowell simply states that virtue ethics is not ‘codifiable,’ meaning that it is not reducible to universal rules like utilitarianism and Kantianism. Rather, “occasion by occasion, one knows what to do, if one does, not by applying universal principles but by being a certain kind of person: one who sees situations in a certain distinctive way.”Footnote 49 In this way, the virtuous person must use practical wisdom to judge situations on a case-by-case basis. Somewhat differently, Hursthouse contends that virtue ethics might be “weakly codifiable,” in that it may give rise to “v-rules” (or rules of thumb) such as “do what is charitable,” “do what is courageous,” and so on; but it is not “strongly codifiability,” as it does not yield (a) “decision procedures for determining what the right action was in any particular case,” or (b) rules that “a non-virtuous person could understand and apply … correctly.”Footnote 50 In other words, even if there are v-rules, they do not amount to decision-making procedures along the lines of the utility maximizing principle and the categorical imperative, and they are not comprehensible without the use of a situation-sensitive faculty of practical wisdom. In considering these distinctive features of virtue ethics, it becomes clear that this theory is more compatible with variantism compared to the alternatives.
These considerations suggest that virtue ethics and situationism are an ideal pairing, since virtue ethical judgments are essentially variantist, i.e., situation-sensitive. I propose that we test this theory using a familiar thought experiment. Following Philippa Foot, imagine that a trolley is hurtling out of control, and heading toward five people tied to the track. You can pull a lever to divert the trolley to another track, but then it will kill one other person. Either way, you will kill someone. What do you do? According to the standard interpretation, if you are a consequentialist, you will pull the lever, since this action maximizes utility, whereas if you are a Kantian, you will abstain from intervening, since pulling the lever would treat the resulting victim as a mere means—i.e., it would violate the humanity formulation of the categorical imperative. But what would the virtue ethicist do? As Flanagan emphasizes, the distinctive feature of virtue ethics is that it respects agent-relative reasons such as concern for one’s loved ones. With this in mind, imagine that the single person on the side-track is one’s child. For the utilitarian and the Kantian, this factor makes no difference. (This is why these theories are referred to as impartial and agent-neutral.) Since virtue ethics is capable of giving special moral status to the parent-child relationship, it agrees with the normative-variantist view that we do, and should, privilege close relationships. Thus, in this specific case of moral judgment, virtue ethics and variantism coalesce, while the other two theories diverge. This illustrates the stronger affinity between virtue ethics and variantism.
These considerations support the conclusion that situationism is more compatible with virtue ethics than with the three alternatives: utilitarianism, Kantianism, and situational ethics. This is because virtue ethics places emphasis on moral perception and practical wisdom, which allows us to respond to the unique features of each situation. By contrast, the other theories force us to apply a single, invariant principle in every situation, regardless of context. Thus, although Doris explicitly rejects Aristotelian virtue ethics, in this section, I suggest that he ought to embrace some version of virtue ethics.
4. Concluding Remarks
In this paper, I have addressed the question of whether situationism has the implications that Doris thinks it has. I have argued that it does not, and that, in fact, the opposite implication holds: situationism is more compatible with virtue ethics, on a certain conception, than with the alternatives, because virtue ethics is primarily dispositional and committed to variantism through the operation of practical wisdom. This position is supported by an appropriate understanding of the psychology of expertise, and insight into the underlying psychological mechanics of virtue ethics. If we combine these arguments with Merritt’s defense of a low-MSC Humean virtue ethics, we have a solution to Doris’ dilemma: a virtue ethical theory that recognizes the importance of social relationships and settings, and is dispositional and non-codifiable in nature. Although the compatibility problem has been outstanding since the publication of Merritt’s article in 2000, a solution is urgent in light of emerging research, which uncritically combines situationism with moral education strategies, without providing a prior metaethical case for their compatibility. Examples of this work include Heather Battaly’s, “Acquiring Epistemic Virtue: Emotions, Situations, and Education,”Footnote 51 Francesca Gino’s, “Sidetracked: Why Our Decisions Get Derailed and How We Can Stick to the Plan,” and Daniel Kahneman’s, “Thinking, Fast and Slow.” These accounts use situationist research as a basis for drawing ethical conclusions, but they require the arguments of this paper to support the presumption of compatibility.
