Skip to main content
Open AccessArticle

The Situationist Critique and Early Confucian Virtue Ethics*


This article argues that strong versions of the situationist critique of virtue ethics are empirically and conceptually unfounded, as well as that, even if one accepts that the predictive power of character may be limited, this is not a fatal problem for early Confucian virtue ethics. Early Confucianism has explicit strategies for strengthening and expanding character traits over time, as well as for managing a variety of situational forces. The article concludes by suggesting that Confucian virtue ethics represents a more empirically responsible model of ethics than those currently dominant in Western philosophy.

Over two hundred years ago, David Hume, impressed by the growing explanatory power of the natural sciences of his time, called upon philosophy to join in the trend toward empirical inquiry, abandoning armchair speculation and a priori abstraction: “Men are now cured of their passion for hypotheses and systems in natural philosophy, and will hearken to no arguments but those which are derived from experience. It is full time that they should attempt a like reformation in all moral disquisitions; and reject every system of ethics, however subtle or ingenious, which is not founded on fact and observation.”1 Due, no doubt, to a widespread disciplinary self-conception that relegated the empirical to the intellectually and ethically irrelevant realm of “heteronomy,” Hume’s call to arms largely fell upon deaf ears, and it is only in the past decade or two that a new movement emphasizing “empirically responsible” philosophy has begun gaining momentum.2 This movement, which in its latest iteration was inaugurated in the early work of Owen Flanagan and Mark Johnson, has argued that philosophical speculation needs to be informed and constrained by our current best empirical accounts of how the human mind works.3 It encompasses positions as diverse as Johnson’s efforts to restore philosophical standing to embodied aesthetics, the work of “neo-Humeans” such as Jesse Prinz, and the so-called experimental philosophy movement spearheaded by Stephen Stich and his students.4

Another manifestation of this “empirical turn” is the work of a group of philosophically minded psychologists and scientifically literate philosophers who argue that evidence about the nature of human cognition emerging from cognitive science, cognitive linguistics, neuroscience, social psychology, and primatology calls into question the psychological plausibility of the “cognitive control” models of ethics—deontology and utilitarianism—that have dominated recent Western ethical thought.5 The apparently foundational importance of emotions, automatic and unconscious processes, and embodied analogical reasoning has led several commentators to suggest that this body of evidence might lend weight to the proponents of the revived “virtue ethics” model of moral reasoning and education that has also in the past several decades been gaining ground in academic philosophy, which—unlike deontology and utilitarianism—emphasizes the ethical importance of social roles, emotions, habits, and imaginative extension.6 If deontology and utilitarianism require us to think or behave in manners that are simply not possible or sustainable in quotidian life, these modern defenders of virtue ethics contend, this should temper our enthusiasm for adopting them as moral ideals.7

Several proponents of the “empirical turn” in philosophy are, however, much less sanguine about the empirical viability of virtue ethics. As far back as Flanagan’s Varieties of Moral Personality, there have been suggestions that social psychological findings concerning the apparent weakness of character traits in the face of situational pressures might call into question the very possibility of “virtues” as stable character traits. This criticism has become even more focused and pointed in the work of Gilbert Harman and John Doris, who argue that the very notion of moral “character”—the bedrock of any virtue ethic—has been empirically discredited.8 It has therefore become quite clear that any contemporary attempt to defend virtue ethics on empirical grounds must address this “situationist” critique.9

Several philosophical defenders of virtue ethics have risen to the challenge, questioning both the interpretation and the significance of the social psychological research cited by Harman and Doris, the notion of “character” that is under attack, and the degree to which historically accurate models of virtue ethics are, in fact, vulnerable to the “lack of character” argument.10 Here I would like to build upon these efforts in a manner that takes on more directly the viability of the empirical claims being made, as well as the degree to which they can be seen as fatal to virtue ethics.

My argument against the situationist critique will unfold in two stages. To begin with, I will question both the empirical and the conceptual foundations of what I call the “strong” situationist position, and I will attempt to demonstrate to my colleagues in philosophy that the supposedly fatal situationist argument is not nearly as lethal as advertised. Personality traits are alive and well, which means that the cognitive foundation of virtue ethics is, in fact, in rather good shape. I will go on to suggest that, even if we acknowledge that traditional notions of character set an extremely high bar for the virtues, there are features of the early Confucian virtue ethics tradition that can be seen as effective and empirically plausible responses to even this more significant challenge.11 Framed in terms of the high bar metaphor, Section I of this article argues that people do, in fact, have the natural capacity to jump (i.e., character traits do exist), while Section II then explores the manner in which early Confucian moral training simultaneously boosts this natural capacity through training and lowers the bar several notches by means of situational controls. It is this combination of enhanced jump and lowered bar that may—and I leave this as an open empirical question—make something like Confucian ethics a psychologically realistic model of virtue cultivation and therefore a potentially valuable resource for contemporary ethical theorists and educators.

I. Personality Traits Are Alive and Well

The entire person versus situation debate was kicked off in 1968 by Walter Mischel with his landmark study, Personality and Assessment, and the decade or so that followed saw an explosion of the sorts of classic studies, for example, Darley and Batson’s study of Princeton theological seminarians and Isen and Levin’s study of the effect of finding a dime in a pay phone on subsequent helping behaviors, that feature so prominently in the work of Harman and Doris.12 The demise of the public pay phone and the depreciation of the dime are not, however, the only things that have changed since the 1970s: personality psychologists, spurred on and informed by the situationist critique, have developed more nuanced models of what a trait might be, as well as more sophisticated methods for exploring the connection between personality traits and behavior. The result has been a massive body of evidence documenting the existence of a robust and diverse set of personality traits. In a recent and important special issue on the person-situation debate, David Funder declares that the debate “ended as a serious scientific conversation decades ago.”13 This is partially because the dichotomous nature of the debate has been recognized as fundamentally mistaken: persons and situations are no more separable than genes and environments, and a strong form of the person-situation contrast is as conceptually muddled as a strong form of the nature-nurture debate. In this regard, the work of Mischel and others has performed a valuable service in debunking early, naively strong views of character as invariant and immune to situational effects.14 Of course, as with gene-environment interactions, much hangs on the relative causal efficacy one attributes to the two factors, and the thrust of Funder’s argument is that the causal efficacy of personality traits can no longer be plausibly denied.

A. Strong Situationism: The Antiglobalist Argument

The one fundamental disagreement that still remains, and the crux of the position that I will be calling “strong situationism,” is whether relatively “broad” personality traits can be seen as having any predictive efficacy, or if it is rather best to see persons as characterized by a motley collection of quite narrow, “local,” extremely situation-sensitive traits. In his recent commentary on the person-situation debate, Mischel maintains as the central thrust of his work a critique of “the classic assumption of high cross-situational consistency in trait-relevant behavior,” and his current position is that individual consistency in “character” consists of a relatively stable “signature” of quite narrow “if … then …” dispositional tendencies, such as the tendency of a given child to be verbally aggressive when chastised by an adult on the playground but unaggressive when approached by a peer.15 This version of situationism forms the backdrop of Harman’s claim that local traits “do not count,” as well as Doris’s “antiglobalist” position that local traits are ethically “fragmented,” cohering only in “evaluatively disintegrated” loose associations.16 There are at least two sets of problems with this critique of the “classic” conception of character, one involving the empirical data on broad character traits and the other a conceptual confusion about the local-global distinction.

B. Empirical Issues

The vast bulk of the situationist literature demonstrating that broad character traits have negligible predictive power is based upon “one-off” assessments of subjects’ behaviors in a particular experimental environment. The problem with this approach is that it misses the “aggregation effect”: the extent to which a clear correlation between character traits and behavior may only begin to emerge over repeated observations over a long period of time. In the 1970s and the 1980s, when the situationist literature had its heyday, there were few large-scale longitudinal studies of the sort that could pick out aggregation effects, primarily because such studies are difficult and quite expensive to pull off. However, in part as a reaction to situationism, personality psychologists in the past few decades have accumulated a wealth of evidence from longitudinal studies demonstrating the reality of broad character traits. For instance, few now would dispute that the so-called Big Five personality traits—openness, conscientiousness, extraversion, agreeableness, and neuroticism—are “real” in the sense that they are stable over time, although they are not entirely unalterable or context insensitive; they appear to have a considerable genetic component; and they predict substantive life outcomes, such as mortality, health, marital satisfaction, divorce, and occupational success.17

To his credit, Doris addresses the aggregation effect, but he dismisses its importance on the grounds that “observers want to predict and explain not only general trends but also particular behaviors” and that the sorts of predictions given by personality traits leave us “completely in the dark” about what a particular person’s behavior might be on any given specific occasion.18 This brings us to another major problem with the strong situationist critique, one that straddles the empirical and conceptual: a severe underestimation of the power and pragmatic usefulness of relatively small correlations. One of the key arguments in Mischel’s groundbreaking 1968 work was that the correlation between broad personality traits (such as “conscientiousness”) and behavior and also within broad personality traits (such as “honesty”) across a variety of trait-relevant situations never seems to exceed 0.3 in any given observation—this is a degree of correlation dismissed by Mischel as “weak.”19 This 0.3 figure became famous, or infamous, as the so-called personality coefficient, and the essence of the argument against broad character traits hangs on the claim that a correlation coefficient of 0.3 is of negligible significance.20

It is not, and unfortunately philosophers who are simply told that it is often lack the formal knowledge of statistics or intuitive understanding of probability to critically evaluate the claim. To begin with, it is important to see that correlation coefficients in the 0.3 range are not a unique feature of personality research. Gregory Meyer and colleagues point out in an important meta-analysis that few correlations in psychology exceed 0.3 and that this is not even a particular weakness of psychology as a discipline: comparing psychological studies with correlations established in a wide variety of fields, they found that effect sizes in psychology are similar to, for instance, those used to justify medical interventions in the health sciences.21 Indeed, the link between consuming aspirin and warding off heart attacks and also that between chemotherapy and positive outcomes in breast cancer hover around 0.02 or 0.03 (that’s an extra zero), and yet this is deemed significant enough by the medical community to make recommendation of these interventions standard practice. Moreover, other large-scale meta-analyses have found that the “situation” effects obtained in the situationist literature, when translated into a common metric, give a correlation coefficient of 0.2–0.3—that is as “weak” if not weaker than the supposed “personality coefficient.”22 If a correlation of 0.3 is genuinely so weak as to be negligible, then nothing predicts behavior.

Fortunately, 0.3 is actually quite good: Robert Rosenthal and Donald Rubin, for instance, observe that a 0.3–0.4 correlation is enough to predict dichotomous outcomes 65–75 percent of the time.23 Perhaps most helpful for the statistically challenged (including myself) is an example that spells out the significance of correlation coefficients in terms of an everyday baseball analogy.24 Ted Williams is considered the greatest batter who ever lived, whereas Bob Uecker, while widely admired for his later efforts as a color commentator, is generally regarded as one of the worst hitters to ever play in the major leagues. Williams’s lifetime batting average was .344 as compared to Uecker’s .200, a difference of only .144. As Robert Abelson points out, the supposedly “weak” correlation of 0.3 is actually three to twenty-seven times more predictive of whether, say, a conscientious person is likely to behave in a conscientious fashion on any given occasion than the difference between Williams’s and Uecker’s lifetime batting averages would predict the likelihood of their getting a hit when at bat. And yet, with two out and the bases loaded in the bottom of the ninth, only a fool would want Bob Uecker rather than Ted Williams coming up to bat for their team. In contemporary major league baseball, differences in batting averages (or earned run averages for pitchers) much less than .144 dictate enormous variations in the salary and prestige of players—this economic evidence suggests that even quite small correlation effects can have very significant practical implications.

C. Conceptual Issues

This debate about the significance of correlation coefficients of a particular magnitude begins, of course, to straddle the empirical and the conceptual. On a more purely conceptual level, there are additional points upon which the strong situationist position can be criticized.

To begin with, the philosophical bite of the “antiglobalist” critique derives from the claim that “local” traits, whose existence no one would dispute, are not really traits at all (i.e., as Harman puts it, that “narrow dispositions do not count”) or that they are so ethically and evaluatively “fragmented” that they cannot perform any of the conceptual lifting that virtue ethics would require of them.25 This point, in turn, is entirely dependent on the implicit assumption that we know what we are talking about when we contrast “local” with “global” traits, that is, that there is a clear, principled distinction between the two. This is not at all the case. When offered as an analytic dichotomy, the “local” versus “global” distinction is simply not tenable because any truly “local” trait would not be a “trait” at all but merely a single occurrence: John performed behavior X in situation Y at this particular time and place. The sorts of local traits that Doris thinks worthy of our attention count as “traits” because they are already abstract to various degrees: it is not merely that John behaved in an honest or extroverted fashion in the presence of Joe on April 24, 2009, at 3:45 p.m., but that he reliably behaves honestly on tests or is extroverted with friends. Looked at in contrast to a truly one-off report of behavior, local traits are already quite broad in their predictive claims. Anything that we can call a “trait” or a “disposition” is already more or less global or broad, which seriously undermines the blanket dismissal of “local traits” or “narrow dispositions” as ethically irrelevant.26

It is thus clear that “local” and “global” mark off opposite ends of a range of abstraction rather than the sort of analytic dichotomy that the antiglobalist critique needs to possess to have any real traction. Once we realize this point, it becomes equally clear that what degree of locality or globality “counts” fundamentally depends on one’s pragmatic needs: the range of behavior one is interested in making predictions about and the degree of predictive reliability one feels comfortable with. As David Funder argues, extremely fractionated traits do give us high fidelity but “at the cost of narrowing the bandwidth” to the point that they are often pragmatically useless.27 The degree of abstraction in trait formulation that strikes one as desirable depends, in the final analysis, on what kind of trade-off between fidelity and broad usefulness one is willing to accept, an observation that leads Richard Lucas and Brent Donnellan to conclude that the “local” versus “global” trait debate is best seen as a debate among personality psychologists with different pragmatic goals, not one between personologists and situationists.28

To understand the force of this point, it is helpful to consider the study by Hugh Hartshorne and Mark May that is often cited as the locus classicus of the “local” versus “global” critique.29 Following the behavior of eight thousand school children over a variety of situations involving what they perceived as “honesty,” the authors found a very high degree of correlation (in the 0.7 range) within specific behaviors such as (i) cheating on written or puzzle test; (ii) cheating on homework, faking a record in athletic competition, faking or cheating in party games; (iii) stealing from a box of money left out; and (iv) lying about any of these. In contrast, they found “little” correlation—0.227, to be precise—between these four local tendencies. In other words, if one views 0.227 as an unacceptably weak correlation, the conclusion is that the children in this study exhibited local “honesties” (“honest when taking a test”) but no such thing as “honesty” in the way we would normally use the term.

As I hope I have established above in the discussion of correlation coefficients, there is no reason to reach this conclusion—0.227 can do quite a bit of work for us. While it is wonderful and potentially quite useful to know that we can predict with great certainty that little Sarah is going to cheat on a test, what we are often interested in doing is extrapolating from one specific type of behavior to another that is perceived as relevant. Assume that I am a youth camp counselor and that all that I have at my disposal at the moment is information about which kids tend to cheat on tests and which kids tend to cheat in games. I now have to decide whom to leave in charge of the donation jar at an important fundraising event. I am an underpaid counselor at an underfunded youth camp, quite concerned about making sure that none of this badly needed money goes unaccounted for, rather than an academic psychologist interested in achieving extremely high correlations to impress journal referees. I do not see how it can be denied that 0.227 is good enough for me: I’m going to pass over little Sarah, whom I saw cheating on the exam, and little Johnny, who habitually cheats at Monopoly, when I choose my donation jar monitor, and this is an entirely rational decision given my available information and pragmatic needs. It is quite reasonable to assume that the origin of “global” trait terms like ‘honest’ or ‘brave’ arose precisely out of this pragmatic need for global prediction: over time, correlations in the 0.2–0.3 range are extremely significant, and we would expect that people interested in extrapolating from one behavioral tendency to another related one would latch onto such significant connections and invent labels for them.

It is, nonetheless, the case that if we could follow Hartshorne and May’s children into adulthood, we would be likely to find genuinely weak correlation between their broad “honesty” with regard to their professional behavior and their honesty with regard to, say, displaying fidelity to their spouse. The idea that this is a problem for virtue ethics hinges on another conceptual misunderstanding—one that begins to nudge us toward our final historical case example because it revolves around the intended scope of traditional virtue terms.

At one point in Lack of Character, Doris makes the helpful observation that “if your mechanic is honest in working on your car, you can commend her honesty to potential customers without worrying that she cheats on her taxes.”30 We could note that Doris fails to emphasize how global this attribution of professional honesty already is—an “honest” mechanic not only refrains from adding spurious items and services to your bill but also refrains from stealing valuable objects from your car, replacing parts that still have some useful life in them, and in general putting her own financial interests above that of her clients. Nonetheless, he is probably correct that even a quite robust attribution of “honest” to a mechanic would fail to helpfully predict, for instance, the degree to which she refrained from cheating sexually on her spouse.

When presented as a critique of traditional virtue ethics, however, this observation fails to take into account the fact that traditional virtue terms, such as ‘honesty’ or ‘bravery’, were actually used in a relatively narrow context when compared to their modern folk usage. Consider the closest thing we could find to a term for ‘honesty’ in Warring States (sixth century to third century BCE) China: the virtue term xin , usually translated as “trustworthiness” or “reliability.” There is always a bit of debate concerning the precise connotation of traditional virtue terms such as this, and their usage also has varied somewhat from thinker to thinker and over time, but what is beyond dispute is that the scope of xin is confined to a gentleman’s professional behavior toward his colleagues, superiors, and/or inferiors, and the idea that it should also encompass an elite male’s sexual fidelity to his wives and concubines would have been incomprehensible to the early Chinese.31

What is true for xin is also true for other traditional Confucian virtue terms: because the concerned actors form a relatively small subset of the population and because the social realities involved are much more clearly structured, Confucian virtue terms inevitably have a narrower scope than their English translations do in contemporary usage. I suspect that this is also true of virtue terms derived from Aristotle or later Western virtue ethicists, although I am not qualified to assert this with any confidence. In any case, the broader issue here is that a small set of virtue words inherited from traditional, highly structured societies are now arguably being asked to bear too much weight in the sort of unstructured, complex environments that characterize modern life in the industrialized world. The idea that the social psychology literature demonstrates that there is no empirical support for the existence of character traits (e.g., Harman in “Moral Philosophy”) picked out by such words as ‘honesty’ or ‘courage’ really depends on one’s personal predilection regarding trade-offs of accuracy versus usefulness. In any case, even if defensible, such a claim should be seen not so much as an indictment of traditional virtue ethics as a symptom of a problem with modern English: it has failed to innovate linguistically as the structure of society has radically changed.

However, even this more limited critique is probably overblown: it is not at all clear to me—and this is something that could and should be empirically investigated—that the actual, contemporary “folk” have any real expectation that there would be a correlation between their “honest” mechanic’s professional behavior and the degree of sexual fidelity she observes in her private life. When we talk about “cheating” spouses, we are talking about something quite different than “cheating” one’s customers; the fact that the same word is used in both cases is conceptually interesting and requires explanation,32 but this in no way proves that the two domains have equivalent moral valence or structure in folk psychology. Critiques of the predictive usefulness of such labels as “honest” or “courageous” might then be seen not as a fatal blow to our fundamental concept of character but rather as an indication of how loosely we use virtue terms in contemporary discourse—a looseness that is permissible because our implicit, shared background knowledge allows us to call up the appropriate social and ethical frameworks required for comprehending any particular use of a term.

II. Early Confucianism Virtue Ethics and the “High Bar” Argument

There is another important feature of traditional virtue terms such as ‘honest’ or ‘courageous’, however, that may present more of a hurdle for virtue ethics. Even if we grant that traditional virtue terms have a relatively narrow scope of application, they nonetheless seem to demand more than even the most robust current conception of personality traits can provide. In characterizing a correlation coefficient on the magnitude of 0.3 as “negligible,” situationists such as Doris are not denying it any predictive power but rather are contrasting it with the closer to 1.0 correlation that traditional notions of the virtues seem to require.33 For instance, Ted Williams’s formidable batting prowess was not diminished by the fact that he occasionally failed to make a hit or even struck out. On the other hand, we would be hard-pressed to characterize as “faithful” a spouse who manages to resist extramarital sexual temptations only much of the time or to laud as “courageous” a warrior who drops his weapons and flees from the enemy only somewhat less frequently than his peers. The core claim of the situationist critique—what we might call the “high bar” argument—is that virtue ethics demands a correlation between virtue possession and actual behavior of close to 1.0 and that anything short of that is a fatal problem.

One way around this problem is to argue that the early Confucians had a much weaker notion of the virtues than, say, Aristotle, and that the “high bar” argument therefore simply does not apply to them. Unfortunately, this does not seem to be the case, at least when it comes to moral exemplars. The Analects, for instance, portrays the perfected Confucian gentleman as impervious to the influence of reputation or material temptations. This internal moral autonomy is perhaps best expressed in Confucius’s famous declaration that he could go and live among the Eastern barbarians and not only maintain his moral perfection but in fact transform his social environment through his moral influence (Analects 9.14).34 It would thus seem that the early Confucian notion of virtue—like the Aristotelian and modern folk notion—is of a more or less 100-percent-reliable, environmentally impermeable character trait, which would mean that it faces the same “high bar” challenge posed by the situationist critique.

I wish to argue that the early Confucian model of virtue education gets them over this hurdle—but not entirely in the manner that the early Confucians themselves envisioned.35 To begin with, their emphasis on intensive, life-long, highly regimented training gives them a higher jump, as it were: the virtues that they ask the gentleman to rely upon are not untutored natural gifts but rather intensively cultivated dispositions, which can be expected to be much more reliable than the traits typically studied by social psychologists. In addition, the manner in which they continuously bolster these cultivated traits with a host of situational buffers—ranging from strict social regulations to careful modulation of one’s physical and interpersonal environment—effectively lowers the bar several notches. I will address both of these points in turn.

A. Moral Training and Dispositional “Extension”

To begin with, although there are occasional suggestions that some extraordinary individuals may come into the world with already well-developed and fully robust character traits,36 the dominant position in early Confucianism is that whatever positive traits we may possess “naturally” at the beginning of the process of self-cultivation are relatively weak and require long-term, intensive training in order to become genuinely reliable—that is, in order to become true virtues. Relating this to some of the empirical literature reviewed above, we can observe that, even if one would want to insist that the sort of local “honesties” that characterize the children in Hartshorne and May’s study are not correlated to a degree that would warrant the global label “honest” as a predictive term, a perfectly reasonable response would be that perhaps this is because they are children and have not yet learned to integrate these local traits into a broad and reliable disposition. Looked at this way, local traits remain ethically fragmented only to the extent that they remain untutored. Most, if not all, traditional virtue ethics envision some sort of cultural training as a necessary prerequisite to the acquisition of virtue, and one way to understand this training is that it involves the extension and integration of ethically related, but naturally fragmented, local traits.

Several responses to Doris’s critique have made this point, but one of the most intriguing is that of Nancy Snow because it draws upon current research in social psychology to demonstrate the plausibility of local trait extension and modification.37 For instance, Snow takes the example of a negative local trait, such as an undesirable racial stereotype—a “deep-seated psychology construct … whose activation often occurs automatically and outside of the agent’s conscious awareness”—and discusses research suggesting that a process of self-regulation can lead to some degree of conscious control over the otherwise automatic activation of social stereotypes.38 The same is presumably true in the case of positive behavioral tendencies. Snow remarks that this body of research suggests that “it is possible, with effort, to inhibit and control negative traits and cultivate and extend desired ones,” which means that “though our virtues might start out by being local, they need not remain so.”39

Interestingly, Snow also speculates about how one would go about extending a positive behavioral tendency limited to a very small domain of objective triggering conditions—say, compassion for small, cuddly animals—to a broader domain of sentient beings including, for example, one’s friends, relatives, and even unrelated strangers. This process of modifying or extending, under the guidance of an expert, various naturally given, but overly local, traits is in fact precisely the central strategy of early Confucian moral education. The process of compassion extension that Snow envisions recalls a famous dialogue in the early Confucian text, The Mencius (fourth century BCE), which describes precisely this sort of identification of a desirable but overly local trait and a strategy for extending this trait through reflection and imaginative work. Because this passage, Mencius 1:A:7, is such a masterwork of psychological insight and because it displays so paradigmatically the manner in which virtue ethical training might proceed, I would like to describe it in some detail, informed by the considerable body of scholarship that has formed around it, especially regarding the concept of “extension” (tui ; literally, “pushing”) that it introduces.

The passage opens with Mencius in dialogue with a notoriously selfish and brutal king, King Xuan of Qi, who oppresses his people and shows no apparent concern for traditional Confucian morality. When Mencius suggests to the king that he can change his ways and become a true, compassionate Confucian ruler—one who protects and nourishes his people rather than oppressing them—the king is dubious: “What makes you think that a person like myself could be capable of this?” Mencius replies by relaying a story that he heard from one of the king’s retainers:

The King was sitting up upon his elevated throne in the Great Hall when an ox was led past him. The King saw it and asked, “Where is that ox being taken?” The reply was, “It is being taken to be ritually slaughtered so that its blood can be used to consecrate a newly-forged bell.” The King said, “Let it go! I cannot bear its look of terror, like that of an innocent man being led to the execution ground.” “Should we then abandon the consecration ritual?” “How could we abandon the ritual? Substitute a sheep in its place.”

“Did this really happen?” Mencius slyly inquires. The king admits that it did, and Mencius then proceeds to lead him through the process of identifying the emotion that he felt—one that “is sufficient in and of itself to make one a true King”—and to distinguish it from other possible motivations. He first mentions that some of the king’s subjects, noting the substitution of the sheep for the ox, speculate that the king was simply trying to economize on ritual expenditures, sheep being considerably cheaper than oxen. The king indignantly denies this—Qi is a small state, but he can certainly afford an ox—and repeats that he was motivated solely by the look of terror on the ox’s face.

“The King should not be surprised that the common people took him to be cheap,” Mencius replied. “You exchanged a small animal for a large one, what were they to make of it? If the King were truly pained by the expression like that of an innocent man headed to the execution ground, then why spare the ox and sacrifice the sheep?”

The King smiled uncomfortably, saying, “What, indeed, was my feeling at that moment? I certainly was not worried about the expense, and yet I did put the sheep in the ox’s place. It is no wonder that the people think me cheap.”

Mencius replied, “There is no harm in this—in fact, it is precisely the feeling that you had that is the method of benevolence. You saw the ox, but had not yet seen the sheep. The gentleman’s attitude toward animals is thus: having seen them alive, he cannot bear to see them killed; having heard their cries, he cannot bear to eat their flesh. This is why the gentleman keeps his distance from the kitchen.”

This injunction for the gentleman to keep his distance from the kitchen has been understood as an expression of monumental hypocrisy,40 but the basic sentiment is quite understandable, and it seems to be as follows. Human beings universally react with compassion to the sight of a suffering animal. In a society where vegetarianism is not even a notional possibility, this means that those who wish to preserve this feeling must avoid exposure to the slaughter and processing of animals—a clear instance of situational management. Presumably those charged by society with performing these necessary functions, such as butchers and tanners, will learn to suppress this sort of compassion, or will simply become desensitized to animal suffering, which is why they were accorded a lowly status in Confucian society. In any case, Mencius holds up as the “method of benevolence” this isolated expression of compassion for an animal that the king had actually seen and heard—a very local and certainly extremely ethically fragmented trait, considering that the king had no compunction about having some anonymous sheep slaughtered in its place, and of course behaves in a notably nonbenevolent manner toward his subjects.41

Mencius’s task as a moral educator is to get the king to see the ethical relevance of this narrow and isolated feeling of compassion for the ox: to understand its nature and to begin to see how it might be relevant to broader ethical life. This next step begins after the king admits that Mencius has seen into his own heart and “taken its measure” better than the king himself had been able to do. He had indeed felt a twinge of compassion, motivated by the sight of suffering. However, in what way, he asks Mencius, is this momentary feeling relevant to the task of becoming a truly benevolent king?

Mencius replied, “If someone came to you and said ‘My strength is sufficient to lift five hundred pounds, but not sufficient to lift a single feather,’ or ‘My eyesight is sharp enough to distinguish the tip of an autumn hair, but unable to perceive a cartful of wood,’ would you accept such words?

The King said, “No.”

“Now, your compassion is abundant enough to reach even a lowly beast, and yet your bounty does not even extend to the common people of the realm. How is this any different?”

In other words, the king is apparently capable of a relatively difficult task (having compassion for a lowly sacrificial animal, in Mencius’s mind not an obvious object of compassion) and yet has shown himself to be incapable of what should be a much easier task, showing compassion and kindness toward the common people—generally conceived of in early Chinese political discourse as the metaphorical children of the ruler and thus as natural objects of compassion.

There is considerable debate in the literature about what Mencius thinks he has accomplished by pointing out this apparent inconsistency to the king.42 However, it seems most likely that his goal is, as Phillip Ivanhoe has argued, to set up an “analogical resonance” involving “emotional resonance not cognitive similarity” between the “local” feeling of compassion for a suffering animal, which the king acknowledges having experienced, and another “local” feeling, that of compassion for his suffering people, which the king has for some reason yet to experience.43 Arguably, it is precisely this sort of cross-situational “resonance” that is captured in global terms such as ‘compassion’ or ‘honesty’.

Having gotten the king to see, by examining his own emotional reactions, that he has the “sprout” of true benevolence (ren ) within him and that there is some analogical relation between suffering animals and suffering commoners, Mencius’s task is then to turn the king’s quite narrow—and, to Mencius’s mind, at least, ethically irrelevant—disposition to feel empathy for an animal into a broader disposition to feel empathy for suffering humans.44 This is to be accomplished through gradual strengthening or “extension” (tui ) of the local disposition, a process of sympathetic projection and emotional training guided by metaphor and analogy: “Treat the aged of your own family in a manner that respects their seniority, and then cause this treatment to reach the aged of other families. Treat the young ones of your family in a manner appropriate to their youth, and then cause this treatment to reach the young of other families. Once you are able to do this, you will have the world in the palm of your hand.” The passage in question, 1:A:7, is actually quite thin when it comes to spelling out precisely how “extension” is to be accomplished—Mencius immediately moves on to another argumentative tack with King Xuan—but the remainder of the text, and the early Confucian cultural context, help us to flesh this out. It clearly involves imaginative work and emotional analysis, directed at previously experienced emotions, as in the case of King Xuan, or imagined emotions, as when Mencius famously invites all of us to consider the feeling of “alarm and distress” we would experience, at least for a moment, upon seeing a baby about to experience immanent physical harm (2:A:6). This emotional analysis is to occur under expert guidance, such as that of Mencius himself, which serves to highlight for the novice as yet unperceived analogical resonances and nuances of the emotion, as well as offer suggestions about how “fragmented” local responses might be integrated into more coherent, broad and ethically useful dispositions.

Two points need to be made in conclusion to this section. To begin with, it may certainly be the case that, even with extensive training, nothing anywhere near a 1.0 correlation between character traits and behavior is attainable. If it turns out that even intensely cultivated character traits cannot break through the 0.3 correlation coefficient barrier, this may very well end up being a fatal problem for virtue ethics of any stripe. This, however, is emphatically not what is demonstrated by the existing situationist literature, which looks only at completely untutored character traits. The degree of reliability in character traits attainable through deliberate training is a mostly unexplored and promising topic for future empirical inquiry, one that must be pursued by anyone interesting in establishing the empirical plausibility of the virtue ethical model.

Second, it is important to realize that the entire process of Confucian character training was portrayed as occurring in a context of intense environmental manipulation, accomplished through immersion in and submission to traditional cultural forms. It is, moreover, not at all clear that—whatever the rhetoric about perfected sages and illustrious gentlemen—it was envisioned that even a fully trained individual was expected to function reliably outside of this buffer of environmental control. We might thus conclude that the function of such environmental buffering was to take up any slack between the reliability of even fully trained virtues and the close to 1.0 reliability expected of a true gentleman—to lower the bar just enough notches to allow the trained jump of the Confucian gentleman to clear it. This leads us to our next point, the role of the “situation” in Confucian self-cultivation.

B. Confucian Self-Cultivation as Manipulation of the Situation

The most important of the traditional cultural forms advocated by the early Confucians is ritual (li ). In the Warring States Confucian context, li referred to a set of cultural scripts governing a broad range of behaviors, from ancestral sacrifice and diplomatic ceremonies to details of one’s personal comportment, such as the manner in which one dresses, takes one’s meal, enters a room, or takes one’s seat. Confucius himself was the first to argue that, by submitting to and internalizing these ritual forms, an aspiring gentleman would be able to restrain improper inborn tendencies, acquire the means to “take his place” (li ) among other adults in society, and thereby acquire full virtue and win the favor of heaven. As Ivanhoe notes, although Mencius, with his faith in the potential goodness of human nature, appeared to have viewed ritual forms more as guidelines or supports to help direct incipient moral tendencies, both Confucius and his late Warring States follower Xunzi conceptualized ritual as a tool for reshaping otherwise crude innate dispositions.45 As Xunzi puts it, “[Sorrow and joy] are emotions that are firmly rooted in the nature that people have at birth. If these emotions can be trimmed or stretched, broadened or narrowed, augmented or decreased, categorized and thereby put to their full use, embellished and beautified, so that the root and branch, beginning and end match together seamlessly, and are worthy to serve as a model for ten thousand generations—this, then, is ritual.”46

Training in ritual was to proceed alongside other forms of behavioral modification, including the appreciation and performance of music and dance, archery, riding, and calligraphy. The product of this immersion in cultural forms was to be the perfectly culturally refined (wen ) gentleman, who in every aspect of his physical deportment would reflect the aesthetic-moral ideals of the past Golden Age.47 This sort of fully cultivated gentleman would then, in turn, exert an influence on others by his mere physical presence, an idea expressed most strongly by Confucius in the passage mentioned above, Analects 9.14. Frustrated by his failure to be recognized by the rulers of his day, Confucius expresses a desire to go live among the barbarian tribes of the Eastern seaboard. When someone asks him how he—such an eminently cultured man—would be able to endure the uncouthness of barbarian life, Confucius replies, “If a true gentleman were to dwell among them, what uncouthness would there be?” The early Confucians had a specific term of art for this type of personal, often unconscious influence of the gentleman upon others: de , or “charismatic Virtue.” Metaphorically compared to the power of the Pole Star to attract the other stars in the sky and keep them in their proper orbits (2.1), or the ability of the wind to bend the grass below it (12.19), de might be seen as a form of interpersonally exerted situational control par excellence.48

Finally, an account of Confucian sensitivity to the situational effects cannot neglect their attention to one’s conceptual environment. For instance, the literature on social stereotype priming highlights the manner in which specific social role terms can automatically and unconsciously have an impact on behavior: subjects primed by the social role “professor” performed significantly better on a general knowledge task than nonprimed subjects, while subjects primed with the “soccer hooligan” role performed more poorly.49 Similarly, alternate verbal framings of situations can significantly alter subjects’ behaviors: higher levels of generosity in economic games, for instance, can be obtained simply by framing the exercise as a “Community Game” as opposed to a “Wall Street Game.”50

One way to view the Confucian practice of “learning” (xue )—a process of intensive study and memorization of textual classics, which describes in great detail the exemplary thoughts and conduct of the ancient sages—was that it served as a form of ever-present conceptual priming. A fully learned Confucian would have always at the forefront of his mind the exemplary behavior and words of ancient paragons, and it is not too great a stretch to see this as designed to increase the probability that he would act in accordance with these models. In Analects 2.2, Confucius remarks of the Book of Odes, a collection of poetry that records the sentiments of the ancients and the deeds of the ancient sage-kings, that “its poems number several hundred, but can be judged with a single phrase: ‘They will not lead you astray.’” Moreover, the early Confucians were also very much concerned with the regulation of language use: while there is some scholarly disagreement concerning the exact nature of the Confucian practice of “rectifying names” (zhengming ),51 it was clearly intended to provide normatively desirable frames for behavior. In Analects 13.3, Confucius is asked by a disciple what his first priority would be if he were given control of a state, and Confucius replies “I would rectify names.” The disciple is surprised by this answer, and Confucius elaborates:

If names are not rectified, speech will not accord with reality; when speech does not accord with reality, things will not be successfully accomplished. When things are not successfully accomplished, ritual practice and music will fail to flourish; when ritual and music fail to flourish, punishments and penalties will miss the mark. And when punishments and penalties miss the mark, the common people will be at a loss as to what to do with themselves. This is why the gentleman only applies names that can be properly spoken, and assures that what he says can be properly put into action.

In another related passage, 12.11, a certain Duke Jing of Qi asks Confucius about how best to govern a state, and Confucius replies simply, “Let the lord be a true lord, the ministers true ministers, the fathers true fathers, and the sons true sons.” This advice appears to be two-sided. Words should be applied only to their proper objects: someone who is not a true son, for instance, should not be called “son”—a message that many commentators feel is specifically directed to the Duke of Qi, who had passed over his elder son for succession and caused discord among his children. At the same time, words such as ‘father’ or ‘ruler’ bring with them certain positive social norms, and their very invocation should inspire a certain type of model emulation. The power of such words is, in turn, linked to the broader conceptual training provided by textual study: it is the paradigmatic behavior of the ancients as recorded in the classics that set the standards for what a “son” or “father” is and that provide the idealized cultural models that are to be activated by proper language use.

The early Confucians clearly believed in the possibility of developing robust, global character traits that could endow an individual with a degree of independence from situational forces. The perfected Confucian gentleman is thought to possess an expansive compassion that would reliably produce benevolent behavior with regard to his inferiors and people in his charge; a degree of moral rectitude and inner strength that would convey stoic invulnerability to external temptations and vicissitudes such as social reputation, wealth, or sickness; a degree of wisdom and ritual propriety that would allow him to stand apart from and judge the cultural practices of his contemporaries; and a forthright courage in the face of corruption or immorality that would allow him to speak out against social superiors and those in power. To put this another way, the early Confucians seemed to believe that fully cultivated virtues should produce a close to 1.0 correlation with behavior. To the extent that even intensive, extended training is incapable of producing something in the neighborhood of such a correlation—something that remains to be examined empirically—we might therefore be forced to conclude that the early Confucians were as deluded as the Aristotelians about the power of character.

However, whatever their explicit claims, the specific social practices and institutions prescribed by Confucian thinkers suggest that—consciously or not—early Confucianism saw the need to bolster even fully trained dispositions with situational support. It is important to note that immersion in carefully designed cultural forms was not seen as coming to an end when the individual “finished” his training and attained the status of gentleman—indeed, the Confucians arguably saw the process of training as never coming to an end.52 In their rhetoric concerning the incorruptible, lofty gentleman or sage, the early Confucians may have been as vulnerable as Aristotle to the “fundamental attribution error”;53 indeed, social psychology suggests that such attribution is a basic, deeply engrained human cognitive tendency. What I hope to suggest here is that, whatever their explicit claims or assumptions, Confucian practices as envisioned by early thinkers did not rely entirely upon fully internalized, 100-percent-reliable, and environmentally impervious character traits. At some level, the Confucians were exquisitely aware of the power of the situation, and their methods of self-cultivation—their techniques for producing relatively independent character traits—thus focused heavily on the manipulation of all aspects of the learner’s physical, linguistic, and social environment.

In his response to Doris’s work, Eric Hutton has also focused on this aspect of Confucianism,54 and he concludes by citing a passage from the Xunzi that captures quite well this attention to—even anxiety about—situational factors, conceptualized metaphorically as an external surface that one “rubs up against” (mi ):

Even if you possess a fine nature and character, as well as a discerning and well-informed mind, you still need to find a worthy teacher and devote yourself to his service, as well as a group of noble friends to befriend. If you obtain a worthy teacher to serve, then what you hear will be the ways of [the sages] Yao, Shun, Yu, and Tang. If you obtain a noble friend to befriend, then what you see will be conduct that is dutiful, trustworthy, respectful, and deferential. In this way, you in your own person will make daily progress toward benevolence and rightness without even being aware of it. This is because it is what you are rubbing up against that is making it so. Now, if you live among those who are not good, then what you hear will be trickery, deception, falseness, and hypocrisy, and what you see will be conduct that is foul, arrogant, perverse, deviant, and greedy. In this way, you in your own person will come to suffer punishment and execution, without you even being aware of what is happening. This is because it is what you are rubbing up against that is making it so. A saying goes, ‘‘If you do not know your son, observe his friends. If you do not know your lord, observe his companions.’’ It is simply a matter of what you rub up against! It is simply a matter of what you rub up against!55

I lack the expertise to judge whether or not the Aristotelian version of virtue ethics was based upon an empirically unjustifiable confidence in the robustness of character traits, independent of situational forces. What I hope to have suggested here is that the early Confucians, at least, combined a quite reasonable faith in trained, culturally modified, somewhat stable and independent character traits with a profound sensitivity to the power of “the situation.”56 Their example not only provides us with a picture of what an empirically defensible virtue ethic might look like but also with a wealth of insights into how situational effects can be drawn upon and utilized in the process of moral education.

III. Conclusion

Over the past decade or so, one of the areas of research upon which Walter Mischel and his colleagues have been focusing attention is the interaction of the “hot,” automatic, emotional “know how” systems and the “cool,” conscious, “knowing that” systems, demonstrating that “willpower,” that is, the ability to resist adverse situational influences, can be enhanced through various forms of conceptual priming and training.57 Mischel concludes: “The French philosopher Descartes a few hundred years ago famously proclaimed ‘cogito ergo sum,’ I think, therefore I am, opening the way to what a few hundred years later became modern psychology. With what is now becoming known about personality, we can change his assertion to say: ‘I think, therefore I can change what I am. Because by changing how I think, I can change what I feel, do, and become.’”58

This observation might serve as a nice expression of the essence of Confucian virtue ethics. To be sure, early Confucian thinkers clearly perceived the ethical importance of rationality and will power. For instance, they emphasized the importance of willpower when it comes to choosing one’s general life priorities, as well as the role of the “heart-mind” (xin ) as the proper “ruler” of the self, charged with moral decision making and the enforcement of those decisions on the rest of the self.59 However, the education system that they established reveals a recognition—conscious or not—of the limits of what contemporary cognitive scientists would call on-line cognitive control. Although the decision to devote oneself to the Confucian Way was often portrayed as a momentary act of will, the actual process of following this Way involved the development of stable, gradually broadened dispositions, proper perceptual habits, and culturally constructed moods—all to be trained and maintained in a carefully controlled physical, conceptual, and social environment.

One way to look at Confucian virtue ethics is as a kind of “time-delayed” cognitive control that functioned by embedding higher-level desires and goals in lower-level emotional and sensory-motor systems. The strength of this approach—and its great advantage over models of ethics such as deontology and utilitarianism that have dominated recent philosophical ethics in the West—is that it avoids the sharp limitations of on-line cognitive control that recent work in social psychology and cognitive science have made clear.60 There is, moreover, a nascent but growing body of empirical evidence that this kind of dispositional education actually works. In addition to Mischel et al.’s work on the “agentic, proactive self,” one might also consider the review by Jonathan Cohen of the relationship between emotions and prefrontal-cortex-mediated cognitive control in human behavior.61 Cohen notes, for instance, that “the specialized training given to doctors and soldiers involves the cultivation of mechanisms for averting or overcoming strong emotional responses that may interfere with their professional functions. These mechanisms may not rely directly on the prefrontal cortex; instead, they may involve the training of other lower-level mechanisms specific to the particular circumstances involved. Importantly, however, the social structures that devised and support the training procedures almost certainly did rely on the prefrontal cortex.”62

The sort of training undergone by doctors and soldiers—designed to instill specific forms of courage and calmness under pressure as stable character traits—employs a set of strategies designed by the prefrontal cortex in order to overcome its own limitations in on-line, “hot” cognition situations: in other words, time-delayed cognitive control. Moreover, at least with regard to a subset of professionally relevant situations, this medical and military training seems to do its job. With regard to more conceptual virtues, Nancy Snow, as noted above, argues that recent work on stereotype modification gives us reason to believe that even quite automatic and unconscious conceptual habits can, through gradual training, be brought into conscious attention and thereby modified in socially desirable directions. Similarly, Patrick Hill and Daniel Lapsley survey a set of contemporary approaches aimed at developing moral personality in ways not dissimilar to the early Confucian strategy.63 This remains a very promising direction for future research on the empirical viability of virtue ethics.

One of the great contributions of the situationist research agenda has been to make clear the immense power of subtle, and often unnoticed, aspects of situations and environments to shape human behavior. Like many scientific insights, it is somewhat counterintuitive. It certainly presents a challenge to the models of ethics that have recently dominated modern Western philosophy, deontology and utilitarianism, which rely on rational agents making explicit decisions grounded in a transparent chain of reasoning under fully conscious control. It also exposed deep problems with early, naive models of personality traits as rigid, situation insensitive, and invariant, forcing personality psychologists to develop much more nuanced models of personality traits that take into account situational cues and developmental change. As championed in philosophical circles by scholars such as Harman and Doris, it has also forced defenders of virtue ethics—which, since at least the time of Mencius (fourth century BCE), have seen psychological realism as their unique strength—to confront an important body of empirical data that calls into questions basic folk assumptions about the nature of the self.64

However, philosophers need to recognize that reports of the death of character have been greatly exaggerated. Pace observers such as Harman and Doris, recognizing the importance of the situation could actually strengthen the empirical plausibility and appeal of virtue ethical models of moral reasoning and moral education, at least in something resembling their traditional Confucian forms. Unlike most modern Western ethicists, the early Chinese Confucians paid a great deal of attention to the power of the embodied situation—social role, dress, ambient color, and sound—to effect human dispositions and behavior. They designed a sophisticated set of technologies to structure such environments in a way that would be conducive to morality, as well as a body of self-cultivation techniques to integrate potentially “fragmented,” untutored dispositions into more robust ones that would—automatically and largely unconsciously—produce reliable, ethically desirable behaviors across a broad range of situations. The combination of the higher jump given by training and the lower bar provided by situational supports might very well be enough to allow the early Confucians to clear the main situationist hurdle—something close to a 1.0 correlation between virtue and behavior—albeit at the expense of bracketing some of the Confucians’ own explicit claims.

Such bracketing will always be necessary when adapting ancient modes of thought to the modern world. No contemporary advocate of Confucianism would endorse their undeniable misogyny or their particular historical or metaphysical views. Adopting anything even remotely resembling the Zhou rituals endorsed by Confucius or Xunzi would of course be absurd, and the hierarchical and patriarchal society envisioned by the early Confucians would not seem to sit well with any sort of modern, liberal democracy. Moreover, we have to admit that these may be fatal problems: it is an open question whether or not one could achieve an effective enough combination of virtue training and situational control within the context of modern, secular democracy.

I will therefore keep my conclusion fairly modest: the early Confucian form of virtue ethics seems as if it could survive even the strongest and most plausible form of the situationist critique, which means that proclamations of the death of virtue ethics are rather premature. We can frame this a bit more strongly by observing that our current understanding of human cognition suggests that “high reason,” cognitive-control-based models of ethics, such as deontology or utilitarianism, appear profoundly psychologically unrealistic, which suggests that some form of virtue ethics is our best hope—if, that is, empirical plausibility is deemed a desirable feature when it comes to ethics. At the very least, we can say that, as we learn more about how the human mind works, ethical traditions such as early Confucianism help us to fill in enormous blind spots—the importance of the body, emotions, cultural training, the unconscious, and the social environment—that have hindered modern Western ethical thinking for the past several hundred years. They are thus of more than merely antiquarian interest, and they can potentially help us in developing an ethic that will meet the challenges of the twenty-first century.65


1David Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals (1777), 3rd ed. (Oxford: Clarendon, 1976), 174–75.

2This earlier attitude is well exemplified in Kant’s indignant rejection of the “slack, or indeed ignoble, attitude which seeks for the moral principles among empirical motives or laws,” as well as his claim that the purity of moral philosophy depends upon it being “the authoress of her own laws” rather than “the mouthpiece of laws whispered to her by some implanted sense or by who knows what tutelary nature” (Groundwork of the Metaphysics of Morals [1785], trans. H. J. Paton [New York: Harper Torchbooks, 1964], 93).

3Owen Flanagan, Varieties of Moral Personality: Ethics and Psychological Realism (Cambridge, MA: Harvard University Press, 1991); Mark Johnson, Moral Imagination: Implications of Cognitive Science for Ethics (Chicago: University of Chicago Press, 1993). This view is most succinctly and famously expressed in Flanagan’s “Principle of Minimal Psychological Realism”: “Make sure when constructing a moral theory or projecting a moral ideal that the character, decision processing, and behavior prescribed are possible … for creatures like us” (Flanagan, Varieties, 32).

4Mark Johnson, The Meaning of the Body: Aesthetics of Human Understanding (Chicago: University of Chicago Press, 2008); Jesse Prinz, “Passionate Thoughts: The Embodiment of Moral Concepts,” in Grounding Cognition: The Role of Perception and Action in Memory, Language and Thinking, ed. Diane Pecher and Rolf A. Zwaan (Cambridge: Cambridge University Press, 2005), 93–114, and The Emotional Construction of Morals (New York: Oxford University Press, 2007). A recent, representative collection of the “experimental philosophy” work can be found in Joshua Knobe and Shaun Nichols, Experimental Philosophy (New York: Oxford University Press, 2008).

5See, e.g., Antonio Damasio, Descartes’ Error: Emotion, Reason, and the Human Brain (New York: Putnam’s, 1994); and Jonathan Haidt, “The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgment,” Psychological Review 108 (2001): 814–34. A helpful review of this literature can be found in Jonathan Haidt, “The New Synthesis in Moral Psychology,” Science 316 (2007): 998–1002.

6Paul Churchland, “Toward a Cognitive Neurobiology of the Moral Virtues,” Topoi 17 (1998): 83–96; William Casebeer, Natural Ethical Facts: Evolution, Connectionism and Moral Cognition (Cambridge, MA: MIT Press, 2003); Edward Slingerland, “‘Of What Use Are the Odes?’ Cognitive Science, Virtue Ethics, and Early Confucian Ethics,” Philosophy East and West 61 (2011): 80–109.

7Of course, more recent proponents of both deontology and utilitarianism have acknowledged an important role for intuitive, implicit cognitive processes, sometimes confining explicit algorithmic reasoning to a critical meta level, which may only be invoked when conflicts arise or justifications need to be provided. While certainly more psychologically realistic, this still begs the question of how the behavioral desiderata arrived at through deontological or utilitarian reasoning are to be built into automatic everyday cognition, which is an issue that I would argue virtue ethics uniquely and explicitly addresses.

8Gilbert Harman, “Moral Philosophy Meets Social Psychology: Virtue Ethics and the Fundamental Attribution Error,” Proceedings of the Aristotelian Society 99 (1999): 315–31; John Doris, Lack of Character: Personality and Moral Behavior (New York: Cambridge University Press, 2002).

9It is worth noting that Flanagan’s original cautions were relatively mild and that he has since become quite critical of what he refers to as “a small band of mischievous hyperbolists, really just two” (i.e., Harman and Doris) who “have had their fun for too long making ontological mischief” (55) among psychologically and statistically underinformed philosophers. See Owen Flanagan, “Moral Science? Still Metaphysical after All These Years,” in Personality, Identity, and Character: Explorations in Moral Psychology, ed. Darcia Narvaez and Daniel Lapsley (Cambridge: Cambridge University Press, 2009), 54–65, for a critique of situationism that resonates with the arguments that I will be making below.

10See, e.g., Joel Kupperman, “Naturalness Revisited,” in Confucius and the Analects, ed. Bryan W. Van Norden (New York: Oxford University Press, 2002), 39–52; Rachana Kamtekar, “Situationism and Virtue Ethics on the Content of Our Character,” Ethics 114 (2004): 458–91; John Sabini and Maury Silver, “Lack of Character? Situationism Critiqued,” Ethics 115 (2005): 535–62; Diana Fleming, “The Character of Virtue: Answering the Situationist Challenge to Virtue Ethics,” Ratio 19 (2006): 24–42; Eric Hutton, “Character, Situationism, and Early Confucian Thought,” Philosophical Studies 127 (2006): 37–58; and Nancy Snow, Virtue as Social Intelligence: An Empirically Grounded Theory (New York: Routledge, 2010).

11For a defense of Confucianism as “virtue ethic,” see Stephen Wilson, “Conformity, Individuality, and the Nature of Virtue: A Classical Confucian Contribution to Contemporary Ethical Reflection,” Journal of Religious Ethics 23 (1995): 263–89; Philip J. Ivanhoe, Confucian Moral Self Cultivation, 2nd ed. (Cambridge, MA: Hackett, 2000); Edward Slingerland, “Virtue Ethics, the Analects, and the Problem of Commensurability,” Journal of Religious Ethics 29 (2001): 97–125; and Bryan Van Norden, Virtue Ethics and Consequentialism in Early Chinese Philosophy (New York: Cambridge University Press, 2007). I take early Confucian thought as my model of virtue ethics both because it is the only form of virtue ethics that I am qualified to responsibly discuss and because, as I will argue in Sec. II, it may be less vulnerable to the situationist critique than the Aristotelian form that has played a more prominent role in the revival of virtue ethics.

12John Darley and C. Daniel Batson, “From Jerusalem to Jericho: A Study of Situational and Dispositional Variables in Helping Behavior,” Journal of Personality and Social Psychology 27 (1973): 100–119; Alice Isen and Paula Levin, “The Effect of Feeling Good on Helping: Cookies and Kindness,” Journal of Personality and Social Psychology 21 (1972): 384–88.

13“Personality and Assessment at Age 40: Reflections on the Past Person-Situation Debate and Emerging Directions of Future Person-Situation Integration,” ed. M. Brent Donnelan, Richard Lucas, and William Fleeson, special issue, Journal of Research in Personality 43 (2009). The reader is referred to this issue for a much more detailed account of the state of the field than can be provided here. Although one might feel that the Journal of Research in Personality hosting a survey of the personality vs. situation debate is a bit like Pravda at the height of the Cold War devoting an issue to the relative merits of capitalism vs. communism, the editors of this special issue made a concerted effort to be evenhanded, and the full spectrum of opinions on the topic is represented. David C. Funder, “Persons, Behaviors and Situations: An Agenda for Personality Psychology in the Postwar Era,” in ibid., 120–26, at 120.

14See Brent Roberts, “Back to the Future: Personality and Assessment and Personality Development,” Journal of Research in Personality 43 (2009): 139–40, on the subject, what are personality traits, anyway?

15Walter Mischel, “From Personality and Assessment (1968) to Personality Science, 2009,” Journal of Research in Personality 43 (2009): 282–90, 284; Walter Mischel, Yuichi Shoda, and Rodol Denton, “Situation-Behavior Profiles as a Locus of Consistency in Personality,” Current Directions in Psychological Science 11 (2002): 50–54.

16Doris, Lack of Character, 64.

17Regarding ”Big-Five” and “not entirely unalterable or context insensitive,” see Jack Block and Jeanne H. Block, “Venturing a 30-Year Longitudinal Study,” American Psychologist 61 (2006): 315–27; and Dan P. McAdams and Jennifer L. Pals, “A New Big Five: Fundamental Principles for an Integrative Science of Personality,” American Psychologist 61 (2006): 204–17. Regarding “have a considerable genetic component,” see Auke Tellegen, David Lykken, Thomas Bouchard, Kimerly Wilcox, Nancy Segal, and Stephen Rich, “Personality Similarity in Twins Reared Apart and Together,” Journal of Personality and Social Psychology 54 (1988): 1031–39. Regarding “predict substantive life outcomes,” see Paul Costa and Robert McRae, “Personality in Adulthood: A Six-Year Longitudinal Study of Self-Reports and Spouse Ratings on the Neo Personality Inventory,” Journal of Personality and Social Psychology 54 (1988): 853–63; Robert Hogan, “In Defense of Personality Measurement: New Wine for Old Whiners,” Human Performance 18 (2005): 331–41; Daniel Ozer and Verónica Benet-Martinez, “Personality and the Prediction of Consequential Outcomes,” Annual Review of Psychology 57 (2006): 401–21; and Brent Roberts, Nathan Kuncel, Rebecca Shiner, Avshalom Caspi, and Lewis Goldberg, “The Power of Personality: The Comparative Validity of Personality Traits, Socioeconomic Status, and Cognitive Ability for Predicting Important Life Outcomes,” Perspectives on Psychological Science 2 (2007): 313–45.

18Doris, Lack of Character, 73–74.

19In another of the early situationist classics, Lee Ross and Richard Nisbett granted that the correlations might be as high as 0.4 (Lee Ross and Richard Nisbett, The Person and the Situation: Perspectives of Social Psychology [Philadelphia: Temple University Press, 1991]).

20Always sensitive to potential counterarguments (a character trait?), Doris does acknowledge that one person’s “weak” correlation is another person’s “suggestive” one (Lack of Character, 38), but he fails to further question this dismissal of 0.3 as insignificant.

21Gregory J. Meyer, Stephen E. Finn, Lorraine D. Eyde, Gary G. Kay, Kevin L. Moreland, Robert R. Dies, Elena J. Eisman, Tom W. Kubiszyn, and Geoffrey M. Reed, “Psychological Testing and Psychological Assessment: A Review of Evidence and Issues,” American Psychologist 56 (2001): 128–65.

22David Funder and Daniel Ozer, “Behavior as a Function of the Situation,” Journal of Personality and Social Psychology 44 (1983): 107–12; F. D. Richard, Charles Bond, and Juli Stokes-Zoota, “One Hundred Years of Social Psychology Quantitatively Described,” Review of General Psychology 7 (2003): 331–63. Richard et al.’s “One Hundred Years,” a massive meta-study of a century of social psychology literature encompassing 25,000 studies of 8 million subjects, concluded that social psychological effects averaged a “Pearson product-moment coefficient” (r), which measures the linear dependence between two variables, of 0.21, with a standard deviation of 0.15.

23Robert Rosenthal and Donald Rubin, “A Simple, General Purpose Display of Magnitude of Experimental Effect,” Journal of Educational Psychology 74 (1982): 166–69.

24From Robert Abelson, “A Variance Explanation Paradox: When a Little Is a Lot,” Psychological Bulletin 97 (1985): 129–33 (cited in Sabini and Silver, “Lack of Character?”).

25Harman, “Moral Philosophy”; Doris, Lack of Character, 64. As Nancy Snow observes, Doris’s strong antiglobalist view of local traits is even more extreme than the model currently embraced by Walter Mischel and his colleagues, the “cognitive-affecting processing system (CAPS),” which encompasses the individual’s subjective interpretation of situations as well as objective situational features and thus is much more amenable to modification or extension (Snow, Virtue as Social Intelligence).

26To be fair, for all his dramatic talk of “lack of character,” the argument of Doris, at least, is not based upon a blanket denial of global traits, or an absolute distinction between local and global, but merely the claim that empirically defensible traits are not global enough to do the work that virtue ethics requires of them. I will address this “high bar” argument—the claim that nothing short of a nearly 1.0 correlation between traits and behavior is adequate to get a virtue ethical system off the ground—in Sec. II.

27David C. Funder, “Persons, Behaviors and Situations: An Agenda for Personality Psychology in the Postwar Era,” Journal of Research in Personality 43 (2009): 120–26, at 122.

28Richard E. Lucas and M. Brent Donnellan, “If the Person-Situation Debate Is Really Over, Why Does It Still Generate So Much Negative Affect?” Journal of Research in Personality 43 (2009): 146–49.

29Hugh Hartshorne and Mark May, Studies in the Nature of Character, vol. 1, Studies in Deceit (New York: Macmillan, 1928).

30Doris, Lack of Character, 115.

31It was pointed out by an anonymous reader that xin is a crucial virtue in interactions with one’s you , a term typically translated as “friends,” which might seem to undermine my claim that xin is restricted to one’s professional behavior. It is important to realize that, in the Confucian context, you refers not to random acquaintances or childhood buddies, but rather to a subset of the professional colleagues of a given “gentleman”—the cultivated scholar-politician who is the target of Confucian education—who are more or less of the same rank/seniority and with whom the gentleman, to borrow a concept from Aristotle, shares a vision of the “Good.” You marks out those colleagues whom one finds personally amenable and who also share one’s moral aspirations.

32My first-run guess would be that it involves a perceived metaphorical resonance between the financial harm caused by professional compact with a client and the emotional harm caused to one’s spouse by violating a social compact.

33As mentioned above, this can be seen as the core of Doris’s argument against virtue ethics, and I thank both Doris (personal communication, 2009) and one of the anonymous readers for this journal for clarifying this point and focusing my attention upon it.

34See John Knoblock’s Xunzi, chap. 2 (“Cultivating the Self”), for a more extended expression of this sentiment (Xunzi: A Translation and Study of the Complete Works, vol. 1 [Stanford, CA: Stanford University Press, 1988], 154–55).

35This article is intended more as a response to the situationist critique than as a full introduction to early Confucian ethics, so I will do no more here than briefly sketch out a few relevant aspects of the early Confucian picture of self-cultivation. For more in-depth discussions, the reader is referred to Ivanhoe, Confucian Moral Self Cultivation; Edward Slingerland, Effortless Action: Wu-wei as Conceptual Metaphor and Spiritual Ideal in Early China (New York: Oxford University Press, 2003); and Van Norden, Virtue Ethics, as well as the Confucian response to situationism described by Eric Hutton, “Character, Situationism,” 37–58.

36Confucius declares in Analects 16.9 that “those who are born understanding it are the best,” and the portrayal of the disciple Yan Hui in that text suggests that he was such a person (see esp. 2.9, 5.9, 6.7, and 11.4); cf. the portrayal of the sage-king Shun in Mencius 7:A:16.

37See, e.g., Kamtekar’s observation (“Situationism and Virtue Ethics”) that narrow dispositions can be extended cross-situationally through analogical reasoning. Snow, Virtue as Social Intelligence.

38Snow, Virtue as Social Intelligence, 34. An example of the research Snow discusses is Patricia Devine and Margo Monteith, “Automaticity and Control in Stereotyping,” in Dual-Process Models and Themes in Social and Cognitive Psychology, ed. Shelly Chaiken and Yaacov Trope (New York: Guilford, 1999), 339–60.

39Snow, Virtue as Social Intelligence, quotations, respectively, at 38 and 37.

40A roughly contemporaneous text that is very critical of Confucian morality, the Zhuangzi, contains a passage where a butcher cutting up an ox in front of his ruler, presumably for just such a sacrifice, is presented as a perfected sage and his butchering as a model for proper living. It is possible that this story is partially intended to mock the fastidious hypocrisy of Mencius 1:A:7.

41For Mencius, “benevolence” (ren ), although only one of four cardinal virtues, is the most important, and it often stands in metonymically for the other three, which accounts for his eagerness to demonstrate its existence in even such a figure as King Xuan.

42See Ivanhoe, Confucian Moral Self Cultivation, for an overview of this literature.

43Phillip J. Ivanhoe, “Confucian Self-Cultivation and Mengzi’s Notion of Extension,” in Essays in the Moral Philosophy of Mengzi, ed. Xiusheng Liu and Phillip J. Ivanhoe (Indianapolis, IN: Hackett, 2002), 221–41, 226; cf. David Wong, “Reasons and Analogical Reasoning,” in Liu and Ivanhoe, Mengzi, 187–220; and Van Norden, Virtue Ethics, 234–38.

44This account of Mencian self-cultivation is derived from Ivanhoe, Confucian Moral Self Cultivation, to which the reader is referred for a more complete account.

45See ibid. for an outline of the various strategies found in early Confucian self-cultivation.

46Xunzi, Discourse on Ritual, chap. 19. In this passage, “sorrow and joy” are standing in metonymically for all innate human emotions.

47For more on the moral significance of “style” for the early Confucians, see Nicholas Gier, “The Dancing Ru: A Confucian Aesthetics of Virtue,” Philosophy East and West 51 (2001): 280–305; Kupperman, “Naturalness Revisited”; Amy Olberding, “The Educative Function of Personal Style in the Analects,” Philosophy East and West 57 (2007): 357–74.

48On this topic, see also the discussion by Hagop Sarkissian on the apparent importance of “agent-introduced situational effects” in Confucius’s Analects (Hagop Sarkissian, “After Confucius: Psychology and Moral Power” [PhD diss., Department of Philosophy, Duke University, 2008]). For contemporary discussions on the importance of situational factors for virtue ethics, see Maria Merritt, “Virtue Ethics and Situationist Personality Psychology,” Ethical Theory and Moral Practice 3 (2000): 365–83, on virtues as local traits sustained by social relationships and settings; or Steven Samuels and William Casebeer, “A Social Psychological View of Morality: Why Knowledge of Situational Influences on Behavior Can Improve Character Development Practices,” Journal of Moral Education 34 (2005): 73–87, on the role of proper training environments in facilitating virtue acquisition.

49Ap Dijkhuizen and A. Van Knippenberg, “The Relation between Perception and Behavior, or How to Win a Game of Trivial Pursuit,” Journal of Personality and Social Psychology 74 (1998): 865–77. For a review of some of this literature, see Ap Dijkhuizen and John Bargh, “The Perception-Behavior Expressway: Automatic Effects of Social Perception on Social Behavior,” in Advances in Experimental Social Psychology, vol. 40, ed. Mark P. Zanna (San Diego: Academic Press, 2001), 1–40.

50Lee Ross and Andrew Ward, “Naive Realism in Everyday Life: Implications for Social Conflict and Misunderstandings,” in Values and Knowledge, ed. Edward Reed, Elliot Turiel, and Terrance Brown (Mahwah, NJ: Erlbaum, 1996), 103–35.

51For more on zhengming, see Hui-Chieh Loy, “Analects 13.3 and the Doctrine of ‘Correcting Names,’” Monumenta Serica 51 (2003): 19–36.

52To take merely a few examples from the Analects, Confucius himself was reluctant to declare himself perfected, noting that “what can be said about me is no more than this: I work at it without growing tired and encourage others without growing weary” (7.34); on a similar note, he warned his disciples to “learn as if you will never catch up, and as if you feared losing what you have already attained” (8.17). Master Zeng in 8.7 notes that the journey of the gentleman ends “only with death,” and even the supposedly “good-by-birth” disciple Yan Hui is portrayed in Analects 9.11 as lamenting of the Confucian Way, “the more I look up at it the higher it seems; the more I delve into it, the harder it becomes. Catching a glimpse of it before me, I then suddenly find it at my back.”

53Harman, “Moral Philosophy.”

54Hutton, “Character, Situationism.”

55Xunzi, chap. 23 (“Human Nature Is Bad”), adapted from Hutton’s translation, with an additional line cited at the beginning.

56As one anonymous Ethics referee observes, one might argue that such an ethic should be seen as “situationism with a touch of virtue theory thrown in” rather than as a genuine virtue ethic. However, since situationism as it has typically been formulated leaves essentially no causal room for character traits, it would seem that, when it comes to early Confucian ethics, we are still clearly under the ambit of “virtue ethics,” although the relative roles being played by enhanced character traits and situational controls is an open question.

57See Mischel, “From Personality and Assessment,” for a helpful overview of this work on the “agentic, proactive self.”

58Ibid., 288.

59Xin refers to the organ of the heart, which by mid–Warring States was perceived as the locus of distinction making, language use, reasoning, and free will, as well as the locus of certain moral emotions. It thus does not correspond neatly to either “mind” or “heart.”

60On this topic, see Edward Slingerland, “Toward an Empirically Responsible Ethics: Cognitive Science, Virtue Ethics, and Effortless Attention in Early Chinese Thought,” in Effortless Attention: A New Perspective in the Cognitive Science of Attention and Action, ed. Brian Bruya (Cambridge, MA: MIT Press, 2010), 247–86, and “Of What Use?”

61Jonathan Cohen, “The Vulcanization of the Human Brain: A Neural Perspective on Interactions between Cognition and Emotion,” Journal of Economic Perspectives 19 (2004): 3–24.

62Ibid., 19.

63Patrick Hill and Daniel Lapsley, “Persons and Situations in the Moral Domain,” Journal of Research in Personality 43 (2009): 245–46.

64One of Mencius’s primary critiques of the consequentialist, rationalist Mohist school was that their ethical demands and extreme voluntarism went against basic human cognitive and emotional capacities and were therefore psychologically infeasible; Xunzi similarly criticized Mohism for ignoring basic human emotional tendencies and the ability of cultural forms to reshape these tendencies in an ethically desirable way.

65For more on this topic, see Donald Munro, A Chinese Ethics for the New Century: The Ch’ien Mu Lectures in History and Culture, and Other Essays on Science and Confucian Ethics (Hong Kong: Chinese University of Hong Kong Press, 2005); and Slingerland, “Of What Use?”