Ethics Discussions at PEA Soup: Edward Slingerland’s “The Situationist Critique and Early Confucian Virtue Ethics” with commentary by Rachana Kamtekar

We are pleased to present our next installment of PEA Soup's collaboration with Ethics, in which we host a discussion of one article from each volume of the journal.  The article selected from Volume 121, Issue 2, is Edward Slingerland’s “The Situationist Critique and Early Confucian Virtue Ethics” open access copy here).  We are very grateful to Rachana Kamtekar for starting our discussion; her commentary follows beneath the fold...

 

In ‘The situationist critique and early Confucian virtue ethics’, Edward Slingerland (henceforth ‘S.’) defends the empirical viability of virtue ethics. According to philosophical situationists like Gil Harman and John Doris, empirical psychology shows that people’s behaviour is much better predicted by features of their situations than it is by alleged features of their character: in similar situations, people tend to behave similarly; and variance in people’s behaviour is better explained by variance in their situations (or perceived situations) than by variance in their character-traits (such as whether they are honest or not, courageous or not, etc.) While people’s behaviour is highly consistent across narrowly-construed situations—so, for example, people who do not cheat on one exam will not cheat on others, although they may very well tell a lie or keep a found wallet—Harman and Doris do not count narrowly situation-specific dispositions as character traits: it is honesty, not honesty-in-exams, that is a character-trait. S. first criticizes the Doris-Harman position on empirical and conceptual grounds and then, having identified as the ‘core’ situationist complaint that virtue ethics ‘sets the bar for virtue too high’, argues that the practices prescribed by early Confucian virtue ethics, of life-long practice and regulation (of who one’s friends are, how one dresses, etc.), both ‘enhance the jump’ and ‘lower the bar’ to enable individuals to cross the hurdle of virtue. This is an intriguing perspective, but the role of this piece is to spark discussion, so I will focus on what seem to me problems with the paper—first in its diagnosis of situationism’s errors and second on its use of Confucian virtue ethics.

According to S., the only live issue in the person-situation debate is whether people have broad dispositions such as the disposition to behave aggressively, or only narrow dispositions such as the tendency to be verbally aggressive when chastised by an adult on the playground but not when approached by a peer.

Against the situationists’ finding that the correlation between one trait-relevant behaviour and another, in a relevantly similar situation, is a low 0.3, S. argues (1) that repeated trait-relevant behaviours are better predictors of behaviour in a relevantly similar situation, (2) that personality psychology’s ‘Big 5’ personality traits—openness, conscientiousness, extraversion, agreeableness and neuroticism—show considerable stability, and (3) that the correlation between chemotherapy and positive outcomes is 0.02 or 0.03, and the difference between the batting averages of the best and worst hitters in baseball history is 0.144: these ‘low’ correlations, he says, are good reasons to undergo chemotherapy or prefer the best over the worst hitter, so why isn’t a 0.3 correlation good reason to entrust my wallet to the student who hasn’t cheated on exams?

Point (1) above is correct and was made by Epstein in the 1980’s. Points (2) and (3) do not help S.’s case. Against (2): the ‘Big 5’ stable personality traits do not seem like the ingredients of virtue (especially not if their stability is a function of genetics, as S. suggests). Against (3): whether a correlation gives us good reason to act or not depends in part on the value of the outcomes to us: given that surviving cancer or a Pro baseball victory are hugely desirable outcomes, even a low correlation between a course of action and success counts as a good reason to prefer that course of action. (However, the 0.144 difference in batting averages is a difference between a player batting at 0.344 and another batting at 0.200—so the better player is batting nearly 75% better than the worse!) But the psychologists’ concern with correlations is theoretical, viz., how predictive is behaviour A in situation S1 of behaviour A in situation S2? (And if the expectation was 1, then 0.3 is low; also, to be correct we should say, instead of ‘P is honest’, ‘P is honest 30% of the time’ or ‘P is 30% honest’). As moral agents we are not going to be satisfied with a 0.3 correlation either: if we want to be honest, or for our friends to be honest, we want honesty all the time, across situations.

Or do we? S. also makes a ‘conceptual’ (or to my mind, ‘historical’) point. The traditional virtue terms, S. says, originally had their home in a narrow context, and had a narrow scope, because they arose and functioned in highly structured societies—unlike their folk psychological counterparts today. So for example, in early Confucian writers, xin, trustworthiness, is the trait a gentleman shows in his professional behaviour towards his colleagues, superiors and inferiors—but carries no implication that the trustworthy gentleman is trustworthy in his sexual relations. This is an important and complex issue. Is it that the virtue terms have lost their moorings? On the one hand, as S. observes, even today, ‘honest’ said of a mechanic bears on the mechanic’s professional conduct but not his sexual morals. On the other hand, if a man who happens to be a mechanic is called honest, that could be challenged on the basis of his sexual morals. But this is not a modern phenomenon: Plato’s use of dikaiosunê in the Republic perennially raises the question with readers whether ‘justice’ or ‘righteousness’ or ‘morality’ is the appropriate translation. Perhaps virtue-concepts are always open to extension out of their original sphere of application.

To show that at least Confucian virtue ethics can address the worry that ‘the bar for virtue is set too high’, S. argues that virtue ethics can ‘enhance the jump’: intensive, life-long, highly regimented training such as that advocated by early Confucians would make it possible for those who engage in it to behave with greater cross-situational consistency than can those with only untutored natural dispositions (so the idea may be that the Big 5 traits yield 0.3 consistency and training can increase it). As S. recognizes, this is an empirical claim, and he calls for empirical investigation. It is worth raising the issue of whether what Confucian virtue ethics produces is what Maria Merritt has called ‘situationally sustained’ virtue: dispositions that depend for their continued existence on situational and especially social factors. S. also thinks virtue ethics can ‘enhance the jump’ because it is possible to generalize narrow virtuous dispositions. S. cites work discussed by Nancy Snow on successes with conscious control over prejudice; it’s worth noting, however, that Snow does not think virtue can be like this—and neither should S., given his opening comments on the demise of cognitive control models of ethics. (Instead, according to Snow, virtue is a habit, consisting of virtue-related goals, the representations of which are automatically activated into an appropriate behavioural response when the agent is presented with a situational trigger.)

Although I am only able to consult a translation, I question S.’s reading of Mencius’ encounter with King Hsüan of Ch’i (Mencius I.A.7) as a case of generalizing local virtues. To show King Hsüan that he in fact has the capacity to feel compassion towards his subjects, Mencius recalls that when the King saw a distressed ox off to slaughter he ordered that a sheep be substituted in its place. Mencius identifies the king’s response as compassion, adding that if the king can feel for an ox he can certainly feel for his people, just as if you can lift a heavy weight you can lift a feather, and if you can see a fine hair you can see a cartload of firewood. S. says what Mencius has done is 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.’ But Mencius does not characterize what he is doing as broadening a narrow disposition; rather, he is pointing out that the King (who had asked about his ability to bring peace to the people) already has the capacity to be compassionate to his people—indeed, Mencius follows the analogy between the cartload and the hair by distinguishing between an inability to act and a refusal to act, suggesting that the king’s inaction is willful, perhaps involving some hypocrisy or self-deception.

A more open-ended question about Confucian virtue ethics concerns S.’s idea that attention to social role, dress, color, sound, correct names, and so on, indicates Confucian sensitivity to situational effects on our dispositions and behaviour. But is the Confucian attention to these things instrumental to producing good moral dispositions and behaviour or is it that for a Confucian, good dispositions and behaviour consist in part in behaving as one’s social role demands, dressing appropriately, etc.?

Comments

[Dan Boisvert]

I found Edward's article interesting. I do have several questions, a couple of which it appears Rachana has anticipated, but I'll go ahead and include them anyway.

One of the objections to virtue theory discussed in the paper derives from the apparently weak correlation (rarely above .3) between a variety of trait-relevant situations and the exemplification of the relevant traits. Edward tries to explain that what appears to be a small correlation is actually explanatorily more powerful (or perhaps a more powerful predictor?) than what we might ordinarily think and that we even have some intuitive grasp on the explanatory power of these apparently small correlations, as evidenced by the recommendations we would make in certain cases, e.g. whether to recommend a certain cancer treatment or who we recommend come to bat in the ninth inning of a particular baseball game. First, I may misunderstand correlation statistics (I confess I'm innumerate), but isn't saying that there is a .3 correlation between the exemplification of, say, honesty in a variety of honesty-relevant situations tantamount to saying that an "honest" person is dishonest 70% of the times in which honesty is called for? If so, I don't see how this fits with the kind of person a virtue theorist has in mind when appealing to the notion of a virtuous person.

The appeal to the cancer and baseball cases doesn't seem to help me. That we would recommend a particular cancer treatment in a particular case even given a low probability that the treatment will work in that case just seems to be a reflection of the importance of saving one's life in the absence of other more effective treatments, not a reflection of how effective we think the treatment really is across cases. (And similarly to whether we would want Ted Williams or Bob Ueker at bat on a particular occasion.)

About the "lowering the bar" side of the defense of Confusian virtue ethics. If I understood things right, this part of the defense is that Confusian virtue ethics espouses that one shape one's environment in various ways to make it more likely that one will behave well. But isn't this to give up the game to situationism? That is, if Confusian virtue ethics espouses that one needs to shape one's environment so that one's environment would cause one to behave in a way that is consistent with virtue, then one who behaves accordingly isn't behaving that way out of a virtuous character, but rather because of certain situational features. Or at least that's what it seems like to me.

[David Shoemaker]

Thanks, Rachana, for the insightful and provocative kick-off to the discussion. I too had some questions about the precise nature of Confucian ethics S. thoughtfully discusses in the paper, and I also wondered about the interpretation of the Mencius passage along the lines you have raised.

A different question I will raise goes to the former issue. On p. 412, S. points to the Confucian practice of "learning," which involves study and memorization of classics describing the thoughts and conduct of sages. It "served as a form of ever-present conceptual priming," so that a person steeped in Confucianism "would have always at the forefront of his mind the exemplary behavior and words of ancient paragons." This would ostensibly increase the odds that he'd act in accordance with their models.

But there's a kind of virtuous trait necessary to get the ball rolling here (perhaps a meta-virtue), "mindfulness," if you will. The Confucian scholar would need to be disposed to keep at the forefront of his mind the details of the sage's behavior. But then this raises a regress worry: what would prime such mindfulness? That is, what would increase the probability that the Confucian would keep these details at the forefront of his mind? Without the relevant sort of priming at this foundational point, we may not have a plausible story in place of the "master" virtue needed to "enhance the jump" in the first place.

[Steve Angle]

Here are some thoughts in response to points made by Dan and David. First, about the idea that Confucianism's recognition of the importance of situations means that it has given up the game to situationism. I think it is indeed the case that Confucians recognize that situations play important roles in shaping our behavior, as Ted says. But I don't think that things are as clear-cut from the psychology end as Dan suggests. Social-cognitive approaches to moral character and moral behavior emphasize the interdependence of trait and situation: one can't move from "situations are important" to "traits don't matter." It is true that a view according to which situations are irrelevant for all stages of moral learners is a non-starter. But beyond that, we have a complex picture in which both matter, though perhaps in different ways as one develops morally.

This leads me to a second point. As I understand it, the general Confucian view is that structured situations (paradigmatically, "rituals," which are understood to be quite ubiquitous) have two effects: they influence our behavior in the moment, and they encourage the development of more robust, relatively-situation-independent virtues. It is important to remember that no respectable virtue ethicist, and certainly no Confucian, thinks that highly virtuous people are common. Confucians understand people to be somewhere on a continuum of moral progress. Sages, if there are any, are extremely rare. So there is no statistical evidence relevant to the cross-situational behavior of the rare person in the upper spectrum of moral development.

Finally, in response to David's worry about regress: as it happens, I'm giving a paper that is relevant to this topic Thursday night at the Central; all are welcome, if you're in town! The idea is to look at the ways that psychological literature on the development of moral self-identity meshes with Confucian ideas concerning moral perception. The key point related to David's question is this: one can develop the tendency to chronically view things in moral terms without consciously developing this tendency. One doesn't need the virtue of mindfulness to motivate priming: it is (typically) the unconscious result of something like Confucian learning.

[Manyul Im]

Rachana, I think your point about Mencius 1A7 is exactly right. And there's nothing tricky about the passage in Chinese. The compassion the king feels for the ox, on Mencius's analysis, shows that the king actually possesses an excess of compassion -- the analogies to lifting and seeing things reveal the assumption that it's actually more difficult to feel compassion for oxen than for humans. Admittedly, misapplication of the 1A7 passage doesn't kill Slingerland's argument, but certainly doesn't help.

On a different note, I'm not entirely sure what type of life-long practices of early Confucianism are meant to be included as candidates for the development of consistent character. There are really two levels of moral practice that Confucius and his followers cared about. One is what you can expect of your ordinary farming peasant family -- the bulk of the population. For those people, the luxury of most of the Confucian practices (ritual training, study of historical documents, and interpretation of poetry, among other things) were practically out of reach. Most of the stuff Mencius says about "the people" has to do with making sure they are being taken care of economically and making sure the dire situations (wars, famines, and pestilence)in which they end up acting badly don't occur. The other level of moral practice is the one the nobility has the opportunity to pursue. But there's no evidence that very many people ascended to high levels of virtue based on the Confucian curriculum of traditional ritual and textual study and careful self-scrutiny. In fact, worries about literary genre aside, it's clear that "the sage" or "the true man" who epitomizes the globally consistent virtuous person is a rare exception, almost to the point that individuals like that are thought only to occur on historically epochal intervals.

[Brad Cokelet]

I really enjoyed this paper and, like other posters, was left wondering about the self-cultivation practices the Confucians recommend and how effective they can be.

I think the comparison with Buddhist practices might be interesting here.

For example, Buddhists commend specific concentration and mindfulness practices that are aimed to (i) enable one to keep a text in mind and apply its recommendations in one's experience, and (ii) develop an initially narrow compassion into one of broader scope. The first might be what David has in mind, I suppose.

The second is particularly interesting in the light of what others have said about the Mencius quote. Buddhists quite explicitly propose that you learn to develop compassion for people you initially hate by engaging in a meditation that starts by triggering your compassion for those you love (e.g. your parents).

More generally, I think Buddhists doubt that rational argumentation or advice from an adviser will enable one to "jump up". As Damien Keown has argued in his interesting book "Buddhist Ethics", many Buddhists think that our ability to profit from reflection, analysis, and advice depends on our uprooting various bad "non-cognitive" mental dispositions, and they think those can be best uprooted by concentration meditations.

In a related vein, this article brought to mind the booming psychological research on mindfulness meditation. In particular, there seems to be good evidence that mindfulness practices can help people overcome akratic smoking and drinking better than practices in which people try to bring to mind their reasons for not smoking and then resist their urges through strength of will.

So I wonder, first, whether the Confucians have analogous practices or whether they might not simply adopt these Buddhists one (with a less impartialistic form of compassion as the goal, of course). Or are they simply more optimistic about the fruit of things like reflection, reading, advice, and analysis?

Finally, perhaps Aristotle would agree with Buddhists that our ability to profit from reflection, analysis, and advice depends on our uprooting various bad non-cognitive dispositions - or not having those in the first place. Unlike the Buddhists and Confucians, however, they seem less optimistic about our ability to reform these dispositions if they are malformed by our poor education and culture. Just an impression - I am no ancient scholar!

[hagop sarkissian]

Great paper and discussion! I haven't had a chance to look at all the comments carefully, but I would like to quickly respond to some of the questions above. First, Rachana poses an open-ended question about Confucian virtue ethics:

“S.’s idea that attention to social role, dress, color, sound, correct names, and so on, indicates Confucian sensitivity to situational effects on our dispositions and behaviour. But is the Confucian attention to these things instrumental to producing good moral dispositions and behaviour or is it that for a Confucian, good dispositions and behaviour consist in part in behaving as one’s social role demands, dressing appropriately, etc.?”

On my view, the answer to both of these questions is ‘yes’. Confucians maintained that behaving as one’s role demands, dressing appropriately, using correct forms of address, etc., constitutes good behavior and reflects laudable dispositions. However, Confucians rather explicitly note that such behaviors also serve as ‘signals’ to others and prompt them (perhaps without them even knowing it) to reciprocate in kind and thus enhance one’s immediate environment. Analects 8.4, 14.42, and 1.12 are pretty direct statements of this idea. In other words, I believe there is clear evidence in the early texts that part of the rationale for adopting certain ritual forms was an awareness of how they might impact one’s immediate situation (including anyone present) and affect those in one’s sphere of influence.

Along similar lines, Dan Boisvert notes the following:

“If I understood things right, this part of the defense is that Confusian virtue ethics espouses that one shape one's environment in various ways to make it more likely that one will behave well. But isn't this to give up the game to situationism? That is, if Confusian virtue ethics espouses that one needs to shape one's environment so that one's environment would cause one to behave in a way that is consistent with virtue, then one who behaves accordingly isn't behaving that way out of a virtuous character, but rather because of certain situational features.”

I think a Confucian would say two general things here. First, drawing a sharp distinction between oneself and one’s environment is, for a Confucian, misleading. One’s behavior is continuously affected by one’s situation—for example, by the other individuals present, what they say and do, how they’re dressed, what time of day it is, etc. This is a brute fact of human life. Moreover (and really, this is a just a consequence of this general view) one is also a source of situational influence oneself.

Now, insofar as one proactively shapes one’s immediate environment (by being mindful of one’s impact on others and adjusting one’s behavior accordingly), one can shape and steer one’s situations in ways that are amenable to achieving virtuous ends. For Confucians, this is not giving up the game. Rather, it is exercising effective moral agency by harnessing situational variables and fostering an environment amenable to cooperative and accommodating behavior. (I should say that I associate these ideas with some early texts more than others; for example, there is less of this stuff in the Mencius than in, say, the Analects of Confucius or the Xunzi.)

Put another way, harnessing situational forces is 'giving up the game' to situationism only insofar as one considers being kind, conscientious, and accommodating to be 'giving up the game', if such qualities happen to create an environment that is conducive to achieving virtuous ends and sustaining virtuous behavior.

For those interested in these questions, I have much more to say about this in a recent article in Philosopher’s Imprint, which may be found here

[Dan Boisvert]

Steve and Hagop, thanks for those replies.

Just to amplify a bit, I didn't intend to move from "situations are important" to "traits don't matter." I meant instead to move from "situations are especially important" to "traits are relatively unimportant." The reason was that I focusing on this line from Edward's paper:

"Confusians … focused heavily on the manipulation of all aspects of the learner's physical, linguistic, and social environment" (p. 414).

If this is true, then consider just a few of the especially minor situational features of one's physical environment that the empirical research suggests one will have to manipulate in order to exemplify relevant virtuous traits when called for—e.g. fragrance, background noise, temperature, weather, lighting. Now first, if just these few minor physical features would need to be manipulated, then a person's behavior would appear to be much more greatly affected by situational features than by traits of character. Second, once we begin manipulating situational features to this extent, it looks like we are just in the business of doing what the situationist would think we should be doing: "encouraging" certain sorts of behavior by focusing much more on manipulation of situational features (to ensure right actions) rather than focusing so much on the development of a person's character. (Again, I don't mean to imply that a Confusian virtue ethicist—or a situationist, for that matter—wouldn't encourage both to some extent, but rather that the much greater emphasis would likely have to be on manipulation of situational features.)

So this is the sort of worry that I had in mind when wondering whether a Confusian virtue ethicist would be "giving up the game" to the situationist.

[Brad Cokelet]

Couldn't the Confucian grant that Dan is right about the influence of some situational factors (while adopting Hagop's line on those), but insist that there are other situational factors by which the more virtuous will characteristically *not* be affected?

For example, the Confucian could point out that the virtuous will not be affected by peer-pressure and others opinions in the way that ordinary people are - their sense of self-esteem and their motivation to do what is right will not be affected by others attitudes as easily. This seems to be an idea in the Analects -- and the Stoics, Buddhists, etc.

On this view the virtuous might still be affected by hot-and-cold-cup experiments (and the like) but not by Milgram like experiments. And moral cultivation would involve (a) developing Hagop-style approaches to some situational factors and (b) moral character that *insulates* one from being swayed by one's (less than virtuous) peers.

This seems to turn Dan's conclusion that when it comes to moral cultivation, "much greater emphasis would likely have to be on manipulation of situational features [than on the development of character]", into an interesting, open question about which situational affects are more important to resist. I think that being characteristically insulated from peer effects is more important than being adept at manipulating the situational factors from whose influence one cannot be insulated, but I am not sure the Confucians would agree - and of course the two probably can't be separated in practice.

[David Shoemaker]

Steve, thanks for the comments. You say in response to my regress worry:

"[O]ne can develop the tendency to chronically view things in moral terms without consciously developing this tendency. One doesn't need the virtue of mindfulness to motivate priming: it is (typically) the unconscious result of something like Confucian learning."

I didn't intend for the worry to be about conscious development of tendencies, only that the tendency to view things in moral terms seems to come across in S's paper as itself a character trait, i.e., a virtue. To the extent that it is, we're still owed a story about the relevant prompts for it that doesn't itself involve a particular sort of virtuous tendency (which will ostensibly need its own prompts). (Presumably, development of more specific/lower-order virtues may also be done unconsciously.)

On another matter, I find Brad's possible story in response to Dan's worries highly suggestive, but I agree that hope of finding a principled way to categorize distinct types of situational influences is likely impossible. We can easily think of cases in which being insulated from influences by dimes lying around or fart smells in the room could be more important than being insulated from peer effects, for instance.

[Edward Slingerland]

Thanks so much for all of the helpful comments about the piece, and my apologies for my delay in responding, I’ve just gotten back from being on the road, and unfortunately I only have time for a few quick responses before the one-week mark is upon us (I’m entirely new to blogging, and I understand that about a week is the usual lifetime for these sorts of discussions), and I’m currently playing a lot of catch-up.

Regarding the interpretation of Mencius 1:A:7, it is indeed the case that, as Rachana remarks, the feather/cart image sends a voluntarist message—just do it—rather than the gradualist one of extension that I focus on in the discussion, and I should have seen and dealt with this, but I disagree with Manyul that this means that I “misapplied” 1:A:7. As I’ve argued elsewhere, when analytic philosophers approach Chinese texts (or any text written by someone other than another analytic philosopher) they tend to want to extract a single, univocal and logically unambiguous argument from it, but this is not how normal argumentation works: Mencius, like any normal person (i.e., any non-analytic philosopher), has a general point he wants to make (the king has the capacity to be compassionate to the people, he should be compassionate to the people), and gets us there by mixing a bunch of images that have slightly different, but overlapping, entailments: voluntarist images like the feather/cart one, but also gradualist ones like learning to care for your own family first before then being able to care for other people’s families, or learning to “extend” what one is capable of. The way to get these two apparently incompatible schemas to cohere is to take the more voluntarist passages as being targeted at motivating people reluctant to even start the process of cultivation (look, you can do it, just do it), which is how the passage starts off (the King asks if such a one as him is capable of being a true King), while the stepwise extension schema more accurately describes how one goes about doing it once one has resolved to begin, and this is more what Mencius focuses on as the passage progresses. But I should have been clearer about this.

Regarding another of the issues raised in the commentary, how much predictive power we, as moral agents, want from our character-trait attributions, one of the things that I cut from the original ms. in the interest of space (because I realized that I needed to say more about it but didn’t have room) was a reference to Ahadhi and Diener 1989 (Multiple determinants and effect size, Journal of Personality and Social Psychology 56 (3):398-406), which argued that, even in experimental situations where only a few possible traits were involved that could cause particular outcomes, about .3-.4 was the about the theoretical best predictive force that you could expect from any single trait, given the complexities of multiple determinants (additive effects, conflicts, etc.). So, considering how noisy any real-life situation is, .3 is actually a surprisingly good predictive score for a single measure, like how X rated on some particular item of some character assessment test you gave them 3 weeks ago. There is a tendency to jump to the conclusion, aha, so situation is responsible for the remaining .7 (the remaining “noise” is all situational), but the noise could be produced instead by other, competing character traits, and it’s simply that we mis-framed the particular scenario where we were interested in predicting behavior (either because we really do misunderstand the scenario, or because the framing depends on who is doing the framing): we thought it was about honesty (and maybe the particular sub-type of honesty assessed in our original measure) and are solely focused on protecting the donation jar, but maybe it’s about compassion or courage or some combination of the two (the subject really needs a bit of money to help a friend in need, the organization from which it’s being nicked is not a worthy one, etc.). I still think that the reason people make character assessments of others has got to be about prediction of behavior at some level, but it’s an open empirical question (and I urge us to join the experimental philosophers and go out and check rather than rely on our own personal intuitions here) what percentage of the time we make virtue-specific judgments about people (he’s a fiscally-honest person) vs. general judgments like “he’s a good guy, I can trust him.” If it’s the latter, it may be because we, the “folk,” know at some level that framing situations precisely in perspective-neutral way is tough, and therefore a global judgment that X is likely to do “the right thing” is more reliable than specific predictions about responses to a narrowly-defined type of scenario.

Also remember that these experiments where “honesty” and similar traits are being measured are radically artificially simplified. Of course, I realize that that is what experiments aim to do (simplify situations to try to isolate individual factors), but it may also be the case that general character judgments are much more accurate in real-life, ecologically-valid situations, where framing, again, is tricky, and probably multiple and competing character traits are involved.

Finally, one my main points was that all of this work on “ethical” character traits has been done on randomly selected subjects with, as far as we know, absolutely no relevant training. Manyul and Steve, among others, have noted that the Confucians seem rather dubious about how widely-distributed positive character traits are (a disturbingly elitist, but perhaps empirically accurate, judgment), which is why they really only focus on a particular sub-set of the population. More encouragingly, though, there are reasons to believe that the training itself could get us significantly above .3, whether that is by “strengthening” the relevant character trait or just harmonizing peoples’ framings of particular types of situations (enemy soldiers shooting at you is a scenario that calls for courage and loyalty to the people right next to you in the trench, not prudence or compassion for your mother’s feelings or loyalty to your family back home that depends on you financially). There has been basically no work done on this, which would the most relevant since all forms of Confucianism (despite the voluntarist strains in the Mencius and other texts) argue that training on a variety of fronts is crucial to the development of reliable virtue.

And, yes, Rachana is right in noting that cognitive control of something like prejudice is not really what we're looking for, what I should have said is that that sort of conscious recognition can set the stage for "time-delayed" cognitive control, where you engage in practices that will give you new dispositions more in line with your conscious values.