Haidt, The Righteous Mind, ch5&ch6


5.1  WEIRD people (Western, Educated, Industrialized, Rich, Democratic people) are statistical outliers in the group of humans – and therefore bad samples for generalizing about the group of humans.  They are perhaps most obviously outliers in that, at least in cases not involving other-harming or unfair action, they resist inferring from feelings of disgust upon considering a social situation to that situation being morally bad or involving someone doing something morally wrong.  For example, they are much less inclined to say that someone having sex with a chicken carcass and then eating it is (universally, morally) wrong.  Similarly for other “harmless taboo” cases.  Therefore, good empirical moral psychology should not sample only WEIRDos (e.g., university students in the United States – hard to get much WEIRDer).

5.2 An explanatory account of WEIRDness:  such folks tend to think of the social and non-social world in terms of separate objects.  For this reason, Westerners do better than non-Westerners in the “absolute” version of the framed-line test, displaying a tendency to focus on, take in, and remember the absolute magnitude of the line as well as its magnitude in relation to the square box that it is in.  And, for this reason, Western folks tend to see themselves in terms of their preferences and life-plans rather than in terms of their role-relationships with others (the “I am…” test; e.g., ‘adventurous’ versus ‘a good father’).  And, regarding putative moral norms, they integrate their moral systems around the protection of individuals (hence, the focus on the claims that individuals have against each other and the state), not the protection of groups or society.

5.3 Richard Shweder’s three-fold distinction in kinds of moralities (or elements in a morality):  the ethic of autonomy (as dominates the WEIRD folks), the ethic of community, the ethic of divinity.  The ethic of autonomy is associated with caring/helping and rights.  The ethic of community is associated with the concepts duty, hierarchy, respect, reputation, patriotism.  The ethic of divinity is associated with the concept of sanctity and sin, purity and pollution, elevation and degradation.

5.4  How Haidt, an opinionated, partisan, atheist liberal became an “ethical pluralist.”  Studying the culture in the city of Bhubaneswar in Orissa, India by living with folks (anthropological method), he liked the people he was spending time with, began to actually feel the different values/norms of community and divinity – things started striking him as impure, degrading, or lacking respect for tradition.  Though he could equally see/feel the sexism (as when the women were not allowed to eat with the men), the non-WEIRDo, non-ethic-of-autonomy values became equally real.  And so, when he got back to the U.S., conservatives no longer seemed utterly crazy (and certain sorts of objects, like copies of important books, seemed holy).

5.4a  Descriptively, this kind of diversity in moral thinking seems quite important.  Even WEIRDos “sanctify” things and seem to intrinsically value respecting (certain sorts of) authority, for example, even though this is not given much voice in their ethical thinking.  However, for all that Haidt has said so far, the elements of the ethics of community and divinity (as characterized) have no normative purchase at all.  He needs to show that his “enculturation” experience gave rise to sensitivity to morally valuable things that he would otherwise have been blind to…



6.1  Haidt is trying to explain the diversity in moral thinking, but thinks we need to integrate evolutionary psychology rather than simply describe different approaches that seem to hang together (Shweder’s approach with his three-fold distinction in moralities.)  Using this kind of approach, we might view morality like tastes in cooking/cuisine.  The standards for cooking/cuisine, though varied, have to take into account the types of taste buds we have and how they interact or fit together.  So you end up with quite a bit of variation, but within certain confines (not a crazy scope of variation).  And, normatively, you evaluate a cuisine or a dish of food relative to our types of taste receptor (“this is sweet and it plays off the salty bits nicely”) but not just anything passes muster as a good dish.

6.2  What normative moral theory looks like if we don’t have a view like this, if we are ethical monists:  Bentham and utilitarian thinking (harm), Kant and deontological thinking (rights).  Thinking Bentham is right is like thinking that the best sort of restaurant would have to only have sweet stuff (say, because neural signals from sweet receptors are the strongest).  Bentham was autistic, Kant was borderline autistic.  Ultra WEIRDos.

6.3  If morality is pluralistic and we want to have a rigorous theory describing it, then we need a theory of what our “moral taste buds” are like.  More precisely, we need an account, in evolutionary-psych terms, of the basic functional elements that, like our taste receptors, provide the basis for cobbling together reasonable moral practices and evaluating them.  Haidt borrows the concept of cognitive modules to describe our “moral taste receptors.”  “Modules are like little switches in the brains of all animals.  They are switched on by patterns that were important for survival in a particular ecological niche, and when they detect that pattern, they send out a signal that (eventually) changes the animal’s behavior in a way that is (usually) adaptive…  An evolved cognitive module… is an adaptation to a range of phenomena that presented problems or opportunities in the ancestral environment of the species” (Sperber and Hirschfeld).  For example, evolution has provided humans with a face-recognition cognitive module.  And most mammals have a snake-recognition-and-avoidance module.  So you might start to get a good solid theory here by describing the “moral taste buds” and mapping them onto a cross-cultural list of common virtues associated with how the “moral taste buds” function.

6.4  The basic framework for describing what our “moral taste buds” are and how they function:  adaptive challenge, original trigger, current trigger, characteristic emotions, associated virtues.  This is the framework for “moral foundations theory.”  For example:  the care/harm “foundation” (or set of cognitive modules) goes like this.  Adaptive challenge: protect and care for children, Original Triggers: suffering or distress, or neediness expressed by one’s child, Current Triggers: baby seals, cute cartoon characters, Characteristic Emotions: compassion, Relevant Virtues: caring, kindness.

6.4a  Haidt officially endorses a “Humean” anthropological and psychological view of normativity (and hence moral normativity).  But, not being a philosopher, he does not really motivate why or how, from all the descriptive stuff, normativity is supposed to pop out.  The material about taste receptors, their affect, and how they function is suggestive (it is intuitive that a good dish has to “fit” how our taste receptors work in some way in order to be good – and intuitive that this is part of what it is for a dish to be good).  Perhaps this is even a good model for modern-day Humeans.  But Haidt does not really say at all how the normative bit is supposed to go.  What he says is just schematic:  what a good moral practice is, like what a good dish is, is constrained by relevant basic affective/motivational functional elements, not an anything-goes affair.

6.4b  In my view, the key to any plausible Humean view is the evaluation of desires and thereby the actions that they tend to produce.  Not all desires, not even all natural desires, are or remain part of the motivational/functional basis for evaluating other, more derivative desires and behaviors  – some, like planets in the wrong orbit, get flung out of the system.  Even granting this kind of Humean view of how normativity works, Haidt has not shown the none of his moral foundations, either generally or in a specific sort of social circumstance, are or become things that jeopardize a relevant sort of social functionality that is beneficial to us (rather than being something that, causally and functionally, tends to get “thrown out of orbit” and normatively should be thus flung away).  For all he has said, having a sanctity/degradation element is incompatible with the sort of moral functionality that best provides for the things we care most about or that benefit us the most.

6.4c  Later, Haidt says that he accepts utilitarianism (he seems to be convinced that deontology is superstitious rule-following).  But it is really unclear how he gets this from his Humean foundations (in fairness, he does just assert that he accepts something like welfare consequentialism).  Oddly, as far as I can tell, Haidt rejects deontological elements in morality partly because they are “part of the elephant” not a product of our powers of reasoning (as Greene’s applied philosophy results seem to show).  But this makes no sense.  Though we would need reason to discover this if it is true, why not hold that the urgent emotional element that corresponds to pushing the fat guy onto the tracks to stop the train (but not some action that changes the path of the train, an equal number of lives being saved) does not, along with relevant causal and functional elements, give rise to formally deontic reasons not to do certain sorts of things to others (that would, roughly speaking, constitute failing to respect their agency, welfare, dignity, etc.)?

32 thoughts on “Haidt, The Righteous Mind, ch5&ch6

  1. I originally wrote an intemperate and dismissive response to this under the influence of some alcohol (Plato would say that was just my true character coming through), but I decided that it was not a worthwhile contribution to the discussion. On a second reading, I have a few questions:

    1. You write that WEIRD people “are much more inclined to say that someone having sex with a chicken carcass and then eating it is (universally, morally) wrong.” Is that right? That seems to be directly contradictory of the preceding claim that such people tend to “resist inferring from feelings of disgust upon considering a social situation to that situation being morally bad or involving someone doing something morally wrong.” Did you mistype, or am I missing something?

    2. Do you think Schweder’s distinction between moralities is adequate? I’ve no trouble thinking of how one might embrace rather a lot of one of these kinds of morality and not much of the others, but I find it difficult to believe that actual morality among actual people tends to fit neatly into one or the other of these types to the exclusion of the others. Elements of all three sit together in various combinations in different strands of ancient Greek thought (though of course there’s little talk of ‘autonomy’ or ‘rights,’ there is plenty of focus on harm and caring), and I simply wouldn’t be able to fit the ethics of Homer, Hesiod, Aeschylus, Thucydides, or even Plato into one of these categories. It’s not just that they all tend to give some non-trivial role to items in each category; it’s that I don’t think we can rightly say even that the items in one or another of the categories tend to be dominant. Maybe the Greeks are weird, but while I’m much less well acquainted with it, I have a similar impression about classical Christian thought as one finds it in Augustine and Aquinas. Perhaps we shouldn’t take philosophers and intellectuals as our guide, but even if we restrict ourselves in the ancient Greek case to poetry and oratory, which both reflect popular values, it still seems hard to think that these categories illuminate them.

    3. To touch on the subject of my original, intemperate post: is there anything in Haidt’s book that is actually of value for philosophers thinking about ethics? From your summaries, he seems to help himself to a lot of false dichotomies and not to address the issues that almost any philosopher these days would recognize as crucial. Is this a false impression? One can’t expect him to be doing ethics like a philosopher, but when it’s all said and done, do ethicists have anything to learn about ethics (or meta-ethics) from this book?

    Liked by 1 person

    • Coincidentally, while I haven’t been reading Haidt (just pestering you and David about it without bothering to read it), I did recently read, for mostly unrelated reasons, a recent piece by Peter Railton, ‘The Affective Dog and Its Rational Tail,’ which takes up some of the claims of Haidt and other psychologists working with the dual-process model: https://www.law.yale.edu/system/files/documents/pdf/Intellectual_Life/LTW-Railton.pdf. Ultimately I didn’t find it very satisfying, but it does seem to show pretty clearly that there are no direct inferences to be made from the psychology to any anti-realist, anti-rationalist conclusions, or any ambitious conclusions of any sort, really. I was reminded of this by your concluding question about the emotional intuitions in trolley cases. If nothing else, Railton certainly seems to show that Haidt and others have drawn some very dubious conclusions from the studies of particular moral intuitions, and he makes a good case for thinking that our affective intuitions can be quite rational in a sense that is not obviously captured by prominent formulations of the dual-process model.

      Liked by 1 person

    • David,

      Since I just wrote a comment endorsing Michael’s summaries, I’ll reply to your point one: Yes, good catch, there must be a missing “not” there. The characteristic WEIRD response to the chicken case is something like, “It’s his chicken, nobody is getting hurt, so it may be perverted but it’s not wrong.” The WEIRDer the sample, the more this sort of response is given.

      Liked by 1 person

    • David R –

      Regarding your second comment: I’m uncertain of the descriptive value of Shweder’s distinction between moral systems of autonomy, community, and divinity. It seems like he meant to claim that most moral practices tend to fall squarely into one category or the other. I share your suspicions that this story is a bit too simple. However, we might treat these categories as features (or sets of features) that any given moral practice might instantiate to some extent. Haidt would view his moral foundations framework as having superior descriptive power (at least at the level of describing features of moral practice, if not at the level of usefully categorizing moral practices in a natural-kinds sort of way).

      Regarding your third comment: I find this sort of empirical work to help me get a better bead on the sense and reference of MORAL and MORALITY and a better sense of how the moral mode of human interaction functions (what it does and how). This latter is relevant, I think, to developing a Humean theory of reasons/normativity and morality.


      • Interesting. I do think that, at least from what you and David have written, Haidt’s deployment of the moral categories is potentially insightful, in large part because it does not seem to require that any given ‘morality’ fall into one or the other category, but simply identifies some clusters of concerns or considerations that any morality might display to some degree — and at least plausibly distinguishing them as they do on the grounds that the items in each category tend to cluster together but do not necessarily cluster with the items in the other categories. I’m still suspicious of the schema insofar as it seems to lend itself to oversimplification, but if the goal is simply to identify some distinct types of moral consideration that can be combined in various ways, then at the very least it seems to be useful for thinking about what we’re encountering.

        It’s funny, I tend to think that ‘moral’ and ‘morality’ have no stable sense or reference and that we’d be better off if we completely abandoned that language (that’s one way that I am a revisionist!). I think we should ideally just talk about what is good and why, what we have reason to do and why, and what traits or characteristics of persons are good and why. Though I’m generally a committed anti-Humean, I think this part of my view is neutral about Humeanism, depending on how we understand that. I take my view of ‘moral’ and ‘morality’ from Anscombe, Williams, MacIntyre, and the early Nussbaum (of the Fragility of Goodness period) — Williams is arguably a Humean in at least some respects, and MacIntyre in his pre-Thomistic phases has not implausibly been accused of Humeanism of a sort, whereas Anscombe and Nussbaum, in their very different ways, could hardly be more anti-Humean. I do not see what is lost by refusing to use the language of morality; therefore I do not see what is gained by using it, unless we count obscurity and confusion as a gain.


        • That said, I sometimes (as in these comments) use the language of morality as a nod to convention. I am, however, proud to say that I manage not to use it at all in my book on Aristotle except when (a) contesting the distinction between ‘morality’ and ‘prudence,’ (b) as a conventional shorthand for what Aristotelian scholars often call ‘moral’ virtue and ‘moral’ education — in hindsight I’d have chosen to use the language of character instead — and (c) in a half-ironic line about ‘our moral discourse.’ The sheer fact that it is possible to write a nearly 300 page book about Aristotle without using the word ‘moral’ in a sense that cannot be readily eliminated in favor of an alternative is itself testimony to just how dispensable that language is (assuming, of course, that I’ve got Aristotle right, which of course I have!).


          • Interesting. I view Haidt’s work as providing a key sort of evidence for the idea that moral behavior is a distinctive sort of behavior, the behavior of holding each other to account (in distinctive sorts of ways, from disapproval to punishment, to achieve distinctive sorts of ends concerning larger-scale action-coordination perhaps in order to achieve distinctive sorts of ends). From the distinctive functionality (and from the right sort of motivational attachment to elements of this functionality) arise a distinctive set of norms (that are motivated, action-guiding, normative-reason-entailing). And this suggests that, though we might be able to have a fine-grained-enough account of what we have reason to do (or what it is to lead a good life) that captures the information about how to best hold each other to account (and with regard to what), I think this runs the risk of ignoring the distinctive functionality of holding each other to account – and hence the distinctive sorts of reasons that we have as a function of this functionality. The distinctiveness and importance of our reactive attitudes speaks to the deep motivational and functional features involved in our being set up to hold each other to account.

            (I think Haidt, in focusing solely on the individual seeking to self-justify and avoid social penalty in the face of such social pressure, gives short shrift the the functional – and normative – texture that holding each other to account exhibits. There are important patterns in the ways in which and ends for the sake of which we typically hold each other to account. And I think that most normative views, not just the Humean ones, require having an accurate bead on such “deep” functional information about humans if they are to provide actionable norms. Again, the distinctiveness, motivational power, and social-functional impact of our reactive attitudes is indicative of how important this functional information about human nature is.)


            • There’s something to that, though I wonder whether in distinguishing ‘moral’ from sorts of broadly practical considerations it sets up an artificial distinction. Of course, if that’s so, it wouldn’t be unusual, so it wouldn’t be a special problem. But while there is probably no important sort of practical consideration that never comes into contact with ‘holding each other to account,’ aren’t there plenty of considerations that one might consider ‘moral’ and yet are not primarily about justifying ourselves to others, but simply about living well? If I read through Aristotle’s treatment of the virtues in Nicomachean Ethics III-IV, I’ll be hard pressed to find many virtues discussed there that are primarily about justifying ourselves to others or holding others to account. Nobody has come up with an entirely satisfactory story about just what unifies that particular list of virtues, but one view that has a lot going for it is that each virtue seems to govern some important area of life — fear and confidence, bodily appetites, anger, honor and respect from others, wealth and resources, general social interaction, truth in speech, humor. Some of these can only really come into play in social contexts — so-called ‘friendliness’ in the ‘interchange of words and deeds’ — and all of them often do, but most of them also govern areas of life that are only indirectly connected to justifying oneself to other people. Why draw a line through the other-regarding, justifying-and-holding-to-account dimensions of these and their self-regarding, living-a-life-that-is-good-for-oneself dimension? Aristotle himself says that justice is a part of each of these virtues, and that might look like drawing a distinction between the moral and the prudential and claiming that virtue involves both. But I’m inclined to think it’s really rather more unified than that, and in any case Aristotle’s ‘justice’ does not map very well onto ‘morality’ in any of its various uses. What’s left over if you abstract the ‘justice’ part of these virtues isn’t these virtues at all; there’s a distinction, but not into two fundamentally different kinds of reason.

              We might draw a similar conclusion from Haidt’s emphasis on the different kinds of morality. Considerations of purity will certainly enter into the practices of justifying oneself to others and holding others to account in a society whose members care about purity; but is what people care about when they care about purity really primarily a matter of holding others to account or justifying themselves to one another? That stuff seems to come into play for Haidt in explaining why moral reasoning takes the broad shape it does, not what morality is centrally about; part of his account is that the reasoning is secondary and functions to justify and serve the non-rational concerns that are what morality is fundamentally about. Of course I might just be misunderstanding the role the theory gives to these different parts. But his ideas seem to point more in the direction of a pluralistic conception of morality on which holding each other to account is important but not quite fundamental in shaping the moral concerns people have.

              Quite apart from Haidt, I wonder just what you have in mind when you talk about “a distinctive set of norms (that are motivated, action-guiding, normative-reason-entailing).” What are these distinctive norms such that they really are distinctive, i.e., importantly different in kind from motivated, action-guiding, normative-reason entailing norms that aren’t grounded in the behavior of holding each other to account? And just how distinctive are they? Why distinguish a whole different set of norms generated in this way rather than seeing the practice of holding each other to account as one among many features of human life that generates and constrains the reasons that people have (or take themselves to have)?


          • Perhaps there are reasons for me, and everyone else, to hold you to account so as to encourage you to be a courageous person (for example). And reasons that concern how this (you or any given other person being courageous) impacts all of us. Of course, we need to ask why you being courageous – something that you have independent reason to do – is of interest to all of us (and of interest in such a way that we should hold you to account on this count). But this seems like a promising way of integrating something like (some) reasons of personal excellence with our reasons for holding each other to account (in the group-centric moral way or such that group-centric moral ends are more likely achieved). Does this kind of idea or approach address your concerns at all, David?


            • Not really, but I’m no longer sure I can adequately formulate just what my concerns are. Certainly what you say here seems unobjectionable. But perhaps my muddy worry is something like the thought that we can’t, or at least shouldn’t, try to treat these as two kinds of reasons or as generating two kinds of norms. Intuitively, it seems better to say that we should hold each other to account because we have reasons to, reasons that are grounded in the shared, intersecting, interlocking interests we have in common, which are not neatly separable from the interests each of us has individually — better to say that than to think either in terms of one set of interests, reasons, and norms that are group-centric and another set of interests, reasons, and norms that are individual-centric or in terms of the practice of holding each other to account generating such norms. But this is all so vague that I’m not sure it amounts to much of a difference.

              I think what I’d be most interested in is seeing an account along the lines you’re suggesting filled out a bit. In one way it might look something roughly like the kind of thing Derek seems to be up to, and that Rawls, in his own way, does, the idea being that the demands of mutually beneficial co-operation yield a set of norms that are distinctive both in their content and in their origin. In Rawls’ (and perhaps Derek’s, though it’s unclear so far) case, one misgiving I have is that the norms of justice (or ‘moral’ norms, if the term is roughly equivalent) end up looking more distinct from other sorts than I think they are, and they end up that way because we begin with too thin a conception of individual interests (though Rawls complicates that issue considerably in part III of ToJ, where he comes at least very close to endorsing substantive, non-subjective interests to which justice bears something like a constitutive, non-instrumental relation). Crudely put, my thought is that this sort of account is misleading because mutually beneficial co-operation is not merely something we need because we are mutually vulnerable and unable to achieve our ends securely without it, but intrinsic to most or all of the goods that constitute human well-being.

              But even if that worry is rightly applied to Rawls and similar projects in the liberal individualist tradition, it might be either misplaced with respect to the kind of idea you’re sketching or have only relatively trivial implications, more a matter of emphasis than substance. In any case, if you’d like to develop the thought further in a series of posts, it’d be interesting, whether or not Haidt’s and similar work bears on it.

              Sorry if this is all too barely intelligible; I really should be in bed at this hour rather than trying to work out thoughts that I already know are fuzzy.


  2. Michael,

    Once again, I still haven’t managed to read the chapters you’re commenting on, but as before I don’t think I really need to. When I do finally get to them in the next day or so, if it turns out I’ve gotten something wrong, I will announce it. Your summaries match my own understanding pretty well. It’s your comments that often seem off to me.

    Nearly all your comments take it that Haidt’s theory is supposed to have some “normative purchase,” that “normativity is supposed to pop out” of it somehow. I’m sure this is incorrect. See the second page of “The Moral Mind” (2007), by Haidt and Joseph, for example (which is worth reading anyway). It reads, “Specifically, we link our view of moral innateness with virtue theory … We are not proposing that virtue ethics is the best normative moral theory. We speak only descriptively…” [emphasis theirs]. See also TRM, 114. Haidt is a psychologist, not a philosopher. He understands the difference. He does not think that the empirical methods of psychology constitute a new way of doing moral philosophy. He does not think he can discover what is good or right by developing a psychological theory of people’s judgments of what is good or right. Indeed, if anything it’s the opposite. Just as Nietzsche’s investigation of the genealogy of morals was intended to undercut Christian morality’s claims to universality and validity, Haidt’s investigation of the evolutionary and cultural “genealogy” of moral judgments has a similar impact. Intentionally, I think. More on this below.

    Meanwhile, this brings us to Haidt’s “Humeanism.” When Haidt makes cracks like, “Hume was right,” this shouldn’t be taken in any strong or literal way, as it might be if Haidt were a philosopher. One of the parade of dissertation advisors I had in philosophy grad school had the text “Hume was right” bouncing around his computer monitor as a screen saver. (Remember those?!) This guy—the most lovable of my dissertation advisors—really meant it, I think; much more seriously than Haidt does. So with respect to Hume’s theories of the passions, desire and motivation, practical reason, the relation of reason to the passions, and all of that—you can just forget it as far as Haidt goes. I don’t think he means to be endorsing anything more than Hume’s noncognitivism.

    Haidt is saying that everyday moral judgments—which is nearly all the ones that take place outside the philosophy seminar room—are not based on discernment of moral reality. Think about Kant’s view that reason more or less exists just so we can discern the moral law. (Reason isn’t strong enough to serve as a guide to desire satisfaction and happiness; therefore, its role is to make us good, not happy. See the Groundwork, 4:396 [about seven paragraphs into Section I].) Haidt is saying that, as a descriptive model of moral reasoning, this is a rationalist fantasy. What actually drives everyday moral judgment is elephantine intuitions, not rational discernment of the moral law. (Note that it does not follow that there is no Kantian moral law. Again, Haidt’s theory is merely descriptive.) Nor is there any other good, whether Platonic, Moorean, utilitarian, or Aristotelian, that ordinary people are perceiving or figuring out when they make moral judgments. Rather, the foundations of moral judgment are internal promptings of the elephant in response to certain kinds of stimuli. As an elementary but realistic example, think of some people’s feelings of disgust when they hear of certain sexual practices. People are inclined to moralize these feelings and say, “That’s wrong!” Note, again, that this doesn’t make it actually wrong, and Haidt is not saying that it does. But he is saying that this is paradigmatic of the origins of our moral judgments—all of them, not just judgments of sexual morality. The way in which this view can be called Humean is obvious.

    Hume likens moral judgments to judgments of taste and beauty (see for example the Enquiry, Section I). Haidt for his part makes a running analogy between judgments of gustatory taste and moral judgments. Thus, our judgments of taste are founded on our taste buds, of which by the usual modern count there are five types: sweet, salty, bitter, sour, and umami. There are good evolutionary reasons why we have these particular receptor types. For example, our ancestors, who were attracted to sweet things and salty things, more often survived to raise their children to adulthood than others, who weren’t. Of course, this does not confer normative status on our taste buds. Our sweet receptors are not discerning the good or the healthy or the nutritious or the caloric. Nor are our sweet receptors necessarily a reliable guide to any of these things. Indeed, our taste receptors may often lead us quite astray. Rather, we are left to figure out their function and evolutionary origins through reason. By doing so, we can learn to evaluate and be guided by their promptings more wisely. In the meantime, we should observe that the five basic taste sensations are only the foundations—the building blocks—of our judgments of taste. Our taste judgments result from a complex automatic computation from the array of all five taste sensations, together with our personal likes and dislikes and associations from childhood memories, recent menu history, hunger level, peer judgments, cooking shows, and God knows what all. Our taste judgments are also to some degree culturally determined. For instance, France, Spain, Italy, and Germany all have markedly different national cuisines and accordingly different national taste preferences, which can hardly be attributed to different genetic constitutions. Rather, they are due to different accidents of geography and environment, history, and individual decision making.

    Haidt thinks all this maps onto moral judgment pretty exactly, even to there being five foundational sources of judgment (though I see that in a later chapter he will add a sixth). For moral judgment, the five are: (1) Harm/Care; stimulated by harm, distress, or threat to one’s kin (especially one’s children), it stimulates the compassionate desire to help; (2) Fairness/Reciprocity; stimulated by interaction with nonkin peers, it produces (a) an impulse to equality of distribution, turn-taking, and other rigidly reciprocal relations, and (b) an impulse to jealously guard one’s prerogatives in such relations and to be wary of and angry at cheaters; (3) In-group/Loyalty; stimulated by group undertakings, it produces a desire to maintain group cohesion and a willingness to subordinate one’s own desires to the group; (4) Authority/Respect; stimulated by hierarchical relations (adult–child, chief–subordinate, elite–commoner), it produces respect and awe in subordinates; (5) Purity/Sanctity; stimulated by waste products, rot, and disease among other things, it produces disgust and a desire to avoid and purify oneself of such things. Haidt thinks these are innate, genetically programmed impulses. Although in our evolutionary history they were triggered by a relatively restricted range of stimuli, that range bleeds easily out to encompass the wider range of stimuli that confront us in the modern world. Originally triggered by crying babies, compassion today is easily triggered by the thought of baby seals and by those little figurines of children with pug noses and saucer-sized eyes. (This is actually a point of disanalogy with taste buds, whose range of stimuli is much more hardwired.) Our intuitive moral judgments result from complex automatic computations from these five foundational intuitions together with such explicit moral principles as we’ve imbibed, our personal attitudes, the associations that have built up over the years of our social and cultural experience, our recent personal experiences and hobbyhorses, the mood we’re in, and many other factors.

    Also, our intuitive moral judgments depend heavily on culture. Just as there are different world cuisines, so there are different world moralities. Just as a cuisine is a contingent product of geography and environment, history, and individual decision making, so exactly is the morality of a culture. Western culture and morality has evolved in a distinctive way. Most of the world’s cultures have quite different prevailing moral views from ours. In particular, ours is radically stripped down, since we tend to think of Fairness/Reciprocity as the whole of morality, with maybe some Harm/Care thrown in. But the notion of Purity/Sanctity seems alien to us, we take a jaundiced view of In-group/Loyalty, and we positively deprecate the idea of Authority/Respect as a moral principle. In these ways, however, we are the odd ones. If morals were determined by majority vote, (especially if we were counting by distinct religio–linguistic cultures, not by total population,) we would lose on these issues.

    Finally, of course the relation between the five moral foundations and what is really good or right will be oblique at best. Genetic evolution works over millennia. It does not respond quickly to the emergence of new environmental conditions. Therefore, the trends and conditions by which it shapes us must be highly general. The same In-group/Loyalty intuition that produces the heroism of a soldier falling on a grenade for the sake of his comrades probably also produces the genocide of ethnic cleansing. So, whatever is the way forward or guidance that can be provided from knowledge of the foundations of our moral intuitions to knowledge of real morality, it will hardly be anything so simple as, “genetic evolution has programmed us to have such-and-such moral intuition, so such-and-such must surely be right/good.”

    This goes likewise for the prevailing moral norms of any culture. Cultural evolution, however, responds to changing conditions much faster than genetic evolution. It is also more sophisticated in that it shapes and transmits ideas, skills, habits, norms, and traditions, not just anatomy and physiology. Culturally evolved moral systems, therefore, have a much better claim to be adaptive than the five intuitive moral foundations. (I should say that this is my own observation; I have no idea whether Haidt says this.) Still, just because a moral system prevails in a cultural does not mean that it necessarily is adaptive. (See Joseph Henrich [2009], “The Evolution of Costly Displays, Cooperation, and Religion: Credibility Enhancing Displays and Their Implications for Cultural Evolution.” An overview of this research program is Maciej Chudek and Joseph Henrich [2011], “Culture-Gene Coevolution, Norm-Psychology, and the Emergence of Human Prosociality.”) Therefore—to say it for the umpteenth time—the normative work of moral philosophy is in no way superseded by any of this empirical work, nor do any of its authors (I’m thinking specifically of Haidt and Henrich) think otherwise.

    But then, what is the value or interest of any of this for moral philosophy? Surely its interest is considerable. People do not begin the serious study of morals in an intellectual vacuum. When Socrates began to investigate moral questions, he necessarily began from the moral judgments and values that were common in his cultural milieu. And nothing has fundamentally changed since his day. We now have a vast body of systematic thought about moral questions to add to the mix, of course. But we are still dependent on pretheoretic moral judgments and attitudes to alert us to the fact that there is even such a domain as “morals” to study. So it behooves us to have some idea of the origins of our moral intuitions and some idea of their epistemic and normative status. Any theory of morals, if it is to be worth anything, should situate itself with respect to our ordinary moral intuitions. Whether it draws on those intuitions for suggestions or attempts to establish normative credentials independently of them, it should explain its choice, and especially it should explain why our moral intuitions are as they are, if the proffered normative moral theory is true. For instance, if a given moral theory says our moral intuitions are mostly wrong, it should give an account of the errors and say why, nevertheless, the theory is correct. Indeed, if it differs so much from our moral intuitions, it should show how it nevertheless counts as a moral theory. On the other hand, if it largely vindicates our moral intuitions, it should have an account of how the blind force of natural selection managed to produce a correct moral theory.

    In general, the development of a worthwhile moral psychology means that moral philosophers can no longer take moral intuitions simply for granted. Once having been given some scientific insight into their origins and status, a moral philosophy that wishes to be taken seriously, especially in the long run, will have to deal with it. The stopper is off the bottle, and the genie won’t be going back in.

    One particular consequence of a scientific and evolutionary theory of moral intuitions is that it raises questions about the function of morality that I don’t think moral philosophers are accustomed to ask. Another is that it more or less decisively puts paid to the notion that moral intuitions per se can provide epistemic warrant for moral judgments. If moral intuitions are elephantine judgments that have been determined by genetic and cultural evolution together with personal preferences, associations, and history, all of which is only obliquely related to even a biological conception of the human good, much less any more comprehensive one, then moral intuitions per se may not be taken as warrant for believing that any particular moral judgment—or any elaborated moral theory—is correct. Perhaps our moral intuitions can be shaped by our moral reasoning. As we grow in maturity and reflect on moral life, we gain insights and knowledge that becomes incorporated into the elephant. Perhaps we can in effect train the elephant over time, as Aristotle suggested, and then our moral intuitions would be more trustworthy. Yes, but then that training depends on having some standard of evaluation that goes beyond mere intuitions. In particular, the notion, harbored by some moral philosophers, that moral philosophy just consists of massaging the collection of all moral intuitions into “reflective equilibrium,” where the standard of equilibrium is merely formal (consistency, conservatism, simplicity, etc.)—this will have to be given up. If anything like Haidt’s theory is right, then moral intuitionsper se assume the status of hints, suggestions, heuristics, and no more. They cannot be the source of a normative moral theory.

    Not that we needed Haidt’s help to know that intuitions per se are epistemically worthless. “Intuition” most of the time is just a fancy word for opinion. Philosophers who like to talk about their “intuitions” should try the experiment of substituting “opinion” whenever they feel the urge to say “intuition.” Then ask yourself whether your argument/claim thereby loses anything epistemically. My bet is you’ll find that epistemically it has lost nothing, but rhetorically it has lost a lot. The lower the ratio, the more illusory are your supposed reasons for your claim.

    It isn’t all bleak. In “The Moral Mind,” Haidt and Joseph suggest that if we want a normative moral theory that is compatible with how people actually reason morally, then Aristotelian virtue theory is the ticket. Moreover, they seem to suggest that Haidt’s theory of ordinary moral judgment and intuitions can be integrated with the sort of attempt to put moral principles on a scientific foundation that I advocated in “Morals and the Free Society,” just as I speculated a month ago. Here is what they say:

    Virtue theory posits a particular kind of organization of moral competence, one in which perception, motivation, action, and reasoning correspond to demands placed on the person by features of situations. Naturally, the objectivity of these demands, and the moral relevance of features of situations, are to some degree dictated by the culture, by the moral concepts, social structures, and narratives that are current in the immediate social context. But this does not mean that the content or structure of a virtue is completely culturally relative. As Aristotle pointed out, and as current virtue ethicists have elaborated (Nussbaum 1993 [“Non-relative Virtues: An Aristotelian Approach”]), what it means for a personality characteristic to be a virtue, and not simply a behavioral regularity, is largely that it consists in functioning well in a specific “sphere of existence.” And what Aristotle and Nussbaum mean by “sphere of existence” is similar to what evolutionary biologists would recognize as persistent adaptive challenges and other types of environmental constraint. Virtues are therefore quite at home in a scientific theory of moral functioning based on evolutionary psychology and cultural psychology.

    This sounds enticingly like my own program. I admit that the statement is not completely unambiguous—as to whether the “science” they’re talking about includes normative elements—but the philosophers they cite were certainly doing normative virtue theory, which implies an answer in the affirmative.

    Finally, on your last comment, 6.4c, and Haidt’s supposed utilitarianism, I guess you are referring to TRM, 316, where he says he’s a “Durkheimian utilitarian.” Whatever that means exactly, it definitely means that he rejects any one-dimensional hedonistic or preference-based utilitarianism of the sort that is so common in contemporary social science. He does not mean it to apply to personal ethics; there he reaffirms that virtue ethics fits human nature most closely. He also restricts the appropriateness of this “utilitarianism” to decisions about legal structures and social institutions in liberal Western democracies that encompass substantial ethnic diversity. This is a pretty limited brief for utilitarianism! We had probably best wait until we reach that part of the book to try to evaluate this.

    Liked by 1 person

    • I’ll respond to more of these points later, but right now just this one: that Haidt is embracing noncognitivism about moral judgment. He does seem to think that the point or purpose of moral reasoning and judgment is something like negotiating the practice of our “holding each other to account,” not the discernment of reality. But this is consistent with the nature of moral judgment being, at least in part, cognitive. Moral judgment sure seems cognitive and there is little reason to think that Haidt embraces noncognitivism as a metaethical thesis.

      Perhaps when we say things like “Breaking one’s solemn oath is wrong,” we are saying that we have conclusive reasons (of the right kind) to hold each other responsible for not breaking solemn oaths. Such judgments might also practically in various ways including to motivate others (and ourselves) not to break solemn oaths.

      I view the sort of descriptive work that Haidt does as potentially shedding light on the following conceptual point: what sorts of practices (and motivations and perhaps normative reasons) count as moral?


      • He does seem to think that the point or purpose of moral reasoning and judgment is something like negotiating the practice of our “holding each other to account,” not the discernment of reality.

        If by point or purpose, you mean the function that morality actually serves in human life, then yes, it seems pretty clear that he thinks morality serves to make us social beings; for instance, able to cooperate in large groups of nonkin. I think he will take up this matter much more in part 3 of the book, and there we may get some more insight into how the blind moral promptings of the elephant might be integrated with cognitive moral understanding, if and when this is achieved.

        Moral judgment sure seems cognitive and there is little reason to think that Haidt embraces noncognitivism as a metaethical thesis.

        Well, what do you make of the passage on page 135, where he quotes Hume approvingly to the effect that, “Moral Perceptions, therefore, ought not to be class’d with the Operations of the Understanding…”? Moreover, if moral judgment is derived from moral “taste buds,” how can it be cognitive? Taste buds aren’t cognitive processors. The whole point of the taste bud analogy is to stress the contrast with rational judgment.

        However, moral judgment can easily become cognitive, on Haidt’s view—or so it seems to me. For example, in Haidt’s system, there will be a virtue of Loyalty. As a child grows up, he learns gradually to value loyalty and to perceive which occasions are appropriate for its exercise and which are not. A mature adult with a proper virtue of loyalty will exercise this virtue as a kind of social skill, knowing the right conditions in which to exercise loyalty, the right degree of loyalty to be exercised, the right people to whom loyalty is owed, and so forth. But all this is a mere matter of socialization, not understanding, and “right” in the previous sentence just means in accordance with prevailing social standards, not right in the sense of being in accordance with some actual moral law or promoting some actual good. I mean, it may in fact promote some actual good—and let’s suppose it does—but our citizen has no knowledge of this. His moral thinking is noncognitivistic. But now suppose some great moral teacher comes along and discovers moral truth. He discovers the real function of the virtue of loyalty and establishes what moral good is served by loyalty. This might enable him to make a few reforms in the prevailing standards on loyalty in this community. But whether reform is needed or not, surely having real knowledge of how and why loyalty is good must make people’s exercise of loyalty more firm and reliable. When people learn this knowledge, then, they become actually wise in their use of this virtue. Their behavior may not change much; nevertheless, they become cognitivistic moral thinkers.


        • Well, if tastebuds aren’t cognitive, but moral judgments can become cognitive, then the analogy becomes pretty thin, doesn’t it? Tastebuds can perhaps be cognitive — my brother, who spent some years as a cook and went to culinary school, can discern fine-grained differences in the ingredients of food that I certainly can’t. But in any case, whatever Haidt says, what should we make of the idea that the emotions at the bottom of morality on his account are non-cognitive? Consider Paul Rozin’s research on disgust, as discussed by Nussbaum in Upheavals of Thought. In one study, he asks the test subjects to sniff two vials that in fact contain the same substance in the same amount and therefore smell identical, but he tells them that one vial contains cheese and the other vial contains feces. In most cases, the subjects claim to enjoy the smell of what they think is cheese and to be disgusted at the smell of what they think is feces. Rozin takes this and similar evidence to show that disgust depends on people’s beliefs about the objects they’re encountering and cannot be regarded as a simple non-cognitive response to sensory input, and that does indeed seem to be the only reasonable conclusion to be drawn from the evidence. It is less easy to say just what sorts of beliefs or ideas about objects render them disgusting, but Nussbaum follows Rozin in thinking that the core idea is that we will be contaminated by coming into contact with the object; upon noticing that all objects of disgust are things regarded either as animals, animal products, or something that has been in contact with them, she suggests that the “the core idea in disgust is a belief that if we take in the animalness of animal secretions we will ourselves be reduced to the status of animals,” and that since “we also have disgust reactions to the spoiled or decaying…Disgust thus wards off both animality in general and the mortality that is so prominent in our loathing of our animality” (203). Whatever we think about that particular proposal, the evidence points strongly to disgust depending on beliefs, beliefs that can be true or false and are not mere expressions of disapproval, let alone brute sensory responses to certain kinds of stimulation. It is not just that, as with flavors, our reaction to certain sensory input can vary a great deal across cultures and individuals. It is that the same person encountering exactly the same thing can be disgusted at it if she believes one thing about it but not if she believes another thing about it.

          It’s possible that when Haidt approves of Hume’s claim that ‘moral perceptions’ should not be regarded as ‘operations of the understanding,’ all he means is that we should not regard them as the conclusions of rational inference, the sort of thing the rider engages in. That seems perfectly correct for our intuitive emotional responses, ‘moral’ or not. But Hume and genuine non-cognitivists mean much more than that. They mean that there is, at bottom, no cognitive content to emotions, and hence to moral approval or disapproval; approval and disapproval is all there is. It sure doesn’t sound like that’s really Haidt’s view, at least not consistently. But whatever he says, it sure doesn’t seem true.


          • Findings of this sort concerning the emotions are familiar. They would be especially so to Haidt, who is a frequent co-author of Rozin’s. It is a commonplace that emotions depend on our cognitive recognition of how we take reality to be. Whether you take the object in the dim corner of the tent to be coiled rope or a snake affects your emotional reaction. I would be astonished if Hume and nearly everybody else who has thought seriously about the emotions wouldn’t agree. But this doesn’t make the evaluative part of the emotion cognitive. It doesn’t mean that the emotion itself is a belief. I doubt that Rozin thinks so, though I can’t say for sure (I’ve only ever read one of his disgust papers). Certainly nothing in the research you report shows anything like this. The evaluative aspect is that the disgusting object is “revolting” and to be avoided, expelled, removed from contact with oneself. There is no reason to think this is a belief as opposed to a direct, noncognitive matter of feeling and motivation. The fact that disgustingness can attach to an object by “contagion” (for instance, by having touched something disgusting) doesn’t change this. We can seek to interpret what disgust must “mean” by examining the wide variety of its triggers and the pattern of its action tendencies (what it makes us want to do), as Nussbaum apparently does, but that interpretation, even if correct, does not mean that the emotion is really a belief. Disgust presumably evolved genetically to serve some function, and certainly its total function has been shaped further by cultural evolution. This function will be what the interpretation captures, I suppose. But just because there is a function does not mean that the function is encoded as a belief; in general, it almost never is.

            Haidt himself does not seem to think that disgust (or other emotions) are cognitive in the way I’m discussing. See TRM, 172–173 and “The Moral Emotions” (p. 857 of the printed version). So, I think Haidt qualifies as a “genuine noncognitivist” about ordinary moral judgment.


          • I should add one more thing. We should remember that the moral judgments of the elephant needn’t be emotional at all for Haidt. A moral judgment might take the form of a direct intuition, that, say, X is just wrong. This is also noncognitivist, because Haidt doesn’t think that the elephant cognizes moral reality.


            • I don’t think anybody contests the claim that people often have intuitions that X is ‘just wrong.’ I’m not sure how sensible it is to think that such intuitions are direct — they would seem more intuitively (!) to be generated by emotions. I’m also not sure that we have good grounds to think that these judgments are non-cognitive; that depends on what we take the content of ‘just wrong’ to be, and whether it is anything more than an expression of a non-cognitive distaste. But neither of these points really matters, because cognitivists have no need to deny that people ever use moral language to express non-cognitive sentiments. Haidt cannot hope to defend non-cognitivism about moral judgment by showing that some moral judgments are just expressions of non-cognitive sentiments. That’d be like trying to show that nobody can speak French because lots of people can’t.


          • I’m not aware of anyone who holds that emotions are beliefs; certainly I didn’t suggest that they are. Nussbaum, whose cognitivism about the emotions is about as strong as can be, holds that emotions are judgments, and she thinks that such judgments have a linguistically formulable propositional content, but not that the subject must represent that content to itself linguistically (else children and non-linguistic animals could have no emotions); the radicalism of her cognitivism lies mainly in her claim that the judgment is the sole necessary constituent of the emotion. Less radical cognitivist views likewise do not identify emotions with beliefs, though many hold that beliefs are components of emotions, others simply that they are necessary to the formation of emotions. I suppose there’s probably someone out there I haven’t read who claims that emotions are nothing more than beliefs, but that’s not central to cognitivism. What is central is rather that beliefs or belief-like appraisals are essential to emotions, either as components of the emotion or as part of the process whereby one comes to have an emotion. This is why they are subject to rational criticism on the grounds that the propositional attitudes that either (partially) constitute them or generate them are false or otherwise unreasonable. The sensation I have when I put a spoon full of salt into my mouth is not similarly subject to criticism, nor similarly sensitive to my beliefs about salt; neither is my liking or disliking that sensation. The fact that motivation is essential to emotion does nothing to gainsay cognitivism, since there is no reason to believe that motivations necessarily lack cognitive content, propositionally formulated or otherwise. To suggest that one can embrace all of these claims and still hold a non-cognitive view of emotion or the subset of emotions that constitute moral judgments can only be to use ‘non-cognitive’ in a highly non-standard way.

            Hume is ordinarily regarded as a non-cognitivist because he is ordinarily read as rejecting these claims. He explicitly denies that passions can be reasonable or unreasonable and argues that only passion and never reason moves us to act. It is hardly an eccentric interpretation of these claims to take him as holding that moral ‘judgments’ are mere sentiments with no cognitive content whatsoever and so not the sort of thing that can be true or false in any sense. Unsurprisingly, this interpretation has been contested (see http://plato.stanford.edu/entries/hume-moral/#ear), but it is sufficiently plausible that emotivists, who do affirm such views, have frequently taken Hume as an inspiration. It would of course not be unsurprising if Hume also frequently says things that are inconsistent with this picture; Hume is full of contradictions, even embarrassingly obvious ones (that’s why he didn’t commit his own writings to the flames), and in any case he would by no means be the first person to fail to fit all of his observations into a Procrustean theory.

            But interpretation of Hume isn’t really the issue here. The issue is whether Haidt’s theory is genuinely non-cognitive and, if so, whether it is plausible. I’m ultimately not married to any particular use of the expression ‘[genuinely] non-cognitive,’ so I’m happy to let you use that term as you like. But only if it’s clear that it’s not being used in the way that is standard in philosophical theories of moral judgment and both philosophical and psychological theories of the emotions. And whatever terms we use, the fact that emotions are so sensitive to beliefs, and that beliefs are so sensitive to reasoning, seems to raise the same dilemma I’ve been harping on for a while now: either it renders Haidt’s theory fairly trivial and without serious implications for moral philosophy or it renders it seriously defective.


            • [Hume] explicitly denies that passions can be reasonable or unreasonable

              Nonsense. See Treatise, II.iii.3.6. A very famous passage, not easy to miss. The sense in which he says passions can be unreasonable is just the same as you indicate for “cognitivists”: that the beliefs on which the passion is founded are unreasonable.

              By your relaxed definition of cognitivism, I doubt you could find a noncognitivist anywhere. Haidt is a noncognitivist in a perfectly recognizable sense, and I don’t propose to change my way of talking or pretend there’s anything nonstandard about it.

              either it renders Haidt’s theory fairly trivial and without serious implications for moral philosophy or it renders it seriously defective.

              Actually, you have yet to make any point at all that seriously cuts against the plausibility of Haidt’s view or its distinctiveness.

              Your unreasoning hostility to this project is getting to be a bit of a drag, David.


              • I don’t particularly want to play the role of referee in discussions here, especially if I haven’t been participating in one, but I’ve read enough of this one to reject the characterization of Riesbeck’s posts as “unreasoning” or “hostile.” Like Riesbeck, I so far haven’t seen anything in defense of Haidt that suggests that his views are overwhelmingly plausible, interesting, or distinctive. The defenses of Haidt seem to me to oscillate between saying that Haidt’s project is distinctive and interesting in its current form, and saying that there are distinctive insights in Haidt that are worthy of further development. The one claim is a long way from the other.

                I don’t think any one passage from Hume’s Treatise, famous, obscure, or otherwise, is going to resolve, once and for all, where Hume stands on cognitivism. In the same section of the Treatise as the one you cite, Hume tells us that the only work that reason does is to obey and serve the passions. So even granting your interpretation of the unreasonability of Humean passions, the resulting view is neither obviously a case of cognitivism, nor particularly clear. On Hume’s view, a passion is unreasonable if the belief on which it is “founded” (or “based on”) is unreasonable. But the same account tells us that beliefs “serve and obey” passions in the first place, and makes a deliberate effort to blur the distinction between reason and passion. I don’t think it takes “unreasoning hostility” to wonder what all of that is supposed to mean or amount to, whether in Hume or in Haidt.

                The Treatise is a book book, and Treatise II.3.3.6 is a long text that says a lot of things that don’t easily cohere with anyone’s conception of cognitivism. That’s why Hume has traditionally been regarded as a non-cognitivist. Maybe he’s a cognitivist after all, but one citation to one paragraph in the Treatise isn’t going to clinch the case.

                Liked by 1 person

          • Nonsense?

            “Reason is the discovery of truth or falshood. Truth or falshood consists in an agreement or disagreement either to the real relations of ideas, or to real existence and matter of fact. Whatever, therefore, is not susceptible of this agreement or disagreement, is incapable of being true or false, and can never be an object of our reason. Now it is evident our passions, volitions, and actions, are not susceptible of any such agreement or disagreement; being original facts and realities, compleat in themselves, and implying no reference to other passions, volitions, and actions. It is impossible, therefore, they can be pronounced either true or false, and be either contrary or conformable to reason.” (Treatise III.1.1).

            As for my use of the terms ‘cognitivism’ and ‘non-cognitivism,’ they aren’t relaxed and they aren’t non-standard, unless everything I’ve read on the topic is non-standard. Here’s some very basic stuff for you:


            “The cognitive theories contend that the early part of the emotion process includes the manipulation of information and so should be understood as a cognitive process. This is in contrast to theories that state that the generation of the emotion response is a direct and automatic result of perceiving the stimulus—these non-cognitive theories are discussed below…Non-cognitive theories are those that defend the claim that judgments or appraisals are not part of the emotion process. Hence, the disagreement between the cognitive and the non-cognitive positions primarily entails the early part of the emotion process. The concern is what intervenes between the perception of a stimulus and the emotion response. The non-cognitive position is that the emotion response directly follows the perception of a relevant stimulus. Thus, instead of any sort of evaluation or judgment about the stimulus, the early part of the emotion process is thought to be reflex-like.” You will find numerous examples of psychologists and philosophers holding each kind of view.


            “Non-cognitivism is a variety of irrealism about ethics with a number of influential variants. Non-cognitivists agree with error theorists that there are no moral properties or moral facts. But rather than thinking that this makes moral statements false, non-cognitivists claim that moral statements are not in the business of predicating properties or making statements which could be true or false in any substantial sense. Roughly put, non-cognitivists think that moral statements have no substantial truth conditions. Furthermore, according to non-cognitivists, when people utter moral sentences they are not typically expressing states of mind which are beliefs or which are cognitive in the way that beliefs are. Rather they are expressing non-cognitive attitudes more similar to desires, approval or disapproval.

            Cognitivism is the denial of non-cognitivism. Thus it holds that moral statements do express beliefs and that they are apt for truth and falsity. But cognitivism need not be a species of realism since a cognitivist can be an error theorist and think all moral statements false. Still, moral realists are cognitivists insofar as they think moral statements are apt for robust truth and falsity and that many of them are in fact true.” Again, plenty of examples of philosophers holding each sort of view.

            The bibliography makes plain that we will not in fact be hard pressed to find people who reject cognitivism as I have used that term, either as a theory about emotions or as a theory about moral judgments. Perhaps you are mistaking rejecting cognitivism with acknowledging and trying to account for the phenomena that cognitivists rely on in arguing for cognitivism; that seems like a silly mistake, but so does not knowing how the terms ‘cognitivism’ and ‘non-cognitivism’ are standardly used and accusing someone who uses them that way of using them ‘loosely.’ Meanwhile, your own use of the terms remains entirely murky, and from what you’ve written I honestly cannot tell what you think they mean. Occasionally it seems as though you might even think that a theory counts as non-cognitive simply because it holds that the relevant phenomenon (emotion, moral judgment) does not get at truth. But that is not at all what cognitivist theories are about, as a cursory reading of the introductory treatments I have just cited makes clear.

            You say that my “unreasoning hostility” is “getting to be kind of a drag.” I have been honest and tried to be courteous in setting forth reasons why I am confused about what the theory says and skeptical of its truth or significance on certain interpretations of what it says. Your responses have not satisfied me because I sincerely cannot see how they adequately address my questions and objections, and in exchange for that I get insulted? If thinking of my posts as “unreasoning hostility” helps you feel better about failing to give a clear and consistent account that dispels my confusions and shows my worries to be misplaced, then feel free to give your rider free rein. But spare me the insult and condescension.


            • Nonsense?

              Yes, and quoting a different passage of Hume from the one I cited, that does not contradict the one I cited, will not make the one I cited go away.

              As for my use of the terms ‘cognitivism’ and ‘non-cognitivism,’ they aren’t relaxed and they aren’t non-standard,

              Well, let’s see. You think taste buds might be cognitive. You think that if beliefs “or belief-like appraisals” play any role at all in producing an emotion, it counts as cognitive. You think the only way emotions could be noncognitive is if they are beyond rational criticism, like taste buds (apparently forgetting your earlier remark).

              By this standard, none of the writers described as noncognitivist in the IEP article you linked except Prinz can be counted as noncognitivist, and neither can Hume. In that article, it seems that writers are considered to be noncognitivist if they believe that the cognitive evaluation (of there being a threat in the case of fear, for example) is automatic or modular (in Fodor’s sense), as opposed to conscious. Even the claim that the cognitive evaluation in emotion is automatic is restricted by nearly all these authors to only certain emotions. The one exception to this, Robinson, hedges her theory—which is necessary, of course, if it is to be at all plausible—with the claim that “cognitively complex information” (beliefs, basically) can feed into the automatic emotion-generating module. So all these people are really cognitivists by your standards (and they are partially cognitivists by the IEP author’s standards).

              Prinz’s case is interesting. He really seems to think that certain emotions can be generated directly by certain stimuli. A coiled shape or a slithering motion in the grass produces fear as a direct response. It’s by no means implausible for certain cases. I don’t know how far he pushes it. (James, discussed with Prinz, is not a noncognitivist by the IEP definition—or at least not a thorough one—though you can’t tell it from the article. See James, Psychology: The Briefer Course, p. 243.)

              So we’ve found a single author who can be considered noncognitivist to some extent according to your way of talking. I would call that nonstandard. And I would say that my previous remark,

              It is a commonplace that emotions depend on our cognitive recognition of how we take reality to be. Whether you take the object in the dim corner of the tent to be coiled rope or a snake affects your emotional reaction. I would be astonished if Hume and nearly everybody else who has thought seriously about the emotions wouldn’t agree.

              has been vindicated.

              But I don’t really understand how we got to be talking about emotional noncognitivism at all. You are the one who introduced the topic. If Haidt takes any particular stand on this question, I don’t know about it. I also don’t know what difference it makes. When I said, in response to your introduction of the topic, that Haidt seems to be a noncogntivist “in the way I’m discussing,” I was alluding to what I said immediately after the remark just quoted, namely, “But this doesn’t make the evaluative part of the emotion cognitive. It doesn’t mean that the emotion itself is a belief.” I admit that this is pretty unclear, and I’m not sure I know how to be any clearer now. I meant the approval/disapproval, approach/avoidance aspect of the emotion. I meant that this evaluative aspect is noncognitive. This is a substantive claim. If you judge something to be good or bad, I would take that to be cognitive. I think that Haidt thinks emotional approval/disapproval is not cognitive. It’s a feeling, on the basis of which we might later form cognitive judgments. For instance, when we judge something that we emotionally disapprove of to be bad. That is all I meant to say.

              Concerning moral noncognitivism, I can see now that if the real core of noncognitivism is the claim that moral statements have no truth conditions, then it is confusing to call Haidt’s theory, even of ordinary moral judgment, noncognitivist. (I can’t forbear adding, though, that my descriptions of Haidt’s theory have been tolerably clear, so that what is being claimed and what is not being claimed shouldn’t be that hard to figure out by a smart person who actually desires to know.) Ordinary moral judgments are still judgments. Haidt doesn’t claim they have no truth-values. Haidt does not claim that there is no moral reality. The claim is only that ordinary moral judgments are not formed by cognizing moral reality. (And this is the specific claim that I took Michael to be denying in the comment to which I responded by quoting Haidt’s quote of Hume.) Rather, they are formed by various tricks of the elephant—by emotions and intuitions that are not formed by perceiving moral facts. Just as taste buds (the noncognitive kind) do not detect nutritional value but only a certain large class of chemical stimuli, so moral emotions and intuitions do not detect moral reality but only respond to certain stimuli. There is a certain range of stimuli that prompt the emotion of compassion, which urges us to take certain forms of action and make certain types of moral judgment. We have apparently been shaped by evolution to feel compassion at the sight of suffering, especially the suffering of children. But that is all there is to it. Compassion does not result from the perception of moral truths.

              Seriously, is this what you find so hard to understand? Is this what is “puzzling” to you and has you “confused”?

              You do not write like somebody who is puzzled. People who are puzzled and confused ask questions. They offer suggestions and alternatives about what might be meant, in order to help the conversation along and promote understanding. They seek to contrast the new view with views they already know in order to get clearer about it. (That is, they positively look for contrasts; they don’t seek to prove that no contrasts exist.) But I do not see these things from you. What I see is a torrent argument, parries, and thrusts attempting to show that Haidt’s views are all either trivial or wildly implausible. All you do is argue. You do not seem to be seeking to know. My impression is that your main concern is to justify your pre-existing aversion to Haidt and probably to the whole program of moral psychology. This is what I meant by “unreasoning hostility to this project,” and I said it because I think it is true and it is interfering with our ability to have any sort of fruitful discussion. I don’t see how else to explain your pattern of behavior. It’s unfortunate. But as long as it goes on, discussion with you is just not very worthwhile. Because as Haidt says—and this is doubly true of you, because you are quite clever—reason (the rider) will generally be capable of proving whatever it wants. And what is the use of that?


          • Well, let’s see. I said that Hume explicitly denies that passions are reasonable or unreasonable. You claimed that was “nonsense” and cited a passage in which he allows for us to make some sense of the ordinary way of talking about them as reasonable or unreasonable. I then cited a passage in which he does in fact explicitly say what I claimed that he explicitly said. You respond by reaffirming that my claim is nonsense. Given that I also originally noted that the interpretation of Hume’s considered view on the question is a matter of dispute and that I was not committed to its being the right one, I fail to see why you think that you have said anything to undermine my point. We might note that the passage you cite seems perfectly consistent with the one I’ve cited because the sense we can make of calling passions reasonable or unreasonable is not anything like what Hume regards as a strict sense. But I don’t want to bicker about interpretation, since, as with Kant and Rawls, it’s clear that you and I aren’t going to get anywhere on that topic, if on any.

            I’ll try to clarify myself one last time, though I have to say that you hardly give ample evidence of being a careful or charitable reader. But since you obviously think that you’ve been clear on things that I find anything but, I’ll try to give you the courtesy that I wish you have been giving me. After all, the likelihood of miscommunication in a medium like this is not exactly small, particularly when the discussion revolves around complex and potentially ambiguous matters.

            I raised the question about whether tastebuds might be cognitive in an effort to get clearer on what we are and aren’t saying in calling something cognitive or non-cognitive. There seems to be a sense in which they are or can become capable of fine-grained discriminations of real differences in the world. If that is enough to make them cognitive, then even assuming a very tight analogy between tastebuds and moral attitudes, might not moral attitudes be cognitive as well? I did not intend to endorse the claim that they are cognitive (in the sense I have been using the term, they aren’t plausibly so regarded), but to try to get clearer on what you, Haidt, or anyone else means by the term. So my later comments do not undermine my former, and I did not forget them.

            I did not intend to suggest that the only plausible sense in which emotions can be subject to any kind of rational criticism is if they are cognitive. What I said was “what is central [to cognitivism about the emotions] is rather that beliefs or belief-like appraisals are essential to emotions, either as components of the emotion or as part of the process whereby one comes to have an emotion. This is why they are subject to rational criticism on the grounds that the propositional attitudes that either (partially) constitute them or generate them are false or otherwise unreasonable.” That is to say, cognitivism is supposed to be better able to explain the way in which emotions are subject to rational criticism, viz., criticism on the grounds that the emotion itself is either (partially) constituted by a false propositional attitude or is an expression of a false propositional attitude (to be more precise, I should have qualified the bit about ‘propositional’ attitudes by adding that many cognitive theories do not require that the cognitive content be propositional, only that it be propositionally formulable, even if the subjects — infants, animals — cannot formulate such propositions). So no, by this standard none of the authors regarded as non-cognitivist in the IEP entry, nor Hume on the common interpretation of him that I’ve mentioned, counts as a cognitivist. They all allow that cognitive processes can have some kind of impact on our emotions; what they deny is that any cognitive process or any propositional or quasi-propositional attitude is either an essential component of an emotion or a necessary part of the process that generates an emotion. You’re of course right that most of these theories recognize that some emotions can be cognitive in the strict sense, and perhaps even that some kinds of emotion do require cognitive input, but they still deny that anything cognitive is essential to emotion or everything that many people want to call emotion (Griffiths’ view that ’emotion’ is not a natural kind seems not implausible). This is, I take it, a non-trivial difference.

            It is not, of course, that the non-cognitivists cannot make sense of claims like, “your fear of that object in the corner is really unreasonable, because it’s actually just a bundled rope, not a snake” or “your fear of that object in the corner is really quite reasonable, because it is a python.” But just consider how differently Nussbaum will understand the sense in which the fear is reasonable or unreasonable. Nussbaum will want to say that my fear is reasonable or unreasonable because my fear is a true or false judgment. Less radical cognitivists will deny that my fear is a judgment, but nonetheless hold that a judgment or belief is a proper part of it or at least that my fear is an expression of a judgment or belief. It is not just that in this case the thing I’m afraid of happens to be such that my fear is fitting or not, but that my fear or a crucial aspect of it has truth conditions that are or are not met.

            I brought cognitivism about the emotions into the discussion for two reasons. First, because as I understood Haidt’s theory, it seemed as though the idea was supposed to be roughly that moral judgments are non-cognitive because they are fundamentally products of emotion rather than reason; second, because Michael suggested in his first post on Haidt that Haidt takes the emotions to have an important cognitive aspect (https://irfankhawajaphilosopher.com/2016/06/30/haidt-the-righteous-mind-chs-1-2/). Now, I hope it is tolerably clear how if Haidt’s view really is that (1) moral judgments are non-cognitive because ultimately grounded in emotion, but (2) emotions are essentially either judgments, partially constituted by judgments, or expressions of judgments (or whatever else we prefer to call the propositional or quasi-propositional attitudes involved), then there is at very least a tension in the theory. For (3) if to say that moral judgments are non-cognitive is to say that they are not even truth-apt, but (4) the judgments essential to emotions are propositional or quasi-propositional attitudes, then not only does (1) not follow, it is in fact inconsistent with (2)-(4). I have never supposed that the matter was so clear cut; but surely you can see what the apparent problem is and why I raised it. If it isn’t this clear cut, the thought went, then I must be misunderstanding one or another part of the theory.

            Unfortunately, I still do not know for sure which part I am misunderstanding, but I take it that at least two important points are: (a) Haidt does not clearly think emotions are cognitive in the standard sense of essentially involving truth-apt quasi-propositional attitudes, (b) Haidt’s view is in fact not that moral judgments are not truth-apt, just that, given how they work, we have no reason to think that they hit upon truth except perhaps by sheer accident. An additional wrinkle is that perhaps emotions have nothing to do with it, and moral intuitions can be direct, non-cognitive feelings in response to certain stimuli. Is that roughly right?

            If it is roughly right, I don’t find the theory very plausible. I am highly sympathetic to the view that ’emotion’ is not a natural kind and that some of what we typically class as emotions are essentially cognitive and others are at least sometimes not. But it seems to me that enough of them are cognitive often enough, and that the judgments they embody or express are often enough unproblematically true, that we have ample reason to think that most human beings have or can readily get a basic cognitive grip on value, on what is basically good or bad for us, and that we are capable of reasoning critically on this basis and thereby arriving at true or at least more adequate beliefs about what we should and should not do. I don’t offer this statement up as an argument against Haidt or anyone else who doubts it; doing that would be a makros logos. I offer it up in the hopes that you will see why I have so far heard nothing about Haidt’s theory that strikes me as casting serious doubt on what I’ve just said. What I have just said is perfectly consistent with a great deal of moral discourse being primarily the rationalization of unreasoned emotional or quasi-emotional reactions and with thinking critically about goodness, justice, and virtue being an awfully difficult thing in which we are often led astray by our prejudices and passions. I don’t deny any of that.

            But neither have most moral philosophers in the Western tradition. This has been the second aspect of this discussion that I have found puzzling. The dialectic seems to have gone this way: A. I say that on a certain construal of what Haidt is saying, the theory seems problematic or implausible; B. You point out that the theory does not have those implications and is in fact entirely consistent with what I’ve just said; C. I then wonder how the theory is supposed to have non-trivial implications for moral philosophy if it is consistent with what virtually every major moral philosopher has thought and differs only in emphasis; D. You respond that it isn’t so consistent; E. I then wonder how, on that construal of the theory, it is supposed to be especially plausible, and we’re back at square one. From my point of view, your exposition of Haidt’s theory has been a constantly moving target: it has important implications until I express some skepticism about it, at which point you effectively deny that it has those implications.

            You seem to be perfectly satisfied that you have been more than adequately clear about all this, and I do not intend to suggest that I have not misunderstood you at various points; you have certainly misunderstood me. As I said, the opportunities for misunderstanding in a medium like this are many. But you have not been remotely clear on these issues; you have instead continually used key terms in ambiguous ways in defiance of their standard usage in the fields of your own supposed professional competence, have ignored many of my key points, have attributed claims to me that I did not make, and have interpreted what I’ve written far less charitably than you might. And now you’ve taken to dismissing what I say as an expression of an irrational hostility toward Haidt, an accusation you reiterate in your latest post. You say that people who are trying to understand ask questions and suggest alternatives; I have done plenty of that. You suggest that people who are trying to understand do not argue; I’d have thought it would be obvious to a philosopher that one way to gain understanding of a view is to try to argue with it! I have tried to be nothing but good spirited in this discussion, and we’ve reached the point where you’re psychologizing me and accusing me not of missing the point or even of being stupid or misinformed, but of being dishonest and merely contentious.

            Well, you’re not a very good psychologist, I’m afraid, and in any case I will tolerate no more of that from you. I’m not sure I can say much more than I’ve already said apart from potentially clarifying what I’ve already said, so I don’t think any further progress is really possible here. But if you want to have a go and you can play like a big boy instead of attacking my character when you don’t successfully persuade me, then have a go.


        • Just replying to David P’s point about basic moral affect/motivation/desire (“the moral taste buds”) – I’ll wade into the discussion between the Davids latter if I have time.

          (a) I was addressing noncognitivism about moral judgment and did not even consider whether cognitivism regarding moral judgment (or normative judgement generally) requires perception/intuition states that are themselves cognitive. This, and the question of whether and in what sense such basic affective/motivational/desire states are or can be cognitive, are interesting, but I’m not sure I have much of a firm opinion. They might be cognitive, but in a sense that is more general and different from the sense in which sensory perception is cognitive.

          (b) On a standard, plain-Jane Humean view, the moral taste buds theory provides basic, objective normative standards: as a primitive matter, we have (objective, normative) reason to eat sweet things and take care of other human beings (if they are cute enough – i.e., if they have the right features, features similar enough to some features of infants). Even though Haidt’s focus is descriptive, he seems to endorse this Humean idea. How we come to grasp that we have such reasons is another matter (and not, I think, something that Haidt registers an opinion on at all). If simple, plain-Jane Humeanism (the desire-based theory of normativity) is true – and perhaps as well if some more sophisticated Humean view is true – it does seem hard to resist the idea that our basic affective/motivational/desire responses do not also function, along side something like a basic concept of a reason for action, to ground judgments to the effect that one has reason to exhibit the behavioral responses thus motivated (eating sweet things, taking care of human-infant-like cute beings).


          • Regarding the substance of the exchange between the Davids:

            (i) I thought the original issue was whether Haidt endorses or is committed to metaethical noncognitivism. I think we are agreed that he is not?

            (ii) I’m not sure why or how we ended up discussing ‘cognitivism’ and ‘noncognitivism’ as these terms apply to emotions, affective and motivational responses, etc. As far as I can tell, Haidt is a moral realist, and is so based on a desire-based or Humean view of how we get normativity or reasons for action in the world, but does not say how or relate this view to much of what he says. The sense and extent to which things like disgust emotions and responses are cognitive seems to me to be a topic in moral epistemology and I don’t think Haidt really addresses this topic any more than he addresses the detailed normative theory that would take him from a Humean theory of normativity to the sort of agent-neutral consequentialist type of general ethical theory that he endorses. He knows that he is not a philosopher.

            (iii) David R: without getting into the details, I view both Haidt’s summary of the state of the art in descriptive moral psychology (and cognitive psychology and other related areas) and his specific “moral foundations” descriptive hypothesis (and its endorsed evaluative or normative upshot that we appropriately evaluate moral practices as good or bad approximately as we appropriately evaluate cuisine as good or bad) as providing *plausible, powerful causal/functional/motivational explanations* of why it is that much of our reasoning is emotion-driven rationalization and why it is so hard to reason in an unbiased evidence-sensitive way that is effective. If this is right, we’ll have a better idea than we had before of what to do about or with the parts of human nature that make these things so. For knowing these things would enable us to train ourselves to have specific values and follow specific rules that are more sensitive to how our reason-giving mechanisms actually work. If some version of Humeanism is correct, then the motivational and functional information in such theories will be even more relevant to what values and norms we should have. That’s all pretty general, but here is an example: if, as Haidt supposes, the easiest and most reliable way to test hypotheses or opinions with disconfirming evidence (at least on ethical and political and religious topics) is via friendly arguments between mutually trusting and respectful elephants, this is a strong explanatory-type reason in favor of establishing trusting, respectful political communities for the sake of facilitating robust disagreement. This seems to speak not only to how to achieve what is valuable (or to what we antecedently intrinsically desire or value), but also to what items to intrinsically desires or value – to what our ideals should be or what specific form they should take.


          • Sorry I missed your response when you posted it. Better late than never, I suppose.

            Your characterization of Haidt as (a) not a non-cognitivist about moral judgment and (b) a moral realist seem to me to conflict with what I thought DP was saying about the matter. I thought the idea was supposed to be that because of the way moral intuitions and reasoning work, we have no good reason to think that they get at some moral reality. But if anything is clear at this point, it’s that it’s not clear to me what Haidt is saying. I think at this point, given how poorly my attempts to enter into the discussion have gone, the better thing would be for me to allow you and DP and perhaps others who have actually read the book to discuss it. I don’t think there’s generally anything wrong with trying to participate in a discussion about a book I haven’t read, but in this case it’s created problems that I can’t overcome without reading the book, which, alas, I will not be able to do anytime soon because I woke up this morning and realized I only have two weeks before my classes start. Yikes.

            One general thought, though, is that I wouldn’t ordinarily count the sort of view you attribute to Haidt — plain-Jane Humeanism — as a form of moral realism. If things are good, right, etc. only because they contribute to the satisfaction of our desires, where even the fact that desire-satisfaction is what ultimately makes things good, right, reasonable, etc. is itself true by virtue of our desires, then there doesn’t seem to me to be enough objectivity involved to call the view a realist one. But that’s perhaps just a pure matter of convention. After all, on the sort of view you’re sketching, claims about what is good, right, or reasonable can be straightforwardly true; it’s just that the truth-makers are contingent facts about our desires and nothing less subjective than that. So it’s not mere non-sense to describe the view as a realist one, even if I’d prefer to reserve that term for views that endorse some sort of objective conception of the good or the (‘morally’) reasonable. I don’t quite think that classificational issues are a sheer waste of time, but provided we all know what each other means I don’t see any reason to insist on one classificational scheme or another, at least in this case. What’s noteworthy, though, is that while the view you describe can be called realist at least in a certain sense, some of the other views that have been associated with Hume cannot. Surely non-cognitivism (whether Hume was a non-cognitivist or not), on which claims about what is good or right cannot be true or false, cannot be realist in any sense; but neither can any form of so-called ‘error’ theory according to which such claims can be true or false, but they’re all false. Perhaps the most famous error theorist of all, John Mackie, held a view of rationality very much like the plain-Jane Humeanism you describe, and in fact appealed to it in defense of the error theory. Again, the classifications ultimately don’t matter, but the example of Mackie suggests that you might want to consider whether the sort of view you describe is really best described as a realist one.

            The way I tend to think of these things, plain-Jane Humeanism is not realist, and if it were the true theory of practical rationality then moral realism would be false. But I incline towards strongly anti-Humean views on these things: I don’t think the fact that I desire to phi is ever sufficient to give me a good reason to phi (it gives me reason either-to-phi-or-rid-myself-of-the-desire-to-phi, and which one I should do depends on further facts, not all of them desires), I think well-being is objective and that nothing that is an aspect of well-being is so because someone desires it, and so on. That doesn’t mean you can’t call your plain-Jane Humean a realist, but it might help to explain why I find that a counter-intuitive label.

            But enough of this for now. I’ve got classes to prepare!


          • Michael,

            (a) I was addressing noncognitivism about moral judgment and did not even consider whether cognitivism regarding moral judgment (or normative judgement generally) requires perception/intuition states that are themselves cognitive. This, and the question of whether and in what sense such basic affective/motivational/desire states are or can be cognitive, are interesting, but I’m not sure I have much of a firm opinion. They might be cognitive, but in a sense that is more general and different from the sense in which sensory perception is cognitive.

            I get the impression that you are concerned with nailing down whether Haidt is a cognitivist or a noncognitivist on the big question of whether there is an ultimate moral reality that we can discover by reason. Is that right? If that is the question, then my view is that Haidt has not committed himself to any answer either way. I wouldn’t be surprised if his attitude was that, as a psychologist, it’s not his place to say. Also, I would bet money that, if he does believe there is a moral reality, he has only the vaguest notions as to its nature. You have read more of the book, so maybe he will say more before he’s done, but so far of course he insists that he is only presenting a descriptive theory of ordinary moral judgment, and he has drawn no normative conclusions from that theory, much less any conclusions about moral reality. And it’s hard to see how it could be otherwise, since his descriptive theory says that people do not form their ordinary moral judgments by cognizing moral reality. Are you meaning to deny this?

            The one place I have found where he seems to indicate how his theory might be combined with some sort of objectively based claims about the good is in the passage I quoted from “The Moral Mind,” (co-authored with Craig Joseph). And really the only grounds for interpreting the passage the way I would like to is its statement that “what it means for a personality characteristic to be a virtue, and not simply a behavioral regularity, is largely that it consists in functioning well in a specific ‘sphere of existence.’ And what Aristotle and Nussbaum mean by ‘sphere of existence’ is similar to what evolutionary biologists would recognize as persistent adaptive challenges and other types of environmental constraint.” This is tantalizing, but obviously it’s pretty vague. It seems to suggest that we will find the true human good in good functioning, but that’s about all. He could easily think that the various moral systems which cultures have evolved, including systems of hierarchy, authority, and spiritual purity, are adaptive in the sense of promoting effective social functioning, and in that sense they are hooked into the human good. Of course, there may be as many different such moral systems as there are cultures, and within limits Haidt might think all are about equally so hooked. If this is his view—which seems entirely likely—would you consider him to be a cognitivist? It seems to me that the answer to this ought to be yes. But I don’t know if it matches the way you are thinking about him.

            (b) On a standard, plain-Jane Humean view, the moral taste buds theory provides basic, objective normative standards: as a primitive matter, we have (objective, normative) reason to eat sweet things and take care of other human beings (if they are cute enough – i.e., if they have the right features, features similar enough to some features of infants). Even though Haidt’s focus is descriptive, he seems to endorse this Humean idea.

            It would be very difficult for me to make out what you’re saying here if you hadn’t later characterized “plain-Jane Humeanism” as desire-based normativity. My own view, which I hold pretty strongly, is that desires are motives, not normative reasons. Therefore, desires per se cannot supply ultimate normative reasons for anything. The only thing that can supply a normative reason is, well, a norm. We never have a normative reason to do something until we have shown why it is good or right. (Did Hume really think otherwise? I would interpret him as simply not believing in normative reasons.)

            But that’s just me. As far as Haidt is concerned, note that the intuitions he’s talking about are normative. We have moral intuitions about what is good, bad, right, wrong. I don’t think desires have anything much to do with his theory of ordinary moral judgment. I definitely don’t think he takes our moral intuitions as giving us objective, normative reasons for action. The whole idea of his theory (as with Hume’s) is to separate the intuitive/emotional basis of ordinary moral judgments from reason, not to pretend that what is irrational is somehow rational after all. There certainly can be no simple move from “I have the moral intuition that X is right” to “X is right” or “I ought to do X.” After all, for Haidt, X includes things like beheading apostates and committing genocide. Nor does Haidt give any indication I’ve seen that he thinks we can find the objectively true morality by massaging our moral intuitions into reflective equilibrium. It would be astonishing if he thought anything like this, given the indirect, haphazard, irrational sources of our moral intuitions, according to him. Rather, the story is going to be (I think) something like what I quoted him (and Joseph) above as saying.


  3. There seems to me to be a tension here between the claim that Haidt’s theory is purely descriptive and the notion that it has important implications for ethics. On the one hand, as purely descriptive, the theory is not supposed to tell us anything about what is good, right, etc. On the other, it is supposed to be able to undermine, or at least seriously challenge, common claims about the cognitive character and potential objectivity of moral judgment, and with it the truth of certain moral judgments. There is of course no contradiction here; the theory can consistently refrain from making any claims about what is good, right, etc. while also entailing that certain claims about what is good, right, etc. are false or even non-truth-apt. But a theory need not make positive moral claims in order to be morally revisionist, and insofar as it is morally revisionist, it has a normative dimension and is therefore not purely descriptive. This is a general problem for non-cognitivist theories that profess to be non-revisionary. But in Haidt’s case I’m confused about whether or not it is supposed to be revisionary and whether it is supposed to be genuinely non-cognitivist, and in any case I don’t see how it could have some of the implications you suggest.

    The comparison with Nietzsche’s Genealogy of Morals certainly suggests a revisionist intent. Occasional appearances notwithstanding, Nietzsche does not commit himself there to any positive claims about what is good or right. But I strongly doubt that he would regard his theory as purely descriptive and having no definite normative implications; quite the contrary, he seems to take it as entailing that all moral claims, at least of the slave morality variety, are false (I take no position on whether Nietzsche endorses any objective positive normative claims, as there seems to me to be evidence on both sides). Haidt is plainly not quite so dramatic or ambitious as Nietzsche, but if the comparison is apt then there is a straightforward sense in which he is doing ethics as well as psychology. I’m all for doing ethics and psychology together, so that’s not an objection. But if true, it would mean that he can’t evade philosophical questions about ethics by disclaiming any intention to do ethics or philosophy. Yet it is just those disclaimers that seem to suggest occasionally that the comparison with Nietzsche is not at all apt. At various points the impression seems to be that the theory not only refrains from positive claims about what is good or right, but has no implications whatsoever about what is good or right; at most it tells us what kinds of psychological constraints we will have to recognize in identifying any moral theory that people can regularly put into practice. But if that‘s the idea, then the theory seems to have no bearing on ethical theory as such; it tells us only about the limits of the widespread application of whatever the true moral theory is. We can easily see how the psychology, when coupled with independent philosophical views about the nature and content of morality, could have far-reaching normative implications. But those would not be implications of the theory itself.

    But I really can’t tell if that’s the idea or not. I’m particularly puzzled about whether Haidt’s theory is supposed to be non-cognitivist. You’ve described it that way several times, and that seems to be the point of the rhetorical invocation of Hume, but Michael’s earlier post reported that Haidt’s theory of the emotions is explicitly cognitivist, and certainly much of what he talks about under the rubric of the five moral foundations seems to fit a cognitive theory better than a non-cognitive one. One of the fundamental points of standard non-cognitive theories of moral judgment is that moral judgments, being non-cognitive, cannot be true or false. The analogy with taste is usually offered to illustrate this point. When you say “brussel sprouts are good!” and I say “no, they’re terrible!” we are, contrary to the surface grammar, not attributing an objective property, intrinsic or relational, to brussel sprouts, but merely expressing our positive or negative feelings about brussel sprouts. Hence our gustatory discourse is not truth-apt at all. If all moral judgments are in fact non-cognitive in this way, then there is no moral truth, and not even any moral falsity, strictly speaking. But such views are dramatically revisionist and cannot but entail that most moral theories as well as conventional moral judgments are not true. If this is how Haidt’s theory is supposed to go, then it is no good to fall back on the claim that it is merely descriptive or that it is a piece of psychology and not of philosophical ethics; if a non-cognitivist theory of moral judgment is true, then it does in fact follow that there is no Kantian moral law, despite the fact that the psychological theory makes no first-order normative claims. It would also follow that there is in fact no Aristotelian objective flourishing, and no objective Millian principle of utility. It would not follow that there can be no true claims about what we have reason to do, and in principle those claims might line up with what a Kantian, a utilitarian, or an Aristotelian would say. But given the malleability and variety of human preferences, that is only a logical possibility.

    I think the prospects for general non-cognitivism are pretty bleak, but as I say, it’s not clear to me that Haidt really is a non-cognitivist in the relevant sense. But if he is not, then I don’t see how his theory can have any of the implications you suggest. Again, even your contrast with Kant would fail. Kant does not believe that most people’s judgments are based on discernment of the moral law; he thinks most people’s ordinary judgments are based on inclination, and he agrees with Hume that inclination is, at bottom, non-cognitive. So if Haidt is claiming not that we can’t discern moral truths via reason (whether because there aren’t any such truths or because we are simply not capable of knowing them), but simply that most people’s moral judgments are not in fact based on such discernment, then the most that follows from his theory is an empirical disagreement with Kant about how often ordinary people form judgments based on discernment of moral truth via reason. On the non-non-cognitivist interpretation, Haidt’s theory creates even fewer problems for most other moral realists. Even the Christianized Aristotelian theory of natural law, which holds that all ordinary human beings have innate moral knowledge, does not confuse innate moral knowledge with whatever the deliverances of our passions happen to be; a major part of that tradition is a theory of the corruption of our moral judgment through the rationalization of our emotions. But most moral realists, whether Aristotelian or utilitarian, do not believe that we have innate moral knowledge, but simply that we can, through rational reflection, arrive at true moral judgments. So if Haidt’s theory isn’t non-cognitivist, then I fail to see how it has any impact whatsoever on ethical theory apart from the secondary question of how moral theory is to be put into practice. By calling it ‘secondary,’ I don’t mean to imply that it’s trivial, but only that it has no bearing on what moral claims or theories are true.

    I don’t even think Haidt’s theory can have the impact you suggest it should on methodology, regardless of whether it is genuinely non-cognitive. For one thing, reflective equilibrium is perhaps especially well suited to non-cognitive theories; if the main objection to coherentism is that coherence is too loosely related to truth, a theory that doesn’t regard fundamental moral judgments as true will hardly be bothered by that fact. The most prominent defenders of reflective equilibrium in ethics are anti-realists anyway (though not non-cognitivists about moral judgments); Rawls, for example, need not be troubled one bit if it turned out that all of our most basic moral intuitions are merely subjective, given that his theory is explicitly grounded in a subjectivist theory of the good and, especially in its later reiteration, makes no pretense to being true. But if Haidt’s theory is not non-cognitive, then it poses no problems for moral realists who defend the method of reflective equilibrium. Realist defenders of reflective equilibrium routinely insist that we need wide reflective equilibrium, for which the criteria are not merely formal but include substantive explanatory considerations. At most the coherentist realist will be wary of assuming that certain intuitions are not mere prejudices — but that is hardly a lesson that Haidt needs to teach us, as it is virtually a platitude among philosophers, however bad some of them are at recognizing when they are relying uncritically on intuitions; after all, the whole rationale for coherentist epistemologies is the notion that there are no self-evident truths, that every belief stands in need of justification or is at least open to challenge that calls for defense. There may be good reasons to find fault with the method of reflective equilibrium or with its widespread abuse, but Haidt’s theory, on either construal, simply has no bite against it so far as I can tell.

    So I’m still confused about just what the theory is supposed to be, but I can’t see why it should have any significant implications for ethics. If it’s genuinely non-cognitivist, it will need to address the formidable objections against non-cognitivism from philosophers and psychologists. If it’s not genuinely non-cognitivist, then it will leave debates in moral theory exactly as they are. I’ll let Michael clarify what he was trying to get at with his question about how the theory can get any normative purchase, but my initial assumption was that he wasn’t wondering how Haidt thinks it gets normative purchase, but how it possibly could — the assumption being that if it can’t, that’s a problem.


    • David,

      On the question of “moral revisionism,” I’m not sure what you mean by that phrase, but the general way in which Haidt’s theory has implications for moral theories is simple. Moral theories rely on premises about human nature. Whether those premises are true or false is a matter of descriptive fact. So a descriptive theory has implications for moral theory. And this is why, in the most general way, moral philosophers ought to be interested in psychology in general and in moral psychology in particular. As a gross and unrealistic example, if you have a moral theory that says the good consists in pleasure and a descriptive theory that says there is no such thing as pleasure, then if the latter is true, the former is false. Again, if you have a moral theory that says that moral truth is discerned by reason in a certain way and a descriptive theory that says reason doesn’t work that way, then if the descriptive theory is true, the moral theory is false.

      Now, Haidt’s specific psychological theory is a theory of ordinary moral judgments. It therefore has implications for moral theories that have something to say about how people acquire and deploy moral knowledge. Kant is one of these. His moral theory is wrapped up in a moral psychology of how the common man makes moral judgments and can become morally adept, which is also crucial to his account of why morality matters. He thinks that the moral law is fully available even to the most unassuming person and that in fact reason is constantly telling us and reminding us of the moral law, if only our inclinations didn’t draw us away from it. All of this is a part of Kant’s moral theory, and an important part. It is all false, if Haidt’s theory is correct. But, obviously, the categorical imperative might still really be the moral law.

      With regard to noncognitivism, it is only Haidt’s theory of ordinary moral judgments that could be noncognitivist, because that is the only theory he has. I have explained the way this is supposed to be so. Nonordinary moral judgments, on the other hand, might perfectly well be cognitivistic and true sometimes. Haidt seems to think there might well be some moral truth out there, but that isn’t his bailiwick.

      Re methodology, if there’s no problem, what is Railton so worried about? I did not find his examples at all convincing, by the way. In each case, he argues that the elephant is right after all. What an obedient lawyer! But in every case, his reasoning is quite thin. He also argues that the elephant is wise—much wiser than the elephant is according to the psychologists on whose research Railton draws.

      You say Rawls needn’t be bothered to discover that intuitions have certain genetic and cultural evolutionary origins and roots in an irrational elephant, roots and origins that don’t necessarily have anything to do with anyone’s good, but I doubt that. One point of reflective equilibrium is to achieve a theory that the subject can rationally believe in because it survives inquiry. It’s hard to see how this strategy survives the revelation that equilibrium is achieved only between irrationalities.


      • A theory or view is morally revisionist if it claims or implies that some important aspects of moral discourse and belief are false (or at least not true). As such, revisionism obviously comes in degrees, from the trivial to the momentous. A theory on which moral judgments are not even truth-apt would be momentously revisionary because it is part of ordinary moral belief and discourse that our moral judgments can be true and false. Nothing that you or Michael has said about Haidt’s theory seems to give us any reason to think that moral judgments — even ordinary moral judgments — can’t be true or false, and I’m still not clear about whether that is even part of Haidt’s view (you say that you’ve explained the way that the theory of ordinary moral judgments is noncognitivist, but I’m thick, so I still don’t understand exactly in what sense it is supposed to be noncognitivist, how that it is supposed to fit with his claim that emotions are cognitive, and so on; I’m just in the dark, perhaps unsurprisingly since I still haven’t read the book). Certainly none of the features you’ve emphasized seems to show that they cannot be true or false. Perceptual beliefs are also mostly formed via fast, automatic, intuitive, largely unconscious processes; they are also often affectively laden; they also vary depending on culture and training (the perceptual beliefs that a hunter-gatherer forms in the forest are not the beliefs I would form in the forest, nor are the perceptual beliefs that a car mechanic forms when he opens the hood of my car the beliefs I would form, let alone the beliefs that a hunter-gatherer would form). But perceptual beliefs can obviously be true or false. Nor does the emotional character of many moral judgments seem to tell in favor of their being non-cognitive, since, minimally, cognitive theories of the emotions are not obviously mistaken; more ambitiously, it is quite difficult to see how most of our emotions could fail to have a strong cognitive component, one that is entirely consistent with their not being formed on the basis of slow, deliberate, conscious reasoning but that is inconsistent with a genuinely non-cognitive analysis. The most I can see following from the core claims of the theory as I understand them is that we have no good reason to believe that ordinary moral judgments are reliably true, precisely because they are formed the way they are formed and because they vary so much by culture.

        If Haidt’s theory is supposed to apply only to ‘ordinary’ moral judgments but not to all of them, then he owes us some non-arbitrary criteria for distinguishing the ordinary from the non-ordinary. Perhaps he has some such criteria; if you or Michael have said what those are, I’ve missed it. There’s no reason why we should exclude a mixed cognitive/non-cognitive theory; perhaps disgust- and purity-based judgments are genuinely non-cognitive but harm/care judgments aren’t. I certainly find it plausible that many things that pass for moral judgments are in fact rationalized disgust reactions (though I also don’t believe that disgust is a non-cognitive phenomenon, but that’s just complicating things). I don’t find it plausible that judgments about benefit and harm or justice and fairness are non-cognitive, even though I believe that many people’s actual judgments of these matters are rationalizations designed to serve their pre-rational interests (some people’s judgments about the plausibility of evolutionary theory are rationalizations designed to serve their pre-rational interests; judgments about the plausibility of evolutionary theory are not non-cognitive, but as truth-apt as anything can be). But if we do have a mixed theory, we need to know what kinds of judgments are supposed to be non-cognitive and what sort aren’t. I can’t tell from what you’ve said, but the general drift seems to be that for Haidt the non-ordinary judgments would be very unusual. If this is supposed to mean that it is only rarely that someone makes a moral judgment that is truth-apt, then I, at least, will require a lot more evidence than I’ve seen so far to find that conclusion plausible, let alone true.

        I’ll leave Kant alone except to say that I’m still not convinced that Kant’s theory is false if Haidt’s is correct; that depends on how unusual non-ordinary moral judgments are supposed to be. Kant would certainly have to revise his assessment of how ordinary purely rational moral judgments are, but unless Haidt thinks that they are so rare as to be unworthy of attention, then I can’t see how anything Haidt says is inconsistent with Kant.

        On methodology, Railton worries about it because Railton is a moral realist. If moral judgments are non-cognitive or if intuitions cannot be cultivated so as to be sensitive to genuine reasons, then that would be a problem for Railton. I don”t read him as claiming that the elephant is right in every case; I read him as claiming that the elephant can be and often is genuinely reason-responsive. I’m not sure why you find his examples unconvincing; I take his discussion of the incest example and his variation on the Knobe case regarding intention to be pretty convincing provided that what they’re supposed to show is that the intuitions can be genuinely reason-responsive (though I also think the examples never showed what folks like Knobe and Haidt take them to show, given that people’s responses to the questions clearly differ in many cases depending on their reflection about the cases — not always reflection leading to what I would regard as an improvement, but surely enough to show that the tail can wag the dog).

        As for Rawls and other anti-realist proponents of the method of reflective equilibrium, I think all I can say here without getting excessively detailed is that you misunderstand Rawls. Have a look back at ToJ, especially Part III. Rawls defines a person’s good in terms of what it is most rational for her to want, where rationality plays a purely formal role in making our preferences consistent and empirically informed and directing means to ends. Rawls is a thorough-going subjectivist. In ToJ, he helps himself to some controversial psychological generalizations about what people want and enjoy, most prominently in his so-called Aristotelian Principle and in the claim that we all have a desire to express our rational nature. In his later work, he backs off from any such claims and is all too happy to acknowledge that this is just a historically and culturally contingent ideal of the person. Throughout his work, the survival of rational scrutiny is entirely a matter of formal coherence, empirical adequacy, and consistency between means and ends. All the substantive work is done by preferences, which he increasingly comes to see as culturally contingent. I don’t put this forward as a strength of Rawls’ view, just as a description of what his view is. If value is at the bottom just a matter of preferences, then reflective equilibrium is probably the right methodology. At the very least, the standard objection to it — that reflective equilibrium has no reliable connection to truth — would be simply irrelevant.


        • Haidt does not commit to noncognitivism about moral judgment. His central concern is moral reasoning. One of his central claims is that moral reasoning usually functions to help one make one’s way in a social world in which we hold each other to account. Though moral reasoning can function to discern reality (and though there is a moral reality to discern – leave it to the philosophers to say precisely what this is), it is hard for us to reason morally in this way. My suggestion/elaboration is that reason-giving, though deeply keyed to evidence, truth, and making good decisions, often functions – especially socially and especially in moral reasoning – to “justify oneself” or avoid the punishment-type consequences involved in being held to account by others (and oneself); surprisingly often, reasoning, especially moral reasoning, does not function to canvass evidence in a thorough, unbiased way.

          If this is right, then a lot of what we think of as unbiased reasoning, especially moral reasoning, really is not (at least not fully). It is, in large part, self-protective rationalization. Yet most people take the deepest functional nature of belief and reasoning (it being keyed to evidence, etc.) to be the controlling factor or what we are trying to do when we are expressing opinions and bringing reasoning to bear on them. This is such a natural, intuitive thing to believe! So the central claim here is pretty radically revisionary relative to what is ordinarily supposed. Also, depending on how deep-seated these self-protective motivations are, this kind of view may be pretty revisionary with regard to how confident self-proclaimed rationalists and sophisticated theorists should be in their belief that, however stupid and irrational humans can be, they are largely exempt (hence Haidt’s emphasis on the social process of mutually calmed or trusting elephants engaged in disagreement – mutual disconfirmation – in getting to the truth of matters).

          This kind of view need not impugn the value of unbiased reasoning, truth, etc. It does if the deep (even constitutive) way that belief and reasoning is tied to evidence is denied. And it does to the extent that individual, responsible people striving to be unbiased discoverers of important truths plays no role in the social or individual production of good reasoning and knowledge. But you don’t need to include either of these elements in the picture. Haidt does not emphasize them and, in this, can come off as sensationally allying himself with those people and ideas who relish – sometimes for perverse reasons – taking reason down a peg (or two or three). I don’t think this is his intent, but allowing this inference does help grab attention. He wants his book to sell.

          Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s