Wednesday, January 16, 2008

Rationalizing Emotions and The Moral Behavior of Kantians

I've recently been enjoying Joshua Greene's "The Secret Joke of Kant's Soul" (penultimate manuscript available here). Greene's research suggests that the moral judgments of Kantian deontologists (who focus on such things as rights, duties, and "respect for persons") tend largely to be rationalizations of evolutionarily-selected emotional responses, while the moral judgments of utilitarians and consequentialists (who focus on such things as maximizing the good of everyone) tend to be more rationally driven (or at least less driven by emotional "alarm systems"). The sorts of cases on which Kantians and consequentialists tend to disagree are cases where maximizing the good violates what we might perceive as someone's rights. Should you push someone in front of a runaway trolley, thereby killing him, if that's the only way to save five other innocent people? Should you smother your baby to death if that's the only way to prevent yourself, your baby, and several other people from being found and killed by Nazis? The Kantian impulse (with caveats and complications, of course) is to say no in such cases, the consequentialist to say yes.

Now if (a.) Greene is right about Kantianism as principally a post-hoc rationalization of evolutionarily selected emotions -- and needless to say it's very controversial! -- and if (b.) the apparently widespread view is correct that Kantians behave less well than consequentialists, and finally if (c.) emotional reactions tend to be more self-serving than do consequentialist principles, then, well, maybe (a) and (c) together explain (b).

Let me stress that I myself have no beef against Kantian deontology or Kantian deontologists and that conditions (a) and (b) are highly speculative. Finally, the only small bit of direct empirical evidence I have on the moral behavior of Kantians versus deontologists (the rate at which ethics books are missing from academic libraries) suggests that patrons are no more likely, and maybe even a bit less likely, to misappropriate Kantian than utilitarian texts.


Anonymous said...

The Nietzschean gloss on moral reasoning — i.e., it's a rationalization of one's will to power, healthy or sick, as the case may be — is a kind of rhetorical performance that's very difficult to coherently defend. Greene's "Secret Joke" is a case in point: We can simply reverse his argument by claiming that such things as maximizing the greatest good is a rationalization of the evolutionarily-selected instinct to promote the survival of the species. At any rate, incommensurability of values will remain a problem so long as we can evaluate reasons from more than one perspective: an agent-centered (personal) POV and an outcome-based (impersonal) POV.

Regards, Kevin

MT said...

I think a carefully spoken scientist would call the book losses "empirical data" not "evidence," which outside courtrooms usually implies "of something." Anyway, my hypothesis is the Kantians are simply better thieves, being zealots to the principle. Members of both groups know the world would be a better place if wrong-headed notions of morality were to disappear or diminish.

Brad C said...

Hi Eric,

A couple - ok 3 - points:

(1) I wonder whether it makes sense to assume that evolution would select for self-serving emotional responses. The "evolution explains altruistic behavior" crowd would prbly have doubts about this as an assumption.

(2) In a related vein -- it is worth noting that, if I remember rightly, Richard Joyce makes a plausible case for the view that belief in, and substantive compliance with, deontic constraints (and overriding external moral reasons) can be explained by evolution. The rough view would be that Kantian morality is in fact based on illusions (deontic constraints "depend on" false beliefs) but that it is in our collective interest to keep the illusion alive, so to speak.

Hope that makes some sense and is true to the animating spirit of his view.

(3)Finally I was talking to Julia Driver about your work on the behavior of ethicists -- specifically I was thinking about a possible link between vegetarianism and utilitarianism. I suggested that this might be a good case to study and show that moral thinking can produce better behavior - or at least bring behavior in line with perceived moral rightness.

But her story, and a brief survey of some other vegetarians and utilitarians suggests to me that the causation may go the other way: feelings that are pro-vegetarian lead to sympathy with, say, Singer's arguments.

Of course this is NOT to say that the view is not rational or rationally adopted.

Any way, I will check out the Greene - sounds like an uncharitable smear attack on Kantians, but maybe it is not.

Why are Kantians irrational for adopting views that *happens* to accord with evolutionarily shaped emotional responses?

Maybe nature was just not that step-motherly after all and gave us emotions that reliably track the right!

I miss commenting here!


Brandon said...

Greene's argument seems to me to overstep the bounds of his evidence; what his research shows is that moral judgments of Kantian deonotlogists largely tend to converge with evolutionarily-selected emotional responses, while the moral judgments of utilitarians and consequentialists will often diverge from them. And it's entirely possible for Kantians to think this good and consequentialists to think it worrisome. For Kant himself, of course, it would not be a problem; his system requires that something like this be true, that the rational and the natural converge (for their different reasons). It's arguably (a version of) a postulate of moral reason, in fact. But that gets into features of Kant's system that aren't always accepted today even by Kantian deontologists. Likewise, many of the early utilitarians would be worried about any major divergences, for similar reasons. (And ditto about the relation between their views and what is accepted today by modern utilitarians).

But assuming (a), (b), and (c) (I have doubts about all three), it does seem reasonable to take (a) and (c) as explanatory for (b).

Eric Schwitzgebel said...

Thanks for the comments, everyone!

Kevin: Greene might not actually disagree with your claim about the evolutionary basis of consequentialism, but he does think that there's a disanalogy, revealed by brain-imaging studies and reaction times in the type of cognition involved in the two theories and the exact way in which emotion is involved.

MT: I'm not sure I follow your suggestion. Is your idea (perhaps meant tongue-in-cheek?) that the Kantians are stealing the consequentialist texts -- to dispose of them, presumably? (If so, it's a cute thought, but it seems unlikely to me!)

Eric Schwitzgebel said...

Welcome back, Brad! We've missed your insight.

I agree about (1) -- but really all that's needed for my point (I think!) is that the emotional reactions selected by evolution are more self-serving than consequentialist principles are; and that seems plausible, even if we grant (as Greene also grants) that some evolutionarily selected emotional reactions are altruistic.

On (2): I'll have to read the Joyce, which I'm afraid I'm not familiar with. On the face of it, it seems reasonable to suppose there might be some rough relationship between what's in our best inclusive fitness and what makes for a good deontic ethics, but I'd bet that a closer look would reveal substantial divergences -- for example in our tendency to react more strongly to viscerally gruesome violations or to nearby violations. (Though maybe deontologists can take care of those sorts of mismatch through reflective equilibrium?)

On (3): That's an interesting speculation about the causal direction. I bet there's a lot to it. I'd love to find a good way to study rates of vegetarianism and charitable giving among philosophers!

In response to your one of your final remarks: Adopting views due to evolutionarily selected emotions is not per se irrational -- but neither is it to be driven by the processes of reason, at least in a fairly restricted sense of "reason". I agree that it's a point to get clearer on and a distinction that can become elusive when you push on it.

Eric Schwitzgebel said...

Thanks for the comment, Brandon. Greene does make the point that this may play out worse for contemparily naturalistic Kantian ethics than for earlier, historical positions, so he might not entirely disagree with you.

Brad C said...

Hi Eric-

Thanks for the speedy response! I just read (VERY quickly) through the Greene and have a more specific worry.

On major claim is that consequentialist responses require more cognitive processing while deotological ones require less and are based more on emotion. The methodology is to identify responses typical of the two types of views and to then do brain imagining and to measure response times. Quick and dirty emotional reactions track "deontic intuitions" - evidenced by brain activity in the emotional centers and by shorter response times.

My worry is that he is not running experiments for cases in which we would suspect (from the armchair) that consequentialist views will track intuitive responses - and hence emotions - and in which Kantian will run counter to conventional views and emotions. (By the way this worry is re-inforced when we see him casting the consequentialist as revising intuitive morality and the kantian as confirming conventional moral views -- he tenuously links this to certain trends in meta-ethics too...but back to the point).

As an example, consider cases such as the classic Kantian murderer at the door. Ignoring recent work by Kant interpreters like Wood, we can take the CI procedure as telling us that it is wrong to lie in this case. But conventional morality would not agree. I bet that this is a case where consequentialist outcomes would track emotion-quick-response time.

In general I think we can challenge his conclusions by picking out cases where deontological morality runs up against common conventional morality and consequentialism supports the conventional view. I bet that emotions will track the conventional and consequentialist views in these cases.

This all makes sense given how we normally teach intro to ethics - each approach captures some parts of conventional morality but they are each revisionary in some respects.

My guess is that emotional reactions and quick responses will track the overlap, but cognitive processing will increase in the places where each theory is revisionary. Emotion tracks convention - not one theory rather than another.

Ok -- sorry for the long post, ending in a speculative hypothesis!

Hope I did not mis-charicaturize his argument because of my quick read-through.


Brad C said...

Just to follow up:

On page 40 Greene seems to anticipate my worry but he tries to diffuse it by saying that contemporary Kantians aim to revise the theory so it no longer has revisionary import (at least in the Lying at the door case). Fair enough. But the same could be said about consequentailism - think of all the rule (and indirect, motive, and character) consequentialists out there. To paraphrase Greene, if you want to know what part of consequentialism will get revised by contemporary theorists, follow the emotion.

The fact is that many philosophers now employ something like the method of reflective equilibrium. But that does not show that they have no revisionary import. And that is not something that is characteristic of Kantianism or deontology per se.

Finally, I still think that in the cases in which Kantians support revision, their theory will not be found to comport with quick emotional processing.

KenF said...

Well, one way to look at it is that it would be very good to have a theory that provides a clear explanation of our naturally evolved moral inclinations. Such a theory might be extended to work on trickier cases, helping us figure out what to do, when our intuitions are unclear.

The analogy is mathematics, of course. We naturally evolved as creatures that can count, reason gemoetrically, etc. The study of mathematics has allowed us to apply these naturally evolved capacities into much more varied situations.

Starting with our natural moral sense and developing a theory that can consistently explain that isn't necessarily a bad thing. It could in fact be very useful, if the mathematical analogy "works".

Anyway thanks for linking to the article it is very interesting and opened up a lot of new ideas to me I hadn't encountered before!

Roman Altshuler said...

I wanted to second Brandon's concerns--that Kant would likely welcome findings that showed Kantian moral judgments to track emotional responses.

But a further point (and maybe Greene deals with this): Kant claims that we are beings who, by virtue of being rational, have a certain moral incentive. This incentive is emotional, in some sense--we respond to rational concerns because we experience them as superceding other concerns (this is Kant's point about respect as the moral incentive). The difference between moral and immoral maxims, for Kant, is not that the latter involve emotions and the former do not, since without some affective pull on the side of pure reason, it would have no motivating force for us.

So really, to experimentally distinguish between Kantian and non-Kantian moral thinking, we'd need to figure out precisely which part of the brain processes respect, and which part of the brain processes other motivational states. And that would seem to require a kind of reductive naturalism that would a priori rule out Kantian morality as anything more than evolutionarily programmed emotional responses.

Do meta-ethical questions like this really lend themselves to experimental testing?

Genius said...

haha but utilitarian texts are so much more worth stealing! for the good of humanity of course :)

As to the main point I have a lot of sympathy for the position.

One of my basic theories of philosophy is that if it matches "intuition" I should be very suspicious of it. And if I find myself biting a bullet it means I have probably weighed the pros and cons fairly well.

Genius said...

BTW I think I am genetically predisposed towards being utilitarian or if not just a plain every day enabler. So maybe some people just tend to be both vegetarian AND utilitarian rather than there being a strong causal relationship (things that might go with this is being collectivist and empathic).

Anibal said...

kantians= post hoc rationalizations guided by rules, norms...

utlitarians= in practical rterms act or do always what provide yu with the most favourbale or greatest return.

Kantians are mapped in the brain in areas such as hippocampus or ACC (an area associated with detection of erros[ a deviation form a rule or norm])On the other hand utilitarians are mapped in the brain in the frontal cortex an are that integrates many input form the rest of the brain and is associated with utility calculations
I think what Greene tries to dmonstrate is that kantians by means of evolution have "internalized" emotions as reasons only that they operate not as concurrent gut-feelings but as less vivid and rationalized gut-feelings

Eric Schwitzgebel said...

Wow, what a thoughtful set of comments this post has provoked!

Brad: Your criticism is very interesting. I'd be curious to see how Greene would address it. I think it would be better to stay away, though, from the case of lying to murderer at the door and try to come up instead with a case that more deontologists can agree on that diverges from our commonsense judgment. Hm, but what case would that be? If there are no cases where mainstream deontologists generally go against folk intuition, then Greene's point may still stand.

Kenf: The analogy to mathematics is a nice one. Greene might argue that what we are responsive to and extend through reasoning in the mathematics case, though, has some independent validity (what validity, exactly, is a tough issue in philosophy of math!) while what we are responsive to in our moral emotions and extend through reasoning has only limited value and validity -- except perhaps something very general like an aversion to seeing others harmed, which we then extend through consequentialist reasoning.

Eric Schwitzgebel said...

Hi Roman! It doesn't seem like you post as much as you used to....

I agree that Kant doesn't want to separate emotion and moral reasoning, so *if* emotional areas in the brain are highly active when the Kantian reflects, that in itself doesn't make Greene's case. Kantians might simply be, as you point out, feeling respect, or they might be filled with passionate awe at the beauty of the moral law. So the more striking aspect of Greene's results is that the deontological reasoning seems to be associated with less activity in the reasoning areas and with quicker reaction times suggesting that not only is there more emotion but there is less reasoning. This I'm not as sure a naturalistic, 21st century Kantian with no reason to expect harmony between automatic emotional judgments and the moral right should happily accede.

Genius: I see your point about being suspicious about what is intuitive -- because maybe, then, it's not really grounded in philosophical reasoning! But then it might be hard to justify a metaphilosophy that stands intuition and "reflective equilibrium" near the center of philosophical method. Where would ethics (or metaphysics, or logic) go if somewhere down the line our judgments weren't tacked down in "intuition"?

Eric Schwitzgebel said...

I don't know, Anibal. In this essay at least, Greene emphasizes emotion-related areas more than the areas you mention. I'm also not sure that he thinks the emotions in Kantians are tamped down due to their reasoning. Are you getting this from another essay of his, maybe? Or am I missing something in this essay?

Genius said...

"Where would ethics (or metaphysics, or logic) go if somewhere down the line our judgments weren't tacked down in "intuition"?"

That is OK with me - just as long as it is a long long way down the line. ... like 1+1=2 is down the line from quantum mechanics.

Roman Altshuler said...

Hi Eric,

Yes, I haven't been posting much because that pesky dissertation has been keeping me busy.

Greene's results are certainly interesting, and I agree that they could be a serious concern for some contemporary Kantians. But as a sort-of Kantian myself (not a naturalistic one, though), I always want to take things back to Kant--and it is Kant, and not just contemporary Kantians, that Greene takes himself to be attacking. So just one point in that regard:

Kant differs fundamentally from the British empiricists and utilitarians in his conception of reason. And in his moral philosophy especially he is very skeptical of the empiricist conception of reason, i.e., as a mere faculty of inferring from given data (such as desires) to conclusions (maxims). Kant stresses the human tendency to provide rationalizations for what we want (and to make excuses for ourselves) and he offers the categorical imperative--as he says--not as a decision procedure, but primarily as a corrective to that rationalizing tendency. He thinks, in fact, that we generally know what the right thing to do it (or, at least, what the right principle is, though there is no set procedure for deducing specific actions from that principle). And the big problem, from his standpoint, is that people abuse reason and employ it as a rationalizing tool against what they already know to be right.

So Greene's results (and I confess I haven't had the time to read his paper in detail) might be taken as ones Kant would appreciate seeing: they show that consequentialists tend to use the calculative capacity of reason more than deontologists. But this could just be because they are struggling more to rationalize conclusions that, as moral beings, they know to be wrong. It may seem like the parts of the brain involved in reasoning are less active in deontologists. But of course this could just be because those are the parts of the brain involved in calculation and rationalization; the part of the brain involved in what might be called the constitutive aim of reason might just be the emotional part.

Sorry about the tangent.

Anibal said...

Eric, perhpas is my own idiosyncratic reading of the essay, but let me quote Greene himself: page 2 "i will argue that deontological judgements tend to be driven by emotional responses and that deontological philosophy rather than being grounded in moral reasoning, is to a large extent an exercise in moral rationalization" i would like to empashize his latests words moral rationalization
Based in his model of the dualistic properties of the mind (in rough manner emotion VS cognition) as a psychological kinds or natural kinds instantiated in our brains, he defends the idea that primarily emotion responses are those specifically tuned to moral decision and behaviour. In this sense, kantians operate according to his model with emotional responses, only that this kind of moral creature after emotional responses rationalize "emotions2 converting them in rules, norms, rights embeddded in what i suggest brain areas such as the hippocampus or brain areas which store permanent information and detect cogntive error to that stored information (ACC) But i know that these brain areas are not mention widely in the text.
As i see Greene, Greene is a modern-way and sohisticated type of utilitarian, that tries to subsume rivals.

Eric Schwitzgebel said...

Genius: That makes sense, I suppose -- at least if the 1+1=2 intuitions are much more solid and immovable than the later-down-the-road intuitions.

Roman: I agree that there are passages in Kant that warn against the abuse of reason for the purposes of rationalization; and there's his famous delight at a critic's complaint that nothing he says overturns the ordinary view. But do you think Kant really would have thought (if this is what you're implying) that intellectual reasoning about moral problems would tend to lead away from the moral truth -- or at least toward consequentialist conclusions and thus away from the away from the truth in those sorts of cases?

Anibal: Yes, I agree with your characterization of Greene's characterization of Kantians (though his emphasis is on different brain areas, as you mention); and as you mention in your earlier comment, he thinks utiliarians do it differently. I think, though, that there's a further point you're making that I'm still not getting....