An Evolutionary Theory of Right and Wrong
November 1, 2006 8:03 AM   Subscribe

Moral Minds, a new book by Marc Hauser, is based on research by Hauser and colleagues such Josh Greene and John Mikhail. In it, he posits that an innate moral sense is analogous to "universal grammar"[Wiki] from Chomskyan linguistics. As reviewed by a Science Times staff member. ...And a philosopher.
posted by Arthur "Two Sheds" Jackson (23 comments total) 6 users marked this as a favorite
 
Rortyl has long sided with the social/cultrual construct/humanities perspective and been on the attack against claims made by science, evolutionary biology...this is a split that has been around now for some time, and it is a shame that he is called in to evaluate something that from the onset he opposses. Perhaps a moral sense would have helped him in this, no matter where that sense came from
posted by Postroad at 8:27 AM on November 1, 2006


Postroad: Rortyl has long sided with the social/cultrual construct/humanities perspective and been on the attack against claims made by science, evolutionary biology...

Of course, not all aspects of science are on the same side of this issue, or even agree that there are two sides in opposition. Most of the theoretical work I've seen suggest that most human behavior is the product of multiple factors.
posted by KirkJobSluder at 8:41 AM on November 1, 2006


What about the Morlocks? Will they have the moral sense?
posted by No Robots at 9:06 AM on November 1, 2006


Anything Chomsky likes, I'm against. Anything Rorty dislikes, I'm for.

So I'm in a quandary here.
posted by languagehat at 10:12 AM on November 1, 2006


Ya languagehat, looks like you might actually have to think about the issue. ;)
posted by Arthur "Two Sheds" Jackson at 11:25 AM on November 1, 2006


Yes, because we all know that science journalism accurately represents the responsible scientific community and never, ever makes poorly substantiated sensationalist claims out to be scientifically legitimate.

Nevertheless, researchers’ idea of a good hypothesis is one that generates interesting and testable predictions. By this criterion, the proposal of an innate moral grammar seems unlikely to disappoint.

Astrology generates interesting and testable predictions.

This doesn't, actually. What kind of testable prediction can be made from the claim "humans have innate morals [which happen to correspond to those of my culture, gee howdy]"? The pushing the fat man on the tracks one? Please show me an experimental design which would test for innate moral sense.

First, the subjects would have to represent every human culture as well as include a child raised by wolves.
Second, the experiment would have to involve an actual moral choice, not a thought experiment (since allegedly this operates on a pre-reflective level).
Third, this moral choice would have to be serious enough to trigger the innate moral sense (since it's pretty obvious that forced expropriation, for instance, changes its moral status based on culture. even rape does, to some extent).
Fourth, even if the subjects demonstrated a statistically significant constancy in their response, there is no proof that this isn't a result of extraneous factors such as danger avoidance or non-innate feelings of empathy.
...I think Hauser has found a way to solve the obesity epidemic: kill lots of fat men in fallacious experiments.

Like nearly everything in the steaming shitpile of evolutionary psychology, this argument is fucking stupid.
posted by nasreddin at 11:50 AM on November 1, 2006


i think i heard an Excellent publicradio piece about these guys on monday. it was an hour of these guys and some other people that was some of the best radio i've ever heard. did anyone else hear that?
posted by localhuman at 1:05 PM on November 1, 2006



"Anything that even hints of hereditary behaviors is just crazy talk.CRAZY TALK I say. I'm not listening to your crazy talk... (fingers in ears) na na na la la la buh buh buh..."

"You should read the study, it sa..."

"CRAZY TALK. Buh! buh buh buh...
posted by tkchrist at 1:13 PM on November 1, 2006


that was some of the best radio i've ever heard. did anyone else hear that?

Yup it was excellent.
posted by tkchrist at 1:15 PM on November 1, 2006


"Anything that even hints of hereditary behaviors is just crazy talk.CRAZY TALK I say. I'm not listening to your crazy talk... (fingers in ears) na na na la la la buh buh buh..."

"You should read the study, it sa..."

"CRAZY TALK. Buh! buh buh buh...


When you get tired of knocking down straw men, my argument's over here if you want to actually engage in debate.
posted by nasreddin at 1:29 PM on November 1, 2006


Yeah, nasreddin?

This quote:

Like nearly everything in the steaming shitpile of evolutionary psychology, this argument is fucking stupid.

...was quite entry to an intelligent debate. Your all ears, I'm sure.

Read the links. You obviously DIDN'T. And then listen to the radio show. MAYBE we'll talk then.
posted by tkchrist at 1:43 PM on November 1, 2006


I for one welcome our new genetics-based moral imperatives.

Springtime for Hitler and Germaneeee!
posted by No Robots at 1:48 PM on November 1, 2006


I have the right to be vocal and uncharitable when advocating a point of view. However, my comment actually contained an argument, which you have failed to respond to. Please demonstrate how the links (which I've read) prove me wrong.

Read the links. You obviously DIDN'T. And then listen to the radio show. MAYBE we'll talk then.
Oh mighty tkchrist, wilst thou condescend to debate with me if I perform the ritually appropriate actions?

The classic argument of the fighting keyboarder: "I can't argue with you until you do this and this, and if I still can't argue with you I'll deny that you did it."
posted by nasreddin at 1:54 PM on November 1, 2006


Always happy to read Rorty. And great to see something by him on the front page.
posted by bardic at 2:43 PM on November 1, 2006


This doesn't, actually. What kind of testable prediction can be made from the claim "humans have innate morals [which happen to correspond to those of my culture, gee howdy]"? The pushing the fat man on the tracks one? Please show me an experimental design which would test for innate moral sense.

There are any number of testable predictions, depending on just how you choose to operationalize the notion of an innate moral grammar. If, for example, you unpack the notion as "there are certain patterns of behavior that are coded for in our genome and express themselves in the same way in virtually all known human environments", then that's easily testable using behavioral genetic approaches. In fact, there's overwhelming evidence that traits associated with variation in moral behavior (e.g., psychopathy, antisocial personality, etc.) are highly heritable.

First, the subjects would have to represent every human culture as well as include a child raised by wolves.

Not really. Science is inductive. Saying you need to look at every token to draw inferences is absurd. There's clearly an innate difference in men and women's heights; do you need to examine every individual on the planet in every possible environment to conclude as much? No. You need a reasonable sample size and a reasonable (ideally, statistical) way to draw inferences with a reasonable degree of certainty.

Second, the experiment would have to involve an actual moral choice, not a thought experiment (since allegedly this operates on a pre-reflective level).

No, the point is that there are both deliberative and non-deliberative influences on moral choice. But to the degree that they're non-deliberative, there can also be an underlying structure there. In the same way, you don't consciously deliberate over the syntax of the words you choose, yet the fundamental existence of a universal grammar is hardly in dispute any more.

The concern that what people say they would do in hypothetical scenarios doesn't necessarily have to map onto their actual behavior is a valid one. But again, it's a perfectly testable hypothesis. You can, for example, ask people whether they would take $10 sitting in a jar if no one would find out about it, and then conduct an analogous experiment. Now obviously, you're not going to go around killing people in scientific experiments. But you don't have to. The entire point of a 'moral grammar' is that a relatively small number of parameters should account for a vast array of moral behaviors, from the egregious to the petty. There are lots of scientific hypotheses we don't test in the real world because of ethical concerns; that doesn't mean we don't bother to test those we can!

Third, this moral choice would have to be serious enough to trigger the innate moral sense (since it's pretty obvious that forced expropriation, for instance, changes its moral status based on culture. even rape does, to some extent).

You misunderstand the point of Hauser's argument (presumably because you haven't bothered to actually read it). Hauser's claim isn't that there's a universal moral code we all obey independent of culture. Far from it. The point is that moral behavior is similar to language: there are a (relatively small number) of parameters that can be 'set' to various levels, and those can vary across cultures. The novel point is the idea that these principles can be relatively simple yet generate a vast array of moral systems when places in specific environmental contexts. This is analogous to Chomsky's argument, which isn't that there's a single innate language (ridiculous!), but that the possible variation in languages is finite and can be well-characterized with a surprisingly small number of parameters.

Fourth, even if the subjects demonstrated a statistically significant constancy in their response, there is no proof that this isn't a result of extraneous factors such as danger avoidance or non-innate feelings of empathy.

Yes, these are called 'confounds', and experimenters take great pains to try to rule such influences out experimentally or statistically. It may come as a surprise to you that people who spend their careers working on such problems might have worried about these issues ahead of you, but there it is.

Like nearly everything in the steaming shitpile of evolutionary psychology, this argument is fucking stupid.

Perhaps, but you've hardly presented a convincing argument to that effect.
posted by heavy water at 4:13 PM on November 1, 2006


heavy water: thank you for engaging with what I've said.

There are any number of testable predictions, depending on just how you choose to operationalize the notion of an innate moral grammar. If, for example, you unpack the notion as "there are certain patterns of behavior that are coded for in our genome and express themselves in the same way in virtually all known human environments", then that's easily testable using behavioral genetic approaches. In fact, there's overwhelming evidence that traits associated with variation in moral behavior (e.g., psychopathy, antisocial personality, etc.) are highly heritable.

I don't see how unpacking it this way is valuable. I don't think anyone really questions that there are certains patterns of behavior that are coded for, etc. The question, to my mind, is why certain behaviors are perceived as moral or immoral, and how moral choices can be weighed. Hauser is arguing that moral weighing operates on a genetic level, to some extent. I argue that this is not testable. Sociopaths are a different question: I think the principal issue in this case is whether antisocial personality, etc results in a) changes to the innate moral grammar or b) ability to adjust behaviors and beliefs to social expectations.

Not really. Science is inductive. Saying you need to look at every token to draw inferences is absurd. There's clearly an innate difference in men and women's heights; do you need to examine every individual on the planet in every possible environment to conclude as much? No. You need a reasonable sample size and a reasonable (ideally, statistical) way to draw inferences with a reasonable degree of certainty.

I know this, and I'm not saying you need to look at every individual. But if you run a study with just Westerners and Native Americans, say, what you're proving is only the overlap between their moral systems. The problem here is that to test the hypothesis, we need to disentangle socially-inculcated morality from innate morality, and that's very difficult to do without a reasonably complete selection of cultures to adjust for cultural variables.

the fundamental existence of a universal grammar is hardly in dispute any more.

I am not an expert on this, but languagehat would probably disagree. I don't know enough to dispute this point.

You can, for example, ask people whether they would take $10 sitting in a jar if no one would find out about it, and then conduct an analogous experiment. Now obviously, you're not going to go around killing people in scientific experiments. But you don't have to. The entire point of a 'moral grammar' is that a relatively small number of parameters should account for a vast array of moral behaviors, from the egregious to the petty. There are lots of scientific hypotheses we don't test in the real world because of ethical concerns; that doesn't mean we don't bother to test those we can!

My argument was that instinctive moral valuations would probably play out in radically different ways depending on the proximity and intensity of the moral question (i.e. we can sit and talk about lions on the internet without my instinctive fight-or-flight response being triggered), and it would be difficult to find a testing scenario that can account for this fully.

You misunderstand the point of Hauser's argument (presumably because you haven't bothered to actually read it). Hauser's claim isn't that there's a universal moral code we all obey independent of culture. Far from it. The point is that moral behavior is similar to language: there are a (relatively small number) of parameters that can be 'set' to various levels, and those can vary across cultures. The novel point is the idea that these principles can be relatively simple yet generate a vast array of moral systems when places in specific environmental contexts. This is analogous to Chomsky's argument, which isn't that there's a single innate language (ridiculous!), but that the possible variation in languages is finite and can be well-characterized with a surprisingly small number of parameters.


I read the linked articles (not the book, though). I read Hauser not only as making this argument, but also making some claims as to the nature of these moral parameters. If his claim is ony that morals are composed of a small number of innate parameters, then it is fairly trivial and was comprehensively mapped out by Aristotle two millennia ago. The more difficult question, which Hauser's argument seems inadequate at explaining, is why the variation in moral systems is so much greater than variation in genetics and environment between cultures. Hauser leaves this, and the equally important question of why different cultures seem to exclude different categories of people from their core moral rules, as an exercise to the reader. But if he doesn't answer these questions, his whole system is pretty pointless.

Yes, these are called 'confounds', and experimenters take great pains to try to rule such influences out experimentally or statistically. It may come as a surprise to you that people who spend their careers working on such problems might have worried about these issues ahead of you, but there it is.

Don't assume I am more ignorant than I am. My point is that there is no way statistically to distinguish innate moral grammar from cultural and other evolutionary artifacts, especially since on the genome level the putative moral grammar probably overlaps substantially with other characteristics.

Perhaps, but you've hardly presented a convincing argument to that effect.

My argument was poorly articulated, but I am willing to defend it.
posted by nasreddin at 5:33 PM on November 1, 2006


the fundamental existence of a universal grammar is hardly in dispute any more.

I am not an expert on this, but languagehat would probably disagree.


Languagehat disagrees fundamentally and vociferously. This is Chomsky's quasi-religious dogma; no matter how many counterexamples are given, he keeps desperately propping it up. It's bullshit.

Not that that has anything to do with this, of course. I take no position on the universality of morals, and I doubt it can be proven one way or the other.
posted by languagehat at 6:03 PM on November 1, 2006



Which culture believes that it is praiseworthy and delightful to kill happy, health kindergartners for fun-- and which culture would consider it most honorable that if given the choice of which happy, healthy five-year-old to kill for fun, an individual would choose his own?

If you can name one, you have disproven universal human moral grammar and disproven that there are evolutionary roots to our moral sense.
posted by Maias at 6:32 PM on November 1, 2006


Which culture believes that it is praiseworthy and delightful to kill happy, health kindergartners for fun-- and which culture would consider it most honorable that if given the choice of which happy, healthy five-year-old to kill for fun, an individual would choose his own?

This is a loaded way of putting the question. What is more reasonable is to admit that some cultures and/or individuals have found various reasons to believe that killing some five-year-olds is honourable.
posted by No Robots at 7:43 PM on November 1, 2006


Good heavens, I agree with Rorty. I must go and flaggelate myself forthwith. Gosh, I have never agreed with him before. I'd really like to hear Alasdair MacIntyre's opinion, or Kieran Egan's, or even David Stove's (oh, sorry, can't do that; he's dead.)

Please don't tell me Rorty is actually right on something.

But seriously; it seems (if the reviews are true to the book's actual content, of course) that the author is saying something that is either a tautology: "Lots of baseline ethical content is determined by the way our central nervous system is constructed." or something, if not unprovable, at least unproved: "We can prove that this PARTICULAR part of ethics is hardwired, and this PARTICULAR part is not."
posted by Topkid at 8:01 PM on November 1, 2006


I would suspect that there is a lot about learning in life that involves our brains flipping on and off switches somewhere inside, morality, language or whatever.

How else do we process all of the data in the world?
posted by taursir at 9:02 PM on November 1, 2006


The question, to my mind, is why certain behaviors are perceived as moral or immoral, and how moral choices can be weighed. Hauser is arguing that moral weighing operates on a genetic level, to some extent. I argue that this is not testable.

Why not? You can make all sorts of novel and testable predictions about what types of moral judgments people ought to make (if your theory is right) in completely novel situations they haven't considered before. For example, take the trolley cart problem: it's not a trivial thing to explain why people have little compunction diverting a train and killing one person to save five, but find it reprehensible to push a person off a cliff to save five people. The explanation "it's the element of personal involvement and feelings of responsibility it engenders!" is non-trivial, and allows you to make plenty of predictions about what people ought to do under specific circumstances. You can experimentally manipulate the degree to which a particular element is present in a situation (hypothetical or not) and measure the effect it has. I fail to see how the notion that there are universal principles underlying moral behavior is any less testable than most other psychological hypotheses. Of course, it may require a measure of ingenuity to test adequately, but that's no reason not to try.

The problem here is that to test the hypothesis, we need to disentangle socially-inculcated morality from innate morality, and that's very difficult to do without a reasonably complete selection of cultures to adjust for cultural variables.

I don't think that's the appropriate test. The prediction Hauser makes is that a small set of principles will explain the interaction between genes and culture. The claim is not that there's a free-floating moral grammar that dictates moral behavior across all cultures. So, for example, the covariance between specific cultural variables and specific types of moral behavior can be harnessed to test predictions. For example, do all cultures with element X present also show moral behaviors Y and Z? (Where Y and Z are non-trivially linked to X, obviously).

Note also that it's not like these predictions only apply at the broad cultural level. To some degree, moral judgment and behavior should be modifiable (within certain bounds) using targeted manipulations. There are already experimental demonstrations of this. For example, people given a positive mood induction are more likely to say pushing someone off a cliff to save 5 other people is ok--presumably because positive mood buffers one against the negative affect associated with the thought of pushing someone to their death. There's no reason why plenty of other predictions can't be similarly tested.


My argument was that instinctive moral valuations would probably play out in radically different ways depending on the proximity and intensity of the moral question (i.e. we can sit and talk about lions on the internet without my instinctive fight-or-flight response being triggered), and it would be difficult to find a testing scenario that can account for this fully.

Fair enough, but that's something a comprehensive model can take into account as well. (E.g., if you think the degree of emotional involvement in a situation is likely to modulate one's judgment, you can experimentally manipulate that factor to whatever extent is possible, even if you can't have subjects facing off against lions.) More generally, this is a concern that applies equally to the bulk of psychological research (the issue of external validity). As a psychologist, you always worry that the lab tasks you use to test your hypotheses are not good analogues of real-world behavior. But again, that's no reason to throw in the towel before you started.

The more difficult question, which Hauser's argument seems inadequate at explaining, is why the variation in moral systems is so much greater than variation in genetics and environment between cultures.

Huh? How are you quantifying 'variation'? This is a vacuous statement unless you can somehow quantitatively compare the variation in moral systems to the variation in genetics and environment in the same metric. Frankly, it's not at all clear to me how you would even begin to make this judgment. The closest thing I can think of is heritability estimates using behavioral genetic methods. But when you do that, it turns out that the vast majority of traits (including most personality dimensions, political attitudes, and moral views) are highly heritable, with the residual variance being due primarily to unique environmental influences (rather than shared environmental influences such as home environment).

Hauser leaves this, and the equally important question of why different cultures seem to exclude different categories of people from their core moral rules, as an exercise to the reader.

Well, it's a little unfair to demand that Hauser articulate not only an interesting idea about the possibility of a universal moral grammar and some general principles, but also a fully-fleshed out model, complete with an explanation for the above! That said, what makes you think a moral grammar, should one exist, can't handle this issue too? In other words, why isn't the question as to who sets of morals are applied to also a question for scientific inquiry to settle?
posted by heavy water at 9:36 PM on November 1, 2006



No Robots, my argument about killing "happy healthy five year olds" for fun was not a loaded way of putting the question-- I phrased it the way i did to stress biological variables.

In other words, the "morally obvious" is biologically driven. It is morally obvious that killing for fun is always wrong-- killing to propitiate a god or something like that is seen as a sacrifice and not something to be done lightly and the times when it is possibly acceptable vary culturally.

It is morally obvious that killing obviously healthy children who are not in some sort of stigmatized "out group" (ie, the children of the enemy) is always wrong, again, defining the enemy is culturally variable but it is always the case that killing healthy children (not newborns or fetuses) is more unacceptable than killing the post-reproductive. It is morally obvious that death before reproductive age is especially tragic. It is morally obvious that it is more difficult to advocate the killing of your own child than that of someone else.

The reason for the moral obviousness is the commonality of our evolved moral reasoning and its relation to our biology as social mammals-- the exceptions are what varies by culture, which is exactly what you would expect of such a system.
posted by Maias at 3:43 PM on November 2, 2006


« Older “Nobody. Myself.”   |   Save Studio 60 Newer »


This thread has been archived and is closed to new comments