« Overcoming Bias: Hobby, Virtue, or Moral Obligation? | Main | Anonymous Review Matters »

August 02, 2007


Sounds like someone needs to examine his bias re Alabama bar patrons.

Paul, I looked up a list of the most religious states in the US. But if you actually go into an Alabama bar and say it, I'll change the post (not recommended).

I'm not about to put my money where my mouth is on that one

I thought this site would be the last place I'd see criticism of the "suicide bomber as cowardly" notion. Under some definitions, sure, doing something you expect to result in your death, in pursuit of a higher goal, necessarily counts as courage. However, it would be justifiable to say they are intellectually cowardly. That is, rather than advance their ideas through persuasion, and suffer the risk that they may be proven wrong and have to update their worldview; rather than face a world where their worldview is losing, they "abandoned" the world and killed a lot of their intellectual adversaries.

It is an escape. There is, after all, no "refutation" for "I'm right because I'm blowing up myself and you".

It's for the same reason one might apply the "coward" label to a divorced, jealous husband, who tries to "get back" at his ex-wife by killing her or their child. He, too, exposes himself to immense risk (incarceration, or if they defend themselves). He too, is pursuing a broader goal. Yet in that case, my calling him a coward is not an artifact of my disagreement with his claim that he has legitimate grievances -- in fact, I might very well be on his side (i.e., that the courts did not properly adjudicate his claim).

So yes, it might be the "American" thing to say terrorists are cowardly -- but that doesn't make the claimant biased or wrong.

Are you breaking your advice to not use contemporary politics in examples?

Good posts. This series is the first thing in a while to make me really glad to participate here.

I think that the stereotype of Alabama bars is pretty reliable. OTOH, the stereotype of suicide bombers is much much less so. If you read the rhetoric of radical Islam, or for that matter if you read ancient mythology such as Homer or the Egyptian Book of the Dead, you will see people who are occupying a VERY VERY different moral universe from us Platonized Christianized (that includes the secular children of "modern orthodox" Jews) post-Enlightenment Westerners.

In terms of realistic psychology fitting neither the SSSM nor the Evolutionary Psychology brand (which you really should spend more times reading non-leftist criticisms of), the Muslims who flew planes into the World Trade Center undoubtedly saw themselves as heros, but in some sense that we would have a VERY hard time empathizing with or relating to. They are NOT a mirror reflection of ourselves, but genuinely something that has to be understood with empiricism, not empathy and wishful thinking.

Good point, Tarleton - although I'm still hard-pressed to think of a better example that isn't directly a religious belief. If you only use the obvious religious examples, people will fall into the standard trap of thinking they've achieved perfection as rationalists because they're not religious - I wanted to use something that would actually strike a sympathetic chord and let people see how the belief-as-attire effect extends beyond religion. Got a better suggestion?

Vassar, also a good point, although I'm skeptical that I would have difficulty empathizing - these are humans we're talking about, not aliens, and the WTC hijackers were mostly educated Saudi Arabians, not Yamomano. They saw themselves as heroes in the support of causes, such as sexual decency = woman's de-emancipation, which are not American causes; they believed and maybe even anticipated 72 virgins; they fought in guardianship of ancient perfection; they carried out the will of God revealed in perfect scripture. None of this strikes me as a significant barrier to understanding. Can you say specifically what you think presents the barrier to empathy?

Ignorance. I may think I understand their minds, but that does not prove that I do understand their minds.

All you know is that you have a mental model of their minds which seems credible to you. Have you tested this model, and if so, how?

All I am reasonably sure of is that they did not see their act as evil and cowardly. Doubtless the same was true of Jack the Ripper and the Boston Strangler, but that tells me nothing about the differences between them and everyone else. After all, I only think that is true of them is because it seems to be true of most people.

It is really just an assumption.

This might be an especially easy category of bias to identify. Just ask yourself if you feel proud that this belief associates you with some group with which you want to be associated. If so, weaken your confidence in this belief.

Pseudonymous, I confess that it is only a guess, just a more plausible guess than the American one. And Jack the Ripper might well have been a monster - there are such things as sociopaths.

Robin, I'd say "recalculate your reasons" not "weaken your confidence". You can't literally shift a probability estimate because it makes you feel proud.

I should mention my old short essay, Are Beliefs Like clothes?.

Eliezer, it seems to me that you can and should shift your probability estimate because it makes you feel proud. Of course you might do even better than that by recalculating your reasons, but that approach will often not be cheap or reliable.

it would be like dressing up as a Nazi for Holloween.

"So remember kids, dressing up like Hitler in school isn't cool."

Eliezer, your lack of familiarity with "the other side" on the topic of terrorists is all too obvious from your crude attempt at a characterization of it. All you appear to know about it is a few platitudes. Often I get the feeling from this site that it is not so much about overcoming your own biases as it is about coming up with new excuses to dismiss views that you don't agree with by applying the genetic fallacy over and over (e.g. "suchandsuch belief is a product of suchandsuch bias").

Brilliant post, in my opinion. Clarifying and edifying. I'm looking forward to where you're going to go with this analysis of how bias and belief operate.

Silas, My opinion: you seem invested in using "muslim terrorists" for in-group/out-group construction, and I think it's coloring (biasing?) your analysis.

Michael, great criticism of an element of Eliezer's post.

Hopefully_Anonymous: You seem invested in labeling people as using "muslim terrorists" for in-group/out-group construction, and I think it's coloring (biasing?) your analysis.


Constant, this blog has warned against the genetic fallacy before. What do you think would be a good characterization of the "other side"? Eliezer's characterization describes a large minority of Americans very well. (He's clearly not intending it to be descriptive of everyone who thinks Islamic extremism is a serious threat, if that's what you're thinking.)

I think questioning the Alabama bar analogy is useful within the context of this post. Whose attire is a belief in the value of giving primacy skepticism, critical thinking, etc.? According to Eliezer's performance in the OP, it's not the attire of either Alabama bar patrons or "muslim terrorist suicide bombers" -and both of those may signal more generally, the losers of the American Civil War and non-white brown people. In short, I think there may be a gentrification of critical thinking: it's reserved for an in-group, perhaps in particular northeastern anglo-saxon and ashkenazi jewish male intellectuals, or an even more narrow archetypal definition that might be reducible to zero actual people. I'm interested in the degree to which our behavior might be governed by aligning with and contesting these archetypes. Including which beliefs as attire to wear (it's perhaps an archetype alignment for Steven Hawkings and Richard Dawkins to claim to be skeptical about religion. It would probably not be an archetype alignment for Oprah to publicly wear such belief attire, even if in fact she was a crypto-skeptic).

This post may meander a bit but I think Eliezer's post (and some of the criticisms of it) are thought provoking and may be getting us closer to a more real world, real time model of how bias and belief is operating in the world we live in.

Empathy is hard. Cultures differ. We Americans (especially secular Americans?) really don't have a clue what it feels like to (for instance) feel an obligation to kill our daughters or our sisters in order to preserve our family honor. Some actions in the name of causes may be psychologically modular, but some really aren't. What's the parallel for honor killing? Pressuring one's schizophrenic philosophy post-doc son to go to law school where he thinks he'll be miserable for the bragging rights? Sending your kids to Hebrew School or Day Camps they hate because your parents made you do it? It just doesn't work.
Even within a culture, I have no idea what it's like to identify with a sports team and very few people can relate to the horror that I feel at some Psychological data or philosophical ideas.
You once pointed out that most of us can no longer even understand why the Psycho shower scene was once considered terrifying. I would recommend Silvia Plath's diary for what are to me stranger attitudes than those.

Actually, I think that much of www.xkcd.com including the current one can be though of as an enumeration of feelings that people who aren't either quite young or quite nerdy have not analogues to. Philosophy is full of others, such as existential despair and satori.

Michael, I think your example is interestingly rooted in an implied in-group/out-group construction that construction Americans in a flattering way. Consider that you contrast honor killings with forcing kids to go to law school or day camp -that won't necessarily result in their death. It's a flattering contrast that I think constructs America as Western and honor killers as culturally Middle Eastern. But, if we contrasts cultures that approve of state-sanctioned killing of people for moral transgressions, America and the nations of the honor-killers are now in the same group, with Western Europe (and much of the rest of the world) in the other group. Incidentally, I'm not opposed to state-sanctioned killing, but I think it would be more rational for the penalty to start with doing it to to individuals to the extent it will prevent future great economic waste/increase in existential risk, rather than to punish premeditated murder of a small number of people or purported extramarital/premarital sex.

HA: I chose my examples carefully to to try to match as closely as possible as many of the categories, relationship types, motivations, etc as I could, and the examples I came up with are both pervasively American and truly ugly from my perspective in the closest way that I could think of (matching type of motive, e.g. content of emotional state, not degree of emotional state or degree of ugliness) to honor killings. My point was that we don't have any very close matches.
Your examples still don't match the intensity of honor killings, but more importantly, the emotional quality is utterly utterly different. Anger, retaliation, maintenance of public order, prevention of repetition, deterrence, The motivations for execution are simple and easy to understand, as is the balance calculation which compares the costs and benefits implicitly, even if a more careful calculation would disagree. At the most visceral level, honor killings are not in retaliation for some harm nor are they motivated by preventing a harm. This article is probably worth looking at for everyone who is still reading this thread, by the way.


By the way
Honor killings != state sanctioned killing.

Michael, how about the point that you're (rather explicitly now) picking a point upon which to manufacture in-groups and out-groups. In-group: those of us who get motivations for execution. Out-group: those who get honor killings.

The in-groups and out-groups change if the point to get is abrahamic monotheism, or if the point to get is state-sanctioned punitive killings. It seems to me that you're picking one that's particularly salient either to you or to what you imagine your audience to be. I think this gets to the belief as attire/beliefs as cheers for teams. It's an attempt to pick teams, but I think the implied in-groups and out-groups are at least in theory contestible.

A bit tangentially, I think teams themselves can be an effective (the most effective?) way to construct hierarchical privilege. The people on the field vs. the people on the bench (or the people regulated to the audience) of the two teams.

In terms of overcoming bias, I think understanding and when necessary countering these phenomena is important primarily to the degree that they warp decision-making or increase economic waste/existential risk.

HA: It seems to me that you think I have changed the topic. I agree with all of the sentences of your most recent comment, except for the first, but they don't seem to be about what I was saying.

Likewise, I agreed with Eliezer's post, but I thought that his analogy was, well, lacking in appreciation of the difficulties involved in analogy.

Basically, I think that Douglas Hofstadter's writings on the difficulties of natural language translation, the proper translation of literature, etc, are all in relevant to the issue of the translation of inferred emotion. Evolution provides a scaffold for our brains to develop, biasing us, strongly or weakly in certain directions,
sometimes strongly enough that translation *is* at least almost always possible (maybe not with the Piraha?) and we can speak of human universals, and sometimes weakly enough that we can talk about Liberal insensitivity to 3 of Haidt's 5 domains of morality. When the biases are strong or when atypical emotional mixes emerge and propagate memetically, empathy ceases to be useful and "the other" must, in some respect or another, be understood empirically, e.g. without the benefit of our specialized social processing capabilities. In such cases, our anticipation suffers, but it suffers even more if we force bad analogies and continue to use our social processing capabilities in inappropriate circumstances.

All curiosity exists to destroy itself; there is no curiosity that does not want an answer.

Vassar, it seems important to you that you not be able to understand certain acts - a badge of pride. I don't think I'm having trouble understanding an honor-killing. Someone else rapes your sister, it stains the family honor, she has to die, QED. It's not the way I think, but that doesn't stop me from modeling it.

In proof of this, I ask you, what virtuous mode of thought, or even mode of thought that you are not particularly indignant at, do you think yourself unable to understand across cultures?

I have thought of another sort of belief that is not an anticipation-controller. Sometimes, I hear quite smart young people (who don't just wear beliefs as attire) profess to a belief in physicalism about qualia, or in libertarianism, or in the virtues of the scientific method, or in anti-pseudoscience (a la Martin Gardner), or in global-warming skepticism (a la Bjorn Lomborg), or in consequentialist egoism, or some similar broad philosophical or political doctrine. When I talk to these people, I find out that they can give a number of good arguments for why someone should follow their position, but that they have little to say in response to arguments for why people should follow alternative positions. For example, they might be able to clearly state various arguments for libertarianism, and to respond well to counters to those arguments. Yet when I tell them various arguments in favor of alternative positions (e.g. democratic socialism), their attempted rebuttals are much weaker in quality than their positive arguments for the position they claim to hold.

This usually occurs because these people have read good books or articles advocating physicalism, libertarianism, and egoism, and have learned (and been convinced by) the arguments contained therein. After reading these books, these people want to talk to others about what they've read and show that they can understand and reconstruct difficult arguments. They could just say to their acquaintances something like "I've read this book and I'd like to discuss some of the arguments in it with you". But, for various reasons, they often instigate such conversations by saying:
"I don't think there should be any income tax at all. nor any other taxes. I'm a libertarian."
From here, a heated discussion may ensue about the merits of libertarianism, in which the neophyte can relate all his carefully reconstructed arguments to his audience. This allows the new libertarian to look clever (since he can relate good arguments) and well-read (since he can quote Nozick's views on politics). It also provides the libertarian with practice in thinking and arguing on the spot, and in articulating difficult ideas.

I don't count this as belief as anticipation-controller. [I'll leave aside questions about whether beliefs in political or ethical doctrines can have empirically testable consequences. The point I'm making works just as well with global-warming skepticism as it does with libertarianism or dualism.] The person who calls himself a libertarian or a global warming skeptic after reading a couple of books and a few articles arguing for libertarianism or global-warming skepticism will often acknowledge (if honest) that if he'd started by reading books advocating alternative views, then he would not have come to be a libertarian or global-warming skeptic. He knows that he hasn't made an attempt to hear views from both sides of contentious issues, despite their being very smart and thoughtful people advocating opposing.

Yet this lack of balance in his reading is not a problem for him. He is not actually going to act on his belief in any serious way. He's not about to give his money and time to support global-warming skepticism or libertarianism. He would probably not bet money on these doctrines being true, unless the bet was a small enough fraction of this wealth that it would be worthwhile to garner more attention. When he says "I'm a libertarian", what he means is "I can articulate a number of good arguments for libertarianism, along with replies to common objections to these arguments".

I think that some professional philosophers hold beliefs in a similar way. The philosopher might come up with some clever arguments for position X. Instead of writing up a paper or blog post that simply relates these clever arguments, he will probably write an article or book that gives lots of arguments for position X, including his new clever arguments (e.g. on anti-dualism). In spending lots of time studying all the good arguments for X (and in refining his own clever arguments), he will end up with an impressive ability to make arguments in support of X. Yet he probably won't have given the same open-minded and lengthy study to arguments in support of alternatives to X (i.e. not-X). Hence, when he says to people that he is an Xist or that he believes in X, what he means is "I can give lots of really sophisticated arguments for X".

(This is not true of all philosophers, as some seem to strive to criticize their own positions. Also, it is not just philosophers that are guilty of this. Some scientists will spend their lives doing experiments that provide evidence for some view X, and they won't have invested as much time in learning about the experiments and arguments of people who have been trying to show that not-X.)

Bob, I take it you're not the deceased kiwi atmospheric scientist Robert "Bob" Unwin. But very high quality commentary. I hope that you start an blog to consolidate your observations under this name/pseudonym (as I have done with mine).

Bob: Great post.

Eliezer: I was not saying anything cannot be understood, but rather that using our specialized "empathic" capabilities for understanding human behavior in terms of our own hypothetical behavior is counterproductive to understanding many instances of human behavior when the humans in question are from different cultures or otherwise very different from one's self. It's easy to model it, possibly even to model it well (Chronicles of a Death Foretold by Gabriel Garcia Marquez tries to), but next to useless to model it by reference to your own feelings. If we couldn't model it at least somewhat we couldn't even form the concept and talk about it, but we don't have the special advantages here that we have when modeling a hungry person eating or the like.

For a less politically charged example, fairly young (7 or 8, maybe 9?) could learn to model sexual desire, but they can't empathize. Adults cannot empathize with a child's enjoyment of TV shows targeting young children *even though they have been children and may remember watching the same shows* and enjoying them. Actually, since enjoyment is virtuous, that example answers your question, though across ages rather than across cultures. For 'virtuous', cross cultural, and impossible to empathically model (ignoring that to some degree there is a contradiction here, as without the ability to model the state it is hard to be confident of its intrinsic virtuousness, only of the virtuousness of the actions it brings about, which could also have been pursued for utilitarian or other deeply generally human reasons such as caring. The symmetry with abhorrent actions is broken in this respect) there are surely many different types of meditation or other altered mental states, enjoyment of a vast number of foods (I can understand liking kumiss via a "comfort food" schema, but not *kumiss AS kumiss), entertainments, and art forms, and probably more subtle and general feelings having to do with attachment to the land, etc, though I can model these empathically to some degree.

Most generally of all, I already had given examples, in citing Haidt's 5 moral domains.

Hmm. Criticism of Haidt's theory. Haidt, and most other people, probably see conservatives and liberals as having equal lack of understanding of one another. (the point of his theory is that he lacks such empathic understanding, hence the need of an empirically derived theory). However, his theory suggests that conservatives should easily understand liberals. The magnitude of one's disagreement shouldn't be the cause of empathy failure. Rather, empathy failure should follow from the apparent pointlessness of the action being criticized. For instance, it's probably easy for us to empathize with Joseph Mengele's actions while still strongly disapproving, as scientific curiosity is a shared motivation and the difference between his actions and actions we would approve of is because of his not applying enough weight to caring/sympathy considerations that we consider important. The opposite is true of one's experience reading about Isaac Bashevais Singer's father in "My Father's Court" or other good works of anthropology dealing with rich cultures. We are bemused by the apparent pointlessness of all of the ritual details that this silly man takes so very seriously, but he is harmless and we are not indignant at all.

Bob, a very high quality comment, but at 800 words it is too long for a comment. Please everyone, let's try to keep comments under about 400 words - longer items should be their own post or essay somewhere, which you can of course link to in a comment.

Re: the Alabama bar, when that same criticism was leveled by Neil Young, the response was, "A Southern man don't need him around anyhow". Apart from the fact that it came in the form of hit song, the reply is notable in that it's not something along the lines of, "them's fightin' words!" Though you may be right about the South's religious and political attitudes, I think you misunderstand how and when violence is used in that culture.

Anyway, back to the issue. The mindset of Mohamet Atta, et al, was elegantly described by Eric Hoffer in _The True Believer_. I don't believe it takes any unusal emotional insight to understand Atta's psychology, if it's seen in those terms.

Konrad: Not to repeat myself yet again, but no, understanding psychology never requires unusual emotional insight. It takes analytical ability, but it gives a different type of understanding from that which emotional insight gives.

Bob wrote "The person who calls himself a global warming skeptic... after reading a couple of books and a few articles arguing for [such skepticism] will often acknowledge that if he'd started by reading books advocating alternative views, then he would not have come to be a global warming skeptic..." This is one mechanism, but sometimes positions just "feel right" to people, i.e. in agreement with their predisposed visions, or traits.

Also it seemed to me that by asking of people that they examine as many arguments opposed to their view as they examine in alignment with their view, you would also be demanding a similar objectivity from scientists. But as has been said often, scientists are only human. They pursue their hunches (conjectures); and natural selection knew what it was doing when it made all of us normally tend to do the same.

This is not to strongly discount a goal of overcoming bias, but is to confirm a point doubtlessly made here before, that not only does bias exist for a reason but can in many instances be optimal for achievement or survival. Admittedly, truth seeking and achievement may be at odds with one another at times.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30