« Measuring Optimization Power | Main | Efficient Cross-Domain Optimization »

October 28, 2008


where this conflict is clearest

I don't view this conflict is clearest in mate selection. (Perhaps I am less superficial than Dr. Hanson!)

I think it is clearest in applying discipline to get the things that we want. At least for me, my primary conflict in life is trying to get myself to apply constant effort to achieve my goals instead of settling for instant gratification. This struggle is constant, while you only face major mate selection decisions from time to time.

Here is a fascinating video interview with Rodolfo Llinás related to this post. In the interview, he describes being a subject in an experiment where magnetic coil was used to induce inwards leg movement.

Llinás decided beforehard that he would try to make outwards leg movement when his brain was activated. Afterwards, he said he felt like he had changed his mind and decided to move his leg inwards. The test was repeated a dozen times with the same result.

Every time he was unable to remember trying but unable to achieve the result. Instead, he feels like he changed his mind.


There's a part X of me that wants A and a (possibly overlapping) part Y that wants part X to want B instead, but in many cases the reason part Y wants part X to want B is that part Y itself wants B. If part Y happens to be the one that is speaking, it can therefore legitimately say "I want B".

Mikko, that is an exceptional link. It is related to a quarter of the posts on the whole site, and I am surprised that I have not seen something similar (or perhaps so succinct). For those who would prefer reading a page to listening for an hour, there is a PDF transcript, read all of page 10.

It seems tempting but wrong to characterize it as, "I can change your mind by putting a magnetic coil by the right part of your head." A more accurate rendering would be that your brain decides what to do before the "you" of your mind gets involved, then the "you" of your mind adopts that as "I decided." "You" are the observer, not the driver, of your brain's functioning.

On page 14, he describes something more powerful and specific: wiring the brain via 30-nanometer-thick wires through the circulatory system. He discusses direct brain interfaces and human external memory storage, but he does not swing back to link that to the experiment whereby electrical stimulation can cause actions directly and also cause the person performing those actions think, "I decided to do that."

If you will pardon me, I need to go update.

"Having a high ability to tell how smart someone is says good things about you - in general it says you are pretty smart too."

According to this article by Robert Burton (http://www.salon.com/env/mind_reader/2008/09/22/voter_choice/index.html), it doesn't really matter if you can tell HOW smart someone else is; if you generally believe others to BE smart, then you are probably smart, even if you are wrong about how smart others are. He says that the smartest people tend to overestimate the smartness of other people: "In short, smart people tend to believe that everyone else 'gets it.'"

it is not at all obvious that what we want to want is better morally that what we actually want

I really have to ask: If what you want isn't morality, and what you want to want isn't morality, then what higher criterion are you appealing to here?

Robin: You might be overestimating anyone's ability to tell how smart another person is. By talking to people, you learn where their minds are, not how fast they got there. You can't know how good you are at estimating smarts until you've seen IQ test results of people you have previously rated. A personality trait that one might have perceived as the sparkle of intelligence might very well turn out to be something else.

You can't know how smart a person is unless you actually challenge them with tasks that measure IQ, a test that is quite difficult to administer in casual conversation.

That said, in an evolutionary setting, it seems as though a well-tuned body would be more beneficial to reproduction than a well-tuned mind equipped with a problematic body. Intelligence makes a difference only in the long run, whereas a healthy body makes a difference in whether you survive the winter or giving birth or the famine or the drought.

In other words, we prefer smart people, if they have healthy bodies.

"... serving your interests, evolution may well have arranged your mind to fool others into thinking that... "

- Is there any evidence to support this hypothesis?

A potentially simpler explanation is this: our brains are a bit of a mess, evolved systems are usually full of slightly unintended and evolutionarily irrelevant side-effects, so we should expect our ability to introspect on our desires to be faulty.

On a related note, I recently saw an talk by professor Steven Stitch and found that there have been lots of studies showing that "what we want" or "what kind of behaviour we condone as ethical or condemn as unethical" varies according to seemingly irrelevant environmental factors. Some of this research is detailed in:


"Isen and Levin (1972) discovered that people who had just found a dime
were twenty-two times more likely to help a woman who had dropped some
papers than those who did not find a dime (88 per cent v. 4 per cent)."

There are some even more alarming studies that Stitch lectured about, but which are not yet published. For example, sitting at a messy, dirty disgusting desk makes people much harsher in their moral judgments than sitting at a clean desk.

I really have to ask: If what you want isn't morality, and what you want to want isn't morality, then what higher criterion are you appealing to here?

Maybe neither 'want' is supposed to win decisively since we evoleved to be somewhat muddled and maybe that's how it works best at least for mate selection. Suppose there was a pill that would make one want that which they want to want instead of the way it is now. Would you take the pill? Would you give away the pill to others who wanted it? Should people be allowed to hack their own brains like that?

Eliezer Yudkowsky: "I really have to ask: If what you want isn't morality, and what you want to want isn't morality, then what higher criterion are you appealing to here?"

- What you want to want to want?

for example: Perhaps Robin wants hot blond bimbos with big tits, wants to want erudite/rich/cultured but not necessarily physically attractive girls, and (after spending far too much time as a cynical economist) wants his second order desires to stop interfering with his first order desires because, um, that would be too "moralistic" and in any case unjustified?

Mikko and Zubon, this sounds like another great argument for "subsumption arch" and against symbolic AI/decision theory: the latter seems to be modeling something that doesn't even occur in humans. :-)

@Roko: I'm sure that's the case for someone, and whoever it is probably wishes that their mind weren't so confused...

denis bider said: You can't know how good you are at estimating smarts until you've seen IQ test results of people you have previously rated. A personality trait that one might have perceived as the sparkle of intelligence might very well turn out to be something else.

You can't know how smart a person is unless you actually challenge them with tasks that measure IQ, a test that is quite difficult to administer in casual conversation.

Re. IQ: I'm a bit confused as to why some people take IQ as being so significant. If you are interacting with someone or following their writing or projects over time, and you find that person (and/or their work) intriguing along some avenue pertaining to your interests, why would you place more of a premium on their IQ score as useful information than on what they were actually saying/doing/accomplishing? Or am I misinterpreting the above statement somehow?

In any case, I am curious as to why, upon learning someone's IQ, you would jump to the presumption that something you'd perceived as being "the sparkle of intelligence" was not, in fact, anything worth calling "intelligence".

Perhaps if the person was highly charismatic in some way that compelled you to rate their (perfectly ordinary) statements as being brilliant you might update your impression of them if you had some numbers to look at, but if a person seems genuinely insightful and/or capable in particular areas, I don't see why you'd downgrade (or upgrade) your perception of that person's intelligence based on their IQ score, unless of course you believe modern IQ tests completely account for all manifestations of "intelligence" worth having.

Going back to Robin's OP, though: personally I don't know on what basis anyone could effectively presume that they must be looking for certain apparent personality traits or cognitive powers in an effort to signal to others that they value things with high social value. Sure, you can come up on the spot with an ev-psych rationalization, but how would an individual know that they were employing it? I've done internal tests on myself along the lines of, "Would you still do this if you were alone on a desert island?" and "Would you do this if everyone else seemed to be doing it?" (so as to check for whether I'm doing something "for attention" or "just to be different"), so perhaps something like that would apply here.

Steven, yes conflicting coalitions are another reasonable way to understand this conflict.

Roko, no doubt random mess also contributes to the conflict. But I also see systematic trends in the difference that suggest a more systematic explanation.

Eliezer, I don't think I'm appealing to a higher criteria. As I discuss here, I can perceive morality, and see that it is part of but not all of what I want, and what I want to want.

Anne, it is not clear how many counterfactuals are minds have been evolved to handle well. Rare enough situations were presumably not handled very well.

Zubon, are people here familiar with Libet et al. 1983 experiments? I could not find any mentions on him in OB.

Vladimir Slepnev, :-)

More on topic:

What evidence is there that we really want to want what we think we want to want?

What if what we want to want is just some random narrative created to deceive ourselves and others of our real motives?

Like in your previous posts Robin, you seem intent on ignoring the irreducably subjective aspects of even purely physical manifestations of attractiveness. Who I am personally most physically attracted to, and who I estimate to have the looks with the highest market value, are two entirely different things. Are my judgements and those of the marketplace completely divergent? Of course not! There is significant overlap as you would expect. Nevertheless they are far from identical. There is plenty of fuzziness in the question of physical attractiveness.

But I digress. Back to the central topic.
I tend to believe that post-hoc rationalizations of *WHY* we are attracted to certain people over others are invariably self-justifying. Our culture seems to regard physical beauty as less essential to TRUE selfhood than it does qualities such as intelligence, emotional maturity and industriousness or 'work ethic.'

It hasn't always been thus. The Ancient Greeks, for example placed an extremely high value on physical beauty AND saw it as spiritually significant without feeling the need to pretend they were admiring something else as modern westerners do when they admire physical attractiveness.

In a society that is to some extent meritocratic and in which our collective economic performance is more dependent on cognitive skills, than on martial prowess, it makes sense that the picture is somewhat different.

We also publically laud those characteristics over which people feel able to exert some degree of control. People gain public respect for exhibiting those qualities that produce positive outcomes for others. If we publically celebrate hard work we might get a higher GDP. If we publically celebrate altruism we can expect more of it. Praising beauty by contrast is unlikely to increase the supply of beautiful people.

If you want to look at why people choose as they do, look to evolutionary explanations. If you want to know how they justify those choices look for cultural explanations.

Robin Said:

"We clearly put a large weight on body features, but we like to think we place more weight on other features, such as mental ones. When we see how much we actually care about bodies we are disturbed,"

I would say, "Speak for yourself." I know what a large weight I put on body features, and I am neither surprised nor disturbed by my preferences. I also suspect — in fact I know — that I am not alone in this. Have you considered that squeamishness about preferences (looks vs smarts) is almost certainly cultural conditioning?

I sense you have an interesting point/question in there. But this is an exceedingly muddled and muddy posting.

For instance: "If there is a pile of fruit and I have a short time to grab some fruit before others take them all, then if I can't see colors well I'll put less emphasize on colors in my choice."

Really? What is your evidence for this? One might just as easily say that you would strain all your attention to pick up the few, fleeting color signals, like a person strains to hear a faint noise in the middle of the night. This post seems more like some ad hoc examples thrown together to give evidence for some vague sentiments about what it's like to be a "smart" person.

I assume the fruit-selecting person in this anecdote isn't calculating where he/she stands on the spectrum of all possible fruit-selecting entities and then selecting. If I was the fruit-selector, and very sharp-eyed in color-detection (for a human), by your line of reasoning would I neglect color signals because I'm not sensitive to the infrared spectrum and therefore have cause to discount my skills?

When aggregated bundles of preferences like finding a desirable spouse (somebody you'll have sex with, build a family with, share life with, etc.) include preferences like sex that are evolutionarily ancient, there will be conflict whenever the conscious mind has preferences that differ from those that evolution hardwired in (either because the ancient desire is ill-suited to the modern world (e.g., sweet things) or because the ancient desire makes sense in isolation but must be modified when aggregated with other desires (e.g., desire to be healthy and sweet things)).

When considering aggregated preferences that don't include ancient desires like sex, food, etc., there is no conflict. We can modify the individual preferences when they are combined so that that the whole is harmonious and the conscious mind is satisfied. When something like sex or food is involved, we are powerless to alter the desire, and so conflict is inevitable. We think about the aggregate and consciously assign greater priority to a harmonious aggregation of the individual preferences, and we know that some individual hardwired preferences (sexual attraction) should be modified in the larger context (e.g., dating or marriage), but there is no way to make that happen (at least in the short term).

Freud talked about this sort of conflict as one of the chief sources of human unhappiness in Civilization and its Discontents. He talked about it more in terms of conflict between individual primitive desires and the restrictions that are placed upon us living in modern society, but I think the underlying insights are relevant. The conflict is inevitable, and there is very little we can do about it, so psychological conflict and suffering are an inevitable part of life, as the Buddhists would say.


Jerome Said:

"We also publically laud those characteristics over which people feel able to exert some degree of control. People gain public respect for exhibiting those qualities that produce positive outcomes for others. ... Praising beauty by contrast is unlikely to increase the supply of beautiful people."

This is an excellent point. Though, I might quibble with the last assertion. Certainly people can't control their genes (at least, conventionally this is the case) and therefore beauty is not influenced by praise. But many aspects of physical beauty are closely linked to behavior, an idea the Greeks themselves were keen to. And I don't necessarily man cosmetics or breast augmentation surgery. It's a revelation to realize how much control you have over your own attractiveness...

Posture is a huge one... over time, we have a huge amount of say in our posture... even scoliosis can be corrected over time. Spending time outdoors and exercising in general influences not only body-fitness but makes for a prettier face (less sallowness and baggy computer-monitor eyes). Proper diet, basic hygeine, clean clothes, etc. etc. The fact that many view these as shallow improvements does not make it so. My morality holds that it is a form of human excellence (though certainly not the only, nor the highest) to be beautiful.

"evolution may well have arranged your mind to fool others into thinking that you put more weight on smarts than you actually do"

But why wouldn't we want intelligence? Isn't it the most valuable human property?

Or is it more important to recall that sadly evolution has given men and women somewhat different interests in mating? Monkey society is of course ranked - let's consider female rank and male rank, leaving aside kin rankings for the moment. These 2 seem to interact in an interesting way. The higher my ranking among females, the higher also it seems my rank with males. Boy monkeys look to see who is most attractive, and they appear to consider not only looks, health & fertility, but also social power.

This is perhaps the origin of the incentive women have to spend so much time looking good to other women in ways that often puzzle men. And alas, intelligence is not high on the list of what women like in other women, generally; other women generally care more about relationship building skills than smarts.

Boy monkeys do not care about fashion/clothes/small talk per se; but they care that girl monkeys display them as an indication of their own high female rank. This is of course because they know in monkey world offspring inherit social rank - monkey society is unfortunately "aristocratic" in this way.

So female mating strategies must include achieving high female rank to cement their attractiveness to men even as they must out-compete each member of their female cohort to stand out to men in the first place.

As many OB commenters lament, most women do not seem to seek smarts - why not? For to do so, by Robin's own analysis, would enhance their social status and thus mating power.

Is there a hierarchy of trait appreciation, on which intelligence is relatively low? Or is it that women know that men look to their female social rank to discern their attractiveness?

In mating, Darwin observed that in many species females were more discriminating; he termed this "female choice." And we see this point of view often here on OB - "if only women weren't so picky!" then the strange shortage of sex would not vex everyone so.

We know you feel, Robin, that this is under-analyzed, because it considered silly and taken for granted. But we should note also that male choice is also possible, and in fact is usually seen in humans.

Male choice generally cuts against overt displays of female intelligence, as has been somewhat confirmed in studies on sexual competition - acting dumb is a more effective strategy for women than men. It's maybe why many girls, href="http://www.msnbc.msn.com/id/25836419/">who have equal math skills, generally stop taking math - a cultural intelligence marker - in junior high or early high school.

Perhaps I must bite an unpleasant bullet: do women have little incentive to display high
intelligence when we have it or to really value it in others? Men don't want to see too much of it on an intersexual level, and women don't value it themselves on an intrasexual level. Displaying much intelligence just doesn't do anything for women in competing for boy monkeys or in getting the female rank that helps them compete for men?

Could this be one reason why we have such a difference between we wish we wanted and what we really want despite the obvious advantages of choosing intelligence?

Happy birthday Katja!

Certainly intelligence is not completely un-weighted; even boy monkeys would be wise to pass on the girl monkey that is unable to make a living in monkey society. And to a certain extent, intelligence and physical attractiveness are correlated (rather strongly, no?). But beyond a common functioning intelligence, being too smart seems a likely liability — being the smartest monkey, and finding more fruit than the other monkeys, isn't always popular. Even if you share.

Re: IQ and appearance. Does anyone else find that the perceived Charisma of others declines as their IQ drops further away from their own? I tend to find women dramatically stupider than myself physically unattractive as well as conversationally challenged. Posture, and body language as well as style and range of facial expression all seem affected by intelligence.

However, brilliant women gain no additional attractiveness points as they pass me in the intelligence stakes.

In summary I regard intelligence as a major contributing factor to attractiveness up to including my own level after which I am indifferent to it.

What are other peoples subjective impressions on this score?

Why do I want to eat pizza and ice cream, but I want to want to eat broccoli and salmon. If I could alter my digestive system to run optimally on junk food I would not need to want to eat the healthy foods. Or if I could alter the ice cream to be as nutritious again I would not need to want to eat the healthy foods. I could also alter my brain so that my reason is never bypassed by emotion and I could make the optimal decisions necessary. That last option is more commonly known as developing a strong will power, and studies have shown that will power is a much better indicator of success in humans then intelligence or physical features. Taken to the extreme, a transhuman or an AI would not have the primal emotions at all, or at least would not be in danger of them overriding reason ever. They would have maximum will power.

When you say 'wanting to want' you are really stating that you wish your reasoning made the final decisions not your emotions. Maybe a higher order 'wanting to want to want' is morality or metaethics or whatever you want to call it--the choosing of reason over instinct.

This is perhaps the best post yet on this blog. Or at least it's my favorite since "Against Disclaimers."

Wouldn't all this be clearer if instead of talking about what you (want to)^n want, you tried to talk about the various programs that are running, each of which has a simpler directive? Most of the conflicts involving "wants that I don't want to want" seem to come from different low-level software conflicting with each other and with the higher-level software that's trying to mediate the process, sort of like System 2 tries to mediate System 1 to overcome bias.

Not claiming the *answers* are clear at all, but this seems a more straightforward, closer-to-what's-actually-going-on way to talk about the issues. "Want to want to want to..." is poetic but each of those "wants" has a different meaning and there's a massive parsing problem. Plus reminiscent of those awful "I Want You to Want Me" type song lyrics.

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