The “Grey Enlightenment” post about Scott’s “Against individual IQ worries” seems to me to have a big mistake at its heart. It quotes Scott’s comparison between “my IQ isn’t very high” and “my family isn’t rich” (both of which are disadvantages in life, neither of which is insurmountable for most purposes) and says
The mistake he is he is conflating IQ (which is biological and thus intrinsic to the individual) with parental wealth (which is not). [...] One can change their socioeconomic standing; one cannot change their IQ.
which misses the point completely. You can change your socioeconomic standing, but you can’t change your parents’ wealth. This is exactly parallel to how you can change (e.g.) how well you understand physics, or how many of Shakespeare’s plays you’ve read, or how effective a debater you are, even though you can’t change your IQ and IQ makes a difference to your ability to learn physics or Shakespeare or debating.
(I am using “IQ” as a shorthand for “overall brainpower as measured by things like IQ tests” and assuming arguendo that it is in fact true that you have negligible ability to change this other than perhaps by improving a specific IQ-test-taking ability that doesn’t transfer to other intellectual tasks.)
I guess the Greyly Enlightened One might want to say: “Sure, you can improve those things, but if your IQ isn’t high enough then some things are completely out of reach to you, whereas bad family circumstances don’t do that. If your IQ is 110 then you are never, ever going to understand string theory.” But it seems to me that (1) there are actually very few things that are completely out of reach to people of good but not outstanding intelligence, which I take it is roughly the group Scott’s mostly addressing, and (2) the information most people have and might possibly worry about is not really of the form “my coefficient of general intelligence is such-and-such” but rather “I took an IQ test once and the result was such-and-such”. To the latter I reply: Richard Feynman, 124. Scott Aaronson, 106. Yes, those results don’t give an adequate description of Feynman’s or Aaronson’s brainpower. That’s the point.
“If your IQ is 110 then you are never, ever going to understand string theory.”
I always wondered about these types of claims. String theory is just math applied in a particular way. So we can try to figure out exactly where the line is:
Can you understand calculus with “IQ 110” (I think clearly so)?
How about analysis of the complex plane?
How about linear algebra, hermitian matrices, etc?
How about group theory?
String theory stuff is very complicated, but it’s just made up of this type of math put together in a particular way. Some of it just takes time to internalize, but there is never any “magic sauce” about any of the specific parts, I don’t think.
One might say smart folks take less long to internalize, but my experience has been truly internalizing complex math is a bit of a slow process for everyone.
I think if Scott Aaronson once took an IQ test and got 106, that should tell you everything you need to know about how good this proxy is for “complex cognition stuff.”
I think folks in the rationality-sphere have a Mensa-like obsession with this number. Folks in Mensa do daily puzzles and worry about their IQ, folks in academia/industry publish papers and contribute to the intellectual conversation or create things and contribute to civilization. I humbly submit the latter is a better use of time.
Plus, a resume is a much better proxy for intelligence than IQ—more bits.
Yeah, I’m also by no means certain that hypothetical-Greyly-Enlightened-One would be right in claiming that an IQ of 110 means you’ll never understand string theory. I think it’s somewhat plausible, though; it feels to me as if there are things in mathematics that I understand quite easily, and things I can get my head around with substantial effort, and things that are just beyond me, and it feels as if the problem with the latter really is a straightforward lack of brainpower. Of course this is all just introspection, and we know how unreliable that is, and my guess is that at least some of the things I think are “just beyond me” I could in fact get my head around if I put in a heroic effort. But, still, it seems pretty damn plausible that some things are just too hard for some people to be realistically able to grasp them, and that in so far as there is any such thing as “your IQ” it’s a decent proxy for that.
(Of course there is such a thing as mathematical talent in particular areas, and there are blind spots in particular areas. I don’t think even Grey Enlightenment Guy would actually deny the possibility that someone might have a lowish measured IQ but specific extraordinary abilities that might enable them to get the hang of string theory or whatever. But—and this is not
I think it’s worth separating the propositions (1) “whatever IQ tests are trying to measure, they don’t measure it anything like perfectly” and (2) “whatever IQ tests are trying to measure, it isn’t perfectly predictive of actual ability to do intellectually difficult things”. The fact that Scott Aaronson once scored only 106 on an IQ test is strong evidence for some combination of those things but leaves it entirely open how much of it is measurement error and how much is weakness of even-ideally-measured IQ.
Strongly agree with the last section. I think a lot of interest in IQ arises from sources we should be skeptical of:
If you have reason to think your own IQ is high, thinking about IQ can feel good. (This is the main source of the Mensa problem, I take it.)
Reducing something complicated to a single numerical measure is appealing.
(I raise this one with great caution, but I do think it is part of the picture in some cases and I think Mr Grey Enlightenment might be one of them.) There is some evidence for correlations between measured IQ and race, and for whatever reason some people are favourably disposed towards anything that bolsters “traditional” views about race.
Of course for each of these there’s a corresponding bad reason for not being interested in IQ. If you’ve had your IQ measured and it didn’t come out very high, you may be motivated to downplay the importance of IQ; just as some people love to simplify and quantify, some people love to complicate and obfuscate; and if you have heard that measured IQ seems to vary with race and dislike racial prejudice, you may want to avoid thinking about IQ in case it leads to some sort of moral corruption. My guess is that the first two of these things, at any rate, are more likely to lead to overemphasis than underemphasis on IQ in the rationalsphere. (I bet the third goes both ways but is more likely to lead to underemphasis.)
The thing about race and intelligence (aside form the fact that it’s a hopelessly toxic topic) is that most folks making claims relating the two can’t possibly have the data to be confident about anything. Intelligence is complicated, very complicated. “Race” is complicated, too. I don’t have to obfuscate anything here, because genetics is inherently weird and messy, and so are brains.
So if folks sound confident, or make strong claims they are either confused or racist or both. The everburning tire fire of slatestar’s comment section (when it comes to this topic) is a prime example of what I am talking about.
I am inclined to agree, except that rather than “confused or racist” I would say “confused or strongly influenced by prior prejudices”, and note that one can have prior prejudices against racial IQ differences as well as for them. That is, if someone declares confidently that of course there are no differences between races in mental qualities, that’s evidence of bad thinking for all the same reasons as if they declare confidently that there are such differences.
A prior prejudice against racial brainpower differences is much nicer than a prior prejudice in favour of such differences, and probably people with the former kind of prejudice are mostly better people than ones with the latter kind, but letting either induce confident strong opinions about race-and-intelligence is the same kind of mistake.
most folks making claims relating the two can’t possibly have the data … So if folks sound confident, or make strong claims they are either confused or racist or both
There is a big difference between “most folks” and “all folks”.
Just because e.g. Twitter is an permanent garbage fire about most complicated topics does not mean that everyone with an opinion about one of these topics is confused (or racist).
I’m puzzled by this comment. Ilya didn’t say anything about “e.g. Twitter” (he did mention the Slate Star Codex comments section, which is a far cry from Twitter), and he didn’t say anything about “everyone with an opinion”.
I am reading “So if folks sound confident” as “So if [all/any] folks sound confident”. I believe Ilya’s point is that no one should have strong opinions on race/IQ and I disagree with it.
The “Grey Enlightenment” post about Scott’s “Against individual IQ worries” seems to me to have a big mistake at its heart. It quotes Scott’s comparison between “my IQ isn’t very high” and “my family isn’t rich” (both of which are disadvantages in life, neither of which is insurmountable for most purposes) and says
which misses the point completely. You can change your socioeconomic standing, but you can’t change your parents’ wealth. This is exactly parallel to how you can change (e.g.) how well you understand physics, or how many of Shakespeare’s plays you’ve read, or how effective a debater you are, even though you can’t change your IQ and IQ makes a difference to your ability to learn physics or Shakespeare or debating.
(I am using “IQ” as a shorthand for “overall brainpower as measured by things like IQ tests” and assuming arguendo that it is in fact true that you have negligible ability to change this other than perhaps by improving a specific IQ-test-taking ability that doesn’t transfer to other intellectual tasks.)
I guess the Greyly Enlightened One might want to say: “Sure, you can improve those things, but if your IQ isn’t high enough then some things are completely out of reach to you, whereas bad family circumstances don’t do that. If your IQ is 110 then you are never, ever going to understand string theory.” But it seems to me that (1) there are actually very few things that are completely out of reach to people of good but not outstanding intelligence, which I take it is roughly the group Scott’s mostly addressing, and (2) the information most people have and might possibly worry about is not really of the form “my coefficient of general intelligence is such-and-such” but rather “I took an IQ test once and the result was such-and-such”. To the latter I reply: Richard Feynman, 124. Scott Aaronson, 106. Yes, those results don’t give an adequate description of Feynman’s or Aaronson’s brainpower. That’s the point.
“If your IQ is 110 then you are never, ever going to understand string theory.”
I always wondered about these types of claims. String theory is just math applied in a particular way. So we can try to figure out exactly where the line is:
Can you understand calculus with “IQ 110” (I think clearly so)?
How about analysis of the complex plane?
How about linear algebra, hermitian matrices, etc?
How about group theory?
String theory stuff is very complicated, but it’s just made up of this type of math put together in a particular way. Some of it just takes time to internalize, but there is never any “magic sauce” about any of the specific parts, I don’t think.
One might say smart folks take less long to internalize, but my experience has been truly internalizing complex math is a bit of a slow process for everyone.
I think if Scott Aaronson once took an IQ test and got 106, that should tell you everything you need to know about how good this proxy is for “complex cognition stuff.”
I think folks in the rationality-sphere have a Mensa-like obsession with this number. Folks in Mensa do daily puzzles and worry about their IQ, folks in academia/industry publish papers and contribute to the intellectual conversation or create things and contribute to civilization. I humbly submit the latter is a better use of time.
Plus, a resume is a much better proxy for intelligence than IQ—more bits.
Yeah, I’m also by no means certain that hypothetical-Greyly-Enlightened-One would be right in claiming that an IQ of 110 means you’ll never understand string theory. I think it’s somewhat plausible, though; it feels to me as if there are things in mathematics that I understand quite easily, and things I can get my head around with substantial effort, and things that are just beyond me, and it feels as if the problem with the latter really is a straightforward lack of brainpower. Of course this is all just introspection, and we know how unreliable that is, and my guess is that at least some of the things I think are “just beyond me” I could in fact get my head around if I put in a heroic effort. But, still, it seems pretty damn plausible that some things are just too hard for some people to be realistically able to grasp them, and that in so far as there is any such thing as “your IQ” it’s a decent proxy for that.
(Of course there is such a thing as mathematical talent in particular areas, and there are blind spots in particular areas. I don’t think even Grey Enlightenment Guy would actually deny the possibility that someone might have a lowish measured IQ but specific extraordinary abilities that might enable them to get the hang of string theory or whatever. But—and this is not
I think it’s worth separating the propositions (1) “whatever IQ tests are trying to measure, they don’t measure it anything like perfectly” and (2) “whatever IQ tests are trying to measure, it isn’t perfectly predictive of actual ability to do intellectually difficult things”. The fact that Scott Aaronson once scored only 106 on an IQ test is strong evidence for some combination of those things but leaves it entirely open how much of it is measurement error and how much is weakness of even-ideally-measured IQ.
Strongly agree with the last section. I think a lot of interest in IQ arises from sources we should be skeptical of:
If you have reason to think your own IQ is high, thinking about IQ can feel good. (This is the main source of the Mensa problem, I take it.)
Reducing something complicated to a single numerical measure is appealing.
(I raise this one with great caution, but I do think it is part of the picture in some cases and I think Mr Grey Enlightenment might be one of them.) There is some evidence for correlations between measured IQ and race, and for whatever reason some people are favourably disposed towards anything that bolsters “traditional” views about race.
Of course for each of these there’s a corresponding bad reason for not being interested in IQ. If you’ve had your IQ measured and it didn’t come out very high, you may be motivated to downplay the importance of IQ; just as some people love to simplify and quantify, some people love to complicate and obfuscate; and if you have heard that measured IQ seems to vary with race and dislike racial prejudice, you may want to avoid thinking about IQ in case it leads to some sort of moral corruption. My guess is that the first two of these things, at any rate, are more likely to lead to overemphasis than underemphasis on IQ in the rationalsphere. (I bet the third goes both ways but is more likely to lead to underemphasis.)
The thing about race and intelligence (aside form the fact that it’s a hopelessly toxic topic) is that most folks making claims relating the two can’t possibly have the data to be confident about anything. Intelligence is complicated, very complicated. “Race” is complicated, too. I don’t have to obfuscate anything here, because genetics is inherently weird and messy, and so are brains.
So if folks sound confident, or make strong claims they are either confused or racist or both. The everburning tire fire of slatestar’s comment section (when it comes to this topic) is a prime example of what I am talking about.
I am inclined to agree, except that rather than “confused or racist” I would say “confused or strongly influenced by prior prejudices”, and note that one can have prior prejudices against racial IQ differences as well as for them. That is, if someone declares confidently that of course there are no differences between races in mental qualities, that’s evidence of bad thinking for all the same reasons as if they declare confidently that there are such differences.
A prior prejudice against racial brainpower differences is much nicer than a prior prejudice in favour of such differences, and probably people with the former kind of prejudice are mostly better people than ones with the latter kind, but letting either induce confident strong opinions about race-and-intelligence is the same kind of mistake.
Why so? Certainly one is much more socially acceptable than the other, but that’s not the usual definition of “nice”.
Is a “prior prejudice” against, say, sex differences in upper body strength “much nicer” than a prior prejudice in favour?
There is a big difference between “most folks” and “all folks”.
Just because e.g. Twitter is an permanent garbage fire about most complicated topics does not mean that everyone with an opinion about one of these topics is confused (or racist).
I’m puzzled by this comment. Ilya didn’t say anything about “e.g. Twitter” (he did mention the Slate Star Codex comments section, which is a far cry from Twitter), and he didn’t say anything about “everyone with an opinion”.
I am reading “So if folks sound confident” as “So if [all/any] folks sound confident”. I believe Ilya’s point is that no one should have strong opinions on race/IQ and I disagree with it.
Yep.
Sure, those are good reasons to suspect motivations, but not good reasons to stay away from IQ issues completely.
Ignoring IQ differences in social sciences (including economics) is likely to lead to… suboptimal conclusions and recommendations.
Crudely speaking, dumb people exist in large numbers and one should not pretend that this is not so.