I think Eliezer is missing the main cause of the uproar in cases like this. The stance of the uproarers is not that “If this was true, it would be horrible, so let’s not believe it.” It’s more like, “believing this is true would cause people to act horribly, so let’s not believe it.”
Claims of innate racial and sexual differences in intelligence have historically been baseless rationalizations that attempt to justify oppressing the group in question. So now when anyone raises the question, they are shouted down because they are tarred with the same brush. The objectors are not saying that if true group intelligence differences would be worse than individual intelligence differences. but that saying there are group differences is worse than saying that there are individual differences, because while individual differences clearly exist, group differences probably don’t and are usually postulated by people who are motivated to invent them. This may be irrational, but not in the way this post focuses on.
because while individual differences clearly exist, group differences probably don’t
Just not true. And obviously not true at that. Was this presented as “one of the crazy beliefs that some insane people have” or as your own position? Hard to keep track in there.
Group differences not existing would be such an overwhelmingly improbable occurrence that it would prompt me to second guess my atheism. The universe isn’t fair. Things just don’t go around being equal to each other without good reason.
It was meant as a joint position of the insane people and myself, but on further consideration I’m abandoning it.
However, I don’t think it’s that unlikely that e.g. racial differences are fairly minimal if they exist at all, at least in terms of genetic rather than cultural/environmental/whatever differences. To the best of my knowledge, races aren’t all that distinct on a genetic level, so I wouldn’t call it “overwhelmingly improbably” that they would turn out to be close to indistinguishable in terms of intelligence.
That might be wishful thinking at play, but it seems sound to me. Not to say that it’s not worth doing a serious investigation of the possibility that there really are such differences.
TTBOMK a grand total of two (2) men whose ancestry is not predominantly West African have ever run 100m in less than ten seconds. If you can come up with some good reasons why selection for g wouldn’t have ancestral group differences that strong I’d be interested to hear them.
If you can come up with some good reasons why selection for g wouldn’t have ancestral group differences that strong I’d be interested to hear them.
I would think that different climates would select different physiologies more intensely than they would select for different “g”—I can think areas where running is a greater advantage, and areas where swimming is a greater advantage, and different musculatures may account for each… but in pretty much all areas generic intelligence is an advantage.
This doesn’t preclude the possibility for differences, but it’s a reason why the differences wouldn’t be as strong.
I can think areas where running is a greater advantage, and areas where swimming is a greater advantage, and different musculatures may account for each… but in pretty much all areas generic intelligence is an advantage.
This doesn’t preclude the possibility for differences, but it’s a reason why the differences wouldn’t be as strong.
No, I don’t think we can really say that unfortunately.
I. Opportunity cost.
To elaborate, even if g is perfectly equally useful in all geographic regions, if other selection criteria vary you can still get pretty strong selection pressures that effect intelligence (say something as simple as heat regulation of the brain or say a different rate of babies surviving birth in regions where fewer parasites are adapted to humans). Pleiotropy also means that these sorts of things may not always be apparent.
II. Speed of adaptation.
Also even given perfectly equal selection pressures on all dimensions, one would expect isolated populations depending on their size to adapt faster or slower to a new equilibrium (depending on which theories you espouse).
You misdescribe your link and so are off by a factor of 8. Your list is of the ~10 people, all west African, who have set records faster than 10s. Here is a list of 80 people to have run that fast. Immediately above that on the page is a list of the 2 or 3 non-Africans and the 2 south Africans to have run that fast.
“believing this is true would cause people to act horribly, so let’s not believe it.”
That’s not just delusional, it’s deluded.
“believing this is true would cause people to act horribly, so let’s not believe that we believe it,” would be merely delusional, and hence less objectionable.
I mostly agree with you; I was just stating my impression of the attitudes of those raising the objections in the first place (note the quotation marks). And to be fair to them, it’s really more, “believing this would cause other people to act horribly, so let’s keep them from believing it.”
It’s more that racism is unfair in a different way than people simply being different from each other. People don’t get upset that some people are cleverer than others because it’s fundamentally different from the unfairness of perfectly competent people having their opportunities crushed because of unfair things that people actually do on purpose. They’re fundamentally different kinds of unfairness, and that’s why they provoke fundamentally different responses in people.
I’m confused that this wasn’t more obvious when this was posted. I’m usually not struck by how obviously wrong something in the Sequences is, and I’m unsure of exactly where the fault lies.
I’m not sure what I was going for there; the whole point of the Chicago resume study was that racist outcomes happened even when nobody involved set out to do racist things. I think I meant “unfair things that people do”, as opposed to unfair things that simply happen.
It’s still not obvious to me. Can you unpack a little here?
What is most confusing to me about your comment is what it means for two examples of unfairness to be fundamentally different, rather than… um… non-fundamentally different (superficially different?). For example, is there some framework I can use to determine whether, say, treating able-bodied people differently from handicapped people is also fundamentally different from treating black people differently from white people, or is fundamentally the same and merely superficially different, or is not different at all, or...?
I’ll try to unpack that, especially since the original post was so sloppy.
The original post said “But why is it that the rest of the world seems to think that individual genetic differences are okay, whereas racial genetic differences in intelligence are not?”. This is incorrect. “The rest of the world” seems to think that individual genetic differences in intelligence exist and are meaningful, whereas racial genetic differences do not. (Does Yudkowsky really think that anti-racist activists believe that black people are inherently less intelligent and that that fact should be ignored? I’m not sure how else to read that.)
The unfairness of individual differences in intelligence is that, well, it happens and it’s unfair. The unfairness of racial differences in intelligence is that they don’t exist, but people act as though they do, and it’s that second part that’s unfair. These are two different kinds of unfairness. For the first, changing people’s minds won’t do a darned thing; it’s like trying to persuade water to run uphill. For the second, changing people’s minds will, following the above logic, reduce the amount of unfairness in the world, because the source of the unfairness is in those people’s minds.
So, treating able-bodied people differently from disabled people is… well, it depends. If you treat someone in a wheelchair like they can’t walk, that’s the first kind. (They actually can’t walk; it’s not anyone’s fault.) If you treat someone in a wheelchair with no obvious signs of mental impairment like they have impaired intelligence, that’s the second kind. (They’re just as clever as anyone else; the unfairness is entirely in the way you’re treating them.)
I agree with you that there’s a difference between treating groups differently based on differences that actually exist, and based on differences that don’t actually exist, and that the second thing involves a kind of unfairness that’s different from the first thing.
That said, it does seem to me that a lot of people not only think that racial genetic differences in intelligence don’t exist, they also think that if racial genetic differences in intelligence did exist, that would be a bad thing, in a way that they don’t think that the existence of individual genetic differences in intelligence are a bad thing.
One thing to say about the study you cited is that I don’t think it was conducted well. The names chosen were steeped in confounding status effects. The “white names” (“Emily” and “Brendan”) were high status and they didn’t include white low status names like Seth, Clint, Cody, Angel, Neveah...I’d better stop, I’m having way too much fun with that list. The “black names” were not high status (“Lakisha” and “Jamal”) and they didn’t choose available black higher status names like Jasmine (sic), Andre, Jeremiah, or Xavier.
A minor nitpick—this isn’t just about perfectly competent people, the study interestingly found a constant relationship between interviews for both perfectly competent white-named/black-named and incompetent white-named/black-named people, with employers 1.5 times as likely to take chance on a poorly qualified “white-named” person as “black-named” person just as they are 1.5 times more likely to give an interview to a qualified “white-named” as “black-named” person.
They’re fundamentally different kinds of unfairness, and that’s why they provoke fundamentally different responses in people.
I think this is incontrovertibly true if reversed, but not as it is; they provoke different responses in people and that effects how we should treat each kind of unfairness, but I’m not sure that aside from that they are so different.
One person is born brilliant, ugly and fat, another good-looking and of average intelligence. Both of the same race, gender, propensity to work hard etc. Both work just as hard. They are given the same scores on their oral exams, do just as well in interviews, and so on. Both do just as well performing their job because the good-looking one does better on collective projects. It’s unfair that the first isn’t rewarded for his or her intelligence or given more opportunities, and this is because of his or her poor appearance, but he or she didn’t earn or deserve his or her intelligence in the first place.
I could easily be persuaded to support treating the different cases of unfairness differently on the mere grounds that humans feel they are different, if an intelligent way to treat them differently is articulated.
From the paper: “We find little evidence that our results are driven by employers inferring something other than race, such as social class, from the names.” Section 5 deals with this; “Carrie” and “Neil” (low-status white) do just as well as “Emily” and “Geoffrey”, while “Kenya” and “Jamal” (high-status black) do just as poorly as “Latonya” and “Leroy”.
A minor nitpick—this isn’t just about perfectly competent people
Absolutely—I should have said “equally competent” or “reasonably competent”.
I don’t have a particularly strong opinion on your example, though; I’ve rolled it around in my head a bit and can’t quite see how to fit it into the same framework. There are, I believe, organizations and affinity groups advocating for better treatment of fat people, at least. I don’t perceive ‘ugly’ or ‘fat’ as being the same sort of grouping as race, though, and I’m not sure where the difference comes from, exactly.
There was a weakness in the method, though. In appendix table one they not only show how likely it actually is that a baby with a certain name is white/black, they show the results from an independent field survey that asked people to pick names as white or black. In table eight, they only measure the likelihood someone with a certain name is in a certain class (as approximated by mother’s education). Unfortunately, they don’t show what people in general, or employers in particular, actually think. If they don’t know about class differences between “Kenya” and “Latonya,” or the lack of one between “Kenya” and “Carrie,” they can’t make a decision based on class differences as they actually are.
I think Eliezer is missing the main cause of the uproar in cases like this. The stance of the uproarers is not that “If this was true, it would be horrible, so let’s not believe it.” It’s more like, “believing this is true would cause people to act horribly, so let’s not believe it.”
Claims of innate racial and sexual differences in intelligence have historically been baseless rationalizations that attempt to justify oppressing the group in question. So now when anyone raises the question, they are shouted down because they are tarred with the same brush. The objectors are not saying that if true group intelligence differences would be worse than individual intelligence differences. but that saying there are group differences is worse than saying that there are individual differences, because while individual differences clearly exist, group differences probably don’t and are usually postulated by people who are motivated to invent them. This may be irrational, but not in the way this post focuses on.
Just not true. And obviously not true at that. Was this presented as “one of the crazy beliefs that some insane people have” or as your own position? Hard to keep track in there.
Group differences not existing would be such an overwhelmingly improbable occurrence that it would prompt me to second guess my atheism. The universe isn’t fair. Things just don’t go around being equal to each other without good reason.
Sorry for the confusion.
It was meant as a joint position of the insane people and myself, but on further consideration I’m abandoning it.
However, I don’t think it’s that unlikely that e.g. racial differences are fairly minimal if they exist at all, at least in terms of genetic rather than cultural/environmental/whatever differences. To the best of my knowledge, races aren’t all that distinct on a genetic level, so I wouldn’t call it “overwhelmingly improbably” that they would turn out to be close to indistinguishable in terms of intelligence.
That might be wishful thinking at play, but it seems sound to me. Not to say that it’s not worth doing a serious investigation of the possibility that there really are such differences.
TTBOMK a grand total of two (2) men whose ancestry is not predominantly West African have ever run 100m in less than ten seconds. If you can come up with some good reasons why selection for g wouldn’t have ancestral group differences that strong I’d be interested to hear them.
I would think that different climates would select different physiologies more intensely than they would select for different “g”—I can think areas where running is a greater advantage, and areas where swimming is a greater advantage, and different musculatures may account for each… but in pretty much all areas generic intelligence is an advantage.
This doesn’t preclude the possibility for differences, but it’s a reason why the differences wouldn’t be as strong.
No, I don’t think we can really say that unfortunately.
I. Opportunity cost.
To elaborate, even if g is perfectly equally useful in all geographic regions, if other selection criteria vary you can still get pretty strong selection pressures that effect intelligence (say something as simple as heat regulation of the brain or say a different rate of babies surviving birth in regions where fewer parasites are adapted to humans). Pleiotropy also means that these sorts of things may not always be apparent.
II. Speed of adaptation.
Also even given perfectly equal selection pressures on all dimensions, one would expect isolated populations depending on their size to adapt faster or slower to a new equilibrium (depending on which theories you espouse).
It’s potentially misleading to quote a statistic like that in isolation without describing the base rate. A quick scan seems to imply that nine men whose ancestry is predominantly West African have ever run 100m in less than ten seconds… which certainly seems to support your point anyway, since less than nine elevenths of the Olympic talent pool is from West Africa.
You misdescribe your link and so are off by a factor of 8. Your list is of the ~10 people, all west African, who have set records faster than 10s. Here is a list of 80 people to have run that fast. Immediately above that on the page is a list of the 2 or 3 non-Africans and the 2 south Africans to have run that fast.
That’s not just delusional, it’s deluded.
“believing this is true would cause people to act horribly, so let’s not believe that we believe it,” would be merely delusional, and hence less objectionable.
I mostly agree with you; I was just stating my impression of the attitudes of those raising the objections in the first place (note the quotation marks). And to be fair to them, it’s really more, “believing this would cause other people to act horribly, so let’s keep them from believing it.”
It’s more that racism is unfair in a different way than people simply being different from each other. People don’t get upset that some people are cleverer than others because it’s fundamentally different from the unfairness of perfectly competent people having their opportunities crushed because of unfair things that people actually do on purpose. They’re fundamentally different kinds of unfairness, and that’s why they provoke fundamentally different responses in people.
I’m confused that this wasn’t more obvious when this was posted. I’m usually not struck by how obviously wrong something in the Sequences is, and I’m unsure of exactly where the fault lies.
I plan on replying to this later as it deserves a full reply and I have no time.
For now let me just say I am suspicious of perspectives on racism that have as an integral component the belief that most racism is on purpose.
I’m not sure what I was going for there; the whole point of the Chicago resume study was that racist outcomes happened even when nobody involved set out to do racist things. I think I meant “unfair things that people do”, as opposed to unfair things that simply happen.
It’s still not obvious to me. Can you unpack a little here?
What is most confusing to me about your comment is what it means for two examples of unfairness to be fundamentally different, rather than… um… non-fundamentally different (superficially different?). For example, is there some framework I can use to determine whether, say, treating able-bodied people differently from handicapped people is also fundamentally different from treating black people differently from white people, or is fundamentally the same and merely superficially different, or is not different at all, or...?
I’ll try to unpack that, especially since the original post was so sloppy.
The original post said “But why is it that the rest of the world seems to think that individual genetic differences are okay, whereas racial genetic differences in intelligence are not?”. This is incorrect. “The rest of the world” seems to think that individual genetic differences in intelligence exist and are meaningful, whereas racial genetic differences do not. (Does Yudkowsky really think that anti-racist activists believe that black people are inherently less intelligent and that that fact should be ignored? I’m not sure how else to read that.)
The unfairness of individual differences in intelligence is that, well, it happens and it’s unfair. The unfairness of racial differences in intelligence is that they don’t exist, but people act as though they do, and it’s that second part that’s unfair. These are two different kinds of unfairness. For the first, changing people’s minds won’t do a darned thing; it’s like trying to persuade water to run uphill. For the second, changing people’s minds will, following the above logic, reduce the amount of unfairness in the world, because the source of the unfairness is in those people’s minds.
So, treating able-bodied people differently from disabled people is… well, it depends. If you treat someone in a wheelchair like they can’t walk, that’s the first kind. (They actually can’t walk; it’s not anyone’s fault.) If you treat someone in a wheelchair with no obvious signs of mental impairment like they have impaired intelligence, that’s the second kind. (They’re just as clever as anyone else; the unfairness is entirely in the way you’re treating them.)
Does that clear things up a bit?
Yes, thanks for clarifying.
I agree with you that there’s a difference between treating groups differently based on differences that actually exist, and based on differences that don’t actually exist, and that the second thing involves a kind of unfairness that’s different from the first thing.
That said, it does seem to me that a lot of people not only think that racial genetic differences in intelligence don’t exist, they also think that if racial genetic differences in intelligence did exist, that would be a bad thing, in a way that they don’t think that the existence of individual genetic differences in intelligence are a bad thing.
Do you disagree?
One thing to say about the study you cited is that I don’t think it was conducted well. The names chosen were steeped in confounding status effects. The “white names” (“Emily” and “Brendan”) were high status and they didn’t include white low status names like Seth, Clint, Cody, Angel, Neveah...I’d better stop, I’m having way too much fun with that list. The “black names” were not high status (“Lakisha” and “Jamal”) and they didn’t choose available black higher status names like Jasmine (sic), Andre, Jeremiah, or Xavier.
A minor nitpick—this isn’t just about perfectly competent people, the study interestingly found a constant relationship between interviews for both perfectly competent white-named/black-named and incompetent white-named/black-named people, with employers 1.5 times as likely to take chance on a poorly qualified “white-named” person as “black-named” person just as they are 1.5 times more likely to give an interview to a qualified “white-named” as “black-named” person.
I think this is incontrovertibly true if reversed, but not as it is; they provoke different responses in people and that effects how we should treat each kind of unfairness, but I’m not sure that aside from that they are so different.
One person is born brilliant, ugly and fat, another good-looking and of average intelligence. Both of the same race, gender, propensity to work hard etc. Both work just as hard. They are given the same scores on their oral exams, do just as well in interviews, and so on. Both do just as well performing their job because the good-looking one does better on collective projects. It’s unfair that the first isn’t rewarded for his or her intelligence or given more opportunities, and this is because of his or her poor appearance, but he or she didn’t earn or deserve his or her intelligence in the first place.
I could easily be persuaded to support treating the different cases of unfairness differently on the mere grounds that humans feel they are different, if an intelligent way to treat them differently is articulated.
From the paper: “We find little evidence that our results are driven by employers inferring something other than race, such as social class, from the names.” Section 5 deals with this; “Carrie” and “Neil” (low-status white) do just as well as “Emily” and “Geoffrey”, while “Kenya” and “Jamal” (high-status black) do just as poorly as “Latonya” and “Leroy”.
Absolutely—I should have said “equally competent” or “reasonably competent”.
I don’t have a particularly strong opinion on your example, though; I’ve rolled it around in my head a bit and can’t quite see how to fit it into the same framework. There are, I believe, organizations and affinity groups advocating for better treatment of fat people, at least. I don’t perceive ‘ugly’ or ‘fat’ as being the same sort of grouping as race, though, and I’m not sure where the difference comes from, exactly.
This makes me think that you are right.
There was a weakness in the method, though. In appendix table one they not only show how likely it actually is that a baby with a certain name is white/black, they show the results from an independent field survey that asked people to pick names as white or black. In table eight, they only measure the likelihood someone with a certain name is in a certain class (as approximated by mother’s education). Unfortunately, they don’t show what people in general, or employers in particular, actually think. If they don’t know about class differences between “Kenya” and “Latonya,” or the lack of one between “Kenya” and “Carrie,” they can’t make a decision based on class differences as they actually are.