Intelligence Tests
April 11, 2013 10:40 AM Subscribe
Is Psychometric g a Myth? - "As an online discussion about IQ or general intelligence grows longer, the probability of someone linking to statistician Cosma Shalizi's essay g, a Statistical Myth approaches 1. Usually the link is accompanied by an assertion to the effect that Shalizi offers a definitive refutation of the concept of general mental ability, or psychometric g."
Myths, Sisyphus and g - "Over the years I have not encountered a single endorser of Shalizi's article who actually understands the relevant subject matter. His article is loved for its reassuring conclusions, not the strength of its arguments. I am sure many 'thinkers' resisted Darwinism, the abandonment of geocentrism, and even the notion that the Earth is a sphere, for similar psychological reasons."
Nuthin' but a 'g' thang
Myths, Sisyphus and g - "Over the years I have not encountered a single endorser of Shalizi's article who actually understands the relevant subject matter. His article is loved for its reassuring conclusions, not the strength of its arguments. I am sure many 'thinkers' resisted Darwinism, the abandonment of geocentrism, and even the notion that the Earth is a sphere, for similar psychological reasons."
Nuthin' but a 'g' thang
So I've always had the intuitive hypothesis that there are different types of intelligence; that different people tend to process information in different ways, whether due to habit or nature.also btw To Make Mice Smarter, Add A Few Human Brain Cells (previously)
But then there are all those people who say that intelligence can be boiled down to a single factor, the mysterious "g" (which I assume stands for either "general intelligence" or "gangsta"). Since this went against years of casual observation, I was somewhat pleased to see the eminent Cosma Shalizi write an essay debunking the notion of "g". But then I saw this blog post defending the notion of "g", and claiming that Shalizi makes a bunch of errors. Basically, the disagreement revolves around the question of why most or all psychometric tests and tasks seem positively correlated with each other. Shalizi points out that this correlation structure will naturally lead to the emergence of a "g"-like factor, even if one doesn't really exist; his opponent points out that if no "g" exists, it should be possible to design uncorrelated psychometric tests, which so far has proven extremely difficult to do.
The latter post, by a pseudonymous blogger calling himself "Dalliard", contains a bunch of references to psychometric research that I don't know about and have neither the time nor the will to evaluate, so I'm a bit stumped. Normally I'd leave the matter at that, shrug, and go read something else, but I realized that my intuitive hypothesis about intelligence didn't really seem to be explicitly stated in either of the posts. So I thought I'd explain my conjecture about how intelligence works.
In a nutshell, it's this: What if there are multiple "g's"? ...just imagine several dozen hyperplanes, and project them all onto one hyperplane... Remember that psychometric tests are simple mental tasks, but most of the mental tasks we do are complex, like computer programming or chess or writing. And for those tasks, learning and practice matter as much as innate skill, or more (for example, see this study about the neurology of chess players). Therefore, everyone can be "smart" in some way, if "smart" means "good at some complex mental task".
Great post, wow. I read Cosma's essay years ago, and have been citing it as gospel. I don't have the time now or the skills to evaluate the claims in the refutations, but I will try soon. Hopefully Dr. Shalizi will respond.
posted by MisantropicPainforest at 10:55 AM on April 11, 2013
posted by MisantropicPainforest at 10:55 AM on April 11, 2013
just imagine several dozen hyperplanes
What's the difference (up to isomorphism) between several dozen hyperplanes and one hyperplane with several dozen times as many dimensions?
posted by delmoi at 10:57 AM on April 11, 2013
What's the difference (up to isomorphism) between several dozen hyperplanes and one hyperplane with several dozen times as many dimensions?
posted by delmoi at 10:57 AM on April 11, 2013
From the 'g' thang link regarding the first link:
Well, I learned something today.
posted by ubermuffin at 11:16 AM on April 11, 2013 [6 favorites]
(Final note: Looking through "Dalliard's" blog, I see that most of it is an attempt to prove that black people are dumber than white people. Sigh. Depressing but hardly surprising. Needless to say, the fact that I addressed a "Dalliard" blog post is not intended as an endorsement of his views or his general interests...)That made me glance back at the author's previous articles, and... yep. Wow. Not necessarily promoting eugenics per se, but a member of the Human BioDiversity (HBD) community, which is apparently the modern-day repackaging thereof.
Well, I learned something today.
posted by ubermuffin at 11:16 AM on April 11, 2013 [6 favorites]
I certainly don't have the chops to get down in the weeds here with these two, but the Daillard article seems to make several errors of logic, at least in the first sections, that don't give me too much confidence. He quotes Shalizi as saying,
posted by OmieWise at 11:18 AM on April 11, 2013 [5 favorites]
By this point, I’d guess it’s impossible for something to become accepted as an “intelligence test” if it doesn’t correlate well with the Weschler [sic] and its kin, no matter how much intelligence, in the ordinary sense, it requires, but, as we saw with the first simulated factor analysis example, that makes it inevitable that the leading factor fits well.and then goes on to say that "Shalizi alleges that there are tests that measure intelligence “in the ordinary sense” yet are uncorrelated with traditional tests, but unfortunately he does not gives any examples." Shalizi certainly didn't allege that in the quoted section, and suggesting that he did paints him in a deliberately bad light, in the sense that it suggests that he can't really back up his own assertions. Daillard then goes on to say:
There are in fact many cognitive test batteries designed without regard to g, so we can put Shalizi’s allegations to test. The Woodcock-Johnson test discussed above is a case in point. Carroll, when reanalyzing data from the test’s standardization sample, pointed out that its technical manual “reveals a studious neglect of the role of any kind of general factor in the WJ-R.” This dismissive stance towards g is also reflected in Richard Woodcock’s article about the test’s theoretical background (Woodcock 1990). (Yes, the Woodcock-Johnson test was developed by a guy named Dick Woodcock, together with his assistant Johnson. You can’t make this up.) The WJ-R was developed based on the idea that the g factor is a statistical artifact with no psychological relevance.Perhaps I am misunderstanding here, but Shalizi isn't saying that tests are designed "with regard to g," he's saying that g is an artifact of how they are designed. These are not the same thing, so arguing as if they are doesn't really show a mistake on Shalizi's part. I really might be misunderstanding here, but it doesn't seem like Daillard understands Shalizi's argument well enough to refute it.
posted by OmieWise at 11:18 AM on April 11, 2013 [5 favorites]
What's the difference (up to isomorphism) between several dozen hyperplanes and one hyperplane with several dozen times as many dimensions?
three infinitely long pieces of string on a table (three 1-d hyperplanes in a 2-d space) are different than one infinitely large box in four dimensional space (one 3-d hyperplane in a 4-d space). they live in different universes with different dimensions.
posted by alk at 11:19 AM on April 11, 2013 [1 favorite]
three infinitely long pieces of string on a table (three 1-d hyperplanes in a 2-d space) are different than one infinitely large box in four dimensional space (one 3-d hyperplane in a 4-d space). they live in different universes with different dimensions.
posted by alk at 11:19 AM on April 11, 2013 [1 favorite]
Not necessarily promoting eugenics per se, but a member of the Human BioDiversity (HBD) community, which is apparently the modern-day repackaging thereof.
The fact that Steve Sailer shows up and that Dalliard actually engages with him tells me all I need to know about the post.
posted by asterix at 11:21 AM on April 11, 2013 [2 favorites]
The fact that Steve Sailer shows up and that Dalliard actually engages with him tells me all I need to know about the post.
posted by asterix at 11:21 AM on April 11, 2013 [2 favorites]
So what are the names of some of the critics of Cosma Shalizi's thesis who are not associated with "Human BioDiversity"?
posted by Sleeper at 11:26 AM on April 11, 2013 [2 favorites]
posted by Sleeper at 11:26 AM on April 11, 2013 [2 favorites]
"Hyperplane" often implies that it has dimension one less than the surrounding space (aka "co-dimension 1"). Each such a hyperplane is associated with its (unit) normal vector, i.e. it picks out a single direction in a multi-dimensional space. That's what choosing to use a single number for "g" does too.
posted by benito.strauss at 11:26 AM on April 11, 2013 [1 favorite]
posted by benito.strauss at 11:26 AM on April 11, 2013 [1 favorite]
(Final note: Looking through "Dalliard's" blog, I see that most of it is an attempt to prove that black people are dumber than white people. Sigh. Depressing but hardly surprising. Needless to say, the fact that I addressed a "Dalliard" blog post is not intended as an endorsement of his views or his general interests...)
and
The fact that Steve Sailer shows up and that Dalliard actually engages with him tells me all I need to know about the post.
This is pretty much the entire problem with this conversation on the Internet, in a nutshell.
Back in the generally interesting and useful early days of reddit, when the site was populated mostly by fairly smart and frequently self-educated tech industry types, a certain species of creepily obsessive "scientific" racism would eventually poison pretty much any given conversation about intelligence, stats, poverty, evolution, etc. I learned to assume, as a consequence, that pretty much anybody arguing for g as a concept would turn out to be a barely-camouflaged white nationalist and a total dickbag to boot.
This is unfortunate, since it seems likely that there's a useful concept of generalized cognitive capacity that it'd be nice if someone without an axe to grind could say useful things about to laypeople like me, but there you go. There's always an axe.
posted by brennen at 11:31 AM on April 11, 2013 [12 favorites]
and
The fact that Steve Sailer shows up and that Dalliard actually engages with him tells me all I need to know about the post.
This is pretty much the entire problem with this conversation on the Internet, in a nutshell.
Back in the generally interesting and useful early days of reddit, when the site was populated mostly by fairly smart and frequently self-educated tech industry types, a certain species of creepily obsessive "scientific" racism would eventually poison pretty much any given conversation about intelligence, stats, poverty, evolution, etc. I learned to assume, as a consequence, that pretty much anybody arguing for g as a concept would turn out to be a barely-camouflaged white nationalist and a total dickbag to boot.
This is unfortunate, since it seems likely that there's a useful concept of generalized cognitive capacity that it'd be nice if someone without an axe to grind could say useful things about to laypeople like me, but there you go. There's always an axe.
posted by brennen at 11:31 AM on April 11, 2013 [12 favorites]
My recollection is that Steve Hsu has been predicting that a genomic test for g is close-at-hand for the past 8 years or so. It's nice that he's finally put a concrete number on it - only 10 years to go.
(It's really sad to look back and see the same bloggers making the exact same points over almost a decade.)
posted by muddgirl at 11:35 AM on April 11, 2013
(It's really sad to look back and see the same bloggers making the exact same points over almost a decade.)
posted by muddgirl at 11:35 AM on April 11, 2013
kliuless: Remember that psychometric tests are simple mental tasks, but most of the mental tasks we do are complex, like computer programming or chess or writing. And for those tasks, learning and practice matter as much as innate skill, or more (for example, see this study about the neurology of chess players). Therefore, everyone can be "smart" in some way, if "smart" means "good at some complex mental task".One interesting takeaway from that study is that master chess players actually use less brainpower to play than lesser players. In essence, they rely on lookup tables, while the rest of us must do much more computation of possible outcomes.
The master could hypothetically have many cognitive functions impaired, be unable to cook or drive a car or remember his nurse's name, but still win games.
So... if "g" is a measure of overall ability to adaptively overcome complex mental problems, the hypothetically impaired master would have a low g, while his opponent, with 1000:1 odds against winning the game, might have a very high "g".
Feels like an argument for an overall "g", but one that certainly runs contrary to many of our preconceptions of intelligence.
posted by IAmBroom at 11:35 AM on April 11, 2013 [2 favorites]
The problem with the whole concept of 'g' is that, as far as I know, there isn't even a coherent biological hypothesis for what it's supposed to be. More dendrites? More grey matter surface area? A physically larger brain?
If you could show that there were at least two physically real, measurable brain structures/chemistry that were both capable of boosting intelligence wouldn't that in fact completely and irrevocably disprove 'g'?
There was a lecture by the head scientist from the "Chinese designer babies" FPP where he said he expected to find thousands of genes that impact intelligence. They hadn't found a single one yet, but they had found something like seven hundreds that impacted height. If that's true, wouldn't it pretty much fundamentally disprove 'g' as well as an actual material, physical thing?
'g' almost seems more like a religious concept then anything else. I mean if you just say that 'g' is just the combination of all the tiny things that might impact intelligence, then you're not really talking about something that actually exists
Thing about it this way. In order for a car to be fast, it needs a lot of different things. Breaks, tires, an engine, transmission, a stiff, light-weight body, etc.
Yet, I'm sure if you were to simply statistically analyze the lap times and speeds of various cars, you would find a lot of correlation between different ways of measuring a cars performance.
Would that mean things like the engine and the materials a car is made out of are really just the result of some mysterious and un-observable 'r' factor?
It's pretty clear that that's a completely insane thing to think. The thing that makes a car go fast are the various parts. Obviously cars are intelligently designed, but the same thing would still be true if you had cars that had parts chosen at random.
So how is this 'g' thing any different then a religious concept? As far as I know there isn't any kind of biological hypotheses for what this is supposed to be. It can't be observed directly. People just insist it exists despite no evidence whatsoever.
Going back to the car metaphor, you can have a car that's fast because it has a huge engine, and you can have a car that's fast because it's super-light (like an Ariel Atom). In fact, engine size and light weight are actually inversely correlated.
___
Like I said, if you could find two different biological factors that actually impact test taking ability but aren't correlated wouldn't that completely disprove 'g'? If not, how is 'g' even actually a thing?
posted by delmoi at 11:39 AM on April 11, 2013 [9 favorites]
If you could show that there were at least two physically real, measurable brain structures/chemistry that were both capable of boosting intelligence wouldn't that in fact completely and irrevocably disprove 'g'?
There was a lecture by the head scientist from the "Chinese designer babies" FPP where he said he expected to find thousands of genes that impact intelligence. They hadn't found a single one yet, but they had found something like seven hundreds that impacted height. If that's true, wouldn't it pretty much fundamentally disprove 'g' as well as an actual material, physical thing?
'g' almost seems more like a religious concept then anything else. I mean if you just say that 'g' is just the combination of all the tiny things that might impact intelligence, then you're not really talking about something that actually exists
Thing about it this way. In order for a car to be fast, it needs a lot of different things. Breaks, tires, an engine, transmission, a stiff, light-weight body, etc.
Yet, I'm sure if you were to simply statistically analyze the lap times and speeds of various cars, you would find a lot of correlation between different ways of measuring a cars performance.
Would that mean things like the engine and the materials a car is made out of are really just the result of some mysterious and un-observable 'r' factor?
It's pretty clear that that's a completely insane thing to think. The thing that makes a car go fast are the various parts. Obviously cars are intelligently designed, but the same thing would still be true if you had cars that had parts chosen at random.
So how is this 'g' thing any different then a religious concept? As far as I know there isn't any kind of biological hypotheses for what this is supposed to be. It can't be observed directly. People just insist it exists despite no evidence whatsoever.
Going back to the car metaphor, you can have a car that's fast because it has a huge engine, and you can have a car that's fast because it's super-light (like an Ariel Atom). In fact, engine size and light weight are actually inversely correlated.
___
Like I said, if you could find two different biological factors that actually impact test taking ability but aren't correlated wouldn't that completely disprove 'g'? If not, how is 'g' even actually a thing?
posted by delmoi at 11:39 AM on April 11, 2013 [9 favorites]
There was an old experiment in which chess masters did no better than control group at recall of board positions made up of random placements of pieces, but dramatically better at recall of positions that come from actual games.
posted by thelonius at 11:39 AM on April 11, 2013
posted by thelonius at 11:39 AM on April 11, 2013
Like I said, if you could find two different biological factors that actually impact test taking ability but aren't correlated wouldn't that completely disprove 'g'? If not, how is 'g' even actually a thing?
I'm not sure I understand this. Why would this disprove g? I'm not disagreeing, I just don't the necessity of the conclusion.
posted by OmieWise at 11:43 AM on April 11, 2013
I'm not sure I understand this. Why would this disprove g? I'm not disagreeing, I just don't the necessity of the conclusion.
posted by OmieWise at 11:43 AM on April 11, 2013
I learned to assume, as a consequence, that pretty much anybody arguing for g as a concept would turn out to be a barely-camouflaged white nationalist and a total dickbag to boot.
This.
posted by localroger at 11:53 AM on April 11, 2013 [4 favorites]
This.
posted by localroger at 11:53 AM on April 11, 2013 [4 favorites]
The chess experiment is a classic
Chase, W. G., and Simon, H. A. (1973). The mind’s eye in chess. In W. G. Chase, ed., Visual information processing, pp. 215–281. New York: Academic Press.
and it's described here.
posted by tsuipen at 11:56 AM on April 11, 2013 [2 favorites]
Chase, W. G., and Simon, H. A. (1973). The mind’s eye in chess. In W. G. Chase, ed., Visual information processing, pp. 215–281. New York: Academic Press.
and it's described here.
posted by tsuipen at 11:56 AM on April 11, 2013 [2 favorites]
The problem with the whole concept of 'g' is that, as far as I know, there isn't even a coherent biological hypothesis for what it's supposed to be. More dendrites? More grey matter surface area? A physically larger brain?
Brain volume is directly correlated with g in adults.
But, even if your statement is taken as true, there is not a "problem" for a brain issue just because there is no coherent biological hypothesis. Migraine headaches have no clear underlying cause. I hope you would agree that migraine headaches exist.
posted by Tanizaki at 12:10 PM on April 11, 2013 [1 favorite]
Brain volume is directly correlated with g in adults.
But, even if your statement is taken as true, there is not a "problem" for a brain issue just because there is no coherent biological hypothesis. Migraine headaches have no clear underlying cause. I hope you would agree that migraine headaches exist.
posted by Tanizaki at 12:10 PM on April 11, 2013 [1 favorite]
I learned to assume, as a consequence, that pretty much anybody arguing for g as a concept would turn out to be a barely-camouflaged white nationalist and a total dickbag to boot.
I am sure you won't mind if anyone arguing for evolution is described as a barely-camouflaged racist and eugenist.
posted by Tanizaki at 12:12 PM on April 11, 2013 [1 favorite]
I am sure you won't mind if anyone arguing for evolution is described as a barely-camouflaged racist and eugenist.
posted by Tanizaki at 12:12 PM on April 11, 2013 [1 favorite]
Brain volume is directly correlated with g in adults.
They are correlated, slightly, but significantly. I don't know what directly means in this case. And they are both confounded by nutrition.
posted by MisantropicPainforest at 12:18 PM on April 11, 2013 [1 favorite]
They are correlated, slightly, but significantly. I don't know what directly means in this case. And they are both confounded by nutrition.
posted by MisantropicPainforest at 12:18 PM on April 11, 2013 [1 favorite]
I learned to assume, as a consequence, that pretty much anybody arguing for g as a concept would turn out to be a barely-camouflaged white nationalist and a total dickbag to boot.
This is very true at least for the parts of the internet I frequent, and I bet the correlation is stronger than that between g and cumulative lifetime earnings or whatever. Of course it's not a particularly good argument against g as a concept.
posted by dhoe at 12:23 PM on April 11, 2013 [1 favorite]
This is very true at least for the parts of the internet I frequent, and I bet the correlation is stronger than that between g and cumulative lifetime earnings or whatever. Of course it's not a particularly good argument against g as a concept.
posted by dhoe at 12:23 PM on April 11, 2013 [1 favorite]
They are correlated, slightly, but significantly. I don't know what directly means in this case. And they are both confounded by nutrition.
"Direct correlation" means that large values of one variable are associated with large values of the other and small with small.
posted by Tanizaki at 12:26 PM on April 11, 2013
"Direct correlation" means that large values of one variable are associated with large values of the other and small with small.
posted by Tanizaki at 12:26 PM on April 11, 2013
Tanizaki: When anyone 'arguing for evolution' has the following, exhaustive, set of blog categories:
Black-White IQ Gap
Caribbean
Colorism
Hispanic-White IQ Gap
HV Global IQ
Nota bene
Psychometrics
Transracial Adoption
Uncategorized
White-East Asian IQ Gap
(Cut and pasted verbatim)
And links approvingly to Satoishi Kanazawa and Steve Sailer, then yes, I will be super okay with calling that person a barely-camouflaged racist and eugenicist.
posted by PMdixon at 12:28 PM on April 11, 2013 [7 favorites]
Black-White IQ Gap
Caribbean
Colorism
Hispanic-White IQ Gap
HV Global IQ
Nota bene
Psychometrics
Transracial Adoption
Uncategorized
White-East Asian IQ Gap
(Cut and pasted verbatim)
And links approvingly to Satoishi Kanazawa and Steve Sailer, then yes, I will be super okay with calling that person a barely-camouflaged racist and eugenicist.
posted by PMdixon at 12:28 PM on April 11, 2013 [7 favorites]
Brain Size and Intelligence. TL;DR there is a small-to-moderate correlation to IQ, specifically
posted by muddgirl at 12:30 PM on April 11, 2013 [2 favorites]
Studies have tended to indicate small to moderate correlations (averaging around 0.3 to 0.4) between brain volume and IQ. The most consistent associations are observed within the frontal, temporal, and parietal lobes, the hippocampi, and the cerebellum, but only account for a relatively small amount of variance in IQ, which itself only shows a partial relationship to the general concept of intelligence and real-world performance... This suggests different mechanisms are involved in a full description of the physiology underlying human intelligence.Of course, not as strong or as definitive as Tanizaki has stated it.
posted by muddgirl at 12:30 PM on April 11, 2013 [2 favorites]
Tanizaki: When anyone 'arguing for evolution' has the following, exhaustive, set of blog categories:
Who cares? The merits of opinions have nothing to do with the speaker stating them. Either g exists or it does not. (for those of you wondering, the consensus is that it does). Whether or not a person holds distasteful views has no bearing on the truth value of his statements. If Darwin and Huxley were proven to be racists, would that evolution any less true? No, of course not.
Denial of general intelligence is a form of creationism.
posted by Tanizaki at 12:39 PM on April 11, 2013
Who cares? The merits of opinions have nothing to do with the speaker stating them. Either g exists or it does not. (for those of you wondering, the consensus is that it does). Whether or not a person holds distasteful views has no bearing on the truth value of his statements. If Darwin and Huxley were proven to be racists, would that evolution any less true? No, of course not.
Denial of general intelligence is a form of creationism.
posted by Tanizaki at 12:39 PM on April 11, 2013
Denial of general intelligence is a form of creationism.
what
posted by OmieWise at 12:44 PM on April 11, 2013 [6 favorites]
what
posted by OmieWise at 12:44 PM on April 11, 2013 [6 favorites]
The merits of opinions have nothing to do with the speaker stating them.
These people are proven racists. Which means they hold views which are outdated, disproven, and vile. I don't trust them to present information to me in an unbiased way, I don't trust them not to lie, I don't trust their citations, I don't trust their interpretations of what other people say, I don't trust them at all.
I don't trust them because they are people who dedicate a significant portion of their life to advancing the idea that black people are dumber than white people. And they do this gleefully.
Or, The Ad Hominem Fallacy Fallacy
posted by MisantropicPainforest at 12:46 PM on April 11, 2013 [5 favorites]
These people are proven racists. Which means they hold views which are outdated, disproven, and vile. I don't trust them to present information to me in an unbiased way, I don't trust them not to lie, I don't trust their citations, I don't trust their interpretations of what other people say, I don't trust them at all.
I don't trust them because they are people who dedicate a significant portion of their life to advancing the idea that black people are dumber than white people. And they do this gleefully.
Or, The Ad Hominem Fallacy Fallacy
posted by MisantropicPainforest at 12:46 PM on April 11, 2013 [5 favorites]
The merits of opinions have nothing to do with the speaker stating them.
See, here's the thing. The FPP isn't an opinion. It's a fact claim. It's a complicated, statistical fact claim. Before I spend the time evaluating it, I am (and you are too! and so is everybody!) going to apply some kind of heuristic to at least determine whether the claim is in good faith.
Here, I apply the following heuristic:
I take the following as facts: The brain has been shown to be incredibly plastic in response to its environment. Crazy plastic. Folks of different racial backgrounds exist in different environments. In fact, many of the differences of environment (nutritition, exposure to vocabulary, status, etc) are of exactly those kinds that the brain is known to be especially responsive to! (I also observe that the usual difference between generic black IQ test taker and generic white IQ test taker is about the same magnitude as the Flynn effect over the last century.)
I observe the following about the speaker: Despite the above, ad arguendo, facts, they're really really really interested in difference of measures of intellect between folks of different racial background being in some non-environmental way hereditary, and primarily so. To me, this seems like a pretty major violation of Occam's Razor. I've also observed that many people are interested in establishing an 'objective' measure in which they are superior to folks of certain other racial backgrounds. I also observe that this person is presumably aware of Occam's Razor.
From all of the above, I conclude that the most likely explanation, from my standpoint, is that they are arguing from a predetermined conclusion, viz, that black folks are dumb. I do not consider that to be operating in good faith, and as such do not have a particular interest in untangling their argument in its entirety.
Fruit of the poison tree, etc.
posted by PMdixon at 12:51 PM on April 11, 2013 [4 favorites]
See, here's the thing. The FPP isn't an opinion. It's a fact claim. It's a complicated, statistical fact claim. Before I spend the time evaluating it, I am (and you are too! and so is everybody!) going to apply some kind of heuristic to at least determine whether the claim is in good faith.
Here, I apply the following heuristic:
I take the following as facts: The brain has been shown to be incredibly plastic in response to its environment. Crazy plastic. Folks of different racial backgrounds exist in different environments. In fact, many of the differences of environment (nutritition, exposure to vocabulary, status, etc) are of exactly those kinds that the brain is known to be especially responsive to! (I also observe that the usual difference between generic black IQ test taker and generic white IQ test taker is about the same magnitude as the Flynn effect over the last century.)
I observe the following about the speaker: Despite the above, ad arguendo, facts, they're really really really interested in difference of measures of intellect between folks of different racial background being in some non-environmental way hereditary, and primarily so. To me, this seems like a pretty major violation of Occam's Razor. I've also observed that many people are interested in establishing an 'objective' measure in which they are superior to folks of certain other racial backgrounds. I also observe that this person is presumably aware of Occam's Razor.
From all of the above, I conclude that the most likely explanation, from my standpoint, is that they are arguing from a predetermined conclusion, viz, that black folks are dumb. I do not consider that to be operating in good faith, and as such do not have a particular interest in untangling their argument in its entirety.
Fruit of the poison tree, etc.
posted by PMdixon at 12:51 PM on April 11, 2013 [4 favorites]
From all of the above, I conclude that the most likely explanation, from my standpoint, is that they are arguing from a predetermined conclusion, viz, that black folks are dumb.
Indeed. One rarely sees such arguments with the hypothesis stated in the reverse direction, where g(black)>g(white), though a priori this is as possible as g(white)>g(black).
posted by monocyte at 12:58 PM on April 11, 2013
Indeed. One rarely sees such arguments with the hypothesis stated in the reverse direction, where g(black)>g(white), though a priori this is as possible as g(white)>g(black).
posted by monocyte at 12:58 PM on April 11, 2013
Fruit of the poison tree, etc.
In your opinion, is the last century of cognitive research the "fruit of a poisonous tree"? Do you really think that the consensus of cognitive scientists is that there is no such thing as general intelligence? Is Science now a crank hothouse journal? ("As discovered by Spearman early in the last century, measures of performance or success in diverse cognitive tests show a pattern of almost universal positive correlation: To some extent at least, the same people tend to perform well in very different tasks.")
I take the following as facts: The brain has been shown to be incredibly plastic in response to its environment.
I have no idea what this is supposed to mean, but you seem out of touch with reality in this regard. So people aren't "born that way"?
posted by Tanizaki at 1:16 PM on April 11, 2013
In your opinion, is the last century of cognitive research the "fruit of a poisonous tree"? Do you really think that the consensus of cognitive scientists is that there is no such thing as general intelligence? Is Science now a crank hothouse journal? ("As discovered by Spearman early in the last century, measures of performance or success in diverse cognitive tests show a pattern of almost universal positive correlation: To some extent at least, the same people tend to perform well in very different tasks.")
I take the following as facts: The brain has been shown to be incredibly plastic in response to its environment.
I have no idea what this is supposed to mean, but you seem out of touch with reality in this regard. So people aren't "born that way"?
posted by Tanizaki at 1:16 PM on April 11, 2013
@muddgirl
"The ironic part is that if I don't understand the relevant subject matter enough to actually endorse Shalizi's article, then I almost certainly don't understand it enough to endorse or even discuss Dalliard's refutation. Is this a fundamental issue with scientific discussions among laymen?"
Yes.
This is why we still have people who don't believe in evolution, or who think vaccines cause autism, or who think Einstein was wrong about special relativity. The scientists who are in these fields are pretty much of one mind, with relatively high rates of confusion in the general population about how strong the evidence really is or where it points.
Hell, I've got a scientific background and I'm not immune. When my first child was born, I worried that there was a link between vaccines & autism. I'm not trained in medicine or medical research, so I can't effectively weigh the evidence, but I'd heard the link had some strong evidence for this. In cases like these, I usually rely on the National Academies of Science. Having checked with them, I can now say that a bunch of people without an agenda and with all the training I'd need to assess it seem to think the data show no link.
Our collective inability to determine the winner of a scientific debate is really easy to exploit if you're a business (like coal, oil, or car companies wrt global warming) that really should have a lot of regulation. It's relatively cheap to hire a few corrupt scientists to muddy the waters, then get some corrupt PR firms to advertise the "need for more research before we start making regulations."
posted by thatnerd at 1:17 PM on April 11, 2013 [3 favorites]
"The ironic part is that if I don't understand the relevant subject matter enough to actually endorse Shalizi's article, then I almost certainly don't understand it enough to endorse or even discuss Dalliard's refutation. Is this a fundamental issue with scientific discussions among laymen?"
Yes.
This is why we still have people who don't believe in evolution, or who think vaccines cause autism, or who think Einstein was wrong about special relativity. The scientists who are in these fields are pretty much of one mind, with relatively high rates of confusion in the general population about how strong the evidence really is or where it points.
Hell, I've got a scientific background and I'm not immune. When my first child was born, I worried that there was a link between vaccines & autism. I'm not trained in medicine or medical research, so I can't effectively weigh the evidence, but I'd heard the link had some strong evidence for this. In cases like these, I usually rely on the National Academies of Science. Having checked with them, I can now say that a bunch of people without an agenda and with all the training I'd need to assess it seem to think the data show no link.
Our collective inability to determine the winner of a scientific debate is really easy to exploit if you're a business (like coal, oil, or car companies wrt global warming) that really should have a lot of regulation. It's relatively cheap to hire a few corrupt scientists to muddy the waters, then get some corrupt PR firms to advertise the "need for more research before we start making regulations."
posted by thatnerd at 1:17 PM on April 11, 2013 [3 favorites]
In your opinion, is the last century of cognitive research the "fruit of a poisonous tree"?
Does the last century of cognitive research conclude that black people are dumber than white people?
posted by MisantropicPainforest at 1:19 PM on April 11, 2013
Does the last century of cognitive research conclude that black people are dumber than white people?
posted by MisantropicPainforest at 1:19 PM on April 11, 2013
g just is the common explanatory factor across different IQ test tasks. There is no further fact that has to exist for g to exist. Scientists don't have to discover SUBSTANCE G being secreted by the pituitary. There is no rule that says that everything that's real must be neatly reducible to a simple biological explanation. Many different biological and environmental factors contribute to general intelligence; one can acknowledge this without abandoning the idea of general intelligence.
posted by grobstein at 1:23 PM on April 11, 2013 [4 favorites]
posted by grobstein at 1:23 PM on April 11, 2013 [4 favorites]
Does the last century of cognitive research conclude that black people are dumber than white people?
That is not the topic of discussion. The discussion is whether or not general intelligence is a myth. I hate to spoil the surprise, but it is not.
By the way, reality gets to be racist. Reality gets to be sexist. Reality gets to be any -ist in the world. Men have larger lung volume and women have more white blood cells. Are either of these things sexist? Who cares? It doesn't change reality.
posted by Tanizaki at 1:34 PM on April 11, 2013 [1 favorite]
That is not the topic of discussion. The discussion is whether or not general intelligence is a myth. I hate to spoil the surprise, but it is not.
By the way, reality gets to be racist. Reality gets to be sexist. Reality gets to be any -ist in the world. Men have larger lung volume and women have more white blood cells. Are either of these things sexist? Who cares? It doesn't change reality.
posted by Tanizaki at 1:34 PM on April 11, 2013 [1 favorite]
Tanizaki,
Can you please point me to some sort of scholarly, but readable writing by anyone who is not affiliated with any racist institute that debunks, criticizes, or rebuts Cosma Shalizi's argument that g is a statistical myth? Preferably someone respected and with some serious academic credentials.
I don't have the skills or knowledge to parse a lot of this stuff, so I'm not going to trust someone with suspect credentials or motives.
If you should point me to someone with the same or similar levels of credentials as Shalizi who disagrees with him I would really really appreciate it. Thanks.
posted by MisantropicPainforest at 1:40 PM on April 11, 2013 [1 favorite]
Can you please point me to some sort of scholarly, but readable writing by anyone who is not affiliated with any racist institute that debunks, criticizes, or rebuts Cosma Shalizi's argument that g is a statistical myth? Preferably someone respected and with some serious academic credentials.
I don't have the skills or knowledge to parse a lot of this stuff, so I'm not going to trust someone with suspect credentials or motives.
If you should point me to someone with the same or similar levels of credentials as Shalizi who disagrees with him I would really really appreciate it. Thanks.
posted by MisantropicPainforest at 1:40 PM on April 11, 2013 [1 favorite]
In your opinion, is the last century of cognitive research the "fruit of a poisonous tree"? Do you really think that the consensus of cognitive scientists is that there is no such thing as general intelligence?
I dunno. If I were interested, I would go do some digging. More into python and distributed computing at the moment, so I won't.
I'm talking about the FPP. That piece is, given it's context, (let's be honest) clearly motivated by the goal of demonstrating that white folk are smarter than black folk. As such, I'm really not interested in further analysis of the FPP.
I have no idea what this is supposed to mean, but you seem out of touch with reality in this regard. So people aren't "born that way"?
Plastic: capable of being shaped or formed, malleable. The usual extreme example is Phineas Gage, the guy who got an iron rod thru his skull. The fact that people survive lobectomies can also be taken as (extreme) examples of this. More prosaically, the rapid pruning of neural connections in toddlers, etc.
Shouldn't feed a derail, but: I assume the last is supposed to be a jab at my (assumed) belief in a genetic basis for homosexuality? As a (formerly card-carrying) gay man, I'll first say that "born this way" is a politically convenient oversimplification (and a terrible song), and further that it seems to me that human sexuality in general is a big mess of genetic, epigenetic, environmental, social, cultural factors. (Can't have a shoe fetish with shoes, eg.) I would say that my own lived experience involves neither any sort of early recognition, nor element of choice. So no, to a first approximation, [most] people aren't born any particular way.
posted by PMdixon at 1:42 PM on April 11, 2013
I dunno. If I were interested, I would go do some digging. More into python and distributed computing at the moment, so I won't.
I'm talking about the FPP. That piece is, given it's context, (let's be honest) clearly motivated by the goal of demonstrating that white folk are smarter than black folk. As such, I'm really not interested in further analysis of the FPP.
I have no idea what this is supposed to mean, but you seem out of touch with reality in this regard. So people aren't "born that way"?
Plastic: capable of being shaped or formed, malleable. The usual extreme example is Phineas Gage, the guy who got an iron rod thru his skull. The fact that people survive lobectomies can also be taken as (extreme) examples of this. More prosaically, the rapid pruning of neural connections in toddlers, etc.
Shouldn't feed a derail, but: I assume the last is supposed to be a jab at my (assumed) belief in a genetic basis for homosexuality? As a (formerly card-carrying) gay man, I'll first say that "born this way" is a politically convenient oversimplification (and a terrible song), and further that it seems to me that human sexuality in general is a big mess of genetic, epigenetic, environmental, social, cultural factors. (Can't have a shoe fetish with shoes, eg.) I would say that my own lived experience involves neither any sort of early recognition, nor element of choice. So no, to a first approximation, [most] people aren't born any particular way.
posted by PMdixon at 1:42 PM on April 11, 2013
By the way, reality gets to be racist. Reality gets to be sexist. Reality gets to be any -ist in the world. Men have larger lung volume and women have more white blood cells. Are either of these things sexist? Who cares? It doesn't change reality.
That's interesting, Tanizaki. So you're just a pure-minded lover of reality and truth, unlike all these other commenters with their fuzzy moral concerns and tribal identities. So why don't you stick to making factual statements and linking to careful arguments that support these facts? So far you've provided "Brain volume is directly correlated with g in adults" and "Denial of general intelligence is a form of creationism" but neither of these are supported by any argument or evidence, or even a link to an argument or evidence. You are just making pronouncements with an air of authority, all while crying that people are too concerned with the personalities behind claims instead of the facts. Please.
posted by leopard at 1:46 PM on April 11, 2013 [2 favorites]
That's interesting, Tanizaki. So you're just a pure-minded lover of reality and truth, unlike all these other commenters with their fuzzy moral concerns and tribal identities. So why don't you stick to making factual statements and linking to careful arguments that support these facts? So far you've provided "Brain volume is directly correlated with g in adults" and "Denial of general intelligence is a form of creationism" but neither of these are supported by any argument or evidence, or even a link to an argument or evidence. You are just making pronouncements with an air of authority, all while crying that people are too concerned with the personalities behind claims instead of the facts. Please.
posted by leopard at 1:46 PM on April 11, 2013 [2 favorites]
Human BioDiversity (HBD)"Human BioDiversity" people aren't eugenicists, they are racists. They don't want to control the gene pool, they just want people to stop being so PC and admit that black people really are genetically inferior as they they keep telling everyone they are.
Speaking of which, there seems to be a very high correlation between people obsessed with 'g' and racists. Charles Murray is a good example, Steve Sailer, etc. Other people as well.
I'm not sure I understand this. Why would this disprove g? I'm not disagreeing, I just don't the necessity of the conclusion.It would prove there is no "one thing" which forms the basis for intelligence, which is what g is supposed to be. As far as I can tell.
I also don't see how you even call a hypothesis scientific if you can't even come up with a plausible biological or physical explanation for what it is, unless it has to do with quantum physics. Is that what you think it is? Is 'g' caused by 'g' particles permeating 'g' fields that can only be characterized using quantum statistics? Penrose style quantum brain?
Because if 'g' is not a physical thing, then how is it even science? And if 'g' is a physical thing then the existence of any other physical thing with a similar effect on intelligence would prove that 'g' is not the underlying physical basis for intelligence.
Without a physical hypothesis 'g' doesn't even need to be disproven, it's not even a coherent idea.
Migraine headaches have no clear underlying cause. I hope you would agree that migraine headaches exist.Migraine headaches are directly observable by the person experiencing them. 'g' is not.
I am sure you won't mind if anyone arguing for evolution is described as a barely-camouflaged racist and eugenist.So? I don't assume people who drive Ferraris are pedophiles or that people who read the NYT like to murder hookers either. I'm not sure what that has to with the fact obsession with 'g' highly correlated with being a racist.
Brain volume is directly correlated with g in adults.So you think on average women, who's main brain size is 1131.1cc brains are stupider then men who's brains are 1273.6cc?
Anyway, you don't seem to actually understand. You can't say that something is correlated with 'g' if you can't prove 'g' even exists. All you can say is that they are correlated with doing well on IQ tests. But 'g' is supposed to be some kind of underlying root cause for all the things that are correlated with doing well on IQ tests.
"Direct correlation" means that large values of one variable are associated with large values of the other and small with small.Wow, that was a breathtakingly innumerate statement. MisantropicPainforest was talking about statistical significance and the size of the effect. They obviously know the difference between direct and inverse correlations.
Who cares? The merits of opinions have nothing to do with the speaker stating them.Who cares if they're racist? Racism a manifestation of intellectual degeneracy, an inability to think logically or rationally. There is a finite amount of time, only a finite number of things to read and listen too. Why waste it on obvious shit, leaking out of the, vile, twisted, sewer-like minds of a racists?
(Other then to mock them, of course)
Either g exists or it does not. (for those of you wondering, the consensus is that it does).The consensus of racist bloggers, sure.
("As discovered by Spearman early in the last century, measures of performance or success in diverse cognitive tests show a pattern of almost universal positive correlation: To some extent at least, the same people tend to perform well in very different tasks.")Which doesn't even suggest a single cause. No single cause, no 'g'.
There is no rule that says that everything that's real must be neatly reducible to a simple biological explanation.There is a rule that says everything real must be reducible to physical processes in the material world. It's called "Science". It doesn't have to be 'biological', but it does have to actually exist, and be able to be measured and detected.
In other words, in order for something exist it has to, you know actually exist, as in be made out of matter, etc.
By the way, reality gets to be racist. Reality gets to be sexist. Reality gets to be any -ist in the world.Reality can only be racist if races existed in reality, which they do not. It's simply a cognitive shortcut necessary for people of limited intellectual capacity.
______
"Hyperplane" often implies that it has dimension one less than the surrounding space (aka "co-dimension 1"). Each such a hyperplane is associated with its (unit) normal vector, i.e. it picks out a single direction in a multi-dimensional space. That's what choosing to use a single number for "g" does too.Hmm, that's true I suppose. But you can have hyperplanes in orthogonal subspaces of a larger space, in which case the normal vectors could be added together to create a vector in the overall space, which could be a normal for a hyperplane in the larger space. And, if you have a bunch of separate, unrelated spaces, S1 ... Sn, with hyperplanes then that should be isomorphic to S1 ⊕ ... ⊕ Sn.
Now that I think about it, you could run into a problem if your hyperplanes are in subspaces of a larger space that are not orthogonal. However, I don't that has anything to do with what that guy was trying to say.
posted by delmoi at 1:46 PM on April 11, 2013 [3 favorites]
That is not the topic of discussion. The discussion is whether or not general intelligence is a myth. I hate to spoil the surprise, but it is not.
I feel like I may have contributed to poisoning this thread above, invoking the typical devolution of the topic by the mention of one of its typical modes. I will only say that yes, whether and how general intelligence is a meaningful/useful concept is a subject of discussion, but even the discussion contained in the FPP links was already, before this thread got started, complicated by the involvement of the usual pack of fancy-vocabulary racists, and suddenly it is very, very hard to talk about the merits of the idea of "general intelligence" without engaging a pre-existing political shitshow.
Which sucks.
posted by brennen at 1:49 PM on April 11, 2013
I feel like I may have contributed to poisoning this thread above, invoking the typical devolution of the topic by the mention of one of its typical modes. I will only say that yes, whether and how general intelligence is a meaningful/useful concept is a subject of discussion, but even the discussion contained in the FPP links was already, before this thread got started, complicated by the involvement of the usual pack of fancy-vocabulary racists, and suddenly it is very, very hard to talk about the merits of the idea of "general intelligence" without engaging a pre-existing political shitshow.
Which sucks.
posted by brennen at 1:49 PM on April 11, 2013
Look, personally, I've got all the reason in the world to believe in this g thing. As a kid I tested pretty highly, and it would be awfully tempting to take pride in that (seems odd to take pride in a personal quality I obviously had nothing to do with, though) and to see it as an indication of some kind of innate biological superiority. Especially since, now as an adult, my son tested so well he became the youngest kid in his school to get invited to participate in the gifted program (and I probably shouldn't but do take a little pride in that).
But you know what? Whatever differences these tests measure are really statistically tiny--teeny tiny. Viewed in context, even granting that IQ tests might actually measure something meaningful, all humans are so much closer to each other in whatever that ability is than not that it's pointless to invest much of any meaning in it. And contrary to the outdated view that IQ is static and predetermined, newer science strongly suggests IQ is much more elastic than we previously thought, with major swings still occurring even into early adulthood. Here's another study exploring how IQ scores can be impacted by environmental circumstances and developmental opportunities both earlier and later in life, for example.
Whatever 'g' is, if it's anything we should care about at all, it's probably not about innate ability or biology--and there being some racially determined component to IQ is just a non-starter, because in modern genetic science, it's become clearer than ever that race is more of a cultural construct than a biological reality anyway (as if the fact that our racial categories are so arbitrary wasn't an obvious enough clue to that, given that they're sometimes based on geographical origins and other times, on a hodge-podge of superficial, phenotypic traits).
posted by saulgoodman at 1:49 PM on April 11, 2013 [1 favorite]
But you know what? Whatever differences these tests measure are really statistically tiny--teeny tiny. Viewed in context, even granting that IQ tests might actually measure something meaningful, all humans are so much closer to each other in whatever that ability is than not that it's pointless to invest much of any meaning in it. And contrary to the outdated view that IQ is static and predetermined, newer science strongly suggests IQ is much more elastic than we previously thought, with major swings still occurring even into early adulthood. Here's another study exploring how IQ scores can be impacted by environmental circumstances and developmental opportunities both earlier and later in life, for example.
Whatever 'g' is, if it's anything we should care about at all, it's probably not about innate ability or biology--and there being some racially determined component to IQ is just a non-starter, because in modern genetic science, it's become clearer than ever that race is more of a cultural construct than a biological reality anyway (as if the fact that our racial categories are so arbitrary wasn't an obvious enough clue to that, given that they're sometimes based on geographical origins and other times, on a hodge-podge of superficial, phenotypic traits).
posted by saulgoodman at 1:49 PM on April 11, 2013 [1 favorite]
The discussion is whether or not general intelligence is a myth. I hate to spoil the surprise, but it is not.
Curiously, the people I know working in cognitive and neuroscience universally agree that it is. Why should I believe you over them?
posted by asterix at 1:49 PM on April 11, 2013 [2 favorites]
Curiously, the people I know working in cognitive and neuroscience universally agree that it is. Why should I believe you over them?
posted by asterix at 1:49 PM on April 11, 2013 [2 favorites]
Curiously, the people I know working in cognitive and neuroscience universally agree that it is. Why should I believe you over them?
I am not asking you to believe me. I am simply a layman stating the expert consensus.
When you understand why Michael Behe's friends should disbelieve him, you will understand why you should disbelieve your friends.
posted by Tanizaki at 1:54 PM on April 11, 2013
I am not asking you to believe me. I am simply a layman stating the expert consensus.
When you understand why Michael Behe's friends should disbelieve him, you will understand why you should disbelieve your friends.
posted by Tanizaki at 1:54 PM on April 11, 2013
But you know what? Whatever differences these tests measure are really statistically tiny--teeny tiny. Viewed in context, even granting that IQ tests might actually measure something meaningful, all humans are so much closer to each other in whatever that ability is than not that it's pointless to invest much of any meaning in it.
This is simply false. General intelligence is the single greatest predictive factor of just about every sociological outcome. It has greater predictive value than socioeconomic status.
posted by Tanizaki at 1:56 PM on April 11, 2013
This is simply false. General intelligence is the single greatest predictive factor of just about every sociological outcome. It has greater predictive value than socioeconomic status.
posted by Tanizaki at 1:56 PM on April 11, 2013
I am not asking you to believe me. I am simply a layman stating the expert consensus.
That merely pushes the question back to, "why should I believe that you are in fact stating the expert consensus?" (Particularly given that my friends and acquaintances are, in fact, not laypeople but actual working scientists in the field.)
posted by asterix at 1:58 PM on April 11, 2013 [2 favorites]
That merely pushes the question back to, "why should I believe that you are in fact stating the expert consensus?" (Particularly given that my friends and acquaintances are, in fact, not laypeople but actual working scientists in the field.)
posted by asterix at 1:58 PM on April 11, 2013 [2 favorites]
>General intelligence is the single greatest predictive factor of just about every sociological outcome.
Which is useless information, because sociological outcomes determine general intelligence.
posted by MisantropicPainforest at 2:01 PM on April 11, 2013 [4 favorites]
Which is useless information, because sociological outcomes determine general intelligence.
posted by MisantropicPainforest at 2:01 PM on April 11, 2013 [4 favorites]
My own personal theory is that there are a probably a large number of potentially independent variables that contribute to intelligence, but these variables exist largely at the neurological level, and are not easily isolated at the activity level. Thus, something like 'g' probably exists as an aggregation of these low-level traits, but we cannot easily tease them out into independent variables since we really don't understand the process (of thinking). The actual traits that determine intelligence probably have many different origins; some biological, some environmental.
Discussions of race and intelligence...a very touchy terrain. Once you start talking about a single variable for intelligence, and suggesting genetic heritability of this factor, it's inevitable that you open the door to the possibility of racial tendencies regarding intelligence. This is one reason (among others) that some people argue so hard against the idea of genetically-determined intelligence or even the concept of "race" as something that has any biological meaning (note: not trying to derail into that topic at this time).
Personally, I have much sympathy for these peoples' point of view, but I cannot eliminate the opposing possibility. On the other hand, I have little sympathy for people who argue in favor of 'g' and the influence of biology as a way of smuggling in their pre-existing racist sentiments, and I think they usually rush to their preferred conclusions, although one cannot simply refute them on the basis of disliking their conclusions. Whatever happened to people just being curious about the truth? I guess there are too many strong agendas to think clearly about subjects like this. That is not to say that proponents of different views of 'g' must fall into one camp or the other; it's just that too many do.
posted by Edgewise at 2:06 PM on April 11, 2013 [2 favorites]
Discussions of race and intelligence...a very touchy terrain. Once you start talking about a single variable for intelligence, and suggesting genetic heritability of this factor, it's inevitable that you open the door to the possibility of racial tendencies regarding intelligence. This is one reason (among others) that some people argue so hard against the idea of genetically-determined intelligence or even the concept of "race" as something that has any biological meaning (note: not trying to derail into that topic at this time).
Personally, I have much sympathy for these peoples' point of view, but I cannot eliminate the opposing possibility. On the other hand, I have little sympathy for people who argue in favor of 'g' and the influence of biology as a way of smuggling in their pre-existing racist sentiments, and I think they usually rush to their preferred conclusions, although one cannot simply refute them on the basis of disliking their conclusions. Whatever happened to people just being curious about the truth? I guess there are too many strong agendas to think clearly about subjects like this. That is not to say that proponents of different views of 'g' must fall into one camp or the other; it's just that too many do.
posted by Edgewise at 2:06 PM on April 11, 2013 [2 favorites]
If your friends have $5, they know where to find me.
I'm hurt.
posted by asterix at 2:06 PM on April 11, 2013
I'm hurt.
posted by asterix at 2:06 PM on April 11, 2013
Yeah Tanizaki, I really don't believe this is remotely "consensus" territory. Do you have some way of showing that this is, in fact, the opinion of most psychologists, neurologists, and cognitive scientists in the intelligence field?
posted by downing street memo at 2:08 PM on April 11, 2013
posted by downing street memo at 2:08 PM on April 11, 2013
(Actually, come to think of it, I'm not even sure they use a single, IQ measure anymore in schools. My son didn't get a single score, but several, with a kind of summary-level overall score that was just characterized as an aggregate of the others).
This is simply false.
No, it's not. You're just misreading me. I didn't claim anything about differences in outcomes. I said the differences in whatever these scores are measuring are insignificantly tiny.
If older tests of general intelligence are culturally biased in the first place, then wouldn't you expect to see the effects of the same cultural biases in later life outcomes? It's garbage in, garbage out.
Also, if the newer science in the area is correct, and IQ isn't static and innate, but can be dramatically shaped by environment and developmental opportunities, then wouldn't you expect the same correlations? People in the best socioeconomic position to start with score better on the tests, and have better life outcomes not because of any mysterious "g" factor, but because the same biases effect both the testing and the outcomes.
Let me put it another way: compared to even the most intelligent non-human intelligent creatures, the differences in human intelligence are statistically insignificant. So why not focus on making sure we all get the same opportunities for development and enrichment instead of tatooing IQs on everybody's foreheads?
posted by saulgoodman at 2:09 PM on April 11, 2013 [1 favorite]
This is simply false.
No, it's not. You're just misreading me. I didn't claim anything about differences in outcomes. I said the differences in whatever these scores are measuring are insignificantly tiny.
If older tests of general intelligence are culturally biased in the first place, then wouldn't you expect to see the effects of the same cultural biases in later life outcomes? It's garbage in, garbage out.
Also, if the newer science in the area is correct, and IQ isn't static and innate, but can be dramatically shaped by environment and developmental opportunities, then wouldn't you expect the same correlations? People in the best socioeconomic position to start with score better on the tests, and have better life outcomes not because of any mysterious "g" factor, but because the same biases effect both the testing and the outcomes.
Let me put it another way: compared to even the most intelligent non-human intelligent creatures, the differences in human intelligence are statistically insignificant. So why not focus on making sure we all get the same opportunities for development and enrichment instead of tatooing IQs on everybody's foreheads?
posted by saulgoodman at 2:09 PM on April 11, 2013 [1 favorite]
Do you have some way of showing that this is, in fact, the opinion of most psychologists, neurologists, and cognitive scientists in the intelligence field?
I second this appeal.
posted by MisantropicPainforest at 2:10 PM on April 11, 2013
I second this appeal.
posted by MisantropicPainforest at 2:10 PM on April 11, 2013
I am simply a layman stating the expert consensus.
You have done this before in threads that touch on this subject, and I don't know that I've ever seen you provide links to this consensus. If you can't or won't provide them, you should stop making this assertion.
posted by rtha at 2:12 PM on April 11, 2013 [3 favorites]
You have done this before in threads that touch on this subject, and I don't know that I've ever seen you provide links to this consensus. If you can't or won't provide them, you should stop making this assertion.
posted by rtha at 2:12 PM on April 11, 2013 [3 favorites]
That is not to say that proponents of different views of 'g' must fall into one camp or the other; it's just that too many do.
Yeah. I apologize for intimating above that anyone reading this who thinks g is a thing is a priori a total dickbag. I have just been repeatedly burned on the subject.
posted by brennen at 2:13 PM on April 11, 2013
Yeah. I apologize for intimating above that anyone reading this who thinks g is a thing is a priori a total dickbag. I have just been repeatedly burned on the subject.
posted by brennen at 2:13 PM on April 11, 2013
General intelligence is the single greatest predictive factor of just about every sociological outcome. It has greater predictive value than socioeconomic status.
I would be extremely surprised if this were true. Can you could point to a study or two showing some comparisons between these variables?
posted by a snickering nuthatch at 2:14 PM on April 11, 2013
I would be extremely surprised if this were true. Can you could point to a study or two showing some comparisons between these variables?
posted by a snickering nuthatch at 2:14 PM on April 11, 2013
This is simply false. General intelligence is the single greatest predictive factor of just about every sociological outcome. It has greater predictive value than socioeconomic status.Just look at George Bush and Paris Hilton!
Also, seriously, do you even know how hyperlinks work?
posted by delmoi at 2:15 PM on April 11, 2013 [2 favorites]
And I apologize as it seems you did recently provide links. Though it would've been nice to know beforehand that one goes to something called stormchan dot org, which I gather is related to stormfront; for future reference, my understanding is that links to stormfront and their ilk is not allowed here - you should check with the mods on that.
posted by rtha at 2:34 PM on April 11, 2013
posted by rtha at 2:34 PM on April 11, 2013
Anecdote time:
Surely you people have noticed that when you meet someone who is very smart at something, they are usually above-average at other mental tasks, too, right?
You've noticed that it's rare for someone to score 300 on one part of the SAT but 700 on the other, right?
Mental abilities are correlated. It's almost like there's a variable you can adjust when designing a human that will slide all of their mental abilities up or down the scale. Perhaps some type of "general" factor.
posted by grahamsletter at 2:34 PM on April 11, 2013 [2 favorites]
Surely you people have noticed that when you meet someone who is very smart at something, they are usually above-average at other mental tasks, too, right?
You've noticed that it's rare for someone to score 300 on one part of the SAT but 700 on the other, right?
Mental abilities are correlated. It's almost like there's a variable you can adjust when designing a human that will slide all of their mental abilities up or down the scale. Perhaps some type of "general" factor.
posted by grahamsletter at 2:34 PM on April 11, 2013 [2 favorites]
And I apologize as it seems you did recently provide links.
Although the second of those links (the one to an article in the Times Magazine) directly contradicts the conclusion Tanizaki draws from it.
posted by asterix at 2:38 PM on April 11, 2013
Although the second of those links (the one to an article in the Times Magazine) directly contradicts the conclusion Tanizaki draws from it.
posted by asterix at 2:38 PM on April 11, 2013
This is simply false. General intelligence is the single greatest predictive factor of just about every sociological outcome. It has greater predictive value than socioeconomic status.
From the NYT piece you linked to in the other thread:
"Regardless of whether the adopting families were rich or poor, Capron and Duyme learned, children whose biological parents were well-off had I.Q. scores averaging 16 points higher than those from working-class parents. Yet what is really remarkable is how big a difference the adopting families’ backgrounds made all the same. The average I.Q. of children from well-to-do parents who were placed with families from the same social stratum was 119.6. But when such infants were adopted by poor families, their average I.Q. was 107.5 — 12 points lower. The same holds true for children born into impoverished families: youngsters adopted by parents of similarly modest means had average I.Q.’s of 92.4, while the I.Q.’s of those placed with well-off parents averaged 103.6. These studies confirm that environment matters — the only, and crucial, difference between these children is the lives they have led.
A later study of French youngsters adopted between the ages of 4 and 6 shows the continuing interplay of nature and nurture. Those children had little going for them. Their I.Q.’s averaged 77, putting them near retardation. Most were abused or neglected as infants, then shunted from one foster home or institution to the next.
Nine years later, they retook the I.Q. tests, and contrary to the conventional belief that I.Q. is essentially stable, all of them did better. The amount they improved was directly related to the adopting family’s status. Children adopted by farmers and laborers had average I.Q. scores of 85.5; those placed with middle-class families had average scores of 92. The average I.Q. scores of youngsters placed in well-to-do homes climbed more than 20 points, to 98 — a jump from borderline retardation to a whisker below average. That is a huge difference — a person with an I.Q. of 77 couldn’t explain the rules of baseball, while an individual with a 98 I.Q. could actually manage a baseball team — and it can only be explained by pointing to variations in family circumstances.
Taken together, these studies show that the issue has changed: it is no longer a matter of whether the environment matters but when and how it matters. And poverty, quite clearly, is an important part of the answer."
posted by rtha at 2:46 PM on April 11, 2013 [6 favorites]
From the NYT piece you linked to in the other thread:
"Regardless of whether the adopting families were rich or poor, Capron and Duyme learned, children whose biological parents were well-off had I.Q. scores averaging 16 points higher than those from working-class parents. Yet what is really remarkable is how big a difference the adopting families’ backgrounds made all the same. The average I.Q. of children from well-to-do parents who were placed with families from the same social stratum was 119.6. But when such infants were adopted by poor families, their average I.Q. was 107.5 — 12 points lower. The same holds true for children born into impoverished families: youngsters adopted by parents of similarly modest means had average I.Q.’s of 92.4, while the I.Q.’s of those placed with well-off parents averaged 103.6. These studies confirm that environment matters — the only, and crucial, difference between these children is the lives they have led.
A later study of French youngsters adopted between the ages of 4 and 6 shows the continuing interplay of nature and nurture. Those children had little going for them. Their I.Q.’s averaged 77, putting them near retardation. Most were abused or neglected as infants, then shunted from one foster home or institution to the next.
Nine years later, they retook the I.Q. tests, and contrary to the conventional belief that I.Q. is essentially stable, all of them did better. The amount they improved was directly related to the adopting family’s status. Children adopted by farmers and laborers had average I.Q. scores of 85.5; those placed with middle-class families had average scores of 92. The average I.Q. scores of youngsters placed in well-to-do homes climbed more than 20 points, to 98 — a jump from borderline retardation to a whisker below average. That is a huge difference — a person with an I.Q. of 77 couldn’t explain the rules of baseball, while an individual with a 98 I.Q. could actually manage a baseball team — and it can only be explained by pointing to variations in family circumstances.
Taken together, these studies show that the issue has changed: it is no longer a matter of whether the environment matters but when and how it matters. And poverty, quite clearly, is an important part of the answer."
posted by rtha at 2:46 PM on April 11, 2013 [6 favorites]
If older tests of general intelligence are culturally biased in the first place, then wouldn't you expect to see the effects of the same cultural biases in later life outcomes?
Also, if the newer science in the area is correct, and IQ isn't static and innate, but can be dramatically shaped by environment and developmental opportunities, then wouldn't you expect the same correlations?
These are conditional statements, not logical arguments.
I would be extremely surprised if this were true. Can you could point to a study or two showing some comparisons between these variables?
Then it pleases me to surprise you. You might try this report, which I have previously linked, for starters. More specifically, general intelligence as the single best predictor of school and work success. Any work, by the way. If you have two bus boys, the smarter one is going to be better at it and if they are far enough apart, experience can never compensate. You also might check out this report that discusses correlations between general intelligence and various factors such as health, violent crime, and several other sociological factors in the context of comparing American states.
posted by Tanizaki at 3:06 PM on April 11, 2013
Also, if the newer science in the area is correct, and IQ isn't static and innate, but can be dramatically shaped by environment and developmental opportunities, then wouldn't you expect the same correlations?
These are conditional statements, not logical arguments.
I would be extremely surprised if this were true. Can you could point to a study or two showing some comparisons between these variables?
Then it pleases me to surprise you. You might try this report, which I have previously linked, for starters. More specifically, general intelligence as the single best predictor of school and work success. Any work, by the way. If you have two bus boys, the smarter one is going to be better at it and if they are far enough apart, experience can never compensate. You also might check out this report that discusses correlations between general intelligence and various factors such as health, violent crime, and several other sociological factors in the context of comparing American states.
posted by Tanizaki at 3:06 PM on April 11, 2013
The great criticism that HBD fans seem unable to address is "if you so smart, why ain't you rich?"
Although the equally zealous support of welfare and affirmative action espoused by rich people who happily live and work in a milieu with very few poor or non-Asian minority friends, neighbors or co-workers also strikes me as a bit contradictory.
posted by MattD at 3:06 PM on April 11, 2013
Although the equally zealous support of welfare and affirmative action espoused by rich people who happily live and work in a milieu with very few poor or non-Asian minority friends, neighbors or co-workers also strikes me as a bit contradictory.
posted by MattD at 3:06 PM on April 11, 2013
Nine years later, they retook the I.Q. tests, and contrary to the conventional belief that I.Q. is essentially stable, all of them did better.
The conventional belief is not that IQ is stable. To the contrary, it is pretty fluid in children. General intelligence stabilizes around age 18, where heritability is at least 0.7. That is why I tell parents not to put too much stock in their 7-year-old's IQ score.
There is no doubt environmental factors. No one denies that. However, it is still largely heritable, just like height. A favorable environment can help someone get closer to their maximum genetic potential, but just as all the nutrition in the world won't make anyone ten feet tall, there are limits to the effect on intelligence. Even taken at face value, kids with really below average scores were brought to just below average. They didn't even get to the median, let alone above average. They were tested at 13 to 15. I would like to see their scores when they are 23 and 25.
The great criticism that HBD fans seem unable to address is "if you so smart, why ain't you rich?"
If that is "the great criticism", HBD fans can rest easy. Below average IQ is more strongly correlated with poverty than above average IQ is correlated with wealth. You may wish to review to the National Longitudinal Surveys of Youth on this.
posted by Tanizaki at 3:15 PM on April 11, 2013
The conventional belief is not that IQ is stable. To the contrary, it is pretty fluid in children. General intelligence stabilizes around age 18, where heritability is at least 0.7. That is why I tell parents not to put too much stock in their 7-year-old's IQ score.
There is no doubt environmental factors. No one denies that. However, it is still largely heritable, just like height. A favorable environment can help someone get closer to their maximum genetic potential, but just as all the nutrition in the world won't make anyone ten feet tall, there are limits to the effect on intelligence. Even taken at face value, kids with really below average scores were brought to just below average. They didn't even get to the median, let alone above average. They were tested at 13 to 15. I would like to see their scores when they are 23 and 25.
The great criticism that HBD fans seem unable to address is "if you so smart, why ain't you rich?"
If that is "the great criticism", HBD fans can rest easy. Below average IQ is more strongly correlated with poverty than above average IQ is correlated with wealth. You may wish to review to the National Longitudinal Surveys of Youth on this.
posted by Tanizaki at 3:15 PM on April 11, 2013
If that is "the great criticism", HBD fans can rest easy.
I just spent a little time reading the "dictionary" at the bottom of that HBD link. I think the "great criticism" of HBD is more along the lines of Christ, what a bunch of assholes.
posted by brennen at 3:38 PM on April 11, 2013 [4 favorites]
I just spent a little time reading the "dictionary" at the bottom of that HBD link. I think the "great criticism" of HBD is more along the lines of Christ, what a bunch of assholes.
posted by brennen at 3:38 PM on April 11, 2013 [4 favorites]
The following set of propositions are not inconsistent:
- g exists.
- g as a concept is frequently defended in online discussions by odious racists who disguise their agenda, and is rarely encountered by laypeople outside this context.
- g is correlated with many as of yet unidentified genetic factors, a small minority of which will turn out to be correlated with ethnic background.
- g is strongly inversely correlated with environmental stressors, especially gestational or in early childhood, that are endemic to many oppressed minorities and people living in less developed countries.
- g is strongly correlated with education
- The causes of differences across ethnic/racial lines in intelligence test results are due in part to design bias and in part to underlying differences in g, but those differences are so dominated by (4) and (5) that differences due to (3) can't be discerned.
- Science will eventually find genetic causal factors for differences in g that have potential racial/ethnic implications, but these implications remain hypothetical because of confounding factors. These implications won't conform to the prejudices of odious racists, who will subsequently drop their pretense of fearless dedication to hard-nosed objectivity and empiricism, becoming overnight obsessed with unmasking a supposed cabal of PC scientists suppressing the truth.
- g is inversely correlated with being a racist dickbag obsessed with g.
brennen: Ew. Ew ew ew ew ew. Also, paging Brigadier General Jack D. Ripper.
Sample definitions:
Devirilisation: Declining values of courage and virility for the sake of political correctness.
Involution: According to Guillaume Faye, the regression of a civilization or species to maladaptive forms that lead to the diminishing of its vital forces. Cultural involution has been stimulated by the decline of education, the regression of knowledge, the collapse of social norms, the immersion of youth in a world of audio/visual play [and] the Africanization of European culture.
Ethnomasochism: Hatred of one's own race.
Racism & Racist: A racist is someone who values truth more than political correctness. Martin Sewell: "So-called racism is a perfectly natural in-group bias which has been stigmatized by the politically correct West.” Peter Brimelow: "The modern definition of 'racist' is someone who is winning an argument with a liberal."
Rasse: German for race.
This is some stormfront level shit.
posted by PMdixon at 3:45 PM on April 11, 2013 [5 favorites]
Sample definitions:
Devirilisation: Declining values of courage and virility for the sake of political correctness.
Involution: According to Guillaume Faye, the regression of a civilization or species to maladaptive forms that lead to the diminishing of its vital forces. Cultural involution has been stimulated by the decline of education, the regression of knowledge, the collapse of social norms, the immersion of youth in a world of audio/visual play [and] the Africanization of European culture.
Ethnomasochism: Hatred of one's own race.
Racism & Racist: A racist is someone who values truth more than political correctness. Martin Sewell: "So-called racism is a perfectly natural in-group bias which has been stigmatized by the politically correct West.” Peter Brimelow: "The modern definition of 'racist' is someone who is winning an argument with a liberal."
Rasse: German for race.
This is some stormfront level shit.
posted by PMdixon at 3:45 PM on April 11, 2013 [5 favorites]
I tend to think of intelligence as a differential equation. "Nature" provides the initial conditions, "Nurture" the formula that they feed into. A simplification, of course, but I think it serves to illustrate how the two interact, and I think the NYT article reinforces the idea.
Someone who is born with better hardware (due to prenatal nutrition, genetics (and don't take that to imply anything racially-correlated), whatever) will diverge from someone with worse hardware when running the same software to grow themselves.
Of two people with the same initial hardware, the one running the more effective software will likewise develop themselves faster. And then there are the viruses that can kick a person off their track.
When someone starts with better hardware, and gets access to better software, that how a supercomputer is built.
posted by I-Write-Essays at 4:07 PM on April 11, 2013 [1 favorite]
Someone who is born with better hardware (due to prenatal nutrition, genetics (and don't take that to imply anything racially-correlated), whatever) will diverge from someone with worse hardware when running the same software to grow themselves.
Of two people with the same initial hardware, the one running the more effective software will likewise develop themselves faster. And then there are the viruses that can kick a person off their track.
When someone starts with better hardware, and gets access to better software, that how a supercomputer is built.
posted by I-Write-Essays at 4:07 PM on April 11, 2013 [1 favorite]
Wow, brennen. That...
That reads like it was written by Goebbels.
posted by [expletive deleted] at 4:14 PM on April 11, 2013 [1 favorite]
That reads like it was written by Goebbels.
posted by [expletive deleted] at 4:14 PM on April 11, 2013 [1 favorite]
However you measure it, or whatever construct for the abstract concept you erect, intelligence matters and has more influence on what kind of control a human will be able to have over her/his environment than anything else. It always shocks me just HOW much of a difference it makes. How is it that There are people who can't tie their own shoes, and there are people who can literally reach into other dimensions and bring back descriptions of phenomenon that change our world (electromagnetism, for example).
posted by Halogenhat at 4:19 PM on April 11, 2013
posted by Halogenhat at 4:19 PM on April 11, 2013
[expletive deleted]: 6) The causes of differences across ethnic/racial lines in intelligence test results are due in part to design bias and in part to underlying differences in g, but those differences are so dominated by (4) and (5) that differences due to (3) can't be discerned.
I think this is the key point, and when discussing genetic factors it's not really worth comparing different classes, but at specific families within a given class. Those variables need to be controlled for, but as has been stated so often, the context of the discussion is not usually to show that there are genetic factors for intelligence, but that those factors are correlated to class.
From my own life experiences, I think it is evident that intelligence is hereditary. However, heredity != genetics; not strictly. Culture is also hereditary. My hypothesis is that intelligent people would tend to select for more intelligent partners, and have more intelligent parenting practices, which results in more intelligent children who select more intelligent partners, and on and on.
In many ways, I think the whole premise of "Nature VS Nurture" is a false dichotomy. Nature IS Nurture, and the two cannot be disentangled.
posted by I-Write-Essays at 4:32 PM on April 11, 2013
I think this is the key point, and when discussing genetic factors it's not really worth comparing different classes, but at specific families within a given class. Those variables need to be controlled for, but as has been stated so often, the context of the discussion is not usually to show that there are genetic factors for intelligence, but that those factors are correlated to class.
From my own life experiences, I think it is evident that intelligence is hereditary. However, heredity != genetics; not strictly. Culture is also hereditary. My hypothesis is that intelligent people would tend to select for more intelligent partners, and have more intelligent parenting practices, which results in more intelligent children who select more intelligent partners, and on and on.
In many ways, I think the whole premise of "Nature VS Nurture" is a false dichotomy. Nature IS Nurture, and the two cannot be disentangled.
posted by I-Write-Essays at 4:32 PM on April 11, 2013
muddgirl, I've been pondering your question about how laypeople form opinions on scientific disputes for some time. The answers that come up seem to be of two forms:
1. Pick some Scientists to trust, and believe what they say. I do not like this approach.
2. Form an opinion like you would any other thing.
I've tried to finagle a third approach, but so far I can't do it. (The best I can do is that sometimes I form trivial or social ones, e.g. "that guy doesn't like potatoes".) based on very little evidence, which is different from when I form Serious Opinions.)
What perhaps stops me, and I suspect many people, is that we generally form opinions by making a decision (based on however much reason and evidence) and then never revisit the reasoning. "1,000 IU of Vitamin D is good for me, that's settled" is easier than "Vitamin D is useful in these doses for these purposes according to these sources who used this reasoning". Most people (including myself) don't have the mental capacity to form very many opinions of the second sort.
What I try to do, and it's only mildly successful, is to store my opinions alongside some minimum of the reasoning I took to get there. I try to remember as much of (to continue the Vit D example) the doses, purposes, sources, and reasoning as I can, with multiple copies if I agree with multiple competing sources. If someone challenges that opinion then I revisit as much of the reasoning as possible, and I try (though often fail) to scale my willingness to admit ignorance or change habits based on how much I remember compared to the quality of reasoning I'm now presented with. We're not wired to do this, but I feel better about myself as a rational being when I do it. I'm still a layman, but this approach allows me to form substantial opinions on matters of science that are complex and disputed. The key is that it requires frequent readjustment of belief, which is good but hard.
With regards to this FPP, this will take days of reading and asking people to even start to form an opinion. So thanks. But there's also a significant degree of frustration that I now have to do all this reading.
posted by daveliepmann at 4:34 PM on April 11, 2013 [3 favorites]
1. Pick some Scientists to trust, and believe what they say. I do not like this approach.
2. Form an opinion like you would any other thing.
I've tried to finagle a third approach, but so far I can't do it. (The best I can do is that sometimes I form trivial or social ones, e.g. "that guy doesn't like potatoes".) based on very little evidence, which is different from when I form Serious Opinions.)
What perhaps stops me, and I suspect many people, is that we generally form opinions by making a decision (based on however much reason and evidence) and then never revisit the reasoning. "1,000 IU of Vitamin D is good for me, that's settled" is easier than "Vitamin D is useful in these doses for these purposes according to these sources who used this reasoning". Most people (including myself) don't have the mental capacity to form very many opinions of the second sort.
What I try to do, and it's only mildly successful, is to store my opinions alongside some minimum of the reasoning I took to get there. I try to remember as much of (to continue the Vit D example) the doses, purposes, sources, and reasoning as I can, with multiple copies if I agree with multiple competing sources. If someone challenges that opinion then I revisit as much of the reasoning as possible, and I try (though often fail) to scale my willingness to admit ignorance or change habits based on how much I remember compared to the quality of reasoning I'm now presented with. We're not wired to do this, but I feel better about myself as a rational being when I do it. I'm still a layman, but this approach allows me to form substantial opinions on matters of science that are complex and disputed. The key is that it requires frequent readjustment of belief, which is good but hard.
With regards to this FPP, this will take days of reading and asking people to even start to form an opinion. So thanks. But there's also a significant degree of frustration that I now have to do all this reading.
posted by daveliepmann at 4:34 PM on April 11, 2013 [3 favorites]
Tanizaki,
Are you every going to provide a link to a respectable scholar who attempts to rebut's Shalizi's argument?
Doing so would convince more people than simply linking to stormfrontesque websites. Thanks!
posted by MisantropicPainforest at 4:52 PM on April 11, 2013 [1 favorite]
Are you every going to provide a link to a respectable scholar who attempts to rebut's Shalizi's argument?
Doing so would convince more people than simply linking to stormfrontesque websites. Thanks!
posted by MisantropicPainforest at 4:52 PM on April 11, 2013 [1 favorite]
To avoid having my only contribution being snark, I did read the linked articles and found the D-dude's arguments rather unconvincing.
By far the largest chunk of his rebuttal is an extended form of " 'proving' impossibility by demonstrating absence-of-success" (in this case, of crafting credible "intelligence tests" that somehow escape the "positive manifold"). In pure form this is unsound logic, but I'd be willing to cut some slack if the conclusions drawn from such evidence were suitably qualified...but they weren't suitably qualified (by a very long shot) and so I don't feel like cutting any slack.
That aside, the secondary issue is that, at least to me, it's not at all surprising that it's been, ahem, difficult to craft tests that're (a) accepted as "intelligence tests" and also (b) not really correlated with other intelligence tests; the problem with (a) is that it's a social and cultural process, and constrains rather dramatically the range of "intelligence tests" that one might construct; it's eminently plausible that having to get one's putative "intelligence test" accepted by other academics as an "intelligence test" is a stringent-enough filter that any "intelligence test" that survives said filter is going to correlate with all the other intelligence tests, perhaps despite its creators' intents.
To an extent this is a matter of opinion -- to my eyes the various tests and so forth the D-dude discusses all do seem plenty close to each other, at least in the big scheme of things, but I suppose it's possible to have a point of view in which they seem dazzlingly diverse -- but at least for me it's like I said, in the absence of an agreed-upon definition of "intelligence" (which would make this discussion generally moot, if only we had it!) the distinction between an "intelligence test" and a mere "test" hinges too much on hazy notions and unstated intuitions to feel robust enough to base as much on it as some might hope.
In this context, it's *extremely* discouraging that the D-dude ends with this summary of the point he's refuting:
posted by hoople at 5:06 PM on April 11, 2013 [1 favorite]
By far the largest chunk of his rebuttal is an extended form of " 'proving' impossibility by demonstrating absence-of-success" (in this case, of crafting credible "intelligence tests" that somehow escape the "positive manifold"). In pure form this is unsound logic, but I'd be willing to cut some slack if the conclusions drawn from such evidence were suitably qualified...but they weren't suitably qualified (by a very long shot) and so I don't feel like cutting any slack.
That aside, the secondary issue is that, at least to me, it's not at all surprising that it's been, ahem, difficult to craft tests that're (a) accepted as "intelligence tests" and also (b) not really correlated with other intelligence tests; the problem with (a) is that it's a social and cultural process, and constrains rather dramatically the range of "intelligence tests" that one might construct; it's eminently plausible that having to get one's putative "intelligence test" accepted by other academics as an "intelligence test" is a stringent-enough filter that any "intelligence test" that survives said filter is going to correlate with all the other intelligence tests, perhaps despite its creators' intents.
To an extent this is a matter of opinion -- to my eyes the various tests and so forth the D-dude discusses all do seem plenty close to each other, at least in the big scheme of things, but I suppose it's possible to have a point of view in which they seem dazzlingly diverse -- but at least for me it's like I said, in the absence of an agreed-upon definition of "intelligence" (which would make this discussion generally moot, if only we had it!) the distinction between an "intelligence test" and a mere "test" hinges too much on hazy notions and unstated intuitions to feel robust enough to base as much on it as some might hope.
In this context, it's *extremely* discouraging that the D-dude ends with this summary of the point he's refuting:
Shalizi’s first error is his assertion that cognitive tests correlate with each other because IQ test makers exclude tests that do not fit the positive manifold....as that's something with a proof nowhere even *attempted*; the entire laundry list is seemingly meant to rebut the far weaker claim that "all intelligence tests are deliberately constructed to correlate with each other", which is not the same thing. that in particular is a bit of a bridge too far in the "drawing stronger conclusions than your stated evidence seemingly warrants", at least for me.
posted by hoople at 5:06 PM on April 11, 2013 [1 favorite]
I have difficulty undertsanding the linked essay. I'm not sure what this says about my intellegence. or sanity.
posted by jonmc at 5:45 PM on April 11, 2013
posted by jonmc at 5:45 PM on April 11, 2013
If that is "the great criticism", HBD fans can rest easy. Below average IQ is more strongly correlated with poverty than above average IQ is correlated with wealth. You may wish to review to the National Longitudinal Surveys of Youth on this.
Well, I guess, at least I won't end up poor. Too bad IQ's not more closely correlated with wealth though. Like 1:1.
I'm really not sure where this discussion ended up, but for the record, while I believe there may be such a thing as general intelligence (with a heavy emphasis on "may"), I'm personally very skeptical that's what I.Q. tests have traditionally measured. Since my own I.Q. is allegedly pretty high (around 148 the last time I took the Stanford-Binet), I must be smart enough to know what I'm talking about, so trust me when I say I.Q. is probably just nonsense. It is possible to score very well on one of those things and be a total idiot. At the same time, I also find it impossible to believe anyone with an I.Q. of 98 could manage a baseball team.
In response to whoever mentioned this up-thread, it's also easier than you might think to score less than 300 on the math portion of the SAT and 700 on the verbal portion, especially if you Christmas tree the math section because you already know you're planning to apply for an English program that doesn't care about the math score anyway.
posted by saulgoodman at 6:21 PM on April 11, 2013
Well, I guess, at least I won't end up poor. Too bad IQ's not more closely correlated with wealth though. Like 1:1.
I'm really not sure where this discussion ended up, but for the record, while I believe there may be such a thing as general intelligence (with a heavy emphasis on "may"), I'm personally very skeptical that's what I.Q. tests have traditionally measured. Since my own I.Q. is allegedly pretty high (around 148 the last time I took the Stanford-Binet), I must be smart enough to know what I'm talking about, so trust me when I say I.Q. is probably just nonsense. It is possible to score very well on one of those things and be a total idiot. At the same time, I also find it impossible to believe anyone with an I.Q. of 98 could manage a baseball team.
In response to whoever mentioned this up-thread, it's also easier than you might think to score less than 300 on the math portion of the SAT and 700 on the verbal portion, especially if you Christmas tree the math section because you already know you're planning to apply for an English program that doesn't care about the math score anyway.
posted by saulgoodman at 6:21 PM on April 11, 2013
I wrote: "At the same time, I also find it impossible to believe anyone with an I.Q. of 98 could manage a baseball team." But on second thought, there are an awful lot of cultural assumptions packed into some of those older I.Q. tests. Some of them assume familiarity with logical operators and typographical conventions--things like representing similes as "A:B :: C:D," etc. And the verbal reasoning portions often require a fairly extensive vocabulary. Those kinds of tests don't seem to be measures of general intelligence, reasoning and pattern-recognition skills so much as tests of familiarity with certain cultural conventions and educational background.
posted by saulgoodman at 6:44 PM on April 11, 2013
posted by saulgoodman at 6:44 PM on April 11, 2013
Since my own I.Q. is allegedly pretty high (around 148 the last time I took the Stanford-Binet)
You poor dear. I have been 4.5 to 5 sigmas out on every standardized test I ever took. If I took any of this shit seriously I would have to widen the doorways in my house to get my head through.
But it's bullshit. I have a pretty good idea of how I got the way I am, and if there was a genetic contribution it was minimal -- as the most cursory look at the rest of my Redneckus Americanus extended family would demonstrate. There may be a few genetic fators that help or hinder, making it harder or easier or more or less likely to grow rich dendrite trees that code for rich, easily searched experience and aggressively grown patterns of invention. But mostly it's experience, practice, more experience, and more practice. Barring something going massively wrong like Down's Syndrome the differences between people due to genetics are trivial compared to the differences due to upbringing, opportunities, nutrition, and possibly to a certain extent a little stimulating trauma at just the right time.
By the time you have acquired enough brain programming to take a test nearly all of your programming has been acquired through experience. The wiring of the brain is an extraordinarily complex thing, easily five orders of magnitude too vast for the genome to have much to do with it at all. That's not conjecture either; that's information theory, a firm science. The genome can't have much to do with wiring the brain because there isn't enough information there to do more than rough cable-laying. The details are acquired through experience; they have to be, because there's nowhere else for them to come from.
Oh, and if IQ was that well correlated with intelligence I'd be Bill Gates. I suspect that sociopathic tendencies are a lot more useful than IQ in that regard lately.
posted by localroger at 6:45 PM on April 11, 2013 [3 favorites]
You poor dear. I have been 4.5 to 5 sigmas out on every standardized test I ever took. If I took any of this shit seriously I would have to widen the doorways in my house to get my head through.
But it's bullshit. I have a pretty good idea of how I got the way I am, and if there was a genetic contribution it was minimal -- as the most cursory look at the rest of my Redneckus Americanus extended family would demonstrate. There may be a few genetic fators that help or hinder, making it harder or easier or more or less likely to grow rich dendrite trees that code for rich, easily searched experience and aggressively grown patterns of invention. But mostly it's experience, practice, more experience, and more practice. Barring something going massively wrong like Down's Syndrome the differences between people due to genetics are trivial compared to the differences due to upbringing, opportunities, nutrition, and possibly to a certain extent a little stimulating trauma at just the right time.
By the time you have acquired enough brain programming to take a test nearly all of your programming has been acquired through experience. The wiring of the brain is an extraordinarily complex thing, easily five orders of magnitude too vast for the genome to have much to do with it at all. That's not conjecture either; that's information theory, a firm science. The genome can't have much to do with wiring the brain because there isn't enough information there to do more than rough cable-laying. The details are acquired through experience; they have to be, because there's nowhere else for them to come from.
Oh, and if IQ was that well correlated with intelligence I'd be Bill Gates. I suspect that sociopathic tendencies are a lot more useful than IQ in that regard lately.
posted by localroger at 6:45 PM on April 11, 2013 [3 favorites]
Dalliard's writing lacks a quality of reflectiveness and self-critique/awareness; I shall conclude the author doesn't possess much g in the first place.
posted by polymodus at 6:49 PM on April 11, 2013
posted by polymodus at 6:49 PM on April 11, 2013
But it's bullshit.
Absolutely. It's a test-taking skills test, at best.
posted by saulgoodman at 7:02 PM on April 11, 2013
Absolutely. It's a test-taking skills test, at best.
posted by saulgoodman at 7:02 PM on April 11, 2013
It's a test-taking skills test, at best.
Absolutely. My family is mostly a bunch of rednecks but Dad got seduced by one of those Atomic Energy Commission recruitment vans they sent around to rural high schools in the postwar era, and then after he got his Ph.D. and became a college physics teacher I grew up in his lab. Dad gave killer multiple choice exams and one of my hobbies as a kid was taking his college students' tests. By the time I was 12 I was usually doing better than his best real students. When I got to the ACT and the SAT they held no mysteries for me.
Anyway, in 1981 I got a 34 on the ACT. Most people don't get these stats but my school told me the numbers: 115,000 students took the ACT in that testing cycle, and in that cycle there were twelve 34's and one 35. Jn those days a 36 was unheard of. (Grade inflation in the Z scores, who would guess?) Translated to Z score and back to IQ that's about 180. Of course the air is pretty thin around numbers like that but occasionally I have figured that a test filtered for college applicants would, if anything, be better at discerning high-bracket differences than a general skills test also useful below 100. That is, if the very idea of such a single-parameter numeric score had any relevance at all.
Which it very obviously doesn't.
posted by localroger at 7:21 PM on April 11, 2013 [1 favorite]
Absolutely. My family is mostly a bunch of rednecks but Dad got seduced by one of those Atomic Energy Commission recruitment vans they sent around to rural high schools in the postwar era, and then after he got his Ph.D. and became a college physics teacher I grew up in his lab. Dad gave killer multiple choice exams and one of my hobbies as a kid was taking his college students' tests. By the time I was 12 I was usually doing better than his best real students. When I got to the ACT and the SAT they held no mysteries for me.
Anyway, in 1981 I got a 34 on the ACT. Most people don't get these stats but my school told me the numbers: 115,000 students took the ACT in that testing cycle, and in that cycle there were twelve 34's and one 35. Jn those days a 36 was unheard of. (Grade inflation in the Z scores, who would guess?) Translated to Z score and back to IQ that's about 180. Of course the air is pretty thin around numbers like that but occasionally I have figured that a test filtered for college applicants would, if anything, be better at discerning high-bracket differences than a general skills test also useful below 100. That is, if the very idea of such a single-parameter numeric score had any relevance at all.
Which it very obviously doesn't.
posted by localroger at 7:21 PM on April 11, 2013 [1 favorite]
According to Dalliard (whom we are already quite certain is a racist/eugenist), Shalizi makes three errors. The first:
The second supposed error is that there is other evidence for g that Shalizi does not address. This is not technically an error, but it stands. I am more convinced by Shalizi's broader points about g than those of the racist, but this "error" does not seem factually incorrect.
The third error Dalliard claims that Shalizi makes is:
Dalliard is admitting that g is an overall "these people do well" explanatory factor (just as Shalizi was criticizing) rather than a causative factor. Dalliard is saying here that even if intelligence is in fact a multitude of sources, with not one but many causes, g would still exist as a way to describe "behavior". But that's precisely useless. He's literally saying, "even if g doesn't exist, we'd still have g because people do differently on tests." If that's the case--if g is just the sum total of people doing well on any correlated measurements of achievement--then as Shalizi points out, why not toss in attractiveness and physical fitness? "I could even attribute them all to a single factor, a (for arete), and start treating it as a real causal variable." If you've got a bucket for all the stuff that makes people good at things, why not include everything correlated with success?
I mean, are we looking for a causative factor that pertains to physical reality or are we just arbitrarily naming a set of results? Are we engaged in science or stamp collecting?
posted by daveliepmann at 8:37 PM on April 11, 2013 [15 favorites]
Shalizi’s thesis is that the positive manifold is an artifact of test construction and that full-scale scores from different IQ batteries correlate only because they are designed to do that.Hogwash. Shalizi specifically states that factor analysis causes "g" to appear regardless of test design. His point is that factor analysis is a flawed method to discover whether g is a cause:
I'm going to show you some cases where you can see that the data don't have a single dominant cause, because I made them up randomly, but they nonetheless give that appearance when viewed through the lens of factor analysis.The closest Shalizi comes to committing the error Dalliard claims he made is in the sentence "Since intelligence tests are made to correlate with each other, it follows trivially that there must appear to be a general factor of intelligence", in which Shalizi is not talking about the intent of the test-designer, but rather the mathematical methodology of the test-designer, as described explicitly in footnote 8.
...
If I take any group of variables which are positively correlated, there will, as a matter of algebraic necessity, be a single dominant general factor, which describes more of the variance than any other, and all of them will be "positively loaded" on this factor, i.e., positively correlated with it.
...
You can get strong positive correlations — even ones with vanishing partial correlations, so it looks like there's one factor — even when all the real causes are about equal in importance and completely independent of one another.
The second supposed error is that there is other evidence for g that Shalizi does not address. This is not technically an error, but it stands. I am more convinced by Shalizi's broader points about g than those of the racist, but this "error" does not seem factually incorrect.
The third error Dalliard claims that Shalizi makes is:
He mistakenly believes that if the sampling model is deemed to be the correct description of the workings of intelligence, it means that there can be no general factor of intelligence.Dalliard is doubly wrong. First, Shalizi does not claim that there can be no general factor of intelligence, but rather that the "sampling model" (as Dalliard calls it) explains the results of factor analysis just as well as the general factor, and therefore factor analysis should not be used as an attempt to prove the existence of g. Of all of Shalizi's points, this is the most clear. Here is one of several repetitions:
...
At a sufficiently basal (neurological, molecular, etc.) level, intelligence necessarily becomes fractionated, but that does not mean that there is no general factor of intelligence at the behavioral level. As explained above, many types of evidence show that g is indeed a centrally important unidimensional source of behavioral differences between individuals.
If, after looking at your watch, you say that it's 12 o'clock, and I point out that your watch has stopped at 12, I am not saying that it's not 12 o'clock, just that your watch doesn't actually give you any evidence about the time. Similarly, pointing out that factor analysis and related techniques are unreliable guides to causal structure does not establish the non-existence of a one-dimensional latent variable driving the success of almost all human mental performance. It's possible that there is such a thing.So Dalliard is factually wrong in his reading of Shalizi. But that's not all--go back and read both paragraphs I quoted of his "third Shalizi error" to understand Dalliard's second failing. It is here that Dalliard runs straight into Shalizi's arete/mens sana in corpore sano argument.
Dalliard is admitting that g is an overall "these people do well" explanatory factor (just as Shalizi was criticizing) rather than a causative factor. Dalliard is saying here that even if intelligence is in fact a multitude of sources, with not one but many causes, g would still exist as a way to describe "behavior". But that's precisely useless. He's literally saying, "even if g doesn't exist, we'd still have g because people do differently on tests." If that's the case--if g is just the sum total of people doing well on any correlated measurements of achievement--then as Shalizi points out, why not toss in attractiveness and physical fitness? "I could even attribute them all to a single factor, a (for arete), and start treating it as a real causal variable." If you've got a bucket for all the stuff that makes people good at things, why not include everything correlated with success?
I mean, are we looking for a causative factor that pertains to physical reality or are we just arbitrarily naming a set of results? Are we engaged in science or stamp collecting?
posted by daveliepmann at 8:37 PM on April 11, 2013 [15 favorites]
One interesting thing is that there is actually a measurable, scientifically verifiable correlation between being a racist and being unintelligent. It's not even a speculative thing, it's actually measurable. Something these 'HBD' types tend to gloss over when demanding "The truth!!!" about IQ.
Go back to the car analogy. Some cars are faster around a race course then others. You'll find that cars that do well on one measure of speed are going to be correlated with good performance on other things. But that's not because there is some mythical "general" "speed factor" - it's due to a combination of lots of different things.
You could do a bunch of math and calculate an 'r' factor for cars, but that's not some intrinsic thing that makes the car fast, it's just the output of a function that takes the properties of a car and uses it to calculate some number. It's an effect, not a cause.
posted by delmoi at 8:45 PM on April 11, 2013 [1 favorite]
Anecdote time:So I take it you don't know what the word "anecdote" means? Why am I not surprised.
Surely you people have noticed that when you meet someone who is very smart at something, they are usually above-average at other mental tasks, too, right? ... Mental abilities are correlated. It's almost like there's a variable you can adjust when designing a human that will slide all of their mental abilities up or down the scale.
Go back to the car analogy. Some cars are faster around a race course then others. You'll find that cars that do well on one measure of speed are going to be correlated with good performance on other things. But that's not because there is some mythical "general" "speed factor" - it's due to a combination of lots of different things.
You could do a bunch of math and calculate an 'r' factor for cars, but that's not some intrinsic thing that makes the car fast, it's just the output of a function that takes the properties of a car and uses it to calculate some number. It's an effect, not a cause.
These are conditional statements, not logical arguments.Why am I not surprised you don't know how logic actually works?
posted by delmoi at 8:45 PM on April 11, 2013 [1 favorite]
Shalizi's arete is actually such a pitch-perfect parody of g that evolutionary psychologists have actually made more-or-less this exact argument in earnest.
posted by en forme de poire at 9:10 PM on April 11, 2013 [1 favorite]
posted by en forme de poire at 9:10 PM on April 11, 2013 [1 favorite]
Well, what I intended with my "Anecdote time" post was to try to pull us back a little from the trees so we could see the forest. Occasionally when you're analyzing something very closely, you should take a quick sanity check.
So it's possible to make very complex arguments for or against the validity of g, but human intuition can still be valuable. There sure seems to be a general factor to intelligence, so we should look at arguments against its existence with a little more skepticism. Especially if so many people seem to want them to be true.
Your car analogy: sure, in that case the 'r' factor is worthless. How about we build a car piece-by-piece. You roll a 20-sided die and get 15. Write that in the "engine" column. You roll again and get a 7, and write that in the "transmission" column, and so on. You ask the eccentric car builder guy what the numbers are for and he says "oh, a higher number gets you a higher-performing part. But first, roll one more die". You roll that die and get an 8, and that number is added to each of your sub-scores. Then you build your car.
If there was no general factor, its possible that different measures of cognitive ability could be correlated, but sheesh, scores on completely verbal tests correlate to scores on completely mathematical tests-- tests that have no other reason to be correlated at all. How do you explain that without some sort of underlying factor?
posted by grahamsletter at 9:26 PM on April 11, 2013
So it's possible to make very complex arguments for or against the validity of g, but human intuition can still be valuable. There sure seems to be a general factor to intelligence, so we should look at arguments against its existence with a little more skepticism. Especially if so many people seem to want them to be true.
Your car analogy: sure, in that case the 'r' factor is worthless. How about we build a car piece-by-piece. You roll a 20-sided die and get 15. Write that in the "engine" column. You roll again and get a 7, and write that in the "transmission" column, and so on. You ask the eccentric car builder guy what the numbers are for and he says "oh, a higher number gets you a higher-performing part. But first, roll one more die". You roll that die and get an 8, and that number is added to each of your sub-scores. Then you build your car.
If there was no general factor, its possible that different measures of cognitive ability could be correlated, but sheesh, scores on completely verbal tests correlate to scores on completely mathematical tests-- tests that have no other reason to be correlated at all. How do you explain that without some sort of underlying factor?
posted by grahamsletter at 9:26 PM on April 11, 2013
grahamsletter: "If there was no general factor, its possible that different measures of cognitive ability could be correlated, but sheesh, scores on completely verbal tests correlate to scores on completely mathematical tests-- tests that have no other reason to be correlated at all. How do you explain that without some sort of underlying factor?"
What is a completely mathematical test? Most math tests aren't just about how well you can crunch arithmetic, but involve logical reasoning, which is related in a very intimate way to verbal reasoning. Math is a language, so I would expect being good at language in general to correlate with being good at math.
There are more basic skills that underlie both disciplines, such as the manipulation of formal symbolic systems. The abilities to commit things to memory, and recall them at appropriate times are other, different abilities that underly both disciplines. Just because they share underlying factors doesn't mean there has to be only one factor.
posted by I-Write-Essays at 9:37 PM on April 11, 2013 [1 favorite]
What is a completely mathematical test? Most math tests aren't just about how well you can crunch arithmetic, but involve logical reasoning, which is related in a very intimate way to verbal reasoning. Math is a language, so I would expect being good at language in general to correlate with being good at math.
There are more basic skills that underlie both disciplines, such as the manipulation of formal symbolic systems. The abilities to commit things to memory, and recall them at appropriate times are other, different abilities that underly both disciplines. Just because they share underlying factors doesn't mean there has to be only one factor.
posted by I-Write-Essays at 9:37 PM on April 11, 2013 [1 favorite]
Granted, but speed of processing (as measured by reaction time), working memory capacity, and spatial reasoning also correlate to each other and to math and language ability.
But yeah, if symbol manipulation ability underlies both verbal skills and math skills (and it probably does, you sound right, at least), then it would seem to be a very important factor.
posted by grahamsletter at 9:49 PM on April 11, 2013
But yeah, if symbol manipulation ability underlies both verbal skills and math skills (and it probably does, you sound right, at least), then it would seem to be a very important factor.
posted by grahamsletter at 9:49 PM on April 11, 2013
grahamsletter: Another explanation is that the math and verbal abilities are in fact completely independent capabilities but are both affected in their development by a common factor (e.g. nutrition or the temperature of your crib or other causes that are also not g).
I wrote the above, then went back to Shalizi to see whether he agreed or not, and discovered that I had committed plagiarism:
posted by daveliepmann at 9:53 PM on April 11, 2013 [3 favorites]
I wrote the above, then went back to Shalizi to see whether he agreed or not, and discovered that I had committed plagiarism:
Suppose for example, that there really were just three distinct skills tapped by intelligence tests — verbal ability, spatial ability, and problem-solving ability. Suppose further (for clarity, not plausibility) that these three abilities were localized in completely distinct parts of the brain, that genes which influenced one had no effect on the others, that they could be trained separately without any transfer of learning, etc. One would then say that these were, indeed, causally distinct and separable talents. They might nonetheless well be positively correlated in a factor analysis, because certain environmental influences would affect them all the same way (nutrition, disease, stress), and even if there was no transfer of training, social processes would tend to correlate verbal schooling with, say, schooling in problem-solving. The higher-level factor associated with their correlations would then just be something like "quality of the developmental environment", and not another, more general mental ability.Sure, there are underlying factors. But it would be quite odd to call childhood nutrition/disease/stress The One g when what we're really talking about is (for the sake of argument) three distinct cognitive capabilities.
posted by daveliepmann at 9:53 PM on April 11, 2013 [3 favorites]
daveliepmann: Correct me if I'm wrong, but the correlations among mental abilities still hold after controlling for environmental factors.
Also, most people say that intelligence is not completely hereditary and not completely environmental. The data seem to bear out the intuition. I think a number given upthread somewhere was 70% genetics (and/or epigenetics or whatever), 30% nurture.
Now, let's say that the three supposedly-completely-distinct cognitive skills are cross-correlated. Unless environmental factors explain 100% of this, there will be an inborn g factor, so the skills can't be totally distinct.
posted by grahamsletter at 10:02 PM on April 11, 2013
Also, most people say that intelligence is not completely hereditary and not completely environmental. The data seem to bear out the intuition. I think a number given upthread somewhere was 70% genetics (and/or epigenetics or whatever), 30% nurture.
Now, let's say that the three supposedly-completely-distinct cognitive skills are cross-correlated. Unless environmental factors explain 100% of this, there will be an inborn g factor, so the skills can't be totally distinct.
posted by grahamsletter at 10:02 PM on April 11, 2013
When controlling for environmental factors in these tests, as I understand it, they control between people taking the test, not between different capacities that cause a single person taking the test to do well across multiple domains.
And no, it does not follow that the remainder after accounting for environmental factors equals g. If we muddle the example a bit, it could be the case that Cognitive Trait A is related (not correlated, which we are positing that all three of A, B, and C are, but in fact causally related) to Cognitive Trait B, and Cognitive Trait B is separate related (again, causally) to Cognitive Trait C. Then where is the unary g? It is still at least two factors: the common cause AB, and the common cause of BC.
To be charitable, you could be saying that g is by definition the entire remainder after accounting for non-heritable factors. That doesn't make g unary, nor interesting, nor a causal factor. It is stamp collecting: putting a name on a batch of unrelated things. That's fine, but you can't then turn around and refer to g as a causal factor (at least not coherently), because it's the result of a set of causes rather than a cause itself.
posted by daveliepmann at 10:21 PM on April 11, 2013
And no, it does not follow that the remainder after accounting for environmental factors equals g. If we muddle the example a bit, it could be the case that Cognitive Trait A is related (not correlated, which we are positing that all three of A, B, and C are, but in fact causally related) to Cognitive Trait B, and Cognitive Trait B is separate related (again, causally) to Cognitive Trait C. Then where is the unary g? It is still at least two factors: the common cause AB, and the common cause of BC.
To be charitable, you could be saying that g is by definition the entire remainder after accounting for non-heritable factors. That doesn't make g unary, nor interesting, nor a causal factor. It is stamp collecting: putting a name on a batch of unrelated things. That's fine, but you can't then turn around and refer to g as a causal factor (at least not coherently), because it's the result of a set of causes rather than a cause itself.
posted by daveliepmann at 10:21 PM on April 11, 2013
I stand corrected, then.
It seems like it will be impossible to determine if there is in fact a single causal g until after the Human Brain Project or something similar is complete. Actually, now that I think about it, there are probably other important factors, since the statistically-derived g only explains around 7/10 of subtest results (I forget what the number is, but I think it's around 0.7).
Until we have a definitive answer, we'll have to go by what we know and what we can infer.
posted by grahamsletter at 10:50 PM on April 11, 2013
It seems like it will be impossible to determine if there is in fact a single causal g until after the Human Brain Project or something similar is complete. Actually, now that I think about it, there are probably other important factors, since the statistically-derived g only explains around 7/10 of subtest results (I forget what the number is, but I think it's around 0.7).
Until we have a definitive answer, we'll have to go by what we know and what we can infer.
posted by grahamsletter at 10:50 PM on April 11, 2013
Tanizaki,
Are you every going to provide a link to a respectable scholar who attempts to rebut's Shalizi's argument?
Doing so would convince more people than simply linking to stormfrontesque websites. Thanks!
Please forgive me, but I have posted links to the journals Science (academic journal of the American Association for the Advancement of Science), American Psychologist (the journal of the APA), Psychological Bulletin (also an APA publication), and Intelligence, the official journal of the International Society for Intelligence Research. (those last three are all links to the same previous comment of mine). If these are your ideas of "stormfrontesque websites", there is no such thing as legitimate research.
By your demand, you have shown that you have not read the links I have provided. These are not the first few Google hits I found. These are articles in some of the world's top journals. Please do not come back until you have read and understood them if you want me to consider assuming that you are arguing in good faith.
I find it novel that the cries are for me to LINK! LINK! LINK! yet those who deny the consensus view of general intelligence bear no such onus.
posted by Tanizaki at 6:46 AM on April 12, 2013
Are you every going to provide a link to a respectable scholar who attempts to rebut's Shalizi's argument?
Doing so would convince more people than simply linking to stormfrontesque websites. Thanks!
Please forgive me, but I have posted links to the journals Science (academic journal of the American Association for the Advancement of Science), American Psychologist (the journal of the APA), Psychological Bulletin (also an APA publication), and Intelligence, the official journal of the International Society for Intelligence Research. (those last three are all links to the same previous comment of mine). If these are your ideas of "stormfrontesque websites", there is no such thing as legitimate research.
By your demand, you have shown that you have not read the links I have provided. These are not the first few Google hits I found. These are articles in some of the world's top journals. Please do not come back until you have read and understood them if you want me to consider assuming that you are arguing in good faith.
I find it novel that the cries are for me to LINK! LINK! LINK! yet those who deny the consensus view of general intelligence bear no such onus.
posted by Tanizaki at 6:46 AM on April 12, 2013
There is not a consensus as you allege.
For example:
See 'The Myth of Intelligence' by Henry D. Schlinger in The Psychological Record.
"This paper argues that a concept of intelligence as anything more than a label for various behaviors in their contexts is a myth and that a truly scientific understanding of the behaviors said to reflect intelligence can come only from a functional analysis of those behaviors in the contexts in which they are observed."
Moreover:
"“When we looked at the data, the bottom line is the whole concept of IQ — or of you having a higher IQ than me — is a myth,” said Dr. Adrian Owen, the study’s senior investigator and the Canada Excellence Research Chair in Cognitive Neuroscience and Imaging at the university’s Brain and Mind Institute. “There is no such thing as a single measure of IQ or a measure of general intelligence." link here.
A quote from a different author of the same study:
"We can all think of people that have poor reasoning and brilliant memories, or fantastic language skills but aren't so hot at reasoning, and so on. Now once and for all we can say there is not a single measure such as IQ which captures all the intelligence that you see in people."
posted by MisantropicPainforest at 7:26 AM on April 12, 2013
For example:
See 'The Myth of Intelligence' by Henry D. Schlinger in The Psychological Record.
"This paper argues that a concept of intelligence as anything more than a label for various behaviors in their contexts is a myth and that a truly scientific understanding of the behaviors said to reflect intelligence can come only from a functional analysis of those behaviors in the contexts in which they are observed."
Moreover:
"“When we looked at the data, the bottom line is the whole concept of IQ — or of you having a higher IQ than me — is a myth,” said Dr. Adrian Owen, the study’s senior investigator and the Canada Excellence Research Chair in Cognitive Neuroscience and Imaging at the university’s Brain and Mind Institute. “There is no such thing as a single measure of IQ or a measure of general intelligence." link here.
A quote from a different author of the same study:
"We can all think of people that have poor reasoning and brilliant memories, or fantastic language skills but aren't so hot at reasoning, and so on. Now once and for all we can say there is not a single measure such as IQ which captures all the intelligence that you see in people."
posted by MisantropicPainforest at 7:26 AM on April 12, 2013
There is not a consensus as you allege.
You know that consensus does not mean "100% agreement", right? You know that there are biologists who are intelligent design advocates, yes? Despite that fact, it is accurate to say that the consensus amongst the experts with relevant expertise is that intelligent design is hokum.
posted by Tanizaki at 8:02 AM on April 12, 2013
You know that consensus does not mean "100% agreement", right? You know that there are biologists who are intelligent design advocates, yes? Despite that fact, it is accurate to say that the consensus amongst the experts with relevant expertise is that intelligent design is hokum.
posted by Tanizaki at 8:02 AM on April 12, 2013
Please forgive me, but I have posted links to the journals Science (academic journal of the American Association for the Advancement of Science), American Psychologist (the journal of the APA), Psychological Bulletin (also an APA publication), and Intelligence, the official journal of the International Society for Intelligence Research. (those last three are all links to the same previous comment of mine). If these are your ideas of "stormfrontesque websites", there is no such thing as legitimate research.
Did you read those papers? They really don't establish a consensus at all. In fact, the APA piece, (not-so-coincidentally in response to The Bell Curve, the modern touchpoint of scientific racism) finishes with the following:
posted by PMdixon at 8:23 AM on April 12, 2013 [4 favorites]
Did you read those papers? They really don't establish a consensus at all. In fact, the APA piece, (not-so-coincidentally in response to The Bell Curve, the modern touchpoint of scientific racism) finishes with the following:
7. It is widely agreed that standardized tests do notPro-tip: This is language that implies the opposite of a consensus.
sample all forms of intelligence. Obvious examples in-
clude creativity, wisdom, practical sense, and social sen-
sitivity; there are surely others. Despite the importance
of these abilities we know very little about them: how
they develop, what factors influence that development,
how they are related to more traditional measures.
In a field where so many issues are unresolved and
so many questions unanswered, the confident tone that
has characterized most of the debate on these topics is
clearly out of place. The study of intelligence does not
need politicized assertions and recriminations; it needs
self-restraint, reflection, and a great deal more research.
The questions that remain are socially as well as scien-
tifically important. There is no reason to think them un-
answerable, but finding the answers will require a shared
and sustained effort as well as the commitment of sub-
stantial scientific resources. Just such a commitment is
what we strongly recommend.
posted by PMdixon at 8:23 AM on April 12, 2013 [4 favorites]
More from that paper:
posted by PMdixon at 8:26 AM on April 12, 2013 [1 favorite]
These complex patterns of correlation can be clar-Doesn't sound like a consensus to me.
ified by factor analysis, but the results of such analyses
are often controversial themselves. Some theorists (e.g.,
Spearman, 1927) have emphasized the importance of a
general factor, g, which represents what all the tests have
in common; others (e.g., Thurstone, 1938) focus on more
specific group factors such as memory, verbal compre-
hension, or number facility. As we shall see in Section 2,
one common view today envisages something like a hi-
erarchy of factors with g at the apex. But there is no full
agreement on what g actually means: [emphasis mine]
it has been described as a mere statistical regularity
(Thomson, 1939), a kind of mental energy (Spearman, 1927),
a generalized abstract reasoning ability (Gustafsson, 1984),
or an index measure of neural processing speed (Reed & Jensen, 1992).
posted by PMdixon at 8:26 AM on April 12, 2013 [1 favorite]
Note for readers, this kind of appeal-to-bafflement is a surefire way to ensure you never learn anything you don't already know:
I'm harping on this because this quote has been bugging me since I read it:
Given that Jensen would, presumably, know more about all of these topics than I do -- I'd like to hope that that's not being too charitable! -- it doesn't speak well of his intellectual efforts that he's resorting to the argument from bafflement and appeal-to-common-sense, at least as quoted here. I'd feel much better if those "surprising" correlations and "surprising" non-correlations were the jumping-off points for further investigation, but at least there are real cognitive scientists out there picking up the slack.
Now, without further ado, the "why forward and backward would be different" explanation: for a forward test it's possible to pass by merely "tape recording" what you heard and repeating it back; a short-term memory for recently-heard sounds seems anecdotally like something almost everyone has operating automatically. For a "reverse" test it'd seemingly be necessary to "parse" that audio into a more-abstract representation (e.g., actual digits as digits, or as the actual number) and then somehow recite the digits backwards by making reference to that representation.
There's some overlap between the capabilities needed between the two tasks -- do the ears work? do you know what numbers sound like? did you have a good breakfast? -- but beyond that shared core the reverse test involves a lot of capabilities that the forward one seemingly does not. (Note that one *could* solve the forward test using a subset of the capabilities in the reverse, but it's not necessary to do so).
Assuming you accept this breakdown, you have your explanation for the "unexpectedly" weak correlation: the correct analysis is not "remember digit sequence" vs. "remember digit sequence + apply 'reverse' function", it's "piggyback on always-on, automatic short-term audio buffers" vs. "parse audio input into alternative representation, then derive reverse sequence via said representation". Analyzed that way there's no a priori reason to expect strong correlation, and I'd say that they're only seemingly "similar" if you give no thought to what each task actually entails.
posted by hoople at 9:10 AM on April 12, 2013 [2 favorites]
scores on completely verbal tests correlate to scores on completely mathematical tests-- tests that have no other reason to be correlated at all...I mean, this has been addressed already in-thread, but still, are you *really* trying to say that if you were tasked with postulating a list of, say, 5+ distinct candidate mechanisms that might explain this fact, that you'd be unable to do so? Not really a good way to distinguish yourself, at least in my book.
I'm harping on this because this quote has been bugging me since I read it:
Another puzzle in terms of sampling theory is that tests such as forward and backward digit span memory, which must tap many common elements, are not as highly correlated as are, for instance, vocabulary and block designs, which would seem to have few elements in common....(which is originally from Jensen). If these digit tests are like the ones I can remember having taken -- various digit sequences are read off to you, then after some delay you're asked to recite them forwards or backwards, perhaps after a delay or after having heard another sequence or two -- then with even an amateur's background in cognitive science it's rather easy to postulate explanations for why forward and backward digit span memory would be weakly correlated (see below); if it's that easy for me to come up with plausible "why they're actually quite different tasks, despite the superficial similarity" (and quite literally instantly, as my "first thought" on reading this quote), then it isn't unreasonable that there're similarly-obvious potential explanations for why "vocabulary" and "block designs" tests are more similar they appear on the surface.
Given that Jensen would, presumably, know more about all of these topics than I do -- I'd like to hope that that's not being too charitable! -- it doesn't speak well of his intellectual efforts that he's resorting to the argument from bafflement and appeal-to-common-sense, at least as quoted here. I'd feel much better if those "surprising" correlations and "surprising" non-correlations were the jumping-off points for further investigation, but at least there are real cognitive scientists out there picking up the slack.
Now, without further ado, the "why forward and backward would be different" explanation: for a forward test it's possible to pass by merely "tape recording" what you heard and repeating it back; a short-term memory for recently-heard sounds seems anecdotally like something almost everyone has operating automatically. For a "reverse" test it'd seemingly be necessary to "parse" that audio into a more-abstract representation (e.g., actual digits as digits, or as the actual number) and then somehow recite the digits backwards by making reference to that representation.
There's some overlap between the capabilities needed between the two tasks -- do the ears work? do you know what numbers sound like? did you have a good breakfast? -- but beyond that shared core the reverse test involves a lot of capabilities that the forward one seemingly does not. (Note that one *could* solve the forward test using a subset of the capabilities in the reverse, but it's not necessary to do so).
Assuming you accept this breakdown, you have your explanation for the "unexpectedly" weak correlation: the correct analysis is not "remember digit sequence" vs. "remember digit sequence + apply 'reverse' function", it's "piggyback on always-on, automatic short-term audio buffers" vs. "parse audio input into alternative representation, then derive reverse sequence via said representation". Analyzed that way there's no a priori reason to expect strong correlation, and I'd say that they're only seemingly "similar" if you give no thought to what each task actually entails.
posted by hoople at 9:10 AM on April 12, 2013 [2 favorites]
These are articles in some of the world's top journals.The first thing that occurred to me was -- well, the bell curve survey article was written roughly ~18 years ago -- and the others are very particular studies without replication (always a bad idea without some general grounding in the field) -- maybe the consensus has changed somewhat?
Not really... and a lot :-) (note that the link to the actual paper in in the article, I decided to link to a summary for a bit of context).
I'm not sure why people are saying 'g' doesn't exist (ok, I can guess), but the updated survey still talks about g factor (and IQ) as valid summary measures. Most of the new information appears to be about how g correlates with other things physical and otherwise. The consensus has gotten much murkier about the correlations, but not really about the existence of a 'g' itself.
I recommend reading the article, it's actually quite interesting to see how things have changed (and remained the same, and gotten more nuanced) in the field.
I think a lot of people here are conflating the existence of a general measure with the consequences of humans using such measure. It's not the measure which is the bad thing -- it's what you do with the results. If you the read the paper I linked, you'll see the measure is not up for a huge debate (ok, there is some), but what its actual properties are *is* (e.g. in terms if heritability within a certain social economic statuses, etc.).
What's more interesting is not 'g', but what the heck you *do* about it? If a group scores lower, what does that actually mean? Does it mean they shouldn't breed (e.g the (sorry) vaguely racist Idiocracy and the most certainly racist Bell Curve), or does it just mean that we're ignoring something very crucial to cognitive development within those groups (eg. SES status)?
posted by smidgen at 3:51 PM on April 12, 2013
smidgen: the thing is, in the local context, Shalizi's article pointed out that a g-like measurement is inevitably going to pop out some collection of tests that're all positively-correlated with each other. So arguing that the g doesn't exist is a fool's errand, which at least I'm not will to undertake (there are plenty of other hills to go die on).
The question is one of interpretation; one the one hand you have people like Shalizi, who take the stance that given that (a) the existence of a g-like factor is a mathematical inevitability and that (b) there are many mechanisms that could produce analogous measurements (the 'sampling' argument), that it is then both foolish to deny the existence of g the statistical metric and foolish to read much into it on its own.
On the other side you have folks like Tanizaki: those who appear to conflate the fact that there's an expert consensus about the *existence* of a g-like factor with the "fact" that the expert consensus also agrees with Tanizaki's personal interpretation of the g-factor; that's, well, a lot less certain, at the least. Or, similarly, the author of the linked article, it would appear.
To the extent I have any sense of the expert consensus it shies far closer to Shalizi than to Tanizaki and the D-dude; not much contention about the existence of the statistical measure, but not much strong consensus on its general interpretation or on any specific point of its interpretation. From where I sit this leaves the non-Shalizi-ites looking a bit distasteful insofar as they seem to be trying to borrow credibility by falsely equating belief-in-g-the-measurement with belief-in-their-own-particular-interpretation.
It's disappointing, too; usually kliuless's posts are full of interesting articles, and I was hoping to get a fresh look at Shalizi's article, which I'd read ages ago; instead I got something full of weak reasoning and appeals-to-authorial-bafflement, the latter of which is *particularly* weak when adjacent to large quantities of the former.
But, that aside, I'd say the policy implications are less interesting than the directions of inquiry the various "surprising" facts all seem to naturally suggest, as above with the forward/backward digit span; I can't envision being interested in the nature of "intelligence" and not wanting to figure out, in detail, exactly what's going on there...but anyways, I think I'll bow out now.
posted by hoople at 5:16 PM on April 12, 2013 [1 favorite]
The question is one of interpretation; one the one hand you have people like Shalizi, who take the stance that given that (a) the existence of a g-like factor is a mathematical inevitability and that (b) there are many mechanisms that could produce analogous measurements (the 'sampling' argument), that it is then both foolish to deny the existence of g the statistical metric and foolish to read much into it on its own.
On the other side you have folks like Tanizaki: those who appear to conflate the fact that there's an expert consensus about the *existence* of a g-like factor with the "fact" that the expert consensus also agrees with Tanizaki's personal interpretation of the g-factor; that's, well, a lot less certain, at the least. Or, similarly, the author of the linked article, it would appear.
To the extent I have any sense of the expert consensus it shies far closer to Shalizi than to Tanizaki and the D-dude; not much contention about the existence of the statistical measure, but not much strong consensus on its general interpretation or on any specific point of its interpretation. From where I sit this leaves the non-Shalizi-ites looking a bit distasteful insofar as they seem to be trying to borrow credibility by falsely equating belief-in-g-the-measurement with belief-in-their-own-particular-interpretation.
It's disappointing, too; usually kliuless's posts are full of interesting articles, and I was hoping to get a fresh look at Shalizi's article, which I'd read ages ago; instead I got something full of weak reasoning and appeals-to-authorial-bafflement, the latter of which is *particularly* weak when adjacent to large quantities of the former.
But, that aside, I'd say the policy implications are less interesting than the directions of inquiry the various "surprising" facts all seem to naturally suggest, as above with the forward/backward digit span; I can't envision being interested in the nature of "intelligence" and not wanting to figure out, in detail, exactly what's going on there...but anyways, I think I'll bow out now.
posted by hoople at 5:16 PM on April 12, 2013 [1 favorite]
smidgen: vaguely racist Idiocracy
Sorry, I know this is completely off-topic, but I'm curious. How is Idiocracy vaguely racist? It's been a while since I watched it, but I didn't sense any racism then.
posted by grahamsletter at 5:40 PM on April 12, 2013
Sorry, I know this is completely off-topic, but I'm curious. How is Idiocracy vaguely racist? It's been a while since I watched it, but I didn't sense any racism then.
posted by grahamsletter at 5:40 PM on April 12, 2013
I haven't seen it, but it's certainly a 'eugenics-ist', the belief that the population is getting stupider because stupid people have more children.
posted by delmoi at 8:47 PM on April 12, 2013
posted by delmoi at 8:47 PM on April 12, 2013
I'm not sure why people are saying 'g' doesn't exist (ok, I can guess), but the updated survey still talks about g factor (and IQ) as valid summary measures. Most of the new information appears to be about how g correlates with other things physical and otherwise. The consensus has gotten much murkier about the correlations, but not really about the existence of a 'g' itself.When I say that 'g' doesn't exist, what I'm simply saying is that I don't think that there is some specific, singular 'g' that is the cause of performing well on tests. Obviously you can have people take a bunch of IQ tests, average the results, and call that number 'g', but what is the point of that, exactly? 'g' doesn't tell you anything that the results of the tests didn't.
Obviously there is the whole epistemic question about what does and does not 'exist'. But it appears that the racists, the "Human Bio-Diversity" people and so on think that 'g' is the cause of the ability to do well on IQ tests, and related primarily to genetics.
On the other hand it seems perfectly obvious that 'ability to do well on IQ tests' would, in fact, be an obvious cause of people doing well on IQ tests.
(a) the existence of a g-like factor is a mathematical inevitability and that (b) there are many mechanisms that could produce analogous measurements (the 'sampling' argument), that it is then both foolish to deny the existence of g the statistical metric and foolish to read much into it on its own.Again, you have to be careful about epistemology here. Just because you can do a bunch of math on some real world data and predict a consistent result does not mean that the result is caused by some real thing in the real world.
It's like Zipf's law. When the law about the distribution of word lengths was discovered, it was believed to indicate something about how the human brain works. Later on it was discovered that a completely randomly generated language would have the same distribution.
In other words, you can have a mathematical result both consistent and at the same time totally meaningless.
Your car analogy: sure, in that case the 'r' factor is worthless. How about we build a car piece-by-piece. You roll a 20-sided die and get 15. Write that in the "engine" column. You roll again and get a 7, and write that in the "transmission" column, and so on. You ask the eccentric car builder guy what the numbers are for and he says "oh, a higher number gets you a higher-performing part.Let's suppose rather then rolling a die and getting numbers, you randomly selected components. You would still have an 'r' factor. And that's the point, 'r' is totally meaningless and tells you nothing at all about how cars are made. It gives you zero new information.
You ask the eccentric car builder guy what the numbers are for and he says "oh, a higher number gets you a higher-performing part. But first, roll one more die". You roll that die and get an 8, and that number is added to each of your sub-scores. Then you build your car.What does this die represent here, a voodoo ritual performed on the car or something?
The problem with this is that it's physically impossible. that might work in a D&D game, but not in the real world. You can't "roll an 8" there has to be some difference. Maybe it's better aerodynamics, maybe it's something exotic like break steer or a new air intake design. The point is the performance of the car is determined by the components of the car, and nothing else.
It should be pretty obvious from the fact that you had to come up with an example that can't even actually happen in the real world that you're on pretty weak ground.
Also, your 'Anecdote' was not an anecdote. It was a hypothetical. An anecdote is a thing that has actually happened in the real world. Anecdotes have at least a little empirical value - you know at least what happened was possible, but not much when you're talking about people. Hypothetical are pretty useless, they can help clarify thinking but they can also make things worse by coming up with hypothetical situations that can't actually happen, like a magical die that improves cars performance without actually doing anything physical to it.
It seems like it will be impossible to determine if there is in fact a single causal g until after the Human Brain Project or something similar is complete.If you have no evidence for a thing, like a causal 'g' then you can assume it does not exist. Science isn't in the business of "disproving" random things people wish existed but can't measure.
In fact, in order to say something in Karl Popper's philosophy of science in order for something to be 'scientific' you need to have some experiment that would disprove it. If you do that experiment and it doesn't disprove it, that's evidence for it's existence (but not proof)
Unless you could come up with an experiment that would disprove a causal 'g' then it's not scientific.
What experiment could you do that would disprove it, in your mind? If you can't think of one, it's not science. (I told you what I thought: if you can find more then one biologically measurable thing that impacts intelligence, then that would prove intelligence has multiple causes, not just one)
posted by delmoi at 9:59 PM on April 12, 2013 [1 favorite]
What does this die represent here, a voodoo ritual performed on the car or something?The last die was supposed to affect the car physically by changing the parts that you get before you build the car. You could then estimate the value of the last die by using statistics. The last die changes the performance of the car by affecting all of the components and can be measured later to get a single number.
Also, your 'Anecdote' was not an anecdote.Well, the first two things I said relied on anecdote (the "surely you have noticed" stuff).
I don't know, I'm obviously not very good at debate, and the position I've taken is hard to defend since any evidence for or against it is rooted in math that I don't understand, so I'm going to sort of sneak away from that.
Here's my position, though:
It seems obvious to me that there are multiple components to intelligence. Obviously some people are better at some things than at other things. But there is still correlation: if someone is really good at math, you should bet that they're above average at other mental skills. They may not be, but it is much more likely that they are than they aren't.
Now, it doesn't immediately follow that there is some general factor underlying cognitive skills, but I thought it was pretty well accepted that there was. This thread actually surprised me. Maybe I'll go back through and take a look at some of the links to see if the controversy is real or just caused by a small minority of scientists a la climate change or evolution.
Even if there were a 'g', it doesn't have to be genetic, but common sense and the studies that have been done indicate that it would be at least partially caused by genetics, and partially environmental. Breast feeding raises IQ by 7 points. Eliminating lead and other toxins has effects like that, too. Having access to pre-k has positive effects that last throughout the lifetime.
But twin studies, adoption studies, and the fact that, come on, people know families that produce loads of smart people and others that don't, all indicate a genetic component.
That is what seems to be so icky here.
Yes, I know that pretty much everyone that goes around blogging about 'g' is a
Intelligence does not measure value. This is the message that needs to get out into society. There may or may not be genetically determined differences between people, but even if there are, it doesn't mean anything. People are people. We are all in this together and we are all of equal worth. Our culture, despite its anti-intellectual streak, is very brain-focused. There is no uproar over any other differences in human traits except for intelligence.
But going to such great lengths to deny that differences could even exist is actually counterproductive. It actually helps reinforce the idea that intelligence measures worth. Equating the idea of g with racism implies that if g turns out to be right, then racist ideas are right. It makes the stakes awfully high and places a huge bet on the idea that g will be discredited, which, admit it, you don't know that it will be when all is said and done.
That's the problem I have with the reflexive anti-g stance. It is well-meaning but wrong. I get the same reaction to it that I do to people who inadvertently "other" people who are different by being overly serious and patronizing, like in this South Park episode. (link)
posted by grahamsletter at 12:10 AM on April 13, 2013 [1 favorite]
What's more interesting is not 'g', but what the heck you *do* about it?
yea! in case you missed some of noah smith's comments, there were some pretty good ones (aimed at the appropriate level(s) ;)
posted by kliuless at 10:10 AM on April 13, 2013
yea! in case you missed some of noah smith's comments, there were some pretty good ones (aimed at the appropriate level(s) ;)
- [W]hat do we actually want to use the measure for? What I want, of course, is smart-ification technology. So for that purpose, the question of "What is g?" is crucially important.
- So all the focus on race is pointless, and smacks of an ulterior motive... guys like "Dalliard" and Steve Sailer just want to keep black guys from banging their daughters & girlfriends... What we need is to discover the physiological mechanisms of intelligence, so we can boost it with technology. That is what Steve Hsu is starting to do.
posted by kliuless at 10:10 AM on April 13, 2013
What I want, of course, is smart-ification technology.
You should be happy then! The Flynn effect is your friend.
posted by asterix at 6:53 PM on April 13, 2013
You should be happy then! The Flynn effect is your friend.
posted by asterix at 6:53 PM on April 13, 2013
Steve Hsu is a core member of the Cognitive Genomics Project run by the Beijing Genomics Institute (BGI). Steven Pinker is also involved in the project. Hsu notes in relation to Shalizi's essay & the response above:
I recommend this well written refutation of Cosma Shalizi's much loved (in certain quarters) g, a Statistical Myth, an attack on the general factor of intelligence. Over the years I have not encountered a single endorser of Shalizi's article who actually understands the relevant subject matter. His article is loved for its reassuring conclusions, not the strength of its arguments. I am sure many "thinkers" resisted Darwinism, the abandonment of geocentrism, and even the notion that the Earth is a sphere, for similar psychological reasons. Some pessimists (speaking, for example, of the quantum revolution in the early 20th century) remarked that science advances one funeral at a time, as the older generation passes away in favor of the next, more open-minded, one. In the case of g it appears we have regressed significantly under relentless attack; social science papers from 50 years ago often seem more clear headed and precise than ones I read today. All battles must be fought and refought again a decade or two later.
The BGI research on the genetic architecture of intelligence is pretty interesting. Hsu has given presentations on the project at Google with the slides here. That's probably a good starting point for the latest research in this area. Also, for a good summary which discusses 'g' see Paul Thompson & Jeremy Gray's paper in Nature on the neurobiology of intelligence.
posted by Boffin77 at 2:09 AM on April 15, 2013 [1 favorite]
I recommend this well written refutation of Cosma Shalizi's much loved (in certain quarters) g, a Statistical Myth, an attack on the general factor of intelligence. Over the years I have not encountered a single endorser of Shalizi's article who actually understands the relevant subject matter. His article is loved for its reassuring conclusions, not the strength of its arguments. I am sure many "thinkers" resisted Darwinism, the abandonment of geocentrism, and even the notion that the Earth is a sphere, for similar psychological reasons. Some pessimists (speaking, for example, of the quantum revolution in the early 20th century) remarked that science advances one funeral at a time, as the older generation passes away in favor of the next, more open-minded, one. In the case of g it appears we have regressed significantly under relentless attack; social science papers from 50 years ago often seem more clear headed and precise than ones I read today. All battles must be fought and refought again a decade or two later.
The BGI research on the genetic architecture of intelligence is pretty interesting. Hsu has given presentations on the project at Google with the slides here. That's probably a good starting point for the latest research in this area. Also, for a good summary which discusses 'g' see Paul Thompson & Jeremy Gray's paper in Nature on the neurobiology of intelligence.
posted by Boffin77 at 2:09 AM on April 15, 2013 [1 favorite]
Everyone believes in general intelligence and that it is heritable when it counts: when making a withdrawal at the sperm bank.
posted by Tanizaki at 6:28 AM on April 15, 2013
posted by Tanizaki at 6:28 AM on April 15, 2013
Tanizaki: Everyone believes in general intelligence and that it is heritable when it counts: when making a withdrawal at the sperm bank.... if by "everyone", you mean that vanishingly small group of people who make withdrawals from sperm banks, and also share a number of other traits like an inability to conceive naturally with their life partner, a fair amount of disposable income, and so on.
posted by IAmBroom at 9:40 AM on April 15, 2013
The last die was supposed to affect the car physically by changing the parts that you get before you build the car. You could then estimate the value of the last die by using statistics. The last die changes the performance of the car by affecting all of the components and can be measured later to get a single number.No, it would be impossible to distinguish that from a situation where N parts are chosen by N dice roles from a situation where N parts are chosen from N+1 dice. In the first case, you have 6N possible outcomes, each equally likely. And in the second case, you have 6N possible outcomes, each equally likely. If you actually have more then N parts then the last die roll is no different from any other, and you have 6N+1 different possibilities, all equally likely.
If you feel otherwise, feel free to show the actual math on how you think divining the value of the (N+1)th die role could be determined. I'll not hold my breath.
But going to such great lengths to deny that differences could even exist is actually counterproductive.No one is arguing some people are not smarter then other people. The argument is about whether or not there is some singular underlying causal effect. If 'g' is just the 'ability to do well on IQ tests' then what exactly is the point? That the ability to do well on IQ tests is correlated with doing well on IQ tests?
The basic problem with these 'g' people is that don't seem to understand the difference between cause and effect. If 'g' is an effect, then it is simply a label on 'ability to perform well on IQ tests'.
In fact, the "causal 'g' theory" is essentially a canonical example of something Occam's razor tells us to remove. If you specify {A, B, C, D, X, Y, Z} all 'cause' 'g', and that and that 'g' causes you to do well on IQ test, you lose nothing by simply stating that {A, B, C, D, X, Y, Z} cause people to do well on IQ tests.
posted by delmoi at 3:29 PM on April 15, 2013
That the ability to do well on IQ tests is correlated with doing well on IQ tests?
Well, it does correlate with neurophysical quantities eg. Thompson & Gray - The Neurobiology of Intelligence and correlates with a number of outcomes.The results show that the tests predict performance and are stable. It also seems it is heritable based on twin/adoption studies and recent molecular studies eg Hsu cites this recent paper:
Although the study failed to find any specific loci that are associated with intelligence, a global fit showed that a significant chunk of the heritability expected from twin and adoption studies is accounted for by SNPs. In other words, genetic similarity is correlated with similarity in g score, even though we don't know which genes are specifically responsible. The results of the study were expected from what we already knew: many genes of small effect, accounting for as much as .6 or so of narrow sense heritability.
posted by Boffin77 at 7:09 PM on April 15, 2013
Well, it does correlate with neurophysical quantities eg. Thompson & Gray - The Neurobiology of Intelligence and correlates with a number of outcomes.The results show that the tests predict performance and are stable. It also seems it is heritable based on twin/adoption studies and recent molecular studies eg Hsu cites this recent paper:
Although the study failed to find any specific loci that are associated with intelligence, a global fit showed that a significant chunk of the heritability expected from twin and adoption studies is accounted for by SNPs. In other words, genetic similarity is correlated with similarity in g score, even though we don't know which genes are specifically responsible. The results of the study were expected from what we already knew: many genes of small effect, accounting for as much as .6 or so of narrow sense heritability.
posted by Boffin77 at 7:09 PM on April 15, 2013
Well, I've abandoned my argument, conceding all points relating to g, but I'll try to do the math, because math is cool. Note that I am no longer trying to argue for the existence of g. I'm just doing a math problem here.
Let's add an abstraction to make this easier: to get the overall performance of the vehicle, we just average all of the values of all of the parts.
There are n m-sided dice for n parts.
Here is the first case, with no extra dice roll:
If we were to run this over many trials, we would expect the performance of the vehicle to conform to this sort of distribution (paste "median 5d20" into the box and click "calculate probabilities"), with the mean being (m+1)/2, the middle of possible values of the dice.
Now, the distribution will get narrower as n increases. If n is pretty large, it will be a very good bet that the performance of the car will be close to (m+1)/2.
That is where our second case comes in.
Second case, extra dice roll value added to all car parts:
This effectively shifts the distribution by whatever the value of that last die is. To estimate the value of the shift, you take the performance of the car and subtract (m+1)/2. If the value of the dice before the last one was added averages to a value close to (m+1)/2 (which it is more likely to do the larger n is), then this will be a very good estimate of the shift, and consequently the last die.
Here it is with concrete numbers:
20 20-sided dice. (n = 20, m = 20)
I generated some random numbers:
{4, 8, 8, 18, 13, 7, 16, 12, 5, 10, 18, 6, 5, 7, 4, 6, 5, 10, 10, 19}
mean = 9.55
Here's the final die: 6
Here's the numbers with the final die added:
{10, 14, 14, 24, 19, 13, 22, 18, 11, 16, 24, 12, 11, 13, 10, 12, 11, 16, 16, 25}
mean = 15.55
Now, to estimate the value of the final die, given only the mean and knowledge of m, we do this:
15.55 - (20+1)/2 = 5.05
Which is not far off from 6.
posted by grahamsletter at 1:42 AM on April 16, 2013
Let's add an abstraction to make this easier: to get the overall performance of the vehicle, we just average all of the values of all of the parts.
There are n m-sided dice for n parts.
Here is the first case, with no extra dice roll:
If we were to run this over many trials, we would expect the performance of the vehicle to conform to this sort of distribution (paste "median 5d20" into the box and click "calculate probabilities"), with the mean being (m+1)/2, the middle of possible values of the dice.
Now, the distribution will get narrower as n increases. If n is pretty large, it will be a very good bet that the performance of the car will be close to (m+1)/2.
That is where our second case comes in.
Second case, extra dice roll value added to all car parts:
This effectively shifts the distribution by whatever the value of that last die is. To estimate the value of the shift, you take the performance of the car and subtract (m+1)/2. If the value of the dice before the last one was added averages to a value close to (m+1)/2 (which it is more likely to do the larger n is), then this will be a very good estimate of the shift, and consequently the last die.
Here it is with concrete numbers:
20 20-sided dice. (n = 20, m = 20)
I generated some random numbers:
{4, 8, 8, 18, 13, 7, 16, 12, 5, 10, 18, 6, 5, 7, 4, 6, 5, 10, 10, 19}
mean = 9.55
Here's the final die: 6
Here's the numbers with the final die added:
{10, 14, 14, 24, 19, 13, 22, 18, 11, 16, 24, 12, 11, 13, 10, 12, 11, 16, 16, 25}
mean = 15.55
Now, to estimate the value of the final die, given only the mean and knowledge of m, we do this:
15.55 - (20+1)/2 = 5.05
Which is not far off from 6.
posted by grahamsletter at 1:42 AM on April 16, 2013
Boffin, heritability and a neurobiological basis for the calculated value of g doesn't tell us whether g exists as a unary causal factor. I mean, for the neurophysical side, the counterclaim here isn't that cognitive ability relies on non-physical properties. That would be silly to any materialist. I think we all agree that the brain is what's doing the thinking. Whether it is one or multiple structures in the brain, or one or multiple genetic factors involved, is the question. We have to distinguish between "g as a specific thing" and "g as a collection of factors that we give a name".
For heritability, well, that doesn't tell us anything about whether g is unary and causal either. As Shalizi notes, Independent factors, taken together for no reason, are heritable:
posted by daveliepmann at 7:52 AM on April 16, 2013 [1 favorite]
For heritability, well, that doesn't tell us anything about whether g is unary and causal either. As Shalizi notes, Independent factors, taken together for no reason, are heritable:
Height, in developed countries, has a heritability around 0.8. Blood triglyceride levels have a heritability of about 0.5. Thus the sum of height and triglycerides is heritable.I like how delmoi put it, because it reminds us where the contention lies: "If you specify {A, B, C, D, X, Y, Z} all 'cause' 'g', and that and that 'g' causes you to do well on IQ test, you lose nothing by simply stating that {A, B, C, D, X, Y, Z} cause people to do well on IQ tests." The debate as I see it is whether g is merely a name for {A, B, C, D, X, Y, Z} or it is its own thing. Or as the original article says:
If there are lots of positive correlations and I want to summarize them, then finding some factors and checking them by decomposing the variance is one reasonable trick. If I want to argue that there must be a preponderant common cause, it's no good to keep pointing out how much of the variance that first factor describes, when plenty of other, incompatible causal structures will give me that too.Taking a step back, I think it would be more productive to identify the specific structures related to specific cognitive abilities (e.g. working memory or the position of one's limbs) while also identifying ways to improve intelligence after someone's genes are determined (e.g. Carol Dweck's work on fixed versus growth mindsets (PDF)). I think any attempt to plant a flag in a specific spot and say "THIS is intelligence!" is misguided, much the way any attempt to plant a flag in a specific spot and say "THIS is consciousness!" is misguided.
posted by daveliepmann at 7:52 AM on April 16, 2013 [1 favorite]
Speaking of specific structures, did we ever discuss this paper ("Fractionating Human Intelligence", Neuron, 2012) on MeFi? (paywall, sorry) I feel like we did but I can't seem to find a link, so maybe I'm confusing a conversation I had somewhere else.
Honestly I think fMRI is still way too coarse grained to really characterize specific circuits in the brain, and I am not in love with their use of PCA here, but it is a pretty interesting study nonetheless. The money quote:
posted by en forme de poire at 8:59 PM on April 17, 2013 [1 favorite]
Honestly I think fMRI is still way too coarse grained to really characterize specific circuits in the brain, and I am not in love with their use of PCA here, but it is a pretty interesting study nonetheless. The money quote:
In behavioral factor analysis, ...the components or latent variables cannot be measured directly. Here, we have an objective measure of the extent to which the tasks are mixed, as we know, based on the functional neuroimaging data, the extent to which the tasks recruit spatially separated functional networks relative to rest. Consequently, it is possible to subdivide “g” into the proportion that is predicted by the mixing of tasks on multiple functional brain networks and the proportion that may be explained by other diffuse factors (Figure 3). ....there was little requirement for a diffuse higher-order “g” factor once the tendency for tasks to corecruit multiple functional brain networks was accounted for.(The independent components they end up identifying are "short-term memory", "reasoning," and "verbal.")
posted by en forme de poire at 8:59 PM on April 17, 2013 [1 favorite]
I had never seen that paper, but I dig it. Particularly this chart.
One of the most fascinating takeaways from the paltry amount of serious cognitive science reading I've done is that cases of brain damage (either accidental (in humans) or calculated (in, say, frogs)) as well as specific neurological disorders fairly clearly show that there are specific brain structures for a number of specific tasks: face recognition, tracking objects in your field of vision unconsciously, tracking objects in your field of vision consciously, keeping a mental picture of our limbs in space, and so on. So I suspect that there are some tasks where we have dedicated wetware, some tasks that are learned and less specific in their location, some simple tasks that are distributed across several distinct areas, all the way up to what the paper (I believe rightfully) calls "higher-order g", that is, looking at all these distinct things as a (made-up, non-causal) unit.
That's part of what makes a causal unary g so hard to swallow: the brain does a lot of things that we've clearly seen are orthogonal. My ability to recognize faces does not affect my ability to mentally rotate objects in space [citation needed].
posted by daveliepmann at 6:29 AM on April 18, 2013 [1 favorite]
One of the most fascinating takeaways from the paltry amount of serious cognitive science reading I've done is that cases of brain damage (either accidental (in humans) or calculated (in, say, frogs)) as well as specific neurological disorders fairly clearly show that there are specific brain structures for a number of specific tasks: face recognition, tracking objects in your field of vision unconsciously, tracking objects in your field of vision consciously, keeping a mental picture of our limbs in space, and so on. So I suspect that there are some tasks where we have dedicated wetware, some tasks that are learned and less specific in their location, some simple tasks that are distributed across several distinct areas, all the way up to what the paper (I believe rightfully) calls "higher-order g", that is, looking at all these distinct things as a (made-up, non-causal) unit.
That's part of what makes a causal unary g so hard to swallow: the brain does a lot of things that we've clearly seen are orthogonal. My ability to recognize faces does not affect my ability to mentally rotate objects in space [citation needed].
posted by daveliepmann at 6:29 AM on April 18, 2013 [1 favorite]
« Older "You don't need that many guitars. Bout two's... | Cardinal Perplexogram Newer »
This thread has been archived and is closed to new comments
The ironic part is that if I don't understand the relevant subject matter enough to actually endorse Shalizi's article, then I almost certainly don't understand it enough to endorse or even discuss Dalliard's refutation. Is this a fundamental issue with scientific discussions among laymen?
posted by muddgirl at 10:48 AM on April 11, 2013 [2 favorites]