[yummyfajitas]

"Intelligence" is a hidden variable g, which comes out of statistical tests of various mental examinations. Basically, give a bunch of people a french test, a math test, a physics test, a driving test, etc.

Run statistical tests for hidden variables, and you will discover this mysterious hidden variable 'g'. It is relatively independent of cultural knowledge: i.e., french is nearly independent of g (1), while physics, plumbing and loading irregularly sized boxes into a truck are highly correlated to g. This all comes out of the statistical analysis, and is not assumed apriori.

Then look for correlations between g and career outcome (and other such things), you'll discover strong correlations there as well. Physicists tend to have high g, janitors low g, etc. For instance, I've never met a math/physics/eng faculty member with an IQ below 120 (though mine is below 100).

So this statistical measure g fits very closely with the intuitive picture of intelligence. It's not cultural, but we don't know what it means computationally either.

(1) Amoung frenchies, people with higher 'g' will score better on french tests. But a low g frenchy will beat a high g brit.

[aswanson]

I've never met a math/physics/eng faculty member with an IQ below 120 (though mine is below 100).

Yummy, you have to be kidding us. Either that, or you were drunk when you took the test, or we really need to rethink what the tests are measuring. There is no way somebody with subnormal intelligence is involved in exchanges like this:

http://news.ycombinator.com/item?id=216701

[DaniFong]

Apparently someone decided the score a bunch of the Caltech faculty on an IQ test, and a surprising number of them turned up with scores < 100. This doesn't prove that the concept is untenable, but clearly some of our tests are missing things.

[rms]

That's interesting... any source?

[DaniFong]

I'm afraid not -- it was just an anecdote circulating around the astro department at Princeton. Feynman was known to have an IQ of 125, and I met a grad student there who admitted to an IQ of around 80 -- a perfectly bright person, to be sure. Again, I don't have the ability to prove these things to you, but I think we should recognize that, occasionally, our methods for measuring intelligence are very very broken.

[tigerthink]

>Feynman was known to have an IQ of 125

That's still more than 1.5 standard deviations above average. If Feynman's IQ was exactly 125, he would be ahead of "only" 95% of the population.

[yelsgib]

I agree with this point, but I think that what you run into is that there are certain important cross-cultural trends which might cause g to be pertinent. What I mean is that perhaps the set of tasks which we "regularly perform" are governed by g - these tasks usually involving "problem solving" of the you-give-me-a-well-defined-problem-and-I'll-solve-it variety. It's a kind of reciprocal process - society will produce lots of jobs/roles which fit mass statistics it can determine, and it will also screen for that mass statistic.

What interests me, and what I'm worried that focus on this sort of gross statistic deprives us from thinking about, is the non-regular. I'm interested in the philosophical, the revolutionary, the artistic. I'm interested in what made Einstein Einstein. I don't think it's "g."

Slightly more abstractly/hypothetically: The way I kind of see intelligence is that we can function within contexts (e.g. functioning within the "Math context" would be doing math problems) and "g" corresponds to our ability to manipulate symbols generally, so that we are generally good at working within contexts. This is a really abstract way of saying g corresponds to our ability to solve problems. However, I don't think g says anything about long-term problem solving (involving abstraction, self-reflection), novel content creation (involving cross-context thinking), etc.

Does this make sense?

[whacked_new]

You are aware that Einstein's brain had some very unusual features, right? I don't pay much attention to this, but I remember it is unusual enough that it is safe to say most people's brains don't have those features.

Relativity was not a well-defined problem when Einstein came up with it. You can analyze this intelligence in however many ways, but it is highly possible that "quantum leaps" in scientific advancement come from unusual thinkers -- unusual in a way that you can't just train very hard in and attain.

[lolb]

"When I use a word, it means exactly what I mean it to mean, no more and no less." -- Humpty Dumpty

"Intelligence is what intelligence tests measure." -- Edward Boring

I'm actually surprised by how few comments have been made regarding the actual nature of intelligence in this thread. yummyfajitas' parent comment is one of those few, and he equates intelligence as we understand it with g, the hypothesized general intelligence posited by psychometricians.

y.f. makes the same two arguments that psychometricians make for the validity of g as a representation of human intelligence,[1] which are supposed to demonstrate g's "internal validity" and "external validity". The former refers to the positive correlations between various (supposed) intelligence tests (described in the jargon as the "positive manifold"), and the latter to the positive correlations between g and supposed success in life. Broadly, the internal validity argument proves the "general" in "general intelligence", whereas the external validity argument proves the "intelligence".

The first wrinkle in these arguments is the claim that all human intelligence tests positively correlate with each other to give a universal intelligence factor. This is false; there exist demographics where g is _not_ the predominant factor explaining inter-demographic test score differences.[2]

Here is another problem. Psychometricians tend to decide whether some test metric measures intelligence by how well it correlates with existing "intelligence tests" that're g-associated. y.f. does this in his comment: "It is relatively independent of cultural knowledge: i.e., french is nearly independent of g (1), while physics, plumbing and loading irregularly sized boxes into a truck are highly correlated to g". (I.e., French tests can't be real tests of intelligence, since they don't correlate with g, but physics, plumbing, and bin packing must be, because they do.) But this renders the internal validity argument circular: its two prongs now become: (1) g must exist because intelligence tests correlate positively, and (2) and intelligence tests are those tests that correlate positively with g. The circle closes.

The external validity argument supposedly rescues the g concept from this trap, by demonstrating that g correlates vaguely with real-life behaviours. Unfortunately, the correlations are nowhere near perfect, that g correlates with success doesn't suffice to show causality, and pointing out the correlations does not eliminate the underlying circularity in g's definition. All in all, the evidence is inadequate to reject the null hypothesis that there is no causal link between g and life success.

[1] Well, I say "human intelligence", but if I recall correctly, at least one psychometrician even tried to argue that g might well be a cross-species phenomenon!

[2] Dolan, C.V. Roorda, W., and Wicherts, J. M. (2004). "Two failures of Spearman's hypothesis: The GATB in Holland and the JAT in South Africa", Intelligence, vol. 32, p. 155-173.

[yummyfajitas]

You seem to misunderstand the definition of g, as g is not defined in terms of intelligence tests.

Take as your statistic the result of many tests: physics, box loading, french, basketball, etc, and then do a PCA or similar test. One of the principal components will be g, provided you have enough data. This is what defines g.

Intelligence tests are simply tests designed to be more highly correlated with g. If we discarded intelligence tests, we could recreate them (or equivalent tests) based on statistical analysis of the other test data. They are certainly not arbitrary measures.

The external argument doesn't rescue g; g is quite safe. The external argument merely claims that 'g' and 'intelligence' are the same thing, or very close. g exists regardless of what you want to call it.

As for the paper you cite, I skimmed it. Unless I misunderstood it horribly, it merely claims that particular IQ tests don't effectively measure g across different groups. That doesn't mean g doesn't exist as a hidden variable, merely that a particular test is poorly correlated with it for some population.

[emmett]

I've posted this at least three times on Hacker News, but I'm just shocked people haven't read it:

http://cscs.umich.edu/~crshalizi/weblog/520.html

This guy is a statistics professor, and has a lot to say about exactly what "g" is. He even runs experiments! I know the article is a bit long but I promise it's worth reading.

[yummyfajitas]

Well, I admit, I just skimmed it, I'll try to read the whole thing later. But near as I can tell, he isn't addressing the validity of g at all, merely claiming it is unproven that genetic differences between population groups is due to genetics or other factors (a claim I don't disagree with).

[emmett]

You're right; sorry, wrong essay. The other one, by the same guy, addressing that point:

http://cscs.umich.edu/~crshalizi/weblog/523.html

[lolb]

Are there any studies showing that basketball performance, French fluency, bin packing, and physics knowledge, all taken together, will produce a meaningfully large general factor? Intuitively it seems unlikely, and it doesn't jibe with what I understand g's definition to be based on the comments of people like Eysenck and Jensen, who justified g on the basis of intelligence test correlations, and only later tried to tie it to biological functions (evoked brain signal potential in Eysenck's case, reaction time in Jensen's) and metrics of social, sports, and job performance.

Regardless, I agree that g exists in at least a limited sense, but I maintain that there is insufficient evidence to demonstrate that it emerges from innate, immutable physical properties of the brain, and you still can't attribute life success to it.