Oh, those gaussians (Was: Twin Studies)

Eliezer S. Yudkowsky (sentience@pobox.com)
Sun, 22 Aug 1999 04:37:39 -0500

phil osborn wrote:
> First, it's nature AND nurture, obviously, else cats and mosquitos would
> both have human intelligence. What Burt did was to heavily tilt the balance
> of evidence to please the British aristocracy. As I recall, he concluded
> that about 80% was heredity, when in fact the data indicated more like
> 50/50.

I beguess that it's almost exactly 50/50, or a Gaussian curve centered very precisely on 50/50 and without too large a standard deviation, either. Speakin' as an external observer, "balance" is the main thing I notice about gaussians (that's as in "gaussian human", somebody who can be anywhere on the curve but *is* on the curve). Your abilities are balanced, not precisely, but within a certain range, so that they don't fall out of synchronization; balance between the problems you face and the means you have to solve them, so that a simple effort of will can make the difference between success and failure; balance between yourselves and the environment, especially the environment created by other gaussians; and, of course, a nearly perfect balance between nature and nurture that reflects the relative evolutionary advantages of perfect optimization versus specialization so that you can be the best at *something*. For the longest time I thought it was terribly ironic that gaussians, who - from my perspective - are whatever they think they are, who could build a self or simply decide on one by fiat, spend all this time trying to "find themselves", while I, whose nature is pretty much hardwired, spend so much time trying to alter it. Now I sort of understand; they're trying to find the mix of native abilities that determines half their selves. It still strikes me as being a little silly, though, since you could easily override any of that by deciding otherwise.

> After three years, however, the six and seven year olds of our first group
> could pick up a newspaper and read it with good comprehension, they could
> find virtually any major named spot on a globe or map (any spot given the
> references), they could diagram any English sentence. They had excellent
> musical relative pitch recognition. They knew quite a bit of Euclidian
> geometry and were starting in algebra. These were not their only skills, by
> any means.

Mmm, I probably could have beaten them at most of that; and without the benefit of that elaborate stimulating environment, just my stimulating neurology. It's like I keep on saying; the human brain contains a lot of slack, and you're going to lose a lot of neurons if they aren't needed in childhood, but once you take up the slack and retain all the neurons you can, you can't go any farther without Specialization or other neurohacking. (Specialization takes up all the slack automatically, of course - if you're specialized to the degree that you're robbing neurons from other abilities, obviously you don't have any neurons to spare.)

> Montessori's goal was to create a revolution by creating powerful
> individuals who knew how to actually use more of their brains - and why
> (such as Milton Friedman, one of the early U.S. Montessori kids, before
> Dewey and Thorndike declared war on Montessori and wiped it out in the U.S.
> for almost four decades). People who would be utterly immune to propaganda
> and fully confident in their own ability to understand and deal with the
> world, who had been brought up to be totally self-motivated (no one tells a
> Montessori child "you VILL do this NOW!", as they do in the German
> Kindergartens. The Montessori kids choose their own activity or not.)

Good for Montessori! I approve(*), even though it's quite obvious that the real revolution will come when you can pack a Montessori education inside a palmtop and airdrop them on primitive villages a la Marc Steigler. But to say it one more time, and to remind those latecomers to the list that I *am* a neurohack (or in less technical terms, a mutant supergenius), THERE ARE LIMITS TO WHAT YOU CAN DO WITH SOFTWARE. I *know* whereof I speak.

(*) = Actually, I don't trust Montessori. I think she probably neglected a few abilities that are going to wind up being sacrificed in an inadvertant software specialization that will wind up being fixed in hardware; I don't trust the kids not to flame out for that and other reasons; and (see _The Nurture Assumption_) the best way to teach kids is with slightly older kids, and I'm not sure the method incorporates that. But if you updated it, it would definitely be better than nothing.

> She also hoped that someone would carry on her work and extend the
> environment into the higher ages. I might point out, in passing, that we
> certainly do know what happens on the other end of the spectrum - early
> sensory deprivation. There have been enought cases of that historically to
> verify that it results in severe intellectual incapacity.

Helen Keller.

> Montessori died in 1952, and, as it was a top-down organization, her death
> ended most innovation in her schools.

Dear me. That's certainly something to keep in mind. You'd think that the whole point would have been for one of her students to replace her, if they're such hotshots.

> The educational equipment for one Montessori pre-school classroom costs over
> $20,000, compared to a few hundred in a typical day care.

I bet you can get 90% of the benefit with 10% of the cost, especially with those newfangled computers.

> There are other major factors I could discuss at some length, such as the
> fact that women, the major factor in childcare in this country, take huge
> pride and virtually identify themselves with strongly genetic factors, like
> their personal appearance, for example.

And the rest of humanity identify themselves with totally genetic factors, like their own emotions. What of it?

> Women - not all of them, obviously, but way too many - are anti-rational and
> consistently resist science and technology.

A certain percentage of women, a certain percentage of men. Which one is greater is quite irrelevant to me. Considering my own immunity to statistics, all I really need to know is that there's a distribution; that defines the cognitive shape of the meme.

> And Amerika
> distrusts people who are too smart, as in "Mad Scientists."

A certain percentage does, just like everywhere else. In such cases, it's not so much who has the largest pie-slice as the actions taken by each pie-slice that counts. I know perfectly well that a certain percentage of the population is going to hate the Singularity, and a certain percentage is going to love it; it doesn't really matter whether the numbers are 10% and 20% or vice versa; what matters is what our people do. That's the fundamental problem with "toning down a message" to "appeal to more people". You can't eliminate the opposition, so why cripple your own memes?

> So it's a hard
> sell culturally and then you have active, dedicated thoroughly evil
> opposition who happen to control the state-financed compulsory education and
> its propaganda apparatus - the schools themselves.

Oh, please. "Thoroughly evil"? Psychologically implausible.

> Bottom line. If any other area of human endeavor were pursued as
> incompetently as childcare, it would be notable for that fact alone.

Philosophy springs to mind.

           sentience@pobox.com          Eliezer S. Yudkowsky
Running on BeOS           Typing in Dvorak          Programming with Patterns
Voting for Libertarians   Heading for Singularity   There Is A Better Way