COMPUTING AT CHAOS MANOR
Tuesday, July 22, 2003
THE COCHRAN FILES
You have met Greg Cochran before. He recently posted this in a discussion I'm part of and I got his permission to put it here. I am collecting his material and copying it to a separate page, but I have not got far along with doing that yet.
Background: I am in a discussion group that includes Greg Cochran. You may have seen his work on evolution and disease: if the theory of evolution is true, then many "hereditary" diseases including schizophrenia must in fact be caused by infectious agents: the evolutionary burden of those disabilities is so large they would be bred out of the race in no time.
Studies have shown that one of the most important variables in education is the intelligence of the teacher: the smarter the teachers, the better the results. This is perhaps unsurprising, but it has serious consequences.
I think there is no serious (other than politically correct) disagreement among informed experts that IQ has a strong hereditary component. Murray and Herrnstein put it at 60%; that's actually considered low.
In our discussion this came up. Published with permission:
Once upon a time, the United States had to deal with enemies who talked about supermen - either Nordics, or the New Soviet Man. But that's all they did about it - talk. That may change in the 21st century.
Today the bell curve limits what any country can do . High IQ increases performance in almost every job, but there are only so many people with high IQs. Two or three percent of European populations have an IQ above 132, maybe 1/10th of one percent have IQs above 148. We have reason to believe that children would get a better education from teachers with higher scores, but there just aren't enough people with such scores, and most of them are already doing something else more lucrative. And we needs lots of teachers - if we found that teachers with IQs three standard deviations above average (> 148) were considerably more effective than typical teachers today, we could not ever take advantage of that fact for most students, because there are only a quarter of a million people in the US who score that high, while there are at least ten times that many teachers in the US.
We could use smarter engineers, smarter doctors, smarter plumbers, smarter soldiers - but right now it can't be done. The best we could do would be to reallocate some smart people currently playing socially destructive roles. Like lawyers.
In fact, with the current demographic trends, we can expect that the fraction of kids with extremely high IQs will be considerably smaller in the next generation. That's called diversity. We're cloning animals, and we could probably clone humans - tomorrow, not generations from now. There is no physical reason why someone, somewhere, could not do this on a large scale. And if they did, you could have a country where 20%, or for that matter 100% of kids had IQs over 150.
If any medium sized country (like France) ever decided to adopt such a policy, it would in one generation have more very smart people than the entire rest of the world. I think this would greatly increase that country's competitive abilities in many fields. it would also make all our current concerns about ethnic differences in cognitive abilities seem pretty unimportant. Not only would this hypothetical country be full of talented people, the powers that be would _know_ who was talented, and could skip much of the expensive and time-consuming sorting processes we use to identify them - when we even care. A reasonable biology professor should be willing to hire a 20-year-old clone of Bill Hamilton without having read his thesis..
I'm sure that lots of people would dismiss this possibility, on the grounds that genetic influences are not all that important in determining IQ, or that IQ itself is not that important. They're wrong, of course, but in fact those choosing who should be Xeroxed are by no means limited to sorting by IQ. Health, emotional stability, even having ideas could all be part of the selection criteria. None of this requires the sequencing of the human genome, or deep understanding of that sequence - Xerox machines don't have to know how to read.
The real question is whether this is any chance that such a radical step will ever be taken - or,, I should say, taken anywhere in the near future. I'd say that it's fairly unlikely, but possible. Certainly there have been individuals within living memory that had enough personal power for enough years to make such a program feasible. I'm thinking of Stalin, but there may be other examples. Anyone with iron control of a country for a generation or so could probably pull it off. It would be hard to do this in any country in Europe, since none of them are really dictatorial right now and it's hard enough to get women to have any kids, let alone someone else's...
But you never know.... It is possible that someone would try this in the service of some familiar goal, such as preservation or aggrandizement of some culture or religion, national resurgence or revanche: Ideally, this would result in a country whose superabundance of talent was organized in incredibly inefficient ways - I say ideally, because this would give us sluggish democratic societies more of a chance and make for interesting history. If these guys were immensely smarter _and_ well organized, they'd just win and that would be the end of it.
We might try to nip this in the bud - although it's hard to say exactly what reason we would give, since our official ideology says that this could never work. We would be far less likely to interfere with anyone trying this who happened to have nuclear weapons. Simple obscurity might also allow success - if the Ethiopians were doing this right now, would we ever hear about it? Foreign-affairs coverage has about vanished from most of the big newspapers and never really existed on network news...
This is not the last, or the most important of the biological innovations that promise to make history more interesting than Frances Fukuyama's worst nightmare. It is just the next one.
We do live in interesting times. And John McCarthy points out that this sort of thing need not be done by a nation. An organization or religion might even try it.
The Boys From Brazil only more so and for real? It would make for a good science fiction novel, but in fact it may not be fiction at all. And as Greg points out, our official ideology says it won't work, so there's no way we could justify interfering, even if we wanted to. China, anyone?
Reading Ullica Segerstrale's book, Defenders of the Truth, I begin to understand some of the issues that get concealed by the politics. I did ESS Theory (the British version of sociobiology) back in the 1980s, and got caught up in the battle. It became strange later, because I found myself in disagreement with the sociobiologists (even though I was one) and in agreement with their critics.
The critics have a real _scientific_ point--if you don't understand the developmental, functional, evolutionary, and/or control mechanisms underlying behavior, you don't understand the causal chain that makes it happen (see Tinbergen). It doesn't matter that you have all sorts of interesting statistical results; you don't really understand it. (By the way, they criticize all 'statistical' fields of science this way.) Your theory has no visible means of support. For example, my dissertation was on sensorimotor behavior in echolocating bats. I tried to connect the model to foundations, but we honestly do not know the following: 1. the detailed neural connections between the ear and the cerebral cortex, 2. how the inner ear functions, 3. how the development of the system is controlled genetically, 4. how sound is encoded, 5. how the system is wired during development, 6. how it evolved, 7. how the system is controlled... Almost everywhere I looked closely, I discovered the current theory was wrong.
Or as I commented elsewhere: "I think that it can be argued that complex systems (in the sense of Rosen, 1985) exist and show emergent higher-level behavior that cannot be predicted by simply modeling lower level processes. However, those lower level processes remain real, and our understanding of the higher-level behavior has to be consistent with them. If we lack valid models of those lower level processes, we will have difficulty formulating higher level models correctly. Or as I kept relearning during my dissertation research:
"It's not what you don't know that will hurt you. It's what you think you know that just ain't so." -- Satchel Paige
A friend comments about models of non-linear processes: "when asked by the King of Belgium, who financed most of his research, whether non-linear dynamics was a marxist theory, Prigogine responded: but it is also a proof of the existence of God ..."
Prigogine's point was (I think) that these models that we love to speculate with (which usually lack firm foundations in the data) are essentially quantized Hegelian dialectic. At the same time, it turns out that we cannot model non-linear systems except in approximation. We cannot create a complete model of a lot of things that we're interested in. The only complete models are the things themselves, that God has made. If Alonzo Church is correct, we _can't_ know those things.
Now to intelligence:
We have statistical results. They suggest that general intelligence (g) is about 50% inherited. We lack an understanding of what general intelligence measures. We don't know how it develops; we don't know how it evolved; we lack a deep understanding of its functions; and we don't know how it is controlled. That suggests we're in danger of making the mistake that Satchel Paige was pointing to. For example, there's some evidence that general intelligence may be controlled by genes inherited from the mother only. <http://www.newscientist.com/ns/970503/features.html>. There's also evidence of maternal effects; for example, it is suspected that fetal alcohol syndrome can be triggered by a single drink at the wrong time. <http://www.aap.org/policy/04358.html>. Third, we don't know how environment interacts with genetics in general, let alone for intelligence, but it's probably highly non-linear and intractable if you try to analyze it. The secular increase in IQ over the last 100 years strongly suggests there's something important going on that we have no understanding of. Not even a clue.
What that means is environment can masquerade as genetics and genetics as environment. Until we tease out the causal chains underlying general intelligence, we should avoid calling our speculations 'scientific', especially when there are politicians, economists, and lawyers in the wings ready to convert those speculations into dumb policy. If you call them 'science fiction', you're doing mankind a favor, because some young researcher is likely to be interested enough to check them out, but please don't call them scientific.
-- --- Harry Erwin, PhD, Computational Neuroscientist (modeling bat behavior), Senior SW Analyst and Security Engineer, and Adjunct Professor of Computer Science, GMU. CV available at: <http://mason.gmu.edu/~herwin/CV.htm>
Mechanistic models are a great deal more satisfying than stochastic equation systems, and often do a better job of prediction: but note that we have not the foggiest understanding of the mechanisms in quantum mechanics, yet the equations predict nicely. No one understands the two-slit phenomenon, and Dick Feynman was at pains to be certain you realize that not only do we not understand the phenomenon, we don't even have the beginnings of a clue, and we may never understand it. Certainly Feynman went to his grave not knowing how the stuff that he got his Nobel for actually works.
There's much the same underlying ignorance in the mechanisms of IQ and success; which is why "g" is often used in scientific conversation, and why serious psychometricians are careful with the word "intelligence". Yes, IQ may be inherited largely through the mother; yes there may be environmental prenatal factors. On that latter point there almost certainly are, but the prenatal factors generally detract from an inherited upper limit. It's that upper limit that we're concerned about.
If what you are saying is that we are only beginning to understand ourselves in any meaningful scientific way, then, yes, of course, and perhaps that message should be stressed often; but I hope you are not saying that because the results are not politically correct we should give up the search?
Having spent a good part of my life doing operations research, I can say with some confidence that stochastic models beat holy hell out of guesswork: that knowing what factors you rely on to make decisions is in general a lot better than merely straining like a gearbox and deciding out of the blue...
Then see http://www.kleinbottle.com/
So, you think clones of a highly intelligent person aren't necessarily going to be highly intelligent? And that we can't predict the result unless we know everything about how genes shape development, the evolutionary origins and purpose of human intelligence, and the secret of Oak Island? Well, that happens not to be the case.
Cloning, is, as far as I know, equivalent to twinning. Identical twins, even those raised apart, have very similar IQs. The correlation between the IQs of identicall twins raised apart is about 0.76. This is higher than we wouild expect from estimates of narrow-sense heritability of IQ; narrow-sense heritability estimates only measure additive effects of alleles. Non additive effects depend on interactions between several alleles, and aren't much transmitted into the next generation, since the sets of alleles get broken up. Identical twins, on the other hand, have the same genomes, and both additive and non-additive effects are duplicated. Maternal genes and paternal genes are of equal importance in determining IQ. The parent-child correlations are almost exactly the same for mothers and fathers. The only way in which clones might be be less similar to their original than identical twins are to each other would be if variations in prenatal environment explained a significant fraction of the variance in IQ This because identical twins share the same uterus, while clones would not. I don't believe this is the case. Certainly kids with the same mother are not significantly more similar than kids with the same father - moreover, all the evidence we have supports the idea that prebirth development is robust, and that a kid's development is insensitive to anything less than a major insult. Whch is why the Dutch draftees who were in utero during the 1944 famine are just as big as other year classes and show no IQ depression.
Next, even if this uncertainly existed, which it does not, it could and would be easily resolved. If a few eccentric rich people clone themselves, we'll know how similar they are to their clones, and we won't need any deeper understanding than that. Where does this idea come from, that we have to understand everything in order to predict or do anything? So we couldn't have domestIcated the horse without first sequencing its genome?
It'll work. If it didn't work, we couldn't have bred border collies to be smart, or developed Guernsey cows that give more milk than any cow every did in Classical times. Medieval farmers didn't have to know a damn thing about the physiology of milk production in order to do this.
It reads like we're talking past each other. I agree IQ appears to have high heritability based on the twin studies, but we lack even a fair understanding of the mechanisms, and we certainly haven't ruled out maternal effects. My experience is that most theoretical speculation does not survive the first encounter with actual experimental results.
The results of Duncan et al. published today in Science suggest that general intelligence is associated with the lateral frontal cortex, a small region of the neocortex on each side that is involved in working memory. Only one of these regions seems to be used in processing verbal problems, while both are activated for spatial problems. Their relatively small size suggests solving problems using general intelligence is not very important to our genetic fitness.
Those regions of the neocortex are known to have continuous neurogenesis in adult primates. In other words, they are continuously rewiring themselves (like the olfactory system). Rewiring is believed to be associated with plasticity. Is there evidence for plasticity in our use of general intelligence? If so, why do we treat it as innate rather than learned?
Cheers, -- --- Harry Erwin, PhD, Computational Neuroscientist (modeling bat behavior), Senior SW Analyst and Security Engineer, and Adjunct Professor of Computer Science, GMU. CV available at: <http://mason.gmu.edu/~herwin/CV.htm>
This came up in a discussion elsewhere, and is published by permission. Carol Iannone is well known to readers of Commentary and other literary publications. Greg Cochran is an engineer you've seen here before:
Message text written by "Carol Iannone"
> Let me second Scott's recommendation of the film "Est-Ouest." I saw it with a Romanian friend who grew up under Communism and whose family life was horribly damaged by the regime. She cried throughout the movie and kept nudging me and saying, It was like that, I remember, they were like that. This was in reference to such things as the awful brutality and trickery and perfidy attributed to the Communists. The funny thing is that the portrayal of the Communists might actually have been seen at one time to be unaesthetic--too extreme, too villainous, too one-sided, lacking "negative capability," but after Solzhenitsyn and all we've learned, it was possible for me too to say, yes, they were like that!
Does anyone understand the unbelievably unbalanced ratio of films, books, documentaries, etc. about the Nazis vs. the tiny number about the Communists? It's a mystery to me why there isn't a flood of films and documentaries like this one. Carol <
Cochran: What's not to understand? The talking classes, in general, think that sympathy and support for the Soviet Union, Maoist China, Castro's Cuba, etc was at the worst an easily excusable error. Any warm-blooded, person with the proper political instincts _should_ have sympathized with them, and anyone who too clearly saw the disadvantages of those totalitarian regimes was surely a bad 'un -- better to be wrong for the right reasons. To some extent, this is also because any other position would involve quite a bit of self-criticism, or at least criticism of those near and dear.
I don't think it has changed much, either. There's a lot of sympathy for Castro in the same circles today. I certainly heard a lot of people talking about how Elian Gonzalez, once back in Cuba, would be 'free' of all those evil forces like McDonald's and Pokemon and would be if anything better off than in Miami. Assuming that they mean what they say, those making such comments must think that left-wing despotism is a positive good, to be preferred to classical political and economic freedom, and even worth a dramatic cut in the standard of living. In theory - I don't think many of these people are sailing their yachts to Cuba.
I think it illustrates a really important point, that in the long struggle between the Soviet Union and the United States, it wasn't just the Soviet Union that was falling apart. It was a race to see what would fall apart first - the Soviet economy (very roughly speaking) or the American will to continue the struggle. It had certainly got to the point, by the 80's, that national defense itself was controversial. Les Aspin was so right-wing that he was almost kicked off the Armed Services committee, while Ron Dellums was just fine. Most university professors thought that any kind of defense work was just dirty. With the coming to full political power of those who had entered politics as part of the anti-war movement ( not my name for it), with the Clinton Administration, even close connections to such regimes were just fine. They almost appointed Johnetta Cole as secretary of Education, even though she was high-up in the American Communist Party and the Venceremos Brigade, until someone at the Forward objected pointing out that an actual member of the Communist Party might be insufficiently supportive of Israel. They sent Dellums as the unofficial envoy to North Korea - or it was he the unofficial envoy _from_ North Korea? I forget. Maybe that's why we keeps their tanks fueled today.
Certainly the establishment attitudes towards the Soviet Union had become comically stupid by the 1980s. Take the attitudes of the middle-of-the-road economists like Paul Samuelson ( and those days, he was close to the middle) - that is was a vulgar mistake to think that people in Eastern Europe were miserable. Or their understanding of East Bloc economics - they knew, the CIA knew, everybody at the New York Times or Time magazine knew that Russians weren't all that much poorer than Americans. .. The CIA's official estimate in 1985 had East German's per-capita income 20% higher than West Germany's. People who thought otherwise, or even doubted, were loons. Don't you remember? Murray Feshbach was some kind of loon. Remember 'convergence'?
Remember all the people who signed on to old-fashioned Front efforts like the 'nuclear freeze'? Like Madeleine Albright?
I remember when even the establishment types started noticing the ground shifting under their feet - very late, of course. In the summer of 1991, Leslie Gelb noted that people were starting to use the "R' word about the Soviet Union. Revolution. About time, I thought - I'd been placing bets on the order in which the SSR's would secede more than a year earlier But Gelb, of course, didn't want it to happen. Only the most reactionary, Neanderthal types would want to see the Communist Party overthrown - so he said - this, coupled with the skeletal evidence of a brain larger than ours, makes me against wonder if Nature missed a bet. And Gelb was a New York Times editor, now some honcho at the Council of Foreign Relations. Not just any old fool. He was an _official_ fool.
There are more ways to fall apart than economic decay. If we ever have another long national struggle, it's important to remember that the political sanity of our talking classes is about as fragile as a Prince Rupert's drop. Any adversary with resources comparable to the Soviet Union's and a less fragrantly shoot-your-own-dick-off domestic policy is likely to give far more trouble than Russia ever did. All it had to do is make symbolic gestures in the direction of the fashionable ideology of the day and useful idiots will spring up like dragon's teeth.
Cochran speaks forcefully and directly, and I fear I have to agree.
I recall at a HACKER'S CONFERENCE in about 1984 being nearly hooted out of the conference because I dared defend ballistic missile defenses: this from high tech people who didn't really believe it wouldn't work, but whose intellectual betters -- at least they acted as if they were their betters -- had told them it was all wrong, and Ronnie Raygun was a jerk, so they mindlessly acted on those premises. Some even apologized later: they had been intimidated by the general atmosphere; I wondered at the time how many knew better but wouldn't say anything because they didn't want to look like idiots to their peers. The odd part is that having hooted at me in the "Star Wars" discussion, there seemed to be no residual from it, and everything else I said on computing and its future was taken seriously. But I never forgot that incident of herd mentality among some otherwise VERY bright people.
And Possony and I were regularly denounced by the State Department and CIA when we said that the USSR had to be spending at least 30% of GNP on weapons. CIA even sent economic "experts" to Aerospace Corp. to denounce us. And RAND Corp had an economic model of the USSR that was pure fantasy but was accepted by nearly everyone. My own daughter, in Army Intelligence, insisted that the East German standard of living was as high as West German -- and she had BEEN there. But all her professors at Cornell and other academic departments of German Studies had convinced her that what she saw wasn't real and relevant. And understand, she's an intelligent lady.
Richard Pipes described the phenomenon in SURVIVAL IS NOT ENOUGH: academia never told the truth about the USSR because to do so was to deny yourself access to the place -- the Soviets wouldn't give a visa to a scholar who told the truth about the place -- and thus to cut yourself off from being 'an expert' since how could you be expert if you never went there? And that was an old story: the Ukranian intentionally induced famine went on under the eyes of western journalists, including people like George Bernard Shaw, and the only one who had much to say at the time was Peregrine Worthstone.
James Burnham in SUICIDE OF THE WEST said "Liberalism is a philosophy of consolation for western civilization as it commits suicide. It allows one to feel good while watching everything one believes in collapse and be destroyed." One sometimes wonders if he didn't know more truth than we like to admit.
Subject: Applied Darwinism
I said, just a bit ago, that hardly anyone 'believes' in evolution, in the sense that electrical engineers believe in Maxwell's equations. Hardly anyone uses our existing theory, the neodarwinian synthesis, to predict things or solve problems. Most biologists don't believe in it. Sure, they'll say that they do, but since in the first place they don't _know_ any evolutionary biology, how could they believe in it? Many know nothing, others know things that aren't so - only a small fraction of biologists take it seriously, probably only a few percent . Let alone anyone else.
It doesn't help that the most prominent popularizer, S. J. Gould, doesn't believe in (or, apparently, know much about) it himself.
Consider medicine. To a good approximation, nobody in medicine knows anything about evolution, thus they feel perfectly free to look for one, or a few alleles that cause common diseases with major consequences. Funny that they never find them. 'Frequently occurring genotypes , or polymorphisms (frequency of 1 percent or more) are unlikely to have a high penetrance for diseases that reduce reproductive fitness: such genotypes would be selected against' says a smarter-than-the-average MD in a recent NEJM article questioning the current genetics-uber-alles enthusiasm. But he is definitely in the minority.
Dean Hamer, like most geneticists, doesn't believe in evolution When he published a report linking an allele on the X-chromosome to homosexuality, he didn't say a word about the problem of how such an allele is supposed to hang around at high frequency. He doesn't think that there _is_ a problem. He's not alone - you could say the same about the people looking for highly penetrant schizophrenia genes. They don't think that biological theories _need_ to make evolutionary sense. My favorite explanation involves a 'young Earth' theory - they must think that natural selection hasn't flushed genes for common conditions that greatly reduce fitness because the human race popped into existence quite recently - probably during the lifetimes of some people alive today. This is also known as the 'born yesterday' theory.
Steve Jones, or Luigi Cavalli-Sforza, must not really believe in evolution, else how could they say that races do not exist? By which, in a practical sense, they are claiming that there can not be socially significant, genetically based differences between people with widely varying geographic origins. But that's untrue; nothing in what we know about the genetics of human populations forbids this, and lots of evidence, like Freedman's work, strongly indicates that such differences exist. So they must not really understand it, right? And what does it mean to say you 'believe' in theories that you fundamentally misunderstand?
Paleontologists don't know much about evolution, and what they know they don't believe (1) .. Most of them hate, with an unyielding passion, Alvarez's theory about the impact extinction of the dinosaurs. Most of them still don't believe in it, despite all the findings of iridium, shocked quartz, sharp transitions in the fossil pollen and plankton record, the impact crater (!) , etc. But if you think about it, considering natural selection, every mass extinction is likely to be caused by some very abrupt event, because of the ability of most species to adapt rapidly to changing conditions. Either the change must have been very large, too great for most species to adapt to, or it involved a smaller physical change that occurred too _rapidly_ for sufficient adaptation. For example, if the Earth lost its atmosphere over a couple of million years - we'll call this a big change - I think we wouldn't see any birds around.. Smaller but significant changes spread over a million years, like the Pleistocene glaciation don't cause mass extinctions. But fast changes can cause mass extinctions - 'fast' meaning in less time than a few generations of your target species. Possibly in one day. People like Gould argue against the possibility of fast evolutionary change under strong selection - this despite the fact that we have _seen_ it, for God's sake, consider resistance to insecticides - because they worry that it might validate racism - or, perhaps, cause they's just stupid. So, whatever caused the Permo-Triassic, the mother of all mass extinctions, the one that wiped out whole orders of _insects_, probably happened in less than a few tens of insect generations. Say less than 40 years. The paleontologists were saying a few million, now maybe as little as few hundred thousand, or even less than 8,000 - but it was less than 100, for sure.
(1) I think they don't 'believe in' Newtonian physics either; anyone who knows it realizes that gravitational perturbations are occasionally going to drive asteroids into the inner Solar system, and can calculate or simulate the collision cross-sections.
Toxoplasma gondi is a small intracellular protozoan parasite that can infect any warm-blooded animal. So says the Merck Manual, and imagine being the guy who proved _that_. It invades and multiplies asexually within the cytoplasm of nucleated host cells. With development of host immunity, multiplication slows and tissue cysts form. Sexual multiplication occurs in the intestinal cells of cats ( and apparently only cats) ; long-lasting oocytes are formed and shed in the stool.
Toxoplasmosis is dangerous in individuals with weak or incompletely developed immune systems. It can be devastating before birth, and is a big problem in people with AIDS. It likes to get into the brain - in animals other than cats.
This is a standard pattern - a parasite with a complex life cycle, that infects an intermediate host and enters the definite host ( cats ) when the definitive host eats the intermediate host. When this happens , manipulation of the intermediate host can help the parasite - manipulation that increases the chance of the intermediate host being eaten. I mentioned toxoplasmosis in a talk earlier this year and suggested that it might manipulate intermediate hosts, and it seems that it does. Recent work shows that uninfected rats fear and loathe the smell of cat urine. Infected rats are at best neutral - some actually seek out cat urine. Pretty obviously, toxoplasmosis is pushing buttons in the rat brain that increase its chances of ending up in a cat stomach. It doesn't appear to do much else - the rats act reasonably normally otherwise, don't look sick, etc.
I think that this supports two points I've been pushing - one, that persistent infections that don't cause obvious pathology may still be doing something, two, that we have to consider the possibility that infectious agents might alter behavior - including human behavior - as a part of their evolutionary strategy.
The third point is that having toxoplasma cysts in your brain may not be a good thing. Only a few billion people do. The infection rate varies a lot from place. It's about 20% in the U.S., about 87% in France ( which might explain a lot).
Thinking about this a bit more, it just might be that the state of being a cat-lover has a simple cause and cure.
Erwin on IQ and Prions:
The consensus seems to be that IQ is not additive in heredity and environment; it seems to be a non-linear random function of both, plus luck and upbringing.
We don't generally know the connection between genes producing specific proteins and intelligence. We do know the brain is too complicated in structure to be encoded fully in the genome.
We can't yet separate the roles of heredity and environment. Last week I ran into a suggestive report that prions play a major role in controlling the structure of proteins. It turns out that the genome of fungi is insufficient to define a working fungus cell--there are lots of prions that are required. The mammal genome seems to be simplified relative to most vertebrates, suggesting two things: the constancy of the maternal environment is important, and that prions are likely to be important. Guess where we get our prion conformations-- not from the genome. Mostly we get them from our mother. So we're back to maternal effects--environmental factors that masquerade as genomic.
Finally, success needs specialized intelligences. Al Gore is not that intelligent (in terms of an IQ test), and George W. Bush seems to be dull normal, but both are very successful. But there's a cost to be paid. High analytical intelligence and high creativity both seem to be correlated with mental illness. I suspect something similar will eventually be found for high social intelligence and high general intelligence.
By the way, the prion result is disturbing...
-- --- Harry Erwin, PhD, Computational Neuroscientist (modeling bat behavior), Senior SW Analyst and Security Engineer, and Adjunct Professor of Computer Science, GMU. CV and papers available at: <http://mason.gmu.edu/~herwin/CV.htm>
[Following comments by Jerry Pournelle]: Oh now, really. First, Bush and Gore are 120 or above, which is "officer material"; they are unlikely to be 140 or above which is considered 'genius'. Actually discrimination above 130 is difficult in part because we don't have enough cases to examine; it's up there that accidentals affect what you know, which affects our ability to measure. Kids with lots of books in the house will score higher than those without, at least until the kids without can get to the library; and the tests are pretty much verbal and have to be.
Gore wrote a wretchedly bad book full of nonsense, but because it blathers on about important subjects everyone is willing to give this confabulation some kind of intellectual credit. Have you READ Earth In The Balance? Do so. It may tell you a lot. Give Gore and Bush IQ tests and I bet they aren't all that far apart, but it sure proves that IQ and judgment are not quite the same thing.
So yes, there's more to life than IQ. But we all knew that. As a child I was supposed to have a really high IQ, but I was in on the computer revolution from the start, and I didn't go make billions and billions. I'm sure I am a disappointment to the testers of my youth. On the other hand, if you're tabulating things, I guess I can be counted a success.
As to mechanisms, Minsky told me 20 years ago to give up trying to learn brain models. They'll be wrong, and you will have wasted your time. He was right; whether he still is right or not I am not sure, but I suspect so. I think it's not yet time for brain models. What I do know is operations research, and something about predictions.
IQ max is set by heredity. Chemicals and drugs can lower it. Protein deprivation in youth will lower it. Fetal alcohol can lower it. That's "environment". But as the Minnesota Twin studies (Bouchard) showed, high IQ makes its own environment, and the twins raised apart tended to have VERY close IQ's: the girl adopted by the iron miners went to the library, the twin adopted by the physician grew up with books in the house, and both ended up at about the same place. Interesting.
People forget: IQ is a construct, something we measure; it's not the underlying cause, it's what we can observe. Different things. We can use IQ to predict, but we shouldn't confuse the map with the territory. But then all old Van Vogt readers know that...
Tell me more about prions, but I warn you, I tend with Minsky to believe it's too early for mechanisms. We're still gathering operations data. [JEP]
These results begin to suggest some ways that maternal effects can play a role in development. I had known earlier that the brain was too complex to be encoded in the genome, but this result begins to clarify why the genome need not be that complex. We probably inherit via other mechanisms, too, and the uniform conditions in the mammalian womb limit the environmental complexity that has to be dealt with during early development. Start messing around with either, and baby is likely to be hurt, because the homeostatic mechanisms that ensure successful development in egg-laying embryos have probably been lost in mammals.
Remember, identical twins have shared a developmental environment during the period they were probably most vulnerable to insult.
-- --- Harry Erwin, PhD, Computational Neuroscientist (modeling bat behavior), Senior SW Analyst and Security Engineer, and Adjunct Professor of Computer Science, GMU. CV and papers available at: <http://mason.gmu.edu/~herwin/CV.htm>
regarding Erwin's comments: He's all wrong, what can I say? The structure of the brain is as far as we know determined by the DNA. It is not a blueprint, but is more of a recipe. and the world is full of really complex things that start out from a simple recipe. Like number theory, a complex subject springing from an extremely simple definition of the integers. As for traits being transmitted by means other than DNA, as far I know it does not happen, at least not in a way that messes up cloning. I say this because clones of mammals are being produced and don't seem to have any obvious problems - hard to argue with that, I should think. Not only that, there are weirder examples that show that utterly nonstandard cells can successfully differentiate and develop into competent individuals. There exists a weird kind of cancer called a teratoma, whose cells seem to think that they are in an embryo. These cancers develop hair, teeth, skin, all manner of messy things. They exist in humans and animals. Some very odd guy wondered if teratoma cells, which seem to want to be an embryo, would actually become one if given a favorable environment. He implanted teratoma cells into an early-stage rat embryo ; the teratoma cells there experienced the proper chemical cues and developed into part of a rat. He ended up with a piebald rat - some of the cells had a regular rat mom and dad, while other parts were descended from a cancer propagated in a tissue culture. The rat was fine. I first ran into this report the very same afternoon that I read _The Boys from Brazil_.