June 01, 2003
Steven Pinker Says Human Genetic Engineering Unlikely

Steven Pinker, author of The Blank Slate: The Modern Denial of Human Nature (see some reviews here), has written an essay arguing that genetic enhancement of humans is too unlikely to worry about.

Why are technological predictions usually wrong? Many futurologists write as if current progress can be extrapolated indefinitely-committing the fallacy of climbing trees to get to the moon. They routinely underestimate how much has to go right for a development to change our lives. It takes more than a single ''eureka!'' It takes a large number of more boring discoveries, together with the psychological and sociological imponderables that make people adopt some invention en masse. Who could have predicted that the videophones of the 1960s would sink like a stone while the text messaging of the 1990s would become a teenage craze?

I find all of his arguments on this subject to be questionable and am surprised to read them coming from such an able mind.

Pinker is right that linear trends can not be extrapolated indefintely. But the problem with his argument is that the rate of biotechnological advance is more likely to accelerate than decelerate. See the FuturePundit Biotech Advance Rates archive for some of the reasons why.

Pinker argues that since many traits such as personality and IQ are due to complex interactions between multiple genes it will be very difficult to puzzle out how changes to genes will affect traits of the mind. Well, there are several billion people in the world and therefore many existing combinations of genetic variations already existing in the human population. Once it becomes cheap and fast to sequence the DNA of each person then it will become possible to sequence the DNA of millions of people and do a massive comparison of mental traits (e.g. IQ, personality tests, and various facts from life histories such as arrest records, mental illnesses, and involvement in various hobbies and forms of recreation) and genetic variations to tease out complex relationships between genetic variations.

Scientists are gaining the ability with computers to collect and analyse absolutely massive amounts of data. The ability of scientists to study interactions in complex systems will only become greater with time. The complexity of human biology will not prevent scientists from figuring it out in enormous detail.

Then Pinker comes to the fact that while identical twins are a lot more like each other mentally than they are like fraternal twins they can still differ in significant ways. Environment still matters. To Pinker's way of thinking this is somehow supposed to be a disincentive to human genetic engineering. But think about it from a prospective parent's position. Given two sets of genetic combinations if one set dramatically increases the odds of getting a child with a certain personality type as compared to the other combination even if the outcome is not guaranteed parents may elect to go with the combination that tilts the odds. Right now some people are choosing to try the Microsort service for sex selection of offspring even though the outcome is less than 100% guaranteed to be successful. Also, by opting for a route that includes the use of IVF they may be increasing the risk of birth defects in their children. Yet they are doing so along with many others who are using IVF to start pregnancies.

Pinker tries to argue that most people will be turned off by the idea of reproducing via unnatural means. This is a curious argument to make in an age when some women schedule the date to hormonally induce labor or to have a Caesarean section because a chosen date is more convenient than waiting for natural processes to bring the pregnancy to an end. Pinker points to in vitro fertilization as a procedure that people resort to only when they are unable to reproduce in the more conventional way. But there are already couples using it in order to have babies to use as donors of tissue to treat another child for a genetic disorder. The utilitarian attitude that couples bring to such a decision ought to give pause to anyone who thinks the wider public will not embrace biotechnological manipulations of early stage fetuses.

Also, and more importantly, there are already a growing number of people reproducing using artificial means in place of sexual relations because they want to select the sex of their offspring (again, see the Microsort service for details). They are using artificial means to control If it becomes possible to control more characteristics besides sex is there any reason to suppose that no portion of the populace will be attracted to "unnatural" interventions when it becomes possible to have them? For instance, many genetic variations that contribute to intelligence will be found and most will be found to not have severe side effects. Why won't many prospective parents be attracted to the idea of providing those genetic variations for their offspring?

We live in an era when millions of women terminate pregnancies every year using the rather unnatural procedure of abortion. We live in an era when millions of people undergo an increasing assortment of plastic surgery procedures, take recreational drugs, use drugs to alter their moods to treat mental illness, and give drugs to their kids to make them concentrate better in school. Where is the popular fear of the biologically unnatural? For cosmetic purposes Isolagen now offers (at least in assorted Western countries where it has approval - it is current seeking FDA approval) a service to take a person's cells, grow them outside of their body, and then to inject those cells back into locations under facial skin in order to provide a more enduring facial enhancement than is provided by collagen shots. An affluent public does not shrink from the latest appearance-enhancing biotechnology. Why will they shrink from biotechnology used on humans for other purposes?

Pinker also argues that since people are repulsed by genetically engineered soybeans they will not opt to have genetically engineered children. Well, 74% of the US soybean crop is genetically engineered to be resistant to the Roundup pesticide. If revulsion was such a big problem how could all this soybean be getting sold and consumed? Again, where is the big revulsion to biotechnology that provides the basis for his expectation that people will not opt to genetically enhance their offspring?

It is my estimation we are within at the very most 20 years of the time when a large assortment of gene therapies will be available to genetically enhance fetuses. Once it becomes possible there will be a stampede for genetic enhancement. Gene therapy will have a very positive image with the public when a large assortment of gene therapies to cure illnesses and to enhance performance of adults become available. The public, accustomed to having gene therapy done on themselves for benefical purposes as adults, will not shrink from considering gene therapy options when planning to have children.

Share |      Randall Parker, 2003 June 01 09:14 PM  Biotech Society


Comments
Dienekes Pontikos said at June 2, 2003 12:13 AM:

>> Once it becomes cheap and fast to sequence the DNA of each person then it will become possible to sequence the DNA of millions of people and do a massive comparison of mental traits (e.g. IQ, personality tests, and various facts from life histories such as arrest records, mental illnesses, and involvement in various hobbies and forms of recreation) and genetic variations to tease out complex relationships between genetic variations.

Essentially this means that we'll be able to figure out how humans work by looking at a great sample of people (ideally ALL people) and figuring out patterns that certain types of people exhibit. Unfortunately, this is not guaranteed (or even likely) to work, because extant humanity is an infenitestimal fraction of potential humanity. This means that patterns observed in the small sample of extant humanity may not generalize to potential humanity.

It's well known that in order to figure out how a particular system works, you can take two routes: (i) break it apart, figure out how each component works, then figure out how the various components fit together, all the way to the full system, (ii) try the system with different inputs (in this case, genes), measure the output (e.g., "beauty") and build a model from this data that you have suspect that will hold in general.

Case (i) is unlikely to happen any time soon, mainly because it might require a conceptual breakthrough, but more importantly because it's computationally intractable and will remain so for the foreseeable future. Case (ii) which is the approach favored by those enthusiastic about genetic engineering will fail for the reason that I mentioned, namely that your model will be built using only on 6 billion pieces of input-output data, and will have to generalize to a much larger set of inputs.

Randall Parker said at June 2, 2003 1:18 AM:

Dienekes, We will not be able to figure out every effect of every genetic variation by comparing people. But we will be able to figure out a great deal.

However, there are really two other approaches besides comparison: Lab Experiments and Simulations.

Experiments will become steadily more easy and more automated to do. More techniques for intervening will be developed and better sensors for watching what happens as a result will be developed as well.

Also, computers will increase in speed by orders of magnitude while more software will be developed. So more will be able to be figured out using simulations.

Between the comparison of the existing population's DNA sequenes and characteristics, better ways to do experiments, and more advanced simulations we will witness an ever accelerating pace at which the meaning of our genetic code is elucidated.

Dienekes said at June 2, 2003 2:37 AM:

>> Between the comparison of the existing population's DNA sequenes and characteristics, better ways to do experiments, and more advanced simulations we will witness an ever accelerating pace at which the meaning of our genetic code is elucidated.

It is undeniable that knowledge will increase, no one is questioning that. The main issue is if it is possible (in the foreseeable future) to acquire a level of knowledge that will allow us to create predictable improvements to humans with a reasonably high probability of success. Until that time, designer babies as part of the culture will just have to wait on their designrers' imaginations.

Bob said at June 2, 2003 10:06 AM:

Dienekes,

There is a difference between identifying the ultimate combination of genes and identifying improvements from a random combination of genes. For instance, many problems in database optimization are N-P complete, but nevertheless one can create algorithms that provide most of the benefit by approximating an optimal solution for a tiny fraction of the cost of identifying the exact optimal solution.

It is also important to realise that genetic traits are not uniformly distributed. Some families may have genetic traits that are incompatible with a given genetic intervention, but those traits will cluster in families. Once genetic interventions become widespread, it is entirely possible that some currently very rare congenital conditions will increase in frequency, but that will only create an incentive to identify the causes of those conditions. Assuming the technologies Randall assumes and that we can easily foresee, the cost of identifying those causes will fall to within the means of those families--if natural selection has not already wiped out some component of the incompatible gene combination.

Randall Parker said at June 2, 2003 10:35 AM:

Dienekes,

You are missing one basic point: There are many improvements already in the human population that most people do not have. The risks of including some of those improvements are no higher than the risks of mating with person X rather than with person Y. Humans take similar risks every time they mate with someone. Actually, they take greater risks since there are harmful recessives that one might share with the person one chooses to mate with.

Certainly there are combinations of variations that are harmful that are not harmful individually. But many of those combinations will already be found in the human population by sequencing millions of people. Others will be found the hard way by genetic engineering offspring. But still others will be found thru future conventional matings.

Cheap DNA sequencing will enable us to discover many effects of many variations. I predict millions of people will gamble on trying to put genetic variations into their children that other humans have that they themselves do not have. After all, a woman decides to do that every time she chooses a man to make her pregnant.

Patrick said at June 3, 2003 2:02 AM:

To argue that human genetic engineering is too difficult to do, you'll have to explain why it is so much harder than engineering plants or animals. Because that already happens. You can go buy some genetically engineered plants RIGHT NOW.

Do we fully understand the workings of the complete soybean DNA? No. Not even close. Does this stop us from genetically engineering them to have desirable qualities? No. Not even close.

So explain why humans would be different. (I mean technically here, not morally.)

And as for Pinker's videophone analogy, I find this as flawed as his pro-beastiality ethics. If he bothered looking at at TV or magazine he would see that videophones are a vibrant and growing section of the market, although curiously only for mobile phones.

(SOrry, that was an ad hominiem comment condeming Pinker for his moral attitudes rather than his technical ability, but he is... eewww :p )

Dienekes said at June 3, 2003 3:38 AM:

>> To argue that human genetic engineering is too difficult to do, you'll have to explain why it is so much harder than engineering plants or animals.

Plants and animals lack the emergent qualities that humans possess. Additionally, genetically engineered plants and animals cannot report the physiological effects that they are experiencing.

Also, in response to another post, of course child-bearing is always a risk. E.g., x% of pregnancies may result in spontaneous termination or problematic offspring. If genetic engineering can guarantee offspring without problems at a better rate, then some people might be willing to give it a try. That is a big "if". In any case, since the success or failure of offspring is determined over the whole lifetime (i.e., a screwy intervention on the human genome might manifest itself late in life, just like many diseases have a late onset), people will be conservative about adopting such techniques until they've been proven to work.

In other words, even in the best case, it's unlikely that we'll see wide application of genetic engineering in any country in less than seventy years. Personally, I doubt it will ever happen for any of the reasonably complex properties of humans.

Randall Parker said at June 3, 2003 12:41 PM:

Patrick,

Yes, genetic engineering is happening right now. You can bet that people will put genes into dogs and monkey and assorted other species to raise their intelligence.

Pinker and Beastiality? Are you sure? I think you might be confusing him with Peter Singer.

Dienekes, Many people do not approach their lives in a conservative fashion. Given the chance to give their kids a 30 or 40 point IQ point boost I bet plenty of 90 or 100 IQ women will figure that it is worth the risk to try it.

Dienekes, You seem to be missing my point that existing genetic variations can be used. There is far less risk in putting IQ-boosting genes into fetuses if those genetic variations can already be shown to be helping assorted existing adult geniuses to be smart enough to win Nobel Prizes and make other big accomplishments.

Bob said at June 3, 2003 2:43 PM:

Dienekes,

Emergent qualities? Like what? I have met intelligent dogs. (Haven't I, Randall?) In fact, I still know a few. I have also met highly moral dogs. I really cannot think of any quality that is truly unique to humans except in degree.

If you are suggesting that people are cautious when it comes to playing with nature, you apparently have not met many bodybuilders and you are totally ignoring the fen-phen phenomenon. If someone came up with a genetic manipulation that would turn male offspring into football or basketball stars, they would have to beat the parents back with a stick. If there were a genetic manipulation that would prevent a child from ever becoming obese, millions of obese parents would go for it just to save their child from the abuse they experienced.

If there were a genetic manipulation to treat huntingon's chorea... that one will happen way before 70 years from now.

Regards,
Bob

Patrick said at June 3, 2003 4:04 PM:

Oops I was totally wrong about the Pinker/Singer thing. I meant to check that but was rushed. Totally my fault, I really should never accuse anyone of something that horrible without being sure.

Getting back to the genetic engineering of desirable traits: Sure something like the musical talent of Mozart may be an extremely finely balanced emergent property that could never be replicate with our crude technology. But a lot of desirable properties are most certainly not.

Making your Sons 6 foot 3 rather than your own 5 8, increased muscle density, lower fat content, larger breasts (for the daughters naturally). These are all exactly the sort of things that are being developed for animal targets. These are all simple properties that have been bred in animals for thousands of years and that there should be no trouble in introducing to humans.

In many cases they are simple the increased or decreased expression of a couple of protiens or hormones. This is the sort of thing we can just about do from scratch (when I say we, I mean genetisists, I personally wouldn't be able to distinguish the DNA from the Lab cleaners lunch). And it is the sort of thing that we can do externally by pharmaceutical means. (Growth hormone, anabolic steroids)

There is no reason this can't be done with humans right now. Or rather, no technical reason. Morally it is the same as force feeding steroids to little children, ie. Not Acceptable except to cure disease.

Dienekes said at June 3, 2003 4:40 PM:

>> Dienekes, Many people do not approach their lives in a conservative fashion. Given the chance to give their kids a 30 or 40 point IQ point boost I bet plenty of 90 or 100 IQ women will figure that it is worth the risk to try it.

People are conservative about important decisions. People might be tempted to give their children a 30 point IQ boost, but if it carries a high enough probability of screwing up their kid, most won't. In any case, the whole discussion is irrelevant, as it is yet to be determined that such an IQ boost can be effected with genetic means.

>> Dienekes, You seem to be missing my point that existing genetic variations can be used. There is far less risk in putting IQ-boosting genes into fetuses if those genetic variations can already be shown to be helping assorted existing adult geniuses to be smart enough to win Nobel Prizes and make other big accomplishments.

Again, you are assuming that such "IQ-boosting" genes will be a success. Even if that turns out to be the case, introducing a certain gene to an adult AFTER maturation and introducing it to a foetus are two different propositions, since it is not clear how a gene will interact with other genes during growth.

>> Emergent qualities? Like what? I have met intelligent dogs.

I doubt that you have met many dogs that have the ability to speak, or that have a culture.

>> If you are suggesting that people are cautious when it comes to playing with nature, you apparently have not met many bodybuilders and you are totally ignoring the fen-phen phenomenon.

Indeed, and how much of the population are those? 0.01%? There is no doubt that some Utopianist kooks will be willing to try anything, but that is not equivalent to genetic engineering having an impact in the world-at-large.

>> These are all simple properties that have been bred in animals for thousands of years and that there should be no trouble in introducing to humans.

Breeding properties into animals via artificial selection in no manner guarantees that it will be "easy" to introduce such properties to humans via genetic engineering - a completely different proposition. Any genetic modification has inherently unpredictable effects to the organism, unless one understands the full function of the organism, as well as its interaction with the environment. Little tweaks like removing sickle-cell trait will doubtlessly be possible. But, even then, if one did not understand the benefits of sickle-cell trait, like we do, they would be potentially damaging their patients.

For the more complex human traits, it will not only be difficult to figure them out, but also more difficult to figure out how to use them. As for the higher human properties, like "talent", these will doubtlessly remain obscure for a long time.

Bob said at June 3, 2003 5:47 PM:

>> Emergent qualities? Like what? I have met intelligent dogs.

>I doubt that you have met many dogs that have the ability to speak, or that have a culture.

Dogs have culture and society. Dogs mouths and throats prevent speech, but dogs can understand speech. And they can communicate. Some other animals--not all of them primates--are capable of complex communication using sound and other means. This quality is not unique to humans--we differ only in degree.


>> If you are suggesting that people are cautious when it comes to playing with nature, you apparently have not met many bodybuilders and you are totally ignoring the fen-phen phenomenon.

>Indeed, and how much of the population are those? 0.01%?

I can assure you that orders of magnitude more than 0.01% of the American male population have abused anabolic steroids at one time or another and the incidence is increasing.

"Six million Americans took some Fen-Phen or Redux combination." Assuming a 300 million population, that's 2% of the population.

Remember too that the 2% is just the thin end of the wedge. Consider IVF. Pioneering work was done on in vitro fertilization in 1969 with the first successes less than a decade later. At first, people worried about all sorts of Frankenstein scenarios. Less than thirty years later the use of IVF was quite widespread in fertility clinics and widely accepted by the general population.

You are predicting 70 years before wide application of technology that is already widely applied. Even if we ignore the sloppiness and only consider the application of this technology for human development, we can predict that early successes will happen in the next decade or two with wide acceptance by the general population following shortly thereafter.

The first interventions probably will not interest the general population (huntington's chorea, sickle cell, cystic fibrosis etc.), but the big money is in those interventions that will interest the general population.


>> These are all simple properties that have been bred in animals for thousands of years and that there should be no trouble in introducing to humans.

>For the more complex human traits, it will not only be difficult to figure them out, but also more difficult to figure out how to use them. As for the higher human properties, like "talent", these will doubtlessly remain obscure for a long time.

Who said anything about talent? What about something simple like the shin flexibility that puts a small tribe in Kenya way ahead in races? Or what if we discover a simple chemical difference that affects behaviour? For instance, what if someone determines that a specific gene variant causes me to have anger reactions that fire suddenly and strongly while a different variant causes Randall to have almost no temper at all?

Or what about a genetic variant that simply makes brains more efficient? One brain might apply that efficiency to music while another might apply it to chess or another to physics. Another brain might not apply the efficiency to anything worthwile at all, but that won't stop parents from wanting the best for their children.

Dienekes said at June 3, 2003 8:25 PM:

>> Dogs have culture and society. Dogs mouths and throats prevent speech, but dogs can understand speech.

There was an article on either Science or Nature on human speech recently. It turns out that the basic component differentiating human communication from animal communication is recursion.

>> I can assure you that orders of magnitude more than 0.01% of the American male population have abused anabolic steroids at one time or another and the incidence is increasing.

Interesting. 2% is still negligible in terms of its influence to the culture at large. Perhaps, Americans will be more hasty in adopting genetic engineering as well, just as they were more hasty in adopting genetically modified food. We'll have to wait and see what the effects are.

To conclude, genetic engineering of human beings is still a potentiality that needs to overcome numerous technical and social obstacles to become a reality. People are usually ignoring the technical aspects by resorting to the argument that "we'll have enough computing power". Strong AI was predicted for today, and despite a remarkable predictable increase of computing power following Moore's Law it's nowhere near to being a reality. It was also predicted that we'd be on Mars today, but that hasn't happened either. We'll just have to wait and see how the "genetically engineered humans" proposition plays out.

Bob said at June 4, 2003 4:00 PM:

Only 0.5% of the population actually use in vitro fertilization, but IFV has become accepted by the general population as a normal reproductive technology for infertile couples nonetheless.

Fen-phen was not around all that long, or the percentage would have been much higher than 2%. And remember, we are talking about the thin end of the wedge.

We already have sufficient computing power for gene discovery. What we need are faster sequencers, larger gene arrays, more mature nanotech etc. We also need safe, reliable vectors for introducing the genetic changes we want.

The predictions for AI and Mars were always highly speculative, ignored market forces (What the hell do we need on Mars?), and were based on no discernable facts. We know what genes are. We know what some genes do. We have developed technologies for disabling and for introducing specific genes, and we have already used those technologies successfully on non-humans.

Moore's law enables us to make accurate predictions within limits. For instance, it would allow someone playing Alternate Reality on an Atari 1040ST in 1985 to predict video games with the features of Duke Nukem or Quake at some point in the not too distant future. It might even allow that person to predict when those features would become feasible to within a year or two. Moore's law does not allow anyone to make predictions about unknown quantities like when someone will invent an effective technique for AI. Nor can it allow anyone to predict when someone will actually want to use an existing available technology. To travel to Mars for instance.

The current state of the art in genetic modification is much more like Alternate Reality on an Atari 1040ST than it is like AI. We know genes are out there. We know we can identify their function. We know we can manipulate genes in living organisms. It is only a question of when. Because the technologies are not as mature as semiconductor technology was in 1985, and because medical treatments are heavily regulated, it is not possible to predict things down to the year, but I think it is unreasonable given the current state of the art to assume another 70 years before GM humans are common enough for wide acceptance.

amy greenwood said at June 4, 2003 6:55 PM:

I think one of Pinker's main points bears repeating. The fact individual genes have multiple functions, and that these functions are highly modulated by genetic background, makes the prospect of altering genes in embryos an extremely risky prospect.

Let's say we do find the gene or genes responsible for bone flexibility.... I don't think anyone who is well informed would be willing to create a human person from modified cells, not knowing whether that individual's genome contained the proper regulatory elements, modifiers, suppressors, etc. True, lots of transgenic animals have been created in the laboratory. But more times than not there are many unexpected phenotypes, including embryonic lethality, and there are also plenty of examples where a mutation will have radically different effects depending on genetic background.

I think all of Pinker's other points are just to make the argument that even if a few people do use alter their embryos, its not likely to have the "dehumanizing" effect that worry-warts like McKibben, Kass, and others say it will have.

Personally, I thought it was an excellent article.

Randall Parker said at June 4, 2003 11:05 PM:

Amy,

Genes governing bone flexibility are relatively less important. Genetic engineering of the mind is far more important.

The purposes of all genes will be discovered in the next couple of decades. The cost of DNA sequencing will fall by several orders of magnitude in the next two decades and then we will have the ability to discover the significance of most human genetic variations just by comparing the health, intelligence, behavior, and personality of millions of people along with their DNA sequence differences. The answers will jump out at us.

The most interesting modifications will be in genes that affect the brain. Intelligence boosting and personality altering genetic modifications are far more important than ones that will make bones stronger or more flexible. The amount of variation already existing in the human population in genes that affect intelligence and personality will allow people to discover all sorts of variations they can give to their offspring using genetic engineering that will make their offspring have more abilities, different mental states (e.g. less susceptibility to depression or less shyness), and behavioral tendencies that the parents do not have.

As for risk: There is no reason to think that the level of risk will remain as high as it is today. New techniques for germ line genetic engineering will be developed. The ability to add a gene to an embryo with extremely limited side-effects will be developed.

In a nutshell:

1) We will discover the significance of all existing genetic variations.

2) Many people will want to give their offspring genetic variations to give them features (in the brain, resistance to disease, athletic abilities, physical attractiveness, etc) that they do not themselves have.

3) People will want to avoid passing down harmful mutants. Just about every person has harmful genetic variations because there are always new variations popping up that take many generations to get selected out.

4) The procedures for modifying genes in embroys will become more advanced, easier to do, safer, and cheaper.

5) Faced with lower costs, declining risks, and all sorts of desireable outcomes lots of people (millions) will elect to have genetic engineering done on their offspring.

6) Eventually new genetic variations as well as genes from other species will be discovered that will be useful to introduce into humans (e.g. Aubrey de Grey's proposal to add fungal hydrolases to clean out lipofuscin and other accumulated junk). Therefore as time progresses there will be ever more advantages to doing genetic engineering to offspring.

As for whether the results will be dehumanizing: it depends on what features the future parents choose. An attempt to say that McKibben's argument is lame because no one is going to use genetic engineering is itself even more lame than McKibben's argument. Pinker's article was very disappointing. He's smart enough and knowledgeable enough to know better.

Any biological scientist who thinks we are just going to learn all about our genetic endowment and use it only for limited medical purposes and not seek to improve on it ought to keep in mind cyper-punk science fiction writer William Gibson's great dictum: "The streets find their own uses for technology".

amy greenwood said at June 5, 2003 9:22 AM:

Randall,

I agree with you on points 2-6, but I think the biology behind point 1 (the fundamental point) is more complicated than even the best bioinformatics approach can address. Let's take your example of a mutation that influences IQ. Suppose we sequence DNA from a million people and find that there is a mutation X that correlates with higher IQ. What is likely is that all of those people who carry X also have a constellation of other genes that allow that mutation to 1. have the IQ enhancing effect, and 2. (more importantly) provide a context in which that gene is not lethal or deleterious. I agree that it will be technically possible to modify a human locus by homolgous recombination in the foreseeable future. But...from what we know about genetic engineering in mice, modifying a locus in a human embryo that has to go through its full development is not a wise risk if you need to be reasonably sure that there won't be nasty side effects. The reason is that even if we knew completely the function of each gene, there is nothing we can do about the fact that gene X, the IQ gene, ALSO happens to have an important role in blood vessel development. Happens all the time.

On the other hand, embryo selection based on microarrays..sure. Gene modification in adult tissues...sure. Cell replacement..sure and let's hope so. Even embryonic gene therapy to fix single hit deleterious mutations in genes that EVERYBODY has (e.g., hemoglobin, insulin)..sure. And, I also agree with your Gibsonesque scenario that some people are going to try more. But I'm saying that a lot of them are going to be sorry.

Also, I think Pinker was trying to make the point that limited use of such techology would not be dehumanizing for society as a whole. Of course it would be dehumanizing for a child if the parents chose to make the kid a freak. Very sad, but not the end of civilization.

Randall Parker said at June 5, 2003 12:08 PM:

Amy,

The biggest limitation I can see with the bioinformatics approach of comparing large numbers of people for DNA sequence and IQ is that such things as methylation patterns and epigenetic state will not be controlled for in such an approach. Even if one could measure methylaton patterns in a cell sample from an adult the reason that some guy is a genius might be due to a transient methylation pattern that sometimes occurs randomly in embryonic development. Also, there are nutritional and other environmental factors that will be other variables that are harder to control for. However, I really expect a great many genetic variations will contribute to IQ and that these will be detectable with a large enough sample.

One reason I think that IQ-linked genes will be discoverable is that there is such a large range of human IQ. In choosing a sample of humans to analyse it would make sense to use a larger proportion of high IQ people than they are as a percentage of the population as a whole. There are going to be genetic variations that just do not occur in them or groups of variations that together do occur in them.

Whether every variation in a group correlated with higher IQ is necessary to achieve the IQ-boosting effect is besides the point if they can all be introduced into an embryo. However, just having a group of SNPs that occur together and which are strongly correlated with high IQ will help a great deal in pinning down which variations are relevant.

As for the constellation of other genes: I expect them to be detected as a group. I do not see why they would not be. Granted, if a number of variations all occur together on the same chromosome it will be hard to tell which ones just were there by chance already in some person thousands of years ago had some mutation that boosted their IQ and which were necessary for it to work. But there are high IQ people from a number of parts of the world and so if we also use high IQ people from many locations that ought to help to control for that problem.

As for the problem of an IQ-boosting gene also affecting angiogenesis: well, if someone has the variation that boosts IQ and still has wounds that heal okay what are we to worry about? If an IQ-boosting mutation is of fairly recent origin and occurred in a population that was under heavier selective pressure to boost IQ it will be more likely to have deleterious side-effects (e.g. the dystonia mutation among the Ashkenazim - I think Gregory Cochran has a paper coming out on the selection of IQ-boosting genetic variations in the Ashkenazim that have deterious side-effects btw). But surely there are plenty of mutations that boost IQ that do not bring undesireable side effects with them.

As for whether some who genetically engineer their progeny will be sorry: Yes, of course. But that reckless willingness to take risks (which is obviously genetically influenced as well) will accelerate the discovery of which variations have deleterious side effects and what those side effects are.

As for whether genetic engineering will be dehumanizing: in my view it depends very heavily on what the specific effects are of each genetic change. I have a big post coming up on that later today.

Post a comment
Comments:
Name (not anon or anonymous):
Email Address:
URL:
Remember info?

                       
Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright