Jackdaws love my big sphinx of quartz

Jackdaws love my big sphinx of quartz - Modality and morality

Scott

[	userinfo	\|	livejournal userinfo	]
[	archive	\|	journal archive	]

Modality and morality

[Dec. 9th, 2012|12:04 am]

Scott

[

Tags

science

]

[Epistemic Status | All I know about this is from blogs and private conversations with smart people. If there are mistakes, I hope people will correct me.]

Recently I have been excited to learn about a new branch in genetic research focusing on mutational load and the "modal human" (in the sense of mean, median, and mode).

Two major sources of genetic diversity are polymorphism and mutation. A polymorphism is normal large-scale variation in a population. A classic example would be light skin versus dark skin in humans. There are large groups of light-skinned humans having light-skinned babies, and large groups of dark-skinned humans having dark-skinned babies. Polymorphisms usually exist because of different selection pressures in different groups; for example selection favors light skin in polar areas and dark skin in equatorial areas. Or they can exist because a beneficial mutation hasn't had time to spread; for example, lactose tolerance is only a few thousand years old and hasn't reached the entire population. Or they can exist because different genes give different costs and benefits that add up: for example, the sickle cell gene both produces a dangerous anaemia and protects against malaria, so some people have it and some people don't and the resulting genetic diversity is "good for the species". Something usually gets dignified with the "polymorphism" label if it's in somewhere between 1% and 99% of the population, although this is pretty arbitrary.

Another source of genetic diversity is mutation. Most people are born with a few hundred mutations just from normal genetic copying errors. Most of these mutations are bad, for the same reason that if you made a hundred random changes in a computer program, most of them would make the program work less effectively. Unliky computer programs, humans are pretty robust, so even with a few hundred deleterious mutations most people do just fine. There are a few well-known mutations that cause total disasters, and these are your genetic diseases like muscular dystrophy. Because people with these diseases rarely reproduce, evolution does a good job of selecting against them most of the time; because the mutations keep cropping up, you get some muscular dystrophy sufferers in each generation anyway.

After failure to explain some traits through polymorphism, geneticists have turned to the idea of mutational load. The reasoning goes like this: we know things like intelligence are heritable because we can correlate parent intelligence with child intelligence in adoption studies and get high numbers. But every attempt to look for an "intelligence gene" has failed; if you take a bunch of smart people and a bunch of stupid people and look for a gene different in group A than in group B, you either find nothing or find something with such a small effect it's not worth your while. Enter mutational load, the idea that everyone has a different set of harmful mutations and different people may have different numbers.

Imagine that there are 1000 genes involved in making your brain run smoothly. And imagine two people, one of whom has mutational load of 2% and the other of whom has mutational load of 1%. The first person will have 20 brain genes working suboptimally, and the second person will have 10 working suboptimally. None of these mean either person won't have a brain or will be clinically retarded. It just means the second person's brain will work a bit better than the first person's. But if you take a thousand people and try to find the one gene that distinguishes smart people from stupid people, you will fail. Miserably. Every single stupid person's stupidity may be caused by a different set of mutations. What's important is less the exact genes and more the total number of flaws.

But mutations are pretty straightforward in a way that polymorphisms aren't. Polymorphisms don't really involve value judgments: I have light skin, you have dark skin, we're both part of the beautiful tapestry of human diversity. Mutations are more like 999 people have the gene one way, 1 person has the gene another way that works worse, and that last person is wrong. Which invites the question: if a lot of human differences are caused by who has which genetic mistakes and how many genetic mistakes they have, can we imagine a perfect person with no genetic mistakes? Someone who, at each genetic locus, has the gene variant which is most common in the human population as a whole?

Such a person, sometimes called "the modal human", could be be taller than Shaq, smarter than Einstein, and better than us in a bewildering variety of unexpected ways. Gregory Cochran writes:

"What would a spelling-checked person, one with no genetic typos, be like? Since no such person has ever existed, we have to speculate. I figure that kind of guy would win the decathlon, steal your shirt and your girl – and you still couldn’t help liking him."

Behold, I teach you the overman! Man is a thing to be surpassed!

If the "modal human" school of genetics is right, this would have some pretty interesting philosophical implications.

At the very least, the Platonists and Aristotelians would be thrilled. It turns out all this stuff about humans being flawed copies of a perfect human form has something to it after all, even if it's not what anyone expected. And this sort of captures something I've been finding more and more often lately, which is that although value isn't objective, per se, it is less random than we tend to expect. There's no little XML tag on people saying [goodness rating="5"], but a lot of the things we think of as "good" vary together and have simple biological correlates. This is true in nutrition, where to a first approximation some foods are "good" and other foods are "evil" no matter whether you want to avoid diabetes or heart disease or obesity or just have lots of energy. It's true in psychology where the "bad" personality traits very often vary together. And apparently it's true in genetics, where all of the things we like - attractiveness, intelligence, charisma, health - now seem to be groundable as Objectively Good Things in the sense of low mutational load.

It also makes the war over designer babies - currently just a potential war, but it seems like the sort of thing that will happen - much more interesting. The bioethics orthodoxy has long been that it's okay to use biotechnology to cure disease, but that if we try to use it to make people "better than normal" we will inevitably end up with a dystopia of a million identical creepy staring blond kids. ("But I was just trying to make people less susceptible to diabetes by changing metabo..." "DOESN'T MATTER! IT WILL LEAD TO ONE MILLION IDENTICAL CREEPY STARING BLOND KIDS!")

But if the theory of mutational load is right, then imperfection is a disease. Trying to make our children smarter isn't changing from one polymorphic variant to another, it's curing a hundred little micro-diseases that they have.

Even more interesting, different people have different amounts of mutational load. If someone gets tested and finds they have a mutational load of 2% compared to a population average of 1% - and we would expect such a person to be obviously defective, in the sense of ugly and unintelligent and so on - then it would be reasonable for them to want some of those mutations cured in their children. It's hard to imagine even a bioethicist saying "You're provably biologically worse off than the rest of the population, we know it's going to give your kids an insurpassable obstacle that will keep them lower class forever, but we're going to forbid you from fixing it." But it's also hard to imagine bioethicists remaining remotely credible if they were to say "People with more load than average are allowed to bring their load down to average, but no one can go any further than that." Plus as long as load correction technology is allowed to exist it would be almost impossible to prevent people from "abusing" it.

(whoa. I think I just realized why ugliness is unattractive. If there is a standard alignment for facial features, then deviation from that alignment will correlate with high mutational load, and it will be a useful signal for the opposite sex that you are genetically inferior and should not be mated with. I assume someone else has figured this out before.)

The final interesting philosophical implication is that if you could test your mutational load, this would look like an objective rating of how good a person you were. I say "look like" beacuse it would probably be only a medium-level correlation with success, in the same way that IQ is only a medium-level correlation with success. And observed characteristics always correlate with observed characteristics better than genotype will - if you want to test someone's intelligence, asking them whether they watch Keeping up with the Kardashians is a cheaper AND more effective test than taking their mutational load. But our culture seems to be pretty into desperately avoiding the idea that an objective ranking of people is at all possible: IQ tests are pretty much allowed to exist only because we can shrug and say they're culturally biased.

So even though measurement of mutational load would be a much worse measure of human goodness than dozens we have available right now, it would have more dignity to it, it would look more final, people couldn't dismiss it as easily as they do everything else. I don't think this would end in a GATTACA-style genetic faux-meritocracy; the government is pretty "good" at blocking those sorts of things. But I could easily imagine more subtle effects.

Imagine what would happen if different groups of people had different mutational loads. There's some empirical evidence this might be the case, although I think the whole field is kind of preliminary. For example, West Hunter presents evidence that load is affected by parental age; older parents pass more mutations on to their children, so cultures in which people marry very young should have lower load than culture that encourage people to marry when they're older (yikes! Mohammed was right!) If it is found that different regions, religions, or ethnicities have different mutational loads, then imagine what happens when someone from a mutationally disadvantaged group demands access to genetic correction technology and claims discrimination when they're refused. I can not tell you how much money I would pay to watch a full-on fight between the social justice people and the bioethicists.

A few months ago, I asked if it was possible to design a eugenics policy that didn't immediately set off "obviously evil" sensors:

"I can imagine a lot of contrived and not very interesting policies, like only people who get a certain IQ score can reproduce, or criminals can't reproduce, and things like that. But I feel like there should be something more interesting and less obviously 'The government is going to subvert this to kill off dissidents, isn't it?' Something that gives eugenics the same kind of 'Huh, didn't expect that would work' factor that prediction markets give decision-making, or that public-key cryptography gives security.

People gave some interesting answers, but with a few exceptions they didn't make me think they were clever in the same way prediction markets are clever. They weren't aikido-ish, in the way my aikido teacher keeps talking about how aikido teaches you philosophically as well as physically to find some quasi-magical approach that makes resistance inconceivable.

The idea of mutational load seems aikido-ishly good. It's objectively measurable and value-neutral, and so hard for a government or individual to game. It would lead to improvements in the human population that would pattern-match our natural idea for what it means for a population to improve, not some statement drawn up by a committee with its own agenda.

And most important, it would allow the preservation not only of human diversity, but of human randomness. If we correct harmful mutations while leaving polymorphisms alone, then we can improve people holistically without doing any damage to concepts like race or even within-group variation. Creating the genetic superman wouldn't have to be a matter of a million identical creepy blond kids. You could take an embryo and just correct the obviously harmful genetic mutations, leading to a kid with twice the IQ and perfect health - but she would still have her mother's eyes, her father's skill at basketball, the musical talent that runs in her family, et cetera. Since race is a polymorphism and not a mutation, it would be completely race-independent; there could be modal white people, modal black people, modal Asians, and so on.

You would be subtly correcting the mistakes of Nature, not running rough-shod over it.

link

Comments:

Page 1 of 42
<<	[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42]	>>

From:

alicorn24
2012-12-09 05:25 am (UTC)

I want a modal baby.

How close are we to this tech?

(Reply) (Thread)

From:

squid314
2012-12-09 05:50 am (UTC)

I think people could probably calculate the modal genotype now, although I don't think anyone's ever tried it and maybe it would take a few years.

I imagine the process would be something like someone reading a random egg cell from you and a random sperm cell from your husband, putting it in a computer, spell-checking it, then printing out the genotype for the baby which gets inserted back in the egg. Even the back-in-the-egg insertion part would require something like cloning, which we can't do for humans yet, but the real trouble would be printing out DNA. This is way beyond what we can do now, but genetics seems kind of Moore's Law-y and I could imagine genetic printing only taking a couple of PCR-level insights.

I predict we will get better at this very quickly, but I'm not even a knowledgeable amateur in any of the appropriate fields. I would say (wild guess) 33% chance it happens in your reproductive lifetime, 66% that it happens in your actual lifetime.

One dyslexic mistake

dk

wholesale price

where can i buy a cheap beats by dre headphone online