How to create a genetically high IQ population while avoiding regression to the mean as much as possible?

First, putting all morality issues aside, IQ can be well defined, and in fact it is, by current intelligence tests. For the purpose of this question simply divorce yourself from the notion that the IQ test has any meaning other than the score ON THIS TEST, it doesn't have to correlate with any kind of life success, survival skill, or moral worth. Consider it like any other breeding program: I want to breed chickens to grow fast and produce the most consumable calories in a given amount of time, like 12 months: Along with a few other economic conditions, like no great increase in disease, injury aggression or self-harm, at least before harvest at 12 months.

To help separate this from knee-jerk responses to controlling humans, let us consider breeding mice for this trait. We will devise a mice IQ test (already been done, I know). Ours will focus on the ability of mice to solve puzzles and make distinctions in patterns.

High IQ would be treated as a good in itself (something we want without having to justify it), and testing would be mandatory before mating would be allowed. Pregnancies without testing would be aborted. Testing at puberty would result in sterilization; or perhaps temporary sterilization (an easy surgery in males, a pill in females) if we wish to allow multiple chances to allow for having a bad testing day.

Breeding could be allowed by choice, just a restricted choice: A female with a desire to mate could be allowed to choose her mate from acceptable mates, for example. We might want a trait in which females prefer smarter mates, and select for that, too.

You could also apply human levels of intelligence to develop the genetic science of what it takes to score high on the IQ test: Correlated genes and gene combinations, for example. Note that we don't have to know exactly what the genes do, in order to use them in the selective practice: We can just assign gene sets to groups of "acceptable IQ" and "unacceptable IQ", and then abort or genetically engineer embryos that have combinations of genes that we have seen in the unacceptable set but do not see in the acceptable set.

Regression to the mean is not some magical property, it is a result of random combinations. By analogy, consider the dice game Yahtzee: A roll of 5 of exactly same number is a Yahtzee. Thus there are six possible Yahtzees. If two Yahtzees are "parents" and we mix 2 dice from one and 3 from the other, odds are only 1 in 6 pairings will result in another Yahtzee: The other 5 pairings result in a full-house (3 of one kind, 2 of another), which is 50 times more common than a Yahtzee.

Thus reducing the randomness of gene combinations can improve the odds of getting mice with high IQ scores on the test we devised.

I will repeat: This is a scientific approach divorced from any morality of controlling the mating freedom of mice, or whatever the subject. Also repeated; there is no guarantee that our test is "well-formed" in the sense that the traits it tests for improve the odds of mouse survival, mouse longevity, or mouse quality of life.

In fact, if you believe evolution, in the last few hundred million years, has already balanced the benefits of mouse intelligence with mouse survival and reproduction, then increasing mouse intelligence probably comes at the cost of reducing mouse survival and reproduction. In evolution there is a chance of stumbling upon a free lunch (as plants did with photosynthesis, and as we humans did with abstract intelligence), meaning one that increases survival and reproductive chances; but if the free lunch is exceedingly rare then the chances of finding it are, correspondingly, exceedingly low.

Of course engineering is in the business of making things that would be nearly impossible to achieve by chance either certain or for more probable: If I am allowed to turn the dice, I can have a Yahtzee every time, instead of averaging just 1 in 1296 throws.

Thus selective breeding and selective genetics can avoid regression to the mean as much as possible, by reducing the randomness of gene combinations to those known to work. In the end, given some very high scoring on the IQ test, simply cloning the winner's DNA will eliminate the chance of DNA being the reason for low scoring, and allow the breeders to move on to other factors (like infant and child nutrition, education, exposure to pollutants and disease, removing stressors known to reduce IQ performance like isolation, frequently interrupted sleep due to noise or other irritants, frequently changing environmental conditions, frequent incidents of physical pain, etc).

posted about 7 years ago

Amadeus‭

31 reputation 1 114 3 0

How to create a genetically high IQ population while avoiding regression to the mean as much as possible?

To avoid the leveling drift (regression to the mean) in a population, what you need to do is Hybridize:

A hybrid may occasionally be better fitted to the local environment than the parental lineage and as such natural selection may favor these individuals.

The Bell curve is considerably flattened and shifted to the right - as most hybrids don't survive, but those who can thrive, will.

So a problem occurs if you wish to genetically isolate a particular population - regression to the mean is the least of your worries if there is no selective pressure to improve the population's intelligence.

But I guess you figured that out, so pressure selecting for inteligence to allow for reproduction then:

You'd be looking at a form of licensed fertility in the wider population based on whatever test of inteligence you deem appropriate. So let's test people (Vis a vis: Hunger Games like?) at an age before they can reproduce and depending on the outcome, sterilize them (I posit that enforced death is a form of reproductive sterilisation) or support them to breed.

The various interactions between intelligence and personality traits are more complex than simple hope would wish. It's kinda difficult to select for just intelligence, intelligent aggressive people will likely always have an edge over the agreeable intelligent ones in any kind of spontaneous combat, or competative environment - will you end up with a society of warlords/conquerors or can you develop a methodology that selects for other characteristics?

I feel that rather than answer the question here, I've re-asked it at a slightly higher resolution. Does this framing suit the OP? If not, it's a subject that interests me, let's go further.

posted about 6 years ago

Mr Bumble‭

11 reputation 4 74 1 0