[ExI] What if humans were twice as intelligent?

[Anders Sandberg]

On 14/10/2013 13:19, BillK wrote:
> On Mon, Oct 14, 2013 at 12:19 PM, Anders Sandberg  wrote:
>> The effect is fairly weak. While the image of Idiocracy is vivid, when you
>> actually sit down and simulate genetic equilibria of numerous small
>> IQ-related genes you will find that the anti-intelligence selection effect
>> is not doing much.
>
> Hmmnnn. Not simulations, but real world studies have found a correlation.

Yes, but so what? I am not denying that there might be a lower fertility 
at the high end of the IQ spectrum (although the data is somewhat tricky 
- cohort effects, delayed childbearing, smaller samples etc.) However, 
this does not lead to dysgenic pressure.

See http://www.jstor.org/stable/2781579 or Julian Simon's excellent 
simpler explanation: 
http://www.juliansimon.com/writings/Articles/PRESTON3.txt
The low-IQ people have low fertility, so the population tends to be 
"pushed away" from too low IQs even if there is some push away from high 
IQs too.


I did a small simulation myself (since I could not find my old 
simulation): assume 100 alleles that can be in two states, each one in 
the "one" state adding a fixed N(0,1) normally distributed number of IQ 
points to all individuals who have it (this is pretty close to Plomin's 
current result for common IQ-related alleles). Some boost IQ, some 
decrease it. Assume a fitness that is a piecevise linear function: 0 for 
IQ 0, 0.1 for IQ 50 (probably a major overestimate), rising to 1 for IQ 
80, 1 at IQ 110, going down to 0.5 for IQ 150 and 0 at IQ 400 (these 
numbers were chosen by guesswork during a non-connected train-ride, but 
I think they are not too crazy compared to the Preston and Campbell 
data). Select parents with a probability proportional to their fitness, 
have them generate offspring using cross-over of their respective genes.

This model produces the expected behavior: the population does maintain 
an equilibrium pretty far away from the lower limit, staying stable 
because of the upper limit. If I remove the high-IQ fitness penalty the 
population drifts upwards very slowly, with diminishing speed as more 
alleles get fixed. Same thing (by symmetry) for removing the low IQ 
limit. In practice, we can likely rely on our real-world IQ distribution 
having a shape roughly in line with what the fitness constraints have 
been during most of the past - and the number of generations it takes to 
affect the frequency of an allele contributing 1 IQ point is pretty 
large, given that its fitness effects gets mixed up with 99 others.

-- 
Anders Sandberg,
Future of Humanity Institute
Oxford Martin School
Faculty of Philosophy
Oxford University