[ExI] What if humans were twice as intelligent?

rex rex at nosyntax.net
Tue Oct 22 10:20:28 UTC 2013

BillK <pharos at gmail.com> [2013-10-22 01:45]:
>On Mon, Oct 21, 2013 at 11:52 PM, rex wrote:
>> http://www.nosyntax.net/cfwiki/index.php/Differential_Breeding
>> http://www.nosyntax.net/cfwiki/index.php/Preston_Campbell_1993
>It looks like there are a lot of differing opinions on this subject.
>Some argue that IQ is just not a heritable characteristic. i.e.
>Regression to the mean within populations.
>But populations are not constant.  There is immigration and emigration going on.
>Others claim that education and poverty are more significant factors.
>Obviously severe poverty or pollution can badly affect a child's
>developing brain. And wealth enables a child's development.
>Others say that it is the number of children in a family that affects
>a child's development. I.e. fewer children with a large gap in years
>between children enables the parents to give more time to caring for
>the child.
>There seem to be just too many factors involved.

Hello Bill,

As you know, IQ is highly politicized. A good read on the history of
the debate is _Born that Way_. For over a generation, the "nurture is
all" group dominated, but they have been on the retreat for years and
except for the diehards, it's recognized that both nature and nurture
play important roles.

The separated identical twin studies are strong evidence that IQ is
highly heritable, though some authors claim heritability is low in low
SES homes. Recent studies have shown that most personality traits that
were once thought to be entirely environmental have a significant
genetic component.

Regression to the mean will occur unless heritability is 1, and
the expected value of trait in the offspring is given by:

O = M + h^2(P-M)

Where M = population mean, h^2 is the narrow heritability, and P
is the mean of the parents. As an illustration of how accurate
this is, Terman started a longitudinal study of over 1500 children
with IQs > 140 in 1922 (it's still going on). The mean IQ was
152, and the mean of their spouses was estimated as 125 (assortative
mating in action) by Terman, so P = (152+125)/2 = 138.5. Using
Jinks and Fuller's estimate of h^2 as 0.71, the predicted mean IQ
of the offspring is 127.33. The measured IQ of 1525 offspring was
132.7 with SD = 16.5. The higher than predicted value is likely due
to the environmental advantage of being raised by high-IQ parents
who also had a higher than average SES. (This is from Jensen's book,
_Educability and Group Differences_, pg 171.)

The Flynn Effect doesn't have a satisfactory explanation, and 
there are other debates still going on, but I'm interested in 
simulation, and in particular in simulating the startling results
of the Preston-Campbell paper. They are highly counterintuitive, to me
anyway, though Julian Simon's explanation quoted on Anders' site
is clear.

Since humans mate assortatively, I'd like add add it to Anders'
simulation, but it's not clear to me how to do that with a gene
simulation. Perhaps Anders can think of a way to add it.

Computers run on smoke. If it leaks out they stop working.

More information about the extropy-chat mailing list