Stephen Jay Gould on “The Bell Curve”

Stephen Jay Gould on “The Bell Curve”
Abagond, Sat 10 Jul 2010

Disclaimer: I have not read “The Bell Curve” for myself. The following is based instead on Stephen Jay Gould’s take on it. As a professor of biology at Harvard he knows the science way better than I do – like some of the important stuff “The Bell Curve” left out. He wrote about it in the 1996 edition of “The Mismeasure of Man”.

“The Bell Curve” (1994) by Richard J Herrnstein and Charles Murray is suppose to give the science proving that blacks in America have less intelligence in general than whites – and always will because intelligence, as measured by IQ, is mostly inborn or genetic. Further, they say this lack of intelligence is why things like crime, unemployment and illegitimacy are so high among blacks – and why throwing tax money at them will make little difference. In short: warmed-over social Darwinism.

As it turns out, even if you go by their numbers, IQ is weakly correlated with illegitimacy and so on – the numbers match up badly. So badly, in fact, that when they draw the lines on their graphs to show the relationships, they leave out the dots, the scatter of data points the lines are based on!

Further, correlation is not cause. After all, there is a much stronger correlation between your age and the national debt and yet neither causes the other.

And even Herrnstein and Murray admit that IQ is only 60% genetic. So only somewhat over half of these weak correlations-which-are-not-even-causes have anything to do with the inborn intelligence of black people.

And it gets worse:

All this is based on only one set of data with the numbers worked a certain way. But there are other sets of numbers, which they overlook – but would not if they had a strong case. And there are other ways to work the numbers. In fact, you can even show that it is impossible to measure intelligence by a single number!

That is no small point: Their argument depends on expressing intelligence as a single number. Not all scientists agree you can. The authors admit as much but do not make it plain why anyone would say that, a point which would cast their whole book into doubt – and make it clear that there are other ways to read the numbers. By leaving this out you do not see how weak their case is.

Their argument also depends on assuming that IQ is largely genetic therefore unchangeable. Not true.

Just because something is genetic does not mean it is set in stone. Height seems to be even more genetic than intelligence yet it is heavily affected by what you eat growing up – as people from India who bring up their children in America discover.

The same with IQ: in some countries it has gone up more in the past 50 years than the difference between blacks and whites in America. There is even a name for it: the Flynn Effect.