There’s been a great deal of fuss about the Bundesbank director Thilo Sarrazin’s book, in which he argues that the “upper layers” of German society ought to be encouraged to breed for fear of Muslims, etc, etc. The SZ points out here that he confesses to just making up his numbers:
Es ging um die Frage, woher Sarrazins viel zitierte, im Brustton der FaktizitÃ¤t vorgetragene Behauptung eigentlich kommt, dass siebzig Prozent der tÃ¼rkischen und neunzig Prozent der arabischen BevÃ¶lkerung Berlins den Staat ablehnten und in groÃŸen Teilen weder integrationswillig noch integrationsfÃ¤hig seien. Sarrazin gab zu, dass er keinerlei Statistiken dazu habe. Er gab zu, dass es solche Statistiken auch gar nicht gibt.
But I’m not sure if anyone has pointed out quite how strange Sarrazin’s thinking is.
FÃ¼r ihn ist die Unterschicht sowieso schon lange abgeschrieben, der Genpool degeneriert. Denn bereits seit dem 19. Jahrhundert sei die deutsche Gesellschaft immer durchlÃ¤ssiger geworden, “auffallende Hochbegabungen” hÃ¤tten damals in PreuÃŸen bereits die MÃ¶glichkeit bekommen, das Gymnasium zu besuchen. “Das bedeutet aber, dass die Entleerung der unteren Schichten von intellektuellem Potential bei uns weiter fortgeschritten ist als in Gesellschaften, deren DurchlÃ¤ssigkeit sich erst spÃ¤ter entwickelte.”
He thinks, or at least claims to think, that because the German (and specifically Prussian) education system has given the lower classes the opportunity to go on to higher education since the 19th century, Germany has a problem – the masses have been emptied of “intellectual potential” too early.
What strikes me as telling here is that it’s not just that Sarrazin’s political thought is trapped in the Wilhelmine era – his understanding of genetics is, too. This post of Razib Khan’s on the great early-20th century debate between the biologists who rediscovered Gregor Mendel’s work, and the biometricians, who had been trying to link data gathered on the range of human traits with the Darwinian inheritance, explains why.
The biometricians were essentially trying to operationalise Darwinism with early statistical methods. This gave them a problem; a lot, but not all, of the variation in biological traits at least seemed to be nice and smooth, movement along a well-behaved curve. With no other model of inheritance available, they assumed that genetics was a simple process of blending – children were an average of their parents. This had wide-ranging consequences; it implied that regression to the mean would apply to people. We would all eventually be average. From there, it wasn’t hard to predict that we would all, eventually, be mediocre and that racial degeneration was inevitable.
This is one of the great intellectual accident black spots – a nauseous gap in the barrier by the roadside. Experimental work, like Mendel’s, showed that something else was happening. One of the problems was that statistics itself needed to advance to resolve the debate. There is a very good reason why Francis Galton was both an important early statistician and a eugenist, and why it would eventually be a statistician, R. A. Fisher, who demonstrated that a Mendelian process was observable in the biometric data.
But by that time, the original mistake had set off a great avalanche of analogies. Social Darwinism and everything that followed from it was out there. It’s a horrific thought that its consequences have a lot to do with statistical methods, and it’s telling that Fisher published in 1918. The important point about Mendelian genetics is that it’s discontinuous – it doesn’t blend down to the average. Variation is conserved; not only will the German working class continue to produce bright kids, the elite will occasionally toss out a Sarrazin.