Intellect as a trap

A fascinating read:

He found that liberals and conservatives scored roughly equally on average, but the highest scoring individuals in both groups were the most likely to display political bias when assessing the truth of various political statements.

In a further study (replicated here), Kahan and a team of researchers found that test subjects who scored highest in numeracy were better able to objectively evaluate statistical data when told it related to a skin rash treatment, but when the same data was presented as data regarding a polarizing subject—gun control—those who scored highest on numeracy actually exhibited the greatest bias.

As a case in point, human intelligence evolved less as a tool for pursuing objective truth than as a tool for pursuing personal well-being, tribal belonging, social status, and sex, and this often required the adoption of what I call “Fashionably Irrational Beliefs” (FIBs), which the brain has come to excel at.

Since we’re a social species, it is intelligent for us to convince ourselves of irrational beliefs if holding those beliefs increases our status and well-being. Dan Kahan calls this behavior “identity-protective cognition” (IPC).

By engaging in IPC, people bind their intelligence to the service of evolutionary impulses, leveraging their logic and learning not to correct delusions but to justify them. Or as the novelist Saul Bellow put it, “a great deal of intelligence can be invested in ignorance when the need for illusion is deep.”

This reminds me of how artistic excellence is correlated with schizophrenia - and how engineering correlates with autism.


The big problem is that if you use the same data for different alleged topics, people have different knowledge of factors relating to the correlation-causation fallacy. The more numerate likely have greater exposure to those facts.


(Emphasis mine)

If engineering correlates with autism, why is the U.S. not experiencing an epidemic of engineering?



External pressures of wokeism destroying engineering. If the US graduated only five engineers next year and one of them was autistic, there would technically be a correlation.

And, of course, there are issues of the expansion of the definition of autism and of overdiagnosis.


And there is always the deeper issue of defining what engineering is. However you decide that, the US is sinking fast in that community. Engineering builds stuff, and we have quit building stuff.


Another autism factor could be that assortative within-profession breeding between engineers has grown significantly in the past two generations.

Some genes are like salt: a little bit is great, too much just ruins the dish and eliminates otherwise worthwhile genes from the population.


Just as I hypothesized Circa 2000 when I began investigating #autismspectrumdisorder there may be a reason why the incidence of this disorder correlated so strongly (as my extensive research demonstrated) with the geographic conjunction of immigrants from India and northern Europeans. The susceptibility to pathological side effects will be greater in populations less adapted to unsanitary conditions.

This is not to say that my hypothesis has been borne out. It is merely to say that the 82% objection to the increase in H-1B immigration expansion by Americans Circa 2000 may be justified on more than merely the basis of consent of the governed.

We don’t know.

But it is certainly reasonable to permit people who do not consent to this kind of experimentation on their human ecologies to separate themselves and their children. That the government overrides the consent of more than a supermajority of its people in this potential destruction of the Next Generation by attacking them as “white supremacists”, after it has done so for half a century, is more than justification to treat its de facto hostility as irremediable via such an unresponsive political system.

PS: The aforelinked study was admittedly amateurish but was forced on me after the Berkeley epidemiologist responsible for investigating the rise in autism told me that a cluster in a Silicon Valley company I was working for at the time (about 100 people, with 3 families that had at least one child diagnosed as “autistic” within a 3 year period after an incident in which shit was smeared all over the walls of the men’s bathroom), was not worthy of being investigated. My resources were limited not only because I was not an epidemiologist, but by virtue of the fact that the H-1b flood, prior to the DotCon bubble burst, managed to hold onto their positions through the purge while the old timers not high enough in the pecking order to be served by toadies and sycophants, were summarily ejected for the hinterlands – which is why my study commenced after I moved to a small town and fixed computers for subsistence.

So, one may well critique my study on the basis that “correlation doesn’t imply causation” – and, indeed, that’s when I discovered with what facility that trope was selectively trotted out. Indeed, that is why I began to become “interested”, shall we say, in exactly how it is that “causation” is inferred from data. That’s when I intuited Algorithmic Information Theory and then discovered that it had been known for 50 years how to properly derive causal models by Ray Solomonoff et al in Algorithmic Information Theory. Hence my 2005 proposal for what became known as “The Hutter Prize”.

You can’t possibly imagine how pissed off I am at the machine learning world, let alone the social pseudoscience world.

At least dynamical systems theory may yet end up inadvertently getting the ML world back on track, thence trapping the social pseudosciences (including epidemiology) into looking at causal models that otherwise would be greeted with all the standard, selectively invoked, statistical critiques.