Thursday, November 08, 2007

The Bell Curve and Logical Positivism

Two more today.

Continuing on in the line of favorite former students, bkriplur asks,

What do you think of Herrnstein & Murray's The Bell Curve? You mentioned it in one of your earlier posts and I was curious about your opinion on it.
It's a good thing you asked me, bkriplur, because according to Charles Murray a couple weeks ago, those of us from the Jewish community are innately smarter.

The Bell Curve is a massive book designed to provide support for an updated version of Herbert Spencer's social Darwinism. The idea is that intelligence as measured by IQ is a functional fact that determines one's ability to contribute to the society, that IQ is both (a) fixed and immutable and (b) a heritable property, and that there is a significant evolutionarily derived difference in mean IQ across different subpopulations, so that certain groups of people are smarter than others and therefore inherently better contributors to society. As such, the distribution of wealth and social power that we see is a socio-biologically explainable (and therefore morally irrelevant) fact and social programs designed to raise the poor will necessarily fail because it would be throwing money at people who simply are genetically inferior, designed so that they cannot as a group contribute to society as well as other groups.

One must avoid the temptation to poison the well and simply argue that the political agenda behind the argument undermines it. There is no doubt that it informs it and guides it, but the question is whether there is independent warrant to believe it.

Turns out there is significant trouble at virtually every major point in the argument. The idea to begin with that IQ is a meaningful property, much less one that is one-dimensional and measurable like weight, is deeply problematic. Let's start with the idea that there is one kind of smart. Surely some people are innately more gifted in certain sorts of intellectual tasks. Some people pick up languages like nobody's business, others have complex problems solved before their mind's eye by just looking at them, yet others are capable of deep insights that cut to the foundations of our beliefs, and others still are creative in ways that boggle the mind. But most who excel in one sort of intellectual endeavor are hardly superior in the others. Insightful critics are often terrible artists, those who are incredible big picture folks tend to be sloppy with details, while those who are wizards with the nitty gritty are frequently impatient with analyses of foundational conceptual concerns. Which of these, then, ought we privilege and try to measure as IQ? All are factors that allow for positive social contribution, which one alone do we mean by intelligence?

Even if we could answer that question, the problem of immutability and measurement both create problems. Can't one increase their abilities to more quickly acquire understanding with work and training? To use an analogy, I am hardly a naturally gifted athlete, yet I figured out ways to compete at the level of NCAA division I athletics, playing respectably against the best players of my generation (some days more than others...). In terms of "Athletic Quotient," I was far below those against whom I was competing, yet I was able, through work and training, to contribute on the lacrosse field. Why shouldn't the same be true of those who need to work harder to develop problem-solving, creative, or conceptual skills?

Measurement is yet another classic black hole. For a wonderful history of the social biases in measuring native intelligence read Stephen Jay Gould's The Mismeasure of Man, one of the best pop science books ever written.

Then we come to the questions of heritability of significantly differing IQ levels (if we take IQ to be meaningful for the sake of argument). Is intelligence or social properties like aggressiveness something that is inherited or socially influenced. One of the temptations we frequently fall into these days is to conflate the notions of biological and genetic. Not everything that is biologically explainable is the result of genetic determination or even predisposition. It is certainly true that behavior is a result of brain activity and brain activity is a function in part of neuroanatomical factors, some of which are innate. But the brain is shaped by all sorts of environmental factors, starting in utero (think of fetal alcohol syndrome in which the mother's drinking will effect the intellectual capacities of the child). The brain is such a complex thing so deeply effected by environmental stimuli that separating out the genetic from the environmental may not be possible, much less meaningful.

And then there's the claim that they are not only heritable properties, but that differences between subgroups exist and are the result of evolutionary pressures. This claim is problematic for a number of reasons. Evolutionary change is a slow gradual process, especially when it comes to complex, subtle things like the working of the brain. Humans are slow to reproduce and the division of the "races" is a fairly recent occurrence in our history. There is good reason to think that even if the claim that these differences exist is meaningful that there has been sufficient time for natural or sexual selection to account for them. If you compare social reasons (human behaviors can be changed incredibly rapidly by social factors alone as anyone in marketing knows) with biological ones, the former gives a much more likely explanation.

So, The Bell Curve is, like Ayn Rand's objectivism, a flawed attempt to justify in the minds of smart privileged white people why they shouldn't feel guilty about the inequities of society and would be wrong to lift a finger to help those who need it.

SteveD asks,
What are the differences between the logical positivists and the logical empiricists?
The logical positivists were a subset of the logical empiricists. The distinction is meant to differentiate what Moritz Schlick, Rudolf Carnap and gang were doing in Vienna from the broader contemporary movement of scientific philosophy that included the goings on in Warsaw, Prague, France, Scandinavia, and especially Berlin.

Logical Positivism can be sketched as having four central pillars. (1) A criterion of cognitive significance, that is, a rule that separates sentences that mean something from those that sound like they do, but don't, (2) The synthetic/analytic distinction, that is a rule that separates the meaningful sentences into those whose truth or falsity can be shown by nothing more than formal logic and those that require observation, (3) A philosophy of math/logic to show how to tell the true analytic sentences from the false ones that followed from the logicist approach of Russell and Whitehead, and (4) a philosophy of science to show how to tell the true synthetic sentences from the false ones that allowed for the reduction of complex empirical claims to logical combinations of simple observation reports.

If we take Hans Reichenbach in the early to mid 20s as an example of a Logical Empiricist who was not a Logical Positivist, he bought into (1) thinking that we could meaningfully distinguish the real questions from the pseudoquestions, but his conventionalism posited a much more slippery line between analytic and synthetic sentences, not an absolute partition, his philosophy of mathematics was much more in the Hilbert formalist camp than the Russellian logicist one, and his philosophy of science made observation reports theory-dependent in an interesting neo-Kantian way that is clearly not there in Carnap. Subtle differences, but there and show why Reichenbach, while a major figure in Logical Empiricism was not a Logical Positivist.

Further, once problems arose for all four pillars -- (1) they never succeeded in actually formulating a functional criterion of cognitive significance, (2) Quine's argument in "Two Dogmas of Empiricism" undermines the notion of a strict meaningful synthetic/analytic distinction (3) Godel's Incompleteness theorems show the sort of complete and sound formal criterion for sorting true from false mathematical and meta-mathematical statements impossible, and (4) the problems of induction and evidence (e.g., Hume's, Goodman's, and Hempel's) give fits to the attempts to formalize scientific inference in the way Carnap and company were trying -- the movement still persisted attempting to figure out ways around the problems. It was therefore a continuation of the project at large without the strict positivist commitments, so it needed a different broader name.

Fun ones tomorrow...