Lehrer on Reductionism

by John Derbyshire

Several readers have sent me a link to this Jonah Lehrer article in Wired, presumably in hopes of tempering my too-ardent zeal for scientific inquiry. (Though perhaps, more charitably, because I mentioned a personal encounter with Jonah in my November Diary.)

The kinds of things I’d want to say about the piece are mostly said in the comment thread that follows it (here, for instance). For a good brief survey of issues relating to reductionism, see Lecture 23, “Reduction and Progress,” in Prof. Jeffrey Kasser’s Philosophy of Science series from the Teaching Company.

These issues keep coming up because science is now tackling exceedingly complex systems — the brain, the climate, the genome — all intricately knotted with feedback loops and self-transforming capabilities. It’s now 25 years since we worked out a complete wiring diagram for the brain of C. elegans (a hermaphroditic worm), and we still don’t altogether understand how the darn thing works. That doesn’t mean that reductionism is whack, only that these are really, really hard problems on which progress is slow.

The title of the piece, which is not likely Jonah’s doing, is dumb. The only people “failed” by science are those with unrealistic expectations of it, a category of persons far, far more prevalent outside science than inside it.

The only thing I’d add which nobody else seems to have picked up on, is that to grumble about a naive approach to causality in the hard sciences is straining at gnats while society at large, and our entire political leadership, is gulping down camels by the … whatever the collective noun is for camels.

For real correlation-equals-causation idiocy, the social sciences can’t be beat. Poverty causes crime! Bad schools cause poverty! Parenting shapes personality! Social pressures keep women out of math and science! Racism causes test-score gaps, incarceration disparities, drug addiction, single motherhood, high blood pressure, … ! Compared with this stuff, the physical and biological sciences are paragons of epistemic modesty.

Also worth mentioning every time the correlation-is-not-causation theme comes up is the converse point: While correlation does not necessarily imply causation, total absence of correlation makes it highly unlikely that any causation is going on.