Theory and Evidence: The Development of Scientific Reasoning

Rating:
Summary: Profound Defense of Humans' Inductive Capacities
Review: It seems that two groups of cognitive scientists are moving in opposite directions: straight cognitive psychologists routinely show that adults can't reason themselves out of a bag, whereas developmental cognitive psychologists tell us that we were born scientists. If you think there's something goofy about this, you're right. And, as Barbara Koswolski shows, what's goofy is cognitive psychologists' model of science--i.e., logical positivism, the notion (inter alia) that scientific induction begins by checking our prior knowledge and theories at the lab door.

Following this Victorian notion of science, cognitive psychologists typically test adults' ability to find covariation by presenting them with tasks that--like logical positivism--fail to represent the features that scientists really face. That is, the tasks rule out the use of background knowledge; any causes of the variation to be identified is already known to be present among a clearly defined set of alternatives; all these potential causes are *stipulated* to be equally possible; and any hypotheses made about covariation must be rejected *in toto* rather than modified. In other words, these canned science experiments are about as representative of science as those cheesy tasks you find in "Hands-On Science" museums.

Faced with these tasks, adults ignore whether variables are confounded, refuse to gather information exclusively about covariation, fail to search for disconfirming evidence when testing hypotheses, cling to theories in the face of "disconfirming" evidence, ignore base-rate information, and are swayed by illusory correlations. Based on this portrait of adults, it is a wonder that we've managed to function at all, let alone land on the moon.

But, as Koslowski shows in a series of 16 elegant and ingenious experiments, the problem is not in ourselves but in our tasks. How adults and adolescents reason about covariation is perfectly sensible from a scientific perspective--they earnestly look for causal mechanisms. And by doing so, they don't look like logical positivists in search of Humean indices. That is, they ignore theoretically and causally trivial covariations and confounds, but look out for unexpected causal mechanisms, and they consider whether variation can be equally explained by rival alternative accounts (even for their own pet hypotheses).

To be sure, Koslowski doesn't whitewash adults' wrong-headed causal theories (like the notion that being an extrovert causes people to eschew library science). But she does find great reason for hope in even American adolescents' scientific capacities, and what's more--she delivers a devastating critique of positivism, a ringing endorsement of "theory-laden" science, and a profound defense of real human rationality.

One last bit of praise: Koslowski underscores her thesis stylistically--she treats the reader as if he is rational. She answers questions as they arise, formulates conclusions so they neither say too much nor too little, and always treats the positions of other researchers fairly. With so many popular books in the cognitive sciences these days, I hope that her book reaches an audience that isn't quite so used to these virtues.