Darwin against the philosophers - Thinking About Science

According to David Hull, "evolutionary theory seems capable of offending almost everyone." Readers of this column will surely be familiar with the sort of offense that has come from creationist quarters ever since the publication of Darwin's Origin of Species. Of course, some of Darwin's contemporary colleagues also raised scientific objections to the theories of common descent and natural selection. But such objections were quickly overcome, and evolution was soon considered as controversial in biology as quantum mechanics is in modern physics.

However, reading Hull's book Darwin and His Critics, one discovers that an entirely different sort of attack caused quite a bit of trouble for Darwin and his fledgling theory: he got himself caught in the middle of the great induction debate in the then-nascent field of philosophy of science. And he seemed to be on the losing side. I will tell the story in somewhat less technical terms than Hull's, though his book is a must read for anybody interested not only in this episode, but more generally in the early reception of the theory of evolution by Victorian intellectuals.

The great induction debate unfolded between two of the major British philosophers of Darwin's time, John Smart Mill and William Whewell. Both were attempting to improve on Francis Bacon's notion of induction. Bacon, by many accounts the originator of what today we call philosophy of science, wanted to establish the scientific method on firmer foundations than what he thought (erroneously, it turns out) Aristotle had done. The latter was known (and still is) for his deductive method, which he exemplified with famous syllogisms such as: (premise 1) All men are mortal; (premise 2) Socrates is a man; therefore (conclusion) Socrates is mortal.

The problem with deduction, as Bacon and the players in the induction debate saw it, was that it arrived at specific conclusions starting from an unjustified leap to generalizations. Bacon therefore proposed that a better method was the inductive one, by which he meant the cautious construction of general conclusions based on an increasing number of observations. But, it was quickly realized, this wouldn't do, since it amounted to an unwarranted extrapolation from known to unknown, with no independent assurance that the new observations would emerge from the same causes as the old ones (see "Thinking about Science" in SI, May/June 2003).

Both Mill and Whewell tried to do better than Bacon. Mill thought that scientists could use two kinds of induction, which would mutually reinforce each other. Enumerative induction was a strengthened version of Bacon's, where the process of extrapolation was made sound by the Law of Universal Causation (i.e., the idea that all phenomena have causes, and that it is parsimonious to attribute similar causes to similar phenomena). Eliminative induction was an operation by which the causes of natural phenomena were discovered by successive elimination of unsuitable alternatives, based on tests as stringent as could be devised.

Whewell, on the other hand, thought that scientific investigation has to start with hypotheses, which have heuristic value in guiding one's inquiry. When a hypothesis is confronted with the data, one knows if it is true because of the feeling of "consilience" (we would say congruence) generated by the match between data and hypothesis. Whewell also retained the term "induction" for his view, probably because of the patina of prestige lent by Bacon.

There were problems with both views. Mill did not like the idea of judging the truth of hypotheses based on how well they matched the data, because of the well-known problem that more than one hypothesis could fit the observations equally well. On the other hand, neither enumerative nor eliminative induction are logical guarantors of truth, because they are themselves based on assumptions about the world. (It turns out that there is no logical guarantor of truth about nature, if one does not count mathematical proofs; scientists just have to do with very likely maybes.)

What does have all of this to do with Darwin? The problem is that Mill and Whewell--regardless of their internal disagreements--saw Darwin's work in Origin as an egregious example of deduction. Darwin had, based on a few observations, jumped to the generalization of natural selection, and had then deduced a series of consequences, which he spent twenty years trying to back up with observations. This, apparently, did not qualify as induction, and was therefore unscientific. As Hull puts it, the philosophers were "equally adamant in their conviction that the Origin of Species was just one mass of conjecture."

It turns out, however, that the philosophers were wrong and the scientist on the mark. Ironically, an analysis of Darwin's work (and of how he himself describes it) reveals that his method was actually very close to the one proposed by Whewell (though in clear opposition to Mill's). In Darwin's own words: "How odd it is that anyone should not see that all observation must be for or against some view if it is to be of any service!" That is, Darwin was applying Whewell's method of consilience, or inference to the best explanation, though he probably did not buy into the philosophical basis (idealistic rationalism) that inspired Whewell. Was this a case in which philosophy slowed down the progress of science? Or did Darwin's awareness of the philosophical issues surrounding his research make him more careful and, ultimately, more successful? To the historian of science it is a nontrivial task to figure this one out.