< previous page page_158 next page >

Page 158
At this point I make a conjecture: I theorize a Rule for which that sack contains beans and all the beans in the sack are white, and I try to consider the Result that I have before my eyes as a Case of that Rule. If all the beans in the sack are white and these beans come from that sack, it's natural that the beans on the table are white.
Peirce observes that the reasoning for Abduction is typical of all "revolutionary" scientific discoveries. Kepler learns from those who preceded him that the planets' orbits are circular. Then he observes Mars in two different positions and reveals that they touch two points (x and y) that cannot be two points of a circle. This Case is strange. It would no longer be strange if one were to assume that the planets trace an orbit that can be represented by another type of curve (noncircular). Kepler must, therefore, find a different rule. He could imagine that the orbits are parabolic or sinusoidal. . . . It doesn't matter to us (in this paper) why he thinks of an ellipse (he has his good reasons). Thus he makes his Abduction: if the orbits of the planets were elliptical and the two positions of Mars revealed (x and y) were a Case of this rule, the Result would no longer be surprising. Naturally, at this point he must test his Abduction by a new Deduction. If the orbits are elliptical (if at least the orbit of Mars is elliptical) one must wait for Mars at a point z, which is another point of the ellipse. Kepler waits for it, and finds it. In principle, the Abduction is proven. Now one must simply test and prove that the hypothesis cannot be negated, Naturally, I abbreviated and summarized the phases of the discovery. The fact is that the scientist does not need ten thousand inductive tests. He makes one hypothesis, perhaps daring, very similar to a bet, and he puts it to the test. As long as the test gives positive results, he has won.
Now, a detective does not proceed any differently. Rereading the declarations of method by Sherlock Holmes, one discovers that, when he (and with him Corian Doyle) talks of Deduction and Observation, in effect he's thinking about an inference similar to Peirce's Abduction. 4
It's strange that Peirce used such a term as "abduction." He formulated it in analogy with Deduction and Induction (and also in reference to some Aristotelian terms). But we cannot forget that in English "abduction" also means kidnapping. If I have a strange Result in a field of phenomena not yet studied, I cannot look for a Rule in that field (if there were and if I did not know it, the phenomenon would not be strange). I must go and "abduct," or "borrow,'' a Rule from elsewhere. You could say that I must reason by analogy.
Let's reconsider the Abduction about the white beans. I find a fistful of beans on the table. On the table is a sack. What tells me that I need to

 
< previous page page_158 next page >