Harman’s paper begins our discussion of confirmation, which will focus on the confirmation of theories. Harman himself does not discuss theories, and his examples, to the extent that they concern confirmation in science, seem closest to the confirmation of empirical generalizations. My excuse for including the paper is, first of all, the importance of the idea of inference to the best explanation for discussions of confirmation generally and, secondly, the fact that the close association of theories and explanation means that inference to the best explanation may be the conception of confirmation that is most easily applied to theories.

Harman’s paper is divided into two sections. The first concerns “enumerative induction,” which is the sort of induction that Russell had in mind. Since this section will be the one we focus on the most, it would be a good idea to look back at Russell’s discussion (see the handout for 9/4, especially paragraphs 6.15-6.16) to freshen your memory of Russell’s account for comparison to Harman. (If you want to get a sense of how someone like Russell might respond to Harman, look at Russell’s ¶6.14.)

Although we may not give it much attention, Harman’s second section is quite interesting. The issues regarding knowledge raised in the short paper by Gettier that is referred to in Harman’s note 4 were quite new when Harman wrote this, but they have come to dominate philosophical discussion of knowledge for the 50 years since Gettier’s work was published. The Gettier problem applies more directly to ordinary talk of knowledge than to scientific knowledge, so Harman’s approach to it is of interest to us mainly for filling out his discussion in section I. In that regard, the key idea in the second section is the notion of intermediate “lemmas.”