The paper is somewhat analogous to the second paper by Goldman that we read. It modifies an account of knowledge in an earlier paper (by Lehrer alone) and does so in a way that introduces a concept that has continued to be used in accounts of knowledge. Although the defeasibility theories of which this paper provides the first example are not the only alternatives to the tradition of reliabilism initiated by the second paper by Goldman, reliabilism and defeasibility are probably the most popular ways of responding to the Gettier problem.
• After an introductory paragraph, section I presents an analysis of non-inferential knowledge, what Lehrer and Paxson call “basic knowledge.” Notice that they leave open the question of what makes such beliefs justified, and much will be left open regarding justification throughout the paper.
• Section II, which presents an account of non-basic knowledge, is the heart of the paper. The actual analysis of knowledge is given early on (p. 227); but it employs the concept of defeasibility, and the rest of section is devoted to working out an account of this concept (which is completed to the authors’ satisfaction only on p. 231).
• Sections III and IV offer criticisms of alternative accounts of knowledge. I have assigned only section IV since the accounts criticized in section III are less important for our purposes and would probably be ignored in our discussion in any case. In thinking about section IV, you should pay most attention to the discussion of Goldman’s causal theory; but Lehrer and Paxson consider only his causal account of perceptual knowledge, so you should think also about how he well he could deal with the examples they devise in section II and how well their defeasibility account could handle the examples he considers in both of the papers you have read.