A-EVIDENCE AND EVIDENTIAL PARTICULARISM

is true. Now, suppose that given that John ate contaminated fish, it is very probable that John got sick, and, more im­portant, it is very probable that he got sick because he ate contaminated fish.

(2) e is (potential) evidence that h.

In this case, e is evidence for hypothesis h, even though h does not entail e (nor does e entail h). Also, in this case, e is evidence for h by itself—not for h in conjunction with other hypotheses that together with h entail e. So, using the con­cept of (potential) A-evidence, together with the proposed idea of “correct explanation,” we obtain a version of eviden­tial particularism.

According to the contrasting deductive-explanatory ac­count of evidence, e above would be evidence that h only if h entails e. It doesn't. In this case, we want the reverse inference from e to h. So let us exchange h and e, as follows:

e': John got sick.

h': John ate contaminated fish.

Suppose we want to claim that e' is evidence that h'. On the de­ductive explanatory account of evidence, to get e' deductively from h', we need to add various assumptions to h'. For example:

(3) John ate fish that was unrefrigerated for 6 days and fish under these conditions accumulate lots of bacteria.

(4) Whenever one eats fish on which lots of bacteria have accumulated for 6 days, the fish becomes contaminated and one who eats it gets sick.

Proposition e' is entailed by h' together with (3) and (4), not by h' alone. So, on the deductive-explanatory view, e' is not evidence for h' by itself, but for the “system” of hypotheses consisting of h' and (3) and (4). It is this system, not h' by itself, that entails e'.

On the account of evidence and explanation I have proposed, whether e' is evidence that h' depends on whether, given e', it is probable that there is an explanatory connec­tion between h' and e'. This would be satisfied if, given e', it is probable that the reason John got sick is that he ate contaminated fish. The claim that it is probable is an empir­ical one, not an a priori one. And, like any empirical claim, evidential or otherwise, it can be defended by appeal to other empirical claims—e.g., by appeal to claims such as (3) and (4). The justificatory claims (3) and (4), however, do not need to be considered part of the empirical claim that e' is evi­dence that h. On the deductive-explanatory account of ev­idence, however, they must be so considered, since e' is not deductively entailed by h' alone. So e' cannot be considered evidence for h' alone.

In view of the difficulties with a deductive-explanatory view of evidence, which are avoided by my own explana­tory view, I prefer the latter, which supports evidential par­ticularism. The holist, however, will not agree that I have eliminated holism simply by a definition, whatever its advantages. The holist will say that I have simply shifted the problem to one of justification. To this I now turn.

9.