Skepticism and the Probability of Nonbasic Statements (I): On Sufficient Conditions for Absolute Probabilities

Abstract

The two tasks that confront us now are to examine the claims that no nonbasic statements are initially acceptable (premise III.1), and that none are probable relative to basic reports (premise III.2). Although it may seem natural to consider first whether any nonbasic empirical statements are initially acceptable, I shall, instead, begin with the second claim, because points developed in evaluating it will also prove helpful in evaluating the first claim. It should be noted, incidentally, that a nonbasic statement that is probable relative to basic-reports may be initially acceptable rather than inferentially acceptable. What is important for distinguishing the two species of acceptability is that, as definition D2 states, initial acceptability of p requires, and inferential acceptability of p forbids, that p would be acceptable even if no other statement should justify it sufficiently to be acceptable. So if we discover that some nonbasic sentences are probable relative to basic-reports that are certain for someone at some time, we will not yet know whether they are initially or noninferentially acceptable. But, I would claim, we would know something more important; we would know that they are acceptable. Of course, we may find that no nonbasic statements would be made inferentially acceptable by basic-reports, but, nevertheless, some might be initially acceptable and others inferentially acceptable because of them.