Abstract

Certain examples involving negatively relevant evidence are trouble for reliabilists, since they show that reliability is not sufficient for justification. Two approaches for dealing with these examples within the reliabilist framework have been taken. Neither approach, however, can account for all cases involving nre. This I will argue. I will explain the two approaches briefly, then describe a counter example which calls for a difference approach. To handle the case I describe, one needs torequire that the agent's belief be ‘internally acceptable’ as well as reliable.Many examples show that reliability is insufficient. Consider a modification of an example by Goldman, which he used to make a different point. Driving through the country, Andy looks to the right, sees a barn, and comes to believe that the object he sees is a barn. His belief, however, is not justified because he possesses nre, viz., his false, but justified, belief that the majority of bam-like objects in the area are barn facsimiles which, under the present visual circumstances, are almost indistinguishable from real barns.