Roderick Chisholm suggests that he and other epistemologists have been arguing backwards. If knowledge is justified true belief, then the concept of justification should not be explained by such other concepts as ground, basis, evidence, and probability: rather, these concepts should be explicated by reference to a proper understanding of justification.

A belief may be justified in several senses, but in epistemology the relevant question is: Is the belief reasonable? If we are able to define 'reasonable' as thus used we will then have the basis for defining other terms of epistemological interest, including 'knowledge'.

Chisholm approaches 'reasonable' by concentrating on its comparative form. It is reasonable for someone to accept a proposition if accepting it is more reasonable than accepting its contrary (i.e., if it is probable), or if accepting it is more reasonable than not accepting it (i.e., if it is beyond reasonable doubt). More importantly for his analysis, he defines a proposition as 'evident' when accepting it is more reasonable than not accepting some other proposition [which we believe to be reasonable].

The next step is to analyze what propositions may be said to confer or transmit probability to other propositions. Chisholm holds that one proposition tends to confirm another if two conditions are met: first, that it is possible for the second proposition to appear probable, and second, that if the relation between the two propositions is such that the first is evident to someone, then the second is either probable, or else, beyond reasonable doubt, not probable. He allows. therefore, that confirming evidence can be disregarded if the conclusion is preposterous. The same formula applies to 'tends to make evident' as to 'tends to confirm'. One proposition tends to make another evident if the second could appear evident to someone and if, the first being evident, the second also becomes evident (unless it is undoubtedly not evident for other reasons). In the absence of all such other reasons, we omit the qualifier and say that one proposition does make the other evident.

As a final preliminary qualifying point, Chisholm goes on to take into account cases of non-knowledge that seem to be true belief. Following E. Gettier's lead, he acknowledges that there are some cases of 'defectively evident' propositions. He defines a proposition as defectively evident when it is made evident by something else, but at the same time every thing that makes it evident also makes evident something false.

With all this behind us, we can now define 'epistemic justification'. A proposition is justified if it is evident for us (in the sense defined above) and, if it is defectively evident, then it is implied by propositions each of which is evident (but not defectively evident).

We know something, therefore, when it is true, when we accept it, and when we are justified in accepting it.

This still leaves problems unsolved. That a person knows does not entail that he knows that he knows. Furthermore, true beliefs may be unjustified, and justified beliefs may be untrue. The best we can do is to strive to believe reasonably and to avoid believing unreasonably. But there is no guarantee that we therefore attain truth.

Chisholm adds that the idea of 'external' justification can be defined simply in this way: S is externally justified in accepting a proposition if the proposition is true and S is a thinking subject. Other attempted definitions have made external justification only a subspecies of internal justification.

Contents

Index