....Previous

This argument by Meyers and Stern does not impress us. If a person believes a proposition to be false, then, of course, he cannot use it as an evidence for another proposition. Bui if the proposition which he takes for an evidence turns out to be false the evidential value of the proposition, as we have noted, does not diminish. In this connection we would like to refer to J. A. Hart and R. Dees who offer a counterexample in which they intend to show that there is a flaw in the argument given by Meyers and Stern.³⁶

They declare that Meyers and Stern's use of the term 'justifica­tion' is not in consonance with ordinary usage of the term. We usually say that the ancient sailors were justified in not sailing too far on the basis of their belief that the earth is flat which belief happens to be false. Hence Meyers and Stern's use of the term 'justification' is a technical one. However, this recommendation for technical usage may be accepted if it helps one to clear up the problem posed by Gettier counterexamples. But this account of Meyers and Stern happens to be both over-restrictive and over-inclusive. It is over-inclusive because it cannot rule out the accidental or coincidental cases. Gettier examples arise not because there is the possibility of false evidence, but because there is the possibility that a subject S may have evidence from which a proposition h can be inferred but which is, in fact, unrelated to the truth of h. This makes S's acceptance that h accidental. Principle (B) cannot take care of these cases.

In addition to being over-inclusive Meyers and Stern's, analysis proves to be over-restrictive. The following counterexample makes this evident. A research chemist is making efforts to develop a low-cost method for producing synthetic fuel. After several experiments Wilma, the research chemist obtained the result that a certain mixture of chemicals X and Y will produce Z which is a viable substitute for natural fuel oil. X and Y are very common and inexpensive chemicals and hence this is an important disco­very. Dees and Hart imagine that Wilma has been greatly influ­enced by von Wrong's Logic of Scientific Method. In this book the author sets forth various principles of justification one of which just happens to be identical with Meyers and Stern's principle (B). But one of these principles which is other than principle (B) slates that in order to be justified in believing a generalization, one must have at least 5000 confirmatory instances. Consequently Wilma performs 5000 experiments and five more to be on the safe side. She finally comes to believe that h (which is the proposition that X and Y mixed in proper amounts produce Z) on the basis of p (which is the proposition that she has performed 5005 experiments in each of which X and Y have produced Z).

In the last ten experiments, however, Wilma had by mistake mixed M and N, the rare chemicals that appear chemically identical with X and Y as they arc already known to yield Z. In a hurry to prepare a paper by the given deadline Wilma came to the research laboratory on a holiday when the stock room was closed. She took bottles marked "X" and "Y" from a colleague's desk. The bottles actually contained M and N and not X and Y. Her colleague had forgotten to change the old labels of the bottles. Wilma performed her last ten experiments which corroborated her earlier findings and sent off her results to the journal for publication.

The question now is: does Wilma know that X and Y would produce Z? The answer seems to be yes. But the evidence in this case proves to be false, for she believes wrongly that in all 5005 experiments she has mixed X and Y to produce Z. Thus principle (B) is not satisfied, yet there is knowledge and consequ­ently justification. If she were told that in the last ten cases she has not combined X and Y but M and N she would not concede that the 4995 experiments were enough to prove that X and Y produce Z because of her strong commitment to the von Wrongian principle, viz., that at least 5000 experiments must be conducted before one can arrive at a conclusion. But if she sends a postscript to the editor of the journal telling him that her results arc not correct for she has reached her conclusion on the basis of 4995 experiments only, the editor might not agree with her and would be ready to give her full credit for her experimental findings. In this case she believes h and she believes it on the basis of the experimental evidence. But she does not believe that she is justified because of her von Wrongian attitude. Yet we would think that she had knowledge.

This counterexample employed by Dees and Hart strikes at the very root of Meyers and Stern's attempt to uphold principle (B): If p justifies S for h, then p. In this task, Dees and Hart are also supported by ordinary usage. They emphasise that Gettier counterexamples are generated not as a result of there being false justifiers, but due to the fact that the true belief happens to be true just due to chance or luck. Thus even when false justifiers are absent, Gettier-like counterexamples, may be constructed, as has been attempted by B. Skyrms and A. Goldman.

Finally, Meyers and Stern's attempt to introduce the concept of well-takenness to replace, in some cases, the concept of justification is against ordinary usage. It is natural to understand the distinction between the case of Miss Information and that of Mr. Fancy as between the presence and absence of justification. Miss Information, we would say, has justification for her belief while Mr. Fancy does not have any such justification. There is no reason why we should introduce here a technical concept whose meaning is determined by stipulation.

There is no doubt, therefore, that Meyers and Stern had a wrong thesis. In other words, Gettier's third presupposition stands unassailed. There may be false justifiers. It is not thus true that if p justifies h for S then p. Even if p justifies h for S, p may not be true. The failure to reject any of Gettier's presuppositions seems to boost the morale of the protagonists of the Gettier problem.

Next..