[bowman] fundamental PROCESSES IN HISTORICAL SCIENCE 561 



IV. THE ORIGIN AND RESULTS OF THE PREVAILING 



USE OF REASONED PROBABILITY AS A POSITIVE 



HISTORICAL CRITERION. 



(a). Origin. 



L Improper extension of proper function. The leading part 

 played by reasoned probability in the process of investigation and 

 accumulation of evidence obscures the fact that, when this process 

 is complete and all available evidence has been gathered, its rôle should 

 be reversed, and, instead of forming the positive basis for final con- 

 clusions, it should be used only as a negative force to test whether the 

 conclusions which then seem to be scientifically necessary are actually 

 free of reasonable doubt. At previous stages of the investigation, 

 conclusions that were not scientifically necessary, but only reasonably 

 probable, were frequently overthrown by subsequent additions of 

 evidence, and there is no ground to suppose that the same would not 

 occur to many conclusions which seem reasonably probable at the 

 final stage, if further additions of evidence should in fact become avail- 

 able; hence experience and correct scientific principles alike require 

 that they should be left in abeyance and excluded from the final, 

 published, results. 



2. Necessity versus certainty in historical conclusions. A failure 

 to distinguish between necessity and certainty in conclusions is another 

 source of the prevailing use of reasoned probability. If one suggests 

 that the conclusions of historical science should be necessary and not 

 merely probable, one of the first points that one may expect to be 

 raised is, "If certainty is required in history, will there be much of 

 history left ?" In science and in actual intercourse, however, certainty 

 and necessity are not thus interchangeable terms. A conclusion may 

 be necessary because it is certain, but it may also be necessary, i.e., 

 require one's assent as a reasonable man, because it is free of reasonable 

 doubt; and ordinarily, as was shown in the introduction to the present 

 part of this paper, scientific conclusions even in the exact science of 

 mathematics are necessary, not because they are certainties, but 

 because they are prima facie correct, and as such, apart from counter- 

 vailing evidence, free of reasonable doubt. Certainty, indeed, could 

 be established, if need be, for a considerable percentage of conclusions 

 in history. The easiest way for an historian to do this is to abbreviate 

 his narrative till it contains only the broadest features of the events 

 under narration; and Lt would not be impossible to devise further 

 special rules to the same end. The resulting narrative would not be 

 wholly devoid of interest, but the sacrifice of material involved is 

 neither desirable nor necessary. The substitution of scientifically 



Set. I and II. 1<^15— 37 



