[bowman] fundamental PROCESSES IN HISTORICAL SCIENCE 571 



reverse of this criterion and therefore improbabilities ought to be 

 excluded. But an improbability, though it has fewer favorable chan- 

 ces than a probability has more favorable chances than a possibility. 

 Thus, if there be 100 favorable chances and only three possible events, 

 80 chances favorable to one event would constitute it a probability 

 (in the ordinary sense), 19 chances favorable to another event would 

 constitute it an improbability (in the ordinary sense), and 1 chance 

 favorable to the third event would constitute it a possibility. Upon 

 what reasonable ground, then, can a historian (1) narrate probabilities, 

 (2) exclude improbabilities, and (3) include possibilities which have 

 fewer favorable chances (and therefore are less likely to be correct) 

 than the improbabilities which he has excluded ? 



4. Abnormal developments. An error, not recognized as such, 

 tends to increasing extremes. Thus we find in a well-known work by 

 a regius professor of modern history in an English university the 

 statement: "The circumstances, if they were known to us, though 

 they could never excuse such a proceeding, might perhaps partially 

 palliate it." In other words, it is no longer sufficient for the historian 

 to suggest what he deems to be probable on the evidence that he has, 

 but we are to have conjectures upon evidence which he has not. A 

 professedly scientific work, in referring to gaps in our records for a 

 certain period of history, remarks that the sources themselves are so 

 meagre that "a use of conjecture, or of what has been called the his- 

 torical imagination, is inevitable." Under the given conditions, what 

 is the matter with silence? To a person who realizes that "one of 

 the most essential characteristics of all science is that it supplies as- 

 sured knowledge," it will seem scarcely possible that a professedly 

 scientific scholar could have seriously made this suggestion to use the 

 imagination not merely as a positive criterion of conclusions but even as 

 a substitute for evidence itself. The statement, however, which is by 

 a Scottish professor, was made in a serious connection, and is seriously 

 intended. Again, a brief summary of an historical question in a stand- 

 ard work of reference some decades ago, contained but 4 probable 

 statements; but in a similar summary of the same question in a recent 

 work of the same character, the number of probabilities, possibilities, 

 presumptions, and conjectures, has grown to 11. In this same recent 

 standard work, the number of probable statements, veiled and open, 

 averaged, in 50 pages, approximately 5 per 1,000 words of text; and 

 in more specifically historical parts, this quota rose at points to as 

 high as 8 and even 15 per 1,000 words. Thucydides is accounted 

 generally, even by moderns, as the ablest and best historian in his own, 

 and every other, age; and Caesar, not without reason, is regarded 

 from the purely secular stand-point as the greatest intellect that the 



