570 THE ROYAL SOCIETY OF CANADA 



bolder spirits who wish to know the reason why, or these decisions 

 are overthrown completely by additions of evidence. Additions of 

 evidence usually bring a like fate to any consensus of opinions resting 

 on balanced probable conclusions or other mere probabilities; and as 

 often as not, this addition of evidence is developed at the moment 

 when the wideness of the consensus makes the humiliation of the ex- 

 posure most complete. Such a consensus of opinions is liable also 

 to mortifications of another sort. Where a leading scholar, or 

 several of them, have built up, in the face of an opposing group, a 

 large following of lesser lights, such a leader may suddenly reverse his 

 opinion on a cardinal point, accepting the view that he formerly re- 

 jected, and rejecting the view that he formerly accepted. This 

 volte face need not be due to increased evidence, but may occur at any 

 time and simply because he chooses erratically to balance his prob- 

 abilities a little differently. By his new friends, this addition to their 

 ranks, though it adds not a jot scientifically to the strength of their 

 case, will be greeted with acclaim; by the old, his apostasy will be 

 viewed with vexation and regret; but to onlookers generally, and 

 especially to thoughtful men, such kaleidoscopic changes, occurring 

 from time to time, suggest that where coats are turned so easily, one 

 side must be about as good as the other, and neither side can 

 be worth any too much. Under such conditions a general and in- 

 creasing scepticism is a healthful development: it marks a stage on 

 the road from worse to better; it is a recognition that a radical defect 

 exists; a challenge to the science involved to locate and remedy the 

 evil. 



3. Logical Inconsistencies. A feature of the present method is 

 its logical inconsistencies. A general instance of these is the attitude 

 towards non-contemporaneous narratives in any period where they 

 are the only sources available. Professing to regard trustworthiness 

 as a sine qua non of trustworthiness, the prevailing method ought 

 to strike every such record and period from its reckoning. It refuses 

 to make the sacrifice. Instead it repeats the entire narrative, with a 

 question mark or a multiplicity of these. A scientific historian has 

 no right to narrate untrustworthy history, either with or without 

 question marks. Another instance of these logical inconsistencies in the 

 the present method is its frequent combination of "possibilities" with 

 "probabilities." This combination is so common in Bernheim and 

 elsewhere, that, in defining a "probability," he feels it necessary to 

 join with it a definition of a "possibility" (a fact against which there 

 is no direct or indirect evidence, and no positive evidence in its favor). 

 Now if probability is to be a positive historical criterion and prob- 

 abilities accordingly are to be included in history, improbability is the 



