568 THE ROYAL SOCIETY OF CANADA 



test is that, in actual intercourse, if a man makes statements conflicting 

 and yet true, he can, on being questioned, give a reasonable explana- 

 tion from his knowledge of the attendant circumstances. The explan- 

 ation, however, must be in accordance with these actual circumstances, 

 i.e., truthful. Harmonization of conflicting statements in records is 

 attempted, on the contrary, only where (and because) the attendant 

 circumstances are unkyiown. Were these circumstances known, it 

 would be clear of itself whether the conflicting statements are both 

 correct, or one of them erroneous. The harmonizer, therefore, at- 

 tempts his harmonization by a pure effort of the imagination, without 

 knowledge of the actual circumstances, and thus necessarily without 

 regard to them. If such a test is proper and legitimate, then a person 

 making conflicting statements in actual intercourse may also properly 

 and legitimately account for them without regard to the actual cir- 

 cumstances, i.e., by any plausible, fictitious explanation that his 

 imagination can suggest. Once this initial weakness in the present 

 test is realized, it becomes clear that the practice is not really defen- 

 sible. In the above mentioned paper, moreover, its invalidity was 

 definitely shown, because, where statements conflicting and yet true 

 were located in actual intercourse, and were recorded without mention 

 of the attendant circumstances, a reasonable harmonization of the 

 statements seemed impossible in nearly half of the cases, and in the 

 remainder a reasonable harmonization seemed easy, but proved in 

 every instance to be contrary to the facts. This demonstrates that, 

 where the attendant circumstances are unknown, (1) the impossibility 

 of suggesting a reasonable harmonization does not prove that the 

 discrepancy cannot be reconciled in fact, and (2) that such harmoniza- 

 tions, even where possible, have no scientific value in themselves. 



(&), Results of Prevailing Use of Reasoned Probability as a 

 Positive Historical Criterion. 



1. Decision by authority. Where an investigator, such as an 

 historian, makes an investigation, there are fundamentally but two 

 grounds on which others can accept his reported results. The first 

 ground is proof by his established findings ( = necessary conclusions), 

 the second, his opinion in points not established by his findings 

 ( = probabilities). A reader or other person using an historical nar- 

 rative cannot ordinarily repeat the investigation, and therefore such a 

 person must depend largely, if not entirely, on the historian to deter- 

 mine what points are established by his findings as necessary conclu- 

 sions. If, however, the historian, in making this determination, uses 

 reasoned probability purely as a resisting, negative force to overthrow. 



