564 THE ROYAL SOCIETY OF CANADA 



product of correct processes — apart from countervailing evidence — 

 excludes reasonable doubt; and in this connection, it is of interest to 

 note that, of the three above dictionaries identifying moral certainty 

 with a high degree of probability, the Standard offers no illustrative 

 citation of the sort of probability which establishes such certainty and 

 thus excludes reasonable doubt; Webster cites, in illustration, "as, 

 there is a moral certainty of his guilt;" and the Century, "as, there is 

 a moral certainty that the sun will rise to-morrow." In the courts, 

 moral certainty of guilt is established by circumstantial evidence, in 

 which reasoned probability is used, not as a positive ground for con- 

 viction, but as a negative force, which seeks to prevent a conviction by 

 explaining otherwise, if possible, the circumstances pointing to the 

 guilt of the accused; and if this attempt at a reasonable counter- 

 explanation fails, reasoned probability thereby assists in the exclusion 

 of reasonable doubt and in establishing as a moral certainty the guilt 

 of the accused, because this failure shows that the established circum- 

 stances are inconsistent with any other reasonable supposition. 

 Webster's citation therefore points to the correct use of probability 

 for the purpose in question, while the illustrative citation in the 

 Century Dictionary is only an unusually strong case of formal prob- 

 ability; the daily rising of the sun is the result of a natural process, 

 against which a failure has never been registered, and until such a 

 failure is registered, this daily occurrence is to be anticipated as the 

 necessary product of this natural and automatically correct process. 



5. Strength of a probability of 2:1. The prevailing regard for 

 reasoned probability as a positive historical criterion, and the accept- 

 ance of balanced probable conclusions on even a small margin of 

 favorable reasons over unfavorable, is due largely to the view that a 

 probability of 2:1 establishes a remarkably strong case for a conclu- 

 sion. This view in turn is based on the idea that a statement by one 

 trustworthy record or witness, if confirmed by another trustworthy 

 record or witness, is therefore twice as probable, and that where 

 a statement by two independent trustworthy records (or witnesses) is 

 accepted over a contradiction by a third independent trustworthy 

 record, this manifestly proper acceptance rests on a probability of only 

 2:1. This popular view, however, is a scientific error. The probabil- 

 ity that two independent records in agreement are correct is not the 

 sum, but the product, of their separate probabilities; and as shown 

 in the previous illustration of the function of pure probability in his- 

 torical science (p. 552), if the probability that any one of three in- 

 dependent records is correct be taken as 49/50 or 49:1, the acceptance 

 of the two records in agreement, as against the third, contradicting 

 record, is based on a probability, not of 2:1, but 51:1. If the average 



