574 THE ROYAL SOCIETY OF CANADA 



No custody, however remarkable in itself, is improper if it be 

 proved to have actually had a legitimate origin, or if there be proof of 

 actual circumstances which would make it a place where the record 

 might naturally be. In this respect, as in all others, proof by correct 

 processes — formal probability — has full right of way, first and last, 

 against all other grounds. 



II. Is the record genuine? 



This point should be determined, wherever possible, by specific 

 proof, i.e., by examination of all available internal and external evi- 

 dence. This examination, however, should be made on the basis of 

 correct processes, not by reasoned probability used as a positive cri- 

 terion. If the available evidence so examined will not establish a 

 necessary conclusion for or against genuineness, then (a) improper 

 custody constitutes prima facie ground for rejecting the record as not 

 genuine, and (b) proper custody constitutes prima facie ground for 

 acceptance as genuine. The scientific reason for (b), i.e., for accepting 

 a record in proper custody as genuine, unless the contrary be proved, 

 constitutes another case (like that for accepting the statement of two 

 independent trustworthy records in agreement against the contradic- 

 tion of a third, independent, trustworthy record, discussed on p. 551 

 above) in which pure probability supplies a scientific basis for a rule of 

 method applicable to constantly recurring situations in historical 

 research. There are cases not a few, and some of them remarkable 

 enough, of the fabrication of records and documents; but for one such 

 fabricated product there are hundreds of the genuine, hence, apart 

 from all evidence whatsoever, the pure probability that any one record 

 or document is genuine is more than 99/100: and since the occasions 

 for applying this rule constantly recur in historical investigation, the 

 investigator who applies it systematically will develop in this series of 

 applications (apart altogether from the first and invariable test by 

 internal and external evidence wherever this will afford a necessary 

 conclusion) an average of correct results higher than the average of 

 essential correctness required in historical science, whether this re- 

 quired average be 24/25, or 49/50, or 99/100. 



Genuineness and fabrication here are not a matter of trustworthi- 

 ness or untrustworthiness, but simply the question whether, e.g., 

 a note of hand or bank note in actual circulation represents an actual 

 debt or is only a forgery. Thus, historically speaking, an exceedingly 

 untrustworthy letter or pamphlet may be the genuine product of an 

 author, acknowledged or anonymous, i.e., it may be an honest but 

 extremely prejudiced presentation of a subject, or even, in the extrem- 

 est sense of untrustworthiness, a genuine effort to deceive. 



