506 THE ROYAL SOCIETY OF CANADA 



engagements. But if the cannon-pinion of such a watch becomes too 

 loose, the time-piece may stop for a half an hour or an hour and then 

 start of itself and run correctly for several weeks before a similar 

 lapse occurs. Thus the owner may suddenly find the watch wound up 

 and apparently going properly, yet fully a half an hour or more behind 

 time; and though, if the hands be reset, it may keep time accurately 

 for several weeks without a similar occurrence, the owner should not, 

 after one such lapse or failure, and he will not, after several of them, 

 be governed by the time-piece in any important engagement until 

 the watch ( = the process) has been adjusted to prevent similar 

 failures. The principle here illustrated corresponds to the 4th method- 

 ic principle, and among the millions of watches in use, the disorder in 

 question, though one of the less frequent, is a common and daily 

 occurrence; hence a principle concerning correct mechanisms corres- 

 ponding to the 4th methodic principle is shown to be, like the 4th 

 fundamental principle, in continual operation. 



II. EXPERIMENTAL AND THEORETICAL TEST OF PROB- 

 ABILITY AS A CRITERION OF CONCLUSIONS IN 

 HISTORICAL SCIENCE. 



What is Probability ? 



In any standard dictionary a definition of probability will be 

 found in the following or equivalent terms: "having more evidence for 

 than against"; "supported by evidence which inclines the mind to 

 belief but leaves some room for doubt;" "having more than half the 

 chances favorable;" "a preponderance of argument on one side, in- 

 clining the mind to receive it as the truth, but leaving some room for 

 doubt." This is the ordinary meaning of probability as a term in 

 general use. It is also the ordinary meaning of probability in his- 

 torical science. Bernheim, a supporter of the prevailing historical 

 method, says, (page 201): {Translation) ("In historical investigation 

 we call a fact probable if the reports or other reasons for believing that 

 it occurred outweigh the reports or other reasons for believing that it 

 did not occur, although its non-occurrence remains always a possibil- 

 ity." 



In standard dictionaries probability is defined further in a speci- 

 fic logical and mathematical sense as the "amount of rational confidence 

 with which a contingent event may be expected or a problematic 

 statement may be accepted as true. In this sense an event has some 

 degree of probability if there be any chances at all in its favor, even 

 though the majority of the chances may be against it." "Were 5 



