APPLIED TO HISTORY 233 



the historic data. From this point of view there is a 

 somewhat close analogy between history and other forms 

 of scientific investigation. Given the facts, the scientific 

 historian endeavours to trace the operation of principles 

 in them. But this reflection only leads on to the state- 

 ment of a point in which historic data differ very widely 

 from the data of other sciences, especially the physical 

 sciences. For the purposes of history every fact is 

 unique, i. e. is fettered in place and time : for the 

 purposes of physical science every fact is potentially 

 independent of particular place and time, and every 

 explanation of a fact is potentially of universal applica- 

 tion. Thus if I want to investigate the nature, say, 

 of the sunrise, I start perhaps from some instance which 

 I may have observed : but the explanation, if I find it, 

 is valid of every single instance since the creation of the 

 world, and for all future instances so long as the solar 

 system continues to exist. And any one of these 

 instances would have done as well as any other to 

 serve as the starting-point of my investigations. This 

 is not the case with the other events just mentioned. 

 I cannot infer from the rule that certain political con- 

 ditions tend to produce tyrannicide (supposing, for a 

 moment, I am justified in assuming this rule to be true), 

 even that Hipparchus or Julius Caesar or Alexander II 

 was killed in point of fact : still less that the event 

 occurred on any particular day or in any particular 

 conditions : whereas, if my explanation of to-day's sun- 

 rise is sufficient, I can infer that the sun must have 

 risen at a calculable time on any day in the past history 

 of the world. Such laws as we reach by the investiga- 

 tion of history do not enable us to infer the historic 

 occurrence of any given fact or the conditions of it. 

 They are generalizations rather than laws. Thus facts 



