242 SCIENTIFIC METHOD AS 



That is, we cannot for the purposes of scientific reasoning 

 begin from a universal doubt. But this does not necessarily 

 imply that no investigation has gone to the formation of 

 the premisses whatever they are. One science will, 

 frequently, take from another the premisses with which 

 it proposes to begin. So, though the interlocking of the 

 various interests is, in the present case, more intricate 

 than usual, we may perhaps assume that historical science 

 will take over from archaeology or from palaeography 

 some fixed starting-points in the way of dated documents. 

 The historian will collect and examine (as Langlois and 

 Seignebos have pointed out) the whole mass of documents 

 concerned with his period : he will find out exactly what 

 they say : he will reject manifest contradictions and 

 anachronisms. Incidentally, by this process a provisional 

 picture will grow in his mind of the history of the period : 

 he will begin to understand the coherence and the drift 

 of it : he will be getting into the frame of mind in which 

 he can bring his conception of the period to bear upon 

 particular statements of individual authorities. All this 

 process is described in considerable detail by MM. Lang- 

 lois and Seignebos. But I venture to think that this is, 

 logically speaking, prior to the asking of the real historical 

 question : Can I believe this historical statement, and why? 

 What really has to be settled when this question is raised 

 at last is the value of the authority, and this is a 

 particularly difficult question to answer. All we can ever 

 do in regard to it is to reach the result : ' The statement 

 or statements of such and such an authority may be 

 trusted.' In regard to this point, I venture to differ in 

 some degree from the two authors to whom I have 

 referred : I should attach more value than they seem to 

 do to the character of the authority. The great source 

 of the difficulty of history altogether is the presence of 



