244 SCIENTIFIC METHOD AS 



actual experience and the dramatic adoption of the tone 

 of an eye-witness. And we shall test our authors as far 

 as may be in connexion with statements of fact of which 

 there is the fullest independent evidence. For this will 

 enable us most clearly to distinguish the psychological 

 characteristics of a given author, and attain some notion 

 of his general credibility. If all goes well, we may arrive 

 at last at a picture of the events concerned in which the 

 various strains of evidence fit in together, or at least 

 define themselves with sufficient clearness to be compared 

 and weighed against one another ; and in which the 

 contents of their evidence form something like a rational 

 and coherent whole. If to this can be added the fact 

 that further investigation down the line of history more 

 and more obviously runs back upon a view of the world 

 such as we have derived from our estimate of the 

 evidence, we shall have the best reason that can be 

 given us for accepting the statements of our authorities 

 as true, or for trusting those whose statements build up 

 the coherent picture, and distrusting, on the whole, those 

 which are eccentric, or which have already been dis- 

 credited under examination. 



I have struggled vigorously to set out my views of 

 these processes in abstract form, avoiding, as far as 

 possible, any side-glances at particular instances, be- 

 cause it appears to me that in this way the process of 

 reasoning becomes clearest. If I have been right in my 

 exposition, it follows that there are, as it were, two ele- 

 ments in historical belief: the acceptance of statements 

 on the authority of witnesses, in^wh^rrTwe have confi- 

 dence, and the partial verification of these statements by 

 comparison with others, on the assumption that in ordin- 

 ary cases the chances of misplacement of confidence will 

 gradually be eliminated if a sufficient variety of testimony 



