An Introduction to a Biology 



And the second point is this. Supposing you start by 

 postulating the existence of a phenomenon A, and in order 

 to find the cause of it cast about for some other phenomenon 

 with which it is associated. Suppose you find this other 

 phenomenon B ; if A is usually but not invariably associated 

 with B, you will be fairly satisfied, and will conclude that A 

 probably exists and that B is probably the cause of A. But 

 if A is invariably associated with B your satisfaction will be 

 complete ; you will conclude that you have not only estab- 

 lished the existence of A but determined the cause of it B. 

 A little reflection about the toad, however, will show you 

 that you are wrong. Complete association, instead of in- 

 creasing your sureness, should diminish it. For it 'shows 

 that there may be an alternative interpretation namely, 

 that there is no such thing as A, but that what you thought 

 was A was really B. So you must make some other test ; 

 if possible a crucial one ; and then, maybe, you will dis- 

 cover that there really was no such thing as A, but that 

 it was B all the time as in my case. 



Biologists who aim at interpreting things may be classed 

 under three headings. The classes are arranged in descending 

 order of bulk and in ascending order of merit. Class I. con- 

 tains those who do not bother about interpretation at all. 

 They do it, but they never stop to wonder what they are 

 doing. Class II. includes those who bother about it, but 

 can get no further than accepting the view of one or other 

 authority as to what they are actually doing. Class III. 

 the smallest and the best of all contains those who have 

 the desire and the intelligence to find out for themselves what 

 they really are doing when they are interpreting nature. 



You will please conclude, therefore, that what I want 

 you to do is not to accept my interpretation of inter- 

 pretation, but to think it out for yourselves. By such 

 means alone will you be able to reduce to a minimum the 

 errors which attend the process of interpretation. 



We have discussed the sources of error in description 



R *57 



