VERIFICATORY TECHNIQUES 225 



is sometimes primarily a search for forms, and at other 

 times primarily a verification of previously suggested forms. 

 Both movements are essential to science. 



A similar distinction may be made between experiments 

 for discovery and experiments for verification. As was 

 pointed out in Chapter VI, the more usual sorts of experi- 

 ment belong in the latter class. The essential difference 

 between the two is that experiments for discovery are 

 unguided, while experiments for verification are guided. The 

 guiding factor in the latter case is a preconception as to 

 the nature of the outcome. In the inductive experiments 

 there is no foreseen knowledge of the result; manipulations 

 are made in a more or less random way, and the results of 

 the changes are noted and recorded. But in verificatory 

 experiments there is an anticipation of the result; it is 

 predicted that a change of a specified character should 

 bring about another change of a specified character. In 

 the former case one is merely learning from nature; in the 

 latter he is, in a sense, legislating for nature, since he is 

 using his acquired knowledge of her behavior in other realms 

 to predict her character in a given realm. He is saying to 

 nature, in effect, "If you have not been deceiving me you 

 must behave in such-and-such manner in the present sit- 

 uation." Verificatory experiments are therefore much more 

 complex than inductive experiments, for one's previous ex- 

 perience with nature suggests myriads of ways in which 

 changes may be introduced into it. In pure observation 

 one merely listens to nature, in -experiments for discovery 

 he prods nature to make it speak, in verificatory experi- 

 ments he asks nature questions — in fact, he asks questions 

 which are definitely leading, and he is justly surprised when 

 he receives negative answers. Consequently verificatory 

 experiments are much more decisive than inductive ex- 

 periments. 



The distinction which is commonly made between ob- 

 servational and experimental verification is largely relative. 

 For example, it may be maintained that the proposition 



