THE SCIENTIFIC MOOD. 



21 



ment in simpler terms, or in another sense, in more 

 generalised terms; and to effect this analysis is in 

 itself a scientific problem. 



(d) Hypothesis. — There is no doubt that some 

 conclusions have arisen in the mind as if by a flash 

 of insight, but even these have perhaps been due to 

 processes of unconscious cerebration. In the ma- 

 jority of cases, the process is a slower one, the scien- 

 tific imagination devises a possible solution — an hy- 

 pothesis — and the investigator proceeds to test it. In 

 other words, he forges intellectual keys and then tries 

 if they fit the lock. If the hypothesis does not fit, it 

 is rejected and another is made. The scientific 

 workshop is full of discarded keys. 'Nov can it be 

 forgotten that even those conclusions which com- 

 mend themselves at first sight have to submit to the 

 process of testing like those which were tried with 

 less confident fingers. It matters little, except to 

 the logician, whether the hypothesis was reached as 

 an induction from many particulars or as a deduc- 

 tion from some previously established conclusion; 

 in either case the result is a provisional hypothesis, 

 which has then to be tested. 



N^ewton said in his Principia that he did not make 

 hypotheses (Hypotheses non fingo)^ and yet he, like 

 all great scientific workers, certainly did, for in- 

 stance in his corpuscular hypothesis of light, which 

 has turned out to be erroneous. The fact is that 

 there are different kinds of hypotheses, — there are 

 guesses at truth which have no experimental basis, 

 which are usually prompted by some big conclusion 

 dominating the mind of the guesser, such as Sweden- 

 borg^s nebular hypothesis; and there are scientific 

 hypotheses which are more or less carefully con- 

 structed systems, harmonised with existing knowl- 



