AS A MENTAL OPERATION 23 



cations. Sometimes it is induction and deduction of 

 facts, and sometimes also of causes. Sometimes it is 

 deduction from facts to causes, and sometimes deduction 

 from causes to facts. Sometimes it is hypothetical, but 

 not so often as is nowadays believed. The facts, again, 

 may precede the hypothesis advanced to explain them ; 

 or the hypothesis may be advanced to deduce conse- 

 quences to be afterwards verified. It may be remarked 

 also that there is a consilience of induction and deduc- 

 tion; and that of two kinds. We may find a fact 

 empirically and then deduce the same fact ; or we may 

 deduce it and then find it empirically to be a fact. As 

 Aristotle said, reason seems to confirm phenomena, and 

 phenomena reasoning. Moreover, as he also said, the 

 very power of deducing many consequences from a 

 premiss confirms it by a kind of reaction. Everything 

 proves that induction and deduction are inextricably 

 interlaced in science : witness the conspicuous example 

 of Newton first inducing 'axioms', next deducing con- 

 clusions from, but beyond experience, and then using 

 those conclusions as though they were particular in- 

 stances to induce the universal gravitation of ponderable 

 bodies. 



So various and subtle is this mixed method that it 

 is difficult to classify it. It can hardly be confined within 

 Mill's explanatory and verificatory methods, at any rate 

 in the narrow sphere within which he restricted the 

 former, and with the hypothetical turn which he gave 

 to both methods. But are not explanation and verifi- 

 cation, even at their widest extent, merely varieties of 

 the mixed method of induction and deduction, always 

 interacting and also confirming one another in many 

 intricate ways ? In conclusion, the lesson of all scientific 

 method, and especially of mixed method, is this : start 



