THE METAPHYSICS OF SCIENCE. 369 



induction from observation, then admitting (contrary to 

 the fact) that the same absolute certainty could be reached, 

 the very process of generalization assumes still the same 

 principle, that homogeneity of phenomena implies simi- 

 larity of cause. Hence, when we look to general induction 

 for the validity of the principle that like results proceed 

 from like causes, we find that the induction itself assumes 

 beforehand the validity of the principle; and our effort 

 is simply a case of reasoning in a circle. As general in- 

 duction cannot, therefore, validate the principle which 

 validates general induction, it follows that the principle 

 is validated either by deduction or by the direct sanction 

 of pure reason. But it is not a deductive conclusion, for 

 the principle itself, possessing the highest degree of gen- 

 erality, is not the result of an analysis. We discover no 

 account of the validation of the principle except in the 

 sanction of the same rational authority as speaks to us in 

 affirming a correlation between certain conscious states 

 and external realities. Here, then, in the first step which 

 science takes in formulating a general concept or scientific 

 doctrine, it is absolutely necessary to rely on the universal 

 validity of a principle which cannot be established by sci- 

 entific processes, nor indeed by any formal logic whatever. 



So, it may be added, the whole search after general 

 laws, or the unification of human knowledge, is prompted 

 and guaranteed by the intuitive conviction that unity 

 exists among the diversified phenomena of nature. If no 

 ulterior unity existed, or if reason were not furnished 

 with the knowledge of its existence, the search for general 

 laws and deeper causes would never be undertaken, or, 

 if undertaken, would be fruitless. 



It is extremely easy for the scientific investigator to 

 24' 



