258 DEVELOPMENT AND PURPOSE CHAP. 



criticism of categories, because it requires the systematic 

 test of experience. But to obtain proof we must go a step 

 further, and frankly base our beliefs upon experience itself. 

 But simple as this sounds, and familiar as the method is in 

 the trivial operations of every-day life, to carry it through 

 as a theory of knowledge, and to make experience as a 

 whole the basis of our view of reality as a whole, is the 

 most complex of all tasks, requiring the maximum of self- 

 criticism in the use of the method, and open at many points 

 to the charge of paradox and self-contradiction. If, 

 indeed, as some of the critics of the hypothetical method 

 have supposed, the object of science were only to describe 

 what we see, the theoretical difficulty would disappear. 

 But if its business is to generalise and infer, be it only to 

 the past and future of our experience, the case is quite 

 altered. Such inference, we have admitted, must be based 

 on a measure of insight into the real causal processes 

 whereby things are determined. But, to assume for the 

 moment that experience gives us reality, how are we to 

 know that it gives us enough of the reality for this pur- 

 pose? Consider only the relativity of perception. By 

 means of the microscope we know enough now to be sure 

 of the negative truth that the causes of zymotic diseases 

 could not be discovered by any analysis or synthesis of data 

 yielded by the unassisted senses. What reason have we 

 to think that the larger scope afforded by the microscope 

 will carry us any further in the way of ultimate laws ? If 

 we rely on observation we never observe the whole of any 

 phenomenon, and there is always the possibility that what 

 is necessary for our purpose resides wholly or in part in the 

 processes which are unobservable. We are brought back, 

 in short, to the initial difficulty, that the world we can touch 

 and see is but a fragment. The results of real processes 

 are visible therein, but we cannot assume that the process 

 3S a whole comes within our limits. We may be able con- 

 ceptually to construct a reality which would yield our 

 results, and this, in fact, is what the inverse method 

 attempts. But to invert the process again and make the 

 results the basis of the construction is a much harder task. 

 It can be fulfilled only if we can answer the two questions 



