i EXPERIENCE AND REALITY 241 



in theory could yield results which might be made 

 correct within any assignable limit of error. Thirdly, in 

 vindicating itself against the criticism of its theoretical 

 basis, mathematical analysis advances beyond the discrete 

 treatment, and renders the continuous without error or 

 inaccuracy. Analysis when pushed through corrects its 

 own deficiencies. 



These results may be stated generally. A method may 

 be sound for certain purposes though not for others. It 

 may yield a partial appreciation of reality which is just, 

 though it cannot be applied to a final interpretation of 

 reality without contradiction. Thus, methods which 

 enable us to determine that a ball will hit a target, may be 

 vitiated with contradictions if we apply them to interpret 

 the nature of motion. They are founded on certain aspects 

 of motion to the disregard of others. But, secondly, when 

 the flaw is detected, thought is not necessarily helpless. 

 On the contrary, the disclosure of a contradiction is a 

 stimulus to new efforts to overcome it. Thought then 

 at any stage may give us certain facets of reality, and may 

 yet be required to reconstruct its methods in order to deal 

 with other facets, and a fortiori with reality as a whole. 

 It is certain that if we are to grasp space and time as 

 wholes our conception of them must undergo a modifica- 

 tion. Without pretending to say in what direction that 

 modification lies, we may revert to an old suggestion in 

 order to illustrate the manner in which it might be effected 

 without destroying the accuracy of our ordinary reasoning. 

 Suppose, in accordance with this image, that space is such 

 that straight lines, simply because they are drawn in space, 

 have an exceedingly minute curvature. It is clear that our 

 calculations, based on the assumption of their straightness, 

 might be accurate within the limits of observable error to 

 indefinite extent. They would only not be absolutely 

 accurate, and only when the'ir inaccuracy became important 

 would serious error arise. Suppose, in corresponding 

 fashion, that time, instead of being uniform, has, in reality, 

 an exceedingly small amount of difference affecting its 

 passage as such. Inferences involving the indifference of 

 time would not be affected unless we were considering 



Q 



