LAW, CAUSE 359 



the developments since John Stuart Mill are of particular 

 importance. Mill and his predecessors seem to have believed 

 that universal laws could be logically justified if one could 

 only find the proper premise. This premise was commonly 

 located in a feature of nature which was called its uni- 

 formity. Granted that nature is uniform, then what is true 

 in a few cases will be true in all cases, provided the cases in 

 question are typical. But this is precisely the difficulty; how 

 can one determine whether the cases are usual or unusual? 

 The fact seems to be that nature exhibits disuniformity as 

 well as uniformity, and one cannot tell in any given case 

 which is revealed. What is required, therefore, for the 

 logical deduction of the principle of induction, is not merely 

 the principle of the uniformity of nature but a further 

 principle which will tell whether the cases observed are such 

 as to come under the general assertion of uniformity. 



This need seems to have turned the attention of investi- 

 gators into new directions. The focal center of the problem 

 of induction has shifted, particularly in the present century, 

 from necessity to probability. It is now seen that the prin- 

 ciple of the uniformity of nature, being itself an inductive 

 generalization, can hardly afford a basis for grounding any 

 particular law. Attention is therefore turned to the actual 

 cases, with a view to determining whether they are such as 

 to permit generalization. Progress on the inductive problem 

 can be made if principles can be formulated describing the 

 effect which such evidence as the number of instances of the 

 association and the degree of similarity of the various in- 

 stances has upon the probability of the law. No law is stated 

 to be necessarily true. Every law has a higher or lower 

 degree of probability, and probability is meaningless without 

 reference to the evidence on which it is based. The two most 

 important contributors to the problem in recent years have 

 been Jean Nicod and J. M. Keynes. A brief reference may be 

 made to the essential contribution of each. 1 



1 J. Nicod, Foundations of Geometry and Induction (New York: Harcourt, Brace, 

 1930); J. M. Keynes, Treatise on Probability (London: Macmillan, 1929). 



