LAW, CAUSE 349 



functioning in the law. A law may be tentatively described 

 at the empirical level, then, as a statement of a repeated 

 association of events. The examination of this formulation 

 will require a more detailed consideration (1) of the kinds of 

 events which may be correlated, (2) of the nature of the 

 correlation, and (3) of the meaning of "repeated." 



(1) In general, any event whatsoever may be correlated 

 with any other event. The coming of spring is associated 

 with the appearance of robins; the ring around the moon 

 with stormy weather; the excessive size of an elephant with 

 excessive weight; freezing weather with bursting water 

 pipes; drinking arsenic with death; the equality of angles 

 in a triangle with equality of sides; black hair with dark 

 skin; and so on. Events, in other words, tend to gather into 

 clusters, and the clusters tend to repeat. In fact, as was seen 

 in the last chapter, the ordinary conception of "thing'' is 

 precisely such a cluster of events, having a certain unitary 

 character which permits inference from one of its aspects to 

 another. The statement of the correlation of these aspects 

 constitutes a law. Whether it is events or aspects which 

 are correlated is of no importance, since the difference is 

 merely in mode of speech. Strictly speaking, only events are 

 correlated, though when events are united into complexes 

 one tends to describe the elemental events as aspects of the 

 complex event. 



What requires special mention at this point is the fact 

 that on the strictly empirical level the events which are 

 correlated are usually described in qualitative rather than in 

 quantitative terms. It should not be insisted dogmatically, 

 however, that an empirical law cannot be quantitative. 

 Much of the literature of science, for example, distinguishes 

 between two types of quantitative law. On the one hand are 

 empirical laws which are expressed not as functional rela- 

 tions between variables but as tables of crude measured 

 values which seem to exhibit a semblance of proportionate 

 variation, as in the measured values of the spaces covered 

 by a ball rolling down an inclined plane correlated with 



