THE LAWS OF SCIENCE 55 



sense of a pair of events occurring in an invariable 

 sequence) is a relation of peculiar significance to scientific 

 laws which are based on experiment. Its significance 

 for such laws is very much less than is generally believed. 

 n fact, it would hardly be too much to say that science 

 seeks to avoid entirely the necessity of recognizing such 

 causal relations, even when it is dealing with events 

 which actually do occur in invariable sequences^} Con- 

 sider, for example, a body falling to the ground. Each 

 position of the body invariably follows those higher up 

 and precedes those lower down. We might describe 

 the motion by saying that each higher position is the 

 cause of the lower positions and that the lower positions 

 are the effects of the higher. But actually we do not 

 adopt such a description. We regard the passage of 

 the body through the whole sequence of positions as a 

 single process which is not to be analysed at all ; it is 

 something which, as a whole, may have a cause (such 

 as the presence of the earth which attracts the body) 

 or an effect (such as the noise finally produced by its 

 impact), but in the process itself cause and effect are| * 

 not involved. This elimination of the causal relation, j ( 

 and its replacement by the conception of a naturally 1 { 

 occurring process, is characteristic of all the more \ 

 advanced sciences. 



But, if the relation whirfi laws ^sfa.'h]^ between events 

 or properties or other things is not that of cause and 

 effect, what is it ? That is a -very interesting question, 

 but it is too difficult and needs too much detailed know- 

 ledge of science for any attempt to be made .here to 

 answer it. I think there are many slightly different 

 relations characteristic of laws ; the differences are 

 important and suggestive ; but they all agree in the 

 feature on which such stress has been laid already. They 

 may all be described as various forms of " invariable 

 association"""; and it is because they are all characterized 



