96 WHAT IS SCIENCE? 



not mechanical ; it does not trace an analogy between the 

 production of such beings and the operation of a piece 

 of mechanism. Can we find any feature that is common 

 to all theories that have proved to be true, or must we 

 (as in the case of laws) rest content with several distinct 

 but well-defined types which have all proved successful 

 and yet display no common characteristic ? 



I think we can find such a feature. The explanation 

 offered by a theory (that is to say, the part of the theory 

 which does not depend simply on the deduction from it 

 of the laws to be explained) is always based on an analogy, 

 and the system with which an analogy is traced is always 

 one of which the laws are known ; it is always one of 

 those systems which form part of that external world 

 of which the subject-matter of science consists. The 

 theory always explains laws by showing that if we 

 imagine that the system to which those laws apply 

 consists in some way of other systems to which some 

 other known laws apply, then the laws can be deduced 

 from the theory. Thus our theory of gases explains 

 the laws of gases on the analogy of a system subject 

 to dynamical laws. The theory of evolution explains 

 the laws involved in the assertion that there are such- 

 and-such living beings by supposing that these living 

 beings are the descendants of others whose characters 

 have been modified by reaction to their surroundings 

 in a manner which is described by laws applicable to 

 living beings at the present day. Again the immense 

 theory involved in the whole science of geology explains 

 the structure of the earth as it exists to-day by supposing 

 that this structure is the result of the age-long operation 

 of influences, the action of which is described by laws 

 observable in modern conditions. Q In each case the 

 " explaining " system is supposed to operate according 

 to known laws, but it is not a system of which those 

 laws can be asserted as laws, because it is, by the very 



