Fundamental Principles of Scientific Inquiry. 379 



In particular, it often happens that common-sense state- 

 ments, including scientific statements, are so vague that the 

 problem o£ deciding whether they are self-contradictory or 

 not is one of no small difficulty. This is especially true of 

 beliefs about the external world of everyday life — the world 

 of chairs and tables beloved of philosophers, the world of 

 polished door-knobs and coloured glass windows. The only 

 possible guide is the analysis of that large aggregate of 

 individual judgments that we call common-sense. Those 

 that survive have to be re-stated in more abstract terms. 

 It is of no use to object that we must define what we mean 

 by common-sense. That is our starting-point. We can 

 either leave it undefined, as we suggest, or we can attempt 

 to define it in terms of some other notion. There is no 

 third possibility. It is a matter of great difficulty to find 

 any more fundamental concept in terms of w T hich to define 

 it. As the notion is in general well understood, we shall 

 not attempt to define it and shall take it as a starting-point 

 in our theories. 



Special care must always be taken that any notion 

 we leave undefined shall be recognizable by some one or 

 other of our faculties : and, further, that any method 

 suggested shall be capable of being carried out. It will bo- 

 ot no relevance to the question of scientific validity that 

 our results are logically coherent and intrinsically plausible, 

 if we cannot recognize the entities to which they are 

 supposed to apply or if we cannot carry out the processes 

 suggested as means of adding to our knowledge. 



TJie Scientific Status of Simplicity. 



A large fraction of known physical laws are expressible 

 in simple mathematical forms, and we have already indicated 

 the need for some examination of the kind of advantage that 

 simplicity confers on a physical law. Obviously the simple 

 law is always the most convenient to work with in theoretical 

 investigations ; and often one is adopted, for this reason 

 alone, in place of a known law that is more complex, but 

 more accurate, when further development of the con- 

 sequences is contemplated. But this is far from being 

 the only reason for adopting such laws ; though several 

 eminent writers have maintained that it is *. It must he 

 realized that there are cases where a law of supreme 



* Bertrand Russell, ' Mysticism and Logic,' p. -04: E. Macfc, c La 

 Oonnaissance et l'Erreur,' p. 37o : Karl Tearson, ' Grammar of Science/ 

 p. 96. 



