8o REMARKS ON CERTAIN [XL 



veniently distinguished as ideal and real. Practically it is found 

 that they are seldom sharply discriminated ; often both kinds 

 occur combined in one and the same theory : nevertheless they 

 should be clearly distinguished. The ideal theory seeks to 

 explain phenomena by any arbitrarily chosen principle, quite 

 apart from the question whether the principle has any actual 

 existence or not \ The ideal theory only seeks to show that 

 there are h3^potheses on which the phenomena in question are 

 explicable. Real theories however are not content with 

 plausible hypotheses, but endeavour to include only those 

 which possess some degree of probability : they attempt to 

 give not merely a formal solution, but, if possible, the correct 

 one. Sir William Thomson has attempted to explain the dis- 

 persion of rays of light, by imagining the existence of molecules 

 which are composed of concentric hollow spheres, arranged one 

 inside the other and connected together by springs. But this 

 distinguished physicist never for a moment believed in the 

 existence of real molecules, provided with springs ; he wished 

 to show that existing conceptions were capable of rendering 

 intelligible the phenomena of dispersion. Obviously Darwin's 

 pangenesis was conceived in this spirit, and was therefore called 

 by him ' provisional ' ; although in later life he may have come to 

 attach real worth to the theory. I consider the gemmules to 

 be a deliberate invention, like Sir William Thomson's mole- 

 cules provided with springs, which have no claim to reality : 

 the gemmules merely serve to show the sort of suppositions 

 we must make in order to understand the phenomena of 

 heredity. 



Ideal theories are by no means useless. They are the first 

 and often the indispensable steps which we must take on our 

 way to the understanding of complex phenomena. They form 

 the foundation upon which real theories can gradually be raised. 

 Above all, they supply the impulse to re-examine again and 

 again the phenomena they attempt to explain. I should pro- 

 bably never have been led to deny the inheritance of acquired 

 characters, if Darwin's pangenesis had not shown me that 

 the belief in such transmission involved an assumption so 



^ The two philosophers Herbart and Lotze have named these two 

 types of theory '■fiction ' and ' hypothesis ' : the former term agrees with 

 ideal in expressing the consciousness of unreahty. 



