106 PART II. SOME IMPORTANT METHODOLOGICAL TERMS. 



trated in the gigantic collection of facts pertaining to topo- 

 graphy, the charting of the seas, and weather lore. 



43. We must touch upon two other aspects. The uni- 

 formity of nature is a fundamental assumption of science. 

 When, therefore, we have some pure nitrogen before us in a 

 flask, we are bound to assert that whatever holds good of this 

 sample in the given conditions will necessarily hold good of 

 all pure nitrogen under identical conditions. In instances of this 

 character, as in the problems of mechanics and chemistry, no 

 question of generalising arises, for he who would assume that 

 a particular body in motion would, but for friction and other 

 opposing forces such as gravitation, continue for ever in motion, 

 without implicitly positing that this was true of all bodies, 

 would appear to be necessarily confused in his thought. 1 When, 

 however, we venture on an assertion concerning a complex 

 matter, say a man's conduct, it is patent that unreasoned gene- 

 ralising in regard to that matter and the species to which it 

 belongs on the mistaken assumption that one member of a 

 class exactly resembles all its companions is unwarranted. 

 Undoubted truths are, therefore, as yet very limited, and we 

 should think of most of our generalisations as contingent rather 

 than as necessarily true. 



This brings us to our second point. As we have just seen, 

 in the case of a chemical element or a primary mechanical 

 property, the existing simplicity renders elaborate generalising 

 superfluous and void of meaning. At the other end of the scale, 

 we discern such diversity that we cannot speak of classes in 

 the ordinary sense of the word. To reason that since Washing- 

 ton, Paris, London, Berlin, and Rome are capitals, therefore 

 their general architectural style is identical, would be to fly in 

 the face of the facts. Still, two landscapes may be much alike, 

 in which case we form the two into a class; or a certain 

 number of species resemble one another markedly, and we form 

 them into a genus. In other words, if we extend deliberately, 

 then all extension is generalisation, and no form of extension 

 should be deemed to fall outside the process of generalising. 



44. In science generally the number of facts e.g., in 

 biology is so prodigious that simple enumeration is commonly 

 precluded. Nevertheless, in many directions complete or perfect 

 inductions, to use the antiquated expression for generalisations, 

 may be obtained, and there they are in place. The present 

 writer was once, as a visitor, in a school-room, and asked the 



1 Animal bodies, on account of their instability and their possible self- 

 determined motions, raise a doubt as to the legitimacy of the generalisation ; 

 and if even the elements should be undergoing incessant change, and electri- 

 cal and other influences should be all-pervasive, the doubt would be uni- 

 versally extended in respect of this proposition. We assume, however, that 

 any such internal changes or external influences are absent, and that we 

 discount them if they are present. 



