SECTION 23 GENERALISATION. 337 



of facts, and, if we have been successful, we may extend the 

 generalisation in every direction so far as circumstances permit. 

 Thus Kopp, examining the specific gravity of liquids at their 

 boiling point, not only detected regularities among their specific 

 volumes, but also in the boiling points of related substances, 

 the temperature of ebullition, and the character of the com- 

 pounds; to-day the fact that uranium and radium change into 

 other elements, has suggested the subtle question whether all 

 elements are in process of decomposition and whether this is 

 also true of compounds ; the explanation of the inertia of electri- 

 city has suggested that somehow the inertia of matter might 

 have the same cause, that indeed matter may be composed of 

 electrons ; and the fixation of nitrogen by bacteria, may suggest 

 that various other useful substances are fixed by them, as 

 water in an arid climate, and that all such bacteria may be 

 artificially multiplied. The present custom is to generalise 

 when something accidentally suggests this course to be de- 

 sirable. We ought, however, to make it a habit to generalise. 

 It may be that we shall thus reach only a second or a third 

 fact, or one class, or, on the other hand, a very comprehensive 

 generalisation: it does not matter, so long as we have ex- 

 hausted the possibilities. Such a habit will develop our powers 

 of generalising and avoid our neglecting to generalise when 

 we ought to do so, and it will also prevent our reasoning to 

 a second fact alone or only to the nearest class. This process 

 presupposes substantially original, definite, and scientifically 

 arrived at statements respecting facts or classes ; because, to 

 attempt to generalise, for example, every statement in every 

 article or book which crosses our path, would be folly and 

 would lead to disgust of generalising. It also assumes habitual 

 resort to verification, without which the process proposed is 

 destitute of sense and value. 



We may now offer a few illustrations in regard to methodical 

 procedure. In connection with a series of particulars relating 

 to the senses, I deliberately seek for something to generalise. 

 Noting a reference to the relative rapidity of one sense, I 

 determine to discover rapidity of apprehension in all the senses. 

 After this, I methodically arrive at the term Time, the most 

 comprehensive class to which rapidity belongs, and decide to 

 render definite all the propositions so far as they relate to 

 Time. Continuing the process of extension, [ pass from Time 

 to the other aspects named in our second part of the table of 

 Primary Categories, and successively endeavour to utilise these. 

 The limit of extension is then reached regarding the problem 

 which we set out to examine. Again, I observe in my notes 

 that some sense is assisted by some other sense. I amplify 

 this into the question: "How far is each of the senses assisted 

 by each of the other senses?" And, after studying this question 

 with a view to extending it to the furthest bounds, I obtain, 



22 



