260 DEVELOPMENT AND PURPOSE CHAP. 



elements together and draw therefrom an inference as to 

 the world beyond my immediate consciousness. To 

 understand why we can make this assumption let us pass 

 to the second class of generalisations, (b) We can gener- 

 alise any relation of terms if we know that the one holds of, 

 i.e. is related to the other c as such.' This is the assump- 

 tion underlying the syllogism, which compels us, if we 

 argue from A to B in a given case, i.e. if we take A as a 

 sufficient ground for the assertion of B, to admit the major 

 premiss any A is related to B in the specific manner stated. 

 What is true of a term as such is true of it universally, is 

 the general postulate of critical reasoning. But what does 

 c as such ' mean ? Whatever else it may mean, it includes 

 this negative, that the relation which holds between two 

 terms as such is not dependent on any third term. It 

 follows that any given relation of two terms not dependent 

 on a third term is universal, and it results further that if 

 any relation between two terms A-B is given in experience 

 only two alternatives are possible. One is that the rela- 

 tion holds of one or other of the terms as such. The other 

 is that it is dependent on a third term. To distinguish 

 the two cases is the problem of scientific induction. Lastly, 

 it will be seen that another way of stating the same assump- 

 tion is that every element of experience is related to some 

 other element of experience as such. 



If B is true of A as such, A is called the ground and B 

 the consequent. There is nothing in the propositions laid 

 down to show that the same consequent may not have 

 several grounds. Thus A-B may be universal while B-A 

 is not so. But if B-A holds in one case and not in another, 

 it must be because there is a third term C in the first case 

 which changes to D in the second. Hence the full conse- 

 quent of A is not B but BC, and the relation A-(BC) is 

 convertible, while some other ground E of B has as its full 

 consequent BF. Different grounds, therefore, have as 

 such different consequents, though these different conse- 

 quents may happen to agree in some point. These assump- 

 tions, all thus traceable to a single root, can be shown to 

 apply to the whole of the functions of thought, and to be 

 the sole assumption made in thinking, (i) They include 



