IDEAS. 25 



qua lion to all grades in the development of thought ; for by 

 it alone can we compare idea with idea, and thus reach ever 

 onwards to higher and higher levels, as well as to more and 

 more complex structures of ideation. As to the history of 

 this development we shall have more to say presently. 

 Meanwhile I desire only to remark two things in connection 

 with it. The first is that throughout this history the develop- 

 ment is a dcvclopinent : the faculty of abstraction is every- 

 where the same in kind. And the next thing is that this 

 development is everywhere dependent on the faculty of 

 language. A great deal will require to be said on both these 

 points in subsequent chapters ; but it is needful to state the 

 facts thus early — and they are facts which psychologists of 

 all schools now accept, — in order to render intelligible the 

 next step which I am about to make in my classification of 

 ideas. This step is to distinguish between the faculty of 

 abstraction where it is not dependent upon language, and 

 where it is so dependent. I have just said that the faculty 

 of abstraction is eveiywhei'e the same in kind ; but, as I 

 immediately proceeded to afiirm that the development of 

 abstraction is dependent upon language, I have thus far left 

 the question open whether or not there can be any 

 rudimentary abstraction without language. It is to this 

 question, therefore, that we must next address ourselves. 



compounding of abstractions, "when, comparing a number of objects, we seize on 

 their resemblances ; when we concentrate our attention on these points of 

 similarity. . . . The general notion is thus one which makes us know a quality, 

 property, power, notion, relation, in short, any point of view under which we 

 recognize a plurality of objects as a unity." Thus, there may be abstraction 

 without generalization ; but inasmuch as abstraction has then to do only with 

 particulars, this phase of it is disregarded by most writers on psychology, who 

 therefore employ abstraction and generalization as convertible terms. Mill says, 

 " By abstract I shall always, in Logic proper, mean the opposite oi concrete ; by an 

 abstract name the name of an attribute ; by a concrete name, the name of an 

 object " {Logic, i. § 4). Such limitation, however, is arbitrary — it being the same 

 kind of mental act to "concentrate attention upon a particular object" as it is to 

 do so upon any "particular quality of an object." Of course in this usage Mill is 

 following the schoolmen, and he expressly objects to the change first introduced 

 (apparently) by Locke, and since generally adopted. But it is of little consequence 

 in which of the two senses now exjilained a writer chooses to employ the word 

 "abstract," provided he is consistent in his own usage. 



