102 THE SCIENCE OF LOGIC. 



But, while this implication or non-implication of existence, in the definition, 

 is perfectly intelligible and in itself all-important, it may well be doubted whether 

 it can form the basis for a clear division of definitions into two distinct classes ; 

 and also whether it does mark the actual line of distinction between definitions 

 that are set forth as nominal and others that are set forth as real. It scarcely 

 can furnish such a basis ; for all definition even the most distinctly nominal 

 carries with it a reference to reality from which language cannot be divorced, and 

 by which it must be constantly moulded and directed ; l while no definition, 

 even the most decisively real, carries with it, of itself, a proof, but only a more 

 or less strong presumption, of the existence of the thing defined. 2 And no 

 matter how strong the implication of such a presumption may be, its presence 

 &quot; cannot be collected from the mere form of the expression &quot;. 3 Accordingly, 

 some logicians would be inclined &quot; to finally discard this distinction from 

 logic &quot; on the ground that &quot; the province of logical definition is not to verify, or 

 to disprove, this presumption [of the existence of the thing defined] &quot;. 4 



It is, we believe, permissible to retain the distinction : regard 

 ing as nominal those definitions whose primary aim is obviously 

 (a) to fix the connotation we attach to the terms we use, without 

 raising any possibly contentious question as to whether, or in 



Y Cf. WELTON, op. cit., p. 120 ; VENN, op. cit., p. 274-6. 



2 &quot; My own view,&quot; writes Dr. Venn, &quot; is that we shall do best if we rid Logic alto 

 gether of this distinction between real and nominal definition. The best general 

 account of the matter would then be this. All language, as a broad primd facie pre 

 sumption, carries with it the implication that the speaker believes in the reality . . . 

 of the things corresponding to the words which he uses. People do not speak with 

 an intention to mislead, nor do ordinary adults talk habitually of non-entities. The 

 mere use of a word, therefore, raises the presumption of the . . . reality of the objects 

 answering to the word. . . . 



&quot; We may say then that every definition . . . raises a presumption of the . . . reality 

 of the objects to which it refers. But it certainly ought not to claim more than to 

 raise such a presumption (unless, of course, reality is formally incorporated into it). 

 Nor indeed is it easy to see how a definition could intimate such a claim. . . . 



&quot; A definition is merely the interpretation of a name. In and by itself it has no 

 warrant to convey one kind of reality rather than another, nor has it any known 

 means of doing so. It stands in fact on precisely the same footing in this respect as 

 a term or name. If one of these is uttered we have to judge, by the context of the 

 subject-matter, to what order of existences it belongs, and we must do the same in 

 the case of definitions. . . . 



&quot; It is, therefore, strictly correct to refer the Definition to the name, that is, to 

 make all Definitions nominal, provided we do, what we have expressly undertaken to 

 do, viz. refer the name to the thing. The name contains the limited group of 

 attributes which always is, or in careful thought should be, present to the mind. 

 This |is subjective or conventional; and is all that we can possibly undertake to 

 expound in any formal kind of science such as logic. The logical Definition, there 

 fore, confines itself to this analysis. But it can only safely trust itself to do so, if 

 it is to be an instrument for inductive research and judgment, so long as we always 

 strenuously assert that the names have an archetype behind them.&quot; VENN, op. cit., 

 pp. 275-8. We have italicized the passages which lay down the conditions on which 

 it may be lawful to regard all definitions as nominal. 



3 MILL, ibid. 4 WELTON, ibid. 



