n8 A N T I E N T M E T A P H Y S I C S. Book If. 



But the found of language, however mufical it may be, and plea- 

 fant to the ear, is only fubfervient to the purpofe for which lan- 

 guage was intended ; and that is the expreffion of cur ideas, and 

 the con^munication of them to one another. A language of art, 

 and which is proper for the invention and cuhivation of other arts 

 amd of fciences, nuift exprefs all the things in the heavens or the 

 earth, of which we have any notion or conception. We muft give 

 names to accidents as well as to fubftances, alfo to the feveral rela- 

 tion« of things to one another ; and we muft exprefs what they fl<5? 

 and what the fuffer. Thefe things are of number infinite. But 

 fcience here fets bounds to infinity, as we have (hown it does in the 

 matter of our ideas of genufes and fpeciefes. This is done by the 

 di'vifion of wMDrds into what is called the Parts of Speech, which 

 may be called the Categories of Language; — comprehending, like the 

 categories, the whole of things, but divided and arranged in fuch a 

 vv'ay as toferve the purpofe of fpeech; and not confidering the nature 

 of things abftradly, as they are confidered in the categories, but 

 with reference otdy to their ufe in language. They admit, howe- 

 ver, of that general divifion of the categories into fuhjlance and ac- 

 cidents \ and accordingly Plato and Ariftotle have divided the parts 

 cf fpeech into thefe two, calling the one of them exprefling Sub- 

 ftances a noun, and the other exprefling Accidents a •verb *. 



Words, however, though they admit of this divifion into clafles, 

 appear ftill to be infinite in number, as well as the ideas which they 

 exprefs. They would feem, therefore, to be altogether incompre- 

 henfible by our memories, and confequently not fit for the ufe of 

 language ; and they certainly would be fo, if all things were to be 

 exprefled by words having no connetSion with one another. But 

 fcience has contrived three ways of connecting words together both 

 'by found and fenfe, fo that the knowledge of one word naturally 



leads 



* See vol. II. of the Oii^in of Language, p. 2S. 



