THE PHILOSOPHY OF LANGUAGE. 347 



not a convention or an institution. The most superficial dissection of the 

 familiar forms of words, as we have them in our grammars, distinctly shows this. 

 A mo amas means merely love I, love thou; the Sancrit asmi, the Latin sum, 

 and the Greek dpi, being merely the two interflowing elements which are 

 presented separately in the Gaelic tlta mi and the English / am. So the case 

 terminations in Greek and Latin are merely significant attachments expressive 

 of local relationship which have grown into the root in these terminational 

 languages, but of which the meaning stands clear in that detached form 

 which these agglutinated postpositives present as independent propositions ; 

 the sign of the genitive in English of being manifestly = off, Greek cbrd, Latin ah, 

 away from. It matters nothing that we cannot in all cases, or in the majority 

 of cases, distinctly put our fingers on the original significant form of the abbre- 

 viated or polished case termination ; enough that man is a reasonable animal, 

 and that from his reasonable proceeding in known cases we can certainly divine 

 it in where the formative action is hidden from our view. Words as we have 

 them, especially terminations, conjunctions, and other such frequently used and 

 much abused elements of the vocal currency of a country, are like old shillings 

 from which the image and superscription has been defaced, but which certainly 

 was there, as it lies in the very nature of a coinage to bear some stamp and 

 authoritative signature on its face. 



VIII. That some modifications made in the root are without separate signi- 

 ficance, and may without impropriety be called arbitrary and conventional, I 

 think we must admit ; and so Horne Tooke's rule, like other rules, will have its 

 exceptions, and must not be pressed urgently in all cases. Any child could tell 

 how rubefacio signifies to make red ; it is merely two words run into one, in the 

 same way that the Greek use woieco in aproTroios, a baker ; but no man can tell 

 me how fell came to signify to cause to fall, or how the plural of man should 

 be men. No doubt in this latter case you may say that the a of the singular 

 was changed into the e of the plural by the reflex contagion of the e in the plural 

 termination Manner ; but this is merely the description of a process of contagion 

 or infection taking place between two contiguous emissions of articulated breath, 

 not the laying bare of any natural significance in the change which has taken 

 place. There is nothing in the word fell that should cause it to mean to cause 

 to fall ; it is a pure matter of convention — an ingenious device, let us say, to 

 make one word serve two purposes, as faces have been made by ingenious 

 draughtsmen representing two different persons, according as you look at them 

 from this side or from that. In the same way no conceivable reason can be 

 given why ivilre in German, were in English, should be the subjunctive mood of 

 was ; or, what is similar, why the a of the indicative of the Sanscrit or Greek 

 should be softened into 77 in the subjunctive. It is for the sake of variety and 

 distinction alone that such changes are made ; and they are in this view perfectly 



