374 -APPENDIX.— On GREEK ANALOGS 



but the vowel founds ferving for the enunciation of the confo- 

 nants. It is altogether an error, therefore, to ftate thefe as of 

 themfelves figniricant : but the framers of the analogical theory- 

 were miiled by this circumftance, that as the vowel founds were 

 neceffary for the enunciation of the confonants ; as all founds 

 were meant to exprefs fome fpecies of action ; and as all action 

 neceffarily implies motion, or the ceffation of motion • it feemed 

 therefore not unnatural, that the neceffary component of the fig- 

 niricant found fhould be fuppofed to denote the equally necef- 

 fary component of the action or idea fignirled. 



Having afcertained the application of the vowel founds, the 

 next object is to confider the application of the confonants. In 

 regard to this, it may be obferved, that whatever is thought of 

 the hypothecs of the duads, it at lead implies nothing abfurd or 

 improbable to fuppofe, that when men began to name objects 

 from their action, as experienced in their effects, the fame or fi- 

 milar feelings would naturally be expreffed by the fame or near- 

 ly fimilar founds ; that the expreflive part of thefe founds was 

 the confonant, and that, confequently, when the defire to indi- 

 cate a feeling or idea once fignified recurred anew, this would be 

 done by a repetition of the fame confonant found which had firft 

 been employed to make it known. Hence it will follow, that 

 each different confonant, when enunciated, would foon come to 

 denote a particular range of ideas, agreeing among themfelves in 

 fome common quality, diftinct from what were expreffed by 

 any of the others. What particular clafs of ideas each confonant 

 was to denote, was purely arbitrary and accidental. This, there- 

 fore, can never be afcertained by theoretical conjecture : if it be 

 known at all, it muft be traced by the obfervation of facts, as 

 they actually occur in the exifling primitive languages. 



Whether, by fuch a mode of investigation, the actual for- 

 mation of language could be fully and fatisfactorily developed, 

 I cannot pretend to determine. I muft own, however, that it 



feems 



