On Comparative Philology. 91 



express the cardinal numbers, family relations, parts of the human body, the 

 principal objects of nature, domestic animals, personal pronouns, and certain 

 familiar actiom or jassions. 



But here, if the positive argument in favour of the connection of dialects 

 which have these in common, he conceded to be ever so strong, the dis- 

 proof arising from a defect in these is not so strong. It is certain that in 

 Arabic eighty words are alleged for honey, and two hundred for a lion ; 

 both ideas very familiar. Indeed, the commonest words seem sometimes 

 to have most tendency to change their names.* In a note will be found a 

 very small selection, (for space will allow no more,) to illustrate the exten- 

 sive changes in very familiar words, happening in dialects which differ so 

 little as to be almost one and the same language in every thing beyond 

 what is familiar : for Hellenic and Romaic have their vocabularies con- 

 siderably different, while conversation runs on every-day topics, but may be 

 called identical upon higher subjects. 



It seems then that the non-agreement of two languages in their most 

 familiar material, is no very decisive disproof of their real connection by 

 pedigree ; and this leads to remark, that equal uncertainty is hereby thrown 

 on the argument, that two languages must needs be akin, which have some 

 of these familiar elements in common. For if they are in any proximity, 

 and if it have been proved that men are not untractably tenacious of their 

 own vocabulary in these matters, nothing hinders that one language should 

 borrow them from the other. Especially is it easy to borrow numbers ; 

 for in two conterminous or mixed races, each is sure to learn the numera- 

 tion table of the other. 



When, however, such agreement is found between languages spoken by 

 nations which can have had no intercourse with each other from the re- 

 motest ages; and after deducting all resemblances which can be with any 

 shadow of reason called uncertain, if but a dozen words are left in com- 

 mon ; if these are all found among the primary elements of speech, they 

 are too many to be supposed accidental coincidences. Such appears to be 

 the case made out by Dr. Prichard for the connexion of the Celtic lan- 

 guages with the Sanskrit. If we resist the notion of a common origin, we 

 must apparently believe that the Celts learned these names, directly or in- 

 directly, from the Sanskrit; suppose, directly from the Goths. 



The second point to be considered, is the agreement between two lan- 

 guages as to the mode of inflexion. And here also experience assures us, 



Hellenic. Romaic. Hellenic. Itoinaic. 



• Black /icXac fiavpo^ Cat mXoupoe, or 



White XlVKOQ uiXTrpoc yaXtwnig .» ydroe 



Gold XP""^^" ixc'tXnyfia House ....oIkoq (nrin 



Horse Vn-Trof dXoyo Bread u(>Tog i//w/ii 



Dog Kuoii' <TKv\dKi Water .... i)(?wj) vtpo 



Ass vvoc yrti<"«po A Rose jx'i^or T(iiavTd(pv\\ov 



Pig ffOf fimt^TiQo 



