246 



It is not difficult to get some idea as to the shape of the 

 uvularized forms in some earlier period. It lies near at hand 

 to assume that they have all originated in the same way, na- 

 mely after the analogy of ateq and those other words where 

 we have parallel forms to compare with in the West Eskimo 

 dialects*). In these dialects, as we have seen, the inflected 

 forms too represent the stage when the vowel of the stem 

 has not yet been uvularized and the metathesis has 

 not yet been carried out. 



But I think it is possible to go one more stage further back. 



There is an indication of the original conditions in one 

 of the more central dialects, namely in the Mackenzie River 

 dialect. For the numeral 2, Pet i tot gives for the Churchill 

 River dialect the form malpoJi, but for the dialect that he 

 himself had heard he gives a form which he spells maUepok 

 or mdlœpoJï] Greenlandic has marXuk. Analogously, then, I 

 think it may be assumed that arqit originally had the shape 

 *ateqit, which would be the regular plural form of ateq^ cf. 

 plur. inuit (men) < sing, innk (man), plur. mvuit < sing, loruh 

 (evening) etc. And Petitot actually gives a plural form for 

 this word from the ^Mackenzie dialect which could be the next 

 stage of the development of the original form, namely atépe'it. 



Petitot has still more Avords that show that the present 

 Greenlandic forms must originally have been one syllable 

 longer, namely: 



M. L. W Or. 



[/sg] malœpit^ waves \y\ maggit^ [X'] maX:it, in the sing. ma^Æ 



apveneJœpit six » aripiniX'it 

 С 



[pl\ ti g iligoyapk thief \qiI\tigiliktok » ti/.'if'oq 

 ^ In M. also »lalit. in L. mallit. 



Yet at the same time I shall not omit to remark that we may risk 

 making mistakes when we try to reconstruct without qualification the 



