282 



a labial. On looking to the more distant dialects for some help 

 in explaining the origin of these differences, I found that / 

 was commonly found in these words everywhere except in 

 the Labrador dialect; but the matter also proved to be more 

 complicated than I had at first supposed. For the West Eskimo 

 forms presented a new. third type, where those pairs of con- 

 sonants that occur in the Greenlandic forms are found in the 

 reverse order; and as against the Greenlandic homogeneous 

 consonant-combinations (r/, w/, /•), the Labrador dialect has 

 heterogeneous assimilations. 



To which division of the language is Ibis last dialect to be 

 assigned? Its [p] and [/] (p. 229) may just as well have come of 

 r/ as of //•. The question must be decided through such cases 

 where the usual process of assimilation has not been carried 

 out and where the original form has therefore not become 

 obscured. As against the constant Greenlandic [>/], there are, 

 in all, three formations in the Labrador dialect [r^, fr. ■(]. I 

 feel convinced that these formations were originally quite similar 

 and that they only indicate different stages of development. As 

 long as nothing points to the contrary, I shall assume that the 

 difference between them has developed within the Labrador 

 dialect itself in this way: an original group consisting of r -\- 

 a front consonant (presumably /), has first passed into ;g ; this 

 has further become assimilated in some words, either directly, 

 partly to ^, partly to ^, or else in every case first to p, which 

 has thfn been fronted in some few words and changed to ^. 

 But the uvular was the first of the two consonants; the inver- 

 sion has taken place; so in this particular, the Labrador 

 dialect stands on the same side as Greenlandic. 



On the other hand, in its treatment of the group labial 

 -f / (or another consonanti, the Labrador dialect leans in the 

 direction of the western dialects. The metathesis which has 

 taken place in wÅ etc. in Greenlandic does not appear to have 

 occurred in Labrador. — Yet this bond is not nearly so strong 



