88 TRANSACTIONS OF THE CANADIAN INSTITUTE, [VoL. VI. 
kwiskwis, a word invented on the opposite side of the continent to 
imitate the grunt of the animal thereby designated. 
Another legitimate deduction from this example which applies to 
many other cases is that the only permanent, and therefore the really 
important, consonants are those which commence a word or at least a 
syllable.* | Non-initial consonants, though generally more immutable 
than the vowels, have but a relative importance. 
The consonants are then the most important element in the formation 
of words. But even among them there are some which are convertible 
with others to such an extent relatively to the various dialects that they 
are practically one and the same. This convertibility may manifest 
itself in three different ways: first, within the same dialect, as is the case 
with @d and ¢, g and &, etc., within each of the Déné idioms which cannot 
detect the slightest difference between, say, Za and da, “lip”; kw and gu 
“worm,” etc. Secondly, between related dialects or dialects belonging 
to the same linguistic group so that, though not changing the sense of 
the word, it indicates the nature of the idiom ; such are the aforesaid 
letters with regard to most Aryan languages compared with one another; 
for instance “dance” is favz in German; the Latin dens is tard in several 
germanic tongues, etc. Consonants of this second class, besides those 
already mentioned, are many and varied. Thirdly, we might extend 
this convertibility to another category of consonants, a category wherein 
cognate consonants in words from heterogeneous stocks, as the Aryan 
and the Turanian, do service in connection with words originally the 
same. Such are the f, 4, and fof the Sanskrit (Aryan) pzza, “father,” 
the Syriac (Semitic) da¢ara, the Zend or old Persian (Aryan) fedre and 
the Déné (American) pa, pzp, etc., all of which terms have the same 
signification. To be brief; some consonants are convertible with corres- 
ponding letters within the same dialect, others’ commutability manifests 
itself from dialect to dialect, while others again are commutable from 
stock to stock, that is between unrelated languages. 
The German philologist, Jakob L. Grimm, was the first to formulate 
the law which bears his name and which regulates the interchange 
of consonants in the corresponding words of the different Aryan langua- 
ges. American phonetics are quite peculiar, as is well known, and in 
connection therewith Grimm’s law not only does not cover the whole 
ground, but in several instances it is positively at fault. A prerequisite 
to safe comparisons between words from stocks of the old and of the 
new worlds would then seem to be the acquisition of some principle 
* For an apparent exception, see my paper ‘‘ Déné Roots,” Trans. C. I. vol. III, p. 151. 
