92 TRANSACTIONS OF THE CANADIAN INSTITUTE, [Vot. VI. 
comparison, to be of any ethnologic value, should be between the Déné 
-word—supposing it to be a root—and the synonymous term in the 
original language from which modern French is derived. The French 
of to-day say a dt when their ancestors said “a dict; but between the 
formation of the dialect of the latter and the abandonment of its parent, 
Latin, an important change had taken place whereby two distinct verbs, 
habere and dicere, had been combined in one, so that the Déné aaz should 
be compared, not with the modern French phrase, but either with its 
original two verbs, HAde¢ DICTwm, or with the exact Latin synonym, d@zx7z, 
which then lacks the equivalent for the first syllable of the American 
term. 
We may classify root-words under two heads: There are those which are 
roots by reason of their standing from a grammatical point of view, and 
those which owe their position as immutable words to their signification, 
or rather to the importance in all languages of the objects they represent. 
By the first class I mean especially the numerals and the pronouns, 
which, it is well-known, generally have a kind of family air in cognate 
dialects. Asto the pronouns. I think that hardly any qualificative reser- 
vation is necessary ; but it is not so with all the numerals. By “all” I 
should be understood as referring only to the first ten numbers where the 
system is decimal. Thus in the Northern Déné dialects the words for 
seven, nine and even ten have no linguistic importance ; but the first 
four numbers especially are of the greatest moment as a means of detect- 
ing philological affinities. 
All the other roots are comprised within my second class as defined 
above. These are either monosyllabic or polysyllabic. In the first case 
their degree of immutability is generally greater, while with the second, 
except in the Semitic languages, which are remarkable for the triliter- 
ality of their radicals, there is very often allied to the original root, a sort 
of increment, accretion, or accidental alteration of a primitive element, 
which it is, of course, very important to discern. This remark applies 
not only to really ancient roots of simple import, but also, especially in 
the Aryan languages, to such words as were originally one throughout 
the whole stock, but which have grown distinctive of the particular nation 
by which they are used. In other words, in all such terms there is the 
radical and what may be considered a mere accident, whatever may be 
its place in. the structure of the word. Of course in such cases the 
radical only has any weight in the balance of the comparative philologist ; 
the accidental part of the word has no other value than that which may 
result from its being the means of identifying the particular dialect to 
which it belongs. 
ma 
a 
ala 
