64 INTRODUCTION TO THE STUDY OF INDIAN LANGUAGES. 
that when the key is once found the word-puzzle may be taken in pieces as easily as it 
was put together. Indeed, it is a requirement of the Indian languages that every 
word shall be so framed as to admit of immediate resolution to its significant elements 
by the hearer. It must be thoroughly self-defining, for (as Max Miiller has expressed 
it) ‘it requires tradition, society, and literature to maintain words which can no longer 
be analyzed at once.” . . . In the ever-shifting state of a nomadic society no 
debased coin can be tolerated in language, no obscure legend accepted on trust. The 
metal must be pure and the legend distinct.* The more cumbrous and unwieldy the 
structure, the greater is the necessity for exact adjustment of its parts; and the laws 
of verbal composition are well-established, admitting no exceptions. 
How far such an analysis as I have suggested can be successfully carried need 
not now be inquired. Every step taken in that direction will be something gained, 
will lead to more exact knowledge and to positive results. To determine and classify 
the primary verbs in any one language would be to bring a larger contribution to lin- 
guistic science than has often been made by students of the American tongues. Back 
of these verbs and of the primary demonstratives are the ultimate roots. These may 
not now be, possibly they never will be, attainable; yet I do not hesitate to express 
my belief that through the study of the American languages scholars may as nearly 
arrive at a solution of the great problem of the genesis of speech, in determining the 
character and office of its germs, as by any other avenue of approach. All attempts 
to establish relationship between the several great linguistic families by the identifica- 
tion of roots, may indeed be regarded as hopeless; for few will be disposed to question 
Professor Whitney’s conclusion (Language and the Study of Language, p. 392) that 
“the difficulties in the way of a fruitful comparison of roots are altogether overwhelm- 
ing”; and probably no one is yet “so sanguine as to expect to discover, amid the blind 
confusion of the American languages, where there are scores of groups which seem to 
be totally diverse in constituent material, the radical elements which have lain at the 
basis of their common development.” But if order is ever to be brought out of this 
blind confusion—if any satisfactory classification of the hundreds of languages and 
dialects now so loosely grouped is to be established—if the genetic relation of one of 
these to another is to be demonstrated even in those cases where, on grounds independ- 
ent of language, the probability of such relation is greatest—analysis must first do its 
work, until, at least, it shall have determined and classified the earliest traceable con- 
stituents of speech, though compelled to stop short of the discovery of ultimate roots. 
If the method I have indicated is the true one, the collection of materials for the 
critical study of an American language should begin, not with the translation into it 
of a given number of English names, but by looking out its simplest, 7. e., least compos- 
ite words, and fixing their meanings,—by detaching from the constant roots or themes 
terminations and formatives which are merely grammatical,—and by translating from 
the Indian to the English, provisionally and subject to correction by more rigid analysis, 
the syntheses which discharge the office of concrete names, by conveying concise defi- 
nitions or specific descriptions of the objects to which they are severally appropriated. 
Among the words and elements of words which claim earliest attention, may be 
mentioned— 
1. The Pronouns, separable and inseparable, and pronominal suffixes: with which 
may be included the demonstratives. 
* Lectures on the Science of Language, 1st series, pp. 292, 293, 
