226 michelson: American Indian languages 



tive singular for a stems by the influence of the od ablative singu- 

 lar of o stems. In one sense the proof is not absolute, but it is 

 as absolute as it is possible to give when dealing with prehistoric 

 phenomena. It must be admitted that at times even Indo- 

 European philology is at sea, and that pure'y subjective specula- 

 tion may come into play. Happily these instances are rare. 

 These remarks are inserted because although the facts are well- 

 known to Indo-Europeanists, they are largely unknown to 

 Americanists. 



The bearing the above has on the problems of American 

 linguistics is this: since American Indian languages have not 

 been transmitted to us in the manner that Indo-European 

 languages have, we do not know what has disturbed phonetic 

 laws in many given cases, and for the methodical reasons out- 

 lined above, we are not justified in assuming that the same 

 influences have been at work in American Indian languages as 

 in Indo-European languages. Herein lies our difficulty in formu- 

 lating phonetic laws that are entirely satisfactory to the Indo- 

 Europeanist. For example, n becomes c in Fox before i which 

 is a new morphological element ; it remains if the i is not such an 

 element. There are some specific grammatical categories in 

 which the law does not work. A study of several related dialects 

 shows that this change also takes place in them, and hence must 

 be very old. At the same time the apparent exceptions have not 

 been explained. Whether they ever will be, is questionable. 

 Yet an Americanist does not object to the formulation of the 

 law as it works in practice. The Indo-Europeanist will object 

 vigorously to such a formulation as it is contrary to his accepted 

 canons. If the canons of Indo-European philologists be accepted, 

 it is quite evident Americanists can not reconstruct the parent 

 languages of American Indian stocks in an entirely satisfactory 

 manner. However, most Americanists are far more interested 

 in observing actual phonetic correspondences and the like be- 

 tween the different dialects of linguistic stocks as they actually 

 occur, than in speculations which from the nature of the case 

 must rest upon rather slim foundations. It may also be noted 

 in this connection that Indo-Europeanists have begun to interest 



