michelson: American Indian languages 225 



archaic is too complicated to be given here, as it would of 

 necessity be altogether disproportionate to the length of this 

 paper.) Naturally, in some instances, absolute proof would be 

 wanting. Thus in certain cases Ojibwa nd corresponds to Cree 

 and Menominee % Fox t. It can be easily shown that the Fox t 

 in this case is in all probability unoriginal; and most philologists 

 will assume that Ojibwa nd is more archaic than Cree and Me- 

 nominee 't. But probability and assumption are not the same 

 as proof. 



A further obstacle to the reconstruction of the parent languages 

 of American Indian stocks is our frequent inability to formulate 

 phonetic laws in a manner such as is demanded by all modern 

 Indo-European philologists. These contend that phonetic laws 

 in themselves admit of no exceptions ; and that wherever we find 

 apparent exceptions, there is some extraneous reason, or reasons. 

 The historical study of the individual Indo-European languages 

 shows that analogy and the like have been potent factors in 

 transforming them, and are fully as important as the actions of 

 phonetic laws. For example late Latin potebam is not a phonetic 

 transformation of Latin poteram, but is due to the influence 

 of other imperfects in bam preceded by a long vowel. For this 

 reason we are justified in extending the principles derived in this 

 manner to the prehistoric period, to harmonize discrepancies 

 among the historical languages which cannot be accounted for 

 by phonetic laws. Thus the Italic languages have an ablative 

 singular of a stems in ad (retained in Oscan and early Latin; 

 final d lost in classical Latin by phonetic law). The collective 

 study of Indo-European languages shows conclusively that the 

 ablative singular of a stems was the same in form as the genitive. 

 Since the same study demonstrates that o stems in the Indo- 

 European parent language had an ablative singular in od (pre- 

 served in early Latin; d lost phonetically in classical Latin), 

 and that no other stems in the Indo-European parent language 

 had a special case form for the ablative singular; and since we 

 know that in historica' Indo-European languages analogy has 

 been a potent transforming factor, we have an entirely legitimate 

 right to assume that the Italic languages deve'oped an ad abla- 



