4 
Short a before and after y may be replaced by i, and intermediate 
sounds (short e- vowels) occur; thus the second person singular conjunct 
ends in -yan or -yin ; ayahtsiyiniw : a Blackfoot, iyahtHyiniw. 
Short a before w is rounded; the usual pronunciation of a word like 
mayaw : straight might be symbolized by mayow. More complex is the 
phrase nama' wiya : not, nama' wya, namo'ya. In other positions the 
rounding or backing of short a is less marked than in other Central Algon- 
quian ; final -ak after w, as in ndpdwak : men, comes nearest to this. 
The combination -iw at the end of a word is spoken with now one, 
now the other, element syllabic, or perhaps with balanced syllable stress; 
apiw : he sits there, apyuw, apyw. 
Final -dw has the second element rather open, and probably often 
balanced stress; ndpdw : man, ndpeo. 
Initial uy- similarly appears often as wiy- ; uydkan : bowl, wiydkan. 
The combination wd after consonant is probably often uy ; mwdstas : 
afterwards, muystas. 
The combination iyi seems often to be merged to I ; kohtd'wiyinaw : 
our father, kohtd'wlnaw ; in this instance and in some others the place of 
the stress presupposes % rather than iyi. 
The sound here represented by a is a long mid front vowel (as in Ger- 
man Tee); the symbol a is used for the sake of uniformity with other 
Algonquian texts; e would be better. Indeed, the vowel is, to my ear, 
quite close to 1, and I have no doubt sometimes confused the two. In some 
words the variation of d and I is probably real and not merely my mis- 
hearing, as kdkway : something, beside klkway ; kdtahtawdh : at a point 
in time, beside kdtahtawdh. 
The sound s varies freely between normal and abnormal sibilant. It 
is often spoken long, especially after stressed vowels (i.e., after the ante- 
penultimate vowel), e.g. in asiniy : stone, asamdw : he gives him to eat, 
misiwdh : all over. It is tempting to view this long s as the reflex of old 
consonant groups that still appear in Menomini and are represented in 
Ojibwa by unvoiced s (as opposed to voiced z): Menomini a’ sen : stone, 
ahsdmdw: he gives him to eat. This hypothesis, however, was not verified : 
the difference between shorter and longer s seems to be non-distinctive, for, 
firstly, all gradations between the shortest and longest forms seem to 
occur; secondly, both short and long forms are used in the same word; 
and, thirdly, the longer form is not confined to words that once had the 
consonant group, but seems to occur in any and all, as misiwdh : all over 
(Menomini mesewd, but here Cree ss might be analogically due to rnisi- : 
big, Menomini me’ si-), isih : thus (Menomini is, is eh). I soon found it 
impracticable to try to record the apparent quantitative fluctuations of s. 
Medially k is often voiced, as, uspwdkan : pipe; this has not been 
indicated, as it is not distinctive. 
There seems to be some fluctuation between final -h (whether histori- 
cally old h OT SL glide, is indifferent in Cree) and -hk ; thus the conjunct 
ending for he . . . it, as dh-utinahk : when he takes it, dh~utinah. Similarly 
between final -s and -sk ; piyis : at last, piyisk. In some words there 
seems to be no such fluctuation, as, askihk : kettle, mdnisk : earthwork. 
If this observation be correct, the phenomenon is due to sandhi variants 
analogically turned loose; words with obvious inflectional parallels 
(askihkwak : kettles) would be protected. 
