1. The data for this paper were collected during the summer of 1952 at the Summer Institute of Linguistics in Caronport, Saskatchewan, Canada. Three informants, ranging in age from fifty to sixty-six years were used. Two were from La Rouge, Sask., and were sisters : Mrs. Angelique Ross, and Mrs. Sarah Ballentyne. The other informant, Mr. Peter Wuttannee, was from Sweet Grass, Sask. Of the three informants, Mrs. Ross seemed to use the fronted [æ] allophone of / a / more than any other informant; while Mr. Wuttannee seemed to use it the least. Probably all the data collected here represent a rather precise style of spech which should be compared with more rapid speech forms. This paper pretends to do nothing more than to describe certain correlations of quality and quantity in the sort of data here analyzed. However, on the basis of sporadic observations of more rapid styles — in recorded text and in overheard conversation between informants — I would hazard the conjecture that the qualitative-quantitative variation in inverse proportion as here described holds in more rapid styles as well. General rapid rate of utterance seems to accentuate qualitative differences between long and short vowels except in syllabes receiving intonational stress and lengthening — in which case the opposite tendency sets in.

2. Bloomfield did not apparently attempt to describe any specific conditioning factors for the Cree allophones that he describes (See introduction to Sacred Stories of the Sweet Orass Cree, National Museum of Canada, Ottawa, 1930). Notice his description of / i / (p. 7) : “i short, high front vowel, as in English pin, varying all the way to the type of French fini”. However, since the various allophones of / i / do not vary freely with each other in all positions, it seems imperative that some attempt be made to state more precise conditions. Bloomfield's summary statements regarding Cree vowels bear evidence to his dependable linguistic intuition.

3. Since the writing of the main body of this paper some evidence has accumulated that vowel quality / e / may also occur as long or short phonemically. Notice the following pair of words : e:ka, now, eha?, yes. However, since the short vowel phoneme / e / has no marked qualitative variants its discussion is not particularly pertinent in this paper, where we are attempting to correlate such qualitative variants with variations in quantity. Bloomfield listed only long / e : / (with symbol ä) in Sweet Grass Cree (p. 6).

4. Bloomfield's statement concerning the occurrence of stress is rather over-simplified : “A non-significant stress accent falls on the third syllable from the end of words or close-knit phrases.” Actually a high pitch (part of the unanalyzed intonational system) occurs a good deal more regularly on the antepenultimate syllable than the stress itself which may occur on either the antepenultimate or penultimate syllables.

5. Although the stress seems to fall regularly on the last syllable of disyllabic words, the pitch of the last syllable may be either lower or higher than that of the first syllable. With intonational pattern low-high, the utterance seems to be more emphatic; while with intonational pattern high-low it is more casual : àstám, come ! (urgently), àstám, come ! (somewhat casually).

6. No attempt is made here to illustrate with examples these impressionistic observations about phonetic variations in vowel length. Such phonetic minutiæ are difficult to symbolize without employing a notation which gives the impression of more exactness than I would care to imply.

7. The hyphen after the syllable eh in the phonemic transcription of this example symbolizes a phonological juncture indicated by the occurrence of the secondary stress in a position where we would not normally except it — since stress normally tends to occur in alternating syllables.

The geminated [p] in the phonetic transcription and any such geminated consonants in phonetic examples in the balance of this paper are allophones of the respective non-geminated consonants. Almost any Cree consonant (except perhaps/s/, /y/, /w/, and /h/ — although we have even recorded a few doubtful examples of these) will tend to geminate in intervocalic position provided the preceding vowel is quite short phonetically (not merely short phonemically). This is another inverse proportion involving vowel length and consonant length. Various phonetic degrees of phonetic gemination occur; in this paper we have only written the gemination where it sounded to us rather prominent phonetically. Syllable break comes in the center of such a long geminated consonant. The onset of primary or secondary stress has also been recorded medial in such consonants as well.