¹ John B. McIntosh, Huichol phonemes, IJAL 11.31-5 (1945). McIntosh lists the following phonemes: p t c č k k^w ? (in this paper written q); h z; m n; l r; w y; i a (in this paper written ï) u e a; stress. For reasons that will appear, č and l should be deleted from the list; for morphophonemic reasons, h should be classified with w and y rather than with z. The present state of our analysis indicates that there may be three or four stress phonemes and at least three pitch phonemes, the latter occurring largely or entirely in morphemes of intonation.

Huichol is spoken in three dialects: eastern, central, and western; this study deals with central Huichol, but the conclusions appear to be valid for the western dialect as well, perhaps also for the eastern. The field work on which the study is based was carried out between 1952 and 1954 under the auspices of the Summer Institute of Linguistics. I wish to express my gratitude for helpful criticisms of this paper contributed by Kenneth L. Pike and other members of the Summer Institute of Linguistics, and by Bernard Bloch.

² Phonetic symbols used in this paper include [^y] palatalization, [r] retroflex reverse flap, [r] alveolar flap, [r̄] alveolar trill, [1] lateral reverse flap, [l] lateral continuant, [m] voiceless bilabial nasal, [z s] voiced and voiceless retroflex sibilants, [Z S] voiced and voiceless alveolar sibilants, [q] glottal stop, [e] lower mid-front vocoid, [ï] unrounded high-back vocoid, and an upward-pointing arrowhead to indicate raising of a vowel.

³ A man, a woman, and a four-year-old child were shown the pictures of ceremonial objects in Carl Lumholtz, Symbolism of the Huichol Indians (New York, 1900). Their comments to each other were recorded on tape. In this fragment of their conversation, the phrases are numbered 1 to 5. The woman spoke phrase 4, the man spoke the rest.

⁴ The alternation of consonants in Class 1 and Class 2 is paralleled by alternation between higher and lower front vowels. In phrases with sounds of Class 1, the more frequently occurring front-vowel phones are [i] and [æ]; in phrases with sounds of Class 2, the more frequently occurring front-vowel phones are [i] and [e^a]. In this paper, only the consonant are discussed.

⁵ Paul L. Garvin, Kutenai I. Phonemics, IJAL 14.87-90 (1948), and Hans Wolff, Osage I. Phonemes and historical phonology, IJAL 18.63-8 (1952), both recognize more than one style on the phonological level in the languages which they describe, but give no details concerning the nature of those styles. Bernard Bloch, Studies in colloquial Japanese IV. Phonemics, Lg. 26.86-125 (1950), esp. 88, limits his description to one style, implying that there are others.

⁶ Kenneth L. Pike, Grammatical prerequisites to phonemic analysis, Word 3.155-72 (1947); More on grammatical prerequisites, Word 8.106-21 (1952). Pike has suggested that grammatical boundaries may constitute relevant environments for phonological units.

⁷ This type of solution in some respects parallels the solution of the problem of tempo offered by Charles F. Hockett, Peiping morphophonemics, Lg. 26.63-85 (1950), §9. However, Hockett's phoneme of speed conditions the distribution of allomorphs, not of allophones as in the present problem.

⁸ In the substitution of Class-2 sounds for Class-1 sounds, in controlled and uncontrolled conversations, in the reciting of texts, and in the eliciting of direct responses, there was never any indication that two utterances differing only in their alveolar phonemes might refer to different physical objects or otherwise have different referential meanings. The same speaker would call the same small white flower both [tseïye] and [ts^yeïye].

⁹ Charles C. Fries and Kenneth L. Pike, Coexistent phonemic systems, Lg. 25.29-50 (1949), esp. 45-7. The general approach here followed is outlined in that article.

¹⁰ Bernard Bloch offers another definition of style, Linguistic structure and linguistic analysis, Report of the Fourth Annual Round Table Meeting on Linguistics and Language Teaching 42 (Washington, 1953): ‘the style of a discourse is the message carried by the frequency-distributions and transitional probabilities of its linguistic features, especially as they differ from those of the same features in the language as a whole.‘ This is adequate for aspects of style in which the elements are already known to be in contrast, but not for those aspects which contrast only stylistically.