¹ Presented in part at the summer meeting of the Linguistic Society of America, 1 August 1947, at Ann Arbor, Michigan.

² The Mazateco tribe is located in the northern tip of the State of Oaxaca, along the southern edge of the State of Puebla, and in the municipio of Zongolica, State of Vera Cruz, Mexico. The tribe numbers perhaps 60,000, with 55,743 monolingual speakers (not including children under five years of age) according to the 1940 census figures. It is a member of the Popoloca-Mazateco language family, which also includes Ixcateca and Chuchon.

The writer, under the auspices of the Summer Institute of Linguistics, lived in Huautla de Jiménez, Oaxaca, during the winters of 1943–44 through 1946–47. The data for the following texts were gathered at Río Santiago, a ranch pertaining to the municipio of Huautla de Jiménez and located three hours east of it. Modesto García, a lad of about 17 years, was the informant. The texts were later checked with Fortino Cortés, a boy of 14 years, who lived in Huautla.

³ The same phenomenon is reported among the Zapotecos of Yatzachi el Bajo by Otis Leal of the Summer Institute of Linguistics, and among the Chinantecos by R. J. Weitlaner of the National School of Anthropology and History, Mexico City. Both the foregoing are tonal languages of Oaxaca, Mexico.

George Herzog, in Speech Melody and Primitive Music, Musical Quarterly 20.454 (1934), refers to a similar phenomenon observed in Africa. ‘The use of musical instruments for communication over long distances is very common in Africa. . . . studies indicate that the musical elements of speech are transferred to the instrument and that this is the essence of the technique.‘

⁴ The Mazateco language is homogeneous so far as sex is concerned, there being no difference between the speech of men and of women.

⁵ Observed and reported by Eunice V. Pike, Summer Institute of Linguistics.

⁶ On one occasion Eusebio Martínez was talking with Erasto Jiménez, a local carpenter of Huautla de Jiménez. Erasto talked so steadily that Eusebio was unable to interrupt and make his contribution to the conversation. He was observed however humming very abbreviated remarks into the conversation while Erasto kept on talking aloud—apparently without noticeable interference with Erasto's flow of speech.

⁷ Herzog, op.cit. 455–6, speaking of similar phenomena in the African drum calls, says: ‘On some signalling instruments communication is restricted to a limited number of signals, perhaps from ten to fifty; on others there is practically unlimited conversation between two players over a long distance. That this should be feasible with the apparently restricted means of the system, and with the great number of words sharing the same pitch, can easily be, and has been, doubted; especially since in West African languages, as in Chinese, the majority of the words consist of only one syllable. It would lead us too far afield, however, to go into detail on the question within the frame of this article. I must be content merely with testifying that it is done.‘

⁸ Observed by Eunice V. Pike.

⁹ Observed by Eunice V. Pike.

¹⁰ Mazateco has four phonemic registers, indicated by raised numbers after the vowels as follows: ¹high, ²semi-high, ³semi-low, ⁴low.

¹¹ For an analysis of these glides, with examples from spoken Mazateco, see Kenneth L. Pike, Tone Languages 95–165 (Ann Arbor, Mich., 1948).

¹² The phonemes of spoken Mazateco are as follows: (Huautla dialect) voiceless stops t, k, ?; voiceless affricates ¢, č, c; voiced nasals m, n, ñ; voiceless fricatives s, š, h; voiced fricative v; lateral l; glide y; flap r; oral vowels i, e, a, o; nasalized vowels į, ę, ą, ǫ. For further description of these phonemes see Kenneth L. Pike, Immediate Constituents of Mazateco Syllables, IJAL 13.78–91, §3 and §5 (1947). In the transcriptions, a hyphen sets off clitic elements.

¹³ For a statement of what constitutes a phonemic syllable in spoken Mazateco see Kenneth L. Pike, Immediate Constituents of Mazateco Syllables, IJAL 13.78–9, §2.

¹⁴ This is confirmed also by the experience of Eunice V. Pike, who, when first learning to whistle, was not always able to strike the same key as her informant. Her informant, realizing her difficulty, would shift pitch, accommodating his pitch to hers.

¹⁵ It is of interest to note that the spoken language has a retroflexed sibilant phoneme š, which, with some speakers, is frequently whistled in normal speech. This retroflexed tongue whistle is never used in the conversational whistling described in this article, nor is there ever any significant pitch to it.