Editing Edward Sapir (section)

====Early fieldwork====
[[File:Tilohash.jpg|thumb|Tony Tillohash with family. Tillohash was Sapir's collaborator on the famous description of the Southern Paiute language]]
Sapir's first fieldwork was on the [[Upper Chinook language|Wishram Chinook language]] in the summer of 1905, funded by the Bureau of American Ethnology. This first experience with Native American languages in the field was closely overseen by Boas, who was particularly interested in having Sapir gather ethnological information for the Bureau. Sapir gathered a volume of Wishram texts, published 1909, and he managed to achieve a much more sophisticated understanding of the Chinook [[phonology|sound system]] than Boas. In the summer of 1906 he worked on [[Takelma language|Takelma]] and [[Shasta Costa|Chasta Costa]]. Sapir's work on Takelma became his doctoral dissertation, which he defended in 1908. The dissertation foreshadowed several important trends in Sapir's work, particularly the careful attention to the intuition of native speakers regarding sound patterns that later would become the basis for Sapir's formulation of the [[phoneme]].<ref>Darnell 1990:23</ref>

In 1907–1908 Sapir was offered a position at the [[University of California, Berkeley|University of California at Berkeley]], where Boas' first student [[Alfred Kroeber]] was the head of a project under the California state survey to document the Indigenous languages of California. Kroeber suggested that Sapir study the nearly extinct [[Yana language]], and Sapir set to work. Sapir worked first with Betty Brown, one of the language's few remaining speakers. Later he began work with Sam Batwi, who spoke another dialect of Yana, but whose knowledge of Yana mythology was an important fount of knowledge. Sapir described the way in which the Yana language distinguishes grammatically and lexically between the speech of men and women.<ref>Darnell 1990:26</ref>

The collaboration between Kroeber and Sapir was made difficult by the fact that Sapir largely followed his own interest in detailed linguistic description, ignoring the administrative pressures to which Kroeber was subject, among them the need for a speedy completion and a focus on the broader classification issues. In the end Sapir didn't finish the work during the allotted year, and Kroeber was unable to offer him a longer appointment.

Disappointed at not being able to stay at Berkeley, Sapir devoted his best efforts to other work, and did not get around to preparing any of the Yana material for publication until 1910,<ref>Sapir, Edward. 1910. ''Yana Texts''. University of California Publications in American Archaeology and Ethnology, vol. 1, no. 9. Berkeley: University Press. ([https://archive.org/details/yanatexts00sapirich Online version] at the [[Internet Archive]]).</ref> to Kroeber's deep disappointment.<ref>Darnell 1990:24–29</ref>

Sapir ended up leaving California early to take up a fellowship at the [[University of Pennsylvania]], where he taught Ethnology and American Linguistics. At Pennsylvania he worked closely with another student of Boas, [[Frank Speck]], and the two undertook work on [[Catawba language|Catawba]] in the summer of 1909.<ref>Darnell 1990:29–31</ref> Also in the summer of 1909, Sapir went to Utah with his student [[J. Alden Mason]]. Intending originally to work on Hopi, he studied the [[Colorado River Numic language|Southern Paiute language]]; he decided to work with [[Tony Tillohash]], who proved to be the perfect informant. Tillohash's strong intuition about the sound patterns of his language led Sapir to propose that the [[phoneme]] is not just an abstraction existing at the structural level of language, but in fact has psychological reality for speakers.

Tillohash became a good friend of Sapir, and visited him at his home in New York and Philadelphia. Sapir worked with his father to transcribe a number of Southern Paiute songs that Tillohash knew. This fruitful collaboration laid the ground work for the classical description of the Southern Paiute language published in 1930,<ref>{{cite journal | last1 = Sapir | first1 = Edward | title = The Southern Paiute language | journal = Proceedings of the American Academy of Arts and Sciences | volume = 65 | issue = 1| pages = 1–730 | doi=10.2307/20026309| jstor = 20026309 | year = 1930 }}</ref> and enabled Sapir to produce conclusive evidence linking the [[Shoshonean languages]] to the [[Nahuan languages]] – establishing the [[Uto-Aztecan languages|Uto-Aztecan language family]]. Sapir's description of Southern Paiute is known by linguistics as "a model of analytical excellence".<ref>Darnell 1990:34</ref>

At Pennsylvania, Sapir was urged to work at a quicker pace than he felt comfortable. His "Grammar of Southern Paiute" was supposed to be published in Boas' ''[[Handbook of American Indian Languages]]'', and Boas urged him to complete a preliminary version while funding for the publication remained available, but Sapir did not want to compromise on quality, and in the end the ''Handbook'' had to go to press without Sapir's piece. Boas kept working to secure a stable appointment for his student, and by his recommendation Sapir ended up being hired by the Canadian Geological Survey, who wanted him to lead the institutionalization of anthropology in Canada.<ref>Darnell 1990:42</ref> Sapir, who by then had given up the hope of working at one of the few American research universities, accepted the appointment and moved to Ottawa.