The idea of a special relation between Takelma and the Kalapuyan languages was first developed by Leo Frachtenberg, who listed 55 lexical correspondences between Takelma and Central Kalapuya. Based on Frachtenberg's observations, Edward Sapir included both Takelma and Kalapuyan in his extended version of the Penutian "stock", listing them however as individual members without positing a special relationship between the two. The first explicit proposal for a family comprising only Takelma and Kalapuyan was made by Morris Swadesh in a lexicostatistic study, who found a lexical similarity of 48% between Takelma and Kalapuyan, although this figure was based on rather bold assumptions about lexical matches. Shipley made the first attempt towards establishing regular sound correspondences by strictly applying the comparative method, and listed sixty-five preliminary reconstructions for "Proto-Takelman". Further lexical cognate sets were given by Berman, while Kendall presented phonological and morphological correspondences.
Reception
considered the proposed Takelma-Kalapuyan hypothesis "highly likely, if not fully demonstrated", and listed them as a single subgroup in an overview of Native American language families that is otherwise characterized by a very critical stance against wide-range proposals. Takelma-Kalapuyan is also tentatively accepted by Victor Golla. In an unpublished, but widely cited conference paper, Tarpent and Kendall critically evaluated the evidence for Takelma-Kalapuyan, and concluded that a grouping which exclusively comprises Takelma and Kalapuyan is not justified, and that features shared between the two have to assessed in a wider Penutian context. Marianne Mithun accepts Tarpent and Kendall's findings, but remains sceptical about the validity of the Penutian hypothesis, and therefore lists Takelma as a language isolate and Takelman as a primary language family. Grant maintains that even though the relation between Takelma and Kalapuyan is not as close as previously assumed, Tarpent and Kendall's discussion does not invalidate the hypothesis itself, and he considers Takelma and Kalapuyan "each other's closest genetic relatives", albeit only at an extremely distant level.
Prehistory
The Kalapuyan languages and Takelma were spoken in two discontiguous areas: while Kalapuyan speakers inhabited the Willamette Valley, Takelma speakers lived in the Rogue Valley in the southernmost part of Oregon. Inbetween, the AthabaskanUpper Umpqua language was spoken. This suggests that Takelma was initially spoken in the direct neighborhood of the Kalapuyan area, being separated from the Kalapuyans, and pushed southwards by intruding Athabaskan speakers. Golla suggests that Takelma replaced an earlier Karuk-like language in the Rogue Valley, based on aeral features shared by Takelma and Karuk. This event is possibly related to archaeological evidence of a transition in the material cultures of the Rogue Valley that occurred around 300 CE, from the "Glade Tradition" to the "Siskiyou Pattern".
Proto-language
Proto-Takelma–Kalapuyan reconstructions by William Shipley :
