Both are based on work carried out in 1906 with language consultant Frances Johnson (Takelma name Kʷìskʷasá:n),[2] who lived on to become the last surviving fluent speaker.

An English-Takelma dictionary is currently being created on the basis of printed sources with the aim of reviving the language.

Writing in 1909, Sapir stated that "the Takelma language represents one of the distinct linguistic stocks of North America".

[6] He later revised his opinion, and assigned Takelma to the hypothetical Penutian language family,[b] a grouping that at present is not generally considered established.

[9] Consonants /s/ and /ts’/ have optional alveopalatal allophones [ɕ] and [tɕ’], which occur mainly in word-initial position before a vowel, and intervocalically.

[12] Sapir also notes the existence of close /e:/, as in [la:le:tʰam] "you became", [kane:hiʔ] "and then", which he considers to be an (apparently unconditioned) allophone of /i:/.

As Sapir comments, "it not infrequently happens that the major part of a clause will thus be strung along with-out decided stress-accent until some emphatic noun or verb-form is reached", as in the sentence: kane:hiʔand thentɛwɛnxatomorrowla:le:it becamehonoʔagainpʰɛlɛ̀xaʔthey went out to warkane:hiʔ tɛwɛnxa la:le: honoʔ pʰɛlɛ̀xaʔ{and then} tomorrow {it became} again {they went out to war}"and then, the next day, they went out to war again""All that precedes the main verb form /pʰɛlɛ̀xaʔ/ 'they went out to war' is relatively un-important, and hence is hurried over without anywhere receiving marked stress [or pitch]".

[16] The transcription system used by Sapir (1909, 1912) is the then current version of the Americanist phonetic notation, which has long since been superseded.

(Phones marked with * occur in interjections and in sound-symbolic forms, but not in normal speech; [ə] is also epenthetic.)

Takelma, like many Native American languages, is polysynthetic meaning that one can link together many different morphemes to form a word.