Hokan hypothesis, proposed but controversial and largely abandoned grouping, or phylum, of American Indian languages. Different versions of the Hokan hypothesis include different members, most of them spoken in California and the U.S. Southwest, though several of them extend into Mexico and beyond.
The original hypothesis, made by Roland Dixon and Alfred Kroeber in 1913, contained only five language groups in California; the name was based on a similarity of the word for ‘two’ in some of those languages to something like hok. Soon various scholars were proposing possible Hokan affinities for many other languages. The following language families and isolates have commonly been associated with Hokan: Chumashan (six languages), Palaihnihan (two languages), Pomoan (seven languages), Salinan (two languages), Shastan (four languages), Yanan (two languages), and Yuman (10 languages; in California, Arizona, and Baja California, Mexico), Tequistlatecan (three languages in Oaxaca, Mexico), and Jicaquean (two languages, spoken in Honduras), plus the language isolates Chimariko (extinct), Esselen (extinct), Karok, Washoe, and Seri (northwest Mexico). Tlapanecan (one living language in Guerrero, Mexico, and an extinct one in Nicaragua) was once thought to be connected but is now known to belong to the Otomanguean family of languages. The Hokan hypothesis was also sometimes broadened to include an array of languages in Texas, Louisiana, and northeast Mexico, called the Hokan-Coahuiltecan hypothesis, with Comecrudan (extinct), Coahuiltecan (extinct), Karankawa (extinct), Tonkawa (extinct), and others. An even broader proposed extension was to join it with Siouan and others, called Hokan-Siouan. Proposals even linked Yurumanguí and Quechua of South America, but they received little following.