Hokan Languages
Hokan languages are a number of languages grouped together not as a language family, but as a linguistic stock, a series of hypotheses about distant genetic relationships among language families and isolates. The Hokan stock was first proposed in 1913 by Dixon and Kroeber, remotely relating five language families and isolates in Northern California: Chimariko, Shastan, Pomoan, and, with some caveats, Yana and Karuk. Later that same year, they added Esselen and Yuman. Earlier, in 1905, Dixon had linked Shasta to Achumawi (Palaihnihan). The first Hokan hypotheses were mainly based on five presumed cognate sets for eye, tongue, water, stone, and sleep. The stock was named after the Atsugewi word hoqi, meaning “two.” Later, in addition to cognate sets, Dixon and Kroeber observed some general structural characteristics of the Hokan languages, such as the absence of plural in most nouns, verb suffixes indicating plurality, elaborate sets of instrumental verb prefixes, and local suffixes. Over the next decades, Dixon, Kroeber, Sapir, and others expanded the stock. In 1925, Sapir included a total of fourteen language families and isolates in the stock, subdivided into three branches: (a) Karuk, Chimariko, Shastan, Palaihnihan, Yana, and Pomoan; (b) Esselen and Yuman-Cochimí; and (c) Washo, Salinan, Seri, Chumash, Tequistlatecan (Chontal), and Subtiaba-Tlapanec. These branches extended geographically from northern California to Nicaragua. Kaufman later evaluated the proposed hypotheses in Kaufman 1989 (cited under Later Proposals) and came out in favor of including the following sixteen language families and isolates in the Hokan stock, with some caveats: Pomoan, Chimariko, Yana, Karuk, Shastan, Palaihnihan (Achumawi and Atsugewi), Washo, Esselen, Salinan, Yuman, Cochimí, Seri, Coahuilteco, Comecrudan, Tequistlatecan (Chontal), and Jicaque. However, Kaufman did not present any evidence in his article. In 1997, Campbell assessed in detail the validity of the Hokan hypotheses and pointed out several problems associated with the proposals (see Campbell 1997, cited under General Overviews). The possibility of a Hokan stock has generated wide interest ever since it was first proposed in 1913. It has also been the grounds for many discussions and criticisms due to the difficulties in finding supporting evidence for the stock. Some of the critics point out that the Hokan proposals were established when little data was available on the languages and that the hypotheses rely on data that lacks phonetic and phonological accuracy, a key for comparative historical linguistics. Moreover, some critics stress that the data on which the hypotheses are based are unreliable, as they were collected from semi-fluent speakers who had not used their language in decades, and who were speakers of multiple local languages, thus possibly confusing their vocabularies. Furthermore, Mithun points out that the languages hypothesized to belong to the Hokan stock have been in close contact for centuries, making it difficult to distinguish cognates from ancient loans. One of the main problems with the Hokan stock, however, lies in its antiquity. It is estimated to be as old as Proto-Indo-European, which complicates establishing genetic relationships among the languages. Overall, the main problem with Hokan lies in that it is based on very little evidence.