The internal classification of Mixtec is controversial. Many varieties are mutually unintelligible and by that criterion separate languages. In the 16th century, Spanish authorities recognized half a dozen lenguas comprising the Mixtec lengua. (See #Classical Mixtec.) It is not clear to what extent these were distinct languages at the time. Regardless, the colonial disintegration of the Mixtec nation and resulting isolation of local communities led to the rapid diversification of local dialects into distinct languages. Below are some attempts at Mixtec classification by various scholars.
Geographic divisions
Josserand (1983:106) lists 5 major geographic (not linguistic) divisions of Mixtec, which together cover a total of about 25,000 square kilometers. Enclaves of Amuzgo, Trique, Cuicatec, Ixcatec, and Chocho speakers are scattered nearby.
Puebla Mixtec
Guerrero Mixtec
Mixteca Baja
Mixteca Alta
Mixteca de la Costa
Colonial divisions
De los Reyes, in his Arte de Lengua Mixteca (1593), spoke of half a dozen lenguas in the Mixtec lengua. To these, his contemporaries added the dialects of Guerrero:[1]
Josserand found that native mundane writing of the colonial era corresponded well to de los Reyes; based on phonological and orthographic consistencies, she divides the dialects into five groups, as follows:
the Baja area around Huajuapan (though there were multiple varieties in Baja, more than de los Reyes recognized)
The following classification is given by William R. Holland (1959), as cited in Josserand (1983:134-135). This preliminary classification is a glottochronological study of the dialects of 22 Mixtec and 4 Cuicatec towns.
Holland (1959) also gives 3 areal groupings for these zones.
Costa: Zones 1, 2
Alta: Zones 3, 4, 5, 6, 7
Baja: Zones 8, 9, 10, 11
However, Josserand (1983) states that these groupings are based on flawed methodologies, including a faulty conception of the geographical layout of the Mixteca. Many towns that Holland listed as Baja are in fact Alta, and vice versa.
Arana (1960)
The following classification is given by Evangelina Arana-Osnaya (1960:257), as cited in Josserand (1983:137).
Cornelia Mak and Robert Longacre (1960) is the first reconstruction of Proto-Mixtec, which is the ancestor of Mixtec proper as opposed to Mixtecan. Below is a classification inferred from Mak & Longacre (1960) by Josserand (1983:142). 9 groups and a total of 28 towns are given.
The following classification, based on "archaeological, ethnohistorical and modern information in his delimitation of interaction spheres within the Mixteca",[5] is given by Richard Spores in The Mixtec Kings and Their People (1967), as cited in Josserand (1983:128). A total of 18 dialects are given.
Apoala, Apasco, Sosola; eastern frontier with Chinantec, Cuicatec, and Zapotec
The following classification is given by C. Henry Bradley (1970), as cited in Josserand (1983:132). A total of 11 languages are given. His classification was most likely based on SIL International's mutual intelligibility surveys.
Egland & Bartholomew find 29 groups at a 70% mutual-intelligibility level. The towns they tested are the following, grouped at 60% intelligibility; a question mark indicates that intelligibility testing had not been done with non-neighboring varieties.
The classification of Ethnologue is largely based on Egland & Bartholomew. There is no sub-classification, only a list of 52 varieties,[6] though these are reported to have a great range of intelligibility, from essentially none[7][8] to 90% or higher.[9][10]