[Paper presented at the Annual Meeting of the Society for the Study of the Indigenous Languages of the Americas (SSILA), in Chicago, January 2008]
Although Macro-Jê remains one of the most controversial language groupings of South America,
recent studies have made substantial contributions towards determining the precise scope of the stock and uncovering new pieces of evidence for the genetic relationship among several of the language families included therein (Ribeiro & van der Voort 2005, Adelaar 2005, Ribeiro 2005). This paper provides an example of how recent advances in Macro-Jê historical linguistics may contribute to a better understanding of synchronic phenomena, offering a diachronic explanation for the emergence of a morphological pattern of root-internal consonant alternation in Karajá. Karajá was one of the last families to be included in the stock (Davis 1968), but the evidence then presented was less than conclusive. Ongoing research, however, has revealed a number of additional cognates, most of which corroborate the phonological correspondences detected by Davis in his pioneering work. Although Karajá, with its fairly complex verb morphology, is superficially rather different from mostly isolating Jê languages, most of its “core” inflectional and derivational morphological inventories have likely cognates in Proto-Jê, including personal prefixes and nominalizing suffixes.
As in Jê, the distribution of the morphemes in Karaja follows a rather linear, agglutinative pattern, with one major exception—Karajá presents, among its nominalizing strategies, a process of consonant replacement, by which a velar stop in the last syllable of a verb root is replaced with /r/ in the corresponding nominal form: koka ‘to grind’ > kora ‘the action of grinding’; ka 'to dig' > ra 'the action of digging'; ky 'to eat' > ry 'the action of eating', etc. As it turns out, this alternation can be traced back to the suffixation of a nominalizing morpheme *-r, cognate with a Proto-Jê nominalizing suffix *-r , which is synchronically productive in most Jê languages (cf. Proto-Jê *ku 'to eat', ku-r 'the action of eating'). As illustrated below by the root ky ‘to eat’, a cognate of Proto-Jê *ku, a succession of independently-demonstrable diachronic processes can account for the emergence of the current pattern of consonantal replacement in Karajá:
(1) a process of vowel copying (“echo vowel”), common throughout Macro-Jê: *ky-r > *kyry
(2) a process of (vowel weakening leading to) cluster formation: *kyry (> *kAry) > *kry
(3) a process of cluster simplification: *kry > *ry
Unlike its cognates in most Jê languages, *-r is not a productive morpheme in Karajá, surviving
only as a relic in examples involving the pattern of consonant alternation described in this study. By "encasing" the original suffix in the root, the phonological processes described above preserved a further piece of evidence for the genetic relationship between Karajá and the Jê family, as part of the Macro-Jê stock.
References:
Adelaar, Willem. 2005. Relações externas do Macro-Jê: O caso do Chiquitano. To appear in Stella Telles (editor), Atas do V Encontro Macro-Jê.
Davis, Irvine. 1968. Some Macro-Jê relationships. IJAL, 34:42-47.
Ribeiro, Eduardo. 2005. A reconstruction of Proto-Jê (and its consequences for the Macro-Jê hypothesis). Paper presented at the International Symposium on Historical Linguistics. Belem (Brazil): MPEG & UFPA.
Ribeiro, Eduardo & Hein van der Voort. 2005. The inclusion of the Jabuti family in the Macro-Je stock. Paper presented at the International Symposium on Historical Linguistics. Belem (Brazil): MPEG & UFPA.