Although final consonant correspondences between Proto-Jê and Maxakalí tend to be straightforward, in some cases the comparative data seem to suggest that Maxakalí has actually undergone a major innovation, through the accretion of final consonants to originally vowel-final roots. This hypothesis is further corroborated by the study of loanwords from both Portuguese and Old Tupí.
In my comparative studies of the Macro-Jê stock (Ribeiro 2006, 2007, etc.), I've noticed that, when compared to Proto-Jê, final consonants in Maxakalí fit into three different categories:
(1) Consonants which correspond clearly with final consonants in Proto-Jê and actual Jê languages:
(1a) Proto-Jê *prãm 'hunger' :: Maxakalí putup
(1b) Proto-Jê *j-õt 'to sleep' :: Maxakalí -yõn
(2) Consonants corresponding with final consonants which, although absent in most Jê languages, were actually present in Proto-Jê (as suggested by morphophonemic alternations found in daughter languages). For instance, while most Jê languages have codaless reflexes of Proto-Jê *njo(p) 'to hang' (cf. Apinajé (Northern Jê) ɲdʒo; Oliveira 2005:41), Kaingáng has three different reflexes: sa, sàv, and sam (Wiesemann 2002); sàv can arguably be seen as the underlying form, whereas sam is likely the result of the suffixation of the causativizer *-n to sàv. Combined with such internal evidence, the occurrence of a final consonant in the Maxakalí cognate helps corroborate the reconstruction of a final *p for the Proto-Jê form:
(2a) Proto-Jê *njo(p) 'to hang' :: Maxakalí xup
(3) Consonants which seem to have been added to an otherwise vowel-final root. That would have been the case of kutex 'to sing' and yõg 'possession marker':
(3a) Proto-Jê *ŋrɛ 'to sing, dance' :: Maxakalí kutex
(3b) Proto-Jê *j-õ 'possession marker' :: Maxakalí yõg
While examples fitting categories (1) and (2) are, from a comparative viewpoint, rather straightforward, examples of the last category (3) are somewhat challenging. After all, accretion of phonological material is harder to argue for than its opposite. If consonant accretion is to be seen as a regular diachronic process, one needs to detect the underlying rules determining which (and, hopefully, how, and why) consonants were inserted, demonstrating that consonants were not added at random, but rather following regular phonological rules.
A major problem is the fact that, given the current level of knowledge of Macro-Jê, one cannot in principle be absolutely sure whether examples of the 3rd type actually involve accretion in Maxakalí, rather than being another instance in which Maxakalí turns out to be more conservative than (most) Jê languages, such as in (2). Fortunately enough, however, there are loanwords which seem to clearly prove that, historically, consonants were indeed added to vowel-final words in Maxakalí. Although the study of loanwords has played an important role in the analysis of Maxakalí phonology, I'm unfamiliar with any previous work dealing with the insertion of final consonants1 (that may very well be a gap in my knowledge of the linguistic literature on Maxakalí).
Loanwords commonly discussed in the literature involve Portuguese paroxítonas words (i.e., those whose stress falls on the penultimate syllable), whose final vowels (if any) are eliminated in the Maxakalí corresponding forms: e.g. hetanat 'picture' (< Portuguese retrato). The examples which will concern us here, on the other hand, involve vowel-final words whose stress falls on the last syllable (either because that was the situation in the donor language (e.g. 'blanket'), or because the stress was switched to the last syllable (e.g. 'money', 'horse'), a rule which seemed to have operated with old loans). That includes a few Tupí loans (apud Ribeiro 2009), in addition to Portuguese ones:
(4) ãmãnex 'priest' (< Tupí abaré)
(5) ãmix 'needle' (< Tupí abĩ)
(6) tayũmak 'money' (< Tupí itajúba)
(7) kamanok 'horse' (< Portuguese cavalo, possibly via Tupí cabarú)
(8) kapex 'coffee' (< Portuguese café)
(9) komenok 'blanket' (< Portuguese cobertor, probably via colloquial cobertô)
(10) kapitõg 'chief' (< Portuguese capitão)
Even within such a limited sample, a few tendencies can be observed—words ending in front vowels tend to add a final palatal consonant, words ending in back oral vowels tend to add a final velar stop, etc. More importantly still, loanword behavior tends to mirror our hypothetical accretion of final consonants to inherited vowel-final roots (cf. kapex 'coffee', kutex 'to sing'; kapitõg 'chief', yõg 'possession').
Although many questions remain unanswered (e.g., is the process synchronically productive?), the loanwords listed above (4-10) seem to strongly corroborate the hypothesis that consonants were indeed added to vowel-final stems in Maxakalí (probably to fulfill syllabic constraints), their properties being determined by the quality of the preceding vowel. Besides the comparative value of such information, it may contribute to reveal yet another facet of the oft-mentioned "intimacy" between consonantal and vowel features in this fascinating language.
References:
Oliveira, Christiane Cunha de. 2005. The language of the Apinajé people of Central Brazil. Ph.D. dissertation, University of Oregon.
Popovich, A. Harold, and Frances B. Popovich. 2005. Dicionário Maxakalí–Português. Cuiabá: Sociedade Internacional de Lingüística.
Ribeiro, Eduardo Rivail. 2006. A reconstruction of Proto-Jê (and its consequences for the Macro-Jê hypothesis). Paper presented at the symposium "Advances in Native South American Historical Linguistics", during the 52nd International Congress of Americanists (Seville, Spain).
Ribeiro, Eduardo Rivail. 2007. Eastern Macro-Jê: A hypothesis on the internal classification of the Macro-Jê stock. Ms.
Ribeiro, Eduardo Rivail. 2009. Tapuya connections: language contact in eastern Brazil. Liames, 9, p. 61-76.
Wetzels, W. Leo. 2009. Nasal harmony and the representation of nasality in Maxacalí: evidence from Portuguese loans. In Calabrese, Andrea & W. Leo Wetzels, Loan phonology, p. 241-270. Amsterdam: John Benjamins.
Wiesemann, Ursula. 2002. Dicionário bilíngüe Kaingang–Português. Curitiba: Editora Evangélica Esperança.
(Página criada em 1303402648|%e %b %Y, %H:%M %Z|agohover. Última edição em 1308975679|%e %b %Y, %H:%M %Z|agohover; comentários e informações adicionais são sempre bem-vindos.)