CC BY-NC-ND
14
Audrey Nikulin
University of Brasilia; andre.n.guzman@gmail.com
A reconstruction of Proto-Je phonology and lexicon1
In this work, I examine the sound correspondences between Proto-Cerrado (Nikulin 2017) and Proto-Southern Je (Jolkesky 2010) and offer a phonological reconstruction of Proto-Je, the proto-language of the most diverse subgrouping within the Macro-Je language stock. I reconstruct 11 consonants and 19 vowels for this proto-language. I also claim that */CrVC/ was the maximal syllable structure in Proto-Je with some further restrictions on its complex onsets (only */pr, mr, kr, qr/ were allowed). I reconstruct a shielding allophony pattern to Proto-Je, according to which nasal onsets would have had post-oralized allophones before oral nuclei. The discussion on Proto-Je phonology is followed by a sample of Proto-Je lexicon.
Keywords: Je languages, Macro-Je languages, language reconstruction, comparative method.
1. Introduction
Since the pioneer work of Davis (1966), which remains the only published attempt at a reconstruction of Proto-Je to the present day, numerous descriptive studies on individual Je languages have been carried out, contributing to a significant improvement in the state-of-the-art of Je linguistics. Recent comparative work on low-level branches of Je, such as Southern Je, Northern Je, Central Je, and Cerrado (Wiesemann 1978, Jolkesky 2010, Nikulin & Salanova forthc., Nikulin 2016, 2017), has enabled the use of intermediate reconstructions in comparative studies on higher levels. In this paper, I offer a revised reconstruction of Proto-Je, taking into account the research on Je produced over the last 50 years.
I follow the principles of bottom-up reconstruction, which means that my reconstruction of Proto-Je is based on the reconstructions of its daughter proto-languages, Proto-Cerrado (Nikulin 2017) and Proto-Southern Je (Wiesemann 1978, Jolkesky 2010); the former, in turn, is based on the reconstructions of Proto-Northern Je (Nikulin 2016, cf. also Nikulin & Salanova forthc.) and Proto-Central Je (Nikulin 2017). This allows to filter out recent phonetic, morphological, and semantic innovations, which is especially important in light of the conservativeness traditionally attributed to Northern Je (cf. Ribeiro & Voort 2010: 560, Carvalho 2016: 64, Pache 2018); it is well known that over-reliance on the data of a given
1 I gratefully acknowledge the financial support of CAPES (Coordenaçâo de Aperfeiçoamento de Pessoal de Nivel Superior) during my doctoral studies. I also thank Andrés Pablo Salanova, Mário André Coelho da Silva, and Matthias Pache for valuable input that helped improve this paper. Furthermore, its text has benefited from the comments of Juho Pystynen, Mikhail A. Zhivlov, and Lev Michael, to whom I am grateful. I warmly thank my Jê-speaking friends, especially Bepo Metyktire and Eneida Xerente, for teaching me bits of their languages. Finally, I thank the audience at Amazónicas VI (Tabatinga), where a much earlier version of this paper was presented on May 27th, 2016.
I use the following abbreviations throughout the paper: PCerr = Proto-Cerrado, PCJ = Proto-Central Jê, PJ = Proto-Jê, PNJ = Proto-Northern Jê, PSJ = Proto-Southern Jê, ABL = ablative, ACT = active, ADESS = adessive, ADVT = advertence prohibitive, ALL = allative, CAUS = causative, COP = copula, DU = dual, ERG = ergative, HABIT = habitual, INSTR = instrumental, INTR = intransitive, LOC = locative, MALEF = malefactive, NEG = negative, NF = non-finite, NMLZ.AG = agent nominalization, PL = plural, SG = singular, STAT = stative, ATR = advanced tongue root.
Journal of Language Relationship • Вопросы языкового родства • 17/2 (2019) • Pp. 93-127 • © The authors, 2019
language or language group has repeatedly led historical linguists to wrong reconstructive solutions.2 It is important to emphasize that bottom-up reconstruction does not exclude occasional use of external material, which may be invoked in situations when internal data offer conflicting or incomplete evidence.
My working model of the Jê phylogenetic tree (without Xakriaba and Akroa, whose position within Central Jê is uncertain) is presented in Figure 1.3
Figure 1. Internal structure of the Jê family
The remainder of this paper is organized as follows. Section 2 contains an outline of the phonological systems reconstructible to Proto-Cerrado (subsection 2.1) and Proto-Southern Je (subsection 2.2). In section 3, I examine the sound correspondences between Proto-Cerrado and Proto-Southern Je and come up with a reconstructive proposal for Proto-Je. Onsets (subsection 3.1), nuclei (subsection 3.2), and codas (subsection 3.3) are dealt with separately. A list of the most reliable Je etymologies is provided in section 4. Section 5 summarizes the paper and lists several unresolved issues in comparative Je linguistics that should be subject to future research.
2 One noteworthy example is the bias towards Finnic data in Uralic comparative studies, which prevented the scientific community from providing a correct account of several important vowel correspondences for decades; only very recently (Aikio 2012, 2015) did it become clear that Finnic vocalism is innovative in a number of respects. Another well known case was the over-reliance on Sanskrit data in mid-nineteenth century Indo-European comparative studies, which led scholars to a long-lasting misconception that Proto-Indo-European vocalism was identical to that of Sanskrit (Benware 1974: 66-81).
3 The tree in Figure 1 was initially obtained through the application of StarlingNJ, a distance-based lexicostatistical method (see Kassian 2015: 5 for details), to 110-item wordlists of 16 extant Je varieties (including several dialects of Kaingang and Mebengokre but excluding Krikati and Ingain). The annotated wordlists are scheduled to be published online at <http://starling.rinet.ru/new100> in 2019 as a part of the Global Lexicostatistical Database. For each node identified by the StarlingNJ algorithm, shared lexical, phonological and/or morphosyntactic innovations can be presented. For reasons of space, I do not discuss the internal classification of Je in what follows, hoping to revisit the subject in future publications.
2. Proto-Cerrado and Proto-Southern Je
In this section, I describe in detail the Proto-Cerrado (PCerr) and Proto-Southern Je (PSJ) reconstructions adopted in this paper.
2.1. Proto-Cerrado
The Cerrado branch of the Je family is constituted by two subbranches, Northern Je4 and Central Je. I use the reconstruction of Proto-Cerrado by Nikulin (2017), according to which the phonological inventory of this proto-language comprised 12 consonants and 15 vowels. PCerr consonantal phonemes are shown in Table 1 below.
Table 1. Proto-Cerrado consonants.5
labial coronal palatal velar
oral stops *p *t *c *k
nasal stops *m *n *ñ
approximants *w *r *j 6
In oral environments, underlying nasal onsets were realized as postoralized (*mb-, *nd-, *nf-, *yg-). In nasal environments, the contrasts *j/*n and *u/*y were neutralized in favor of *n and respectively. All these allophonic processes are reflected in the transcription system used in this work. The pronunciation of PCerr consonants in coda position is uncertain, but at least *m is likely to have had a preoralized allophone *[-bm] in oral environments (unless followed by an echo vowel, for which see below).
The 15 vowels of PCerr are listed in Table 2 below.
4 For convenience, I will use "Northern Je" and "Proto-Northern Je" ("PNJ") as synonyms of "Northern Je proper" and "Proto-Northern Je proper" (that is, excluding Panara/Southern Kayapo) throughout this paper. I will refer to Northern Je lato sensu as "Northern Je-Panara" whenever necessary.
5 In this work, I diverge from my earlier practice of representing linguistic data using the Unified Transcription System (UTS) and adhere to a modified version of the International Phonetic Alphabet (IP A) instead. The main differences between the transcription system used in this work and IPA are: r = IPA r; n = IPA j; c stands for any voiceless palatal, alveo-palatal or postalveolar stop, or affricate; f stands for any voiced palatal, alveo-palatal or postalveolar stop or affricate; x stands for a voiceless alveo-palatal or postalveolar fricative; 3, o, e, 3, o, e are open-mid or simply mid; 3, o, e are close-mid; 3, 3, y, y are central or back unrounded vowels (open-mid, close-mid, near-close, and close, respectively); i, u, y are falling lowering diphthongs. In Karaja data, the circumflex stands for the [+ATR] feature. The apostrophe in PCerr reconstructed forms is used to indicate that the echo vowel is absent (see Nikulin 2017: 168-169 on this phenomenon).
6 Although the reflexes of this segment in all Cerrado languages are stops (MBG g, API, TIM, TAP, KSJ, PNR, XER k, XAV ?), it patterns phonologically with PCerr *j and *w in a number of respects (see Nikulin & Salanova forthc. for discussion), suggesting that at least phonologically it belonged to the natural class of approximants in Proto-Cerrado. Fortition of approximants, especially in stressed syllables, is a recurrent trend across Cerrado languages: the development *w > *b is known from Proto-Northern Je-Panara, and the development *j > *f is hypothesized for Proto-Cerrado (in all environments), for Proto-Northern Je-Panara (in stressed syllables), and for Proto-Trans-Tocantins (in unstressed syllables preceding *3, *u). That way, it does not seem entirely impossible that the segment in question was an approximant not only phonologically, but also phonetically in Proto-Cerrado.
Table 2. Proto-Cerrado vowels.
oral
nasal
front unrounded non-front unrounded7 back rounded
close *i *y *u
close-mid *ê *9 *ô
open-mid *e *9 *o
open *a
front non-front back
unrounded unrounded rounded
*
An already rich vocalic system reconstructible to Proto-Cerrado became even richer in ProtoNorthern Je, where *a and *u emerged due to positional nasalization.
Apart from the monophthongs, at least two complex nuclei, *wa and *ja,8 are reconstructible to Proto-Cerrado. Note that new important etymologies have been identified since the publication of the previous proposal (Nikulin 2017: 163), corroborating the reconstruction of the diphthongs in question. I list these etymologies in (1) below.
(1) New Cerrado etymologies involving a diphthong
a. PCerr *cwa/*cwa-r' 'to ask' > PNJ *cu/*cwd-r, PCJ *wa/*wa-ri
b. PCerr *jwan' 'nmlz.ag' > PNJ *jwm, PCJ *-kwaj // *-kwa
c. PCerr ^ajwa^jajwa-r' 'to lay.PL, to spill' > PNJ *jaju/*jajwd-r, PCJ *c-a(j)kwa/*c-a(j)kwa-ri
d. PCerr *jgrwa 'moriche' > PNJ *jgrwa ~ *jgru, PCJ *wa:-bu 'moriche stem' (< PCerr *jgrwa-pu > PNJ *jgrwa-pu)
e. PCerr *nokwa 'sternum' > PNJ *nokwa, PCJ *nowa 'in front of'
f. PCerr *norkwa 'home' > PNJ *nurkwa, PCJ *norowa
g. PCerr *krwat 'beak' > PNJ *krw3t, PCJ *wata // *wada
h. PCerr *rwaj-ci 'rib' > PNJ *rwd-ci, PCJ *waj-hi
i. PCerr *rwaj' 'moon' > PNJ *mbyt-rwd, PCJ *waj // *wa
j. PCerr *twam' 'fat' > PNJ *tw3m, PCJ *wam // *wa
k. PCerr *a-ygia/*ygja-c 'to enter.PL' > PNJ *a-qgi ~ *a-rjgje/*ygje-c, PCJ *d-ja/*ja-ci 'to enter.DU'9
l. PCerr *rgja/*rgja-n' 'to enter.caus.pl' > PNJ *jgi ~ *rgie/*rgie-n, PCJ *jaj // *ja 'to enter.CAUS.DU'
m. PCerr *krja/*krja-r' 'to bring up' > PNJ *kri/*kje-r, PCJ *ja/*ja-ri
7 It is uncertain whether non-front unrounded vowels were phonetically central or back.
8 Nikulin (2017) adopted the reconstruction *je. The change in reconstruction is suggested by the near-
symmetry of the reflexes of PCerr *ja and *wa in daughter languages. In Proto-Northern Je, these two diphthongs
yielded *je and *wd in closed syllables (as well as in open syllables that go back to PCerr syllables ending in *-j), but *je ~ *i and *wa ~ *u in originally open syllables (with a so far unclear variation that is sometimes preserved in Proto-Northern Je, see Nikulin 2017: 178-179). In Proto-Central Je, the reflexes of these diphthongs are *ja and *wa, and any PCerr onset is eliminated before the diphthong (except that PCerr *jwa > PCJ *kwa).
9 The etymologies (1k) and (1l) had been erroneously conflated in Nikulin 2017. Although related, they are two clearly distinct verbs. The same semantic and morphological relation holds for their singular counterparts: PCerr *a-jd/*jap-r 'to enter.SG' > PNJ *a-]d/*p-r, PCJ *3-fe/*feb-re; PCerr *jd/*ja-r 'to enter.CAUS.SG' > PNJ *]d/*p-r, PCJ *ne/*ne-re.
n. (?) PCerr *kwa/kwa-r' 'to crack nuts' > PNJ *kukw3/*kukw3-r, PCJ *wa/*wa-ri o. (?) PCerr *krwaj 'parrot' > PNJ *krw3j, PCJ *waj // *wa, *waj-hdrd
In Proto-Cerrado, the onset position could be filled with any single consonant or by a cluster formed by a peripheral (labial or velar) stop and *-r-: *pr, *m(b)r, *kr, *y(g)r. Strikingly, the occurrence of onsetless syllables appears to have been very limited: only pretonic syllables that contained *a or *i as its nucleus could lack a consonantal onset.
Syllables could also have a coda in Proto-Cerrado, which could be followed or not by a so called echo vowel (only word-finally). In my notation, the occurrence of an echo vowel after a consonantal coda is unmarked, whereas the absence of a word-final echo vowel is marked with the apostrophe. The following codas can be reconstructed for Proto-Cerrado: *p, *t, *c, *k, *m, *n, *r, *j, *pr, *m', *n', *n', *r', *j\ *pr'. The complex codas *pr and *pr' occur exclusively in non-finite verb forms, when the non-finite suffix *-r is adjoined to a verbal stem that ends in an underlying labial consonant. Very few examples are available that could substantiate the reconstruction of Proto-Cerrado *c, *n, *y, *j, *n', *n' in coda position.10 The reconstruction of Proto-Cerrado *d, as proposed by Nikulin (2017: 169), is even more problematic in that no *d can be reconstructed for the onset position.
One major issue in Cerrado historical phonology remains entirely unresolved. Namely, some Central Je stems unexpectedly contain nasal nuclei in stems whose cognates in Northern Je (and in other Macro-Je languages, whenever available) display oral nuclei. Examples include PCJ *ne/*ne-re 'to enter.CAUS.SG' (cf. PNJ *fd/*fd-r), *c-3/*c-3-ri 'to stand.CAUS.SG' (cf. PNJ *fa/*fa-r), *nd/*n§m-n 'to weave' (cf. PNJ *cy/*cy-r), *fatd/*fatd-n 'to send' (cf. PNJ *jando/ *jando-r), *c-aco/*c-acom-ri 'to hang' (cf. PNJ *janfd/*janfd-r), *re/*re-me 'to leave.SG' (cf. PNJ *re), *to/*to-rt 'to urinate' (cf. PNJ *ij-tu/*tu-r), *p5/*p3-ri 'to kill.DU' (cf. PNJ *pa/*pa-r 'to kill.SG'), *t3j // *t3 'rain' (cf. PNJ *nda). The origins of nasalization in Central Je have not yet been identified, and it is uncertain how the respective PCerr forms should be reconstructed. I leave this question open and provisionally represent these unclear cases by underlining the nuclei in my PCerr reconstructions (e.g. *jd/*jd-r, *ja/*ja-r', *cy/*cyp-r', *jando/*jando-r', *janfd/*janfdp-r', *re, *tu/*tu-r', *pa/*pa-r', *ndaj').
2.2. Proto-Southern Je11
The reconstruction of PSJ consonantism does not pose serious difficulties, and the systems reconstructed by Wiesemann (1978) and Jolkesky (2010) are almost identical. I accept their reconstruction with minor changes in notation.
Although underlyingly nasal, voiced stops were subject to postoralization in oral environments (*mb-, *nd-, *nf-, *yg-); this allophony pattern is active in both Southern Je languages and is reconstructible to Proto-Cerrado. The preoralization of underlying nasal codas is also attested both in Kaingang and Laklano, though this process is not equally robust in all Southern Je varieties: depending on the dialect and on the place of articulation, either the oral phase or the nasal phase may be omitted. Although clearly reconstructible to Proto-Southern Je, I chose not to represent this allophony in my transcription.
10 Known examples include PCerr *rjgjac 'to enter.PL.NF', *mbyn 'tail', *porj 'arrow, bamboo', *mborj 'lake', *krwaj 'parrot', *mbjan' 'husband', *rgja-n' 'to enter.CAUS.PL.NF', *jwan' 'NMLZ.AG', *j3n' 'sweet', *rgwan' 'feather', maybe also *mbec 'good' and *mben 'liquid'.
11 In this paper, I do not consider Ingain data due to their ill-transcribed nature. Hereinafter the terms "Southern Je" and "Proto-Southern Je" are used in their narrow definition (that is, "Southern Je proper" and "Proto-Southern Je proper", excluding Ingain). Ingain data will need to be subjected to the procedure of restitution (Constenla Umana 2000) before they can be meaningfully used in the reconstruction of Proto-Je.
Table 3. Proto-Southern Jê consonants.
labial coronal palatal velar glottal
oral stops *P *t *c *k *?
fricative *e
nasal stops *m *n *ñ
approximants *w *r *j *h
Regarding the vocalic system of Proto-Southern Je, I have previously argued (Nikulin 2015) that Jolkesky's (2010) reconstruction of PSJ vowels is seriously flawed and put forward a proposal backed up with morphophonological evidence. Note that while my reconstruction of PSJ oral vowels coincides with that of Wiesemann (1978), our reconstructions of PSJ nasal vowels are radically different. 12
Table 4. Proto-Southern Jê vowels.
oral
nasal
front un- non-front back rounded unrounded rounded
close *i *y *u
close-mid *ê *9 *ô
open-mid *e *9 *o
open *a
front unrounded
non-front back unrounded rounded
-i-
The phonotactic restrictions are identical in Kaingang and Laklânô and are easily traced back to Proto-Southern Jê. The maximal syllable structure reconstructible to PSJ is *CrVC. Just like in Proto-Cerrado, only labial and velar stops could form complex onsets with the rhotic: *pr, *m(b)r, *kr, *q(g)r. As for the codas, only voiced segments could occur in that position; approximant codas were obligatorily followed by a non-phonemic echo vowel (phonetically a copy of the nucleus), whereas nasal codas did not trigger the occurrence of any vocalic segment. I leave PSJ echo vowels unmarked in my transcription because their occurrence and quality are fully predictable.
12 Jolkesky (2010: 173) reconstructs yet another nasal vowel for Proto-Southern Je, tracing the correspondence KGG 3 ~ LKL u back to PSJ *y (*m in his transcription). Nikulin (2015: 286) reinterprets it as *o (*5 in the transcription of the original paper). Note, however, that this sound correspondence occurs exclusively in grammatical morphemes, such as the interrogative pronoun KGG n3 ~ LKL nu, the situational marker KGG k3 ~ LKL ku and the subject marker KGG w3 ~ LKL vu. It now seems possible that this sound correspondence may have arisen as a conditioned development of PSJ *3 (in this case Laklano would have innovated by raising its usual reflex *o > u) or PSJ *u (in this case Kaingang would have innovated by unrounding and lowering its usual reflex *u > 3) in unstressed grammatical morphemes.
3. Proto-Jê phonology
In this section, I examine the correspondences between Proto-Cerrado and Proto-Southern Jê onsets (3.1), nuclei (3.2), and codas (3.3), offering in each case a Proto-Jê reconstruction. Some Proto-Jê elements appear to have been preserved in only one constituent branch; in such cases, external evidence from other Macro-Jê languages has also been discussed.
3.1. Onsets
The correspondences between the onsets reconstructible for PCerr and PSJ are mostly trivial except for the palatal place of articulation, as shown in Table 5.
Table 5. Proto-Jê onsets and their reflexes in Proto-Cerrado and Proto-Southern Jê.
Proto-Jê Proto-Cerrado Proto-Southern Jê
*/p/ *p */p/ *p */p/ *p
*/m/ *mb, *m */m/ *mb, *m */m/ *mb, *m
*/w/ *w */w/ *w */w/
*/pr/ *pr */pr/ *pr */pr/ *pr
*/mr/ *mbr, *mr */mr/ *mbr, *mr */mr/ *mbr, *mr
*/t/ *t */t/ *t */t/ *t
*/n/ *nd, *n */n/ *nd, *n */n/ *nd, *n
*/r/ *r */r/ *r */r/ *r
*/c/ */c/ *c */e/
*/n/ *nf */n/ *nf */c/ *c
*j */j/ *j */j/ */j/ *j
*n *n */n/ *n
*/k/ *k */k/ *k */k/ *k
*/kr/ *kr */kr/ *kr */kr/ *kr
*/0/ *0 (?) ? ?
Most of the correspondences in Table 5 are identity correspondences, thus posing no difficulty for the reconstruction of the respective Proto-Je segments. For reasons of space, the identity correspondences between Proto-Cerrado and Proto-Southern Je onsets will not be examined in this section; numerous examples are available in Section 4.
Regarding the origin of the infrequent PCerr *uj, it is unknown which segment should be reconstructed to Proto-Je due to the fact that the tokens that contain this phoneme in Proto-Cerrado lack known cognates in Southern Je. Nevertheless, it is certain that these tokens are of Je origin, given that they have secure external cognates in other Macro-Je languages, where PCerr *u corresponds to glottal segments (Maxakali h; Krenak h; Jabuti ?; cf. Nikulin & Silva, forthc.). For lack of a better solution, I reconstruct *0 (> PCerr *j) for Proto-Je, but it should be kept in mind that other possibilities (e.g. or *h) cannot be totally ruled out at present. Note that not a single onsetless monosyllabic root morpheme reconstructs to Proto-Cerrado, making it possible to stipulate an epenthetical origin for PCerr j
Palatal onsets. I claim that Proto-Cerrado is conservative with respect to the palatal onsets of Proto-Je. Conversely, Proto-Southern Je would have undergone a chain shift: as shown in (2-3) below, PJ *nj > PSJ *c, whereas PJ *c > *Q.
(2) Proto-Je *nj > Proto-Southern Je *c
a. PJ *njy 'bee, bumblebee' > PCerr *njy, PSJ *cy
b. PJ *njo 'to hang.SG' > PCerr *njo/*njop-r', PSJ *c3/*ca-j (vb. *c3-m)
c. PJ *nju 'to heat, to dry' > PNJ *nju/*nju-n 'to dry in the sun'; PSJ *cu-r (vb. *cu-n) 'to heat by the fire', *jaj-cu/*jar-cu-r, *jaj-cu-r (vb. *jaj-cu-n) 'to roast'
d. PJ *nje 'to tickle' > PNJ *nje/*nje-r, PSJ *ku-ce/*ku-ce-j
e. (?) PJ *nja 'to bite' > PCerr *nja/*nja-r', PSJ *ca 'game (animals); war'
f. (?) PJ *njum' 'dirty' > PNJ *njum 'dirty', PSJ *coj 'egg yolk; larvae's food'
g. (?) PJ *nji 'mother' > PNJ *nji, PSJ *ci 'old'
(3) Proto-Je *c > Proto-Southern Je *d
a. PJ *cy 'to weave' > PCerr *cy/*cyp-r', PSJ *Qy
b. PJ *cym 'seed' > PCerr *cym, PSJ *Qy (vb. *Qy-n)
c. PJ *coj 'leaf' > PCerr *coj', PSJ *dej (vb. *de-n)
d. PJ *jace ~ *jVmce 'nest' > PCerr *jace, PSJ *j3jQe (vb. jjde-n)
e. PJ *-ci(C) 'bandage, sling' > PNJ *ja-ci, *ka-ci, PSJ *-6i (vb. -di-n)
Note that PJ *c- is also reconstructible as the allomorph of the third person marker used with *j-initial stems. As suggested by Ribeiro (2004: 95-96, 2011: 109, fn. 8), Proto-Southern Je appears to have fossilized an earlier third person prefix *c- in some stems.13
f. PJ *j3, 3 (= third person) *c-3 'bitter' > PCerr j, PSJ *&3 (vb. *da-j)
h. PJ *jara, 3 *c-ara 'wing; armpit' > PNJ *jara 'wing', *jara-kre 'armpit', PSJ *Qar 'wing' (but *jara 'armpit', vb. *jara-n)
13 It is worth observing that the stems that underwent this fossilization process are those that typically have an inanimate internal argument. This semantic peculiarity may have contributed to an elevated frequency of occurrences of these stems with a third person prefix in Proto-Je discourse when compared to other *j-initial stems, thus enabling the reanalysis. Another comment to be made is that this fossilization trend continued in Kaingang even after the dissolution of Proto-Southern Je: cf. PJ *ji, 3 *c-i 'to lay.SG' > PSJ *ji, 3 *9-i > KGG fyi, LKL ji, 3 d-i. In other instances, on the contrary, Kaingang generalized the uninflected form: PSJ *jama-j, 3 *0-ama-j 'to listen' > KGG jeme-j, LKL nama-j, 3 d-ama-j; PSJ *jo, 3 *6-o 'for' > KGG jo, LKL jo, 3 d-d (Wiesemann 1978: 209).
i. PJ *ju, 3 *c-u 'pus' > PCerr *jup-r, PSJ *6o (vb. *6o-m)
j. PJ *je, 3 *c-e (~ *jen, 3 c-en) 'thread, knot, cloth' > PNJ *je, PSJ vb. *6e-n 'to spin thread'
k. PJ *ja, 3 *c-a 'to put vertically.SG' > PCerr *ja/*ja-r', PSJ *6a (vb. *6a-r) 'to be put vertically'
In addition, Proto-Southern Je depalatalized PJ *n (the nasal allophone of */j/) to PSJ *n, merging it with PSJ *n < PJ *n, as shown in (4) below.
(4) Proto-Je *n > Proto-Southern Je *n
a. PJ *ni 'meat' > PCerr *m, PSJ *ni
b. PJ *ny 'to sit.SG' > PCerr *ny/*nyp-r, PSJ *ni (vb. *ni-m)
c. PJ *nijd 'smoke' > PCJ *nije, PSJ *nijd (vb. *nija-y)
d. PJ *nija 'nose' > PCerr *nija-kre, PSJ *nija
e. PJ *numje(C) 'female breast' > PNJ *nomje, PSJ *nuyje 'female breast, milk'
f. PJ *ut/*nut 'to sleep' > PCerr *rSot/*not, PSJ *nur
g. PJ *ni(m)- 'hand (in compounds)' > PCerr *ni(m)-kra 'hand', PSJ *ni- (as in *ni-pe 'to wash hands')
The fact that Proto-Cerrado retains the original configuration of the inventory at the palatal point of articulation is suggested by the typology of the directionality of sound changes (from a cross-linguistic perspective, it is more plausible to posit a fricativization development such as *c > *6 than an unconditional fortition of a fricative), as well as by external data from Maxakali, Krenak, and Karaja, in which languages PJ *c corresponds to stops or affricates (Maxakali tp, Krenak kj; Karaja d, palatalized tf; cf. Nikulin & Silva, forthc.).
Word-medial clusters. Sometimes voiced segments are found in Proto-Southern Je stems where, judging by their Proto-Cerrado cognates, voiceless consonants would be expected. I hypothesize that these apparently aberrant correspondences instantiate a sound law according to which voiceless stops became voiced (*/t, c/ > */n, j/) Proto-Southern Je if preceded by an etymological coda (the coda itself may disappear). This rule ceased to be productive and no longer applies in new formations (compounds, reduplicated plurals etc.). Some examples are provided in (5).
(5) Voiceless consonants yield nasals in clusters in Proto-Southern Je
a. PJ *nuctd 'tongue' > PCerr *nojto, PSJ *nuna
b. PJ *k(r)Vmtym' 'capybara' > PCerr *kumtym', PSJ *kryrndyr
c. PJ *kapnja 'to chew' > PCerr *kapnja/*kapnja-r', PSJ *kaja
d. PJ *nje 'to tickle' > PNJ *nje/*nje-r, PSJ *ku-ce/*ku-ce-r, pl. *ku-r-je/*ku-r-je-r
In a number of instances, it is unknown whether the Southern Je words have a Je etymology, but alternations attested in derived forms appear to stem from the same diachronic process:
e. PSJ *ku-cir (vb. *ku-cin) 'to be roasted in fire' ^ pl. *ku-r-jir (vb. *ku-r-ji-n)
f. PSJ *ku-cor (vb. *ku-con) 'to be pierced' ^ pl. *ku-r-jor (vb. *ku-r-jo-n)
g. PSJ *cin 'to glean' ^ antipassive *jdr-jin
h. PSJ *ce/*ce-r 'to tie' ^ *kar-je/*kar-je-r, pl. *kyr-je/*kyr-je-r 'to tie, to make a knot'; *jyr-je/*jyr-je-r 'to mend clothes'; *ar-je/*NP jdr-je 'trap with a maize bait', (?) *krir-je/*krir-je-r 'to scratch; to go berserk'
3.2. Nuclei
The correspondences between PCerr and PSJ nuclei are significantly less straightforward than those between the onsets of these reconstructed languages. They are listed in Table 6 below; tentative reconstructions of Proto-Je segments are included.
Table 6. Proto-Je nuclei and their reflexes in Proto-Cerrado and Proto-Southern Je.1
PJ PCerr PSJ PJ PCerr PSJ PJ PCerr PSJ
*a *a *a
*e *e *e *9 *o *a *o *wa *a
*e *e *e *9 *9 *9 *o *o *9
*i *ja *e *y *9 *9 *o *u *o
*i *i *i *y *y *y *u *u *u
¡<u ■k ¡<u ■k *1
*1 *1 *1
! ^ ■k *9 ! ^ k
■k *y, *9+ *1
*o *o ! ^ k
*u *o *u
+ *y after a palatal onset, *S elsewhere
As can be seen from Table 6, I reconstruct a system with five contrastive vowel heights for Proto-Je, in contrast with the systems of Proto-Cerrado and Proto-Southern Je with only four vowel heights. Although systems with five contrastive heights are cross-linguistically rare (Ladefoged & Maddieson 1990: 95-96),15 it is of notice that one modern Macro-Je language, Karaja, has also been documented to have five contrastive vowel heights (Ribeiro 2012: 86), though in featural terms Karaja vowel heights are better understood as an interaction between a ternary height contrast and a privative [ATR] feature.
If this reconstruction is accepted, the evolution of PCerr and PSJ nuclei can be described in the following way.
• From Proto-Je to Proto-Cerrado (ordering uncertain):
(a) chain shift: *y > *d; *d > *o; *o > *wa;
(b) diphthongization of *i > *ja;
(c) raising of *u > *u;
(d) lowering of *y > (except after palatals) and of *u > *o (unconditionally).
14 I remind the reader that the transcription system adopted here differs from IPA in important ways (see fn. 5). Most importantly, the characters <y, y, y> stand here for unrounded non-front (central or back) vowels.
15 Known cases include Austro-Bavarian as spoken in Amstetten (Traunmuller 1982) and Kensiu (Bishop 1996).
• From Proto-Je to Proto-Southern Je (order a ^ b ^ c; ordering of d and e uncertain):
(a) lowering of *d > *a;
(b) unrounding of *o/*o/*o > *a/*a/*a;
(c) lowering of near-close vowels: *i > *e; *u > *o; *r > *d (accompanied by a push chain lowering: *d > *d; *d > *a; *a > *a);
(d) raising of *e > *l;
(e) fronting of *y > *i.
I do not reject other possible interpretations of the sound correspondences exposed above; the scenario detailed here is the most parsimonious I was able to find. Future studies of external correspondences will undoubtedly contribute to corroborate or reject at least some of my claims.
It must be noted that Proto-Southern Je close-mid and open-mid oral vowels are further lowered (*a, *o, *e, *3, *o, *e > *d, *o, *e, *a, *a, *a, respectively) in syllables with nasal codas as a result of a synchronically active process. For space reasons, this process cannot be discussed here at length; the reader is referred to Nikulin (2015) for details.
Cognate sets that instantiate the aforementioned sound correspondences are provided below in (6-24).
(6) Proto-Je *a > Proto-Cerrado *a, Proto-Southern Je *a
a. PJ *par 'foot' > PCerr *par, PSJ *pan
b. PJ *pan 'arm, branch' > PCerr *paj', PSJ *pa
c. PJ *mba 'to hear, to understand' > PCerr *mba/*mba-r', PSJ *ma/*ma-r
d. PJ *mba 'liver' > PCerr *mba, PSJ *td-ma
e. PJ *-mbar' 'to be afraid' > PCerr *p3-mba, PSJ *ka-mar, *mu-mar
f. PJ *wa 'to walk' > PNJ *wa 'to walk, to live', PSJ *wa 'stat' (aspect marker)
g. PJ *ra 'to hit' > PNJ *ku-ra/*ku-ra-n, PSJ *ra/*ra-r/*ra-n
h. PJ *kapnja 'to chew' > PCerr *kapnja/*kapnja-r', PSJ *kaja
i. PJ *jara 'armpit; wing' > PNJ *jara 'wing', *jara-kre 'armpit', PSJ *jara (vb. *jara-n) 'armpit', *6ar 'wing'
j. PJ *ja 'to stand.SG' > PCerr *ja/*ja-m', PSJ *ja/*ja-r
k. PJ *ja 'to put vertically.SG' > PCerr *ja/*ja-r', PSJ *6a (vb. *6a-r) 'to be put vertically' l. PJ *nija 'nose' > PCerr *nija-kre, PSJ *nija
m. PJ *jan 'to eat.INTR, food' > PCJ *caj // *ca, PSJ *ja/*ja-r/*ja-n (vb. *ja-n)
n. PJ *kaj' 'basket' > PNJ *kac, PSJ *kaj (vb. *kan)
o. PJ *kra 'offspring' > PCerr *kra, PSJ *kra (vb. *kra-y)
p. PJ *jandd 'to send' > PCerr *jando/*jando-r', PSJ *jana/*jana-r
q. PJ *jare (~ *jaren) 'root' > PNJ *jare, PSJ *jare (vb. *6-are-n)16
16 In examples (6q-s), one could be tempted to reconstruct PSJ *-a- rather than PSJ *-a- in unstressed (nonfinal) syllables, given that the vowel in question is reflected as Kaingáng o (jdre 'root', jon-ko 'door', ko-ka 'wind') rather than e (usually PSJ *a, *a yield Kaingáng o, e, respectively). However, Laklano data (jare 'root', d-an-ka 'door') show that PSJ *a must be reconstructed for these tokens. I propose to consider the following sound law: PSJ *a > KGG e in stressed (final) syllables as well as in unstressed (non-final) syllables if the stressed syllable also contains PSJ *a > KGG e (cf. PSJ *jana > KGG jene 'to send'; PSJ *jara > KGG jere 'armpit'). Conversely, PSJ *a > KGG o in unstressed (non-final) syllables if the stressed syllable contains a different vowel (cf. PSJ *jare > KGG jore 'root'; PSJ
*jan-ka > KGG jon-ko 'door'; PSJ *jard > KGG jora 'saliva'). Note that o (~ a) and e merged in all dialects other than Paraná Kaingáng and South-Eastern Kaingáng (Wiesemann 1978: 203-204); in some varieties, a new contrast between o and e emerged, whereby the use of o is associated with round, compact objects and the use of e is associated with long, diffuse objects (D'Angelis 2002). The Kaingáng forms given in this paper are from the Paraná dialect, where the etymological contrast is retained.
r. PJ *jar-ko 'mouth' > PCerr *jad-kwa, PSJ *jan-ka 'door'
s. PJ *k6k ~ *ka-k6k 'wind' > PCerr *k6k, *ka-k6k, PSJ *ka-kd (vb. *ka-kd-n)
(7) Proto-Je *d > Proto-Cerrado *o, Proto-Southern Je *a
a. PJ *nuctd 'tongue' > PCerr *nojto, PSJ *nuna
b. PJ *jandd 'to send' > PCerr *jando/*jando-r', PSJ *jana/*jana-j
c. PJ *nddm 'eye' > PCerr *ndom, PSJ *kd-na (vb. *kd-na-n) 'eye, fruit'
d. PJ *mbrd 'ashes' > PCerr *mbro, PSJ *mra
e. PJ *td 'to fly.SG' > PCerr *to/*top-r, PSJ *ta (vb. *ta-m)
(8) Proto-Je *d > Proto-Cerrado *a, Proto-Southern Je *d
a. PJ *mbd 'to grab' > PCerr *mbd/*mbd-r'; PSJ *mbd/*mba-j, *mbd-r (vb. *mba-n) 'to hold, to carry (a short object)'
b. PJ *tat 'strong, hard' > PCerr *tdt, PSJ *tar (vb. *tan)
c. PJ j, 3 *c-3 'bitter' > PCerr *p, PSJ *da (vb. *da-j)
d. PJ *nijd 'smoke' > PCJ *nije, PSJ *nijd (vb. *nija-j)
(9) Proto-Je *y > Proto-Cerrado *a, Proto-Southern Je *d
a. PJ *prY 'coal, ember' > PCerr *pra 'ember', PSJ *pra-n
b. PJ *jY 'urine' > PCerr *jd 'urine', PSJ *jd-n 'to urinate'
(10) Proto-Je *y > Proto-Cerrado *y, Proto-Southern Je *y
a. PJ *pry (~ *pryn) 'road' > PCerr *pry, PSJ *a-pry/*NP jd-pry
b. PJ *mbyt 'celestial body' > PCerr *mbyt 'sun', Ingain pyr 'moon'
c. PJ *mbyn 'tail' > PCerr *mbyn, PSJ *mby
d. PJ *ty 'to die' > PCerr *ty(r)/*tyk, PSJ *ty 'to die.STAT'
e. PJ *tyk 'black, dark' > PCerr *tyk 'black', PSJ *ku-ty (vb. *ku-ty-j) 'dark, night'
f. PJ *k(r)Vmtym' 'capybara' > PCerr *kumtym', PSJ *kryjndyj
g. PJ *cy 'to weave' > PCerr *cy/*cyp-r', PSJ *dy
h. PJ *cym 'seed' > PCerr *cym, PSJ *dy (vb. *Qy-n)
i. PJ *njy 'bee, bumblebee' > PCerr *njy, PSJ *cy
(11)Proto-Je *o > Proto-Cerrado *wa, Proto-Southern Je *a
a. PJ *tom' 'fat' > PCerr *twam', PSJ *taj
b. PJ *ron 'celestial body' > PCerr *rwaj' 'moon', PSJ *ra 'sun'
c. PJ *jo 'tooth' > PCerr *jwa, PSJ *ja
d. PJ *kVjo 'salt' > PCerr *kVjwa, PSJ *kdja (vb. *kdja-j) 'salty, sour'
e. PJ *jar-ko 'mouth' > PCerr *jad-kwa, PSJ *jan-ka 'door'
f. PJ *kVnko 'sky' > PCerr *kdjkwa, PSJ *kdnka
g. PJ *jgon' 'feather, hair' > PNJ *jgwm 'feather', PSJ *jgan 'animal hair'
(12)Proto-Je *d > Proto-Cerrado *o, Proto-Southern Je *d
a. PJ *poc 'to leave.PL' > PCerr *poc, PSJ *pa (vb. *pa-n)
b. PJ *wo 'to untie' > PNJ *b6/*bd-n, PSJ *kd-wd/*kd-wa-j/*kd-wa-n 'to let loose, to untie'
c. PJ *nj6 'to hang.SG' > PCerr *nj6/*nj6p-r', PSJ *cd/*ca-j (vb. *cd-m)
d. PJ *j6 'to tear, to rip' > PNJ *ka-j6/*ka-j6-n; PSJ *jd/*ja-j, *ja-r (vb. *ja-n)
e. PJ *k6p 'fly, mosquito' > PCerr *k6p 'fly', PSJ *kd 'mosquito'
f. PJ *k6m 'tree, horn' > PCerr *k6m, PSJ *kd 'tree', *ni-kd 'horn'
g. PJ *k6k ~ *ka-k6k 'wind' > PCerr *k6k, *ka-k6k, PSJ *ka-kd (vb. *ka-kd-n)
h. PJ *jg6 'louse' > PCerr *jg6, PSJ *jg3 (vb. *jga-n) 'louse, flea'
(13)Proto-Je *u > Proto-Cerrado *u, Proto-Southern Je *o
a. PJ *mbrum' 'ant' > PNJ *mbrum, PSJ *mbror-jy 'Argentine ant' (with nasal lowering)
b. PJ *tum' 'old' > PCerr *tum' 'old, tall'; PSJ *tor 'dry (of plants)' (with nasal lowering)
c. PJ *pondv 'bad or crooked' > PNJ *pundu 'bad', PSJ *pando (vb. *pando-n) 'crooked'
d. PJ *ju, 3 *c-u 'pus' > PCerr *jup-r, PSJ *6d (vb. *6o-m)
e. PJ *jum' 'father' > PNJ *ju(m), PSJ *jor (with nasal lowering)
f. PJ *kv 'to eat' > PCerr *ku/*ku-r', PSJ *ko 'to eat, to use'
(14)Proto-Je *u > Proto-Cerrado *u, Proto-Southern Je *u
a. PJ *tu 'to carry' > PCerr *tu/*tu-r', PSJ *tu 'to carry on one's back, to wear'
b. PJ *ru 'to transport liquid or to pour' > PNJ *ru/*ru-n 'to spill, to pour', PSJ *ru/*ru-r/ *ru-n 'to fetch water'
c. PJ *peju 'to hide' > PNJ *puju/*puju-r, PSJ *peju
d. PJ *nju 'to heat, to dry' > PNJ *nju/*nju-n 'to dry in the sun'; PSJ *cu-r (vb. *cu-n) 'to heat by the fire', *jar-cu/*jar-cu-r, *jar-cu-r (vb. *jar-cu-n) 'to roast'
(15)Proto-Je *e > Proto-Cerrado *e, Proto-Southern Je *e
a. PJ *mbec 'good' > PCerr *mbec, PSJ *mbe 'habit', *a-mbe (vb. *a-mba-n) 'good weather'
b. PJ *mbre > PNJ *mbre 'sibling-in-law', PSJ *jd-mbre 'cousin, son-in-law'
c. PJ *re 'to leave, to abandon' > PCerr *re, PSJ *re/*ra-r
d. PJ *kre 'hole' > PCerr *kre, PSJ *kre
e. PJ *rgre 'egg' > PCerr *rgre, PSJ *rgre 'egg, penis'
f. PJ *rgre 'to dance' > PCerr *rgre/*rgre-r, PSJ *rgre (vb. *rgre-n)
(16)Proto-Je *e > Proto-Cerrado *e, Proto-Southern Je *e
a. PJ *mben 'liquid' > PCerr *mben, PSJ *mbe (vb. *mbe-n)
b. PJ *jare (~ *jaren) 'root' > PNJ *jare, PSJ *jare (vb. *6are-n)
c. PJ *-re(C) 'to weed' > PNJ *ka-re/*ka-re-r, PSJ *ku-re/*ku-re-r 'to weed, to cut clean'
d. PJ *jace ~ *jVmce 'nest' > PCerr *jace, PSJ *jdr6e (vb. *jar6e-n)
e. PJ *nje 'to tickle' > PNJ *nje/*nje-r', PSJ *ku-ce/*ku-ce-r
f. PJ *je (~ *jen) 'thread, knot, cloth' > PNJ *je, PSJ vb. *6e-n 'to spin thread'
g. PJ *ket 'neg' > PNJ *ket, PSJ *ker 'advt'
(17)Proto-Je *i > Proto-Cerrado *ja, Proto-Southern Je *e
a. PJ *mbin' 'husband' > PNJ *mbjen, PSJ *mben (with nasal lowering)
b. PJ *kij 'pit' > PNJ *kX 'earth oven', PSJ *kej 'tomb'
c. PJ *ki ~ *kij ~ *kij' 'to split' > PNJ *ko-ki/*ko-kje-r, PSJ vb. *ke-n
d. PJ *krin 'thigh' > PCerr *k(r)jaj', PSJ *kre
e. PJ *rgi 'to enter.PL' > PCerr *a-rgja/*rgja-c, PSJ *rge (vb. *rge-m)
(18)Proto-Je *i > Proto-Cerrado *i, Proto-Southern Je *i
a. PJ *peti 'to dream' > PNJ *pyti/*pyti-r, PSJ *peti/*peti-r
b. PJ *ji 'to lay.SG' > PCerr *ji/*ji-r', PSJ *ji/*ji-r
c. PJ *jiji ~ *niji 'name' > PCerr *niji, PSJ *jiji ~ *jyjy (vb. *jiji-n ~ *jyjy-n)
(19)Proto-Je *d > Proto-Cerrado *d, Proto-Southern Je *d
a. PJ *pram' 'hungry, to want' > PCerr *pram', PSJ *prar 'hungry, year'
b. PJ *td 'instr' > PCerr *td, PSJ *td 'erg, instr'
c. PJ *nd 'mother' > PCerr *n§, PSJ *n§
d. PJ *r^(C) 'to push against, to grind' > PNJ *r^/*r^-n, PSJ *r^/*r^-r
(20)Proto-Je *y > Proto-Cerrado (*y after palatals), Proto-Southern Je *i
a. PJ *ny 'to sit.SG' > PCerr *ny/*nyp-r, PSJ *ni (vb. *ni-m)
b. PJ *kryn 'head' > PCerr *krdj', PSJ *kri (vb. *kri-n)
c. PJ *jVkryn 'knee' > PCJ *hikrdj // *hikr§, PSJ *jdkri
(21)Proto-Je *o > Proto-Cerrado *o, Proto-Southern Je *d
a. PJ *po 'to rub, to clean' > PNJ *po/*po-n, PSJ *p§/*pd-j 'to clean a field'
b. PJ *no 'to lie.SG' > PCerr *no/*nop-r, PSJ *nd
(22)Proto-Je *u > Proto-Cerrado *o, Proto-Southern Je *u
a. PJ *pru 'wife' > PCerr *pro, PSJ *pru (vb. *pru-j)
b. PJ *mu 'to go.PL' > PCerr *md/*md-r', PSJ *mu (vb. *mu-n)
c. PJ *tu 'NEG' > PCJ *to, PSJ *tu/*tu-r (vb. *tu-r)
d. PJ *ut/*nut 'to sleep' > PCerr *rot/*not, PSJ *nur
e. PJ *ju 'to push against, to crumble' > PNJ *ka-ro/*ka-ro-n, PSJ *ru/*ju-r
f. PJ *jru (~ *jrun) 'toucan' > PNJ *jro, PSJ *jru
g. PJ *nuctd 'tongue' > PCerr *nojto, PSJ *nuna
h. PJ *numje(C) 'female breast' > PNJ *nomje, PSJ *nujje 'female breast, milk'
(23)Proto-Je *e > Proto-Cerrado *e, Proto-Southern Je *i
a. PJ *me 'pl; with' > PCerr *me, PSJ *mi 'pl'
b. PJ *te 'to go.SG' > PCerr *te/*tem', PSJ *ti/*ti-j (vb. *ti-n)
c. PJ *re 'to throw.PL' > PNJ *re/*re-n, PSJ *ri/*ri-r/*ri-n
(24)Proto-Je *l > Proto-Cerrado *i, Proto-Southern Je *i
a. PJ *pim 'tree, wood' > PCerr *pim, PSJ *pi 'fire, firewood'
b. PJ *-ci(C) 'bandage, sling' > PNJ *ja-ci, *ka-ci, PSJ *-6i (vb. -di-n)
c. PJ *ni 'meat' > PCerr *ni, PSJ *ni
d. PJ *ni(m)- 'hand (in compounds)' > PCerr *ni(m)-kra 'hand', PSJ *ni- (as in *ni-pe 'to wash hands')
e. PJ *jri(C) > PNJ *ku-jri 'to gather in a bundle', *-jri 'to make packages', PSJ *jri/ *jri-j, *jri-r (vb. ji-n) 'wrap'
f. PJ *nija 'nose' > PCerr *nija-kre, PSJ *nija
g. PJ *nijd 'smoke' > PCJ *nije, PSJ *nijd (vb. *nija-j)
In non-final (unstressed) syllables, divergent correspondences may be observed. One such correspondence involves the PCerr formative that is reflected in PNJ as *py-/*pu- word-initially (harmonizing in roundness with the nucleus of the root, cf. Salanova 2011a: 66) or as *-p-word-medially, and in PCJ as *pi-. In some stems, this formative corresponds to PSJ *pa- (25a-b); in others, it corresponds to PSJ *pe- (25c-d), thus evidencing a merger that would have occurred historically in PCerr. I assume that PSJ displays here the same reflexes of PJ vowels that are found in stressed syllables, whereas in the history of PCerr the unstressed vowels *o and *e were reduced to zero (probably via *u/*i, since high vowels are more likely to be syncopated). The consonant clusters created by this development were subsequently undone via different processes in PNJ (resyllabification if available, epenthesis of *y/*u elsewhere) and in PCJ (epenthesis of *i).
(25)Proto-Je *o, *e > PCerr zero in unstressed syllables
a. PJ *pondv 'bad or crooked' > PCerr *p3ndu (PNJ *pundu 'bad'), PSJ *pand6 (vb. *pando-n) 'crooked'
b. PJ *potu 'horsefly' > PCerr *ftu (PNJ *pucu, PCJ *pidu), PSJ *patu
c. PJ *peju 'to hide' > PCerr *p3ju/*p3ju-r' (PNJ *puju/*puju-r), PSJ *peju
d. PJ *peti 'to dream' > PCerr *p3ti/*p3ti-r' (PNJ *pyti/*pyti-r), PSJ *peti/*peti-y
There appear to be no counterexamples to the development exemplified in (25): the usual reflexes of PJ *o and *e (PCerr *wa > PNJ *wa, *wê, PCJ *wa; PCerr *e > PNJ *e, PCJ *ê) do not occur in prefixes or in the unstressed syllables of underived stems in the Cerrado languages.
Only one well-known cognate set violates the regularities outlined in this section: as shown in (26), it features a correspondence between PCerr *ô and PSJ *e. Judging by the external cognates, such as Maxakali -cyc (orthographically <-xux>) or Krenak jdt, PCerr is conservative here: the correspondence between Maxakali y, Krenak d, and PCerr *ô is well-attested (these vowels usually correspond to PSJ *d), whereas PSJ *e usually has entirely different correspondences: Maxakali e, Krenak i, PCerr *e (Nikulin & Silva, forthc.).
(26)An exceptional correspondence between PCerr *ô and PSJ *e
PJ *côj 'leaf' > PCerr *côj', PSJ *6ej (vb. *6e-n)
I hypothesize that PSJ *e in *6ej is a result of fronting of pre-PSJ *d (the regular outcome of PJ *ô) before a palatal coda. The development *-dj > *-ej may have been regular: the rhyme *-dj does not occur abundantly in Proto-Southern Jê lexicon, the only known example being *wdj 'dawn' without a known external etymology. However, in the absence of supporting examples the hypothesis in question remains rather speculative.
3.3. Codas
The correspondences between the codas of PCerr and PSJ, as well as the respective Proto-Jê reconstructions are presented in Table 7 below. I remind the reader that the apostrophe stands for a suppressed echo vowel (word-finally). For Southern Jê, in addition to the codas found in underived stems, I provide information regarding the allomorphy of the verbalizer/causative suffix. Four codas (/*p', *t', *c', *k'/) are reconstructed only for the underlying level of PJ.
Cognate sets that instantiate the aforementioned sound correspondences (excepts those going back to PJ underlying codas, for which see below) are provided below in (27-40).
(27)Proto-Jê *p > Proto-Cerrado *p, Proto-Southern Jê *0
a. PJ *kôp 'fly, mosquito' > PCerr *kôp 'fly', PSJ *kd 'mosquito'
(28) Proto-Jê *t > Proto-Cerrado *t, Proto-Southern Jê *r (vb. *-n)
a. PJ *mbyt > PCerr *mbyt 'sun', Ingain pyr 'moon'
b. PJ *tat 'strong, hard' > PCerr *tdt, PSJ *tar (vb. *tan)
c. PJ *nt 'to look' > PCerr *nt, PSJ *rlr
d. PJ *nut 'to sleep.NF' > PCerr *not, PSJ *nur
e. PJ *kët 'stone' > PCerr *kët, Ingain kër
f. PJ *krot 'chin, beak' > PCerr *krwat 'beak', Ingain jat-krar 'chin'
(29)Proto-Jê *c > Proto-Cerrado *c, Proto-Southern Jê *0 (vb. *-n)
a. PJ *pôc 'to leave.PL' > PCerr *pôc, PSJ *pa (vb. *pa-n)
b. PJ *mbec 'good' > PCerr *mbec, PSJ *mbe 'habit', *a-mbe (vb. *a-mba-n) 'good weather'
c. PJ *nuctd 'tongue' > PCerr *nojto, PSJ *nuna
(30) Proto-Jê *k > Proto-Cerrado *k, Proto-Southern Jê *0 (vb. *-n) or *j (after *ê)
a. PJ *tyk 'black, dark' > PCerr *tyk 'black', PSJ *ku-ty (irreg. vb. *ku-ty-rj) 'dark, night'
b. PJ *kôk ~ *ka-kôk 'wind' > PCerr *kôk, *ka-kôk, PSJ *ka-kd (vb. *ka-kd-n)
Table 7. Proto-Jê codas and their reflexes in Proto-Cerrado and Proto-Southern Jê.
PJ PCerr PSJ PSJ (vb.)
followed by echo vowels
*P *P *0
*t *t *r
*c *c *0
*k *k *0, *j+ *-n
*m *m
*n *n *0
*n *j'
*r *r *n *-n
*j *j' *j *-n
PJ PCerr PSJ PSJ (vb.)
no echo vowels
*/P'/ *0, *pi *-m
*0 */t'/ *0 *-n
*/c'/ *0 *-n
*/k'/
* m' *m'
*n' *n' *n *-n
*n' *n' *n *-n
*r)' (?)
*r' *d *n *-n
*j' *c *j *-n
+ after *e17 (no vb. attested, but probably *n); X in non-finite forms preceding the suffix *-r
c. PJ *pek 'to fart.NF' > PNJ *pe-k, PSJ *pej (but PCJ *pi-ri)
d. PJ *kujek 'vein' > PNJ *kujek, PSJ *kujej
(31)Proto-Je *m > Proto-Cerrado *m, Proto-Southern Je *0 (vb. *-n)
a. PJ *pim 'tree, wood' > PCerr *pim, PSJ *pi 'fire, firewood'
b. PJ *nddm 'eye' > PCerr *ndom, PSJ *kd-na (vb. *kd-na-n) 'eye, fruit'
c. PJ *cym 'seed' > PCerr *cym, PSJ *6y (vb. *6y-n)
d. PJ *kom 'tree, horn' > PCerr *kom, PSJ *kd 'tree', *ni-kd 'horn'
(32)Proto-Je *n > Proto-Cerrado *m, Proto-Southern Je *0 (vb. *-n)
a. PJ *mbyn 'tail' > PCerr *mbyn, PSJ *mby
b. PJ *mben 'liquid' > PCerr *mben, PSJ *mbe (vb. *mbe-n)
(33)Proto-Je *n > Proto-Cerrado j, Proto-Southern Je *0 (vb. *-n)
a. PJ *pan 'arm, branch' > PCerr *paj\ PSJ *pa
b. PJ *jan 'to eat.iNTR, food' > PCJ *caj // *ca, PSJ *ja/*ja-y/*ja-n (vb. *ja-n)
c. PJ *kryn 'head' > PCerr *krdj', PSJ *kri (vb. *kri-n)
d. PJ *krin 'thigh' > PCerr *k(r)jaj', PSJ *kre
(34)Proto-Je *r > Proto-Cerrado *r, Proto-Southern Je *n a. PJ *par 'foot' > PCerr *par, PSJ *pan
17 I thank Mario André Coelho da Silva for pointing out the possibility that PJ *-êk > PSJ *-êj could be a regular sound change.
(35)Proto-Je *j > Proto-Cerrado *j', Proto-Southern Je *j (vb. *-n)
a. PJ *c6j 'leaf' > PCerr *c6j', PSJ *6ej (vb. *6e-n)
(36)Proto-Je *m' > Proto-Cerrado *m', Proto-Southern Je
a. PJ *pram' 'hungry' > PCerr *pram\ PSJ *prSj 'hungry; year'
b. PJ *mbrum' 'ant' > PNJ *mbrum, PSJ *mbroj-jy 'Argentine ant'
c. PJ *k(r)Vmtym' 'capybara' > PCerr *kumtym', PSJ *kryjndyj
d. PJ *tom' 'fat' > PCerr *twam', PSJ *taj
e. PJ *tum' 'old' > PCerr *tum' 'old, tall'; PSJ *toj 'dry (of plants)'
f. PJ *numje(C) 'female breast' > PNJ *nomje, PSJ *nujje 'female breast, milk'
(37)Proto-Je *n' > Proto-Cerrado *n', Proto-Southern Je *n
a. PJ *mbm' 'husband' > PNJ *mbjen, PSJ *mben
(38)Proto-Je *n' > Proto-Cerrado *n', Proto-Southern Je *n
a. PJ *jgon' 'feather, hair' > PNJ *jgw3n 'feather', PSJ *jgan 'animal hair'
(39) Proto-Je *r' > Proto-Cerrado *d, Proto-Southern Je *n
a. PJ *jar-ko 'mouth' > PCerr *jad-kwa, PSJ *jan-ka 'door'
(40)Proto-Je *j' > Proto-Cerrado *j', Proto-Southern Je *j (vb. *-n)
a. PJ *kaj' 'basket' > PNJ *kac, PSJ *kaj (vb. *kan)
The reconstruction of PJ is supported by only one PSJ token, PSJ *mdj 'honey bee', which lacks known cognates in Cerrado languages but corresponds externally to Maxakali pyk (orthographically <puk>) and Krenak pdj. As Nikulin and Silva (forthc.) show, the point of articulation of Proto-Je codas systematically matches that of Maxakali, and the manner of articulation of Proto-Je codas systematically matches that of Krenak.
My reconstruction of Proto-Je underlying codas (/*p', *t', *c', *k'/) is crucially based on morphophonological evidence from Central and Southern Je and is corroborated by external comparison with Maxakali and Krenak (Nikulin & Silva, forthc.). No contemporary Je language shows any reflex of these codas in underived environments, which probably means that they surfaced as zero in underived Proto-Je words. In Central Je, only PJ /*p'/ left any trace: in non-finite forms of verbs, it combined with the non-finite suffix *-r, yielding PCJ *-b-rV. In Southern Je, the quality of the underlying codas is made manifest in derived denominal verbs or causatives, where underlying /-n/ acquires the point of articulation of the underlying coda consonant (see Cavalcante 1987: 51-52 for a synchronic analysis for Sao Paulo Kaingang). Examples are given in (41-44).
(41)Proto-Je */p'/ > Proto-Cerrado *0 (nf *-p-r), Proto-Southern Je *0 (vb. *-m)
a. PJ *td */tap'/ 'to fly.SG' > PCerr *to/*top-r, PSJ *ta (vb. *ta-m) cf. Maxakali /tup-a/, irr /tup/ 'id.'
b. PJ *no */nop'/ 'to lie.SG' > PCerr *no/*nop-r, PSJ *n~3 cf. Maxakali /nup/ 'to lie.PL'
c. PJ *cy */cyp'/ 'to weave' > PCerr *cy/*cyp-r', PSJ *6y cf. Maxakali /cap/ 'id.'
d. PJ *c6 */cop'/ 'to eat soft food, to suck' > PNJ *c6/*c6-r, PSJ *pe-6d (vb. *pe-6a-m) 'to suck breast'
cf. Maxakali /cyp/ 'to suck, to lick'
e. PJ *nj6 */nop'/ 'to hang.SG' > PCerr *njo/*njgp-r', PSJ *cd/*ca-j (vb. *cd-m) cf. Maxakali /cyp/ 'id.'
f. PJ *ny */jyp'/ 'to sit.SG' > PCerr *ny/*nyp-r, PSJ *ni (vb. *ni-m) cf. Maxakali /nyp/ 'id.'
g. PJ *ju 7jop'/ 'pus' > PCerr *jup-r, PSJ *6o (vb. *6o-m)
h. PJ *u ~ *o /*up' ~ *op'/ 'to give' > PCerr *yo/*ndp-r' cf. Maxakali /hup/ 'id.'
(42)Proto-Je */t'/ > Proto-Cerrado *0, Proto-Southern Je *0 (vb. *-n)
a. PJ *jare (~ *jaren) */jaret' ~ jaren/ 'root' > PNJ *jare, PSJ *jare (vb. *6-are-n) cf. Maxakali /nip-catit/ 'id.'
b. PJ *je (~ *jen) */jet' ~ jen/ 'thread, knot, cloth' > PNJ *je, PSJ vb. *6e-n 'to spin thread' cf. Maxakali /cit/ 'id.'
c. PJ *jiji ~ *niji */jijit' ~ jijit'/ 'name' > PCerr *niji, PSJ *jiji ~ *jyjy (vb. *jiji-n ~ *jyjy-n) cf. Maxakali /-cit/ 'id.'
d. PJ *ko */kot'/ 'to dig' > PNJ *kwd/*kw§-n, PSJ *ka/*ka-y (vb. *ka-n) cf. Maxakali /kut/ 'id.'
e. PJ *ygd */q6t'/ 'louse' > PCerr *ygd, PSJ *yga (vb. *yga-n) 'louse, flea' cf. Maxakali /kyt/ 'id.'
(43)Proto-Je */c'/ > Proto-Cerrado *0, Proto-Southern Je *0 (vb. *-n)
a. PJ *prr */pryc'/ 'coal, ember' > PCerr *prd 'ember', PSJ *prd-n18
b. PJ *jr */jyc'/ 'urine' > PCerr *jd 'urine', PSJ vb. *jd-n 'to urinate' cf. Maxakali /cyc/ 'id.'
c. PJ *jace ~ *jVmce 'nest' */jacec' ~ jamcec'/ > PCerr *jace, PSJ *jdy6e (vb. *jdy6e-n)
(44)Proto-Je */k'/ > Proto-Cerrado *0, Proto-Southern Je *0 (vb. *-n)
a. PJ *tu */tuk'/ 'neg' > PCJ *to, PSJ *tu/*tu-rS (vb. *tu-rj) cf. Maxakali /nuk/ 'to end, to run out'
b. PJ *kra */krak'/ 'offspring' > PCerr *kra, PSJ *kra (vb. *kra-r) cf. Maxakali /ktuk/ 'id.'
I do not represent these underlying codas in my reconstructions throughout this paper; the examples (41-44) above include almost every cognate set for which the presence of an underlying coda may be established with any certainty.
4. Je etymologies
In this section, we present the most reliable cognate sets identified so far. The data are sorted by the onset of the final (stressed) syllable, then by its nucleus, then by its coda, then by any preceding material, following the order /p, pr, m, mr, w, t, n, d, r, c, n, j, k, kr, q, qr, a, a, a, a, y, y, y, o, o, 6, o, u, u, e, e, e, i, i, i/. I list first the etymologies that have reflexes both in Cerrado and in Southern Je languages, and then proceed to the reconstructions that are based only on Cerrado or Southern Je reflexes (as well as their external cognates elsewhere in Macro-Je or Tupian19), respectively.
18 It is unclear why the PSJ reflex looks like a verbalized form.
19 Although I do not currently regard the relation between Macro-Je and Tupian as conclusively proven, I still find the hypothesis quite promising. Some novel lexical evidence may be found in Nikulin & Silva forthc. (as well as in this section). The Proto-Tupian reconstructions are my own and follow the principles outlined by Carvalho and Nikulin (in prep.).
For reasons of space, I do not provide the data of individual Northern Je languages, limiting myself to Proto-Northern Je reconstructions. Detailed information on the reflexes of ProtoNorthern Je items in individual Northern Je languages may be found in Nikulin & Salanova (forthc.). Ingain and Southern Kayapo data are only given for cognate sets that lack known reflexes in Kaingang/Laklano and Panara, respectively.
Reflexes in both branches
PJ *par 'foot':
PCerr *par > PNJ *par; pnr pa:; PCJ *para // *para (xav para // para, xer pra) PSJ *pan > KGG pen; lkl pan PJ *pan 'arm, branch':
PCerr *paj' > PNJ *pa; pnr pa; PCJ *paj-no 'arm' (xav pan-no, xer paj-no), *paj-hi (xav paj-hi); (?) PCJ *pa-krata // *pa-kra:da 'root' (xer pa-krta // pa-krda), (?) *wede-pa 'root' (xav wedepa, xer wde-pa) PSJ *pa > KGG pe; lkl pa PJ *pim 'tree, wood, firewood':
PCerr *pim > PNJ *pi; pnr pi; PCJ *mimi // *mi (xav mimi // mi, xer mmi) PSJ *pi 'fire, firewood' > KGG pi; lkl pe PJ *po 'to rub, to clean':
PCerr *pd/*po-n' > PNJ *pd/*pd-n PSJ *p§/*pd-j 'to clean a field' > KGG pd/pd-j PJ *p6c 'to leave.PL':
PCerr *p6c > PNJ *to=p6j/*to=p6c 'to extract.PL'; PCJ *pu/*pu-ci 'to leave.DU' (xav pu-ci, xer pu/pu-s(i))
PSJ *pd (vb. *pd-n) > KGG pa (vb. pa-n); lkl po PJ *pe/*pe-k 'to fart':
PCerr *pe/*pe-k > PNJ *ij-pe/*pe-k; PCJ *pi/*pi-ri20 (xav pi/pi-ri) PSJ *pej > KGG pej PJ *pr§m' 'hungry, to want':
PCerr *pr5m' > PNJ *pr5m; pnr pjmi 'to want'; PCJ *mrdm // *mrd 'hungry' (xav, xer mrSm // mra) PSJ *prdj 'hungry; year' (probably via 'hunger season') > KGG prsj; lkl ploj 'year' PJ *prr 'coal, ember':
PCerr *prd 'ember' > PNJ *prd; PCJ *pre (xer pre-hika 'glowing embers', pre-nizuri 'spark',
pre-zapdo 'ember') PSJ *prm > KGG prdn; lkl plan PJ *pry (~ *pryn) 'road':
PCerr *pry (~ *pryn) > PNJ *pry; pnr pjy PSJ *a-pry/*NP jd-pry > KGG e-pry/NP ja-pry PJ *pru 'wife':
PCerr *pro > PNJ *pro; PCJ *mro 'spouse, to marry' (xav, xer mro) PSJ *pru (vb. *pru-j) > kgg pru (vb. pru-j); lkl plu (vb. plu-j) PJ *mba 'to hear, to understand':
PCerr *mba/*mba-r' > PNJ *mba/*mba-r; pnr impa-ri; PCJ *wa-pa/*wa-pa-ri (xav wa-pa/wa-pa-
ri, xer wa-pa/wa-pa-r(i)) PSJ *ma/*ma-j > KGG me/me-j; lkl ma/ma-j
20 The expected non-finite form would be **pi-ki // **pi. Northern Jê is considered to be more conservative than Central Jê here because the Southern Jê cognate appears to correspond to the Northern Jê non-finite form.
PJ *mba 'liver':
PCerr *mba > PNJ *mba; pnr impa; PCJ *pa (xav, xer pa) PSJ *ta-ma > KGG ta-me; lkl to-ma PJ *-mbay'21 'to be afraid':
PCerr *p3-mba > PNJ *pymba; pnr s-umpa; PCJ *pipa (xav, xer pipa) PSJ *ka-may, *mü-may > KGG ka-mey, mü-mey; lkl ko-may, mö-mäy PJ *ma 'DAT':
PCerr *ma > PNJ *ma; pnr ma; PCJ *ma (xav, xer ma) PSJ *ma > KGG ma; lkl mo PJ *mba 'to grab, to carry':
PCerr *mba/*mba-ñ > PNJ *mba/*mba-ñ 'to grab'; pnr impa-ri 'to carry'; PCJ *kwa-pe/*kwa-pe-j
// *kwa-pe 'to carry.DU' (xav ?wa-pe/?wa-pe-j // ?wa-pe, xer kwape) PSJ *mba/*mba-y, *mba-r (vb. *mba-n) 'to hold, to carry (a short object)' > KGG mba/mö-y/mö-n''; lkl mbo/mba-y, mbol (vb. mba-n) PJ *mbyn 'tail':
PCerr *mbyn 'tail, penis' > PNJ *mby 'penis', *ja-mby 'tail'; pnr s-ampy 'tail'; PCJ *mana // *ba
(xav mana // ba, xer mna // ba) PSJ *mby > KGG, lkl mby PJ *mbyt 'celestial body':
PCerr *mbyt 'sun' > PNJ *mbyt; pnr impyti 'time'; PCJ *bata // *ba:da (xav bata // bada, xer bta // bda)
Ingain22 pyr 'moon': Lista (pirihí), Pedro, Cosme Román (puirí), María Antonia (puiré), Vogt1 (pyri^ Vg (pyry) PJ *mbec 'good':
PCerr *mbec > PNJ *mbec; pnr impe 'real'; PCJ *pécé // *pé (xav pécé // pe 'well; to recover,
to get better', xer pésé // pé) PSJ *mbe 'habit' > KGG, lkl mbe; PSJ *a-mbe (vb. *a-mba-n) 'good weather' > KGG e-mbe (vb. e-mo-n) PJ *me 'PL; with':
PCerr *me > PNJ *me; pnr -me-ra 'pl (nominal number)', me= 'du (indexed on verbs)'; PCJ
*me 'with' (xav me, xer me) PSJ *mi 'pl' > lkl me PJ *mbén 'liquid':
PCerr *mbén > PNJ *mbé; PCJ *pini // *pi (xav pini // pi23 'honey', xer to-pi 'eye rheum') PSJ *mbé (vb. *mbe-n) > KGG mbé (vb. mbe-n); lkl mbé PJ *mbm' 'husband':
PCerr *mbjan' > PNJ *mbjén; pnr impin-pja24 PSJ *mben > KGG, lkl mben
21 PSJ in underived stems usually goes back to PJ *-m' and corresponds to PCerr *-m'. The correspondence between PSJ *-r¡ and PCerr zero is unparalleled; I provisionally reconstruct PJ
22 Hereinafter, Lista refers to data from Lista 1883; Pedro, Cosme Román, and María Antonia are the names of Ambrosetti's (1896) consultants who provided the respective pieces of data; Vogt1, Vogt3 refer to two original word-lists contained in Vogt 1904.
23 The expected utterance-final allomorph would be **pi. One possible explanation for this irregular reflex is an analogical extension of nasality from the regular utterance-medial allomorph, pini.
24 It is unclear whether the Panará reflex of PCerr *-ja- is regular.
PJ *mu 'to go.PL':
PCerr *mo/*mo-r' > PNJ *mo/*mo-r; pnr md-ri; PCJ *mo/*mo-ri 'to go.SG' (xav mo/mo-ri, xer mo/mo-r(i))
PSJ *mu (vb. *mu-n) > kgg mu (vb. mu-n); lkl mu PJ *mbra 'ashes':
PCerr *mbro > PNJ *mbro; PCJ *wede-pro 'coal' (xav wede-pro, xer wde-pro), *ka jaj-mpro 'foam' (xav ?a jaj-pro, xer kd zai-pro ~ kd zam-pro), *jadaj-mpro 'saliva' (xav jadaj-pro, xer zdai-pro), (?) *pro 'to burn' (xer pro) PSJ *mra > KGG mrej;25 lkl mla PJ *mbrum' 'ant':
PCerr *mbrum' > PNJ *mbrum PSJ *mbroj-jy 'Argentine ant' > KGG mbroj-jy PJ *mbre 'relative by marriage' (kinship term):
PCerr *mbre > PNJ *mbre 'sibling-in-law'; PCJ *md-pre-baba // *md-pre-wa (xav ma-pre-baba //
ma-pre-wa 'mother/father-in-law', xer ma-pre-wa) PSJ *ja-mbre > KGG ja-mbre 'cousin, son-in-law'; lkl jo-mble 'brother-in-law, father-in-law, mother-in-law' PJ *wa 'to walk':
PCerr *wa > PNJ *ba 'to walk, to live'; pnr pa PSJ *wa 'stat' (aspect marker) > KGG we; lkl va PJ *wy 'to take, to carry':
PCerr *wy/*wy-r' > PNJ *by/*by-r; pnr py-ri
PSJ *wa/*wa-j/*wa-n ~ *wy-n26 'to hold, to carry (a long object)' > KGG wa/wo-j/wo-n ~ wy-n; lkl va/vd-j/vd-n ~ vu-n PJ *w6 'to untie':
PCerr *w6/*w6-n' > PNJ *b6/*b6-n; Southern Kayapo (tipo) (likely p6) PSJ *ka-wa/*ka-wa-j, *ka-wa-r (vb. *ka-wa-n) 'to let loose, to untie' > KGG ka-wa/ka-wo-j, ka-wa-r (vb. ka-wo-n) PJ *ta 'to fly.SG':
PCerr *to/*top-r > PNJ *to/*to-r; PNR to:/to:-j 'to fly, to dance'; PCJ *tob-ro (xav tob-ro
'to descend', xer tb-ro 'to cross a body of water') PSJ *ta (vb. *ta-m) > KGG te (vb. te-m); lkl ta 'to approach.SG' PJ *nucta 'tongue':
PCerr *nojto > PNJ *noto; pnr s-oto; PCJ *nojto (xav nojto ~ notto, xer nojto) PSJ *nuna > KGG nune; lkl nuna PJ *ta 'INSTR'
PCerr *ta > PNJ *ta 'loc'; pnr ta 'all' or ha/(r)aha 'adess'; PCJ *na (xav, xer na) PSJ *ta 'ERG, INSTR' > KGG ta; LKL to PJ *tat 'strong, hard':
PCerr *tat > PNJ *tac; pnr tati; PCJ *tete // *te:de (xav -tete // te:de, xer tet(e) ~ tte, -tet(e) ~ -tte // -tde)
PSJ *tar (vb. *ta-n) > KGG tar (vb. to-n); lkl tol PJ *ty 'to die':
25 The accretion of Kaingang *-j is irregular.
26 The relation of PSJ *wy-n to the remaining forms is not a regular one; it might be the case that two distinct etymologies are conflated here. Moreover, the Laklano reflex unexpectedly shows a rounded vowel.
PCerr *ty(r)/*ty-k27 > PNJ *ty/*ty-k; pnr ty; PCJ *dara/*da-ka // *da (xav dara/da-?a // da, xer
dara/d-ka // da) PSJ *ty 'to die.STAT' > KGG ty 'numb'; lkl ty PJ *k(r)Vmtym' 'capybara':
PCerr *kumtym' > PNJ *kümtüm;28 pnr intyrj; PCJ *kumdam // *kumda (xav ?umda,
xer kumdam // kumda) PSJ *kryjndyr > kgg kryjndyr; (?) lkl klefun PJ *tyk 'black, dark':
PCerr *tyk 'black' > PNJ *tyk; pnr ka-ty:; PCJ *daka // *da (xav ?ra-da, xer dka 'dark') PSJ *ku-ty (vb. *ku-ty-j) 'dark, night' > KGG, lkl ku-ty (vb. ku-ty-j) PJ *tom' 'fat':
PCerr *twam' > PNJ *twam; (?) pnr tuma; PCJ *wam // *wa (xav wam // wa, xer rom-wa) PSJ *taj > kgg töj; lkl taj PJ *tum' 'old':
PCerr *tum' 'old' > PNJ *tüm; pnr tuj; PCJ *dum // *du 'tall' (xer dum), *dum-krata // *dum-
kra:da 'elder sibling of the opposite sex' (xav dub-?rata // dub-?ra:da, xer dum-krda) PSJ *toj 'dry (of plants)' > kgg toj; lkl tuj PJ *tu 'to carry':
PCerr *tu/*tu-r' > PNJ *tu/*tu-r; pnr tu-ri 'to carry in a basket.SG'; PCJ *du/*du-ri 'to carry.SG'
(xav du/du-ri, xer du/du-r(i)) PSJ *tu 'to carry on one's back, to wear' > KGG, lkl tu PJ *potu 'horsefly':
PCerr *p3tu > PNJ *pucu;29 pnr pu:su; PCJ *pidu (xav, xer pidu) PSJ *patu > KGG pötu; lkl patu PJ *tü 'NEG':
PCerr *tö > (?) pnr rö 'privative; negation in non-finite clauses'; PCJ *tö30 (xav tö 'prohibitive; negation in purpose clauses', xer tö 'privative') PSJ *tü/*tü-j (vb. *tü-j) > KGG tü/tü-j (vb. tü-j); LKL tü/tü-j PJ *te 'to go.SG':
PCerr *te/*te-m' > PNJ *te/*te-m; pnr te/te-ri 'to leave, to fall'; PCJ *ne/*ne-m // *ne 'to go.DU'
(xav ne/ne-m // ne, xer ne/ne-m(a) // ne) PSJ *ti/*ti-j (vb. *ti-n) > KGG ti/ti-j (vb. ti-n); lkl te/te-j (vb. te-n) PJ *peti 'to dream':
PCerr *p3ti/*p3ti-r' > PNJ *pyti/*pyti-r; Southern Kayapo (iupintin) (likely pyti-j) PSJ *peti/*peti-j > kgg peti/peti-j; lkl van-mbi[j]ti PJ *janda 'to send':
PCerr *jando/*jando-r' > PNJ *jando/*jando-r; pnr s-anto-ri; PCJ *fatö/*fatö-n (xav fatö/fatö-ri,
xer zatö/zatö-r(i)) PSJ ^jana/^iana-j > kgg jene/iene-j; lkl jana/jana-j PJ *ndam 'eye':
PCerr *ndom > PNJ *ndo(p-); pnr into; PCJ *tömö // *to (xav tömö/to, xer tmö/to) PSJ *ka-na (vb. *ka-na-n) 'eye, fruit' > KGG ka-ne; lkl ko-na
27 Northern Je languages point to PCerr finite *ty; Central Je languages point to PCerr finite *tyr.
28 The rounding of PCerr *y in PNJ is irregular.
29 Medial *-c- in place of the expected **-t- is irregular.
30 It is unclear why PCerr *t yields PCJ *t in a nasal environment here. The expected reflex would be PCJ **no.
PJ *na 'mother':
PCerr *na > PNJ *na; pnr na-pja; PCJ *na (xav na) PSJ *na > KGG na; lkl no PJ *no 'to lie.SG':
PCerr *no/*nop-r > PNJ *no/*no-r; pnr no; PCJ *nom-ro (xav nom-ro, xer nom-r(o) ~ nm-ro) PSJ *na > KGG na; lkl no PJ *pondv (~ *pondun) 'bad or crooked':
PCerr *p3ndu (~ *p3ndun) 'bad' > PNJ *pundu
PSJ *pandd (vb. *pando-n) 'crooked' > KGG pondo (vb. pondo-n); lkl pandd PJ *ra 'to hit':
PCerr *-ra/*-ra-n' > PNJ *ku-ra/*ku-ra-n PSJ *ra/*ra-r/*ra-n > KGG re/re-r/re-n; lkl la/la-r/la-n PJ *jara 'wing; armpit':
PCerr *jara > PNJ *jara 'wing', *jara-kre 'armpit'; pnr s-aja 'wing, feather', s-a-kre 'armpit' PSJ *6-ar 'wing' > KGG tyer; lkl d-ar; PSJ *jara (vb. *jara-n) 'armpit' > KGG jere (vb. jere-n); lkl jala PJ *ron 'celestial body':
PCerr *rwaj' 'moon' > PNJ *mbyt-rwa; PCJ *waj // *wa (xer waj // wa) PSJ *ra 'sun' > KGG ro; lkl la 'sun, day' PJ *ru 'to transport liquid or to pour':
PCerr *ru/*ru-n' > PNJ *ru/*ru-n 'to spill, to pour' PSJ *ru/*ru-r/*ru-n 'to fetch water' > KGG ru/ru-r/ru-n; lkl lu-n 'well' PJ *re 'to leave, to abandon':
PCerr *re31 > PNJ *re/*re-(r); PCJ *re/*re-me (xav re/re-me, xer re/re-m(e) ~ r-me) PSJ *re/*ra-r > KGG re/ro-r PJ *re 'to throw.PL'
PCerr *re/*re-n' > PNJ *re/*re-n
PSJ *ri/*ri-r/*ri-n > KGG ri/ri-r/ri-n 'to carry.PL'; lkl le/le-r/le-n PJ *jare (~ *jaren) 'root':
PCerr *jare (~ *jaren) > PNJ *jare; pnr s-are PSJ *jare (vb. *6-are-n) > KGG jore; lkl jale (vb. d-ale-n) PJ *-re(C) 'to weed':
PCerr *-re(C)/*-re(C)-r > PNJ *ka-re/*ka-re-r
PSJ *ku-re/*ku-re-r 'to weed, to cut clean' > KGG ku-re/ku-re-r; lkl ku-le/ku-le-r PJ *rit 'to look':
PCerr *nt > PNJ *rit; PCJ *riti // *ri:ni (xav riti // rf.ni 'to look for', xer riti ~ rti) PSJ *rir 'to wake up, to be alive' > KGG rir; lkl lel PJ *cy 'to weave':
PCerr *cy/*cyp-r' > PNJ *cy/*cy-r; PCJ *na/*nam-ri (xav na/nam-ri, xer nam-r(i)) PSJ *6y > KGG $y; LKL dy PJ *cym 'seed':
PCerr *cym > PNJ *cy; pnr sy; PCJ *nama // *ja (xav nama //ja, xer za) PSJ *6y (vb. *6y-n) > KGG fyy (vb. 4>y-n); lkl dy PJ *co 'to eat soft food, to suck':
PCerr *co/*co-r' > PNJ *cd/*co-r; pnr sdw-ri
PSJ *pe-6a (vb. *pe-6a-m) 'to suck breast' > KGG pa-tya32 (vb. pa-tyd-m); lkl pe-do (vb. pe-da-m)
31 It is unclear how the non-finite form of this verb should be reconstructed.
PJ *côj 'leaf':
PCerr *côj' > PNJ *cô; pnr parï-sô; PCJ *cuj // *cu (xav wê-cuj-ra, -cuj // -cu, xer su) PSJ *Qej (vb. *de-n) > kgg tyej; lkl dej PJ *jacê ~ *jVmcê 'nest':
PCerr *jacê > PNJ *jacê; pnr s-asê; PCJ *faci (xav faci, xer zasi) PSJ *jajdê (vb. *jajde-n) > KGG jajtyê (vb. jajtye-n) PJ *-cï(C) 'bandage, sling':
PCerr *-cï(C) > PNJ *ja-cï, *ka-cï
PSJ *-Qï (vb. -dï-n) 'bundle, bandage' > KGG -tyï (vb. -tyï-n); lkl -dë (vb. -dë-n) PJ *kapnfa 'to chew':
PCerr *kapnfa/*kapnfa-r' > PNJ *kapnfa/*kapnfa-r; PCJ *waca/*waca-ri (xer wasa/wasa-r(i)) PSJ *kaja > KGG kajë; lkl ka[j]ja/ka[j]ja-j PJ *nfy 'bee, bumblebee':
PCerr *(am-)nfy > PNJ *(am-)nfy; PCJ *am-fa (xav am-fd 'Tetragona clavipes', xer am-zd
'Brazilian wasp (Protonectarina sylveirae)') PSJ *cy > KGG xy; LKL cy PJ *nfô 'to hang.SG':
PCerr *nfô/*nfôp-r' > PNJ *nfô/*nfô-r; PCerr *ja-nfô/*ja-nfôp-r' 'to hang.PL' > PNJ *ja-nfô/
*ja-nfô-r; PCJ *fa-cô/*fa-côm-rï (xavfa-cô/fa-côm-rï, xer za-sô/za-sôm-r(ï)) PSJ *ca/*ca-j (vb. *ca-m) > KGG xa/xa-j (vb. xa-m);33 lkl co/ca-j PJ *nfê 'to tickle':
PCerr *nfê/*nfê-r' > PNJ *nfê/*nfê-r PSJ *ku-cê/*ku-ce-j > KGG ku-xê/ku-xe-j; lkl ku-cê/ku-ce-j PJ *ja 'to stand.SG':
PCerr *ja/*ja-m' > PNJ *fa/*ja-m; pnr s-a:-j; PCJ *fa/*fam // *fa (xavfa/fam // fa, xer za ~ da) PSJ *ja/*ja-j > kgg jë/jë-j; lkl ja (vb. ja-j) PJ *ja 'to put vertically.SG'
PCerr *ja/*ja-r' > PNJ *fa/*fa-r; pnr sa-rï; PCJ *ca/*cà-n (xav ca/ca-rï, xer sa/sa-r(ï)) PSJ *ja (vb. *ja-j) 'to be put vertically' > KGG ty-ë (vb. ty-ë-j); lkl vb. ja-j, 3 d-a-j PJ *nïja 'nose':
PCerr *nïja-kre > PNJ *nïja-kre; (?) PCJ *nïci-krê (xav nïci-krê, xer n-krê) PSJ *nïja > KGG nïjë; lkl nëja PJ *jan 'to eat.iNTR, food'
PCerr *faj' > PCJ *caj // *ca (xav caj // ca, xer saj // sa) PSJ *ja/*ja-j/*ja-n (vb. *ja-n) > KGG *jë/*jë-j/*jë-n (vb. jë-n); lkl ja-n PJ *ja 'bitter':
PCerr *ja > PNJ *fa; PCJ *fe (xav fe, xer ze) PSJ *Q-a (vb. *d-a-j) > kgg ty-a (vb. ty-ô-j); lkl d-o (vb. d-a-j) PJ *nïja 'smoke':
PCerr *nïja > PCJ *-nïfe (xav -nïfe, xer -nïze) PSJ *nïja (vb. *nïja-j) > kgg nïja (vb. nïjô-j); lkl nïjo PJ *jr 'urine':
PCerr *ja > PNJ *fa; PCJ *fe (xav fe, xer ze 'bladder') PSJ *ja-n 'to urinate' > KGG ja-n; lkl ja-n
32 The expected form would be KGG *pe-tya (vb. *pe-tyô-m), given both the Laklânô cognate and the Kaingang plural pi-j-tya (vb. pi-j-tyô-m).
33 Stative KGG xaw 'to hang' is apparently a non-etymological formation, back-derived from xa-m.
PJ *ny 'to sit.SG':
PCerr *ny/*nyp-r > PNJ *ny/*ny-r; pnr si-r/si-ri; PCJ *nam-ra (xav nam-ra, xer nam-ra) PSJ *ni (vb. *ni-m) > KGG ni (vb. ni-m); lkl ne (vb. ne-m) PJ *jo 'tooth':
PCerr *jwa > PNJ *jwa; pnr swa; PCJ *kwa (xav, xer kwa) PSJ *ja > KGG jo; lkl ja PJ *kVjo 'salt':
PCerr *kVjwa > PNJ *kajwa; PCJ *kVkwa ((?) xav 7i?wa-wa:ha, xer kakwa-ra) PSJ *kaja (vb. *kaja-r) 'salty, sour' > KGG kajo (vb. kajo-r) PJ *jd 'to tear, to rip':
PCerr *-jd/-jd-r' > PNJ *ka-jo/*ka-jo-n
PSJ *ja/*ja-r, *ja-r (vb. *ja-n) > KGG ja-r (vb. jo-n); lkl da-jo/da-ja-r, da-jol (vb. da-ja-n) PJ *ju 'pus':
PCerr *jup-r > PNJ *jur; PCJ *jubruj // *jubru (xav jubruj // jubru, xer zbruj // zbru) PSJ *6-o (vb. *6-o-m) > KGG ty-o (vb. <p-o-m); lkl d-o (vb. d-u-m) PJ *ju, postposition:
PCerr *ju > PNR su: 'looking for'; PCJ *jo34 'looking for; for' (xav jo, xer zo) PSJ *jo 'in front of' > KGG jo 'in front of; else'; lkl jo 'in front of' PJ *jum' 'father':
PCerr *jum' > PNJ *ju(m); pnr jum-pja PSJ *jor > kgg jor; lkl jur PJ *peju 'to hide':
PCerr *pju/*pju-r' > PNJ *puju/*puju-r PSJ *peju > KGG peju; lkl peju PJ *nju 'to heat, to dry':
PCerr *nju/*nju-n' > PNJ *nju/*nju-n 'to dry in the sun'; Southern Kayapo (timu^unkuatu,
kuataungun) (likely nsu-r) PSJ *cu-r (vb. *cu-n) 'to heat by the fire' > KGG cur (vb. cu-n); lkl cul 'cooked half'; PSJ *jar-cu/*jar-cu-r, *jar-cu-r (vb. *jar-cu-n) 'to roast' > KGG jor-xu/jor-xu-r, jor-xu-r (vb. jor-xu-n); lkl jar-cu/jar-cu-r, jar-cu-l (vb. jar-cu-n) PJ *ut/*nut 'to sleep':
PCerr *rdt/*ndt > PNJ *rdr/*ndt■; pnr s-otV; PCJ ^noto // ^norno (xav noto // nam, xer mUo) ~ nto) PSJ *nur > KGG nur; lkl nul PJ *je (~ *jen) 'thread, knot, cloth': PCerr *je (~ *jen) > PNJ *je; pnr se PSJ vb. *6e-n 'to spin thread' > KGG tye-n; lkl de-n PJ *numje(C) 'female breast':
PCerr *mmje(C) > PNJ ^nomje
PSJ *nurje 'female breast, milk' > KGG nurje; lkl nurnje PJ *kujek 'vein':
PCerr *kujek > PNJ *kujek PSJ *kujej > KGG, lkl kujej PJ *ji 'to lay.SG':
PCerr *ji/*ji-r' > PNJ *ji/*ji-r; PCJ *hi/*hi-ri (xav hi/hi-ri, xer hi/h(i)-ri) PSJ *ji/*ji-r > kgg ty-i/ty-i-r; lkl ji/ji-r
34 The expected reflex would be PCJ *ju. Note that the vowel o is extremely rare in Central Je, its occurrence being restricted to a few function words.
PJ *jiji ~ *ñíji 'name':
PCerr *ñíji > PNJ *ñí-ji; pnr issi; PCJ *ñíci // *ñí:ci (xav ñíci // ñí:ci, xer nísi-ze) PSJ *jiji ~ *jyjy (vb. *jiji-n ~ *jyjy-n) > KGG jiji ~ jyji ~ jyjy (vb. jiji-n); lkl jyjy, 3 d-yjy (vb. jyjy-n, 3 d-yjy-n) PJ *ñí 'meat':
PCerr *ñí > PNJ *ñí; pnr ñí; PCJ *ñí (xav ñí, xer ní) PSJ *ní > KGG ní; lkl né PJ *ñí(m)- 'hand (in compounds)'':
PCerr *ñí(m)-kra 'hand' > PNJ *ñí-kra ~ *ñy-kra; pnr s-i-kja; PCJ *ñíp-kra (xav ñíp-?rata // ñíp-?ra:da, xer níp-kra)
PSJ *ní- > KGG ní- (as in ní-pé 'to wash hands', ní-ju/ní-ju-j 'to show with one's hand', ní-jge (vb. ní-jo-n) 'hand'); lkl (as in né-ju-jo 'index finger', né-jga 'hand') PJ *kaj' 'basket':
PCerr *kac > PNJ *kac; (?) pnr kaj PSJ *kaj (vb. *kañ) > KGG kéj; lkl kaj PJ *kyñ 'skin, bark', *jar-kyñ 'lip':
PCerr *kyj' 'skin, bark' > PNJ *ka; pnr ka; PCJ *hdj // *hd (xav ha, xer kaj ~ hé // ha); PCerr *jaj-kyj' 'lip' > PNJ *ja-ka; pnr s-a-ka; PCJ *jaj-haj // *jaj-ha (xav jaj-haj // jaj-ha, xer zaj-ha 'buttocks') PSJ *jan-ky (vb. *jan-ky-n) 'mouth' > KGG jén-ky (vb. jén-ky-n); lkl jan-ky PJ *ko 'to dig':
PCerr *kwa/*kwa-ñ' > PNJ *kwa/*kwa-ñ; pnr kwa-j PSJ *ka/*ka-j (vb. *ka-n) > lkl ka/ka-j (vb. ka-n) PJ *jar-ko 'mouth':
PCerr *jad-kwa > PNJ *jar-kwa; pnr s-a-kua; PCJ *jada-wa (xav jada-wa, xer zda-wa 'mouth, door')
PSJ *jan-ka 'door' > KGG jon-ko; lkl d-an-ka PJ *kVñko 'sky':
PCerr *kajkwa > PNJ *kajkwa; PCJ *hajwa (xav hajwa, xer hajwa ~ hewa) PSJ *kañka > KGG kañko; lkl koñka PJ *kop 'fly, mosquito':
PCerr *kdp 'fly' > PNJ *kop; PCJ *kupu // *ku:bu (xav ?upu // ?u:bu, xer kpu // kbu) PSJ *ka 'mosquito' > KGG ka; lkl cé-to-ko 'black fly' PJ *kom 'tree, horn'; *ñí-kom 'horn':
PCerr *kom > PNJ *ko; PCJ *komo // *ku 'horn' (xav ?omo // ?u, xer kmo // ku); PCerr *ñí-kom
'horn' > PNJ *ñí-kd PSJ *ka 'tree' > KGG ka; lkl ko; PSJ *ní-ka 'horn' > KGG ní-ka; lkl né-ko PJ *kok ~ *ka-kdk 'wind':
PCerr *kok, *ka-kok > PNJ *kok; PCJ *wa-kuku // *wa-ku (xav -wa-?u?u // -wa-?u, xer wa-kku) PSJ *ka-ka (vb. *ka-ka-n) > KGG ko-ka (vb. ko-ka-n) PJ *ku 'to eat':
PCerr *ku/*ku-r' > PNJ *ku/ku-r 'to eat.PL'; pnr ku-ri; PCJ *hu/*hu-ri (xav hu/hu-ri) PSJ *ko 'to eat, to use' > KGG, lkl ko PJ *két 'stone':
PCerr *két > PNJ *kén; pnr kíéj; PCJ *kété // *ké:né (xav ?été // ?é:né, xer kté // kné) Ingain kér: Lista (queré, quené), María Antonia (keré), Cosme Román (kiné), Vogt3 (kiré) PJ *két 'NEG':
PCerr *két > PNJ *két
PSJ *kér 'advt' > KGG kér; lkl kél-o
PJ *kij 'pit'
PCerr *kjaj' > PNJ *ki 'earth oven'; pnr kje 'earth oven' PSJ *kej 'tomb' > kgg kej PJ *ki ~ *kij ~ *kij' 'to split'
PCerr *-kja/*-kja-r' > PNJ *ko-ki/*ko-kje-r
PSJ vb. *ke-n > KGG ke-n; PSJ vb. *ka-ke-n 'to split; canoe' > KGG ko-ke-n; lkl ka-ke-n PJ *kra 'offspring':
PCerr *kra > PNJ *kra; PCJ *kra: // *kra (xav ?ra: // Ira, xer kra) PSJ *kra (vb. *kra-y) > KGG kre (vb. kre-y); lkl kra PJ *krat 'macaw':
PCJ *krata // *kra:da (xav ?rata // ?ra:da, xer krda)
Ingain kla(r): Pedro (klua), Maria Antonia (klan), (?) Vogt1 (kakladein), Vogt3 (kla) PJ *kryn 'head':
PCerr *kraj' > PNJ *kra; pnr kja; PCJ *kraj // *kra (xav ?raj // ?ra, xer kraj ~ kre // kra) PSJ *kri (vb. *kri-n) > KGG kri (vb. kri-n); lkl kle PJ *jVkryn 'knee':
PCerr *jikraj' > PCJ *hikraj // *hikra (xav hi?ra-ti, xer hi-kraj-ti ~ hi-kre-ti, hi-kraj-) PSJ *jakri > KGG jakri; lkl jokle PJ *krot 'chin, beak':
PCerr *krwat 'beak' > PNJ *krwat; PCJ *wata // *wa:da (xav wata // wa:da 'chin, beak', xer wda) Ingain jat-krar 'chin': Lista (amincrara), Maria Antonia (miet krara), Vogt3 (jitkyrara) PJ *kre 'hole':
PCerr *kre > PNJ *kre; pnr kre; PCJ *kre 'vagina, anus' (xav ?re, xer kre), *am-kre 'hole' (xav
am-?re, xer am-kre) PSJ *kre > KGG kre; lkl kle PJ *kre 'to plant':
PCerr *kre > PNJ *kre; pnr kre; PCJ *kre (xav ?re, xer kre)
PSJ *kre (vb. *kra-n) > KGG kre (vb. kro-n); lkl klel35 (vb. kla-n); PSJ *a-kre/*ja-kre 'plantation' (vb. *a-kra-n/*ja-kra-n) > KGG e-kre/ja-kro-n PJ *krin 'thigh':
PCerr *krjaj' > PNJ *kje; PCJ Jaj // Ja (xav faj // fa, xer zda) PSJ *kre > KGG kre; lkl kle PJ *ya(C) 'to push against, to grind': PCerr *ya/*ya-n' > PNJ *ya/*ya-n PSJ > KGG ya
PJ *ygon' 'feather, hair':
PCerr *ygwan' > PNJ *ygwan 'feather'; pnr iykway PSJ *ygan 'animal hair' > KGG yon 'hair'; lkl ygan 'horsehair' PJ *yg° 'louse':
PCerr *ygo > PNJ *ygo; pnr kja-yko; PCJ *ku (xav ?u) PSJ *yga (vb. *yga-n) 'louse, flea' > KGG yga (vb. yo-n); lkl ygo (vb. -yga-n) PJ *yu 'to push against, to crumble':
PCerr *-yd/*-yd-n' > PNJ *ka-yo/ka-yo-n PSJ *yu/*yu-y > KGG yu/yu-y
35 Final -l in the Laklano form is apparently non-etymological. It might have arisen as a back-derivation from the regular verbalized form kla-n.
PJ *jgi 'to enter.PL':
PCerr *a-rgjal*rgja-c > PNJ *a-ygî/*ygjê-c, pnr ijkjd; PCJ *d:-jalja-ci 'to enter.DU' (xav ?§:-jal
fa-ci, xer za-s(i)) PSJ *jgê (vb. *qgê-m) > KGG jgê (vb. jgê-m); lkl jgê (vb. jgê) PJ *rjru (~ *rjrun) 'toucan':
PCerr *jro > PNJ *jrô; pnr irkjô-pepetï PSJ *jru > KGG jru; lkl jlu PJ *jgre 'egg':
PCerr *jgre > PNJ *jgre; pnr ijkre; PCJ *krê (xav ?rê, xer krê) PSJ *jgre 'egg, penis' > KGG jgre 'penis'; lkl jgle PJ *jgre 'to dance':
PCerr *rgrel*rgre-r > PNJ *jgrel*rgre-r; pnr ijkre:; PCJ *aj-krël*ci-krë-në (xav aj-?rëlaj-?rë-në ~ ci-?rë-në)
PSJ *jgre (vb. *jgre-n) > KGG vb. jgre-n; lkl jgle (vb. jgle-n) PJ *jn(C) 'wrap':
PCerr *-jn(C) > PNJ *ku-jrï 'to gather in a bundle', *-jrï 'to make packages'
PSJ *rrïl*rrî-j, *jrï-r (vb. *jrï-n) > KGG rrîlrrî-rlrrî-n; lkl jlëljlë-j, jlë-r (vb. jlë-n)
Semantic issues
PJ *pu:
PCerr *pu > PNJ *pu 'tube (in compounds)'; PCerr *jgrwa-pu 'moriche stem' > PNJ *jgrwa-pu,
PCJ *wa:-bu (xav wa:-bu, xer wa-bu) PSJ *pu (vb. *pu-j) 'handle' > kgg pu (vb. pu-j); lkl vb. pu-j PJ *prït:
PCerr *prït > PNJ *prït 'pequi'
PSJ *prïr > KGG prïr 'araucaria resin'; (?) lkl plël 'reddish' PJ *wë(C):
PCerr *wë(C) > PNJ *bël*bë-r 'to show', *bë-r 'speech'; pnr pëjlpëj-n 'to say' PSJ *wï (vb. *wï-n) 'to speak' > KGG wï (vb. wï-n); lkl vë PJ *më:
PCerr *mël*më-n' 'to throw.SG' > PNJ *mël*më-n; pnr më-nlmë:-rï; PCJ *mël*më-j ll *më (xav
mëlmë-j ll më, xer më) PSJ *mïl*mï-j/*mï-n 'spill' > KGG mïlmï-j/mï-n PJ *tit:
PCerr *tjat 'to burn' > PNJ *tjêrl*tjêt; pnr titï; PCJ *jata ll *ja:da (xav fata ll ja:da, xer zata) PSJ *têr (vb. *te-n) 'to die.ACT' > KGG têr (vb. te-n); lkl têl (vb. te-n) PJ *rdt:
PCerr *-rot > PNJ *nï-rot 'crooked' PSJ *rar 'sharp-ended; thorn' > KGG rër; lkl lal PJ *ro:
PCerr *rwal*rwa-k > PNJ *rûl*rwd-k 'to descend' PSJ *ra 'to enter.SG, to begin.SG' > KGG rô; lkl la PJ *kvrdt:
PCerr *kurôt > PNJ *kurôt 'skin irritation' (only Apinajé) PSJ *kôrar > KGG kôrar 'frog' PJ *rê:
PCerr *rêl*rê(p)-r > PNJ *rêl*rê-r 'to cross'
PSJ *rê (vb. *rê-m) 'to descend.PL' > KGG *rê (vb. rê-m); lkl lê (vb. lê-m)
PJ *nfa 'to bite':36
PCerr *nfa/*nfa-r' 'to bite' > PNJ *nfa/*nfa-r; pnr insa-ri; PCJ *ca/*ca-ri (xav ca/ca-ri, xer ca/ca-r(i)) PSJ *ca 'game (animals); war' > kgg xe PJ *nfum':
PCerr *nfum' > PNJ *nfum 'dirty'
PSJ *coy 'egg yolk; larvae's food' > KGG xoy; (?) lkl cuy 'bee sp. (Plebeia sp.) PJ *nfi(C):
PCerr *nfi(C) > PNJ *nfi 'mother'; (?) PNR si-pja 'wife' PSJ *ci 'old' > KGG xi; lkl ci PJ *ygre:37
PCerr *ygre > PNJ *ygre 'few'
PSJ *ygre/*ygre-y 'to sieve' > KGG ygre/ygre-y; lkl ygle/ygle-y
Phonetic issues
PJ *pu ~ *pu (if PSJ is exceptional) ~ *po (if PCerr is exceptional) 'to wrap': PCerr *ku-pu > PNJ *ku-pu; PCJ *ku-bu (xav ?u-bu) PSJ *pa (vb. *pa-y) > KGG po (vb. po-y); lkl pa (vb. pa-y) PJ *wYr' (if PSJ is exceptional) ~ *war' (if PNJ/PCerr is exceptional) 'tree sp.':
PCerr *wad ~ *wad 'tree' > PNJ *bar; pnr pari; PCJ *wede // *we:de (xav wede // we:de, xer wde) PSJ *wan 'bamboo' > KGG won; lkl van PJ *tYm' (if PSJ is exceptional) ~ *tam' (if PNJ/PCerr is exceptional) 'new': PCerr *tam' ~ *tam' > PNJ *tam 'raw'; PCJ *tem // *te (xav, xer tem // te) PSJ *tay > KGG toy; lkl tay PJ *tut (if PSJ is exceptional) ~ *trn' (if PNJ/PCerr is exceptional) 'pigeon': PCerr *tut > PNJ *tut PSJ *ton > KGG tdn
PJ *jar-ra (if PSJ is exceptional) ~ *jar-rY (if PNJ/PCerr is exceptional) 'saliva':
PCerr *jad-ra ~ *jad-ra > PNJ *jar-ra; PCJ *fare (xer zdare me 'to spit' = 'to throw saliva') PSJ *jara (vb. *jara-n) > KGG jora (vb. joro-n); lkl jalo PJ *jVmV 'cheek':
PCerr *jamV > PNJ *jama PSJ *jama > KGG jame; lkl joma PJ *ka (if PSJ is exceptional) ~ *ky (if PNJ/PCerr is exceptional) 'bad smell': PCerr *ka > PNJ *ka PSJ *ky > KGG, lkl ky PJ *ndon (if PSJ is exceptional) ~ *ndun (if PNJ/PCerr is exceptional) 'snail': PCerr *ndwan' > PNJ *ndwan; pnr intdw PSJ *ndun > KGG ndon nune, ndun; lkl ndun 'caterpillar' PJ *kum' (if PSJ is exceptional) ~ *ygum' (if PNJ/PCerr is exceptional) 'mist, cloud':
PCerr *kum' ~ *ygum' > PNJ *kum 'smoke, mist'; PCJ *hum // *hu; PCerr *ka-kum' 'cloud, dry season' > PNJ *ka-kum 'cloud'; (?) pnr a-kuy 'dry season'; (?) PCJ *wa-hum // *wa-hu 'dry season, year'
36 Despite the semantic divergence between the Cerrado languages and Kaingang, which makes the comparison dubious, the Proto-Cerrado root may safely be identified as a retention from Proto-Jê, given that a secure external cognate exists in Maxakali: /ca/ 'to bite, to sting'.
37 If this comparison is correct, the Proto-Jê word could have meant 'small' and correspond externally to Maxakali /ktïk-nâk/ 'small'.
PSJ *ygoy 'cloud' > KGG ygoy; lkl yguy PJ *yg°j (if PSJ is exceptional) ~ *yguj (if PNJ/PCerr is exceptional) 'water'
PCerr *ygoj ~ *yguj > PNJ *ygo; pnr iyko; PCJ *kuj // *ku 'still water' (xav ?uy // ?u, xer kuj-) PSJ *ygoj (vb. ygon) > KGG ygdj (vb. ygon); lkl ygdj
Reflexes only in PCerr
PCerr *por/*pok 'to ignite' > PNJ *por/*pok; pnr po 'to burn'
Possible PJ reconstruction: *pok. External cognate: Maxakali /pyk/ 'to burn (vi)', Rikbaktsa pok, Proto-Tupian *prnk.
PCerr *a-pe/*jV-pe-n' 'to make' > PNJ *a-pe/*ja-pe-n 'to work'; pnr s-a-pe 'to work'; PCJ *ap-pi/*nlp-pi (xav ap-pi/nlp-pi 'to cook', xer ni-pi)
Possible PJ reconstruction: *-pe. External cognate: Krenak pi. PCerr *kupe/*kupe-n' 'to touch' > PNJ *kupe/*kupe-n; PCJ *kupi (xav ?upi, xer kupi)
Possible PJ reconstruction: *-pe. External cognate: Maxakali /REL-pit/. PCerr *pra(j') 'feather, corn husk' > PNJ *pra
Possible PJ reconstructions: *pra, *praj, or *pran. External cognate: Maxakali /ptyc-nak/ 'bird'. PCerr *mbra38 'to walk.PL' > PNJ *mbra/*mbra-r; PCJ *kma-pra/*kma-pra-ba 'to take away' (xav ?ma-pra/?ma-pra-ba, xer kma-pra/kma-pra-ba ~ kma-pra-b ~ kma-pr-ba), *pra-ba 'to run' (xav pra-ba, xer pra-ba ~ pra-b ~ prb-a 'to dance') Possible PJ reconstruction: *mbra(C). External cognate: Rikbaktsa parak 'to walk, to run.SG'. PCerr *mbwa/*mbwa-r' 'to cry' > PNJ *mbu/*mba-r
Possible PJ reconstruction: *mbo. External cognates: Maxakali /pu-t-a, IRR pu/ (Ritual Maxakali /pu-t, irr pu/), Krenak pu-k, Karaja bu, Proto-Jabuti *mbo. PCerr *-mra(m) 'to bathe (vt.)' > PNJ *kumra
Possible PJ reconstruction: *-mry or *-mrym. External cognates: Maxakali /mnyp/ 'to submerge', (?) Rikbaktsa para (unless related to PSJ *mbro 'to swim, to bathe' > KGG mbro, lkl mbld). PCerr *mro(C) 'to dive, to soak; to cook' > PNJ *mro
Possible PJ reconstruction: *mru(C) or *mro(C). External cognates: Rikbaktsa muru 'to bathe (vi)', Proto-Jabuti *mra 'to bathe (vi)'. PCerr *wd/*wa(p)-r 'to smell' > PNJ *bd/*ba-r; (?) Southern Kayapo (tipen) (likely pa)
Possible PJ reconstruction: *wa. External cognates: Maxakali /cy-pyp/ 'nose', /ca-pyp/ 'pig', Ritual Maxakali /byp/ 'to smell', /ca-byp/ 'pig', Krenak wap 'to smell, to kiss'. PCerr *woc 'to arrive' > PNJ *bdj/*boc; pnr pow/pow-ri; (?) PCJ *wi/*wi-ci 'to arrive.SG' (xav wi/wi-ci, xer wi/w(i)-si)
Possible PJ reconstruction: *woc. External cognate: Proto-Tupian *wmc 'to arrive (Juruna), to go out (Karo, Purubora)'. PCerr *we 'abl' > PNJ *be 'abl, malef, cop'; pnr pe:; PCJ *wi (xav, xer wi) Possible PJ reconstruction: *we. External cognate: Proto-Tupian *wi. PCerr *jawe 'to love, to respect' > PNJ *jabe 'to love, to be melancholy'; PCJ *fawi (xav fawi, XER zawi)
Possible PJ reconstruction: *jawe. External cognate: Maxakali /capit/ 'to call, to invite'. PCerr *wi/*wi-r 'to kill.SG' > PNJ *bi/*bi-r; pnr pi-ri; PCJ *wi/*wi-ri (xav wl/wi-ri, xer wl/w(l)-ri) Possible PJ reconstruction: *wl. External cognates: Maxakali /mi-k, irr mi/ 'to make' (Ritual Maxakali 'to kill'), Proto-Tupian *wl 'to kill'.
38 It is unclear how the non-finite form of this verb should be reconstructed.
PCerr *toj' 'brother' > PNJ *to; pnr to; PCJ *noj // *no 'younger sibling of the same sex' (xav noj // no)
Possible PJ reconstruction: *tuj, *toj, *tun, or *ton. External cognate: Maxakali /nuc/. PCerr *te 'gen, erg' > PNJ *te 'erg, gen (material, stimulus)'; PCJ *te (xav te 'Isg.erg', wa-te 'Ipl.erg', (te-)te '3.ERG', xer te) Possible PJ reconstruction: *te. External cognates: Maxakali /te/, (?) Krenak ti 'I', ho-ti 'you'. PCerr *tik 'belly' > PNJ *tik; PCJ *diki // *di (xav di?i // di, xer dki // di)
Possible PJ reconstruction: *tik. External cognates: Maxakali /tek/. PCerr *ndap 'sour' > PNJ *ndap 'sour, ripe'; PCJ *wam-tapa 'sour, bitter' (xer wam-tap(a) ~ wam-tpa)
Possible PJ reconstruction: *ndap. External cognate: Proto-Tupian *ndap. PCerr *ndep 'ripe' > PNJ *ndep
Possible PJ reconstruction: *ndep. External cognate: Tupari tep (if from Proto-Tupian *ndep). PCerr *p3rdk 'to look like' > PNJ *pyrak
Possible PJ reconstruction: *pVrrk. External cognates: Maxakali /pytyk/. PCerr *r3(r) 'flower' > PNJ *r3; pnr ij3; PCJ *ni-r3r3 ~ *ni-r3n3 (xav ni-r3r3, xer ni-r3n(3) ~ ni-rn3)
Possible PJ reconstruction: *r3 or *r3r. External cognate: Maxakali /-dyt/. PCerr *kucym 'fire' > PNJ *kucy; pnr issy; PCJ *kun3m3 // *kujd (xav ?un3m3 // ?ufa, xer kunm3 // kuz3)
Possible PJ reconstruction: *kucym. External cognates: Maxakali /kycap/, Karaja he-kody. PCerr *ci 'bone' > PNJ *ci; PCJ *hi (xav, xer hi)
Possible PJ reconstruction: *ji or *ci. External cognates: Krenak jek, Karaja di, Ofaye hi ~ hih, Rikbaktsa ek 'leg', Jabuti *ji, *i, Chiquitano -i 'leg, plant', -pa-?i 'bone'. PCerr *j3n' 'sweet, tasty' > PNJ *j3n'; PCJ *jej // *je (xav jej // je, xer zej // ze)
Possible PJ reconstruction: jn'. External cognates: Maxakali /cyc-pek/, Tupari hoc (if from Proto-Tupian *joc ~ *joc). PCerr *jwan' 'nmlz.ag' > PNJ *jw3n, PCJ *-kwaj // *-kwa
Possible PJ reconstruction: *jon'. External cognate: Karaja /-du/. PCerr *jajwa/*jajwa-r' 'to lay.PL, to spill' > PNJ *jaju/*jajw3-r; pnr s-asw3-ri 'to spill'; PCJ *c-a(j)kwa/*c-a(j)kwa-ri (xav c-a?wa/c-a?wa-ri, xer s-ajkwa/s-ajkwa-r(i)) Possible PJ reconstruction: *-jo. External cognate: Maxakali /ny=cu-k, irr ny=cu/. PCerr *jwa(m)/*jwa(m)-r 'to bathe' > PNJ *ju/*jw3-r; pnr sw3-ri
Possible PJ reconstruction: *jom. External cognate: Krenak jum, Proto-Jabuti *d^o. PCerr *pj 'one' > PNJ *pyji; PCJ *pici (xer pici 'only')
Possible PJ reconstruction: *pVji. External cognates: Maxakali /pycet/, (?) Krenak pucik. PCerr *nVt 'faeces, guts' > PNJ *nin; PNR ni:; PCJ *n3n3 // *n3:n3
Possible PJ reconstruction: *nVt. External cognates: Maxakali /nut/, Chiquitano -a?a. PCerr *kwar 'manioc' > PNJ *kw3r; pnr kwy
Possible PJ reconstruction: *kor. External cognate: Maxakali /kut/. PCerr *kwa(C)/*kwa(C)-r 'to defecate' > PNJ *ij-ku/*kw3-r
Possible PJ reconstruction: *ko(C). External cognates: Karaja ku. PCerr *kryt 'metal, flint' > PNJ *kryt; PCJ *hdtd // *hd:dd (xav hddd, hdtd-ra 'ax', xer hdd)
Possible PJ reconstruction: *kryt. External cognate: Krenak krak 'metal, knife'. PCerr *kroj' 'rotten' > PNJ *kro; PCJ *kroj // *kro (xav ?roj // ?ro, xer kroj // kro)
Possible PJ reconstruction: *krdj or *kran. External cognates: Maxakali /ktuc/, Rikbaktsa horo.
PCerr *kre/*kre-r 'to eat.SG' > PNJ *kre/*kre-r; PCJ *kre/*kre-ne (xav ?re/?re-ne, xer kre/kr(e)-ne)
Possible PJ reconstructions: *kre or *krey'. External cognate: Maxakali /knyk/ 'to have sex with'.
PCerr *rgryk 'angry' > PNJ *rgryk; PCJ *haka // *ha (xav ha?a // ha)
Possible PJ reconstruction: *ygryk. External cognate: Krenak ygray 'angry, snake'.
PCerr *^a/*jap-r 'to roast' > PNJ *ga/*fa-r; PCJ *feb-re (xav feb-re, xer deb-r(e))
Possible PJ reconstruction: *a (*/ap'/). External cognates: Maxakali /mu=hap/ 'to roast, to burn', Krenak op.
PCerr *yo/*nop-r' 'to give' > PNJ *yo/*no-r; pnr s-o-ri; PCJ *co/*com-rl (xav co/com-rl, xer so/som-r(l)) Possible PJ reconstruction: *u or *o (*/up' ~ op'/). External cognates: Maxakali /hup/, Krenak um, Karaja o, Proto-Jabuti *u.
Reflexes only in PSJ
PSJ *-pe/*-pe-y 'to wash' > KGG -pe/-pe-y; lkl -pe/-pe-y
Possible PJ reconstructions: *-pe or *-pi. External cognates: Maxakali /pi-k, irr pi/, Rikbaktsa pik, Proto-Jabuti *pi.
PSJ *pra/*pra-y 'to bite' > KGG pra/pro-y; lkl plo/pla-y
Possible PJ reconstructions: *pro, *prop, *prom, *pra, *prap, or *pram. External cognates: Maxakali /ptup/, Karaja ro, Rikbaktsa boro.
PSJ *m§y 'bee, honey' > KGG may; lkl moy
Possible PJ reconstruction: *may\ External cognates: Maxakali /pyk/, Krenak pay, Ofaye tyok ~ fyak ~ fydk ~ $dgn ~ tyok-tdj ~ tydg-tdj.
PSJ *mbey 'ax' > KGG, lkl mbey
Possible PJ reconstructions: *mbem' or *mbim'. External cognate: Maxakali /pip-/.
PSJ *wa 'rotten' > KGG wa
Possible PJ reconstruction: *wYm. External cognates: Krenak wam, (?) Ofaye fya:?.
PSJ *rey-re (vb. *rey-re-y) 'two' > KGG rey-re (vb. rey-re-y); lkl ley-le (vb. ley-le-y) Possible PJ reconstructions: *-re or *-ri. External cognate: Maxakali /tik/.
PSJ *rir 'soft (of corn)' > KGG rir
Possible PJ reconstruction: *rit. External cognate: Maxakali /te, IRR te-t/ 'to prepare'.
PSJ *da/*da-y 'to wash (clothes)' > KGG tya/tyd-y; lkl do/da-y
Possible PJ reconstructions: *jo(C), *ja(C), *co(C), or *ca(C). External cognate: Ofaye xoh.
PSJ *du 'vagina' > KGG $u; lkl du
Possible PJ reconstructions: *ju or *cu. External cognate: Krenak jo, Ofaye xa.
PSJ *jur 'to arrive' > KGG jur; lkl jul
Possible PJ reconstruction: *jut. External cognate: Maxakali /mu-ca, IRR mu-ca-t/ 'to arrive/leave.PL'.
PSJ *ki 'LOC' > KGG, LKL ki
Possible PJ reconstruction: *ki(C). External cognate: Karaja ki.
PSJ *yga 'earth' > KGG yga, lkl ygo
Possible PJ reconstruction: *ygo(C) or *yga(C). External cognates: Chiquitano ky-, Proto-Tupian *krnc.
5. Conclusions
Above I have exposed what I believe to be the first modern proposal regarding the reconstruction of Proto-Je phonology and basic vocabulary.
Some of the ideas advanced in this paper may appear to be somewhat radical, such as the reconstruction of five contrastive vowel heights and of two coda types; these were, however,
the most parsimonious explanations I could find. It is nevertheless possible that an entirely different reconstruction will emerge in the future that will also account for the data.
Future research will have to tackle the issue of the unexpected nasalization in Central Jê, which remains completely unexplained in my current framework.
Another issue that remained beyond the scope of this paper is the reconstruction of Proto-Jê morphosyntax. It appears possible to project many morpho syntactic features shared by Northern Jê and non-Jê Macro-Jê languages, such as Maxakalí, onto the Proto-Jê level, even though a detailed study that would take into account Central Jê and Southern Jê data has not been carried out so far. These features include:
• AOV/SV, dependent-head constituent order;
• split-S alignment in clauses headed by finite verbs; ergative-absolutive alignment in clauses headed by non-finite verbs (cf. Castro Alves 2010, Nikulin & Silva to appear);
• obligatory expression of preposed internal arguments, leading to the occurrence of "expletive" third person markers in case of arguments dislocated from their canonical position;
• a clear-cut distinction between inflectable and non-inflectable nouns, whereby non-inflectable nouns must be preceded by a possessive classifier or a genitive postposition in order to be possessed, among other features.
Inflectional and derivational morphology of Proto-Jê also deserves to be explored in greater detail. Inflectional markers must have included person prefixes (encoding the possessor in nouns, the complement in postpositions, the O/So argument in finite verbs, and the O/S argument in non-finite verbs) and non-finiteness suffixes; derivational morphemes definitely included, but were hardly limited to, so called formatives (also known as transitivity prefixes, cf. Oliveira 2005, Salanova 2011b, Nikulin & Salanova forthc.). These topics should be covered in future research.
Data provenance
Bardagil-Mas 2018, Dourado 2001, Lapierre ms., Vasconcelos 2013 Barbosa 1918
Estevam 2011, Hall et al. 1987, Lachnitt 1987
Krieger & Krieger 1994, Sousa Filho 2007, Souza 2008, Eneida Brupahi Xerente p. c.,
Mário André Coelho da Silva p. c. Wiesemann 2002
Alves Jr 2014, Bublitz 1994, Gakran 2015, Jolkesky & Gakran ms.
Ambrosetti 1896 (Pedro, Cosme Román, María Antonia), Lista 1883 (Lista), Vogt 1904 (Vogt1, Vogt3)
The PNJ reconstructions are based on data from numerous sources specified in Nikulin & Salanova (forthc.). For sources on Macro-Jê languages other than Jê, the reader is referred to Nikulin & Silva (forthc.).
References
Aikio, Ante (Luobbal Sámmol Sámmol Ante). 2012. On Finnic long vowels, Samoyed vowel sequences, and Proto-
Uralic *x. In: Tiina Hyytiainen, Tiina Jalava, Janne Saarikivi, Erika Sandman (eds.). Per Urales ad Orientem. Iter
polyphonicum multilingue. Festskrift tillagnad Juha Janhunen pa hans sextioarsdag den 12 februari 2012: 227-250.
Helsinki: Suomalais-Ugrilainen Seura. (Suomalais-Ugrilaisen Seuran Toimituksia, 264.)
Panará
Southern Kayapó
Xavánte
Xerénte
Kaingáng
Laklânô (= Xokléng) Ingain
Aikio, Ante (Luobbal Sámmol Sámmol Ánte). 2015. The Finnic 'secondary e-stems' and Proto-Uralic vocalism. Suomalais-Ugrilaisen Seuran Aikakauskirja 95: 25-66.
Alves Jr, Ozias. 2014. Parlons xokleng / laklänö. Langue indigène du sud du Brésil. Paris: L'Harmattan.
Ambrosetti, Juan Bautista. 1896. Materiales para el estudio de las lenguas del grupo Kaingangue (Alto Paraná). Boletín de la Academia Nacional de Ciencias de Córdoba 14: 331-380.
Barbosa, Alexandre de Souza. 1918. Cayapó e panará. Uberaba: manuscript. Available online at: <http://etnolinguistica.wdfiles.com/local--files/biblio%3Abarbosa-1918-cayapo/barbosa_1918_cayapo.pdf>
Bardagil-Mas, Bernat. 2018. Case and agreement in Panará (Naamval en congruentie in het Panará). PhD. Groningen: Rijksuniversiteit Groningen.
Benware, Wilbur A. 1974 [1995]. The Study of Indo-European Vocalism in the 19th century: From the beginnings to Whitney and Scherer. Amsterdam, Philadelphia: John Benjamins. (Studies in the History of the Language Sciences, 3.)
Bishop, Nancy. 1996. A preliminary description of Kensiw (Maniq) phonology. Mon-Khmer Studies 25: 227-253.
Bublitz, Terezinha. 1994. Análise fonológica preliminar da língua Xokléng. MA thesis. Brasilia: Universidade de Brasilia.
Carvalho, Fernando O. de. 2016. On the Development of the Proto-Northern Jê Rhotic in Panará Historical Phonology. Anthropological Linguistics 58(1): 52-78.
Carvalho, Fernando O. de, Andrey Nikulin. In preparation. Internal classification of Tupian.
Castro Alves, Flávia de. 2010. Evolution of Alignment in Timbira. International Journal of American Linguistics 76(4): 439-475.
Constenla Umaña, Adolfo. 2000. La restitución: un método lingüístico reconstructivo sincrónico. Filología y Lingüística 26(2): 161-180.
D'Angelis, Wilmar da Rocha. 2002. Género em Kaingang? In: Ludoviko dos Santos, Ismael Pontes (orgs.). Línguas Jê: estudos vários: 215-242. Londrina: Eduel.
Davis, Irvine. 1966. Comparative Jé phonology. Estudos Lingüísticos: Revista Brasileira de Lingüística Teórica e Aplicada 1(2): 20-24.
Dourado, Luciana Gonçalves. 2001. Aspectos morfossintáticos da língua Panará (Jê). PhD. Campinas: Universidade Estadual de Campinas.
Estevam, Adriana Machado. 2011. Morphosyntaxe du xavante. Langue jê du Mato Grosso (Brésil). PhD. Paris: Université Paris Diderot (Paris 7).
Gakran, Nanblá. 2015. Elementos fundamentais da gramática Laklänö. PhD. Brasilia: Universidade de Brasilia.
Hall, Joan, Ruth Alice McLeod, Valerie Mitchell. 1987 [2004]. Pequeno dicionário Xavante-Português, Português-Xavante. Damreme pibuzé hä a'uwë mreme, warazu mreme na te te ïsaprï mono zém na duré warazu mreme, a'uwë mreme na te te ïsaprï mono zém na. Cuiabá: Sociedade Internacional de Lingüistica.
Jolkesky, Marcelo Pinho de Valhery. 2010. Reconstruçâo fonológica e lexical do Proto-Jê Meridional. MA thesis. Campinas: Universidade Estadual de Campinas.
Jolkesky, Marcelo Pinho de Valhery, Nanblá Gakran. Ms. As classes verbais em Xokleng. Available online at: <https://www.academia.edu/5278237/As_classes_verbais_em_xokleng>.
Kassian, Alexei. 2015. Towards a formal genealogical classification of the Lezgian languages (North Caucasus): testing various phylogenetic methods on lexical data. PLoS ONE 10(2): e0116950.
Krieger, Wanda Braidotti, Guenther Carlos Krieger (orgs.). 1994. Dicionário escolar: Xerente-Português; Português-Xerente. Rio de Janeiro: Junta das Missöes Nacionais da Convençao Batista Brasileira.
Lachnitt, Georg. 1987. Romnhitsi'ubumro. A'uwë mreme = waradzu mreme. Dicionário xavante = português. Campo Grande: Missao Salesiana de Mato Grosso.
Ladefoged, Peter, Ian Maddieson. 1990. Vowels of the world's languages. Journal of Phonetics 18: 93-122.
Lapierre, Myriam. Ms. A phonological analysis of Panära. Berkeley: University of California.
Lista, Ramón. 1883. El Territorio de Las Misiones. Buenos Aires: Imprenta «La Universidad», de J. N. Klingelfuss.
Nikulin, Andrey. 2015. Apofonia e sistema vocálico do Proto-Jê Meridional: contribuiçao para estudos comparativos das linguas Jé. Revista Brasileira de Linguística Antropológica 7(2): 275-306.
Nikulin, Andrey. 2016. Historical phonology of Proto-Northern Jê. Journal of Language Relationship 14(3): 165-185.
Nikulin, Andrey. 2017. A phonological reconstruction of Proto-Cerrado (Jê family). Journal of Language Relationship 15(3): 147-180.
Nikulin, Andrey, Andrés Pablo Salanova. Forthc. Northern Jê verb morphology and the reconstruction of finiteness alternations. International Journal of American Linguistics 85(4).
Nikulin, Andrey, Mário André Coelho da Silva. Forthc. As línguas Maxakalí e Krenák dentro do tronco Macro-Jê. Cadernos de Etnolingüística (Estudos de Lingüística Sul-Americana) 7.
Nikulin, Andrey, Mário André Coelho da Silva. Forthc. A evoluçao do alinhamento na língua Maxakalí. In: Christiane Cunha de Oliveira (org.). Memorias do II Encontro dos Americanistas no Cerrado. Goiânia: Universidade Federal de Goiás.
Oliveira, Christiane Cunha de. 2005. The language of the Apinajé people of Central Brazil. PhD. Eugene: University of Oregon.
Pache, Matthias. 2018. Contributions to Chibchan Historical Linguistics. PhD. Leiden: Universiteit Leiden.
Ribeiro, Eduardo Rivail. 2004. Prefixos relacionais em Jê e Karajá: um estudo histórico-comparativo. Línguas Indígenas Americanas 4: 91-101.
Ribeiro, Eduardo Rivail. 2011. Prefixos relacionais como evidência histórico-comparativa: os casos Chiquitano e Jabutí. In: Ana Suelly Arruda Cámara Cabral, Aryon Dall'Igna Rodrigues, Jorge Domingues Lopes, Maria Risolêta Silva Juliäo (orgs.). Línguas e Culturas Macro-Jê, vol. 2: 105-120. Campinas: Curt Nimuendajú; Brasília: Laboratório de Línguas Indígenas.
Ribeiro, Eduardo Rivail. 2012. A grammar of Karajá. PhD. Chicago: University of Chicago.
Ribeiro, Eduardo Rivail, Hein van der Voort. 2010. Nimuendajú was right: The inclusion of the Jabutí language family in the Macro-Jê stock. International Journal of American Linguistics 76(4): 517-570.
Salanova, Andrés Pablo. 2011a. A flexäo de terceira pessoa nas línguas Jê. Línguas Indígenas Americanas 11: 75-114.
Salanova, Andrés Pablo. 2011b. Reduplication and number in Mêbengokre. In: Gale Goodwin Gómez, Hein van der Voort (eds.). Reduplication in Indigenous Languages of South America: 252-273. Leiden: Brill.
Sousa Filho, Sinval Martins. 2007. Aspectos morfossintáticos da língua Akwê-Xerente (Jê). PhD. Goiânia: Universidade Federal de Goiás.
Souza, Shelton Lima de. 2008. Descriçâo fonético-fonológica da língua akwen-xerente. MA thesis. Brasília: Universidade de Brasília.
Traunmüller, Hartmut. 1982. Der Vokalismus im Ostmittelbairischen. Zeitschrift für Dia-lektologie und Linguistik 49: 289-333.
Vasconcelos, Eduardo Alves. 2013. Investigando a hipótese Cayapó do Sul-Panará. PhD. Campinas: Universidade Estadual de Campinas.
Vogt, Federico. 1904. Die Indianer des Obern Paraná: II. Die Indianer der Coroados-Gruppe. Mitteilungen der Anthropologischen Gessellschaft in Wien 34(4): 353-377.
Wiesemann, Ursula. 1978. Os dialetos da língua Kaingáng e o Xokléng. Arquivos de Anatomia e Antropologia 3: 199-217.
Wiesemann, Ursula Gojtéj. 2002. Dicionário Kaingang-Português, Português-Kaingang. Curitiba: Esperança.
А. В. Никулин. Реконструкция фонологии и словарного состава пра-же языка
В статье рассматриваются регулярные фонетические соответствия между праязыками нагорной (Никулин 2017) и южной (Жолкески 2010) ветвями языковой группы же, самой глубокой генетической единицы в составе южноамериканской семьи макро-же. На основании этих соответствий предлагается реконструкция фонологии праязыка семьи же. Автор восстанавливает 11 согласных и 19 гласных фонем. Максимальная структура слога, восстанавливаемая для пра-же, — */СгУС/; при этом имеется ряд ограничений на сложные инициали (восстанавливаются только */рг, тг, кг, цг/). Автор восстанавливает правило, согласно которому носовые инициали имели в пра-же постора-лизованные аллофоны перед неносовыми централями. Статья завершается списком этимологий же.
Ключевые слова: языки же, языки макро-же, языковая реконструкция, сравнительно-исторический метод.
