Nawa River

Timothy Usher, Santa Fe Institute

The Nawa River family consists of two languages, Kaure and Kosare, spoken to either side of the upper Taritatu (Idenburg) river, mostly west of the Nawa and Sobger rivers, in the hill country east of the Lakes Plains in the Airu and Kaureh districts of the Jayapura regency in Indonesiaˈs Papua province (Voorhoeve 1975: 422-423, Wurm 1982: 197, Dommel and Dommel 1991: 2, Dommel 1992: 89-91, Wambaliau 2006: 2-4, q.v. Menanti and Rumaropen 2009: 3.)

The internal classification of Nawa River is as follows:

Nawa River

Kaure

Kosare

Voorhoeve (1975: 45, Smits and Voorhoeve 1994: 9-10) identifies a third language, Narau, following the vocabularies of the Anceaux collection in which smatterings of Narau forms, said to have been collected in the Lakes Plains region, are excerpted from Giël (1959.) Narau is the name of a river which is nowadays seasonally occupied by Kapauri and Kosare people from Pagai and Naira villages respectively (Wambaliau 2006: 2;) the Kaure call this group Mbalu /mbaɺu/ (Dommel, Dommel, Auri and Pokoko 1991: 84.) While there are not enough terms in Anceauxˈs excerpts to arrive at a definitive conclusion, there is little in the vocabulary to suggest that Giëlˈs Narau is anything but a variety of Kaure, with Dommel and Dommel (1991: 2) stating that the two are mutually intelligble.

Galis (1956) 114 comparative terms for Kaure

Giël (1959) Narau (unobtained)

Smits and Voorhoeve (1994: 18-266) 125 comparative terms for Kaure after D.A. Mallo, for two varieties of Kaure after Anceaux and for Narau after Giël

Voorhoeve (1975: 114-115) 39 comparative terms for Kaure after Bromley (n.d.) and 12 comparative terms for Narau following Anceaux after Giël

Barr and Walker (1978: 11-14) 110 comparative terms for Kaureh of Soar and Kasu (Lereh) villages

Heeschen (1978: 41-44) 97 comparative terms for Kosare

Dommel and Werner (1985) survey of Kaure (unobtained)

Dommel and Dommel (1991) phonology of Harna-Masta dialect of Kaure

Dommel, Dommel, Auri and Pokoko (1991) comprehensive vocabulary of Harna Kaure

Auri, Dommel and Pokoko (1991) Kaure conversations and (pp. 107-116) grammar sketch

Dommel (1992) Kaure kinship

Dommel (1997) Kaure kinship (unobtained)

Clouse (1997) 202 comparative terms for Kaure after Dommel and Wener (1985)

Wambaliau (2006: 24-30) 260 comparative terms for Kosare of Muara Nawa and Naira villages

Voorhoeve (1971: 51, 77-78,) working from Galisˈ (1956) vocabulary, lists Kaure as an unclassified language having no evident historical relationship to any of its known neighbors. Soon thereafter he gained access to Anceauxˈs collection of vocabularies and by (1975a: 45) placed the Kaure Family, consisting of Kaure and Narau (see comments on Narau above,) in a Kaure Stock alongside Kapori (Kapauri) and Sause, the Kosare language having not appeared in the literature until Heeschen (1978.) This was further asserted to constitute a Subphylum-level Stock within the much larger Trans-New Guinea Phylum (1975b: 422-423.) No arguments were presented in support of this proposal save for what is visible in the accompanying 40-term wordlists (1975a: 114-115.)

Wurm (1982: 197) expanded Voorhoeveˈs Kaure Family to include the recently-attested Kosare as coequal to Kaure and Narau, otherwise maintaining Voorhoeveˈs classification unchanged, and characterized the degree of relationship between the three as “on the lower medium to high family level” in contrast to the “low stock level” relationship uniting the Kaure Family, Kapori and Sause.

Hammarström (2010: 188, reiterated in Pawley and Hammarström (2017: 81) … “Voorhoeve (1975b) grouped Kapauri with the Kaure languages based on some lexical correspondences. However, a newer evaluation of the lexical relationships reveals figures that are only in the range of 5–6%, shedding considerable doubt on a geneticrelation between the Kaure languages and Kapauri (Rumaropen 2006:13).”

Foley (2017: …) …

We find that Kaure and Kosare form one group, designated here as Nawa River, and Kapauri and Sause another which is related most closely to Masep, Isirawa and Apauwar River (i.e. Kwerba) and has no special relationship to Nawa River. This fundamental division is supported by Dommel and Dommel (1999: 1-2,) who counted only 1% lexical similarity between Kaure and Sause of Badrun village, and by Wambaliau (2006: 12-13) and Menanti and Rumaropen (2009: 13-14,) who found only 5-6% lexical similarity between Kosare and Kapauri village, most of which can be attributed to loans (below.) Unfortunately, none of the SIL researchers are known to have compared Kaure with Kosare or Kapauri with Sause, but a fair accounting of the resemblances within each of these two pairs would yield a much higher figure, as demonstrated in the sections which follow.

[under construction]

Proto-Nawa River had 11 or 12 consonants and perhaps 8 vowels as follows:

In addition to the simple vowels given above, at least four diphthongs are found as follows:

Voiced stops /*b *g/ have not been found to occur intially. Unoccluded voiceless fricative /*h/ has not been found to occur medially. This assymetry suggests that one or more pre-Kaure initial voiceless stops became /*h/, while initial voiced stops were devoiced.

There is no apical voiced stop /*d/, suggesting non-stop /*ɽ/ to be its systemic equivalent; this is the case also in the neighboring Kapauri-Sause family. Initial /*ɽ/ is very uncommon, in keeping with the absence of initial voiced stops /*b *g/; presumably this pre-Kaure /*d/ would be reflected as initial voiceless /*t/.

Supporting these notions is the fact that the Orya language immediately to the north, along with Orya-Tor River more generally, has undergone the same changes from proto-West North New Guinea; for example bilabial voiced prenasalized /*mb/ is devoiced to /*p/ in Orya-Tor River and then deoccluded to /h/ in Orya, with apical /*nd/ likewise becoming Orya-Tor /*t/.

Only two archiphonemic consonants are distinguished finally, nasal /*N/ and stop /*C/:

Dommel and Dommel (1991: 7-15) describe a tonal distinction in Kaure which must be reconstructed for the Nawa River level as well, as it is plainly contrastive and not reducible to any segmental changes within the family. These are given by Dommel and Dommel as high tone and high-mid falling tone, but can be viewed systemically as high tone against low tone, the former being significantly more common. Unfortunately, this distinction is not indicated in Wambaliauˈs Kosare or in Dommel, Dommel, Auri and Pokoko (1991) or Auri, Dommel and Pokoko (1991,) only in Dommel and Dommelˈs (1991) phonology, which discusses only a selection of attested roots and even so is not entirely consistent in this respect.

Initial consonants and consonant clusters correspond as follows:

Medial consonants and consonant clusters correspond as follows:

Final consonants correspond as follows:

Vowels and diphthongs correspond as follows:

These correspondences are examplified as follows, with Kaure attestations drawn from Dommel, Dommel, Auri and Pokoko (1991, Dommel and Dommel 1991,) Kosare from Heeschen (1978: 41-44) and Kosare of Muara Nawa and Naira villages from Wambaliau (2006: 24-30.)

Kaure stress and tone are indicated only where they are known from Dommel and Dommelˈs (1991) phonology; where their tone assignations are inconsistent, both versions are given. A more vexsome problem is the indication of front mid vowels, as Dommel and Dommel ditinguish high mid /e/from low mid /ɛ/, but this is not indicated in their practical orthography. This distinction is not correlated with that between Nawa River /*e̝/ vs. /*ɛ/, so comparison to Kosare is of no use. However, Dommel and Dommel's high mid /e/ is uncommon, and is found only root-finally (pp. 41, 63-65.) Diphthong /ɛⁱ/ is realized as [e] in some contexts, where it is likewise written as <e> in the dictionary, leading to further questions. We have specified /ɛⁱ/ where it is given in the phonology, but otherwise can do little to address the overall problem other than to state that the effect upon the veracity of the reconstruction presented here is necessarily limited. Where the phonology specifically indicates either /ɛ/ or /e/, we present these as <ɛ e̝> respectively, leaving unmarked <e> to symbolize attestations found only in the dictionary, where its value cannot be determined.

Any consonant except for voiced stops /*b *g/ can occur initially, though initial apical non-stop /*ɺ/ is only very thinly attested.

Initial bilabial nasal /*m/, unless followed by a nasal consonant (below,) is reflected as voiced prenasalized stop /mb/ in Kaure. According to Dommel and Dommel (1991: 31,) in some words this is in free varation with [m]. One example in Heeschenˈs Kosare gives voiced stop [b]:

When followed by a nasal consonant in the next syllable or a nasalized vowel reflecting final nasal archiphoneme /*N/ (below,) initial /*m/ is retained as such in Kaure:

Initial apical nasal /*n/ … According to Dommel and Dommel (1991: 32-33,) in some words this is in free varation with [n]. One example in Heeschen and one in Wambaliauˈs Kosare of Muara Nawa give voiced stop [d]:

… /*n/ …:

In one example, initial /*n/ is dropped entirely in Kosare with consequent nasalization of the following vowel; this should probably be viewed as a further development of the reflex given as [ⁿ] (above):

… /*n/ … It may be seen in Dommel, Dommel, Auri and Pokokoˈs (1991: 140-143) vocabulary that kin terms are disproportionately represented among Kaure words beginning with [n]. This is likely because kin terms are obligatorily possessed, either by an inalienable prefix or by a personl name, and both first and third person inalienables begin with /n/ (Dommel 1992: 94) … :

Initial bilabial voiceless stop /*p/ is retained as such in both Kaure and Kosare:

Initial apical voiceless stop /*t/ is retained as such in both Kaure and Kosare:

Initial velar voiceless stop /*k/ is retained as such in both Kaure and Kosare:

Initial laminal voiceless fricative /*s/ is retained as such in both Kaure and Kosare:

Initial unoccluded voiceless fricative /*h/ is dropped altogether in Kosare:

In one example, initial /*h/ is unexpectedly retained as such in Kosare, perhaps indicating a doublet from Kaure:

Another may be a loan from Kaure or from some third source:

Another seeming exeption is a loan from Sause (below,) in which /h/ is the regular reflex of Kapauri-Sause /*s/:

Initial bilabial non-stop /*w/ is retained as such in both Kaure and Kosare:

Initial apical non-stop /*ɽ/ has been found in only one root, where it is realized as lateral flap /ɺ/ in Kaure just as are medial and final /ɽ/ (below):

The only known example of initial palatal non-stop /*j/ is a widespread northwest New Guinea regional term (below) and probably a loan into the Nawa River languages:

Any consonant except unoccluded voiceless fricative /*h/ can occur medially.

Medial bilabial nasal /*m/, unlike initial /*m/ (above,) is retained as such in Kaure as well as Kosare regardless of whether the following vowel is nasalized:

Medial apical nasal /*n/ is merged with the reflex of apical non-stop /*ɽ/ [ɺ] in Kaure, regardless of whether the following vowel is nasalized:

Where word-medial apical /*n/ is the first consonant of a root in a compound, it is retained as such in Kaure:

Medial bilabial voiceless stop /*p/ is retained as such in both Kaure and Kosare:

Medial apical voiceless stop /*t/ is retained as such in both Kaure and Kosare:

Medial velar voiceless stop /*k/ is retained as such in both Kaure and Kosare:

In at least two examples, Kaure medial /k/ corresponds to zero in Wambaliauˈs Kosare. The Kaure word given as /ɺóˈkó/ “hole” is found also as /ɺoʔ/ “ravine, gorge” (Dommel, Dommel, Auri and Pokoko 1991: 138) suggesting the intervening vowel below to have been interpolated. In support of this notion, Dommel and Dommel (1991: 19) state that /k/ never occurs syllable-finally. We interpret the correspondences in both of these instances as those of medial cluster /*kC/:

The second of these contrasts directly with /*po̝k[o̝]ɽɛ/ “salt” above, in which both consonants are preserved in Kosare, indicating the presence of an underlying vowel between them.

Medial laminal voiceless fricative /*s/ is retained as such in both Kaure and Kosare:

Unoccluded voiceless fricative /*h/ has not been found to occur root-medially, although it does synchronically in Kaure.

Medial bilabial voiced stop /*b/ is devoiced and merged with the reflex of voiceless /*p/ in Kaure:

Medial velar voiced stop /*g/ is uncommon. It is devoiced and merged with the reflex of voiceless /*k/ in Kaure:

One example in which the term for “mouth” (cf. Kaure /sɛ-ɺóˈkó/, literally “chin-hole”) has been suffixed with /*ki/ “skin” to yield the meaning “lip(s)” suggests /*ki/ to be underlyingly /*gi/, with voiced /g/ surfacing as such only in medial position:

Medial bilabial non-stop /*w/ is generally retained as such in both Kaure and Kosare:

Medial apical non-stop /*ɽ/ is realized as apical flap /ɺ/ in Kaure:

Medial palatal non-stop /*j/ …

…

Only two archiphonemic consonants are distinguished finally, nasal /*N/ and stop /*C/.

Final nasal archiphoneme /*N/ …:

Final stop archiphoneme /*C/ is realized as glottal stop /ʔ/ in Kaure. It is not clear that sporadic glottal stop in Kosare is meaningful as it is found also following final vowels and final /*N/ (above):

Among the vowels and diphthongs, only low central /*a/ occurs initially, and even so only in a few roots, two of which are kin terms. Thus, very few Nawa River roots are vowel-initial. Any vowel can occur medially or finally.

High front vowel /*i/ is retained as such in both Kaure and Kosare:

(Wambaliau's [y] “elbow” follows the segment [kʷ] in a compound.)

High back rounded vowel /*u/ is fronted to central rounded /ʉ/ in Kaure:

High mid front vowel /*e̝/ is raised and merged with the reflex of high front /*i/ in Wambaliauˈs Kosare but not in Heeschenˈs:

High mid back rounded vowel /*o̝/, variably realized as [o u ɤ ɯ] in Kaure (Dommel and Dommel 1991: 42-44,) is raised and merged with the reflex of high back /*u/ in Wambaliauˈs Kosare but not in Heeschenˈs:

When high mid-back /*o̝/ is followed by a front vowel or diphthong /*ɛ *ɛⁱ/ in the next syllable, it is reflected as [wi] in Kosare:

Low mid front vowel /*ɛ/ is retained as such in both Kaure and Kosare:

When low mid /*ɛ/ is followed by high back rounded /*u/ in the next syllable, Kosare merges it with the reflexes of /*u/ just as it would with diphthong /*ɛᵘ/ (below):

Low mid back rounded vowel /*ɔ/ is retained as such in both Kaure and Kosare:

Low front vowel /*æ/ is merged with the reflex of low central /*a/ in Kaure and with those of low mid front /*ɛ/ in Kosare:

Low central vowel /*a/ is generally retained as such in both Kaure and Kosare:

…

…

…

… /*aⁱ/ …

…

…

[under construction]

… with Kaure attestations drawn from Auri, Dommel and Pokoko (1991: 108) and Dommel, Dommel, Auri and Pokoko (1991: 82,) Kosare from Heeschen (1978: 42-43) and Kosare of Muara Nawa and Naira villages from Wambaliau (2006: 24):

In addition to …:

[under construction]

…

[under construction]

…

… Kapauri of Pagai village drawn from Menanti and Rumaropen (2009: 23-30) and Sause of Sebum village drawn from Wambaliau (2008) with Mamberamo River forms presented for comparison:

A smaller number of words can probably be identified as loans from Kapauri-Sause into Kaure and/or Kosare. In the case of “long,” this is mandatory on phonological grounds as Nawa River initial /*h/ is dropped in Kosare (above):

One Kapauri-Sause word which looks very similar to the Nawa River form and would likely have been included in Wambaliauˈs (2006: 12-13) and Menanti and Rumaropenˈs (2009: 13-14) lexical similarity figures is conceivably a legitimate retention from proto-Foja Range; cf. Grime River /*min/:

In a number of words thusfar found only in Kapauri and Kaure, the direction of borrowing cannot be determined upon distributional grounds:

The Kaure words above include several marginal sounds, namely medial /h/, which has no known Nawa River origin, medial /nd/ and laminal voiced affricate /dʒ/. Kaure /áhóˈɺîʔ/ “skin/bark” appears to have replaced Nawa River /*ki/ (above) which is now given as “rind, pod, peel” (Dommel, Dommel, Auri and Pokoko 1991: 135) or “shell” (Dommel and Dommel 1991: 10.)

There are also a number of words which have been found only in Kapauri and Kosare:

The Kosare term for “leaf” is likely a loan from Kapauri, where /su/ is given as “hair” (cf. /*haⁱ/ above;) Naira [suʔ] is given as “roof.” The term for “paddle” is found alongside the reflex of the native term /*tɽo̝wa/ (above.) The term for “kunai grass” is phonologically easier to explain as a loan from Kapauri into Kosare than vice-versa, as is that for “tired,” where there is no known native source for Kosare medial /h/. It seems then probable that most to all of the words listed above are loans from Kapauri into Kosare, which makes a good deal of sense in light of the recent history of Kosare resettlement as described by Wambaliau (2006: 4.)

… Sause and Kaure … The word for “black cockatoo” is found also in Anceaux's Orya of Sigi village as [bargwe] (Smits and Voorhoeve 1994: 249) and probably originates with Kaure since otherwise there is no origin for final nasalization:

Deapite Schiefenhövelˈs (1978: 49) assertion that the Kosare take Sirkai (Ketengban) wives, there is little lexical influence from the Mek family, which is allied first to Momuna and then to the Digul River-Ok languages to the south of the cordillera:

Another example would seem to be a loan from Kaure into Mek, as while the prefixation of [se- s-] (from /*se̝N/) upon words for mouth-related terms is typical of both Kaure and Kosare, there is no such prefix in Mek, where the initial variation would be irregular (Ketegban from Sims, Sims, Basini, Difur and Uropka 1990: 22, Yale Kosarek from Heeschen 1992: 136):

… regional … Kaure's two terms for “tobacco” are doublets, Anceaux's [sambuãŋ] (Smits and Voorhoeve 1994: 105-106) coming directly from Orya [sabwan]: