South Pauwasi River

Timothy Usher, Santa Fe Institute

[under construction]

The South Pauwasi River family consists of four languages spoken … Its nearest linguistic relative within the greater Pauwasi River family is Yetfa (Biksi,) which was formelry wrongly classificed as a member of the invalid Sepik-Ramu Phylum (CITE.)

The internal classification of South Pauwasi River is as follows:

South Pauwasi River

Kimki

Lepki-Murkim

Kembra

Lepki

Murkim

[under construction]

Langeler (1915) 15 terms for …

Reimer and Songkilawong (1980) … comparative terms for Kimgi (Kimki)

Rumaropen (2004) … comparative terms for Kimki of Batom and Sabi villages

Wambaliau (2004: 22-28) 258 comparative terms for Murkim of Mot and Milki villages

Anderson (2007: 72) 64 terms for Lepki

Donohue (n.d.) … comparative terms for Lepki

Konel (n.d.) … Kembra

The South Pauwasi River languages were recognized only in recent times and only the scantest documentation of them has been published.

Silzer and Heikkinen (1984: 55,) in possession of Reimer and Songkilawang's (1980) Kimki data, write of the language they call Sukubatong (Aipki, Kimgi) “May be the same as Biksi (301) or Pyu (315) language.” Yetfa (Biksi) and Pyu are basically unrelated to one another, but It is true that Kimki is closely related to Yetfa, and there are a number of loans shared with Pyu (below.)

Wambaliau (2004: 13) in her survey of Murkim writes, “Tingkat kesamaan leksikal di antara bahasa Murkim dengan bahasa Kimki dan Yetfa adalah 8% samapai 14%. Biasanya tingkat kesamaan leksikal seperti ini menunjukkan bahasa bahwa bahasa-bahasa ini berhubungan, tetapi, karena penutur bahasa Murkim kadang-kadang mencampurkan percakapan mereka dengan bahasa Kimki dan Yefta, tingkat kesamaan leksikal ini oleh karena pinjaman kata-kata.” In other words, the lexical similarity figures between Murkim and Yetfa and Kimki respectively are high enough to indicate a relationship, but they are attributed here to borrowing due to multilinguialism.

Hammarström (2010: 188-189, 189, 190) rebuts an unattributed (Grimes ed., though tracing back to at least Silzer and Heikkinen) placement of Kimki with Yetfa in particular, which (p. 193) he considers an isolate, while treating also Lepki and Murkim as isolates, writing of Lepki, “Wherever it appears, the language is listed as “unclassified” (Lewis 2009; Silzer and Heikkinen-Clouse 1991). The label “isolate” is more appropriate, as the wordlist shows no significant relationship to any of the neighboring languages” and of Murkim, “It has some lexical matches with neighboring languages, but they look more like loans than the outcome of genetic inheritance.”

Wichmann's (2012: 329-330) Automated Similarity Judgment Program (ASJP) correctly placed Lepki with Murkim based upon a small number of mostly valid comparisons. Foley (2017: 463-469) likewise places Lepki and Murkim and possibly Kembra together in a Lepki family while maintainng Kimki and Yetfa as isolates, though his actual text about Kimki is palpably equivocal about a possible relationship with Lepki-Murkim. (p. 468): “The Kimki plural pronouns in Table 37 share the formative –me with those ofMurkim of the Lepki family (section 4.11), but in Kimki this is spread throughoutthe pronoun paradigm except for first person singular. The stems of the first and second person plural pronouns are also the same as in Murkim. There are also a few similar shared lexical forms, compare ‘moon’, ‘ear’, ‘give’, and ‘hair’ with Murkim. All of this may indicate a genetic relationship with Lepki family languages, but the bulk of the evidence suggests otherwise; there are too many usual cognates not shared, like ‘tooth’ (‘tooth’ may be shared with Yetfa) and ‘water’. If Kimki and Murkim are not genetically related, the pronoun resemblances are either due to chance or to borrowing, and the high degree of similarity suggests the latter to me.”

The classification presented here was developed in the early to mid-2000s as Paul Whitehouse (Usher and Whitehouse 2007) assembled unpublished vocabularies of these languages obtained from the Summer Institute of Linguistics. Whitehouse (2006: 268) refers to an early version of this classfication, writing, “Tim Usher (pers. comm.) has suggested that the Pauwasi family of central New Guinea should be expanded to include a number of other languages hitherto considered isolates of various degree, namely Molof, Tofanma, Usku, Biksi and Yuri…To this group may also be added a couple of recently discovered neighbors, Kimgi and Lepki.”

[under construction]

Proto-South Pauwasi River had at least 11 consonants and … vowels as follows:

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

…

…

…

Initial consonants correspond as follows:

Medial consonants corresopnd as follows:

Final consonants correspond as follows:

Vowels and diphthongs correspond as follows:

These correspondences are exemplified as follows, wtih Kimki attestations drawn from Rumaropen (2004: pages,) Kembra from Konel (n.d.,) Lepki from Donohue (n.d.) and Murkim from Wambaliau (2004: 22-28.) Due to the brief and prephonemic nature of the underlying data, some details of the reconstructions presented below, especially vowel qualities, are necessarily tentative.

Initial consonants ….

Initial bilabial nasal /*m/ is retained as such in all descendants:

Initial apical nasal /*n/ …:

Initial bilabial voiceless stop /*p/ …

Initial apical voiceless stop /*t/ is retained as such in all descendants:

Initial laminal voiceless /*s/ …:

Initial velar voiceless stop /*k/ is generally lost in Kimki:

Apparent exceptions in which root-initial /*k/ is retained in Kimki are attributable to the roots appearing only as the second memebr of compounds:

Initial bilabial voiced stop /*b/ is retained as such in all descendants:

Initial apical voiced stop /*d/ …:

Initial bilabial non-stop /*w/ …:

Initial apical non-stop /*ɺ/ …:

Initial palatal non-stop /*j/ when followed by a non-nasal consonant or only a vowel is lost in Kimki, often after the fronting and raising of the following vowel:

One apparent exception in which initial /*j/ is retained in Kimki may be a loan from Lepki:

When followed by a nasal /*m *n/, initial /*j/ …:

Medial consonants …

Medial bilabial nasal /*m/ …:

Medial apical nasal /*n/ is retained as such in all descendants:

Medial bilabial voiceless stop /*p/ …:

Medial apical voiceless stop /*t/ is retained as such in all desscendats:

Medial laminal voiceless /*s/ …:

Medial velar voiceless stop /*k/ is generally retained as such in all descendants, unlike initial /*k/ which is lost in Kimki (above):

Medial consonant cluster /*km/ is simplified to /m/ in Kimki:

… /*kn/ …:

Medial bilabial voiced stop /*b/ …:

Medial apical voiced stop /*d/ …:

Medial bilabial non-stop /*w/ …:

Medial apical non-stop /*ɺ/ …:

When/*ɺ/ is a member of a conconsant cluster //*ɺC/ or /*Cɺ/, it is realized as …:

Medial palatal non-stop /*j/ …:

… /*j/ …:

Final consonants …

Final bilabial nasal /*m/ is generally retained as such in all four lagnuages:

Final apical nasal /*n/ …:

Final bilabial voiceless stop /*p/ …:

Final apical voiceless stop /*t/ …:

Final laminal voiceless /*s/ …:

Final velar voiceless stop /*k/ is lost in Lepki:

Final bilabial voiced stop /*b/ …:

Final apical non-stop /*ɺ/ …:

… /*ɺ/ …:

Vowels and diphthongs …

High front vowel /*i/ …:

…

…

…

…

…

…

Free pronouns for South Pawuasi languages are given as follows, with Kimki forms drawn from Rumaropen (2004,) Kembra from Konel (n.d.,) Lepki from Donohue (n.d.) and Murkim from Wambaliau (2004:22.) It is doubtful that any of the Kembra or Lepki forms given here are accurate:

Foley (2018: 468) obseves “The Kimki plural pronouns in Table 37 share the formative –me with those of Murkim of the Lepki family (section 4.11), but in Kimki this is spread throughout the pronoun paradigm except for first person singular. The stems of the first and second person plural pronouns are also the same as in Murkim.” The suggestion would seem to be that Murkim has borrowed its first and second person plural pronouns from Kimki, in keeping with the concept that all of the lexical resemblances between Kimki and Murkim are the result of borrowing. This is a priori possible, but would require am extreme level of language mixture which has not otherwise been found in New Guinea.

Evaluating the situation with the pronouns is hobbled by the fact that the reliability of both Konel's Kembra and Donohue's Lepki pronouns are highly questionable. For example, Donohue's second person singular [jojo] would appear to mean “that”, while [jiris] is given for both the third person singular and the first person plural inclusive, while Konel's forms seem to bear no relationship to anything else at all. If taken at face value, both Kembra and Lepki would likewise need to be isolates, with their numerous lexical resemblances to Murkim attributed to loans.

Besides the first and second person plurals, Kimki's second person singular has obvious cognates elsewhere in the Pauwasi River family, most immediately to Yetfa [po(-wo) ~ pu-wo] (Kim 2006) as well as to Afra and Tebe-Towei forms; Yetfa's first and second person plurals would also appear to be cognate (q.v. Foley 2018: 165.) The picture then is while complicated not as dismal as it first appears, and will likely be improved by more extensive and reliable data, especially for Kembra and Lepki.

No information about the verbal morphology of any South Pauwasi River languages is currently available to us.

….

The most problematic loans are very often between those of genuinely related languages because their mutual retention is a priori plausible and because there are sometimes no violated sound change rules which flag the comparisons as loans. Wambaliau (2005: 13) suggests that the high level of lexical resemblance between Murkim and Kimki and Yetfa respectively is attributable to loans due to multilingualism, but the sheer volume of comparisons and the number of non-trivial sound changes which affect especially Kimki as well as the fact that many are shared also with other Pauwasi languages including Lepki militate against this interepretation. Still, the following word is so lengthy as well as uncannily identical that skepticism is warranted:

It might also be asked if there are loans in South Pauwasi languages to or from the closely related Yetfa. …

…

Kimki shares at least a few words from the basically unrelated Pyu to the east, which helps to explain Silzer and Heikkinen's (1984: 55,) suggestion that Kimki might belong with or even be the same language as Pyu, with Pyu attestations drawn from Conrad and Dye (1975: 25-29.) “Fire/wood” has a convincing cognate in Arai and Samaia Rivers /*ka/ “tree” and is certainly not Pauwasi (cf. /*jaᵘ/ above) so is most likely a loan from Pyu, especially as Kimki drops initial /*k/ in native words (above.) “House” and “sago palm” on the other hand both have cognates in Murkim (above) and no known analogues in Arai and Samaia Rivers, though this wouldn't rule out the possibioity of mutual loans from Pyu:

There is also one very salient comparison between Pyu and Lepki-Murkim, although again here the existence of a local cognate would seem to suggest Pyu to be the borrower:

… regional …:

… Tok Pisin … both are also found in Yetfa as [sot] and [dais] respectively…: