East Keram River

Timothy Usher, Santa Fe Institute

[under construction]

The East Keram River family consists of two languages, Ambakich and Ap Ma, spoken along the Keram river, in the swamps and grasslands to the west and to the east along the Porapora (Bien) river, in the Angoram subdistrict of Papua New Guinea’s East Sepik province (q.v. ….)

The internal classification of East Keram River is as follows:

East Keram River

Ambakich

Ap Ma

Laycock (1971) 280 comparative terms for Aion (Ambakich) of Angang village and 246 comparative terms for Kambot (Ap Ma) or Kambaramba village (digitalized by Russell Barlow)

Laycock and Z'graggen (1975: 758-759) pronouns for Kambot (Ap Ma)

Ross (CITE) pronouns for Aion following Laycock (1971)

Wade (1981: 15-29) 163 comparative terms for Botin (Ap Ma) of Samban, Punyaten, Simbir, Raten, Yamen, Lol, Wom, Kekton, Krogopa, Bobten, Kambot, Kambaramba, Bobaten, Pamban and Kambuku villages

Wade (1983) phonology of Botin of Samban village

Wade (1983) dictionary of Botin of Samban village

Wade (1984) grammar of Botin of Samban village

Wade (1994) sketch phonology of Botin of Samban village

Abbot (1985: 329-336) 95 comparative terms for Kambot of Kambaramba village

Pryor and Farr (1989) deictics of Botin of Samban and Paiñiten villages (Central dialect)

Pryor (1990) Botin deictics

Adam (n.d.) vocabulary of Kambaramba (Ap Ma) provided in spreadsheet form by Paul Whitehouse via the Summer Institute of Linguistics

Sanders (n.d.) vocabulary of Aion provided in spreadsheet form by Paul Whitehouse via the Summer Institute of Linguistics

Potter, Lambrecht, Alemán and Janzen (2008: 60-65) 156 comparative terms for Ambakich of Yaut, Agurant, Ombos, Pangin and Arango villages

Barlow (2019) 757 comparative terms for Ambakich of Yaut village and Ap Ma of Yamen village

Barlow (2021) phonology of Ambakich of Yaut village

(We thank Pioneer Bible Translators of Madang and Martha Wade for access to her unpublished materials.)

[under construction]

Proto-East Keram River had 13 consonants and 6 vowels as folows:

…

…

Only a limited subset of consonants occurs finally:

…

Initial consonants in monosyllables correpsond as follows:

Initial consonants in roots with more than one syllable correspond as follows:

Medial consonants correspond as follows:

Final consonants correspond as follows:

Vowels correspond as follows:

…

These correspondences are exemplified as follows, with ….

Any consonant besides perhaps apical non-stop /*ɺ/ (below) can occur initially. Initial consonants in monosyllables are generally retained as such in both languages.

Initial bilabial nasal /*m/ in monosyllables is retained as such in both languages:

All three known examples of initial apical nasal /*n/ in monosyllables are followed by high front vowel /*i/, which induces palatalization to [ɲ] in both languages:

Bilabial voiceless stop /*p/ has not yet been found word-initially, though it seems probable that one existed.

Apical voiceless stop /*t/ has been firmly established in only one root, in whcih it /*t/ is palatalized and affricated to /tʃ/ [tʲ tʃ] in Ambakich:

Initial laminal voiceless /*s/ in monosyllables is retained as such in both languages:

Initial velar voiceless stop /*k/ in monosyllables is lost in Ap Ma, a change shared with Ulwa of the West Keram River family:

No clear example of initial bilabial prenasalized voiced stop /*mb/ in monosyllables has been found.

Initial apical prenasalized voiced stop /*nd/ in monosyllables is becomes non-stop /ɺ/ in Ap Ma before merging with palatal /j/ in the Kambaramba dialect:

Initial laminal prenasalized voiced /*ndz/ in monosyllables is reflected as /ndʒ/ in both languages:

Initial velar prenasalized voiced stop /*ŋg/ in monosyllables is retained as such in both languages:

Initial bilabial non-stop /*w/ in monosyllables is probably retained as such in both languages but the only known example, based upon comparison to West Keram, is not attested for Ap Ma:

No examples of initial apical non-stop /*ɺ/ in monosyllables have been found

Initial palatal non-stop /*j/ in monosyllables is occluded to affricate [dʒ] in Ambakich:

In roots with more than one syllable, not only the initial consonant but also the following vowel is dropped in Ap Ma, an areal process which affects also most Ramu River languages.

Initial bilabial nasal /*m/ in roots with more than one syllable is dropped in Ap Ma along with the following vowel:

No clear examples of apical nasal /*n/ in roots with more than one syllable have been found, though it's probable that it would follow the same pattern as does /*m/ (above.)

Bilabial voiceless stop /*p/ has not yet been found word-initially, though it seems probable that one existed.

No clear examples of apical voiceless stop /*t/ in roots with more than one syllable have been found, though we can probably assume that it is palatalized and affricated to /tʃ/ [tʲ tʃ s] in Ambakich and is lost along with the following vowel in Ap Ma.

No clear examples of laminal voiceless /*s/ in native roots with more than one syllable have been found, though it's probable that it would follow the same pattern as do other consonants. One example in which apparent initial /*s/ is not dropped in Ap Ma is a regional loan, and its reflex in Ambakich /tʃ/ suggests /*t/ (cf. medial /*t/ below):

Initial velar voiceless stop /*k/ in roots with more than one syllable is dropped in Ap Ma along with the following vowel. This /k/ is often heard in Ambakich as uvular [q] or glottal stop [ʔ] with which it appears to be in free variation (Barlow p.c. 2020, q.v. Potter, Lambrecht, Alemán and Janzen 2008: 60-65):

One example in which Ambakich also drops the vowel of the initial syllable may be due to the fact that it is a trisyllable, or it is a loan from Ap Ma, in which case the initial would be ambiguous between /*k/ and /*ŋg/ (below):

In one example, Barlow's Ap Ma though not Wade's or Laycock's preserves initial /*kV/ because it is the second member of a compound:

Other seeming exceptions are presumed to be loans; the first from … and the second from Waran (Banaro [koɲim] “spirits of nature” Z'graggen 1972):

Initial bilabial prenasalized voiced stop /*mb/ in roots with more than one syllable is generally dropped in Ap Ma along with the following vowel. Like all initial prenasalized voiced stops, it is often heard as a plain voiced stop [b] in Ambakich:

There are two puzzling exceptions in which the initial syllable is preserved in Ap Ma , perhaps because they are loans or due to some phonological feature which isn't yet understood:

Initial apical prenasalized voiced stop /*nd/ in roots with more than one syllable would presumably follow the same pattern as /*mb/ (above) but the only known example, based upon ccomparison to West Keram, is not attested for Ap Ma. Like all initial prenasalized voiced stops, it is often heard as a plain voiced stop [d] in Ambakich:

Initial laminal prenasalized voiced /*ndz/ has not yet been found in roots with more than one syllable.

Initial velar prenasalized voiced stop /*ŋg/ in roots with more than one syllable is dropped in Ap Ma along with the following vowel. Like all initial prenasalized voiced stops, it is often heard as a plain voiced stop [g] in Ambakich, but can also be heard as voiceless [k] though unlike initial /k/ (above) is never deoccluded:

It is sometimes impossible to distinguish initial /*ŋg/ from /*k/ (above) because the reconstruction is based upon a comparison between Ap Ma and West Keram where the distinction has been neutralized:

Initial bilabial non-stop /*w/ has not yet been found in roots with more than one syllable.

At least two examples suggest that apical non-stop /*ɺ/ may have occurred initially. Here the Kambaramba and Ap Ma reflexes are expected but Ambakich's ouright loss of /*ɺ/ is not; the correspondence is different from that of initial /*j/ (below):

Initial palatal non-stop /*j/ in roots with more than one syllable follows a different pattern from other consonants in this position as the first syllable is dropped in not only Ap Ma but also Ambakich; its former presence is discernable by comparison to West Keram River and by the fact that /*j/ when followed by nasals or prenasalized voiced stops in the onset of the following syllable (below) is not entirely dropped:

One example is irregular in both languages and is most likely a loan:

When initial /*j/ is followed by a nasal or a prenasalized stop in the onset of the following syllable, it becomes velar nasal [*ŋ], a sound otherwise not found in proto-East Keram River. If the following consonant is /*ŋg/ this [*ŋ] disappears in Ambakich, otherwise it is realized as /ŋg/ [k ŋg] typically in a cluster with the following consonant. Like other consonants in roots with more than one syllable, both intiial /*j/ and the following vowel are dropped entirely in Ap Ma. That these roots originally had /*j/ is evident by comparison with West Keram River, where /*j/ in this environment is usually realized as apical nasal [n], and in the case of “mosquito” by comparison to numerous languages across the region:

Any consonant can occur medially.

Medial bilabial nasal /*m/ is generally retained as such in both languages:

In one example medial /*m/ is occluded to /mb/ [b] in Ambakich, presumably because it has formed a cluster with following apical non-stop /*ɺ/:

In another medial /*m/ has disappeared in Ap Ma after forming a cluster with following velar prenasalized voiced stop /*ŋg/:

Medial apical nasal /*n/ is generally retained as such in both languages:

When preceded by high front vowel /*i/, medial /*n/ is palatalized to [ɲ] in at least Ap Ma:

Medial bilabial voiceless stop /*p/ is retained as such in both languages:

Medial apical voiceless stop /*t/ is palatalized and affricated to /tʃ/ [tʲ tʃ s] in Ambakich:

Medial laminal voiceless /*s/ is retained as such in both languages:

Medial velar voiceless stop /*k/ is generally retained as such in both languages:

One exception in which medial /*k/ is dropped in Ap Ma may be due to the simplification of derived cluster [*ɺk] or to being both preceded and followed by low central vowel /*a/; it's not possible to say without more examples:

Medial bilabial voiced prenasalized stop /*mb/ is retained as such in both languages, though is often denasalized to plain voiced [b] in Ambakich:

Medial apical voiced prenasalized stop /*nd/ is retained as such in both languages:

Medial laminal prenasalized /*ndz/ is known from only one example, in which it is reflected as apical plain voiced [d] in Ambakich and as [ndʒ] in Ap Ma:

Medial velar prenasalized stop /*ŋg/ is generally retained as such in both languages:

In one example in which medial /*ŋg/ is followed by the sequence /*ij/ it is palatalized to /ndʒ/ in Ap Ma:

Medial bilabial non-stop /*w/ …:

… medial /*w/ …:

… medial /*w/ …:

Medial apical non-stop /*ɺ/ …:

… /*ɺ/ …:

Palatalization does not occur if word-medial /*ɺ/ is the initial consonant of a suffix:

… /*ɺ/ …:

… /*ɺ/ …:

Medial palatal non-stop /*j/ …

This loss is not discernible if /*j/ was preceded or followed by high front vowel /*i/:

Any consonant except for prenasalized voiced stops /*mb *nd *ndz *ŋg/ can occur finally.

Final bilabial nasal /*m/ is retained as such in both languages, though is not directly discernible in Ap Ma when a nominal is followed by article /-ndɨma/:

Final apical nasal /*n/ is generally retained as such in both languages, though is sometimes realized as velar /ŋ/ in Ap Ma for unknown reasons and is not directly discernible in Ap Ma when a nominal is followed by article /-ndɨma/:

One example which has final /*n/ in West Keram and Ramu River has an unexplained final high central vowel /ɨ/ in Ambakich:

Final bilabial voiceless stop /*p/ is retained as such in both languages:

Final apical voiceless stop /*t/, like medial /*t/ (above,) is palatalized and affricated to /tʃ/ in Ambakich:

Final laminal voiceless /*s/ is found in only one example which is based upon comparison to West Keram and is not attested for Ambakich:

Final velar voiceless stop /*k/ is retained as such in both languages:

Final bilabial non-stop /*w/ is lost in Ambakich:

Final apical non-stop /*ɺ/ is retained in Ambakich, unlike medial /*ɺ/ which is palatalized and deoccluded to /j/ (above,) and lost in Ap Ma:

One exception in which /*ɺ/ is retained in Ambakich may be a loan from West Keram River, although the initial syllable is dropped as expected:

Final palatal non-stop /*j/ is lost in Ambakich:

… vowels and diphthongs …

High front vowel /*i/ …:

High central vowel /*ɨ/ …:

High back rounded vowel /*u/ …:

… /*u/ …:

…:

…:

…:

East Keram River free pronouns are reconstructed as follows, with attestations drawn from Laycock (1971,) Wade (1984) and Barlow (2019):

[under construction]

…

[under construction]

…