Language phonologies

Askopan

The title of the translation of the Catholic catechism according to Glottolog is Deo puronemaita oirana totoirana. This is not evidence for /d/, since Deo is a clear Romance loan, but the existence of /p t m n r/ seems secure, meaning that Askopan, unlike Konua and Central Rotokas, contrasts nasals and voiced oral consonants.

The consonant characters in the Askopan wordlist, in descending order of frequency, are t n k p r m s d β ŋ b l y. Some of these can be eliminated:

• l appears once, in lapoto "rock". This likely represents /r/.
• y appears once, in torɛrɛ nyimpoːnoi "stand". It can be discarded.
• Except for abokan "old woman", b only occurs word-initially, which β never does. The same applies to d and r.
• s only occurs before i, an environment in which t never occurs. While a contrast between /s/ and /t/ is likely now phonemic in Askopan, as it is in Rotokas, due to loan introduction of the syllable ti and the consonant s before other vowels, Askopan can be assumed to have inherited assibilation.

This leaves a likely consonant inventory of /p t k β r m n ŋ/.

Keriaka

The consonant characters in the Keriaka wordlist, in declining order of frequency, are k n t r m p b s β g x y ʔ l ɸ w d. Some of these can be eliminated:

• d appears once, in ndaʔa, the 1SG pronoun.
• w appears twice, in inum-wano "nose" and kakaraw "frog". In both cases it probably represents a back vowel; cf. Rotokas kakarau "frog".
• ɸ appears nine times: in ɸuakxa "bad", ɸutaleoka "bird", ɸuɸurua "black", kotuɸi-ano "knee", totoɸai "spit", aʊɸoia "stone", and tutuɸuɸai "swim". This last may be cognate to Konua tutupu- "dive", with an affix -ɸai also appearing in uniupai "come", aitɛpapai "drink", and totoɸai "spit". This may be an allophone of /p/, or even, considering kokoβai "laugh", of a /β/ distinct from /b/.
• l appears in forms such as isipalo "mouth" (cf. the common [inalienable possessive?] suffix -ano in e.g. nobe-ano tongue), but also in loβoβesi "one", which likely represents an erroneous merger of two forms, rovo "one" and vesi "two".
• y is presumably either excrescent or a front vowel.
• x only appears in the suffix -kxa ~ -kʰa ~ -ka.
• s appears before ɛ e ɪ i, which presumably represent /e i/. As in Rotokas, the syllable ti does not occur, but te does, e.g. in atekaʔa "father" and ateate-ano "liver".

This leaves a likely consonant inventory of /p t k ʔ b r g β s m n/. There is no obvious way to unify b and β; these may be allophones of a single phoneme, but it's not clear from the wordlist. SIL notes the existence of a Keriaka title Bie Turipa Ovaka 1: Rovo Potaka, which suggests the existence of contrastive /b β/.

Konua

The headings in Müller's 1954 wordlist are a b e h i k l m n o p r s t u; however, the SIL OPD gives an inventory of /p t k ʔ β ɾ ɣ s h/, and notes that /β ɾ/ may be pronounced [b l~d] utterance-initially. Müller lists only three words starting with /n/: nǎbe ~ rǎbe "mouth; beck" (sic), nābusa "wave", and nabútu-ai "my leg; specially lower part of leg".

Rotokas

Central Rotokas has the six consonants /p t k v r g/. Aita Rotokas adds the three nasals /m n ŋ/, but lacks long vowels.

Correspondence sets

p - h - ? - p < *p

Finger may instead be assignable to the following set.

v ~ p ~ p/b ~ v < *b

'Star' might be a loan; cf. Proto-Papuan Tip *vituqun and Proto-Central Papuan *vitiɣu, and forms like Magori vitiriu and Roro bisiu.

v ~ b ~ b ~ v < *v

Note that Konua orthographic b represents [b ~ β], essentially the same sound as Rotokas v.

v ~ b ~ m ~ m < *m

? ~ b/v ~ p ~ ? < *b₂

? ~ p ~ b ~ p < *p₂

t ~ t ~ t ~ t < *t

Tuna may be an error for a form like tuua, or may be an unrelated form.

s ~ s ~ s ~ s < *s (= *ti)

Only before /i/, so assibilation is reconstructible to PNB.

TODO: Doesn't Capell have nejuna for Konua? And the syllable su appears. Maybe *tiu > *siu > su.

r ~ t ~ r ~ r < *d

TODO: Where did ita come from? My spreadsheet has roira.

r ~ r ~ r ~ r < *r

r ~ r ~ n ~ n < *n

* Also Rotokas roroo ovi "breast milk".

** TODO: Are Keriaka 'new' and 'yesterday' really homonyms? Likely an error, possibly mine.

k ~ k ~ k ~ k < *k

k ~ 0? ~ k ~ 0 < *q

g ~ g ~ ʔ ~ ŋ < *ŋ

g ~ g ~ g ~ ? < *g

The Rotokas form is the 1INCL.PL.

g ~ k ~ k ~ k < *g₂

ei ~ ? ~ i ~ oi < *əi

TODO: Konua glottal stop

Reconstructions

References

• Askopan wordlist
• Keriaka wordlist
• Konua wordlist
• Konua SIL OPD

Posts

• Is North Bougainville real? (2019-07-10)
• Allen and Hurd say it is, but don't give evidence (2019-07-18)
• The old Rotokas endonym Asigoao (2019-08-13)
• Is North Bougainville real? (2022-01-05)
• North Bougainville numerals (2022-01-06)
• Askopan consonants (2022-01-06)
• Affixes in Askopan (2022-01-06)
• North Bougainville possessive affixes (2022-01-06)
• Ramopa consonants (2022-01-06)
• North Bougainville correspondences (2022-01-17)
• North Bougainville correspondences: errata, pronouns, and loans (2022-01-17)
• Proto-North Bougainville reconstructions (2022-01-17)
• Once more on the old Rotokas endonym Asigoao (2022-01-17)