====== Rumu ======

Timothy Usher, Santa Fe Institute

===== Situation =====
Rumu, also called Kairi or Kibiri (not the same as Kibiri-Porome,) is spoken by over 700 people (1992) living along the middle reaches of Kikori River and its tributaries, inland from Kerewo of the Kiwai family and from Kibiri-Porome, in Papua New Guinea's Gulf province (Petterson 1992: 1.)

===== Sources =====
Bevan (1890) Kairi (unobtained)

Chinnery (1917-1918) 130 comparative terms for Kibiri

MacDonald (1973: 132) phonological inventory of Kairi

Franklin ed. (1973: 573) 100 comparative terms for Kairi

Petterson (1989) dictionary of Rumu (unobtained)

Newman and Petterson (1990) Rumu tones (unobtained)

Petterson (1992) sketch phonology of Rumu

Petterson (1999) grammar notes and dictionary of Rumu

Petterson (2014) verbal inflections for Rumu

Murepe (n.d.) survey vocabulary of Rumu provided in digitalized form by Paul Whitehouse via the Summer Institute of Linguistics Ukarumpa

===== Phonology =====
Petterson (1992, q.v. 1999 ibid.) gives 10 consonants and 7 vowel qualities for Rumu:

|< - 60px 60px 60px 60px >|
|  m  |   n   |     |     |
|  p  |   t   |     |  k  |
|     |  [s]  |     |  h  |
|  w  |   ɾ   |  j  |     |


|< - 60px 60px 60px >|
|  i  |     |  u  |
|  e  |     |  o  |
|  ɛ  |     |  ɔ  |
|     |  a  |     |

Petterson's orthogaphy represents high mid vowels /e o/ and low mid vowels /ɛ ɔ/ as <ë ö> and <e o> respectively.

Apicals /n t ɾ/ are specified as dental.

Fricative /s/ occurs only in loans, and can be discounted for historical comparative purposes.\\
\\

Rumu lacks a contrast between voiceless (aspirated) and voiced (plain) stops because erstwhile aspirated stops /*p *t *k/ are reflected as /h ɾ ø/ by processes which also affect languages of the Omati River subgroup, while voiced stops /*b *d *g/ have been devoiced to [p t k] in a trend affecting languages of the northern Papuan Gulf and Bosavi regions generally.\\
\\

Neither consonant clusters nor final consonants occur, the latter having been lost or, in the case of /*b/ augmented with a final vowel, in another regional trend shared with (at least) Suki-Aramia River and the eastern members of the Fly River family as well as Mowase.\\
\\

Like Barikewa and Mowase, Rumu has lost constrastive nasalization on vowels.\\
\\

Four root-level tones, rising, falling, level, rising-fallng, are orthographically indicated on the final vowel as follows:

|< - 60px 60px 60px 60px >|
|  í   |  ì   |  ī   |  î   |
|  é   |  è   |  ē   |  ê   |
|  ɛ́  |  ɛ̀  |  ɛ̄  |  ɛ̂  |
|  á   |  à   |  ā   |  â   |
|  ɔ́  |  ɔ̀  |  ɔ̄  |  ɔ̂  |
|  ó   |  ò   |  ō   |  ô   |
|  ú   |  ù   |  ū   |  û   |

===== Pronouns =====
Petterson (1999: 160, ibid.) gives pronouns for Rumu in four case forms as follows:

|< - 100px 100px 100px 100px 100px >|
|     |absolutive|ergative|emphatic   |possessive|
|1 sg.|í         |ɛnɛ̂    |nɛì, nɛi-ɾì|nà        |
|2 sg.|ikí       |ɛkɛ̂    |kɛì, kɛi-ɾì|kà        |
|3 sg.|á ≈ áé    |á-ne    |ɛì, ɛi-ɾì  |à         |
|1 pl.|na-mé     |na-mè   |na-me-ɾà   |na-mē     |
|2 pl.|ka-mé     |ka-mè   |ka-me-ɾà   |ka-mē     |
|3 pl.|a-mé      |a-mè    |a-ma-ɾà    |a-mē      |
|1 dl.|na-tí     |na-tì   |na-t-à     |na-tī     |
|2 dl.|ka-tí     |ka-tì   |ka-t-à     |ka-tī     |
|3 dl.|a-tí      |a-tì    |a-t-à      |a-tī      |

The second form of the third person singular absolutive is from the Sirebe dialect (Petterson 1999: 17;) outcomparison shows it to be original.

The plurals and duals are derived from singular possessive bases /na ka a/ by the addition of /-me/ plural or /-ti/ dual. The dual might be assumed to to have been equivalent to /taí/ "two" at some stage, while dual emphatic /-tà/ might be thought a contraction of /-ti-ɾà/. The absolutive, ergative/possessive and possessive forms of the non-singulars differ from one another only in word-level tone, suggesting the former presence of since-elided segments in at least two of the three.

The alternations of the first and second person singular case forms are idiosyncratic. If the first person singular ergative were assumed to have once been /*iní/, probably via /*ĩ́/ and the subsequent loss of nasalization, this would make their development entirely parallel through all four cases.

Historically, the key points are that the non-singular forms are all secondary, the bases of the unmarked forms are first person /n/V, second person /kV/ (reflecting /*gV) and third person /a/, and the variations between the singular case forms cannot easily be further reduced by reference to internal evidence.\\
\\