Wambon

Timothy Usher, Santa Fe Institute

Wambon, also known as Dumut, consists of anywhere from three to nine closely-related languages spoken mostly east of the Digul river and west of the Kao river in Indonesia's Papua province, somewhat to the west of the the Papua New Guinea border, with a small enclave just across the Fly River in Papua New Guinea's Western province.

There are several reasons why we use the term Wambon rather than Dumut to refer to this family. Dumut is the native word for the Mandobo river, which isn't geographically representative of the region as a whole. Drabbe (1959) called all of these Awyu languages in distinction to Kati (i.e. Muyu of the Lowland Ok family,) using Dumut to refer only to varieties of Mandobo. The designation Dumut was extended northward by Healey (1970) to include Drabbe's Wambon of Waniktit village (i.e. Agayop) in contrast to the Awyu languages west of the Digul, and then by Voorhoeve (2000) to include de Vries' (1986, 1989, 1992) Digul Wambon (North Wambon,) becoming less and less appropriate with each extension. In contrast, the term Wambon has been used to refer to at least some portions of all three languages, and according to Jang (2003: 24 after Farneubun 2002) “people in Mandobo language region insist that their language is part of the Wambon languages.” As the name of an ethnic group, the use of Wambon is parallel to the use of Awyu.

The southern Wambon languages have been significantly influenced by the languages of the Lowland Ok family immediately to the east, while North Wambon has borrowed a number of terms from the Central Ok languages to the north.

The internal classification of Wambon is as follows:

Wambon

North Wambon

Ketum

Kenon-Kenyam

Kenon

Kenyam

South and East Wambon

East Wambon

Agayop

Kokenop

Mandobo

Upper Mandobo

Central Mandobo

Mariam Mandobo

As no Kenyam or Mariam Mandobo data is currently available to us, their independence and placement are based upon the lexical similarity figures presented in Jang (2003: 19-27.)

De Vries and de Vries-Wiersma (1992: 79) assert that North Wambon and Yonggom Wambon (i.e. East Wambon) are viewed by their speakers as dialects of a single language, but the lexicostatistical figure presented at 47% does not support this, nor does Voorhoeve's (2000: 362) figure of 50% vs. 62% for Yonggom Wambon and Mandobo (below.)

Voorhoeve (2000: 362) provides lexicostatistical figures for Dumut as follows:

(Digul = Kenon, Yonggom = Agayop, Mandobo = Kaeti)

Healey (1970: 1009) :provides a substantially higher figure of 77% between Yonggom Wambon and Mandobo.

Drabbe (1954: 211- 214) grammar sketch of Oàmbon of Anggimòp village and Runggoanjàp and (pp. 214-229) 419 comparative terms and (pp. 231-255) 100 comparative terms for Oàmbon Dumùt (Upper Mandobo)

Drabbe (1959) grammar and (pp. 162-181) 430 comparative terms for Kaeti (Central Mandobo) and Wambon (Agayop) of Waniktit village

Healey (1970) proto Awyu-Dumut

Stokhof ed. (1982) Holle vocabularies with (pp. 120-121) 71 comparative terms for Upper Uwimerah [sic]/Upper Digul (North Wambon, possibly Kenyam,) (pp. 129-130) 44 comparative terms for Uwimerah (observation post) (Kokenop) and (pp. 141-150) 429 comparative terms for Digul Mandobo (Upper Mandobo) after van Doorn

de Vries (1986) Wambon relators

de Vries (1989) Wambon and Korowai (unobtained)

de Vries and de Vries-Wiersma (1992) grammar of Digul Wambon (Kenon) of Sawagit village

de Vries (1994) Awyu (i.e. Central Digul River) number systems

de Vries (2010) Dumut conjoined clauses

de Vries, Wester and van den Heuvel (2012) Greater Awyu (Digul River) comparative pronouns and verbal morphology

Voorhoeve (2000) revision of Healey's (1970) Awyu-Dumut, including (pp. 372-381) 108 reconstructed terms

Farneubun (2002) Central Digul survey (unobtained)

Farneubun (2002) Mandobo survey (unobtained)

Jang (2003: 39-47) 239 comparative terms for Ketum of Ater village, Kenon of Klofkamp and Kokenop of Bukit

Jang (2007) phonology of Kenon-Wambon of Manggelum village

Jang (2008) grammar of (Kenon) Wambon of Klopkam (Klofkamp) and Manggelum villages

Jang and Jang (2008) dictionary of (Kenon) Wambon

Jang and Jang (2009) orthography of (Kenon) Wambon

Hughes (2009: 27-37) 239 comparative terms for (Kenon) Wambon of Manggelum

Lebold, Kriens and Susanto (2013: 36-47) 239 comparative terms for (Central) Mandobo of Kaisah village and (pp. 48-50) 60 comparative terms for Subur and Aiwat

Wester (2014) Awyu-Dumut comparative morphology including (pp. 195-210) 430 comparative terms for (Kaeti) Mandobo, and Yonggom Wambon (Agayop) after Drabbe (1959: 162-181) and for Digul Wambon (Kenon) after Jang (p.c.)

[under construction]

[under construction]

The presentation which follows is inevitably deeply indebted to the reconstruction presented in Healey (1970.)

Proto-Wambon had 15 consonants and 9 vowels as follows:

…

Only a restricted subset of consonants occurs finally:

…

The most intractable uncertainty in the reconstruction of vowels is the distinction between low back /*ɒ/, which yields low central [a] in Mandobo, and low mid back rounded /*ɔ/ which is retained as such. That these were distinct is mandated by the contrast between /*ɔk/ “water” on the one hand and /*ɒm/ “breast”, /*ɒɾ/ “excrement” on the other. However there are many examples in which reflexes differ between Mandobo varieties, as well as some which lack attestations in Mandobo, thus are ambiguous between /*ɒ *ɔ/.

…

Healey (1970: 1002-1005) gives 12 consonants and 8 vowels for proto-Dumut in original transcription as follows:

Healey specifies <ü ö> as high front rounded and mid central rounded vowels respectively.

Voorhoeve (2000: 366-369) expands Healey's inventory to 13 consonants to account for de Vries' Digul Wambon, confirming the existence of Wambon /*p/ in non-final position and tentatively adding laminal voiced affricate /*dʒ/, but the reconstrution is otherwise basically identical save for a few added comparisons and the inclusion of North Wambon reflexes.

The most significant differences between these reconstructions and that presented here are, first, the recognition of rounded velar stops /*kʷ *ŋgʷ/, which are uncommon and sometimes discernible only through their effects upon neighboring vowels but correspond to rounded velars in other families, and, more sweepingly, the rearrangement of Healey's vowel qualities, which deserves further explanation: The first difference is that front mid-vowels /*e̝ *ɛ/ are distinguished, as they are in Awyu as well as North Digul River and Mountain Ok; this distinction is neutralized in most descendants, so would not have been obvious to either Healey or Voorhoeve, but is visible in the details of Mandobo dialect reflexes and supported by outcomparison to the aformentioned groups. The second is the reinterpretation of Healey's high central rounded vowel <ü> as high back rounded /*u/, which undergoes fronting across a very large portion of southwest New Guinea, excepting only the margins of the fronting region, namely Mandobo and South Awyu. Concomitantly, Healey's <u> are reinterpreted as high mid, low mid and low back rounded /*o̝ *ɔ *ɒ/ respectively.

To Healey's (1970: 1003) inventory of finals we add /*ɾ/, of which the North Wambon reflex is /t/ rather than /j/. Healey had no way of discerning this because finals /*ɾ *j/ have merged in East Wambon and Mandobo, upon which his reconstruction was based. It's not clear why Voorhoeve overlooked the North Wambon distinction; for example Voorhoeve's (2000: 377) /*wakot/ “moon” is inexplicably reconstructed with final /*t/ in contravention of Voorhoeve's own correspondences (p. 367.)

…. Healey's values (1970: 1002-1005, 1007-1008) …

Initial consonants correspond as follows:

Medial consonants correspond as follows:

Final consonants correspond as follows:

Vowels correspond as follows:

These correspondences are exemplified as follows. Kenon attestations are drawn from de Vries and de Vries-Wiersma (1992,) Voorhoeve (2000) quoting de Vries (p.c.,) from Jang (2008,) from Hughes (2009: 27-37,) or from a digitalized vocabulary provided by Dekker and Jang (2013,) in which instances it is indicated [hj], [jh] and [rw] respectively. Ketum Wambon is drawn from Jang (2003: 39-47.) Agayop forms are drawn from Drabbe's (1959: 162-181, ibid) Wambon. While Drabbe's vocabularies are intended to be phonemic, as is de Vries' and Jang's Kenon, materials from the surveys of Jang (2003,) Hughes (2009) and Lebold et al. (2013) are phonetic.

Any consonant can occur initially.

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

Initial apical nasal /*n/ is much less common than initial bilabial /*m/ (above.) It is retained as such in all descendants:

Initial bilabial voiceless stop /*p/ [ɸ] is fricated and deoccluded to /h/ in Kenon and dropped in Ketum and in East Wambon. In the dialects of Mandobo, it is either reflected as [w] or dropped under unknown conditions:

These developments are shared with Nakai to the north and Lowland Ok to the east, with Ninggirum reflecting initial /*p/ as [h] while Tangko, Iwur and Muyu drop it altogether. The Marind-Yaqay family to the south likewise deoccludes Fly River /*ɸ/ to /*h/. We can identify this change as part of a broader phenomenon affecting all the families of the Digul River corridor.

Initial apical voiceless stop /*t/ followed by vowels /*ɵ̝ *ɛ *ɔ *a *ɒ/ is retained as such in all descendants:

When followed by high and high mid front vowels /*i *e̝/ or high back tounded /*u/, apical voiceless stop /*t/ is merged with laminal /s/ [ts s] in North Wambon:

Initial laminal voiceless /*s/ is merged with apical stop /t/ in Mandobo, a development shared with Muyu of the Lowland Ok family to the east:

The following loan is likewise affected:

Initial velar voiceless stop /*k/ is retained as such in all descendants:

Initial rounded velar voiceless stop /*kʷ/ is merged with velar /*k/ in North Wambon and in East Wambon:

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

In one apparently irregular example, initial bilabial voiced prenasalized stop /*mb/ is reflected as nasal /m/ in North Wambon:

Initial apical voiced prenasalized stop /*nd/ is retained as such in all descendants:

Initial laminal voiced prenasalized /*ndz/, found clearly in only a single example, is like laminal voiceless /*s/ merged with apical voiceless stop /t/ in Mandobo:

In another example …:

Initial velar voiced prenasalized stop /*ŋg/ is retained as such in all descendants:

One example …:

Initial rounded velar prenasalized stop /*ŋgʷ/ is merged with velar /*ŋg/ in North Wambon and in East Wambon:

Initial bilabial non-stop /*w/ …:

When followed by high front vowel /*i/, initial /*w/ is dropped in Mandobo:

Initial apical non-stop /*ɾ/ is retained in all decendants. In Kenon, its realization in all positions varies between lateral and rhotic [l l̪ ɺ ɾ] (de Vries and de Vries-Wiersma 1992: 6):

Initial palatal non-stop /*j/ is generally retained as such in all descendants:

When followed by high mid front vowel /*e̝/, /*j/ is dropped in North Wambon and in Mandobo:

…:

In one example, initial /*j/ followed by low central vowel /*a/ is unexpectedly lost in Mandobo, with concomitant fronting and raising of /*a/ to high mid /*e̝/:

In another example …:

… medial consonants …

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

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

Medial bilabial voiceless stop /*p/ is lenited to /β/ [β w] in Wambon, merged with non-stop /w/ in East Wambon and with either non-stops /*w *j/ or zero /*ø/ in Mandobo depending upon the qualities of adjacent vowels:

An alternative analysis would hold that /*p/ has simply been dropped in Mandobo. This may be historically accurate, however synchronically there are otherwise no sequential vowels, so interpreting these contours as /*w *j/ avoids having to posit them.

In one example known to be a loan, /*p/ is preserved as /p/ in Agayop and as a phoneme /*h/ in Mandobo:

The term for “canoe” is a loan from Marind /*iawun/, as it is in Awyu:

Medial apical voiceless stop /*t/ when followed by vowels other than high vowels /*i *u/ is retained as such in all descendants:

When followed by high front and central vowels /*i *u/, voiceless apical stop /*t/ is fricated to laminal /s/ in North Wambon:

Medial laminal voiceless /*s/ is uncommon. It is merged with apical /*t/ in Mandobo:

Medial velar voiceless stop /*k/ is lenited to voiced [g] in East Wambon and Mandobo and fricated to [ɣ ʁ] in North Wambon:

North Wambon's frication of medial /*k/ is shared with the Central Ok languages, and is viewable as an intermediate stage preceding the loss of medial /*k/ in the Lowland Ok languages.

Medial rounded velar voiceless stop /*kʷ/ is merged with velar /*k/ in North Wambon and in East Wambon:

Medial bilabial voiced prenasalized stop /*mb/ is retained as such in all decendants:

Medial apical voiced prenasalized stop /*nd/ is retained as such in all decendants:

There are only two known examples of medial laminal voiced prenasalized /*ndz/:

Medial velar voied prenasalized stop /*ŋg/ is retained as such in all decendants:

Medial rounded velar voiced prenasalized stop /*ŋgʷ/ has been found in only one non-compound example:

Medial bilabial non-stop /*w/ …:

Medial apical non-stop /*ɾ/ is very common. It is realized as /l/ [l l̪ ɾ r] in Kenon:

Medial palatal non-stop /*j/ …

…

We assume segments [ia ija] to be underlyingly /*ija/ because this allows us to state that there are no root-internal vowel sequences:

Due to mergers of Digul River-Ok final consonants in proto-Digul River, only a restricted subset of consonants occurs finally: nasals /*m *n/, voiceless stops /*p *t *k/, possibly also rounded velar voiceless /*kʷ/, and apical and palatal non-stops /*ɾ *j/. Finals /*p *t/ are significantly more common than final /*k/ because Digul River-Ok final anterior voiced prenasalized stops /*mb *nd/ were merged with /*p *t/ while velar /*ŋg/ was merged with apical nasal /*n/.

Final bilabial nasal /*m/ is sporadically merged with /n/ in Agayop and in varieties of Mandobo under unknown conditions:

Accordingly, it's possible that some forms reconstructed with final apical nasal /*n/ (below) and based upon only Agayop and Mandobo really had final bilabial /*m/.

Voorhoeve (2000: 367) notes that “the appearance of n in the Kaeti forms is due to Drabbe's decision to give in his word list the primary stems' of the Kaeti verbs in which the /n/ seems to be an artifact of his analysis. The 'secondary stems' have /m/ instead of /n/. It should be noted that in Kaeti word-final /m/ and /n/ sometimes alternate (Drabbe 1959: 5.)”

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

In one example, …:

Final bilabial voiceless stop /*p/ is retained as such in all descendants:

One apparently irregular example cannot be solved by outcorrespondence, as besides Ndeiram River and Awyu which regularly drop final stops, no final stop is found in Becking and Dawi River or Sawuy cognates:

Final apical voiceless stop /*t/ …:

Final velar voiceless stop /*k/ is retained as such in all descendants:

It may be seen that finals /*p *t/ are singificantly more frequent than final /*k/, while in Ok they are roughly equally frequent. This is likely attributable to Digul River's merger of Digul River-Ok finals /*mb *nd/ with /*p *t/, but /*ŋg/ with /*n/ rather than with /*k/.

There is only one known example of a possible final rounded velar voiceless stop /kʷ/, which both Healey and Voorhoeve reconstruct as /*p/ in parallel with the augments found on other pronouns:

Final apical non-stop /*ɾ/ is merged with the reflex of /*t/ in North Wambon and merged with the reflex of /*j/ in East Wambon and in Mandobo:

Final palatal non-stop /*j/ is generally retained as such in all descendants:

In one example …:

Due to the merger of finals /*ɾ *j/ in East Wambon and Mandobo, the final qualities of several roots cannot be determined without more evidence:

Vowels ….

High front vowel /*i/ is retained as such in all descendants:

High back rounded vowel /*u/ is fronted and merged with high front /i/ in North Wambon and in East Wambon:

High mid front vowel /*e̝/ …:

High mid central rounded vowel /*ɵ̝/ is merged with the reflexes of high mid back /*o̝/ (below) in North Wambon and in East Wambon:

High mid back rounded vowel /*o̝/ is raised to /u/ in North Wambon and in East Wambon.

High mid back rounded /*o̝/ is fronted to /i/ in Kenon and in East Wambon when followed by high back /*u/ in the next syllable:

In one example …:

Low mid front vowel /*ɛ/ …:

Low mid back rounded vowel /*ɔ/ is generally retained as such in all descendants:

Low mid back /*ɔ/ is raised to /u/ in North Wambon and in East Wambon when followed by high mid rounded vowels /*ɵ̝ *o̝/ in the next syllable:

Several roots of the form /*CɔCaC/ are reflected as /CaCaC/ in Kokenop:

When followed by /*j/, low mid back rounded /*ɔ/ is raised and fronted to /i/ in Wambon and gains an offglide to become /oᵘ/ in Mandobo; we presume the difference in Kokenop reflexes to result from …

Low back rounded vowel /*ɒ/ is uncommon. It is merged with the reflexes of low mid /*ɔ/ in North Wambon and in East Wambon and with those of low central /*a/ in Mandobo:

Low central vowel /*a/ is generally preserved as such in all descendants, excepting some Mandobo dialects which …, a development which postdates proto-Mandobo:

Low central vowel /*a/ is fronted and raised to /ɛ/ in Mandobo when preceded by /*j/: