Bulaka River

Timothy Usher, Santa Fe Institute

Timothy Usher 2014. “Bulaka River Consonants” Journal of Language Relationship 12: 31-50.

The Bulaka River language family consists of two closely-related languages, Yelmek and Maklew, spoken to the east of Marianne Strait, across from Kolopom (Frederik Hendrik) island, in the Merauke district of Indonesia's Papua province. Drabbe (1950: 549) counted 350 speakers of Yelmek in four villages, Yelwayab on the Wanam River, from which his material was collected, Woboyu, Dudaling and Bibikem, and 120 speakers of Maklew in a single village on the Welbuti River, a tributary of the Bulaka (q.v. Boelaars 1950: 19, 25.) More recent figures from the Summer Institute of Linguistics give a Welbuti population of 226 (Lebold, Kriens and de Vries 2010: 8.)

Like other languages of the Marianne Strait region, both Yelmek and Maklew have been significantly influenced by the unrelated Marind language, the intrusion of which from the north separates them from the Morehead River languages near the Papuan border (Nevermann 1952 ibid.) Maklew in particular has coparticipated in a number of sound changes characteristic of Marind dialects. A smaller amount of more basic vocabulary is shared with the equally unrelated Kolopom family to the west (q.v. Geurtjens 1933: 198-433, Greenberg 1971: 830-833.) There is no trace of either Asmat or Awyu influence.

Geurtjens (1933: 398-433) 455 comparative terms for of Jab(sch)

Drabbe (1950) grammar and (pp. 566-574) 422 comparative terms for Jélmèk and Makléw

Boelaars (1950: 19-28) English language recension of Drabbe (1950) grammars

Voorhoeve (1975: 95) 40 comparative terms for Yelmek and Maklew following Drabbe (1950)

Nevermann (1952) language notes (pp. 81-82) up to 90 comparative terms for Jab(ga) of Bibikem and Imbáke villages, Galum (Nggarum) of Wobui and Dudalem villages, Dib(ga) of Dib village, Ilwajab village, and Makleu(ga) of Welbuti village

Lebold, Kriens and de Vries (2010: 46-52) 239 comparative terms for Maklew of Welbuti

The first to document Maklew, and hence to recognize its close relationship to Yelmek, was Drabbe (1950.) For the broader region, Drabbe (p. 548) draws a typological distinction between languages with suffixes (Paniai Lakes, Asmat-Kamoro, Central Digul and Mombum,) those with prefixes (Marind-Yaqay,) those with both (Yei, Kanum, Moroari and Bulaka River,) and those with neither (insular Kolopom.)

Voorhoeve (1968: 8-9,) measuring a 58% lexicostastical resemblance between Yelmek and Maklew, includes Bulaka River as a stock-level family among five first-order subgroups of his proposed South and Central New Guinea Phylum, alongside Kolopom, Yey-Kanum-Moraori, Morehead River, and the South and Central New Guinea Stock itself. This placement was based upon lexicostatistical figures of 9% between Yelmek and the Gawir dialect of Marind and 8% between Yelmek and Yaqay. There could hardly be worse choices for evaluating Bulaka River's relationship to the family which would be expanded to form Trans-New Guinea, as both Yelmek and Maklew are full of Marind loans, while Yaqay-Warkay is Marind's nearest relative.

Wurm (1971a: 574-577) reiterates Voorhoeve's classification and figures under the name Central and South New Guinea Phylum, to which Agöb (Pahoturi River) is added. In addition to Voorhoeve's evidence, he provides a list of typological features said to characterize the phylum (pp. 581-582,) but these show almost no intersection with those of Bulaka River.

Wurm (1971b: 166) is more specific, suggesting a special relationship between Bulaka River and Trans-Fly upon the basis of an average 9% lexical similarity and, citing Boelaars (1950,) shared morphemes and typological traits, though which ones he finds probative aren't specified. No such relationship is asserted in Wurm ed. (1975: 357,) in which Bulaka River is listed as a subphylum-level family, related to the Trans-Fly phylum level stock only at the Trans-New Guinea level, but the Trans-Fly-Bulaka River suphylum level superstock is reasserted in Wurm (1982: 179-180.) This would then make its way into derivative woks such as Ruhlen (1987: 358) and versions of the Ethnologue (Grimes 2000: 721,) becoming the default classification by a small number of high-profile repetitions.

Greenberg (1971: 829-830) includes Bulaka River in his Jei subfamily of South New Guinea. an unfortunate collection of a number of nearby but unrelated languages including most of those presented in Geurtjens (1933) and Nevermann (1939): Yei, Kanum and Arammba, Kolopom, South Marianne Strait and Bulaka River. South New Guinea as a whole, the outlines of which are broadly similar to Wurm's Trans Fly-Bulaka River, is supported by 70 comparisons, most of which don't involve Bulaka River.

Pawley, Ross and Osmond (2005: 24-26) propose a South-Central Papuan family including Yelmek-Maklew, Morehead-Upper Maro and Pahoturi, recapitulating the westernmost portions of Wurm's Trans-Fly-Bulaka River, based upon comparison of reconstructed personal pronouns.

Proto-Bulaka River had as many as 16 consonants and probably 6 vowels as follows:

The consonant system is formally symmetrical, if we allow that /*w *l *j *ɣ/ form a series of non-stops as is found in (for example) Guhu-Oro. Fricative /*s/ is found only in loans. Phonemic velar nasal /*ŋ/ is very common, yielding /ŋ/ in both Yelmek and Maklew. Palatal stops /*c *ɟ/ and voiced velar non-stop /*ɣ/ are relatively rare, as is, somewhat suprisingly, apical nasal /*n/. Palatal nasal /*ɲ/ is supported by only one example.

We tentatively assume Drabbe's qualities [e ɛ ə] and [o ɔ ə] to be allophones of two mid vowels /*e *o/, in the absence of clear and recurring patterns to the contrary, and the general sense that Drabbe reflexively draws these contrasts in most languages he described, giving the impression of a standardized phonetic rather than phonemic transcription. Geurtjens' transcriptions are less reliable, but show no pattern of agreeing with Drabbe's distinctions and often contradict them.

Consonants correspond as follows:

These correspondences are exemplified as follows, with attestations drawn from Drabbe (1950: 566-574, ibid.) Geurtjens (1933: 398-422) and Lebold, Kriens and de Vries (2010: 46-52.)

Bilabial nasal /*m/ is retained as such in all descendants, and is found in all positions:

Apical /*n/ is retained as such in all descendants. It has not been found to occur finally:

The scarcity of /*n/ relative to peropheral nasals /*m/ (above) and /*ŋ/ (below) suggests that a pre-Bulaka River /*n/ has become something else in most or all instances, a possibility to be borne in mind when comparing Bulaka River to other New Guinean families.

The existence of a palatal nasal /*ɲ/ is supported by only one known example, in which is is found initially. We tentatively reconstruct /*ɲ/ here by analogy to the dispositions of palatal stops /*c *ɟ/ (below,) and our skepticism of its originality applies equally here. Lebold, Kriens and de Vries give Maklew [ɲ] to Drabbe's [ŋ]:

Velar nasal /*ŋ/ is retained as such in all descendants, and is found in all positions:

Bilabial voiceless stop /*p/ is retained as such in all descendants, and is found in all positions:

Apical voiceless stop /*t/ is retained as such in all descendants. It is less common than peripheral stops /*p/ (above) and /*k/ (below.) Like apicals /*n/ (above) and /*d/ (below,) it has not been found to occur finally:

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

Bilabial voiced stop /*b/ is retained as such in all descendants, and is found in all positions. Initially, it is occasionally realized as prenasalized [mb]:

Apical voiced stop /*d/ is retained as such in all descendants. Like apicals /*n *t/ (above.) it has not been found to occur finally.:

Velar voiced stop /*g/ is retained as such in all descendants, and is found in all positions:

The qualities of the protosounds shown here are /*c *ɟ/ may not be possible to determine. We tentatively designate them as palatal stops, because they yield apicals /t d/ in Yelmek and velars /k g/ in Maklew, and do not appear to be conditioned reflexes of /*t *d/ or /*k *g/. Neither is common, but /*ɟ/ is less common than /*c/, a distributional pattern similar to /*k *g/:

… /*ɟ/ …:

It's striking that of nine examples, six are followed by /*e/, seven if one considers “smoke” to reflect /*ace aja/ as it likely does. This strongly suggests that there was originally a vocalic component to these sounds, indicating perhaps /*tiV *diV/ or /*kiV *giV/, neither of which, despite the frequency of sequential /*iV/, is otherwise yet known to occur. Nor are there any known examples of /*itV *ikV/, leaving many segments which /*c *ɟ/ could reasonably be thought to really be, but with no way of deciding between them.

Lateral /*l/ is preserved in all descendants. It does not occur root-initially, except on a few verbs which are prefixed in both Yelmek and Maklew:

Maklew drops medial /*l/ from some clusters:

While not directly attested in either Yelmek or Maklew, a voiced velar fricative /*ɣ/ is reconstructed to account for a correspondence in which Drabbe's Yelmek /ŋ/ is answered by Maklew /h/, rather than by /ŋ/, the aspiration and deocclusion of /*ɣ/ being shared with Marind's central dialects (q.v. Drabbe 1933: 14.) This is further supported by Guertjen's Jab [g], distinct from the reflexes of /*ŋ/ (above):

A direct indication of equivalance to the Marind sound is found in this loan:

As in Marind, glides /*w *j/ are distinguished from sequential vowels /*u *i/ by stridency, with the former but not the latter fricated to /h s/ in Maklew, merging with the reflexes of /*ɣ *s/. For whatever reason, /*w/ is much more common than /*j/:

… /*w/ …:

Maklew's fortitions further develop changes shared with the West and Atih dialects of Marind, in which /*j *w/ yield /hʲ hʷ/ (q.v. Drabbe 1933: 14.) The equivalence to and coevolution with Marind qualities is evident in the behavior of loans therefrom:

Sequential vowels /*i *u/ are not fricated:

While high vowels may join sequences with any other vowel, sequences of two non-high vowels /*ea *eo *ae *ao *oe *oa/ do not occur. Where a non-high vowel is adjacent to a high vowel, the high vowel is usually perceived as a contour unless it immediately follows a consonant. When two high vowels are in sequence, the first is perceived as a contour unless it immediately follows a consonant. It might be remarked that this distribution and behavior is not unlike what would be expected in a system with glides and no vowel sequences, and we would have reconstructed /*j *w/ had these not already been assigned to the correspondences which yield aspirates /s h/ in Maklew and codevelop with their Marind equivalents. A reconstruction of fricatives /*ʝ *β/ alongside vocalic /*j *w/ would match our data equally well, but with formally more complicated inventories.

Fricative /*s/ is found only in loans, primarily from Marind, in which it is found in all positions. It is defricated to /t/ in Yelmek:

This change is shared with Yaqay to the north, which likewise reflects Yaqay-Warkay /*s/ as /t/, and with Central Kolopom to the west, in which /*s/ is realized as [tʲ].

On a number of roots, Yelmek /k/ is answered by Maklew /ŋ/. In a few instances, one or the other of these is zero. Since this variation occurs only on items which are conceivably adjectives, and is otherwise irregular, we tentatively conclude it morphological in origin:

The nominative forms of the personal pronouns (q.v. Drabbe 1950: 550-551) are as follows. Maklew's second person nominatives differ from Yelmek's, which are the same as oblique bases of all descendants; it's conceivable that this difference was original. It's not clear whether Maklew has added a suffix to the third person plural or if the Yelmek form is an irregular reduction:

The oblique forms are derived from the nominatives by the suffixation of /*-el ≈ *-ol/ to the first and second persons and /*-i/ to the third persons. Here the originality of Yelmek's /-i/ is assumed in order to account for the difference in third person base vowels; our hypothesis is that Maklew generalized the third person obliques to the nominative and then reinforced them with /*-el/ by analogy:

The possessive forms are derived from the nominatives by the suffixation of /*-a[u]/ followed by the adjectival Yelmek /k/, Maklew /ŋ/. The Yelmek first person plural appears to be defective:

[under construction]

Drabbe (1950: 552-561) …

[under construction]

… Marind … (q.v. Geurtjens 1983: 398-411, Nevermann 1952 ibid.):

… “tobacco” …