Families‎ > ‎

Enga-Southern Highlands

    Timothy Usher, Santa Fe Insitute


    The Enga-Southern Highlands family, called Engan in the work of Karl Franklin, is comprised of fourteen languages spoken throughout Papua New Guinea's Enga Province and in the central portion of Southern Highlands Province, with enclaves in East Sepik and Western Highlands Provinces. The region straddles New Guinea's central cordillera, with the northernmost portions draining into the Sepik River and southern portions into the Papuan Gulf by way of the Kikori and Purari Rivers, while in the northwest the Lagaip reaches the Gulf via the Strickland River.
Enga-Southern Highlands includes several of New Guinea's most well-documented and populous ethnicities, such as Enga (Draper and Draper 2002: 160,000,) Huli (Lomas 1988: 1 after Kloss and McConnel 1981: 65,000) and Kewa (Franklin and Franklin 1978: 50,000.)


    The internal classification of Enga-Southern Highlands is as follows:

    Enga-Southern Highlands
        Huli-Southern Highlands
            Southern Highlands
            East Trans-Enga

    Huli and Southern Highlands are considerably more divergent from one another than are any Trans-Enga languages.

    [Lexicostatistics – Franklin and Franklin (1965: PAGE after Wurm 1960, reprinted in Franklin and Franklin 1978: 379) Wurm (1971: 556),

    The lexical basis for the divisions between these subgoups may be illustrated by the following differences, as well as those in the pronouns further below:

 Huli  S. Highlands  Trans-Enga
 name  mi-ni  *[i]mbi  *ŋge
 fire/tree  iɾa  *ti  *ita
 moon  ege  *eke, *jumba  *kana
 four  ma-  *mala  *tumenda
 path  haɾiga  *pota  *kaita
 stand  ha  *ka  *kata
 cassowary  jaɾi  *jati  *laima
 skin  doŋgo-ne  *joŋgale  *jan[o/u]

    History of classification

    [under construction]

    Wurm (1960: 127) includes the Enga-Huli-Pole-Wiru Family as a subgroup of his East New Guinea Highlands Stock. This is equivalent to Enga-Southern Highlands, excepting that Nete, Bisorio and Lembena were not known at this time, and the mistaken inclusion of Witu:

    Enga-Huli-Pole-Wiru Family
        Enga-Ipili Sub-Family (Trans-Enga)
        Huli Sub-Family
        Mendi-Pole Sub-Family (East Southern Highlands)
        Wiru Sub-Family (Witu)

    The other family-level members of Wurm's proposed Stock were Gadsup-Auyana-Awa-Tairora (Kainantu,) Gende-Siane-Gahuku-Kamano-Fore (Goroka,) Hagen-Wahgi-Jimi-Chimbu (Simbu-Western Highlands) and Duna.
    Wurm (1965: 385-388, reiterated in 1971: 550-551) provides the following full tree, said to be based upon lexicostatistical comparison, for what was henceforth dubbed the West-Central Family, now including Lembena (Lemben,); nodes named in brackets are displayed but not labeled in Wurm's text:

    West-Central Family
        Wiru (Sub-Family)
            Mendi Sub-Family
            [non-Mendi Sub-Family]
                Huli Sub-Family
                [non-Huli Sub-Family]
                    Enga Sub-Family

    Franklin (1974: PAGE, reprinted in Franklin and Franklin 1978: 89-90) offers comments upon Wurm's divisions, proposing instead the following subclassification, leaving Lembena unplaced:

                Enga (including Kyaka, which Franklin considers a dialect)
                    Huli (including Huliduna, which Franklin asserts to be the same language)
                    Kewa (including Pole, which Franklin considers a dialect)
                    Mendi (including Augu, which Franklin considers a dialect)

    Excepting the inclusion of Witu and the placement of Sau with Huli, Franklin's proposal is entirely in accord with our own findings as outlined in the preceding section.

    Wurm (1982: 125-126) presents a heavily revised subclassification which is now very similar to Franklin's, but with Wiru reclassified as a family-level Isolate within Wurm's East New Guinea Highlands Stock:

    West-Central Family
        Enga Sub-Family
        Huli sub-family level Isolate
        Angal (Mendi)-Kewa

    Excepting the placement of West-Central itself in the East New Guinea Highlands Stock, for which we assert there is no valid basis, Wurm's revision is fully in accord with our own subclassification as outlined in the preceding section.

    Historical phonology

    [under construction]

    Proto-Enga-Southern Highlands had perhaps 11 consonants and 5 vowels as follows:

*m *n
*p *t *k
*mb *nd *ŋg
*w *l *j

*i *u
*e *o

    Unlike the families to the south, there is no /*s/; /s/ in descendants reflects palatalized [kʲ] reflexes of /*k *ŋg/ (Enga,) palatalized frications of /*t/ adjacent to /*i/ (Nete-Bisorio, East Trans-Enga, Angal-Kewa) or occlusions of /*j/ following /*i/ or followed by /*u/ (Southern Highlands.) While palatals [ɲ ndʒ ʎ] occur in Trans-Enga languages, these are in all instances reducible to apicals /*n *nd *l/ influenced by neighboring vowels.
    All consonants occur initially and medially, except /*nd/ which does not occur initially. Neither final consonants nor consonant clusters occur.
    Typically of New Guinean families in which the fundamental prosodic distinction between stops is the presence or absence of prenasalization, initial prenasalized /*mb *nd *ŋg/ had prevailing allophones [b d g] root-initially; synchronic examples of initial [NC] are invariably attributable to /*[C]VNC/ where the initial syllable has been elided. As is also typologically common, several descendants have subsequently devoiced these initial allophones, in some instances leading to mergers with the reflexes of plain stops.
    The persistence of medial prenasalization sharply distinguishes Enga-Southern Highlands languages from the families to the west and south, where, excepting only the heavily Southern Highlands-influenced Witu, the members of this series correspond to plain voiced or voiceless stops.
    Frication of the voiceless stops is frequent, but no family-wide generalizations can be made. Nete and Bisorio typically aspirate and fricate voiceless stops, probably under the influence of the Sepik Hills languages into whose ranges they've migrated. Excepting Nete-Bisorio, /*p/ is not fricated as it is in Kikori River-Lake Kutubu and Folopa to the south. In a development shared with Duna,, Foe, and Tua River, Huli, Angal Heneng and Sau fricate Southern Highlands initial /*k/ to /h/, with Angal Enen and West Kewa neutralizing this to zero.
    At least two tone contours are distinguished in monosyllables, and at least four in longer roots; these are probably attributable to the intersection of tone and stress. Vowel length is not contrastive.
    Enga-Southern Highlands did not distinguish nasalized vowels as do the families to the south and southwest, but Huli, Kewa and Sau have each developed phonemic nasalization, concomitant with the loss of nasal or prenasalized segments, under their influence.

    A very common trend among Enga-Southern Highlands languages is the imposition of height agreement between successive vowels /i u/ and /e o/ respectively.

    Sequential vowels occur, but in only a few patterns and are best viewed as diphthongs. Observed diphthongs include /*ai/, /*ae/ and /*au/.

   Enga-S.H.  Huli  E. S. Highlands  Trans-Enga


    [under construction]

    The basic (absolutive) pronouns of Enga-Southern Highlands' three subgroups are sharply distinct:

   Enga-S.H.  Huli  S. Highlands  Trans-Enga
 1 sg.    ĩ́    *na
 2 sg.    ĩ̀    *n[i/e]
 3 sg.    ibu    *[e]
 1 pl.    ina    *nani
 2 pl.    tʰì    
 3 pl.    tʰī    
 1 dl.    ija    *nali
 2 dl.    ɭibù    
 3 dl.    ɭībū    

    Verbal morphology

    [under construction]