Ndom

Timothy Usher, Santa Fe Institute

Ndom is spoken by approximately 300 people (1949) living in six small villages in the west portion of Kolopom (Frederick-Hendrick) island. Drabbe names these villages as Kariràm (Kalilam,) Fatàgh (Kambrada,) Murùba, Wétaghu, Pembrö and Sabèndar. They call themselves the Ndom-wer /ndom-wer/ (“Ndom people”,) in contrast to other Papuans who are called Yegiri-wer /jekiri-wer/ and foreigners who are called Kurik /ku'rik/, and their language Ndom-bana /ndom-baː'na/ (“Ndom language”) (Drabbe 1949: 1-2, 3.)

Drabbe (1949) brief description of and (pp. 14-24) 469 comparative terms for Ndom

Drabbe (1950) …

Drabbe (1954: 231-255) 100 comparative terms for Ndom

Voorhoeve (1975: 366) phonological inventory and grammar notes for NDom following Drabbe (1949)

Voorhoeve (1975: 96) 40 comparative terms for Ndom after Drabbe (1949)

Susanto (2001) 233 comparative terms for Kalilam (Ndom) of Wetau village

Drabbe (1949: 4) gives 14 consonants and 7 vowels for Ndom as follows:

Velar nasal [ŋ] is generally found only in clusters with a following velar consonant, where it is presumably an allophone of apical /n/. However, Susanto gives [ŋ] in one loan [paˈraŋa] from Indonesian /paraŋ/, suggesting that it may be needed for a modern synchronic description.

Bilabial voiceless stop /p/ is attested only in Susanto's vocabulary, where it is found a handful of recent loans, Kolopom /*p/ having been fricated to labiodental /f/ in all positions.

Apical voiclesss stop /t/ is often fronted to dental [t̪] in Susanto's vocabulary. Finally, it is typically aspirated to [tʰ t̪ʰ]. Medial /t/ is uncommon and most known examples of it are compounds.

Velar voiceless stop /k/ is found primarily but not exclusivesly when followed by high ftont vowel /i/ or apical non-stop /r/ in a cluster. Since velar not-stop /x/ has not been found followed by /i/, it is clear that these were originally allophones of a single phoneme /*k/, which outcomparison shows to reflect both Kolopom /*k/ and voiced prenasalized /*ŋg/. However, synchronically there are a small number of examples in which /k/ is followed by other vowels, necessitating a phonemic distinction in Ndom. In the only known final example, /k/ follows /i/ and is given as aspirated [kʰ] in Susanto.

Drabbe's voiced bilabial plain and prnasalized stops [b mb] appear to be conditioned allophones of the same phoneme, with the latter occuring when following a nasal in the preceding syllable. It's not obvious whether this phoneme should be represented as /b/ with conditioned prenasalization or as /mb/ with conditioned denasalization; the latter interpretation would parallel apical voiced /nd/ which is almost always prenasalized.

Voiced apical prenasalized stop /nd/ remains prenasalized in most positions, the only known exception being a cluster given as [md] which might well be underlyingly [mnd]. … According to Drabbe (1949: 4,) it is sometimes realized as voiceless prenasalized [nt].

A clear example of plain voiced /d/ has been found in only one word, Susanto's [dɛˈdim] “chewing betel”, which we presume to be a loan.

There is no velar voiced plain or prenaalized stop /g ŋg/ because the distinction between Kolopom /*k *ŋg/ has been neutralized with both yielding either voicless /k/ or fricative /x/ depending upon the following vowel.

Fricative /f/ is invariably given as labiodental [f] in Susanto's vocabulary, in contrast with bilabial [ɸ] as is often found in the Marianne Strait languages from the same survey. It reflects Kolopom bilabial voiceless stop /*p/, accounting for the lack of a /p/ in Ndom.

Laminal /s/ is realized as an affricate [tʃ] in Susanto's Ndom, just as it is in Kimaghama and Riantana as well as in Konorau of the Marianne Strait family .

Drabbe (1949: 3-4) states of Ndom but not of Kimaghama that the sound <gh> is pronounced like Dutch <g>, suggesting voiceless [x]. This would be in keeping with its dual origin in Kolopom voiceless and voiced prenasalized /*k *ŋg/, with /*ŋg/ presumably having been lenited to non-stop [*ɣ] before merging with the fricated[x] reflexes of /*k/. This sound has not been found to occur before high fton vowel /i/, where /k/ is found instead. In Susanto's Ndom, it is retained as glottal stop [ʔ] medially and dropped initially and finally, though prior to its loss raises following mid back rounded vowel /o/ to [u].

….
….
….
In addition to the above, it seems probable that a rounded bilabial/labiodental series /mʷ mbʷ fʷ/ is needed to explain the fact that initial clusters [mw bw mbw fw], while uncommon, exist while no other consonant can be paired with [w] in a cluster.

Thus, a conservative version of Ndom's consonant inventory would include only 12 consonants, plus three rounded bilabials/labiodentals for a total of 15 consonants:

However, it is evident that latter-day distinctions such as /n ŋ/, /p f/, /k x/ and /d nd/ are needed for a valid synchronic description, for as many as 19 consonants, with the most recent shown in brackets.

Drabbe's high central rounded vowel <ö> /ʏ/, corresponding to [ʉ y] in Susanto's transcription, is nearly always followed by word-final apical non-stop /r/, or, in a few instances, by nasals /m n/, suggesting it to be not a phoneme but a component of syllabic resonants /m n r/ [m̩ n̩ r̩]. Other final clusters exist, but unless the final is a resonant, no preceding vowel is given; conversely there are no final clusters with resonants as their second members, except those in which a high central rounded has been interpolated. Thus, we conclude that this vowel is epenthetic and not phonemic.

…

Accordingly, we propose a simpler system with only 5 vowels for Ndom

Generally, any consonant can occur initially, medially or finally except for bilabial and palatal non-stops /w j/, bilabial /w/ having not been found finally and palatal /j/ having not been found either medially or finally, and rounded bilabials/labiodentals /mʷ mbʷ fʷ/ which have been found only root-initially.

While Susanto's Ndom of Wetau village unquestionably represents the same language as Drabbe's, Susanto's transcription differs from Drabbe's in a few noteworthy respects. It is unclear to what degree these reflect differences in dialect, the passage of over fifty years' time or differences in elicitor interpretation.

Bilabial nasal /m/ occurs in all positions unchanged:

Final /m/ is uncommon:

Apical nasal /n/ occurs in all positions unchanged:

One apparent example of velar nasal [ŋ] in Susanot's vocabulary reflects a final cluster /nx/:

Bilabial voiceless stop /p/ is attested only in Susanto's vocabulary, where it is found a handful of recent loans, Kolopom /*p/ having been fricated to labiodental /f/ (below) in all positions.

Apical voiceless stop /t/ is often fronted to [t̪] in Susanto's elicitations:

It is conspicuous that most examples of medial /t/ are known to be compounds, possibly suggesting that the others are, too. Most likely, original medial /t/ has changed to some other sound, perhaps /r/ (below) as this is very common and there is at least one example, “five”, in which Ndom medial /r/ corresponds to /t/ in Riantana and Kimaghama.

Final /t/ is often aspirated [tʰ t̪ʰ] in Susanto's vocabulary:

Velar voiceless stop /k/ occurs primarily but not exlusively before high front vowel /i/. …:

…:

Final /k/ has been found in only one example, where it follows high front vowel /i/. Like final apical voiceless /t/ (above,) it is aspirated in Susanto's elicitations:

There are reasons to believe Drabbe's bilabial voiced plain and prenasalized stops /b mb/ to be allophones of the same phoneme. Initially, or medially or finally when not following a nasal /m n/ in the preceding syllable, only plain voiced [b] is found:

When following apical nasal /n/ in the preceding syllable, including a preposed first person singular possessor as seen in “friend”, and probably also bilabial /m/ though we have no examples of this, prenasalized [mb] is found instead:

This analysis is supported by comparison to other Kolopom languages, in which /*mb/ is denasalized or not according to the conditions laid out above. There are a few apparent exceptions in each vocabulary, so perhaps a synchronic distinciton between /b mb/ will prove necessary, but if so it is not historical and its functional load is very marginal.

Apical voiced prenasalized /nd/ is uncommon. Unlike bilabial /mb/ (above,) it generally retains prenasalization across contexts:

One apparent expection involves an intiial cluster which is probably underlyingly [mnd]:

An irreducible exception, Susanto's [dɛˈdim] “chewing betel”, is probably a loan (below.)

According to Drabbe (1949: 4,) /nd/ is sometimes realized as voiceless [nt]. Generally, this looks to be the case only root-finally:

One medial example is likely a compound with the third person plural pronoun (above):

Another medial example is also likely a compound:

One root meaning “grandparent” is given with root-intial [nt] in Drabbe, but this sound is not prenasalized when compunded with the preceding root /if/ “woman.” Cognates including Moraori /tat/ and Kimaghama, [tad-u] suggest Ndom's prenasalization to be secondary:

Relatedly, there is no phonemic velar voiced plain or prenasalized stop /g ŋg/ because the distinction between Kolopom /*k *ŋg/ has been neutralized, with both yielding either voicless /k/ or fricative /x/ depending upon the following vowel. However, there is one example in whch voiced prenasalized [ŋg] is given in both Drabbe's and Susanto's elicitations:

Here we assume Susanto's [ⁿda-] to reflect the first person plural possessor, Based upon analogy with “grandfather” (above,) we propose that inalienable possessors at least a have a final naasal element which surfaces in some instances; note also the example of /mbʷin/ friend“ (above, probably really “affine”) in which root-intiial /mbʷ/ somewhat unexpectedly surfaces as [mbʷ] rather than as [bʷ] when following a possessor. Drabbe's assertion that Ndom lacks both prefixes and suffixes then may be somewhat of an exaggeration, as the presence of prefixed inalienable possessors on kin terms and a restricted number of kin-like nominals, noting that language is inherited from one's speech community, is quite typical even of Trans New Guinea languages which have lost inalienable possession in other contexts.

Labiodental voiceless fricative /f/ occurs in all positions unchanged:

Interdental voiceless fricative /θ/ is sometimes given as voiced [ð] in Susanto's transcriptions:

… <t> … <th> /θ/:

Drabbe's laminal /s/ is invariably answered by Susanto's [tʃ], affrication of /s/ being characteristic of Riantana and Kimaghama as well as Konorau of the Marianne Strait family:

Drabbe's velar fricative <gh> /x/ has in Susanto's forms disappeared initially and finally, sometimes remaining medially as a glottal stop [ʔ]:

… /w/ …:

… One solution would be to treat these sequences as underlying vowel /u/ followed by another vowel. However, it is conspicuous that all these examples involve a preceding bilabial or labiodental consonant. This suggests rounding to be a feature of the consonants themselves, presumably /mʷ (m)bʷ fʷ/:

…:

…:

Initially, apical non-stop /r/ is not common, but definitely exists. Medially and finally, it is far and away the most common consonant in the Ndom language:

Palatal non-stop /j/ has been clearly found only root-initially:

… /j/ …”

… consonant clusters …:

Two probable clusters involving labiodental fricative /f/ followed by apicals … In the first of these, the modifying root is /if/ “woman”, but [f] is absent from both elicitations; the second is likely the root for “bow” folowed by an element of unknown meaning:

In monosyllables with initial high front vowel /i/, Susanto's elicitations have a preposed palatal non-stop [j]. This should probably be viewed as a property of non-phonemic vowel length and stress:

High front vowel /i/ …:

Drabbe's high central rounded vowel <ö> /ʏ/ is usually answered by [ʉ y] in Susanto's transcription. As is evident below, it is nearly always followed by word-final apical non-stop /r/, or, in a few instances, by nasals /m n/, suggesting it to be not a phoneme but a component of syllabic resonants /m n r/ [m̩ n̩ r̩]:

In one example, the medial segment is ambiguous between [bʏr] and consonant cluster [br], lending support to the notion that there is no underlying vowel:

In a few examples, Drabbe's /ʏ/is answered by Susanto's [u ʊ]. Since these are found in similar environments to the examples above, we can probably attribute the variation to Susanto's elicitation:

These patterns contrast with high rounded back vowel /u/ in segmentally identical environments:

Unlike central rounded [ʏ ʉ y], high back /u/ is not restricted to any particular environment. In destressed positions, it is sometimes given as schwa [ə] in Susanto's vocabulary:

In two examples, /u/ in presumably destressed positions is given as schwa [ə] in Susanto's vocabulary:

As elsewhere throughout his vocabularies of southwest New Guinea, Drabbe distinguishes high mid fron vowel <é> /e/ from low mid <e> /ɛ/. Meanwhile Susanto distinguishes low mid friont [ɛ] from lowfront [æ], though for a number of words both are given. … It is quite conceivable that a distinction between high mid and low mid or low front vowels existed for proto-Kolopom, but if so its difficult to see where this distinction would be drawn for Ndom:

When Drabbe's mid back rounded vowel /o/ is preceded by velar fricative /x/, it is reflected as high /u/ in Susanto's vocabulary:

There is only one other example of Drabbe's /o/ corresponding to Susanto's [u], and here either Susanto's vowel is in error, as low central /a/ is not backed and rounded as it usually owuld be when followed by /u/ in the next syllable (below, ) or Drabbe's <ò> was meant to read <ö> /ʏ/ (above):

Thus we conclude that there is no basis to propose a distinction between high mid and low mid back rounded vowels /o ɔ/ as is found in many other Trans-New Guinea languages and families.

… In the first two examples, the root is most likely /u/ “fire”:

Otherwise, Drabbe's /o/ is found as [o ɔ ɤ] in Susanto's vocabulary:

Drabbe's low central vowel /a/ is backed and rounded to [ɔ o] in Susanto's vocabulary when followed by bilabial non-stop /w/ or high rounded back /u/ in the next syllable :

However there is at least one apparent exception to this rule:

It is not clear if this occurs with Drabbe's central rounded vowel /ʏ/ (Susanto's [ʉ y], above):

In a few instances …:

When immediately following velar voiceless fricative /x/, low central /a/ is most usually reflected as schwa [ə] or dropped altogether in Susanto's vocabulary:

…:

Otherwise, /a/ occurs in at least medial and final positions unchanged:

Drabbe (1949: 5-6) gives pronouns for Ndom in three case forms as follows, with Susanto's (2001: 4) nominative forms presented for comparison:

According to Drabbe (1949,) Ndom, like Kimaghama and Riantana, is a nearly isolating language with no verbal morphology.

…Indonesian …:

…cannot be from Indonesian /aⁱr/ …:

…:

…: