First posted: 19 May 2014
Last revised: 20 Aug 2018
1. Introduction
The Southwest Maluku Regency (Kabupaten Maluku Barat Daya, traditionally also known as "Tenggara Jauh") is located between Timor Island to the west and the Tanimbar Islands to the east, and comprises a number of small islands and island groups, with a population of around 70.000. The largest island is Babar with an area of approximately 800 km², and which is the eastern-most island in Southwest Maluku.
On Babar island and the smaller neighboring Masela island, we can find a subgroup of the Austronesian language family which is highly innovative with respect to its phonological history. This group was called South Babar by Mark Taber in his survey of the languages of Southwestern Maluku (Taber 1993); this term is also used in the Ethnologue. Next to a paper by Hein Steinhauer (2009) that deals with the phonological history of the Southeast Babar language, Taber's survey is until now the only data source for the South Babar languages.
Seven South Babar lects are documented by Taber, which can grouped into three branches:
a. Central Masela
b. Masela-Southeast Babar
East Masela
Serili
SE Babar
c. Southwest Babar
Emplawas
Tela-Masbuar
Imroing
Taber included Central Masela in his "Masela-SE Babar Cluster". Because of several differences in phonological history, Central Masela is better put into a branch of its own.
The sound changes in this subgroup are quite far-reaching, leading to reflexes of Proto Austronesian etyma that are hardly recognizable. Examples from Emplawas, the most innovative South Babar language, shall illustrate this:
Although it is not obvious at all at the first sight, each of the Emplawas reflexes here is the result of regular and predicatable sound change.
2. Note on Taber's transcriptions
Although the data collected by Taber is written down phonetically without any further attempt for a narrower analysis, we can still make a few careful steps towards a phonemic interpretation, at least for certain languages in the list. Again, I will illustrate my approach to the data by taking the example of Emplawas.
2.1. Final glottals
Final glottal stops are exclusively found with the noun marking suffix -ɛˀ. This ending also appears as -e without a glottal stop. The distribution of these two variants is quite random. The variant -ɛˀ usually occurs with noun roots ending in -i, often with an intervening glide j:
osi-ɛˀ, uvij-ɛˀ, usi-ɛˀ, wusij-ɛˀ, lalkij-ɛˀ.
One of the few counterexamples is iririj-e. For noun roots ending in a consonants however, the ending -e dominates over -ɛˀ:
- ass-e, kul-e, kal-e, lis-e, nup-e, kutt-e, kɛss-e;
- rap-ɛˀ, lar-ɛˀ, wos-ɛˀ, op-ɛˀ.
For some roots, we have a coda which is identical with regards to its phonological history, and yet either -e or -ɛˀ is employed:
Based on this, I take the glottal stop a non-phonemic, and also the difference in vowel height. Thus, the noun marking suffix will be written phonemically as /-e/.
2.2. Gemination
In a few entries, we find gemination of a medial consonant. This mostly occurs with s and t. As will be discussed below in detail, geminate tt always goes back to an earlier cluster:
On the other hand, Proto South Babar simple *t is represented as t: *punti > *uti > uti-e. For this reason, I will keep the geminate notation as phonemic.
In the case of geminate s, there is no contrasting source for a single s, since Emplawas s(s) is always derived from an earlier cluster. For this reason, I only use single s in the phonemic transcription.
2.3. Miscellaneous
Taber's transcription distinguishes between v and w. Judging from the occurence of both v and w in warviɛri "new", which goes back to Proto South Babar reduplicated *varviari ~ *vari-vari, I take the contrast as non-phonemic and use /v/ for the single phoneme.
The glide j which is often found between root final -i and the noun ending -e (see 2.1. above) is certainly non-phonemic and will not be spelled in phonemic transcription. Other occurences of j clearly represent phonemic /j/.
3. South Babar historical phonology
The historical phonology of the South Babar languages and the reconstruction of their common ancestor, Proto South Babar, is best understood in the context of Southwest Maluku as a whole.
3.1. Proto Southwest Maluku
Even though the genetic unity of the Southwest Maluku languages still remains to be proven, we can yet reconstruct the phonological system of their immediate common ancestor, the putatative Proto Southwest Maluku. Proto Southwest Maluku can be reconstructed based on evidence from the Luangic languages (Leti, Luang, Wetan), Kisar, Roma, East Damar, Teun, Nila and Serua.
The sound changes from PMP to Proto Southwest Maluku can be summarized as follows:
PMP *q, *h and *p are lost without trace. This has resulted in monosyllabic roots in the case of *ma-nipis > *m-nis and *tuhud > *tur.
In stressed (= penultimate) position, all four vowels remain distinct. PMP *ə became PSWM *e. In unstressed closed syllables, *ə and *a have merged into *a, e.g. *maqitəm > *metam, *tələn > *telan.
In final position (including secondary final due to loss of *q), *i and *u remain distinct. *a and *aw merged into *a, while *ay and *ə (< *-əq) merged into *e.
A very similar sound system was reconstructed by Roger F. Mills for "Proto Barat Daya" (Mills 1991) and "Proto Luangic-Kisaric" (Mills 2010), based on evidence from Luangic languages, Kisar, Roma and Teun-Nila-Serua. The major difference is my reconstruction of *R (< PMP *R/*j) as separate form *r (< *d), based on the fact that in Teun, *R and *j are mostly reflected as zero.
3.2. Proto South Babar
(See the Appendix for a full list of Proto South Babar reconstructions.)
Proto South Babar can be reconstructed as a direct descendant of Proto Southwest Maluku, as reconstructed solely based on evidence from the languages listed above in 3.1. This clearly indicates that the South Babar languages are nested within the Southwest Maluku subgroup, and not coordinate to the latter (as proposed in the lexicostatistical classification by Taber 1993).
Proto South Babar is characterized by some far-reaching shifts and mergers:
Proto Southwest Maluku *ˀ and *R became zero in in Proto South Babar.
PSWM stressed *i, *e remained unchanged, while *a and *u underwent conditioned splits/changes.
3.3. Split of *u into *o and *u
Stressed *u underwent a split in Proto South Babar conditioned by the presence or absence of a following unstressed *i (< *i, *e, *u; see below). The vowel *u remained unchanged if followed by *i, otherwise it was lowered to *o.
'banana'
'body hair'
'vomit'
'knee'
'thick'
Emplawas uti-e
Emplawas vusi-e
Emplawas mok-e
Tela-Masbuar korn-e
Emplawas vokl-i
3.4. Conditioned raising of *a
In non-final syllables, Proto Southwest Maluku *a has a raised reflex in all South Babar languages in absolute initial position, or following a word-initial nasal:
For Proto South Babar, I reconstruct *ə here.
'dog'
'child'
'fire'
'name'
'laugh'
'tongue'
SE Babar uth, C. Masela eti-e, Emplawas uti-e
SE Babar un, C. Masela en-e, Emplawas un-e
SE Babar uwi, C. Masela ej-ei, Emplawas uj-e
SE Babar non, C. Masela nen-e, Emplawas nun-e
Serili -mon, C. Masela -melmiel, Emplawas -mun
SE Babar nom, C. Masela nemn-ei, Emplawas nup-e
In all other cases, PSWM *a is reflected as a (= Proto South Babar *a).
'road'
'stone'
'hot'
Serili hall-e, C. Masela all-ei, Emplawas as-e
SE Babar waxai, C. Masela waij-ei, Emplawas vaki-e
E. Masela palpant, C. Masela pant-i, Emplawas patt-i
Note that the split became phonemic after the loss of *ˀ at the (pre-)Proto South Babar stage, as a result of which *a can also appear in absolute initial position and following a nasal.
'we (ex.)'
'itchy'
SE Babar am, Emplawas am
C. Masela mal, Emplawas makl-i
3.5. Reflexes of Proto Southwest Maluku vowels in the ultimate syllable
Unstressed vowels in ultimate syllables underwent drastic reduction from Proto Southwest Maluku to Proto South Babar, leaving a zero reflex in many South Babar lects. Based on the material in Taber's wordlists, we can make the following reconstrucions for Proto South Babar:
- *a became zero, both in open and closed final syllables;
- *i, e, and *u merged into *i. Closed finals with *i underwent metathesis: *-iC# > *-Ci#.
Complete loss of *a resulted in final clusters in Proto South Babar, which in some lects were broken up again by an epenthetic echo vowel:
'sago'
'kunai grass'
'black'
SE Babar rohol (cf. Emplawas rokl-e)
SE Babar wevel (cf. Emplawas vevl-e)
SE Babar -mekem (cf. Tela-Masbuar mekəm-e)
Proto South Babar *i in ultimate syllables is best preserved with nouns in Emplawas (and also the other Southwest Babar lect), since all nouns carry the noun marking suffix -e.
'stone'
'dog'
'banana'
'wind'
'sky'
'liver'
'body hair'
'brain'
However, if not supported by the noun marking suffix -e, final *i was lost in Emplawas. Some adjectives and transitive verbs have a suffix -i, which is clearly not a reflex of PSB *i, but most probably a third person marker, as is also found in other languages of Maluku.
3.6. Proto South Babar *l and *n
The contrast between *l and *n has become subject to neutralization of varying degree in the South Babar lects, ranging from two distinct reflexes in Central Masela (*l > l, *n > n), to complete merger in Tela-Masbuar (*l/*n > n).
In SE Babar, Emplawas, and Imroing, *l and *n merged to l, unless adjacent to earlier PSB *ə, in which case the reflex is n.
'sand'
'sago'
'name'
'wind'
'back'
SE Babar ili, Emplawas ili-e, Imroing ili-e
SE Babar rohol, Emplawas rokl-e, Imroing rokl-e
SE Babar non, Emplawas nun-e, Imroing nun-mu
SE Babar un, Emplawas usi-e, Imroing unn-e
Emplawas unir-e
3.7. Proto South Babar *k
Proto South Babar *k in medial position underwent further lenition to x/h in Southeast Babar, ˀ in Serili, and zero in Central Masela. In the latter, lenition only affects part of the vocabulary.
'black'
'sago'
'eye'
'louse'
'die'
3.8. Fortition of sonorants in Emplawas
One of the most unusual sound changes in the Southwest Maluku area is the fortition of the sonorants *l, *n, *m and *r in the neighborhood of *n/*l in Emplawas. The rule is quite simply stated: first *l and *n merged into *l, then the following changes occurred:
'moon'
'road'
'ear'
'sleep'
'inside'
'hand'
'thin'
'knee'
'nose'
4. The position of North Babar
North Babar is spoken in the northernmost part of Babar Island, and was grouped by Taber together with Dai and Daweloor into a "North Babar" subgroup, based on lexico-statistical scores. However, closer inspection of qualitative - especially phonological - evidence yields a different result.
4.1. Split of *u into *o and *u
North Babar shows perfect agreement with South Babar languages in the split of *u into o and u.
'banana'
'body hair'
'vomit'
'knee'
'thick'
4.2. High vowels in ultimate position
North Babar agress with the South Babar lects in the complete merger of *u and *i into i in ultimate syllables. In Dai and Daweloor, *u and *i were partially kept distinct and subsequently subject to fortition (*u > ˀ/k, *i > s), which is shared with West Damar.
'egg'
'breast'
'banana'
4.3. Reflexes of *t
Proto Southwest Maluku *t is has three reflexes in North Babar, viz. t, k and ˀ. k and ˀ appear to be in complementary distribution: k in initial position, and ˀ elsewehere. We are thus dealing with a two-fold split into t and k/ˀ.
In initial position, and in medial position with open final syllable, we find the following dstribution of reflexes:
'deaf'
'burn'
'seven'
'we (in.)'
'earth'
'three'
'egg'
'knee'
'dead'
'vomit'
Based on these reflexes, we can tentatively posit a conditioned split: *t was preserved as t in the neighborhood of the high vowels *i and *u, otherwise it shifted to k/ˀ. Lowering of *u must have predated this split, as can be seen in the reflexes korne and -moˀo.
4.4. Conclusion
North Babar clearly shows common features with the South Babar languages. As for the merger of ultimate *u and *i and the split of stressed *u, these are however areal features also found in Wetan, a Luangic language spoken in the western part of Babar island. This sound change however shows that *u and *i did not undergo fortition as in Dai, Daweloor and West Damar. More significant is the shift *t > k, although it only occured partially in North Babar.
North Babar does not share with South Babar the raising of *a, nor the drag-shift from *d (< PSWM *d, *s) to *t, thus agreeing with Dai, Daweloor and West Damar.
Based on these pieces of evidence, we have to place North Babar as an isolate within a higher-level Babar subgroup, not included in Damar-Daweloor nor in South Babar.
References
Mills, Roger F. 1991. Tanimbar-Kei: An Eastern Indonesian Subgroup. In Robert Blust (ed.), Currents in Pacific Linguistics: Papers on Austronesian Languages and ethnolinguistics in Honour of George W. Grace, 241-263. Canberra: Research School of Pacific and Asian Studies, Australian National University.
Mills, Roger F. 2010. Three common misconceptions about Proto-Lettic (Luangic-Kisar). In Chlenova, Svetlana and Fedorchuk, Artem (eds.), ''Studia Anthropologica: a Festschrift in Honor of Michael Chlenov'', 297-326. Moscow-Jerusalem: Gesharim.
Steinhauer, Hein. 2009. The sounds of Southeast Babar. In Adelaar, K. Alexander and Pawley, Andrew (eds.), Austronesian historical linguistics and culture history: a festschrift for Robert Blust, 399-409. Canberra: Research School of Pacific and Asian Studies, Australian National University.
Taber, Mark. 1993. Toward a better understanding of the indigenous languages of southwestern Maluku. Oceanic Linguistics, 32:2, 389-441
van Engelenhoven, Aone. 2010. Tentatively locating West-Damar among the languages of Southwest Maluku. In Chlenova, Svetlana and Fedorchuk, Artem (eds.), ''Studia Anthropologica: a Festschrift in Honor of Michael Chlenov'', 297-326. Moscow-Jerusalem: Gesharim.
The Southwest Maluku Regency (Kabupaten Maluku Barat Daya, traditionally also known as "Tenggara Jauh") is located between Timor Island to the west and the Tanimbar Islands to the east, and comprises a number of small islands and island groups, with a population of around 70.000. The largest island is Babar with an area of approximately 800 km², and which is the eastern-most island in Southwest Maluku.
On Babar island and the smaller neighboring Masela island, we can find a subgroup of the Austronesian language family which is highly innovative with respect to its phonological history. This group was called South Babar by Mark Taber in his survey of the languages of Southwestern Maluku (Taber 1993); this term is also used in the Ethnologue. Next to a paper by Hein Steinhauer (2009) that deals with the phonological history of the Southeast Babar language, Taber's survey is until now the only data source for the South Babar languages.
Seven South Babar lects are documented by Taber, which can grouped into three branches:
a. Central Masela
b. Masela-Southeast Babar
East Masela
Serili
SE Babar
c. Southwest Babar
Emplawas
Tela-Masbuar
Imroing
Taber included Central Masela in his "Masela-SE Babar Cluster". Because of several differences in phonological history, Central Masela is better put into a branch of its own.
The sound changes in this subgroup are quite far-reaching, leading to reflexes of Proto Austronesian etyma that are hardly recognizable. Examples from Emplawas, the most innovative South Babar language, shall illustrate this:
*hangin
*tələn
*daləm
*anak
*hikan
*maRi
*manipis
*tələn
*daləm
*anak
*hikan
*maRi
*manipis
> usi-e
> kes-e
> rap-e
> un-e
> il-e
> pe-e
> pilit-i
> kes-e
> rap-e
> un-e
> il-e
> pe-e
> pilit-i
Although it is not obvious at all at the first sight, each of the Emplawas reflexes here is the result of regular and predicatable sound change.
2. Note on Taber's transcriptions
Although the data collected by Taber is written down phonetically without any further attempt for a narrower analysis, we can still make a few careful steps towards a phonemic interpretation, at least for certain languages in the list. Again, I will illustrate my approach to the data by taking the example of Emplawas.
2.1. Final glottals
Final glottal stops are exclusively found with the noun marking suffix -ɛˀ. This ending also appears as -e without a glottal stop. The distribution of these two variants is quite random. The variant -ɛˀ usually occurs with noun roots ending in -i, often with an intervening glide j:
osi-ɛˀ, uvij-ɛˀ, usi-ɛˀ, wusij-ɛˀ, lalkij-ɛˀ.
One of the few counterexamples is iririj-e. For noun roots ending in a consonants however, the ending -e dominates over -ɛˀ:
- ass-e, kul-e, kal-e, lis-e, nup-e, kutt-e, kɛss-e;
- rap-ɛˀ, lar-ɛˀ, wos-ɛˀ, op-ɛˀ.
For some roots, we have a coda which is identical with regards to its phonological history, and yet either -e or -ɛˀ is employed:
*zalan
*bulan
*bulan
> ass-e
> wos-ɛˀ
> wos-ɛˀ
Based on this, I take the glottal stop a non-phonemic, and also the difference in vowel height. Thus, the noun marking suffix will be written phonemically as /-e/.
2.2. Gemination
In a few entries, we find gemination of a medial consonant. This mostly occurs with s and t. As will be discussed below in detail, geminate tt always goes back to an earlier cluster:
*mapanas
*qijung
*masu-n
*qijung
*masu-n
> *pant
> *irn
> *mətni
> *irn
> *mətni
> patt-i
> itt-e
> mutti-e
> itt-e
> mutti-e
On the other hand, Proto South Babar simple *t is represented as t: *punti > *uti > uti-e. For this reason, I will keep the geminate notation as phonemic.
In the case of geminate s, there is no contrasting source for a single s, since Emplawas s(s) is always derived from an earlier cluster. For this reason, I only use single s in the phonemic transcription.
2.3. Miscellaneous
Taber's transcription distinguishes between v and w. Judging from the occurence of both v and w in warviɛri "new", which goes back to Proto South Babar reduplicated *varviari ~ *vari-vari, I take the contrast as non-phonemic and use /v/ for the single phoneme.
The glide j which is often found between root final -i and the noun ending -e (see 2.1. above) is certainly non-phonemic and will not be spelled in phonemic transcription. Other occurences of j clearly represent phonemic /j/.
3. South Babar historical phonology
The historical phonology of the South Babar languages and the reconstruction of their common ancestor, Proto South Babar, is best understood in the context of Southwest Maluku as a whole.
3.1. Proto Southwest Maluku
Even though the genetic unity of the Southwest Maluku languages still remains to be proven, we can yet reconstruct the phonological system of their immediate common ancestor, the putatative Proto Southwest Maluku. Proto Southwest Maluku can be reconstructed based on evidence from the Luangic languages (Leti, Luang, Wetan), Kisar, Roma, East Damar, Teun, Nila and Serua.
The sound changes from PMP to Proto Southwest Maluku can be summarized as follows:
Without change
PMP
*m
*t
*s
*l
*w
*m
*t
*s
*l
*w
PSWM
*m
*t
*s
*l
*w
*m
*t
*s
*l
*w
One-to-one reflexes
PMP
*b
*d
*z
*k
*b
*d
*z
*k
PSWM
*v
*r
*c
*ˀ (only retained in Kisar, Luang and Dai)
*v
*r
*c
*ˀ (only retained in Kisar, Luang and Dai)
Mergers
PMP
*n/*ng
*R/*j
*nt/*nd
*mp/*mb
*n/*ng
*R/*j
*nt/*nd
*mp/*mb
PSWM
*n
*R
*d
*p
*n
*R
*d
*p
PMP *q, *h and *p are lost without trace. This has resulted in monosyllabic roots in the case of *ma-nipis > *m-nis and *tuhud > *tur.
In stressed (= penultimate) position, all four vowels remain distinct. PMP *ə became PSWM *e. In unstressed closed syllables, *ə and *a have merged into *a, e.g. *maqitəm > *metam, *tələn > *telan.
In final position (including secondary final due to loss of *q), *i and *u remain distinct. *a and *aw merged into *a, while *ay and *ə (< *-əq) merged into *e.
A very similar sound system was reconstructed by Roger F. Mills for "Proto Barat Daya" (Mills 1991) and "Proto Luangic-Kisaric" (Mills 2010), based on evidence from Luangic languages, Kisar, Roma and Teun-Nila-Serua. The major difference is my reconstruction of *R (< PMP *R/*j) as separate form *r (< *d), based on the fact that in Teun, *R and *j are mostly reflected as zero.
3.2. Proto South Babar
(See the Appendix for a full list of Proto South Babar reconstructions.)
Proto South Babar can be reconstructed as a direct descendant of Proto Southwest Maluku, as reconstructed solely based on evidence from the languages listed above in 3.1. This clearly indicates that the South Babar languages are nested within the Southwest Maluku subgroup, and not coordinate to the latter (as proposed in the lexicostatistical classification by Taber 1993).
Proto South Babar is characterized by some far-reaching shifts and mergers:
Without change
PSWM
*p
*m
*n
*l
*r
*p
*m
*n
*l
*r
PSB
*p
*m
*n
*l
*r
*p
*m
*n
*l
*r
One-to-one reflexes
PSWM
*t
*c
*t
*c
PSB
*k
*h (only retained in Serili)
*k
*h (only retained in Serili)
Mergers
PSWM
*s/*d
*w/*v
*s/*d
*w/*v
PSB
*t
*v
*t
*v
Proto Southwest Maluku *ˀ and *R became zero in in Proto South Babar.
PSWM stressed *i, *e remained unchanged, while *a and *u underwent conditioned splits/changes.
3.3. Split of *u into *o and *u
Stressed *u underwent a split in Proto South Babar conditioned by the presence or absence of a following unstressed *i (< *i, *e, *u; see below). The vowel *u remained unchanged if followed by *i, otherwise it was lowered to *o.
'banana'
'body hair'
'vomit'
'knee'
'thick'
PMP
*punti
*bulu
*mutaq
*tuhud
---
*punti
*bulu
*mutaq
*tuhud
---
PSWM
*udi
*vulu-n
*muta
*tur-n
*vutan
*udi
*vulu-n
*muta
*tur-n
*vutan
PSB
*uti
*vulli
*mok
*korn
*vokn
*uti
*vulli
*mok
*korn
*vokn
Emplawas uti-e
Emplawas vusi-e
Emplawas mok-e
Tela-Masbuar korn-e
Emplawas vokl-i
3.4. Conditioned raising of *a
In non-final syllables, Proto Southwest Maluku *a has a raised reflex in all South Babar languages in absolute initial position, or following a word-initial nasal:
Central Masela:
Masela-SE Babar:
SW Babar:
Masela-SE Babar:
SW Babar:
e (with two exceptional cases of o)
u in absolute initial position, o following a nasal
u (with two exceptional cases of o in Tela-Masbuar in absolute initial position)
u in absolute initial position, o following a nasal
u (with two exceptional cases of o in Tela-Masbuar in absolute initial position)
For Proto South Babar, I reconstruct *ə here.
'dog'
'child'
'fire'
'name'
'laugh'
'tongue'
PMP
*asu
*anak
*hapuy
*ngajan
**malip
---
*asu
*anak
*hapuy
*ngajan
**malip
---
PSWM
*asu
*anaˀ
*api
*naRan
*mali
*nama-n
*asu
*anaˀ
*api
*naRan
*mali
*nama-n
PSB
*əti
*ən
*əi
*nən
*məli
*nəm-n
*əti
*ən
*əi
*nən
*məli
*nəm-n
SE Babar uth, C. Masela eti-e, Emplawas uti-e
SE Babar un, C. Masela en-e, Emplawas un-e
SE Babar uwi, C. Masela ej-ei, Emplawas uj-e
SE Babar non, C. Masela nen-e, Emplawas nun-e
Serili -mon, C. Masela -melmiel, Emplawas -mun
SE Babar nom, C. Masela nemn-ei, Emplawas nup-e
In all other cases, PSWM *a is reflected as a (= Proto South Babar *a).
'road'
'stone'
'hot'
PMP
*zalan
*batu
*mapanas
*zalan
*batu
*mapanas
PSWM
*calan
*vatu
*panas
*calan
*vatu
*panas
PSB
*hall
*vaki
*pant
*hall
*vaki
*pant
Serili hall-e, C. Masela all-ei, Emplawas as-e
SE Babar waxai, C. Masela waij-ei, Emplawas vaki-e
E. Masela palpant, C. Masela pant-i, Emplawas patt-i
Note that the split became phonemic after the loss of *ˀ at the (pre-)Proto South Babar stage, as a result of which *a can also appear in absolute initial position and following a nasal.
'we (ex.)'
'itchy'
PMP
*kami
*makatəl
*kami
*makatəl
PSWM
*ˀami
*mˀatal
*ˀami
*mˀatal
PSB
*ami
*makl
*ami
*makl
SE Babar am, Emplawas am
C. Masela mal, Emplawas makl-i
3.5. Reflexes of Proto Southwest Maluku vowels in the ultimate syllable
Unstressed vowels in ultimate syllables underwent drastic reduction from Proto Southwest Maluku to Proto South Babar, leaving a zero reflex in many South Babar lects. Based on the material in Taber's wordlists, we can make the following reconstrucions for Proto South Babar:
- *a became zero, both in open and closed final syllables;
- *i, e, and *u merged into *i. Closed finals with *i underwent metathesis: *-iC# > *-Ci#.
Complete loss of *a resulted in final clusters in Proto South Babar, which in some lects were broken up again by an epenthetic echo vowel:
'sago'
'kunai grass'
'black'
PSWM
*rotan
*vevan
*metam
*rotan
*vevan
*metam
PSB
*rokn
*vevn
*mekm
*rokn
*vevn
*mekm
SE Babar rohol (cf. Emplawas rokl-e)
SE Babar wevel (cf. Emplawas vevl-e)
SE Babar -mekem (cf. Tela-Masbuar mekəm-e)
Proto South Babar *i in ultimate syllables is best preserved with nouns in Emplawas (and also the other Southwest Babar lect), since all nouns carry the noun marking suffix -e.
'stone'
'dog'
'banana'
'wind'
'sky'
'liver'
'body hair'
'brain'
PSWM
*vatu
*asu
*udi
*anin
*lanit
*ate-n
*vulu-n
*ralus
*vatu
*asu
*udi
*anin
*lanit
*ate-n
*vulu-n
*ralus
PSB
*vaki
*əti
*uti
*ənni
*lanki
*əkni
*vulni
*ralti
*vaki
*əti
*uti
*ənni
*lanki
*əkni
*vulni
*ralti
Emplawas
vaki-e
uti-e
uti-e
usi-e
lalki-e
ukli-e
vusi-e
ralti-e
vaki-e
uti-e
uti-e
usi-e
lalki-e
ukli-e
vusi-e
ralti-e
However, if not supported by the noun marking suffix -e, final *i was lost in Emplawas. Some adjectives and transitive verbs have a suffix -i, which is clearly not a reflex of PSB *i, but most probably a third person marker, as is also found in other languages of Maluku.
3.6. Proto South Babar *l and *n
The contrast between *l and *n has become subject to neutralization of varying degree in the South Babar lects, ranging from two distinct reflexes in Central Masela (*l > l, *n > n), to complete merger in Tela-Masbuar (*l/*n > n).
In SE Babar, Emplawas, and Imroing, *l and *n merged to l, unless adjacent to earlier PSB *ə, in which case the reflex is n.
'sand'
'sago'
'name'
'wind'
'back'
PMP
*qənay
---
*ngajan
*hangin
*talikud
*qənay
---
*ngajan
*hangin
*talikud
PSWM
*ene
*rotan
*naRan
*anin
*aliˀur
*ene
*rotan
*naRan
*anin
*aliˀur
PSB
*ili
*rokn
*nən
*ənni
*əlir
*ili
*rokn
*nən
*ənni
*əlir
SE Babar ili, Emplawas ili-e, Imroing ili-e
SE Babar rohol, Emplawas rokl-e, Imroing rokl-e
SE Babar non, Emplawas nun-e, Imroing nun-mu
SE Babar un, Emplawas usi-e, Imroing unn-e
Emplawas unir-e
3.7. Proto South Babar *k
Proto South Babar *k in medial position underwent further lenition to x/h in Southeast Babar, ˀ in Serili, and zero in Central Masela. In the latter, lenition only affects part of the vocabulary.
'black'
'sago'
'eye'
'louse'
'die'
PMP
*maqitəm
---
*mata
*kutu
*matay
*maqitəm
---
*mata
*kutu
*matay
PSWM
*metam
*rotan
*mata-n
*kutu
*mate
*metam
*rotan
*mata-n
*kutu
*mate
PSB
*(me)mekm
*rokn
*mək-n
*oki
*məki
*(me)mekm
*rokn
*mək-n
*oki
*məki
SE Babar
memehem
rohol
mox
ohoi
moxoi
memehem
rohol
mox
ohoi
moxoi
Serili
memˀemi
roˀol
moˀo
hoˀoje
-moˀo
memˀemi
roˀol
moˀo
hoˀoje
-moˀo
C. Masela
memem
ron
monei
ojei
-mokei
memem
ron
monei
ojei
-mokei
Tela-Masbuar
mekəmi
rokne
mukni
oke
muki
mekəmi
rokne
mukni
oke
muki
3.8. Fortition of sonorants in Emplawas
One of the most unusual sound changes in the Southwest Maluku area is the fortition of the sonorants *l, *n, *m and *r in the neighborhood of *n/*l in Emplawas. The rule is quite simply stated: first *l and *n merged into *l, then the following changes occurred:
*ll
*lm/*ml
*rl
*lm/*ml
*rl
> s/ss
> p
> t/tt
> p
> t/tt
'moon'
'road'
'ear'
'sleep'
'inside'
'hand'
'thin'
'knee'
'nose'
PMP
*zalan
*bulan
*talinga
---
*daləm
*lima
*manipis
*tuhud
*qijung
*zalan
*bulan
*talinga
---
*daləm
*lima
*manipis
*tuhud
*qijung
PSWM
*calan
*vulan
*tlina-n
*n-nina
*ralam
*lima-n
*mani(s)nis
*tur-n
*irun
*calan
*vulan
*tlina-n
*n-nina
*ralam
*lima-n
*mani(s)nis
*tur-n
*irun
PSB
*hall
*voll
*klinn
*nnin
*ramn (?)
*lim-n
*mninit
*korn
*irn
*hall
*voll
*klinn
*nnin
*ramn (?)
*lim-n
*mninit
*korn
*irn
Emplawas
as-e
wos-e
klis-e
-sil
rap-e
lip-e
pilit-i
kutt-e
itt-e
as-e
wos-e
klis-e
-sil
rap-e
lip-e
pilit-i
kutt-e
itt-e
4. The position of North Babar
North Babar is spoken in the northernmost part of Babar Island, and was grouped by Taber together with Dai and Daweloor into a "North Babar" subgroup, based on lexico-statistical scores. However, closer inspection of qualitative - especially phonological - evidence yields a different result.
4.1. Split of *u into *o and *u
North Babar shows perfect agreement with South Babar languages in the split of *u into o and u.
'banana'
'body hair'
'vomit'
'knee'
'thick'
PMP
*punti
*bulu
*mutaq
*tuhud
---
*punti
*bulu
*mutaq
*tuhud
---
PSWM
*udi
*vulu-n
*muta
*tur-n
*vutan
*udi
*vulu-n
*muta
*tur-n
*vutan
PSB
*uti
*vulli
*mok
*korn
*vokn
*uti
*vulli
*mok
*korn
*vokn
North Babar
udvi-ja
wunɲ-ai
-moˀo
korn-e
votenne
udvi-ja
wunɲ-ai
-moˀo
korn-e
votenne
Daweloor
urs-ol
wulk-ol
-mut
tur-ol
vutl-el
urs-ol
wulk-ol
-mut
tur-ol
vutl-el
4.2. High vowels in ultimate position
North Babar agress with the South Babar lects in the complete merger of *u and *i into i in ultimate syllables. In Dai and Daweloor, *u and *i were partially kept distinct and subsequently subject to fortition (*u > ˀ/k, *i > s), which is shared with West Damar.
'egg'
'breast'
'banana'
PMP
*qatəluR
*susu
*punti
*qatəluR
*susu
*punti
PSWM
*teluR
*susu(-n)
*udi
*teluR
*susu(-n)
*udi
PSB
*keli
*tutni
*uti
*keli
*tutni
*uti
North Babar
kenj-ai
dutni-je
udvi-ja
kenj-ai
dutni-je
udvi-ja
Daweloor
telk-ol
dudk-ol
urs-ol
telk-ol
dudk-ol
urs-ol
4.3. Reflexes of *t
Proto Southwest Maluku *t is has three reflexes in North Babar, viz. t, k and ˀ. k and ˀ appear to be in complementary distribution: k in initial position, and ˀ elsewehere. We are thus dealing with a two-fold split into t and k/ˀ.
In initial position, and in medial position with open final syllable, we find the following dstribution of reflexes:
'deaf'
'burn'
'seven'
'we (in.)'
'earth'
'three'
'egg'
'knee'
'dead'
'vomit'
PMP
(*tuli)
*tunu
*pitu
*kita
*tanəq
*təlu
*qatəluR
*tuhud
*matay
*mutaq
(*tuli)
*tunu
*pitu
*kita
*tanəq
*təlu
*qatəluR
*tuhud
*matay
*mutaq
PSWM
*tilu
*tunu
*itu
*wutu
*ita
*tane
*tave-n
*telu
*teluR
*tur-n
*mate
*muta
*tilu
*tunu
*itu
*wutu
*ita
*tane
*tave-n
*telu
*teluR
*tur-n
*mate
*muta
PSB
*kili
*kul(i)
*iki
*vuki
*ik
*kal
*kavni
*keli
*keli
*korn
*məki
*mok
*kili
*kul(i)
*iki
*vuki
*ik
*kal
*kavni
*keli
*keli
*korn
*məki
*mok
North Babar
-tini
tunni
iti
-wuti
ita
kanjai
kaunija
kini
kenjai
korne
maˀi
-moˀo
-tini
tunni
iti
-wuti
ita
kanjai
kaunija
kini
kenjai
korne
maˀi
-moˀo
Based on these reflexes, we can tentatively posit a conditioned split: *t was preserved as t in the neighborhood of the high vowels *i and *u, otherwise it shifted to k/ˀ. Lowering of *u must have predated this split, as can be seen in the reflexes korne and -moˀo.
4.4. Conclusion
North Babar clearly shows common features with the South Babar languages. As for the merger of ultimate *u and *i and the split of stressed *u, these are however areal features also found in Wetan, a Luangic language spoken in the western part of Babar island. This sound change however shows that *u and *i did not undergo fortition as in Dai, Daweloor and West Damar. More significant is the shift *t > k, although it only occured partially in North Babar.
North Babar does not share with South Babar the raising of *a, nor the drag-shift from *d (< PSWM *d, *s) to *t, thus agreeing with Dai, Daweloor and West Damar.
Based on these pieces of evidence, we have to place North Babar as an isolate within a higher-level Babar subgroup, not included in Damar-Daweloor nor in South Babar.
References
Mills, Roger F. 1991. Tanimbar-Kei: An Eastern Indonesian Subgroup. In Robert Blust (ed.), Currents in Pacific Linguistics: Papers on Austronesian Languages and ethnolinguistics in Honour of George W. Grace, 241-263. Canberra: Research School of Pacific and Asian Studies, Australian National University.
Mills, Roger F. 2010. Three common misconceptions about Proto-Lettic (Luangic-Kisar). In Chlenova, Svetlana and Fedorchuk, Artem (eds.), ''Studia Anthropologica: a Festschrift in Honor of Michael Chlenov'', 297-326. Moscow-Jerusalem: Gesharim.
Steinhauer, Hein. 2009. The sounds of Southeast Babar. In Adelaar, K. Alexander and Pawley, Andrew (eds.), Austronesian historical linguistics and culture history: a festschrift for Robert Blust, 399-409. Canberra: Research School of Pacific and Asian Studies, Australian National University.
Taber, Mark. 1993. Toward a better understanding of the indigenous languages of southwestern Maluku. Oceanic Linguistics, 32:2, 389-441
van Engelenhoven, Aone. 2010. Tentatively locating West-Damar among the languages of Southwest Maluku. In Chlenova, Svetlana and Fedorchuk, Artem (eds.), ''Studia Anthropologica: a Festschrift in Honor of Michael Chlenov'', 297-326. Moscow-Jerusalem: Gesharim.
No comments:
Post a Comment