This grammatical sketch gives an introduction to the Sula language. While this sketch contains ample interlinearly glossed example sentences and, in several places, statements written to the linguist, I have done my best to write these pages in a way that they will also be accessible to the non-linguist language learner. I am writing this document in the process of earning my PhD in linguistics, but it is my sincere hope that it will not take a PhD in linguistics to understand and make use of it; with luck, even the lowly aerospace engineer or intellectually impoverished neuroscientist will be equipped to divine meaning from these pages. I also hope that these pages might be accessible enough assist Sula community members with establishing language programs as the tongue is increasingly threatened by external pressures.
By and large language learners should feel free to skip past theoretical discussions and focus on the descriptions of what each structure means in practice and the examples demonstrating how to use the structures. When possible, I try to briefly explain important linguistic terminology and concepts inline, and in case you trip up on a bit of terminology, the dissertation’s glossary includes definitions for most if not all of the grammatical jargon used throughout. While this sketch does not pretend to be a pedagogical grammar, it is my hope that by writing it as accessible as possible, it may be of use for aid workers, language planners, and other researchers and interested individuals who may no have formal linguistic backgrounds.
1. Phonology
Every spoken language is built from sounds, but languages all differ in which sounds they use and which sounds are treated as unique versus sounds belonging to a group that share the same function. That is: phonemes are the smallest sounds in a language that can alter the meaning of a word, and many of a language’s phonemes will be grouped into subconscious categories that the speaker population consider to be in some sense ‘the same’ as one another. The different sounds in these groups are known as allophones.
This section describes the basic sounds of the Sula language and how they are grouped. If you are reading this as a non-linguist Sula language learner, it might be helpful to only briefly skim it, as this section necessitates using the International Phonetic Alphabet to differentiate similar sounds, whereas subsequent sections use the Roman alphabet with pronunciations based on the Indonesian practical orthography. Also, as a language leaner, it is probably easiest to absorb much of a language’s phonology at a subconscious level through speech interaction, as it is quite a lot to try to memorize a language’s phonological rules and structures.
1.1 identification of phonemes and allophones
Table 33. consonants
bilabiallabioalveolarpostpalatalvelarglottal
dental alveolar
plosive p, b t̪,d k, g ʔ
nasal m n ŋ
trill r
fricative f s (ʃ) h
approximant j
lateral approximant l
1.1.1 Approximant and affricates
/w/ (voiced, labial velar, approximant)
/t̪ʃ/ (voiceless dental palato-alveolar affricate)
/d̪ʒ/ (voiced dental palato-alveolar affricate)
Table 34. Vowels
front central back
Sula has a five vowel system like many other Austronesian languages, but /o/ in Sula was introduced independently during a process described in the previous chapter in which words that had two /u/ vowels separated by a consonant or a /u/ and a subsequent /i/ separated by a consonant, the first /u/ became an /o/. Introduction of /o/ paved the way for other instances of the vowel to be borrowed into Sula from other languages. The word “Oreo” for example—now a perfectly acceptable Sula word—was borrowed as is without needing to delete the /o/ sounds or substitute a native Sula phoneme in their place.
In languages with small vowel inventories, there is often a broad range of pronunciation variations that still count as the same vowel, and this is also true of Sula’s five vowels. That is: the language’s five phonemic vowels each cover a broad phonetic area. That said, four of the five vowels, i, u, o, ɑ, are usually fairly easily identified. The fifth vowel could be argued as e, or ɛ, but ‘e’ was chosen in the subsequent sections of this sketch simply for the sake of orthographical simplicity. Similarly, /ɑ/ will be written as ‘a’ in subsequent sections.
1.2 Phonemic tally
To determine the approximate frequency of each phoneme across the lexicon, a phonemic tally was conducted on a sample of 1711 unique word instances. Words represent natural Sula language as spoken today on Sanana—this includes loan words from Indonesian. This sampling excluded repeated word instances (e.g. those found in compound words). In the sample, 8,040 total phoneme instances were found, and they break down as follows:
Table 35. Instances of each phoneme
Out of 8,040 total phoneme instances found in 1,711 word sample.
a: 2001
i: 731
u: 465
n: 434
o: 403
b: 377
k: 373
m: 370
t̪: 355
s: 307
e: 295
l: 287
p: 250
h: 215
d: 193
g: 181
r: 165
f: 160
ŋ: 120
j: 113
ʔ: 92
w: 83
t̪ʃ: 38
d̪ʒ: 30
ʃ: 2
1.2.1 Diphthongs/vowel clusters
As sula has a five vowel phoneme inventory, there are twenty possible two-vowel combinations: ae, ai, ao, au, ea, ei, eo, eu, ia, ie, io, iu, oa, oe, oi, ou, ua, ue, ui, uo. A 1,994 item list of non-duplicated entries was referenced in searching for these possible combinations. The instances found were then corrected so that only one instance would be represented for each unique morpheme. This was necessary as some frequently occurring morphemes contained otherwise rare vowel combinations.
Of the twenty possible combinations, ie, oe, ue, and uo were not found to occur at all, while the combinations, ae, ao, eo, io, iu, and ou were found to occur only four, nine, one, two, eight, and two times respectively. Of those, eo was only found in the word baeo ‘bad’ which is a sporadic version of baeu that was only produced by a handful of speakers, and they were together in the same room.
Only ten of the possible combinations were found to occur at least ten unique times across the vocabulary set. These are: ai (50), au (30), ea (14), ei (11), eu (14), ia (47), oa (17), oi (26), ua (35), and ui (12).
Table 36. Possible diphthongs in the Sula language
Asterisks indicate combinations with fewer than ten occurrences.
frontback
mid ea, ei, eo*, eu oa, oi, ou*
low ae*, ai, ao*, au
Because of pronunciation variation among speakers (and sometimes even within the same speaker), it is difficult to determine which if any of the double vowels are phonemic. Further complicating the issue is the fact that many of the vowel combinations are the result of historical and ongoing deletion of intermediate consonants; these deletions are not uniform across all geographic regions or speaker demographics. Essentially, the above table is a best guess, about possible diphthongs across the Sula language community, but if the language does indeed possess phonemic diphthongs, it is extremely unlikely that any speaker’s idiolect would represent all of them. An accurate determination regarding the diphthong status of Sula’s vowel clusters and their distribution will require a focused, study that takes geographical and sociolinguistic variation into consideration.
1.3 Possible consonant clusters
Sula phonology does not permit consonant clusters except (1) in borrowings, (2) at compound-word or reduplication boundaries, (3) in cases of possible prenasalization, and (4) at morpheme boundaries following the prefixes man-, mat-, bis-, bal-, and mak-. This is especially the case in the Sanana dialect, where many final vowels were deleted.Table 37. Consonant clusters present in Sula words
no. ofSula
cluster instances example loan English
dy (1) pad-yaŋa ‘attic’ (under the roof)
fm (1) af-mai ‘ashes’
ft̪ (1) af-tuka ‘ashes’
gy (1) yog-yog ‘almost’
hn (1) basahn ‘angry’ (general)
(This form was received from only one speaker)
kd̪ʒ (1) [mak-d̪ʒad bakai] ‘betrothal’
kk (1) hakkot ‘hug’
kl (1) soklat (L) ‘brown’,‘chocolate’
kr (1) kramat (L) ‘holy place’
(the original Indonesan form contains an /e/, so it is a safe bet that there is a whispered vowel in this form, as it would be the only instance of /kr/ in Sula)
ks (1) saksi (L) ‘witness’
kt̪ (2) pak-tatoto ‘pounder’
lb (3) pamasi kakolbi (L?) ‘harvested rice’
lf (2) bau bal-faa ‘produce’ (make)
lm (1) almanak (L) ‘calendar’
ln (1) bilnaka ‘steal’
mb (14) mboya ‘tail’ (prosthetic /m/)
mt̪ʃ (1) [pohi limt̪ʃui] (partial loan?) ‘lemon’
mf (1) mam-fuwa ‘young’
mp (15) mam-pai ‘bitter’
mpr (1) pampres (L) ‘diaper’ (‘nappy’)
ms (1) samsi beu (L?) ‘headcloth’ (for men)
nb (3) wai nboni tuka ‘saliva’ (prosthetic /n/)
nt̪ʃ (8) [t̪ʃin-t̪ʃin] ‘a small parakeet’
nd (19) ponda (L?) ‘pandanus’
nf (5) nfat ‘stone’ (prosthetic /n/)
nh (1) kau nhal ‘flower’ (prosthetic /n/)
nd̪ʒ (3) [berd̪ʒand̪ʒi] (L) ‘promise’
nl (1) genli ‘like’ (‘in the manner of’)
nn (1) binnaka (L?) ‘steal’ (variant of bilnaka)
np (3) man-parika ‘to divorce’ (to divorce)
ns (2) sinsara (L?) ‘difficult’
nt̪ (27) baku bantu ‘to help’
nj (10) [menjanji] (L) ‘to sing’
ŋg (9) [jiŋgo] (L) ‘beard’
ŋk (8) [ŋkoli] ‘skin’ (prosthetic /ŋ/)
ŋl (1) [paŋlu] (L?) ‘bullet’
ŋm (1) [noŋmarur] ‘embers’
ŋs (1) [baŋsa] ‘tribe’
pf (1) ap-fe ‘smoke’
rt̪ʃ (2) [pert̪ʃaya faa basah] ‘superstition’
rg (1) sorga (L) ‘the upper-world’
rd̪ʒ (2) e.g.[hard̪ʒa] ‘to knit’
rk (1) laka berkema (L) ‘to go camping’
rp (1) harp (L?) ‘to want (something)’
rt (4) martel (L) ‘hammer’
sp (2) despat ’break (bone)’
st (4) pastina ‘coral reef’
t̪r (1) trus (L) ‘continue’
t̪ʔ (1) [putʔana] ‘ghost’
Table 38. Phonotactic environments of each phoneme
phoneme examples
/p/
Initial: [‘poa] ‘blood’
Final: [nap] ‘head/hair’, [ba’map] ‘to cook’, [ba’nap] ‘to shoot’, [‘ʔap] ‘fire’, [hi’dup] ‘to live’
Intervocalic (syllable initial or final): [sa.na’pau] ‘shoulder’, [ma.ta’pia] ‘human’, [‘ta.pa] ‘left’, [ba.pi’kir] ‘to think’
/b/
Initial: [ba.ka’lɛŋ] ‘to lie down’, [ba’wɛl] ‘to climb’, [‘bɛu] ‘to tie’, [bis’nau] ‘to sew’, [ba’nap] ‘to shoot’ ‘to shoot’, [‘bau] ‘to hit’, [‘bag] ‘thick’, [‘bot] ‘white’, [‘buaˌkɛu] ‘to spit
Intervocalic (syllable initial or final): [‘ba.ba] ‘father’, [‘so.ba] ‘wing’
/t/
Initial: [’til] ‘ear’, [‘tua] ‘husband’, [‘toŋ.ka] ‘stick’, [man’tɛl] ‘egg’
Final: [‘fat] ‘stone’, [‘bɛt] ‘day’, [ʔi’staɾ.ʔat] ‘road’
Intervocalic (syllable initial or final): [ma.ta’pia] ‘person’, [‘fa.ta] ‘wife’ [‘jo.ta.haiˌwan] ‘to hunt’ [‘bauˌma.ta] ‘to kill’
/d/
Initial: [da’gat] ‘narrow’, [da’lɛ.na] ‘wide’
Intervocalic (syllable initial or final): [hi’dup] ‘to live’
/k/
Initial: [‘kul] ‘right’, [‘kol] ‘skin’, [‘kag] ‘to fear’
Final: [‘ʔɛk] ‘neck’, [‘mak] ‘tongue’, [‘sak] ‘to pierce’
Intervocalic (syllable initial or final): [‘buaˌkɛu] ‘to spit’, [‘ʔuka] ‘to bite’, [ba’kɛd] ‘to hear’, [ba.ka.fa’ʔo.ja] ‘to yawn’, [‘la.ka] ‘to walk’
/ɡ/
Initial: [‘ɡɛ.hi] ‘to stand’, [ɡa’wai] ‘to scratch’, [‘ɡɛm] ‘to hold’, [‘ɡɛ.ka] ‘to dig’ [‘ɡas] ‘salt’
Final: [‘baɡ] ‘thick’,
Intervocalic (syllable initial or final): [da’ɡiŋ] ‘meat’, [da’ɡis] ‘painful’, [ba’ɡi.la] ‘clean’, [ba’ɡoa] ‘cold’, [da’ɡat] ‘narrow’
/ʔ/
Initial: [‘ʔɛk] ‘neck’, [‘ʔu.ka] ‘to bite’, [‘ʔo.ja] ‘oya’, [ʔi’staɾ.ʔat] ‘road’, [‘ʔuma] ‘house’, [‘ʔas] ‘dog’, [sa’ʔafa] ‘rat’, [‘ʔu.ja] ‘rain’, [‘ʔap] ‘fire’, [‘ʔa.ja] ‘big’
Final: X
Intervocalic (syllable initial or final): [ba.ka.fa’ʔo.ja] ‘to yawn’, [ma’ʔa.na] ‘man’, [‘manˌkau ‘ɡo.ɡa’ʔinˌfoa] ‘feather’
/m/
Initial: [‘mak] ‘tongue’, [‘mu.hi] ‘to smell’, [‘mɛ.na] ‘to cry’, [ma.ta’pia] ‘person’, [ma’ʔa.na] ‘man’, [ba’map] ‘to cook’, [‘mo.ma] ‘to hit’, [‘ma.ta] ‘to die’
Final: [‘ɡɛm] ‘to hold’, [sa’mam] ‘to chew’
Intervocalic (syllable initial or final): [‘ha.ma] ‘eye’, [sa’mam] ‘to chew’, [‘wa.ma] ‘to breathe’
/n/
Initial: [‘na.na] ‘child’, [‘na.hu] ‘long’
Final: [man’tɛl] ‘egg’, [ka’fin] ‘mosquito’, [‘bun.ɡa] ‘flower’, [‘lan] ‘sky’, [ko’kon] ‘yellow’
Intervocalic (syllable initial or final): [hai’wan] ‘animal’, [‘kɛ.na] ‘fish’, [‘kauˌsa.na] ‘branch’, [‘si.na], ‘moon’, [ma’nip] ‘thin’
/ŋ/
Initial: [ŋan] ‘to boil’
Final: [ba.ka’lɛŋ] ‘to lie down’, [‘toŋ.ka] ‘stick’, [da’ɡiŋ] ‘meat’
Intervocalic (syllable initial or final): [faŋ’ʔaɾa] ‘cloud’
/r/
Initial: [rasa] ’emotional’, [rek] ‘to count’
Final: [bɛ.hi’for] ‘lungs’, [ba.pi’kir] ‘to think’, [ma.ka’hor] ‘worm’
Intervocalic (syllable initial or final): [faŋ’ʔaɾa] [‘mo.ɾa] ‘cloud’, [ba’ɾa.sa], ‘thin’, [‘bau.muˌna.ɾa] ‘to work’, [ʔi.’sta.ɾat] ‘road’
/f/
Initial: [‘fɛu] ‘new’, [‘fi.na] ‘female’, [‘fa.ta] ‘wife’, [‘foa] ‘fur’, [‘fua] ‘fruit’, [‘fat] ‘stone’
Final: [kof] ‘coffee’, [baˈtif] ‘steep’ (e.g. a hill)
Intervocalic (syllable initial or final): [sa’fi.la] ‘lightening’, [‘ba.fɛ] ‘smoke’, [bɛ.hi’for] ‘lungs’, [ba.ka.fa’ʔo.ja] ‘to yawn’, [‘bau.fa.ʔaˌsɛl] ‘to grow’, [sa’ʔa.fa] ‘rat’, [ka’fin] ‘mosquito’, [ka.sa’fo.hi] ‘grass’
/s/
Initial: [‘so.ba] ‘wing’, [sa’ʔa.fa] ‘rat’, [‘si.na] ‘moon’, [‘sa.ʔoa] ‘thunder’
Final: [‘ɡas] ‘salt’
Intervocalic (syllable initial or final): [‘kauˌsa.na] ‘branch’, [‘ho.sa] ‘leaf’, [ka.sa’fo.hi] ‘grass’, [ba’ra.sa] ‘thin’
/h/
Initial: [‘hoi] ‘bone’, [‘ha.ma] ‘eye’, [hi’dup] ‘to live’ [hai’wan] ‘animal’, [ma.ka’hor] ‘worm’, [‘ho.sa] ‘leaf’, [‘hai] ‘land’
Final: [besah] ‘off’ (Also: [basah]. e.g. spoiled food), [beˈdah] ‘powder’
Intervocalic (syllable initial or final): [‘da.hi] ‘correct’, [‘ɡɛ.hi] ‘back’, [bɛ.hi’for] ‘lungs’, [‘ni.hi] ‘tooth’, [‘mu.hi] ‘to smell’, [‘jo.ta.haiˌwan] ‘to hunt’ [sa’mo.ha] ‘needle’, [ka’hik] ‘grass’, [ka.sa’fo.hi] ‘grass’, [ma’ho.ka] ‘dull’, [‘na.hu] ‘long’
/j/
Initial: [‘jo.ta] ‘short’, [‘ja.na] ‘to see’ [‘jo.ta.haiˌwan] ‘to hunt’
Final (non-intervocalic): X
Intervocalic (syllable initial or final): [‘ʔo.ja] ‘to throw’, [‘bo.ja] ‘tail’, [‘ʔu.ja] ‘rain’, [‘ʔa.ja] ‘big’, [ba’jon] ‘mouth’, [‘ɡa.ja] ‘to eat’, [‘so.ja] ‘to say’
Cluster: [‘nja.ja] ‘mother’
/l/
Initial: [‘li.ma] ‘five’, [‘lal] ‘liver’, [‘lua] ‘vomit’, [‘la.ka] ‘to walk’, [‘lɛi] ‘to swim’, [‘li.ka] ‘to choose’, [‘la.wa] ‘spider’, [‘lan] ‘sky’, [‘lɛa] ‘sun
Final: [‘kuɫ] ‘right’, [‘koɫ] ‘skin’, [‘tɛ.naˌlal] ‘intestines’, [‘tiɫ] ‘ear’, [ba’wɛɫ] ‘to climb’, [‘sɛɫ] ‘to plant’, [‘bauˌfa.ʔa.sɛɫ] ‘to grow’, [‘buɫ] ‘to squeeze’, [man’tɛɫ] ‘egg’
Intervocalic (syllable initial or final): [sa’fi.la] ‘lightening’, [ba’ɡi.la] ‘clean’, [da’lɛ.na] ‘wide’, [‘tɛ.naˌlal] ‘intestines’, [ba.ka’lɛŋ] ‘to lie down’, [sa’fi.la] ‘lightening’
/w/
Initial: [‘wa.ma] ‘breathe’, [‘win] ‘drink’, [‘wa.ka] ‘root’, [‘wai] ‘water’
Final (non-intervocalic):
Intervocalic (syllable initial or final): [ba’wɛl] ‘to climb’, [‘jo.ta.haiˌwan] ‘to hunt’, [ɡa’wai] ‘to scratch’, [hai’wan] ‘animal’, [‘la.wa] ‘spider’, [ku.ma’wai] ‘sand’
/t̪ʃ/
An nearly exhaustive list of Sula words found to contain the phoneme is: /limu t̪ʃui/ ‘lemon’ (/limu/ is the general term for citrus fruit), /man t̪ʃamo/ ‘heron’, /mont̪ʃa/ ‘type of beete’, /nt̪ʃ(e,i)li/ ‘fruit’ (Mangon dialect), /nt̪ʃumi/ ‘nose’ (Mangon dialect), /rit̪ʃa/ ‘chili pepper’ (L?), /t̪ʃahana/ ‘pants’, /t̪ʃapalong/ ‘top’, /t̪ʃato/ ‘a gift’, /got̪ʃifa/ ‘raft’, /karit̪ʃa/ ‘parakeet’, /onat̪ʃe/ ‘to braid’, /pant̪ʃona/ ‘torch’, /pot̪ʃi/ ‘pot/container’ (L?).
/ʃ/
Just two instances of this phoneme were encountered, and both were within the same, reduplicated word: nish-nish which is a synonym of yota-yota ‘piece’. This is unexplained, likely onomatopoeia or a borrowing.
/d̪ʒ/
Like /t̪ʃ/, this phoneme is incredibly rare except in words that are clearly loans. A near exhaustive list is: /d̪ʒao/ ‘god’ and /d̪ʒoh/ ‘god’(from Hebrew, /Yahweh/ or Ternate /d̪ʒou/ ‘high-ranking person’?), /manabad̪ʒa/ ‘to chat up’, /od̪ʒo/ ‘scramble/mix’, /hard̪ʒa/ ‘to knit’ (interestingly, the Indonesian form, merajut, also contains /d̪ʒ/), /d̪ʒahi/ ‘curtain’ (possibly a Ternate loan), /d̪ʒer(e,i)/ ‘sacred place’ (likely a Ternate loan), /d̪ʒub/ ‘bow and arrow’, /makd̪ʒad bakai/ ‘betrothal’.
1.3.1 Phonological Processes
Aside from an apparently unnatural morphophonological intervocalic devoicing rule, Sula’s phonological processes are for the most part quite straightforward. That said, I suspect there could yet be interesting processes of prothesis, epenthesis, and paragoge to be discovered, but these must await future, targeted research efforts. Following is a list of processes so far encountered:
1.3.1.1 Final Devoicing
Syllable final plosives are often devoiced. This is process varies among speakers.
[-continuant] → [-voice] / _#
- [ba.kɛd̥] ‘to hear’ (in isolation)
compare: [ba’ked.ha’bar] ‘to hear the news’ - [kag̥] ‘to fear’ (in isolation)
compare: [‘ʔak ‘kag a ‘ʔas.ne.ka] ‘I fear that dog’ - [nib̥] ‘to sit’ (in isolation)
compare: [‘ʔak ‘nib a ‘kau.sa.na ‘ʔik.ki] ‘I sit on this branch.’
1.3.1.2 Morphophonologically conditioned intervocalic fortition
The subject of naturalness is widely discussed in the field of phonology, but the term does not yet have a consensus for a definition. Herein I use the term in the manner of the Neogrammarians, i.e. that language change is an unconscious process that results from speakers falling short of their target when trying to reproduce a sound. An analogy to the Neogrammarian view on naturalness is a child learning to throw a ball. The child sees a ball thrown and attempts to reproduce the throw. The child likely won’t be successful at first, but the result of the child’s attempts will be natural in that they are a logical result of the physical motions of attempting the throw. It would be natural for the ball to go straight ahead but not far enough. It would be natural for it to go far enough but a bit to the left or the right. It would be natural for the ball to go too far ahead. But, it would be unnatural for a ball to go directly backwards or straight to the left or right, because those directions of motion don’t result from the body motion to make a typical ball throw.
Similarly with language, your tongue, lips, jaw, palate, and vocal folds all move in particular directions and sequences when we say words. Sometimes we move these parts too much, too little, or in a mixed up sequence, and when that happens, the word pronunciation changes in a natural way. An example of this could be a word with an /n/ after a long vowel. The /n/ sound requires opening the passageway to the nose and sinus, but the vowel sound requires keeping it closed. When speakers mix up the order and open the nasal passageway while still saying the vowels, a different vowel sound is made, and over time this new vowel sound can become a second accepted version of that vowel. So a hypothetical word, [aon] could become [ãõn] and then the phonemes, [a] and [o] would each add an allophone and become [a, ã] and [o, õ]. This sort of change would be very natural, but it would be totally unnatural for the word [aon] to change into [abn], because there is nothing about the speech movements needed for [aon] that could even accidentally end up as a [b] sound.
Many linguists debate whether unnatural sound changes or phonological processes exist, and how and why they would come to pass if so. When apparent examples of unnatural changes and processes are found, they tend to be of particular interest.
There is a strange and apparently unmotivated morphophonological process present in all dialects of Sula: b, d, g in onset position optionally become voiceless when morphology puts them in intervocalic position—i.e. following prefixation, serialization, or compounding (initial b,d,g > [-voice] / V_V). This alternation is productive in all documented dialects of Sula, and it is verified by acoustic phonetic analysis. An important note in reading this grammatical sketch: this is a phonetic alternation whereas the working orthography is phonemic. Outside of this section, the alternation is not usually discernible in the example glosses, because speakers don’t typically perceive the segments as voiceless even when they are produced that way.
1.3.1.2.1 Observed conditions
Voicing alternation occurs only to plosives that are underlyingly voiced
Table 39. Voicing alternation examples
Underlying Alternation example
b/p
- basel basel fia-kau neka ‘plant that tree’
‘to plant’ i-pasel fia-kau neka ‘3SG plants that tree’ - pia pia makata bau munara ‘the doctor is working’
‘live/person’ i-pia moya ‘3SG is not alive’
d/t
- detdet moya kau neka ‘don’t cut that tree’
‘to cut’ a-tet kau neka ‘I cut that tree’ - tedted lepa nui fat hia ‘lift one coconut’
‘to hold’ ak dad a-ted mua-mua ‘I can hold all of them’
k/g
- gayagaya moya pel ‘don’t eat anymore’
‘to eat’ a-kaya ‘I eat’ - kagkag moya pel ‘don’t be afraid anymore’
‘to fear’ a-kag moya ‘I’m not afraid’
b ~ p
- baked ‘to hear’: a-paked ‘I hear’
- bama ‘to split’: a-pama ‘I split’
- basel ‘to plant’: a-pasel ‘I plant’
- bis ‘full, satisfied’: ak a-kaya moya, sebab a-pis pel
‘I don’t want to eat because I’m full’ - bisloi ‘look for’: a-pisloi ‘I look for’
- bisnau ‘to sew’: a-pisnau ‘I sew’
- bisnoya ‘to talk’: ta-pisnoya ‘1PL.INCL.AGR says’
d ~ t
- det ‘to cut’: a-tet ‘I cut’
- donu ‘to burn’: ba-tonu ‘to ignite’ (Mangon)
- duf ‘hole’: bau-tuf ‘make a hole’
- dona ‘to burn’: ba-tona ‘to ignite’ (Sanana)
1.3.1.2.1.1 dona ~ tona in conversation
A young female speaker produced morpheme-initial /d/ as [d] even in a V_V environment (i.e. she does not always produce the devoicing):
- ba-dona. kit kata soya ba-dona bo ak bahasa
CAUS-burn 2PL word say CAUS-burn in 1SG language
‘‘Ignite’. In my language, we say ‘ignite’.’
An older male speaker seems to always devoice morpheme-initial b, d, g when it is put in a V_V environment:
- dona, dona, ba-tona, a a ba-tona…
burn, burn CAUS-burn, yeah yeah CAUS-bur
‘‘Burn’, ‘burn’. ‘ignite’, yeah, yeah, ‘ignite’…
1.3.1.2.1.2 g ~ k
a. kalea-keka
‘kidney disease’
- gem ‘to hold’: mana-kem ‘to hold (in the manner of firmly grasping another’s arm)’
- gaya ‘to eat’: a-kaya ‘I eat’
1.3.1.2.1.2.1 gaya ~ kaya in conversation
A young female speaker produces both [g] and [k] variants in an identical phonetic environment in adjacent sentences
- ak bit a- gaya fia-sinaŋa hia
I want 1SG.AGR eat banana-fried one
‘I want to eat a fried banana.’ - ak a- kaya fia-sinanga hia
I 1SG.AGR eat banana-fried one
‘I eat a fried banana.’
1.3.1.2.2 Devoicing examples
1.3.3.2.2.1 Devoicing in compound words
The word geka ‘ill, injured’ becomes [keka] when compounding places the initial segment in intervocalic position. That is, kalea ‘kidney’ plus geka ‘ill’ is often if not typically produced as, kalea-keka ‘kidney disease’. Compare this to the Sanana dialect form of the compound word meaning ‘headache’. In this form, the initial segment of the same word, geka, is placed in non-intervocalic position. In this case the resulting form is nap-geka ‘headache’, and it is seldom if ever produced with a voiceless segment. The Mangon dialect of Sula still retains many final vowels that were dropped in Sanana, including a final u on the form napu ‘head’. This gives corroborating evidence of the alternation, because the Mangon variant of the same compound causes an intervocalic environment, and expectedly, the resulting form is devoiced, napu-keka ‘headache’.
1.3.3.2.2.2 Devoicing with pronominal prefixes
Verbs that begin with a b, d, or g undergo the same alternation due to agreement morphology. The alternation has been observed with several of the pronominal prefixes, but there may be differences in the frequency of occurrence across the different markers. Complete paradigms will be needed to fully explain the phenomenon, and that will require a well-planned study of numerous speakers that considers speaker demographics, speech context, and broader phonetic environment considerations.
1.3.3.2.2.2.1 1st person singular:
det ‘to cut’: (ak) a-tet ‘I cut’.
Compare this example to ted ‘to hold, carry’ which begins with a voiceless segment and which shows no alternation following prefixation: (ak) a-ted Sua ‘I carry Sula’.
1.3.3.2.2.2.2 1st person plural:
bisnoya ‘to say, talk’: (kit) ta-pisnoya ‘(we) 1PL.INCL-say’
Compare this example to pahu ‘to appear’ which begins with a voiceless segment and which shows no alternation following prefixation: (kit) ta-pahu gan gatel para bo uma bet ik ‘we seem to be just three at home today’.
1.3.3.2.2.2.3 3rd person singular:
basel ‘to plant’: i-pasel fia-kau neka ‘3SG-plant that tree’
Compare this example to pia ‘to live’ which begins with a voiceless segment and which shows no alternation following prefixation: i-pia moya ‘s/he is not alive’.
1.3.3.2.2.3 Devoicing with verbal prefixes
Verbal prefixes can also result in intervocalic environments that trigger the alternation.
- baked ‘to hear’: baka-paked ‘to message’ (direct digital messaging or conveying a message on another’s behalf)
- ban ‘intercourse’: mana-pan ‘to initiate sex’
- donu ‘to burn’: ba-tonu ‘to ignite’ (Mangon)
- duf ‘hole’: bau-tuf ‘make a hole’
- gem ‘to hold’: mana-kem ‘to grab’
- gehi ‘to stand’: baka-kehi ‘to erect, construct’
Compare these forms to forms made with the prefix, mat-, which does not result in an intervocalic environment. The mat- prefix therefore does not trigger a voicing alternation: for example, mak-bobai ‘dating (courtship)’ and mak-dahi ‘to meet’.
1.3.3.2.2.4 Devoicing in serial verb (auxiliary verb?) constructions
It is unclear whether Sula’s laka constructions should be called auxiliary or serial constructions, but regardless, they trigger the voicing alternation as do the more clearcut instances of serialization.
- gaya ‘to eat’: laka kaya ‘go eat’
- basel ‘to plant’: laka pasel ‘to go planting’
- bisloi ‘look for’: laka pisloi ‘to go searching’
- gega ‘to awaken’: nona-kega ‘the routine of going to bed and getting up’
1.3.3.2.2.5 Devoicing in reduplicated forms
Although it doesn’t seem to be as frequent as with other triggers, the alternation can also be triggered by reduplication.
Compare the underlyingly voiceless p in pia ‘safe’: pia-pia ‘safe’ (with emphasis).
1.3.3.2.2.6 devoicing in numeral prefixes?
This is the only instance of the voicing alternation that appears to be mandatory rather than optional. This is intriguing in that the forms have been shown to trace back to Proto-Austronesian (ref. Blust and Trussel ongoing). This might be coincidental, but the pattern matches perfectly with the other, still productive, alternations present in the language. In concert with the historical changes Blust (2005) identified in Kiput and Berawan, and the reconstructable PAN numeral prefixes, Sula’s still-productive alternation suggests a complex historical voicing problem in the fabric of the Austronesian language family.
Table 40. Devoicing of numeral prefix
ga- ‘cardinal prefix’ pa-ka– ‘ordinal prefix’
ga-tel ‘three’ pa-ka-tel ‘third’
ga-reha ‘four’ pa-ka-reha ‘fourth’
ga-lima ‘five’ pa-ka-lima ‘fifth’
1.3.3.2.3 Evidence
1.3.3.2.3.1 Acoustic evidence
Spectrograms often show quite clearly that the b, d, g variants are completely voiced and the devoiced variants are completely voiceless, but the difference is not always binary; often the intervocalic segments are just less voiced. This suggests either a sound change in progress and/or a tug-of-war between the intervocalic devoicing process and the more natural, universal tendency to voice intervocalic segments. The following spectra illustrate the most clear difference between voiced and voiceless variants:
Figure 19. Spectrogram of voiced segment.
A young, female speaker in Capuli village did not produce the alternation in the word dona following prefixation.
ba-dona
Figure 20. Spectrogram of voiceless segment.
An older, male speaker in Capuli village did produce the alternation in the same word following prefixation.
ba-tona
1.3.3.2.3.2 Minimal pair evidence
A voicing contrast in the language prevents analysis of the alternating segments as underlyingly voiceless. If a phonological voicing process were in place that caused initial voiceless segments to become voiced in citation form, Sula would not have other words that begin with p, t, k; but there are many. For example, words like kag ‘to fear’ would be *[gag] when no prefix is applied, but they are not: (1) a-kag ‘I’m afraid’ (2) kag moya pel ‘don’t be afraid again/anymore’
1.3.3.2.3.2.1 b/p
- a. basa ‘bad’
b. pasa ‘before’ - a. bahu ‘quick’
b. pahu ‘to appear’ - a. bia ‘from’
b. pia ‘life’ - a. baka ‘CAUS’
b. paka ‘ORDINAL’ - a. bo ‘LOC’
b. po ‘ten’
1.3.3.2.3.2.2 d/t
- a. dena ‘aboard’
b. tena ‘stomach’ - a. dad ‘can’
b. tadu ‘horn’ - a. da ‘bye’
b. ta 1PL.INCL - a. dua ‘to push’
b. tua ‘old’ - a. duk ‘to come’
b. tuka ‘intestines’
1.3.3.2.3.2.3 g/k
- a. goi ‘buttocks’
b. koi ‘don’t’ - a. gena ‘to hear’
b. kena ‘fish’ - a. gan ‘in the manner of’
b. kan ‘injury’ - a. gaya ‘to eat’
b. kaya ‘wealthy’ - a. gau ‘two’
b. kau ‘tree’
1.3.3.2.4 Peculiarities: A more natural, intervocalic voicing process
Some speakers frequently voice word-final stops when the following segment is +voiced or +continuant. This is never mandatory and it is certainly much more natural than the previously discussed alternation. For example, ak ‘I’ is frequently produced as [ag] when followed by a voiced or continuant segment. Examples:
- dok ag gon camera
give I belong camera
‘give me MY camera’ - ag na Ismael
I name Ismael
‘my name is Ismael’ - ki’i bo a(k,g) lug
s/he LOC I front
‘he is in front of me’ - ki’i bo ag gehi
s/he LOC I back
‘he is behind me’ - ki’i bo ag ila-pon
s/he LOC I beside
‘he is beside me’
1.3.3.2.5 An exception?
duf ‘hole’ seems to sometimes be produced with [t] even when not V_V , e.g. sak (d,t)uf til ‘pierce ear’ (It could be that the one of the phonemic forms is actually /tuf/ or /saku/)
Incomplete, pronoun-specific inflection systems that are superficially similar have been proposed for some Central Maluku languages (e.g., the neighboring Soboyo; Fortgens 1921:20-22). A similar analysis could not explain the Sula data for several reasons. (1) The proposed systems do not describe alternations in intervocalic position that are exclusively limited to voicing. (2) The hypothetically inflected forms do not occur in pro-drop constructions (where V_V is not present). (3) The alternation is also sometimes present in affixed forms and intervocalic environments in compound words independent of person-inflection. (4) The alternation appears to be exceptionless in Sula; initial-segment devoicing is available to all verbs beginning with voiced stops when a target environment is present, and not, as inflection proposals describe, a system which affects only some words.
An intervocalic environment is the only apparent condition that is consistent in all situations, and this makes an active phonological process the most likely explanation for the alternation. The data show a process occurring in all dialects of Sula I have worked with so far, and this universal distribution suggests the process has been productive for several generations at least.
1.4 Syllable stress
Where words have light final syllables, Sula has penultimate stress (e.g. [‘gɛ.hi] ‘back’). Stress is final in words with heavy final syllables (e.g. [ba.pi’kir] ‘to think’). A moraic analysis fits the Sula data well (where closed syllables and open syllables with a diphthong or long vowel are bimoraic). Under such an analysis, stress would predictably fall on whichever syllable contains the penultimate mora.
In compound words, primary stress is located on the stressed syllable of the first word, and secondary stress occurs on the stressed syllables of the following words.
- [‘t˭ɛ.naˌlal] ‘intestines’
- [‘sa.naˌpau] ‘shoulder’
2 description of orthography
This section provides a description of the Sula orthography and the phonetic value of the characters used in this grammatical sketch.
Virtually all Sula speakers are bilingual Malay speakers, and most are already literate in Malay. Because of this, it makes most sense to stick with what people already know and use rather than inventing a new orthography that is completely foreign to speakers. The Sula language can be represented well with Indonesian orthographical conventions and with only a few small differences from standard Bahasa; these differences are shown below.
2.1 How words are pronounced
Table 41. Sula to Indonesian and English approximate pronunciations
Sula Indonesian English
letter Example Example
p [p] apa apple
b [b] lebih scuba
m [m] umur name
f [f] – wifi
t [t̪] utara town
s [s] rasa dense
d [d] padat trade
n [n] makanan manage
r [r, ɾ] dari dari
l [l̴] – ball
k [k] makanan flaky
y [j]yangyear
ɡ [g] keluarga omega
ng [ŋ] dingin swinging
‘ [ʔ] tidak ‘apple
h [h] bahan aha
w [w] awalnya towing
c [t̪ʃ] coklat (some dialects) choose
j [d̪ʒ] juta jump
3 greetings and introductions
As is common in the region, Sula speakers tend to call out greetings to any unexpected passerby. At first, I assumed this to be the curiosity of a few bored strangers, but to my surprise, it turned out to be a widespread and surprisingly nuanced greeting formality.
Sula greetings are structured around (a) where one is going, and (b) what one is doing. One is generally expected omit locational details except when they contradict what would otherwise be assumed, and this holds for greetings as well (see section 9, directional system and space). Unfortunately for outsiders, there is no general assumption about one’s actions or destination, so there is an expectation of specificity, and this can make an otherwise ten minute walk turn into an all-day affair if specific, detailed answers are given to each bystander while walking through town.
Out of fatigue and a need to get from point A to point B within a reasonable amount of time, I started experimenting with ways to interrupt the expected pattern of greeting. Doing so began to reveal an intricate etiquette that parallels the specificity defaults in the language’s positional grammar. There are two basic greeting formats for situations in which there is an implicit assumption as to the other’s goings on. One exchange (a) is used when two people meet and at least one of them is en route to a destination. The other (b) is used when people meet who are not en route to a destination.
3.1 Exchange type a (when both participants are en route)
- Person 1:
laka s(a(h))oa (neka)
walk where (that)
‘where are you going?’ (‘where is it you’re going?’) - Person 2:
laka s(a(h))oa (neka)
walk where (that)
‘where are you going?’ (‘where is it you’re going?’)
In this exchange, neither participant need answer the question. This is much like the English greeting exchange of Hiya doin’? answered with Hiya doin’? in that neither question is interpreted as a literal request for information. However, like in English, it is perfectly acceptable for the Sula speaker to answer the question literally if s/he desires to engage in conversation.
3.2 Exchange type a (when only Person 2 is en route)
- Person 1:
laka s(a(h))oa (neka)
walk where (that)
‘where are you going?’ (‘where is it you’re going?’) - Person 2:
a- laka (direction)
1SG.AGR walk (direction)
‘I’m going that way’
In this exchange, person 2 does not respond by repeating Where are you going? back to Person 1, since s/he is not en route anywhere. In this case, the most common response is for Person 2 to reply in the most general terms by stating the direction toward which s/he is walking. Again though, it is perfectly acceptable for Speaker 2 to answer the question more specifically if s/he desires to engage in a conversation.
3.3 Exchange type b (when neither person is en route)
- Person 1:
bau hapa (neka)
do what (that)
‘what are you doing?’ - Person 2:
bau hapa
do what
‘what are you doing?’
Like in exchange type (a) above, in this exchange, neither participant need answer the question, but it is perfectly acceptable to do so if a speaker desires to engage in conversation.
There is also a secondary greeting (c) which translates roughly to ‘how are you’. Like exchange type (b) this tends to be directed at a person who is not en route to a destination.
3.4 Exchange type c
Replies to this question vary more, but a common reply is:
The greeting etiquette’s complexity begins to surface when one of the participants is unexpected at the location or time of the exchange. In these situations, there is an ongoing back-and-forth between information requests and evasion strategies—ways to vaguely answer while remaining within discourse boundaries.
My assistant, Marlia, was often aggravated at people’s nosiness as we walked back and forth to our various consultants’ houses. At first I suggested she just ignore people and walk on, but that was not an option for her, because ignoring greetings from bystanders is itself a rude and aggressive action. Marlia had a clever (if somewhat passive aggressive) strategy of dealing with unwelcome greeting questions.
The basic discourse expectation for a greeting is one round of exchange. If that exchange begins with a ‘where are you going’, the reply, ‘where are you going’ is only available if the first asker is also en route. However many people in Sula spend significant portions of every day sitting in front of homes or on public sitting platforms. And that means that for passers by, they cannot simply respond back with, ‘where are you going’; I have made that mistake numerus times, and it leads the conversation initiators to look at me as though I am either mentally deficient or incredibly discourteous before replying in a stern, deliberate tone: ak nib. laka sahoa? ‘I’m sitting. Where are you going?’
When prompted, ‘where are you going’, by a person who is not in motion, the generally expected answer is, ‘I’m going that way’, where “that way” is substituted for one of Sula’s four directions (tema ‘inland’, fai ‘seaward’, lepa ‘up’, neu ‘down’). If the answerer desires, s/he can instead respond with a general statement of intention, e.g. a-laka bihu ‘I’m returning’. More specific answers are sometimes provided when the action or destination is already obvious. For example, someone walking toward a water source carrying a wash basin and soap might say a-laka nan ‘I’m going to bathe’, or someone walking toward their home might answer a-laka bo uma ‘I’m going home’. Although such replies might seem passive-aggressively evasive, they suffice the discourse expectation. However, for outsiders and even insiders in unexpected situations (e.g. someone walking in the heat of the day during Ramadan, someone dressed funny, or someone carrying a bizarre item as they walk by), the expectation is that they provide a specific answer, and a reply of ‘I’m going (that way)’ will generally be met with a repeat of ‘where are you going’ in an assertive tone.
As walking with me, an outsider, was inherently an unexpected action of my assistant, Marlia began to preempt the question in order to avoid having to explain our goings on to everyone we passed. This proved tricky though, because the sitters and the passers by see each other from far away, and if one speaks while there is still too much distance, the other party can ignore the question as though it were not heard and then proceed to reinitiate the discourse when they get closer. When Marlia timed it right (i.e. before the other spoke but close enough that the other couldn’t feign not hearing her question), the sitters would be locked into a greeting exchange initiated by Marlia, and the bystander would not be able as easily able to then ask us where we were on our way to, since the discourse parameters defaulted to just one round of greeting exchange. This often caused unmistakable frustration on the faces of bystanders as we passed by. Such an exchange tends to go as follows:
3.5 Preempted exchange
- Marlia:
bau hapa neka?
do what there
‘what are you doing?’ - bystander:
nib-nib (para)
sit-sit (just)
‘I’m (just) sitting.’
The single-round discourse expectation is evident both by the frustration on the faces of bystander who were unable to inquire as to our destination and also by Marlia’s reaction to the occasional brash individual who would follow up the exchange with, laka sahoa? ‘where are you going’. In those circumstances, the ever-polite Marlia would begrudgingly answer the question, and then turn to me red faced as we walked on and proclaim, “he is so rude!”
To further nuance the art of greeting in Sula, all of the above interactions can be substituted with anticipated answers. For example, as one walks inland, a bystander might call out, laka tema? ‘walking inland?’ which is the expected answer to laka sahoa? To this question, the passer by can either repeat, laka tema. ‘walking inland.’ or answer, iyo. nib-nib para? ‘Yeah. Just sitting?’ at which point the bystander can either say ‘yes’ or let the conversation fade. Likewise, this entire transaction could be reversed with the passerby first inquiring, nib-nib? ‘are you sitting’.
And to condense the interaction even further, either party can preemptively declare what the other is doing with declarative intonation. For instance, when passing by, Marlia frequently called out to bystanders, nib-nib. ‘sitting’ in declarative intonation. This effectively cancelled the entire discourse and answered for the bystander what s/he is doing. To this, the polite options are (a) for the sitting party to say, ’yes’, (b) to say, ’walking.’ (as in, ‘you are walking’) or, (c) say, ‘Yes. Walking.’ (as in, ‘Yes I am, and you are walking’).
Sula’s greeting formalities might seem like a lot to remember, but they are quite logical and straightforward in their own right. While It could take awhile to fully master the etiquette, one can learn to function politely in a short amount of time.
4 Counting
The language of everyday trade and commerce in the Sanana markets is a mix. People coming from the remote villages tend to speak Sula when communicating with other Sula trade partners. Those who are from Sanana city proper are fully able to understand basic transactional Sula, but they are more likely to code switch and reply in a mix of Sula and local Malay. Those from Sanana city are much more likely to count in Malay (even though most are able to count in Sula), while many older people from the villages will tend to count in Sula. Financial transactions and money counting on the other hand are almost exclusively in Malay. It seems as though the currency bill denominations are named by their Malay numerical values, and they are spoken of more than counted as such. In small villages there are no markets as such, but Sula is much more likely to be spoken throughout trade and financial interactions there than in the city.
Table 42. Cardinal Numbers
Sanana Mangon English
gane ganei ‘six’
gatahua gaatua ‘eight’
gatsia gatasia ‘nine’
poadogatel podigatelu ‘thirteen’
po(a)gatel pogatelu ‘thirty’
po(a)gatel do hia pogatelu di gia ‘thirty-one’
po(a)gatel do gahu pogatelu di guu ‘thirty-two’
po(a)gatel do gatel pogatelu di gatelu ‘thirty-three’
po(a)galima pogalima ‘fifty’
4.1 Counting Objects
Sula has a common, general classifier (counting word), fat (fatu in Mangon), which has a meaning similar to ‘pieces’ in its numeral sense. fat also means ‘stone’ when not used in counting. The following pattern is used to count objects: fat hia, fat gahu, fat gatel (‘one piece’, ‘two pieces’, ‘three pieces’). fat is also used to state numbers of items. For example:
- ana-mehi-fina dok ana-mehi-maana weuw fat gatel
child-small-female give child-small-male mango CLF three
‘The girl gave the boy 3 mangoes’
4.2 Ordinals and multiplicatives
When items are numbered in order, numerical instances are formed with the multiplicative and ordinal prefix pa-. The ka on the word pakahia ‘first’/‘once’ indicates that the ga- cardinal number prefix may indeed still be an active prefix rather than a fossil from Proto–Buru–Sula–Taliabo as is speculated in Collins (1981). Evidence for this is found in the Sanana word hia ‘one’—on cardinal numbers two through nine, ga- is present, but hia lacks this prefix. The form gahia does, however, exist meaning ‘alone’. Because the ordinal and multiplicative form for the number one is not *pa-hia, and because the form for two is not *paka-ga-hu (and likewise for three–nine), it seems likely that pa- is the ordinal and multiplicative prefix, and ga- is the still somewhat productive cardinal prefix—a prefix whose g is devoiced to k due to the language’s bizarre process of intervocalic fortition discussed in the phonology section. It must be noted for historical linguists, that paka- corresponds exactly to a similar prefix that marked multiplicative numerals in Proto-Malayo-Polynesian and in several other Austronesian languages. It is likely that pa- and ka-/ga- in Sula are derived directly from *paka- but that the syllables were either reanalyzed as separate morphemes in Sula or that Sula preserves an earlier ordinal/multiplicative prefix relationship that was lost in other Austronesian languages who reanalyzed their cardinal prefix as part of their multiplicative prefix.
Table 43. Ordinal and multiplicative counting
Sanana ordinal multiplicative
pa-ka-hia ‘first’ once
pa-ka-hu ‘second’ twice
pa-ka-tel ‘third’ thrice
pa-ka-reha ‘fourth’ four times
pa-ka-ne ‘sixth’ six times
pa-ka-pit ‘seventh’ seven times
pa-ka-tahua ‘eighth’ eight times
pa-ka-tsia ‘ninth’ nine times
pa-ka-poa ‘tenth’ ten times
pa-ka-poadohia ‘eleventh’ eleven times
pa-ka-poadoga’u ‘twelfth’ twelve times
pa-ka-poadogatel ‘thirteenth’ thirteen times
pa-ka-poga’u ‘twentieth’ twenty times
pa-ka-pogatel ‘thirtieth’ thirty times
pa-ka-pogatel do hia ‘thirty-first’ thirty-one times
pa-ka-pogatel do gahu ‘thirty-second’ thirty-two times
pa-ka-pogatel do gatel ‘thirty-third’ thirty-three times
pa-ka-pogalima ‘fiftieth’ fifty times
pa-ka-ota ‘hundredth’ hundred times
pa-ka-cahia ‘one thousandth’ one thousand times
pa-ka-caga’u ‘two thousandth’ two thousand times
- ordinal:
paka-hia son
ordinal-one more
‘one more time’ - multiplicative:
paka- pila pel gu- duk bo Sua?
ordinal- how.many already 2SG.AGR come LOC Sula
‘How many times have you visited Sula?’
5 Word order
A limited number of discourse strategies like topicalization can produce sentences with non-Subject-Verb-Object word order, but Sula is for the most part an SVO language. The language’s optional pronominal prefixes help to limit ambiguity—allowing noun phrases to be reordered—but this occurs infrequently during non-elicited speech.
5.1 Left-dislocation and word order
5.1.1 Transitive verbs
- Non-topicalized transitive
ak a- dota mon
1SG 1SG.AGR- strike you
‘I punched you.’ - Topicalized transitive with pronoun
mon, ak a- dota
you, 1SG 1SG.AGR- strike
‘It’s you I punched.’ - Topicalized transitive without pronoun
mon, a- dota
you, 1SG.AGR- strike
‘It’s you I punched.’
5.1.2 Intransitive verbs
- Non-topicalized unergative verbs
ak a- laka bo Waitebi
1SG 1SG.AGR- walk LOC Waitebi
‘I go to Waitebi.’ - Topicalized unergative verbs
bo Waitebi, ak a- laka
LOC Waitebi, 1SG 1SG.AGR- walk
‘To Waitebi I go.’ - Non-topicalized unaccusative verbs
ak nib bo Waitebi
1SG sit LOC Waitebi
‘I live in Waitebi.’ - Topicalized unaccusative verbs
bo Waitebi ak nib
LOC Waitebi 1SG sit
‘In Waitebi I live.’
5.2 Double objects (ditransitives) and word order
Sula permits optional double-object constructions. In non-double-object constructions, the oblique NP and the direct object can be optionally reordered. In double-object constructions; however, the NP order is obligatorily.
- Non-double-object construction with ‘give’. The indirect object is linked by a locative rather than word order or pronominal prefix.Ikbal i- dok bunga bo FatimaIkbal 3SG.AGR- give flower OBL Fatima
‘Ikbal gave a flower to Fatima.’ - Non-double-object construction with ‘give’ and reordered NPski’i i- dok bo ak kastela3SG 3SG.AGR- give OBL 1SG potato
‘He gave to me kastela.’ - Double-object construction with ‘give’: DO1, DO2Ikbal i- dok Fatima bungaIkbal 3SG.AGR- give Fatima flower
‘Ikbal gave Fatima a flower.’ - *Double-object construction with ‘give’: DO2, DO1 *Ikbal i- dok bunga Fatima
Ikbal 3SG.AGR- give flower Fatima
‘Ikbal gave a Fatima a flower.’
Sula permits alternate word orders to some degree, but it is apparently forbidden in double-object constructions.
- Non-double-object construction with ‘show’a- bakatoya ak handfon bo mon1SG.AGR- show 1SG phone LOC you
‘I showed you my mobile.’ - Non-double-object construction with ‘show’a- bakatoya mon ak handfon1SG.AGR- show you 1SG mobile
‘I showed you my mobile.’ - Non-double-object construction with ‘show’a- bakatoya bo mon ak handfon1SG.AGR- show LOC you 1SG mobile
‘I showed you my mobile.’ (increased emphasis compared to (f)) - *Double-object construction with ‘show’ *a- bakatoya ak mon handfon
1SG.AGR- show 1SG you mobile
‘I showed you my mobile.’
6 verbal indexing
Sula possesses a pronominal indexing system in which persons may be optionally indexed on the verb via a prefix. There is optional dropping of pronouns and pronominal prefixes—and even both simultaneously. The conditions on pronominal indexing are not well understood, but may include (a) increased signal robustness or emphasis (when both pronoun and pronominal prefixes are used), (b) indicate reference disambiguation and reference switching, (c) indicating aspectual information. These topics are discussed further in the word order and verbs sections.
The overlapping functions of the pronominal prefixes along with intra-speaker variation and disagreement regarding grammatical judgment tests make it difficult to decipher the system with a great deal of specificity, and indeed scholars working on nearby languages have even described similar systems as, incomplete verbal agreement systems (##CITATION##). It could take an entire dissertation to begin to reveal the nuances of these systems, but some broad generalities can be made and are presented in the aforementioned sections.
The high frequency of pronoun omission and pronominal prefix omission make coming up with a full paradigm very difficult to get from one speaker, so this section will present the pronominal indexing syntax as it was agreed upon and presented to me by a group of around a half dozen, middle-aged Sanana speakers (a mix of men and women) from the Fagudu and Falahu tribes who worked with me for about an hour in Pasar Bamboo to come up with a paradigm they could all agree on. It is important to note that there is a lot of geographical variation even among Sanana speakers regarding the pronunciation of pronouns and pronominal prefixes, but none of the variants encountered were different enough to obscure meaning, so the following paradigm should suffice for a beginning Sula learner hoping to communicate anywhere on the island of Sanana.
Table 44. Pronouns
PronounPronominal prefix
Second-person singular mon gu-
Third singular (human) ki’i i-
Third singular (nonhuman) name of thing i-
First plural (exclusive) kam ka-
First plural (inclusive) kit ta-
6.1 Example sentences:
- First-person singular
ak a- dota Bob.
1SG 1SG.AGR strike Bob
‘I struck Bob.’ - Second-person singular
mon gu- dota Bob.
2SG 2SG.AGR strike Bob
‘You struck Bob.’ - Third singular (human)
ki’i i- dota Bob.
3SG 3SG.AGR strike Bob
‘S/he struck Bob.’ - Third singular (nonhuman)
as neka i- dota Bob.
dog that 3SG.AGR strike Bob
‘That dog struck Bob.’ - First plural (exclusive)
kam ka- dota Bob.
1PL.EXC 1PL.EXC.AGR strike Bob
‘We (without you) struck Bob.’ - First plural (inclusive)
kit ta- dota Bob.
1PL.INC 1PL.INC.AGR strike Bob
‘We (including you) struck Bob.’ - Second-person plural
kim gu- dota Bob.
2PL 2PL.AGR strike Bob
‘Y’all struck Bob.’ - Third-person plural
ihi i- dota Bob.
3PL 3PL.AGR strike Bob
‘They struck Bob.’ - Formal
kim gi- dota Bob.
F.PRONOUN F.PRONOUN.AGR strike Bob
‘You struck Bob.’ / ‘s/he struck Bob.’
7 affirmative statements
7.1 Confirmations
Confirmation statements in Sula can be formed with the particle, iyo ‘yes’. If questioned, for example:
One might answer:
- iyo, (kena) neka ak gon
yes (fish) that 1SG belong
‘yes, that (fish) is mine’, - iyo, ak gon
yes 1SG belong
‘yes, mine’
Or commonly, people simply answer: iyo ‘yes’ or ak gon ‘mine’. The following examples demonstrate an additional question and confirmation variation.
- neka kim kena te?
that you.F fish Q?
‘Is that your fish?’ - Person 2 confirmation variations
a. iyo, (neka) ak -gon.
yes (that) 1SG belong
‘Yes, (that) belongs to me.’b. iyo, neka ak kena.
yes that 1SG fish
‘Yes, that’s my fish.’c. iyo, ak kena.
yes 1SG fish
‘Yes, my fish.’d. iyo
yes
‘Yes.’e. ak kena
1SG fish
‘My fish.’
Yes/no questions with verbs are answered similarly.
- Person 1
gu- laka bo uma te?
2SG.AGR walk LOC home Q?
‘Are you going home?’ - Person 2 (a)
iyo laka bo uma.
yes walk LOC home
‘Yes, I’m walking home’Person 2 (b)
laka bo uma.
walk LOC home
‘Walking home’Person 2 (c)
iyo.
yes
‘Yes.’ - Person 1
mon gu- nan bet-ik te?
2SG 2SG.AGR bathe now Q?
‘Are you bathing now?’ - Person 2 (a)
iyo a- nona.
yes 1SG.AGR bathe
‘Yes, I’m bathing’Person 2 (b)
iyo nona.
yes bathing
‘Yes, bathing.’Person 2 (c)
nona.
bathing
‘Bathing.’Person 2 (d)
iyo.
yes
‘Yes.’
Affirmative statements with less certainty use the pahu gan construction. These statements have a meaning similar to it seems in English.
- pahu gan gatel
appear with three
‘there seem to be three’ - pahu gan baba duk moya bet’ik
appear with father come NEG today
‘It seems that father will not come home today.’ - bet’ik pahu gan baba duk moya
today appear with father come NEG
‘It seems that father will not come home today.’
8 negative statements
Non-imperative negative statements are formed with the word moya ‘no’/‘not’.
8.1 Negating verb phrases
moya must come after the entire verb phrase in order to negate the action’s occurrence.
- a- laka bihu.
1SG.AGR walk return
‘I’m returning’ - a- laka bihu moya.
1SG.AGR walk return NEG
‘I’m not returning’ - a- laka moya bihu (a- bena sapeda).
1SG.AGR walk NEG return 1SG.AGR climb bicycle
‘I’m not returning on foot (I’m riding a bicycle)’ - Yanto baha kena damet (neka).
Yanto buy fish yesterday (that)
‘Yanto bought a fish yesterday.’ - Yanto baha kena moya damet (neka).
Yanto buy fish NEG yesterday (that)
‘Yanto did not buy a fish yesterday.’ - *Yanto baha moya kena damet (neka).
Yanto buy NEG fish yesterday (that)
‘Yanto did not buy a fish yesterday.’ - *Yanto baha kena damet moya (neka).
Yanto buy fish yesterday NEG (that)
‘Yanto did not buy a fish yesterday.’
The first asterisked example above is indeed a grammatical sentence, but it has a different meaning than expected. Placing moya immediately after the verb negates the verb alone. So in this case, it would mean, ‘It was not PURCHASING that Yanto did with a fish yesterday’ (e.g. ‘he CAUGHT one’). Likewise, the second asterisked example is grammatical, but here moya negates the time word immediately preceding it. This sentence means ‘YESTERDAY is not when Yanto bought the fish’ (e.g. ‘he bought it TODAY’).
Moya directly follows the sentence element to be negated, so negating preferences is just like negating active verbs.
- ak lal- suka suglela bal.
1SG inner like play ball
‘I like to play ball.’ - ak lal- suka moya suglela bal.
1SG inner like NEG play ball
‘I don’t like to play ball.’
Placing moya between suglela and bal negates only the verb, suglela, rather than the entire verb phrase.
- ak lal- suka suglela moya bal.
1SG inner like play NEG ball
‘It is not playing that I like to do with balls.’
To negate a noun, moya occurs directly after the noun being negated, so by placing moya after bal, the example sentence becomes, It’s not ball that I like to play.
In predicate nominatives, the noun is negated that is on the same side of the equation as moya.
9 commands
As in English, imperative constructions in Sula have an implied subject.
Imperative with null subject
- Ø gaya muamua
you eat everything
‘Eat everything’ - Ø gu- gaya muamua
you 2SG.AGR- eat everything
‘Eat everything’ (the “gu” emphasizes “you”) - Ø gaya choklat muamua
you eat chocolate everything
‘Eat all the chocolate’
Imperatives do not typically cooccur with pronominal prefixation on verbs in the affirmative. Negative commands can mark progressive with pronominal prefixation though.
Negative imperative with null subject and null pronominal prefix
Negative imperative with null subject and pronominal prefix
Additionally, Sula permits a null direct object in double-object constructions.
Imperative with double object
Imperative with null subject and null DO1
Imperative with null subject and null DO2
10 directional system and space
As in English, Sula speakers make use of multiple frames of reference. Intrinsic and relative descriptions pull from the following vocabulary:
Table 45. directional terminology
Sula English
tapa ‘left’
kul ‘right’
samana ‘there (near addressee)’
bangana ‘there (distant)’
(mak) han, (baka)han ‘by’ / ‘near’
yau ‘far’
goga ‘on’
tina ‘on’
lepa ‘on’ / ‘over’
heha ‘under’
mena ‘front’
gabalil ‘around’
lug de lug ‘opposite’ (face to face)
gehi ‘back’ (behind)
manehaneha ‘together’
Additionally, there are two words hifa and hafa/hata that function as directions, but seem to be un-mapable to any constant bearing or direction. They seem to function similarly to the English terms to and fro and, similarly, they seem to occur only in figurative or poetic speech, such as in the folk song, Hai Sua:
- lif (hifa hata) mai -a pil hia moya e
turn (to and fro) but 1SG.AGR see one NEG emphasis
‘I look both directions, but I’m unable to see anybody’
10.1 Intrinsic frame of reference
Object location can be described in relation to a part of another object. Intrinsic frames of reference can be used to locate objects in relation to other objects perceived as having intrinsic sides.
- kii bo uma lug
3SG LOC house front
‘He is in front of the house.’ - komodo bo mon gehi
monitor lizard LOC 2SG back
‘The monitor lizard is behind you.’
10.2 Absolute frame of reference
Sula has a robust vocabulary of fixed bearings. The basic terms are tema ‘inland’, fai ‘seaward’, lepa ‘up’, neu ‘down’. The language also uses the Malay cardinal directions, timur ‘east’, barat ‘west’, utara ‘north’, selatan ‘south’.
The words lepa ‘up’ and neu ‘down’ also have horizontal directional associations, however the cardinal direction mapped to ‘up’ and ‘down’ varies depending where on the islands one is discussing. On the island of Sanana, laka neu corresponds to walking counterclockwise around the island, and laka lepa corresponds to walking clockwise around the island. This means that in the town of Sanana on the Eastern coast, heading neu is north whereas the opposite of this is true across the island in the town of Malbufa. However, along the inhabited southern coast of Mangon island to the north, laka lepa corresponds to walking counterclockwise around the island while laka neu corresponds to walking clockwise around the island, and the cardinal directions they map to are east and west respectively. Unlike in some languages where these would function as relative directions based on the orientation of the speaker while talking, lepa, and neu are fixed cardinal directions in Sula, but the meaning of the directions is opposite depending whether one is talking about locations on Mangon or Sanana.
Figure 21. Absolute directional system
10.3 Relative frame of reference
Relative frames of reference can be used for any situation (even with objects that are perceived as having intrinsic sides).
- kii bo kau lug
3SG LOC tree front
‘He is in front of the tree.’ - mankau neka la bo lida lepa
bird that fly LOC mountain over
‘The bird flies over the mountain.’
10.4 directions of motion
sai maneha neha / laka terus ‘go straight ahead’
lif (ila) tapa ‘turn left’
lif (ila) kul ‘turn right’
lif gehi ‘turn around’
sai gabalil ‘go around’ (‘circumnavigate’)
sai lepa ‘go up’ (e.g. a an inclined road)
bena lepa ‘climb up’ (e.g. a ladder)
sai neu ‘go down’ (e.g. a an inclined road)
bua neu ‘fall down’ (e.g. from a tree)
10.5 Adpositions & spatial relationships
Positional information is structured parallel to Mandarin: locative particle + location NP + adposition. Unlike in Mandarin however, Sula drops position words—adpositions are typically only present for emphasis, in situations where the location contradicts situational expectation, and in situations where there is no corresponding expectation. For example, to express that a pair of shoes is by the door, the statement would be shoes LOC door Ø, while in the less-common scenario that a pair of shoes were atop a door (e.g. hanging by the laces to dry), the construction would be, shoes LOC door on, and in a scenario where, say a new organism was found living on the bottom surface of leaves, an adposition following the NP would also be required, because there would be no prior expectation of the organism’s location. Typically, the locative particle in Sula is situated to the left of the noun phrase and positional information is situated to the right.
The most common positional word is the locative particle bo. It is used to mean roughly at, to, on, or in when there is no inherent ambiguity in the sentence’s meaning; bo is always situated before the NP.
- tasoya ‘banana’ bo basa Sua
say banana (Eng.) in language Sula
‘How do you say “banana” in Sula?’ - Bahasa hapa yang gu- pak bo hai Mangon
Language what REL 2SG speak on island Mangoli
‘Which languages are spoken on Mangole island?’ - da’i kof bo meja
exist coffee on table
‘There is a coffee on the table.’ - kastela deha bo piga lal bo meja
potato stay LOC bowl in LOC table
‘There are potatoes in the bowl on the table.’ - piga deha bo meja
bowl stay LOC table
‘There is a bowl on the table.’ - kastela deha bo piga (lal)
potato stay LOC bowl (in)
‘There are potatoes in the bowl.’
Sula has very few prepositions, but the lexeme, bia may be one. It serves an ablative function similar to ‘from’ in English, and it is situated before the NP.
- bia sahoa
from where
Where (are you) from? - a- laka bia uma
1SG- go from home
I’m coming from home. (e.g. presently) - koi bihu bia Waibao
don’t return from Waibao
Don’t come back from Waibao.
Where there is no contextually-expected location for an object or where the object’s actual location differs from expectation, the spatial information is provided after the noun phrase. In the following examples, he is in front of me, and he is behind me are both logically plausible, and there is no default expectation as to the location. As such, the post nominal position morpheme is necessary.
- ki’i bo ak lug
He LOC 1SG front
‘He is in front of me.’ - ki’i bo ak gehi
He LOC 1SG back
‘He is behind me.’
Complicating Sula locational system is at least one positional morpheme which can occur in the absence of bo and which interestingly, can come either pre or post nominally. Fittingly, this all-around morpheme means ‘around’.
- ak a- laka hai Sua gabalil
I 1SG.AGR- walk island Sula around
‘I walk around Sanana Island.’ - ak a- laka gabalil hai Sua
I 1SG.ARG- walk around island Sula
I walk around Sanana Island.
11 question forming
Yes/no questions are formed using the sentence-final question particle te. te also functions as the conjunction, ‘or’, and when functioning as a question particle, it is an abbreviation of the phrase, te moya ‘or not’.
- mon gu- laka bo Wai Ipa te-moya
You 2SG.AGR walk LOC Wai Ipa or–not
‘Will you go to Wai Ipa or not? - mon gu- laka bo Wai Ipa te
You 2SG.AGR walk LOC Wai Ipa or
‘Are you going to Wai Ipa?
11.1 “wh” words
Generally “wh” question words are in situ, that is they are left in the place of the word being substituted. For example, instead of what are you eating? as English speakers say, Sula speakers say, You are eating what? which parallels the statement, You are eating (bananas).
As in English, the question word can sometimes change places to alternate focus. e.g.: hapa neka ‘whats that’ vs neka hapa ‘that’s WHAT’
11.1.1 hapa ‘what’
11.1.2 bet pila ‘when’
- bet-pila gu laka nan
time-how.much 2SG.AGR walk bathe
‘When will you bathe?” - gu laka nan bet-pila
2SG.AGR walk bathe time-how.much
‘You will bathe when?”
11.1.3 han ‘who’
11.1.4 sahoa ‘where’
Also, saoa, soa, and sibo (Mangon)
11.1.5 bagahoa ‘why’ / ‘how’
Also bagaoa, bagoa, goa bal, and ganoa and ganoki in Mangon.
11.1.6 ga pila ‘how many/much’
Also eb gahoa, eb gaoa, and eb goa.
- matapia ga-pila nib bo mon uma
people how many sit LOC 2SG home
‘How many people live in your home?’
11.1.7 bahoa ‘which’
Also, baoa, boa, (hia baoa ‘which one’).
12 requests and offers
Most commonly, requests are made as imperatives (commands) where the request is implied though not explicitly stated. For example, dok pip ‘give money’ is received as, ‘Can I have some money?’ More polite requests can be formed with dad ‘can’ / ‘may’ and heka which means ‘to want’ in the context of requests (bit and mau are the common forms of ‘to want’ in other contexts).
12.1 Requests with dad
12.1.1 Agree
12.1.2 Refuse
- dad moya
can NEG
‘No, you can’t.’
note: pinjam is a Malay loan, a Sula native term is pak pai, but this is less commonly used, at least among speakers in and around Sanana city.
12.2 Requests with heka
- Kim gi- heka bit laka do ak
2SG.F 2SG.F.AGR want AUX walk with 1SG
‘Would you like to walk with me?’
12.2.1 Agree
12.2.2 Refuse
13 modifying
Sula modifiers occur after the word being modified. They can modify nouns, verbs, and other modifiers. And, complementation is attributive and does not make use of a copula—that is, speakers say, you tall rather than having to say, you ARE tall.
13.1 Nominal modification
- ak suka kof mit
1SG like coffee black
‘I like black coffee’ - ak suka kof mit mota
1SG like coffee black hot
‘I like hot black coffee’
13.2 Verbal modification
- nau neka sena lepayau
cat that jump high
‘That cat jumps high.’ - nau neka sena lepayau bahu-bahu
cat that jump high quick-quick
‘That cat jumps high quickly.’
13.3 Modifier modification
- ki’i =in nap nahu kadiga
3SG =POSS head long strong
‘His/her hair is very long.’ - ak baba kag hebat
1SG father fear serious
‘My father is seriously scary.’
note: there are phrase structuring constraints regarding the ordering of modifiers. One such example can be seen in the following:
- ak bit win kof mit mota
1SG want drink coffee black hot
‘I want to drink hot black coffee.’ - *ak bit win kof mota mit
1SG want drink coffee hot black
‘I want to drink hot black coffee.’ - matapia aya baeo
person big bad
‘big bad person’ - *matapia baeo aya
person bad big
‘big bad person’ - matapia bakatai aya baeo
person dirty big bad
‘big bad dirty person’ - *matapia aya bakatai baeo
person big dirty bad
‘big bad dirty person’
14 Expressing time
14.1 Telling time
Most Sula speakers tell time in Malay, or at least a mixture of Sula and Malay. This is likely because clocks and the concept of a 24-hour day were borrowed from external sources. That said, when elicited, there is a straightforward system of telling time that multiple people have independently produced for me. Sula people use both 12-hour and 24-hour time, and as in English, 12-hour time is ambiguous as to the time of day unless a qualifying word accompanies the time.
Table 46. Time telling vocabulary
Sula English
menit ‘minute’
lea neu ‘afternoon’ / PM
ba’uhi ‘night’
sabatena ‘midnight’
14.1.1 Examples
Swapping lea neu / subur flips the time of day.
1:00 (a.m. or p.m.)
1:30 (a.m. or p.m.)
4:00 (a.m.)
4:00 (p.m.)
4:01 (p.m.)
- befa gareha do menit pakahia lea (neu)
o’clock four and minute first sun down
‘a minute past four p.m.’
4:15 (p.m.)
- befa gareha do menit pakapoa do galima lea (neu)
o’clock four and minute tenth and five sun down
‘fifteen minutes past four p.m.’
4:30 (p.m.)
- befa gareha do menit pakapoa gatel lea (neu)
o’clock four and minute tenth three sun down
‘thirty minutes past four p.m.’ - befa gareha dunnis lea neu
o’clock four half sun down
‘four thirty p.m.’
4:45 (p.m.)
- befa gareha do menit pakapoa
o’clock four and minute tenthgareha do galima lea neu
four and five sun down
‘four forty-five p.m.’ - befa galima kurang menit pakapoa
o’clock five minus minute tenthdo galima lea neu
and five sun down
‘fifteen to five p.m.’
4:54 (p.m.)
- befa gareha do menit pakapoa
o’clock four and minute tenthgalima do gareha lea neu
five and four sun down
‘four fifty-four p.m.’ - befa galima kurang menit ganei lea neu
o’clock five minus minute six sun down
‘six minutes to five p.m.’
14.2 Saying the date
Sula borrowed the Malay names for the twelve months, but Sula uses its words for ‘moon’, fashina, and ‘day’, bet, to mean, ‘month’ and ‘day’. A prototypical date statement is formed, ‘day’ + NUMBER + ‘month’ + MONTH-NAME.
In practice the above structure comes across as terribly specific, so people normally just say: gapit Juni ‘June seventh’.
Table 47. The Sula/Malay Months
Sula/Malay English
Januari January
Februari February
Maret March
Mei May
Juni June
Juli July
Agustus August
Oktober October
Desember December
14.3 Talking about the past and future
Sula does not obligatorily mark past or future tense. A simple sentence like, a-laka bo uma (1SG.AGR-walk LOC house) ‘I walk home’ is ambiguous with respect to tense; it can mean ‘I walked home’, ‘I walk home’, or ‘I will walk home’. This is not as confusing as it might seem though, since context almost always makes it clear when an action takes place, and for the occasions where context does not make it clear, Sula speakers just add the necessary information to the sentence to clear up the ambiguity. For instance, if it is not evident from context that the walking home happened yesterday, a speaker will say, a-laka bo uma dabet ‘I walked home yesterday’.
The temporal word can occur at the beginning or end of the phrase as in the following example.
- (dabet) [ak] (dabet) [[a- dok] [in hal]]
(yesterday) 1SG (yesterday) 1SG.AGR give POSS flower
S V OBJ (dabet) [bo [ak nyaya]] (dabet)
(yesterday) to 1SG mother (yesterday)
OBL
‘I gave a flower to my mother yesterday.’
When used as a modifier, a temporal word can go almost anywhere within the phrase as long as it does not split a compound word, an affix, or break up a sub-phrase, such as the negation phrase in the following example.
- (dabet) [lab (dabet) [sama
(yesterday) [shirt (yesterday) [same(*dabet) moya]] (dabet)
(*yesterday) NEG]] (yesterday)
‘Today’s shirt is different (from yesterdays)’
14.3.1 Future auxiliary construction
English speakers use the auxiliary verb, will, to talk about events in the future. Like in English, Sula can also use an auxiliary to talk about events in the future without specifying precisely when they happen. Sula’s future auxiliary is a secondary function of the verb ‘to want’, bit (or mau in Mangon dialects). Being a close parallel to English, this is a simple construction to learn. However unlike in English, only context can differentiate whether a sentence is a future auxiliary construction and means something will happen or whether the speaker is wanting something.
- ak bit a- laka bo sekolah
1SG AUX/want 1SG.AGR walk to school
‘I want to go to school.’ / ‘I will go to school.’ - mit gamam bit/mau uya
black dark AUX rain
‘It has gotten dark, and it is going to rain.’
14.4 Aspect
Sula uses aspectual particles to describe how an event or state is carried out over time. Aspect is similar to tense except instead of describing when something happened, it specifies the completeness of an action or the speakers temporal view of a situation. Aspect places focus on the endpoint of an action, on the carrying out of an action, on the beginning of an action, or on the event as a whole. For example, in English there is an aspect difference between the following sentences even though they are all in the present tense:
Present tense (no aspect): You go.
Present tense (progressive aspect): You are going.
Present tense (perfect aspect): You have gone.
Present tense (perfect & progressive aspect): I have been eating.
Although it it can take a bit to wrap one’s head around the theoretical differences between aspect and tense, the good news is that it is not necessary to completely understand the difference, because aspect in Sula is easy to master. The speaker just needs to put an aspect word at the end of a sentence clause with a verb.
14.4.1 Examples (with pel ‘already’ / ‘finished’)
- a- laka pel
1SG.AGR walk finished
‘I finished walking.’ - a- laka pel dabet
1SG.AGR walk finished yesterday
‘I finished walking yesterday.’ - a- laka pel dawika
1SG.AGR walk finished tomorrow
‘I will finish walking tomorrow.’ - gaya mua pel
eat all finished
‘eat it all up’ - gaya mua pel dawika
eat all finished tomorrow
‘eat it all up tomorrow’
14.4.2 Sula’s most common aspect words
14.4.2.1 (completive aspect/perfect aspect) pel ‘already’ / ‘finished’
14.4.2.2 (inceptive aspect, commonly directed toward oneself) pai ‘from now’
14.4.2.3 (inceptive aspect, commonly directed toward others) ol ‘from now’
14.4.2.4 (habitual inceptive aspect) son, bihu/pihu ‘again from now’
- gaya son
eat again.from.now
‘Start eating.’ - a- laka pihu
1SG.AGR walk return.from.now
‘I shall return now’
14.4.2.5 (habitual completive aspect) bihu/pihu pel ‘completed again’
14.5 Time situating words
Sula also has non-aspect words that help situate events in time.
14.5.1 ‘never’ pernah moya
- (declarative) ak a- gaya faf ihi pernah moya
1SG 1SG.AGR eat pig meat ever not
‘I never eat pork’ - (imperative)pernah moya gaya faf ihi
ever not eat pig meat
‘Never eat pork’
14.5.2 ‘not yet’ moya da
- ak ana-mehi-nana sai bo Ternate moya da
1SG baby go LOC Ternate never
‘My baby hasn’t been to Ternate yet.’
14.5.3 ‘after’ sonlulu
- sonlulu jam hia, Telkomsel Internet dad
after hour one Telkomsel Internet can
‘In an hour, Telkomsel Internet will work.’
sonlulu pai ‘later’ (Adding the aspect marker, pai to sonlulu forms a phrase meaning ‘later’.)
Table 48. Vocabulary for orienting days
Sula English
bet ‘day’
dawika ‘tomorrow’
dabet hia tuna ‘the day before yesterday’ (yesterday one before)
banai ‘the day after tomorrow’
sonlulu bet gatel ‘in three days’ (after day three)
Table 49. Vocabulary for orienting weeks
Sula English
ahat ‘week’
ahat dawika ‘next week’
ahat banai/ ahat pihu ‘the week after next’
ahat gatel (‘week three’) ‘three weeks from now’
da ahat ‘last week’
(da ahat) hia tuna ‘(last week) one before’ ‘the week before last’
Table 50. Vocabulary for orienting months
Sula English
fasina ‘month’
fasina pihu ‘next month’
fasina gatel ‘three months from now’
da fasina ‘last month’
fasina hia tuna ‘the month before last’
Table 51. Vocabulary for orienting years
Sula English
taun ‘year’
tahun pihu ‘next year’
tahun dabet ‘last year’
Table 52. Additional time vocabulary
Sula English
befa ‘time’
befa ga pila ‘at what time?’
dawika ‘future’
pasa ‘past’
sering ‘often’
manen manen ‘seldom’
tahun paka-poa ‘decade’
15 reflexives
Reflexive constructions describe someone or something doing an action to themselves. For example, Bob hits Bill, is not reflexive, because Bob is doing something to another person. Bob hits himself is a reflexive sentence though, because the action reflects back to Bob. Reflexive constructions in Sula are made using a reflexive pronoun, dit. There is also an auxiliary verb, bal, that is often included. The precise function of the auxiliary is still unknown, but it seems to only be present in active sentences and carry no semantic content.
- ki’i (bal) i- dota dit
3SG AUX 3SG.AGR hit self
‘He hit himself’ - ki’i (bal) i- dota ki’i dit
3SG (AUX) 3SG.AGR hit 3SG self
‘He hit himself’
As shown in (b) above, the pronoun can be repeated, but some speakers claim it must come immediately prior to the reflexive pronoun. In conversation, I have recorded instances of the reverse, but it is unclear if that was speech error, or a strategy such as topicalization that the speakers who reported it as ungrammatical did not have in mind.
- (grammatical) ak (bal) a- dota dit
1SG (AUX) 1SG.AGR hit self
‘I hit myself.’ - (grammatical)ak (bal) a- dota ak dit
1SG (AUX) 1SG.AGR hit 1SG self
‘I hit myself.’ - (grammaticality unclear)*? ak (bal) a- dota dit ak
1SG (AUX) 1SG.AGR hit self 1SG
‘I hit myself.’
Like in most (if not all) languages, it is not grammatical for a pronoun to come before an antecedent. That is, you can say, he hit himself but not himself hit he.
- (grammatical)Aamira (bal) i- dota ki’i dit
Aamira (AUX) 3SG.AGR hit 3SG self
‘Aamira hit herself’ - (grammatical)Aamira (bal) i- dota Aamira dit
Aamira (AUX) 3SG.AGR hit Aamira self
‘Aamira hit herself’ - (grammatical)Aamira (bal) i- dota Aamira
Aamira (AUX) 3SG.AGR hit Aamira
‘Aamira hit herself’ - (grammaticality unclear)*ki’i (bal) i- dota Aamira dit
3SG (AUX) 3SG.AGR hit Aamira self
‘herself hit Aamira.’
When there is another direct object, it must come before the reflexive pronoun.
- (grammatical)ak (bal) a- gut ak nap foa (ak) dit
1SG (AUX) 1SG.AGR scissors 1SG head fur (1SG) self
‘I cut my hair myself.’ - (grammaticality unclear)*ak (bal) a- gut (ak) dit ak nap foa
1SG (AUX) 1SG.AGR scissors (1SG) self 1SG head fur
‘I cut my hair myself.’
Alternatively, the reflexive pronoun can come directly after the subject to emphasize that an action was done alone.
- Yanto bau fota (ki’i) (in) cahana yota
Yanto CAUS tear (3SG) (POSS) pants short
‘Yanto tore his shorts’ - Yanto dit bau fota (ki’i) (in) cahana yota
Yanto self CAUS tear (3SG) (POSS) pants short
‘Yanto himself tore his shorts’ - *ki’i (dit) bau fota (Yanto) (in) cahana yota
3SG self CAUS tear (Yanto) (POSS) pants short
‘He himself tore Yanto’s shorts’
16 comparative constructions and degree
Degree words or phrases are what we use to rate things in relation to other things. A comparative construction is a phrase that measures the similarity or dissimilarity between (usually) two things. A superlative construction identifies the thing that is at the greatest extreme between multiple things. For example, I might say that tap water is colder than tea (comparative), but ice water is the coldest (superlative).
16.1 The same / similar
- O’riorio cap biskuit sama
O’riorio brand cookie samegan Oreo cap biskuit
with Oreo brand cookie
‘O’riorio brand cookies are the same as Oreo brand cookies.’
- ak nyaya sama do ki’i in nyaya
1SG mother same with 3SG POSS mother
‘My mother is the same as her/his mother.’
16.2 Comparisons
16.3 Preferences (more)
16.4 Preferences (less)
- ak lal suka (pasa) teh kurang bia kof
1SG inside like (pasa) tea less of coffee
‘I like tea less than coffee.’
16.5 Degree / amount
Sula uses several degree words to make degree statements. Some of these like eb, foloi, and totoya which mean, ‘much’ / ‘many’, are adverbial words. They are different from verbs, because they cannot take the subject prefixes, and they are different from adjectives in their meaning. Whereas adjectives describe properties of nouns, these words only identify the amount of a noun, verb, or adjective. There is no distinction between countable and mass nouns in Sula, so these words have a meaning including both ‘much’ and ‘many’.
16.5.1 eb
- dahii pip eb
explicative money much
‘Look there, it’s a lot of money.’ - dahi nui eb
exist coconuts much
‘There are a lot of coconuts.’
Additionally, there are words like kadiga ‘strong’ which can function as adjectives, or adverbs.
16.5.2 adjectival function of kadiga
16.5.3 adverbial function of kadiga
16.5.4 sequential adverbial and adjectival function of kadiga
Sometimes reduplication also indicates degree, but this does not seem to be grammatically predictable, and the words that are reduplicated and their manner of reduplication must be learned on a word-for-word basis.
baha ‘spicy’ (hot)
For added emphasis, the reduplicated adjective can be used with one or more degree words. (See the section on reduplication for more.)
baha-ha foloi
hot much
‘very hot’
Even degree words can be reduplicated for emphasis.
- baha-ha kadiga-diga
hot-hot strong-strong
‘very very very hot’ - baha-ha eb-eb
hot-hot very-very
‘very very very hot’ - baha-ha eb kadiga-diga
hot-hot very strong-strong
‘very very very very hot’ - baha-ha eb-eb kadiga
hot-hot very-very strong
’very very very very hot’
Speakers reported it to be ungrammatical to reduplicate both degree words.
Although eb before kadiga is acceptable, kadiga before eb is not.
16.6 None, all
- (ak) a- gaya biskuit moya
(1SG) 1SG.AGR eat cookies NEG
‘I did not eat the cookies.’ - (ak) a- gaya biskuit gapila (para)
(1SG) 1SG.AGR eat cookies some (just)
‘I ate some of the cookies.’ - (ak) a- gaya biskuit mua-mua (pel)
(1SG) 1SG.AGR eat cookies all-all (ASPECT)
‘I ate all the cookies.’
17 listing: events, directions, instructions
Listing is typically expressed with a pitch raise, vowel lengthening, and a pause between items. The last item in the list takes a lowered intonation, indicating it is the final item. If however a speaker wants to imply that the list is ongoing, s/he will not lower intonation on the final item.
17.1 complete list
- ak lal suka fua gatel
1SG inner like fruit threepara: weuw↑ fia↑ dahia↓
only mango banana durian
‘I only like three fruits: mangoes, bananas, and durian.’
17.2 truncated list
- ak lal suka fua mua-mua:
1SG inner like fruit everyweuw↑ apel↑ fia↑ dahia↑
mango apple banana durian
‘I like all fruit: mangoes, apples, bananas, durian (etc.)’
It is grammatical to insert do ‘and’ between any or all of the list items, but this is not typically done.
18 Indicating possession
18.1 Word order determined
Although Sula does have a possessive marker, it is hard to come up with a rule for how and when to use it. Most of the time, possession is indicated by word order: the possessor comes before the object of possession. This is also the primary pattern for English but the opposite of the Indonesian pattern. In fact, English and Sula are both somewhat atypical in this regard, because it is more common for languages whose verbs (V) come before noun phrases (NP) to have possessors (P) come after the possessed noun phrase (e.g. Comrie 1989).
verbs come before objects:
(Sula)
(Indonesian)
(English)
possessed objects come after possessors:
(Sula)
(Indonesian)
- bapak saya
NP P
my father
(English)
- my father
P NP
my father
- 18.2 Possessive marker in
Sula has a linker, =in, which often functions to indicate possession. The rules determining when this marker is or is not used have been difficult to pin down and will require more data. I have been able to determine from various consultants’ contributions that the =in marker is used primarily (or exclusively) with third-person possessors.
18.2.1 First-person possessors do not permit =in
(good)
ak nyaya
1SG mother
‘my mother’
18.2.2 Second-person possessors do not permit =in
- (good)
mon nyaya
2SG mother
‘your mother’ - (ungrammatical)
*mon =in nyaya
2SG =POSS mother
‘your mother’
18.2.3 Formal pronoun second-person possessors probably do not permit =in
- (good)
kim nyaya
2SG.F mother
‘your mother’ - (ungrammatical)
*?kim =in nyaya
2SG.F =POSS mother
‘your mother’
18.2.4 Third-person singular and plural possessors sometimes require =in
singular
- (ungrammatical)
*ki’i nyaya
3SG mother
‘her/his mother’ - (good)
ki’i =in nyaya
3SG =POSS mother
‘her/his mother’
name
- (ungrammatical)
*Mustafa nyaya
Mustafa mother
‘Mustafa’s mother’ - (good)
Mustafa =in nyaya
Mustafa =POSS mother
‘Mustafa’s mother’
object
plural
(good)
ihi =in nyaya
they =POSS mother
- It also appears that second-person plural possessors probably do not permit =in, but some speakers do not determine it ungrammatical.
exclusive
inclusive
- (ungrammatical) *?kit =in nyaya
- Many related languages have an alienability contrast in their possessive grammar, where items (such as body parts and family members) are indicated differently from items whose possession can be easily transferred. Some degree of an alienability contrast is likely present in Sula too, but it has been elusive if so, because speakers vary with regard to the boundary between alienable and inalienable. The following examples show that distribution of =in probably does not relate to alienability.
- *ak =in pip
1SG =POSS money
‘my money’ - ak a- gawai ak tena bag
1SG 1SG.AGR scratch 1SG stomach thick
‘I scratch my fat stomach.’
- *ak a- gawai ak =in tena bag
1SG 1SG.AGR scratch 1SG =POSS stomach thick
‘I scratch my fat stomach.’ - a- laka do ak yai bal bit laka
1SG.AGR walk and 1SG leg AUX want walk
‘I go where my legs take me.’ - *a- laka do ak =in yai bal bit laka
1SG.AGR walk and 1SG =POSS leg AUX want walk
‘I go where my legs take me.’
mon pip
2SG money
‘your money’
*ki’i pip
3SG money
‘my money’
Possession in general is seldom used figuratively (i.e. with non-concrete possessors).
- *ika bet’ik lab
this today shirt
‘this is today’s shirt’ - *ika bet’ik =in lab
this today =POSS shirt
‘this is today’s shirt’
These same concepts are instead expressed without possessive constructions.
- dabet lab sama moya gan (lab) bet’ik
yesterday shirt same NEG with (shirt) today
‘Yesterday, the shirt is different from today.’
The concept can also be expressed using the non-concrete possessor as a modifier.
- ika lab bet’ik
this shirt today
‘this shirt today.’ - lab dabet sama moya gan lab bet’ik
shirt yesterday same NEG with shirt today
‘The yesterday shirt is different from the today shirt.’
Some consultants indicate that in is acceptable between ‘yesterday’ and ‘shirt’ in example (9), but they explain that in does not indicate ‘belonging’ in this situation. Rather, it seems to be functioning in its other role—as a syntactic explicative prefix (EXP) that is empty in meaning but sets focus in the sentence.
- *?dabet in- lab sama moya gan lab bet’ik
yesterday EXP shirt same NEG with shirt today
‘Yesterday, the shirt is different from the today shirt.’
Likewise, the same explicative function of in- is occasionally used in possessive constructions with non-third-person possessors, but these sentences are judged as less correct, and they usually include a pause before in-.
first person
(best) ak pat
1SG smell
‘my smell’
third person
- (less correct) ki’i pat
3SG smell
‘his smell’ - (less correct) ki’i… in- pat
3SG EXP smell
‘his, the smell’ - (best) ki’i =in pat
1SG =POSS smell
‘his smell’
It seems likely that the =in marker is a vestige from a genitive system that had been present at an earlier time (perhaps in Proto Central Maluku). In present day Sula its possessive use seems to have disappeared from the language for all but third-person possessors. In other words, Sula only sometimes allows the =in marker, and it is only when the possessor is he/she/it and concrete rather than abstract. In such cases, =in marks a possessor.
=in cannot be used to mark an agent.
- *fina neka =in bau munara
woman that =POSS CAUS work
‘that woman’s working’
=in cannot be used to mark a theme.
=in cannot be used to mark a time.
For the above examples, Sula instead uses simple sentences:
- fina neka bau munara ‘that woman is working’
fina neka bal mata ‘that woman dies’
There is a possessive in both Sula and English that makes use of the word, of. Think of the sentence, you are not the boss of me. As compared to my boss, the possessor has been demoted to an oblique phrase. Whereas some find the periphrastic passive nonstandard in English, this construction is accepted in Sula.
18.4 Possessive questions
Possessive questions can be formed three ways: with rising intonation; by substituting han ‘who’ in place of a human possessor; or by adding the question phrase, te (moya) ‘or (not)’ at the end of the sentence.
han pena?
who pen
‘Whose pen is it?’
To form a question, han can be substituted in place of any human possessor in a statement—both subject and object alike.
non-question, first person
non-question, third person
- [ki’i [i- gawai [ki’i =in tena]]].
[1SG [1SG.AGR scratch [1SG =POSS stomach]]]
S V OBJ
‘s/he scratched her/his stomach.’
question, subject (null prefix)
- [han [Ø- gawai [ki’i =in tena]]]?
[who [ scratch [1SG =POSS stomach]]]
‘Who scratched her/his stomach?’
question, object (null possessive marker)
question, subject and object (null prefix and possessive markers)
18.5 gon constructions
Sula speakers can also express possession with statements of belonging. These are formed with a word meaning ‘belong’ that has to forms, =gon and =non. =gon is used with first-person possessors, and =non is used with all others. =non is probably cognate with the Buru possessive forms, nang / nan / nam, but Sula lacks the complex pattern of inflection found in Buru.
- ak tonka neka
1SG stick that
‘That is my stick.’ - ak =gon, tonka neka
1SG =belong stick that
‘It is my belonging, that stick.’ - tonka neka ak =gon
stick that 1SG =belong
‘That stick belongs to me.’ - dok ak patu
give 1SG hoe
‘Give me my hoe.’ - dok ak =gon patu
give 1SG =belong hoe
‘Give me the hoe that belongs to me.’ - dok mon =non patu
give 2SG =belong hoe
‘Give me the hoe that belongs to you.’ - patu iki ak =gon
hoe this 1SG =belong
‘This hoe belongs to me.’
When forming questions, han precedes =gon/=non. Either =gon or =non is reportedly grammatical, but, =non always seems to be produced in questions. This is likely because from a logical standpoint, asking with =gon would be answering the question within the question (i.e. I cannot be asking who something belongs to, if I am identifying the notion of belonging with a word that assigns the object to myself).
- han =non tongka neka?
who =belong.1SG stick that
‘It is whose belonging, that stick.’ - *non han tongka neka?
belong who stick that
‘It is whose belonging, that stick.’ - tongka neka han =non?
stick that who =belong
‘This stick belongs to who.’ - *tongka neka non han?
stick that belong who
‘This stick belongs to who.’
19 reference switching through pronouns and pronominal prefixes
Sula does not have a full switch-reference system, but verbal prefixes are used to disambiguate and switch reference. The mechanics of reference switching are not yet fully understood, but it is clear that a pronominal prefix in the absence of a pronoun can indicate a change in subject.
There are two basic structures for transitive sentences. The prototypical citation form is SUBJECT + PREFIX + VERB + OBJECT
- ki’i i- dota ak
3SG 3SG.AGR- strike 1SG
SUBJECT PREFIX VERB OBJECT
‘He hits me.’ - ak a- dota ki’i
1SG 1SG.AGR- strike 3SG
SUBJECT PREFIX VERB OBJECT
‘I hit him.’
In the first structure, the subject is overt, and the verb is marked to agree with the subject. This prototypical structure is typically chosen to introduce or transition between subjects. Thereafter, pronoun drop is common, as in the following examples.
- Ø i- gaya fia
Ø 3SG.AGR- eat banana
SUBJECT PREFIX VERB OBJECT
‘He eats banana.’ - Ø a- gaya fia
Ø 1SG.AGR- eat banana
SUBJECT PREFIX VERB OBJECT
‘I eat banana.’
Since the pronominal prefix inherently carries subject information, switching pronominal prefixes indicates a change in subject even without restating an overt subject (proper noun or pronoun). In the following sentence, the third person’s verbal prefix, i- is present in the absence of the third-person pronoun. This indicates a switch to a third-person subject. This sentence would take place in a conversation where the third person was previously overtly identified (i.e. “he” had already ben named).
Similarly in the following example, the first-person’ verbal prefix, a- is present in the absence of the first-person pronoun, indicating a shift to a first-person subject.
- ki’i i- mata mai a- pusi moya
3SG 3SG.AGR die COMP 1SG.AGR headache NEG
‘That he died, doesn’t concern me.’
In the following example, the subject alternates from third person to first person and back to third person, and switching is indicated by marking first-person or third-person pronouns on verbs.
- Yanto nau, bahwa a- duk, tapi
Yanto know COMP 1SG.AGR come but
‘Yanto knows, that I’m coming, buta- nau moya, (bahwa)
1SG.AGR know NEG (COMP)
I don’t know whetheri- duk atau moya
3SG.AGR come or NEG
he will come or not.’
20 Determiners
Sula possesses the demonstrative determiners iki/ika and neka which are roughly equivalent to ‘this’ and ‘that’ respectively. Determiners come to the right of NPs.
20.1 Demonstrative determiner iki/ika
20.2 Demonstrative determiner neka
20.3 Determiner word order
- Ikbal i- gaya man ika
Ikbal 3SG.AGR eat chicken this
‘Ikbal is eating this chicken.’ - *Ikbal i- gaya ika man
Ikbal 3SG.AGR eat this chicken
‘Ikbal is eating this chicken.’
Additionally, iki/ika and neka have a discourse-marking function. In these constructions, the words bear little to no semantic weight but rather serve to bind a statement contextually. Instances of discourse deixis are syntactically optional.
20.4 Discourse marking
- Yanto baha kena moya damet (neka).
Yanto buy fish NEG yesterday (that).
‘Yanto did not buy a fish yesterday (is what it is).’
21 Passives
Sula appears to make use of a paraphrastic passive construction in which the subject is demoted to oblique and the direct object becomes the new subject. However while oblique agents can often be omitted, they appear to be mandatory in Sula.
21.1 Passive construction with ‘punish’
- (active)
Yanto =in baba naha ki’i
Yanto =POSS father punish him
‘Yanto’s father punished him.’ - (passive)
Yanto dahi naha bia =in baba
Yanto receive punish OBL =POSS father
‘Yanto was punished by his father.’
21.2 Passive with ‘punish’ & null oblique?
22 Morpheme and word classes
22.1 verbs & nouns
While there are generally clear semantic differences between verbal and nominal words, there is no good evidence that a grammatical distinction exists defining separate classes for each.
Any word can be a verb, if it makes logical sense. For example in the section on reflexives, we had an example that used the ‘scissors’ as a verb. gut ‘scissors’ is semantically nominal (most likely a loan from the Indonesian noun, gunting). However, when functioning as a verb, it can take the pronominal prefixes.
Likewise, words that are semantically verbal can function seamlessly as nouns without any derivational morphology.
- (verbal) a- mata moya
1SG.AGR die NEG
‘I won’t die.’ - (nominal) a- kag ak mata
1SG.AGR fear 1SG death
‘I fear my death.’ - (verbal) kalo dahi moya, Alif i-
if exist NEG Alif 3SG.AGRnoi tabak bia (ki’i) =in ponoida
request tobacco from (3SG) =POSS friend
‘If Alif doesn’t have any tobacco, he requests it from his friends.’ - (nominal) Alif =in noi bau naha
Alif =POSS request CAUS anger(ki’i) =in ponoida
(3SG) =POSS friend
‘Alif’s asking irritates his friends.’
The ease by which semantically nominal and verbal words can function in either role challenges the usefulness of analyzing nouns and verbs as independent word classes. Syntactically speaking, these words are not inherently nominal or verbal, but they can function in either role. As stated regarding adjectival modifiers in section 12: they are their own thing, and do not fit neatly into pre-existing categories.
22.2 modifiers
Sula has both adjectival modifiers and adverbial modifiers. Adverbial modifiers are distinct from adjectival modifiers in that they cannot take pronominal prefixes. (see section “15 comparative constructions and degree” for a description)
22.3 pronouns
Pronouns are surprisingly difficult to nail down in Sula. There seems to be some pronoun, but it is unclear if this is due to community variation or individuals overthinking elicitation questions and confusing themselves. The precise meaning and usage of the formal pronoun kim is especially tricky as some speakers claim it to be only a second-person singular or plural pronoun, some claim it to be used for second and third-person singular or plural, and others claim it to be person independent—i.e. available for second and third-person pronouns, singular and plural, and, in the case of one consultant, even first-person plural pronouns as long as the group includes a member older than the speaker. As a practical matter, I did not observe the pronoun used in this way, but the possibility exists that (at least for some speakers), this is the case. There is also significant variation in pronoun pronunciation, particularly with the third-person singular pronoun ki’i.
It is also difficult to independently elicit an identical paradigm from any two people. That said, I elicited the pronouns and agreement markers from numerous people, compared the lists, and used trial and error during targeted elicitations. Of course I also carefully observed natural conversation samples to parse the paradigm that (as mentioned earlier) I used as the basis for a fairly intensive elicitation and grammaticality judgment session with a group of consultants who evaluated my determinations. There is likely more to the story, but I am fairly confident that the following table represents the pronouns and their verbal prefixes as they are most widely used on Sanana.
Table 53. Pronouns & Pronominal prefixes
pronounsPronominal prefixes
Second-person singular mon gu-
Third singular (human) ki’i i-
First plural (exclusive) kam ka-
First plural (inclusive) kit ta-
22.4 demonstratives
neka ‘that’
22.5 locatives
bia ablative
bo dative
22.6 deictic locatives
isuka, isuna, saiki, saik ‘here’
bangana, neka ‘there (near addressee)’
samana ‘over yonder’
22.7 interrogatives
sahoa, sa’oa, soa, subo ‘where’
bagahoa, bahoa,
baga’oa, bagoa, bagano ‘why’ / ‘how’
bit pila, betu pila,
bet pila, bet pila hoa ‘when’
tiba, hia bahoa,
hia baoa, hia boa ‘which’
22.8 numerals
(see section 3 counting)
22.9 numeral classifier (counting word)
Sula seems to have just one counting word, fat(u), which derived from the word ‘stone’. As there is just one, it would be incorrect to call this word a classifier, but grammatically it functions similar to a classifier in other languages. fat is used whenever a quantity of objects is stated unless it can be replaced with a more specific term. For instance, one cannot say (a) one cooked rice, but one can say, (b) one fat of cooked rice, or (c) one plate of cooked rice. Although plate replaces fat in the third option, it is not a counting word; the word plate itself must be counted using fat (d).
*bird hia
rice one
‘one cooked rice’
piga fat hia
plae CLF one
‘one plate’
- Table 54. locational adpositions
Sula English
gehi ‘behind’
lepa ‘over’ / ‘up’
neu ‘down’
heha ‘under’
lug ‘across from’ / ‘opposite’
do ‘with’ / ‘near’
Table 55. conjunctions
Sula English
do ‘and’
te ‘or’
mai ‘but’
tapi ‘but’
son ‘because’
Table 56. particles
Sula English
pel ‘already’ / ‘finished’
pai ‘from now’ (said about oneself)
ol ‘from now’ (said about others)
Table 57 clitics
Sula English
=in POSSESSIVE
=gon ‘belongs to’ (first-person possessors)
=non ‘belongs to’ (non-first-person possessors)
22.10 causative prefixes
Sula has three common causative prefixes. While they are in complimentary distribution, I have been unable to identify any rule to describe the conditions for when one prefix is chosen over the others.
Table 58. causative prefixes and examples
prefix example word gloss English
bau- bau-mata make-die to kill
bau-munara make-labor to work
bau-pia make-safe to make peace
baka- baka-nau make-know to message
baka-baret make-stop to stop
baka-toya make-tell to point, to show
mak- mak-dahi make-exist to meet
mak-bobai make-significant other to date
22.11 other prefixes (some likely fossilized)
As with the causative prefixes, I have been unable to discover any semantic or phonological conditions that explain the difference in how Sula’s prefixes are used. Whereas I suspect some of the causative prefixes could be fossilized onto their forms, I am inclined to believe interesting but elusive grammatical processes are at play among the other prefixes. This topic should be a high priority for future study.
Table 59. other verbal prefixes and examples
prefix example word gloss English
ba- ba-lea dit ?-dry self to dry oneself
ba-map ?-? to cook
ba-lela ?-? to laugh
ta- ta-pisnoya ?-speak to speak
ta-bicara ?-speak to speak
ta-toya ?-say truth
ta-linga ?-ear to listen
ma- ma-lia ?-? to swallow
ma-hara ?-hate to hate
ma-hoka ?-dull dull
pa- pa-laka ?-walk journey
pa-nau ?-know funeral
pa-masi ?-(nasi ID) rice
mana- mana-yop ?-kiss kiss
mana-pan ?-intercourse sex
mana-tota ?-hit fight
22.12 auxiliaries and modality
Auxiliary verbs in Sula can occur independently as verbs in their own right, and they can work in concert with other verbs to help indicate the amount that an action is certain, permitted, or desired. Additionally, modality helps to differentiate statements of conviction from measured claims (e.g. Cameron 2007). Further research will likely reveal more specific syntactic classifications for the words in the following table, But at this early stage in Sula language research, it suffices to simply treat these words as auxiliaries.
Table 60. auxiliary verbs and modals
Sula English
awa still
bal must
bit want, shall, will, should
dad can
lela enjoy
mai dad could, may, might
mau, mot mau want, shall, will, should
mot follow
musti must
suka like
Table 61. other, non-modal auxiliary verbs
Sula English
mau future AUX (Mangon)
dahi existential verb or AUX
bal active sentence AUX
mat possible active sentence AUX
(also a contraction of matua ‘old’ as in matnana ‘elder’)
paka /
pake /
pakai AUX meaning ‘to use’
23 Noun phrases
As was shown in section 2.1.1, modifiers follow the head noun in Sula. This is the expected pattern in SVO languages. One way to differentiate noun modification from compound words, is that the order of stems (conjuncts) in compounds is the opposite of this order. For example, a Sula speaker would describe a tree as large with the phrase, kau aya (‘tree big’), but the word for banana tree is, fia kau (‘banana tree’) rather than *kau fia. Likewise, one would describe long fur as, foa nahu (‘fur long’), but the word for a person’s hair is, nap foa (‘head hair’) rather than *foa nap.
Noun phrases can be grouped into attributively modified nouns, descriptively modified nouns, verbally modified nouns, quantifier modified nouns. Attributively modified nouns are modified by words that are semantically adjectival (these words can take pronominal prefixation morphology). Descriptively modified nouns are modified by words that are semantically nominal (these words do not take pronominal prefixation morphology). Verbally modified nouns are followed by a verb without pronominal prefixation morphology. Quantifier modified nouns are followed by a degree word or a quantifier.
23.1 examples basic structure
- matapia gatel nib bo sahoa
[person three] sit LOC there
‘Three people are sitting over there.’ - ak bau mata ak sap aya bet-ik
1SG make dead 1SG [cow large] today
‘I’ll kill my large cow today.’
23.2 examples attributive modification
maana birahi
man handsome
‘handsome man’
matapia-bau-awa
person-make-field
‘farmer’
- ki’i i- patana mataia gapit
3SG 3SG.AGR pay person seven
‘he paid seven people’ - ak bit win teh eb
1SG want drink tea much
‘I want to drink a lot of tea’ - (generalized quantifier)
nui galima para
coconut five only
‘only five coconuts’
24 Verbs
24.1 Alignment
There is significant intra-speaker variation, and it will require a targeted study to confidently and comprehensively determine the rules governing verbal alignment in Sula. For the purpose of this section, I primarily analyzed the speech of an individual consultant, Ismael Duila, of Waibau village. The analysis is based primarily on Mithun (2008), O’Grady (2010), and also referenced for areal comparisons was Holton (2008).
The data show Sula to have a split pronominal prefixation system: first-person subjects follow an Active–Stative classification like that proposed by Dixon (1979), but intransitive verbal marking is split along agentive lines rather than strictly along unergative/unaccusative lines. non-first-person, human arguments follow an Ergative–absolutive system, and non-first-person, non-human arguments trigger no pronominal prefixation. Evidence of this split is revealed in the presence/non-presence of pronominal prefixes.
In its basic, non-aspect-marked form, the first-person subject’s verbal marker is present in all transitive constructions. Its presence or non-presence in intransitive constructions however depends whether the verb takes an agentive subject (one that is volitional); stative verbs are not marked. Precisely pinpointing how the system works, is complicated because Sula’s pronominal prefixes serve alternative functions as well: along with indexing pronouns on verbs, the markers can indicate progressive aspect as well as disambiguating and changing of reference.
The consultant produces a consistent pattern of alignment, and it seems to conform with the patterns of many other Sula speakers. Whether this pattern represents the language as a whole is an open question. The pattern shows a system of split alignment and animacy, where first-person subjects follow an Active–Stative classification in which transitive verbs and agentive intransitive verbs are marked for pronominal prefixation. non-first-person, human subjects are marked along ergative–absolutive lines, and verbs are not marked to agree with non-human subjects.
Table 62. Subject and agreement markers
basic alignment, with no aspectual or reference functions
1st 2nd 3rd Non-Human
Subject of transitive verb a- gu- i- Ø-
Subject of agentive intransitive verb a- Ø- Ø- Ø-
Subject of non-agentive intransitive verb Ø- Ø- Ø- Ø-
Direct Object Ø- Ø- Ø- Ø-
24.2 Transitive verbs are marked to agree with human subjects
In basic, non-aspect-marked form, the first-person subject’s verbal marker is present in transitive constructions. Its presence or non-presence in intransitive constructions depends on whether the verb takes an agentive subject; stative verbs are not marked for prefixation. Precisely pinpointing how the system works is complicated because Sula’s pronominal prefixes serve alternative functions: markers are used both to indicate progressive aspect and to disambiguate reference.
First person
Non-first person
Non-human
as uka Yanto
dog bite Yanto
‘The dog bites Yanto.’
- Intransitive verbs are marked to agree with agentive first-person subjects but not with non-first-person subjects (except when indicating progressive aspect or disambiguating reference in discourse when an overt pronoun is not present)
First person
Non-first person
Non-human
Intransitive verbs are typically not marked to agree with non-agentive subjects.
First person
ak bua bia kau
I fall LOC tree
‘I fall from the tree.’
- Non-first person
Non-human
Objects are grammatically linked by word order; there is no object marking on verbs.
First person
Non-first person
ak a- dota mon
I 1SG.AGR- strike you
‘I hit you.’
- Non-human
24.3 Aspectual function of pronominal prefixes
As discussed in section 13 Expressing time, Sula has aspect markers that indicate the completeness of an action.
unmarked
completive
In addition to marking aspect with overt aspect markers, progressive aspect can be conveyed with a secondary function of the language’s subject prefixes. This aspectual function of pronominal prefixation complicates deciphering Sula’s alignment system, because when marking progressive aspect, non-first-person subjects are also marked on intransitive verbs. This occurs across verb classes.
Transitive verbs with overt objects are reportedly most grammatical with pronominal prefixes:
Direct object, pronominal prefix:
- Ikbal i- gaya man ika
Ikbal 3SG.AGR– eat chicken this
‘Ikbal eats this chicken.’/’Ikbal is eating this chicken.’
Pronominal prefixes are optional when direct objects are not overt. In these sentences, subject-marked verbs convey progressive aspect:
Without prefixes, no aspectual information is conveyed:
Ikbal gaya
Ikbal eat
‘Ikbal eats.’
- With prefixes, progressive aspectual information is conveyed:
Ikbal i- gaya
Ikbal 3SG.AGR– eat
‘Ikbal is eating.’
- Intransitive, unergative verbs with non-first-person subjects:
Without prefixes, no aspectual information is conveyed:
Yanto nona (tetapi) matapia-basa baumunara.
Yanto sleep (but) people-other work
‘Yanto sleeps while others work.’
With prefixes, progressive aspectual information is conveyed:
Intransitive, unaccusative verbs with non-first-person subjects:
Without prefixes, no aspectual information is conveyed:
With prefixes, progressive aspectual information is conveyed:
25 Causatives
Causatives are formed with the prefixes bau-, baka-, and mak- (see also section 22.10). There is no predictable difference in distribution between the prefixes, and in fact sometimes multiple prefixes can be chosen to form words with different meanings. No subtle semantic differences have thus far been identified, but there are definite word-specific differences in distribution. bau– is found in the most forms and seems to be most productive:
bau-mata ‘kill, murder’ (lit. ‘make dead’)
- *baka-mata ‘kill, murder’ (lit. ‘make dead’)
- *mak-mata ‘kill, murder’ (lit. ‘make dead’)
- bau-munara ‘to work’
- *mak-munara ‘to work’
- baka-toya ‘point, show’
- *bau-toya ‘point, show’
- *mak-toya ‘point, show’
- Some words are able accept multiple prefixes to affect the same meaning:
bau-gehi ‘construction’ (lit. ‘make standing’)
- baka-gehi ‘construction’ (lit. ‘make standing’)
- *mak-gehi ‘construction’ (lit. ‘make standing’)
- Some words are able accept multiple prefixes to affect the different meanings:
- baka-nau ‘to message’ (direct digital messaging or conveying a message on another’s behalf) (lit. ‘make known’)
- *bau-nau ‘to message’ (direct digital messaging or conveying a message on another’s behalf) (lit. ‘make known’)
mak-nau ‘to introduce’ (lit. ‘make known’)
- Causative constructions are straightforward, simple sentences with a causative prefix on the verb where subject prefixes would otherwise go.
25.1 non-causative construction
- ana-mehi-nana mena
child-small-small cry
‘The baby cried.’
25.2 causative construction with bau-
- nyaya bau- mena ana-mehi-nana
mother make cry child-small-small
‘The mother made the baby cry.’ - Yanto bau- nona ana-mehi-nana
Yanto make sleep child-small-small
‘Yanto made the baby sleep.’
25.3 causative construction with baka-
- nyaya baka- mena ana-mehi-nana
mother make cry child-small-small
‘The mother made the baby cry.’ - Yanto baka- nona ana-mehi-nana
Yanto make sleep child-small-small
‘Yanto made the baby sleep.’
There are also semantic causatives—constructions whose verb is inherently causative and thus do not require a causative prefix.
- Yanto soya ana-mehi-maana moma as
Yanto tell child-small-male hit dog
‘Yanto told the boy to hit the dog.’ - Yanto soya ana-mehi-maana i- moma as
Yanto tell child-small-male 3SG.AGR hit dog
‘Yanto told that the boy hit the dog.’
26 Reduplication
This section is a preliminary investigation into Sula’s system of reduplication. Reduplication is a process in which a word or part of a word is repeated in succession. While this section is not comprehensive or conclusive, it provides groundwork for future studies into the topic at greater depth. I was unable to conduct a detailed survey of the language’s reduplication patterns, but by comparing field notes against Charles Grimes’s (1992) thorough analysis of the related Buru language, it was still possible to identify several patterns in Sula and their likely functions.
26.1 Types of reduplication present in Buru
Grimes (1992) identifies the following reduplication types in Buru: (1) phrasal duplication, (2) multiple repetition, (3) full reduplication of stem, (4) reduplication of root, (5) reduplication of first CVC, (6) reduplication of the first CV, (7) reduplication of the first C. Grimes also states that phrasal question words become indefinite when reduplicated. (Grimes 1992: p. 130)
26.2 Phrasal reduplication
Although a handful of instances were identified where a phrase is successively repeated, there is no evidence as to the grammatical affect in Sula.
- nona-gega nona-gega wakdab hia para
sleep-wake sleep-wake, miss one just
‘I sleep and wake and sleep and wake; I just miss someone.’
26.3 Reduplication of Verbs
Grimes draws a distinction between active and non-active verbs in his analysis of Buru. The active category includes verbs capable of taking an actor subject, whereas the non-active category includes verbs that take only one core argument, and whose subjects must be in the role of undergoer. Active verbs are further split into transitive and intransitive, the intransitive of which have only one core argument and can optionally take an undergoer subject.
The transitive active verbs have two core arguments, the subject of which being restricted to the actor role. According to Grimes, Buru has no canonical adjectives, and all attributive modifiers are derived of verbs. Although this grammatical sketch does not analyze Sula verbs the same way, for the purposes of making a parallel comparison, this section follows Grimes’s analysis.
26.4 Reduplication of active verbs
As for multiple repetition: Grimes (1992) describes reduplication of active verbs as producing an iterative aspect wherein the duration of the aspectual quality is contextually based. Though far from conclusive, this analysis appears to apply to reduplicated active verbs in Sula as well.
26.5 active intransitive
i- laka
3SG.AGR walk
‘S/he walks.’
i- noi
3SG.AGR ask
‘S/he asks.’
26.6 active transitive
- mon gu- dota ak bib
2SG 2SG.AGR strike 1SG goat
‘You hit my goat.’ - mon gu- dota-dota ak bib
2SG 2SG.AGR strike-strike 1SG goat
‘You keep hitting my goat.’
26.7 Reduplicated modifying, non-active verbs (adverbs of manner)
Grimes analyzes Buru adverbs as derivational products of underlyingly non-active verbs. Sula does not have derivational morphology for adverbs, but Sula adverbs can be reduplicated, as shown below in the modification: verb2 ➞ verb example. The following example might represent an instance of verbal modification by a reduplicated noun, but this is uncertain. lea-lea has been translated as ‘slowly’/’leisurely’/’calmly’, but the non-reduplicated form, lea is translated as a nominal ‘sun’ and ‘place’ (in dialects that have dropped the intervocalic h in leha). It is unknown whether lea-lea is derived from either noun or if it is an independent lexeme. The modification: noun2 ➞ verb example below represents verbal modification by a reduplicated base that seems to already function as a modifier prior to reduplication, and which seems unable to function as a main sentence verb (further data collection will be necessary to verify these observations).
[modification: verb2 ➞ verb]:
[modification: noun2 ➞ verb]:
modification:
[modifier ➞ verb]
[modifier2 ➞ verb]
[without verb]
*koi bahu
NEG.IMP fast
‘Don’t be quick.’
According to Grimes, when non-active verbs are not modifying another verb but are themselves the main sentence verb, a reduplication process “indicates intensity.” When the reduplicated word is the sentence’s main verb, the intensity is intrinsic, whereas the intensity is transferred to the main verb when the reduplicated word is in modifier rather than primary-verbal position. (following ex. re-glossed from 129: Grimes, p. 76)
26.9 Buru intensity
haa-haa ‘really big (big-big)’
- em-pei=em-pei ‘really sick (sick-sick)’
- It is unlikely that Sula’s verbal categories are an exact parallel to Buru, but it does appear that descriptive words in Sula are intensified through reduplication, and when a receiver (nominal or main verb) is present to the left, the descriptive feature as well as the reduplication’s added intensity transfers to the receiver as an adjectival or adverbial quality. In the example below, mota-mota describes the ‘sagu’ as ‘hot’ and bahaha describes the ‘chili sauce’ as ‘quite-spicy’, but dab-dab is received as a nominal, ‘chili sauce,’ even though the only recorded translation of dab is at least semantically adjectival, ‘dizzy’ or ‘drunk’.
- gaya sa mota-mota do dab-dab baha-ha
eat sagu hot-hot with chili.sauce spicy-spicy
‘Eat hot sagu-palm starch with chili sauce that’s quite-spicy.’
The following example represents the type of sentences needing grammaticality judgements and accurate translations to determine how to correctly parse reduplicated modifiers and reveal their syntactic implications (e.g. description-transfer hierarchies, and/or stacking-limitations on reduplicated modifiers). In the example, bahu-bahu has been introduced to modify gaya, resulting in: ‘very-quickly eat,’ however after this point, it is unclear how to proceed in mapping the sentence’s descriptive features. Is the [[very-white sagu] very-hot], or is the [sagu very, very [white-hot]]? ponoida-ponoida indicates that more than one ‘friend’ (perhaps ‘each friend’) is present, but if ‘friends’ are modified by dabu-dabu, the result would be: [quite-spicy [very-drunk friends]], but if bahaha modifies dabu-dabu, the result will be [[very-spicy ‘chili-sauce’] friends] (which is rather nonsensical). At present, insufficient fieldwork has been conducted to determine how to parse reduplicated modifiers that have been stacked.
26.10 Additional-Descriptors
- *?gaya bahu-bahu sa bot-bot mota-mota do
eat fast-fast sagu white-white hot-hot withponoida-ponoida dab-dab baha-ha
friend-friend spicychili.sauce spicy-
‘…’
26.11 Distributive Noun Reduplication
Of the reduplication types described in Buru, the distributive nominal is the most readily observable in Sula. However, while this process in Buru results in a semantic value close to ‘each noun’ rather than ‘nouns’, the process seems to result in a value closer to plurality in Sula.
In the next example, everyone is derived from the lexeme mua (‘complete’/’all’). However while sanohi-sanohi logically represents more than one individual in the following example, my data provide no indication that the reduplicated form references all members distributed across the category of ‘family’; that is, it seems more likely that the speaker plans to buy coto Makassar for some plurality of family members rather than for everyone who is in the speaker’s family (the category of ‘family’ in Sula encompasses a great many individuals).
26.11.1 Distribution effect
- sai yau bia hai Sua bakareha yaya do baba
go far from land Sula separate mother and father
‘I went far from Sanana island and was separated from my mother and father,sanohi ponoida bubai do mua-mua bo hai Sua
family friend darling and everyone LOC land Sua
‘my family and friends, my darling, and everyone else on Sula.’
26.12 Plural effect?
- bo Falahu ak baha gaya coto
LOC Falahu 1SG buy eat stewMakassar buat sanohi-sanohi
Makassar for family-family
‘I buy Makassar-style stew in Falahu for my family.’
26.13 Native vs. loan reduplication
Two patterns emerge among reduplicated loanwords. Following the pattern of native Sula vocabulary, we would expect uang-uang (‘money-money’) in the next example to mean either ‘monetary bills’ or ‘each monetary bill’, but in the following example, we see that it instead means ‘money’ just as in the source language.
26.13.1 Reduplicated loan [reduplicated prior to import]
- uang-uang duk saik ak
money-money come here 1SGbit bihu bo Sua pai
want return LOC Sula ASPECT
‘Money, come my way! I want to return to Sula.’
On the other hand, sedikit-sedikit in Indonesian means ‘a little bit’, whereas in the next example it intensifies the smallness of pip (‘money’), resulting in ‘very little money’. This raises an interesting possibility that loanwords introduced in reduplicated form import the semantic effect of reduplication in the source language with them, while loanwords that are reduplicated after introduction follow the native-Sula model. Though this is a logical hypothesis, a much broader survey of the phenomenon is needed.
26.13.2 Reduplicated loan [reduplicated following import]
- ak bit faa moya sebab
1SG want buy NEG becauseak pip dahi sedikit-sedikit
1SG money exist little-little
‘I don’t want to go shopping, because I have very little money.’
Comparing Sula reduplication to Grimes’s (1992) Buru description, it is clear that while there are some similarities, significant differences exist between the two languages regarding principle reduplication classes and their syntactic and semantic functions. With regard to the principal classes of reduplication and the corresponding syntactic and semantic functions; unambiguous examples were found of Sula reduplication for three of the structures identified in Buru: (1) full reduplication of stem, (2) reduplication of root, and (3) reduplication of the first CV. There is also a possible additional structure in Sula: reduplication of final CV (e.g. baha ‘spicy’ ➞ baha-ha ‘quite spicy’).
Sula and Buru share properties associated with reduplication of active verbs and non-active, modifying verbs (adverbs of manner). Likewise, Sula shares the intensity transfer associated with reduplicated non-active main verbs in Buru, but this may extend beyond the verbal category in Sula. Lastly, Sula was found to share the distributive-noun reduplication process of Buru.
Regarding native versus loan reduplicated forms, there is some indication that loanwords can optionally enter Sula in their matrix form or in an already reduplicated form—in which case the reduplicated forms retain the semantic functions associated with reduplication in the source language. Based on this information, it is likely direct transference of Indonesian reduplications is permitted because modern Sula speakers are raised bilingual Indonesian speakers, and because Indonesian is the dominant language of most day-to-day speech domains.
27 Complex sentences
It is hard to come upon a definition of sentence that satisfies all linguists, but for the purposes of sketching this grammar, I am working with the following definition: a sentence is the basic verbal unit representing a complete thought. A sentence is a structural element that, in practice, usually includes (or at least implies) a subject and a verb and often direct or indirect objects and other complements.
27.1 relative clauses
Sula relative clauses mirror noun modification. That is, the head noun is followed by the clause that modifies it. Typically either a pronoun, pronominal prefix, or the borrowed Indonesian relativizer, yang is employed, but the same effect can also be achieved with a pause in speech. Because restrictive relative clauses seem to be formed only with yang, it is likely that Sula did not possess a native relative clause construction.
27.1.1 non-restrictive, possible RC with pronoun
- ak suka moya kapala-desa
1SG like NEG village-headik ki’i babua kadiga
DET 3SG crazy strong
‘don’t like the village leader, who is totally nuts.’
27.1.2 non-restrictive, possible RC with pronominal prefix
- ana fina bo hai Mangon i-
child female LOC island Mangon 3SG.AGRlaka nan bo Wai-tebi
walk bathe LOC Wai-tebi
‘the girl from Mangoli Island, who goes to bathe in the Waitebi river.’
27.1.3 restrictive RC with yang
27.1.4 possible RC with pause
- a- bamap sa (pause) bot do mina
1SG cook sagu (pause) white and tasty
‘I cook the sagu paste that’s white and tasty.’
27.2 Complements
Sula complements can be formed without an overt complementizer or with the borrowed Indonesian complementizer, bahwa.
27.2.1 Examples without overt complementizer
- ak nap-geka mai moya i- mata
1SG headache but NEG 3SG.AGR- die
‘It doesn’t surprise me that he died.’ - ki’i i- mata mai a- pusi moya
He 3SG.AGR- die but 1SG.AGR- headache NEG
‘That he died, doesn’t surprise me.’
27.2.2 Examples with overt complementizer
- Yanto nau bahwa a- duk,
Yanto know COMP 1SG.AGR- cometapi a- nau moya
but 1SG.AGR- know NEGbahwa i- duk atau moya
COMP 3SG.AGR- come or NEG
‘Yanto knows that I’m coming, but I don’t know whether Yanto will come or not.’
27.3 Conjunctions
Conjunctions in Sula closely mirror English: words, phrases, and clauses can all be conjoined; coordinating conjunctions (when present) are situated between connected parts of a sentence; subordinating conjunctions introduce the subordinate clause; and when there is a dependent and independent clause, the independent clause comes first in a prototypical subordinately conjoined sentence.
The three most common conjunctions are do ‘and’, te ‘or’, and mai ‘but’.
27.3.1 do ‘and’
27.3.2 te ‘or’
27.3.3 mai ‘but’
- ak lal suka wew moya mai
1SG inside like mango NEG butak lal suka fia
1SG inside like banana
‘I don’t much like mangoes, but I like bananas a lot.’
The following lists includes additional common conjunctions (including synonymous conjunctions borrowed from Indonesian):
Table 63. Sula conjunctions
Sula English
do ‘and’
te ‘or’
mai ‘but’
dan ‘and’
atau ‘or’
tapi ‘but’
tetap ‘but’
tetapi ‘but’
son ‘because’
bia ‘of’ / ‘more than’
non-conjoined sentences
Two simple, non-conjoined sentences:
- ak lal suka kof. ak lal suka teh.
1SG inside like coffee 1SG inside like tea
‘I really like coffee. I really like tea.’
27.3.4 coordinating conjunctions
laka te moya?
walk or not
‘to go or not?’
- kim bit (gi)- win
2SG.F want (2SG.F.AGR)- drinkkof te (gi)- win teh?
coffee or (2SG.F.AGR)- drink tea
‘Do you want to drink coffee or to drink tea?’
- ak lal suka kof do teh
1SG inside like coffee and tea
‘I really like coffee and tea.’ - ak lal suka kof do lal suka teh
1SG inside like coffee and inside like tea
‘I really like coffee and really like tea.’ - ak lal suka kof do ak
1SG inside like coffee and 1SGlal suka teh moya
inside like tea not
‘I really like coffee and I don’t really like tea.’
- i- mata mai a- pusi moya
3SG.AGR- die but 1SG.AGR- headache NEG
‘That he died, doesn’t surprise me.’ (doesn’t cause me headache) - ak a- laka gaya dumon
1SG 1SG.AGR- walk eat ?(tetapi) a- gaya kena moya
(but) 1SG.AGR- eat fish NEG
‘I will go to dinner with you, but I don’t want to eat fish.’ - Yanto nona (tetap) matapia basa bau-munara
Yanto sleep (but) person other CAUS-work
‘Yanto slept, but the others worked.’ - ak lal suka kof tapi
1SG inside like coffee but ak lal suka teh moya
1SG inside like tea not
‘I really like coffee but I don’t really like tea.’
27.3.5 Subordinating conjunctions
- koi nib bo nui-kau,
don’t sit LOC coconut-treeson nui bona mon
because coconut fall you
‘Don’t sit under a coconut tree, because a coconut will fall on you.’ - ak a- gaya moya
1SG 1SG.AGR- eat NEGsebab a- bis pel
because 1SG.AGR satisfied ASP
‘I don’t want to eat, because I’m full.’ - koi laka kaka-hoi sebab uya
NEG.IMP walk outside because rain
‘Don’t go outside, because it’s raining.’
kalo, kalau
- a- laka yota- haiwan
1SG.AGR walk hunt- animalbo awa kalo faf uka kasbi
LOC field if pig bite cassava
‘I go hunting in the fields, if pigs eat the cassava. - a- baha fa’a moya
1SG.AGR- buy thing NEGkalau ak pip dahi moya
if 1SG money exist NEG
‘I don’t buy things, if I have no money.’
27.3.6 Correlative conjunctions
Sula does not seem to have correlative conjunctions that consist of conjunction pairs joining alternative phrases together—for example either… or in I’d like either coffee or tea. Sula speakers form similar statements using a single conjunction.
- ak bit win kof te teh
1SG want drink coffee or teh
‘I’d like to drink coffee or tea.’ - ak lal suka kof bia teh
1SG inside like coffee of tea
‘I like coffee more than tea.’
28 Sula language sentences and texts
(Still compiling. To be added prior to defense)
29 Conclusion
While far from comprehensive, the previous pages have identified many basic grammatical structures in the previously undocumented Sula language. It is my hope that these pages will be of use to future researchers and aid workers desiring to learn the language’s fundamentals in advance of their projects. It will be an added bonus if these pages also provide answers to other linguists and a foundation for future exploration and documentation of the language.