Word classes

Doromu-Koki has three open word classes (nouns, complex verbs and adjectives), two semi-open classes (simple verbs and adverbs) and 13 closed classes (pronouns, demonstratives, locatives, postpositions, interrogatives, polar question markers, affirmative, negatives, degree adverbs, conjunctions, interjections, vocative and discourse markers). The noun class comprises approximately 43.8% of lexemes, while 37.5% are complex verbs and 7.2% are adjectives; the simple verbs comprise 5.5% and adverbs 1.1%, such that less than 1% are found in the closed classes. The open classes can accept new members through borrowings or derivations.

Members of the closed classes are listed below, followed by some discussion of the open and semi-open classes, grouped conceptually. (Refer to the various lexemes in the main section of the dictionary to see example sentences of each one.)

There are three types of pronouns found in the language: 1) transitive/intransitive subject/object, 2) possessive and 3) reflexive. The distinction between singular and plural is neutralised for second and third person in all three sets. Distinction can be indicated for the first two sets through the use of vene ‘people’ as in ya vene (2 people) ‘you (PL)’, ya vene di (2 people GEN) ‘your (PL)’ or ina vene (3 people) ‘they/them’, ina vene di (3 people GEN) ‘their(s)’.


Table 1: Pronouns

  Transitive/intransitive subject/object   Possessive   Reflexive
  Singular (SG) Plural (PL)   SG PL   SG PL
First na una   nai uni   naike uniye/unike
Second ya   yi   yaike
Third ina   ini   inaike/inike/iniye

An alternate reflexive form is kaya as in na kaya (1SG self) ‘myself’. Reciprocal is also indicated by use of reduplication of the A/S/O form + kaya, as in ya kaya ya kaya (2 self 2 self) ‘yourselves/each other’ or in the case of first or third person plural only through reduplication of the appropriate reflexive pronoun, as in uniye uniye (1SG.REFL-RED) ‘each other’.

In addition to these, two other word class items also function as impersonal pronouns. The noun amiye ‘person’ as an impersonal pronoun means ‘someone/anyone’ while the interrogative kaere ‘who’ as an impersonal pronoun means ‘whoever/whomever’.

There are two main generic demonstratives: mina ‘this’ and mirona ‘that’ which are further differentiated for greater specificity. Aside from these, two others to indicate greater distance are indicated in the table below. The demonstratives function both pronominally and adnominally.


Table 2: Demonstratives

  Specific   Generic   Greater distal
Elevated         isena  
          ‘that (one) up over there’  
Distal yomirona   mirona   yorona  
  ‘that (one) right there’   ‘that (one)’   ‘that (one) over there’  
Proximal yomina   mina      
  ‘this (one) right here’   ‘this (one)’      

The locatives follow the same structure as the demonstratives with the generic mini ‘here’ and mironi ‘there’, however, they have several further distinctions (indicated in bold) which are not correspondingly found with the demonstratives.


Table 3: Locatives

  Specific Generic Greater distance Further still
Elevated yoiseni forero iseni (HM) isefu
  ‘right up over there’ ‘up there’ ‘up over there’ ‘away up over there’
Distal yomironi mironi yoroni/soroni  
  ‘right there’ ‘there’ ‘over there’  
Medial yomirodu   mirodu yoisefu
  ‘right over there’   ‘over there’ ‘away right over there’
Proximal yomini mini midu gaima
  ‘right here’ ‘here’ ‘over here’ ‘far away’

There are five semantic subclasses of postpositions as indicated in the following table. Because of the many borrowed words this class is considered semi-closed. Many can additionally be used with the postpositional clitic =ri ‘at, in, on’, as in etofaro=ri (outside=at) ‘outside’, and in fact several have grammaticalised forms incorporating =ri, as in fogori ‘amongst, during’ or the rapid speech form (ro) of rofu, as in vaifuro ‘above’.


Table 4: Postpositions

Relative location Directional
fafau ‘above, on (top of)’ etafari ‘away from’
odoro ‘above, over, on (top of)’ gutuna ‘from, out of’
vaifuro ‘above’ urana ‘to(ward)(s), for’
egona ‘below, lower down’ vaitani (HM) ‘from, according to’
gabire ‘under(neath), below’ Temporal
nefau ‘in front (of)’ fogori ‘amongst, during’
adina ‘beside, close/next to’ fuofuori ‘while, during, when’
atafu ‘near, close/next to, beside, aside’ lalonari (HM) ‘during, while, when’
kefe ‘beside’ duakau ‘during, when, whilst’
negau ‘near, close (by)’ neganai (HM) ‘when, whilst’
etofaro ‘outside’ neitua ‘little while, during’
ide ‘inside, within, in’    
Case relationship Clitics  
di ‘genitive’ =ri ‘at, in, on’
dudu ‘instrument’ =sa ‘accompaniment, with’
nufa ‘possessive’ =u ‘by, on, instrument’
rofu ‘benefactive, recipient, ablative, purpose’    
totona (HM) ‘purpose’    

A few of the postpositions are heterosemous, also functioning as clause linkers; they are listed in the table below.


Table 5: Heterosemous postpostional/clause linking forms

Form Postpositional meaning   Clause linking meaning
adina ‘beside, close/next to’   ‘because, meaning of’
dudu ‘instrument, with’   ‘according to, thus’
fafau ‘above, on (top of)’   ‘concerning, basis of, about’
=ri ‘at, in, on’   ‘simultaneous different subject (while)’
rofu ‘for, at, to, with, from’   ‘in order to, so that’

The eleven interrogatives are shown in the following table, divided into those which start with the indeterminate form go- on the right and those that do not on the left.


Table 6: Interrogative words

Item Gloss(es) Item Gloss(es)
kaere ‘who’ gokai ‘how’
beika ‘what’    
beika resi[1] ‘why’ gokai resi ‘why, how’
    gokaisanu ‘how many/much’
beikadinare ‘when’ goivaka ‘when’
    goina ‘which (one)’
    goidu ‘where (direction)’
    goini ‘where (location)’

As previously mentioned, kaere ‘who’ also functions as the impersonal pronoun ‘whoever/ whomever’. Alternative forms include reduplication, as in kaere kaere ‘those who’ and from beika ‘what’ the alliterative beika baika ‘whatever’.

There are two different polar question words, ba and eni, which can also be used together as eni ba or ba eni. In addition the second one can be used in combination with the complex verb ‘tell’.

For an affirmative response the word yo ‘yes’ is used; it can also be used to indicate further affirmation to a statement. There are five negatives, shown below. As a response, the first term is said as ide ‘no’, while as a means of negating a predicate it is realised as de.


Table 7: Negatives

Term Gloss(es) Functions
i(de) ‘no, not’ Predicate/response
dia ‘not’ Clausal
ga ‘PROH’ Mood: Prohibitive
kode ‘not yet’  
koina ‘nothing’ Completive-type mood

There are six degree adverbs, shown in the following table. The first two (iniye and gade ‘very’) can be reduplicated for increased intensity. The second, gade and third, vayavaya ‘very’, have limited distribution, with gade only modifying tora ‘big’ or tau ‘many’ and vayavaya being reserved for buni ‘good’. The adjective tora ‘big’ also functions as a degree adverb, indicating quantitative intensification for uncountable items, as in goroka tora (voice great) ‘great/loud voice’. Similarly, the final form vanu ‘every’ is reserved to modify tau ‘many’, again indicating further quantitative intensification.


Table 8: Degree adverbs

Term Gloss(es)
iniye ‘very’
gade ‘very’
vayavaya ‘very’
tau ‘many, plenty, a lot, all’
tora ‘great, much’
vanu ‘every’

Like the postpositions, several conjunctions are also borrowed. There are three types each of coordinating and subordinating conjunctions. Often the borrowed terms have autochthonous counterparts.


Table 9: Conjunctions

Coordinating Subordinating
Additive Cause-effect
ma/bona (HM) ‘and’ ye ‘so, and’
=ka ‘also, too, as well as’ ine ‘so, because’
kumo ‘and/so (then)’ dada ‘so (that), because’
ma(mo) ‘at once, (and) even/then, until’ mina dada/ resi ‘therefore, for this reason, because’
mo (HM) ‘but, at once, and, then’ rofu ‘so that’
kami(ni) ‘(and/so) then’    
Alternative Result-reason
ba ‘or, and’ adina/badina (HM) ‘because’
o ‘or’[2]  
Contrastive Contingent
idu ‘but, yet, nevertheless, even though’ vonisi[3] ‘if, even though’
to (HM) ‘but’ bema (HM) ‘if’
    baebu ‘lest, possibly’
    baeko ‘might, maybe, probably’

The eight interjections found in the language can be seen below; they can occur in isolation or initiate a clause.


Table 10: Interjections

Item Gloss(es)
aee ‘oh, o, ho’
ai ‘hey, wow’
aiso ‘wow’
aiyo ‘oh, o, behold’
amiyo ‘wow, behold’
ao ‘wow, oh’
e ‘oh, ah, o’
oe ‘hey, yes’

There are two discourse markers, the topic marker bi and the differential subject marker ya(ku). They are very commonly used, and the second in particular does not have an English equivalent.

As well as being used to highlight a topic, bi functions as the ‘copula’ in a verbless clause construction, as in na bi buni (1SG TOP good) ‘I am well’.

The differential subject marker ya(ku) marks the most salient A/S of the discourse, as in na yaku iruku mar-aka (1SG DSM food give-1SG.PST) ‘I gave (him) the food’. A sentence is still grammatical without it, as it instead has a pragmatic effect.

Nouns and their modifiers

There are seven subclasses of nouns: mass, count, proper (naming), plural, temporal, compound and adjectival. The first three subclasses (mass, count and proper) are exemplified in the tables below.


Table 11: Mass nouns

Type Examples    
Emotions ada ‘happiness’   badu ‘anger’
Liquids dava ‘blood’   ayaka ‘saliva’
Semi-liquids dikina ‘mud’   bidoka ‘mucus’
Free-flowing particles ramena ‘salt’   budoka ‘dust’
Complements baku ‘finding’   esika ‘pain’
Terms for days agiya ‘yesterday’   aineka ‘day after tomorrow’


Table 12: Count nouns

Type Examples    
Generic plants oruma ‘grass’ tauga ‘banana’  
Specific plants adafa ‘mushroom species’ koyotu ‘yam species’  
Generic animals auna ‘dog’ dona ‘pig’  
Specific animals ramuro ‘bird of paradise’ bau ‘lizard species’  
Objects yava ‘house’ fore ‘stone’  
Environmental features omuna ‘mountain’ koru ‘water, river’  


Table 13: Proper nouns

Type Examples Type Examples
Places Keininomu Buridobu Pools Ayaka moka Tabu moka
Men Rabona Dagere River mouths Gatama ema Muya ema
Women Imore Bravo Streams Efa buruka Seme buruka
Mountains Rofaika Vauraika Dances Kariku Kitoro
Rivers Vaya koru Tarua koru Spirits Aire sori Aneru

There are four limited subclasses of plural nouns: special/close kinship terms, more distant kinship terms and two others with only one member each: oyena ‘fish (sg)’ vs. oyevani ‘fish (pl)’ and amiye ‘person’ vs. vene ‘people’. Kinship terms have distinct forms as shown below.


Table 14: Kinship terms

Singular Gloss(es) Plural Gloss(es)
aufa ‘grandparent, grandchild’ aufakai ‘grandparent(s), grandchildren’
baba ‘father’ bakai ‘father(s)’
vada ‘uncle, nephew’ vadakai ‘uncles, nephews’
nono ‘mother’ nokai ‘mother(s)’
asini ‘friend’ asiyaka ‘friends’
dubuini ‘brother’ dubuiyaka ‘brothers’
edini ‘in-law’ ediyaka ‘in-laws’
madini ‘father, daddy’ madiyaka ‘fathers, daddies’
meraini ‘uncle’s child’ merayaka ‘uncle’s children, cousins’
radini ‘in-law’ radiyaka ‘in-laws’
rasini ‘sister’[4] rasiyaka ‘sisters’
tobaini ‘sister’[5] tobaiyaka ‘sisters’

The only other means of indicating plural number is through reduplication; this is, however, limited to just eighteen nouns in three categories: countable items, mass and distributive. There is a rather fine distinction such that the distributive could also be considered countable.


Table 15: Reduplication of nouns

Countable   Mass
Plural Gloss(es)   Plural Gloss(es)
bado bado ‘pieces’   baku baku ‘really finding’
duma duma ‘stealing a lot’   dogo dogo ‘really preparing’
gagani gagani ‘many places’   esika esika ‘great pain’
gauka gauka ‘much illness’   maina maina ‘searching and searching’
isaka isaka ‘lots of crying’   sufa sufa ‘very great/big bush’
mida mida ‘children’   Distributive
ofa ofa ‘many lies’   ida ida ‘different ways/roads’
ofi ofi ‘many young women’   roka roka ‘different names’
sina sina ‘many stories’   yabo yabo ‘different trees’
usa usa ‘much praying’      

Temporal nouns are of five types: calendrical, cyclical, terms for days and relative temporal nouns. Three are seen in the following table.


Table 16: Temporal nouns

Calendrial Gloss   Terms for days Gloss
meda ‘day (sun)’   gua ‘today, now’
fura ‘week’ (HM)   agiya ‘yesterday/tomorrow’
eyo ‘month (moon)’   aineka ‘day before yesterday, day after tomorrow’
lagani ‘year’ (HM)   neureka ‘three days before/after today’
      maitoka ‘four days before/after today’
      didoka ‘five days before/after today’
Cyclical     Relative  
uriyaku ‘morning’   biyatete ‘long time’
uriyenau ‘afternoon’   ruika ‘right now’
urusa ‘night’   sena(gi) ‘long time ago’

Compound nouns are formed by juxtaposition of two nouns considered to be one grammatical unit. There are over 200 currently identified in the language. The most common are Root – Endocentric, a case of the first element specifying the second. There is only one type of synthetic (endocentric) compounds. All the identified types can be seen in the examples found in the table below.


Table 17: Compound nouns

Endocentric Glosses Free translation
Doromu ago Doromu word ‘Doromu language’
koru.gena fou gun (lit. ‘water bamboo’) shot ‘gunshot’
uka esika mokena stomach pain thinking ‘heartache’
sioni iruku white-man food ‘store-bought food’
kani konagi mustard stem ‘mustard stem’
Koki gagani Koki place ‘Koki land’
koru seri water shore ‘riverbank’
ekalesia amiye church person ‘Christian’
furisi vene police person ‘police’
ita rafu sina fire fellowship story ‘argument’
koru gena water bamboo ‘gun’
ada fore head stone ‘ignorance’
ita ruvena firewood small.firewood ‘bundle of firewood’
nono baba mother father ‘parent(s)’
omuna ika mountain summit ‘mountain top’
sero.redo/rena amiye selling does/doing person ‘seller’

In male-female pairs, the female is always first (i.e. nono baba [mother father] ‘parents’ and rema rumana [woman man] ‘people’. One noun, vegu ‘life’ also functions as an adjective ‘green/living’.

Adjectives form the smallest open word class, with around 200 members, functioning as complement in verbless complement clauses as well as modifying nouns. They typically follow the noun they modify and can themselves be modified. Some (of the ‘difficulty’ semantic type) are polysemous, as seen in the following table.


Table 18: Polysemous adjectives covering the ‘Difficulty’ semantic type

Adjective Primary gloss Alternate glosses
amuta ‘peaceful’ ‘smooth, soft, nice, cool, simple, comfortable, kind, gentle, calm’
buni ‘good’ ‘well, nice, beautiful, easy, reasonable, sensible, rational, logical, worthy’
gira ‘hard’ ‘firm, strong, stiff, solid, thick, tough (rough), difficult’
toe ‘heavy’ ‘weighty, suffering, judged, problematic, loaded, difficult’

There are 13 semantic types of adjectives found in Doromu-Koki: Dimension, age, value, colour, physical property, human propensity, similarity, qualification, quantification, position and number.

There is a small set of polar adjectives; they cannot be paired together to form compounds. They are shown below – note that categories are somewhat arbitrary.


Table 19: Polar adjectives

Positive/lesser Gloss(es) Negative/greater Gloss(es)
aita ‘lightweight’ toe ‘heavy’
amuta ‘peaceful, smooth’ kako ‘rough’
buni ‘good’ no ‘bad’
gabu ‘black’ feyo ‘white’
gira ‘hard, firm, strong, stiff’ yau ‘soft, weak, faint’
itu ‘cold’ vava ‘hot’
kakaita ‘narrow, tight, small’ rafo ‘flat, wide’
keika ‘little, small’ tora ‘big’
kiku ‘immature’ kiri ‘dry/mature’
mami ‘tasty, delicious, savory’ ru ‘bitter’
roko ‘dry, arid’ dou ‘wet, damp’
rorobo ‘straight, just’ kevo ‘bent, crooked’
ruaka ‘new, clean, fresh, young’ ugava ‘old’
ruana ‘young’ kori ‘old (things)’
tumu ‘short’ doba ‘long, tall’

There are two subclasses of adjectives: pre-nominal and multi-class functioning. This second group consists of adjectives which can function as nouns, adjectival modifiers or manner adverbs. They include aita ‘light (weight)’, berou ‘other’, buni ‘good’, gira ‘hard’, isivaga ‘powerful’, keika ‘little’, no ‘bad’, toe ‘heavy’ and tora ‘big’.

Verbs and their modifiers

There are only 179 simple verbs in the language, divided into two classes. The majority of verb stems are two syllable (52.5%), and then three (25.1%), then one (17.3%), and a few with four (5%), which are mainly compounded.

Class I is made of those verb stems ending in ar, and amounts to approximately 11% of Doromu-Koki verbs. The r is only realised when inflected with suffixes beginning with a vowel [e.g. mar-aka (give-1SG.PST) ‘I gave’ vs ma-dedi (give 3PL.PRS) ‘they give’].

Class II includes all the verb stems with all the remaining vowels (e, i, o and u), and comprises 87% of the verbs in the language.

Doromu-Koki has a rich inflectional morphology. Verbs primarily take suffixes, though there are also three prefixes and compounding. The suffixes indicate non-spatial setting, while the prefixes vary. There is no indication of object roles on the verb; they are encoded syntactically. Only S/A are indicated in the tense suffixes, such that on verb forms without tense, these are then determined by context.

In a list of all attested words in the language, there are over 1,700 different verb forms; for the verb ame- ‘stay’ alone there are 81 different inflected forms, such as ame ‘stay (sg)’, amebi ‘be staying (sg)’, amebifa ‘be staying (pl)’, amebifo ‘be staying (polite pl)’, amebigedi ‘they will be staying’, etc. The table below shows verbal affixation grouped according to type and order. The actual components themselves will be described in the applicable sections following.


Table 20: The structure of verbal word

  Prefix Stem Suffix(es)      


    -Aspect -Mood  
  -Reality status -Tense -Switch-reference

Mood is the most productive suffix type, having six types (imperative, polite imperative, hypothetical, possible, purposive and conative) while Aspect has four (perfective, past and future continuative and iterative) and Tense only three (past, present and future); Reality status has only two (indicative and potential). Tense, however, is the most common, as it is obligatory whenever its slot is not already filled by Mood. There are only a total of four prefixes, as well as four switch-reference markers. All are shown in the tables below.


Table 21: Mood and modality markers in Doromu-Koki

Slots Type Singular Plural
Reality status Indicative -ø  
  Potential -bo  
Other moods Imperative    
  1   -nadi
  2 -fa
  3 -aine(di) -agane(di)
  Polite imperative 2 -vo -fo
  Hypothetical -agadu  
  Possible -adu -aidu
  Purposive -gi  
  Conative -gika  


Table 22: Aspect markers in Doromu-Koki

Aspectual suffix Meaning
-gam ‘past continuative’
-bi ‘future continuative’
-RED ‘iterative’


Table 23: Tense suffixes

  Past     Present     Future  
Number Singular Plural   Singular Plural   Singular Plural
1 -aka -afa   -da -sifa   -gida -gifa
2 -o -adi   -sa -dedi   -giya -gedi
3   -do   -go


Table 24: Switch-reference suffixes

  Same subject Different subject
Subsequent -si -ma
Simultaneous -gasa -ri

There are a few uninflected verbs, which have been classified as verbs of intention and of cognition. The first class only has one member (va ‘try’), whilst the second has three (diba ‘know’, toto ‘do not know, forget’ and ura ‘want, wish, like’).

There are many more ways of expressing an action with a verb in Doromu-Koki; these include single word serial verbs, transitivity alternations or forming a complex predicate. The language has various single-word verbal compounds, primarily of two types: 1) Those beginning with the verb (as v1) ni- ‘say’ and 2) a few others. While the meanings of the whole are compositional, they are actually verbal compounds.

Verbs of speech: The verb ni- ‘say’ is quite commonly used in complex verbs after the complement, as in yoga ni- ‘laugh’ (lit. ‘laugh say’), but also often as a prefix before another verb or complex verb, forming a compositional verbal compound, realised as one grammatical word, as in ni-oteima (say-tell) ‘teach’.

There are other compounds which have undergone grammaticalisation, such as dei ne- ‘die’ [composed of de- ‘come’ and ne- ‘go (down)’], ‘move around, travel’ [composed of di- ‘go (around)’ + bo- ‘go (over)’ + associated verb re-], ourefeide- ‘lead, direct, guide, reign/rule over, precede’ (composed of oure- ‘be first/eldest/oldest/ next’ and feide- ‘leave, depart’) and youfeide- ‘surrender, give up, quit, concede’ [composed of you- ‘throw (away)’ and feide- ‘leave, depart’].

Approximately 66.5% of Doromu-Koki verbs are transitive, 26.1% are intransitive, 4.0% are ambitransitive and 3.4% are ditransitive. The verb rafe- ‘wash, bathe, swim’ is reflexive without an overt object; there is no indication of transitivity on the verb. More often than not, a third person transitive object is implied.

Doromu-Koki is able to make a transitive verb “ambitransitive” (S=A) through reduplication of the stem, often used as a complex verb nominal with the verb re- ‘do’. To date seven such forms have been attested: feide-feide re-yafa (cook-red do-1PL.PST) ‘we did cooking’, goe-goe (dig-RED) ‘digging’, iri-iri (eat-RED) ‘eating’, moke-moke (think-RED) ‘thinking’, neide-neide (hear-RED) ‘understanding/listening’, vari-vari (plant-RED) ‘planting’ and ve-ve (see-RED) ‘sight’.

There are five ditransitive verbs, mar- ‘give’, nimar- ‘commend’ (lit. ‘say-give’), nioku- ‘testify, inform’ (lit. ‘say-break’), nioteimar- ‘teach’ (lit. ‘say-tell’) and oteimar- ‘tell, show’. Four of them have to do with speech, three being ni- ‘say’ compounds. There are three special subtypes of verbs as well: 1) motion verbs (shown below), 2) posture [amei-nu (stay-STAT) ‘stay, remain’] and 3) causation verbs.


Table 25: Verbs of movement

Come Go  
  bo- Up
bae- di- Over
de- ne- Down

Complex verbs in Doromu-Koki consist of a complement and an associated simple verb. They mainly make use of the associated verbs re- ‘do’, ni- ‘say/become’, ri- ‘make’ and to a lesser extent 26 others: ae- ‘put’, ari- ‘arrange’, bae- ‘come’, bo- ‘go’, dadi- ‘get up’, di- ‘go (around)’, fere- ‘leave’, gurau- ‘suffer’, imi- ‘pierce’, iri- ‘eat’, mar- ‘give’, moi- ‘get’, ne- ‘go down’, nimar- ‘commend’, nioku- ‘testify’, nugar- ‘cut’, ode- ‘break’, ori- ‘burn, cook’, oure- ‘be first’, u- ‘hit’, vadi- ‘weave’, vai- ‘burn’, vari- ‘plant’, ve- ‘see’ and youfeide- ‘surrender’. New members can be created through borrowings, which most often make use of the associated verb re- ‘do’, but can also use others dependent on the semantics.

There are two subclasses; the larger (~70%) encompasses frozen forms in which the complements cannot be used on their own without the associated verb ( ‘designate’), and the smaller subclass (~30%) in which the complements are nouns (abata re- ‘flood’), adjectives (bere re- ‘be full’) or postpositions (negau ni- ‘draw near’), which can be used on their own in other contexts. The larger subclass, the non-compositional type, are indicated here by a full stop (hyphen in the main section of the dictionary) between the components, whilst the smaller subclass, the separable type, are written as consisting of separate words.

Some are double complex in that they require another associated verb or other complement (‘secondary’) at the beginning. Those with a secondary complement associated verb then function as a serial verb construction as in of ae- ‘put’ + ‘store away’ or ae- + ‘gather/store up’.

Also included in this type are echo-compounds, which carry an extended meaning. Most often they indicate intensity, as in ‘preparing and preparing’ versus dogo re- ‘prepare’. The noun complement dogo ‘preparation’ on its own forms a typical reduplication (i.e. dogo-dogo ‘preparations’). For another, the meaning is extended to repetition and intensity: ‘turn’ becomes ‘be very busy’ (i.e. ‘turning this way and that’). For yet another a further action is indicated: ‘chop’ becomes ‘chop and bring’. One has no other form or use than with the associated verb vari- ‘plant’: nikito.nakito.vari- ‘last planting’.

Another includes a tertiary complement ( ‘be fogged in’), in this case the noun seuya ‘cloud’. The secondary complement, fati is most likely modifying the noun, perhaps with a meaning such as ‘fog’, though since it cannot occur on its own, it has no individual meaning. Fono-re- means ‘cover’.

As alluded to above, one strategy for producing new complex verbs is through serial verb constructions. Another is through various means of forming single word verbal compounds. By far the most productive technique involves the verbs of speech. Ni- ‘say/become’ is the second most common simple verb used to form a complex verb, and is also realised as a verbal prefix, forming a single-word compound composed of ni- + another verb or an adjective or noun, or ni- + a complex verb. Some examples are seen below. Those above the line are those for which there is a known separable complement/component, as shown in the complement/component column. Those below the line all involve speech verbs, so that we can ascertain that the ni- component is analysable, but the other component does not have (or possibly no longer has) a corresponding separable component; they are non-compositional, lexicalised as such.


Table 26: Single-word verbal compounds analysed as verbs of speech

Item Gloss(es) Complement/component ‘command, sanctify, honour’ akeke ‘special, sacred, taboo’ (adj) ‘rebuke, scold’ besenai ‘rebuke, scolding’ (n)
niedadi- ‘wake (someone) up’ edadi- ‘get up/moving’ (v)
nifeide- ‘send, release, set free’ feide- ‘leave, depart’ (v) ‘rebuke, scold’ ‘discuss, debate, talk about, gossip’
nirabe- ‘credit, declare’
nivake- ‘worship, exalt, praise, serve’

More productive still are SVCs which have moi ‘get’ as their first component, of the form v1 + v2 (+ v3) where v2 can be a compositional or non-compositional complex (or even double complex) verb. Only one example for the full set of three verbs has been observed to date. With the 70+ Verbs of speech compounds and the 200+ of these ‘get’ type SVCs, the total possible number of verbs in the language stands at over 1,400, but it seems plausible that many more could be generated than have been observed so far. SVCs will of course add more.


Table 27: ‘Get’ serial verb constructions

v1 + v2 Gloss(es) Components
moi bae- ‘bring (up)’ moi- ‘get’ + bae- ‘come’
moi rama ae- ‘fulfil, make true’ moi- ‘get’ + rama ‘true’ + ae- ‘put’
moi vata.bae- ‘fill (up)’ moi- ‘get’ + vata.bae- ‘fill, limit’
moi rorobo ‘erect’ rorobo ‘straight’ + ‘stand (up)’
v1 + v2 + v3   moi- ‘get’ + etagae.ri-

‘move/go away/over’ + ae- ‘put’

moi etae.ri ae- ‘put/set aside’

Complex verb formation is a very productive strategy for generating new verbs in the language. When a term is borrowed, the associated verb in a complex verb will be re- ‘do’ unless the semantics require another, such as ni- ‘say/become’ as in ‘sing song/hymn’ or ‘become bald’. Many borrowed complex verbs are separable, but not all. Perhaps it has to do with the way it is first understood and used when it comes into the language. If it is understood merely as an action, then it becomes a non-compositional complex verb. If the particular thing is in focus, it can stand on its own. Table 28 is a listing of some borrowed complex verbs, separated according to type (non-compositional first, classified according to associated verb re- ‘do’, ni- ‘say’ and ni- ‘become’, followed by separable, in the same fashion, according to frequency of associated verb).


Table 28: Borrowed complex verbs

Non-compositional Gloss(es) Source language ‘believe, trust’ HM abidadama henia ‘trust’ ‘check, verify’ HM sekea ‘check’ ‘sing song/hymn’ HM ane abia ‘sing’ ‘be x o’clock’ English ‘become bald’ HM kebere ‘coconut shell, cup, bald’ ‘win, be victorious’ HM uini ‘win’


Separable Gloss(es) Source language
abata re- ‘flood’ HM abata ‘flood, tide’
babatiso re- ‘baptise’ HM bapatiso ‘baptism’
guriguri ni- ‘pray’ HM guriguri ‘pray’
bero ni- ‘be wounded/injured’ HM bero ‘wound, scar’
biyaguna ni- ‘inherit’ HM biaguna ‘owner’
dabua ri- ‘wear clothing’ HM dabua ‘clothes, clothing’
fuse ri- ‘bag (up)’ HM puse ‘bag, sack’
babatiso mar- ‘baptise’ HM bapatiso ‘baptism’
meino mar- ‘pacify’ HM maino ‘peace’
guri ae- ‘bury’ HM guri ‘hole in ground, pit’
susu iri- ‘breastfeed, suckle’ HM susu ‘sap, liquid’

The vast majority are from Hiri Motu (84.6%), although 27.3% of those came through English; altogether 38.4% have come from English on their way into the Doromu-Koki language.

One of the three means of causative formation provides a method of deriving verbs in the language, as the causative prefix e- is added to a non-verbal element, such as an adjective [ ( ‘make silent’], locative [e-gaima.ri- (I.CAUS-far.away.make-) ‘distance’] or postposition [ ( ‘draw near’]. This becomes the complement in a non-compositional complex verb, which is then followed by the appropriate associated simple verb.

When a verb takes the -na ‘nominaliser’ suffix it typically behaves as a noun [moke-na (think-NMLZ) ‘thinking’] of a special subtype. It also functions as deverbal nominalisation, denoting activity as in [amei-na (stay-NMLZ) ‘staying’].

There are three types of causatives found in the language; they can be organised according to their frequency: Direct (or intential), Indirect (or partial) and Forceful (or complete). An exhaustive list is shown in Table 29 below. The Semantic categories are based on Dixon 2012:268-280.


Table 29: Causative constructions

Type Form Strength Semantics
Direct moi x- ‘make’ Control (A), intentionally affecting (O)
Indirect e-x ‘cause’ Partial control (A), partial affecting (O)
Forceful u-x ‘force’ Control (A), intentionally, completely affecting (O), with effort


There are two primary types of clauses in Doromu-Koki: verbal and verbless. Verbal clauses can be either main or dependent clauses, while verbless are main clauses only.

Main clauses stand as complete sentences (cf. Dixon 2010a:75). Dependent clauses have either non-final verbs or final verbs with coordinate or subordinate clause linking conjunctions.

Non-final verbs are inflected with switch-reference marking (and non-spatial setting in the case of different subject marking). When same subject marking is used, non-spatial setting (reality status, aspect and tense) is not indicated until the final verb of the clause.

Final verbs are always inflected for non-spatial setting, and typically are the final element in a clause. All clauses can be negated with either ide or dia ‘negative’ or ga ‘prohibitive’ in conjuction with an imperative marking.

A verbal clause has the structure (TEMP) (A/S) (NPOBL) (NPO) (NPE) V. An intransitive clause could minimally contain a verb [e.g. di-yafa (go-1P.PST) ‘we went’] and still be a complete clause and sentence; all the remaining elements are optional. Temporals and locatives have a bit more flexible positions, for pragmatic effects, while other elements do not.

Due to switch-reference, clause linking clauses can have a dependent coordinate linking structure – forming a chain in which they are all dependent on the final clause verbal morphology to identify the transitive/intransitive subject (as well as tense).

Verbless clause structure is: VCS bi ‘topic marker’ VCC. (The topic marker is obligatory in this clause type.) Verbless clauses cannot form dependent clauses or commands; these two require verbs, as in dependent clauses and commands using imperative forms. Negation is straightforward, with addition of the negative to the end of the clause, as in Koki bi ago roka de [(name) TOP word name NEG] ‘Koki is not the name of the language’. There are four types of verbless clauses found in Doromu-Koki: 1) equative, 2) attributive, 3) locative and 4) existential.

The verbless clause subject can also be filled by a pronoun or demonstrative, such as na bi amiye doba (1SG TOP person long) ‘I am a tall man’ or mina bi dona keika (this TOP pig small) ‘this is a small pig’. There are two subtypes of equative: interrogative and possessive.

An interrogative can occupy the slot of the verbless clause complement. Interrogatives retain the same constituent order as an equative (or declarative) clause, as in yi roka bi beika? (2.POSS name TOP what) ‘what is your name?’ or baiya bi goini? (bush.knife TOP where) ‘where is the bush knife?’ versus baiya bi mini (bush.knife TOP here) ‘the bush knife is here’.

A possessive pronoun can also occupy the NP slot in the complement slot, such that it could be considered pertensive as pro + di ‘genitive’, in which case the initial consonant of the genitive is deleted, as well as the final vowel of the original intransitive/transitive subject/object pronoun [except in the case of first person singular] (i.e. na + di > nai; ya + di > yi).

The possessor focus type above could be translated as ‘belongs to/owns…’, and as such some items cannot ‘belong to’ a person [e.g. one cannot say *maruka bi ini (husband TOP 3.POSS) ‘the husband is hers’, but instead one would say bi maruka nufa (TOP husband POSS) ‘she has a husband’]. However, this possessor focus type above can readily be used with objects, as in gugura bi uni (thing TOP 1PL.POSS) ‘the things are ours’. Similarly, states cannot be likewise possessed: *esika bi nai (pain TOP 1SG.POSS) ‘the pain is mine’, whereas these possessive pronouns can be used to modify a noun in an NP: nai esika bi tora (1SG.POSS pain is big) ‘my pain is great’.

In the second type, attributive, the complement is filled by an adjective phrase, the instrumental dudu, often glossed as ‘with’ or an adverbial, with ones such as kamini ‘enough’ or kaini ‘already’. After equative, this is one of the more productive types.

Locative clauses make use of the various locatives found in the language in the verbless clause complement slot. A verbless locative clause complement can also be filled with a postpositional phrase.

There are two types of possessive verbless clauses; with the case relationship postposition nufa ‘possessive’, marking the possessee or the equative, subtype possessive. This type of predicative possession emphasises the relationship between the possessor and the possessee, such that nai foketa bi fon nufa (1SG.POSS pocket TOP phone ACCM) could be translated as ‘(My) pocket had a phone (in it).’ It is not, however, inalienable possession, as one could not say *na bi yuka nufa (1SG TOP leg POSS) ‘I have a leg’ or *na bi nono nufa (1SG TOP mother POSS) ‘I have a mother’, but instead would encounter nai nono ame-do (1SG.POSS mother stay-3SG.PRS) ‘My mother stays/exists’.

There are three types of speech acts; clause types can be distinguished in terms of these three: 1) statements, 2) questions and 3) commands.

Statement is the most common speech act; these types of clauses have general downward pitch contours for each clause, falling at the end of the sentence.

Asking a question, real or rhetorical, is filled with a question word and has a downward pitch contour with a final sharp rise.

In other instances, only a word expressing uncertainty will give a clue to the interrogative nature of the utterance.

Commands are indicated by the use of the imperative verbal affixes, and as applicable, the prohibitive. As with statements, commands also have falling contour patterns. The means of maintaining distinction between the two then falls on the verbal morphology.

In addition to dependent clauses marked with switch-reference or other clause linking devices and main clauses with final verbs, two other types of clauses occur in Doromu-Koki, relative clauses and complement clauses.

A relative clause modifies the common argument in the main clause (Dixon 2010b:314). Doromu-Koki often exhibits external pre-head relative clauses. They are juxtaposed to the head and restrictive. When an intervening optional relativiser mina ‘’ is used as an anaphoric referent, then a post-head non-restrictive RC is formed, to add extraneous information.

Common arguments can be common or proper nouns, pronouns or demonstratives, or headless adjectives when used in a non-restrictive RC; in a restrictive RC only common nouns are permissible.

Relative clauses with S as a common argument are the most common. Also RC ‘subject’ with main clause ‘subject’ are very common (A/A, A/S, S/A, S/S, VCS/A, VCS/S), which highlights the fact that Doromu-Koki is a nominative-accusative language in terms of its syntactic pivot.

The only other relativiser is kaere ‘who’, an English calque, often also post-head, as in a verbless clause subject slot functioning as a relative clause [with recapitulation of the subject] or in a verbless clause complement slot.

Not all uses of bi are exclusive to verbless clauses. The topic marker is quite pervasive throughout the language. Relative clauses are quite productive in all core argument slots.

According to the Accessibility hierarchy (Keenan and Comrie 1977:66), relativisation follows a continuum; that is the higher on the hierarchy that relativisation is found, it will also be found in all the lower levels. The accessibility hierarchy is seen below:

Accessibility hierarchy


The subject is most accessible to relativisation and every language can relativise on it, whereas the object of comparison is the least accessible, and not found in some languages. If a language can relativise the object of comparison then it can also relativise those items preceeding it in the hierarchy. Below we explore how far up the relativisation hierarchy Doromu-Koki goes.

Subject (as defined by the hierarchy above) function of the common argument in the RC is the most common, as verbless clause subject, as intransitive subject (verbless in the RC itself) and as transitive subject.

Direct object position is also quite common, but not as much as transitive/ intransitive/ verbless clause subject.

Indirect object is quite rare.

The oblique slot is filled by a locative construction. It is more common than the previous type.

No examples in the genitive or object of comparison slots have been observed, so it is concluded that oblique is the highest position on the hierarchy available in Doromu-Koki. Relative clauses can be negated just like main clauses, as in [moimai de re-do] rema (work NEG do-3SG.PRS woman) ‘the non-working woman’. They cannot however be made into commands or questions.

Complement clauses may be less frequent than relative clauses, yet productive in the language nonetheless. They fill a whole argument slot, but have only been observed in the verbless clause subject or object slots. By far the most productive strategy is for them to fill a reported speech. Verbs used with complement clauses include those of perception (ve- ‘see’), cognition ( ‘assume’), emotion (ori re- ‘fear’) and speech ( ‘tell’, ni- ‘say’). Due to the pragmatics of such verbs, complement clauses do not fill transitive/intransitive subject slots or other peripherial slots, but only function as verbless clause subject or object. A complement clause can be identified by its ability to optionally have the demonstrative mina ‘this’ occurring after it, as anaphoric reference. Complement clauses are also extremely common in reported speech.

As with main clauses, complement clauses can be negated and be questions. They can also be commands. In some respects complement clauses are more productive than relative clauses, but in most respects they are not.

[1] re-si is literally ‘do-SQ.SS’, so this could be glossed as ‘doing what’ and similarly gokai resi as ‘doing how’; note a similar construction with mina resi ‘therefore’ [lit. ‘doing this’].

[2] Most likely from English.

[3] This appears to be a grammaticalisation of vo+ni-si ‘happen+become-SQ.SS’.

[4] Woman’s parent’s same sex sibling’s daughter; female three generations removed from ego.

[5] Parent’s same sex sibling’s same sex child; female three generations removed from a male, male three generations removed from a female.