Typological Overview

Authored by: John Lynch , Malcolm Ross , Terry Crowley

The Oceanic Languages

Print publication date:  December  2001
Online publication date:  March  2013

Print ISBN: 9780700711284
eBook ISBN: 9780203820384
Adobe ISBN: 9781136749858

10.4324/9780203820384.ch3

 

Abstract

Although the Oceanic languages constitute a fairly well-defined subgrouping within the larger Austronesian family, as described in Chapter 1, they do certainly not constitute a typological unity. At the same time, however, there are certain patterns and structures which tend to recur over large geographical and genetic groupings of Oceanic languages. It is the purpose of this chapter to describe those structural features that are more widely distributed among Oceanic languages. We will also describe those features that are less widely distributed, but which are nevertheless found in languages spread over particular geographical areas.

 Add to shortlist  Cite

Typological Overview

Although the Oceanic languages constitute a fairly well-defined subgrouping within the larger Austronesian family, as described in Chapter 1, they do certainly not constitute a typological unity. At the same time, however, there are certain patterns and structures which tend to recur over large geographical and genetic groupings of Oceanic languages. It is the purpose of this chapter to describe those structural features that are more widely distributed among Oceanic languages. We will also describe those features that are less widely distributed, but which are nevertheless found in languages spread over particular geographical areas.

No detailed typological survey of Oceanic languages has been published, though much of the comparative literature necessarily makes typological comparisons between languages (Pawley 1972, 1973, 1981, Clark 1976, Chung 1978, Lynch 1981, Ross 1988). Lynch (1998) represents an attempt to present a typological survey of the whole group, though for an introductory readership.

The content of this chapter will allow each of the sketches in this volume to be placed in an appropriate typological context. When reading a sketch, if a feature is not described in detail, and it is indicated in this chapter as typical for Oceanic languages in general — or for the languages of that particular area — then it can be assumed that this feature is characteristic of the language in question.

We can generalise, for example, that Oceanic languages for the most part exhibit two main patterns of possessive constructions, which we can refer to as ‘direct’ and ‘indirect’ possession. Therefore, we will not describe this difference in each sketch where this distinction is made. It is only if a particular language — or group of languages — does not make this distinction, or makes it in an unusual way, that a special point will be made. To facilitate comparison, this chapter and Chapter 5, which describes Proto Oceanic, have virtually the same structure as the sketches.

1  Phonology

This is an area in which it is very difficult to make generalisations that are applicable to members of the entire subgroup of Oceanic languages. Although the reconstructed phoneme inventories for Proto Oceanic and a number of intermediate languages are fairly uncontroversial (see Chapter 5), the kinds of phonological changes that have taken place in different languages have been so diverse as to produce a very wide range of different sorts of phonological typologies.

Even so, languages in this subgroup are frequently phonologically less complex than those of many other linguistic groupings in the world. Syllable structures tend to approximate a simple CV type, and phoneme inventories tend to be both fairly small, and characterised by relatively few complex articulations. The major exceptions to the latter generalisation involve the presence of labio-velar or labialised velar stops, nasals and fricatives in many Melanesian and Micronesian languages, as well as the occurrence of phonetically prenasalised voiced stops which contrast with plain oral voiceless stops in many parts of Melanesia.

Stress is generally fully predictable, falling on the penultimate syllable of a word. In languages that have contrastive stress, it usually has a fairly low functional load. Contrastive vowel length is widely distributed in the eastern parts of Oceania, though it is by no means universal. Distinctive vowel length is much less common in western Oceania. Oceanic languages are almost exclusively non-tonal, though a few languages of New Caledonia and the Huon Gulf in PNG have developed phonemically contrastive tone.

2  Nouns And Noun Phrases

2.1  Pronouns

Pronominal systems generally involve a contrast between first, second and third person, with no gender distinctions. First person pronouns almost without exception distinguish between inclusive and exclusive. There is always a distinction between singular and plural, and generally also a distinct dual series. The latter usually contains an element that is historically related in some way to the numeral ‘two’, though there is normally no productive synchronic compounding process involved. Some languages in Melanesia also have a series with an element reflecting ‘three’, and a few have one with an element reflecting ‘four’. The series with ‘three’ is trial in some languages but paucal (expressing the meaning of ‘few’) in others. The series with ‘four’ may be synchronically either paucal or plural (but is probably never genuinely quadral).

The number of languages that do not fit within these generalisations is generally small or geographically restricted. Only Kiribati, a few varieties of Fijian, and some New Guinea languages do not mark an inclusive/exclusive distinction. Languages with only a two-way number contrast in pronouns are for the most part geographically concentrated in the New Guinea area, as well as a scattering of Micronesian languages, and Sye and Nakanamanga in Vanuatu. Separate masculine and feminine third person pronouns are reported in Kilivila (Trobriand Islands), southern New Britain and Maringe (Santa Ysabel).

There are generally several separate paradigms of pronominal forms. Four kinds of paradigm are widespread in Oceanic languages, one of free forms, the others of affixes or clitics:

  1. Independent (i.e. free) pronouns are used in citation and function as noun phrases, i.e. as topic in topic-comment constructions, and as subject, object, possessor or prepositional object.

  2. Possessor suffixes on bound nouns and possessive classifiers indicate the person and number of a possessor; their syntax is described in §2.7 below. In some western Melanesian languages these suffixes also occur on some prepositions and on attributive adjectives indicating the person and number of, respectively, the prepositional object and the head noun.

  3. Subject: most languages have one or more sets of preverbal morphemes — usually clitics, but sometimes prefixes or free forms — which indicate the person and number of the subject. In Melanesian languages these are often portmanteau forms which combine with the expression of the tense/aspect/mood categories of the verb (§3.2; in the grammar sketches these are often handled under the verb phrase).

  4. Object: the fourth kind of paradigm is less widespread than the other three, but is found quite often in Melanesia and Micronesia. It consists of a set of postverbal clitics or suffixes indicating the person and number of the object.

Many languages lack one or more of these sets, and in western Melanesia, there are languages with other sets (see the discussions of Taiof and Kairiru in this volume). There are often partial formal similarities in the shapes of free and bound pronominal forms, especially the independent and object sets, but there are seldom systematic correspondences between the various paradigms. Manam (PNG) pronominal forms are typical:

1 INC

1 EXC

2

3

Independent

SG

ŋa(u)

Ɂai(Ɂo)

ŋai

DL

Ɂitaru

Ɂeru

Ɂanru

diaru

PC

Ɂitato

Ɂeto

Ɂanto

diato

PL

Ɂita

ɁeɁa

Ɂaŋ ∼ ɁaɁamiŋ

di

Possessor

SG

-gu

-ŋ)

-Ø ∼ -na

DL

-daru

-mairu

-minru

-diaru

PC

-dato

-mato

-minto

-diato

PL

-da

-ma

-min

-dia

Subject (realis)

SG

u-

Ɂu-

i-

PL

ta-

Ɂi

Ɂa-

di-

Subject (Irrealis)

SG

m-

go-

rja-

PL

ta-

ga-

Ɂama-

da-

Object

SG

-a

-(i)Ɂo

-Ø ∼ -i ∼ -a

DL

-Ɂitaru

-Ɂamairu

-Ɂaminru

-diaru

PC

-Ɂitato

-Ɂamato

-Ɂaminto

-diato

PL

-Ɂita

-Ɂama

-Ɂamirj

-Ø ∼ -i ∼ -di

From these paradigms, it will be noted that there is a recurring element /ru/ in the dual forms and /to/ in the paucal forms, related historically (though not synchronically) to the numerals /rua/ ‘two’ and /toli/ ‘three’ respectively.

As the absence of dual and paucal forms in the Manam subject paradigms illustrates, in some languages there are differences between paradigms in the semantic distinctions that are made. In Vinmavis (Vanuatu) it is the independent paradigm and the second and third persons of the possessor paradigm that lack dual forms, and where the independent pronouns make typical non-singular inclusive/exclusive and second-/third-person distinctions, these distinctions are missing from the subject prefixes:

1INC

1EXC

2

3

Independent

SG

no

gu

i

PL

get

gemem

gem

ar

1INC

1EXC

2

3

Subject

SG

nV-

u-

i-

DL

er-

er-

ar-

ar-

PL

it-

it-

at-

at-

2.2  Nouns

Nouns are often categorised in two ways in Oceanic languages. Firstly, nouns are either directly or indirectly possessed. Directly possessed nouns occur with a possessor suffix (§2.1), whilst indirectly possessed nouns are unsuffixed. This structural distinction reflects a semantic distinction between inalienable and alienable possession (§2.7).

Secondly, nouns are either personal, local or common. Personal nouns include personal proper names and, in some languages, also kin terms denoting particular individuals relative to the speaker (e.g. ‘my father’). Local nouns include institutionalised place names as well as nouns denoting places so familiar in the environment that they need no further specification (e.g. ‘home’, ‘(own) village’, ‘(own) garden’, ‘bush’, ‘beach’ etc.). This category sometimes also includes directly suffixed locative part nouns (e.g. ‘inside’, ‘above’, ‘beneath’ etc.). The class of common nouns includes all other nouns, including kin and place nouns when not used in the senses just mentioned. Common nouns often fall into a smaller non-count and a larger count subcategory.

A noun's category membership determines its syntactic behaviour (e.g. the prepositions with which it may occur), and sometimes an accompanying article. In the languages of the north Bougainville linkage, this categorisation has become a gender system with articles marking gender and number co-occurring not only with the head noun but also with an attribute (see Taiof in this volume, §2.2).

The only other genuine gender systems, in which there is concord between constituents of a noun phrase and the head, occur in the languages of the Kilivila (Trobriand Islands) group, where a numeral classifier system (§2.4) has undergone functional expansion to become a gender system. Thus, in Kilivila we find:

tau

m-to-na

to-kabitam

ma-bwa-na

bwa-tolu kai

man

this-CL-thi

CL -intelligent

this-CL-this

CL-three tree

‘this intelligent man’

‘these three trees’

The two systems of nominal subcategorisation — directly/indirectly possessed and personal/local/common — are independent of each other, i.e. we cannot predict whether a particular personal (or common, or local) noun will be directly or indirectly possessed. Two pseudo-categorisations also occur, according to the nature of the possessive marker that an indirectly possessed noun co-occurs with (§2.7), and according to the form of the numeral classifier with which a noun co-occurs. These are pseudo-categorisations in the sense that the same noun may occur with several different markers/classifiers, so that the latter function as closed sets of attributes rather than as markers of the noun's category membership.

Nouns are always invariant for morphological case. Oceanic languages generally do not mark number on nouns inflectionally, so that an unmarked noun can usually be used to express both singular and plural meanings. However, there is often a hierarchy of animacy, with more highly animate nouns more likely to obligatorily distinguish singular from plural, and less animate nouns making no formal distinction between singular and plural. It is fairly common for Polynesian languages to distinguish singular and plural forms of a few kinship terms or nouns with human reference by either some form of reduplication, or a root modification pattern involving the lengthening of one or more root vowels. Thus, in Māori, singular tangata ‘person’ and tuahine ‘(man's) sister’ correspond to plural tāngata ‘people’ and tuāhine ‘(man's) sisters’. Languages of central and southeast Papua also tend to have separate plural forms for kinship terms. Oceanic languages also use various phrasal number-marking strategies (§2.4).

Nouns are often productively derived from verbs (and occasionally from other roots). Generally abstract nouns have zero-derivation or a suffix (or, in some languages in western Melanesia, the infix -in-), whilst agents, instruments and locations are derived by means of a prefix. Reduplication also occurs, deriving abstract nouns or instruments. For example, the Tubetube (southeast Papua) verb ‘work’ is paisewa, from which are formed pai-paisewa ‘work (NOUN)’ (reduplication), to-paisewa ‘one who works’ and kaba-paisewa ‘place where one works’. In Tolai (New Britain) we find m-in-omo ‘a drink’ from momo ‘drink’, tu-tutuk ‘a hammer’ from tutuk ‘hammer (VERB)’ (reduplication), tena-papalum ‘one who works’ from papalum ‘work (VERB)’.

2.3  Articles and demonstratives

Many Oceanic languages have articles that precede a noun phrase. These often make a distinction between singular and plural, and between common and proper, and sometimes make a more fine-grained set of semantic contrasts than this. In Fijian, for example, the distinction between common and proper is marked by the preposed articles na and o respectively:

na

koro

o

Tomasi

ART

village

ART

Thomas

‘the village’

‘Thomas’

Noun phrases with generic or locative/temporal reference generally do not appear with any article.

The languages of Manus, mainland New Guinea, Micronesia and Vanuatu generally do not have articles. What was historically an article has in many of the languages of Vanuatu and some of the languages of southwest New Britain been fused with the noun root, being morphologically inseparable in all, or at least most, morphosyntactic contexts (Crowley 1985). When these fused articles are separable, it is most likely to be when a noun appears as the second member of a nominal compound. Thus, in Atchin (Vanuatu), temets ‘ghost’ can be compounded with naleng ‘dance type’ to derive temets-leng ‘madman’, with the historical article (POc *na) being deleted in the compounded form -leng. In the languages of the Admiralties, New Caledonia, and part of Santa Ysabel, however, fusion has progressed so far that the fused element is never separable, and the languages of New Caledonia have subsequently developed new sets of articles, which express a fine range of semantic distinctions. While many of the languages of western Melanesia have articles, there is also a wide scattering of languages that do not.

Demonstratives are often identical to locative pro-forms, and usually make a three-way distinction based on either person (i.e. near speaker, near addressee, or near neither), or relative distance (i.e. close, intermediate distance, and distant), with the form marking intermediate distance being the least marked and often serving as the relative clause marker or occasionally as third person pronoun. More complex systems marking distinctions such as inland/seaward or upper/lower are found in various languages, while others make only a simple proximate/distant distinction.

2.4  Numerals and number-marking

The most widely distributed pattern of numerals in Oceanic languages is based on a decimal system, found throughout Polynesia and Micronesia, as well as in much of Melanesia. These languages often also have separate lexical items for ‘hundred’ and ‘thousand’, and in a few cases in Micronesia, separate lexical items for 10,000, 100,000 and 1,000,000 and so on, up to 1,000,000,000.

However, in Vanuatu and New Caledonia, as well as in a wide scattering of locations further west, quinary systems are often found, with numbers higher than ‘five’ expressed as compounds based on ‘five’ or some other word. Some languages have a combination of quinary and decimal systems, with the numbers ‘six’ to ‘nine’ being compounds involving the form for ‘five’, along with a separate lexical item for ‘ten’. These languages do not generally have separate stems for ‘hundred’ or ‘thousand’.

Numerals sometimes behave like adjectives in that they are postposed to a nominal head. However, outside western Melanesia it is more common for numerals to accept some verbal inflectional morphology, while in others there is some vestigial verbal morphology.

Some languages of Micronesia and the Admiralties, as well as the languages of the Kilivila family (§2.2), are characterised by fairly elaborate systems of numeral classifiers which are either postposed to the numeral, or directly suffixed to it. There can be more than two dozen separate numeral classifiers in Micronesian languages. Note the following from Woleaian:

se-uw

teebol

se-fash

wa

se-faiu

usous

one-CL

table

one-CL

canoe

one-CL

bead

‘one table’

‘one canoe’

‘one bead’

se-mal

mal

se-yal

temaag

one-CL

bird

one-CL

cigarette

‘one bird’

‘one cigarette’

As we noted in §2.2, nouns are not generally inflected for number, but there are other number-marking strategies. In many languages, these are normally used only for more highly animate nouns. One, rather rare, strategy is affixation. For example, Sye (Vanuatu) has a plural prefix ovn-, and Micronesian languages tend to use suffixes to mark definite singular and plural. It is more common for plurality to be marked by adposed independent number markers, which are often identical in shape to the third person non-singular pronouns, e.g. Paamese:

ehon

keil

child

3PL

‘children’

In western Melanesia a plural marker sometimes occurs between the preposed article and the noun. In north Bougainville languages, in New Caledonia and in Polynesia the article itself has distinct singular and plural forms. In many languages the number of a subject or object noun phrase is indicated by the pronominal forms in the verb phrase (§2.1, §3.2), and in western Melanesia number is often marked on an attributive adjective (§2.5).

2.5  Adjectives and nominal modifiers

If an Oceanic language has a class of genuine adjectives at all, it is likely to be a small closed set of forms which is defined by the fact that its members are uninfected and can be used both as a predicate and attributively when they directly follow the nominal head that they modify. Stative meanings are more generally expressed by intransitive verbs. Some languages allow such verb roots to appear as postmodifiers to nouns with no inflectional marking, while other languages require such roots always to be inflected. In such cases, a stative modifier to a noun can only be expressed in the form of a relative clause. It is common in the languages of western Melanesia for adjectives to distinguish the number and person of their referent by means of ‘possessor’ suffixes (§2.1). Thus, in ‘Ala‘ala (central Papua, this volume), we find:

oveka nama-na

dog big-3SG

‘big dog’

oveka nama-ta

dog big-3PL

‘big dogs’

oni eva kau

toɁotoɁo-mu

you:SG TOPIC person

short-2sG

‘You are a short person’

2.6  Basic noun phrase structure

If a language has articles and/or number markers, these usually precede the head noun. This means that in languages which lack both, the head noun is either always or frequently the first item in a noun phrase. Languages in which plurality is marked by an independent 3PL pronoun vary: in some it precedes the head noun, in others it follows it. In most languages, all other noun modifiers follow the noun, but there are a few where a demonstrative or a possessor precedes.

2.7  Possession

The expression of possession is usually one of the more complex aspects of the grammar of Oceanic languages, and this is an area that has been widely studied from a typological and comparative perspective (Lynch 1973, Lichtenberk 1985). There is also a considerable amount of variation in the form of possessive constructions, though the semantic distinction between alienability and inalienability lies at the core of the main structural differences in all systems.

For the majority of Oceanic languages, there are two main possessive construction types, which we refer to as ‘direct’ and ‘indirect’ possession. In the case of direct possessive constructions, a possessor suffix (§2.1) is attached directly to the possessed noun, while with indirect possession, an uninfected possessed noun is either preceded or followed by an independent possessive constituent, which is itself marked with one of the possessor suffixes. Thus, in Fijian we find the following:

na mata-qu

na

no-qu

vale

ART

eye-1SG

ART

POSS-1SG

house

‘my eye’

‘my house’

With both direct and indirect possession, a possessor noun (as against a pronoun) generally follows the possessed noun, but the order is reversed in a number of western Melanesian languages, especially those with SOV clause order. In Micronesian and eastern Melanesian languages, the possessed NP is marked with what is generally referred to as the ‘construct’ suffix, or some other linking morpheme. The construct suffix sometimes coincides in shape with the third person singular pronominal suffix, but the two are frequently morphologically distinct. Thus, compare the following in Anejorh:

etma-k

etma-n

etma-i

natimarid

father-1SG

father-3SG

father-CONST

chief

‘my father’

‘his/her father’

‘the chief's father’

In some PNG languages, however, nominal and pronominal possession are not expressed by different constructions at all. Instead, the possessed NP is marked with a pronominal suffix that expresses the pronominal category of the possessor, and the possessor appears as a full NP. Thus, in Motu (where the possessor is preposed) we find:

(lau) tama-gu

Morea

tama-na

(1SG)father-1SG

Morea

father-3SG

‘my father’

‘Morea's father’

The formal distinction between directly and indirectly possessed nouns generally corresponds to a semantic distinction between inalienable and alienable possession, with direct possession expressing semantic inalienability, and indirect possession expressing alienability. The detailed semantic content of inalienable nouns varies from language to language but, in general, it includes body parts, locative parts (‘inside’, ‘underneath’ etc.), kin terms, and often abstract nouns denoting things done to or said of the possessor. Alienable nouns are all other nouns in the language. Generally, body parts that are in some sense removable, and kin over whom one has authority or who one acquires through marriage, are often indirectly possessed.

Possessive systems making only the simple direct/indirect distinction are widely distributed in the languages of PNG. However, most Melanesian and Micronesian languages have systems that are more complicated than this in that they distinguish different kinds of alienable possession by means of a set of pronominally suffixed possessive constituents or ‘classifiers’. In western Melanesia, there is often a classifier that is used to express the possession of items for consumption (i.e. food and drink, and often items used in procuring them), as distinct from other items. In eastern Melanesia, we find separate classifiers for food and drink possessions, while some languages have up to half a dozen separate classifiers expressing different semantic categories of possession. Compare the following in Lenakel (Vanuatu):

nɨte

nɨki-k

nɨkava

nɨmwɨk

taro

POSS:FOOD-1SG

kava

POSS:DRINK-1SG

‘my taro’

‘my kava’

kolei

ne-k

nimwa

taha-k

sweet.potato POSS: PLANT-1SG

house

POSS:NEUTRAL-1SG

‘my sweet potato’

‘my house’

In Micronesia, as well as in Mussau (to the north of New Ireland) and Iaai (New Caledonia), the number of possessive classifiers is much larger, with many kinds of alienable possession being formally distinguished. While Micronesian languages typically group nouns semantically according to the kind of numeral classifier they are associated with, the semantic categories encoded by the systems of numeral classifiers and possessive classifiers do not coincide (nor do their forms): the numeral classifiers are generally based on physical form, while the possessive classifiers are generally based on function.

The possessive classifiers have sometimes been referred to as reflecting noun class systems. Unlike noun classes, however, association with a particular possessive classifier is often not fixed (Pawley and Sayaba 1990). This means that the same noun can appear in more than one possessive construction, depending on the relationship between the possessor and the possessed NP. Thus, we find examples such as the following in Paamese:

ani

ā-k

ani

ema-k

coconut

POSS:FOOD-1SG

coconut

POSS:DRINK-1SG

‘my coconut (of which I intend to eat the flesh)’

‘my coconut (of which I intend to drink the liquid)’

ani

esa-k

ani

ona-k

coconut

POSS:PLANT-1SG

coconut

POSS:NEUTRAL-1SG

‘my coconut (which is growing on my land)’

‘my coconut (which I intend, perhaps, to sit on)’

There are languages in which nouns may shift fairly freely from the directly to the indirectly possessed category, but in others this mobility is quite constrained and inalienable nouns occur only in the direct possession construction.

Despite the widespread distribution of these different patterns of possession, there are some languages which exhibit rather different patterns. In the languages of Malakula (Vanuatu), there are often possessive pronominal suffixes only with singular possessors. Plural pronominal possession is expressed instead by independent pronouns in association with the construct suffix, in the same way as nominal possession. Thus, in Vinmavis we find the following:

netal-ung

netal-n

get

netal-n

matoro

leg-1SG

leg-CONST

1INC:PL

leg-CONST

old.man

‘my leg’

‘our legs’

‘the old man's leg’

In Western Fijian, there is also a formal contrast between two kinds of inalienable possession, with kin terms accepting direct pronominal suffixes and body parts accepting instead pronominal prefixes, e.g.

o

mna-m

m-ulu

ART

mother-2SG

2SG-head

‘your mother’

‘your head’

In Melanesia, some languages express indirect possession with a preposition which links possessed and possessor. Thus in Nakanai (New Britain):

la bua te Pasi

la bua te

la tahalo

la bua t-egite

ART betelnut PREP Pasi

ART betelnut PREP

ART man

ART betelnut PREP-3PL

‘Pasi's betelnut’

‘the man's betelnut’

‘their betelnut’

The Polynesian subgroup is characterised by a completely different pattern of possessive markers. These are usually referred to in the literature as the a and o forms because possessor and possessed nouns are linked by either the particle a or o, depending on the semantic relationship between the possessed noun and the possessor. Thus, in Samoan, we find:

le paopao o Tavita

le naifi a le

taule ‘ale’a

ART canoe POSS Tavita

ART knife POSS ART

young.man

‘Tavita's canoe’

‘the young man's knife’

With pronominal possessors, there are two sets of preposed possessive classifiers which carry pronominal suffixes, which differ mainly in the fact that the vowel of the root is a or o. Thus, in Samoan:

lo-‘u

paopao

la-‘u

naifi

POSS-1SG

canoe

POSS-1SG

knife

‘my canoe’

‘my knife’

The semantic distinction between the two possessive constructions is often referred to in the literature as ‘dominant’ and ‘subordinate’ possession, with dominant possession being marked by a, and subordinate possession being marked by o. Forms that express possession with a in Polynesian languages generally correspond to forms that participate in indirect possessive constructions elsewhere, while o possession corresponds roughly to directly suffixed constructions. However, there is a greater amount of arbitrariness in the way that nouns in Polynesian languages are assigned to the two possessive constructions.

2.8  Relative clauses

Relative clauses are generally postposed to the nominal head, with the clause being preceded by a marker that often has some kind of broad subordinating function. These languages generally allow relativisation of NPs well down the universal Accessibility Hierarchy. With relativised NPs high on the hierarchy, there may be zero trace at the site of the relativised NP, though in languages in which verbs are obligatorily cross-referenced for subject or object, there will still be marking on the verb for the relativised NP. Thus, we find the following relative clauses in Paamese:

meatin

keke

ēhon

ø-les-i

person

REL

child

3SG.REAL-see-3SG

‘the person who the child saw’

meatin

keke

ēhon Ø-selūs

min-i

person

REL

child

3SG:REAL-speak GOAL-3SG

‘the person who the child spoke to’

With relativised NPs lower on the hierarchy, there is generally some kind of obligatory free form trace. There are many languages, however, which require that a trace be left with all relativised NPs.

3  Verbs and Verb Phrases

3.1  Verbal derivation and inflection

Verbs in Oceanic languages typically do not have extensive patterns of derivational morphology. The causative is generally expressed by a verbal prefix in Polynesian and Micronesian languages, Rotuman and Fijian, as well as in a broad scattering of Melanesian languages. Fijian and the Polynesian languages, as well as many western Melanesian languages, also have a derivational reciprocal prefix. Western Melanesian languages also often have a prefix which derives an intransitive stative from a transitive.

Reduplication is almost universally used in Oceanic verbal morphology, as well as in noun derivation. It expresses a wide range of meanings, including randomness of action, repetition, and plurality of actors or patients. It is sometimes also used to derive intransitive from transitive verbs.

Some Oceanic languages exhibit patterns of root-initial segment mutation corresponding roughly to a distinction between realis and irrealis categories (Lynch 1975, Tryon 1979, Walsh 1982, Crowley 1991). Hote, Buang, Jabêm and Bukawa of the Huon Gulf area, PNG, have patterns in which verbs marked for realis categories reflect the basic form of the root, and irrealis categories are expressed with a verb root that reflects a historical nasal increment from earlier *na ‘irrealis marker’ (which has often lost any phonemic nasal element synchronically) (Ross 1988: 370–372). Such a pattern is also found in Nāti on Malakula and Sye on Erromango (both in Vanuatu).

Most of the languages of central Vanuatu deliver us a typological surprise: they have the opposite pattern, such that it is the irrealis which uses the basic root form, and the realis which uses verb roots reflecting a nasal increment (from earlier *mV-). Most of these languages have two sets of root forms which are distributed according to the morphosyntactic context, though Paamese and Southeast Ambrym have as many as four different mutated forms of the verb root.

In some languages subject and object pronominal morphemes (§2.1) are respectively prefixed and suffixed to the verb: see §3.2 below.

We frequently find some kind of formal marking for transitivity on verbs. Some languages have derivational morphemes, normally a maximum of two, each of which may transitivise an intransitive verb, increasing its valency by adding an object. With the first, which generally has a shape that can be derived from the Proto Oceanic form *-i (see Ch. 5, §3.1), the object is typically a patient. With the second, derived from Proto Oceanic *-aki(ni), the object is an argument such as a location, a goal, an instrument or a cause, i.e. an argument which would otherwise be an oblique noun phrase. In the Oceanic literature these objects are often referred to respectively as ‘close’ and ‘remote’ objects. In Proto Oceanic, these suffixes were generally added to an intransitive root with a final consonant, like *taŋis below, but in most Oceanic languages word-final consonants have been lost, with the result that when the ancient consonant is retained before a transitive affix it is interpreted as part of the suffix, as is Fijian /ð/ here:

Proto Oceanic

*taŋis

*taŋis-i-a

*taŋis-aki-a

weep

weep-TR-3SG:OBJ

weep-TR-3SG:OBJ

Fijian

taŋi

taŋi-ða

taŋi-ðaka

weep

weep-TR:3SG:OBJ

weep-TR:3SG:OBJ

‘weep’

‘cry for’

‘cry because of’

Because there were a number of root-final consonants in Proto Oceanic, an outcome of this process is that the transitive morphemes in Fijian and many other Oceanic languages have acquired a variety of allomorphs, e.g. Fijian (in orthographic rather than phonemic form) -ca, -ta, -ka, -va, -na etc., and similarly -caka, -taka and so on. This in turn has resulted in the occurrence of etymologically unexpected consonants in this position, and, in Fijian at least, in their acquisition of specific meanings, so that from Proto Oceanic *soka ‘stab, spear’ we do find the expected Fijian coka-a ‘spear it’, but expected *coka-aka ‘spear with it’ is replaced by coka-taka. We also find coka-va ‘dive toward it’, coka-ta ‘tackle him’ and coka-raka ‘spear it repeatedly’ (Sch?85:152; also Arms 1974). In other languages, however, these suffixes are often no longer productive, and in many of the languages of Melanesia there is only vestigial marking, or no formal marking for transitivity at all.

Passive constructions are only very rarely encountered in the languages of Melanesia. In some Polynesian and Micronesian languages, there are passive constructions. In Polynesian, this is generally expressed by a morpheme that is historically related to a transitive marker. In those Micronesian languages that have a passive construction, this is marked by a separate verbal affix.

3.2  Basic verb phrase structure

It is in the area of verbal morphology and verb phrase syntax that Oceanic languages generally exhibit the greatest complexity. Oceanic languages generally have preposed verbal morphemes, falling into two basic types, according to whether these morphemes are free or prefixed, as in Raga (this volume) and Paamese (both in Vanuatu) respectively:

ra-m

ban

au-va

3PL-CONT

go

3PL:REAL-go

‘they are going’

‘they are going’

Languages with preposed particles (like Raga) are found throughout Micronesia, Fiji and Polynesia, though there is a significant number in Melanesia as well. Languages with verbal prefixes (like Paamese) are widely distributed throughout Melanesia.

The most transparent pattern of subject and tense/aspect/mood (TAM) marking is that in which these categories are marked by separate free forms which precede an invariant verbal root, with preceding pronominal markers indicating subject (§2.1). The number of preverbal markers can be quite large, with several markers appearing in fixed sequence. There may also be a negative marker interposed between the last of the TAM markers and the verb. This is the kind of pattern that is found with Polynesian and Micronesian languages. Thus in Samoan:

Ou

te

fia

alu

i

Apia.

1SG

TAM

NEG

DESID

go

PREP

Apia

‘I don't want to go to Apia.’

Straddling the divide between free and prefixing systems are those in which a morphologically complex preverbal marker expresses a combination of subject and TAM categories, with some combinations of categories being expressed by portmanteau forms. Raga, illustrated above, is a language where the portmanteau form is not prefixed to the verb. In many of the languages of Melanesia, however, we find extensive patterns of portmanteau prefixation, with varying degrees of morphotactic complexity. In some languages, there are clearly recognisable orders of subject and TAM prefixes, while in other languages there are sets of subject/TAM markers that are essentially not morphologically divisible. A language of the latter type is Paamese. Thus:

lo-loh

lehe-loh

1INC:PL:REAL-run

1INC:PL:DlST.FUT-run

‘we ran’

‘we will run’

Postverbal morphemes are often less tightly bound to the verb than preverbal morphemes; some are better analysed as enclitics than as suffixes, others as clause constituents than as verb phrase constituents.

When a pronominal object is expressed by a bound form, this invariably takes the shape of a postverbal clitic or suffix (§2.1). When there is only a partial set of object forms, these are more likely to be singular than plural forms, and third person than non-third. Languages with bound pronominal object markers are spoken in Micronesia, New Caledonia, parts of southern and central Vanuatu and in many of the languages of western Melanesia.

It is also quite common for a generic object to be incorporated into the verb phrase. In such cases, the verb is syntactically intransitive, there is no bound object marker, and the ‘object’ cannot be separated from the verb. This structure is most obvious in languages with clear verb phrase boundaries like Taiof (this volume) and in languages with ergative/absolutive case-marking like Roviana (this volume) or Samoan. In Iaai (this volume) a verb with an incorporated object may itself be transitivised.

Another commonly occurring category of postverbal morphemes consists of directional enclitics, with at least two members meaning ‘hither’ and ‘thither’. These are often cognate with the verbs ‘come’ and ‘go’ in other languages, and are presumably derived from earlier directional verb serialisations (cf. §3.3). For example, in Babatana (Choiseul):

Ra ko-qisu-me

kavia kuda.

Göi

ma-zo-la Susuka.

I 1SG:REAL-carry-hither

some coconut

s/he

3SG:IRR-walk-thither Susuka

‘I have brought some coconuts.’

‘S/he is going to Susuka village.’

The final element of a verb phrase is often an aspect morpheme, either enclitic or free. In Melanesia it is also common for manner adverbs to be incorporated into the verb phrase. Both features are illustrated in this Motu clause:

E-gwau-heni-gu

dikadika-va.

3SG-scold-give-1SG

bad-PAST:CONT

‘He scolded me badly’

3.3  Verb serialisation

Serial verb constructions of various types are encountered in a wide range of Oceanic languages (Crowley 1987, Early 1993, Sperlich 1993, Hamel 1993, Bradshaw 1993). These are more easily recognisable in languages that have inflectional prefixes and suffixes, as the initial verb in a serial verb construction is the one which typically carries the prefixed markers, while the final verb is the one which typically carries the suffixed markers. Thus, in Paamese, where the discontinuous negative marker ro-/-tei surrounds the verb, we find:

Ni-ro-kan

vīs-tei

ouh.

1SG:DlST.FUT-NEG1-eat

try-NEG2

yam

‘I will not try to eat the yam.’

Even so, series of verbal roots in languages where verbs are uninflected are widely distributed. They can be recognised as serial verb constructions by the fact that they share nominal arguments, and a single set of tense-aspect-mood and, where relevant, negative markers, as well as often having meanings that are not completely predictable from the meanings of their constituent verbs.

Serial verb constructions in Oceanic languages differ in the extent to which the verbs in question are structurally linked. Some languages allow serial verbs to independently choose objects, while other languages only allow a single set of subject and object arguments to a serial construction. In many languages, serialised verbs all have subject prefixes, whilst in others, only the first verb is prefixed. Some languages make a contrast between ‘nuclear’ serialisations, where the verbs are bound together and have only a single set of arguments (i.e. the serial construction behaves just like a single verb), and ‘core’ constructions, where the verbs remain separate words and usually share just one argument, any other argument being the subject or object of just one of the component verbs.

Despite this variety, we can identify certain semantic types of serialisation which recur in Oceanic languages. These types are all core constructions, but some have nuclear equivalents in some languages. The types are illustrated below from Paamese. Although the transitivity of the first verb may be determined by the construction, the transitivity of the second is not so determined.

(1) Directional/Positional: the first verb expresses movement, the second the direction of that movement or the position reached as a result of that movement. There are two syntactic subtypes, depending on whether the first verb is intransitive or transitive. If it is intransitive, the moving object/person is the subject of both verbs,

e.g.

Meatin

kail

a-valus

au-mai.

person

PL

3PL:REAL-row

2PL:REAL-come

‘The people rowed hither.’

If it is transitive, the moving object/person is object of the first verb and subject of the second, e.g.

Kaik

ko-muasi-nau

nau-vā

netan.

you

2SG.REAL-hit-1SG:OBJ

1SG:REAL-go

down

‘You hit me down.’

(2) Sequential: the first verb expresses movement, the second the action that follows the movement. The verbs have the same subject. A purposive relationship between the actions is usually implied:

Ki-ro-vā-tei

kī-hol?

2SG:IRR-NEG1-go-NEG2

2SG:IRR-dance

‘Won't you go dancing?’

(3) Causative: the first verb is transitive, the second expresses the result of the action of the first. The object of the first verb is subject of the second:

Kail

a-muas

vuas

Ø-emat.

they

3PL:REAL-hit

Pig

3SG:REAL-die

‘They killed the pig by hitting it.’

(4) Manner: the second verb expresses how the action of the first verb was performed. The verbs have the same subject:

Kai

Ø-mual

Ø-suai.

he

3SG:REAL-walk

3SG:REAL-disappear

‘He was walking without being seen.’

(5) Ambient: the implicit (third person singular) subject of the second verb is the sub-event expressed by the first:

Ki-hulī-n

ato

kail

he-mal.

2SG:IRR-count-OBJ

chicken

PL

3SG:IRR-be.correct

‘Count the chickens correctly.’

The second verb is often ‘finish’, expressing completive aspect or sequence (‘and then’).

This categorisation is certainly not exhaustive: languages such as Loniu (Admiralties), Jabêm (Huon Gulf, PNG) and Paamese have a variety of other types of serialisation. But despite this, and although serial verb constructions are widely encountered, we generally do not find the same amount of freedom in Oceanic languages that we find, for example, in many of the non-Austronesian languages of PNG, which are more thoroughgoing in the extent of their serialisation patterns. What we typically find in Oceanic languages is that only some categories of verbs can appear as the second member of serial constructions other than (b), or that there is even a lexically determined set of verbs that can be serialised.

Indeed, in some languages, some serialised verbs may even never occur as independent verbs. In certain languages of central Vanuatu, verbs which occupy the second slot in a serial construction appear to be becoming restricted to that structural slot alone. In some of these languages, these forms can be better analysed as adverbial constituents within a structurally expanded verb phrase. In some languages of the Huon Gulf of PNG, this process is complete for causative serialisation, and former verbs must now be analysed as adverbs.

Other languages have what can be analysed as verbal derivational suffixes. In languages such as Paamese and Lewo in central Vanuatu, there is some ambiguity between nuclear serialisation, adverbial postmodification and derivational suffixation with some verbal constructions, which suggests that earlier serialised verbs are currently being morphologised as derivational suffixes.

There are yet other languages in which serial verbs have been further grammaticalised, having evolved into derivational affixes of various kinds. In many languages of southeast Papua and the north coast of PNG, verbs are found with what are referred to as classificatory prefixes. These prefixes are derived from the first verb in an earlier causative serial construction. Thus all verbs expressing hitting actions, or cutting actions, or actions involving the teeth or the feet, may begin with the same prefix, with the following element either occurring independently as a verb with a related meaning, or only occurring in conjunction with one or more of these classificatory prefixes. Compare the following in Tawala, which involve the classificatory prefixes hana- ‘action involving teeth’ and tu- ‘action involving feet’:

hana-hedali

hana-lolova

tu-hedali

tu-loloya

BITE-break

BITE-tear

STEP-break

STEP-tear

‘break with the teeth’

‘tear with the teeth’

‘break with the feet’

‘tear with the feet’

Despite being widespread, serial verb constructions are not universal in Oceanic languages. The languages of southern Vanuatu have evolved a system of echo-subject prefixes which is not found anywhere else in the Oceanic subgroup. The existence of this feature coincides with a lack of serial verb constructions. Thus, in Lenakel we find:

R-ɨm-vin

kani m-ɨm-amnuumw nɨkava

3SG-PAST-go and ES-PAST-drink

kava

‘He went and drank kava.’

In languages that have this construction, switch-reference can be indicated by using the regular subject marker instead of the echo-subject prefix on the second verb. Thus:

R-ɨm-vɨn

kani r-ɨm-amnuumw nɨkava

3SG-PAST-go and 3SG-PAST-drink

kava

‘He1 went and he2 drank kava.’

4  Clause Structure

4.1  Verbless clauses

Equational sentences are generally expressed by simple juxtaposition of noun phrases with no intervening verb. However, in some languages there is a copula. In certain Vanuatu languages, a copula is optional in the present and past affirmative, though with other inflectional categories it becomes obligatory. Thus, compare the following in Paamese:

Mail

(vī)

asuv.

Mail

ro-vi-tei

asuv.

Mail

3SG.REAL:COP

chief

Mail

3SG:REAL:NEG1-COP-NEG2

chief

‘Mail is a chief.’

‘Mail is not a chief.’

4.2  Verbal clauses: core arguments

A wide variety of basic constituent orders is encountered in the Oceanic subgroup. The various constituent orders are distributed geographically as follows. (Note the following abbreviatory conventions: S = subject, T = topic, V = verb phrase, O = object, X = arguments other than topic):

SVO Admiralty Islands, most Markham Valley, Siasi islands, most New Britain, New Ireland, some Bougainville, most southeast Solomons, most Vanuatu, some New Caledonia, most Micronesia.

SOV central and southeast Papua, some Markham Valley, Madang coast, Wewak coast, Sarmi coast, a few Bougainville, some New Britain.

VSO New Georgia, some Santa Ysabel, much of Polynesia, Yapese.

VOS Fijian, Anejo (this volume), Loyalty Islands, Kiribati, many New Caledonia, Gela (this volume).

TVX many Bougainville, Choiseul, some Santa Ysabel.

Note that clauses in which both subject and object are realised as noun phrases are rare in discourse. Clauses often consist only of a verb phrase, with its clitics or affixes coreferencing subject and object.

The presence of an independent pronoun as subject or object marks contrast or focus. However, pronominal objects are more frequently expressed by independent pronouns than are subjects.

It is clear that SVO order is geographically the most widely distributed pattern, as well as being found in the genetically most diverse sample of languages. The SOV order is restricted to certain parts of PNG, and it is generally assumed that this order has arisen as a result of contact with non-Austronesian languages which have this as the dominant constituent order (see Ch. 1, §4.1). There are two subtypes: strict SOV (e.g. Takia, Madang coast, this volume) and ‘leaky’ SOV where some — usually peripheral — constituents may follow the verb (e.g. Tawala, southeast Papua). In Fijian and many Polynesian languages, VOS and VSO respectively are generally treated as the basic word orders, but there is considerable freedom in the order in which constituents can occur.

Most Oceanic languages have fairly fixed basic constituent orders, but generally allow the movement of constituents to clause-initial position in order to express topicalisation. This has become grammaticalised in the TVX order of Bougainville, Choiseul, and Santa Ysabel. If a topic is the ‘framework within which the main predication holds’ (Chafe 1976:50), then T in these languages is a marked topic, a phrase whose referent is introduced for the first time or is re-introduced, often after a gap (an unmarked topic, one that is present in the immediately preceding discourse, is referenced at most by a subject or object morpheme in the verb phrase).

It might be argued that these are simply VSO or VOS languages in which topicalisation occurs, but this is not an adequate characterisation, for several reasons. First, only one argument, the ‘topic’, is permitted before the verb phrase (although it may be preceded by a peripheral — usually temporal — argument). Second, when the subject is not T, then its exact postverbal position is unpredictable. Third, the interrogative phrase in a wh-question is T in these languages, whereas in most Oceanic languages wh-questions have the same order as declaratives. There are a number of variations on the TVX pattern. Maringe (Santa Ysabel), for example, has a special clause-final focus position.

Oceanic languages with VO order are typologically ‘well-behaved’ in that they are generally associated with the occurrence of prepositions rather than postpositions, postnominal adjectives and possessed-possessor order in genitive constructions. Languages with the order OV are less well-behaved in that they associate with postpositions and possessor-possessed genitive constructions, while still having post-nominal adjectives. When languages of this type have inflectionally marked subject and object affixes on the verb, the subject is typically marked by a prefix, and the object by a suffix, which is what we might expect from a VO language. The OV languages of Morobe in PNG, however, are aberrant in exhibiting a mixture of typically VO and OV features, having possessor-possessed genitive constructions, but a mixture of prepositions and postpositions.

Oceanic languages are generally nominative-accusative in their formal marking of core syntactic roles, with the distinction between subject and object being marked by word order and, in some languages, also by cross-referencing on the verb. Some languages have overt marking for subject and object roles. This is more likely to be the case with languages that do not have SVO order.

However, a few Oceanic languages have ergative-absolutive patterns of marking. A few languages of Papua (Motu, Sinagoro, Maisin) have optional ergative and absolutive clitics which follow respectively a transitive subject and an object or intransitive subject. Roviana in Solomon Islands has an absolutive morpheme which precedes an object or intransitive subject. Of the Polynesian languages, those belonging to the Tongic and Samoic subgroups have ergative marking. In these languages, both ergative and absolutive NPs are marked by preposed case markers. In some of these languages, there is a kind of split ergative pattern of marking, with some verbs marking their core nominal arguments ergatively, and other verbs being associated with accusatively marked nominal arguments.

4.3  Verbal clauses: peripheral arguments

Non-core nominal arguments in a clause are generally marked by adposed constituents, these typically being prepositions in the case of VO languages, and postpositions in the case of OV languages. Oceanic languages do not mark peripheral functions with affixed case markers.

Most Oceanic languages have less than half a dozen genuine adpositions, with other peripheral arguments being expressed by means of complex constructions involving an adposition and a possessive noun phrase with a locative element as its head. Thus in Motu, which has postpositions (cliticised as ai, which contracts to i after a), we find forms such as the following:

ruma lalo-na=i

house inside-3SG:POSS-POSTP

‘in the house’ (lit. ‘at the house's inside’)

ruma henu-na=i

house underside-3SG:POSS-POSTP

‘under the house’ (lit. ‘at the house's underside’)

ruma lata-na=i

house top-3SG:POSS-POSTP

‘on top of the house’ (lit. ‘at the house's top’)

The languages of New Caledonia, however, have larger sets of prepositions, with closer to a dozen members.

In those languages which have lost the original transitive suffixes on verbs, there is often a multi-purpose oblique preposition that is used in a semantically fairly empty way to allow an NP to be used as an object of what we might want to call a pseudo-transitive verb. Thus, in Paamese:

Na-musau

en

sauen.

1SG:REAL-sing

OBL

song

‘I sang the song.’

4.4  Negative clauses

The expression of negation tends to be closely related to the expression of subject and TAM categories. In languages that express these categories with free forms, the negative marker also tends to be a free form. It is generally interposed between the TAM markers and the verb, though in some Polynesian languages, negation is marked by a clause-initial marker, while in many PNG languages it is marked clause-finally. In languages like Raga, where the subject/TAM markers are morphologically bound to each other as preposed particles, the negative marker still follows these forms and precedes the verb. In languages that have extensive inflectional prefixation, negation is generally also marked with a prefix. This form is generally morphotactically separable from the subject/TAM markers, appearing between these and the verb root.

There is a recurring tendency in Oceanic languages for negation to be expressed discontinuously. Typically, the first element occupies a preverbal slot, while the second negative element appears postverbally. This pattern is found both with languages that have free form preverbal particles (e.g. Rotuman and some Polynesian languages) as well as languages that have inflectional prefixes (e.g. Takia, Paamese, Lenakel, Vinmavis). Thus, we find the following in Paamese:

Ni-ro-lesi-ko-tei.

1SG:DIST.FUT-NEG1-see-2SG-NEG2

‘I will not see you.’

The bipartite negative markers are generally quite different in shape from language to language, and the patterns are scattered geographically, so these clearly represent parallel innovations. There is even one language, Lewo (Vanuatu; Early 1994) in which there is tripartite negative marking, with a single preverbal constituent and two postverbal negative markers, e.g.

Pe

wii re

poli

NEG1

water NEH2

NEG3

‘There is no water.’

Some languages express negation by means of a negative verb, with the negated verb being expressed as a complement to this. Thus in Southwest Tanna (Vanuatu) the complement is a nominalisation:

k-a-s-apwah

n-⁨vn⁨ŋ-ien

k-a-s-⁨vŋin

3NSG-CONT-PL-not

NOM-eat-NOM

SNSG-CONT-PL-eat

‘They are not eating.’

‘They are eating.’

In Fijian, the negated clause is expressed as a subordinate clause:

Au

na

sega

ni

lako

mai.

1SG

FUT

NEG

SUBORD

come

hither

‘I will not come.’

5  Imperative and Interrogative Sentences

5.1  Imperative sentences

An imperative verb phrase often has no marking at all, or only a preverbal subject morpheme. Prohibitions usually display a different form of negation from declaratives.

5.2  Interrogative sentences

Polar questions are generally expressed simply through an intonation change from declaratives, or a following questioning interjection. Except in TVX languages (§4.2) content questions typically involve an interrogative marker that does not move from the structural slot of the questioned constituent. Generally, interrogatives belong in the same word class as the questioned constituent, though interrogative verbs meaning ‘do what’, and sometimes ‘do how’, are encountered. Thus, we find in Paamese:

Ki-hiteni he-mukave?

2SG:DIST.FUT-say 3SG:DIUST.FUT-do.how

‘How will you say it?’

6  Complex Sentences

Oceanic languages generally do not have especially complicated systems of overt marking of subordination, and subordinate markers often perform other functions in these languages. Relative clause markers, for example, are often similar or identical in shape to demonstratives, and reason clauses are often expressed by means of a causal preposition. There is often a single subordinator that expresses a wide range of subordinating functions. It is not uncommon for clauses to be simply juxtaposed without any linking morphemes at all. The structural relationship between clauses may be shown instead by interdependence in inflectional marking between main and subordinate clauses, with the range of categories that are expressed in subordinate clauses typically being a subset of those encountered in main clauses. Conjoined sentences are generally linked by a small set of conjunctions. There is widespread use of a lexical verb meaning ‘say’ marking subordinate clauses to verbs of locution or perception.

Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.