1

Morphology in Austronesian languages

Theodore Levin and Maria Polinsky

Keywords: Austronesian languages, Formosan languages, Malayo-Polynesian languages,

Oceanic languages, infix, prefix, suffix, reduplication, root, stem, voice, category

neutrality, applicative, causative

Summary

This chapter presents on overview of the major morphological properties of Austronesian

languages. We present and analyze data that may bear on commonly discussed category

neutrality of Austronesian and suggest that Austronesian languages do differentiate

between core lexical categories. The chapter addresses the difference between roots and

stems showing that Austronesian roots are more abstract than roots traditionally

discussed in morphology. Austronesian derivation and inflexion rely on suffixation and

prefixation; infixation is also attested. Austronesian languages make extensive use of

reduplication. In the verbal system, main morphological exponents mark voice

distinctions as well as causatives and applicatives. In the nominal domain, the main

morphological exponents include case markers, classifiers, and possession markers.

Overall verb morphology is richer in Austronesian than nominal morphology. The

chapter also presents a short overview of empirically and theoretically challenging issues

in Austronesian morphology: the status of infixes and circumfixes, the difference

between affixes and clitics, and the morphosyntactic characterization of voice

morphology.

1 Setting the stage

This chapter presents on overview of major morphological properties of Austronesian

languages (AN below). AN languages are usually divided into two branches: Formosan

and extra-Formosan or Malayo-Polynesian. Malayo-Polynesian (MP below) further

divides into two primary branches, Western Malayo-Polynesian (WMP) and Central-

Eastern Malayo-Polynesian (CEMP); see Blust (2013) for a detailed discussion.

WMP languages sre spoken in Madagascar, Malaysia, Indonesia, the Philippines, parts of

Taiwan, Thailand, Vietnam, and Cambodia.

AN languages are predominantly agglutinative. Within Western MP, the isolating

type develops in Chamic under contact with the neighboring Mon-Khmer languages. In

the Oceanic branch, languages spoken further east generally have more impoverished

morphology. The AN family is very large (over 1,200 languages, according to the

Ethnologue) and spans an immense geographical area, from Madagascar and Southeast

Asia to the Pacific, so any generalizations concerning the whole family are bound to

remain rather shallow. Nevertheless, a number of properties are characteristic of AN

languages: the relevance of stems in word-formation and inflection; the predominance of

prefixes over suffixes; productive infixation, extensive reduplication; articulated voice

systems; and articulated systems of possession marking.

2

2 Main morphological units

In the morphological makeup of AN, researchers recognize roots, stems, affixes, and

clitics. Some researchers also distinguish a special class of particles but those are often

co-extensive with clitics. Distinguishing clitics and affixes is not always straightforward,

and different authors do not always agree on the differentiating criteria. In a number of

cases, the distinction between derivational and inflectional morphology is subject to

debate, especially in WMP (see De Guzman 1994 on Tagalog).

2.1 Roots and stems

Following Blust (2013:360), we use the term ‘stem’ to refer to freei morphemes that are

capable of being affixed, and the term ‘root’ for a smaller, more abstract, “submorphemic

unit that is defined by recurrent association but not by contrast”. Stems are also referred

to as ‘bases’ in the literature. This definition differs slightly from the more commonly

adopted definition of root as a morpheme which is not an affix (Baker and Bobaljik 2008:

24).

A typical AN root has the shape –CVC which carries a generalized meaning and can

combine with so-called “formatives” (thematic affixes) to produce stems. Because of their

generalized meaning, most AN roots are category-neutral and can be used to derive

nouns, verbs, adjectives, etc. This category neutrality may be perceived as a sign that AN

languages lack a noun-verb distinction (Broschart 1997; Gil 2000, 2009; Kaufmann

2009). For instance, the Tongan manako can be used verbally (‘to like, want’), nominally

(‘desire’), adverbially (‘willingly’), or adjectivally (‘wanted, desirable’).

However, some operations in AN syntax are sensitive to the noun-verb distinction.

Using Tongan again as an example, the transitivizing suffix -'i can combine only with

those stems that are lexically specified as verbal, cf. the contrast between (1a) and (1b).

For a non-verbal stem to combine with -'i it first needs to be verbalized with the prefix

faka- which forms stative verbs from nouns or adjectives; these verbs can then be

transitivized by adding -'i, (2b).

(1) a. manako ‘want’ -- manako’i ‘desire’ Tongan

b. motu ‘island’ -- *motu'i

(2) a. motu ‘island’ – faka-motu ‘live an island way of life’ Tongan

b. faka-motu -- faka-motu'i ‘separate, cut someone off’

Co-occurrence with the prefix faka- is also category-sensitive; the prefix derives statives

from non-verbal stems and causative verbs from intransitive verbal stems (Churchward

1953: 253-255).

2.2 Affixes

AN languages are head-initial, and many of them, including languages spoken at the

geographical extremes of the family, are verb-initial or predicate-initial, i.e., VSO or

3

VOS. SVO as the neutral word order is found across Malayic and Micronesian languages

and some Melanesian languages. With the exception of some Western Melanesian

languages that have long been in contact with Papuan languages and have developed

verb-final orders, including postpositions (Lynch et al. 2011: 41, 49-50, 87; Crowley

2002: 37; Lichtenberk 1983b), no other word order types are considered basic within the

family. Consistent with the head-initial properties of AN languages, these languages also

have more prefixes than suffixes, especially in the MP branch. AN languages are also

noted for their use of productive infixation.ii

Compounding, though attested, is not a particularly productive form of

morphologically complex word-formation in AN languages. Therefore, we do not discuss

it in any detail here. It is found more frequently in Oceanic and eastern Indonesian

languages (see, e.g., Elbert and Pukui 1979, Crowley 1982 for discussion).

From a morpho-phonological standpoint, the relation between stems and derived

words is formally transparent, with few instances of ablaut (though see Section 5.2 for an

example), umlaut, suprasegmental or subtractive morphology (see Blust 2013: Ch. 6 for

discussion of each of these phenomena in AN). However, the process of affixation can at

time be obscured by other morpho-phonological alternations. The most common

processes observed in morphophonology include metathesis, nasal substitution

(Dempwolff 1934-1938; Blust 2013; Edwards 2018 on Amarasi; Pater 1999), and

reduplication (see 2.3 below). Apocope, vowel reduction, and haplology, though less

well-attested within AN also render certain affixations processes less than transparent.

2.2.1 Prefixation

Most AN prefixes are mono- or bi-syllabic and typically contain the a vowel. The most

common prefixes that appear on verbs across most AN languages mark the following

categories: inchoative, causative, stative, active voice, intransitive (or detransitivization),

instrumental voice, and collective action. As the present paper is intended to serve as a

wide overview of AN morphology a comprehensive list of these morphemes is not

provided (see Blust 2013: 371-381 for individual examples and reconstructions). The

prefix maN-, where N represents a homorganic nasal, is particularly widespread in

western MP as the marker of the active/agent voice, serving a similar function to the infix

<um>, discussed in Section 3.1. Examples of Indonesian meN- are presented below:

(3) a. meng-ajar ‘teach’ Bahasa Indonesia

b. men-curi ‘steal’

c. mem-baca ‘read’ (Sneddon 1996:13-4)

In some western MP languages, the equivalent of maN- is solely the homorganic nasal N-

(e.g., in Balinese, see Arka 2004).

Another widely observed affix is the prefix pa- and its cognates, widely used for

causative formation. In general, voice affixes of almost all Formosan and western MP

languages have both verbal and nominalizing functions, but the verbal functions are

primary (Blust 2013: 456). Valency-changing morphology such as voice and

causativization is discussed in more detail in Section 3 below.

4

Beyond the formation of verbs, AN languages use prefixation to derive ordinal

numerals from cardinal numerals, e.g., Marshallese ka-ruo ‘second’, Seediq (Taiwan)

tege-kingal ‘first’. Possession on nominals is often indicated by prefixes although for a

number of languages it is debatable whether these are actually prefixes or clitics (see

Sections 2.4 and 5.4).

2.2.2 Suffixation

The main suffixes observed throughout the family mark applicative (or circumstantial)

voice on verbs, see Section 3.1, based on the active (agent voice) form of the verb in

Malagasy:

(4) a. n-a-mono / n-a-mono-an' Malagasy

PST-AV-kill PST-AV-kill-APPL

‘killed’

b. n-i-vidi / n-i-vidi-anan'

PST-AV-buy PST-AV-buy-APPL

‘bought’

c. n-a-fina / n-a-fen-an'

PST-AV-hide PST-AV-hide-APPL

‘hid’

Because of the co-occurrence of the active prefix and the applicative suffix some

researchers have proposed to treat the applicative morphology as circumfixal (see fn. ii).

However, the applicative forms are clearly derived in a stepwise process with a fixed

ordering (Blust 2013:399). Across AN, patient voice and future tense suffixes are also

commonly attested. Again, a full list of frequently occurring suffixes is outside the scope

of this brief overview, but see Section 3 for further discussion of valency-changing

morphology.

Another common verbal suffix is the transitivizing -'i, widely attested in Oceanic

languages, see (1) and (2) above and discussion in Section 3.4.

Suffixation can also be found in the nominal domain. In Oceanic, we find suffixes

deriving nouns from verbs, e.g., -Vnga. Examples from the Kwaio language (Solomon

Islands) are given below.

(5) a. eno-nga Kwaio

sleep-NMLZ

‘sleeping’

b. leka-nga

go-NMLZ

‘trip’

c. to’oru-ngaa

live-NMLZ

5

‘living’ (Keesing 1985:77)

The choice of allomorph above is conditioned by the length of the base. Disyllabic stems

are nominalized with –nga; trisyllabic roots are nominalized with –ngaa (Lichtenberk

2011).

Finally, suffixes are commonly used to encode possessors on nouns; see Section 4.4

below for further discussion on possession marking.

2.2.3 Infixation

The two main infixes that recur across AN languages are the marker of the active (actor)

voice, which often has the inchoative meaning as well, and the marker of the perfective

aspect (which sometimes also functions as a nominalizer). The former is reconstructed as

the Proto-AN *-um-, the latter as *-in- (Blust 2013).iii Consider the Tagalog examples:

(6) a. lakad ‘walk’ – l<um>akad ‘to walk’ Tagalog

b. sulat ‘write’ – s<um>ulat ‘to write’

c. basa ‘read’ – b<um>asa ‘to read’

(7) a. lakad ‘walk’ – l<in>akad ‘to walk’

b. sulat ‘write’ – s<in>ulat ‘to write’

c. basa ‘read’ – b<in>asa ‘to read’

Productive infixation of the kind attested above is, for the most part, found in the

languages of Taiwan, the Philippines, and northern Borneo and Sulawesi (Adelaar and

Himmelmann 2005); but see Davis (2003) on the productive use of <in> as a nominalizer

in Hoava (Solomon Islands).

Infixation is sensitive to phonological environment; infixes are found with C-initial

stems. However, when the stem is V-initial the relevant morphemes may appear as

prefixes, cf. (6), (7) and (8):

(8) a. ulán ‘rain’ – um-ulán; in-ulán ‘to rain’ Tagalog

b. aral ‘teach’ – um-aral; in-aral ‘to teach’

This variability of placement, dictated by the phonological environment, has led many to

the conclusion that infixes are not a unique category of affix – distinct from prefixes or

suffixes – but, rather, that infixes should be modeled as prefixes or suffixes whose

realization within the root/stem is conditioned by phonological operations that occur after

affixation, e.g. metathesis (e.g. Halle 2001 and sources therein).iv

Tagalog speakers also show variability in infix placement in loanwords. Native words

in Tagalog lack initial consonant clusters. However, Spanish loanwords retain these

clusters. Some speakers place in the infix after the onset, while others intersperse the

infix between onset consonants (see Zuraw 2007 and references therein).

As the Tagalog data above illustrate, some languages show one allomorph of the

prefix/infix elements <um> and <in> regardless of placement (but see fn. iv for a

possible explanation). Other languages frequently show two (or more) allomorphs

6

depending on the position of the morpheme; most commonly, the infixal form is realized

as -VC-, but the prefixal form is realized as C-. Toba Batak (Sumatra) illustrates this for

<um>, as in taŋis ‘weeping’: t<um>a-taŋis ‘to weep, cry’, but inum ‘drinking’: m-inum

‘to drink’. Still other languages display different patterns of allomorphy for <um> and

<in> despite their identical -VC- form. Paiwan (Taiwan) retains the -VC- of <in> in

prefixal position, even in environments where <um> is realized as m- as in m-alap ‘take,

pick up’: in-alap ‘one’s catch in hunting’ (Blust 2013: 384).

2.3 Reduplication

Another environment where the interaction of morphology and phonology is particularly

clear is that of reduplication. Reduplication thus involves the repetition of all or part of a

root or stem sometimes accompanied by some phonological alteration(s) –As

reduplication is an instance of affixation, it is quite common to find that it interacts with

other instances of affixation occurring before some and after others. The ordering of

affixation can have consequences for the form and placement of the reduplicant.

Reduplication is probably the most pervasive morpho-phonological process in AN

languages, especially in the western part of the family. Full reduplication (repetition of

the entire root/stem) and/or partial reduplication (repetition of a subpart of the root/stem)

are productive processes in almost all western AN languages. The functions of

reduplicative processes are, however, quite variable across languages.

Full reduplication is often used to indicate distributivity or plurality as applied to

entities or to events; habitual or repeated activities; ongoing events; intensity and

emphasis; and increase/decrease of size/amount. Related to these two latter functions is

the use of reduplicated forms as augmentatives (correlating with the marking of an

increasing size or amount) and diminutives or pejoratives (correlating with the marking

of decrease). Some functions of full reduplication are illustrated below. In (9), full

reduplication expresses the plurality of entities, in (9) full reduplication expresses a

diminutive (or possibly a similarity relation), and in (9) it expresses approximation.

(9) a. rumah ‘house’; rumah-rumah ‘houses’ Botolan Sambal

b. anak ‘child’; anak-anak ‘doll’ Bahasa Malay

c. mbiriŋ ‘black’; mbiriŋ-mbiriŋ ‘blackish’ Karo Batak

(Blust 2013: 419-20)

AN languages also show a number of partial reduplication patterns. The full inventory of

attested patterns is too great to discuss in any detail here (see Blust 2013:406-31 for a

comprehensive overview), so we illustrate only with the more pervasive patterns of

partial reduplication below.

So-called Ca-reduplication occurs when the first consonant is copied and prefixed to

the base, along with a default vowel. This pattern of reduplication frequently forms

numerals and/or deverbal nouns. As indicated in the name of the phenomenon, the default

vowel is most often a, but it can also be a schwa or other vowel. Examples include

Puyuma (Taiwan) kədan ‘whet’: ka-kədan ‘whetstone’ (Blust 2013) or Balinese daar

‘eat’ : də-daar-an ‘food’ (Barber 1979, Clynes 1995). In vowel-initial bases this fixed

7

vowel is the reduplicant, as in Thao m-iup ‘blow on’: a-iup ‘tube used to blow on the fire’

(Chang 1998). CaC-reduplication is a variant of Ca-reduplication reported for Taba

(Makian, Kayoa, and southern Halmahera). This pattern also derives instrumental nouns,

as in tek ‘scoop up water’: tak-tek ‘water scoop’ (Bowden 2001).

CV(C)-reduplication – also called monosyllabic or syllable reduplication – occurs

when the reduplicant consists of a syllable prefixed to the base. This syllable can be

either light (monomoraic; necessarily CV) or heavy (bimoraic). In Philippine languages,

it is common to have both heavy and light syllable reduplication. Tagalog displays both

light and heavy CV(C)-reduplication, as in bili ‘buy’: pag-bi-bilí ‘selling’ and bili ‘buy’:

bí-bilí ‘will buy’. In light CV-reduplication, if the first syllable of the base is heavy, the

reduplicant is lightened as in Tagalog pag-la-lakbáy ‘travelling’ where the coda of the

stem syllable lak is not copied into the reduplicant. Conversely, in heavy CV(C)-

reduplication, if the first syllable of the base is light, the vowel in the reduplicant will be

stressed, as in the Tagalog example above, or lengthened,v or the syllable will be closed;

see Thurgood 1997 on Bontok (Luzon). As with Ca-reduplication, vowel-initial stems

copy only the vowel in the surface realization, as in Bunun (Taiwan) ma-a-asik ‘keep

sweeping’ or ma-u-uktic ‘keep cutting’ (Blust 2013: 425).vi The examples above show

that the reduplicant is usually a copy of some base-initial segment. However, there are

also examples where the reduplicant copies other parts of the base. In Madurese, the

base-final syllable is copied as in les-toles ‘write (more than once)’ or ku-buku ‘books’

(Davies 1999). A particularly striking instance of CVC-reduplication is discontinuous

reduplication where the reduplicant form is determined by segments of the base that are

not contiguous. For example, in Palawan, diminutive partial reduplication is obtained by

copying the first CV of the base and the final C of the base: bajuʔ ‘clothing’: bäʔ-bajuʔ

‘child’s clothing’, libun ‘woman’: lin-libun ‘girl’ (Revel-Macdonald 1979). Similar

patterns are reported for Nakanai (New Britain) (Johnston 1980) and Ulumuar Malay

(Nuger 2010).vii

Lastly, CV(C)CV-reduplication also called disyllabic or foot reduplication occurs

when the reduplicant copies a combination of two syllables (i.e. a foot) prefixed to the

base. The second syllable of the reduplicant is frequently monomoraic, regardless of the

shape of the corresponding syllable in the base. Lauje (Sulawesi) illustrates this contrast

in maale-alenda ‘rather long’ (Adelaar and Himmelmann 2005: 123). Given the size of

the reduplicant, foot reduplication can be indistinguishable from full reduplication unless

a base longer than two syllables is used. However, a fuller range of data shows that the

two patterns do diverge. In a number of AN languages, foot reduplication occurs in

addition to full base reduplication. Consider Lauje again, as in mong-ontong-ontong

‘watch (for some time)’. In full reduplication the second syllable is closed, unlike in foot

reduplication. Palauan exhibits a complex pattern of prefixal foot reduplication. The

reduplicant is C1eC1V(C2)- as in mə-saul ‘tired’: mə-sesu-saul ‘sort of tired’, mə-dakt

‘afraid’: mə-dedək-dakt ‘sort of afraid’ (Josephs 1975). Suffixal foot reduplication is

attested in Hawaiian (Pukui and Elbert 1971), for example, aloha ‘love, affection’: āloha-

loha ‘express affection’. Suffixal foot reduplication is also reported in Paiwan (Ferrell

1982) and Manam (Lichtenberk 1983b).

Given the prevalence of prefixation in the AN languages, it is unsurprising that most

reduplicants are prefixes, as well. However as the Hawaiian foot reduplication data above

show, reduplicant suffixation is also attested. Suffixal CV(C)-reduplication is reported in

8

Chamorro, as in ña.laŋ ‘hungry’: ñá.la.-laŋ ‘very hungry’ (Topping 1973) and Yapese

qa.thuk- ‘to mix’: maq.thuk.-thuk ‘mixed up’ (Jensen 1977). Thao, Rukai (Taiwan),

Pazeh (Chang 1998), Balinese (Clynes 1995), and Siraya (Taiwan) (Adelaar 2000) are

also reported to display suffixal reduplication. Reduplicative infixation is also reported.

CV-reduplicants appear infixed to the stem in Thao (Blust 2013) and Xârâcùù (New

Caledonia; Moyse-Faurie 1995). Examples fromcl the latter include atîrî ‘count on’:

atî<tî>rî ‘have confidence in’, and jikiè ‘rich’: ji<ki>kiè ‘very rich’ (see also Elbert

1988 on Rennellese).

Finally, some western AN languages have been claimed to permit triplication: the

affixation of a reduplicant two times. Examples from Thao include m-apa ‘carry’: apa-

apa-apa-n ‘be carried’, shkash ‘fear’: makit-shka-shka-shkash ‘slowly overwhelmed by a

sense of apprehension or foreboding’ (Blust 2001). Blust also provides the example

zumzum ‘hold in one’s mouth’: za-za-zumzum ‘keep holding in one’s mouth’. Here, Ca-

reduplication appears to have occurred twice. (Müller-Gotama 2001 provides possible

examples from Sundanese.)

2.4 Clitics

Traditionally, clitics are characterized as morphemes that show contradictory behavior in

their syntactic and phonological properties. From the point of view of syntax they are

independent elements; from the point of view of phonology, their realization depends on

another word or phrase.viii

A distinction can made between peripheral and second-position clitics. Peripheral

clitics appear at the beginning or end of the constituent to which they belong. These

clitics are found in almost all western AN languages. Second-position (Wackernagel)

clitics usually occur after the first constituent of the phrase to which they belong, and are

only attested in Formosan and western MP languages. Consider the following data from

Tagalog. Its basic word-order is verb-initial; therefore, second position clitics normally

occur immediately after the verb (10a,b). However, when a negative or other adverbial

element appears in pre-verbal position, as in (10c), the clitics will also precede the verb.

(10) a. Ibinigay=ko=na ang pera kay Charlie Tagalog

IV.PFV.give=1.SG.GEN=PRF CM money DAT Charlie

‘I already gave the money to Charlie.’

b. Nagtatrabaho=rin=ho=ba=kayo roon?

AV.IPFV.work=also=HON=Q=2.PL.NOM there

‘Are you working there too, sir?’

c. Hindi=pa=man=lamang=tuloy nakakapag-almusal si Juan.

NEG=IPFV=even=only=as.result eat-breakfast NOM Juan

‘As a result, Juan hasn’t even had breakfast yet.’ (Kroeger 1998:1-2)

Both peripheral and second-position clitics are usually unstressed and form a prosodic

unit with either the following word (proclitics) or the preceding word (enclitics). Because

of their positional variability, it is generally easy to distinguish second position clitics

from affixes. This distinction is less clear in the case of peripheral clitics. There are three

9

types of evidence employed to distinguish clitics from affixes: (i) Clitics generally do not

trigger morphophonological alternations of the stem. (ii) Clitics tend to be less selective

than affixes with regard to the category of their hosts. (iii) Clitics can show variability

with regard to their position within morphologically complex elements. The order of

affixes, however, is often rigid. Despite these diagnostics, there are some cases where the

clitic-affix distinction remains uncertain. With respect to pronominal clitics Blust (2013:

403) notes, “In general, atonic monosyllabic pronouns that attach to an independent word

are treated as clitics rather than affixes, although the basis for this decision is rarely

explicit.”

In terms of their meaning and functions, AN clitics represent the following main

types: (i) pronominal, (ii) aspectual/modal, (iii) negative, (iv) determiner, (v) emphatic,

and (vi) clitics which encode question markers,politeness markers, as well as mirativity

and evidentiality. In Tagalog, pronominal clitics are seen in =ko `1SG.GEN’ (10a) and

=kayo ‘2PL.NOM’ (10b); see also Section 5.3 for a fuller discussion of pronominal forms.

Aspectual clitics =na (10a) and =pa (10c) indicate perfective and imperfective aspect

respectively. Finally, the clitic =ba encodes a question in (10b).

3 Valency-changing morphology

Valency alternations involve the addition of logical arguments to or subtraction of those

arguments from a clause, as well as manipulations of the mapping of logical arguments

into different clausal constituents.

3.1 Voice and transitivity

3.1.1 Voice

Perhaps no morphosyntactic category has been as much investigated in AN languages as

voice. AN languages have constructions that resemble passives of more familiar Indo-

European languages, as well as antipassive constructions. However, so-called symmetric

voice systems are the primary focus of interest, and their analysis is still quite

controversial, as discussed below.

English-like passives are found in numerous AN languages, though they are rare in

Melanesian languages. The logical object is realized as subject, the logical subject is

demoted to an oblique, adpositional phrase, and passive morphology appears on the verb:

(11) a. saya di-jemput (oleh dia). Indonesian

1SG PASS-meet by 3SG

‘I was met by him.’ (Sneddon 1996:248)

b. i koohete-tia a Pani e Huia Māori

T/A scold-PASS PERS.DET Pani by Huia

‘Pani was scolded by Huia.’ (Bauer 1993:396)

10

Some languages allow such passivization freely and extend it to intransitive predicates as

well, cf. in Hawaiian:

(12) a. ua komo-hia ka mana'o i loko ona Hawaiian

PRF enter-PASS DET thought to inside 3SG.POSS

‘A thought occurred to him.’

b. ua hae-hia ka 'īlio

PRF bark-PASS DET dog

‘The dog was angry.’ (Elbert and Pukui 1979: 86)

In some cases the passive morphology on the verb is zero. For example, alongside the

marked passive as in (11a), Indonesian has a so-called bare passive (Chung 1976).

Altogether, in Indonesian and Malay, researchers distinguish morphologically-marked

active (with the prefix maN-), morphologically-marked passive with di-, bare active, and

bare passive. As a consequence of null morphology, it may be difficult to distinguish

passives from null argument constructions. For example, subjects and objects can be

dropped in Marshallese for some verb classes, making the analysis of the ‘passive’ in

(13) unclear (similarly in Hoava, see Davis 2003, and Fijian, with Schütz and Nawadra

1972 arguing against the passive analysis and Kikusawa 1998 arguing in its favor).

(13) John e=naaj mwij~mwij rainin. Marshallese

John 3SG.AGR=FUT operate-INTR today

(a) ‘John will operate today.’

(b) ‘John will be operated on today.’ (Willson 2010: 233, citing Bender 1969)

Even in languages with overt passive morphology, several strategies may be used (as in

Marshallese, Willson 2010: 238), or the marking may be ambiguous with other verbal

categories, such as transitive and perfective. (See Willson 2010 for Marshallesee, Cook

1996, Mosel and Hovdhaugen 1992: 198-204, 729-743 for Samoan.)

In the antipassive construction, it is the logical object that is demoted and may be left

unrealized, and again special morphology often appears on the verb (Polinsky 2017).

Antipassives are less commonly discussed amongst AN languages, in part because

different researchers often use different operational criteria by which they define the

antipassive and in some instances antipassives are conflated with intransitives or

durative/imperfective forms. Chamorro displays a transitive-antipassive alternation,

whereby the logical object is realized within an adpositional phrase and special

morphology (man- or fan- depending on mood) appears on the verb (Cooreman 1988,

Chung 1998). Even in languages where the presence of antipassives is not contested, their

formation is lexically restricted (cf. Chung 1998: 37-39 for Chamorro).

(14) a. Un-patek i ga'lagu. Chamorro

2S.SBJ-kick the dog

‘You kicked the dog.’

b. Mamatek hao gi ga'lago.

ANTIP.kick 2S.NOM LOC dog

‘You kicked at the dog.’ (Cooreman 1988: 578)

11

Pseudo noun-incorporation constructions and constructions with oblique object in

Polynesian languages, like Niuean and Tongan, have also been treated as antipassive

constructions (Polinsky 2017), as they have actor voice clauses in symmetrical voice-

systems (Aldridge 2012; see immediately below for discussion).

The symmetrical voice systems of Western AN languages are not limited to more

familiar active-passive or transitive-antipassive alternations, at least not

straightforwardly.ix In these systems, there is no morphologically unmarked voice, and no

argument is demoted to an oblique (Kroeger 1993, Foley 2008, Pearson 2005). Instead,

symmetrical voice systems are characterized by the fact that a single argument of the

clause – possibly a non-core argument – is privileged in certain ways. This argument may

be in a certain linear position or receive a particular morphological marking. Dedicated

morphology on the verb indicates which argument of the verb was chosen for this special

status. Furthermore, this privileged argument is often the only one that can be wh-

questioned, relativized, or focused—the restriction often referred to as the subject-only

constraint (Schachter and Otanes 1972, Keenan 1976, Schachter 1976).

Consider the symmetrical voice alternation in Squliq Atayal (Taiwan):

(15) a. M-aniq qulih qu' Tali'. Actor Voice

AV-eat fish QU Tali

‘Tali eats fish.’

b. Niq-un na' Tali' qu' qulih qasa. Patient Voice

eat-PV GEN Tali QU fish that

‘Tali ate the fish.’

c. Niq-an na' Tali' qulih qu' ngasal qasa. Locative Voice

eat-LV GEN Tali fish QU house that

‘Tali eats fish in that house.’

d. S-qaniq na' Tali' qulih qu' qway. Instrumental Voice

IV-eat GEN Tali fish QU chopsticks

‘Tali eats fish with chopsticks.’

(Erlewine et al. 2017: 374; citing Liu 2004)

In each example, one argument of the verb (in italics) is in sentence-final position

preceded by the marker qu’. Voice morphology on the verb (in bold) reflects this choice

of argument.

The number of voices in a given language varies. Philippine and Formosan languages

have been reported to have four or five distinct voices. Malagasy is usually analyzed as

having three voices. Malay/Indonesian languages are frequently reported to have just two

symmetrical voices, though they also have a more diverse inventory of active-passive

alternations, as well.

3.1.2 Transitivity

Generally speaking, transitivity refers to a predicate’s ability to take a direct object, or

“close object” as it is referred to in the literature on some Oceanic languages (Lynch et al.

2011). The issue of how transitivity is marked in AN languages, especially Formosan and

12

Western MP languages is particularly controversial and is connected to the status of

symmetrical voice systems (see also Section 5.1 on case-alignment). Some treatments

take voice affixes to be indicative a verb’s transitivity, while others maintain that voice

morphology is orthogonal to the issue of transitivity. This latter position is most clearly

supported in Split-S systems, like Balinese (Arka 2004) and Achenese (Legate 2014),

whereby verbs that are clearly intransitive – they cannot take a direct object – can appear

with either actor or patient voice forms depending on the thematic role of the sole

argument, for example,

(18) a. Ia ulung Balinese

3 OV.fall.down

‘(S)he fell down.’

b. Ia ng-eling

3 AV-cry

‘(S)he cried.’ (Arka 2004: 40)

Unergative predicates bear actor voice (18b). Unaccusatives bear patient/object voice

(18a).

The distinction is more clear-cut in Oceanic languages where symmetrical voice

systems do not arise. Oceanic languages frequently display formal marking for

transitivity on verbs. Transitive predicates are marked with a suffix that can be derived

from the Proto-Oceanic form *-i; see the Tongan examples in (1) and the following

examples from Tigak (Papua New Guinea):

(19) a. nonol ‘be thinking’ – nol-i ‘think about’ Tigak

b. visvis ‘fight’ – vis-i ‘hit’

c. kalkalum ‘look’ – kalum-i ‘see’ (Lynch 1998:85)

In Proto Oceanic, *-i was generally added to an intransitive root with a final consonant.

However, most Oceanic languages have lost these final consonants. Therefore, when the

consonant is retained before a transitive affix, it is treated as part of the suffix, for

example, consider Fijian:

(20) taŋi / taŋi-ða Fijian

weep / weep-TR.3SG.OBJ

'weep' 'cry for' (Lynch et al. 2011:44)

In Nakamanaga (Papua New Guinea), the transitivizing suffix is –gi, and in Lenakel

(Vanuatu) it is –in (Lynch 1998). Ulithian (Micronesia) shows considerable allomorphy

in the transitive suffix, including /-ðI/, /-lI/, /-ŋI/, /-sI/ and /-xI/. These forms also show

vowel harmony whereby /i/ becomes /u/ when affixed to a stem ending in schwa or a

back vowel. The language also shows the forms /-fi/, /-mi/, and /-ri/ which are insensitive

to the preceding vowel. All these variations are claimed to be lexically idiosyncratic

(Lynch et al. 2011).

In Xȃrȃcùù it is reported that a single suffix, –ri, is used to form both transitive and

applicative verbs (Moyse-Faurie 1995). A similar pattern is found in Saliba (Papua New

Guinea; Margetts 1999).

13

Many Oceanic languages also display a transitivity alternation that employs

reduplication (Lynch et al. 2011, Blust 2013). Here, intransitive forms are

morphologically more complex than their transitive counterparts, and are marked with

partial reduplication. This is illustrated in the Tolai language (Papua New Guinea);

compare iu ‘to wash’: iuiu ‘to bathe’, tumu ‘to write down’: tutumu ‘to write’, or kal ‘to

dig up’: kakal ‘to dig’ (Blust 2013: 684-685) and in Fijian, as in cula ‘sewn’: cula-cula

‘sew’ or rabe ‘kick’: rabe-rabe ‘do a lot of kicking’ (Dixon 1988).

3.2 Causativization

Causative constructions are valency-increasing constructions that serve to indicate that

the subject causes someone/something to bring about an event, causes a change in state,

or allows for the occurrence of an event or change of state. Cross-linguistically,

causativization can target both transitive and intransitive predicates, thought there are

language-specific constraints on the predicates that can be causativized.

Causativization is frequently signaled by reflexes of the proto-AN prefix *pa/paka-

(Blust 2013), as in Amis (Taiwan) pa-, and Roviana (Solomon Islands) va-.

(16) a. Pa-rakat kaku t-u paliding. Amis

CAUS-walk 1S.NOM DAT-CN car

‘I drive the car.’

b. Pa-ka’en kaku ci panay-an t-u pawli.

CAUS-eat 1S.NOM PPN Panay-DAT DAT-CN banana

‘I feed Panay banana.’ (Wu 2006: 298)

(17) a. Va-mae-a sa magu. Roviana

CAUS-come-3SG DEF knife

‘Give me the knife!’

b. Keke totoso va-la-i-u ri pa Vira Haba.

One time CAUS-go-TR-1SG LOC Vira Harbor

‘One time, they made me go to Vira Harbor.’ (Corston-Oliver 2011:482)

Zeitoun and Huang (2000), suggest that *paka- (attested in, e.g., Paiwan paka-,

Chamorro faha-, and Fijian vaka-) may be analyzed as multi-morphemic, consisting of

the causative prefix *pa- ‘causative’ and the stative prefix *ka- (see Blust 2013: 372-380

for further discussion).

3.3 Applicativization

Like causativization, applicativization is a valency-increasing operation. However, in this

case it is an oblique argument of the verb that is promoted to function as a grammatical

object also called the applied argument. The applied argument, sometimes called a

remote object in the Oceanic literature (Lynch et al. 2011: 44), can bear one of many

thematic roles including benefactee, malefactee, goal, location, instrument, etc. Some

languages utilize distinct morphemes across a number of applicative constructions.

14

Applicatives are divided into two groups cross-linguistically: high and low

(Pylkkänen 2008). High applicatives can occur with intransitive predicates, implicate no

relationship between the applied argument and the logical object and can treat either

argument as the grammatical object for the purposes of syntactic operations. Low

applicatives only occur with transitive predicates, they implicate a relationship between

the applied argument and the logical object (frequently one of possession), and the

applied argument is privileged for syntactic operations reserved for the grammatical

object.

In Oceanic languages, applicativization is signaled by the use of a suffix related to the

Proto Oceanic *-aki(ni) (Harrison 1982, Lynch et al. 2011), cf. in Niuean:

(18) a. Kua hele aki tuai e ia e titipi e falaoa. Niuean

PERF cut APPL PERF ERG 3SG ABS knife ABS bread

‘He has cut the bread with the knife.’

b. Kua hele tuai e ia e falaoa aki e titipi.

PERF cut PERF ERG 3SG ABS bread with ABS knife

‘He has cut the bread with the knife.’ (Seiter 1980: 243-4)

It is common for the form of the applicative affix to bear formal similarity to adpositions

in the same language. (18b) shows a prepositional oblique marked with aki. whereas.

(18a) shows that aki can appear within the predicate. When it does, the instrument, like

the theme, is marked with absolutive case.

It has also been suggested that additional symmetrical voices beyond agent and

patient can be analyzed as applicative constructions (Aldridge 2004). In these cases, an

adjunct or indirect object is introduced as the highest internal argument. The co-

occurrence of agent and/or patient voice morphology along with applicative morphology

is particularly clear in Malay/Indonesian languages, e.g. Balinese (Arka 2004) and

Bahasa Indonesia (Sneddon 1996). Javanese illustrates a pattern where the applicative

suffix –ake can appear with both active and passive constructions; this suffix freely

combines with active verbs, indicated by the homorganic nasal prefix N-, (19), or passive

verbs, with the prefix di, (19).

(19) a. Ani n-ulis-ake Tono layang Javanese

Ani ACT-write-APPL Tono letter

‘Ani wrote Tono a letter’

b. Tono di-tulis-ake Ani layang

Tono PASS-write-APPL Ani letter

‘Tono was written a letter by Ani’ (Nurhayani 2012:1-2)

15

4 Verbal agreement and related phenomena

4.1 Verbal agreement

Descriptively, agreement is a pattern of morphological variation on verbal elements

conditioned by the grammaticalfeatures (person, number, gender, case, etc.) of noun

phrases in a clause.

AN languages with agreement generally mark agreement with the subject (regardless

of their case-marking alignment; see Section 5.1 on alignment), and that marking is

achieved using two basic strategies: affixal verbal agreement, common, for example, in

Melanesian languages (see Yamada 2006 for representative languages and examples),

and clitics (in Micronesian, some dialects of Fijian, Rotuman, Polynesian, as well as

some Melanesian languages). Micronesian languages and some Melanesian languages

also show agreement with the object (see Song 1994 for an overview of the Nuclear

Micronesian data). Compare in Puluwat (Micronesia), which illustrates a common pattern

of agreement:

(20) Wuŕumwo ya yákékkél-ee-ŕ átekkit mákk Puluwat

Wurumwo 3SG.SUBJ teach-TRANS-3PL.OBJ children writing

‘Wurumwo taught the children writing.’ (Elbert 1974: 86)

In Melanesian languages subject-agreement markers are often portmanteau forms which

combine with the expression of the tense/aspect/mood categories of the verb (Lynch et al.

2011). In other languages, while subject agreement is independent of tense/aspect/mood,

its form is nevertheless conditioned by such specifications, e.g. Palauan (Georgopoulous

1991, Nuger 2016).

Verbal agreement is less well-attested in western MP languages, though see Legate

(2014) and sources therein on Acehnese; however a number of researchers have argued

that symmetrical voice morphology (discussed in Section 3.1.1), in fact, reflects a form of

agreement, namely case-agreement. On such accounts, voice reflects the outcome of

independent syntactic processes of case-assignment and syntactic movement (e.g.,

Pearson 2005; Rackowski 2002; Rackowski and Richards 2005).

Two questions are of interest in relation to the morphosyntax of agreement in AN.

First, a number of AN languages are pro-drop, but it is not yet clear if there is a

correlation between the availability of pro-drop and the availability of agreement. For

example, Micronesian languages, which have relatively rich agreement, also have pro-

drop, but so do agreement-poor Melanesian languages such as Cheke Holo (Palmer

2009a). Presumably, the conditions on subject drop or topic drop may differ depending

on whether or not a given language has agreement, but these conditions still need to be

better studied. The potential relationship between pro-drop and rich agreement has been

very widely investigated in non-AN languages, and AN languages have the potential to

inform the debate.

The second issue has to do with the categorical nature of agreement markers: are they

affixes or clitics?x For subject markers, it is generally assumed that they are clitics, often

on the basis of separability from the verb. For Rotuman, which on the surface seems to

employ subject suffixes, den Dikken (2003: Ch. 6), following Vamarasi (2002), argues

16

that these are also clitics. His main argument is that the apparent “pronominal suffixes”

attach to any element on their immediate left and do not select for a particular category of

host. He also cites diachronic evidence that some pronominal suffixes develop from

clitics. However, that does not necessarily mean that a clitic cannot change its category.

Object markers are generally assumed to be suffixes (cf. Song 1994).

As mentioned above, the distinctions between affixation and cliticization have lately

generated a lively debate in theoretical linguistics. We would like to emphasize the rich

empirical potential offered by AN languages in this area.

4.2 Wh-Agreement

Wh-agreement is a phenomenon whereby canonical verbal agreement is obscured or

altered when the argument cross-referenced by agreement is relativized, topicalized, or

focused. The pattern is found in a variety of AN languages, mostly in Western MP

languages, and has been well described for Chamorro (Chung 1982, 1994, 1998), Palauan

(Georgopoulos 1991), Tukang Besi (Sulawesi; Donohue 1999), and Malagasy (Pearson

2005).

The wh-agreement patterns of Chamorro can be used as a representative example of

the effect. In Chamorro transitive, realis clauses, subject agreement is realized as a prefix

on the verb as in (21), where ha- cross-references the subject si Juan. However, when the

subject is wh-questioned (21), ha- is not realized on the verb. Instead, the verb carries the

infix <um>. Similarly, if the object is wh-questioned (21), ha- again is not realized;

instead, the infix <in> appears.

(21) a. Ha-fa'gasi si Juan i kareta. Chamorro

3SG-wash DET Juan DET car

‘Juan washed the car’

b. Hayi f<um>a'gasi i kareta?

who WH.SUBJ.wash DET car

‘Who washed the car?’

c. Hafa f<in>a'gasése-nña si Henry pära hagu?

what WH.OBJ.wash.RED-3S DET Henry for you

‘What is Henry washing for you?’ (Pearson 2005:410)

Wh-agreement shows a clear relationship to other patterns of ‘anti-agreement’

phenomena cross-linguistically (see, e.g., Ouhalla 1993 on Berber and Baier 2017 for a

recent comprehensive overview). Furthermore, Pearson (2005) has argued that wh-

agreement systems are related to symmetrical voice systems (see Section 3.1.1). The

voice morphology of a language like Malagasy or Tagalog may, in fact, be wh-agreement

morphology. This connection is particularly clear in Chamorro. The Chamorro subject

wh-agreement marker <um>/mu- is cognate with the Tagalog actor voice marker <um>,

and the object wh-agreement marker <in> is cognate with the patient voice marker <in>

(Topping 1973, Donohue and Maclachlan 1999, Pearson 2005). The ubiquity of voice in

many Western MP languages is then attributed to the ubiquity of topicalization (or focus)

in every clause in these languages. Wh-agreement remains an active area of research in

AN linguistics.

17

5. Nominal morphology

5.1 Case

Generally AN nouns do not inflect for case. Case-marking is expressed by particles, if at

all, which can be analyzed as either case-marking clitics or prepositions. A number of

researchers specifically argue that they are prepositions (see Broschart 1994, which also

includes a review of earlier research). Two arguments support this conclusion. First, case-

marking particles are often homophonous with the actual prepositions used in a given

language. For example, the prepositions in Māori (New Zealand) are i ‘in, to’, ki ‘toward,

at’, e ‘from, by’, and a ‘of’. Of these, i also marks direct objects, ki, indirect (possibly

dative) objects, and e passive by-phrases. The second argument in favor of treating these

markers as prepositions and not pure case-markers, comes from the fact that the oblique

case forms do not combine with any prepositions, an unexpected distributional pattern if

some forms with case-markers are to be governed by prepositions. There is no consensus

in the literature on what these elements are, and their status is unlikely to be uniform

across different languages.

In Philippine, Formosan, and some Oceanic languages, the exact form of nominal

marking can be affected by definiteness and/or specificity and noun type. Frequently,

unique forms are used to mark proper names, and in some cases indefinite or non-specific

nouns are marked differently than definite/specific ones.

AN languages show a variety of case-alignment patterns. The majority of AN

languages show a nominative-accusative alignment: the subject of an intransitive verb

and the subject of a transitive verb appear in the same case (nominative), to the exclusion

of the object (accusative). Tahitian in (22) is representative. Subjects are unmarked, and

objects occur with the accusative particle 'i.

(22) a. te ma'ue nei te mau manu. Tahitian

ASP fly ASP DET PL bird

‘The birds are flying.’

b. 'ua 'ite te tamaiti 'i te mau manu.

PERF see DET child ACC DET PL bird

‘The child saw (the) birds.’

Numerous AN languages show neutral alignment: noun phrases show no overt case

marking. In such languages, grammatical functions are distinguished by agreement and/or

word order alone. Lewo (Vanuatu) is an example:

(23) a. sira Ø-puyu Ø-pa ne Ø-tol metava. Lewo

girl 3SG.SBJ-climb 3SG.SBJ-go now 3SG.SBJ-reach above

‘The girl climbed on up to the top.’

b. omami me-muni wii

1PL.EXCL 1PL.EXCL.SBJ-drink water

‘We drank water.’ (Early 1993: 70, 73)

18

Some AN languages are clearly ergative; the subject of an intransitive verb and the object

of transitive verb appear in the same case (absolutive), to the exclusion of the transitive

subject, which is marked ergative, for example,

(24) a. na'e 'alu 'a Sione ki he ako. Tongan

PST go ABS John to OBL school

‘John went to school.’

b. na'e tua'i 'e Sione 'a Mele.

PST call ERG John ABS Mary

‘John called Mary.’

Ergative alignment is found in Western Polynesian languages, Roviana (Corston 1996,

Corston-Oliver 2011), several languages of New Caledonia (Bril 1997, 2002; Moyse-

Faurie 1983; Moyse-Faurie and Ozanne-Rivierre 1983), and Melanesian languages such

as Motu (Lister-Turner and Clark 1930; Dixon 1994: 58), Hula (Pat 1996, Ball 2007),

and Sinaugoro (Tauberschmidt and Bala 1992).

Still other languages show a split-intransitive alignment whereby intransitive subjects

are not uniformly marked. The distinction frequently correlates with the

unergative/unaccusative distinction. Unergative subjects are marked like transitive

subjects, while unaccusative subjects are marked like transitive objects. Split-intransitive

analyses have been proposed for Acehnese (Durie 1985, 1987), Balinese (Arka 2004),

and Tukang Besi (Donohue 1999) among others.

A number of AN languages show alignment patterns that are difficult to describe in

these established terms. Philippine, Formosan languages, and Malagasy have been

particularly subject to controversy as to what kind of alignment they represent. Consider

the following examples from Tagalog:

(25) a. b<in>ili ng babae ang isda Tagalog

<ASP>buy NG woman ANG fish

b. b<um>ili ang babae ng isda

<ASP>buy ANG woman NG fish

‘The woman bought fish.’

c. d<um>ating ang babae

<ASP>arrive ANG woman

‘The woman arrived.’

The status of the markers ang (see also (10) above) and ng, which occur with common

nouns, is subject to an ongoing debate concerning the relationship between verbal

morphology and nominal-argument marking. When the verb takes the infix –in-, the

marker ang appears on the object; when the verb takes the marking –um-, ang appears on

the subject. Some researchers suggest that Tagalog and related Philippine languages are

ergative; the marker ang marks the absolutive case, while ng marks ergative and/or

oblique case (Payne 1982; De Guzman 1988; Manning 1996; Wechsler and Arka 1998;

Aldridge 2004, 2006, 2008; see Erlewine et al. 2017 for a recent review). On the other

hand, Rackowski (2002) and Rackowski and Richards (2005) propose that Tagalog is a

nominative-accusative language. Under this approach, the markers ang and ng are not

19

indicators of case at all.xi Instead, the main function of ang is then to mark the highest

structural argument in a given configuration. As mentioned in Section 4.2, a similar

approach is proposed for Malagasy by Pearson (2005).

5.2 Number and gender

Plurality of nominals can be indicated in a number of ways across AN languages. These

include affixation; reduplication, as in (26); and the use of special pluralizer words,

which appear to be similar to classifiers (Dryer 1989; Lynch at al. 2011: 174), cf. (26):

(26) a. pulau / pulau-pulau Bahasa Indonesia

island / RED-island

‘island’ / ‘islands’ (Sneddon 1996: 21)

b. e faiaoga / e tau faiaoga Niuean

ABS teacher / ABS PL teacher

‘(the) teacher’ / ‘(the) teachers’ (Seiter 1980: 38)

Polynesian languages distinguish singular and plural forms of nouns that refer to humans

by root modification. For instance, in Māori, the vowel undergoes ablaut to indicate

plurality; singular tangata ‘person’ and tuahine ‘(man’s) sister’ lengthen to form plural

tiingata ‘people’ and tuiihine ‘(man’s) sisters’ (Bauer 1993: 106; Lynch et al. 2011: 38).

However such pluralization strategies are optional. Given the appropriate context, plural

interpretations can be achieved in the absence of this morphology.

Very few AN languages have gender/noun class. In those languages that have gender

distinctions, the agreement in gender is registered on the determiner. For example, Teop

(Papua New Guinea) has three noun classes: e-class (personal names, people with high

social status, pets), a-class (all other humans, vertebrates, landmarks), and o-class (plants,

amorphous masses) (Mosel 2007). Some AN languages also show gender distinctions in

pronominal form. The Windesi dialect of Wandamen (Western New Guinea) is reported

to make an animate/inanimate distinction (Anceaux 1961). In the Mataram-Selong dialect

of Sasak (Lombok), the low register second person distinguishes male and female

addressee as ante (male), and kamu (female) (Austin 2000). In addition, Sellato (1981)

has reported a ‘three gender’ system of personal pronouns used by several groups of

nomadic Punan who inhabit the Muller-Schwaner Mountains of southeast Borneo.

5.3 Pronominal forms

Most AN languages make a distinction between inclusive and exclusive forms in the 1st

person plural. The former groups the speaker and the addressee, the latter groups the

speaker and others to the exclusion of the addressee(s). This distinction is found in all

paradigms of person pronouns (subject, object, possessor, etc.). Some exceptions to this

generalization can be found in Western MP languages (see Stevens 1968, Donohue and

Smith 1998, Robson 2002). In languages without an inclusive/exclusive distinction,

pronominal number distinctions have for the most part collapsed entirely, as in Javanese

and Madurese (Blust 2013).

20

Western MP pronominal systems usually do not include dual or trial forms – referring

to groups of two and three, respectively. Instead, they frequently make only a

singular/plural distinction if one is made at all. Many Oceanic languages, however, do

make use of dual and sometimes trial/paucal pronouns. Paucal pronouns pick out small

groups, usually between 3 and 10. In most Oceanic languages, the dual is derived by

adding the numeral ‘two’, and the trial/paucal by adding the numeral ‘three’ to the

pronominal stem, although this can be obscured by phonological processes. The Boumaa

Fijian pronominal paradigm presented below illustrates both clusivity distinctions in the

1st person plural as well as a 4-way number distinction:

(27) Boumaa Fijian pronominal paradigm

SINGULAR DUAL PAUCAL PLURAL

INCL. EXCL. INCL. EXCL. INCL. EXCL.

1ST Yau 'eetaru 'eirau 'etatou 'eitou 'eta 'eimami

2ND i'o 'emudrau 'emudou 'emunuu

3RD 'ea (i)rau (i)ratou (i)ra

(Dixon 1988; Harley and Ritter 2002:494)

Other AN languages known to recognize more than a singular/plural pronominal number

distinction are found in central and western Borneo. Most languages in the area have a

dual form, and many have a trial form as well. Most striking of all are Kenyah languages

that show a five-way person distinction – singular, dual, trial, quadral and plural numbers

– in addition to having an inclusive/exclusive distinction (Wong and Mantenuto 2017;

Smith 2013, and further references therein).

5.4 Possession marking

AN languages commonly categorize entities as directly vs. indirectly possessed.

Semantically, direct possession corresponds roughly to inalienable possession, also

referred to as obligatory, inherent, subordinate, or realized possession. Beyond truly

inalienable entities like body parts, the semantics of (in)alienability is not entirely

predictable; it has been subject to much discussion in the literature (see Lynch 1973,

1997, Lichtenberk 1983a, 1985, 2005, Wilson 1982, Bickel and Nichols 2013, Nichols

and Bickel 2013, and references therein). Indirect possession includes everything that can

be alienably possessed, and is also referred to as dominant or unrealized possession.

Morphologically, the distinction in possession types is marked in diverse ways across

AN. In Drehu (New Caledonia), inalienable possession is marked by an affix on the head

noun indicating that the possessor and alienably possessed nouns have only a

freestanding possession marker:

(28) a. la pengö-ng / keme-hun Drehu

DET manner-INAL.1SG father-INAL.3PL

‘my manner’ ‘their father’

b. la ihnim i angeic

DET love PRP 3SG

21

‘his love’ (Moyse-Faurie 1983: 60-1)

Inalienable possession marking may be obligatory, with 3rd singular typically being the

default, citation form. Alienable possession marking is never obligatory.xii

In Polynesian languages, the distinction between inalienable and alienable possession

is represented as the contrast between two series, the o series corresponds roughly to

inalienable possession, and the a series corresponds roughly to alienable possession, for

example,

(29) a. te Pukapuka a Heremaia Māori

DET book A Jeremiah

‘The Book of Jeremiah’ (written by him)

b. te Pukapuka o Hōhua

DET book O Joshua

‘The Book of Joshua’ (written about him) (Bauer 1997: Ch. 12)

Within alienable possession, many Oceanic languages further distinguish several

categories based on salient properties of objects (see Lichtenberk 1983a for an overview

and Bender and Beller 2006 for an overview and historical reconstruction). The most

common, and rather simple, system is one that divides entities into food, drink, and

everything else. Micronesian languages have a more articulated classification (cf. Dyen

1965, Benton 1968 for Chuukese; Rehg 1981 for Pohnpeian; Lee 1975 for Kosrae).xiii

The classification into categories such as ‘food’, ‘drink’, ‘general’, etc., is encoded by

freestanding expressions inside the noun phrase which are indexed for the person and

number of the possessor; we will gloss them as CLF. Compare in Iaai (New Caledonia):xiv

(30) a. bele-n kәiә Iaai

CLF.DRINK-3SG.POSS water

‘his/her water’

b. hanii-ɲ wɔɔ

CLF.FOOD-3SG.POSS fish

‘his/her fish’

c. aɲi-n meie

CLF.GEN-3SG.POSS fire

‘his/her fire’ (Ozanne-Rivierre 1976: 189)

The actual category of these classificatory expressions has been subject to debate. Most

researchers agree that they are syntactically heads; their order in the noun phrase follows

the general headedness principles of a language. In most AN languages they precede the

noun denoting the possessum, although in VSO Micronesian languages they follow the

noun. Most researchers agree that these expressions are different from sortal and measure

classifiers familiar from such languages as Chinese or Thai (see, however, den Dikken

2003: Ch. 2 for an argument for them being more similar to the familiar classifiers than

one would assume).

22

Unlike the better-known Southeast Asian classifiers, AN classifiers are not obligatory

in counting, their inventory is more limited than that of familiar classifier languages, and

most importantly, they do not serve to individuate and atomize nouns (cf. also Palmer and

Brown 2007: 203).

Some researchers suggest that these words are a special closed class of nouns which

take pronominal possessive marking and nominal dependents (Palmer and Brown 2007;

Palmer 2009b). This approach relies on the parallelism between the classifiers and

inalienably possessed nouns, which are also indexed for the person and number of their

possessor using the same marking. The alternative, proposed by Lichtenberk (1983a,

2009), is that these expressions should be considered “relational classifiers”, thus

functional elements, whose main purpose is to individuate the relation between the

possessor and possessum under indirect possession. Lichtenberk’s main morpho-syntactic

argument against treating these words as nouns comes from the fact that they are

typically monosyllabic/monomoraic, while all other lexical nouns in Oceanic are

disyllabic and/or bimoraic. Thus, “classifiers” do not meet the minimal nominal word

criterion (Lichtenberk 2009: 385).

Further readings

An excellent overview of AN morphology, with extensive historical reconstructions, can

be found in Blust (2013: Chapter 6). Lynch et al. (2011: Chapter 3) offer an overview of

morphology in Oceanic languages, followed by several grammar sketches. Adelaar and

Himmelmann (2005: Chapter 5) presents an overview of morphology in Formosan and

Western MP languages, also followed by several grammar sketches.

Abbreviations

Abbreviations follow the Leipzig Glossing Rules. Additional abbreviations: AV—agent

voice, CM—case marker, CN—common noun, HON—honorific, INAL—inalienable,

IV—instrumental voice, LV—locative voice, OV—object voice, PERS—personal,

RED—reduplicated, SV—subject voice.

Acknowledgements

We would like to thank Marcel den Dikken, Michael Yoshitaka Erlewine, and Paulina

Lyskawa for comments on previous versions of this chapter. Unless stated otherwise,

illustrative examples are from the authors’ field notes. This work was supported in part

by NSF grant BCS-1619857.

References

Adelaar, K.A. 2000. Siraya reduplication. Oceanic Linguistics 39: 33–52.

Adelaar, K.A., and N. Himmelmann, eds. 2005. The Austronesian languages of Asia and

Madagascar. London and New York: Routledge.

Aldridge, E. 2004. Ergativity and word order in Austronesian languages. Ph.D.

dissertation, Cornell University, Ithaca, NY.

Aldridge, E. 2006. Absolutive case in Tagalog. In Bunting, J., Desai, S., Peachey, R.,

Straughn, C., and Tomkov Z. (eds.), Proceedings from the 42nd Annual Meeting of

23

the Chicago Linguistic Society, Vol 1: The Main Session, 1-15. Chicago: Chicago

Linguistic Society.

Aldridge, E. 2008. Generative approaches to syntactic ergativity. Language and

Linguistics Compass 2/5: 966–995.

Aldridge, E. 2012. Antipassive and ergativity in Tagalog. Lingua 122: 192-203.

Anagnostopoulou, E. 2003. The syntax of ditransitives: Evidence from clitics. Berlin:

Mouton de Gruyter.

Anceaux, J.C. 1961. The linguistic situation in the islands of Yapen, Kurudu, Nau and

Miosnum, New Guinea. VKI 35. The Hague: Nijhoff.

Arka, I. W. 2004. Balinese morphosyntax: A lexical-functional approach. Canberra;

Pacific Linguistics.

Austin, P.K., ed. 2000. Sasak: Working Papers in Sasak, vol. 2. Melbourne: Department

of Linguistics and Applied Linguistics, The University of Melbourne.

Baier, N. 2017. Anti-agreement. Doctoral dissertation, University of California, Berkeley.

Baker, M. C. and J. D. Bobaljik. 2008. Generative Morphology. Ms., Rutgers University

and the University of Connecticut.

Ball, D. 2007. On ergativity and accusativity in Proto-Polynesian and Proto-Central-

Pacific. Oceanic Linguistics 46: 128-153.

Barber, C.C. 1979. Dictionary of Balinese-English. 2 vols. Aberdeen University Library

Occasional Publications, No. 2. University of Aberdeen.

Bauer, W. 1993. Māori. London: Routledge.

Bauer, W. 1997. The Reed reference grammar of Māori. Auckland: Reed.

Bender, A., and S. Beller. 2006. Numeral classifiers and counting systems in Polynesian

and Micronesian languages: Common roots and cultural adaptations. Oceanic

Linguistics 45: 380-403.

Bender, B. 1969. Spoken Marshallese: An intensive course with grammatical notes and

glossary. Honolulu: University of Hawai‘i Press

Benton, R.A. 1968. Numeral and attributive classifiers in Trukese. Oceanic Linguistics 7:

104-146.

Benton, R.A. 1971. Pangasinan reference grammar. Honolulu: University of Hawaii

Press.

Bickel, B., and J. Nichols. 2013. Obligatory possessive inflection. In M. Dryer, and M.

Haspelmath, Martin (eds.) The World Atlas of Language Structures Online. Leipzig:

Max Planck Institute for Evolutionary Anthropology. Available online at

http://wals.info/chapter/58, Accessed on 2017-10-05.

Blust, R. 2001. Thao triplication. Oceanic Linguistics 40: 324-335.

Blust, R. 2002. Notes on the history of ‘focus’ in Austronesian languages. In F. Wouk

and M. Ross (eds.), The history and typology of Western Austronesian voice systems.

Canberra: Pacific Linguistics.

Blust, R. A. 2013. The Austronesian languages. Revised edition. Canberra: Australian

National University.

Bowden, J. 2001. Taba: Description of a South Halmahera language. Canberra: Pacific

Linguistics.

Bril, I. 1997. Split ergativity in the Nêlêmwâ language. In C. Ode and W. Stockhof (eds.).

Proceedings of ICAL-7, 377-394. Leiden: Leiden University.

24

Bril, I. 2002. Le nêlêmwa (Nouvele-Calédonie): Analyse syntaxique et sémantique.

Louvain: Peeters.

Broschart, J. 1994. Praepositionen im tonganischen: Zur Varianz und Invarianz des

Adpositionsbegriffs. Bochum: Brockmeyer.

Broschart, J. 1997. Why Tongan does it differently: Categorial distinctions in a language

without nouns and verbs. Linguistic Typology 1, 123-165.

Chang, M.L. 1998. Thao reduplication. Oceanic Linguistics 37: 277-297.

Chung, S. 1976. On the subject of two passives in Indonesian. In C.N. Li (ed.), Subject

and topic, 57-98. New York: Academic Press.

Chung, S. 1994. Wh-agreement and “referentiality” in Chamorro. Linguistic Inquiry 25:

1-44.

Chung, S. 1998. The design of agreement. Chicago: University of Chicago Press.

Churchward, C.M. 1953. Tongan grammar. London: Oxford University Press.

Clynes, A. 1995. Topics in the phonology and morphosyntax of Balinese based on the

dialect of Singaraja, North Bali. PhD thesis, Australian National University.

Cook, K. 1996. The Cia suffix as a passive marker in Samoan. Oceanic Linguistics 35:

57-76.

Cooreman, A. 1988. The antipassive in Chamorro: Variations on the theme of transitivity.

In M. Shibatani (ed.) Passive and voice, 561-594. Amsterdam: John Benjamins.

Corston, S. 1996. Ergativity in Roviana, Solomon Islands. Canberra: Australian National

University.

Corston-Oliver, S. 2011. Roviana. In Lynch et al., 467-497.

Crowley, T. 1982. The Paamese language of Vanuatu. Canberra: Pacific Linguistics.

Crowley, T. 2002 Serial verbs in Oceanic: A descriptive typology. Oxford: Oxford

University Press.

Davies, W. 1999. Madurese. Munchen: Lincom Europa.

Davis, K. 2003. A grammar of the Hoava language, western Solomons. Canberra: Pacific

Linguistics.

De Guzman, V.P. 1988. Ergative analysis for Philippine language. In R. McGinn (ed.)

Studies in Austronesian linguistics, 323-345. (Ohio University Monographs in

International Studies, Southeast Asia Series, No. 76.) Athens, Ohio: Ohio University

Center for International Studies.

De Guzman, V.P. 1994. Verbal affixes in Tagalog: inflection or derivation? S.I.: no

publisher.

Dempwolff, O. 1934-1938. Vergleichende Lautlehre des austronesischen Wortschatzes. 3

vols. Berlin: Reimer.

den Dikken, M. 2003. The structure of the noun phrase in Rotuman. Munchen: Lincom

Europa.

Dixon, R.M.W. 1988. A grammar of Boumaa Fijian. Chicago: University of Chicago

Press.

Dixon, R.M.W. 1994. Ergativity. Cambridge: Cambridge University Press.

Donohue, M. 1999. A grammar of Tukang Besi. Berlin: Mouton de Gruyter.

Donohue, M., and A. Maclachlan. 1999. What agreement in Chamorro? In C. Smallwood

and C. Kitto (eds.) Proceedings of AFLA VI, 121-132. Toronto: University of Toronto

Working Papers.

25

Donohue, M., and J. C. Smith. 1998. What’s happened to us? Some developments in the

Malay pronoun system. Oceanic Linguistics 37: 65–84.

Dryer, M. 1989. Plural words. Linguistics 27: 865–895.

Durie, M. 1985. A grammar of Acehnese on the basis of a dialect of North Aceh.

Dordrecht: Foris.

Durie, M. 1987. Grammatical relations in Acehnese. Studies in Language 11: 365–399.

Dyen, I. 1965. A sketch of Trukese grammar. Washington, DC: American Oriental

Society.

Early, R. 1993. Layer serialization in Lewo. Oceanic Linguistics 32: 65-93.

Edwards, O. 2018. The morphology and phonology of metathesis in Amarasi.

Morphology 28: 25-69.

Elbert, S.H. 1974. Puluwat grammar. Canberra: Australian National University.

Elbert, S.H. 1988. Echo of a culture: A grammar of Rennell and Bellona. Honolulu:

University Press of Hawaii.

Elbert, S.H., and M. Pukui. 1979. Hawaiian grammar. Honolulu: University Press of

Hawaii.

Erlewine, M., T. Levin, and C. van Urk. 2017. Ergativity and Austronesian-type voice

systems. In J. Coon, D. Massam, and L. Travis (eds.) The Oxford handbook of

ergativity, 373-391. Oxford: Oxford University Press.

Ferrell, R. 1982. Paiwan dictionary. Canberra: Pacific Linguistics

Foley, W.A. 2008. The place of Philippine languages in a typology of voice systems. In

P. Austin and S. Musgrave (eds.) Voice and grammatical relations in Austronesian

languages, 22-44. Stanford: CSLI.

Georgopoulous, C. 1991. Syntactic variables: Resumptive pronouns and A’-binding in

Palauan. Dordrecht: Kluwer.

Gil, D. 2000. Syntactic categories, cross-linguistic variation and universal grammar. In

P. Vogel and B. Comrie (eds.) Approaches to the typology of word classes, 173–216,

Berlin: de Gruyter

Gil, D. 2009. Austronesian nominalism and the thinginess illusion. Theoretical

Linguistics 35: 95-114.

Halle, M. 2001. Infixation vs. onset metathesis in Tagalog, Chamorro, and Toba Batak. In

M.J. Kenstowicz (ed.) Ken Hale: A life in language, 153–168. Cambridge, MA.: MIT

Press.

Harizanov, B. 2014. Clitic doubling at the syntax-morphology interface: A-movement

and morphological merger in Bulgarian. Natural Language and Linguistic Theory 32:

1033-1088.

Harley, H., and E. Ritter. 2002. Person and number in pronouns: A feature-geometric

analysis. Language 78: 482-526.

Harrison, S. P. 1982. Proto-Oceanic *aki(ni) and the Proto-Oceanic periphrastic

causatives. In A. Halim, L. Carrington and S.A. Wurm (eds.) Papers from the Third

International Conference on Austronesian Linguistics, I: Currents in Oceanic, 179-

230. Canberra: Pacific Linguistics.

Jensen, J.T. 1977. Yapese reference grammar. Honolulu: University Press of Hawaii.

Johnston, E. L. 1980. Nakanai of New Britain: The grammar of an Oceanic language.

Canberra: Australian National University.

26

Josephs, L.S. 1975. Palauan reference grammar. Honolulu: The University Press of

Hawaii.

Kaufmann, D. 2009. Austronesian nominalism and its consequences: A Tagalog case

study. Theoretical Linguistics 35: 1-49.

Keenan, E.L. 1976. Remarkable subjects in Malagasy. In C. N. Li (ed.) Subject and topic,

247-301. New York: Academic Press.

Keesing, R.M. 1985. Kwaio grammar. Canberra: Pacific Linguistics.

Kikusawa, R. 1998. A formal analysis of the so-called ‘passive’ in Fijian. Journal of

Asian and African Studies 56:111-139.

Kramer, R. 2014. Clitic doubling or object agreement: The view from Amharic. Natural

Language and Linguistic Theory 32: 593-634.

Kroeger, P. 1993. Phrase structure and grammatical relations in Tagalog. Stanford:

CSLI.

Lee, K.-D. 1975. Kusaiean reference grammar. Honolulu: University Press of Hawai

Legate, Julie Anee. 2014. Voice and v: Lessons from Acehnese. Cambridge, MA: MIT

Press.

Li, P.J.K., and S. Tsuchida. 2001. Pazih dictionary. Taipei: Institute of Linguistics,

Academia Sinica.

Lichtenberk, F. 1983a. Relational classifiers. Lingua 60: 147-176.

Lichtenberk, F. 1983b. A grammar of Manam. Honolulu: University of Hawaii Press.

Lichtenberk, F. 1985. Possessive constructions in Oceanic languages and Proto-Oceanic.

In A. Pawley and L. Carrington (eds.). Austronesian linguistics at the 15th Pacific

Science Congress, 93-140. Canberra: Australian National University.

Lichtenberk, F. 2005. Inalienability and possessum individuation. In Z. Frajzyngier, A.

Hodges and D. S. Rood (eds.) Linguistic diversity and language theories, 339-362.

Amsterdam—Philadelphia: John Benjamins.

Lichtenberk, F. 2008. A grammar of Toqabaqita. Berlin: Mouton de Gruyter.

Lichtenberk, F. 2009. Oceanic possessive classifiers. Oceanic Linguistics 48: 379-402.

Lichtenberk, F. 2011. Nominalizations in Toqabaqita and closely related languages. In F.

H. Yap, K. Grunow-Hårsta, and J. Wrona (eds.), Nominalization in Asian Languages:

Diachronic and typological perspectives, 687-720. John Benjamins; Philadelphia.

Lister-Turner, R., and J.B. Clark. 1930. A grammar of the Motu language of Papua. 2nd

edition. Sydney: Government Printer.

Liu, A. K-L. 2004. On relativization in Squliq Atayal. Master’s thesis, National Tsing

Hua University.

Lynch, J. 1973. Verbal aspects of possession in Melanesian languages. Oceanic

Linguistics 12: 69-102.

Lynch, J. 1997. On the origins of the possessive markers in Central Pacific languages.

Oceanic Linguistics 36: 227-246

Lynch, J. 1998. Pacific languages: An introduction. Honolulu: University of Hawaii

Press.

Lynch, J., M. Ross, and T. Crowley. 2011. The Oceanic languages. New York:

Routledge.

Manning, C. D. 1996 Ergativity: Argument structure and grammatical relations.

Stanford: CSLI.

27

Margetts, A. 1999. Valence and transitivity in Saliba, an Oceanic language of Papua

New Guinea. (MPI Series in Psycholinguistics, 12.). Nijmegen: MPI.

Mosel, U. 2007. Teop sketch grammar. Ms. Kiel University. http://www.linguistik.uni-

kiel.de/mosel_publikationen.htm#download

Mosel, U., and E. Hovdhaugen. 1992 Samoan reference grammar. Oslo: Scandinavian

University Press.

Moyse-Faurie, C. 1983. Le drehu: Langue de Lifou (Iles Loyauté). Paris: Société d'Etudes

Linguistiques et Anthropologiques de France.

Moyse-Faurie, C. 1995. Le xaracuu: langue de Thio-Canala (Nouvelle-Caledonie) –

Eléments de syntaxe. (SELAF: Langues et Cultures du Pacifique 10.) Paris: Peeters.

Moyse-Faurie, C., and F. Ozanne-Rivierre. 1983. Subject case markers and word order in

New Caledonia and Loyalty Islands languages. In Papers from the Third

International Conference on Austronesian Linguistics, vol. 4: Thematic Variation,

113-152. Canberra: Australian National University.

Müller-Gotama, F. 2001. Sundanese. München: Lincom Europa.

Nevins, A. 2011. Multiple agree with clitics: Person complementarity vs. omnivorous

number. Natural Language and Linguistic Theory 29: 939-971.

Nichols, J., and B. Bickel. 2013. Possessive classification. In M. Dryer and M.

Haspelmath (eds.) The World Atlas of Language Structures Online. Leipzig: Max

Planck Institute for Evolutionary Anthropology. Available online at

http://wals.info/chapter/59, Accessed on 2017-10-07.

Nuger, J. 2010. Discontinuous reduplication in a local variety of Malay. In R. Mercado,

E. Potsdam, and L. Travis (eds.) Austronesian and theoretical linguistics, 45–64.

Amsterdam: John Benjamins.

Nuger, J. 2016. Building predicates. Dordrecht: Springer.

Nurhayani, I. 2012. Javanese applicative construction. In Coyote Papers 19.

Ouhalla, J. 1993. Subject-extraction, negation and the anti-agreement effect. Natural

Language and Linguistic Theory 11:477–518

Ozanne-Rivierre, F. 1976. Le Iaai: Langue mélanésienne d'Ouvéa (Nouvelle-Calédonie):

Phonologie, morphologie, esquisse syntaxique. Paris: Société d'Etudes Linguistiques

et Anthropologiques de France.

Palmer, B. 2009a. Clause order and information structure in Cheke Holo. Oceanic

Linguistics 48: 213-249.

Palmer, B. 2009b. Kokota grammar. Honolulu: University of Hawai’i Press.

Palmer, B., and D. Brown. 2007. Heads in Oceanic indirect possession. Oceanic

Linguistics 46: 199-209.

Pat, F. 1996. Transitive constructions in Hula. In J. Lynch and F. Pat (eds.) Oceanic

studies: Proceedings of the First International Conference on Oceanic Linguistics,

271-292. Canberra: Australian National University.

Pater, J. 1999. Austronesian nasal substitution and other NC effects. In R. Kager and H.

van der Hulst (eds.) The prosody-morphology interface, 310-343. Cambridge:

Cambridge University Press.

Payne, T. E. 1982. Role and reference related subject properties and ergativity in Yup’ik

Eskimo and Tagalog. Studies in Language 6: 75–106.

Pearson, M. 2005. The Malagasy subject/topic as an A’-element. Natural Language and

Linguistic Theory 23: 381-457

28

Polinsky, M. 2017. Antipassive. In J. Coon, D. Massam, and L. Travis (eds.) The Oxford

handbook of ergativity, 308-331. Oxford: Oxford University Press.

Preminger, O. 2009. Breaking agreements: distinguishing agreement and clitic-doubling

by their failures. Linguistic Inquiry 40: 619–666.

Pukui, M. K., and S. H. Elbert. 1971. Hawaiian dictionary: Hawaiian-English, English-

Hawaiian. Honolulu: University of Hawaii Press.

Pylkkänen, L. 2008. Introducing arguments. Cambridge, MA: MIT Press.

Rackowski, A. 2002. The structure of Tagalog: Specificity, voice, and the distribution of

arguments. Ph.D. dissertation, MIT.

Rackowski, A., and N. Richards. 2005. Phase edge and extraction: A Tagalog case study.

Linguistic Inquiry 36: 565-599.

Rajemisa-Raolison, R. 1971. Grammaire malgache. Fianarantsoa: Librairie

Ambozontany

Rehg, K.L. 1981. Ponapean reference grammar. Honolulu: The University Press of

Hawaii.

Revel-Macdonald, N. 1979. Le Palawan (Philippines): phonologie, catégories,

morphologie. (Langues et civilisations de l’Asie du sud-est et du monde insulindien,

No. 4.) Paris: SELAF.

Robson, S. 2002. 2002. Javanese grammar for students. Second, revised edition.

Victoria, Australia: Monash University Press.

Ross, M.D., and S. Teng. 2005. Formosan languages and linguistic typology. Language

and Linguistics 6: 739- 781.

Schachter, P. 1976. The subject in Philippine languages: Topic, actor, actor-topic, or none

of the above? In C.N. Li (ed.) Subject and topic, 491-518. New York: Academic

Press.

Schachter, P., and F.T. Otanes. 1972. Tagalog reference grammar. Berkeley: University

of California Press.

Schütz, A., and T. Nawadra. 1972. A refutation of the notion of “passive” in Fijian.

Oceanic Linguistics 11: 88–109.

Seiter, W. 1980. Studies in Niuean syntax. New York: Garland.

Sellato, B.J.L. 1981. Three-gender personal pronouns in some languages of central

Borneo. Borneo Research Bulletin 13: 48-49.

Smith, A.D. 2013. Reconstructing Proto Kenyah pronouns and the development of a true

five number system. Paper presented at ICAL-13, Taipei, Yaiwan.

Sneddon, J. N. 1996. Indonesian: A comprehensive grammar. New York: Routledge.

Song, J. J. 1994. The verb-object bonding principle and the pronominal system: With

special reference to Nuclear Micronesian languages. Oceanic Linguistics 33: 517-

565.

Stevens, A.M. 1968. Madurese phonology and morphology. (American Oriental Series,

52.). New Haven: American Oriental Society.

Tauberschmidt, G., and A. Bala. 1992 Transitivity and ergativity in Sinaugoro.

Language and Linguistics in Melanesia 23 : 179-191.

Thurgood, E. 1997. Bontok reduplication and prosodic templates. Oceanic Linguistics 36:

135-148.

Topping, D.M. 1973. Chamorro reference grammar. Honolulu: University Press of

Hawaii.

29

Vamarasi, Marit 2002 Rotuman subject suffixes and stative-active distinction. Paper

presented at the 5th International Conference on Oceanic Languages. Canberra,

January 2002.

Wechsler, S., and W. Arka. 1998. Syntactic ergativity in Balinese: An argument structure

based theory. Natural Language and Linguistic Theory 16: 387-442.

Willson, H. 2010. Passives and resultatives in Marshallese. Oceanic Linguistics 49: 233-

258.

Wilson, W.H. 1982. Proto-Polynesian possessive marking. Canberra: Australian National

University.

Wong, D., and I. Mantenuto. 2017. Quanifiers in Kenyah Uma Baha. In D. Paperno and

E.L. Keenan (eds.) Handbook of quantifiers in natural language. Volume 2, 939-962.

Cham, Switzerland: Springer.

Wu, J. 2006. The Analysis of pa- verbs in Amis. In H. Chang, et al. (eds.), Streams

converging into an ocean: Festschrift in Honor of Professor Paul Jen-Kuei Li, 279-

321, National Tsing Hua University, Taiwan.

Yamada, F. S. 2006. The pronoun system in Galeya: Arguments against a clitic analysis.

Oceanic Linguistics 45: 474-490.

Zeitoun, E., and L.M. Huang. 2000. Concerning ka-, an overlooked marker of verbal

derivation in Formosan languages. Oceanic Linguistics 39: 415-427.

Zuraw, K. 2007. The role of phonetic knowledge in phonological patterning: Corpus and

survey evidence from Tagalog. Language 83: 277-316.

Zwicky, A. 1985. Clitics and particles. Language 61: 283-305.

Zwicky, A., and G. Pullum. 1983 Cliticization vs. inflection: English n’t. Language 59:

502-513.

Theodore Levin Maria Polinsky National University of Singapore University of Maryland tedlevin@nus.edu.sg polinsky@umd.edu

ii A final affixal category that we do not discuss in detail here is the circumfix/confix, a

prefix-suffix unit. It is not immediately obvious whether these elements should be treated

as single morphemes or as combinations of two, prefix and suffix, that occur together

frequently or always. Some ‘true’ circumfixes are claimed to exist in the AN literature.

Schachter and Otanes (1972) describe Tagalog (the Philippines) pag- … -an as an

Object-Voice counterpart of Actor-Voice verbs prefixed by mag-, and Malagasy

(Madagascar) demonstratives (as discussed in Rajemisa-Raolison 1971, a.o.) have been

described as phrasal circumfixes.

iii See Blust (2013: 389-92) for discussion of the plural infix *-ar- (which is significantly

less well represented in modern AN languages).

iv This may, in fact, be an over-simplification of the Tagalog data because Tagalog has

been argued to lack true V-initial stems. Such stems are said to begin with a glottal stop

30

[Ɂ] that is not represented in the orthography (Schachter and Otanes 1972). Nevertheless,

other languages with unambiguously V-initial roots, like Chamorro (Guam, Northern

Mariana Islands) and Toba Batak, confirm this generalization.

v Typically AN CVC-reduplication is analyzed as involving length, not stress.

vi See Benton 1971 on a notable exception in Pangasinan (Luzon). In this language, the

first CV is reduplicated even if the stem is V-initial, yielding an infixed reduplicant, as in

amigo ‘friend’: a<mi>migo ‘friends’.

vii It is rare that both CV- and Ca-reduplication apply to the same set of lexical bases in a

given language (but see Chang 1998 and Blust 2001 on Thao (Taiwan) and Li and

Tsuchida 2001 on Pazeh (Taiwan)). Therefore, it is possible that Ca-reduplication and

CV-reduplication might be variants of one another.

viii In recent literature, a distinction has been made between phonological and syntactic

clitics (see e.g. Zwicky and Pullum 1983, Zwicky 1985, Anagnostopoulou 2003,

Preminger 2009, Nevins 2011, Kramer 2014, Harizanov 2014). For our present purposes,

we restrict our discussion to phonological clitics.

ix The ‘voice system’ of Western AN languages is also referred to as the ‘focus system’,

among other terms. See Blust (2002, 2013) and Ross and Teng (2005) for an overview of

terminological use in the literature.

x See Section 2.4 for additional discussion of cliticization in AN.

xi Regardless of the treatment of ang and ng, and related markers in other Western AN

languages, these languages also display case-markers for oblique elements.

xii The complete absence of possessive markers is unusual, and such languages are rare;

Toqabaqita (Solomon Islands) is a language with no such marking (Lichtenberk 2008).

xiii It seems that the more articulated Micronesian classification is subject to attrition. For

example, Benton and Dyen recognize over two dozen classificatory expressions in

traditional Chuukese. However, in her fieldwork with Chuukese conducted in 2000, the

second author of this chapter found only the following: general, inanimate mobile,

inanimate able to grow, small/intimate, drinkable, edible (raw), edible (cooked), animate

female, and animate male.

xiv The form of the possessive morpheme is phonologically conditioned (Ozanne-Rivierre

1976: 149); the allomorph -ɲ appears after the long e and after the long/short i, and unless

the consonant preceding i is palatalized, the allomorph -n appears elsewhere.