Home Possession in Patani
Article Open Access

Possession in Patani

  • Linn Iren Sjånes Rødvand EMAIL logo
Published/Copyright: September 18, 2023

Abstract

This paper provides the first description of the possessive system in Patani, a South Halmahera-West New Guinea language (Austronesian). Possession in Patani involves the interplay of several parameters. Syntactically, there are two broad possessive constructions: direct, and indirect. Morphologically, both constructions make use of pronominal proclitics and possessive suffixes which express the person and number of the possessor. In the direct construction, this possessive marking attaches directly to the noun, whereas the indirect construction makes use of a possessive particle (a relational classifier) on which the possessive marking occurs. Semantically, all nouns which are directly possessed are inalienable nouns expressing kinship, part-of-whole relations, and close association. In the indirect construction, a semantic distinction is made between items related to eating and drinking on the one hand, and general possession on the other. This distinction is signaled through the choice of one of two possessive classifiers. At the same time, there are some idiosyncrasies which are not explained by the main semantic pattern. When both direct and indirect constructions are used together in a sentence, each part maps a different semantic relationship. Typologically, the Patani order is possessor-possessed, as is common in the region. Finally, both the direct and the indirect construction may be used phrasally (‘my house’), or predicatively (‘I have a house’).

1 Introduction

Patani belongs to the South Halmahera-West New Guinea (henceforth SHWNG) subgroup of Austronesian. Although there is uncertainty concerning the exact subgrouping of SHWNG languages, it is widely accepted that Patani belongs to the South Halmahera subgroup together with the languages shown in Figure 1 below (building on Kamholz [2014]).

Figure 1: 
Languages of the South Halmahera subgroup of SHWNG.
Figure 1:

Languages of the South Halmahera subgroup of SHWNG.

Most SHWNG languages have two broad syntactic possessive constructions: a direct and an indirect construction, the latter involving a possessive particle. On the basis of these two main constructions, a distinction is often made between inalienable and alienable nouns. Inalienable nouns in SHWNG languages generally refer to body parts, kinship terms, locative nouns, and ‘name’ (van den Berg 2009: 351). Some SHWNG languages code body parts and kin terms differently, thereby having two constructions associated with inalienable nouns (see Arnold, this issue). However, this is not attested for the subgroup of South Halmahera languages. Other SHWNG languages make a distinction between general and alimentary/edible possession among alienable nouns. In these languages, the possessive particle in indirect constructions reflects this semantic distinction and thus functions as a classifier (Gasser et al. forthcoming). Finally, it is widespread within SHWNG that the same construction can be used to express both adnominal and predicative possession, although this is highly uncommon cross-linguistically (see Section 6).

Although there are great similarities in the area of possession among the South Halmahera languages, there are a few notable differences even within this small subgroup. Taba and Gane have only one adnominal possessive construction, whereas the other SH languages have two. Furthermore, only the Central-Eastern South Halmahera (CESH) languages make a distinction between alimentary and general possession.

Until now, very little documentation has been available for Patani. A wordlist including some notes on the grammar is available in Stokhof (1980), and the paradigm of the direct possessive construction is given in Kamholz (2014). This paper is the first to describe the possessive system in Patani in detail. It shows that Patani generally patterns with the other CESH languages in the possessive system, having two broad syntactic possessive constructions: direct and indirect. Furthermore, the indirect construction makes use of a possessive classifier which signals an alimentary/general distinction. However, there are exceptions to the main pattern, and these are carefully described in this paper. Finally, comparisons with the other South Halmahera languages will be made throughout the paper.

The data upon which this paper is built consists of transcriptions of recorded natural texts, translated material from Malay, indicated by ‘(transl.)’, and elicited material, indicated by ‘(elic.)’. Most of the data were gathered during three months fieldwork in Patani Barat, but some translation and elicitation has subsequently been carried out online (through Messenger or WhatsApp). Examples from the recordings are indicated with the name of the recording (e.g. 015-BM3), as well as the relevant time. These recordings will become available in PARADISEC[1] once my PhD project is finished. Loanwords are given in italics in glossed examples. Since all Patani speakers are also speakers of Indonesian and/or North Moluccan Malay (NMM), there are many loanwords from these varieties. Words which are found in both Indonesian and NMM are indicated as Malay for the sake of simplicity. A comma is used to indicate a pause.

This paper is organized as follows: A brief overview of the main traits of Patani grammar is presented in Section 2. Section 3 presents the main characteristics of the possessive system. Section 4 is dedicated to direct possession, and discusses formal aspects (Section 4.1) as well as semantics (Section 4.2). Section 5 deals with the indirect possessive construction, in which a distinction between general and alimentary possession is made. Nouns occurring with the alimentary classifier are the topic of Section 5.1, whereas Section 5.2 discusses nouns which are attested with both classifiers. Finally, Section 6 demonstrates that both the direct and the indirect construction can be used to express predicative possession, alongside a formally similar, designated predicative construction.

2 Patani: a brief grammatical overview

The phoneme inventory of Patani is given in Table 1 below. The orthographic representation is given in parentheses where this deviates from IPA. Additionally, Patani has geminate consonants. Orthographically, this is indicated with a double consonant. Geminate (unvoiced) obstruents occur across morpheme or word boundaries (e.g. through subject prefixation f-fan ‘2pl-go’ or vowel deletion ti ta > tta ‘that one’), whereas geminate l and nasals also occur in monomorphemic words (e.g. lollo ‘inside’, nnau ‘seawards’). The syllable structure in Patani is (C1)(C2)V(C3). Across morpheme and word boundaries, epenthetic vowels (copying the vowel quality of the preceding vowel) are inserted to avoid three-consonant clusters.

Table 1:

Vowels and consonants in Patani.

VOWELS Front Back CONSONANTS Labial Alveolar Palatal Velar
Close i u Plosive p b t d k g
Close-mid e o Nasal m n ɲ(ny) ŋ (ng)
Open-mid ɛ (é) ɔ (ó) Fricative f s
Open a Affricate ʧ(c) ʤ(j)
Trill r
Approximant w l j (y)

Patani has little obligatory inflection, but verbs normally bear subject marking, distinguishing seven person-number combinations (cf. Table 2). Within the domain of the noun phrase, plural number is optionally marked with the plural enclitic =si. NPs may also be marked for definiteness, through demonstratives (distal ta, proximal ja), or the definite enclitic =a, which attaches to the last word in the NP. Most noun modifiers follow the head noun, but possessor NPs precede the possessed head. To illustrate this difference, examples (1)–(2) below show two possible ways of expressing ‘hen egg’ in Patani. In the noun-noun compound in (1), the modifier (‘hen’) follows the head. In the possessive construction in (2), the possessor NP precedes the (possessed) head.

Table 2:

Patani pronouns.

Singular Plural
1.INCL it(i)
1.EXCL aya am(a)
2 aw(a) mew(e)
3 i si
(1)
tól takalé
egg hen
‘hen egg’
(2)
takalé tól-ó
hen egg-3sg.poss
‘hen egg’ (lit. ‘hen’s egg’)

The NP can also be headed by a pronoun, and the Patani personal pronouns are given in Table 2. The pronominal proclitics appearing in possessive constructions (cf. Tables 3 and 6) are clearly derived from the free pronouns.

Table 3:

Possessive marking occurring on inalienable nouns.

Singular Plural
Optional Obligatory Optional Obligatory
1.INCL iti= -(V)r
1.EXCL a= -(V)g ama= -(V)mam
2 a= -(V)m me= -(V)me
3 i= -(V)Ø si= -(V)re

Derivational morphology in Patani often involves partial reduplication. Nominalization is expressed through copying the onset and the coda consonant of the following syllable (CiC-), and pluractionality through copying the following coda consonant (faC-). Applicativization of verbs involves copying the vowel quality of the preceding syllable (-V). Other derivational affixes include the causatives fa- and fi-. Syntactically, Patani has nominative-accusative alignment, signaled through the subject prefix on the verb, and word order (SVO). Arguments may be elided when they are retrievable from the context.

3 Main characteristics of possession in Patani

Patani nouns can be divided into two classes depending on how they are possessed. Some nouns occur in a direct possessive construction, while all other nouns occur in an indirect construction. We may call the first class inalienable nouns, and the second class alienable nouns. The possessive marking attaches directly to inalienable nouns (hence the term ‘direct’), see example (3),[2] whereas alienable nouns are preceded by a possessive classifier to which the possessive marking attaches (hence ‘indirect’). There are two possessive classifiers in Patani, which reflect an alimentary/general distinction. The alimentary classifier is used for items that are consumable or associated with eating and drinking. The classifiers are marked in bold in examples (4) and (5)[3] below.

(3)
afyeleg
a=fyel-eg
1sg=nose-1sg.poss
‘my nose’ (elic., 016-BF2_BM5 03.23)
(4)
a ni k atom
a=ni-k atom
1sg=cl.gnrl-1sg.poss pen
‘my pen’ (elic., 015-BM3 01.38)
(5)
a na k in
a=nó-k in
1sg=cl.ali-1sg.poss fish
‘my fish’ (elic., 015-BM3 10.13)

The locus of possessive marking is the possessed noun in the direct construction, and the classifier in the indirect construction. Hereafter, I use the term possessive target as a way of referring to the locus of marking across the two main constructions (direct and indirect). The possessive marking is fundamentally the same across these two main types. The marking in the examples above consists of obligatory suffixes and optional proclitics. The suffixes only occur in possessive constructions and will therefore be referred to as possessive suffixes throughout this article. Importantly, the possessive suffixes are functionally equivalent across the two syntactic constructions: they are obligatory, and indicate the person and number of the possessor. Furthermore, there is great formal similarity between the suffixes appearing in the direct and indirect constructions (cf. Tables 3 and 6).

The optional proclitics also express the person and number of the possessor, but some of these forms are attested outside of possessive constructions, e.g. attached to a verb. Compare the proclitics in (3) through (5) above to the proclitic in bold preceding the 1sg subject prefix i- and the verb sebe ‘fall’: a= i-sebe ‘I fall’ (lit. ‘1sg=1sg-fall’). I therefore refer to these as pronominal proclitics. The possessor may alternatively be expressed in the form of a free pronoun (see the bold text in 6), or a full NP (see the bold text in 10). Such preposed possessor expressions (proclitics or NPs) are extremely frequent in the material, attested in 111/115 indirect constructions and 114/120 direct constructions. However, prenominal expression of the possessor is not obligatory; compare (6) and (7), and (8) and (9). What governs the choice of preposed possessor expression is not fully known at present. As can be seen from (10), the ordering of the possessor and the possessed is Possessor Possessed when the possessor is expressed as an NP. This is also the order found in most SHWNG languages (Gasser et al. forthcoming).

(6)
aya ntig isó le
aya nti-g isó le
1sg child-1sg.poss one only
‘I have only one child’ (001-MF5 00.25)
(7)
ntig
nti-g
child-1sg.poss
‘my child’ (001-BF9 00.50)
(8)
a nak yofa
a=nó-k yof=a
1sg=cl.ali-1sg.poss sago=def
‘my sago’ (011-MF1_MF7 02.24)
(9)
nak yof
nó-k yof
cl.ali-1sg.poss sago
‘my sago’ (elic., 016-BF2_BM5 12.40)
(10)
tentarasia niri simat lól
tentara=si=a ni-ri simat lól
soldier=pl=def cl.gnrl-3pl.poss person big
‘the soldiers’ superior’ (007-BM5 01.23)

The possessive marking always attaches to the possessive target, and not to any other element in the NP. This is illustrated in (11) below, where the possessive suffix (in bold) attaches to the main noun kmo ‘parent-in-law’ and not to the NP as a whole.

(11)
[akme g nyangan] NP
a=kmo-g nyangan
1sg=parent.in.law-1sg.poss married.woman
‘my mother-in-law’ (BM1, transl.)

4 Direct possession

The following sections deal with different aspects of the direct possessive construction, both formal and semantic. Section 4.1 presents the possessive marking in direct constructions, and explains morphophonological changes in directly possessed nouns. Section 4.2 goes systematically through the semantics of directly possessed (i.e. inalienable) nouns in Patani, and shows that these mainly refer to parts of a whole or kinship relations.

4.1 Formal aspects of direct possession

In SHWNG languages with an alienable/inalienable split, the person and number of the possessor is marked directly on inalienable nouns (Gasser et al. forthcoming). In Patani, this marking consists of optional proclitics and obligatory suffixes, which are given for all person-number combinations in Table 3 below. The vowel variant of the suffix (V) appears with inalienable nouns ending in a consonant, and the vowel quality of the preceding vowel is copied. Note that this vowel copying is confined to Patani, as the other CESH languages add vowels of a fixed quality. 3sg is unmarked when the noun ends in a vowel. For ease of exposition, I indicate the 3sg possessive suffix as ‘-Ø’.

One of the main differences between the possessive marking in direct and indirect constructions is found for 1sg possessors (-g vs. -k respectively, cf. Tables 3 and 6). According to Kamholz (2014, 130), the form in the direct construction is a result of voicing of historical *-k to -g. Note that the more conservative form -k is retained in the indirect construction. This is true of Patani, Maba and Sawai, whereas Buli has kept -k in the direct construction as well.

It is not uncommon for inalienable nouns to be obligatorily possessed, in the sense that they only occur in possessive constructions (see Dixon 2010). Such a strong generalization cannot be made for Patani, since there are examples of inalienable nouns in their root form, without possessive markers, e.g. kóm ‘hand’, and fiw ‘vagina’. At the same time, not all inalienable nouns occur outside of a possessive construction in the available data. For instance, no speaker was able to give me the root form of ‘fruit’; they all used a possessive construction, e.g. yaya pyono ‘a tree’s fruit’, or ipyono ‘its fruit’. My attempt to say *pyon(o) in isolation was not recognized as a Patani word. Thus, it remains an open question whether some directly possessed nouns are bound roots.

4.1.1 Paradigmatic vowel alternation for 1sg and 2sg

1sg and 2sg stand out in the possessive paradigm as they trigger vowel alternation in certain noun roots and in the alimentary possessive classifier. These have different vowel qualities for 1sg and 2sg compared to the rest of the paradigm.[4] Table 4 below illustrates the vowel alternation by providing the full possessive paradigm of three nouns. The main point is to show that the root vowel is different with 1sg/2sg possessors compared to the rest of the paradigm. More specifically, a back vowel is fronted for 1sg/2sg forms: u becomes i, o becomes e, and ó becomes a. It is interesting to note that the counterpart of ó is not é as might be expected since both are open-mid vowels.[5]

Table 4:

Possessive paradigm of three inalienable nouns exhibiting vowel alternation.

Gloss ‘child’ ‘wish’ ‘cross-sex sibling’
1sg a=nti-g a=wle-g a=rima-g Fronted vowel
2sg a=nti-m a=wle-m a=rima-m
3sg i=ntu i=wlo i=rmó Back vowel
1pl.incl iti=ntu-r iti=wlo-r iti=rmó-r
1pl.excl ama=ntu-mam ama=wlo-mam ama=rmó-mam
2pl me=ntu-me me=wlo-me me=rmó-me
3pl si=ntu-re si=wlo-re si=rmó-re

Vowel alternation is only attested with monosyllabic noun roots ending in a vowel. However, not all inalienable nouns matching these criteria exhibit vowel alternation. For instance, ‘shoulder’ keeps the same vowel quality throughout the paradigm. Furthermore, one of the vowel-alternating nouns, fyó ‘thigh’, is reported to have two possible forms for 1sg.poss: the expected a=fya-g as well as a=fyó-g.

The same vowel alternation is found with the alimentary classifier (used with alienable nouns) as well: The classifier is realized as na for 1sg and 2sg, cf. the examples below, where the classifier is given in bold.

(12)
a na k wóyó
a=nó-k wóyó
1sg=cl-1sg.poss water
‘my water’ (BM5, elic.)
(13)
mnóm
nó-Ø mnóm
cl.3sg-poss food
‘his/her food’ (001-MF5 01:46)

Sawai is also reported to have vowel alternation for some inalienable nouns and the possessive classifier (see Whisler [1992, 1996]), whereas Buli has only the classifier form na throughout the paradigm (Maan 1951).

4.1.2 Vowel deletion in noun roots

For a few noun roots, the vowel of the final syllable is deleted in the direct possessive construction. In my data, this is the case for the nouns ngósón ‘name’, wóngót ‘flesh’, bangal ‘wing’. This can be seen when comparing the first and the second lines in (14) and (15) below.[6] Note that these nouns end in a consonant, and therefore, the -V variant of the possessive suffix, which copies the quality of the preceding vowel, is used (cf. Table 3).

(14)
iti wóngt ór
iti=wóngót-Vr
1pl.incl=flesh-1pl.incl.poss
‘our flesh’
(15)
a ngósn óg
a=ngósón-Vg
1sg=name-1sg.poss
‘my name’

These three noun roots are disyllabic, end in a consonant, and in their root form have no medial consonant-cluster. In the data, there is one more root which fits all these criteria, namely kinun ‘tail’. It is subject to vowel deletion (in addition to a change in vowel quality) by some speakers (illustrated in 16), but not by all, which is shown in (17) below.

(16)
i kunn u
i=kinun-VØ
3sg=tail-3sg.poss
‘its tail’ (BM6, elic.)
(17)
kabal kinun u
kabal kinun-VØ
goat tail-3sg.poss
‘goat’s tail’ (012-MM1 24.24)

For disyllabic stems with a medial consonant cluster, e.g. tiktuk(u) ‘elbow/knee’, siksak(a) ‘thigh’, deletion of the root-final vowel does not apply, arguably because this would result in an unattested three-consonant cluster: *tiktku-g, *sikska-g. However, some of these are already shortened by some speakers by reducing the first syllable: a= tik tuk-ug > a= t tuk-ug ‘my knee’, a= sik sak-ag > a= s sak-ag ‘my thigh’.

4.2 Semantics of directly possessed nouns

The semantic distinction between directly possessed nouns and indirectly possessed nouns can be described in terms of inalienability. Indeed, some works discussing possession in SHWNG use the terms inalienable versus alienable possession (e.g. van den Berg [2009]) instead of direct versus indirect possession. ‘Direct’ versus ‘indirect’ are terms normally used within Oceanic linguistics, and reflect the morphological marking of a possessed noun instead of its semantics. Cross-linguistically, the semantic categories which are described as inalienable are typically kin terms, body parts, part-whole relations, spatial relations, and culturally basic possessed items (Chappell and McGregor 1996: 8) – broadly speaking, items belonging to the personal domain, as defined by Bally (1996 [1926]). Importantly, however, in many languages only a subset of nouns belonging to these semantic categories are expressed as directly possessed nouns.

Within SHWNG, nouns which occur in the direct construction always include (at least some) body parts and/or kin terms. Furthermore, many languages “treat associative nouns (‘name’, ‘breath’, ‘shadow’, ‘feeling’), locational nouns (e.g. ‘top’, ‘side’), and parts of wholes (e.g. ‘edge’, ‘fruit’) as inalienable” (Gasser et al. forthcoming). This is also true for Patani. All nouns which are directly possessed in Patani belong to the aforementioned semantic categories. At the same time, not all nouns in these categories take direct possessive marking. This means that the class of inalienable nouns is grammatically defined, on the basis of which possessive construction they occur in (namely, the direct construction), and not on the basis of their semantics per se.

4.2.1 Subgroups of inalienable nouns in Patani

I have identified 47 directly possessed nouns in the data, which are presented in Table 5 below, arranged by semantic category. For the nouns which exhibit vowel alternation, the 1sg form is also provided (cf. Section 4.1.1).

Table 5:

Directly possessed nouns in Patani.

Body parts (incl. organs and body products) bangal ‘wing’ ngang ‘tooth’
bok ‘head’ ngól(ó) ‘chin’
‘shoulder’ pléw ‘tongue’
fsi ‘penis’ plu ‘body hair’
fyel ‘nose’ siksak(a) ‘thigh’
fyó (1sg fya-g) ‘animal thigh’ sum ‘mouth’
gyé ‘foot’ tang ‘ear’
kanna ‘skin’ tiktuk(u) ‘knee’
kilkol(o) ‘digit’ usnu ‘face’
kinun ‘tail’ wló (1sg wla-g) kiykaya ‘heart’
kóm ‘arm/hand’ ‘neck’
long ‘bone’ wóngót ‘flesh’
mimi ‘urine’ wyoco ‘liver’
mom ‘bone’ yél(é) ‘back’
mtó (1sg mta-g) ‘eye’

Other parts of a whole/locational nouns kówó ‘seed’ pyono ‘fruit’
lollo ‘inside’ tól ‘egg’
mdi ‘lower part; stem’ tubu ‘end’
pipo ‘top’ wlu ‘leaf’

Kinship terms fón ‘niece/nephew’ kmo (1sg kme-g) ‘parent-in-law’
ftén ‘same-sex sibling’ ntu (1sg nti-g) ‘child’
gbu (1sg gbi-g) ‘grandchild’ rmó (1sg rima-g) ‘cross-sex sibling’

Close association gów ‘place’ ngósón ‘name’
li ‘voice’ wlo (1sg wle-g) ‘seat of character, emotions, and values; wish’

Only four inalienable nouns in my data are not parts of a whole or kinship terms: gów ‘place’,[7] li ‘voice’, ngósón ‘name’, and wlo ‘seat of emotions, values’. These concepts can all be thought of as being particularly closely associated with or inherently linked to their possessor (cf. Dixon 2010: 285). Note that the original, literal meaning of wlo is not known at present, but I suspect that it fits within one of the other inalienable semantic categories.[8]

In the following sections, body parts are treated in detail in 4.2.2, before we consider other parts of wholes in 4.2.3. Finally, kinship terms are discussed in 4.2.4.

4.2.2 Body parts

In my data, body parts generally take direct possession, but six exceptions are identified: byeket ‘sweat’, dong ‘navel’, lifléfén ‘eyebrow’, liló ‘blood’, sus ‘breast’, and utu ‘hair’, e.g. a=ni-k utu ‘my hair’, a=ni-k liló ‘my blood’. That utu ‘hair’ is alienable must be considered a lexical idiosyncrasy, since the closely related plu ‘body hair’ is inalienable (see example 18 below). Similarly, that liló ‘blood’ and byeket ‘sweat’ are alienable is not explained by them being body liquids, since mimi ‘urine’ is inalienable (see example 19 below). Note that many of these exceptions are also attested in Sawai: leló ‘blood’, mameéket ‘sweat’, and bebnaé ‘eyebrows’ are all alienable (Whisler 1996). However, so is mome ‘bones’, which is inalienable in Patani.

(18)
iplu
i=plu-Ø
3sg=body.hair-3sg.poss
‘his/her body hair’ (transl., 012-BM5 03.04)
(19)
amimig
a=mimi-g
1sg=urine-1sg.poss
‘my urine’ (BM6, transl.)

The examples in my data of designated animal body parts are all directly possessed: bangal ‘wing’, kinun ‘tail’, and fyó ‘animal thigh’. Patani seems to diverge from Sawai here, which is reported to often use the indirect construction if the possessor of a body part is an animal (Whisler 1996: 49).

4.2.2.1 Body parts in double possessive constructions

Some body parts are essentially parts of a larger body part, such as fingers (part of a hand), and toes (part of a foot). These body part terms are expressed through what we may call a double possessive construction in Patani. Grammatically, the animate possessor is the possessor of the larger body part, which in turn is the possessor of the smaller part. Example (20) below shows that ‘finger’ is expressed as ‘digit of possessor’s hand’:

(20)
akómóg kilkolo
a=kóm-Vg kilkol-Ø
1sg=hand-1sg.poss digit-3sg.poss
‘my finger’ (lit. ‘my hand’s digit’) (elic., 015-BM3 14.23)

The noun kilkol(o) can mean both ‘finger’ and ‘toe’, and ‘my toe’ is expressed as ‘digit of my foot’: a=gyé-g kilkolo. Note however that this specification does not have to be made, as the possessor can be expressed directly on kilkol(o): a=kilkolo-g ‘my digit’.

Wló kiykaya ‘heart’ is also expressed as a double possessive, the animate possessor of ‘heart’ being indexed on wló. Neither wló nor kiykaya has been attested outside of this body part term, and their individual meaning is not clear at this point.

(21)
awlag kiykaya
a=wló-g kiykaya-Ø
1sg-?-1sg.poss ?-3sg.poss
‘my heart’ (lit. ‘kiykay(a) of my wló’) (BM1, transl.)

In all examples of the double possessive construction, the inanimate possessor (i.e. the larger body part) is 3sg and the possessive marking is -Ø. It is therefore reminiscent of a part-whole possessive construction found in some SHWNG languages (e.g. in Biak, as described by van den Heuvel [2006]), in which there is simple juxtaposition and no possessive marking (Emily Gasser, p.c.). This is also true of the part of wholes discussed in 4.2.3. We would need examples of a non-singular inanimate possessor in order to decide whether the constructions in (20)–(21) and 4.2.3 are in fact examples of a juxtapositional possessive construction without possessive marking. More data are therefore needed to conclude on this matter.

4.2.2.2 Body parts with an external possessor

Inalienable nouns denote concepts of the personal domain. An interesting question, then, is how body parts belonging to other entities than the possessor are treated grammatically. There are not many examples of this kind naturally occurring in the recordings. However, when asked to translate ‘I have one fish head’, accompanied with a picture of a fish head on a plate, my main consultant replied:

(22)
anak iboko isó
a=nó-k i=bok-VØ si-só
1sg=cl.ali-1sg.poss 3sg=head-3sg.poss num.cl-one
‘I have one fish head’ (lit. ‘its head that is in my possession (to eat) is one’) (BM6, transl.)

In (22),[9] ‘head’ is directly possessed by a 3sg possessor (signaled with i= and the possessive suffix -VØ), i.e. referring to the fish. The possessive relation to the speaker, on the other hand, is expressed through an indirect possessive construction, utilizing the alimentary classifier to signal that the possessum is meant for eating. A similar example results from a picture stimuli task, in which two speakers have a set of pictures each. One picture in each set depicts slices of fish, but the fish is cut on different planes. In one picture the fish is cut on the transversal plane, i.e. across the fish spine, and on the other, it is cut on the mid-sagital plane, along the spine. This difference is expressed through a direct possessive construction in Patani. The first type is called in longo ‘fish bones’ (lit. ‘the fish’s bone’), and the other type is called in wóngtó ‘fish meat’ (lit. ‘the fish’s meat)’. In the example below, this possessive expression is embedded in an indirect possessive construction, which indicates that the pieces of fish belong to the speaker, and that the fish is meant for eating.

(23)
anak in wóngtó mettel
a=nó-k in wóngót-VØ met-tel
1sg=cl.ali-1sg.poss fish meat-3sg.poss num.cl-three
‘I have three pieces of fish’ (lit. ‘I have three pieces of a fish’s meat (to eat)’) (elic., 016-BF2_BM5 21.52)
4.2.2.3 Body parts from nominalization

Six body part terms are formally similar in an interesting way. Kilkol(o) ‘digit’, kiykaya ‘?’ (cf. Section 4.2.2), lifléfén ‘eyebrow’, mimi ‘urine’, siksak(a) ‘thigh’, and tiktuk(u) ‘elbow/knee’ can all be analyzed as consisting of a prefix CiC- which is partially reduplicated from the following syllable. This is a nominalization prefix in Patani, e.g. yibyób ‘long hook’ from the verb yób ‘fetch’. Similar reduplication is also attested for body part terms in Taba and Buli, where it seems to be even more pervasive. Both Bowden (2001: 176) and Maan (1951: 21, footnote 18) recognize the nominalizing function of the reduplication strategy, and both assume that the root was originally a verb. The hypothesis is that these body parts are derived from the action with which they are closely associated. The putative verb roots are generally not found synchronically, but the Patani verbs mi ‘urinate’ and tuku ‘fold (tr.)’ seem to support this hypothesis; mimi ‘urine’ and tiktuk(u) ‘knee/elbow’ are closely associated with the action of urinating and folding respectively.

4.2.3 Other parts of wholes

Reference to different parts of an inanimate possessor is also made through the direct possessive construction. Examples (24) and (25) below illustrate how the top and the foot of a mountain are distinguished by alternating the possessed noun. An epenthetic vowel is inserted in (24) due to the subsequent consonant cluster.

(24)
loloso mdi
lolos-V mdi-Ø
mountain-ep.v stem-3sg.poss
‘foot of the mountain’ (transl., 012-BM2 07.28)
(25)
lolos pipo
lolos pipo-Ø
mountain top-3sg.poss
‘mountain top’ (transl., 012-BM2 07.28)

Reference to different parts of a fruit is also made by alternating the possessed noun: lilit kanna ‘mango peel’ (lit. ‘mango’s peel’) versus liliti wlu ‘mango leaf’ (lit. ‘mango’s leaf’). Seen from another perspective, the possessor serves to specify a subclass of peel and leaves respectively; we are talking about mango leaves, not banana leaves. This means that in Patani, the direct construction may function as a classificatory label for a certain subclass (cf. Koptjevskaja-Tamm [2004]). In such cases, the possessor is always non-specific (cf. Ross [1998]), which is signaled through the lack of definiteness marking in Patani.

Two further examples of part-whole relations are given in the following. In (26), the direct possessive construction (in bold) is embedded in another direct construction (lit. ‘my foot’s end’).

(26)
ncayanga na a gyé g tubu filea
n-sayang=a na a=gyé-g tubu-Ø file=a
3sg-love=1sg loc  1sg=foot-1sg.poss end-3sg.poss still=def
rómbe na abokog
rómbe na a=bok-Vg
till loc  1sg=head-1sg.poss
‘He loves all of me’ (lit. ‘he loves me from the end of my foot to my head’) (009-BF2 04.04)

Finally, location inside something is expressed through the direct possessive construction following the preposition naLOC’: na um/ruangan/desa lollo inside a house/room/village’ (lit. ‘at house’s/room’s/village’s inside’).

4.2.4 Kinship terms

Of the 23 kinship terms in my data, 6 are directly possessed and thus inalienable. Noun class is not decided purely on semantic grounds for kinship terms. Whether a term is borrowed also seems to play a role. In my data, all nouns borrowed from Malay take indirect possession (kaka ‘older sibling’,[10] ade ‘younger sibling’,[11] orang tua ‘parents’, mama ‘mother’, papa ‘father’, tete ‘grandfather’, etc.). Furthermore, morphologically complex kinship terms take the same possessive construction as their head, irrespective of the meaning of the complex term.

With the exception of loanwords and morphologically complex kinship terms, the following generalization emerges from the (admittedly sparse) data: The inalienable kinship terms are non-borrowed nouns expressing: 1) consanguineal kinship relations where the possessum is of the same or a younger generation than the possessor, or 2) parents-in-law:

ftén ‘same-sex sibling’ gbu ‘grandchild’
rmó ‘cross-sex sibling’ fón ‘niece, nephew’
ntu ‘child’ kmo ‘parent-in-law’

Native nouns which do not fit into these semantic categories are indirectly possessed: a=ni-k tamay ‘my sibling-in-law’. Similarly, morphologically complex kinship terms which do fit into the semantic categories may still be indirectly possessed according to their head. This is the case for a=ni-k délé wyó ‘my younger sibling’, since dél ‘family member’ is indirectly possessed.

The split within kin terms with regards to possessive construction is also noted for closely related languages, and the generalization seems to be the same as what is claimed here for Patani. In Sawai, descendants, siblings, and parents-in-law are expressed through the direct possessive construction, whereas ancestors and siblings-in-law are indirectly possessed (Whisler 1996: 49–50). In Gebe, mam ‘father’ and didi ‘mother’ are indirectly possessed, whereas ntu ‘child’ and knó ‘sibling’ are directly possessed (Bax 2019).

5 Indirect possession

Across SHWNG, there is more morphosyntactic variation in how alienable nouns are possessed compared to inalienable nouns. A typical pattern, however, is that the possessor NP and the possessed NP are separated by a possessive particle or verb (typically of the form nV) which is inflected to mark the person and number of the possessor (Gasser et al. forthcoming). The indirect construction in Patani follows this main pattern, but unlike most SHWNG languages, the possessive particle is best described as a classifier (cf. Section 1). Possessive classifiers (general vs. alimentary) are only attested in CESH languages and Waropen. In Gebe, a South Halmahera language which is not in the Central-Eastern subgroup, there seems to have been a distinction similar to what we find in CESH, but which is now lost; both the forms ni and na are attested in the possessive system, but these are described as being in free variation instead of mapping different possessive relations (Bax 2019: 88).

In the Patani indirect possessive construction, the possessed noun is obligatorily preceded by a possessive classifier, ni or nó/na (the vowel alternation is explained in 4.1.1). As noted in Section 3, the possessive classifiers are the target of possessive marking in the indirect construction, and they thus inflect for the person/number of the possessor. Table 6 shows this inflection, i.e. the pronominal proclitics and the possessive suffixes. Note that for 1pl.excl, there is no classifier to which the possessive marking attaches.

Table 6:

Possessive marking in the indirect construction.

Singular Plural
Optional Obligatory Optional Obligatory
1.INCL it= -r
1.EXCL a= -k (a)mam
2 a= -m me= -mi/mó
3 i= si= -ri/ró

In general, nó/na is used for items of food and drink, as well as implements used in food and drink preparation and consumption, whereas ni is used otherwise. This is exemplified in examples (27) and (28) below, where the possessive classifier is indicated in bold. However, the distinction between alimentary and general is not made with 1pl.excl possessors, as can be seen in examples (29) and (30).[12]

(27)
i kui
i=nó-Ø kui
3sg=cl.ali-3sg.poss cake
‘his/her cake’
(28)
i ni tarapesa
i=ni-Ø tarapesa
3sg=cl.ali-3sg.poss chair
‘his/her chair’
(29)
mam yafa
1pl.excl.poss lime.powder
‘our lime powder’
(30)
mam sigi
1pl.excl.poss mosque
‘our mosque’

The Patani classifier functions as a relational classifier in the sense of Lichtenberk (1983: 174), since the “nature of the relationship of one entity to another” may determine the choice of classifier. The relationship between a given possessor and a given possessum can vary, and therefore, the same noun can be attested with different relational classifiers. For instance, a noun such as takalé ‘chicken’ can refer both to the living bird, as well as to the meat, and Patani has the opportunity to express this difference solely through the possessive classifier. This is illustrated in (31) and (32) below, where the classifiers are given in bold.

(31)
a ni k takalé
a=ni-k takalé
1sg=cl.gnrl-1sg.poss chicken
‘my (living) chicken’ (elic., 015-BM3 12.39)
(32)
a na k takalé
a=nó-k takalé
1sg=cl.ali-1sg.poss chicken
‘my chicken (to eat)’ (elic., 015-BM3 12.57)

Even though the main generalization is that the classifiers distinguish between items associated with food and drink versus everything else, it is not always clear why certain nouns are used with one classifier rather than the other. Furthermore, some nouns are attested with both classifiers without the difference in meaning witnessed in (31) and (32). The following sections list the nouns which behave differently from the main generalization, as attested in the data so far. Nouns which somewhat surprisingly occur with the alimentary classifier are discussed in 5.1, whereas nouns occurring with both classifiers are dealt with in 5.2.

5.1 Nouns attested with the alimentary classifier

Nó/na is used with all food and drinks. It is also used for things that are not digested, but which are chewed or sucked on, such as tobacco, as illustrated in the examples below. Hereafter, ‘food’ also includes tobacco, for the sake of simplicity.

(33)
a na m kofi
a=nó-m kofi
2sg=cl.ali-2sg.poss coffee
‘your coffee’ (elic., 016-BF2_BM5 06.22)
(34)
na k tabako
nó-k tabako
cl.ali-1sg.poss tobacco
‘my tobacco’ (elic., 016-BF2_BM5 10.33)

Table 7 below gives an overview of all nouns which do not denote food or drinks per se, but which occur with the alimentary classifier in the available data.

Table 7:

Nouns not denoting food and drinks which are classified using the alimentary .

Food utensils asbak ‘ashtray’ (< Malay < Dutch)
dandang ‘rice steamer’ (< Malay)
galas ‘glass’ (< Malay < Dutch)
gów ‘plate’
kantong ‘plastic bag’ (< Malay)
lógóy ‘woven plate’
pang ‘saucepan’ (< Malay < Dutch)

Nature/body (food associated) bét ‘garden’
liliti wlu ‘mango tree leaf’
sus ‘breast’

? lifléfén ‘eyebrow’
kimón ‘axe’

Plastic bags are classified as ‘food utensils’ since they are used for bringing food home from ritual events like weddings. Similarly, it is not surprising that bét ‘garden’, which is where food is grown, is classified with . This is also attested for Buli (Maan 1951: 55) and Sawai (Whisler 1996: 50). However, at this point it is unclear why lifléfén ‘eyebrow’ and kimón ‘axe’ take the alimentary classifier. When it comes to ‘axe’, we might say that this is an implement which can be used for obtaining food. However, yété ‘knife’, which is also often used in obtaining or preparing food, occurs with the general classifier ni; compare a= na -m kimón ‘your axe’ and a= ni -m yété ‘your knife’. For now, I have to consider it an idiosyncratic fact that kimón ‘axe’ and lifléfén ‘eyebrow’ take the alimentary classifier (see also Lichtenberk [1983] on idiosyncrasies in the possessive systems). It should be noted, however, that the alimentary classifier is used for ‘eyebrows’ and certain tools, (perhaps ‘axe’?) in Buli as well (Maan 1951: 55).

5.2 Nouns attested with both classifiers

In Section 5, we saw that the choice of classifier may depend on the relation between possessor and possessed. However, this is not always the case. The nouns which occur with both the general and the alimentary classifier in my material are given in Table 8 below. Here, I also indicate whether or not the possessive relation can explain the choice of classifier.

Table 8:

Nouns occurring with both classifiers in the Patani material.

Patani noun Gloss What governs the choice of classifier?
takalé ‘chicken’ Relation possessor–possessum
yaya mdi ‘tree trunk’ Unclear, no difference in meaning
meja (< Malay < Portuguese) ‘table’ Unclear, no difference in meaning
lampur (< Malay < Dutch) ‘lamp’ Unclear, no difference in meaning
sém kówó ‘nutmeg seed’ Unclear, no difference in meaning
gilógó ‘stone’ Unclear, no difference in meaning
mpin ‘female; wife’ Unclear, no difference in meaning

This table illustrates that most occurrences of nouns attested with both classifiers cannot be explained in terms of the relation between possessor and possessum. On the contrary, there is often no obvious difference in meaning. This is illustrated in (35) below, where the same speaker uses the general classifier in (a), but the alimentary classifier in (c) (both in bold). In both cases the referent in question is a depicted tree trunk (yaya mdi).

(35)
a. ni m ó yaya mdi aifise
ni-m ó yay mdi-Ø ai-fis
cl.gnrl-2sg.poss hes tree stem-3sg.poss num.cl-how.many
BF2: ‘How many tree trunks do you have?’
b. yaya mdi aisó
yay-V mdi-Ø ai-só
tree-ep.v stem-3sg.poss num.cl-one
BM5: ‘One tree trunk’
c. aya na ksi ja yaya mdisia ailu
aya nó-k=si ja yay mdi-Ø=si=a ai-lu
1sg cl.ali-1sg.poss=pl dem tree stem-3sg.poss=pl=def num.cl-two
BF2: ‘I have two tree trunks’ (lit. ‘Mine here are two tree trunks’) (elic., 016-BF2_BM5 21.23-21.28)

Turning to the next noun in the table, meja ‘table’ can be used for referring to a dining table as well as a desk. In Gasser et al. (forthcoming), Patani’s close relative Sawai is reported to distinguish the two types of table through the classifier (edible vs. general respectively). However, this is not the case in the Patani data. Though most occurrences of ‘table’ have a desk as the intended referent, the alimentary classifier is still used. My main consultant confirms that both classifiers are acceptable with the noun meja, but the data show that the classifier is not used to convey a difference in meaning (i.e. between dining table and desk).

For sém kówó ‘nutmeg seed’ (lit. ‘nutmeg’s seed’), the situation is somewhat different. During an elicitation task, two speakers (BM5 and BF2) discuss nutmeg seeds. The two speakers have a set of pictures each, and are told that both sets depict the same items, but that these items differ in number. The task is to find out how many of each item the other person has, without looking at each other’s pictures. When discussing the number of nutmeg seeds, one speaker consistently uses ni, while the other uses . This is illustrated in the examples below. In (36), the male speaker BM5 uses the alimentary classifier (in bold) in (a), whereas the female speaker BF2 replies with the general classifier in (b).

(36)
a. na m sém kówó pifis
nó-m sém kówó-Ø pi-fis
cl.ali-2sg.poss nutmeg seed-3sg.poss num.cl-how.many
‘How many nutmeg seeds do you have?’ (BM5)
b. aya ni k sém kówó pifót
aya ni-k sém kówó-Ø pi-fót
1sg cl.gnrl-1sg.poss  nutmeg seed.3sg-poss num.cl-four
‘I have four nutmeg seeds’ (elic., 016-BF2_BM515.12-15.17)

The referent is the same for both speakers, and it is therefore surprising that they use different classifiers. However, similar examples are described for the neighboring language Buli, where the alternation between classifiers has an additional function.

In Buli, the alimentary classifier is considered more ‘polite’, whereas the use of the general classifier can be regarded as vulgar in the sense that it may have sexual innuendo in combination with certain nouns, at least when uttered by a male speaker. For instance, nouns denoting long objects may be interpreted as ‘penis’ if modified by the general classifier. Similarly, nouns denoting containers can be interpreted as ‘vagina’, and nouns denoting small, roundish objects as ‘testicle’ (Bubandt 2014: 196–197). For instance, the noun tol ‘egg’ – which normally takes the alimentary classifier – can mean ‘testicle’ when modified by the general classifier and uttered by a man (Nils Bubandt p.c.).[13] In order to avoid this sexual innuendo, the alimentary classifier is used.

There are two indications that something similar is going on in Patani as well. First of all, kówó ‘seed’ can have the meaning ‘testicle’. Secondly, in my data, only the male (older) speaker uses the alimentary classifier with sém kówó ‘seed’. This speaker is also kepala desa, the head of the village. In the Buli context, we would expect the older men – and especially those who have high status and are used to speaking in public – to be particularly careful with their wording in a recording context, and in front of a young woman such as myself.

If the classifiers can have the same, additional function in Patani as they have in Buli, this would explain why BM5 chooses the alimentary classifier with a noun that can mean ‘testicle’, whereas BF2 can use the general classifier. The wish to avoid sexual innuendo might also govern the choice of classifier with the noun gilógó ‘stone’, which has the same shape as sém kówó ‘nutmeg seed’. In the picture task described above, the female speaker (BF2) starts out with the general classifier, whereas the kepala desa (BM5) replies with the alimentary classifier.

Another, perhaps related, observation is that mpin ‘female’ is often referred to with the alimentary classifier when it means ‘wife’. My impression is that this normally occurs with a 1sg (male) possessor. This is illustrated in (37) below, where the alimentary classifier is indicated in bold. In contrast, món ‘male; husband’ is not attested with the alimentary classifier.

(37)
mpin ta a na k i
mpin ta a=nó-k i
female dem  1sg=cl.ali-1sg.poss 3sg
‘That woman is my wife’ (BM6, transl.)

More data are needed to uncover the full range of meanings associated with the alimentary classifier, and to reach firm conclusions on whether it can be used to avoid sexual innuendo in Patani.

6 Predicative possession

Most of the examples considered so far are instances of attributive possession, where the relation of possession is presupposed and possessor and possessum are part of the same NP. However, the possessive relation can also be the main assertion of the sentence, which is the case in the example ‘Agoes has a motorcycle’. This is an instance of predicative, or clausal, possession. The Patani possessive constructions considered so far can be used to express predicative possession as well as attributive possession. This is very uncommon cross-linguistically (Heine 1997: 25–26), but is attested in other languages in the region, e.g. in Ambel (Arnold 2018: 354) and the Yapen SHWNG languages (Emily Gasser, p.c.). Two examples of predicative possession are exemplified with the indirect construction in (38b) and (38c) below, which is taken from a recording where the speaker is describing the house we’re in by telling us what can be found there. He thus provides many examples where the possessive relation is the main assertion of the sentence.

(38)
a. ja um i
dem house 3sg 
‘This is a house’
b. um ja ni lampu
um ja ni-Ø lampu
house dem cl-3sg.poss lamp
‘This house has a lamp’
c. um ja ni luntubu
um ja ni-Ø lentubu
house dem cl-3sg.poss door
‘This house has a door’ (014-BM4_1 02.05-02.45)

Another example of clausal possession is given in (39) below, again with the indirect construction.

(39)
awa nim leptop
awa ni-m leptop
2sg cl-2sg.poss laptop
‘Do you have a laptop?’ (BM1, transl.)

In (38) and (39), only the context determines whether the constructions in questions are interpreted as attributive or predicative. However, Patani also has designated constructions for predicative possession. These constructions are derived from the direct and the indirect construction by adding re- to the possessive target.[14] The possessive suffixes are identical to those given in Tables 3 and 6 for the direct and indirect construction respectively. Note that a pronominal proclitic cannot occur in this construction, but a possessor NP may precede the re-prefixed possessive target, as can be seen in (40) and (41) below.

Importantly, re-constructions are never found attributively, i.e. functioning as arguments of a predicate. An example of the predicative re-construction (the indirect version) is given in the following example, which was uttered between (38b) and (38c) above. It is hard to find any difference in meaning between the re-construction in (40) and the regular possessive constructions (e.g. 38), strengthening the claim that (38b) and (38c) are indeed examples of clausal possession.

(40)
um ja reni senemdi
um ja re-ni-Ø senemdi
house dem pred-cl-3sg.poss  kitchen
‘This house has a kitchen’ (014-BM4_1 02.49)

An example of the re-version of the direct construction is given below. This is a common way to ask someone if they have children.

(41)
awa rentim to
awa re-ntu-m to
2sg pred-child-2sg.pos already
‘Do you have children?’ (BM1, transl.)

To sum up, the direct and the indirect possessive construction can be used to express both attributive and predicative possession. However, predicative possession is commonly expressed by prefixing re- to the possessive target.

7 Summary

As most SHWNG languages, Patani has two main types of possessive construction: the direct construction, where the possessive marking attaches directly to the noun, and the indirect construction, where the marking attaches to a possessive particle. Unlike most SHWNG languages, however, the possessive particle is best described as a possessive classifier. Morphologically, the possessive marking consists of an optional pronominal proclitic and an obligatory suffix in both types of possessive construction. Nouns entering into the direct construction are called inalienable nouns as these denote parts of a whole, kinship terms, or concepts tightly associated with a person. Nouns which are directly possessed are called alienable nouns, and their relation to the possessor is signaled through one of two possessive classifiers: The alimentary classifier is mainly used with food and drinks, whereas the general classifier is used otherwise. However, some Patani nouns appear to be idiosyncratic. For example, it is not clear why some body part terms are alienable nouns, or why certain nouns are used with the alimentary classifier. However, in some instances, the use of alimentary classifier might be related to the wish to avoid sexual innuendo. Both the direct and the indirect construction can express attributive as well as predicative possession, but Patani also has designated predicative constructions, signaled by prefixing re- to the possessed noun or the possessive classifier.


Corresponding author: Linn Iren Sjånes Rødvand, Department of Linguistics and Scandinavian Studies, University of Oslo, Oslo, Norway, E-mail:

Acknowledgment

I am grateful to many people for contributing to this paper. First and foremost, thanks to the people of Patani Barat for welcoming me into their homes and sharing their knowledge and language with me. Specifically, I want to mention those of my language helpers who wanted to be acknowledged by name: Ibu Adema, Ibu Asfa, Ibu Astin, Ibu Fatima, Ibu Hadija, Ibu Jaenam, Ibu Leha, Ibu Masni, Ibu Mayati, Ibu Muslika, Ibu Nun, Ibu Raena, Ibu Safia, Ibu Samsia, Ibu Sumia, Ibu Sunnat, Pak Hasan, Pak Kepala Desa Taha, Pak Malik, Pak Muawwiyah, Pak Rais, Pak Sale, Pak Umar, Triya, and last but not least, Udin. My visit to Indonesia would not have been possible without the help of my research partner institution, Universitas Khairun in Ternate, and especially Roswita M. Aboe, and Dr. Farida Maricar. I am also grateful to several fellow linguists who helped me in one way or another: Laura Arnold for useful discussions on body part nouns, Steve Pepper for important insights about non-anchoring possessive constructions, my supervisor Åshild Næss, Antoinette Schapper, Emily Gasser, and two anonymous reviewers for useful comments on earlier drafts. The fieldwork was supported financially by the Norwegian Research Council, project no. 275243.

Glossary

1

first person

2

second person

3

third person

ali

alimentary

cesh

Central-Eastern South Halmahera

cl

possessive classifier

def

definite

dem

demonstrative

excl

exclusive

ep.v

epenthetic vowel

gnrl

general

hes

hesitation

incl

inclusive

loc

locative

nmm

North Moluccan Malay

np

noun phrase

num.cl

numeral classifier

pl

plural

pmp

Proto-Malayo-Polynesian

poss

possessive

pred

predicative

sg

singular

sh

South Halmahera

shwng

South Halmahera-West New Guinea

References

Arnold, Laura. 2018. A grammar of Ambel. An Austronesian language of Raja Ampat, west New Guinea. Edinburgh: The University of Edinburgh (dissertation).Search in Google Scholar

Bally, Charles. 1996 [1926]. L’expression des idées de sphère personelle et de solidarité dans les langues indo-européennes [The expression of concepts of the personal domain and indivisibility in Indo-Europan languages]. In Franz Fankhauser & Jakob Jud (eds.), Festschrift Louis Gauchat. Aarau: Verlag Sauerländer. Translated and reprinted in Hilary Chappell & William McGregor (eds.), The grammar of inalienability, 31–61. Berlin: Mouton de Gruyter.10.1515/9783110822137.31Search in Google Scholar

Bax, Marton. 2019. A sketch grammar of Gebe: A language of North Maluku. Leiden: Leiden University (MA thesis).Search in Google Scholar

Blust, Robert A. & Stephen Trussel. 2020. Austronesian comparative dictionary. http://www.trussel2.com/ACD/ (accessed 12 March 2021).Search in Google Scholar

Bowden, John. 2001. Taba: Description of a South Halmahera language. Canberra: Pacific Linguistics.Search in Google Scholar

Bubandt, Nils. 2014. The empty seashell. Ithaca, NY: Cornell University Press.10.7591/9780801471971Search in Google Scholar

Chappell, Hilary & William McGregor. 1996. Prolegomena to a theory of inalienability. In Hilary Chappell & William McGregor (eds.), The grammar of inalienability, 3–30. Berlin: Mouton de Gruyter.10.1515/9783110822137.3Search in Google Scholar

Dixon, Robert, M. W. 2010. Basic linguistic theory, vol. 2 Grammatical topics. Oxford: Oxford University Press.Search in Google Scholar

Donohue, Mark & Charles E. Grimes. 2008. Yet more on the position of the languages of eastern Indonesia and East Timor. Oceanic Linguistics 47(1). 114–158. https://doi.org/10.1353/ol.0.0008.Search in Google Scholar

Gasser, Emily, Laura Arnold & David Kamholz. forthcoming. The languages of Halmahera and West New Guinea. In Alexander Adelaar & Antoinette Schapper (eds.), The Oxford guide to the Western Austronesian languages. Oxford: Oxford University Press.Search in Google Scholar

Heine, Bernd. 1997. Possession: Cognitive sources, forces, and grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511581908Search in Google Scholar

Kamholz, David. 2014. Austronesians in Papua: Diversification and change in South Halmahera–West New Guinea. Berkeley, CA: UC Berkeley (dissertation).Search in Google Scholar

Koptjevskaja-Tamm, Maria. 2004. Maria’s ring of gold: Adnominal possession and nonanchoring relations in the European languages. In Yury A. Lander Kim Ji-yung & Barbara H. Partee (eds.), Possessives and beyond: Semantics and syntax, 155–181. Amherst, MA: GLSA Publications.Search in Google Scholar

Lichtenberk, Frantisek. 1983. Relational classifiers. Lingua 60(2–3). 147–176. https://doi.org/10.1016/0024-3841(83)90072-4.Search in Google Scholar

Maan, Govert. 1951. Proeve van een Bulische spraakkunst [Sketch grammar of Buli], vol. 10 Verhandelingen van hetKoninklijk Instituut voor Taal-, Land- en Volkenkunde. ’s-Gravenhage: Martinus Nijhoff.10.26530/OAPEN_613349Search in Google Scholar

Rødvand, Linn. 2023. Grammatical description of Patani. Manuscript in preparation.Search in Google Scholar

Ross, Malcolm. 1998. Possessive-like attribute constructions in the Oceanic languages of Northwest Melanesia. Oceanic Linguistics 37(2). 234–276. https://doi.org/10.2307/3623410.Search in Google Scholar

Stassen, Leon. 2009. Predicative possession. Oxford: Oxford Uniersity Press.10.1093/oso/9780199211654.001.0001Search in Google Scholar

Stokhof, A. L. (ed.). 1980. Holle lists: Vocabularies in languages of Indonesia vol. 2: Sula and Bacan islands, North Halmahera, South and East Halmahera. Canberra: Pacific Linguistics.Search in Google Scholar

van den Berg, René. 2009. Possession in South Halmahera–West New Guinea: Typology and reconstruction. In Alexander Adelaar & Andrew Pawley (eds.), Austronesian historical linguistics and culture history: A festschrift for Robert Blust, 327–358. Canberra: Pacific Linguistics.Search in Google Scholar

van den Heuvel, Wilco. 2006. Biak: Description of an Austronesian language of Papua. Amsterdam: Vrije Universiteit Amsterdam (dissertation).Search in Google Scholar

Whisler, Jacqueline. 1996. A grammar of Sawai. [120 pages. Unpublished Manuscript].Search in Google Scholar

Whisler, Ronald. 1992. Phonology of Sawai. In Donald A. Burquest & Wyn D. Laidig (eds.), Phonological studies in four languages of Maluku, 7–32. Dallas, TX: Summer Institute of Linguistics and the University of Texas at Arlington Publications in Linguistics.Search in Google Scholar

Published Online: 2023-09-18
Published in Print: 2023-09-26

© 2023 the author(s), published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Downloaded on 25.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/stuf-2023-2016/html
Scroll to top button