The Papuan Languages of Timor, Alor and Pantar: Volume 1 9781614515241, 9781614517238

Citation preview

Antoinette Schapper (Ed.) The Papuan Languages of Timor, Alor and Pantar: Volume 1

Pacific Linguistics

Managing editor Paul Sidwell Editorial board members Wayan Arka Mark Donohue Bethwyn Evans Nicholas Evans Gwendolyn Hyslop David Nash Bill Palmer Jane Simpson Andrew Pawley Malcolm Ross

Volume 644

The Papuan Languages of Timor, Alor and Pantar: Volume 1

Sketch Grammars

Edited by Antoinette Schapper

ISBN 978-1-61451-723-8 e-ISBN (PDF) 978-1-61451-524-1 e-ISBN (EPUB) 978-1-5015-0115-9 ISSN 1448-8310 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. 6 2014 Walter de Gruyter, Inc., Boston/Berlin Photo courtesy Emilie T.B. Wellfelt Typesetting: RoyalStandard, Hong Kong Printing and binding: CPI books GmbH, Leck ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Table of contents Abbreviations

vii

1

Antoinette Schapper 1 Introduction

2

Gary Holton Western Pantar

3

Marian Klamer 97 Kaera

4

Hein Steinhauer 147 Blagar

5

Laura C. Robinson & John W. Haan 221 Adang

6

Antoinette Schapper 285 Kamang

7

František Kratochvíl 351 Sawila

8

Antoinette Schapper & Rachel Hendery 439 Wersing

Index

505

23

Abbreviations Leipzig Glossing Rules (http://www.eva.mpg.de/lingua/resources/glossingrules.php) are used throughout this volume. The table below summarizes the glosses used. 1 2 3 4 A ABL ADD ADDR ADJ AGT AGR ALIEN ALL ANAPH

AP APPL ARG ART ASP ASSOC AST ATT ATTR AUX CAUS CLF CMN CMP COMPL CONC CONJ CONT

first person second person third person fourth person agent-like argument of transitive verb ablative additive addressee adjective agentive agreement alienable allative anaphoric Alor-Pantar languages applicative argument article aspect associative plural assistive attenuated affectedness attribute auxiliary causative classifier common person complement relations completive connector conjunction continuative

CONTR CSEQ DAT

DC DEF DEV DEM DET DIR DISC DIST DISTR DU EMPH EXCL EXCLAM EXT FIN FOC GRAD GRP GEN HIGH HUM IMP INAL INC INCP INCL INCL + INCOMPL INCORP INDEP

contrastive consequential relations dative deictic center definite direct evidential demonstrative determiner directive discourse marker distal distributive dual emphasis exclusive exclamative existential final focus gradation marking group genitive elevation higher than DC human imperative inalienable inchoative inceptive inclusive all inclusive incompletive incorporated independent

viii

INT INTER INV IPFV ITER LEVEL LIM LOC LOW LMT LNK MAN NEG NFIN NMLZ

NP NMBR NUM NVIS OBJ OBL

P PART PAT PFV PL POSS POT

Abbreviations

intentional interjection inverse imperfective iterative elevation level with DC limitation marking locative elevation lower than DC limited affectedness linker manner marking negative non-final nominalizer noun phrase number-marked numeral non-visible object oblique patient-like argument of transitive verb particle patientive perfective plural possessive potentive

PP PRED PRIOR PROSP PROG PROH PROX PSR QUANT RC REAL RECP REFL REL RESTR RDP

S SBEN SBJ SET SG SPEC SPKR TAG

TAP TOP VIS

postpositional phrase predicate priorative prospective progressive prohibitive proximal possessor quantifier relative clause realis reciprocal reflexive relative restrictive reduplicant single argument of intransitive verb self-benefactive subject setting relations singular specific speaker question tag Timor-Alor-Pantar languages topic visible

Antoinette Schapper

1 Introduction to The Papuan languages of Timor, Alor and Pantar. Volume I 1 Overview The Papuan languages of Timor, Alor and Pantar are a small family comprising roughly 30 languages spoken by around 300,000 people in south-eastern Indonesia and Timor-Leste (East Timor). Members of the family predominate on the islands of the Alor archipelago, consisting of Alor, Pantar and numerous small islands in the Pantar Strait. Other members are found interspersed among Austronesian languages on the islands of Timor and Kisar.1 Lying some eight hundred kilometres from the New Guinea mainland as the crow flies, TimorAlor-Pantar (TAP) languages are the western-most of the Papuan languages (Map 1). “Papuan” is a negative label indicating any language that is found in the area of New Guinea that does not belong to the Austronesian language family (Foley 1986: 1). This book is the result of an eventful last decade in TAP linguistics (see section 4). The volumes were conceived as a way for numerous researchers to bring the major descriptive results of their work together in a cohesive and focussed forum. All sketches are written by expert linguists based on their own original fieldwork into previously undescribed TAP languages, and together provide an authoritative and comprehensive overview of the grammars of this Papuan “outlier” family.

2 The language scene On Timor, there are four TAP languages: Bunaq straddling the border between Indonesian West Timor and independent East Timor, and Makasae, Makalero and Fataluku occupying a contiguous region at the eastern tip of the island 1 Wurm & Hattori (1982) list Adabe on the island of Atauro just east of Alor and north of Timor as being another Papuan language. Hull (1998: 3–4) suggests that this error came about through a misreading of an ambiguous missionary statement. A visit by myself to Adabe village on Atauro in 2007 yielded no trace of a Papuan language. David Penn (pers. comm.) states that his own survey work on Atauro found only a series of closely-related Austronesian dialects.

Map 1: The extent of Papuan languages (on and around New Guinea)

2 Antoinette Schapper

Introduction to The Papuan languages of Timor, Alor and Pantar

3

(Map 2). Close by to the north, again in Indonesia, Oirata is a TAP language spoken on Kisar Island which is dominated by the Austronesian language, Meher. The Timor languages are far and away the largest in the family: Bunaq, Fataluku and Makasae together make up seventy per cent of TAP language speakers. Although existing under the shadow of more dominant Austronesian languages, the Timor-Kisar languages are vigorous and still being learnt by children in the villages. It is clear that Timor groups have even significantly expanded in the modern period, taking over areas previously occupied by speakers of neighboring Austronesian languages (McWilliam 2007, Engelenhoven 2010, Schapper 2011a). The Oirata language is itself the result of an historical migration of Fataluku speakers from the northern dialect area into Austronesian territory on Kisar.2 The greatest concentration of TAP languages is in the Alor archipelago, with 26 Papuan languages on current counts (Map 3). This classification is based on a combination of mutual intelligibility, ethno-social considerations and non-trivial grammatical differences. Yet the exact number of languages remains difficult to determine. This is due to two factors. Firstly, it is difficult to sort out mutual intelligibility from passive bilingualism in a context such as Alor-Pantar where multilingualism between neighboring groups is the norm. Kamang, Suboo, Moo, Tiee and Manet in central Alor have in earlier classifications been marked as a single language (e.g., as Woisika in Stokhof 1975 on the basis of 117-item wordlists). The languages are indeed closely related (forming the North-central Alor language area, Schapper & Manimau 2011) and speakers typically claim partial passive competence in one or more of the other groups’ speech. Still the variation between lects is not superficial. Recent investigation has revealed significant differences in the details of the segmental phonology and morphology (such as in the number and functions of agreement paradigms) that appear to warrant each of them being regarded as an independent language. Secondly, in Alor-Pantar unique glottonyms that are widely recognized and accepted are lacking for many languages. Instead, a language may be referred to with the name of the clan that speaks it; for instance, Lamma, used by Stokhof (1975) and Nitbani et al. (2001), is a single clan within the larger Western Pantar language area (Holton this volume), or Tanglapui, used by Stokhof (1975) and Donohue (1996), is the name of one of several Kula-speaking clan groupings. Alternatively, a language may be referred to by the name of the village where it is spoken or the geographical location of its speakers: for instance, the Wersing 2 According to Dutch historical records, the Oirata arrived in Kisar in 1721, having fled from Loikera in the Fataluku-speaking area (Riedel 1886: 403). Still today the northern Fataluku dialect is claimed to be, at least partly, mutually intelligible with Oirata (Katrina Langford pers. comm.), though other Fataluku dialects appear to be too divergent (Aone van Engelenhoven pers. comm.).

Map 2: The extent of the Timor-Alor-Pantar family

4 Antoinette Schapper

Map 3: The Alor-Pantar languages

Introduction to The Papuan languages of Timor, Alor and Pantar

5

6

Antoinette Schapper

language is often designated locally as Bol, though this is a general label that can strictly speaking be applied to any coastal group, while Wersing speakers themselves tend to refer to their language as Bahasa Kolana, Bahasa Pureman etc. after their respective villages of origin. In other cases, a language may be designated simply as our language, as is the case with many Blagar varieties (Steinhauer this volume). The largest and most vibrant of the Alor-Pantar languages are Abui and Western Pantar, with 16,000 and 10,000 speakers respectively and with active child acquisition. Other languages such as Nedebang are near-moribund with only a few dozen adult speakers remaining. For the majority of Alor-Pantar languages, the situation is not so dire but language shift is progressing. The typical situation in Alor-Pantar villages may be broadly characterized as follows: many adults still use the local languages regularly but switching to Malay/Indonesian is common particularly in addressing children; some but not all children acquire the local language, but often only passively, and they prefer to use Malay/ Indonesian with their peers.

3 Prehistory of the family The relatedness of the TAP languages has traditionally been assumed by scholars from their proximity. However, it was only very recently that the languages’ genetic unity has been shown by the comparative method (Schapper et al. forthcoming). The primary subgroups of the family, correlating with geography, are the Timor subgroup and the Alor-Pantar subgroup (Figure 1). While the Timor subgroups are easily established (Schapper et al. 2012), the internal relationships of the Alor-Pantar languages are more complex and as yet little worked out. Figure 2 presents a possible tree of the Alor-Pantar subgroup, synthesizing the results of a variety of comparative studies of phonology and morphology (Donohue & Schapper 2007, 2008, Holton et al. 2012, Robinson & Holton 2010, Robinson & Holton 2013, Schapper forthcoming, Schapper & Klamer forthcoming, Steinhauer 1995) as well as ongoing reconstructive work. Within the Alor-Pantar subgroup, the West Alor subgroup and the lower nodes within Central-East Alor and Central-North Pantar subgroups seem relatively secure and well-defined. The high-level nodes for Pantar and Alor are shaky. The position of Western Pantar, Nedebang, Kui-Kiraman and Abui-Papuna is also extremely problematic. Subsequent work will likely revise their constituency within the Alor-Pantar tree, possibly in favour of a flatter structure with ‘linkages’, resulting from layered innovations with entangled patterns of distributions between languages.

Introduction to The Papuan languages of Timor, Alor and Pantar

7

Figure 1: Tree diagram of Timor-Alor-Pantar languages

Despite the genetic unity of the family now being clear, the prehistory of the Timor-Alor-Pantar languages remains largely obscure. The origin and timing of the arrival of the ancestors of TAP-speaking peoples in the Timor area is a matter of particular conjecture. It is generally thought that the TAP languages are the descendants of the languages of the autochthonous peoples preceding the Austronesian arrival after 2000 BCE.3 They may be a remnant of the languages spoken by the first settlers of the area who arrived during the Pleistocene more than 40,000 years ago (O’Connor 2007). However, it may also be that the protoTAP peoples themselves had arrived from New Guinea as part of an expansion powered by the development of taro and banana agriculture in the eastern highlands of New Guinea around 8000 BCE (Pawley 2005). On the back of these agricultural advances, the languages of the Trans New Guinea (TNG) Phylum are thought to have spread along the central cordillera of New Guinea and then possibly on to the Timor region. There are around a dozen pTNG lexical items that have look-alikes in modern TAP languages (Pawley & Hammarström forthcoming). The lexical similarities are not enough to establish the regular sound correspondences demanded by the comparative method, but they do seem to point to some connection between TAP and mainland TNG languages. The weakness of the signal may be a reflection of the great time-depth that presumably separates the TAP languages from pTNG, leaving the TNG hypothesis as it touches on TAP ultimately unfalsifiable. 3 McWilliam (2007) suggests that the Fataluku in eastern Timor may represent a post-Austronesian arrival in Timor. This is, however, at odds with the genetic unity of the Timor-Alor-Pantar family and in particular conflicts with the clear subgrouping of Fataluku with Makasae and Makalero. As argued by Schapper (2011b), the Austronesian cultural patterns observed amongst the Fataluku by McWilliam can readily be explained as the outcome of significant and prolonged contact with a more prestigious neighboring cultural group.

Figure 2: Possible tree diagram of the Alor-Pantar languages

8 Antoinette Schapper

Introduction to The Papuan languages of Timor, Alor and Pantar

9

Robuster evidence of a TNG relationship might be found by comparing pTAP with a lower level, more closely related subgroup within the TNG Phylum. Like TAP languages, the Papuan languages of West Bomberai have been tentatively grouped into the TNG “Western Linkage” by Ross (2005).4 Within the linkage TAP and West Bomberai are seen to group together because of the shared innovation of a 1st person plural pronoun *bi (Ross 2005: 36), reflected as pTAP *pi ‘1PL . INCL ’ (Schapper et al. forthcoming) and Iha mbi ‘1PL . INCL ’ (Mark Donohue pers. comm.). There is also archeological evidence linking the Timor region and the Bomberai peninsula, with a striking similarity observed in rock art motifs in eastern Timor and West Bomberai (O’Connor 2003). The relationship between TAP languages and Papuan languages of West Bomberai deserves serious attention following more complete reconstructions of the two proto-languages, with which the TNG hypothesis can be tested at its lowest level (see Robinson and Holton 2012a for a preliminary attempt at this). Only with this will we bring some serious historical perspective – most importantly the application of the comparative method – to the validity of TNG at its claimed fringes. The outlying position of the TAP languages means that they have had a different history from that of many mainland New Guinea Papuan languages. In particular, they have been in close contact with Austronesian languages for at least 2,000 years and this has wrought many changes on the languages that have only begun to be explored (e.g., in Engelenhoven 2010, Klamer 2012, McWilliam 2007, Schapper 2011a, 2011b). A major question in the history of TAP-Austronesian interaction is how the speakers of TAP languages resisted the incoming tide of Austronesians. The conventional understanding of Austronesian history involves the expansion across eastern Indonesia, overwhelming preexisting sparse populations of Papuan hunter-gatherers and transforming them into speakers of Austronesian languages (Bellwood 1998). Yet the TAP languages survive today, and we may speculate that their speech communities possessed some characteristics that allowed them to maintain themselves in the face of the Austronesian arrival. An incumbent agriculturist population (see Oliveira 2008 on the signs of early Holocene agriculture, such as domesticated taro Colocasia esculenta, in Timor) or a highly organized maritime culture (see Schapper & Huber 2012 for this suggestion) would perhaps allow resistance against the incoming Austronesians. It is hoped that further comparative-historical research may shed light on this issue.

4 The possibility of a TNG link between TAP and Bomberai languages was first floated by Hull (2004).

10

Antoinette Schapper

4 History of study of the family The TAP languages remained in almost complete obscurity until decades into the 20th century. The earliest publication with data from a TAP language was the short wordlists in Anonymous (1914) and in Vatter (1932), but without recognition of the languages’ Papuan character. It was Josselin de Jong’s (1937) monograph that first identified the presence of Papuan languages with Oirata on Kisar and by implication Fataluku on Timor, which Josselin de Jong named as a close relative. Nicolspeyer (1940) provided the first glossed texts of a TAP language and hinted at the Papuan character of the Abui language. This was followed by Capell’s (1943) identification of Bunaq and Makasae on Timor as Papuan. The Papuan classification of the Alor-Pantar languages lagged behind: they were still marked as Austronesian in Salzner’s (1960) Sprachenatlas des Indopazifischen Raumes, but by the time of Wurm and Hattori’s (1982) monumental Language atlas of the Pacific area all were given as Papuan (following Stokhof 1975; see below). Detailed work on the TAP languages began only after World War II. António de Almeida, head of the Portuguese Missão Antropológica de Timor, collected word lists and elicited sentences in most languages of the Portuguese-held part of Timor between 1953 and 1975 (partly published as Almeida 1994). Alfonso Nacher, a priest at the Missão Salesiana in Fuiloro, compiled a dictionary of Fataluku in the period between 1955 and 1968 (published as Nacher 2003, 2004, and then again as Nacher 2012). Louis Berthe started field work on Bunaq in what is today Indonesian West Timor between 1957 and 1959 and published a range of anthropological linguistic materials (Berthe 1959, 1963, 1972). Berthe returned to Timor between 1966 and 1967 leading a multi-disciplinary team, with Henri Campagnolo who produced the first in-depth grammatical description of Fataluku (Campagnolo 1973) and Claudine Friedberg who studied ethno-linguistic classification of plants amongst the Bunaq (1970, 1972, 1974, 1979, 1990). On Alor and Pantar, linguistic documentation took off in the 1970s. A better view of the Pantar languages was gained with the word lists in Watuseke (1973) and the fieldwork completed by James Fox (nd) between 1972 and 1973. Extensive fieldwork was carried out by Wim Stokhof and Hein Steinhauer as part of a project by Netherlands Foundation for the Advancement of Tropical Research (see Steinhauer & Stokhof’s 1976 project report). Out of this project, first came Stokhof (1975) containing 34 word lists taken from a range of locations across the Alor-Pantar languages. Stokhof’s fieldwork was concentrated in Alor and resulted in a significant body of materials on Kamang under the name Woisika (Stokhof 1977, 1978, 1979, 1982, 1983), as well as short texts in Abui (Stokhof 1984) and Kabola (Stokhof 1987). Steinhauer’s fieldwork concentrated on Blagar,

Introduction to The Papuan languages of Timor, Alor and Pantar

11

working chiefly on the Dolabang dialect of Pura Island (Steinhauer 1977, 1991, 1993, 1999, 2010), but he also collected materials on the Bukalabang dialect on Pantar (Steinhauer 1995). No new work was carried out on TAP languages in the 1980s. The 1990s saw the Indonesian national Pusat Pembinaan dan Pengembangan Bahasa (Center for Language Development and Cultivation, later renamed Pusat Bahasa ‘Center for Language’5), conduct research on a range of TAP languages. The center’s fieldwork resulted in a general survey on the languages of Timor (Sudiartha et al. 1994), error-ridden sketches of Bunaq (Sawardo et al. 1996) and of Makasae (Sudiartha et al. 1998), as well as survey word lists of the languages of Alor (Martis et al. 2000) and a problematic sketch of Lamma (Nitbani et al. 2001). Other than these only a few studies appeared: Donohue (1996, 1997) on Kula, Wersing and Kui, and Marques (1990) with a primer of Makasae. The turn of the century saw a huge boom in TAP linguistics, with a surge in descriptive work by a suite of linguists from around the world. Within a decade, modern grammatical descriptions have been produced for over half a dozen TAP languages: Haan’s (2001) Adang grammar, Huber’s (2005) Makasae grammar (later published as Huber 2008), Kratochvíl’s (2007) Abui grammar, Baird’s (2008) Klon grammar, Schapper’s (2010) Bunaq grammar, Klamer’s (2010) Teiwa grammar, and Huber’s (2011) Makalero grammar. Additional materials and several monograph-length treatments of particular topics have also appeared for these TAP languages as well as others. Most significant are: for Abui, an introductory dictionary (Kratochvíl & Delpada 2008a) and texts (Kratochvíl and Delpada 2008b); for Bunaq, a short dictionary (Bele 2009); for Fataluku, a monolingual dictionary (Valentim 2002), primer materials and texts (Valentim 2001a, 2001b, 2004a, 2004b); for Kamang, an introductory dictionary (Schapper & Manimau 2011); for Makalero, a short dictionary (Pinto 2004) and a primer (Pinto 2007); for Makasae, dissertations on space (Brotherson 2003), genres of verbal art (Carr 2004), and phonology (Fogaça 2011), a primer (Hull 2005), and two short dictionaries (Ximenes and Menezes 2002, Hull & Correira 2006); for Oirata, a sketch with Josselin de Jong’s (1937) text glossed (Faust 2005), and a historical phonological treatment (Mandala 2010); for Teiwa, an extended word list (Klamer 2011); for Western Pantar, an introductory dictionary (Holton and Lamma Koly 2008).

5 Basic typological overview Out of this great surge of work, we have a broader picture of the variation and diversity of the TAP languages. This section presents a short, but by no means 5 Subsequent to these publications, this institution was rebranded Badan Pengembangan dan Pembinaan Bahasa.

12

Antoinette Schapper

comprehensive, typological overview of the family. It functions as a taster to the descriptions found in these volumes. TAP languages have crosslinguistically inconspicuous phonologies. Consonantal inventories are moderately small with an average of 13–16 consonant phonemes. The largest consonant inventory is found in Western Pantar with 16 simple plus 10 geminate consonant phonemes, followed by Teiwa with 20 (non-geminate) consonant phonemes. Makalero has the smallest consonant inventory with only 11 phonemes, and is unique in entirely lacking a voicing distinction in its plosives. Glottal stop is universally present in the phoneme inventories of Timor languages (although lost in some individual dialects), but only sporadically found in Alor-Pantar. Velar nasals are absent in Timor languages, but are frequently present in Alor-Pantar, though often have dubious phonemic status (e.g., Western Pantar, Kaera) or are clear allophones of /n/ (e.g., Wersing). Fricatives in most languages are limited to /s/ and /h/, with sporadic occurrences of (bi)labial fricatives (e.g., Teiwa, Makalero, Fataluku), /z/ (e.g., Blagar and Bunaq), /x/ (e.g., Teiwa and Kaera). Unusual consonants are /ћ/ and /q/ in Teiwa, /ɓ/ in Blagar and Reta, and /ʈ / in Oirata. Vowel systems are varied. Most common in the TAP family are systems with five cardinal vowels and a length distinction for all qualities (e.g., Kaera, Kui, Abui and Kamang), though in some languages long vowels are only marginal (e.g., Blagar and Makalero). Five cardinal vowel systems with no length distinction are found in Western Pantar, Wersing, Bunaq, Makasae and Fataluku. Contrasting heights of mid-vowels are rare, found in Adang (seven vowel phonemes with two mid-vowels front and back and no length distinction) and Klon (12 vowels with marginal short /e/ and /ɔ/ contrasting with /ɛ, ɛː/ and /o, oː/). Unusual vowels are Sawila front-rounded /y, yː/ contrasting with five short and long cardinal vowels, and Kula extra-short central vowels /ɪ/ and /ɐ/ contrasting with five full cardinal vowels. Bunaq is atypical in having three phonemic diphthongs that contrast with sequences of the same vowels. The most common syllable structures are (C)V and (C)VC, frequently with restrictions on the consonants permitted in codas. In most languages, consonant clusters are rare, even across syllable boundaries. Where there are underlying clusters, surface phonotactic processes often exist to break them up (e.g., vowel epenthesis in Klon and Wersing). Unusual features are the extensive appearance of echo vowels in Wersing to avoid final codas, and the productive morphophonemic process of high vowel metathesis found Bunaq and Wersing to maintain a CV structure in certain affixation environments. A range of stress systems is found in TAP languages. Non-phonemic stress comes in different forms: penultimate stress is found in Western Pantar, Blagar and Bunaq, final stress in Kaera, and weight-sensitive stress in Klon and Abui. Fully or at least partially phonemic stress is also found in Alor-Pantar languages

Introduction to The Papuan languages of Timor, Alor and Pantar

13

such as Teiwa, Kamang and Sawila. Tone is almost unknown in the TAP languages. A pitch-accent system is reported for one dialect of Fataluku (Stoel 2007), and tone is said to play a limited lexical role in a few domains of Abui. Morphological profiles show significant variation across the family. Makasae is highly isolating with no productive morphology whatsoever. Teiwa and Bunaq have inflectional morphology limited to a single paradigm of person prefixes that appear on some verbs and nouns (Teiwa eight prefixes and Bunaq five), while Makalero has a single 3rd person prefix appearing on a highly restricted number of verbs. At the other extreme, Abui and Kamang have five and seven paradigms of person prefixes respectively. West Alor languages typically have three paradigms, east Alor two paradigms and Pantar one paradigm. Abui and Kamang are also unique in each having several aspectual and dependencymarking suffixes. Elsewhere such morphology is limited: A single dependency suffix is found in both Kaera and Blagar, a realis suffix in Teiwa and Wersing, and aspectual prefixes are known in Western Pantar. Derivational morphology is, with some exceptions, limited in the family. A causative prefix is found in Adang, a causative suffix in Kaera, and both a causative prefix and suffix that are used in concert on some verbs in Blagar. Whilst lacking in many languages (e.g., Western Pantar, Kaera, Blagar, Abui and Bunaq), applicatives are still relatively widespread in the family: Adang and Teiwa have one applicative prefix each, Klon, Kamang and Wersing two, and Sawila three; Fataluku and Makalero have more than a dozen applicative-like locative prefixes that attach to verbs triggering consonant mutations on them. Nominalizing suffixes of some productivity are found in Blagar, Makalero and Fataluku. Fataluku and Oirata are distinct in having a large number of precategorical roots that must be marked with one of two suffixes depending on whether their use is nominal or verbal. Kula and Sawila are unique in the family in having lexicalized forms of many verbs and nouns depending on whether they occur in phrase-final or non-final position. Kaera has a related but distinct phenomenon: vowel suffixes mark phrase-final verbs ending in a consonant. TAP systems of morphological alignment hold especial typological interest due to their diversity and the frequent appearance of elements of semantic alignment. Outside of Alor, TAP languages have one prefixal agreement paradigm with a range of different alignment patterns. Western Pantar has a very fluid alignment system in which S, A or P can be indexed by an agreement prefix on the verb with different prefixation choices reflecting different levels of participant volitionality and affectedness (Holton 2010). Blagar has accusative alignment with only P ever being marked on the verb, while languages such as Bunaq, Teiwa and Kaera have “leaky” accusativity, in which prefixes on most verbs agree with P, but there are also small classes of S-agreement verbs. In western

14

Antoinette Schapper

and central Alor, most languages have multiple agreement paradigms. For instance, Klon has three, Abui five, and Kamang six that can be used for either S or P. Multiple paradigms means that these languages don’t just have one coding ‘split’, but multiple splits in the prefixation of S and P with varying degrees of semantic flexibility versus lexical stipulation in the individual languages. Kula and Sawila have partly lexicalized inverse alignment systems involving an inverse prefix na- in combination with different agreement paradigms. Wersing has one agreement paradigm obligatorily agreeing with S, P or A depending on the lexical verb, but also a second, semantically governed agreement paradigm that exists alongside this encoding “inceptive agents”. TAP languages present a relatively uniform word order profile. All languages have SV/APV word order, postpositions and a basic head-initial NP. Major differences in the word order of the clause can be observed in the verb finalness of the languages: whilst TAP languages typically allow a few elements to appear after the verb, most prominently the clausal negator and TAME markers, eastern Timor languages are strictly verb final with all such elements occurring before the clausal verb. At the other extreme, Bunaq, in central Timor, permits a great many elements to follow the verb, with strings of up to eight post-verbal elements encoding aspect, temporal duration, manner, addition, polarity and information being observed. Within the NP, word order differences are minimal. All TAP languages have basic GEN N, N ATTR 6, N DEM /ART and N NUM word orders. Individual languages may show optional deviations from this. For instance: Abui demonstratives can precede the noun in some limited contexts; Fataluku has a set of demonstrative particles that precede the noun as well as a set of determiners that follow it, as well as distinct pre-posed and post-posed possessor constructions; Bunaq allows some possessors to be post-posed and has a set of deictic “locational” morphemes that precede the noun. In TAP languages, nouns are typically unmarked for case, number, and gender. Exceptions are: Makalero and Fataluku have plural marking suffixes available on a small number of, predominantly kinship, nouns, and Makalero and Kamang have associative plural suffixes; Bunaq has a covert grammatical gender system based on an animate-inanimate distinction expressed on determiners and verbal agreement. Possessive classification in which a lexically determined subset of nouns, typically kinship and body part nouns, gets distinctive possessor coding is found in all TAP languages. Languages such as Teiwa, Wersing and Makalero have only one way of marking possession: most nouns are only optionally marked with the possessor prefix, whilst a small class are obligatorily 6 The label “attribute” is used here rather than “adjective” because most TAP languages lack a distinct class of adjectives. Relative clauses are also often absent.

Introduction to The Papuan languages of Timor, Alor and Pantar

15

marked by it. Other languages have two distinct ways of marking possessors, one for optionally possessed alienable nouns and one for obligatorily possessed inalienable nouns. In languages such as Blagar, Klon and Bunaq, the alienable possessor is encoded “indirectly” by means of an inflected free possessor classifier, while inalienable possessors are encoded “directly” by a means of a prefix on the possessed noun. In other languages such as Abui, two different sets of prefixes encode alienable and inalienable nouns respectively. In some cases, systems are more complex. In Western Pantar and Kamang, for instance, there are the usual two classes that are optional and obligatory possessors encoded in two different ways respectively, plus a third class which is obligatorily possessed, but uses the markers of the optional and not the obligatory class. Amongst the most interesting linguistic phenomena in TAP languages are their intricate deictic systems involving features such as visibility, knowledge, contrastiveness, distance, and elevation. Whilst elevation distinctions are common in Papuan languages, the TAP systems are remarkable for their complexity, incorporating three elevational heights (above, below and level with the deictic center), often with multiple sub-terms within an elevational height distinguishing features such as distance (Abui), steepness of the slope (Western Pantar, Kamang), and directionality (traversing across the slope or directly following it; Makasae, Kamang, Sawila and Kula). Many TAP languages also do not limit their elevational distinction to a single set of elevational terms, but reiterate the elevational distinctions across multiple parts of the lexicon: Blagar marks elevation in a total of 10 paradigms, Western Pantar eight, and Makasae in seven (see Schapper forthcoming for further discussion). A final important aspect of the typology of TAP languages is the prevalence of serial verb constructions. These are used to encode additional participants, cause, manner, direction, result and aspect, to name just a few functions they have in the clause. The distinction between postpositions and participant-adding serial verbs is often not well articulated in TAP languages, and the array of variation in the languages provides fascinating material for the grammaticalization cline between verb and postposition. Bunaq, for instance, has a large set of “verbal postpositions” which have verb-like inflectional properties, but lack the ability to occur as independent clausal predicates. In other languages such as Kamang, we find both true postpositions and a postposition-like light verb that has a reduced, non-inflecting form occurring clause-medially to add participants. Other light verbs in TAP languages have become fused with other serial verbs (Blagar, Teiwa), become obligatory markers of certain roles (Makasae, Fataluku) and in other cases have become prefixes on verbs (Makalero, see Klamer and Schapper 2012 for a treatment of this variation in one TAP serialization construction).

16

Antoinette Schapper

6 This book Despite the profusion of activity in TAP linguistics, there remains little data on the TAP languages in wide circulation among linguists. Published descriptions of TAP languages still only cover a small number of languages and do not capture the full extent of the diversity within the family. The present volumes bring together grammatical descriptions of previously undescribed TAP languages, thus providing a comprehensive overview of the variation across the entire TAP family and bringing to light the many noteworthy features of individual TAP languages. Sketch chapters are highly structured, with each author providing information on the same important topics in the family: phonology, clause structure, noun phrases, verbal morphology in particular person-number prefixes, independent pronouns, serial verb constructions, and TAME marking. Readers with interests in diverse topics such as stress systems, reduplication, demonstratives and deixis, numeral systems, agreement and alignment, as well as verb serialisation and more, will discover aspects of the TAP languages captivating and stimulating. TAP linguistics is a relatively new field and its terminology still shows flux (e.g., pronominal prefixes versus agreement prefixes). The largely parallel structures of chapters, however, allow the reader to directly compare phenomena between languages. Detailed tables of contents at the beginning of each chapter enable the reader to navigate quickly to treatments of desired topics. IPA has been used in the phonology sections of all sketches, but other sections use the individual orthographies of the different authors. Readers should keep in mind these potential differences. This volume, Volume I, presents the first published grammar overviews of seven languages of the Alor-Pantar (AP) subgroup. These cover the full geographic spread of the AP languages: Western Pantar (Holton this volume) located in the south-west of Pantar, Kaera (Klamer this volume) in the north-east, Blagar (Steinhauer this volume) in eastern Pantar and dispersed throughout the straits between Pantar and Alor, Kamang (Schapper this volume) in north-central Alor, and Sawila (Kratochvíl this volume) and Wersing (Schapper and Hendery this volume) situated in the far south-east of Alor. Map 4 illustrates the extent of coverage by the sketches, giving an overview of the major fieldwork sites and places of study in this volume. Volume II presents descriptions of additional Alor languages and the major languages of the Timor subgroup of the family. Together, these volumes represent a significant advancement of descriptive Papuan linguistics. There is still ground to be covered before we have a complete picture of the TAP languages, but with the appearance of these volumes we now have descriptions for around three quarters of the family, making it

Map 4: Major field sites and places of study in this volume

Introduction to The Papuan languages of Timor, Alor and Pantar

17

18

Antoinette Schapper

one of the best described Papuan families. That said, almost all aspects of the TAP languages are in need of future analytic work. Investigations of phonetics, semantics, discourse and sociolinguistics remain all but non-existent in the TAP languages. Comprehensive comparative study of the lexicon and grammar has until recently been impossible. These volumes will also aid in the comparison of TAP with other Papuan languages, and contribute to the construction of a better picture of Island Southeast Asian and Melanesian prehistory.

References Almeida, António de. 1994. O Oriente de Expressão Portugesa. Lisbon: Fundação Oriente e Centro de Estudos Orientais. Anonymous. 1914. De eilanden Alor en Pantar, Residentie Timor en Onderhoorigheden. Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genootschap 31: 70–102. Baird, Louise. 2008. A Grammar of Klon: A non-Austronesian language of Alor, Indonesia, (Pacific Linguistics 596). Canberra: Pacific Linguistics. Bele, Anton. 2009. Kamus bahasa Buna’-Indonesia. Kupang: Gita Kasih. Bellwood, Peter. 1998. The archaeology of Papuan and Austronesian prehistory in the northern Moluccas, eastern Indonesia. In Roger Blench & Matthew Spriggs (eds), Archaeology and language II; Correlating archaeological and linguistic hypotheses, 128–124. London/New York: Routledge. Berthe, Louis. 1959. Sur quelques distiques bunaq (Timor central). Bijdragen tot de Taal-, Land- en Volkenkunde 115: 336–371. Berthe, Louis. 1963. Morpho-syntaxe du Bunaq (Timor central). L’Homme 3(1): 106–116. Berthe, Louis. 1972. Bei Gua, itinéraire des ancêtres, mythes des Bunaq de Timor. Paris: CNRS. Brotherson, Anna. 2003. A spatial odyssey: referring to space in Makasai. Honours dissertation, Australian National University, Canberra. Campagnolo, Henri. 1973. La Langue des Fataluku de Lórehe (Timor Portugais). PhD thesis, Université René Descartes, Paris. Capell, Arthur. 1943. Peoples and languages of Timor. Oceania 14(4): 311‒37. Carr, Felicita M. 2004. Pupuuk, Masu Moru and Data: Three genres of Makasai. Master of Linguistics thesis, The Australian National University, Canberra. Donohue, Mark. 1996. Inverse in Tanglapui. Language and Linguistics in Melanesia 27(2): 101– 118. Donohue, Mark. 1997. Developments of affectedness marking in the languages of Alor (split, splat). Talk handout, University of California Berkeley, 25 September. Donohue, Mark & Antoinette Schapper. 2007. Towards a morphological history of the languages of Timor, Alor and Pantar. Paper presented at The fifth East Nusantara Conference, Universitas Nusa Cendana, Kupang, Nusa Tenggara Timur, Indonesia, August 1–3 2007. Donohue, Mark & Antoinette Schapper. 2008. Whence the Austronesian indirect possessive construction? Oceanic Linguistics 47(2): 316–327. Engelenhoven, Aone van. 2010. The Makuva Enigma: Locating a hidden language in East-Timor. Revue roumaine de linguistique 80(2): 161–181.

Introduction to The Papuan languages of Timor, Alor and Pantar

19

Faust, Vera. 2005. Oirata, a language of Kisar, 1. Grammatical sketch, 2. Myth of Creation (glossed text). MA thesis, Leiden University. Fogaça, Jessé Silveira. 2011. Fonética e fonologia da língua Makasae. MA thesis, Universidade de Brasília. Foley, William A. 1986. The Papuan Languages of New Guinea. Cambridge: Cambridge University Press. Fox, James. n.d. Some notes on the languages of Pantar. Ms., PARADISEC Series AC2-INPANT. Friedberg, Claudine. 1970. Analyse des quelques groupements de végétaux comme introduction a l’étude de la classification botanique bunaq. In J. Pouillon & P. Maranda (eds), Echanges et Communications: mélanges offerts à Claude Lévi-Strauss à l’occasion de son 60ème anniversaire, 1092–1131. The Hague: Mouton. Friedberg, Claudine. 1972. Eléments de botanique bunaq. In J. Barrau (ed), Langues et techniques nature et société. Vol. 2. Paris: Klincksieck. Friedberg, Claudine. 1974. Les processus classificatoires appliqués aux objets naturels et leur mise en évidence: quelques principes méthodologiques. Journal d’agriculture tropicale et de botanique appliqué 21(10): 313–334. Friedberg, Claudine. 1979. Socially significant plant species and their taxonomic position among the Bunaq of central Timor. In Roy Ellen & David Reason (eds), Classifications in their social context, 81–101. London, New York: Academic Press. Friedberg, Claudine. 1990. Le savoir botanique des Bunaq: Percevoir et classer dans le Haut Lamaknen (Timor, Indonésie). Paris: Éditions du Muséum. Haan, Johnson. 2001. The grammar of Adang, a Papuan language spoken on the island of Alor, East Nusa Tenggara, Indonesia. PhD thesis, University of Sydney. Holton, Gary. 2010. Person-marking, verb classes, and the notion of grammatical alignment in Western Pantar (Lamma). In Michael Ewing & Marian Klamer (eds), Typological and Areal Analyses: Contributions from East Nusantara, 101–121. Canberra: Pacific Linguistics. Holton, Gary & Mahalalel Lamma Koly. 2008. Kamus pengantar Bahasa Pantar Barat (Companion dictionary of Western Pantar). Kupang, Indonesia: UBB-GMIT. Holton, Gary, Marian Klamer, František Kratochvíl, Laura C. Robinson & Antoinette Schapper. 2012. The historical relation of the Papuan languages of Alor and Pantar. Oceanic Linguistics 5(11): 86–122. Huber, Juliette. 2005. First steps towards a grammar of Makasae. MA thesis, Zürich University. Huber, Juliette. 2008. First steps towards a grammar of Makasae. München: Lincom. Huber, Juliette. 2011. A grammar of Makalero: A Papuan language of East Timor. Utrecht: LOT. Hull, Geoffrey. 1998. The languages of East Timor 1772–1997: A Literature Review. Studies in Languages and Cultures of East Timor 1: 1–38. Hull, Geoffrey. 2004. The Papuan Languages of Timor. Studies in Languages and Cultures of East Timor 6: 23–99. Hull, Geoffrey. 2005. Makasai. East Timor language profile no. 6. Dili: Instituto Nacional de Linguística. Hull, Geoffrey & Adérito José Gueterres Correira. 2006. Disionáriu badak Makasae-Tetun, TetunMakasae = Disionari gi digara Makasae-Tetun, Tetun-Makasae. Dili: Instituto Nacional de Linguística. Josselin de Jong, J.P.B. 1937. Studies in Indonesian culture I: Oirata, a Timorese settlement on Kisar. Amsterdam: Noord-Hollandsche Uitgevers-Maatschappij. Klamer, Marian. 2010. A Grammar of Teiwa. Berlin: Mouton. Klamer, Marian. 2012. Papuan‐Austronesian language contact: Alorese from an areal perspective. In Nicholas Evans & Marian Klamer (eds), Melanesian Languages on the Edge of Asia:

20

Antoinette Schapper

Challenges for the 21st Century. Language Documentation & Conservation Special Publication No. 5, 72–108. Klamer, Marian & Antoinette Schapper. 2012. The history of ‘give’ constructions in the Papuan languages of Timor-Alor-Pantar. Linguistic Discovery 10(3): 174–207. Klamer, Marian & Amos Sir. 2011. Kosakata Bahasa Teiwa-Indonesia-Inggris (Teiwa-IndonesianEnglish wordlist). Kupang, Indonesia: UBB-GMIT. Kratochvíl, František. 2007. A grammar of Abui: A Papuan language of Alor. Utrecht: LOT Publications. Kratochvíl, František & Benidiktus Delpada. 2008a. Kamus pengantar Bahasa Abui (AbuiIndonesian-English dictionary). Kupang, Indonesia: UBB-GMIT. Kratochvíl, František & Benidiktus Delpada. 2008b. Abui stories from Takalelang. Kupang, Indonesia: UBB-GMIT. Mandala, Halus. 2010. Evolusi fonologis bahasa Oirata dan kekerabatannya dengan bahasabahasa nonaustronesia di Timor Leste. PhD thesis, Universitas Udayana, Denpasar. Martis, Non, Wati Kurnatiawati, Buha Aritonang, Hidayatul Astar & Ferr Feirizal. 2000. Monografi kosakata dasar Swadesh di Kabupaten Alor. Jakarta: Pusat Bahasa, Departemen Pendidikan Nasional. Marques, Tito Lívio Nunes. 1990. Método prático para apreender o Makasae. Lisbon: selfpublished. McWilliam, Andrew R. 2007. Austronesians in linguistic disguise: Fataluku cultural fusion in East Timor. Journal of Southeast Asian Studies 38(2): 355–375. Nacher, Alfonso. 2003. Léxico Fataluco-Português: Primeira Parte. Studies in Languages and Cultures of East Timor 5: 138–196. Nacher, Alfonso. 2004. Léxico Fataluco-Português: Segunda Parte. Studies in Languages and Cultures of East Timor 6: 119–177. Nacher Alfonso. 2012. Léxico Fataluco-Português. Edited by Alberto Fidalgo Castro, Efrén Legaspi Bouza, Frederico Delgado Rosa & Geoffrey Hull. Dili: Salesianos de Dom Bosco. Nicolspeyer, Martha Margaretha. 1940. De sociale structuur van een Aloreesche bevolkingsgroep. Rijswijk: Kramers. Nitbani, Semuel H., Jeladu Kosmas, Sisila Wona & Hilda Naley. 2001. Struktur Bahasa Lamma. Jakarta: Pusat Bahasa, Departemen Pendidikan Nasional. O’Connor, Susan. 2003. Nine New Painted Rock Art Sites from East Timor in the Context of the Western Pacific Region. Asian Perspectives 42(1): 96–129. O’Connor, Susan. 2007. New evidence from East Timor contributes to our understanding of earliest modern human colonisation east of the Sunda Shelf. Antiquity 81(313): 523–535. Oliveira, Nuno V. 2008. Subsistence Archaeobotany: Food Production and the Agricultural Transition in East Timor. PhD thesis, The Australian National University, Canberra. Pawley, Andrew. 2005. The chequered career of the Trans New Guinea hypothesis: recent research and its implications. In Andrew Pawley, Robert Attenborough, Jack Golson & Robin Hide (eds), Papuan pasts: cultural, linguistic and biological histories of Papuanspeaking peoples, 67–108. Canberra: Pacific Linguistics. Pawley, Andrew & Harald Hammarström. forthcoming. The Trans New Guinea Family. In Bill Palmer (ed.), The Languages and Linguistics of New Guinea: A Comprehensive Guide. Berlin: Mouton de Gruyter. Pinto, Tomas Mariano. 2004. Makalero Disionariu. Dili: Timor Loro Sa’e-Nippon Culture Center. Pinto, Tomas Mariano. 2007. Ma’u Fi Makalero Lolo. Dili: Timor Loro Sa’e- Nippon Culture Center. Riedel, J.G.F. 1886. De sluik- en kroesharige rassen tusschen Selebes en Papua. ’s Gravenhage: M. Nijhoff.

Introduction to The Papuan languages of Timor, Alor and Pantar

21

Robinson, Laura C. & Gary Holton. 2010. Mapping the Alor-Pantar languages. Paper presented at the 6th International East Nusantara conference on language and culture, Kupang, Indonesia. August 6, 2010. Robinson, Laura C. & Gary Holton. 2013. Internal classification of the Alor-Pantar language family using computational methods applied to the lexicon. Language Dynamics and Change 2(2): 123–49. Ross, Malcolm. 2005. Pronouns as a preliminary diagnostic for grouping Papuan languages. In Andrew Pawley, Robert Attenborough, Robin Hide & Jack Golson (eds), Papuan pasts: cultural, linguistic and biological histories of Papuan-speaking peoples, 15–66. Canberra: Pacific Linguistics. Salzner, Richard. 1960. Sprachenatlas des Indopazifischen Raumes. 2 vols. Wiesbaden: Otto Harrassowitz. Sawardo, P., F. Sanda, H. Jehane, S.H. Nitbani & S. Kusharyanto. 1996. Fonologi, morfologi, dan sintaksis Bahasa Buna. Jakarta: Departemen Pendidikan dan Kebudayaan. Schapper, Antoinette. 2010. Bunaq, a Papuan language of central Timor. PhD thesis, The Australian National University, Canberra. Schapper, Antoinette. 2011a. Crossing the border. Historical and linguistic divides among the Bunaq in central Timor. Wacana, Journal of the Humanities of Indonesia 13(1): 29–49. Schapper, Antoinette. 2011b. Finding Bunaq; The homeland and expansion of the Bunaq in central Timor. In Andrew McWilliam & Elizabeth G. Traube (eds), Life and land in Timor; Ethnographic Papers, 163–186. Canberra: ANU E Press. Schapper, Antoinette. forthcoming. Elevational systems in Alor-Pantar languages. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Schapper, Antoinette & Marian Klamer. forthcoming. Numeral systems in Alor-Pantar languages. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Schapper, Antoinette & Marten Manimau. 2011. Kamus Pengantar Bahasa Kamang-IndonesiaInggris (Introductory Kamang – Indonesian – English Dictionary). Kupang, Indonesia: UBBGMIT. Schapper, Antoinette, Juliette Huber & Aone van Engelenhoven. forthcoming. The relatedness of Timor-Kisar and Alor-Pantar languages: A preliminary demonstration. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Schapper, Antoinette, Juliette Huber & Aone van Engelenhoven. 2012. The historical relations of the Papuan languages of Timor and Kisar. Language and Linguistics in Melanesia: Special Issue On the History, Contact & Classification of Papuan languages, 194–242. Steinhauer, Hein. 1977. ‘Going’ and ‘Coming’ in the Blagar of Dolap (Pura, Alor, Indonesia). NUSA, Miscellaneous Studies in Indonesian and Languages in Indonesia III, 38–48. Jakarta. Steinhauer, Hein. 1991. Demonstratives in the Blagar language of Dolap (Pura, Alor, Indonesia). In Tom Dutton (ed), Papers in Papuan linguistics, 177–221. Canberra: Pacific Linguistics. Steinhauer, Hein. 1993. Sisters and potential wives: where linguists and anthropologists meet: notes on kinship in Blagar (Alor). In P. Haenen (ed.), Vrienden en verwanten, Liber Amicorum Alex van der Leeden, Leiden/Jakarta, DSALCUL/IRIS: 147–168. Steinhauer, Hein. 1995. Two varieties of the Blagar language (Alor, Indonesia). In Connie Baak, Mary Bakker & Dick van der Meij (eds), Tales from a concave world; Liber amicorum Bert Voorhoeve, 269–296. Leiden: Projects Division, Department of Languages and Cultures of South-East Asia and Oceania.

22

Antoinette Schapper

Steinhauer, Hein. 1999. Bahasa Blagar Selayang Pandang. In Bambang Kaswanti Purwo (ed), Panorama bahasa Nusantara, 71–102. Jakarta: Universitas Cenderawasih and Summer Institute of Linguistics. Steinhauer, Hein. 2010. Pura when we were younger than today. In Artem Fedorchuk & Svetlana Chlenova (eds), Studia Anthropologica. A Festschrift in Honour of Michael Chlenov /Сборник статей вчести М.Ф.Членова, 261–283. Москва/ Иерусалим: Мосты Культуры/ Gesharim. Steinhauer, Hein & W.A.L. Stokhof. 1976. Linguistic research in the Alor Islands. Netherlands Foundation for the Advancement of Tropical Research; Report for the year 1976, The Hague: 27–31. Stoel, Ruben. 2007. The Prosody of Fataluku. Paper presented at the First Workshop on the Languages of Papua, Manokwari, Universitas Negeri Papua. Stokhof, W.A.L. 1975. Preliminary notes on the Alor and Pantar languages (East Indonesia). Canberra: Pacific Linguistics B-43. Stokhof, W.A.L. 1977. Woisika I: An ethnographic introduction. Canberra: Pacific Linguistics. Stokhof, W. A. L. 1978. Woisika text. In Miscellaneous Studies in Indonesian and Languages in Indonesia 5: 34–57. Jakarta: NUSA. Stokhof, W.A.L. 1979. Woisika II: Phonemics. Canberra: Pacific Linguistics. Stokhof, W.A.L. 1982. Woisika riddles. Canberra: Pacific Linguistics. Stokhof, W. A. L. 1983. Names and Naming in Ateita and environments (Woisika, Alor). Lingua 61(2/3): 179–207. Stokhof, W.A.L. 1984. Annotations to a text in the Abui language (Alor). Bijdragen tot de Taal-, Land- en Volkenkunde 140: 106–162. Stokhof, W.A.L. 1987. A short Kabola Text (Alor, East Indonesia). In Donald C. Laycock & Werner Winter (eds), A World of Language: Papers presented to Professor Stephen A. Wurm on his 65th Birthday, 631–648. Canberra: Pacific Linguistics. Sudiartha, I Wayan, Made Denes, I Wayan Tama & R.B. Suprihanto. 1994. Survei bahasa dan sastra di Timor Timur. Jakarta: Departemen Pendidikan dan Kebudayaan. Sudiartha, I Wayan, I Nengah Budiasa, Ni Luh Partami & Anak Agung Putra. 1998. Struktur bahasa Makasai. Jakarta: Departemen Pendidikan dan Kebudayaan. Valentim, Justino. 2001a. Lia Fataluk iha Lautem. Hakerek Kona Ba Timor Loro Sa’e 3. Valentim, Justino. 2001b. Konversasaun iha lia Fataluku. Hakerek Kona Ba Timor Loro Sa’e 4. Valentim, Justino. 2002. Fata-Lukunu i Disionariu. Dili: Timor Loro Sa’e- Nippon Culture Center. Valentim, Justino. 2004a. Waihohonu. Hakerek Kona Ba Timor Loro Sa’e 10. Valentim, Justino. 2004b. Família iha lian Fataluku. Hakerek Kona Ba Timor Loro Sa’e 11. Online: http://www.geocities.jp/hkbtls/hakerek/Hakerek11/fam_lia.htm Vatter, Ernst. 1932. Ata Kiwan; Unbekannte Bergvölker im tropischen Holland. Leipzig: Bibliographisches Institut. Watuseke, F. S. 1973. Gegevens over de taal van Pantar: Een Irian taal. Naschrift door J.C. Anceaux. Bijdragen tot de Taal-, Land- en Volkenkunde 129(2/3): 345–346. Wurm, Stephen A. & Shirô Hattori (ed.). 1982. Language atlas of the Pacific area. Part 1: New Guinea area, Oceania, Australia. Canberra: Pacific Linguistics C-66. Ximenes, Henrique de Oliveira & Elda Maria Guterres Menezes. 2002. Ginigni mu’a Timor watu misa gi ge’e. Hakerek Kona Ba Timor Loro Sa’e 6.

Gary Holton

2 Western Pantar 1

The language scene

25

2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.2 2.3 2.4 2.4.1 2.4.2 2.5 2.6 2.7 2.7.1 2.7.2 2.7.3 2.7.4

27 Phonology 27 Consonants 28 Plosives 28 The fricative /s/ 28 Nasal consonants 30 Liquids /r/ and /l/ 31 Approximants /j/ and /w/ 31 Glottal consonants /h/ and /ʔ/ 32 Geminate consonants 33 Phonotactics 35 Vowels 35 Diphthongs 36 Vowel sequences 36 Syllable structure 38 Stress 41 Morphophonemics 41 Reduction of vowel sequences 41 Non-phonemic gemination 42 Compounding 43 Reduplication

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7

43 Basic clausal syntax 44 Simple verbal clauses 45 Ditransitive constructions 46 Equational constructions 47 Existential constructions 47 Postpositional phrases 49 Adverbs 51 Negation

4 4.1 4.2 4.3 4.4

53 Noun phrases 54 Number 55 Adjectives 55 Quantification Numeral classifiers

56

24

Gary Holton

Demonstratives 57 58 Visibility 58 Specificity 59 Articles 60 Possession 60 Alienable possession 61 Inalienable possession Another inalienable possession construction 65 Genitive construction 68 Locational nouns

4.5 4.5.1 4.5.2 4.6 4.7 4.7.1 4.7.2 4.7.3 4.7.4 4.8

64

5 5.1 5.2 5.2.1 5.2.2 5.2.3

71 Pronouns and person-marking prefixes 74 Actor and undergoer independent pronouns 76 Person-marking prefixes 76 Verbs with obligatory person-marking prefixes 78 Verbs which do not admit person-marking prefixs Verbs which optionally occur with a person-marking prefix

6 6.1 6.1.1 6.1.2 6.1.3 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5

80 Serial verb constructions 81 Symmetrical SVCs 81 Complex events 81 Manner 82 Cause-effect 83 Asymmetrical SVCs 83 Directional SVCs 84 Aspect-marking SVCs 85 Auxiliary verbs 88 Dual constructions with -ilaku 89 Comitative constructions with -tu

7 7.1 7.2 7.2.1 7.2.2

89 Aspect-marking Aspectual particles Aspectual prefixes Inceptive prefix aProgressive prefix i-

8

Discussion

References

95

94

89 91 91 93

79

Western Pantar

25

1 The language scene Western Pantar (ISO 639-3 code: lev) is spoken by an estimated 10,804 people on the southwestern portion of Pantar, west of the Sirung volcanic massif. The dry region of western Pantar is separated from the northern peninsula of the island by a physical barrier consisting of a number of steep-walled canyons. No road yet connects the northern peninsula with the western part of the island. In most of the academic literature the language is referred to as Lamma, though that name more properly refers to only a single dialect of the language. This description is based on first-hand field work by the author between 2004 and 2010.1 There are three mutually-intelligible primary dialects – Lamma, Tubbe, and Mauta – spoken respectively in the western, central, and eastern parts of the region. Phonological differences between dialects are best described as tendencies, such as: the presence of final lateral in Lamma; diphthongization in Tubbe; and retention of final rhotic in Mauta. These phonological differences are exemplified in Table 1 below. Table 1: Phonological differences across Western Pantar dialects Lamma

Tubbe

Mauta

gloss

features

niba nekul nau ginaka yel

niar niaku niu ganiaka dia

nissar nebu niau ganeka(r) si

‘my father’ ‘my younger sibling’ ‘my mother’ ‘see (it)’ ‘go’

loss of -r in L -l in L; diphthong in T variation in diphthongs diphthong in T; -r in M -l in L; diphthong in T

In terms of lexicon the Lamma dialect is most distinctive; however, most speakers have knowledge of lexical synonyms across all three dialects. Table 2 compares lexical differences across the three dialects. 1 I am especially grateful to my principal collaborator Mr. Mahalalel Lamma Koly, in whose house I resided for much of the time I was on Pantar, and who assisted with much of the data collection and analysis. This chapter is much improved thanks to feedback from the editor, two anonymous reviewers, and my colleagues in the EuroBABEL project. Funding for exploratory field work in July 2004 was provided by US National Science Foundation grant #BCS-0408448. Funding for 2006–2007 field work was provided by a Documenting Endangered Languages fellowship from the US National Endowment for the Humanities. Funding for 2008 field work was supported by a field trip grant from the Hans Rausing Endangered Languages Documentation Project. Funding for field work in 2009 and 2010 was provided by the US National Science Foundation grant BCS-0936887, under the auspices of the European Science Foundation EuroBABEL program.

26

Gary Holton

Table 2: Lexical differences across Western Pantar dialects Tubbe

Mauta

Lamma

gloss

bis haweri kalla karani sai sussung tuang

bis haweri kalla karani hoba sussung tuang

salepi bana kisang sinnal hobi kutti talasing

‘mat’ ‘many’ ‘small’ ‘Canarium sp.’ ‘quiver’ ‘throw out’ ‘scrape’

There are few systematic phonetic differences in pronunciation, and most of these differences can be formulated in terms of tendencies which are more or less prominent in one dialect or another. For example, the in all three dialects the phoneme /j/ may occur as liquid [j] or palatalized as [t y] or affricated as [dʒ]. However, the affricate pronunciation is much more strongly associated with the Lamma dialect. The description in this sketch applies equally to all three major dialects; however, the examples and other data draw heavily from the Tubbe dialect. Where a particular feature or phenomenon is restricted to only one or two dialects, this is noted in the text. All examples in this sketch were recorded by the author in Pantar. Examples derived from continuous speech are notated with an identifying text name and line number in square brackets (e.g., [pubila078] refers to line 78 of the text “pubila”).2 Examples which are not so marked are based on elicitation or participant observation. In spite of pervasive multilingualism the use of Western Pantar remains vigorous within the domestic sphere. Children in the region grow up with Western Pantar as their first language, and immigrants gain at least passing fluency in the language. Malay is the preferred language in political, religious, and educational domains. The economy is dominated by subsistence agriculture, and opportunities for cash income are limited. Many residents seek temporary employment in the district capital of Kalabahi, providing an additional source of language contact.

2 Corpus data, including recordings and time-aligned transcriptions, are available from The Language Archive (http://tla.mpi.nl).

27

Western Pantar

2 Phonology Two aspects of Western Pantar phonology stand out as distinctive. First, most consonants exhibit a phonemic length contrast in medial position. While geminates are found in other AP languages, WP is unique in the extent to which it has phonemicized this distinction. Second, the glottal fricative is retained as a reflex of original *h in forms other than pronouns, though it is only weakly articulated. There is no evidence for phonemic pitch distinctions.

2.1 Consonants Western Pantar has a relatively small consonant inventory consisting of 16 consonants: voiceless plosives /p t k ʔ/ and voiced plosives /b d g/; voiceless fricatives /s h/; nasals /n m ŋ/; alveolar liquids /r/ and /l/; and glides /w/ and /j/. Table 3 lists WP consonants in IPA. Practical orthography, where different, is given in angle brackets. Table 3: Western Pantar consonant inventory

plosive

labial

alveolar

velar

glottal

p b

t d s n r l

k g

ʔ

fricatives nasals liquids

m

approximants

w

h ŋ

j

The glottal consonants are exceptional in their limited distribution. The glottal stop occurs only word-medially between vowels. The glottal fricative occurs only in onset position at the beginning of a morpheme. The remaining consonants have a much wider distribution, as discussed below. Voicing contrasts are present at each place of articulation for stops (excepting glottal), but there are no voiced fricatives. This inventory is typical for the languages of the Alor-Pantar group. However, the WP consonant inventory differs somewhat from that found in neighboring Nedebang and Teiwa, both of which have somewhat larger inventories. Nedebang contains both labial and velar fricatives /f/ and /x/, as well as an alveolar affricate. Teiwa includes both voiced and voiceless labial fricatives /v/ and /f/, as well as a uvular stop /q/ and a pharyngeal fricative /ħ/.

28

Gary Holton

2.1.1 Plosives The plosive inventory includes both voiceless and voiced plosives at the bilabial, alveolar, and velar places of articulation. (See below for discussion of the glottal stop.) The voiceless stops lack any significant aspiration. Voice onset time is close to zero for word-initial voiceless stops. In intervocalic position more aspiration may be present, though voice onset time is typically less than 15% of closure duration. Voiced stops are fully voiced throughout. Place of articulation and voicing contrasts in stops are indicated in the following examples. (1)

/pi/ /ti/ /ki/ /bi/ /di/ /gi/

‘our’ ‘tuak container’ ‘maggot’ ‘k.o. drum’ ‘rat’ ‘their (3PL)’

With only a few exceptions, plosives are limited to initial and medial positions (see §2.3 below).

2.1.2 The fricative /s/ The fricative /s/ is a voiceless alveolar fricative and is the only true fricative in the WP consonant inventory. The fricative /s/ occurs in all positions. It may also occur geminated. (2)

/si/ /losi/ /hisnakkuŋ/ /bis/ /gissi/

[si] [losi] [ʰisnakːuŋ] [bis] [gisːi]

ART

‘cucumber’ ‘six’ ‘mat’ ‘bite’

The fricative /h/ is discussed below under glottal consonants.

2.1.3 Nasal consonants In initial and medial positions /m/ and /n/ contrast.

Western Pantar

(3)

29

Contrast between /m/ and /n/ /ma/ ‘come’ /na/ ‘eat’ /tame/ ‘where’ /tane/ ‘face’

The bilabial nasal /m/ is limited to word-initial and medial positions except when it occurs as the enclitic allomorph of the locative postposition /me/. The enclitic allomorph occurs following a word which ends in a vowel (4a); or a word which ends in a nasal (4b), in which case it replaces the nasal. (4)

a.

/bla me/

[bla me ~ blam]

‘at the house’

b.

/habbaŋ me/

[habːaŋ me ~ hab:am]

‘at the village’

The alveolar nasal /n/ is limited to word-initial and medial positions with two identified lexical exceptions, as follows.

(5)

Word-final /n/ /kan/ ‘also’ /an/ RELATIVIZER

The alveolar nasal /n/ occurs in a small number of words of foreign origin in medial coda position.

(6)

Medial coda /n/ /kunda/ ‘shirt’ /hundar/ ‘afraid’ /undi/ ‘stalk’ /sanlak/ ‘ten thousand’

With the exception of the two words in (5) the velar nasal /ŋ/ () is the only nasal which can occur in final position. Conversely, with a very few exceptions, the velar nasal is limited to final position. An exhaustive list of known exceptions where /ŋ/ occurs in non-final position is given in (7). Some of the forms may be historically analyzed as compounds and thus originally word-final. (7)

/giŋtamme/ /iŋbina/ /taŋge/ /siŋlap/ /biŋge/

‘woman’s female cross-cousin’ ‘seed’ ‘middle’ ‘rock lobster’ ‘nautilus’

30

Gary Holton

The bilabial and alveolar nasals can also occur geminated in medial position. (8)

/bunna/ /ammu/

‘smoke’ ‘boy’

In contrast, the velar nasal does not occur as a geminate. Given their near complementary distribution, it may be preferable to treat the velar nasal as an allophone of the alveolar nasal. For expository purposes /ŋ/ is treated here as a distinct phoneme.

2.1.4 Liquids /r/ and /l/ The liquids /r/ and /l/ occur in initial, medial, and final positions. In initial position the rhotic trill /r/ is often preceded by a prothetic short low vowel. In non-initial position /r/ is pronounced as an alveolar flap [ɾ]. The following example illustrates the phonemic and phonetic contrast between initial /r/, initial vowel, and initial /h/. (9)

/ra/ /ara/ /hara/

[ara] [ʔaɾa] [haɾa]

‘fire’ ‘large’ ‘ember’

The phonemes /r/ and /l/ contrast phonemically in all positions.

(10)

Contrast between /r/ /ra/ ‘fire’ /ara/ ‘large’ /tor/ ‘main road’

and /l/ /la/ ‘caldron’ /ala/ ‘interior’, ‘forest’ /dol/ ‘bench’

Word-final /l/ is more commonly preserved as such in the Lamma dialect than in the other dialects.

(11)

Correspondence of Lamma -l and Tubbe -Ø Lamma Tubbe /jabbal/ /jabbe/ ‘dog’ /tubbal/ /tubbe/ ‘name of a clan’ /jel/ /dia/ ‘go’ /taˈmal/ /tamˈe/ ‘tamarind’ /pinnil/ /pinni/ ‘hold’ /maddal/ /madde/ ‘necklace’ /nekul/ /niaku/ ‘my younger sibling’

Western Pantar

31

2.1.5 Approximants /j/ and /w/ The consonant /j/ (orthographic ) is a voiced palatal approximant [j], though it may sometimes be slightly to heavily affricated as [dʒ], especially in the Mauta dialect. It may occur in word-initial or word-medial position, but always as a syllable onset. (12)

/ja/ /weja/

[ja], [dʒa] [weja], [wedʒa]

‘road’ ‘door’

The consonant /w/ is a voiced labio-velar approximant [w]. Like /j/, it may occur in word-initial or word-medial position, but always as a syllable onset. (13)

/wa/ /tawa/

‘go’ ‘ocean’

The approximants /j/ and /w/ contrast in all positions in which they occur. The approximants are distinguished from the high vowels /i/ and /u/ primarily on phonotactic evidence. Vowels but not approximants may occur in coda position. Thus, /ai/ ‘lice’ and /au/ ‘grasshopper’ are permissible but *ay and *aw are not. However, vowels and approximants can be distinguished phonemically as well. The near-minimal pair in (14) distinguishes /i/ from /j/. (14)

/iaka/ /jakal/

‘youngest child’ ‘to jerk’

These phonemes can be further distinguished via dialect correspondences. The vowel sequence /ia/ in the Tubbe dialect corresponds regularly to a mid-front vowel /e/ in the Mauta dialect. In contrast, the phoneme /j/ is realized in Mauta as an affricate. Thus, /iaka/ corresponds to [eka] in Mauta, but never *[dʒaka].

2.1.6 Glottal consonants /h/ and /ʔ/ The glottal fricative /h/ occurs only in morpheme-initial position preceding a vowel. Vowels preceded by /h/ are sometimes referred to as “weak” vowels (see below), as the phoneme /h/ is very lightly articulated, represented here phonetically as a raised h [ʰ]. Word-initial vowels which are not preceded by /h/ are pronounced with a preceding glottal stop. This gives rise to a phonemic contrast between /h/ and Ø (zero) in word-initial position which is realized phonetically as a contrast between Ø ~ [ʰ] and [ʔ], respectively.

32

(15)

Gary Holton

Contrast between /h/ /haŋ/ [ʰaŋ] /hai/ [ʰai] /har/ [ʰar] /habbaŋ/ [ʰabːaŋ] /hu/ [ʰu]

and /Ø/ ‘you’ ‘boat’ ‘for you’ ‘village’ ‘cobra’

/aŋ/ /ai/ /ar/ /abbaŋ/ /u/

[ʔaŋ] [ʔai] [ʔar] [ʔabːaŋ] [ʔu]

‘market’ ‘clothing louse’ ‘root’ ‘shake’ ‘k.o. banana’

In casual speech this contrast is often neutralized and the glottal stop insertion rule is blocked, resulting in apparent homonyms. When the glottal fricative occurs word-medially preceded by a prefix it is always retained as such rather than being deleted. Thus, the morpheme /hapaŋ/ ‘near’, when prefixed with the third person singular person-marking prefix /ga-/ gives rise to [gaʰapaŋ] ‘near her/him’, not *[gaapaŋ] or *[gaːpaŋ]. In addition to the phonetic glottal stop which is inserted preceding wordinitial vowels, some bound morphemes contain initial phonemic glottal stop which may contrast with zero. Thus compare the behavior of the following root forms. The first contains a glottal stop when prefixed, whereas, the second does not. (16)

/-ʔuŋ/ /-ule/

‘head’ ‘neck’

/na-ʔuŋ/ /na-ule/

[naʔuŋ] [naule]

‘my head’ ‘my neck’

Phonemic glottal stop also occurs in word-medial position between two vowels; see the description of vowel sequences below.

2.2 Geminate consonants Consonants /p t k b d g s m n l/ contrast in length with longer geminate counterparts /pp tt kk bb dd gg ss mm nn ll/. (17)

Contrast /sipe/ /gataŋ/ /dala/ /dake/ /duba/ /idia/ /taga/ /hasi/ /gima/ /hina/

between short and long consonants ‘kingpost’ /sippe/ ‘flea’ ‘on it’ /gattaŋ/ ‘his hand’ ‘ripe’ /dalla/ ‘tomorrow’ ‘now’ /dakke/ ‘dry pandanus’ ‘slippery’ /dubba/ ‘to push’ ‘they go’ /iddia/ ‘have gone’ ‘which’ /tagga/ ‘incision’ ‘roof thatch’ /hassi/ ‘bite you’ ‘they come’ /gimma/ ‘their smell’ ‘you (pl.) eat’ /hinna/ ‘to die’

33

Western Pantar

Voiceless geminate stops are typically more than 50% longer in duration than their non-geminate counterparts, while voiced geminates may be only slightly longer than their voiced non-geminate counterparts. Gemination is also an active morpho-phonological process, discussed in §2.7.2 below.

2.3 Phonotactics Table 4 describes the distribution of consonants across syllables. Parentheses indicate consonants which occur only in loanwords. Table 4: Distribution of consonants onset

p t k ʔ b d g s h m n ŋ r l w j

coda

word-initial

word-medial

word-medial

word-final

✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ – ✓ ✓ ✓ ✓

✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ – ✓ ✓ ✓ ✓

(✓) (✓) (✓) – (✓) (✓) (✓) ✓ – – – ✓ ✓ ✓ – –

✓ – (✓) – (✓) – – ✓ – (✓) (✓) ✓ ✓ ✓ – –

Word-final plosives are extremely rare and are restricted largely to borrowings, onomatopoetic forms, and contractions. Borrowings include /hap/ ‘fish’ (perhaps from neighboring Nedebang; compare the native form /keʔe/) and /sanlak/ ‘ten thousand’. Contractions include the enclitic allomorph /=b/ of the relativizing particle /ba/, as in /hinani=b ga-niaka/ (what=REL 3SG-see) ‘what is it that you saw?’.

34

Gary Holton

Distribution of plosives (18) word-initial /pai/ ‘slice’ /ti/ ‘tuak container’ /ki/ ‘maggot’ /bi/ ‘k.o. drum’ /dake/ ‘now’ /gai/ ‘his/her’

word-medial /kapi/ ‘smashed’ /pati/ ‘pay’ /doki/ ‘vaccinate’ /sabi/ ‘pumice’ /wedi/ ‘decant’ /hoga/ ‘untruth’

word-final /hap/

‘fish’

/sanlak/ ‘ten thousand’ /bub~bub/ ‘k.o. quail’

Other forms with word-final plosives result from recent apocope, as evidenced by the existence of corresponding vowel-final forms in other dialects. For example, Tubbe /siŋlap/ ‘rock lobster’ corresponds to Lamma /silapu/. Word-medial coda plosives occur only as part of geminates. Word-final nasals are limited to the velar place of articulation, with just a few exceptions as noted above. A handful of lexical items contain a final alveolar nasal. Final labial nasals do not occur in basic lexical items but result from phonological reduction of the postposition /me/ to the enclitic for [=m]. On the other hand, the velar nasal is limited to coda position. The frequency of word-medial liquids in coda position is quite low, and in some cases the words which have liquids in coda position might be better analyzed as compounds, as suggested in (19).

(19)

Word-medial /wirwiaŋ/ /hulbaraŋ/ /ˈjenarpaŋ/

liquids in coda position ‘barred dove (Geopelia maugei)’ (cf. /wiri/ ‘back’) ‘cardinal fish (Apogonidae)’ (cf. /hule/ ‘neck’) ‘day before yesterday’ (cf. /jer/ ‘still’, /wadːipaŋ /‘yesterday’)

The glottal consonants /h/ and /ʔ/, as well as the glides /w/ and /j/ occur only in onset positions. The remainder of the consonants may occur freely in all positions. In particular, WP freely permits word-final consonants. Consonant sequences occur only in medial position across syllable boundaries. Apparent word-initial clusters may arise in casual speech when an obstruent is followed by a liquid /r/ or /l/. In careful speech such forms are always pronounced with an intervening unstressed low central vowel /a/.

(20)

Apparent word-initial clusters with liquids /bla/ [bˈla], [baˈla] ‘house’ /gliŋ/ [gˈliŋ], [gaˈliŋ] ‘its edge’ /bro/ [bˈro], [baˈro] ‘flour, dust’ /griaŋ/ [gˈriaŋ], [gaˈriaŋ] ‘to care for’ /srau/ [sˈrau], [saˈrau] ‘that.HIGH’

Western Pantar

35

2.4 Vowels The vowel inventory is a fairly typical five-vowel system. A minimal set contrasting the vowel phonemes is given below.

(21)

Contrast between vowel phonemes /bi/ [bi] ‘k.o. drum’ /be/ [be] ‘shadow’ /ba/ [ba] RELATIVIZER /bo/ [bo] ‘ocean wave’ /bu/ [bu] ‘areca nut’

The structure of the vowel system differs somewhat across the dialects, especially with regard to the realization of diphthongs. The description here is based on the Tubbe dialect. Most vowels exhibit unconditioned phonetic height variation. The vowel /e/ varies [e] ~ [ɛ], apparently without contrast. (22)

/bes/ /he/

[bɛs] [ʰe]

‘papaya’ ‘or’

The vowel /a/ is generally expressed as a low central vowel [a] but may occur as [ə] or [ɐ]. (23)

/was/ /ara/ /halia/

[wəs], [wɐs] [ʔaɾa] [ʰaliɐ]

‘day’ ‘large’ ‘water’

Preceding a high back vowel, /a/ occurs as [o], as in /gaume/ [goume] ‘inside’. Preceding a high front vowel, /a/ may be raised to [e], as in /gaiŋ/ [geiŋ] 3SG.UND. Preceding /r/ the vowel /a/ may occur as schwa [ə], as in /war/ [wər] ‘burn’.

2.4.1 Diphthongs Diphthongs are possible between the vowels /a/ and /u/, the vowels /a/ and /i/, and the vowels /e/ and /u/. The phonetic diphthongs [ei] and [oa] are here analyzed as an allophones of /ai/ and /ua/, respectively (see above).

36

(24)

Gary Holton

Diphthongs /au/ /lau/ /ua/ /kuaŋ/ /ai/ /bai/ /ia/ /pia/ /eu/ /eu/ /ue/ /kue/

[la͡u] [ku͡aŋ] [ba͡ i] ͡ [pia] [ʔe͡u] [ku͡e]

‘to bark’ ‘moko drum’ ‘pig’ ‘descend’ ‘woman’ ‘fish spear’

2.4.2 Vowel sequences Diphthongs can be distinguished from sequences of vowels separated by a glottal stop. These vowel sequences are here analyzed as forming distinct syllable nuclei. Thus, the vowel sequences are disyllabic.

(25)

Vowel sequences /laʔu/ [la.ʔu] /muʔaŋ/ [mu.ʔaŋ] /baʔai/ [ba.ʔa͡ i] /siʔa/ [si.ʔa]

‘crawl’ ‘under’ ‘to drink’ ‘finger’

While the intervening glottal stop is only weakly articulated, spectrographic analysis reveals two distinct vowel articulations in the forms in (25). In addition to these vowel sequences which are broken up by a glottal stop, there are also examples of sequences of identical vowels which are not broken up by a glottal stop. These arise as a result of de-gemination of the following consonant. For example, the form gaata ‘already’ is a dialect variant of gatta, in which phonetic vowel length preserves the phonemic consonant length.

2.5 Syllable structure The minimal Western Pantar syllable is (C)V. Phonemically vowel-initial words are pronounced with an initial phonetic glottal stop (see above).

(26)

Mono-syllabic V /u/ CV /bu/ CVV /pia/ VC /ur/ CVC /dol/ CCV /bro/ CCVV /brai/

words ‘k.o. plant, pisang air’ ‘areca nut’ ‘slice’ ‘blowpipe’ ‘bench’ ‘flour’ ‘ringworm’

Western Pantar

(27)

(28)

Di-syllabic words V.CV /a.la/ CV.CV /da.la/ CVV.CVC /mai.jaŋ/ CVC.CVC /mid.daŋ/ CV.CVV /ba.ʔai/

37

‘interior’ ‘ripe’ ‘place’ ‘return from below’ ‘drink’

Tri-syllabic words V.CV.CV /e.no.ni/ CV.CV.CV /bu.wa.re/ CV.CVC.CVC /ga.tag.gaŋ/

‘k.o. moth’ ‘few’ ‘area above it’

Consonant clusters within syllables are found only in syllable onsets. There are no complex codas. Across syllable boundaries consonant clusters (sequences) are extremely limited. Other than geminate consonants only the following types of sequences are found. These are either clusters in which the first member is /r/ or clusters of homorganic consonants.

(29)

(30)

Clusters within native vocabulary /tim.ber/ ‘think over’ /his.nak.kuŋ/ ‘six’ Clusters within non-native vocabulary /kar.bau/ ‘buffalo’ /kan.de.ra/ ‘chair’ /pren.ta/ ‘order’

Most apparent initial consonant clusters are actually broken up by an unstressed vowel in careful speech. These have been here analyzed as CVC sequences, as below. (31)

/baˈla/ /baˈwa/ /baˈka/

[bˈla] ~ [bɐ ˈla] [bˈwa] ~ [bɐ ˈwa] [bˈka] ~ [bɐ ˈka]

‘house’ ‘large’ ‘maybe’

This approach has the admitted disadvantage of requiring the overt specification of lexical stress (in this case ultimate) in the above examples. An approach grounded in epenthesis might thus be better, and this approach is indeed taken in the practical orthography (e.g., /baˈla/ ‘house’ is represented orthographically bla).

38

Gary Holton

Clusters with /r/ are true clusters and are not broken up by epenthetic vowels even in careful speech. (32)

/bro/ /kro/ /brai/

[bro] [kro] [brai]

‘flour’ ‘snot’ ‘ringworm’

2.6 Stress Phonetic correlates of stress include increased intensity, higher pitch, and increased duration. In words without geminate consonants stress falls on the penultimate syllable.

(33)

Penultimate /kabi/ /baʔai/ /betalaku/ /tawaka/

stress [ˈka.bi] [ˈba.ʔai] [be.ta.ˈla.ku] [ta.ˈwa.ka]

‘chief’s house’ ‘drink’ ‘seven’ ‘hot’

Obligatory affixes are within the domain of stress assignment and may be stressed. Thus, the prefixes on the monosyllabic bound roots in (34) receive stress.

(34)

Stressed affixes /galu/ [ˈga.lu] /gane/ [ˈga.ne]

‘expel him’ ‘his face’

Optional affixes attached to unbound roots are outside the domain of stress assignment and do not receive stress.

(35)

Unstressed affixes /taŋ/ [ˈtaŋ] /gataŋ/ [ga.ˈtaŋ]

‘on’ ‘on it’

Clitics are also outside the domain of stress assignment and do not receive stress. Thus, in (36) the proclitic gai does not receive stress, and the enclitic me does not count in the calculation of the penultimate syllable.

Western Pantar

(36)

39

Clitics outside domain of stress assignment /gai=be/ [gai.ˈbe] ‘his shadow’ /kebaŋ=me/ [ˈke.baŋ.me] ‘on the barn’

When the penultimate syllable is closed by a geminate consonant, the penultimate stress pattern is more difficult to hear. The long medial consonant adds prominence to the following syllable, effectively distributing stress evenly across the penultimate and ultimate syllables. This is sometimes interpreted by speakers as ultimate stress but is better described as a lack of penultimate stress, as in (37).

(37)

Even stress with geminate /balla/ [balːa] ‘friend’ /takkaŋ/ [takːaŋ] ‘hit’

Some words which contain a fossilized prothetic vowel have irregular stress patterns. For example, the first four numerals begin with /a/. (38)

/anuku/ /alaku/ /atiga/ /atu/

[a.ˈnu.ku] [a.ˈla.ku] [a.ˈti.ga] [a.ˈtu]

‘one’ ‘two’ ‘three’ ‘four’

For the purposes of stress assignment the initial vowel in these forms is treated as an optional affix and thus outside the domain of stress assignment, thus explaining the irregular ultimate stress on /atu/ ‘four’. Ultimate stress can also be found on some words in which the penultimate syllable is not closed by a geminate consonant and does not consist of a prefix. Some examples are given below.

(39)

Ultimate stress without geminate /daˈgeŋ/ [da.ˈgeŋ] ‘wind’ /kaˈre/ [ka.ˈre] ‘red ant’ /taˈwa/ [ta.ˈwa] ‘ocean’ /maˈta/ [ma.ˈta] ‘k.o. sulphurous creek’ /taˈme/ [ta.ˈme] ‘tamarind’

It is not accidental that the first syllable of each form in (39) contains the vowel /a/. The ultimate stress pattern is always found in words whose penultimate syllable contains the vowel /a/. This suggests that the vowel /a/ may be non-phonemic in these forms, inserted epenthetically in order to break up pho-

40

Gary Holton

nemic consonant clusters. Under this analysis these forms consist underlyingly of initial clusters, e.g., /kre/ ‘red ant’ or /bla/ ‘house’. However, as noted in the preceding section, there exist at least some phonemic clusters which are not broken up via an epenthetic vowel. We can even find contrasts between forms with a phonemic cluster, and forms with unstressed vowel breaking up a cluster.

(40)

Contrast between clusters and non-clusters a. /kro/ [kro] ‘snot’ b.

/kaˈro/

[ka.ˈro]

‘k.o. beach plant’

The two examples in (40) cannot have the same underlying form, or else we would have no way to prevent the rule of cluster simplification from applying to (40a). Hence, (40b) must be analyzed as having phonemic stress on the ultimate syllable. I have not found minimal pairs for the other forms in (39), but speaker intuition also suggests that the cluster analysis is not valid. Even words such as ‘house’ are frequently written by speakers as disyllables. Nevertheless, the uniformity of the vowel quality in the unstressed syllable remains suspicious. In fact, not only is the vowel uniform throughout these forms with ultimate stress, but the vowel is also uniformly reduced, occurring as a slightly raised lower mid vowel [ɐ] (though clearly lower than schwa). It is possible that the cluster analysis is at least historically valid. That is, at one point WP may have been more tolerant of initial phonemic clusters such as CCV. These clusters then came to be phonetically broken up via epenthesis of a low mid vowel CaCV, with stress remaining on the phonemic vowel. These forms have been reanalyzed in the modern language as simple CVCV sequences with final stress. A very few minimal pairs distinguished by stress can be identified (41). Irregular final stress is lexicalised and so given in the phonemic representation of items that have it.

(41)

Minimal /mata/ /tame/ /dama/ /tauwe/ /ala/

pairs for stress [ˈmata] ‘luck, fortune’ [ˈtame] ‘where’ [ˈdama] ‘share’ [ˈtauwe] ‘our ears’ [ˈala] ‘male animal’

/maˈta/ /taˈme/ /daˈma/ /tauˈwe/ /aˈla/

‘k.o. sulphurous creek’ ‘tamarind’ ‘jicama’ ‘snail’ ‘interior’

In the practical orthography unpredictable final stress may be optionally indicated with an acute accent over the stressed vowel, thus, /taˈme/ ‘tamarind’ is represented orthographically as tamé.

Western Pantar

41

2.7 Morphophonemics 2.7.1 Reduction of vowel sequences With bound stems beginning with a vowel other than /u/, the vowel of the prefix is elided. So, for example, the noun /‑eba/ ‘nape of neck’ is obligatorily possessed and must occur with an inalienable possessive prefix, as in (42). Deletion of the prefix vowel results in homophonous singular and plural forms, which can then be optionally distinguished via the use of an alienable possessive proclitic /gai=/ 3SG.POSS or /gi=/ 3PL.POSS.

(42)

Deletion of prefix vowel in /‑eba/ ‘nape’ /ga-eba/ [geba] ‘his nape’ /gi-eba/ [geba] ‘their napes’

Bound stems which begin with a high back vowel exhibit less predictable behavior. In some cases the prefix vowel is retained, resulting in a diphthong, as with /‑ule/ ‘neck’ in (43).

(43)

Retention of prefix vowel in /‑ule/ ‘neck’ /ga-ule/ [gaule] ‘his neck’ /gi-ule/ [giule] ‘their necks’

In other cases the prefix low vowel is retained, while the high prefix vowel is deleted, as with /-ume/ ‘liver’ in (44).

(44)

Deletion of prefix vowel in plural form of /‑ume/ ‘liver’ /ga-ume/ [gaume] ‘his liver’ /gi-ume/ [gume] ‘their livers’

2.7.2 Non-phonemic gemination The progressive prefix /i‑/ triggers gemination of stem-initial consonants when it occurs immediately preceding a short verb stem. A short verb stem is one which consists of a single open syllable. This process is at least in part morphologically conditioned, as not all /i/ preceding a short verb stem trigger such gemination. For example, the vowel of the first person exclusive plural /ni‑/ in (45) does not trigger gemination. However, the vowel of the progressive prefix in (46) does trigger gemination. Here the low vowel of the first-person singular prefix is deleted (see §2.7.1 above), with the result that these two examples become minimal pairs distinguished via gemination of the stem-initial consonant.

42

Gary Holton

(45)

[hoaŋ me nipia] /hoaŋ me ni-pia/ beach LOC 1EXCL-descend ‘We are descending to the beach.’

(46)

[hoaŋ me nipːia] /hoaŋ me na-i-pia/ beach LOC 1SG-PROG-descend ‘I have descended to the beach.’

2.7.3 Compounding Nominal compounding is pervasive throughout the lexicon, and the use of compounds and other types of parallelisms is one of the most salient features of the language.

(47)

(48)

Nominal compounds (N N) /diʔiŋ hauwe/ firestarter rock /eu ittar/ woman basket /ga-uta hagaŋ/ 3SG-foot coconut.shell.interior

‘coal’ ‘basket for fruit’ ‘his/her ankle’

Adjectival compounds (N Adj) /kusi miaka/ nail white ‘cuticle’ /bla kalla/ house little ‘kitchen’ /kamau ara/ cat big ‘thumb’ Possessive compounds are formed with the third-person prefix /ga-/.

(49)

Possessed nominal compounds (N ga-N) /ala ga-ume/ grass 3SG-inside ‘interior (geographic)’ /bla ga-ʔuŋ/ house 3SG-head ‘ridgepole’ In noun-verb compounds the verb follows the noun.

(50)

Noun-verb compounds (N V) /was bauluŋ/ sun fall ‘sunset’ /halia salili/ water flow ‘river’ /mua misiŋ/ earth sit ‘centerbeam of a house’

43

Western Pantar

Verbal compounds are much less common, although some serial verb constructions discussed in §6 might be equally analyzed as compounds. Some examples of verb-verb compounds are given in (51).

(51)

Verb-verb compounds (V V) /lelu sajaŋ/ give.tribute imprecise /ga-llaŋ ga-wanuŋ/ 3SG-search 3SG-focus.on

‘wander without purpose’ ‘head toward an objective’

2.7.4 Reduplication Reduplication may be used with certain motion verbs to indicate continuing aspect, as with lama ‘walk’ in (52). However, this usage is not productive and is limited to a very few lexical items. (52)

Wenang ye lama~lama wa yattu ara haila ye go tree large CLF one man one RDP~walk ‘A man is walking along and without seeing a large tree,’

natar stand

ga-niaka kauwa wang raung pia. NEG exist climb descend 3SG-see ‘bumps into it and keeps going.’ [C16] Full reduplication may also be used to derive certain nominal compounds, as exemplified in (53). As with the motion verb reduplication this usage is not productive. Moreover, the semantic relationship between the original noun and the reduplicated noun is opaque.

(53)

Reduplicated nouns /baliŋ/ ‘axe’ /diŋ/ ‘sea urchin’ /kati/ ‘flea’ /kis/ ‘twig’

/baliŋ~baliŋ/ /diŋ~diŋ/ /kati~kati/ /kis~kis/

‘dragonfly’ ‘earthworm’ ‘corn with few kernels’ ‘k.o. termite’

3 Basic clausal syntax This section provides the important background information about clause structure including basic clausal word order, the expression of negation, presence/ absence of a copular, adpositional marking of NPs etc.

44

Gary Holton

3.1 Simple verbal clauses Word order in isolated (elicited) simple verbal clauses is SOV (or SV/APV), as in the following examples. Aspectual and discourse connective particles may follow the verb within the same intonation contour. (54)

Nang mising. 1SG.ACT sit ‘I am sitting.’

(55)

Gaing pali. 3SG.UND afraid ‘S/he is afraid.’

(56)

Inamme

gang gaing mukkung. 3SG.ACT 3SG.UND swallow ‘There it (the whale) swallowed him.’ [bm009]

DEM

(57)

Gang yattu walli ga. trim PFV 3SG.ACT tree ‘S/he cut down the tree.’

Core arguments may be highlighted using a left dislocation structure which results in an alternate PAV word order, as in (58) and (59). (58)

(59)

Duang ning lu’ung gaterannang kanna snake 1EXCL.ACT cut all.of.them already ‘The snake we cut up completely.’ [tonu.119] Gai ke’e maru si aname ging 3SG.POSS fish PL ART person 3PL.ACT ‘People took those fish of ours.’ [tonuburi073]

ga

ya.

PFV

CONJ

haggi take

kanna. already

Core arguments may also occur post-verbally in a right dislocation structure under a single intonation contour. (60)

I-ga-niaka allang duang. snake 4PL-3SG‑see thus ‘They saw a snake.’ [tonu.110]

Western Pantar

45

In natural discourse it is rare for two core arguments to be expressed as full nouns or independent pronouns. Elision of one or more arguments is common. (61)

Nang ga-niaka. 1SG.ACT 3SG-see ‘I saw him.’

(62)

Naing na-niaka. 1SG.UND 1SG-see ‘[Someone] saw me.’

(63)

Ha-ttang lu’ung hala. FOC 2SG-hand cut ‘Watch out that your hand doesn’t get cut.’

Core nominal arguments are often omitted altogether when reference is clear from the wider discourse context. (64)

Lama ga. walk PFV ‘Let’s go.’

(65)

Aukung ma hos wang a-diakang. exit come outside exist INCP-descend ‘[They] started to come out going down.’ [tonu.110]

3.2 Ditransitive constructions Western Pantar has a single ditransitive construction, formed with the intransitive verb ma ‘come’ and the transitive verb ‑nia ‘give’, inflected with a person-marking prefix indexing the recipient. This construction might be equally analyzed as a serial verb construction (see §6 below); however, it is highly lexicalized, uttered under a single intonation contour. (66)

Nang maggi ma ga-nia. banana come 3SG-give 1SG ‘I fetched him a banana.’

46

Gary Holton

The semantics of this construction emphasizes the path of movement of theme to recipient; example (66) implies that the banana is not immediately at hand but must be retrieved from another location. Hence, the gloss ‘fetch’. A simple ‘give’ construction in (67) does not require the verb ma. (67)

Maggi hissa na-nia. banana CLF:fruit 1SG-give ‘Give me a banana.’

More typically, ditransitive semantics are expressed with serial verb constructions using the verb haggi ‘take’ to mark the theme. (68) Bi haggi gi-sussung i-ti’ang ha ging pi allang. drum take 3PL-throw.out PROG-stay DISC 3PL.ACT descend then ‘(They) took the drum and gave it to them, then they went down.’ [soli041]

3.3 Equational constructions Equational constructions express either the equivalence of a single nominal argument with an entity expressed by a nominal predicate or the inclusion of an argument in a class of entities. The nominal argument precedes the nominal predicate. There is no copula and no inflection for person. (69)

[Nang]NP1 [wenang Amos old.man A. 1SG.ACT ‘I am Mr. Amos’ (child).’

ga’ai]NP2 . 3SG.GEN

(70)

[H-iu saina]NP1 [tabba hoga birang most untruth speak 2SG-mother DIST.VIS ‘Your mother is the biggest liar.’ [tonu045]

wala]NP2 . person

The predicate role in an equational construction can be filled by a demonstrative pronoun. (71)

Taga ba saiga? which REL PROX.VIS ‘Who is this?’

47

Western Pantar

3.4 Existential constructions Existential constructions are formed with wang. This form does not admit personmarking prefixes. (72)

Halia wang? water exist ‘Is there any water?’

(73)

Ning ga-niaka ga masi huppa PFV sulphur.creek also 1INCL.ACT 3SG-see ‘We saw that there is also a sulfur creek there.’

smaume LEVEL .DEF

wang. exist

The existential can be negated with the negative kauwa. (74)

Habbang me si listrik wang village LOC DEM electricity exist ‘There is no electricity in this village.’

kauwa. NEG

Existential wang often predicates existence of an unreferenced third person argument. (75)

Mea galling wang. table edge exist ‘[It is] at the edge of the table.’

(76)

Wang he kauwa? exist or NEG ‘Is it [there] or not?’

3.5 Postpositional phrases The category of postposition is not clearly delineated in Western Pantar. Bound postpositions are similar to obligatorily possessed nouns in that they must be obligatorily inflected with a person-marking prefix. (77)

[Bla ga-ne] wang. house 3SG-front exist ‘It’s in front of the house.’

48 (78)

Gary Holton

Nia marung haggi aukung [bla ga-ume] maiyang. take enter house 3SG-inside put things PL ‘Take these things and put them inside the house.’ [3.87]

Unbound postpositions differ in that they do not require or admit a personmarking prefix. Thus, the unbound postposition gegung ‘under’ in (79) cannot occur with a prefix ga‑. (79)

Tiaku [mea gegung] migang. glass table under place ‘Set the glass under the table.’ *tiaku [mea ga-gegung] migang

Unbound postpositions also share many properties with verbs. In particular, they may be used predicatively, as in (80) and (81). (80)

Wa allang gob hauwe butang ye tang. DISC rock tall one on go then ‘Going and then (being) on top of a tall rock.’

(81)

Biring diggang ga-wenung run take.shelter 3SG-return ‘Run back and take shelter.’

me. LOC

The verby nature of unbound postpositions makes possible an alternative analysis of constructions such as (81) as serial verb constructions (see §6). Complex postpositions may be built up from combinations of both bound and unbound postpositions. In this case the bound postposition precedes the unbound postposition, and the semantics are combinatorial. (82)

Soru lamura [hauwe ga-ling tang] maiyang. place cane straight rock 3SG-surface.edge on ‘A straight cane place on the edge of a rock.’ [pspv035]

(83)

[Kapa ge’e waddang] i-ti’ang. PROG-stand river 3SG-container.edge along ‘(It) was standing along the edge of the river.’

In certain cases postpositional phrases may be discontinuous. In (84) the third-person plural pronoun intervenes between the noun phrase tawa si ‘the ocean’ and the complex postposition me tang.

Western Pantar

(84)

49

Tawa si ging me tang lama kauwa. walk NEG ocean ARTAP 3PL.ACT LOC on ‘They could not walk on/in the water.’ [pubila431]

3.6 Adverbs Adverbs are distinguished from nouns, verbs, and postpositions in that they do not admit pronominal prefixes. Further, adverbs cannot serve as the object of a postposition. Adverbs typically occur immediately preceding the predicate, although this is not required. (85)

Gai bloppa mosang tabba patang. very unable 3SG.POSS weapon sabre ‘His sabre was completely useless.’ [publia070] The largest category of adverbs consists of adverbs of time.

(86)

Nai la marakang iatang wala. 1SG.POSS cauldron last.night disappear DISC ‘My cauldron disappeared last night.’ [bm141] A complete list of temporal adverbs is given in Table 5.

Table 5: Temporal adverbs anaga/ana anata anu’us dalla dalla mabu haya me ir kaleku ir lera ir mabu ir pata mabu markang tialaku tubi umme waddipang was anaga yenarpang

‘now’ ‘shortly, soon’ ‘presently’ ‘tomorrow’ ‘tomorrow morning’ ‘rainy season’ ‘cold season’ ‘daytime’ ‘afternoon’ ‘night’ ‘twilight’ ‘last night’ ‘day after tomorrow’ ‘long ago’ ‘night time’ ‘yesterday’ ‘today’ ‘day before yesterday’

50

Gary Holton

Another important category of adverbs consists of deictic adverbs. The deictic adverbs formally resemble the demonstratives (§4.5). Like the demonstratives the deictic adverbs distinguish proximal, distal, and three levels of elevation. Unlike the demonstratives, the deictic adverbs distinguish only the specific vs. nonspecific dimension; the visible/non-visible dimension is not coded on deictic adverbs. Table 6: Deictic adverbs (Tubbe dialect)

PROXIMAL DISTAL HIGH LEVEL LOW

SPEC

NSPEC

siga sina srau smau spau

iga ina dau mau pau

Though similar in function to demonstratives, the deictic adverbs only indicate location and situation of a proposition but are not pronominal. That is, they do not refer but are instead used as adverbs to indicate a location. In form the deictic adverbs resemble the visible deictic pronouns; they differ in that the deictic adverbs lack the final ‑gu in the elevation terms and they replace the diphthong ai with i in the proximal and distal terms. Also, the deictic adverbs do not distinguish visible and non-visible. Some examples of deictic adverbs are given below: (87) Wo~wo wele~wele mumum mlai prani gai sida dagurung procession applause small.music 3SG.POSS sound thunder mau LEVEL. NSPEC

mudali dali. play sing

‘The sound of music and applause thundered over there from the singing.’ [bm056] (88) Tamakang ga’ai So Balang hali ga’ai srau gaunung T. 3SG.GEN So Balang river 3SG.GEN HIGH.SPEC just ya si. toward this ‘Tamakang has its river So Balang down toward there.’ [tonuburi239]

Western Pantar

(89)

51

Deka Ang dau a-yau hai. Deka Ang HIGH.NSPEC INCP-war EXCL ‘Deka Ang wants to make war down there!’ [pubila049]

In contrast to the demonstrative pronouns, deictic adverbs have scope over the entire predicate and hence can occur in other locations within the clause. For example, (90) could be rendered equally as (91). (90)

(91)

Mantri dau gang gar ti’ang tulis. write official HIGH.NSPEC 3SG.ACT continue stay ‘The official up there just kept on writing.’ [publia121] Dau

mantri gang gar official 3SG.ACT continue ‘The official up there just kept writing.’

HIGH. NSPEC

ti’ang stay

tulis. write

This is not the case with demonstrative pronouns, which always occur following the adjunct noun. Thus, bla saiga ‘the house here’ but not *saiga bla. While deictic adverbs may move around in the clause, the position of the demonstrative pronouns is fixed within the NP.

3.7 Negation Clausal negation is accomplished via the clause-final negative particle kauwa. This same form is used as an exclamation to express a negative response. (92)

(93)

(94)

Anaga was arugga today sun hot ‘It’s not hot today.’ Gang ga-ume 3SG.ACT 3SG-inside ‘He doesn’t like it.’

kauwa. NEG

banang like

kauwa. NEG

Ging naing hukung kauwa. 3PL.ACT 1SG.UND sentence NEG ‘They didn’t put me in jail.’ [publia460]

The negative kauwa may also be used to negate non-verbal predicates, as in (95).

52 (95)

Gary Holton

Halia maro’o kauwa, di hala ba raung. NEG rat FOC REL climb frog ‘It wasn’t a frog, but rather a rat which climbed up.’ [frog054]

The negative kauwa retains limited verbal attributes. It can function as a predicate, as in (96), and may even be inflected for aspect, as in (97). (96)

Nang kauwa hai. EXC 1SG.ACT NEG ‘I’m not (doing that)!’

(97)

I-kauwa

ing ha-ttang i-lu’ung! PROG-NEG then 2SG- hand 4PL-cut ‘Not (being the case) then they will cut off your hands.’ [publia287]

The negative kauwa carries a completive sense indicating an expectation that the proposition it modifies will remain false. For example, if the weather is not hot but there is a reasonable expectation that it will become hot, then (92) above would be infelicitous. Instead, such incompletive negation is expressed with the adverb yadda ‘not yet’. This form occurs in complementary distribution with kauwa. (98)

Nang hori yadda. 1SG.ACT eat not.yet ‘I have not eaten yet.’

The alternative hori kauwa would imply that I have not eaten and do not expect to eat, implying perhaps that I am fasting or starving. The adverb yer may occur with yadda, in which case yer still precedes the verb and yadda follows the verb. (99)

Gang yer ma yadda. 3SG.ACT still come not.yet ‘He still hasn’t come.’

Negative imperative constructions are marked with the clause-final particle gayang. (100)

Horang gayang. make.noise NEG.IMP ‘Don’t make noise’

53

Western Pantar

(101)

Ga-leba kassing wang haliati gayang. exist tease NEG .IMP 3SG-lip split ‘Don’t tease someone with a harelip.’

The negative imperative may be combined with the incompletive yadda ‘not yet’. (102)

Yadda maiyang gayang. not.yet place NEG .IMP ‘Don’t put them down yet.’ [pubila097a]

4 Noun phrases Nouns are here formally delineated by their ability to accept nominal morphology, particularly possessive prefixes. Nouns may head a noun phrase structured as follows.

(103)

Structure of the noun phrase [N ADJ CLF NUM DEM ART]NP

Noun phrases may co-occur in apposition with free pronouns. In this case the pronoun follows the entire noun phrase. (104)

[[Tibbi si]NP gang]NP [[ai hali]NP me]PP ART 3SG.ACT 4SG.POSS water LOC crab ‘The crab took his water down.’ [sirung041]

dekang descend

ga. PFV

The use of a co-referential pronoun is the preferred strategy with names and complex noun phrases. (105)

[[Tabang slave a-raung INCP-climb

alaku two yattu tree

Duinni Maggangkala]NP Duinni Maggangkala ga-ung 3SG-head

[ging]NP]NP 3PL.ACT

misingup. sit

‘The two slaves Duinni Maggangkala climbed up and sat in the tree.’ [tonu057]3 3 Duinni Maggangkala is a single (binomial) name give to the pair of slaves together.

54

Gary Holton

WP nouns are not marked for gender or noun class, nor are such markings indicated on the verb or via possessive prefixes. Gender is not marked on any other word class either. Nouns referring to less time-stable concepts can be derived from nouns via the addition of nominal morphology. For example, the noun ‘running’ in (106) is derived from the verb biring ‘to run’ via the addition of possessive prefix. (106)

Gai biring tabba diala. very fast 3SG.POSS run ‘His running is very fast.’ / ‘He runs very fast.’

4.1 Number WP nouns are not obligatorily marked for number, though the plural marker maru(ng) may be optionally used to indicate plurality of both animate (107) and inanimate referents (108). (107)

Wenang marung ging pia. 3PL.ACT descend old.man PL ‘The old men went down.’ [bm016]

(108)

Pia gi hai sampan marung sing. art descend 3PL.POSS boat/canoe PL ‘(They) went down to their canoes.’ [conflict042]

The plural word cannot occur with numeral (109) or numeral classifier (110). (109) *ke’e fish

kealaku twenty

(110) *ke’e fish

bina/dis

maru

CLF/ CLF

PL

maru pl

The plural word cannot occur independently in the noun phrase slot (111), but it can occur in a noun phrase in apposition with a pronoun, as in (107) above. (111) *marung PL

lama walk

ta IPFV

Western Pantar

55

The human nominalizing morpheme wala may also function as an associative plural for nouns referring to humans. (112) [Deka Ang wala] sura si kang rakki ya wang i-ti’ang. tear road exist PROG-sleep Deka Ang person letter ART hit ‘Deka Ang and his men tore up the letter leaving it lying on the road.’ [pubila045]

4.2 Adjectives Adjectives can be distinguished from both nouns and verbs by a lack of inflectional morphology. In particular, adjectives do not admit person-marking prefixes. Adjectives may be used in both attributive and predicative functions. Attributive adjectives directly follow the noun which they modify. (113)

[Toggo ara ye] i-diakang. python large one PROG-descend ‘A large python was coming down.’

(114)

Ging [sabu miaka] wang putar kanna gob. exist turn already then 3PL.ACT cloth white ‘They bound [them] with white cloth, then.’ [pubila375]

Predicative adjectives pattern as verbs which do not admit person-marking prefixes (see §5.2.2). Adjectives may function to express quantification, as discussed in the following section.

4.3 Quantification Quantification is expressed with adjectives or with nouns. Adjectival quantifiers immediately follow the noun. (115)

Aname buware gaunung person few just ‘(I) saw just a few people.’

(116)

Wakke~wakke haweri wang Tubbe birang kalalang. many exist Tubbe speak know child~PL ‘Most/many children can speak the Tubbe language.’

ga-niaka. 3SG-see

56

Gary Holton

Quantification may also be expressed with nouns. The quantifier der ‘some’ is a noun which occurs in apposition to the noun phrase. (117)

[[Hai bloppa sing]NP der]NP ga-r diakang. some 3SG-with descend 2SG.POSS weapon ART ‘Bring down some of your weapons.’ [publia152]

The quantifier gaterannang ‘all of them’ is a noun which contains a fossilized third-person person-marking prefix and so resembles an adnominal possession construction. Unlike the other nominal quantifier, gaterannang exhibits some distributional variation, occurring both to the left (118) and to the right (119) of the article. (118)

Ha hing gaterannang si. ART yes 2PL.ACT all.of.them ‘Yes, the lot of you.’ [publia304]

(119)

Habbang sing gaterannang gi-mma=m sussung kanna. village ART all.of.them 3PL-nose=LOC throw.out already ‘It could already be smelled by that entire village.’ [bm063]

4.4 Numeral classifiers WP makes extensive use of numeral classifiers describing the shape or structure of an object. Classifiers exist along a cline between lexicon and morphosyntax (Grinevald 2007), and most WP classifiers still retain lexical functions in addition to their classifier functions. Numeral classifiers follow the noun phrase and precede the numeral. (120)

bla (haila) house CLF:area ‘one house’

ye one

Numeral classifiers are generally optional but may be obligatory in cases where they provide additional specifying information, as with mensural classifiers. In (121) the classifier bina ‘detached’ indicates individual fish as opposed to bundles of fish.

57

Western Pantar

(121)

Ke’e bina dinni? fish CLF:general how.many ‘How many individual fish?’

In (122) the classifier gamma ‘nose ridge, point of land’ indicates that the reference is to a stick rather than an entire tree. (122)

Yattu gamma ye boti ge’e tang ti’ang. CLF:sticklike one basket rim on lay tree ‘A stick of wood lying across the rim of a basket.’

Some common numeral classifiers and examples of their use are given in Table 7 below. This list is not exhaustive. Table 7: Common Western Pantar numeral classifiers Classifier

Usage

Meaning of source lexeme

bina haila dawal dis kakka gamma hissa kassi lu’a waya

general classifier objects with areal extent rope-like object (rope, nylon, cable thread, etc.) object strung on a string long, stiff, flat objects sticklike objects fruit, contents split-off objects rounded object flat, flexible objects

‘detached’ ‘base, area’ ‘roll, coil’ ‘stringer’ ‘stiff’ ‘nose, point of land’ ‘fruit’ ‘to split’ ‘rounded, oblong’ ‘leaf’

4.5 Demonstratives WP demonstratives may function as determiners within the noun phrase or as pronouns. The demonstratives index spatial deixis along three mutually exclusive dimensions of geophysical elevation, visibility, and specificity. The forms are given in Table 8. Table 8: Demonstratives visible

proximal (PROX) distal (DIST ) above (HIGH) below (LOW) same level (LEVEL)

non-visible

SPEC

NSPEC

SPEC

NSPEC

saiga saina sraugu spaugu smaugu

aiga aina daugu paugu maugu

sigamme sinamme sraume spaume smaume

igamme inamme daume paume maume

58

Gary Holton

4.5.1 Visibility The distinction between visible and non-visible is addressee based but controlled by the speaker. That is, a speaker uses visible forms when she expects that the addressee is able to see the referent. Similarly, non-visible forms are used when the speaker does not expect the addressee to be able to see the referent. For example, the non-visible demonstrative smaume in (123) is felicitous regardless of whether the referent ‘book’ is located in an adjacent room out of sight of both the speaker and the addressee, or simply located behind the speaker out of sight of the addressee. (123)

Buku smaume haggi. book LEVEL.SPEC.NVIS take ‘Get the book over there (not visible to you).’

(124) Tar pi-gaddi ta aname smaugu haddung pi-galu. 1EXCL-expel what 1INCL-make IPFV person LEVEL.SPEC.VIS steal ‘What can we do to get rid of those people over there (visible) who are stealing?’ 4.5.2 Specificity Specific demonstratives are used to emphasize a particular location rather than a more general, areal location. Often, specific forms indicate a precise location. In (125) the speaker wishes to indicate the location at which he wants to turn into the forest to follow a trail inland. The trail is not yet visible but the speaker knows well the location of the trail, though the addressee does not. (125)

Smaume

ta

LEVEL .SPEC .NVIS

IPFV

pi-mia. 1INCL-ascend ‘Let’s ascend from that (precise) place over there.’

In contrast, non-specific forms are commonly used to express that a speaker knows only the general location, not the precise location of a referent. For example, in reply to a question about the location of a certain thing or person, a speaker may use the non-specific form inam(me) ‘there (somewhere)’.

59

Western Pantar

(126)

Ga-niaka 3SG-see

allang then

bila cliff

kalla small

inam

i-ti’ang.

DIST.NSPEC. NVIS

PROG -stay

ye one

inam

a-lelang

DIST.NSPEC .NVIS

INCP-sheltered

‘(They) saw a small sheltered protrusion sitting there.’ If the precise location is known, such as the specific house where the person or thing is located, then the speaker will use the specific form sinam(me) instead. (127)

Sinam

kaweni ging tang saukang kalung children 3PL.ACT on look here.there ‘They looked around at the children there.’ [bm152] DIST.SPEC.NVIS

allang. then

In contrast to visibility, specificity is speaker-based, in that it is based on the speaker’s assessment of whether the addressee will be able to identify a referent. In narrative contexts initial reference is often made using a non-specific demonstrative, while continuing reference is made using a specific demonstrative, as in the following example. (128)

Gang 3SG.ACT

mittung first

eu woman

gai 3SG.POSS

si

maum

ART

LEVEL .NSPEC. NVIS

bukkus room

bukkus room

me

smaume

LOC

LEVEL. SPEC. NVIS

me LOC

gar continue

halli. cry

‘At first the woman is in a room [then] she cries in her room.’ [marriage146]

4.6 Articles WP has two articles sing and si which may optionally mark identifiable referents of significance to the discourse. The articles sing and si can be loosely described as proximal and distal, respectively, though their precise function is difficult to characterize. Articles may occur with full NPs or names. (129)

Jawa eu manne marung sing ayang ma. PL ART drift come Java woman wife ‘These Javanese married women came in noisily.’ [publia465]

60

Gary Holton

(130)

Peing Lila si ga-kkang. Peing Lila ART 3SG-strike ‘That Peing Lila was killed.’ [darang052]

(131)

Gai ke’e maru si aname ging ART person 3PL. ACT 3SG.POSS fish PL ‘People took those fish of ours.’ [tonuburi073]

haggi take

kanna. already

4.7 Possession Western Pantar makes use of four distinct strategies for expressing nominal possession: alienable possessive proclitic; inalienable possessive prefix; undergoer pronoun; and genitive pronoun. These strategies are exemplified in Table 9 with third-person singular. Table 9: Summary of nominal possession strategies alienable possessive proclitic inalienable possessive prefix undergoer pronoun genitive pronoun

gai bla ‘his house’ ga-wasing ‘his tooth’ gaing ke’e ‘his bone’ bala ga’ai ‘the house of his’

4.7.1 Alienable possession Alienable possession is indicated via the use of an alienable possessive proclitc indexing the person and number of the possessor. These proclitics attach to the possessed noun phrase and may optionally be preceded by an adjunct noun phrase referring to the possessor. Though written here as separate words following the practical orthography, they are bound forms which cannot occur independently. The alienable possessive proclitics have essentially the same phonological shape as the inalienable possessive prefixes (§4.7.2), with the addition of a high front vowel. In contrast to the inalienable prefixes they are never phonologically integrated into the noun stem. As illustrated in (132), the possessive pronoun construction is right-headed, and the proclitic functions as a linker between the possessor and the possessum.

(132)

Possessive pronoun construction (NPPOSSESSOR ) LINKAL NPOSSESSUM

Western Pantar

61

The paradigm of alienable possessive proclitics is given in Table 10. Table 10: Alienable possessive proclitics

1 2 3 4

singular

plural

nai hai gai ai

pi/tai/ni hi gi i

While alienable possessive proclitics may be used optionally in the inalienable possessive construction (see below), the alienable construction is distinguished by the lack of an inalienable possessive prefix. An adjunct noun phrase referring to the possessor may optionally precede the possessive pronoun. (133)

wenang gai man 3SG.POSS ‘the man’s house’

bla house

Nouns which are obligatorily possessed normally occur with the inalienable construction (§4.7.2). However, such nouns may also be alienably possessed using the alienable possessive proclitic. Alienable possession of obligatorily possessed nouns is used to indicate non-intrinsic possession of nouns which would otherwise be inalienably possessed. The prototypical usage is to indicate possession of a body part which is not part of one’s one own body. (134)

ke’e nake’e nai ke’e

‘bone’ ‘my bone’ (in my body) ‘my bone’ (part of an animal being butchered)

Another example can be found lexicalized in the word ai wake ‘orphan’, literally, ‘his/her (fourth person) child (alienable)’. Compare na-wake ‘my child’

4.7.2 Inalienable possession Inalienable possession is expressed morphologically via an inalienable possessive prefix indexing the person and number of the possessor. The inalienable possessive prefixes are formally identical to the person-marking prefixes on verbs (see §5.2). (In addition to nouns and verbs, person-marking prefixes also occur with

62

Gary Holton

postpositions.) Inalienable possessive prefixes may be optionally used in conjunction with alienable possessive proclitics (§4.7.1). An adjunct noun phrase referring to the possessor may also precede the inalienable possessive prefix. Like the alienable possession construction above, the inalienable possession construction is right headed.

(135)

Inalienable possession (NPPOSSESSOR ) (LINKAL ) LINKINAL-NPOSSESSUM

As with other person-marking paradigms, the paradigm of inalienable possessive prefixes includes three forms referring to first person plural, written here in the order inclusive/distributive/exclusive. The fourth-person forms are often used to refer to a distinct third person, similar to an obviative pronoun; however, they may also have a reflexive function, quite different from the obviative usage (see below). Table 11: Inalienable possessive prefixes

1 2 3 4

singular

plural

nahagaa-

pi-/ta-/nihigii-

Inalienable possession is the standard way to express possession of nouns which are obligatorily possessed, including many nouns referring to body parts and kinship and locational nouns. An example is given below.

(136)

Possessive paradigm for -wasing ‘tooth’ (nai) nawasing ‘my tooth’ (hai) hawasing ‘your tooth’ (gai) gawasing ‘his/her/its tooth’ (ai) awasing ‘his/her/its tooth’ (distinct) (pi) piwasing ‘our (inclusive) tooth’ (tai) tawasing ‘each of our (distributive) tooth’ (ni) niwasing ‘our (exclusive) tooth’ (hi) hiwasing ‘your tooth’ (gi) giwasing ‘their tooth’ (i) iwasing ‘their tooth’ (distinct)

Western Pantar

63

Other consonant-initial obligatorily-possessed nouns follow the same pattern of inalienable possession. Some examples with the third-person singular prefix ga- are given in (137).

(137)

Consonant-initial obligatorily possessed nouns -s gas ‘its stomach’ -lebu galebu ‘its tongue’ -’ung ga’ung ‘its head’ -ru garu ‘his wife’ -’ar ga’ar ‘his male cross-cousin’

When a nominal adjunct is present the use of an alienable possessive proclitic in addition to the inalienable possessive prefix is the preferred strategy for animate referents (138) but not for inanimates (139). (138)

kuba gai g-iar old.woman 3SG.POSS 3SG-father ‘the old woman’s father’

(? kuba giar)

(139)

yattu ga-’ung (? yattu gai ga-’ung) tree 3SG-head ‘the head (top) of the tree’

The use of inalienable possession is not restricted to obligatorily possessed nouns but may be used to indicate intrinsic possession of any noun where semantically possible. Moreover, some nouns which might be thought on semantic grounds to be obligatorily possessed do not always occur with a possessive prefix. For example, the noun uwe ‘ear’ may occur in unpossessed form in certain idiomatic constructions such as (140). (140)

Ging mang mising uwe la’ai. sit ear tie 3SG.ACT only ‘They just sat with ears tied (i.e., listening intently).’ [soli027]

This noun can also occur inalienably possessed, for example, nauwe ‘my ear’. In addition, nouns which are usually inalienably possessed may occur without possessive morphology in nominal attribution constructions. (141)

do buka sago trunk/body ‘trunk of a sago tree.’

64

Gary Holton

However, for most non-obligatorily possessed nouns the inalienable possession construction is not possible. For example, *nabla is not acceptable for ‘my house’. Obligatorily possessed nouns occur in citation form with either a thirdperson or distributive prefix. However, the distributive plural prefix ta- is extremely rare in natural discourse and is generally used when individuation is required, as in (142). (142)

Ta-gas be ta-buka yattang. 1DISTR-reflection shadow 1DISTR-body same.length ‘(Each of ) our shadows are as long as (each of ) our bodies.’

4.7.3 Another inalienable possession construction There is one further method of expressing inalienable possession which makes use of the undergoer independent pronouns. While this construction is not possible with all nouns, where it is possible it contrasts in meaning with the usual inalienable possessive construction formed with the possessive prefix. The construction with the undergoer pronoun refers to a particular item or part of an item, while the construction with the prefix refers to an entirety of items. For example, consider the noun ke’e, which can mean ‘bone’ or ‘fish/meat’. (143)

naing ke’e 1SG.UND bone ‘my bone, one of the bones in my body’

(144)

na-ke’e 1SG-bone ‘my bones, all the bones in my body’

Thus, both the undergoer pronoun and the prefix can express inalienable possession of a body part. In contrast, the possessive pronoun can only by used with the meaning ‘fish/meat’, not to express possession of a body part. (145)

nai ke’e 1SG.POSS fish/meat ‘my fish/meat’

Western Pantar

65

Similarly, with the noun killi ‘skin’ the use of the undergoer pronoun (146) refers to a part of skin, for example, a patch of skin which is infected or fungal. The use of the prefix (147) refers to the entirety of one’s skin. (146)

naing killi 1SG.UND skin ‘my skin, a particular part of my skin’

(147)

na-killi 1SG-skin ‘my skin, my entire skin’

4.7.4 Genitive construction An alternative possessive construction is formed using the genitive pronoun rather than the possessive pronoun. The genitive pronoun may occur either independently or in conjunction with an adjunct noun phrase referring to the possessor. Genitive pronouns function as the head of a genitive phrase. In contrast to the possessive construction with the possessive pronoun, the possessor follows the possessum. The possessum noun phrase can also be omitted altogether when the reference is clear from context.

(148)

Word order in genitive construction (NPPOSSESSUM ) (NPPOSSESSOR ) PROGEN

The form of the genitive pronoun follows the same pattern as that of the other pronouns, as shown in the table below. Table 12: Genitive pronouns

1 2 3 4

singular

plural

na’ai ha’ai ga’ai a’ai

pi’i/ta’ai/ni’i hi’i gi’i i’i

The genitive construction differs structurally from the ordinary possessive construction in that the latter is head-final (149b), while the former is head-initial (149a).

66

(149)

Gary Holton

Comparison of genitive and possessive constructions a. ( Jon) gai bla Jon 3SG.POSS house ‘his house’ (‘John’s house’) b.

bla ( Jon) ga’ai house Jon 3SG.GEN ‘the house of his’ (‘the house of John’s’)

Like other NPs, genitive pronouns may be followed by a quantifier. In (150) the third-person genitive pronoun is followed by an article. The resulting noun phrase ga’ai sing then functions as an argument of the transitive verb maggar. (150)

Ga’ai 3SG.GEN

sing art

maggar work.garden

gayang. NEG. IMP

‘Don’t work in his garden.’ (lit. ‘Don’t work to clear that which he owns.’) [pubila061] As indicated above, the genitive pronoun may optionally occur with a noun phrase referring to the possessor and the possessed. In (151) the nominal compounds ber mauwa ‘history’ and yau Pu Bila ‘Pu Bila war’ serve as possessum and possessor, respectively. na-birang. (151) [Ber mauwa]POSSESSUM [yau Pu Bila]POSSESSOR ga’ai words old war Pu Bila 3SG.GEN 1SG-speak ‘I will tell the history of the Pu Bila war.’ [pubila002] Either one or both of the possessum and possessor NPs could be omitted if the referent were already established in discourse. The first example below omits explicit nominal reference to the possessor yau Pu Bila. The second example omits explicit reference to the possessum ber mauwa. The third example omits explicit reference to both the possessum and the possessor. (152)

a.

[Ber mauwa]POSSESSUM words old ‘I will tell its history.’

b.

na-birang. [Yau Pu Bila]POSSESSOR ga’ai war Pu Bila 3SG.GEN 1SG-speak ‘I will tell that (history) of the Pu Bila war.’

ga’ai 3SG.GEN

na-birang. 1SG-speak

Western Pantar

c.

67

Ga’ai na-birang. 3SG.GEN 1SG-speak ‘I will tell about its (history).’

Like any noun phrase, a genitive pronoun may function as a nominal predicate in an equational construction. In this case the genitive pronoun predicates ownership or possession. Since reference to the possessor and possessed may be omitted, a genitive pronoun alone can function as a predicate. (153)

Na’ai. 1SG.GEN ‘mine’ / ‘that of mine’ / ‘It is mine.’

(154)

Bla sing na’ai. house ART 1SG.GEN ‘The house is mine.’ / ‘It is my house.’

(155)

Talle gaterannang ha’ai. ancestor all.of.them 2SG.GEN ‘These are all of your ancestors.’ (lit. ‘All these ancestors are yours.’) [tonu258]

(156)

Nang wenang Amon ga’ai. 1SG.ACT old.man Amon 3SG.GEN ‘I am Mr. Amon’s child’ (lit. ‘I am Mr. Amon’s.’, ‘I belong to Mr. Amon.’)

Genitive pronouns may also occur as part of nominal compounds, as in (157). (157)

Tawa ga-ume ga’ai. ocean 3SG-insde 3SG.GEN ‘sea-life’ (lit. ‘that which the inside of the ocean owns’)

Pragmatically, genitive pronouns differ from possessive pronouns in having a more emphatic reading, asserting the possessive relationship. (158)

Na’ai na-raung. 1SG.GEN 1SG-climb ‘I am the one who will climb.’

68 (159)

(160)

Gary Holton

Na’ai na-birang. 1SG.GEN 1SG-speak ‘Let me speak.’ (lit. ‘It is mine, I will speak.’) Saina=b

ha’ai. 2SG.GEN ‘That is what you must do.’ (lit. ‘That’s what is yours.’) [bela024] DIST.VIS= REL

Genitive pronouns are often used to refer to actions, as in (159) and (160) above. In the following example the third plural genitive refers to something done by a third party. Thus, this example could be literally translated as ‘what shall we do to return theirs?’, a reference to revenge. (161)

Gob tar pi-gaddi ta gi’i as wang pi-golang? thus how 1INCL-do before 3PL.GEN also exist 1INCL-return ‘How will we respond to what they have done?’ [port013]

4.8 Locational nouns Locational nouns (NLOC ) are nouns which designate a location with respect to second referent (NPREF ). The referenced entity may be unexpressed or may be expressed via a pronominal prefix, usually third-person singular ga‑. The prototypical locational noun construction is thus a type of inalienable adnominal possessive construction in which NLOC is the possessum (head) noun and NPREF is the possessor. The structure of the locational noun construction is diagrammed below.

(162)

Locational noun construction ( (NPREF) ga-)NLOC

When NPREF is overtly specified in the locational noun construction, the construction refers to the area located in a specified direction or elevation from NPREF. When NPREF is not specified, the construction refers to the area located in a specific direction or elevation from an already established reference point. (163)

kebang ga-wanang barn 3SG-uphill ‘the area uphill from the barn’

Western Pantar

(164)

ga-wanang 3SG-uphill ‘the area uphill from it’

(165)

Bai [[kakus]NPREF [ga-lawang]NPLOC ] ga-laking. pig toilet 3SG-area.downhill 3SG-tie ‘The pig is tied up in the area downhill of the toilet.’

(166)

Kuri marongga.tree

[[bla]NPREF house

[ga-manang]NPLOC ] 3SG-front

69

natar. stand

‘The marongga tree stands in front of the house.’ (between the speaker and the house) There are eight locational nouns, and they include both terms which refer to geophysical elevation and terms which refer to global elevation, in the sense of Burenhult (2008). Table 13: Locational nouns wanang manang lawang bila / dimmang

‘behind, other side’ (same level) ‘in front, this side’ (same level) ‘downhill’ (below) ‘uphill’ (above)

geophysical elevation

butang taggang pittung muang

‘above’ (not touching) ‘above’ (in contact) ‘below’ (in contact) ‘below’ (not touching)

global elevation

The forms bila and dimmang are synonymous, differing in distribution in that only bila may also occur as a free form without a pronominal prefix. The forms wanang and manang also refer to geophysical elevation, but they assume a reference object (NPREF) which is capable of obstructing the view. The forms butang, taggang, pittung, and muang refer to vertical height above or below and object. Of these, taggang and pittung indicate physical contact, while the forms butang and muang indicate lack of physical contact. Locational nouns alternate syntactically with postpositions. The phrase bla gamanang ‘the region in front of the house’ contrasts with the semantically similar but syntactically distinct postpositional phrase bla gane ‘in front of the house.’ Locational nouns also contrast semantically with postpositions, in that they depend crucially on location of the deictic center for their interpretation. Thus, bla gane refers to an intrinsically determined front of the house: the side

70

Gary Holton

with the front door, opposite the kitchen. In contrast, bla gamanang refers to the area in front of the house from deictic center. In terms of frames of reference (Levinson 2003), constructions using postpositions describe an intrinsic frame of reference, whereas those using locational nouns describe a relative frame of reference. Locational nouns can be used alone as nouns denoting an area, as in (167) and (168). (167)

Ma butang ya. come area.above head.toward ‘Head toward the area above.’

(168)

Taggang lama. above walk ‘Walk on the higher side (of the path).’

Postpositions would not be felicitous in these contexts. On the other hand, locational nouns cannot substitute for postpositions either. For example, neither taggang or butang can be substituted for the postposition tang ‘on’ in the following example. (169)

Tiaku glass

mea table

tang on

migang. set

‘Set the glass on the table.’ *tiaku mea taggang/butang migang However, a locational noun may occur within a postpositional phrase headed by a semantically similar postposition. (170)

A-raung daum luppu butang tang garung. release 4SG-climb LEVEL.NSPEC platform area.above on ‘He climbed up on a tall raised bamboo platform and set her there.’ [bm054]

Both the locational noun and the postposition can occur with a pronominal prefix. The prefix is obligatory when a locational noun occurs in apposition with a noun.

Western Pantar

71

5 Pronouns and person-marking prefixes The system of personal pronouns and person-marking prefixes includes six distinct paradigms indexing person and number of the referent. Free pronouns and bound prefixes function as part of a single referent-tracking system and hence are discussed together here. There are four paradigms of independent (free) pronouns: actor (ACT), undergoer (UND), genitive (GEN), and reflexive (REFL); and two paradigms of bound person-marking prefixes, one marking alienable possession (POSS) and one general prefix marking both inalienable possession on nouns (§4.7.2) and core arguments on verbs and postpositions. Table 14: Personal pronouns and person-marking prefixes

1SG 2SG 3SG 4SG 1INCL 1DISTR 1EXCL 2PL 3PL 4PL

ACT

UND

GEN

REFL

POSS

PREFIX

nang hang gang ang ping tang ning hing ging ing

naing haing gaing aing pi’ing taing ni’ing hi’ing gi’ing i’ing

na’ai ha’ai ga’ai a’ai pi’i ta’ai ni’i yi’i gi’i i’i

neri heri geri eri piri teri niri hiri giri iri

nai hai gai ai pi tai ni yi gi i

nahgaapitanihigii-

The first person distributive actor and undergoer independent pronouns tang and taing, while theoretically plausible forms, are extremely difficult to elicit and extremely rare in natural speech. I have recorded only one example of an undergoer distributive pronoun in spontaneous speech. The same is true of the first person distributive person-marking prefix ta- with verbs. However, distributive ta‑ does occur with nouns and postpositions, particularly in citation form or with generic reference. The fourth-person actor and undergoer pronouns have several functions. First, fourth-person pronouns may be used as a reflexive marker to indicate coreference between two core arguments. (171)

Gang aing daranung. 3SG.ACT 4SG.UND protect ‘He is protecting himself.’

The reflexive construction in (171) is only possible for third person referents. For non-third persons reflexivity must be indicated using a reflexive pronoun, as in (172) and (173).

72

Gary Holton

(172)

Neri na-niaka. 1SG.REFL 1SG-see ‘I saw myself.’

(173)

Nai ber asang si heri kalalang lamala hai. ART 2SG .REFL know certainly DISC 1SG.POSS word say ‘You can understand for yourself what I am saying.’ [adat 2.014]

The reflexive eri can be used with third person referents as well, resulting in a functional overlap with the fourth person pronouns. Thus, both the reflexive pronoun (174a) and the fourth person undergoer pronoun (174b) are acceptable as the marker of reflexivity, though with this particular lexical item (‘to see’) the reflexive pronoun is preferred. (174)

a.

Gang eri 3SG.ACT 4SG.REFL ‘He saw himself.’

a-niaka. 4SG-see

b.

Gang aing 3SG.ACT 4SG.UND ‘He saw himself.’

a-niaka. 4SG-see

In discourse context fourth-person pronouns function as switch-reference markers signaling that a referent is not co-referential with the currently tracked referent. In (175) the fourth-person pronoun does not indicate reflexivity but rather points anaphorically to a previously established referent. (175)

Perung gang aing mukkung. whale 3SG.ACT 4SG.UND swallow ‘The whale swallowed him.’ [bm092]

Similarly, in (176) the fourth-person pronoun references a previously established participant. (176)

Aing ping wa pussur itami hubba yasa. bad 4SG.UND 1INCL.ACT go spatter until wet ‘We were spattered by him until we were completely wet.’ [tonuburi262]

In the following text excerpt about marriage customs the third-person pronoun gang refers to the groom, while the fourth-person ing refers to the bride’s

Western Pantar

73

parents. The use of the fourth-person eliminates potential ambiguity whereby the referent of the second clause could be read as referring to the bride and groom. (177)

Inata so.that

eu girl

sing

Kalau if

kauwa

ART

NEG

gang 3SG.ACT

ing 4PL.ACT

bisa able

ma come

haggi. take

bla house

me, . . . LOC

‘So that he can take (marry) the girl. If not, they (the parents) come to the house, . . .’ [marriage039] Fourth-person prefixes have a reflexive function with third-person referents. (178)

Gang a-kkang. 3SG.ACT 4SG-hit ‘He hit himself.’

(179)

A-ttang kusi tang lu’ung. cut 4SG-hand nail on ‘(He) cut his own fingernails.’

Fourth-person possessive pronouns also have a reflexive function in that they indicate co-reference with an established participant as opposed to introducing a new participant. (180) Re wani dimis ga-r ma ai baa ga-nia dimi. bird honey suck 3SG-with come 4SG.POSS young 3SG-give suckle ‘The bird gathered honey and brought it to suckle its own chicks.’ In (181) the fourth-person possessive ai indicates that the crab retreated to his own river rather than to the river by the people. (181)

Aname person tibbi crab

bunni garden

si ART

me

marung

sing

LOC

PL

ART

gang 3SG.ACT

ai 4SG.POSS

hali water

ayang drift me LOC

ma come dekang descend

ga. PFV

‘The people drifted in and the crab descended to his (own) river.’ [sirung041]

74

Gary Holton

Similarly, in (182) the fourth-person possessive indicates that the reference is to Wenang Kalabasi’s slaves rather than to the grandmother’s slaves. (182) Wenang Kalabasi gang ai tabang alaku Duinni Maggangkala sir Kalabahi 3SG.ACT 4SG.POSS slave two Duinni Maggangkala as tang gi-uwang wa kuba ga-asang. also further 3PL-order go grandmother 3SG-say ‘Sir Kalabasi ordered his two slaves Duinni Maggangkala to go tell grandmother.’ [tonu034] The distribution of fourth-person pronouns is highly sensitive to discourse context; fourth-person pronouns are rare in elicitation contexts.

5.1 Actor and undergoer independent pronouns Actor and undergoer independent pronouns are distributed according to an agentpatient system of semantic alignment (Mithun 1991). More controlling arguments are encoded with actor pronouns (183), while less controlling arguments are encoded with undergoer pronouns (184). (183)

Nang mising. 1SG.ACT sit ‘I am sitting.’

(184)

Naing massa. 1SG.UND tired ‘I am tired.’

In transitive clauses both actor and undergoer pronouns may occur, indexing A and P argument, respectively, as in (185). (185)

Nang gaing 1SG.ACT 3SG.UND ‘I cut him.’

lu’ung. cut

Independent pronouns may occur in apposition with a co-referential noun or noun phrase. In (186) the third-person actor pronoun occurs in apposition

Western Pantar

75

with the name Hen, while in (187) the pronoun occurs in apposition to a full noun phrase aname ara sing ‘this big man’. (186)

Hen gang mising. Hen 3SG.ACT sit ‘Hen is sitting.’

(187)

Aname ara sing gang person large ART 3SG.ACT ‘This big man is sitting.’

mising. sit

Independent pronouns may occur with co-referential person-marking prefixes, as in (188). (188)

Nang na-lama 1SG.ACT 1SG-walk ‘I will go.’

ta. IPFV

In (189) the undergoer pronoun occurs with the noun phrase adjunct kuba ‘old woman’, additionally co-indexed by the person-marking prefix ga-. (189)

Gang kuba gaing si 3SG.ACT old.woman 3SG.UND ART ‘He gave (them) to the old woman.’

ga-nia. 3SG-give

Where varying degrees of control can be attributed to a participant, an alternation between actor and undergoer pronoun is possible. The verb muddi ‘strong’ typically occurs with an undergoer pronoun, as in (190), reflecting less participant control. (190)

Naing muddi. 1SG.UND strong ‘I am strong.’

In (191) the use of the actor pronoun imparts a modal reading, implying a certain degree of control on the part of the participant. This could be felicitous when the participant has the intention to be strong, to endure a challenge. (191)

Nang muddi. 1SG.ACT strong ‘I should be strong.’

76

Gary Holton

Both actor and undergoer independent pronouns may index inanimate referents. (192)

Kauwa,

bunne gang ma. haze 3SG.ACT come ‘No, [instead] haze came.’ [tonuburi053] NEG

Since independent pronouns are semantically aligned it is possible for a single clause to contain two undergoer pronouns indexing core arguments. This may occur when neither core argument is sufficiently controlling, as in (193). In this example the use of the first-person undergoer pronoun implies less control, less agency. Rather than physically taking the person outside, this example implies that other means of persuasion were used to cause the person to go outside. (193)

Naing gaing hoswang haggi. take 1SG.UND 3SG.UND outside ‘I took (coaxed) him outside.’

In contrast, were a first-person actor pronoun substituted the example would imply that the person was physically moved outside. On the other hand, it is not possible for two actor pronouns to occur with a single predicate. (194) *nang 1SG.ACT

gang 3SG.ACT

hoswang outside

haggi take

5.2 Person-marking prefixes The use of person-marking prefixes with noun roots was discussed in §4.7 in connection with nominal possession. WP verbs can be classified according to whether they (i) require an obligatory person-marking prefix; (ii) do not admit a person-marking prefix; or (iii) may optionally occur with a person-marking prefix. (For a more detailed classification see Holton 2010).

5.2.1 Verbs with obligatory person-marking prefixes A small class of verbs require a person-marking prefix. These verbs are always transitive and the prefix indexes the P or R argument.

Western Pantar

(195)

77

Verbs which require a person-marking prefix -nia ‘give’ -ussar ‘hunt’ -niaka ‘see’ -llang ‘search for’ -kkang ‘hit’

These verb stems cannot be uttered in isolation and in citation form occur with a third singular prefix ga-. With animate referents the person-marking prefix may occur with a co-referential undergoer independent pronoun. (196)

Gaing ga-niaka. 3SG.UND 3SG-see ‘He was seen.’

With inanimate referents the person-marking prefix may not occur with a co-referential undergoer pronoun. (197)

Nang 1SG.ACT

bla house

ga-niaka 3SG-see

‘I saw the house.’ *nang gaing ga-niaka ‘I saw it.’ A limited number of verbs which require person-marking prefixes may optionally occur with a second prefix indexing the actor. (198)

Ke’e pi-ga-ussar. fish 1INCL-3SG-catch ‘We are catching fish.’

But the actor prefix may also follow the undergoer prefix, as in the following examples. (199)

Wee ga-i-tiaring. fishpond 3SG-4PL-close ‘They are guarding the fishpond.’

78 (200)

Gary Holton

Nang 1SG.ACT ana day

ana now

ha-asang 2SG-say

dinni how.many

ta before

kauwa NEG

ing gob, if

ha-na-asang? 2SG-1SG-say

‘If I don’t tell you now, how long will it be before I can tell you?’ [adat1_050] 5.2.2 Verbs which do not admit person-marking prefixs Verbs which do not admit person-marking prefixes are primarily stative intransitive verbs. The single argument is indexed by an undergoer independent pronoun rather than a prefix.

(201)

Verbs which do gaing batta gaing kutta gaing maba gaing irpatang

not admit a person-marking prefix ‘he is stupid’ ‘he is fat’ ‘he is cold, well’ ‘he understands’

Holton (2010) reported a number of transitive verbs which also do not admit person-marking prefixes. However, this characterization is not strictly correct. While certain transitive verbs rarely admit person-marking prefixes, under certain conditions they may in fact occur with person-marking prefixes. For example, the verb lu’ung ‘cut’ typically occurs without a prefix, as in (202). (202)

Nang yattu lu’ung. cut 1SG.ACT tree ‘I cut the tree.’ *nang yattu ga-lu’ung

However, while lu’ung may not occur with a person-marking prefix indexing an undergoer, in (203) it occurs with a prefix indexing the actor. (203)

Hi ber tang na-lu’ung. 1SG-cut 2PL.POSS word on ‘I’m going to cut you off.’ [oikos035]

Thus, lu’ung might be better classified with those verbs which may optionally occur with a prefix, to be discussed below. However, lu’ung differs from those verbs in that its use with a prefix is extremely rare and does not occur in elicitation.

Western Pantar

79

5.2.3 Verbs which optionally occur with a person-marking prefix The majority of WP verbs may occur with or without a person-marking prefix. Verbs within this class differ as to which persons and which semantic roles can be indexed by the person-marking prefix when present. Some of these verbs only permit prefixes indexing first and second person referents; third person referents must be indexed with an independent pronoun. (204)

Na-olang ta. 1SG-fall.over IPFV ‘I’m about to fall over.’

(205)

Gang olang ta. 3SG.ACT fall.over IPFV ‘He’s about to fall over.’

(*ga-olang ta)

Other verbs in this class do allow prefixes to index third person referents, but only as the undergoer in a transitive construction (206), not as the single argument in an intransitive construction (207). (206)

Ga-pessing. 3SG-sneeze ‘Sneeze him (wipe his nose).’ Not: ‘He is about to sneeze.’

(207)

Na-pessing 1SG-sneeze ‘I’m about to sneeze.’ Not: ‘Sneeze me (wipe my nose)’.

Some transitive-only verbs also follow this pattern. With these verbs prefixes index first and second person actors and third person undergoers. (208)

Hatua hissa na-kasi. coconut fruit 1SG-split ‘I split the coconut.’

(209)

Nang ga-kasi. 1SG.ACT 3SG-split ‘I split it.’

80

Gary Holton

The examples above give only a hint of the extremely complex nature of the WP prefix system. A more complete description of verb classes and restrictions on prefixes can be found in Holton (2010).

6 Serial verb constructions A serial verb construction (SVC) expresses a single (but possibly complex) conceptual event with a combination of two or more lexical verbs, sharing at least one argument, under a single intonation contour. In the simplest examples the two elements of an SVC express individual components of a complex event. For example, the SVC tipping lama in (210) expresses the sequential events of getting up and then walking. A single shared argument is indexed by the third-person plural pronoun ging. (210)

Ging tipping lama. 3PL.ACT get.up walk ‘They got up and walked.’

A nominal or pronominal argument may intervene between the elements of the SVC indexing an additional argument. In (211) the third-person singular prefix ga- intervenes between the verbs ma and -nia. The latter is a bound verb which requires a person-marking prefix. The argument indexed by the firstperson pronoun nang is the actor argument of both verbs. (211)

Nang maggi ma ga-nia. 1SG.ACT banana come 3SG-give ‘I fetched him a banana.’

SVCs have several characteristics which distinguish them from either compounds or conjoined clauses. Unlike compounds, nominal and pronominal arguments may intervene between members of an SVC. That is, SVCs are separable. SVCs contrast with conjoined verb phrases in that: – the individual members of an SVC share aspectual and negative inflection – the members of an SVC must have one argument in common – the SVC falls under a single intonation contour – no conjoining particle may intervene between the members of the SVC In the remainder of this section SVCs are described according to their semantic properties.

Western Pantar

81

6.1 Symmetrical SVCs Symmetrical SVCs include two or more verbs from open classes. In WP symmetrical SVCs are employed to encode complex events, manner, and cause-effect.

6.1.1 Complex events Symmetrical SVCs may be used to describe complex events consisting of many individual sub-events in an iconic manner, following the temporal sequence of events. In this and the remainder of examples in this section SVC’s are highlighted in boldface. (212)

Yabbe gang hailang ti’ang. stand dog 3SG.ACT crawl ‘The dog crawled (out) and stood up.’ [frog071]

(213)

Ging dul tipping lama. walk 3PL.ACT immediately get.up ‘They immediately got up and left.’ [pubila269]

(214)

Wakke bogga gai aname ging aukung child young.man 3SG.POSS person 3PL.ACT enter ‘The young man’s people go in and ask.’ [marriage155]

banang. ask

Nominal arguments may intervene between sequences of verbs, as in (215). (215)

Pi-mising tua pi-ba’ai. 1INCL-sit tuak 1INCL-drink ‘Let’s sit down and drink tuak.’ [wedding2]

6.1.2 Manner Symmetrical SVCs may serve to elaborate the manner in which one of the verbs is carried out. In (216) the second verb sauke-yabe elaborates the first by indicating how the people are making noise. Equally, it could be said that the first verb elaborates the second by specifying that the dancing was done noisily.

82 (216)

Gary Holton

Habbang mau aname horang sauke-yabe. LEVEL person make.noise dance-lego.lego village ‘Over there in the village people are making noise dancing lego-lego.’ [tonu155]

In (217) the verb dalagar ‘gaze up’ specifies the manner in which the participant is lying down. (217)

Aname person

a-ttang 4SG-hand

a-eba 4SG-neck

lukku pillow

a-nne 4SG-face

dalagar gaze.up

ti’ang. lie.down

‘A person is lying down gazing up with his hands behind his head as a pillow.’ [P19] Verbs which take clausal complements can be expressed as complex SVCs. In (218) the meaning ‘order them to go tell’ is expressed as a complex event: he ordered them, then they went, then they told. (218)

As tang gi-hauwang wa kuba ga-asang. go grandmother 3SG-say also further 3PL-order ‘Then (he) ordered them to go tell grandmother.’

6.1.3 Cause-effect Symmetrical SVCs may specify causation, where the first verb indicates the cause and the second verb indicates the result, or effect. In (219) the slicing of the neck results in death, and in (220) the attack (literally, ‘hit’) results in disappearance. (219)

A-ule pai hinna kanna gaata. finish already 4SG-neck slice die ‘They sliced his neck and killed him.’ [tonuburi197]

(220)

Hing 2PL.ACT

ya road

hailing, crawl.up

ging 3PL.ACT

hi-kkang 2PL-hit

bawang. disappear

‘(If) you crawl (sneak) up the road, they will hit (attack) you till you disappear.’ [conflict094]

Western Pantar

83

6.2 Asymmetrical SVCs Asymmetrical SVCs consist of one verb from a small, closed class and one or more additional verbs from an unrestricted class. The denotation of the construction is provided by the unrestricted verb(s), while the restricted verb provides modificational specification (Aikhenvald & Dixon 2006).

6.2.1 Directional SVCs Directional SVCs combine a motion verb with one or more verbs from a closed class of eight directional verbs specifying path of motion using geophysical elevation. The system of directional verbs in Table 15 distinguishes three levels of elevation, as well as motion away from (TRANSLOCATIVE) or toward (CISLOCATIVE) the deictic center. In addition, the up and down translocative domains distinguish relative slope. Table 15: Directional verbs TRANSLOCATIVE

CISLOCATIVE

LEVEL

wa

go (not far, any direction)

ma

come (not far, any direction)

LOW

pia diakang

descend (gentle slope, westward) descend (steeper)

yang

return from above (east)

HIGH

mia raung

ascend (gentle slope, eastward) ascend (steeper)

middang

return from below (west)

The level terms wa and ma are used only for short distances. For longer distances up or down terms are used, recognizing that any long distance movement must involve a change in elevation. The direction verb occurs following the motion verb. The direction verb middang in (221) specifies the path of motion toward the deictic center and moving upward, while the direction verb ma in (222) specifies the path of motion toward the deictic center with no significant elevation change. (221)

Tubbe Pering gang golang middang. Tubbe Pering 3SG.ACT return up.CIS ‘Tubbe Pering came back home (upward)’ [pubila449]

(222)

Aname ging biring ma. person 3PL.ACT run come ‘The people came running.’ [bm138]

84

Gary Holton

Example (223) has two directional SVCs, each elaborating the motion verb kolang ‘roll’. The second SVC uses three directional verbs to indicate the path. (223) Kolang raung bila ara tang, kolang wa pia diakang. roll climb hill large on roll go down.TRANS down.TRANS ‘It rolled up onto the large hill, (then) rolled away down.’ [motion3.05] Directional verbs may also occur alone without a motion verb, in which case they indicate motion but without specifying the manner of motion. (224)

Hoang pi-pia ta. beach 1INCL-descend IPFV ‘Let’s go (descend) to the beach.’

(225)

Ga-r middang bla=m ta 3SG-with return.from.below house=LOC before ‘We’ll bring it back to the house and then drink.’

(226)

Mia gatta na-yang. 1SG-return.from.above ascend SEQ ‘I’ll go up then come back down’

gob thus

pi-ba’ai. 1INCL-drink

6.2.2 Aspect-marking SVCs A restricted class of verbs may be used in SVCs to indicate aspect. The verb kanna ‘be finished’ functions as a perfect marker, denoting a past event which is of current relevance. (227)

Nang hori kanna. finish 1SG.ACT eat ‘I have already eaten.’ / ‘I have finished eating.’

(228)

Gang mulai gi-ong kaulang ayang ma mising drift come sit 3SG.ACT begin 3PL-area call ‘He began to call them to come in and sit down.’ [pubila118]

kanna. finish

In this usage kanna contrasts with the perfective marker ga (see §7.1). (229)

Ir patta ga. condition dark PFV ‘It’s (already) dark.’

Western Pantar

(230)

85

Ir patta kanna. condition dark finish ‘It has gotten dark.’

The verb yadda ‘be not yet’ occurs in SVCs to mark incompletive aspect. (231)

(232)

Yadda sauke gayang be.not.yet dance NEG.IMP ‘Don’t dance yet.’ [bm073]

ha. DISC

Gi-masala saiga si yadda haulung i-t’iang. 3PL-problem PROX.SPEC.VIS ART be.not.yet hang PROG -lie.down ‘Their problem there is not yet being hung down (resolved)’ [conflict168]

The verb gatta/gaata is used in SVCs to mark completive aspect. It is often used as an explicit linker of sequential events, as (233), where gatta signals that the act of staying occurred prior to the act of running away. (233)

Mis gatta biring wa. stay already run go.TRANS ‘(They) stayed and then afterward they ran away.’ [darang092]

6.2.3 Auxiliary verbs Another type of asymmetrical SVC involves the use of a semantically bleached verb or postposition as an auxiliary or preverb. The set of verbal auxiliaries is a closed class of seven items: me, wang, kang, tang, ‑ong, -r, and ir. While some auxiliaries are homophonous with primary verbs (e.g., wang ‘exist’) or postpositions (e.g., tang ‘on top’), others do not occur outside this class. The function of these auxiliaries is not easily captured in a simple gloss. The auxiliary me is used in SVC’s which indicate a greater degree of affectedness. (234)

tubang asang kle’e using tukka

‘point’ ‘say’ ‘vomit’ ‘pick up’ ‘be short’

me tubang me asang me kle’e me using me tukka

‘point something out, indicate’ ‘tell’ ‘vomit on’ ‘carry on back’ ‘be not enough’

The auxiliary me may also function to create verbalizations from nominal forms.

86

Gary Holton

(235)

pammung ‘grave’

me pammung ‘to inter, bury’

The auxiliary wang forms middle constructions describing an extended event, as exemplified in the following verbal lexemes. (236)

birang golang galu pering pinni ranung garung tuppung

‘to ‘to ‘to ‘to ‘to ‘to ‘to ‘to

speak’ return’ chase’ pour’ hold’ catch falling’ release’ grow’

wang birang wang golang wang galu wang pering wang pinni wang ranung wang garung wang tuppung

‘to ‘to ‘to ‘to ‘to ‘to ‘to ‘to

tell a story’ reply, answer’ be expelled’ splash on’ fasten, to follow orders’ take aim’ imitate’ spill’

The auxiliary kang acts as a caustivizer deriving transitive verbs. (237)

diggi bina

‘to be bright’ ‘to be detached’

kang diggi kang bina

‘destroy’ ‘to wipe out, scatter’

The auxiliary tang often imparts a sense of repetition. (238)

saukang baulang ussing hillang pering pinni golang hori tasing

‘watch’ ‘fall over’ ‘to tie’ ‘to fly’ ‘to pour’ ‘to hold’ ‘to return’ ‘to eat’ ‘to fall down’

tang saukang tang baulang tang ussing tang hillang tang pering tang pinni tang golang tang hori tang tasing

‘watch out for’ ‘pounce’ ‘decide’ ‘to violate’ ‘to pour on’ ‘to grope’ ‘to repeat, add to’ ‘to feed’ ‘to fall on’

The auxiliary ‑ong refers to an area, or to a place being opened up. This auxiliary is usually inflected with a person-marking prefix, and it may inflect in two ways, intransitively or transitively. The prefix vowel a is deleted before ‑ong, while the prefix vowel i merges with the vowel of the auxiliary to yield a high back vowel.

Western Pantar

(239) assing tuppung baulang diggi ta’ai kaulang

‘remove, open’ ‘grow’ ‘fall over’ ‘to be clean, open’ ‘to cut’ ‘to call’

gong assing gong tuppung gong baulang gong diggi gong ta’ai gong kaulang

87

‘unravel, untie’ ‘extend’ ‘give birth’ ‘to clean an area entirely’ ‘to cut away, clear out an area’ ‘to call out to him/her’

Some examples of ‑ong conjugated for other person and number combinations are given below. (240)

Gang ai lamuli marung gaterannang PL all.of.them 3SG.ACT 3SG.POSS hero ‘He called all his heroes together.’ [pubila298]

(241)

N-ong na-ayanung. 1SG-AUX 1SG-be.first ‘I will go first.’

gi-ong 3PL-AUX

kaulang. call

The fourth person forms ong and ung may also function as reciprocal markers. (242)

Tiaku ong pinni. glass 4SG.AUX hold ‘Hold the glasses together.’

(243)

Ging ung baulung. 3PL.ACT 4PL.AUX fall ‘They gathered together.’

The auxiliary -r usually functions as an applicative adding an additional argument. It is obligatorily inflected with a person-marking prefix, most often the third singular ga-. (244)

ma wa middang pati palena natar

‘come’ ‘go’ ‘return from below’ ‘to pay’ ‘to teach’ ‘to stand’

gar ma gar wa gar middang gar pati gar palena gar natar

‘bring’ ‘take’ ‘bring back from below’ ‘to pay for something’ ‘to teach someone’ ‘to erect’

With some roots -r derives intransitive verbs from transitive ones.

88

Gary Holton

(245)

pinni ‘to hold P’

gar pinni ‘to continue, endure’

The auxiliary -r can also be inflected with other person-marking prefixes. (246)

Gaata gi-r mi raung Si Domma me. afterward 3PL-AUX ascend climb Si Domma LOC ‘Afterwards they were brought to Si Domma.’ [soli_hist]

The fourth person plural form ir derives verbs which refer to an abstract condition or place. (247)

kalalang patang lera war

‘to know’ ‘to be unable’ ‘to be light’ ‘to shine’

(248)

Gang ir patang. 3SG.ACT AUX not.able ‘He is stupid.’

ir kalalang ir patang ir lera ir war

‘to understand’ ‘to be stupid’ ‘to be daylight, daytime’ ‘to burn’

6.2.4 Dual constructions with -ilaku The verb -ilaku ‘be two’ (derived from the numeral alaku ‘two’) can be used with person-marking inflection to mark dual number for actor arguments.

(249)

Dual markers from alaku pilaku 1INCL ‘the two of us’ nilaku 1EXCL ‘the two of us (excluding addressee)’ ‘the two of you’ hilaku 2PL ‘the two them’ ilaku 3PL

The resulting dual verbs then function in serial verb constructions to mark dual reference. (250)

Pi-laku pi-lama ta. 1INCL-two 1INCL-go IPFV ‘Let’s (just) the two of us go.’

Western Pantar

89

In the case of first and second person arguments -ilaku must take the same person-marking prefix as the actor prefix on the main verb. However, the form ilaku is used with both third and fourth person arguments. (251)

Ging i-laku kap~kap birang. 3PL.ACT 4PL-two RDP~whisper talk ‘They two of them are whispering.’

6.2.5 Comitative constructions with -tu The verb ‑tu ‘be together with’ functions in serial verb constructions as a comitative to mark an event or state undertaken with the main verb. (252)

(253)

Pi-tu pi-lama 1INCL-together 1INCL-walk ‘Let’s go together.’

ta. IPFV

Na-dia kaweni ni-tu au ni-ga-llang. 1SG-go children 1EXCL-together grasshopper 1EXCL-3SG-search ‘I am going with my friends to look for grasshoppers.’

7 Aspect-marking In contrast to the languages of central and eastern Alor, Western Pantar lacks significant morphology expressing aspect. Tense and modality are also not morphologically marked. Lexical aspect may be optionally indicated via adverbs. The adverb gar in (254) expresses a continuous action. (254)

Mantri dau gang gar ti’ang tulis. write official LEVEL.NSPEC 3SG.ACT continue stay ‘The official up there just kept on writing.’ [publia121]

Situational aspect may be expressed via aspectual particles (§7.1) or aspectual prefixes (§7.2).

7.1 Aspectual particles Aspect is most commonly marked via the particles ta and ga, marking imperfective and perfective, respectively. These particles follow the predicate. Imperfective

90

Gary Holton

ta almost never occurs with predicates whose referent is indexed only via a full pronoun. Similarly, perfective ga almost never occurs with predicates whose referent is indexed via a person-marking prefix. (255)

Na-lama ta. 1SG-walk IPFV ‘I’m going.’

(256)

Gang lama 1SG.ACT walk ‘He left.’

(? nang lama ta)

ga.

(? ga-lama ga)

PFV

Imperfective ta often signals intention. Thus, in (257) there is an intention to build fences, and in (258) there is an intention to check something out, have a look. (257)

Ping 1SG.ACT

dalla tomorrow

pi 1INCL.POSS

badde fence

siga=b there=REL using raise

lama walk

dia go

kalung here.and.there

maum LEVEL. SPEC. NVIS

ta. IPFV

‘Tomorrow we will go there and put up our fences here and there.’ [pubila188] (258)

Na-ga-llang na-ga-niaka ta. 1SG-3SG-look.for 1SG-3SG-see IPFV ‘I will take a look for him.’ [bm032]

The imperfective is also often used in imperative constructions, as in the examples below. (259)

Tigung ta. add IPFV ‘Have some more (food).’

(260) Yasa marung sing ga-r diakang tang gaddi a-akku ta. bad PL ART 3SG-with descend on make INCP-good IPFV ‘Bring down your bad ones [weapons] and fix them up.’ [pubila153] As the examples above suggest, the imperfective marker has a strong functional association with the marking of mood, specifically irrealis mood. A more

Western Pantar

91

complete description of the use of aspectual particles to mark mood would require a detailed examination of discourse data which is beyond the scope of this sketch.

7.2 Aspectual prefixes Two aspectual-marking prefixes a- and i- may occur with verbs, marking inceptive and progressive aspect, respectively. These prefixes are semantically and grammatically complementary, and hence they do not co-occur. Both aspectual prefixes a- and i- are optional. The location of the aspectual prefix depends on the morphological status of the verb root. With bound roots the aspectual prefix precedes the obligatory person-marking prefix, as shown in (261).

(261)

Position of aspectual prefix with bound roots (PRON2 ) ASP PRON1 ROOT

With free roots the aspectual prefix precedes the root and follows the personmarking prefix, if present.

(262)

Position of aspectual prefix with free roots (PRON) ASP ROOT

An example of the progressive aspect prefix with a bound root is given in (263). (263)

John na-i-ga-niaka. John 1SG-PROG-3SG-see ‘I am seeing / have been watching John.’

7.2.1 Inceptive prefix aThe verb prefix a- marks inceptive aspect. Its interpretation is dependent on the implied tense. With past situations this prefix generally marks events which have just begun, just started to happen. With non-past situations this prefix marks events which are about to occur. (264)

Tame a-dia ga’ai? where INCP-go 3SG.GEN ‘Where did she (just) go?’

92 (265)

Gary Holton

A-kanna

ga?

INCP- finish

PFV

‘[Are you] (about to be) finished?’ (266)

Wakke bogga ye kalau eu ye child young.man one if woman one ‘If a young man is to marry a girl,’ [marriage006]

(267)

Aname tawagang mising hala=b a-bloppa ta=b. person middle sit FOC = SEQ INCP-shoot IPFV= SEQ ‘It’s the people sitting in the middle who will start to shoot first.’ [pubila196]

a-haggi INCP-take

ga’ai. 3SG.GEN

The inceptive prefix can also have an inchoative reading, indicating a coming into being. This is the case in the following textual example, in which the character Imam Blegur is introduced. (268)

a.

Imam Blegur Imam Blegur ‘Imam Blegur,’

b.

aname Lamma tapi gang pau=m asa Bara=m Bara=LOC person Lamma but 3SG.ACT LOW=LOC so ‘was from the Lamma clan but he stayed down at Baranusa.’

c.

As Islam ya dia. so Islam toward go ‘So he entered Islam.’

d.

Latuna wala. Latuna person ‘He came from Latuna.’

e.

Raja chief ‘[He was] a chief,’

f.

tapi pau=m Bara=m but LOW=LOC Bara=LOC ‘but he lived in Baranusa.’

mising. sit

mising. sit

Western Pantar

g.

Bara=m mising Bara=LOC sit ‘Living in Baranusa,’

h.

Bara eu haggi as Islam ya a-dia. Bara woman take so Islam toward INCP-go ‘[He] took a Baranusa wife so entered Islam (became a Muslim).’ [conflict061]

93

In (268c) the phrase Islam ya dia occurs meaning ‘enter Islam, go toward Islam’. This same phrase is repeated later in (268h) but with an inceptive prefix emphasizing the inchoative sense.

7.2.2 Progressive prefix iThe verb prefix i- marks progressive aspect. (269)

Depang wang i-dekang si kauwa ging ga-kkang ga. 3PL.ACT 3SG-hit PFV ladder exist PROG-descend go NEG ‘Before he could descend the ladder they assaulted him.’ [tonuburi041]

(270)

N-iu ang me i-golang. 1SG.POSS-mother market LOC PROG-return ‘My mother is returning from the market.’

The progressive prefix is homophonous with the fourth-person plural prefix i-. However, unlike the fourth-person prefix, the progressive prefix triggers gemination of stem-initial consonants of short verb stems. The progressive prefix may co-occur with person-marking prefixes, in which case it follows the personmarking prefix and triggers deletion of the prefix vowel. In particular, with consonant-initial verb stems which lack the progressive prefix there is a contrast between first singular and first exclusive plural person-marking prefixes, as in nama ‘I come’ versus nima ‘we (excl.) come’. However, when the progressive prefix is present this contrast is neutralized, as in nimma ‘I/we (excl.) have come’. Nevertheless, the presence of the progressive prefix is clear from the resulting gemination. (271)

Ni-ar ni-hauwang n-i-mma ganung 1EXCL-father 1EXCL-order 1EXCL-PROG-come just ‘Our father is ordering us to come here.’ [tonu036]

saiga. PROX.VIS

94

Gary Holton

The progressive prefix may have a modal evidential function, asserting the truth of an utterance in spite of a lack of direct evidence, as in (272) and (273). (272)

N-iu n-iaku i-ga-aulang. 1SG.POSS-mother 1SG.POSS-sibling PROG-3SG-bathe ‘My mother is bathing my brother [but I can’t see it].’

(273)

N-i-mising. 1SG-PROG-sit ‘I am sitting [but you can’t see me].’

The preceding examples also demonstrate the contrast in the position of the progressive prefix. With bound transitive roots such as -aulang ‘bathe’ the progressive prefixes precedes the person-marking prefix indexing the less agentlike argument. With free roots (transitive or intransitive) such as mising ‘sit’ the progressive prefix follows the person-marking prefix. In serial verb constructions only the final verb may be marked with an aspectual prefix. This is exemplified in the following excerpt were the SVC wena me ‘get ready’ is repeated in the second intonation unit with an additional verb, thus wena me lamala ‘get ready completely’. The location of the inceptive prefix shifts from me to lamala, accordingly. (274)

Dul immediately hi-wena 2PL-prepare

hi-wena 2PL-prepare me LOC

a-me, INCP-LOC

a-lamala. INCP- do.completely

‘All of you get ready immediately, get ready completely.’ [pubila305]

8 Discussion I summarize here some of the features of Western Pantar which distinguish it from other languages of the Alor-Pantar family. While the phoneme inventory is typical of AP languages, the presence of a complete set of phonemic geminates (§2.2) is unique. Geminates are present sporadically in other languages of the family but do not play such a fundamental role in the phonology. Another unique aspect of WP phonology is the retention of the initial glottal fricative, though weakly articulated (§2.1.6).

Western Pantar

95

In comparison to the languages of eastern Alor, WP is striking for its almost complete lack of morphology. Nouns may be inflected with a proclitic and prefix indexing alienable and inalienable possession, respectively (§4.7), while verbs may be inflected with a single paradigm of person-marking prefixes (§5.2) and one of two aspectual prefixes (§7.2). However, this single paradigm of personmarking prefixes interacts with the pronominal system to create an extremely complex system of grammatical relations not found in other AP languages. Independent pronouns are semantically-aligned, patterning according to an agentpatient system (§5.1), while the indexing of arguments on verbs is at least in part governed by lexical semantics (§5.2). This results in a situation in which a single verb prefix may index any of the A, S, or P macro-roles; not just the P role (or S and P roles), as is the case for most other Alor-Pantar languages. Related to this lack of syntactic alignment is a lack of valency-changing morphology. Finally, though not discussed in detail here, WP has a rich system of elevationbased spatial deixis reflected in several different domains of the grammar, including demonstratives (§4.5), locational nouns (§4.8), directional verbs (§6.2.1), and deictic adverbs (§3.6). The wide distribution of this elevation semplate across distinct syntactic domains, and its pervasive use in everyday discourse, clearly reflects the prominence of elevation in WP grammar (Holton 2011).

References Aikhenvald, Alexandra Y. & R.M.W. Dixon (eds). 2006. Serial Verb Constructions: A crosslinguistic typology. Oxford: Oxford University Press. Burenhult, N. 2008. Streams of words: Hydrological lexicon in Jahai. Language Sciences 30(2–3): 182–99. Grinevald, Colette. 2007. The linguistic categorization of spatial entities: Classifiers and other nominal classification systems. In Michel Arnague, Maya Hickmann & Laure Vieu (eds), The Categorization of Spatial Entities in Language and Cognition: 93–121. Amsterdam: John Benjamins. Holton, Gary. 2010. Person-marking, verb classes, and the notion of grammatical alignment in Western Pantar (Lamma). In Michael Ewing & Marian Klamer (eds), Typological and Areal Analyses: Contributions from East Nusantara: 101–121. Canberra: Pacific Linguistics. Holton, Gary. 2011. Landscape in Western Pantar, a Papuan outlier of southern Indonesia. In David M. Mark, Andrew G. Turk, Niclas Burenhult & David Stea (eds), Landscape in Language (Culture and Language Use: Studies in Anthropological Linguistics), 143–166. Amsterdam: John Benjamins. Levinson, Stephen C. 2003. Space in Language and Cognition. Cambridge: Cambridge University Press. Mithun, Marianne. 1991. Active/agentive case marking and its motivations. Language 67(3): 510–546.

Marian Klamer

3 Kaera 1

The language scene

98

2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.2 2.3 2.4 2.5 2.6 2.7 2.7.1

99 Phonology 99 Consonants 100 Plosives 100 Fricatives 101 Nasals 101 Liquids and approximants 102 Consonant phonotactics 103 Vowels 104 Vowel phonotactics 105 Structure of syllables and roots 106 Stress Interaction between phonology and morphology 107 Person-prefix allomorphy

3 3.1 3.2 3.3 3.4 3.5 3.6

109 Basic clausal syntax Constituent order in verbal clauses 112 Equational clauses 112 Existential clauses 113 Negation 115 Questions 117 Postpositional phrases

4 4.1 4.2 4.3 4.3.1 4.3.2 4.3.3

119 Noun phrases 120 Nominal attributes Number, quantification and enumeration 125 Possession 126 Alienable possession 126 Inalienable possession 127 Possession and nominalization

5 5.1 5.2 5.2.1 5.2.2 5.3 5.4

127 Pronouns and person prefixes 128 Pronouns to encode A, S, P and Possessor 129 The encoding of P 129 Transitive verbs and the encoding of P 133 Labile verbs 134 Encoding of S 136 Possessor pronouns

109

122

107

98

Marian Klamer

6 6.1 6.2 6.3 6.4 6.5

Serial verb constructions and related predicates 137 137 Serial verb constructions (SVCs) 137 SVCs with wang ‘be, exist’ 139 SVCs with pin ‘hold’ and met ‘take’ 139 Causative SVCs with er ‘make, do’ 139 Grammaticalized SVCs with the verb –(e)ng ‘give’

7 7.1 7.2

140 Encoding clausal position and aspect of verbs Distribution of verbal suffixes marking clausal position and 141 aspect 143 Semantics of aspectual suffixes and adverbs

8

Discussion

References

145

145

1 The language scene Kaera (Bahasa Kaera in Indonesian) is spoken by an estimated 5,500 people on the northeastern coast of Pantar island.1 It has not been previously reported as a separate language, and hence has no ISO 639-3 code. Kaera speakers currently live in the following villages (the lexeme abang means ‘village’): Abang Iwang (~1,500), Padangsul and Bibit Gomi (~2000), Matgomi/Weniwa (~50), and Tamal Abang (~2000). In earlier times, Kaera speakers lived in Kaera Abang, Labi Abang, Xraner Abang, Silorang Abang, Uyumau Abang, and Tamai Abang (today’s Tamal Abang). The logonym Kaera is based on the ancestor village Kaera (originally named Xai Er Abang ‘Village at the Xai Er tree’). An alternative logonym is Dorit, after the name of the biggest clan that speaks the language.2 The Kaera speaking region borders with Blagar to the north and south, and with Teiwa to the west. Kaera is lexically more similar to Teiwa than to Blagar. Together with Teiwa, Nedebang and Western Pantar, Kaera belongs to the Pantar subgroup of AlorPantar subgroup (Holton and Robinson forthcoming, Holton et al. 2012, Schapper and Klamer forthcoming). 1 I am grateful to Marianus Waang for his help in the collection and analysis of the materials presented here, and for his hospitality when I visited them in their home in Kampen, The Netherlands. This chapter has benefited from feedback by the editor and two anonymous reviewers. Funding of the research reported here was provided by the Netherlands Foundation for Scientific Research (NWO), through the Vernieuwingsimpuls-project “Language variation in eastern Indonesia” (2002–2007). 2 Other Kaera speaking clans include the clans of Labi, Marggang, Uyumau and Malabo.

Kaera

99

This description is based on primary data collected by the author, working with Marianus Waang, a native and fluent speaker of Kaera.3 The data on which this sketch is based include: (ii) A lexicon of ~900 items, collected between 2005–2007; (ii) Three short narratives: a Frog Story (47 utterances); a Pear Story (32 utterances); and a narrative about palm wine (50 utterances);4 (iii) 450 elicited sentences/utterances, collected between 2005–2007. After the first draft of this sketch was written, we met in Bali in July 2012 and were able to collaborate for several more days, primarily double-checking the language data presented here, and supplementing information relating to the verb classes discussed in §5. Kaera speakers are fluent in Indonesian, but the language appeared to be relatively vigorous when I visited Abangiwang in 2007. Many of the children playing outside were using Kaera amongst themselves, and Kaera and Indonesian are both used in church, mostly depending on who is conducting the service: if it is a Kaera speaker, most of the service will be in Kaera; if it is a speaker from another language, it will be in Indonesian. Kaera is not allowed in school, Indonesian being the language of education.

2 Phonology 2.1 Consonants Kaera has 16 consonants, as laid out in Table 1. Orthographical representations used later in this sketch which differ from IPA are given in angled brackets. The

3 As this sketch is mainly based on work with a single speaker, it is relevant to include some basic sociolinguistic data on that speaker. Born in Matgomi (1970), Waang grew up in a Kaeraspeaking environment in Abang Iwang. From 1986–1989, he lived in Kalabahi (Alor), in an environment where Indonesian, Kaera, and Warsalelang (a dialect of Blagar) were spoken. In 1989, he left Kalabahi to study English in Kupang, and since then has not been living amongst Kaera speakers. In the Netherlands, Waang acquired a working knowledge of Dutch, Latin and Greek. From Sept 2005–Aug 2007 he stayed in The Netherlands to study Theology. During these years, we met nine times to work on Kaera for about half a day. In May 2007, I visited his home town Abang Iwang for about a week, when I was in Pantar for fieldwork on Teiwa (Klamer 2010). I recorded ~12 hours of video-recordings of narrative texts, sermons, conversations and songs, but as most of these materials still need to be annotated and analyzed, they have not been used for this sketch. 4 Metadata Frog story: speaker M. Waang, recorded in Kampen, The Netherlands, Nov 2005; Pear story: speaker M. Waang, recorded in Leiden, The Netherlands, April 2006; Palmwine narrative: speaker A. Puling, recorded in Abangiwang, Pantar, May 2007. All texts were transcribed and annotated in collaboration with M. Waang in The Netherlands.

100

Marian Klamer

brackets around the glottal fricative indicate its marginal status (see §2.1.2). The subsections below provide minimal pairs per consonant class. Table 1: Kaera consonants ALVEOLAR

BILABIAL PLOSIVES FRICATIVES NASALS LIQUIDS

p

APPROXIMANTS

w

b

t s

m

PALATAL d

VELAR k x

GLOTTAL g (h) ŋ

n r l j

The size of the Kaera consonant inventory is fairly typical for the languages of the Alor-Pantar group. Kaera differs from neighboring Teiwa, which has a larger inventory of 20 consonants (Klamer 2010). Atypical for the Alor-Pantar group is the lack of a phonemic glottal stop in Kaera. 2.1.1 Plosives Minimal pairs for plosives are: (1)

/p, b/ /t, d/ /k, g/

/pin-/ /bin-/ /ta/ /da/ /kiri/ /giri/

‘hold’ ‘not’ ‘on’ ‘go up’ ‘bone’ ‘companion(s), other(s)’

2.1.2 Fricatives Of the fricatives, the velar fricative /x/ is a reflex of the proto-Alor Pantar uvular stop *q, which is still found in Teiwa, and became /k/, /g/, /ʔ/ or zero in Western Pantar, Blagar and Adang (Holton et al. 2012). The glottal fricative /h/ has a very minor phonemic status in the language. In my corpus it is attested in only a few words: the exclamations /hɛ/ and /ɛh/ ‘hey!’, onomatopoeic words like /hɔŋ/ ‘woof’, and /hiŋɔ/ which expresses a negative response: (2)

/s, x, h/

/sib/ /xib/ /hiŋɔ/

‘new’ ‘goat’ ‘no, not like that’ (negative response)

Kaera

101

2.1.3 Nasals In initial position, /m/ and /n/ are contrastive. In final position, /m/, /n/ and /ŋ/ are contrastive. (3)

/m, n/ /m, n, ŋ/

/ma/ /na/ /tam/ /pan/ /-taŋ/

‘come’; ‘house’ ‘consume, eat’; ‘thing’ ‘sea’ ‘kemiri (tree/nut)’ (Aleurites moluccana) ‘hand’

There are only a few verbs ending in /n/, examples include pin ‘take’ and nimin ‘die’. Such verbs can be inflected: pin-o ‘take- FIN’, nimin-o ‘die-FIN’. Many verbs end in a final /ŋ/. Most of these verbs cannot take inflectional suffixes. Examples of such verbs are given in Table 18, §5.2.2., where it is suggested that the final /ŋ/ on verbs may be a (fossilized) suffix. Some verbs ending in /ŋ/ can take suffixes, and then the velar nasal changes to [n]. An illustration is the verb –(e)ng ‘give’. When this verb is not suffixed, it ends in /ŋ/, when it is suffixed, /n/ is used, as shown in (4)–(5). (4) Egu met mi boom~boom g-eng gang na. that.one.there take LOC RDP~old.respected.man 3SG-give 3SG eat/drink ‘Take that to the elders to drink.’ (5) Etang xar ut gang ge-saring g-en-o. and ten four 3SG 3SG.POSS-birth.party 3SG-give-FIN ‘And [on] the 40th [day] she gave him5 his birth party.’ 2.1.4 Liquids and approximants Minimal pairs of liquids and approximants are given in (6): (6)

/r, l/ /j, w/

/riŋ/ /liŋ/ /ja/ /wa/

‘wait for someone’ ‘grow’ ‘go down’ ‘leaf’

5 For the third person singulars translated as male forms one can also read the female forms.

102

Marian Klamer

In root- or word-final position, high vowels that follow a vowel are realized as glides. Verbal suffixes like –o /-ɔ/ ‘FIN’ (which encodes the clause-final position of verbs, see §7) only attach to consonant-final verbal roots, as shown in Table 2.6 Roots ending in a glide pattern with the other consonant-final roots by taking a vowel-suffix. Table 2: Verbal bases taking inflectional suffix –o ‘FIN’ 7 C-final verbs taking suffix

V-final verbs not taking suffix

‘buy-FIN’ ‘call-FIN’ ‘hang-FIN’ ‘feed-FIN’ ‘cry-FIN’ ‘sit-FIN’ ‘peel-FIN’ ‘stream out-FIN’ ‘bathe-FIN’ ‘twist-FIN’

/wal-ɔ/ /-pɛk-ɔ/ /ɔr-ɔ/ /-aːs-ɔ/ /bag-ɔ/ /mis-ɔ/ /xɛw-ɔ/ /saw-ɔ/ /-wej-ɔ/ /-jaj-ɔ/

‘eat, drink, consume’ ‘lie down, sleep’ ‘go (from speaker)’ ‘come to deictic center’ ‘go up to deictic center’ ‘go down to deictic center’ ‘pull’ ‘scratch, peel’ ‘carry’

*na-ɔ8 *tɛː-ɔ *wa-ɔ *ma-ɔ *da-ɔ *ja-ɔ *kɛri-ɔ *tuː-ɔ *wrɛ-ɔ

/na/ /tɛː/ /wa/ /ma/ /da/ /ja/ /kɛri/ /tuː/ /wrɛ/

Root-final consonants become the onset glide of any syllables created by suffixes; for example, /xɛw-ɔ/ ‘peel-FIN’ is syllabified as [xɛ.wɔ].

2.2 Consonant phonotactics The distribution of Kaera consonants across syllables is summarized in Table 3. For notes on the distribution of /h/ see §2.1.2; the distribution of the nasals is discussed in §2.1.3. Table 3: Kaera consonant positions in words and roots POSITION

p

b

t

d

k

g

s

x

h

m

n

ŋ

r

l

w

j

#_V

(

(

(

(

(

(

(

(

(()

(

(

–

(

(

(

(

V_V

(

(

(

(

(

(

(

(

–

(

(

–

(

(

(

(

V_#

(

(

(

(

(

(

(

(

(()

(

(()

(

(

(

(

(

6 Roots ending in a velar nasal are the exception to this rule, see §2.1.3 and §5.2.2. 7 Hyphens on verbs signal their bound root status, i.e. that they take an obligatory inflectional prefix or suffix. 8 The consultant notes that these verbs can combine with the (phonetically identical) hortative particle o /ɔ/, as in Na o! ‘Let’s eat!’.

Kaera

103

Kaera has restrictions on consonant clusters. Within a morpheme, sequences of consonants are only allowed if they are word-initial, and involve a bilabial consonant followed by a sonorant /r, l/, as illustrated in (7a), or a voiceless stop followed by a sonorant, as in (7b). A voiceless velar stop can combine with /r, l/, a voiceless alveolar stop only combines with /r/. Other mono-morphemic consonant clusters are disallowed. (7)

a.

/plɔgaŋ/ /prasi/ /bleliŋ/ /bram/ /mrak/ /wrɛ/

‘bunch’ ‘place to store corn’ ‘open’ ‘dust’ ‘sweet’ ‘carry’

b.

/kliki/ /krabis/ /trɛ/

‘sour’ ‘(to) claw’ ‘shake sth. out of sth. else’

2.3 Vowels Kaera has 5 vowels, all of which have a phonemically long counterpart, as represented in Table 4. In section 3 onwards, the long vowels are orthographically represented as double vowels; the mid vowels as /e, o/. Table 4: Kaera vowels FRONT HIGH MID LOW

CENTRAL

i, iː ɛ, ɛː

BACK u, uː ɔ, ɔː

a, aː

(Near-)minimal pairs showing the vowel length contrast are given in (8). Not many of these pairs exist in my corpus. This may be due to the limited size of the corpus, or suggest that phonemic vowel length does not play a major role in Kaera phonology. The fact alone that length is distinctive for all 5 vowels in Kaera is rather unusual in the Pantar context: Teiwa has a length distinction for only 3 vowels, and Blagar, Western Pantar and Adang have no phonemic vowel length at all.

104

Marian Klamer

(8)

/a, aː/

/u, uː/ /i, iː/ /ɛ, ɛː/ /ɔ, ɔː/

/war/ /bak-/ /wat/ /bum/ /tub/ /kir/ /tɛ/ /tɔr/

‘rock’ ‘cut thin’ ‘coconut’ ‘flower’ ‘burn (tr)’ ‘hair comb’ ‘grass, weeds’ ‘main road’

/waːr/ /baːk/ /-waːt /duːm/ /uːb/ /biːr/ /tɛː/ /xɔːr/

‘hot’ ‘long’ ‘live’ ‘bird’ ‘sugarcane’ ‘flat’ ‘sleep’ ‘women’s betelnut and betelvine container’

2.4 Vowel phonotactics Combinations of vowels occur both within a single morpheme, and across morpheme boundaries. The morpheme-internal combinations attested in my database are presented in Table 5. The sequences not attested within morphemes are indicated by the double dash ‘–’. Table 5: Kaera vowel sequences within morphemes V1/V2

a

a

i

u

ɔ

ɛ

30

12

–

–

2

1

2

–

2

i

11

u

10

10

ɔ

–

10

1

ɛ

–

10

1

7 –

Illustrations of words with vowel sequences are given in Table 6, along with their phonetic realizations, which may vary in slow and fast speech as indicated. Some distributional restrictions seem to apply on long vowels, which are mainly found in monosyllabic nouns or verb roots (CV:C, V;C, CV:) (26 of the 29 words/roots with long vowels in my lexicon are monosyllabic), while sequences of vowels are found in both disyllabic and monosyllabic forms. At this stage, I have insufficient data to allow an analysis of the phonology of long vowels in comparison to vowel combinations.

Kaera

105

Table 6: Examples of vowel sequences within morphemes9 Initial vowel Word

Meaning

Surface realization

/a/

/tamai/ /tau/ *aɔ *aɛ /siax/ /miu/ /tiɔkuŋ/

‘tamarind’ ‘bean’

[ta.ˈmaj] [ˈtaw]

‘chicken’ ‘owl’ ‘shake’

/kiɛb/ /nuaŋ/ /xubui/ *uɔ /nuɛlum/ *ɔa /bɔi/ /tɔu/ /bɔɛ/ *ɛa /bɛi/ /xɛu/ *ɛɔ

‘wall’ ‘cloth’ ‘meat’

[si.ˈax] [mi.ˈu] [tjɔ.ˈkuŋ] (fast speech), [ti. ɔ.ˈkuŋ] (slow speech) [ˈkjɛb] [ˈnwaŋ] (fast sp.), [nu.ˈwaŋ] (slow sp.) [xu.ˈbuj]

‘long’

[nwɛ.ˈlum] (fast sp.), [nu.ɛ.ˈlum] (slow sp.)

/i/

/u/

/ɔ/

/ɛ/

‘river’ [ˈbɔj] ‘mug from coconut shell’ [ˈtɔw] ‘maybe’ [bɔ.ˈɛ], *[bɔj] ‘pig’ ‘wind; cold’

[ˈbɛj] [ˈxɛw]

2.5 Structure of syllables and roots The minimal Kaera syllable, and word, consists of a single vowel (V), the maximal syllable is CVVC. Most nominal or verbal roots consist of one or two syllables see (9) and (10). While most tri-syllabic words are poly-morphemic, there are some that are morphologically simple, including /ɔjɔgi/ ‘green, blue’ /bagari/ ‘yellow’, /amara/ ‘many’, /biɛbuŋ/ ‘last night’, and /ugugi/ ‘recently’. Morphologically complex trisyllabic words include inflected verbs, possessed nouns, and compounds like /n-uar sil/ ‘1SG.POSS-ear lobe’ or /kɛi g-om/ ‘rainy season’ < ‘rainy.season 3SG.POSS-inside’.

9 Primary stress is represented as before the stressed syllable, secondary stress as .

106

(9)

(10)

Marian Klamer

Mono-syllabic words V /i/ CV /si/ VV /ɛi/ CVV /sɛi/ V: [not attested] CV: /tɛː/ VC /ur/ CVC /bis/ VVC /auk/ CVVC /rauŋ/ V:C /uːm/ CV:C /duːm/ CCV /wrɛ/ CCVC /bram/

‘sick’ ‘spoon’ ‘canoe’ COMPLETIVE

‘sleep’ ‘moon’; ‘grass field’ ‘mat’ ‘wooden plate’ ‘water melon’ ‘worm’ ‘bird’ ‘carry’ ‘dust’

Di-syllabic words V.CV /aki/ CV.CV /buku/ CV.V /bɔɛ/ V.CVC /ibar/ CV.CVC /banab/ CV.VC /biaŋ/ CV.CVV /tɛnɛi/ CCV.CVV /blibai/ CCV.CVC /blɛliŋ g-om/

[a.ˈki] [bu.ˈku] [bɔ.ˈɛ] [i.ˈbar] [ba.ˈnab] [bi.ˈaŋ] [tɛ.ˈnɛj] [bli.ˈbaj] [blɛ.ˈliŋ ˌg-om]

‘root’ ‘mountain top’ ‘maybe’ ‘dog’ ‘foggy, cloudy’ ‘sleepy’ ‘right’ (opposite of ‘left’) ‘hungry’ ‘outside’ (lit. ‘open 3SG.POSS-inside’)

2.6 Stress Stress is not phonemic in Kaera. The domain of Kaera stress assignment is the root morpheme. Stress is on the final syllable of the root, as illustrated in (11), and adding prefixes or suffixes to a root does not alter the stress pattern. Stressed syllables may be open or closed, and light or heavy. (11)

/xabi/ /isar/ /kɛlikil/

[xaˈbi] [iˈsar] [keliˈkil]

‘spear’ ‘salt’ ‘sour’

Kaera

107

The stress pattern of a root with a possessive prefix is illustrated in (12). When a root is suffixed, stress remains on the ultimate root syllable and does not shift to the final syllable of the word. This is illustrated for a monosyllabic root in (13), and a disyllabic root in (14). (12)

/gɔ-tɔxi/

[go.tɔ.ˈxi]

‘3SG.POSS-egg’

(13)

/jas-ɔ/

[ˈja.sɔ]

‘bad-FIN’

*[jaˈsɔ]

(14)

/patak-ɔ/

[pa.ˈta.kɔ]

‘cut-FIN’

*[pa.ta.ˈkɔ] (see (141))

In noun collocations, primary (=word) stress is on the first lexeme, and secondary stress on the second lexeme, as illustrated in (15): (15)

/g-ɔŋ kul/ /g-ɔŋ wa/

[ˈgɔŋ ˌkul] [ˈgɔŋ ˌwa]

‘3SG.POSS-head shell’ ‘3SG.POSS-head leaf’

‘his skullead’ ‘his hair’

2.7 Interaction between phonology and morphology Various interactions exist between the phonological shape of words and roots, and processes of affixation. First, the shape of the initial syllable of a noun/ verb determines the shape of the prefix that encodes the nominal possessor or the verbal patient. A consonant-initial noun/verb selects a syllabic (CV-) prefix, a vowel-initial noun/verb selects a consonantal (C-) prefix. (Verbal prefixes are presented in Table 7 in §2.7.1, nominal prefixes in Table 14 in §4.3.) Second, the shape of the final syllable of a verb root determines its suffixing potential: except for verbal bases ending in –ng, all consonant-final verbs can host the inflectional suffixes that mark phrasal position and aspect. Verb roots ending in a vowel cannot be such hosts (see §4.1, Table 12; and §5.2.1, Table 17). Third, the class of prefixing transitive verbs encodes their patient-like (P) argument with a prefix that shows vowel harmony with the verbal root vowel, as discussed in the following section. 2.7.1 Person-prefix allomorphy Patient-like arguments in transitive clauses may be marked on the verb with a prefix encoding person and number. While the shape of the prefixes encoding plural Ps is invariable, prefixes encoding singular Ps depend on the phonological shape of the verb. As laid out in Table 7, C(onsonant)-initial verbs take a syllabic (CV) prefix (I or II), while V(owel)-inital verbs take a consonantal prefix (III).

108

Marian Klamer

The syllabic prefixes occur in two shapes: one where the vowel (V) harmonizes with the initial vowel of the verb (prefix I), and one whose vowel is /a/, independent of the vowel in the verb (prefix II).10 Most C-initial verbs take the harmonizing prefix I; prefix II is found on only a few verbs. Illustrations of the harmonic prefix I are given in Table 8, of the a-prefix II in Table 9, and of the consonantal prefix III in Table 10. Paradigm II with invariant vowel /a/ reflects the person prefixes reconstructed for proto-Alor Pantar (Holton et.al. 2012). Note that there are no differences between paradigms in the plural. Table 7: Person prefixes encoding P C-initial verb Syllabic prefix

V-initial verb Consonantal prefix

I

II

1SG

/nV-/

/na-/

III /n-/

2SG

/V-/

/a-/

/Ø-/

3SG

/gV-/

/ga-/

/g-/

1PL.EXCL

/ni-/

1PL.INCL

/pi-/

2PL

/i-/

3PL

/gi-/

Table 8: Harmonic syllabic prefix I

1SG 2SG 3SG

1SG 2SG 3SG

Ci- ‘look after’

Cu- ‘kiss’

Ca- ‘leave, let loose’

/ni- riaŋ/ /i- riaŋ/ /gi- riaŋ/

/nu- muni/ /u- muni/ /gu- muni/

/na- dag/ /a- dag/ /ga- dag/

Cɛ- ‘expel-FIN’

Cɔ- ‘push’

/nɛ- tɛr-ɔ/ /ɛ- tɛr-ɔ/ /gɛ- tɛr-ɔ/

/nɔ- tɔbuŋ/ /ɔ- tɔbuŋ/ /gɔ- tɔbuŋ/

Table 9: Syllabic A-prefix II Ca- ‘talk to’ 1SG 2SG 3SG

/na- tutuk/ /a- tutuk/ /ga- tutuk/

Ca- ‘quarrel with’ */nu- tutuk/ */u- tutuk/ */gu- tutuk/

/na- juraŋ/ /a- juraŋ/ /ga- juraŋ/

10 The prefix vowel /a/ may be reduced to a central vowel in running speech.

*/nu- juraŋ/ */u- juraŋ/ */gu- juraŋ/

Kaera

109

Table 10: Consonantal prefix III

1SG 2SG 3SG

C- ‘feed-FIN’

C- ‘order’

/n- as-ɔ/ /Ø- as-ɔ/ /g- as-ɔ/

/n- iliŋ/11 /Ø- iliŋ/ /g- iliŋ/

Verbs that start with a glide /j, w/ use a combination of the consonantal paradigm III (Table 10) and the harmonizing paradigm I (Table 8). Illustrations are in Table 11: prefixes from paradigm III encode first and third person singular, prefixes from paradigm I encode second person singular. (Before /j/ the plural vowel /i/ is barely audible and may be deleted, so that the surface form of the inflected verb becomes disyllabic, e.g. [nijɔ.kuŋ] ‘we shake’. Such plural vowel reduction does not occur before /w/, e.g. [ni.wɛ.jɔ] ‘we bathe’.) Table 11: Combination of prefix III and prefix II with verbs with initial glide [j, w]

1SG 2SG 3SG 1PL.EXCL 1PL.INCL 2PL 3PL

‘shake’

‘twist, turn around’

‘bathe-FIN’

/n- jɔkuŋ/ /o- jɔkuŋ/ /g- jɔkuŋ/ /ni- jɔkuŋ/ /pi- jɔkuŋ/ /i- jɔkuŋ/ /gi- jɔkuŋ/

/n- jajɔ/ /a- jajɔ/ /g- jajɔ/ /ni- jajɔ/ /pi- jajɔ/ /i- jajɔ/ /gi- jajɔ/

/n- wɛj-ɔ/ /e- wɛj-ɔ/ /g- wɛj-ɔ/ /ni- wɛj-ɔ/ /pi- wɛj-ɔ/ /i- wɛj-ɔ/ /gi- wɛj-ɔ/

3 Basic clausal syntax 3.1 Constituent order in verbal clauses I use the following primitives for core participants of an event in Kaera: S is the single argument of an intransitive predicate (verbal or non-verbal), A is the agent-like argument of a transitive clause, and P the patient-like argument of a transitive clause. Constituent order in intransitive clauses is S V(erb), as in (16), and in transitive clauses it is A P V, as in (17)–(20). P may be encoded as an NP, (17), a free pronoun, (18), or a prefix, (19). A P-prefix can co-occur with a co-referent NP, as in (20a); the prefix is obligatory (cf. (20c)). When there is no 11 The initial vowel /i/ in a verb like /iliŋ/ fuses with the /i/ in the plural prefixes. The result is an identical surface form for 2nd person singular and plural: /Ø-iliŋ/ ‘order you (sg)’ and /i-iliŋ/ ‘order you (pl)’ are both pronounced as [i.ˈliŋ].

110

Marian Klamer

NP, the P-prefix functions as an anaphoric pronoun, as in (20b). For details on verbal prefixing, see §5.2. (16)

Gang bag-i si, abang g-om cry-PFV so village 3SG.POSS-inside 3SG ‘He cried, so all the villagers came down. . .’

(17)

Gang tei patak-o. 3SG wood cut-FIN ‘He cuts wood.’

(18)

Nang gang med. 1SG 3SG take ‘I marry him/her.’

(19)

Gang ga-bang. 3SG-drop 3SG ‘He dropped it/him/her.’

(20)

a.

Gang uxai gu child that 3SG ‘He killed that child.’

b.

Gang ga-bar. 3SG-kill 3SG ‘He killed him/her/it.’

egat all

ya-t. . . go.down-IPFV

ga-bar. 3SG-kill

c. *Gang uxai gu bar. child that kill 3SG Intended: ‘He killed that child.’ In order to topicalize a P, the referring NP can be moved to precede A. This is illustrated in (21), where the P netu egat ‘all my milk’ precedes the A nang ‘1SG’: (21)

Nang gi sei, ne-tu egat nang go COMPL 1SG.POSS-milk all 1SG 1SG ‘I went [and] all my milk I brought along.’

pin hold

gi. go

In imperatives, the addressee (A/S) is omitted, as in (22). Otherwise, the clausal structure of imperatives is identical to that of declaratives, where P precedes V, (23):

Kaera

(22)

Ma g-om mi da! come 3SG.POSS-inside LOC go.up ‘Do come in!’ [invitation; also applies when speaker and addressee are at same level]

(23)

Pagang gu pin-o! basket12 that hold-FIN ‘Hold that basket!’

111

Oblique constituents expressing a location, goal, instrument or displaced theme are expressed as postpositional phrases with mi ‘in’, ‘at’, ‘into, on’ or ta ‘on (top)’. They occur before the predicate, and cannot be moved to a postpredicate position, see §3.6. Adjuncts such as time expressions are clause-initial, as in (24); or, if they are part of a focus constituent, move to a clause-internal position preceding the predicate, as in (25). Adverbs of manner such as kali~kali ‘slowly’ and user~ user ‘quickly’ are also found in pre-predicate position, as in (26)–(27), as are adverbs of modality like masu ‘maybe’, see (39). (24)

Miag ang dumang. yesterday 2SG swim ‘Yesterday you swam.’

(25)

Ui gu gang miag yesterday person that 3SG ‘That person came YESTERDAY.’

(26)

la FOC

ma. come

Ging kali~kali tei baxi gu wang 3PL RDP ~slow tree branch that be ‘Slowly they climb on that tree branch.’

ekeng. climb

(27) Ilwang gang user~user bir bleling g-om mi eser-it. . . 3SG.POSS-inside LOC exit-IPFV Ilwang 3SG RDP~quick run open ‘Ilwang quickly ran outside. . .’

12 A pagang is a long basket made of leaves, with a rope hanging from the head, used to carry e.g. firewood.

112

Marian Klamer

3.2 Equational clauses Equational clauses are verbal, as in (28)–(29),13 or non-verbal, as in (30)–(32). S precedes the predicate. Kaera has no copular verb, but there is a clear intonational break between the S, which has rising intonation and is followed by a pause, and the nominal predicate, which is pronounced with low intonation. In (30), S is expressed as a combination of both an NP and a pronoun, but it is equally grammatical to have S expressed only once. (28)

Ge-kono gu 3SG.ALIEN-shirt that ‘His shirt [is] black.’

(29)

Boom ge-sur respected.old.man 3SG.POSS-word ‘Words of the elders [are] serious.’

(30)

Ui gu (gang) guru. person that 3SG teacher (Mly) ‘That person [is] a teacher.’

(31)

Nang n-en Paulus. 1SG-name Paulus 1SG ‘My name is Paulus.’

(32)

. . . gang lal-ang erang gu see-CONT that.ANAPH that 3SG ‘[. . .] he saw that those [were] bees.’

xan-o. black-FIN

gu that

tamad-o. heavy-FIN

taung. bee

3.3 Existential clauses Existential constructions involve the existential verb wang ‘be, exist’, as in (33), or the locative verb ming ‘be at X’, as in (34). Ming is a mono-transitive verb whose arguments are aligned as A (located entity) and P (location), as illustrated in (34)–(35).

13 For a motivation why Kaera property words are analyzed as verbs, see §4.1.

Kaera

(33)

(34)

(35)

113

Siax wang. chicken be ‘There are chicken(s)’ [we know they are somewhere, not necessarily in sight] A P Nang [ir boi] water river 1SG ‘I am in the river’

ming. be.at

A P [Ir la] boi ming. water FOC river be.at ‘[There] is water in the river’

The existential verb wang ‘be’ is also used to express nominal possession predicatively, as in (36)–(37). (Other possessive strategies are discussed in §4.3.) (36)

N-uax wang sei. Masik rax-o, umux nuk-o. COMPL male two-FIN female one-FIN 1SG.POSS-child be ‘I have children (lit. ‘My children exist’). Two boys, one girl.’

(37)

Masik wal gu gang ge-kriang wang. man that 3SG 3SG.ALIEN-work be ‘That man has work.’ (lit. ‘That man, his work exists’)

3.4 Negation Kaera predicates are negated with the negator bino ‘NEG’, which always follows the predicate. Negative verbal predicates are illustrated in (38)–(39), negative nominal predicates in (40). (38)

Nang na na14 it.thing consume 1SG ‘I won’t eat anything.’

bino. NEG

14 Na ‘consume’ refers to consuming food or liquid, but its basic semantics relate to food; compare na na ‘food eat’ > ‘eat’ vs. ir na ‘water drink’ > ‘drink’.

114

Marian Klamer

(39)

Gang masu ma maybe come 3SG ‘He may not come.’

(40)

bino. NEG

Nang bino, gang bino. 1SG NEG 3SG NEG ‘Not me, not him.’ [e.g. answering the question ‘Who drank my coffee?’]

Existence can be negated with wang bino, as in (41), or with the negative existential verb ning ‘NEG.exist’, as illustrated in (42)–(43). (41)

Xalam nuku gu, xalam g-en old.lady one that old.lady 3SG.POSS-name ‘[There was] an old lady, she had no name.’

wang be

bino. NEG

(42)

Siax ning-o. chicken NEG.exist-FIN ‘There is/are no chicken(s).’

(43)

Lugum di ning-i si uxai masik nuk sepeda tang ma. long.time also NEG.exist-PFV then child male one bicycle(Mly) be.on come ‘Not long [after that] a boy on a bicycle arrives.’

The intransitive negative verb gasaing ‘NEG.can’ expresses inability. It occurs clause-finally with transitive and intransitive verbs, as illustrated in (44)–(45). (44)

Nang ang kup-it 2SG hit-IPFV 1SG ‘I can’t hit you.’

(45)

Liwang gang tar-it swim-IPFV Liwang 3SG Liwang, he can’t swim.

gasaing. NEG.can

gasaing. NEG.can

A third negative verb is -gay- ‘refuse’; this verb can be used in intransitive as well as transitive constructions. When it is used intransitively, S is expressed as both a free pronoun and a prefix, as in (46). The free pronoun is optional, the prefix obligatory. (See §5.3 for a discussion of intransitives that encode S with a prefix.) When -gay- appears in a transitive construction, as in (47), it is the P that is expressed as free pronoun and prefix (see §5.2.1); again the free pronoun is optional, and the prefix obligatory.

Kaera

(46)

(Nang) na-gay-o. 1SG-refuse-FIN 1SG ‘I refuse.’

(47)

Gang (ing) i-gay-o. 3SG 2PL 2PL-refuse-FIN ‘He doesn’t like you (PL).’

*Nang 1SG

115

gay-o. refuse-FIN

*Gang 3SG

ing 2PL

gay-o. refuse-FIN

Negative imperatives are expressed with aki ‘NEG.IMP’, as illustrated in (48). (Gang expresses the S of maxo ‘bitter’, and being co-referential to the clause topic tu gu ‘that milk’ it may be omitted.) Note that aki precedes the predicate. (48)

Tu gu aki er (gang) milk that NEG.IMP make 3SG ‘That milk, don’t make (it) bitter.’

max-o. bitter-FIN

3.5 Questions The identity of persons is questioned with the interrogative pronoun yami ‘who’. Interrogative pronouns always combine with the focus marker la, and stay in the original position of the questioned nominal. Illustrations of questioned A and S arguments are (49)–(50), a questioned P argument is illustrated in (51). (49)

Yami la gareja who FOC church (Mly) ‘Who is in the church?’

(50)

Yami la gang who FOC 3SG ‘Who hit him?’

kup-o? hit-FIN

(51)

Gang yami la 3SG who FOC ‘Who did he hit?’

kup-o? hit-FIN

ming? be.at

Questioned possessors are illustrated by (52a–b). The questions contrast in focus: in (52a) the possessor is focused, in (52b), it is the possessed NP.

116

Marian Klamer

(52) a. Yami la ge-buku ba? b. Yami ge-buku la ba? who 3SG.POSS-book FOC fall who FOC 3SG.POSS-book fall ‘Whose book fell?’ ‘Whose book fell?’ Inanimate entities are questioned with nab- ‘what’, (53)–(54) and locations with ita ‘where’, (55). Note that nab- inflects for its position in a phrase or clause as being final or non-final, just like verbs do (see §7.1). The non-final suffix –i may optionally be omitted. (53)

Egu nab-o? that.one.there what-FIN ‘What is that?’

(54)

Ang nab(-i) la 2SG what-NFIN FOC ‘What do you see?’

(55)

“Liwang. . . Ang ita ming? Liwang. . .” Liwang 2SG where be.at Liwang “Liwang. . . where are you? Liwang. . .”

lal-o? see-FIN

Quantities are questioned with idang ‘how many/much’ as in (56). Idang is also used for questions about time, as in (57), where wedi idang ‘how many days’ acts as the subject of gi ‘go’. The noun wedi ‘day’ may be omitted, probably because the expression has been conventionalized. (56)

Nang gelas idang glass (Mly) how.many/much 1SG ‘How many glasses did/shall I buy?’

(57)

(Wedi) idang gi ang gi? day how.many/much go 2SG go ‘When will you go?’ (Lit. ‘How many days go [then] you go?’)

wal-o? buy-FIN

Questions about reason or method use the word tarang ‘why/how’. Tarang can be fronted as part of a focused constituent, as in (58), or occur post-predicatively, as in (59).

Kaera

(58)

Tarang la gang ma bino? Walsi15 come NEG because why/how FOC 3SG ‘Why didn’t he come? Because he is ill.’

(59)

Siax ga siring tarang? chicken this fry why/how ‘How is this chicken fried?’

gang 3SG

117

i. sick

3.6 Postpositional phrases Kaera has two frequently used adpositions: mi ‘LOC’ (‘in’, ‘on’, ‘at’, ‘into’) and ta ‘on’. Of these, mi is used most frequently and with the widest range of functions, marking oblique constituents of a wide semantic range, as illustrated in (60a)– (66). Oblique constituents must precede the verb; they cannot be moved to a postverbal position, (60b).

(60)

Mi encodes a location a. [Abang mi] ga-dag. village LOC 3SG-leave ‘Leave him/her in the village.’ b. *Ga-dag [abang mi] 3SG-leave village LOC Intended: ‘Leave him/her in the village.’

(61)

(62)

(63)

Mi encodes a goal Ui gu gang [abang mi] village LOC person that 3SG ‘That person goes to the village.’

gi. go

Mi encodes an instrument Ui gu gang [ped mi] tei machete LOC wood person that 3SG ‘That person cut wood with a machete.’

patak-o. cut-FIN

Gang [ga-tang mi] n-uax ga-taring. 3SG 3SG.POSS-hand LOC 1SG.POSS-child 3SG-point.at ‘S/he pointed at my child with her/his hand.’

15 The fixed expression walsi ‘because’ occurs clause-initially. It contains the conjunction si, which on its own can only occur clause-finally.

118

Marian Klamer

Mi encodes the theme argument of a ‘transfer’ event like ‘feed’, ‘give’, or ‘show’: (64)

Gang [naxar mi] 3SG rice LOC ‘S/he fed me rice.’

(65)

Gang [naxar mi] n-eng 3SG rice LOC 1SG -give ‘S/he gave me rice to eat.’

(66)

Gang [ foto mi] photograph (Mly) LOC 3SG ‘S/he showed me a picture.’

n-aas-o. 1SG-feed-FIN

nang 1SG

na. eat

na-taring. 1SG-point.at

A clause can contain two postpositional phrases with mi, as illustrated in (67), where mi encodes both a theme (gong kul ‘his skull’) and a goal (toples gom ‘inside a jar’). (67)

Gang [g-ong kul mi] 3SG.POSS-head shell LOC 3SG ‘He stuck his head into a jar.’

[toples jar (Mly)

gom inside

mi] LOC

uru. stick.into

The postposition mi is etymologically related to the locative verb ming ‘be at’ illustrated in (68) (see also (34)–(35) in §3.2). (68)

Ne-na xas-i ula lemari 1SG-thing split-PFV FOC closet (Mly) ‘My clothes are in the closet.’

ming. be.at

As an independent verb, ming occurs in clause-final position, and unlike mi, it does not need to combine with another verb. The contrast between the postposition mi and the verb ming is shown in the minimally contrastive sentences (69a–b). (69)

a.

Ging abang mi mis-o. village LOC sit-FIN 3PL ‘They stay in the village.’

b.

Ging abang ming gu, mis-o. 3PL village be.at that sit-FIN ‘Those [that are] in the village, [will] stay [there].’

Kaera

119

The other simple postposition in Kaera is ta ‘on’, illustrated in (70). Like mi, ta is etymologically related to a locative verb; tang ‘be on’ in (71): (70)

Gang a-bat ta ga-dag. 3SG.POSS-leg on 3SG-leave 3SG ‘She puts him on her lap.’

(71)

Gang a-bat 3SG.POSS-leg 3SG ‘He is on her lap.’

tang. be.on

More precise locative relations are expressed by combining mi with a possessed noun such as paning ‘front’ in (72) and abung ‘side’ in (73). (In (73), the first mi is optional, the second mi is obligatory.) (72)

(73)

Gang mampelei gu pagang ge-paning mango that basket 3SG.POSS-front 3SG ‘He puts those mangoes in front of the basket.’ Pagang mi n-abung basket LOC 1SG.POSS-side ‘Put the basket next to me.’

mi LOC

mi LOC

mey-o. put-FIN

mey-o. put-FIN

In sum, the two frequently used adpositions mi and ta are related to the locative verbs ming and tang and synchronically encode oblique constituents.

4 Noun phrases Kaera nouns are formally distinguished from verbs by their ability to take possessive prefixes (§4.3). Kaera nouns are not marked for number, gender, noun class or case. A non-possessed NP is composed as in (74). The initial head noun (NHEAD) may be followed by an attribute (ATTR), a numeral or non-numeral quantifier (NUM/QUANT), a demonstrative (DEM) and a demonstrative particle (PART). An illustration is given in (75). If a numeral classifier (CLF) is used, it precedes the numeral, see (87)–(88) below; classifiers do not combine with non-numeral quantifiers.

120

(74)

Marian Klamer

Structure of the Kaera NP [NHEAD ATTR CLF NUM/QUANT

DEM

PART]NP

The demonstrative particle erang ‘that.ANAPH’ refers anaphorically to entities in space, time or discourse. It can combine with a demonstrative particle; ga ‘this’ or gu ‘that’. The demonstrative particles encode location of the referent with respect to the speaker as well as definiteness, as in siax ga ‘this chicken’, siax gu ‘that chicken’. Two demonstrative pronouns are derived from the demonstrative particles: ega ‘this one here’, egu ‘that one there’. The demonstrative pronouns can also function as adnominal attributes, as in botol egu ‘this bottle here’, botol ega ‘that bottle there’. (75)

kunang masik utug erang gu child male three that.ANAPH that ‘those three boys (mentioned earlier)’

In a possessed NP the head noun hosts a prefix encoding person and number of the possessor. It may be preceded by the possessor noun, as in (76). Adnominal possession is further discussed in §4.3.

(76)

[[POSSESSOR Noun POSS-N] bur ge-tag 3SG.POSS-stalk k.o.palm16 ‘this youngest bur palm offshoot’

ATTR iki youngest

PART ]NP

ga this

4.1 Nominal attributes Nominal attributes are nouns, as in (75), or verbs used in modification of a noun head in the NP, as in (77). Kaera has no morpho-syntactic class of adjectives. Words expressing properties like size, color or age, pattern like verbs. Just like verbs, they are grouped into two classes according to their suffixing potential: lexemes ending in a closed syllable, which can take suffixes, and lexemes ending in an open syllable, which cannot. Examples of both groups are given in Table 2 (§2.1) and Table 12 below. Note that the single vowel word i- ‘red’ belongs to the suffix-taking class, while the homophonous verb i ‘sick’ does not.

16 Bur is translated in Indonesian as pohon enau (Arenga saccharifera).

Kaera

121

Table 12: Kaera property verbs Property verbs taking inflectional suffixes ‘ripe’ ‘big’ ‘true’ ‘black, dark’ ‘white’ ‘good’ ‘cold’ ‘red’

tenwadmolxanmiexxowxewi-

Property verbs taking no inflectional suffixes ira kiki xaxui oyogi bagari iki mutu i

‘ripe’ ‘small’ ‘immature, small’ ‘green, blue’ ‘yellow’ ‘youngest’ ‘eldest’ ‘sick’

In (77), xan-o ‘black’ takes an obligatory suffix, encoding it as the final lexeme in the phrase. In (78), kiki ‘small’ is also the final lexeme in the phrase, but does not take a suffix, because unlike xan-, kiki ends in an open syllable. (77)

pagang xan-o (*pagang xan) basket black-FIN ‘(a) black basket(s)’

(78)

pagang kiki (*pagang kiki-o) basket small ‘(a) small basket(s)’

The suffixes that attach to verbs encode the position of the verb in the phrase as final or non-final. Lexemes in phrase-final position take the suffix –o ‘FIN(AL)’, lexemes in non-final position take the suffix –i ‘NON-FINAL’ (see also §7.1). The contrast is illustrated in (79). The intonation of both inflected verb forms differs significantly: the ending –o has rising intonation, while the ending –i has level intonation. Another illustration of the contrast between the final and non-final suffixes is the pair (80)–(81). (79)

a.

[kono xan-o]NP shirt black-FIN ‘(a) black shirt(s)’

b.

[kono xan-i utug]NP shirt black-NFIN three ‘three black shirts’

122

Marian Klamer

(80)

Egu jugut xow-o. that.one.there road good-FIN ‘That one [is] the good road.’

(81)

Nang ega jugut xow-i 1SG this.one.here road good-NFIN ‘I here follow this bad road.’

ga this

g-oyo. 3SG-follow

When a verb is used predicatively it takes the same suffixes, but now these encode its position as final or non-final in the clause rather than the NP, compare (82) with (79). Verbal inflections encoding clausal position are further discussed in §7.1. (82)

Kono gu xan-o. shirt that black-FIN ‘That shirt [is] black.’

4.2 Number, quantification and enumeration Kaera nouns are not morphologically marked for number. Bare nouns can be interpreted as singular or plural. Nouns are explicitly pluralized with the plural number word namung ‘PL’, compare (83a–b). Quantifiers follow the noun, (84). (83)

(84)

a.

na-rat 1SG.POSS-grandchild ‘My grandchildren’

namung PL

b.

na-rat 1SG.POSS-grandchild ‘My grandchild(ren)’

Mauxubar amara yas-it 17 mis-o. frog many bad-IPFV sit-FIN ‘Many frogs are sitting [there].’

Kaera numerals also follow the noun; the cardinals 1–12, 100 and 1000 are given in Table 13. The forms nuk- ‘1’ and (a)rax- ‘2’ take different inflections depending on whether their position in the phrase is final or not (cf. §4.1, 7.1). In their inflectional ability, these two numerals resemble verbs. Higher numerals that have nuk- and rax- in final position are also inflected. This includes yesrax‘7’, which is historically composed of ‘5’ and ‘2’.18 17 Yas ‘bad’ has developed a sense of “very” in combination with amara ‘many’; it also occurs in negative terms of reference, see (95). 18 In all the languages of Pantar the numerals ‘7’ to ‘9’ are historically formed as additive basefive numerals: [5 2] = 7, [5 3] = 8, [5 4] = 9, see Schapper and Klamer (forthcoming).

Kaera

123

Table 13: Kaera numerals 1 2 3 4 5

nuku, nuk(a)rax(i/u)tug ut isim

6 7 8 9 10

tiaam yesraxyentug yeniut xar nuku/nuk-

11 xar nuk beti 19 nuku/nuk12 xar nuk beti rax100 ratu nuku/nuk1000 ribu nuku/nuk-

In (85), the Kaera expressions for the months are given. They all involve numerals, but not all numerals take a suffix when they occur in phrase-final position. (85)

Kaera numerals as part of NPs referring to names of the months (ur ‘moon’) ur nuk-o ‘January’ ur yesrax-o ‘July’ ur rax-o ‘February’ ur yentug ‘August’ ur (i)tug ‘March’ ur yeniut ‘September’ ur ut ‘April’ ur xar nuk-o ‘October’ ur isim ‘May’ ur xar nuk beti nuk-o ‘November’ ur tiaam ‘June’ ur xar nuk beti rax-o ‘December’

Ordinal expressions are created with an (alienable) possessive prefix and a locative adposition mi (§3.6), as in (86). This process applies regularly for numerals three and up. (86)

Uxai ge-mi ut. Uxai child 3SG.ALIEN-LOC four child ‘The fourth child. The tenth child.’

ge-mi 3SG.ALIEN-LOC

xar ten

nuk-o. one-FIN

Ordinals ‘first’ and ‘second’ are expressed with the verbs etu ‘be/go first’ and –murung ‘follow’, as in Ang etu ‘you are/go first’, ang gu-murung ‘you go/are second’ (lit. ‘you follow him’.) Numerals in Kaera do not regularly combine with numeral classifiers, except when they provide additional specifying information, as in (87a–c). No classifier is used to count animals. (87)

a.

wat ipi coconut CLF.fruit ‘four coconuts’

b.

wat bai ut coconut CLF.bunch four ‘four bunches of coconuts’

ut four

19 Beti is an operator word that signifies addition; it is only used in numerals.

124

Marian Klamer

c.

wat er coconut CLF.base ‘four coconut trees’

ut four

Humans may be enumerated with nam ‘CLF.HUM’, as in (88a) and (89). The human classifier is optional, compare (88a–b), and in constructions with nam, the head noun may be omitted, (88c). (88)

a.

ui nam arax-o, person CLF.HUM two-FIN ‘two people, three people’

ui person

b.

ui arax-o, ui person two-FIN person ‘two people, three people’

utug three

nam

utug three

c.

(89)

arax-o, nam two-FIN CLF.HUM ‘two people, three people’

CLF.HUM

masik wal (nam) CLF. HUM man ‘one man; two men’

nuk-o, one-FIN

masik wal man

nam CLF.HUM

(nam) CLF. HUM

utug three

arax-o two-FIN

When the lexeme nam ‘CLF.HUM’ takes a plural prefix and combines with a numeral, the result is a pronominal expression for groups of humans of any number. (90)

i-nam arax-o 2PL-CLF.HUM two-FIN ‘you two’

ni-nam arax-o 1PL.EXCL-CLF.HUM two-FIN ‘we two (excl. you)’

pi-nam arax-o 1PL.INCL-CLF.HUM two-FIN ‘we two (incl. you)’

gi-nam arax-o 3PL-CLF.HUM two-FIN ‘they two’

i-nam tug 2PL-CLF.HUM three ‘you three’

ni-nam tiaam 1PL.EXCL-CLF.HUM six ‘we six (excl. you)’

pi-nam isim 1PL.INCL-CLF.HUM five ‘we five (incl. you)’

gi-nam ut 3PL-CLF.HUM four ‘they four’

Kaera

125

4.3 Possession This section focuses on possession that is encoded by prefixes attached to the possessed noun. Possession is also expressed predicatively with the existential verb wang ‘be’ (‘my children exist’ = ‘I have children’, §3.3), and with a dedicated possessive (‘genitive’) pronoun (§5.4). Possessor prefixes mark person and number of the possessor on the possessed noun. They may combine with a lexical possessor, as in (91a), or a free pronoun referring to the possessor, as in (91b). The free pronoun is optional and encodes emphasis or contrastiveness. (91)

a.

Ilwang g-abat Ilwang 3SG.INAL-leg ‘Ilwang’s leg(s)’

b.

(gang) g-abat 3SG 3SG.INAL-leg ‘his leg(s)’

Nouns are lexically specified as either alienable or inalienable; nouns for body parts and kin terms are typically inalienable. Alienable and inalienable possession is encoded differently, as laid out in Table 14. Inalienable nouns have an obligatory possessor prefix; alienable nouns take optional possessor prefixes. While plural possessors are invariably marked with a prefix containing the plural vowel /i/ and do not distinguish between alienable and inalienable possession, singular possessors do. The singular possessor of an alienable noun is encoded with an e-prefix, while the prefix encoding the singular possessor of an inalienable copies the first vowel of the possessed noun, or, when the noun is vowel-initial, is just consonantal. Table 14: Kaera alienable and inalienable possessor prefixes

1SG 2SG 3SG 1PL.EXCL 1PL.INCL 2PL 3PL

AL(IENABLE)

INAL(IENABLE)

C-initial & V-initial noun neegenipiigi-

C-initial noun nVVgVnipiigi-

V-initial noun nØgni piigi-

A noun cannot switch between the two types of possessor prefixes. For example, g-abat ‘3SG.INAL-leg’ can only take an inalienable possessor prefix: *ge-abat ‘3SG.ALIEN-leg’ is ungrammatical, and cannot refer to e.g. the leg of an animal that has been cut from the body.

126

Marian Klamer

4.3.1 Alienable possession The e-paradigm encoding alienable possession is illustrated in (92). This paradigm is used for both consonant-initial alienables such as ma ‘house’ and vowel-initial ones like abang ‘village’ or ibar ‘dog’. Note that koi ‘skin’ and tu ‘breast’ belong to the alienably possessed noun class despite referring to body parts.

(92)

Nouns with an alienable possessor prefix ‘house’ ‘village’ ‘dog’ ‘skin’

‘breast’

‘woman’

1SG

ne-ma

ne-abang

ne-ibar

ne-koi

ne-tu

ne-umux

2SG

e-ma

e-abang

e-ibar

e-koi

e-tu

e-umux

3SG

ge-ma

ge-abang

ge-ibar

ge-koi

ge-tu

ge-umux

1PL.EXCL

ni-ma

ni-abang

ni-ibar

ni-koi

ni-tu

ni-umux

1PL.INCL

pi-ma

pi-abang

pi-ibar

pi-koi

pi-tu

pi-umux

2PL

i-ma

i-abang

i-ibar

i-koi

i-tu

i-umux

3PL

gi-ma

gi-abang

gi-ibar

gi-koi

gi-tu

gi-umux

When the prefix vowel and the first vowel of the stem are not identical, they project different syllables; for example, /ne-ibar/ ‘my dog’ is pronounced with three syllables: [ne.i.bar]. However, when the prefix vowel and the stem vowel are identical, they are pronounced as a single long vowel, which may be abbreviated in running speech; for example /pi-ibar/ ‘our dog’ is pronounced as [pi:.bar] or [pi.bar]. 4.3.2 Inalienable possession The possession of inalienable nouns is illustrated in (93). Unlike alienables, inalienable nouns must always occur with a possessor prefix. Inalienable nouns that start with a consonant take a syllabic prefix, and the vowel of that prefix is a copy of the first vowel of the possessed noun. The items in (93) illustrate harmony of the five cardinal vowels /a, u, o, e, i/. Vowel-initial inalienable nouns such as uax ‘child, offspring’ take a consonantal prefix. For those items, the second person singular possessor remains unexpressed, as it lacks a consonant.

Kaera

(93)

Nouns with an inalienable possessor prefix ‘father’ ‘intestines’ ‘stomach’ na-mam nu-duyax no-toku 1SG a-mam u-duyax o-toku 2SG ga-mam gu-duyax go-toku 3SG ni-toku 1PL.EXCL ni-mam ni-duyax pi-toku 1PL.INCL pi-mam pi-duyax i-mam i-duyax i-toku 2PL gi-mam gi-duyax gi-toku 3PL

‘tongue’ ne-leb e-leb ge-leb ni-leb pi-leb i-leb gi-leb

‘fontanelle’ ni-dimang i-dimang gi-dimang ni-dimang pi-dimang i-dimang gi-dimang

127

‘child’ n-uax Ø-uax g-uax ni-uax pi-uax i-uax gi-uax

4.3.3 Possession and nominalization Possessive prefixes can be used to create nouns from non-nominal base forms, as in (94). Such forms may be used to express selection from a set, such as when someone asks to take ‘the black one’ from a pile of shirts with different colors. Combined with the word yas ‘bad’, such nominalizations are used as negative terms of reference, as in (95). (94)

(95)

a.

ge-xan-o 3SG.ALIEN-black-FIN ‘(the) black one’

b.

ge-bagari 3SG.ALIEN-yellow ‘(the) yellow one’

c.

ge-wad-o 3SG.ALIEN-big-FIN ‘(the) big one’

d.

ge-kiki 3SG.ALIEN-small ‘(the) small one’

a.

ge-bag 3SG.ALIEN-cry ‘cry baby’

b.

ge-taxau 3SG.ALIEN-steal ‘thief’

c.

ge-akal 3SG.ALIEN-cheat ‘cheater’

yas-o bad-FIN

yas-o bad-FIN

yas-o bad-FIN

5 Pronouns and person prefixes Kaera has free pronouns and prefixes encoding person and number. The free pronoun forms are given in Table 15. Free pronouns may encode a transitive agent (A) (§5.1), a patient (P) (§5.2), an intransitive subject (S) (§5.3), or a possessor (§5.4). The prefixes encoding P are presented in Table 16; the paradigm shapes

128

Marian Klamer

are further discussed in §2.7.1. There are no differences between paradigms in the plural. The possessor prefix forms were presented in Table 9 above (§4.3). Table 15: Free pronouns

1SG 2SG 3SG 1PL.EXCL 1PL.INCL 2PL 3PL

Pronouns encoding A, P, S

Possessor pronouns

nang ang gang ning ping ing ging

neg eg geg nig pig ig gig

‘mine’ ‘yours (sg)’ ‘his/hers/its’ ‘ours’ ‘ours’ ‘yours (pl)’ ‘theirs’

Table 16: Person prefixes encoding P C-initial verbs

1SG 2SG 3SG 1PL.EXCL 1PL.INCL 2PL 3PL

I nVVgV-

V-initial verbs II naaga-

III nØgnipiigi-

The syntactic alignment in Kaera basically follows an accusative pattern, encoding A and S with free pronouns (§5.1, §5.3), while Ps may be encoded as verbal prefixes or lexical NPs depending on which class the verb belongs to (§5.2). Kaera has traces of a split-S system: while the typical function of the verb prefixes is to encode P, a few intransitive verbs (also) encode S as a verbal prefix (§5.3).

5.1 Pronouns to encode A, S, P and Possessor A and P can both be expressed as free pronouns, as in (96). The referent of third person gang may be animate or inanimate. Whether P is expressed as a free pronoun or a verbal prefix depends on the verb class (§5.2.1). When P is a prefix, it may co-occur with an optional pronoun, as in (97). A possessive prefix may also co-occur with an optional pronoun, see §4.3.

Kaera

(96)

(97)

129

A P Nang gang lal-o 3SG see-FIN 1SG ‘I see him/her/it.’ A P P-V Nang (gang) ga-samang. 3SG 3SG-decorate 1SG ‘I decorate it (e.g. a cake)’; ‘I dress him/her up.’

In (98), S is a semantic controlling agent, in (99) it is a non-controlling undergoer. In both cases, S is encoded as a free pronoun preceding the predicate. The examples also illustrate that full pronouns can co-occur with lexical NPs, including personal names, to make their referent explicit. (98)

Ilwang gang user~user bir bleling g-om mi eser-o. RDP ~quickly run open 3SG-inside LOC exit-FIN Ilwang 3SG ‘Ilwang quickly ran outside.’ (lit. ‘. . . ran out to (the) open’s inside’)

(99)

Ui gu gang i. person that 3SG sick ‘That person is sick.’

5.2 The encoding of P The encoding of P depends on the lexical class to which the verb belongs. Kaera has three classes of transitive verbs: Class 1 encodes P only as a free constituent, and never as a prefix. Class 2 takes an obligatory prefix for P. Class 3 expresses P either with a prefix, or as a free constituent. The three transitive verb classes are discussed in §5.2.1. In addition, Kaera has a class of labile verbs, discussed in §5.2.2, which can be used transititively or intransively without changing the morphological shape of their base: with a prefix they are used transitively, without a prefix they are used intransitively. 5.2.1 Transitive verbs and the encoding of P The first class of transitive verbs encodes P only as a free constituent, and never as a prefix. Examples are presented in Table 17.

130

Marian Klamer

Table 17: Transitive verbs that only take a free P-constituent VERB

MEANING

VERB

MEANING

anwalye iswati laltarakxasbakpatakpaygaing er tabag-

‘sell’ ‘buy’ ‘scoop up’ ‘scoop (water, rice)’ ‘hear, listen, obey’ ‘see’ ‘choose, pick (person, thing)’ ‘split, break’ ‘cut in small pieces’ ‘cut down (e.g. tree)’ ‘slice’ ‘order’ ‘do, make’ ‘to bundle’

xewtuu kupxaxsi kikmed na (ei) bli keri aposwre

‘peel’ ‘scratch, peel’ ‘hit (thing, person)’ ‘itch, sting’ ‘bite (to hurt)’ ‘bite (to eat)’ ‘take (thing), marry (person)’ ‘eat, drink, consume’ ‘pull (canoe) on shore’ ‘pull’ ‘sweep’ ‘lift (up)’ ‘carry’

The following sentences illustrate that verbs expressing P only as a free constituent (which can be a pronoun or a lexical NP) do not encode P as a verbal prefix: compare (100a–b), (101a, c, d)–(101b) and (102a, b)–(102c). (100)

a.

Nang gang kup-o. 3SG hit-FIN 1SG ‘I hit him/her/it.’

b. *Nang gu-kup-o. 1SG 3SG-hit-FIN Intended: ‘I hit him/her/it.’

(101)

a.

Nang gang wati. 1SG 3SG hear ‘I hear / listen to him.’

b. *Nang 1SG

c.

Nang ga-mang wati. 3SG-voice listen 1SG ‘I hear his voice.’

d.

Nang presiden 1SG president (Mly) ‘I obey the president.’

b.

Gang tei patak-o. tree cut 3SG ‘He cuts a tree.’

b.

Gang nang patak-o. 3SG 1SG cut-FIN ‘He cuts me.’

(102)

c. *Gang na-patak-o. 1SG-cut-FIN 3SG Intended: ‘He cuts me.’

g-wati. 3SG-hear

wati. hear

Kaera

131

The second class of transitive verbs has an obligatory prefix for P, see Table 18. Table 18: Transitive verbs with an obligatory P-prefix VERB

MEANING

VERB

MEANING

-yay-od -tutuk -pek-wey-aas-tuu -ter-dag -bar -koli -tuki -tub-ada

‘twist, turn around’ ‘throw to’ ‘talk, speak to’ ‘call’ ‘bathe’ ‘feed’ ‘precede’ ‘chase (away), expel’ ‘let go, release (animal, person)’ ‘kill’ ‘let roll’ ‘peck at’ ‘burn, roast, light (up)’ ‘frighten’

-tobung -wal-yurang -iling -riang -yokung -eng-bang -tering -samang -bung -oyo -taring

‘reject, push away (person, thing)’ ‘fill’ ‘quarrel with’ ‘urge, incite, invite’ ‘look after’ ‘swing, shake’ ‘give to’ ‘drop (person, thing)’ ‘let float’ ‘decorate, dress up’ ‘be near (person, thing)’ ‘follow’ ‘point at, show to’

In (103), it is shown that the object of -aas ‘feed’ must be encoded as a prefix; it cannot be expressed as only a lexical NP or pronoun. The same is illustrated for the prefixing verb –iling ‘urge’ in (104). (103)

a.

Nang uxai gu child that 1SG ‘I feed that child.’

g-aas-o. 3SG-feed-FIN

b. *Nang uxai gu / gang aas-o. 1SG child that 3SG feed-FIN Intended: ‘I feed that child / him.’ c. *Nang gang g-aas-o. 3SG 3SG-feed-FIN 1SG Intended: ‘I feed him.’ (104)

a.

Gang n-iling nuang mi 3SG 1SG-urge cloth LOC ‘He urged me to sell cloth.’

an-o. sell-FIN

b. *Gang nang iling nuang mi an-o. 3SG 1SG urge cloth LOC sell-FIN Intended: ‘He urged me to sell cloth.’

132

Marian Klamer

Observe that the verbs in the right hand column of Table 18 all end in –ng, while in Table 17 we have just one verb ending in –ng (gaing ‘order’). This suggests a correlation between the presence of a final nasal and the transitive cum prefixing character of a verb. More specifically, there are indications that final –ng is a (fossilized) causative suffix in at least some of these verbs. The verb –ba-ng ‘drop’ can be analyzed as the causative of ba ‘fall’, compare (107a–b). And in Blagar, the direct neighbor of Kaera, a transparent causative suffix –ng is found (Steinhauer, this volume). (105)

a.

Wat nuk ba ser-o. coconut one fall descend-FIN ‘A coconut fell down.’

b.

Gang e-wat ga-ba-ng. 2SG-coconut 3SG-fall-CAUS 3SG ‘He dropped your coconut.’

For other verbs ending in –ng I do not have similar evidence to suggest that final –ng is a causative suffix. For example, the verb dumang ‘swim’ is a labile verb that is used in both transitive and intransitive constructions, without changing the morphological shape of its base, (106) (see §5.2.2). (106)

a.

Uxai gu dumang pati. CONT person that swim ‘That person is swimming.’

b.

Gang uxai gu gu-dumang. 1SG person that 3SG-swim ‘I made/help/let that person swim.’

In short, the class of verbs taking an obligatory prefix has many members ending in –ng, and this consonant may be a fossilized causative suffix in (some of) the verbs. Synchronically, however, causatives are derived by an analytical construction (see §6.4). The third class of transitive verbs express P either with a prefix, or as a free constituent. Examples include pin ‘hold’ in (107)–(108), and tub ‘burn’ in (109)– (110). A P-prefix can co-occur with a co-referent NP, as in (107) and (109), or occur on its own, (110b).

Kaera

(107)

Nang dur gi-pin-o. mouse 3PL-hold-FIN 1SG ‘I caught (the) mice.’

(108)

Gang gelas pin-o. 3SG glass (Mly) hold-FIN ‘He holds (a) glass.’

(109)

Kabakut gu gu-tub-o. cigarette that 3SG-burn-FIN ‘Smoke that cigarette.’

(110)

a.

Nang lampu tub-o. 1SG lamp burn-FIN ‘I light (the) lamp(s).’

c.

Nang gang tub-o. 3SG burn-FIN 1SG ‘I burn him.’

b.

133

Nang gu-tubo. 1SG 3SG-burn-FIN ‘I burn him.’

5.2.2 Labile verbs Kaera labile verbs can be used in both intransitive and transitive clauses. As intransitives, they take no prefix (see §5.3); as transitives, they take a P-prefix. Examples are given in Table 19, and illustrations in (111)–(113). An exception is -gay- which takes a prefix both when it is used as an intransitive, encoding S (‘refuse’); and when it is used as a transitive (‘not like something’), encoding P. Illustrations are given in (46)–(47) above. Table 19: Verbs that occur in both intransitive and transitive constructions VERB

MEANING

VERB

MEANING

tutuk -tutuk pek-pekobo-obo-

‘talk, speak’ ‘talk, speak to X’ ‘yell’ ‘call X’ ‘return’ ‘return X’

ming -ming dumang -dumang -gay-gay-

‘be’ ‘be at X, be put at X’ ‘swim’ ‘help / let X swim’ ‘refuse’ ‘not like X’

134

(111)

(112)

(113)

Marian Klamer

a.

Ui gu tutuk pati. CONT person that talk ‘That person is talking.’

b.

Ui gu ga-tutuk pati. person that 3SG-talk CONT ‘That person is talking with him.’

a.

Nang pek-o, etang 1SG call-FIN and ‘I yell and jump.’

b.

Nang ge-pek-o, etang 3SG-call-FIN and 1SG ‘I call him and jump.’

a.

Ui masik gu gang obo-t return-IPFV person male that 3SG ‘That man climbs down again. . .’

b.

Nang ge-topi gu g-obo-t 3SG-hat that 3SG-return-IPFV 1SG ‘I return with his hat. . .’

kodok. jump kodok. jump

eser. . . climb.down

pin hold

wa-t. . . go-IPFV

5.3 Encoding of S Most Kaera intransitive verbs encode S with a free pronoun. Illustrations are given in Table 20. The verbs express active or non-active events. Additional examples of intransitive verbs are the deictic verbs in Table 21. The two patterns of marking S are illustrated in (114) and (120). Table 20: Intransitive verbs with free pronoun S VERB

MEANING

VERB

MEANING

mulai bir tarkodok amar

‘play’ ‘run’ ‘swim (to catch fish)’ ‘jump’ ‘go, move’ (e.g. on a bicycle)

ba tee ekeng eser nimin

‘fall’ ‘sleep’ ‘climb up (tree, wall)’ ‘climb down (tree, wall); exit’ ‘die’

Kaera

135

Table 21: Intransitive deictic verbs with free pronoun S MOTION AWAY FROM DEICTIC CENTER UP DOWN SAME LEVEL REMOTE

MOTION TOWARDS DEICTIC CENTER

‘go up from DC’ ‘go down from DC’ ‘go from DC’ ‘go far’

ip mid wa gi

(114)

Nang ekeng. climb.up 1SG ‘I go up.’

(115)

Nang sepeda wang bicycle (Mly) be 1SG ‘I go on/with a bicycle.’

*(Nang) 1SG

da ya ma

‘go up to DC’ ‘go down to DC’ ‘come to DC’

n-ekeng. 1SG-climb.up

amar. move

Kaera also shows remnants of an alignment system where a non-controlling undergoer S is encoded like P, with a verbal prefix. Illustrations of such verbs attested in my corpus are given in Table 22. Examples (116)–(120) show how they are used in context. In a clause, the S must be expressed with a free pronoun alongside the prefix, see (116)–(118). Table 22: Intransitive verbs with S prefix VERB

MEANING

VERB

MEANING

-waat -kuru -lagur

‘live’ ‘be silent’ ‘jump up’

-uabung -asu -iya

‘faint, be unconscious’ ‘think, want, intend’ ‘give birth’

(116)

Ang a-lagur. 2SG 2SG-jump.up ‘You jump up.’

(*A-lagur.)

(117)

Gang gu-kuru. 3SG-be.silent 3SG ‘He is silent.’

(118)

Nang n-uabung. 1SG-faint 1SG ‘I faint.’

(*Gu-kuru.)

(*N-uabung.)

136

Marian Klamer

(119)

Gang g-iya-t gang ge-sil tuning met or-i sei. 3SG 3SG-go.down-IPFV 3SG 3SG-rope placenta take hang-PFV COMPL ‘Giving birth, she took his umbilical cord [and] hung [it up].’

(120)

N-uax nimin-i sei, nang yedi 1SG-child die-PFV COMPL 1SG still ‘My child is dead already, I still live.’

n-waat. 1SG-live

(*Nang yedi waat.) 1SG still live

There is no apparent relation between the semantics of the verb or S, and the ability of S to be expressed as a prefix; compare agentive, controlled ‘jump up’ in (116) and uncontrolled ‘faint’ in (118) which both take a prefix; and -waat ‘live’ in (120), which is prefixed, while nimin ‘die’ in (120) is not.

5.4 Possessor pronouns A nominal possessor is expressed as a prefix on the possessed noun, and may combine with a coreferent lexical NP or free pronoun (see §4.3). In addition, there is a set of dedicated possessive pronouns, listed in Table 23. Illustrations of how these pronouns are used in context are (121)–(122). Free pronouns can combine with possessor pronouns. This is illustrated in (123)–(124). The free pronoun is optional and encodes emphasis or contrastiveness. Table 23: Possessor pronouns 1SG

neg

‘mine’

2SG 3SG

eg geg

‘yours (sg)’ ‘his/her/its’

(121)

1PL.EXCL 1PL.INCL 2PL 3PL

Egu neg, eg that.one.there mine yours ‘That’s mine, not yours.’

(122)

Ang eg erang 2SG yours that.ANAPH ‘Where’s yours?’

(123)

Nang neg erang mine that.ANAPH 1SG ‘Where’s mine?’

bino. NEG

gu that

gu that

ita where

ita where

ming? be.at

ming? be.at

nig pig ig gig

‘ours’ ‘ours’ ‘yours (pl)’ ‘theirs’

Kaera

(124)

137

Boom gu gang gen Bak Leng Tuko his Bak Leng Tuko old.man that 3SG ‘That man’s name was Bak Leng Tuko.’ (contrasting with a woman’s name mentioned in the preceding sentence.)

6 Serial verb constructions and related predicates 6.1 Serial verb constructions (SVCs) A serial verb construction (SVC) expresses a single conceptual event by a combination of two or more lexical verbs, which share at least one argument, and occur under a single intonation contour. Kaera SVCs often combine a verb expressing an action, motion or posture with a deictic verb expressing the direction or purpose of the event. Illustrations are (125)–(127). (125)

(126)

Pagang xan-i nuk pi basket black-NFIN one go ‘Bring a black basket up here.’ Sei

gi, ning ga go 1PL.EXCL this ‘After that, we ran up.’

COMPL

(127)

Ge-topi di 3SG.ALIEN-hat (Mly) also ‘His hat also falls. . .’

bir run

ba fall

da. go.up

da. go.up

ser-o, . . . descend-FIN

Kaera SVCs also function to introduce additional participants into a sentence. Examples include serializations with existential wang ‘be’ (§6.2) and transitive pin ‘hold’ and met ‘take’ (§6.3), causative serializations (§6.4), and grammaticalized serializations with –(e)ng ‘give’ (§6.5)

6.2 SVCs with wang ‘be, exist’ The verb wang ‘be, exist’ (§3.3) combines with intransitive and transitive verbs. In SVCs, wang is developing into an adpositional-like element. The verb ekeng ‘climb up’ in (128) is intransitive. In combination with such a verb, wang intro-

138

Marian Klamer

duces a goal; compare (128a) with (128b), where tei baxi gu ‘that tree branch’ is the goal. (128)

a.

Ging kali~kali ekeng. RDP~slow climb.up 3PL ‘They climb up slowly.’

b.

Ging kali~kali tei baxi gu wang ekeng. . . RDP~slow tree branch that be climb.up 3PL ‘Slowly they climb up onto that tree branch. . .’

The only way to encode a goal with an intransitive verb is in combination with wang, compare (129a–b). (Intransitive verbs can also combine with goals that are part of a postpositional phrase with mi, see §3.6) (129)

a.

Wer bir~bir ming, gang ekeng 3SG climb.up sun RDP~run be ‘At midday, she went up to the village.’

b. *Wer sun

bir~bir RDP~ run

ming, be

gang 3SG

ekeng climb.up

abang village

wang be

abang village

gi. go

gi. go

Wang ‘be’ also combines with transitive verbs. In such constructions, it functions to express that the object is less affected, or less easily available. For example, in (130a–b) it marks the difference between holding a glass in one’s hand and touching a glass that is still on the table. In (131a–b), using wang implies that the candle nuts are less easily available, e.g. because they are in a pile mixed with various other items. (130)

(131)

a.

Gang gelas pin-o. glass (Mly) hold-FIN 3SG ‘He holds a glass (in his hand).’

b.

Gang gelas wang pin-o. glass (Mly) be hold-FIN 3SG ‘He touches a glass (standing on the table).’

a.

Pan tarak-o. candle.nut pick-FIN ‘Pick candle nuts.’

Kaera

b.

139

Pan wang tarak-o. candle.nut be pick-FIN ‘Pick candle nuts [by selecting them from a pile that also contains other stuff]’

6.3 SVCs with pin ‘hold’ and met ‘take’ When used in a serial verb construction, the verbs pin ‘hold’ and met ‘take’ introduce displaced theme participants. For instance, in (132), pin ‘hold’ introduces mauxubar ‘frog’, the item brought home. (132)

Ilwang mi Liwang unang mauxubar gu pin ma mi gi. that hold house LOC go Ilwang OBL Liwang together frog ‘Together Ilwang and Liwang take that frog home.’ (lit. ‘. . . hold that frog go home’

SVCs with met ‘take’ are illustrated in §6.5. Met is the grammaticalized variant of med ‘take’. It only occurs in SVCs and preceding mi ‘LOC’.

6.4 Causative SVCs with er ‘make, do’ The productive synchronic strategy to derive causatives in Kaera is with a SVC that contains the verb er ‘do, make’, as in (133a). (Traces of a morphological causative are discussed in §5.2.1.) (133)

a.

Gang nuang gu er bagari. cloth that make yellow 3SG ‘He made that cloth yellow.’

b. *Gang 3SG

na-bagari-ng 3SG-yellow-CAUS

6.5 Grammaticalized SVCs with the verb –(e)ng ‘give’ Kaera has no underived ditransitive verbs. If an event has three participants, one participant is encoded as an oblique constituent with the postposition mi (§3.6), or it is introduced into the sentence with a separate verb wang ‘be’ or pin ‘hold’, as discussed above (§6.2–6.3).

140

Marian Klamer

Three participant constructions denoting a ‘give’-event are grammaticalized SVCs with the verb –(e)ng ‘give’. The recipient (R) of –(e)ng ‘give’ is encoded with a prefix on the verb, and the displaced theme (T) is marked as oblique with the postposition mi, as in (134). In addition, ‘give’ constructions may involve a particle me, as in (134)–(135). Me is an optional element that may be a further reduced form of the serial verb met which in turn derives from med ‘take’. (134)

Gang buku mi book (Mly) LOC 3SG ‘He gave me a book.’

(135)

Gang kono wad-i la shirt big-NFIN FOC 3SG ‘He gave me a big shirt.’

n-eng. 1SG-give

nuk one

mi

me

LOC

PART

n-eng. 1SG-give

The verb –(e)ng ‘give’ often combines with met ‘take’ and mi ‘LOC’ in fixed expressions, as in (136)–(138). Note that the first clause in (138) contains the independent verb med ‘take’, while in the second clause met is combined with mi and –(e)ng. The function of mi is unclear. (136)

Gang ge-topi gu med a, 3SG 3SG.ALIEN-hat that take CONJ ‘He takes that hat of his, xabi mampelei utug met mi kunang masik namung gu gi-ng. that 3PL-give then mango three take LOC children male PL then takes three mangoes to give to the boys.’

(137)

Egu met mi boom~boom that.one.there take LOC RDP~old.man ‘Take that one there to the elders to eat.’

(138)

Ge-topi gu g-obo-t pin wa-t met mi me g-eng. 3SG.ALIEN-hat that 3SG-return-IPFV hold go-IPFV take LOC PART 3SG-give ‘[They] return with his hat and give it to him.’

g-eng 3SG-give

gang 3SG

na. eat

7 Encoding clausal position and aspect of verbs Kaera verbs take suffixes that mark the clausal position of the verb as final or non-final; most but not all of them also have an aspectual function. In addition,

141

Kaera

aspect is expressed lexically by aspectual adverbs. In §7.1 I first discuss the distribution of the suffixes encoding clausal position of the verb, followed by a sketch of the semantics of aspectual suffixes and adverbs in §7.2. The analysis presented here is preliminary; many details of the position and aspect suffixes and their interaction with aspectual adverbs are still unclear.

7.1 Distribution of verbal suffixes marking clausal position and aspect Kaera verbal suffixes mark the clausal position of the verb to which they attach, as laid out in Table 24. The suffix –o only occurs on clause-final verbs; it signals the end of a clause and has no aspectual function. In contrast to this, the perfective and imperfective suffixes occur on non-final verbs only, but also have an aspectual meaning. The continuative suffix can occur on both final and non-final verbs. Table 24: Inflectional suffixes encoding verbal position -o

-i ‘PFV’

-(i)t ‘IPFV’

-ang ‘CONT’

NON-FINAL

–

(

(

(

FINAL

(

–

–

(

While the suffix –o ‘FIN’ is restricted to occur on clause-final verbs, it is not the case that all clause-final verbs take this inflection. Some verbs do not take inflectional suffixes at all, examples are given in §2.1 (Table 2), §4.1 (Table 12) and §5.2.1 (Table 17, Table 18). The suffix only attaches to verbs that end in a consonant;20 it never occurs on verbs ending in a vowel. There are a few exceptional verbs, an example is tutuk ‘talk, speak to’, which is consonant-final but does not take a suffix. In what follows some of the contrasts encoded by the verbal suffixes are further demonstrated. The verb patak- ‘hit’ in (139) ends in a consonant and occurs in clause-final postion, therefore it takes –o. This clause contrasts with (140)–(141), where the verbs are non-final and have different aspects as also expressed by the different aspectual adverbs (see §7.2). In (142), the verb inflected with continuative –ang is final; such verbs can also be non-final as shown in (143). 20 The exception to this rule is verbs ending in –ng: while they are consonant-final, most of them do not allow suffixing. This reflects the fact that final –ng on verbs may represent a (historical) causative suffix, and is not part of the verb base, see §5.2.1.

142

Marian Klamer

(139)

Ging tei gu patak-o. tree that cut-FIN 3PL ‘They cut that wood.’

(140)

Ging tei gu patak-it pati. 3PL tree that cut-IPFV CONT ‘They are cutting that wood.’

(*sei)

(141)

Ging tei gu patak-i sei. 3PL tree that cut-PFV COMPL ‘They have cut that wood.’

(*pati)

(142)

Ging tei gu patak-ang. 3PL tree that cut-CONT ‘They are cutting that wood.’

(143)

Ang tar-ang erang gu, nang 2SG swim-CONT that.ANAPH that 1SG ‘[While] you were swimming, I saw you.’

ang 2SG

lal-ang. see-CONT

In (144), the first clause ends in a verb yas-o, while in the second clause yasis non-final as it is followed by the aspectual particle bedo ‘INCOMPL’. Therefore it must be inflected with one of the non-final suffixes, in this case –i ‘PFV’. (It is not possible to combine an imperfective verb taking the suffix –(i)t with bedo ‘INCOMPL’.) (144)

Ma ge-ega house 3SG.POSS-this.one.here ‘This house here is broken,

yas-o, bad-FIN

ma ge-memo yas-i (*yas-o; *yas-it) house 3SG.POSS-there bad-PFV the house over there is not yet broken.’

bedo. INCOMPL

Another illustration of how the suffixes differ according to the position of the verb is given in (145)–(147), with the verb wal- ‘fill’. (145)

Ipai botol nuk g-wal-o. oil bottle (Mly) one 3SG-fill-FIN ‘One bottle of oil.’ (lit. ‘Oil filling one bottle.’)

Kaera

(146)

(147)

Botol ega g-wal-i bottle (Mly) this.one.here 3SG-fill-PFV ‘This bottle here is already full.’

sei.

Botol ega g-wal-it bottle (Mly) this.one.here 3SG-fill-IPFV ‘This bottle here is very full.’

yas-o. bad-FIN

143

COMPL

In sum, verbal suffixes in Kaera encode both clausal position and aspect of the verb.

7.2 Semantics of aspectual suffixes and adverbs Kaera aspect is encoded by verbal suffixes (148) and aspectual adverbs (149).

(148)

(149)

Aspectual suffixes -it, -t IMPERFECTIVE -i PERFECTIVE -ang CONTINUATIVE Aspectual adverbs COMPLETIVE, ‘already’ sei CONTINUATIVE, ‘continuously’ pati INCOMPLETIVE, ‘not yet’ bedo

Examples (150)–(153) illustrate the aspectual inflections and adverbs. The imperfective typically combines with the particle pati ‘CONT’, and the perfective with sei ‘COMPL’ or bedo ‘INCOMPL’. (150)

Yami la tar-it pati la who FOC swim-IPFV CONT FOC ‘Who is swimming over there?’

gu? that

(151)

Yami la tar-i sei who FOC swim-PFV COMPL ‘Who has swum over there?’

gu? that

(152)

Yami la tar-i bedo? who FOC swim-PFV INCOMPL ‘Who hasn’t swum yet?’

la FOC

144 (153)

Marian Klamer

Yami la tar-ang la who FOC swim-CONT FOC ‘Who swims over there?’

gu? that

Sentences (154)–(156) are additional illustrations of the different semantics of the adverbs. (154)

(155)

(156)

Nang buku g-obo-t mi g-eng book (Mly) 3SG-return-IPFV LOC 3SG-give 1SG ‘I have already returned the book to him.’ Nang buku g-obo-t 1SG book (Mly) 3SG-return-IPFV ‘I am returning the book.’ Ma ega yas-i house that.one.there bad-PFV ‘That house is not broken yet.’

sei. COMPL

pati. CONT

bedo. INCOMPL

Of the three aspectual suffixes, imperfective –(i)t and perfective –i occur in the corpus much more frequently than continuative –ang. Why this is so remains to be established. Constructions with –ang and –it (pati) are semantically similar: the consultant gave the examples in (157a–b) and (158a–b) as synonyms. (157)

(158)

a.

Gang masu erex-ang maybe scared-CONT 3SG ‘Maybe he is scared, or. . .’

ba only

e. . . eh

tarang how

b.

Gang masu erex-it maybe scared-IPFV 3SG ‘Maybe he is scared, or. . .’

pati

ba only

e. . . eh

tarang how

ba.. only

Gang ging lal-ang ba. Gang gi-wang 3PL see-CONT only 3SG 3PL-be 3SG ‘He just looks at them. He doesn’t ask them.’

takan-i ask-PFV

bino.

Gang ging lal-it ba. Gang gi-wang 3SG 3PL see-IPFV only 3SG 3PL-be ‘He just looks at them. He doesn’t ask them.’

takan-i ask-PFV

bino.

a.

b.

CONT

ba.. only

NEG

NEG

Kaera

145

8 Discussion The phonology of Kaera, and the size of its consonant and vowel inventory is fairly typical for the languages of the Alor-Pantar group. The Kaera velar fricative /x/ is a reflex of the proto-Alor-Pantar uvular stop *q, which is still found in Teiwa, but became /k/, /g/, /ʔ/ or zero in Western Pantar, Blagar and Adang (Holton et al. 2012). In the Pantar context, Kaera is unusual in having distinctive vowel length for 5 vowels: Teiwa has a length distinction for 3 vowels only, and long vowels are completely lacking in Blagar, and Western Pantar. Atypical for Alor Pantar is the absence of a phonemic glottal stop in Kaera. While the syntactic alignment in Kaera basically follows an accusative pattern, as in Teiwa and Blagar (but not in Western Pantar), Kaera shows traces of a split-S where a non-controlling undergoer S is encoded like P – with a verbal prefix (§5.3). In transitive clauses, the encoding of P varies according to the lexical class of the verb: verbs in class 1 encode P only as a free constituent, and never as a prefix; verbs in class 2 always encode P as a prefix; and verbs in class 3 express P either as a prefix, or as a free constituent (§5.2). Kaera differs from its neighbor Blagar in having no productive causative or nominalizing morphology. However, what makes Kaera stand out in comparison to other Pantar languages such as Teiwa (Klamer 2010) and Western Pantar (Holton, this volume) is that it has a relatively complex morphology which shows multiple types of interaction between both syntax and phonology. First, the syntactic position of a verb as clause final, or non-final, is reflected by the verbal suffix. Second, interaction with phonology is evident in the person-prefix allomorphy, where the shape of the initial syllable of a noun/verb determines the choice of allomorph. Third, the shapes of verbs also determine their suffixing properties: verbs that end in a consonant can host aspectual suffixes, while those ending in a vowel cannot. Fourth, in almost all the person prefixes, the vowel harmonizes with the first vowel of the host verb or noun. Kaera and Wersing (Schapper and Hendery, this volume) are the only AP languages with vowel harmony.

References Holton, Gary, Marian Klamer, Frantisek Kratochvíl, Laura C. Robinson & Antoinette Schapper. 2012. The historical relations of the Papuan languages of Alor and Pantar. Oceanic Linguistics 51(1): 86–122.

146

Marian Klamer

Holton, Gary & Laura Robinson. forthcoming. The internal history of the Alor-Pantar language family. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Klamer, Marian. 2010. A grammar of Teiwa. Berlin New York: De Gruyter Mouton. Schapper, Antoinette & Marian Klamer. forthcoming. Numeral systems in the Alor-Pantar languages. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/.

Hein Steinhauer

4 Blagar 1

The language scene

148

2 2.1 2.2 2.3 2.4 2.5 2.5.1 2.5.2 2.5.3 2.5.4

150 Phonology 150 Vowels 152 Consonants 153 Phonotactics 154 Stress Morpho(phono)logy 156 Reduplication 160 Prefixation 161 Suffixation 163 Compounding

3 3.1 3.1.1 3.1.2 3.1.3 3.2 3.2.1 3.2.2 3.3 3.4 3.4.1 3.4.2

164 Basic clausal syntax 166 Verbal predicates 166 Intransitive verbs and adjectives 167 Transitive verbs 169 Encoding three participant events 172 Non-verbal predicates 172 Nominal predicates 172 Postpositional predicates 174 Clausal negation 175 Questions 175 Yes-no questions 176 Questions with question words

4 4.1 4.2 4.3 4.4 4.5

178 Noun phrases 178 Adjective slot 180 Quantity slot 180 Relative clauses 181 Determiners 182 Possession

5 5.1 5.2 5.3 5.4

184 Pronouns 185 Subject pronouns 186 Object pronouns 187 Possessive pronouns 190 Dual pronouns

155

148

Hein Steinhauer

6 6.1 6.1.1 6.1.2 6.2 6.3

Prefixation: agreement, causation and intentionality 193 Causative prefixation 193 Non-verbal roots 194 Verbal roots 196 Intentional prefixation 197 Prefixal defaulting

7 7.1 7.1.1 7.1.2 7.2 7.3

198 The verbal complex 199 Incorporated slot 199 Incorporated mi 200 Incorporated veŋ 203 Aspect slot 204 Appeal slot

8 8.1 8.2 8.3 8.4

205 The –t suffix Evidential marking -t Gradation marking -t Limitation marking -t Manner marking -t

9 9.1 9.2 9.3

210 Serial verb constructions Serialization with =ma, met and met=ma 213 Other light verbs 213 Aspectual serialization

10

Discussion

References

192

206 206 207 209 210

216

218

1 The language scene This sketch deals with the Blagar (ISO 639-3 code: beu) language.1 The name “Blagar” is an exonym and said to be derived from the name of the former kingdom of Balagur. The kingdom covered the north-east of Pantar and the islands in the 1 I thank the Netherlands Foundation for the Advancement of Tropical Research which financed the project Linguistic Research on the Island of Alor (project number W 38-7). I am especially grateful to my main language assistant, Hendrik Daniel Rudolf Gomang, for his patience, enthusiasm, and friendship. I hasten to add that the description below is the result of my (mis)interpretations. Finally, I thank two anonymous reviewers for their useful remarks. In particular, I thank Antoinette Schapper for her stimulating suggestions and penetrating questions on an earlier version of this sketch. Without her this sketch would not have been written.

Blagar

149

Pantar Strait, of which the largest is Pura. The endonym is simply pi abaŋ hur (1PL.INCL.POSS village language) ‘our village language’. This name is much more in line with the actual linguistic state of affairs: Blagar is not a single monolithic language, but a cluster of related dialects often displaying significant variation. Varieties of Blagar are spoken in all but two of the villages on the island of Pura (a dead volcano of over 1,000 m high), on the smaller island of Tereweng south of Pura, and in a number of villages on the east coast of Pantar opposite Pura and Tereweng. Outside this core area there are scattered groups of Blagar migrants, mainly in the regional and provincial capitals of Kalabahi and Kupang. The two non-Blagar villages on Pura with some 1000 inhabitants are Reta (or Retta) speaking, a related but nevertheless distinct language. All Reta speakers, being a minority in frequent contact with the 4,000 Blagar speakers on Pura, were said to be bilingual in Blagar at the time of my fieldwork, the mid 1970’s. A rough estimate then was that the total number of speakers of Blagar varieties did not exceed 10,000 people. The Blagar speaking villages in the core area are not targets for migration and so the communities remain largely homogeneous. But contacts with the outside world are increasing. Already in the 1970s Blagar was no longer used in the church or the mosque, but had been replaced by Indonesian. At school the use of Blagar was discouraged to put it mildly; Indonesian was the sole language of instruction. For secondary and higher education Blagar people must migrate to Kalabahi or further afield. In the early 2000s, electricity reached Pura Island, and with it came television, another source of influence of Indonesian. As a result, bilingualism in Blagar and Indonesian has become generalized. The language is still transmitted to the younger generations, but in October 2012, elder Blagar speakers I met during a short stay in Kalabahi, complained about the decreasing speaking competence of their grandchildren and the increase of Indonesian in their language. The following grammatical sketch is based on data collected in the mid-1970s with a focus on the variety of Blagar spoken in the village of Dolabang, halfway up on the north-western slopes of Pura. The examples in this paper have been elicited or are quoted from recorded stories. To my knowledge the only local written texts in Blagar are: (i) Gomang & Steinhauer (1976), a translation of a comic strip consisting of text with illustrations, distributed by photocopying in the 1970s and three years ago again, (ii) a translation of the Gospel of Mark: Mutu Yesus (2007), and (iii) an illustrated section of the latter, Mark 4:1–20, Mutu Yesus (2009). Both latter texts are written in a dialect different from the one described here, namely the East Pura variety of Apuri. A dictionary of Blagar exists in manuscript form and is currently being prepared for publication (Steinhauer n.d.). Previous academic publications dealing with specific topics in the Blagar language are Steinhauer (1990, 1993, 1995, 2010). Until now, there has been no general description of the Blagar language.

150

Hein Steinhauer

2 Phonology 2.1 Vowels Blagar has a five vowel system with a very marginal phonemic contrast between short and long vowels (geminates), as presented in Table 1. Table 1: Blagar vowels FRONT

CENTRAL

BACK

i iː e eː

HIGH MID

u uː o oː a aː

LOW

Minimal pairs illustrating the contrastiveness of vowel phoneme qualities are given in (1): (1)

/taŋ/ ‘sea’ ≠ /teŋ/ ‘our eye(s)’ ≠ /toŋ/ ‘our head(s)’ ≠ /tuŋ/ ‘year’; /ta/ ‘wear (on waist)’ ≠ /te/ ‘tree, wood’ ≠ /to/ ‘k.o. bowl’ ≠ /tu/ ‘k.o. quiver’; /ad/ ‘fire’ ≠ /id/ ‘star’; /eŋ/ ‘your (SG) eye’ ≠ /iŋ/ ‘2PL’ ≠ /oŋ/ ‘your (SG) head’; /pa/ ‘cave’ ≠ /pi/ ‘we (INCL)’ ≠ /po/ ‘down there’ ≠ /pu/ ‘tree sp.’

For the most part, vowels are short only. However, phonemically long vowels do occur in a limited number of roots. What is more, the distribution of long vowels over the lexicon is subject to dialectal differentiation. Compare the dialect variants in (2).

(2)

Dolabang /eetaŋ/ [ˈ heːtɑŋ] /ba/ [ba]

Limarahing (northeast Pura) /aetaŋ/ [haˈetɑŋ] ‘laugh’ /baa/ [baː] ‘fall’

Nevertheless, minimal pairs do exist in one and the same dialect, as in the Dolabang examples in (3). (3)

/ʔu/ /seŋ/ /aru/

[ʔu] [sɛŋ] [ˈ haru]

‘there’ ‘finished’ ‘two’

≠ ≠ ≠

/ʔuu/ /seeŋ/ /a~aru/

[ʔuː] [sɛːŋ] [ˈ haːru]

‘k.o. tuber’ ‘cent’ ‘both’ (RDP~two)

Blagar

151

Long vowels in Dolabang are also regularly found in derivations from roots beginning with a vowel, notably in causatives (4a) and intentional verbs (4b). (4)

a.

/anamaŋ/ [haˈnamɑŋ] ‘strong’ ≠ /a-anamaŋ/ [haːˈnamɑŋ] (2SG.CAUS-strong) ‘make you strong’; /ele/ [ˈ hele] ‘big’ ≠ /e-ele/ [ˈ heːle] (2SG.CAUS-big) ‘make you big’; /ololiŋ/ [hoˈloliŋ] ‘ticklish’ ≠ /o-ololiŋ/ [hoːˈloliŋ] (2SG.CAUS-ticklish) ‘make you ticklish’; /umunaŋ/ [huˈmunɑŋ] ‘tame’ ≠ /u-umunaŋ/ [huːˈmunɑŋ] (2SG.CAUS-tame) ‘tame you’;

b.

/ihi/ [ˈ hihi] ‘scoop (water)’ ≠ /ʔi-ihi/ [ˈʔiːhi] ‘scoop (water) intentionally’; /osoŋ/ [ˈ hosɔŋ] ‘rub’ ≠ /ʔo-osoŋ/ [ˈʔoːsɔŋ] ‘rub intentionally’; /udok/ [ˈ hudɔk] ‘strut, support’ ≠ /ʔu-udok/ [ˈʔuːdɔk] ‘strut, support intentionally’.

In word-initial position after a pause, vowels have a slightly aspirated onset. This, however, is non-phonemic and opposed to the phonemic glottal consonants /h/ and /ʔ/. Before syllable-final voiced stops, vowels are slightly lengthened. In closed syllables the realization of the mid vowels is relatively low. In open end syllables, word-initially and as first element of a diphthong, they are realized higher. In other positions (i.e. in open syllables within a morpheme) they tend to be relatively low again, as in (5). (5)

/vet/ /verehak/ /meke/ /be/ /ele/ /mol/ /boʔori/ /molo/ /do/ /ololiŋ/

[vɛt] [vɛˈrɛhɑk ~ vɛˈrehɑk] [ˈmɛke] [be] [ˈ hele] [mɔl] [bɔˈʔɔri] [ˈmɔlo] [do] [hoˈloliŋ ~ hoˈlɔliŋ]

‘coconut’2 ‘strew’ ‘ground’ ‘pig’ ‘big’ ‘banana’ ‘yellow’ ‘correct’ ‘up there’ ‘ticklish’

2 Blagar does not mark number in nouns. Examples of nouns are glossed as singular unless the context suggests plurality. Also (in)definiteness is determined by the context.

152

Hein Steinhauer

The realization of the high vowels is high [i, u] in all positions. After another vowel, /i/ tends to become realized as a glide: /ai/ [haj] ‘your (SG) mouth’, /bui/ [buj], a woman’s name, /odoi/ [hoˈdoj] ‘catch you (SG)’, /eteʔiŋ/ [heˈteʔiŋ ~ heˈtejŋ] ‘see you (SG)’. The same holds for /u/ after a preceding non-front vowel: /bau/ [baw] ‘to cry’, /aʔuŋ/ [ˈ haʔuŋ ~ ˈ hawŋ] ‘good’, /paul/ [pawl] ‘divide’, /toroʔu/ [tɔˈroʔu ~ tɔˈrow] ‘bury’, in contrast to its full vocalic realization after /i/ and /u/: /hiu/ [ˈhiju] ‘second level corner beam’, /ebeuŋ/ [heˈbeuŋ] ‘other’. As the examples show, the diphthongic realization of the /Vi/ and /Vu/ sequences does not affect the position of the word stress (see section 2.4). The realization of the low central vowel /a/ is slightly more back in syllables closed with a sonorant, than in other positions. Thus /zam/ [zɑm, ʤɑm] ‘hour’, but /ʔamo/ [ˈʔamo] ‘that/those (yonder, at speaker’s level)’.

2.2 Consonants The Blagar consonant system is relatively simple, but does contain a few peculiarities as compared to other languages of the Timor-Alor-Pantar family. Table 2 presents the 16 consonant phonemes proper and one marginal consonant phoneme. Orthographic representations are given in brackets. Table 2: Blagar consonants

PLOSIVE IMPLOSIVE FRICATIVE NASAL TRILL LATERAL APPROXIMANT

BILABIAL

(DENTAL-) ALVEOLAR

p

t ̪

d

s

z n r l

b ɓ m

PALATAL

VELAR k

GLOTTAL g

ʔ h

ŋ

(j)

The approximant /j/ only occurs in the interjection /jo/ ‘yes’ and in a few borrowings, such as recently adopted Christian names like Johan (also often spelt Yohan), and /rajat/ ‘the people’ borrowed from Indonesian rakyat. In contrast to the other coronal consonants which are alveolar or post-alveolar, /t /̪ is dental. This articulatory distinction is quite common in Austronesian languages in Indonesia (for instance, in Dawan or Kerinci, see Steinhauer 1993b, forthcoming, and Mckinnon 2011 respectively), to mention just a few. The voiced implosive bilabial stop /ɓ/ is a distinctive feature of Blagar. In cognates from the language variety of Bukalabang in the Pantar mountains

Blagar

153

opposite Pura, Blagar /ɓ/ corresponds to a voiced labio-dental fricative /v/ or to zero (see Steinhauer 1995). The phoneme /z/ in the consonant chart above represents the now obsolete pronunciation [z]. In the mid 1970s this phoneme was still realized as [z] by some older speakers. Younger speakers had replaced it with the affricate [ʤ].3 In presenting examples below I shall write {j} instead of {z} for this phoneme, in accordance with the Indonesian orthography that the Blagar are already familiar with. Probably the most distinctive feature of the Blagar phonological system is the high functional load of the glottal stop /ʔ/. Word-initially and word-medially it is opposed to zero.4 Word-medially between unlike vowels its realization is optional, however. Compare: /liaŋ/ [ˈlijɑŋ] ‘k.o. wickerwork’ versus /liʔaŋ/ [ˈliʔɑŋ ~ ˈlijɑŋ] ‘cooked (of rice)’. There is a morpheme /ʔ/ which refers to a third person (see Table 12 below) and is found in cognates where related languages have [g].

2.3 Phonotactics All phonemes in Table 2 occur word-initially and word-medially. The phonemes /ɓ, g, n, z (ʤ), ʔ/ never occur word-finally, while /h/ and /m/ are rare wordfinally. The obstruent phonemes that do occur word-finally are usually unreleased before a pause or a following word beginning with a consonant. The voiced ~ voiceless opposition for oral stops is usually neutralized in word-final position if the following word begins with an obstruent (with the exception of /ʔ/). That is, the voicing harmonizes across consonants. For instance, only in careful speech would /vet toaŋ/ ‘many coconuts’ be distinguished from /ved toaŋ/ ‘many days’. Word-initial consonant clusters in Blagar are practically restricted to onomatopoeic nouns. The number of examples is limited. The clusters are restricted to /sb, sl, sr, bl, ɓl/ and /ɓr/. 3 Locally there is an interesting story on the origin of this change. I was told that the change had started in the late 1940s or early 1950s, when a boarding school for primary education was opened on the island. The most popular girl at the time among the pupils of all ages – she was good at sports, singing, and also quite attractive physically – propagated the change with the argument that Indonesian assimilated the few words of Arabic/Persian descent containing [z] by replacing [z] by [ʤ], so that was how it should be, and [z] was difficult to write anyway. She was imitated by her fellow pupils and subsequently by their illiterate parents, who reasoned that their children had been to school and therefore must know best. 4 In the translation of the Gospel of Mark (Mutu Yesus 2007) a syllable-initial single vowel indicates that it is preceded by a glottal stop, whereas a geminate vowel in that position marks its absence. The phonemes /b/ and /ɓ/ are both written as {b}.

154

Hein Steinhauer

Some examples are given in (6). For examples with reduplication see (11) and (12) below. (6)

/ɓlas/ /ɓruk/ /blek/ /srr/

‘gurgling sound’ ‘sound of walking in mud/water lifting one’s feet; sound of copulating’ ‘(accompanied with gesture of pulling down one’s lower eyelid) ‘forget it!’ ‘sound of corn falling on stones’

Word-medially consonant clusters are rare in roots of native lexemes. Most of the instances appear to be loanwords, as in (7).

(7)

Blagar /parluː/ /valaŋda/ /binaŋta/ /siriŋta/ /daŋpur/ /kaŋpal/

< < < < <
/e~ele/ ‘rather big’ /user/ ‘fast’ > /u~user/ ‘rather fast’ All other roots pattern: (V0)C1V1C2V2. . . /kiki/ ‘small’ /kaʔana/ ‘black’ /eveniŋ/ ‘long’ /uruka/ ‘dry/crunchy’

> > > > >

(V0)~C1V1~C1V1C2V2 . . . , /ki~kiki/ ‘rather small’ /ka~kaʔana/ ‘rather black’ /e~ve~veniŋ/ ‘rather long’ /u~ru~ruka/ ‘rather dry/crunchy’

With verbs indicating a change of place and with the elevational adjectives ʔetela ‘upper’ and ʔololaŋ ‘lower’, partial reduplication indicates movement or position further in the same direction, as in (18). Other cases involving partial reduplication with the deictic verbs of ‘coming’ and ‘going’ are discussed below (see Table 4).

Blagar

(18)

Reduplication of verbs /ʔipa/ ‘go down’ /da/ ‘come up’ /hera/ ‘descend’ /ʔetela/ ‘upper’ /ʔololaŋ/ ‘lower’

of displacement > /ʔi~ʔipa/ > /da~da/ > /he~hera/ > /ʔe~ʔetela/ > /ʔo~ʔololaŋ/

159

‘go further down’ ‘come further up’ ‘descend further’6 ‘higher up’ ‘further down’

A few inconclusive field notes suggest that partial reduplication with adjectives denoting qualities which are more time-bound (as well as with an intransitive verb such as /hera/ ‘descend’) may have the meaning ‘be in the habit of (being/doing)’. For instance: /dira/ ‘ill’, /di~dira/ ‘be ill off and on’; /hera/ ‘descend’, /he~hera/ ‘be in the habit of descending’. With the intransitive verb /bihi/ ‘run’ partial reduplication adds – as with adjectives – the notion of ‘to some extent’: /bi~bihi/ ‘trot’. Partial reduplication of /lamal/ ‘walk’ is /la~lamal/ with the meaning ‘walk without a goal.’ Of /pelela/ ‘laugh’, partial reduplication results in /pe~pelela/ ‘laugh on and on’. These uses of partial reduplication with intransitive verbs remain to be systematically checked. With cardinal numerals partial reduplication has a distributive effect; compare: /hava ɓuta/ ‘four houses’, /hava ɓu~ɓuta/ ‘each of the four houses’.7 With deictic categories, reduplication of the deictic morph (which has in most cases the shape of partial reduplication of the stem) indicates visibility at the time of speaking. Such reduplication may be repeated if the hearer fails to see what can clearly be observed. Similarly with derived deictic adverbs of manner: /do-laŋ/ (up.there-as) ‘like that/those up there (not necessarily visible)’, /do~do-laŋ/ (RDP~up.there-as) ‘like that/those up there (visible)’, /do~do~dolaŋ/ (RDP~RDP~up.there-as) ‘like that/those up there (visible)’. Reduplication of the basic deictic root (see section 9.3) results in a verb meaning ‘be (visible) at the indicated place’, for instance, /do-ʔe/ (up.there-be) ‘be up there (not necessarily visible)’ versus /do~do/ (RDP~(be.)up.there) ‘be visible up there’ (see also Table 16). Partial reduplication can finally be observed in incidental cases of derivation triggered by often physical similarities between the referents of the nominal (or adjectival) stem and the derivation: /kasi/ ‘pawpaw’, /ka~kasi/ ‘k.o. tree with pawpaw-like leaves’; /koa/ ‘raw’, /ko~koa/ ‘green/blue; (wife) whose bride-price has not been paid off yet’; /kelut/ (several kinds of fish), /ke~kelut/ ‘k.o. plant with leaves like the spots on a kelut fish’; /ʔomi/ ‘its inner part’, /ʔo~ʔomi/ ‘bay’. 6 The expression ved hera (sun descend) refers to a time ‘about 3 o’clock in the afternoon’, while the reduplicated form ved he~hera (sun descend.more) denotes a time ‘between 4 and 5 o’clock in the afternoon’ 7 The equivalent for ‘one’ is nu~nuʔu ‘each (one)’ from the obsolete base nuʔu ‘one’.

160

Hein Steinhauer

2.5.2 Prefixation Prefixation in Blagar is inflectional and derivational. As indicated above, inflectional prefixation is confined to verbal agreement forms. Their form and functions will be discussed in section 6. Blagar derives intentional transitive verb stems by prefixation of the first root or stem vowel: (C1)V1. . . > V1-(C1)V1. . . . Causative stems are derived with a similar process with the addition of a suffix [-ŋ] for polysyllabic roots ending in a vowel or [-naŋ] for monosyllabic roots ending in a vowel. Roots ending in a consonant have only the prefix and no suffix. These patterns are common in the north and north-western dialects of Pura and of the opposite Pantar coast. Historically this reduplicated vowel seems to be the result of a process of assimilation of *a- to the first vowel of the root. In the north-eastern and eastern dialects of Pura the prefixed vowel is merely /a-/: (C1)V1. . . > a-(C1)V1. . . . Some examples are given in (18) to illustrate the phenomenon. In section 6 the patterns will be further discussed.

(18)

Examples of copy vowel prefixation a. /miba/ vs. /-i-miba/ Ɂana n-iva miba. 3SG.SBJ 1SG-mother embrace ‘S/he embraced my mother.’ Naŋ di hula ʔ-i-miba. 1SG also want 3SG-INT-embrace ‘I also want to embrace her.’ b.

/papa/ vs. /-a-papa/ Ɂana ʔe jaŋu papa. 3SG.SBJ 3SG.POSS woman search.groping ‘He searched, groping for his wife.’ Boma jasi ʔaŋu ne jaŋu ʔ-a-papa. old.man bad DEM 1SG.POSS woman 3SG-INT-search.groping ‘That naughty old man searched, groping for my wife.’

c.

/mihi/ vs. /-i-mihi/ N-imaŋ mihi. 1SG-father sit ‘My father sits.’ Ɂana n-i-mihi-ŋ. 3SG.SBJ 1SG-CAUS-sit-CAUS ‘He causes me to sit.’

161

Blagar

d.

/kaʔana/ vs. /a-kaʔana/ Ne koŋdo kaʔana. black 1SG.POSS shirt ‘My shirt is black.’ Ɂana ne koŋdo ʔ-a-kaʔana. 3SG-CAUS-black 3SG.SBJ 1SG.POSS shirt ‘S/he made my shirt black.’

e.

/da/ vs. /-a-da/ Ɂini da. 3PL.SBJ come.up ‘They come up.’ Na ʔi-a-da-naŋ. 1SG.SBJ 3PL-CAUS-come.up-CAUS ‘I caused them to come up.’

2.5.3 Suffixation Blagar has two productive suffixes: /-t/ and /-naŋ/. Both only occur with roots ending in a vowel. The suffix /-t/, which functions as a syntactic linker in complex predicates (often glossed as MAN, ‘manner’), and with adjectives before an expression of gradation (glossed as ‘GRAD’, will be discussed in section 8. The suffix /-naŋ/ has two allomorphs: [-naŋ] occurs after monosyllabic verbal roots, and [-ŋ] after polysyllabic ones. After nouns these derivations function as adjectives denoting a permanent quality (‘having an unspecified relation with [verb]-ing’). Examples are given in (20). (20) /na ma-naŋ/ /nehe ma-naŋ/ /ved ba-naŋ/ [ved baŋ] /kuda bihi-naŋ/ [kuda bihiŋ] /ɓil bihi-naŋ/ [ɓil bihiŋ] /hava dira-naŋ/ [hava diraŋ]

‘import’ ‘immigrant’ ‘the west’8 ‘race horse’ ‘stadium’ ‘haunted house’

< < < < <
ʔe mi aru ‘second’). ‘First’ ʔe holomaŋ (literally: ‘its first thing’) deviates from this pattern, having its own special form. A cardinal numeral may be followed by the enclitic =pi, which indicates that the number is the total, ‘all’. The following short phrases show their use as well as the difference between a cardinal and an ordinal numeral: (72)

hava (ele) tuaru hava tuaru=pi hava ʔe mi aru (ʔaŋu)

‘eight (big) houses’ ‘all eight houses’ ‘(that) second home’

4.3 Relative clauses A Blagar relative clause may have the structure of a predicate or a simple clause. The examples in (73) and (74) illustrate the use of such predicates and clauses as relative clauses. Note that Blagar does not have a relativizer morpheme of any kind, marking the head of the RC or otherwise. A relative clause is obligatorily followed by a determiner (ʔaŋu in the examples).

181

Blagar

(73)

[Boma [ʔ-ene Pulamau]RC ʔaŋu]NP naiŋ veŋ that 1SG.OBJ about old.man 3SG-name Pulamau ‘That old man whose name is Pulamau speaks about me.’

tutuk. speak

(74)

[Hava aru [pi meleŋ ɓeli]RC ʔaŋu]NP ele that big house two 1PL.INCL.SBJ yesterday buy ‘Those two houses that we bought yesterday are not big.’

niaŋ. not

4.4 Determiners The determiner slot can be filled by either a demonstrative or one of three other determiners. The forms of the determiners are given in Table 10. Table 10: Blagar determiners Demonstratives

PROX†

DIST‡

Other determiners

unmarked

collective

SPKR ADDR

ʔaŋa ʔaŋu/ ʔauŋ

ʔanaŋa ʔanaŋu/ ʔanauŋ

‘[the one] (just) mentioned’ ‘[the one] just a while ago’

ɓenaŋ (ʔu) vede ʔu

LEVEL

ʔa(ŋ)mo ʔa(ŋ)do ʔa(ŋ)po

ʔana(ŋ)mo ʔana(ŋ)do ʔana(ŋ)po

‘[the one] as was/is visible’

vala

HIGH LOW

† Proximal (‘PROX’) terms are differentiated in terms of their anchoring the speech participants: ‘SPKR’ is closer to the speaker (glossed simply ‘this’), while ‘ADDR’ is closer to the hearer/ addressee (glossed simply ‘that’). ‡ Distal (‘DIST’) terms are distinguished in terms of elevational height: ‘HIGH’ means closer to the mountain than the deictic center; ‘LOW’ means farther away from the mountain than the deictic center; and ‘LEVEL’ means at the same level as the deictic center.

The demonstratives have alternative forms, which do not seem to imply different meanings; the differences seem to be a matter of individual preference. Speakers who use ʔauŋ and ʔanauŋ at all tend to prefer them for anaphoric reference. The forms with inserted -ŋ- are commoner for the ‘collective’ set than for the unmarked one. The determiners are derived by prefixation of ʔa(ŋ)- from a set of deictic particles, which have various functions (see sections 7.3, 9.3 and 10 of this sketch for discussion of different aspects of these). For these particles and verbs see Table 17 in section 10. The deictic interpretation of all these forms depends on the spatial form of reference (see Steinhauer 1990 for details). The glosses indicated in the table are appropriate for the most common speech

182

Hein Steinhauer

situation: referring to a location on a flat or sloping surface between the top of the mountain and the shore.

4.5 Possession Possessor NPs precede the NP which indicates the possessum. Blagar distinguishes alienably and inalienably possessed nouns. Alienable possessors are expressed by means of a free possessive marker and inalienable possessors are expressed by an obligatory person prefix (though the second person singular prefix is zero). Blagar thus preserves the direct vs. indirect possessor contrast that was present in pTAP (Donohue & Schapper 2008). The forms for marking both inalienable (direct) possessors and alienable (indirect) possessors are given in Table 11. The markers for both types of possession (with the exception of the “all inclusive” prefix t- and the “insulting” prefix m-) are also used in nonpossessive contexts (see section 6 for details). Table 11: Coding of possessors NUMBER

PERSON

INALIENABLE POSSESSORS

ALIENABLE POSSESSORS

SINGULAR

1 2 3 1INCL 1INCL+ 1EXCL 2 3

nØʔpit-† niiʔi-

ne e ʔe pi – ni i ʔi

tm-

te –

PLURAL

RECIPROCAL INSULTING

† The “all inclusive” prefix t- (glossed as ‘1PL.INCL+’), not only refers to the speaker and the hearer(s), but also to all members of the group to which they belong.

Alienably possessed nouns are those nouns that do not have to occur with a possessive prefix. Alienably possessed nouns form a large, open class. On these nouns, the possessor is expressed indirectly by means of an independent possessive marker. Examples of one such noun, hava ‘house’, are provided in (75).

Blagar

(75)

hava ne hava e hava ʔe hava pi hava ni hava i hava ʔi hava te hava

house 1SG.POSS house 2SG.POSS house 3SG.POSS house 1PL.INCL.POSS house 1PL.EXCL.POSS house 2PL.POSS house 3PL.POSS house RECP.POSS house

183

‘house’ ‘my house’ ‘your (sg.) house’ ‘his, her, its house’ ‘our (incl.) house’ ‘our (excl.) house’ ‘your (pl.) house’ ‘their house’ ‘each other’s house

Inalienably possessed nouns form a small closed class, characterized by an obligatory person prefix. As in other TAP languages, the class of inalienably possessed nouns includes body parts, such as -ava ‘chin’ (76a), and kin terms, such as -iva ‘mother’ (76b). (76)

a.

n-ava 1SG-chin ‘my chin’

b.

n-iva 1SG-mother ‘my (classificatory) mother’

As Table 11 indicates, apart from the reciprocal and person-marking prefixes, there is a special prefix m-. This only occurs on the inalienable roots for genitals, male (-al ‘penis’) and female (-ar ‘vagina’). Marked with this prefix, these roots are used as insulting and denegrating terms of address to members of the respective sex, roughly translatable as ‘shut up, keep out of it stupid’, as a reaction to which the addressee will reply in the same terms and leave the scene. The full inflectional pattern of each root is given in (77). The m- prefix is not attested anywhere else in Blagar and has to date not been observed in any other TAP language. (77)

a.

n-al Ø-al ʔ-al pi-al t-al ni-al i-al ʔi-al t-al m-al

1SG-penis 2SG-penis 3SG-penis 1PL.INCL-penis 1PL.INCL+-penis 1PL.EXCL-penis 2PL-penis 3PL-penis RECP-penis INSULT-penis

b.

n-ar Ø-ar ʔ-ar pi-ar t-ar ni-ar i-ar ʔi-ar t-ar m-ar

1SG-vagina 2SG-vagina 3SG-vagina 1PL.INCL-vagina 1PL.INCL+-vagina 1PL.EXCL-vagina 2-PL-vagina 3PL-vagina RECP-vagina INSULT-vagina

184

Hein Steinhauer

Blagar allows double possessive marking in contexts where morphologically inalienable nouns are “alienated” from their possessor. For instance, consider the example in (78). Here we have a possessive prefix n- ‘1SG’ on the inalienable noun -oŋ ‘head’ cross-referenced with an alienable possessor marker ne ‘1SG. POSS’. The possessive prefix refers to the whole of ‘me’ to which the -oŋ part belongs, while ne refers to the actual owner from whom the inalienable part nonetheless has been alienated. (78) Na ipar nehe ne n-oŋ met=ma hava=taŋ mea. 1SG.SBJ dream people 1SG.POSS 1SG-head taken=move house=on put ‘I dreamed that people put my head on the house.’ Non-possessive uses of the alienable possessive markers are discussed in section 5.3.

5 Pronouns Blagar has four paradigms of free pronouns (Table 12). It is not possible to completely analyze the forms of pronouns as strings of meaningful morphemes, although the reader will observe that that the pronominal paradigms are all built with similar elements: n- 1st person, non-inclusive; ʔ- vs. its absence for 3rd vs. 2nd person; p- 1st person inclusive; -e- singular; -i- non-singular; -ole dual; -iŋ finally on the “object” series. See also the person prefixes in Table 13. Table 12: Free personal pronouns

1SG 2SG 3SG 1PL.INCL 1PL.EXCL 2PL 3PL RECP

SUBJECT

OBJECT

POSSESSIVE

DUAL

na ana ʔana pi ni ini ʔini –

naiŋ† aiŋ ʔaiŋ piŋ niŋ iŋ ʔiŋ taiŋ

ne e ʔe pi ni i ʔi te

– – – pole nole ole ʔole –

† Before the enclitic =di ‘also’ the singular forms lose the -i-, becoming naŋ=di etc.

The SUBJECT paradigm is used exclusively for encoding subjects. The OBJECT paradigm can be used for encoding almost any non-subject role. The POSSESSIVE paradigm is homonymous with the alienable possessor markers (given in Table

Blagar

185

11) and is used to add a range of participants with peripheral thematic roles to the clause. Finally, the DUAL paradigm is used to signal groups of exactly two. In the following sub-sections I will discuss the functional and distributional differences between the four sets.

5.1 Subject pronouns The subject pronouns are only used for encoding subjects, be it with a transitive, intransitive, or equative predicate, as in (79a–c). A subject pronoun can never be used for a non-subject, such as a transitive object (79d) or the complement of a postposition (79e). In these latter two examples na can only be interpreted as ‘thing(s)’, not as ‘1SG’. (79)

a.

Transitive subject Ɂana n-ataŋ pina. 3SG.SBJ 1SG-hand hold ‘S/he held my hand.’

b.

Intransitive subject Na hava ʔaŋa=mi 1SG.SBJ house this=in ‘I live in this house.’

c.

Equative subject ʔini guru. 3PL.SBJ teacher ‘They are teachers.’

d.

Transitive object ʔana na taʔavi. 3PL.SBJ thing steal ‘S/he stole things.’

e.

Complement of postposition ʔana na=taŋ mihi. 3SG.SBJ thing=on sit ‘S/he sat on things.’

mihi. sit

Subject pronouns also cannot be used as an independent utterance, nor can they be followed by an enclitic or any other kind of element of further specification. These restrictions will be discussed further in relation to the object pronouns.

186

Hein Steinhauer

5.2 Object pronouns Object pronouns are named as such for the fact that they are used to encode transitive objects, as in (80). They also frequently appear encoding the nominal complement of a postposition, as in (81).

(80)

a.

b.

(81)

a.

b.

Transitive object Na iŋ taʔaniŋ 1SG.SBJ 2PL.OBJ count ‘I’ll count you first.’

heʔi. first

Ɂini taiŋ harak. 3PL.SBJ RECP.OBJ meet ‘They met each other.’ Nominal complement of postposition Na tia seŋ=ba, da naiŋ=taŋ 1SG.SBJ sleep finish=so, come.up 1SG=on ‘I lay down, so come up and lay down on me.’

tia. sleep

Ɂini taiŋ veŋ tia. 3PL.SBJ RECP.OBJ with sleep ‘They slept with each other.’

Object pronouns can also be used to encode subjects in some circumstances (with the exception of taiŋ which must be bound by a subject antecedent). There are two contexts in which we observe the subject use of object pronouns, both of which are contexts that subject pronouns are excluded from. An object pronoun is also used when the pronoun stands as an independent utterance or where the enclitic =ba is used in contexts of contrastive emphasis. A subject pronoun in these positions is impossible. Example (82) illustrates this use. (82)

Q.

Nuba=ba ʔila? who=EMPH go ‘Who (is the one who) goes?’

A.

Naiŋ. / Naiŋ=ba. (*na) 1SG.OBJ=EMPH 1SG.SBJ 1SG.OBJ ‘Me’ / ‘Just me; let it be me.’

Blagar

187

The object pronouns also must be used in heading a pronominal phrase, another, strictly seen, subject position. Examples of this are given in (83a) and (83b). The pronouns are specified by ʔemaŋpi ‘all’ and ʔaŋa ‘this’, respectively. In other cases, the use of an object pronoun over a subject pronoun cannot be motivated by morpho-syntax. Rather the object pronoun’s use encoding a subject seems to imply extra effort in executing the action referred to by the verb. This latter meaning would explain why in contrast to the third personal pronouns ʔana, ʔini and ʔole the corresponding forms of the second column in Table 11 are not used for inanimate subjects. Another possible interpretation of the object pronoun as subject seems to be a matter of emphasis; cf. (84c) as opposed to the cleft-construction in (84d). (83)

a.

Ɂiŋ ʔemaŋpi 3PL.OBJ all ‘They all went.’

ʔila. go

(*ʔini) 3PL.SBJ

b.

Ɂaiŋ ʔaŋa jaŋu veŋ 3SG.OBJ this woman with ‘He here is not married.’

c.

Ɂaiŋ jaŋu veŋ niaŋ. 3SG.OBJ woman with not ‘(I don’t know about the other possible subjects, but) he is not married.’

d.

Ɂaiŋ=ba jaŋu veŋ niaŋ. 3SG.OBJ=EMPH woman with not ‘He is the one (in contrast to the others) who is not married.’

niaŋ. not

Further research is needed to determine the exact difference between the two options (subject or object pronoun) in subject position.

5.3 Possessive pronouns The possessive pronouns have a variety of functions. I have already discussed their various uses in nominal constructions encoding possessors in section 4. In this section, I focus on their use to optionally introduce various extra participants into the clause. The role of the participant introduced by a possessive pronoun is determined by the semantics of the clausal predicate.

188

Hein Steinhauer

A possessive pronoun can introduce a participant towards whom the action denoted by the verb is directed. This may be either a recipient with verbs of transfer, such as panatu ‘send’ (84), or an entity the subject of the verb wants to get hold of, such as with ʔaru ‘call’ (85) or bari ‘place a fish trap’ (86). (84)

Na buk tue met=ma 1SG.SBJ book three take=move ‘I send three books to my mother.’

(85)

Na n-iva 1SG.SBJ 1SG-mother ‘I call my mother.’

(86)

a.

Na agal bari. 1SG.SBJ k.o.fishtrap place.fishtrap ‘I place a long rectangular fish trap.’

b.

Na agal met=ma kelut ʔe bari. 1SG.SBJ k.o.fishtrap take=move fish.sp 3SG.POSS place.fishtrap ‘I place a long rectangular fishtrap to catch kelut fish’

ʔe 3SG.POSS

n-iva 1SG-mother

ʔe 3SG.POSS

panatu. send

ʔaru. call

With verbs of motion, a possessive pronoun is used to introduce an entity with respect to which the motion changes the distance. Dependent on the context the interpretation may again be ‘to try to reach . . . by [moving]’, ‘to try to get away from . . . by [moving]’, as in (87). (87)

Na 1SG.SBJ ‘I go up ‘I go up

n-iva ʔe mida. 1SG-mother 3SG.POSS go.up after my mother, trying to reach/overtake her.’ (i) trying to get away from my mother.’ (ii)

The ambiguity in interpretation of the possessive pronouns with verbs of motion explains why the possessive pronoun is also used with the adjectives ʔola ‘far from’ and peaŋ ‘close to’: ʔe ʔola ‘far from him/her/it’, ʔe peaŋ ‘close to him/her/it’. With adjectives expressing an emotion or attitude a possessive pronoun can be used to add the entity towards which the emotion or attitude is directed, as in (88) and (89).

Blagar

(88)

Ɂini te aʔuŋ. 3PL.SBJ RECP.POSS good ‘They are well-disposed to each other.’

(89)

Na e tojaŋ. 1SG.SBJ 2SG.POSS glad ‘I am/act glad towards you (i.e., to get s.t. from you)’

189

With such predicates, there is often a contrast between the use of a possessive pronoun and an object pronoun. In (90a) the possessive pronoun introduces the target of the emotion, while in (90b) the object pronoun in the postpositional phrase is the stimulus of the emotion. Similarly, in (91a) the possessive pronoun is again the target of the emotion, whereas in (91b) the object pronoun indicates the affectedness of the undergoer of the emotion put into action. Apparently the choice of the pronoun entails the interpretation of the verb. (90)

(91)

a.

Ɂana ne ʔiliʔil. 3SG.SBJ 1SG.POSS angry ‘S/he is angry at me.’

b.

Ɂana naiŋ veŋ ʔiliʔil. 3SG.SBJ 1SG.OBJ about angry ‘S/he is angry because of me.’

a.

Ɂana ne pakaŋ. 3SG.SBJ 1SG.POSS have.bad.intentions ‘S/he has bad intentions towards me.’

b.

Ɂana naiŋ pakaŋ. 3SG.SBJ 1SG.OBJ act.foul ‘S/he treats me badly.’

Finally, there are two further, less frequent functions attested for possessive pronouns. First, a plural possessive pronoun (pi, ni, i or ʔi) occasionally appears coreferent with a reciprocal (the object pronoun taing or the reciprocal prefix t-). In this environment, the possessive pronoun functions to explicitly indicate the persons involved in the reciprocal activity. In (92a) the most likely interpretation of t- is 1st person plural (all-inclusive). In (92b), however, the prefix ʔi- before tindicates that the latter marks reciprocality.

190 (92)

Hein Steinhauer

a.

Pi ɓelevaŋ t-oʔal 1PL.INCL.POSS enemy 1PL.INCL+-child ‘Our enemies steal our children.’

b.

Pi ɓelevaŋ ʔi-t-oʔal taʔavi. 1PL.INCL.POSS enemy 3PL-RECP-child steal ‘Our enemies steal each other’s children.’

taʔavi. steal

However, this use of the possessive pronoun is not limited to contexts where there is ambiguity. For instance, in (93) the possessive pronoun seems to be redundant. I have not been able to discover a semantic correlate other than the marginal possibility to disambiguate reciprocal from the all-inclusive t- (see section 4.5 on this distinction). Further research is necessary therefore, also to establish possible distributional constraints. (93)

Ne aru ia mutu ʔi taiŋ veŋ tia. person two junior senior 3PL.POSS RECP.OBJ with sleep ‘The two brothers, the younger and the older, slept together.’

Second, there is a marginal use of the possessive pronoun where it appears to encode that it is the subject’s turn to execute the action of the predicate, as in (94). To what extent such a construction is possibly more than marginal has to be subject to further research. (94)

Ɂi tia, na ʔila. 3PL.POSS sleep 1SG.SBJ go ‘Let them sleep (it’s their turn to sleep), I go.’

5.4 Dual pronouns Like several other TAP languages, Blagar has a paradigm of dual pronouns. On their own dual pronouns are used for subjects, as in (95). In other roles, a dual pronoun always occurs in combination with a plural person marking element (either a free pronoun or a prefix), as in (96) and (97).

(95)

Subject use of dual pronoun ʔole tia. 3DU sleep ‘The two of them slept.’

191

Blagar

(96)

(97)

Dual and object pronoun N-iva ʔole ʔiŋ bue. 1SG-mother 3DU 3PL.OBJ hit ‘My mother hit the two of them.’ Dual and possessive pronoun pole pi hur 1DU.INCL 1PL.INCL.POSS word ‘the words of us two’

In Blagar, dual number is not obligatory. Plural pronouns may be used instead of dual pronouns in reference to only two participants.11 Typically, once dual reference has already been established in the clause, plural pronouns are used. For instance, in (98a) the 3rd person dual pronoun is used initially, but the pair is subsequently referred to with the 3rd person plural (98b). (98)

a.

Ɂole ʔila be ʔ-eteri. pig 3SG-hunt 3DU go ‘The two of them went hunting pigs.’ [. . .]

b.

Ɂini be ʔ-eteri tatalaŋ=di=se, ʔini dapa 3PL.SBJ pig 3SG-hunt how=also=when 3PL.SBJ get ‘How ever they tried to hunt pigs, they did not get one.’

niaŋ. not

The Blagar equivalent of the English expression ‘[Pronoun]SG and [NP]SG’ makes use of the dual pronouns. For instance: (99)

Nole n-iva ɓil pusi=mi 1DU.EXCL 1SG-mother garden dense=in ‘My mother and I went to the forest.’

ʔila. go

Finally, dual pronouns can be suffixed with the morpheme -ve, as in (100). The meaning of this morpheme, however, has not been established and no restrictions on its use have been observed. Further research is required.

11 The existence of the dual in Blagar indicates that the plural is perhaps better referred to as a “non-singular”. However, throughout this chapter these non-singular forms are glossed simply as ‘PL’.

192 (100)

Hein Steinhauer

Ɂole-ve tia ʔaŋu, ʔi-eŋ baina niaŋ. sleep that 3PL-eye tired not 3DU-? ‘While the two of them lay down, their eyes were wide open.’

6 Prefixation: agreement, causation and intentionality There are three kinds of interrelated prefixes in Blagar. They are given in Table 13 hereafter. The reader is referred to section 2 for discussion of the morphophonemics of these prefixal sets. In this section, I will concentrate on their morphosyntactic properties. The appearance of agreement prefixes on verbs and nouns has been discussed already in section 3.1.2 and section 4.5 respectively. As such, their roles in the different word classes will only be repeated here in brief. On verbs, an agreement prefix co-indexes a P argument (but never A or S). On nouns, an agreement prefix marks an inalienable possessor. Importantly, there are two respects in which the agreement prefixes on nouns differ from those on verbs: (i) on verbs there is no “all-inclusive” category, i.e., the prefix for the first person inclusive is only pi-, while t- can only be used with reciprocal meaning, and; (ii) the insulting prefix m- can only be used with inalienable nouns (for genitals) and is entirely absent on verbs. Table 13: Blagar prefixes Agreement prefixes

1SG 2SG 3SG 1PL.INCL 1PL.EXCL 2PL 3PL RECP/1PL.INCL+

nØʔpiniiʔit-

Causative prefixation (monovalent roots) Non-verbal roots

Verbal roots

n-VØ-Vʔ-Vpi-Vni-Vi-Vʔi-Vt-V-

n-V- ~ -ŋ Ø-V- ~ -ŋ ʔ-V- ~ -ŋ pi-V- ~ -ŋ ni-V- ~ -ŋ i-V- ~ -ŋ ʔi-V- ~ -ŋ t-V- ~ -ŋ

Intentional prefixation (transitive verbs)

n-VØ-Vʔ-Vpi-Vni-Vi-Vʔi-Vt-V-

In what follows, I discuss the two remaining kinds of Blagar prefixation: causative prefixation (section 6.1) and intentional prefixation (section 6.2). In addition, I briefly discuss a process of prefixal defaulting to the 3rd person (section 6.3).

Blagar

193

6.1 Causative prefixation The causative prefix is V-, an unspecified vowel that harmonizes with the first vowel of the root to which it attaches. Causative prefixation in Blagar derives a two-place predicate with a causer subject and causee object (co-)indexed by an agreement prefix from a one-place predicate with an undergoer subject. Causative prefixation cannot occur on a two-place predicate (i.e., a transitive verb). Two distinct kinds of causative derivation can be distinguished. 6.1.1 Non-verbal roots Causative prefixation with a non-verbal root derives a predicate ‘cause to be X’ from a predicate ‘be X’. The input for the derivation may be an adjective (101), a noun (102), or a numeral (103).

(101)

(102)

Adjectival causativization a. Stative Jar ɓara. water hot ‘The water is hot.’ b.

Causative Na jar Ɂ-a-ɓara. 1SG.SBJ water 3SG-CAUS-hot ‘I made the water hot.’

a.

Stative Ɂaŋu

hava. house ‘That’s a house.’

DEM

(103)

b.

Causative Na ʔaŋu ʔ-a-hava. 1SG.SBJ DEM 3SG-CAUS-house ‘I made that into a house.’

a.

Stative Ab ʔaŋu ɓuta. fish DEM four ‘There are four fish.’

194

Hein Steinhauer

b.

Causative Ɂana ab ʔaŋu ʔ-u-ɓuta. 3SG.SBJ fish DEM 3SG-CAUS-four ‘S/he caused that fish to become four.’

The causee may also be coreferential with the causer, as in (104) and (105). In this case the meaning expressed by the causative derivation is ‘become, turn into X’, rather than ‘cause oneself to be X’, at least with nouns. My data do not contain unambiguous examples with adjectival roots. (104)

Na n-o-dokter. 1SG.SBJ 1SG-CAUS-doctor ‘I become a doctor.’

(105)

Ib ʔ-o-toka. caterpillar 3SG-CAUS-cocoon ‘The caterpillar turned into a cocoon.’

Causative prefixation differs from the analytic causative which is formed with the verb ʔeniŋ ‘make’. Compare the causatives with the adjective kaʔana ‘black’ in (106). Derived stems have a causative meaning with a perfective connotation in that the implication is that the end result is clearly envisaged (106a). By contrast, the analytic construction suggests indirect causation and does not have a perfective connotation (106b). (106)

a.

b.

Causative prefixation Na ʔe pakiaŋ 1SG.SBJ 3SG.POSS clothes ‘I made his clothes black.’

ʔ-a-kaʔana. 3SG-CAUS-black

Analytic causative Na ʔe pakiaŋ ʔeniŋ kaʔana. 1SG.SBJ 3SG.POSS clothes make black ‘I caused his clothes to become black.’

6.1.2 Verbal roots Intransitive verbal roots can also undergo causative derivation. On these, the causative prefix V- is accompanied – with roots ending in a vowel – by a suffix -ŋ (-naŋ after monosyllabic roots). The resulting circumfix V- ~ -(na)ŋ derives a

Blagar

195

predicate ‘cause to X’ in which the causee object is, as on other root types, indexed by an agreement prefix. Causative derivation of verb roots seems to be limited to a number of semantic types. Most typical are posture verbs, such as in (107) and (108), manner of motion verbs, such as in (109) and (110), and directional motion verbs, such as in (111) and (112). (107)

(108)

(109)

(110)

a.

Jabar mihi. dog sit ‘The dog sat.’

b.

Na jabar ʔ-i-mihi-ŋ. 3SG-CAUS-sit 1SG.SBJ dog ‘I cause the dog to sit.’

a.

Ni 1PL.EXCL.SBJ ‘We slept.’

b.

Ana ni-i-tia-ŋ. 2SG.SBJ 1PL.EXCL-CAUS-sleep-CAUS ‘You caused us to sleep.’

a.

Ɂini ba. 3PL.SBJ fall ‘They fell.’

b.

Ɂini t-a-ba-ŋ.12 3PL.SBJ RECP-CAUS-fall-CAUS ‘They caused each other to fall.’

a.

Ɂini bihi. 3PL.SBJ run ‘They ran.’

b.

Na ʔi-i-bihi-ŋ. 1SG.SBJ 3PL-CAUS-run-CAUS ‘I made them run.’

tia. sleep

12 See footnote 8 on the fact that the suffix is –ŋ instead of –naŋ.

196 (111)

(112)

Hein Steinhauer

a.

Na ja. 1SG.SBJ come.down ‘I came down.’

b.

Na i-a-ja-naŋ. 1SG.SBJ 2PL-CAUS-come.down-CAUS ‘I made you (PL) come down.’

a.

Na e hava=mi mida. 1SG.SBJ 2SG.POSS house=in go.up ‘I came up into your (SG) house.’

b.

Ana e hava=mi n-i-mida-ŋ. 2SG.SBJ 2SG.POSS house=in 1SG-CAUS-go.up-CAUS ‘You (SG) made me come up into your (SG) house.’

6.2 Intentional prefixation The pattern of prefixation of V-, an unspecified vowel that harmonizes with the first vowel of the root, does exist for transitive verbs as well. Semantically, this prefixation adds a sense that the action denoted by the verb root is done with a conscious effort to affect the object referent. Hence, it is called “intentional” prefixation (glossed ‘INT’) here. This is the intuition of native speakers in elicitation; I have no textual evidence to support this. As set out in section 3.1.2, intentional prefixation occurs on Class I transitive verbs, such as loboŋ ‘slurp’ (113) and kavaluŋ ‘embrace’ (114). (113)

(114)

a.

“Neutral” object encoding Na tua loboŋ. 1SG.SBJ palm.wine slurp ‘I slurped the palm wine.’

b.

“Intentional” object encoding Na tua ʔ-o-loboŋ. 1SG.SBJ palm.wine 3SG-INT-slurp ‘I slurped the palm wine intentionally.’

a.

“Neutral” object encoding Ana naiŋ kavaluŋ. 2SG.SBJ 1SG.OBJ embrace ‘You (SG) embraced me.’

Blagar

b.

197

“Intentional” object encoding Ana n-a-kavaluŋ. 2SG.SBJ 1SG-INT-embrace ‘You (SG) embraced me intentionally.’

Intentional prefixation is differentiated here from causative prefixation on the syntactic grounds that it does not cause an increase in valency. However, there is clearly a semantic link between causative and intentional prefixation: both involve increasing agency to the clause, causative prefixation by adding a participant with an agent role (namely A) and intentional prefixation by increasing the agency of the participant already in the agent role. The fact that there is no increase in valency could then simply be ascribed to the nearly absolute restriction whereby no verb in Blagar may have more than two arguments (cf. section 3.1.3 for the only exceptions).

6.3 Prefixal defaulting Verbs which co-index their object with an agreement prefix on the verb may replace the latter with an independent object pronoun followed by the form of the verb with a 3rd person prefix, ʔ- ‘3SG’ for a singular object (115) and ʔi- ‘3PL’ for a plural one (116). (115)

(116)

a.

Prefixal object agreement Jabar ʔaŋu n-eteʔiŋ. dog that 1SG-see ‘That dog sees me.’

b.

Fossilized 3rd person singular object prefix Jabar ʔaŋu naiŋ ʔ-eteʔiŋ. dog that 1SG.OBJ 3SG-see ‘That dog sees me.’

a.

Prefixal object agreement Na i-oboi. 1SG.SBJ 2PL-turn ‘I turned you (PL) around.’

b.

Fossilized 3rd person plural object prefix Na iŋ ʔi-oboi. 1SG.SBJ 2PL.OBJ 3PL-turn ‘I turned you (PL) around.’

198

Hein Steinhauer

This process of defaulting to the 3rd person is not limited to Class II transitive verbs as in the above examples, but also occurs with agreement prefixes on derived causatives. This can be seen, for instance, on the causativized verb root mihi ‘sit’ in (117). (117)

a.

Prefixal object agreement Ɂana n-i-mihi-ŋ. 3SG.SBJ 1SG-CAUS-sit-CAUS ‘S/he caused me to sit.’

b.

Fossilized 3rd person object prefix Ɂana naiŋ ʔ-i-mihi-ŋ. 3SG.SBJ 1SG.OBJ 3SG-CAUS-sit-CAUS ‘S/he caused me to sit.’

Neither distributional constraints nor semantic differences have been found between these two alternatives. The possibility of using the 3rd person agreement prefix for all persons may indicate that there is an incipient process of loss of agreement in Blagar. Further research is required to confirm this.

7 The verbal complex In Section 3, we saw that the template for simple clauses contained a ‘verbal complex’ (VCOMPLEX), but it was left undefined. The structure of the verbal complex is set out in Figure 5. It consists minimally of a verbal or adjectival head (HEADV/ADJ). This can be optionally preceded by an incorporated postposition (mi or veŋ) (INCORP) and optionally followed by one or more aspect markers (ASPECT) and/or a deictic construction with a pragmatic and/or epistemic function (APPEAL),

VCOMPLEX = INCORP HEADV/ADJ ASPECT APPEAL13 Figure 5: Structure of the verbal complex

13 Although the fillers of the APPEAL slot have a clausal scope they are included here for practical reasons.

Blagar

199

In the following sections I shall discuss the incorporated postpositions (section 7.1), the aspect marker slot and the use of deictic verbs as aspect markers (section 7.2), and finally the role of determiners in appealing to the hearer and related to that in expressing degrees of certainty of the speaker (section 7.3).

7.1 Incorporated slot A number of features distinguish an incorporated postposition from a postposition heading its own postpositional phrase. First, whilst the head of a postpositional phrase is cliticized to the preceding NP, an incorporated postposition does not cliticize in this manner. For instance, compare (118) and (119). (118)

Jabar ʔapo=mi bihi. [. . . ʔaˈpo. . .] DEM .LOW=in run dog ‘The dog runs down there.’

(119)

Jabar ʔapo mi bihi. [. . . ˈʔapo. . .] DEM. LOW in run dog ‘That dog down there runs not really hard.’

Second, the behavior in fronting is different. In (120) the postclitic =mi remains part of the noun phrase when that is fronted. In (121) incorporated mi does not follow the fronted noun phrase but remains where it was: immediately preceding the verb. (120)

Jabar ʔapo=mi ʔana bihi. DEM.LOW=in 3SG .SBJ run dog ‘The dog down there, it runs.’

(121)

Jabar ʔapo ʔana mi bihi. DEM. LOW 3SG. SBJ ATT/ in run dog ‘That dog down there, it runs not really hard.’ (i) ‘That dog down there, it runs in (it).’ (ii)

7.1.1 Incorporated mi As a postposition mi has a locative meaning ‘in, at’, which in the presence of a directional verb of locomotion and depending on the wider context can be interpreted as ‘to’ or ‘from’:

200 (122)

(123)

Hein Steinhauer

a.

Ana taʔaŋ=mi mihi? 2SG.SBJ which=in sit ‘Where do you live?’

b.

Ana taʔaŋ=mi ʔila? 2SG.SBJ which=in go ‘Where do you go?’

c.

Ana taʔaŋ=mi hoʔa? 2SG.SBJ which=in come ‘Where do you come from?’

Na Melaŋvala=mi Dolabaŋ=mi ʔila. 1SG.SBJ Melaŋvala=in Dolabaŋ=in go ‘I went from Melangvala to Dolabang.’

Preceding other verbs (as in (121)) mi may still have the postpositional meaning ‘in’, whether or not it is preceded by a noun phrase. But it also may have an attenuating meaning: ‘not really, perfunctorily, for the form’; in short, it weakens the intentionality of the verb. The following is an example. I gloss mi as ATT (attenuating). (124)

ʔake

bau=t ʔulaŋ, na maŋ mi cry=MAN like.that 1SG.SBJ only ATT ‘Don’t cry like that, I only hit perfunctorily’

PROH

bue. hit

7.1.2 Incorporated veŋ As we saw in section 3.2.2, the many interpretations or translations of veŋ (and =veŋ) cannot easily be reduced to one or two meanings. As an enclitic =veŋ can be interpreted as ‘at (time); at, against, touching (location); with (instrumental)’. As a non-cliticizing particle its senses are ‘with (associative), having; via, through; about’. (125a–f ) illustrate these semantic nuances. Especially in these latter cases the common notion seems to be one of indirect relationality or limited affectedness. (125)

a.

ʔaŋu=veŋ ʔana ʔila. that=at 3SG.SBJ go ‘At that time s/he went.’

Blagar

b.

ʔana n-ia=veŋ bue. 3SG.SBJ 1SG-leg=at hit ‘S/he hit at my leg.’

c.

ʔana ʔ-ia=veŋ Dolabaŋ=mi 3SG.SBJ 3SG-leg=with Dolabaŋ=in ‘S/he went on foot to Dolabang.’

d.

ʔana jaŋu veŋ 3SG.SBJ woman with ‘He is already married.’

e.

ʔana via ʔaŋu veŋ ʔila. 3SG.SBJ road that via go ‘S/he went along that road.’

f.

ʔana jaŋu ʔaŋu veŋ 3SG.SBJ woman that about ‘S/he talks about that woman.’

201

ʔila. go

ʔila. already

tutuk. talk

With transitive verbs of Class I (see section 3.1) the object may be optionally followed by non-cliticizing veŋ. Dependent on the meaning of the verb this denotes limited affectedness of the referent of the object or indirect relationality (both glossed here as LMT). Compare (126a and b) and (127a and b). (126) a. Pi ʔila boma guru ʔe haŋi taʔavi. 1PL.INCL.SBJ go old.man teacher 3SG.POSS chicken steal ‘We (incl.) went to steal the teacher’s chicken(s).’ b. Pi ʔila boma guru ʔe haŋi veŋ taʔavi. 1PL.INCL.SBJ go old.man teacher 3SG.POSS chicken LMT steal ‘We (incl.) went to steal from the teacher’s chickens.’ (127) a. Ved kanak ni jaŋu arapiŋ. sun each 1PL.EXCL.SBJ woman search ‘Each day we (excl.) search for women.’ b. Ved kanak ni jaŋu veŋ arapiŋ. sun each 1PL.EXCL.SBJ woman LMT search ‘Each day we (excl.) do research about women.’

202

Hein Steinhauer

Preceding adjectives, incorporated veŋ also conveys the notion of limited affectedness; they indicate only partial presence of the quality denoted by the adjective: (128)

a.

Ne deko ʔ-ia kaʔana. 1SG.POSS trousers 3SG-leg black ‘The legs of my trousers are black.’

b.

Ne deko ʔ-ia veŋ kaʔana. 1SG.POSS trousers 3SG-leg LMT black ‘The leg(s) of my trousers is/are partially black.’

In these examples, (126) to (128), veŋ is not an enclitic and is analyzed as a pre-predicate, incorporated in the following verb or adjective. With some verbs of Class I the latter analysis is the only possibility. These verbs regularly occur without object (as in 129a versus 129b), or – in the case of verbs of saying – are followed by a quote, usually introduced by hula ‘want, say’ as in (130). The incorporated veŋ is glossed here as ‘about’. (129)

(130)

a.

Veŋ meʔe! Veŋ ani, about listen about remember ‘Listen! Remember, don’t forget!’

ʔake

veŋ about

PROH

tejebiŋ! forget

b.

Na ne abaŋ 1SG.SBJ 1SG.POSS village ‘I remember my village.’

a.

ʔana bau-ŋ=mi veŋ 3SG.SBJ weep-NMLZ=in about ‘In tears s/he said (quote . . .)’

b.

ʔana n-amal=mi veŋ banaŋ (hula . . .) say 3SG.SBJ 1SG-voice=in about say ‘In answer to me (lit. to my voice) s/he said (quote . . .)

c.

ʔipa ʔe veŋ banaŋ go.down 3SG.POSS about say ‘ʔipa means ‘go down there’.’

veŋ about

ani. remember

banaŋ say

(hula . . .) say

hula say

ʔapo=mi

DEM. LOW=in

ʔila. go.LOW

Blagar

203

7.2 Aspect slot There are two verbs which can fill the ASPECT slot in Figure 5: ʔila ‘go’ and seŋ ‘finish’. As Table 14 shows they can also be combined: seŋ ʔila. Table 14: Aspect marking At the period of orientation

Longer ago, usually with indirect evidence

Recent, usually with direct evidence

Approximate meaning ‘to . . .’

NON -PAST

~

PAST

ATELIC

~ seŋ

~ (t) seŋ

‘to have finished to . . .’

TELIC

~ seŋ ʔila

~ (t) seŋ ʔila

‘to have completely finished to . . .’

NEUTRAL

~ (t) ʔila

‘to have started already (and possibly finished) to . . .’

Explanatory notes to the table. (i) Any expression of time (whether a moment or a stretch of time – but not a duration) in the non-subordinate context functions as “period” of orientation. If such an expression is lacking, the moment of speaking functions as such. (ii) “Telic” means that for the moment there will be no continuation or repetition of the activity indicated by the verb (there being no return so to speak), whereas for the “atelic” category this is left open. (iii) The suffix –t has an evidential function and will be discussed in connection with other functions of this suffix in section 8. It only occurs after verbs ending in a vowel, hence the parentheses in the table. There are two non-verbal items which can occur in the ASPECT slot: (i) sehi marking progressive aspect (glossed ‘PROG’), and (ii) heŋ marking a range of non-progressive aspectual distinctions (glossed ‘INC’). The progressive marker sehi is used in contexts where there should be no indication about location denoted by the predicate verb. In (131) the second speaker in the answer in (131b) starts indicating the location of the mother at the speech moment, but subsequently keeps his interlocutor away from that location by emphasizing that his mother is continuing to be asleep. If he would have used a deictic progressive aspect marker of the ʔaʔe series (see section 9.3), such as doʔe in the answer in (131c), she would probably have been disturbed.

204

Hein Steinhauer

(131)

a.

N-iva ʔaʔe? 1SG-mother be.here ‘Is my mother here?’

b.

Ø-iva ʔaʔe, aruŋe ʔana tia 3SG.SBJ sleep 2SG-mother be.here but ‘Your mother is here, but she is sleeping.’

c.

sehi. PROG

Ø-iva ʔaʔe, ʔana doʔe tia. 2SG-mother be.here 3SG.SBJ be.HIGH sleep ‘Your mother is here, she is up there sleeping.’

The inchoative marker heŋ ‘nearly, be about to’ is the only aspectual marker which can be used with expressions indicating the future, such as tobaŋ ‘tomorrow’. (132)

N-iva, ni ʔila heŋ. INC 1SG-mother 1PL.EXCL.SBJ go ‘Mother, we (excl.) must go now.’ (the speaker is already making moves to go)

(133)

Unuar hera heŋ. rain descend INC ‘It is starting to rain.’

Heŋ often occurs with the verbal perfective aspect markers, seŋ and seŋ ʔila. (134) is an example. (134) Na pi abaŋ hur veŋ tejebiŋ seŋ ʔila heŋ. 1SG.SBJ 1PL.INCL.POSS village language about forget finish go INC ‘I (nearly) have forgotten the language of our village completely.’ Further examples with the verbal aspect markers will be given in section 9.3.

7.3 Appeal slot The APPEAL slot in Figure 5 specifies different shades of pragmatic and epistemic appeal to the hearer. It can be filled by one of the five deictic particles typical of Blagar or one of the verbal and demonstrative derivations thereof that are given in Table 15. These latter fillers can optionally be preceded by ba, see (135).

Blagar

205

Table 15: APPEAL slot fillers

PROX DIST

SPKR ADDR LEVEL HIGH LOW

Deictic Particles

Demonstratives

Verbs ‘be [there]’

ʔa ʔu mo do po

ʔaŋa ʔaŋu/ʔauŋ ʔa(ŋ)mo ʔa(ŋ)do ʔa(ŋ)po

ʔaʔe ʔuʔe doʔe moʔe poʔe

In (135) the possible deictic fillers of the APPEAL slot are illustrated schematically with ‘down there’ as the deictic element. Needless to say, parallel forms from the other sets of deictics can be equally applied. (135) SUBJECT SUBJECT SUBJECT SUBJECT

PREDICATE PREDICATE PREDICATE PREDICATE

ʔapo po ba poʔe ba ʔapo

‘SUBJECT ‘SUBJECT ‘SUBJECT ‘SUBJECT

PREDICATE PREDICATE PREDICATE PREDICATE

down down down down

there, be careful’ there (as) you know’ there, imagine’ there, you should know’

These constructions imply also certainty about the state of affairs concerned on the part of the speaker. Differences in degree of his/her certainty may be expressed lexically (e.g. by clause-initial ʔoma ‘probably’, deictically neutral), but also by means of demonstrative and particle paradigms of Table 15, before and after the SUBJECT slot of Figure 1: (136)

SUBJECT po PREDICATE ʔapo SUBJECT PREDICATE

‘SUBJECT PREDICATE down there for sure’ ‘Probably SUBJECT PREDICATE down there’

To turn these patterns into real life sentences substitute SUBJECT with n-imaŋ ‘my father’ or bapa ‘crocodile’, and PREDICATE with tia ‘sleep’ or mol na ‘eat a banana’. For more examples the reader is referred to Steinhauer (1990).

8 The –t suffix In this section, I discuss the suffix -t which creates a variety of complex predicative constructions. The suffix can only appear on a root ending in a vowel; roots ending in a consonant do not have a morphological pendant. Although some of its senses are restricted to certain contexts, I have not been able to unite these

206

Hein Steinhauer

senses into one meaning. Table 16 overviews the various functions which have been observed. These are discussed thereafter. Table 16: The functions of the suffix -t in the construction Y-t X Y

X

-t indicates

GLOSS

Verb Adjective Adjective/Verb NPOBJ + Verb Verb

aspect marker expression of gradation extention (duration, frequency, amount) predicate predicate

recent direct evidence of Y X marks degree of Y X marks limitation of Y X marks effect on NPOBJ Y marks manner of X

DEV GRAD LIM LIM MAN

8.1 Evidential marking -t In section 7.2 we saw that before a verbal aspect marker in the ASPECT slot of Figure 5 the suffix -t indicates recent and usually direct evidence about the state of affairs expressed by the preceding verb (glossed as DEV, ‘direct evidence’). The suffix is opposed to Ø (zero); this absence of a suffix (after a verb ending in a vowel) implies that the event was less recent and not directly observed. The nurse coming out of the IC room may inform the waiting family of the death of a relative with (137). Days later the message will reach his more distant relations with (138). (137)

Boma imina-t seŋ. old.man die-DEV finish ‘The old man has died.’

(138)

Boma M. imina old.man M. die ‘Mister M. has died.’

seŋ. finish

This opposition -t vs. its absence is paralleled by the anaphoric markers vala ‘known from observation’ and ɓenaŋ ‘known from being informed’.

8.2 Gradation marking -t In section 4.1 it was indicated that an adjective gets the suffix when it is followed by an expression of gradation, such as in dira-t ʔamara ‘very ill’, dira-t molo

Blagar

207

(ill-GRAD true) ‘really ill’14, dira-t ma=veŋ (ill add) ‘more ill’. Above this suffix was glossed as ‘GRAD’. Also comparison may be considered a matter of degree, as in majaʔa-t haŋi ʔ-udua (white-GRAD chicken 3.SG-egg) ‘white as a chicken egg’.

8.3 Limitation marking -t The set of appropriate referents of an adjective may also be delimited by an expression of frequency or duration, in which case a suffix -t is also obligatory: dira-t mi aru (ill times two) ‘ill two times’. In these cases -t is more properly glossed as ‘LIM’ (limitation). It is the construction following the suffix which indicates the limitation. The following examples illustrate the function of -t ‘LIM’ vs. its absence in determining the syntactic structure of the sentence. In (139a) dira is the head of the predicate and requires the suffix; (139b), however, is an equative sentence, in which ʔ-e luŋ tuŋ aru is not an attribute to dira, so that the suffix stays out. (139)

a.

[Na]SBJ [dira-t ʔe 1SG.SBJ ill-LIM 3SG.POSS ‘I was ill for two years.’

b.

[Ne dira ʔe luŋ]SBJ [tuŋ aru]PRED . 3SG.POSS long year two 1SG.POSS ill ‘My illness lasted two years.’ (lit. ‘my being ill’s duration was two years’)

luŋ long

tuŋ year

aru]PRED . two

Also verbs may be followed by an expression of frequency or duration, or – in case of a transitive verb – amount of the object. Again the suffix –t ‘LIM’ is obligatory. (140) contains an expression of frequency (‘once’), (141) limits the amount of water swallowed as ‘one gulp’, whereas in (142) an expression for the amount of the rice cooked (‘one pot’) is contained within an expression of duration (‘the time needed to cook it’). Finally, in (143), an expression of duration (‘of spittle to not yet dry up’) is contained within another expression of duration (‘of depositing the spittle and letting it dry’).

14 Here the suffix -t also seems to have a similar evidential function as with the aspect markers. In dira molo the evidence is absent, the meaning being ‘as if ill’.

208

Hein Steinhauer

(140)

Ni tirivaŋ na-t mi nu 1PL.EXCL.SBJ fish.sp eat-LIM time one ‘We (excl) ate six-lined perch only once.’

(141)

Na jar muʔu-t muʔu-ŋ 1SG.SBJ water swallow-LIM swallow-NMLZ ‘I swallowed the water in one gulp.’

(142)

Na 1SG.SBJ

tia-t sleep-LIM

[ʔe 3SG.POSS

luŋ long

na thing

kolaŋ. alone

nu. one

liʔaŋ prepare

doa-t cook-LIM

[ara=nu]amount]duration. pot=one ‘I slept as long as it takes to cook a pot of rice.’ (na liʔaŋ ‘cooked rice’) (143) Ini mihi-t [ʔe luŋ puruŋ mea-t [taʔata jeduŋ]duration]duration. not.yet 2PL.SBJ sit-LIM 3SG.POSS long spittle put-LIM dry ‘You (PL) sat down shorter than a glob of spittle needs to dry.’ A transitive verbal head may be followed by an expression of the effect the action has on its object; (144) and (145) are two examples. The obligatory suffix –t in these cases indicates that the action reaches a certain level of intensity, such that it has the indicated effect. With some effort of the mind this –t too can be considered an instance of ‘LIM’. (144) Ɂana uruhiŋ aru ʔ-atapa-t imina. 3SG.SBJ deer two 3SG-shoot.with.arrow-LIM die ‘S/he killed two deer with bow and arrow.’ (145) Na var giri met=ma to=mi batar giri-t kal~kal. 1SG.SBJ stone grind take=move mortar=in maize grind-LIM RDP~fine ‘I ground the maize very fine with the pestle in the mortar.’ More obvious instances of limiting –t are the following examples in which the length of the sarong and the noun hava ‘house’ are the limiting factors. (146)

a.

Na noaŋ ta-t 1SG.SBJ sarong wear-LIM ‘I wear my sarong(s) long.’

eveniŋ. long

Blagar

b.

209

Ɂini t-atara-t hava. 3SG.SBJ RECP-differ-LIM house ‘They differ from each other in their housing.’

8.4 Manner marking -t Especially the deictic verbs of ‘going’ and ‘coming’ are followed by an expression for the pupose/goal of the movement, often consisting of or accompanied by an expression of a location. These deictic verbs do not get a suffix -t as is shown by mida in the following sentence. However, bihi in this sentence is obligatorily suffixed with -t, indicating that bihi specifies the manner of the action of the following verb. Here ‘MAN’ seems to be the obvious gloss. (147) Leki ʔaŋu bihi-t mida(*-t) [ɓil pusi mi ʔ-amajaŋ]AIM. garden dense in 3SG-hide monkey that run-MAN go.up ‘Running the monkey went up to hide in the forest.’ The following sentences are some more examples of this manner-action order. (148)

Ɂana tia-t kakar. 3SG.SBJ sleep/lie-MAN spread-eagled ‘S/he had her/his legs wide apart while sleeping.’

(149)

Basa-t ɓajaul ʔ-eɓele ʔila. 3SG-towards go read-MAN left ‘Read from right to left.’ (lit. ‘go to the left reading’)

(150)

Ne aru ia mutu ɓenaŋ ʔ-eteri-t ʔ-umuruŋ. people two junior senior mentioned 3SG-hunt-MAN 3SG-follow ‘The two brothers, the younger and the older one, followed it hunting.’

(151)

Ɂana unuar hililiŋ=mi hera ʔaŋu doda-t na. roof.edge=in descend that catch-MAN drink 3SG.SBJ rain ‘Catching it s/he drank the rainwater that came down from the edge of the roof.’

In these examples of a manner verb followed by an action verb (147) to (151), the subject of the first verb is also the subject of the second, while the narrated

210

Hein Steinhauer

periods of both actions coincide. A possible aspect marker (which follows the last verb in the string) relates to the combined action. If “verb” is understood as “verb + dependencies” one may call the constructions discussed here serial verb constructions. If not, “serial predicate constructions” would be an alternative. But whichever label one chose, the fact remains that Blagar does not sharply distinguish these constructions from other semantically closely connected syntactic elements. See the next section for some further discussion. To conclude this section I mention a few idiomatic phrases in which verbal elements are overtly connected by the suffix –t, without this being analyzable as in any of the examples above: (152)

a.

ɓeli-t aɓali buy-t sell ‘to trade’

b.

Na ʔ-amal=mi-t hula . . . 1SG.SBJ 3-voice=in-t say . . . ‘In answer to him/her I said . . .’ (lit. ‘I to his voice said . . .’)

c.

mi-t dira in-t ill ‘internally ill’

9 Serial verb constructions Several of the sentences presented as examples in the previous sections contain constructions which can be analyzed as serial verbs. Further research is needed to arrive at a meaningful classification of those constructions. In the following sections I will confine myself to the most salient cases of serialization, namely those which contain grammatically specialized verbs or verb-like lexical units. In section 9.1 I discuss the theme markers =ma, met and met=ma. Section 9.2 describes a small set of inflected verbs which do not form independent predicates. Finally, aspectual serialization is discussed in 9.3.

9.1 Serialization with =ma, met and met=ma The main function of the postposition =ma (which cannot be the head of a predicate) is that it marks the preceding NP as a theme, an entity which is or has

Blagar

211

been moved or transferred possibly by mediation of some other entity. As we saw in section 3.1.3 themes are typical in the expression of three participant events, as for instance in (153). (153)

Na buk=ma e panatu. 1SG.SBJ book=move 2SG.POSS send ‘I sent a book to you.’ (by post or through someone else)

Alternative constructions are: (154)

a.

Na buk met 1SG.SBJ book taken ‘I sent a book to you.’

b.

Na buk met=ma 1SG.SBJ book taken=move ‘I sent a book to you.’

e 2SG.POSS

panatu. send

e 2SG.POSS

panatu. send

The particle met is presumably a contracted form of medi-t (take-MAN). There is no evidence that the enclitic =ma is historically related to ma ‘come.LEVEL’. It is tempting to analyze (154b) as a description of the whole act of sending, encompassing the handling (met), the transport (=ma) and the delivery to the recipient (e). In the two alternative constructions (153, 154a) one of the stages of the process (the handling, or the transport) would then be underexposed. However, all three possibilities are attested in contexts where ‘taking’ and ‘transporting’ are at best metaphorically part of the narrated event, as in (155) in which met=ma can readily be replaced by met or =ma. (155)

Na dar met=ma 1SG.SBJ song taken=move ‘I perform a song for you.’

e 2SG.POSS

meʔe. hear

My field notes contain many examples of comparable constructions with different theme marking. Compare for instance (156) and (157). (156)

Dokter ʔ-ataŋ met=ma ʔ-eŋ doke. doctor 3SG-hand taken=move 3SG-eye gouge.out ‘The doctor gouged his/her eye out with his hand.’

212 (157)

Hein Steinhauer

Na n-imiŋ n-ataŋ tupa 1SG.SBJ 1SG-nose 1SG-hand protuberance ‘I picked my nose with my finger.’

met taken

doke. gouge.out

Many occurrences of met=ma mark an instrument, such as dad ‘fivepronged bamboo arrow’ in (158), na ʔe ul veŋ ‘soap’ in (159), and hemeriŋ ‘knife’ and elut ‘whetstone’ in (160a) and (160b). (158)

Na dad met=ma duŋ ʔatapa. 1SG.SBJ k.o.arrow taken=move bird shoot ‘I shot a bird with a five-pronged bamboo arrow.’

(159) Na na ʔe ul veŋ met=ma n-apaŋ hia. 1SG.SBJ thing 3SG.POSS foam with taken=move 1SG-face wipe.face ‘I wash my face with soap (lit. thing its foam with).’ (160)

a.

Na hemeriŋ met=ma elut=veŋ hia. taken=move whetstone=at whet 1SG.SBJ knife ‘I whetted the knife at the whetstone.’ (manipulating the knife)

b.

Na elut met=ma hemeriŋ=veŋ hia. whet 1SG.SBJ whetstone taken=move knife=at ‘I whetted the knife with the whetstone.’ (manipulating the whetstone)

Both (160a and b) contrast with (160c), in which the manipulating aspect of the event is left unexpressed and the whetstone is merely presented as the location of the action: (160)

c.

Na elut=veŋ hemeriŋ hia. whet 1SG.SBJ whetstone=at knife ‘I whetted the knife at the whetstone.’

Above we saw that human themes do not require a postposition. The use of the postposition met and met=ma with human themes appeared to be unacceptable, perhaps because “handling” people is unusual. However, =ma was accepted. It emphasizes transmission over a distance, as in the following example. (161)

Ɂana ʔe gahiŋ vala=ma 3SG.SBJ 3SG.POSS messenger=move ‘S/he sent his/her messenger to me.’

ne 1SG.POSS

panatu. send

Blagar

213

Further research is required to establish the exact nature of the constraints in the use of these theme markers, and the semantic differences between them as opposed also to their absence.

9.2 Other light verbs Blagar has a few inflecting “light verbs”, which – unlike real verbs – cannot be used as the sole verb of a clause. Some of them function as postpositions, inflecting for the person-number of their NP complement. So far I have identified the following inflecting light verbs in Blagar: (i) -eɓele ‘towards, because of’, (ii) -at ‘be with’, (iii) -aŋ ‘to’ (someone, in a context of asking or selling), (iv) -ot ‘order’, (v) -abuŋ ‘near’, and (vi) -ena ‘on one’s own, self’. A few examples suffice. In (162) and (163) ʔ-eɓele and ʔ-i-aŋ function as postpositions. (162)

Ɂaŋu ʔ-eɓele ni t-abuŋ that 3SG-because.of 1PL.EXCL.SBJ RECP-close.to ‘Because of that they slept close to each other.’

tia. sleep

(163)

Ɂana n-iva n-imaŋ 3SG.SBJ 1SG-mother 1SG-father ‘S/he asked my parents: “. . .”’

hula . . . say . . .

(164)

N-imaŋ n-ot i-at be ʔ-eteri. 1SG-father 1SG-order 2PL-be.with pig 3SG-hunt ‘My father ordered me to hunt pigs with you (PL).’

(165)

Na n-ena jar 1SG.SBJ 1SG-self water ‘I take water myself.’

ʔ-i-aŋ 3-PL-to

taʔaniŋ ask

medi. take

9.3 Aspectual serialization There are two kinds of aspectual serialization in Blagar: (i) with the deictic verbs of Table 14, filling the PROG slot of Figure 1; (ii) with the post-verbal aspect markers filling the ASPECT slot of Figure 5. In section 7.2 the particle sehi was discussed as a marker of progressive aspect. However, progressive aspect in Blagar is usually expressed together

214

Hein Steinhauer

with location by the deictic verbs ʔaʔe ‘be here’ and ʔaʔa ‘be here (visible at the speech moment)’, ʔuʔe ‘be there (near you)’ and ʔuʔu ‘be there (near you, visible at the speech moment)’, etc. (see Table 17). The deictic verbs in question may be used as heads of predicates and as aspectual auxiliaries immediately following the subject NP (and a possible Time expression) in the template of Figure 1. Compare the following examples. In (166) ʔaʔe and ʔaʔa function as regular verbs. Whereas ʔaʔa is inherently bound to the speech moment, ʔaʔe is not; in contrast to ʔaʔa it can therefore be combined with a modal or time expression as in (166c–d). The same differences can be observed in (167) and (168), where the deictic verbs indicate not only location but also progressive aspect of the event expressed by the immediately following verb or predicate. (166)

(167)

a.

Memet Buiveni ʔaʔe. old.woman Buiveni be.PROX ‘Mrs. Buiveni is/was here.’

b.

Memet Buiveni ʔaʔa. old.woman Buiveni PROX.VIS ‘Mrs. Buiveni is here, here she is.’

c.

Memet Buiveni ʔaʔe old.woman Buiveni be.PROX ‘Mrs. Buiveni isn’t here yet.’

d.

Meleŋ memet Buiveni ʔaʔe. yesterday old.woman Buiveni be.PROX ‘Yesterday Mrs. Buiveni was here.’

a.

Ɂana doʔe mihi. 3SG.SBJ be.HIGH sit/stay ‘S/he is/was sitting/staying up there.’

b.

Ɂana dodo mihi. 3SG.SBJ be.HIGH.VIS sit/stay ‘S/he is sitting up there (look!).’

c.

Ɂoma ʔana doʔe mihi. (*dodo) probably 3SG.SBJ be.HIGH sit/stay ‘S/he is/was probably sitting/staying up there.’

d.

Vede ʔaŋa ʔana doʔe mihi. just this 3SG.SBJ be.HIGH sit/stay ‘Just now s/he was sitting up there.’

jeduŋ. not.yet

(*ʔaʔa)

(*ʔaʔa)

(*dodo)

Blagar

(168)

a.

Ɂana poʔe ab aɓali. 3SG.SBJ be.LOW fish sell ‘S/he is/was down there selling fish.’

b.

Ɂana popo ab aɓali. 3SG.SBJ be.LOW.VIS fish sell ‘S/he is down there selling fish (look!).’

c.

Ɂana poʔe ab aɓali niaŋ. (*popo) not 3SG.SBJ be.LOW fish sell ‘S/he isn’t/wasn’t down there selling fish.’

d.

Tobaŋ=se ʔana poʔe ab aɓali. tomorrow=when 3SG.SBJ be.LOW fish sell ‘Tomorrow, s/he will be down there selling fish.’

215

(*popo)

These deicitc verbs cannot be combined with the other aspect markers, discussed in section 7.2. As we saw there, the verb seŋ ‘finish’ indicates that the event expressed by the immediately preceding predicate was finished before the period of orientation reached its end. The verb ʔila ‘go’ in the same ASPECT slot means that the event at least had begun before that moment and possibly was finished. The combination seŋ ʔila finally, indicates the same as seŋ with the additional suggestion that a repetition or resumption of the event is not foreseen. Consider the following dialogue. In (169a) a newly arrived guest is asked whether he would like to eat: the use of ʔila here is the proper aspect choice, since it leaves open the possibility that the guest hasn’t finished his/her eating before coming to the party. If the answer to (169a) is not negative ( jeduŋ ‘not yet’), it is as in (169b): seŋ alone does not exclude the possibility of a second helping, whereas seŋ ʔila implies that the speaker is satisfied. The use of seŋ instead of ʔila in (169a) would be highly inappropriate in the given situation. It could only imply “can I take your plate”. (169)

a.

Ana na na-t ʔila? 2SG.SBJ thing eat-t go ‘Have you eaten already?’

b.

Seŋ. / finish ‘I have.’

Seŋ finish

ʔila. go

(*ʔila)

216

Hein Steinhauer

10 Discussion Within the TAP family, Blagar is distinguished by several features, but at the same time has much in common with the surrounding languages of the Pantar Straits and Bird’s Head. Phonologically, the Blagar vowel system (5 short and long cardinal vowels – section 2.1) is common amongst AP languages. However, the very marginal status of the length distinction in Blagar makes it difficult to reconcile historically with the vowel systems in related languages (see Schapper, Huber & Engelenhoven forthcoming for discussion). Blagar’s bilabial implosive phoneme (section 2.2) stands out, being reported for some varieties of Reta (Laura Robinson pers. comm.), but is otherwise completely absent in the family. It is, however, an apparently areal feature of Lesser Sunda Islands, with implosives found in many Austronesian languages of the region, including Ndhao, Ngadha, Bima and Savu. Still the exact origins of the bilabial implosive in Blagar remain mysterious, since it is not present in the languages Blagar speakers may have been in regular contact with, such as Lamaholot, Makassarese and varieties of Malay. Blagar is also exceptional in the family for the array and productivity of its derivational morphology. Causative prefixes such as Blagar’s (section 6.1) are not uncommon in western Alor languages such as Adang and Klon, but they are typically limited to a handful of verbs (see, e.g., Adang sketch section 6.5). Cognates of the Blagar derivational suffix -(na)ŋ (section 2.5.3) are only found lexicalized on a small set of verbs in Kaera (see Klamer this volume section 5.2.1). On the basis of its similar form and distribution, the inflectional -t suffix occurring on non-final verbs in Blagar appears to be related to the Kaera -(i)t (see Klamer this volume section 7.2), but with a much more frequent appearance and wider semantic range than the Kaera suffix. Finally, although not discussed here in detail, Blagar has a plethora of deictic and elevation terms denoting clusters of spatial and epistemic information about events and situations (Table 17; see Steinhauer 1990 for detailed discussion of them). With a total of 37 elevation-marked forms spread over eleven paradigms, no TAP language has more elevation-marked terms than Blagar (see Schapper forthcoming for a typology of AP elevational systems). The proliferation of elevation terms in Blagar has clearly been achieved by reiterating deictic particles over multiple paradigms of different word classes. Whilst Blagar has been the most exuberant language in doing this, such reiteration of deictic partciles to create multiple paradigms of deictic marked terms is characteristic of Pantar Straits-Bird’s Head languages, such as West Pantar and Adang.

Blagar

217

Table 17: Blagar elevation terms Deictic particles Demonstratives

LEVEL HIGH LOW PROX .SPKR PROX .ADDR

mo do po ʔa ʔu

Manner adverbs Stative verbs

Basic Collective ‘like’

‘be as ‘be as big as’ high as’

‘be as ‘be many/ at’ much as’

‘be at VIS’

ʔamo ʔado ʔapo ʔaŋa ʔaŋu

movaŋ dovaŋ povaŋ ʔavaŋ ʔuvaŋ

monoaŋ donoaŋ ponoaŋ ʔanoaŋ ʔunoaŋ

momo dodo popo ʔaʔa ʔuʔu

ʔanamo ʔanado ʔanapo ʔanaŋa ʔanaŋu

molaŋ dolaŋ polaŋ ʔalaŋ ʔulaŋ

mohukaŋ dohukaŋ pohukaŋ ʔahukaŋ ʔuhukaŋ

moʔe doʔe poʔe ʔaʔe ʔuʔe

Motion verbs

LEVEL HIGH LOW NEUTRAL

‘from DC’

‘to DC’

va mida ʔipa ʔila

ma da ja hoʔa

From a broader typological viewpoint two general observations can be made about Blagar. In the first place, it can be concluded that the distinction between the various word classes is vague. Morphological diagnostics are rare. From adjectives and verbs ending in a vowel nouns can be derived by suffixation of -(na)ŋ; roots ending in a consonant do not have this possibility, with the consequence that many of these roots have both a verbal/adjectival and a nominal meaning. Apart from that there is a considerable degree of syntactic overlap, given the fact that verbs and adjectives (with their dependencies) can readily be used with a gerund-like meaning in positions typical of nouns. Secondly, it will be clear that further research is required to determine the productivity, distribution and semantics of the morphological processes of partial reduplication, and nominalization with -(na)ŋ, and of the syntactic linking suffix -t. All function words, postpositions and their homonyms (if that is what they are), the aspect markers, the deictic particles and their derivations, and the multifunctional possessive pronouns as well as the possessive prefixes require further attention and a more sophisticated semantic analysis than I have been able to offer within the scope of this paper and on the limited basis of my field notes. In short, the above notes can only evoke the picture of a Swiss cheese: holes prevail. I can only implore the reader: (170)

Ɂake sudi-t mida don’t judge-MAN go.HIGH ‘Don’t be too critical.’

hera descend

va go.LEVEL

ma. come.LEVEL

218

Hein Steinhauer

References Donohue, Mark & Antoinette Schapper. 2008. Whence the Austronesian indirect possessive construction? Oceanic Linguistics 47(2):316–327. Gomang, Hendrik Daniel Rudolf & Hein Steinhauer. 1976. Nehe buning. Orang mengong. Suatu cerita dalam bahasa Blagar [Crazy people. A story in Blagar]. Kupang, photocopied manuscript. Klamer, Marian & Antoinette Schapper. 2012. The history of ‘give’ constructions in the Papuan languages of Timor-Alor-Pantar. Linguistic Discovery 10(3): 174–207. Mckinnon, Timothy. 2011. The morphophonology and morphosyntax of Kerinci word-shape alternation. PhD Thesis, University of Delaware. Mutu Yesus Esarenta aaung-hama. Markus esurat mi hurak. Injil Markus dalam bahasa Pura [The story of Jesus written by Mark. The gospel of Mark in the language of Pura]. 2007. Kupang/Jakarta: Unit Bahasa dan Budaya, Gereja Masehi Injili Timor. Mutu Yesus kolil metma nabing meke mi varehak adana. Markus 4:1–20. Bahasa Pura. [Jesus’ parable of the sower. Gospel of Mark 4:1–20]. 2009. Kupang: Gereja Masehi Injili Timor. Pampus, Karl-Heinz. 2001. Mué Moten Koda Kiwan. Kamus Bahasa Lamaholot, Dialek Lewolema, Flores Timur [A dictionary of the Lewolema dialect of Lamaholot]. Frankfurt am Main: Frobenius-Institut. Schapper, Antoinette. forthcoming. Elevational systems in Alor-Pantar languages. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Schapper, Antoinette, Juliette Huber & Aone van Engelenhoven. forthcoming. The relatedness of Timor-Kisar and Alor-Pantar languages: A preliminary demonstration. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Steinhauer, Hein. 1990. Demonstratives in the Blagar of Dolap (Pura, Alor, Indonesia. In Tom Dutton (ed.), Papers in Papuan Linguistics 1, 179–223. Canberra: Pacific Linguistics. Steinhauer, Hein. 1993. Sisters and potential wives: where linguists and anthropologists meet: notes on kinship in Blagar (Alor). In Paul Haenen & Albert Trouwborst (eds), Vrienden en verwanten, Liber Amicorum Alex van der Leeden, 147–168. Leiden/Jakarta: Irian Jaya Studies Project, Department of Languages and Cultures of South-East Asia and Oceania, Leiden University. Steinhauer, Hein. 1995. Two varieties of the Blagar language (Alor, Indonesia). In Connie Baak, Mary Bakker & Dick van der Meij (eds), Tales from a concave world. Liber amicorum Bert Voorhoeve, 265–296. Leiden: Projects Division of the Department of Languages and Cultures of South-East Asia and Oceania. Steinhauer, Hein. 2005. Phonemics and morphophonemics in Nilulat Dawanese. Paper presented at the 3rd International Conference for East Nusantara Linguistics, Leiden, 30 June–1 July 2005.

Blagar

219

Steinhauer, Hein. 2010. Pura when we were younger than today. In Artem Fedorchuk & Svetlana Chlenova (eds), Studia Anthropologica. A Festschrift in Honour of Michael Chlenov. /Сборник статей в чести М.Ф. Членова, 261–283. Мосва/ Иерусалим: Мосты Культуры/ Gesharim. Steinhauer, Hein. n.d. Blagar-Indonesian-English dictionary. Unpublished manuscript, Leiden University. Stokhof, W.A.L. 1983. Names and Naming in Ateita and environments (Woisika Alor). Lingua 61 (2/3): 179–207.

Laura C. Robinson & John W. Haan

5 Adang 1

The language scene

222

2 2.1 2.2 2.3 2.4

224 Phonology 224 Consonants 228 Vowels 230 Syllable structure 231 Stress

3 3.1 3.2 3.3 3.4 3.5 3.6

232 Basic clausal syntax 232 Verbal predicates 233 Nominal predicates 234 Existential constructions Locational and directional constructions 238 Negation 240 Questions

4 4.1 4.2 4.3 4.4 4.5 4.6

242 Noun Phrases 243 Possession 247 Attributes 248 Classifiers 252 Non-numeral quantifiers 253 Relative clauses Demonstratives and determiners

5 5.1 5.2 5.3 5.4

257 Pronouns 258 Subject pronouns 260 Object pronouns 263 Numbered pronouns 265 Alone pronouns

6 6.1 6.2 6.3 6.4 6.5

Pronominal prefixes and valency increasing prefixes 266 Object prefixes 268 Ablative prefixes 268 Allative prefixes 270 Applicative prefix 271 Causative prefix

235

256

266

222

Laura C. Robinson & John W. Haan

7 7.1 7.2 7.3 7.4 7.5 7.6 7.7

Serial verb constructions 273 273 Properties of SVCs 276 Symmetrical SVCs 277 Comitative SVCs 277 Causative SVCs 278 Directional SVCs 278 Instrumental SVCs 279 Theme SVCs

8

Aspect marking

9

Discussion

References

280 282

282

1 The language scene1 Adang (ISO 639-3 code: adn) has about seven thousand speakers living in villages in the Bird’s Head region of Alor, which includes the city of Kalabahi. Adang people are mostly farmers, and they grow corn or rice, or both. They also usually have gardens where they grow bananas, coconuts, candlenut, mangoes, citrus fruits, and other crops. Today the vast majority of Adang speakers are protestant Christians, although there is a small minority of Adang-speaking Muslims. Adang is an endangered language; parents are choosing to teach their children Indonesian (and Alor Malay) instead of Adang. Certain special registers of Adang have disappeared entirely (Haan 2001: 6), and an experimental study shows that even supposedly bilingual young adults are beginning to lose proficiency in Adang (Hamilton et al 2013). Adang is closely related to the language Kabola, also spoken on the Bird’s Head of Alor, and Hamap, spoken in a small pocket just across the bay from Kalabahi. Although we can make generalizations about phonological and lexical innovations that define each language, the languages actually form a dialect chain. To illustrate, we present some newly collected data from one variety of Hamap, three varieties of Adang, and two varieties of Kabola (see Map 4 in Introduction). The dialect spoken in the village of Lawahing, for example, is intermediate between Adang and Kabola, as befitting its location on the border 1 The authors would like to thank the many speakers of Adang who shared their language with us, especially Marlon Adang. We also wish to thank Antoinette Schapper, Sebastian Fedden, and two anonymous reviewers, who all provided valuable suggestions for improving this sketch. Fieldwork was supported by a grant from the US National Science Foundation (NSFSBE 0936887), under the aegis of the European Science Foundation EuroBABEL program.

Adang

223

between the two language groups. Lawahing aligns with dialects of Kabola in merging *r and *l as /l/ (examples (i) and (ii)) and adding epenthetic vowels under certain conditions (examples (ii), (iii), and (vi)). On the other hand, Lawahing aligns with dialects of Adang in reflecting the change of *w > f (examples (iv) and (v)) and in the shift of *j > s (examples (vi) and (vii)). These phonological innovations are best categorized as tendencies rather than rules, as there are numerous exceptions. For example, Lawahing shows *w > w instead of *w > f in the word for ‘bathe’ (viii). Phonologically, all three languages share a loss of proto-Alor-Pantar *g (see Holton et al 2012), and Adang and Hamap further share a loss of *k, as in (xvi), and *w > f, as in (iv), (v), (viii), and (xvii). Kabola differs from the others in adding a final vowel in many words, such as in (ii), (iii), (vi), (xiv), and (xvi). Kabola also frequently has /l/ where Hamap and Adang have /r/, as in (i) and (ii). Phonologically, Hamap differs from the other two in that it often shows /r/ where the other languages show /l/ ((vi), (xv), and (xvi)). The languages can also be grouped by lexical innovations. Adang and Hamap, on the one hand, and Adang and Kabola, on the other, are both about 85% lexically similar. In contrast, Hamap and Kabola are just 75% lexically similar, putting Adang in the middle of the dialect chain lexically. Hamap is geographically isolated from Adang, however, and there are some lexical differences that distinguish Hamap from the other varieties considered here. In the lexicon, for example, Hamap shows innovations in (viii), (xiii), and (xiv). Table 1: Phonological differences of Hamap, Adang, and Kabola

(i) (ii) (iii) (iv) (v) (vi) (vii) (viii) (ix) (x) (xi) (xii) (xiii) (xiv) (xv) (xvi) (xvii)

Wolwal (Hamap)

Kokar (Adang)

Pitungbang (Adang)

Lawahing (Kabola)

Monbang (Kabola)

Poliboʔo (Kabola)

English

dur tarop bɛh foi faʔai asar sob felaŋ ibiŋ doi mari kafeʔ vutu mode -aribuŋ afer ʔai

dur tarɔp beh fɔi faʔai asal sob foil ibiŋ doi mari kafɛ lafuŋ tasoʔ -alibuŋ afɛl ʔai

dur tarɔp beh fɔi faʔai asal sob foil ibiŋ doi mari kafɛ lafuŋ tasoʔ -alibuŋ afɛl ʔai

dul talopo bɛhɛ fui faʔai asala sob wili ib bini nui hanu loŋ tasoʔo -lɛb wɛlɛʔ ʔei

dul tlopoʔ beheʔ wi waʔai jala ʤob wili ib bini nui sono loŋ tejoʔo -leb wele koi

dul tolopo bɛhɛ wui waʔai jala job wili ib bini nui hono loŋ tojoʔo -leb wɛlɛ koi

‘rat’ ‘bone’ ‘leaf’ ‘stone’ ‘soil, earth’ ‘laugh’ ‘sugarcane’ ‘bathe’ ‘star’ ‘mountain’ ‘tears’ ‘grasshopper’ ‘head’ ‘fever’ ‘tongue’ ‘white’ ‘child’

224

Laura C. Robinson & John W. Haan

This grammatical sketch is based in large part upon Haan (2001), which is a grammar of Adang in the form of a dissertation from the University of Sydney. We decided to publish this version here because Haan (2001) is not widely available, and also to present new data and analyses. Where the analysis is different from that of Haan (2001), the new analysis is that of the first author. The data from Haan (2001) are based on the author’s native speaker intuitions, and in this sketch, example sentences with no cited source are from Haan (2001). Data in this sketch come from the Adang villages Pitungbang (most of the new materials) and Kokar (Haan 2001 and a few of the new materials), which are largely similar. The words in Section 2 come from the first author’s lexical database and Haan (2001). Other examples are from the first author’s fieldwork conducted in 2010 and 2011, most of which is derived from speakers responding to various video or photo stimuli. The recordings are to be archived at The Language Archive in Nijmegen, and item names in parenthesis are intended to help the reader locate specific data.

2 Phonology Adang is unlike other Alor-Pantar languages in having an incipient set of postalveolar phonemes arising from a palatalization rule. Adang also has fewer vowel phonemes than most Alor-Pantar languages, with only the cardinal vowels plus a contrast between tense and lax in the mid vowels.

2.1 Consonants Below is the Adang consonant inventory. Phones that are not considered fully phonemic are given in parentheses. Orthographic forms used later that differ from IPA are given in angled brackets Table 2: Adang consonant phonemes

PLOSIVE

LABIAL

ALVEOLAR

pb

td

AFFRICATE NASAL FRICATIVE TRILL LATERAL

m f

n s r l

POST- ALVEOLAR

(ʧ ) ʤ ( ɲ)

VELAR

GLOTTAL

kg

ʔ

ŋ h

Adang

225

The Adang post-alveolar consonants [ʧ ʤ ɲ] arise from a process of palatalization, whereby the alveolar consonants [t d n] are palatalized to their postalveolar counterparts when they follow a diphthong ending in a high front vowel [i], as in (1). (1)

/alveolar/ > [post-alveolar] / [Vi]__

In some cases, the high front vowel which originally conditioned this palatalization has subsequently dropped out (as in (2), leaving minimal pairs between the alveolar and post-alveolar variants. In these cases, then, there is now a phonemic distinction between the alveolar and the post-alveolar variants. (2) *ViCalveolar > VCpost-alveolar The three post-alveolar consonants are in different stages of phonemicization. We consider /ʤ/ (orthographic ) a separate phoneme, but [ʧ] and [ɲ] do not yet have full status as separate phonemes. The voiced post-alveolar affricate /ʤ/ comes from pre-Adang *Vid > Adang Vʤ. Today there is only very occasional evidence of the high front vowels that have given rise to /ʤ/, and the alveolar variant [Vid] is restricted to older speakers. In the recently collected field data, only one 74-year old speaker retained this variant. We consider /ʤ/, then, to be a separate phoneme from /d/ because they contrast for most speakers. Minimal pairs are given in (3). (3)

/falad/ /ud/ /mud/

‘liver’ ‘beeswax’ ‘citrus fruit’

≠ ≠ ≠

/falaʤ/ /uʤ/ /muʤ/

‘sweet potato’ ‘to vomit’ ‘fall’

The voiceless post-alveolar consonant [ʧ] (orthographic ) is not as far along the cline towards phonemicization as /ʤ/. It derives from *Vit > Vʧ. This change is in the process of diffusing through the lexicon and through individual speakers, but there are a handful of cases where it is still possible to get [it] in variation with [ʧ], even among younger speakers. Example (4) is a minimal pair for most speakers. Example (5) shows how the palatalization of Vit > Vʧ is an active process that can occur across morpheme boundaries. Example (6) shows cases where the palatalization is in free variation for speakers, and different forms of the same word may occur, even by the same speaker on different occasions. We can also see in example (6) that an intermediate form retaining the high vowel is possible, where the result is [ViC] instead of the earlier [Vit] or the new [Vʧ].

226

Laura C. Robinson & John W. Haan

(4)

/fet/ ‘pull out’, /feʧ/ ‘sand’

(5)

/ʔai/ ‘child, person’ + /tumɔʔ/ ‘old’ > [ʔaʧumɔʔ] ‘old man, Mr.’

(6)

/baroit/ /apeit/

[baroit] ~ [baroiʧ] ~ [baroʧ] [apeit] ~ [apeʧ]

‘afraid’ ‘tie’

The palatal nasal [ɲ] is derived from *Vin > Vɲ, but this change is not yet complete. Most speakers use both variants, and we know of no words that permit only one or the other variant, as appears to be the case with ‘sand’, shown in example (4). That is, the variant [Vɲ] is almost always found in variation with both [in] and [iɲ], as shown in example (7). (7)

/ahain/ /puin/ /fain/

[ahain] ~ [ahaiɲ] ~ [ahaɲ] [puin] ~ [puiɲ] ~ [puɲ] [fain] ~ [faiɲ] ~ [faɲ]

‘jungle’ ‘hold’ ‘bee’

Because all three variants are possible, we do not regard [ɲ] as a separate phoneme for most Adang speakers. For the second author of this sketch, however, the variants with /i/ do not occur, and /ɲ/ fully contrasts with /n/, as in (8) (Haan 2001: 19). More research is needed to determine whether this change in progress has more fully affected certain villages or speakers. (8)

/fan/ /-tan/

‘to carry’ ‘to ask’

≠ ≠

/faɲ/ /-taɲ/

‘bee’ ‘let go’

In a similar palatalization, the phoneme /l/ is lost when it follows a diphthong ending in a high front vowel /i/ (rule shown in (9)). This is a completed change in certain dialects, such as the dialect of the second author, but the change only affects younger speakers in other dialects (such as the Pitungbang dialect). A few examples are given in (10). (9) (10)

*l > Ø /Vi_# /fail/ /baʔoil/ /ʔuil/

[fail] ~ [fai] [baʔoil] ~ [baʔoi] [ʔuil] ~ [ʔui]

‘burn’ ‘yellow’ ‘skin’

Table 3 shows the distribution of consonants. Since /ʤ/ is derived from preAdang *Vid, a sequence only possible in final position, it only occurs in final

Adang

227

position. In final position, the phoneme /g/ only occurs in onomatopoetic forms. The velar nasal only occurs in syllable-final position. The fricative /s/ does not occur in final position except in loan words, e.g., dos ‘box’. The fricative /f/ does not occur in final position at all. The other consonant phonemes all occur in initial, medial, and final positions. Table 3: Distribution of consonants Consonant

Initial

Medial

Final

p b t d ʤ k g ʔ m n ŋ f s h r l

/par/ ‘mud’ /bala/ ‘soft’ /ti/ ‘tree’ /dɔ/ ‘egg’ – /kafaʧ/ ‘rich’ /gorgor/ ‘owl’ /ʔɔl/ ‘fall over’ /mala/ ‘shy’ /namɛ/ ‘person’ – /feʧ/ ‘sand’ /sah/ ‘bad’ /hafɔʔ/ ‘wash (dishes)’ /raga/ ‘dark’ /lamɛ/ ‘walk’

/alɛpaŋ/ ‘hear’ /aba/ ‘tell a lie’ /tatɔʔ/ ‘cut’ /adɔbɛ/ ‘hard’ – /akɔnaŋ/ ‘massage’ /agɔl/ ‘taro’ /fiʔiŋ/ ‘left (hand)’ /tameʔ/ ‘bedbug’ /bone/ ‘rotten’ /laŋlɛ/ ‘slice (meat)’ /afɛl/ ‘white’ /asal/ ‘laugh’ /mahip/ ‘bitter’ /aroŋ/ ‘swim’ /talaŋ/ ‘six’

/lap/ ‘search’ /tub/ ‘set on fire’ /but/ ‘garden’ /fed/ ‘sun’ /falaʤ/ ‘sweet potato’ /bok/ ‘spear’ /rog/ ‘ko. sound’ /bɔʔ/ ‘stem’ /dum/ ‘smoke (v.)’ /dun/ ‘look’ /darefaŋ/ ‘sweet’ – – /pah/ ‘old’ /tar/ ‘lie down’ /mol/ ‘river’

Most of these distributions can easily be explained with an appeal to the phonological history of Adang. The distribution of the post-alveolar consonants has already been explained. The fricatives /f/ and /s/ are derived from the glides *w and *y, respectively, which did not occur in non-initial position in proto-AlorPantar (Holton et al 2012). The phoneme /ŋ/ comes from proto-Alor-Pantar (pAP) *m and *n, which both became velar in final position. Adang must have subsequently re-introduced /m/ and /n/ in final position. Note that Adang also merged pAP *q, *k, and *g as /ʔ/, with a subsequent re-introduction of /k/ and /g/ via loanwords. While pAP *r did not occur in initial position, there are a small number of words in Adang with initial /r/ today. The following is a list of minimal pairs and near minimal pairs showing the phonemic status of each of these consonants. Voicing contrasts are shown in (11). (11)

/paʧ/ ‘slow’ /tur/ ‘past’ /laka/ ‘headband’

≠ ≠ ≠

/baʧ/ ‘suck’ /dur/ ‘rat’ /raga/ ‘dark’

228

Laura C. Robinson & John W. Haan

Example (12) shows the contrast among the three fricatives. (12)

/afɛl/ ‘white’ /afeŋ/ ‘full’ /sam/ ‘go’

≠ ≠ ≠

/asɛl/ ‘tree stump’ /-aheŋ/ ‘waist’ /ham/ ‘center’

Example (13) illustrates the contrast among the nasals. (13)

/ɛm/ ‘or’ /-tan/ ‘ask’

≠ ≠

/-ɛn/ ‘give’ /-taɲ/ ‘let go’

≠ ≠

/ɛŋ/ ‘sarong cloth’ /taŋ/ ‘sea’

The contrast among the voiceless stops is shown in (14). (14)

/dopaŋ/ ‘stick’ ≠ /botaŋ/ ‘basket’ ≠ /bokaŋ/ ‘to kneel on hands and knees’ /pir/ ‘fruit’ ≠ /kir/ ‘comb’ The contrast among the voiced stops is given in (15).

(15)

/baŋ/ ‘house’ /bar/ ‘bunch, cluster’ /dul/ ‘slippery’

≠ ≠ ≠

/daŋ/ ‘roast’ /gar/ ‘scrape’ /gulu/ ‘squat’

The contrast between /l/ and /r/ is shown in example (16). (16)

/bulaŋ/ ‘patch’ /bɔl/ ‘hollow’

≠ ≠

/buraŋ/ ‘sky’ /-bɔr/ ‘return’

In example (17), we can see that the presence of a glottal stop is contrastive in initial and final positions. (17)

/ʔahaɲ/ ‘cry’ /diʔ/ ‘younger sibling’

≠ ≠

/ahaɲ/ ‘jungle’ /di/ ‘also, too’

2.2 Vowels Adang has seven vowel phonemes, including a distinction between tense and lax mid vowels, as shown in Table 4.

Adang

229

Table 4: Adang vowel phonemes

High Mid

Front

Back

i e

u o ɛ

Low

ɔ a

All the vowels may occur in initial, medial, and final positions, as shown in Table 5. Table 5: Distribution of vowel phonemes

/i/ /e/ /ɛ/ /a/ /ɔ/ /o/ /u/

Initial

Medial

Final

/ifihiŋ/ ‘five’ /elbaroʧ/ ‘afraid of you’ /ɛn/ ‘give you’ /ab/ ‘fish’ /ɔd/ ‘shaky’ /od/ ‘throw a stone at you’ /ul/ ‘moon’

/ʔadid/ ‘split’ /fed/ ‘day’ /lɛt/ ‘far’ /lap/ ‘search’ /lɔl/ ‘climb’ /mol/ ‘river’ /but/ ‘garden’

/ti/ ‘tree’ /bone/ ‘rotten’ /bunɛ/ ‘hot’ /mala/ ‘shy’ /dɔ/ ‘egg’ /to/ ‘breast’ /bu/ ‘betel nut’

The following is a list of minimal pairs showing the phonemic status of each of these vowels. Example (18) shows the phonemic status of each of the front vowels plus /a/. (18)

/bil/ /bel/ /bɛl/ /bal/

‘wild citrus’ ‘dog’ ‘peel’ ‘soak, inundate’

Example (19) shows the phonemic status of each of the back vowels. (19)

/ɔd/ /od/ /ud/

‘shaky’ ‘throw a stone you (with a zero 2SG prefix)’ ‘beeswax’

Adang also has diphthongs. As in many other languages, Adang diphthongs are usually composed of a mid or low vowel followed by a high vowel, as seen in example (20).

230 (20)

Laura C. Robinson & John W. Haan

/ʔafai/ /doi/ /deu/ /afau/ /outi/

‘caustic lime’ ‘mountain’ ‘plant in the ground’ ‘crush, pound’ ‘cassava’

The lax vowels /ɛ/ or /ɔ/ cannot form the nucleus of a diphthong in Adang. In such cases, the vowel sequences are disyllabic. Examples (21) and (22) contrast the diphthong sequences containing tense vowels with the disyllabic sequences containing lax vowels, where a dot . indicates a syllable boundary. (21)

/boi/ /bɔi/

[ˈboi] [bɔ.ˈi]

(22)

/afei/ /ahɛi/

‘pig’ ‘hole’

[a.ˈfei] [a.hɛ.ˈi]

‘fishing hook’ ‘salt’

In initial position, vowels are phonetically lengthened, as in (23). (23)

/ibiŋ/ /ɛi/ /ala/ /ɔd/ /ut/

[iːbiŋ] [ɛːi] [aːla] [ɔːd] [uːt]

‘star’ ‘canoe, boat’ ‘rice’ ‘shaky’ ‘four’

2.3 Syllable structure Syllable structure in Adang is (C)V(C), where the V slot can be filled by a plain vowel or a diphthong. Examples of the various syllable types are given in (24) through (27) below. (24)

(25)

V /a/ /ti/ CV /ba/ /dai/

‘firewood’ ‘tree’

‘necklace’ ‘still, yet’

Adang

(26)

(27)

VC /ab/ /ail/

231

‘fish’ ‘Alor island’

CVC /dal/ /ʔail/

‘sing’ ‘cough’

Adang only allows heterosyllabic consonant clusters, but even these are quite rare. A few of the examples that do occur are given in (28). (28)

/-arfah/ /barpi/ /turlɔ/ /laŋlɛ/

[ar.fah] [bar.pi] [tur.lɔ] [laŋ.lɛ]

‘chin’ ‘all’ ‘eight’ ‘slice meat’

2.4 Stress In contrast to Haan (2001), we analyze stress as phonemic in Adang, as the pairs in (29) through (31) illustrate. Note that we do not usually mark stress in the sketch. (29)

/buˈraŋ/ /taˈburaŋ/

(30)

/ˈmatɛ/ /maˈtɛŋ/

(31)

/ˈfara/ /maˈla/

‘sky’ ‘debate, fight’ ‘large’ ‘language’

‘foam’ ‘shy’

While Haan (2001) considers pairs such as (32) below to contrast in length, we find no systematic difference in length, but there is a noticeable difference in pitch and loudness, which we consider to be stress. (32)

/ʔaˈtaŋ/ /ˈʔataŋ/

‘his/her arm’ ‘get stuck on it/them’

Although vowels are phonetically lengthened in initial position, stress does not necessarily fall on the long vowel, as the examples in (33) illustrate.

232 (33)

Laura C. Robinson & John W. Haan

/iˈbiŋ/ /ɛˈ/

[iː.ˈbiŋ] [ɛː.ˈi ]

‘star’ ‘canoe, boat’

In noun-noun compounds, primary stress falls on the regularly stressed syllable of the first member of the compound, as in (34) and (35). (34)

/ʔap ʔoh/ [ˈʔap ʔoh] shaft arrow ‘arrow shaft’

(35)

/ala mɛʤ/ [ˈala mɛʤ] rice table ‘cooked rice’

In noun-verb compounds, primary stress falls on the regularly stressed syllable of the second member of the compound, as in (36) and (37). (36)

/dil [dil universe ‘night’

pana/ ˈpana] dark

(37)

/paliʔ matɛ/ [paliʔ ˈmatɛ] island large ‘Earth’

3 Basic clausal syntax Like other Alor-Pantar languages, Adang has predicate-final order. Section 3.1 illustrates clauses with verbal predicates, 3.2 discusses clauses with nominal predicates, 3.3 discusses existential constructions, 3.4 discusses locational and directional constructions, 3.5 discusses negation, and 3.6 discusses questionformation.

3.1 Verbal predicates In Adang, word order in simple verbal clauses is SV/APV. Example (38) is intransitive, while example (39) is transitive.

Adang

(38)

Ni 1PL.EXCL.SBJ ‘We sleep.’

(39)

Manu sɛi pataŋ. Manu water boil ‘Manu boiled water.’

233

mop. sleep

Adang allows ditransitive constructions with two objects. Where both recipient and theme are lexical noun phrases, the theme precedes the recipient (examples (40) and (41)). (40)

Rin ahɛi falɛi ta. Rin salt vegetable add ‘Rin adds salt to vegetables.’

(41)

John sɛŋ Rudy ʔ-ɛn. John money Rudy 3.OBJ-give ‘John gave Rudy money.’

Ditransitives can also be formed with a serial verb construction (see Section 7).

3.2 Nominal predicates Equational constructions indicate the equivalence of an entity with a nominal predicate or the inclusion of an entity within a class indicated by a nominal predicate. Equational constructions in Adang are formed by juxtaposing the subject and the nominal predicate. There is no copula or person inflection on the predicate. Examples (42) and (43) show equivalence. Note that word order is important; the first element is the subject and the second element is the predicate. (42)

Ni-maŋ Lukas. 1PL.EXCL.INAL-father Lukas ‘Our father is Lukas.’

(43)

Lukas ni-maŋ. Lukas 1PL.EXCL.INAL-father ‘Lukas is our father.’

234

Laura C. Robinson & John W. Haan

Examples (44) and (45) show inclusion in a class. (44)

Roni guru. Roni teacher ‘Roni is a teacher’

(45)

Sarɛa ab afaŋ nu. mackerel fish type one ‘Mackerel is one type of fish.’

Note that aspectual particles can occur in equational constructions; see section 8.

3.3 Existential constructions The verb -ra may be used to indicate existence (46). The verb -ra is obligatorily prefixed with a pronominal. (46)

Namɛ sarani bitɛ Lahtal person Christian say God ‘Christians say that God exists.’

ʔa-ra. 3.OBJ-be

The verb -ra can also be used to indicate possession, as in (47). (47)

Roni ʔɔ ʔai ʔa-ra. Roni 3.POSS child 3.OBJ-be ‘Roni has children.’ (lit, ‘Roni’s children exist.’)

The verb -ra can also be used to indicate accompaniment. When -ra indicates accompaniment, the pronominal prefix indicates a second participant. (48)

Adi ni-ra. Adi 1PL.EXCL.OBJ-be ‘Adi is with us.’

(49)

Sa sɔ bel ʔa-ra hɛl. 3SG.SBJ 3.REFL.POSS dog 3.OBJ-be descend ‘He put down his dog.’ (lit., ‘He went down with his dog.’; frog story, Baird n.d.)

Adang

235

Predicates with the verb -ra may take aspectual particles (see section 8 for more on aspect marking).

3.4 Locational and directional constructions Location and direction are expressed with verbs in Adang, as with the verb ta ‘be on’ in (50) and the verb lɛ ‘(move) to, toward’ (51). (50)

Tafiʔiŋ buraŋ ta. cloud sky on ‘A cloud is in the sky.’ (lr-adn-20100714-TopoRels)

(51)

Na baŋ lɛ mid. 1SG.SBJ house to ascend ‘I go up to the house.’ (lr-adn-20110420-01-MotionVerbs)

Location and direction verbs can be combined with aspect markers, as in (52) and (53). (52)

Namɛ nu afail abuŋ mi eh. person one fire near in PROG ‘A person is near a fire.’ (lr-adn-20100714-TopoRels)

(53)

Mol sɛi tafoi taŋ lɛ am. river water flow sea to PFV ‘The river water flowed toward the sea.’ (lr-adn-20100820-01-SurreyVerbs)

Location and direction verbs often occur in serial verb constructions (see section 7), as in (54), where the locational verb mi ‘be in, be at’ is combined with the verb mih ‘sit, live’, and in (55) where the locational verb ta ‘be on’ is combined with the dynamic verb lamɛ ‘walk, move, travel’ (54)

Na ʔarabah mi mih. 1SG.SBJ Kalabahi in sit/live ‘I live in Kalabahi.’ (lr-adn-20100714-05-SurreyVerbs1)

(55)

Ɛi matɛ nu taŋ ta lamɛ eh. boat large one sea on walk PROG ‘A large boat is traveling on the sea.’ (lr-adn-20100714-TopoRels)

236

Laura C. Robinson & John W. Haan

Location and direction verbs are transitive, as in (56), or ditransitive, as in (57), but they cannot be intransitive (58), since both a subject and a location must be specified. (56)

Aru banaŋ mi. deer forest in ‘There are deer in the forest.’

(57)

Ince sɛŋ dec mi am. Ince money wallet in PFV ‘Ince put money in the wallets.’

(58) *aru deer

mi in

While location and direction verbs seem to be grammaticalizing as adpositions, they can still occur disjoint from the noun phrases which they are locating in space, as in (59), where the location NP is left-dislocated but the locative verb ta ‘be on’ remains in the predicate. (59)

Mɛja ho, namɛ napah tɛn med ta ol meŋ. red take on lie set table DEF person cloth ‘Someone set a red cloth on the table.’ (or, ‘As for the table, someone set a red cloth on it.’) (lr-adn-20110408-PositionalVs)

Adang also has a set of compound directional and locational verbs derived from nouns or motion verbs plus lɛ ‘to, towards’. They are given in Table 6, along with the roots from which they are derived. Table 6: Adang compound direction and location verbs Directional Root

Semantics

talɛ

ta ‘on, on top of’

up on

madɔŋlɛ

madɔŋ ‘come up’

toward the speaker at a higher elevation, fairly vertical

taʔlɛ

ta ‘on, on top of’

away from speaker to a higher elevation, fairly vertical

adaŋlɛ

adaŋ ‘mountain’

away from the speaker toward the mountain

midlɛ

mid ‘ascend, go up’

away from the speaker to a higher elevation

malɛ

ma ‘come from nearby (level)’

toward the speaker at the same elevation

falɛ

fa ‘go nearby (level)’

away from the speaker at the same elevation

iplɛ

ip ‘descend, not steep’ away from speaker to a lower elevation

lifaŋlɛ

lifaŋ ‘anchor’

away from the speaker towards the sea

hɛllɛ

hɛl ‘descend’

away from the speaker to a lower elevation, fairly vertical

Adang

237

These compound verbs can also modify nouns and co-occur with a demonstrative (60). (60)

Namɛ [bɛ talɛ hɛmɔ]NP fail. person mango up.on that.above sell ‘Someone is selling those mangoes on top (e.g., of the other mangoes).’

Adang has a set of locative deictics ‘here/there’, given in Table 7. These forms contain [ɔŋ] plus an initial consonant indicating deixis, which is the same consonant used in the demonstrative paradigm (section 4.6). These locatives show a two-way division between proximal, or near the speaker, and distal, or away from the speaker. The proximals are based on the paradigmatic consonant [ʔ]. The distals show a further three-way distinction along the vertical dimension. Location above the speaker is indicated with the consonant [t], location below the speaker is indicated with [p], and location level with the speaker on the vertical dimension is indicated with [m]. Table 7: Adang locatives locative deictics proximal distal (above) distal (below) distal (level)

ʔɔŋ tɔŋ pɔŋ mɔŋ

The locative deictics locate an object or an event in space. In example (61), the locative deictic mɔŋ ‘over there’ occurs at the beginning of the predicate to locate it in space. (61)

bana mi tufeh.]PREDICATE [Ti taʔat ho]NP [mɔŋ tree dry DEF there.level forest in stand.PROG ‘The dry stick is standing over there in the forest.’ (lr-adn-20110408-PositionalVs) Locative deictics can also stand alone as predicates (62).

(62)

Aru nu tɔŋ. deer one there.above ‘There is a deer up there.’

238

Laura C. Robinson & John W. Haan

The locative deictics can also modify a noun, often in combination with a demonstrative, to indicate its location with respect to the speaker (63). Because the locative deictics can modify a noun and occur in what may be a serial verb construction (61), they are best considered verbs. (63)

matɛ. [Bel mɔŋ hɛmɔ]NP dog there.level that.level large ‘That dog over there is large.’

3.5 Negation To negate a clause, the negator nanɛ (nɛnɛ for the second author)2 follows the predicate. Example (64) has a verbal predicate, (65) has a locative predicate, and (66) has a nominal predicate. (64)

Ince sam don nɛnɛ. Ince go shop NEG ‘Ince did not go shopping.’ (i.e., it is not the case that Ince went shopping)

(65)

Sopi ʔɛ but mi nanɛ. they NEG garden in NEG ‘They were not in the garden.’ (lr-adn-20120630-01-sketch)

(66)

Heni guru nɛnɛ. Heni teacher NEG ‘Heni is not a teacher.’ (i.e., it is not the case that Heni is a teacher)

To limit the scope of the negator nanɛ/nɛnɛ, it can be combined with the negative particle ʔɛ which precedes the negated element and serves to limit the scope of the negation to that which falls between ʔɛ and nanɛ/nɛnɛ. In example (67), only upapaʔ ‘frog’ is negated, while in (68), only ubaŋ ‘say that, mention’ is negated. When the negative particle ʔɛ precedes the negated element, it must be combined with the negator nanɛ/nɛnɛ. (67)

“Eː na hɔʔɔ ʔɛ upapaʔ nanɛ.” NEG hey thing this NEG frog ‘Hey, this is not the frog!’ (frog story, Baird n.d.)

2 It is not clear how widespread this variant is; it did not occur in any of the newly collected field materials.

Adang

(68)

239

Fali ho ʔɛ u-baŋ nanɛ. price DEF NEG APPL-say NEG ‘The price has not been mentioned.’ (lr-adn-20110408-01-moko)

To express a general prohibition or rule, the negator nanɛ/nɛnɛ is used alone, so that the negation takes scope over the entire clause, as in (69). (69)

Sam don nɛnɛ! go shop NEG ‘No (going) shopping!’

To form a negative imperative, the negative particle ʔɛ is placed after the predicate, as in (70) and (71). It is not combined with nanɛ/nɛnɛ. (70)

Sam don ʔɛ! go shop NEG.IMP ‘Don’t go shopping!’

(71)

Ɂɔ fali u-baŋ ʔɛ. 3.POSS price APPL-say NEG.IMP ‘Don’t talk about its price.’ (lr-adn-20110408-01-moko)

A negative imperative particle haʔai can also be placed in initial position in the clause to make the negative imperative more polite (example (72)). (72)

Haʔai

foi sɛŋ n-ɛn again money 1SG.OBJ-give ‘Please don’t give me money again.’

NEG .IMP

ʔɛ

NEG. IMP

ɛma. please

The negative imperative cannot be used alone. Example (74) is not a grammatical paraphrase of (73). (73)

sam don ʔɛ. NEG .IMP go shop NEG.IMP ‘Don’t go shopping.’ (Marlon Adang, p.c.)

(74)

*haʔai

Haʔai

NEG.IMP

sam go

don shop

The verb aʔai is used to negate possession (75) or existence (76).

240 (75)

Laura C. Robinson & John W. Haan

Ti tree

ɛl CLF

kaʔai small

alɔ two

ʔɔ 3.POSS

beh aʔai bana mi tufeh. leaf NEG.EXT forest in stand.PROG ‘Two small trees without leaves are standing in the forest.’ (lr-adn-20110407-03-PositionalVs) (76)

Guru aʔai. teacher NEG.EXT ‘There is no teacher.’

3.6 Questions In Adang, question words remain in situ, taking the place of the word(s) that they replace. The interrogative anɔ ‘who’ can replace both subject and object arguments and therefore functions as a nominal, as shown in examples (77) and (78), respectively. (77)

Anɔ Nimo bɛh? who Nimo hit ‘Who hit Nimo?’

(78)

Lilo anɔ bɛh? Lilo who hit ‘Whom did Lilo hit?’

The interrogative anɔ ‘who’ may also occur with the subject or object focus marker, as in (79) and (80). (79)

Anɔ so sɛi gelas mi ʔuhuɲ eh? PROG who FOC.SBJ water glass in pour ‘Who is pouring water in the glass?’ (lr-adn-20100607-01-SurreyVideosCore)

(80)

Anɔ fe sa who FOC.OBJ 3SG.SBJ ‘Whom did s/he hit?’

bɛh? hit

Adang

241

The interrogative naba ‘what’ also functions nominally and can replace both subject and object arguments. The subject or object focus markers may occur following naba, as shown in examples (81) and (82), respectively. (81)

Naba (so) a-nɔʔ? what FOC.SBJ 2SG.OBJ-affect ‘What happened to you?’

(82)

Rudi naba (fe) panɛn Rudi what FOC.OBJ do ‘What is Rudi doing?’

eh? PROG

The interrogative tarɔ ‘where’ functions as a nominal, replacing object arguments. Like anɔ and naba, it can occur with the object focus marker (example (83)). (83)

Supi tarɔ (fe) lɛ sam where FOC.OBJ to go 3PL ‘Where have they gone (to)?’

am? PFV

The interrogative tarɔ can be combined with the suffix -ni ‘like’ to question manner, as illustrated by the short dialogue in (84). It functions as an adverb. (84)

A:

kɔd ho harɔt! shirt DEF sew ‘Sew the shirt!’

B:

Harɔt tarɔni? sew how ‘How (to sew it)?’

The interrogative tarɔni is also used to question reason, as in example (85). (85)

Ay tarɔni fe Ay why FOC .OBJ ‘Why is Ay crying?’

ʔahaɲ cry

eh? PROG

The interrogative den is used to question quantity, and it can be adverbial, as in example (86) or predicative, as in (87).

242 (86)

(87)

Laura C. Robinson & John W. Haan

Namɛ pir den lamɛ person CLF how.many leave ‘How many people have left?’

am? PFV

Ɔ ʔai pir den am? 2SG.POSS child CLF how.many PFV ‘How many children do you have?’ (lit., ‘Your children are how many?’)

In (88), it is used to question time and may occur with the object focus marker. (88)

Supi den (fe) lamɛ? when FOC.OBJ leave 3PL ‘When will they leave?’

Yes-no questions in Adang are formed without any special question element. The speaker signals that s/he is asking a question with rising intonation on the final syllable. (89)

Supi lamɛ am? leave PFV 3PL ‘Have they left?’ Tag questions are formed by adding ɛ to the end of the clause.

(90)

Mid ʔ-at ʔoh ɛ? ascend 3.INAL-mouth empty right ‘The next one is Empty Mouth, right?’ (discussing the names of various drums, moving from the least valuable to the most valuable; lr-adn-20110408-01-moko)

4 Noun Phrases The structure of the Adang noun phrase is set out in the template in (91). In Adang, possessive phrases (discussed in section 4.1) precede the head noun, while all other modifiers follow the head noun. After the noun, there may be one or more modifying verbs (section 4.2), a classifier (section 4.3), a numeral, a non-numeral quantifier (section 4.4), a relative clause (section 4.5), and the definite determiner ho or a demonstrative (section 4.5). Note that a modifying verb may occur in two positions, either preceding or following a classifier. (91) (possessor + possessive.pronoun) N V CLF V numeral quantifier REL

DEM /DEF

Adang

243

All the elements in this noun phrase template are optional, including the noun, as in (92), where only the definite marker ho remains of the noun phrase, since the meaning can be inferred from the context. (92)

Ho

so

DEF

FOC.SBJ

tufeh. stand.PROG ‘They are just saved.’ (lit., ‘They stand.’, in reference to very valuable drums which are not sold or traded, lr-adn-20110408-01-moko)

Note that number nu ‘one’ and the definite ho may co-occur, as in (93). In such cases, the referent is definite, and the number nu ‘one’ serves to emphasize that it is singular. (93)

[Bal nu ho]NP [mɛja]NP ta ol eh. table on lie PROG ball one DEF ‘The ball is lying on a table.’ (lr-adn-20110408-PositionalVs)

4.1 Possession In Adang, there are three classes of nouns, depending on their behavior in possessive constructions. One class of nouns is obligatorily possessed, and nouns in this class must occur with possessive pronominal prefixes in all circumstances. We consider this class inalienable. A second class of nouns may never occur with pronominal prefixes. When these nouns are possessed, they occur with independent possessive pronouns. We consider this class alienable. A third class of nouns occurs with pronominal prefixes only when possessed; if they are not possessed, they may occur bare. The set of prefixes that occurs with these optionally possessed nouns is different from the set that occurs with the obligatorily possessed nouns. The possessive markers are given in Table 8. Table 8: Possessives

1SG 1PL.INCL.DISTR 1PL.EXCL 1PL.INCL 2SG 2PL 3 3.REFL

Inalienable possessive prefixes

Alienable possessive pronouns

Optional possessive prefixes

n(a)t(a)nipi(a)iʔ(a)s(a)-

nɔ tɔ ni pi ɔ i ʔɔ sɔ

nɛtɛniɛpiɛɛiɛʔɛsɛ-

244

Laura C. Robinson & John W. Haan

The set of inalienable nouns includes body parts, some kinship terms, and a few other nouns such as -ni ‘name’ and -lɛmɛ ‘place’. The /a/ vowel of the prefix is lost before vowel-initial roots, and the first vowel of a polysyllabic vowelinitial root is lost following the plural prefixes ending in /i/, as illustrated in Table 9. Table 9: Inalienable possessive prefix derivations

3 1PL.EXCL

prefix

example -at ‘mouth’

example -ɔmaŋ ‘father’

example -taŋ ‘arm’

ʔ(a)ni-

ʔat niat

ʔɔmaŋ nimaŋ

ʔataŋ nitaŋ

Example (94) illustrates a typical use of an inalienable possessive prefix to index the possessor of a body part, in this case na-taŋ ‘my hand’. The root taŋ is meaningless without one of the pronominal prefixes from the inalienable possessive set (Table 8). (94)

luh. [Na]NP [na-taŋ]NP 1SG.SBJ 1SG.INAL-hand rub ‘I rub my hands.’ (lr-adn-20100820-01-SurreyVerbs)

In example (95), the reference of the third person prefix is understood from the context. (95)

[Ɂa-taŋ]NP bibil. 3.INAL-hand shake ‘His/her hand shakes.’ (lr-adn-20100820-01-SurreyVerbs)

In example (96), in contrast, the third person pronominal prefix has an explicit antecedent, the noun namɛ ‘person’, forming a complex noun phrase namɛ ʔafɛl ‘a person’s ear’. (96)

mi. [Falɔfaŋ nu]NP [namɛ ʔa-fɛl]NP earring one person 3.INAL-ear in ‘An earring is on a person’s ear.’ (lr-adn-20100714-TopoRels)

Alienably possessed nouns are not prefixed, but if they are possessed, they take independent pronouns from the set in Table 8. Example (97) shows the alienable noun ʔai ‘child’ with a first person singular possessive pronoun, and

245

Adang

example (98) shows the alienable noun fali ‘price’ with a third person possessive pronoun. (97)

Nɔ ʔai ale am. 1SG.POSS child crawl PFV ‘My child is already crawling.’ (lr-adn-20110420-01-MotionVerbs)

(98)

Pi ʔɔ fali u-baŋ nanɛ. 1PL.INCL.SBJ 3.POSS price APPL-say NEG ‘We have not mentioned its price.’ (lr-adn-20110408-01-moko)

A small set of nouns optionally take pronominal prefixes. This set includes body parts such as lafuŋ ‘head’, faʔur ‘knee’, parih ‘shin’, mara ‘buttocks’, and ʔariʔnɛmaŋ ‘hair’. When prefixed, these nouns take a different set of pronominal prefixes from the ones used for obligatorily possessed nouns. In (99), we see the noun lafuŋ ‘head’ occurring without a prefix. In (100), the same noun occurs with a pronominal prefix. (99)

Ti Tree

ɛl CLF

matɛ large

nu one

ʔɔ 3.POSS

ʔa-balɛ 3.INAL-branch

kaʔai small

nu one

ho DEF

faʔai doŋ ta mih eh lafuŋ sɔ hɛl. PROG head 3.REFL.POSS descend earth pot on sit ‘The large tree’s small branch has a clay pot sitting on it with its head down.’ (lr-adn-20110407-03-PositionalVs) (100)

ʔai lɔtɛ nu topi sɛ-lafuŋ mi tamaŋ. person male one hat 3.REFL.POSS-head at put.on ‘A man puts a hat on his (own) head.’ (lr-adn-20110407-01-FuncVs)

Adang also has a set of independent pronouns that can be used to put possession in contrastive focus, given in Table 10 below3. Table 10: Contrastive possessive pronouns 1SG 1PL.EXCL 1PL.INCL 1PL.INCL.DISTR 2SG 2PL 3 3.REFL

ne ni(e) pi(e) te e i(e) ʔe se

3 These pronouns have a final /i/ vowel for some speakers.

246

Laura C. Robinson & John W. Haan

The contrastive possessive pronouns can be used with alienable nouns whether they have a possessive pronoun (101) or not (102). The contrastive possessive pronouns occur before the alienable possessive pronoun if the two co-occur. (101)

Ne nɔ bel mi matɛ. 1SG.POSS.CONTR 1SG.POSS dog than large ‘My dog is larger.’ (e.g., than your dog)

(102)

John John

ʔe 3.POSS.CONTR

ʔai child

farɛkaŋ hard.working

bo ne ʔai ʔɛ farɛkaŋ nɛnɛ. but 1SG.POSS.CONTR child NEG hard.working NEG ‘John’s children are hard-working, but my children are not hard-working.’ Contrastive possessive pronouns can occur with inalienable nouns in conjunction with a pronominal prefix (example (103)). (103)

Ne na-fɛl mi habu. 1SG.POSS.CONTR 1SG.INAL-ear than wide ‘My ears are larger (e.g., than yours).’

These contrastive possessive pronouns can also be used without a possessed noun, as in (104) and (105). (104)

Ne tar eh. 1SG.POSS.CONTR lie.down PROG ‘Mine is lying down.’ (e.g., ‘my child’)

(105)

Sɛŋ ho ne. money DEF 1SG.POSS.CONTR ‘The money is mine.’

In contrast, an alienable possessive pronoun cannot be used without a possessed noun (106). (106) *nɔ 1SG.POSS

tar lie.down

eh PROG

Adang

247

The third person reflexive pronouns are used when a noun co-indexes a subject NP. In example (107), the 3.REFL prefix co-indexes the subject NP namɛ nu ‘a person, someone’. (107)

[Namɛ nu]NPI [topi]NP [sɛi-lafuŋ]NP tamaŋ. person one hat 3.REFL.POSS-head put.on ‘Someone put a hat on his/her (own) head.’ (lr-adn-20100714-TopoRels)

To co-index a possessor, the non-reflexive pronoun is always used. In example (108), the non-reflexive third person prefix is used to co-index a possessor, namɛ nu ‘a person, someone’, to create a complex NP, namɛ nu ʔɛlafuŋ ‘a person’s/ someone’s head’. In example (109), we can see that this applies equally to possessed nouns in the subject NP. (108)

ta apec. Laka tɛn [[namɛ nu]i ʔɛi-lafuŋ]NP cloth red person one 3.POSS-head on tie ‘A red cloth is tied on someone’s head.’ (lr-adn-20100714-TopoRels)

(109)

[Butiliŋ ʔɔ tel]NP.SBJ [butiliŋ]NP.OBJ ta ol eh. bottle on lie PROG bottle 3.POSS cover ‘The bottle’s cover is on the bottle.’ (lr-adn-20100714-TopoRels)

A non-reflexive pronoun is used even if the possessor and possessee are not adjacent within the noun phrase, as in (110), where ʔɔ bel ‘his dog’ is coordinate with ʔai lɔtɛ kaʔai nu ‘a small boy’ inside a larger noun phrase. (110)

[[ʔai child

lɔtɛ male

kaʔai small

nu]NP one

ɛ and

[ʔɔ 3.POSS

bel]NP]SBJ.NP dog

sanaŋ 3PL.NMBR

alɔ two

upapaʔ nu ba botol om mi ho ʔɛ-dun eh. frog one REL bottle inside in DEF 3.ALL-look PROG ‘A small boy and his dog; the two of them are looking at a frog that is inside in a bottle.’ (frog story, Baird n.d.)

4.2 Attributes In the noun phrase template, we saw that verbs could modify the head noun. In example (111), the verb min ‘to die’ is modifying the noun ab ‘fish’.

248 (111)

Laura C. Robinson & John W. Haan

[Supi]NP [ab min]NP taraʔ. fish die collect 3PL ‘They collected dead fish.’

There is no separate adjective class in Adang. Property-denoting words are not distinguished from intransitive verbs. For example, both action-denoting and property-denoting (adjectival) verbs can be modified by biʔ ‘very’. (112)

Sɛi bunɛ biʔ. water hot very ‘The water is very hot.’

(113)

Aru foi batɛ ala ʔa-dɛ biʔ am. deer again corn rice 3.OBJ-eat very PFV ‘Deer have started eating the crops a lot more often.’

In example (114), we see that more than one verb can modify a noun, with both tɛn ‘be red’ and matɛ ‘be large’ modifying kertas ‘paper’. (114)

[Mɛja afɛl nu]NP [kertas tɛn matɛ nu]NP table white one paper red large one ‘A large red paper is lying on a white table.’ (lr-adn-20110407-03-PositionalVs)

ta on

ol lie

eh. PROG

Since locative deictics (115) and directionals (116) are considered verbs (see section 3.4), it is unsurprising to find that they can occupy the same slot in the noun phrase as other modifying verbs. (115)

matɛ. [Bel mɔŋ hɛmɔ]NP dog there.level that.level large ‘That dog over there is large.’

(116)

fail. Namɛ [bɛ talɛ hɛmɔ]NP person mango up.on that.above sell ‘Someone is selling those mangoes on top (e.g., of the other mangoes).’

4.3 Classifiers Adang uses classifiers which have properties of both numeral classifiers and noun classifiers (following Aikhenvald 2000 and Grinevald 2004). They most

Adang

249

typically occur with numerals, especially numerals higher than ‘one’, as in example (117), where we can see that the classifier follows the noun, and the numeral follows the classifier. Since Adang classifiers typically occur with numerals, they could be considered numeral classifiers. (117)

fak pir alɔ bat CLF.fruit two ‘two bats’ (lr-adn-20100820-02-Classifiers)

However, noun phrases with the number nu ‘one’ frequently lack a classifier, as in example (118). (118)

[Duir nu]NP [fui]NP ta. knife one blood on ‘A knife has blood on it.’ (lr-adn-20100714-TopoRels)

Even with higher numbers, it is possible to omit the classifier (119), especially when counting persons or animals (120), although it is not common. (119)

[Ho]NP

[ɛi malahai ut]NP om mi. canoe Malay four inside in ‘It (the Oil Male Pah drum) is worth four Malay canoe (drums).’ (lit, “It has four Eimalahai in it.”, lr-adn-20110408-01-moko)

DEF

(120)

am. [Namɛ alɔ]NP [mok]NP med s-ɛn person two mug take 3.REFL.OBJ-give PFV ‘Two people give each other a mug.’ (i.e., one gives it then the other gives it back) (lr-adn-20110408-02-FuncVs)

Moreover, a modifying verb can intervene between a numeral and a classifier, a feature not known with numeral classifiers (Aikhenvald 2000). In example (121), the verb tɔʔaŋ ‘be short’ intervenes between the classifier bɔʔ and the numeral tou ‘three’. (121)

nu]NP [Ti bɔʔ toʔaŋ tou]NP [lai tree CLF.stalk short three long one ‘Three short logs plus one long (one).’ (lr-adn-20100714-01-SurreyVideosCore)

ta. add

250

Laura C. Robinson & John W. Haan

Modifying verbs may occur either before or after the classifier. With low numbers, especially ‘one’, the modifying verb typically follows the classifier, as in both noun phrases of example (122). In contrast, with higher numbers, the modifying verb typically precedes the classifier, as in (123). In our data, only the numbers ‘two’ and ‘three’ were found with both orders. (122)

[Ti tree

bɔʔ CLF.stalk

kaʔai small

nu]NP one

[ti ɛl matɛ nu]NP u-ʔab tufeh. APPL-lean stand.PROG tree CLF large one ‘A small stick is leaning on a large tree.’ (lr-adn-20110407-03-PositionalVs) (123)

[Bal ball

matɛ large

paʔ CLF

tou]NP three

med take

[kaʔai paʔ tou]NP [faʔai]NP ta ol eh. on lie PROG small CLF three soil ‘Three large balls and three small ones are on the ground.’ (lr-adn-20110408-PositionalVs) Example (122) shows how the same noun, in this case ti ‘tree’, can receive different interpretations when used with different classifiers. In this case, ti means ‘tree(s)’ on its own (as in examples (131), (148), and (152) or when classified with ɛl (as in 122), which is a classifier for rigid, standing objects. When classified with bɔʔ ‘stalk, stem’, it means ‘stick, log’. We have not encountered examples of ‘repeaters’ where a noun is used to classify itself, as ti ti for ‘tree’ or pir pir for ‘fruit’. Adang has many classifiers, which classify nouns depending on their size, shape, and flexibility. The most common classifiers are given in Table 11, along with their literal meaning (if any), the semantics of the classified nouns, and examples of nouns that can occur with that classifier. Note that tuliŋ ‘drop’ is a mensural classifier, indicating the quantity of the classified noun, but all the rest are sortal classifiers, indicating the type or sort of the classified noun. The most common classifier is paʔ, which is a default classifier used for objects of many shapes and sizes, including arrows (124) and moko drums (125), which usually have an hour-glass shape. (124)

ʔaburiŋ paʔ ut CLF four arrow ‘four arrows’ (lr-adn-20100820-02-Classifiers)

Adang

(125)

251

hahai paʔ tou moko CLF three ‘three moko drums’ (lr-adn-20100820-02-Classifiers)

The second-most common classifier in our data is pir, which means ‘fruit’ and is used mainly for small, round objects. Example (126) illustrates use of pir with fruits. (126)

fa pir ut coconut CLF.fruit four ‘four coconuts’ (lr-adn-20100820-02-Classifiers)

Table 11: Adang classifiers Classifier

Nominal meaning

Semantics

Examples

ʔafail

seed

small, rigid objects

corn kernel, rice grain

ʔahang

slice

small, flat, rigid objects

wood, walls

ʔanemeng

sheet

flexible, non-flat objects

clothes, rope, string

ʔir

–

long, cylindrical, jointed objects

bamboo, sugarcane

bar

bunch

short, clustered, hanging objects

coconut bunches, earrings, bells, betel nut

beh

leaf

flat, flexible objects

leaves, money

bɔʔ1

log

long, cylindrical, rigid objects

bamboo, logs

bɔʔ2

–

large, flat, rigid objects

fields

buil/buliŋ

stem

long, clustered, hanging objects

banana bunches, rice

ɛl

–

rigid, standing objects

trees, buildings

kumaŋ

piece

short, cylindrical objects

logs, eels, snakes

paʔ

non-round fruit

default

pir

round fruit

large, round objects

fruits, people, animals

puh

–

hanging objects

corn ears, banana blossoms

tuliŋ

drop

liquids

water, oil, wine

The classifier pir is only used for fruits that are round. Fruits with other shapes, such as bananas (127), occur with other classifiers.

252 (127)

Laura C. Robinson & John W. Haan

mɔʔɔi paʔ itito banana CLF seven ‘seven bananas’ (lr-adn-20100820-02-Classifiers)

The classifier pir is used with many animals, which are counted by their heads. Example (128) illustrates the use of pir with humans and (129) illustrates its use with non-human animals. The classifiers paʔ and pir appear to be expanding in usage at the expense of other classifiers. (128)

ʔɔb tumɔʔ pir ifihiŋ CLF. fruit five woman old ‘five old women’ (lr-adn-20100820-02-Classifiers)

(129)

mon pir alɔ snake CLF.fruit two ‘two snakes’ (lr-adn-20100820-02-Classifiers)

Classifiers seem to be an open word class in Adang, a potential property of both noun classifiers and numeral classifiers (Aikhenvald 2000). Although the classifiers listed in Table 11 are the most common, it appears that a great number of nouns can be used as classifiers. In example (130), for instance, the noun hɛi ‘rope’ is used to classify ba ‘earring’. More work is needed to determine the extent of this phenomenon. (130)

ba hɛi tou earring CLF.rope three ‘three earrings’ (lr-adn-20100820-02-Classifiers)

4.4 Non-numeral quantifiers As shown in the noun phrase template, non-numeral quantifiers follow the noun and any modifying verbs. Example (131) shows the use of the quantifier nun ‘some’ with a modifying verb matɛ ‘be large’. Adang has no overt plural marker, so the quantifier nun ‘some’ is often use to emphasize plurality. (131)

matɛ nun]NP [Pen]NP [ti Pen tree large some ‘Pen cut some large trees.’

ʔa-bɔʔɔi. 3.OBJ-cut

Adang

253

A relative clause occurs after the quantifier, as in (132), where the relative clause ba na fel ‘that I bought’ follows the quantifier barpi ‘all’. (132)

Sa [boi barpi [ba na fel]REL ho]NP aminɛ. REL 1SG.SBJ buy DEF kill 3SG.SBJ pig all ‘He killed all the pigs that I bought.’ (lr-adn-20120630-01-sketch)

Quantifiers can co-occur with mass nouns (133) and with the determiner ho (134 and 135). (133)

[Sɛi nun ho]NP ʔuhuɲ ɛ bɛŋ tanib. pour and other draw.water water few DEF ‘Pour out that little bit of water and get some more from the well.’

(134)

ʔai lɔtɛ sa-dimaŋ [upapaʔ nun ho]NP ʔɛ-dun. some DEF 3.ALL-see child male 3.REFL.OBJ-look.back frog ‘The boy looked back and saw the frogs.’ (frog story, Baird n.d.)

The quantifier barpi ‘all’ can co-occur with a number, as in (135). (135)

fɔi matɛ ifihiŋ barpi ho stone large five all DEF ‘all five large stones’ (lr-adn-20120630-01-sketch)

4.5 Relative clauses Relative clauses in Adang follow the head noun, and any modifying verbs and quantifiers. The relative clause is introduced with the relativizer ba, and the definite marker ho or another demonstrative appears at the end of the noun phrase. Adang employs a gap strategy. In examples (136), the subject of the matrix clause namɛ matɛ pir tou ‘three large people’ is relativized and also functions as the subject of the relative clause. In this instance, it is obvious that ‘three large people’ is the subject of the relative clause since the other noun phrase in the relative clause is an accusative pronoun niri. (136)

[Namɛ person [ba

matɛ large

pir CLF

tou three

niri bɛh]REL ho]NP tɛʔɛŋ am. 1PL.EXCL.OBJ hit DEF run PFV ‘The three large people who hit us have run away.’ REL

254

Laura C. Robinson & John W. Haan

In example (137), it is again the subject of the matrix clause that is relativized, but now it functions as the object of the relative clause. In this case, it is clear that the relativized noun must be object because the relative clause contains a subject pronoun sa. (137)

[Sɛŋ [ba sa tarɔʔ]REL ho]NP suhuɲ DEF disappear money REL 3SG.SBJ save ‘The money that s/he saved has disappeared.’

am. PFV

In (138), we can see that it is also possible to relativize objects in Adang. Here ala mɛj ‘cooked rice’ is relativized and is also the object of the relative clause. (138)

[Na]NP [ala mɛj [ba i 1SG.SBJ rice cooked REL 2PL.SBJ ‘I ate the rice that you cooked.’

dou]REL cook

ho]NP DEF

ʔa-de 3.OBJ-eat

am. PFV

In example (139), the object is again relativized, but only context allows us to interpret its role within the relative clause. This could be interpreted as either ‘the rope that is on the tree branch’ or ‘the rope that the tree branch is on’. (139)

[ʔai person

tumɔʔ old

nu]NP one

[hɛi rope

[ba REL

ti tree

balɛ branch

ta]REL on

ho]NP DEF

Ɂ-abiʔiŋ med. 3.OBJ-pull take ‘An old man pulls off the rope that is on the tree branch.’ (lr-adn-20110409-01-FuncVs) Adang allows both restrictive relative clauses, such as those given above, and non-restrictive relative clauses, such as in example (140). (140)

[Roni [ba guru am]REL ho]NP sam sakolah. DEF go school Roni REL teacher PFV ‘Roni, who has become a teacher, went to school.’

In Adang, it is possible to have a headless relative clause, as in (141), where the relative clause ba a ʔa-hol ‘that you know’ is preceded by a possessive pronoun and followed by the definite determiner, which becomes hɔ when followed by the focusing suffix -rɔ.

Adang

255

(141) Fe

a dai [ʔɔ [ba a ʔa-hol]REL hɔ-rɔ]NP u-baŋ. 2SG.SBJ still 3.POSS REL 2SG.SBJ 3.OBJ-know DEF-FOC APPL-say ‘So, you still (need to) say what it is that you know.’ (lr-adn-20110408-01-moko) FOC .OBJ

Several characteristics distinguish relative clauses from the verbal attributes discussed in section 4.2. First, they occur in a different slot in the noun phrase. This is easiest to see when both a modifying verb and a relative clause occur within the same noun phrase, as in (136), which contains the modifying verb matɛ ‘be large’ and the relative clause ba niri bɛh ‘that hit us’. Second, relative clauses, and not modifying verbs, can have aspect markers, as in (142), where the progressive aspect marker occurs in the relative clause. (142)

[Sapad [ba a lap eh]REL ho]NP mɛja REL 2SG. SBJ look.for PROG DEF table knife ‘The knife that you are looking for is on a table.’

ta. on

Relative clauses in Adang may occur without the relativizer ba, as in (143). (143)

Sa [namɛ [bel bɛh am] ho]NP hukuŋ. punish 3SG.SBJ person dog hit PFV DEF ‘S/he punished the person who had hit dogs’.

Despite the characteristics outlined above which distinguish relative clauses from verbal attributes, it can sometimes be difficult to distinguish the two in practice. In (144), there is a relative clause that contains an aspect marker and ends in a demonstrative. (144)

a [Namɛ [karɛsaŋ eh]REL hɛpɔ-rɔ]NP PROG that.below-FOC 2SG.SBJ people work ‘It is the people working down there that you fed.’

ʔ-ah. 3.OBJ-feed

In (145), we see only the verb, functioning to modify the noun. The only difference is that in example (145), there is no determiner or demonstrative. (145)

Umi [namɛ karɛsaŋ] ʔ-ah. Umi people work 3.OBJ-feed ‘Umi fed working people/people who work.’

256

Laura C. Robinson & John W. Haan

While it might then be tempting to say that a relative clause must have a determiner or demonstrative, example (146) seems to contain a relative clause, since there is a noun baŋ and we do not want to claim that a modifying verb can take an object. On the other hand, it is possible that the focus marker fe fulfills the function of the determiner or demonstrative here. (146)

Umi [namɛ [baŋ ʔa-ten]REL Umi people house 3.OBJ-build ‘Umi fed people building houses.’

fe]REL

FOC.OBJ

ʔ-ah. 3.OBJ-feed

4.6 Demonstratives and determiners The final slot in the noun phrase template can be filled by a demonstrative (147) or the definite determiner ho (148). (147)

[Ti hɔʔɔ]NP matɛ. tree this large ‘This tree is large.’

(148)

[ʔai ʔɔb kaʔai ho]NP nel-baroc. 1SG.ABL-afraid child female small DEF ‘The little girl is afraid of me.’ (lr-adn-20110407-02-SurreyVideos1st&2ndPers)

Adang has a set of demonstratives ‘this/that’ which modify nouns or serve as pronouns, given in Table 12. The demonstrative paradigm begins with a consonant [h] and ends with the vowel [ɔ], with the medial consonants depending on deixis. Like the locative deictics discussed in section 3.4, the demonstratives show a two-way division between proximal, or near the speaker, and distal, or away from the speaker. The proximals are based on the paradigmatic consonant [ʔ]. The distals show a further three-way distinction along the vertical dimension. Location above the speaker is indicated with the consonant [t], location below the speaker is indicated with [p], and location level with the speaker on the vertical dimension is indicated with [m]. Table 12: Adang demonstratives demonstratives proximal distal (above) distal (below) distal (level)

hɔʔɔ hɛtɔ hɛpɔ hɛmɔ

Adang

257

Example (149) shows the use of a demonstrative specifying the location of a head noun. (149)

[Aru hɛmɔ]NP [mɔʔɔi]NP ʔa-dɛ. deer that.level banana 3.OBJ-eat ‘That deer (over there) eats bananas.’

A demonstrative can be the sole element of a noun phrase, as in (150). (150)

[Hɛmɔ]NP [mɔʔɔi]NP ʔa-dɛ. that.level banana 3.OBJ-eat ‘That (over there) eats bananas.’ (e.g., ‘That deer over there eats bananas.’)

The determiner ho can also function as a demonstrative, locating the object in space. Such usage is most apparent in contrastive constructions such as (151). (151)

Ti hɔʔɔ matɛ ɛ ti ho kaʔai. tree this large and tree DEF small ‘This tree is large and that tree is small.’

5 Pronouns Adang has six distinct paradigms of independent pronouns. The two paradigms of independent possessive pronouns were discussed in section 4.1 above, and the rest are listed in Table 13. In general, the singulars are marked with /a/, and the first and second person plurals are marked with /i/. The first person is marked with /n/, while the third person is marked with /s/ or /ʔ/, and the second person is zero-marked. The subject pronoun paradigm, however, is different in structure from the other pronominal paradigms. First, the subject pronouns distinguish between singular and plural in the third person. In the other paradigms, there is a distinction in the third person between pronouns that are reflexive/reciprocal and those that are not, without regard for number, although an additional plural pronoun may optionally be added. The object paradigm also has a first person inclusive pronoun that is distributive.

258

Laura C. Robinson & John W. Haan

Table 13: Adang independent pronouns

1SG 1PL.EXCL 1PL.INCL 1PL.DISTR 2SG 2PL 3SG 3SG.REFL 3PL 3PL.REFL

Subject

Object

Numbered

Alone

na ni pi – a i sa sa supi supi

nari niri piri tari ari iri ʔari sari (supi) ʔari sari

– ninaŋ pinaŋ – – inaŋ – – sanaŋ sanaŋ

nɔlɔ nilɔ pilɔ – ɔlɔ ilɔ sɔlɔ sɔlɔ sɔlɔ sɔlɔ

5.1 Subject pronouns Adang subject pronouns are used for the subjects of intransitives (152) as well as the subjects of transitives (153). (152)

Sa asal eh. 3SG.SBJ laugh PROG ‘She is laughing.’ (lr-adn-20100714-05-SurreyVerbs1)

(153)

Ni mɔʔɔi ʔa-dɛ. 1PL.EXCL.SBJ banana 3.OBJ-eat ‘We eat bananas.’ (lr-adn-20100714-05-SurreyVerbs1)

Unlike Western Pantar, Kula, Kamang, or Sawila, but like Blagar and Kaera, Adang has an accusative alignment system, where the subjects of transitives and the subjects of intransitives are treated alike, without regard to volitionality. Examples (154) and (155) illustrate non-volitional subjects. (154)

Sa tumɔʔ. 3SG.SBJ old ‘S/he is old.’

(155)

Ni Nani 1PL.EXCL.SBJ Nani ‘We know Nani.’

ʔa-hol. 3.OBJ-know

Adang

259

Examples (156) and (157) illustrate the use of a subject pronoun with volitional subjects. (156)

Na piŋ hafɔʔ. 1SG.SBJ plate wash ‘I’m washing a plate.’ (lr-adn-20110409-02-SurreyVideos)

(157)

Ni sam taŋ mi aroŋ. sea in swim 1PL.EXCL.SBJ go ‘We go swimming in the sea.’ (lr-adn-20100714-05-SurreyVerbs1)

Example (158b) illustrates use of a subject pronoun with an animate nonhuman subject. (158)

a.

ʔamɔ ʔai nu n-ɛn cat child one 1SG.OBJ-give ‘Give me a kitten please.’

ɛma. please

b.

Supi matɛ fe a nu med. large FOC.OBJ 2SG.SBJ one take 3PL ‘When they (the kittens) get large, you can take one.’

Subject pronouns are not usually used to refer to inanimates. Example (159) shows that the subject pronoun supi cannot be used to refer to machetes in B’s response to the question asked by A. (159)

A:

Sapad tarɔ mi? machete where at ‘Where are some machetes?’

B: *supi 3PL

mɛja table

ta on

In some limited circumstances, however, a subject pronoun can be used to refer to an inanimate, as in (160) and the dialogue in (161). More work is needed to determine when a subject pronoun can be used to refer to an inanimate.

260 (160)

Laura C. Robinson & John W. Haan

Ɛi canoe

malahai Malay

honihe like.that

ʔɔŋ here

utaŋpe bitter.melon

turlɔ eight

ta on

honi like.that

sa rat nu ʔɛr ifihiŋ. 3SG.SBJ hundred one ten five ‘This Eimalahai ‘Malay Canoe’ (kind of moko drum), with eight bitter melons on it, it is one hundred fifty.’ (lr-adn-20110408-01-moko) (161)

A:

ɛi malahai ʔa-ra ham? canoe Malay 3.OBJ-be same ‘Is it the same as the Eimalahai ‘Malay Canoe’?’

Ho

DEF

B:

Ɛi malahai ʔa-ra ham. canoe Malay 3.OBJ-be same ‘The Eimalahai ‘Malay Canoe’ is the same.’

B:

Supi ʔa-ra ham. 3.OBJ-be same 3PL ‘They are the same (as the others).’ (lr-adn-20110408-01-moko)

5.2 Object pronouns One class of verbs uses object pronouns to indicate the objects of transitive verbs, as in examples (162) and (163). Not all transitive verbs, however, indicate their objects with object pronouns. Some indicate their objects with pronominal prefixes instead (see section 6 for more information on this dichotomy). (162)

ʔai tumɔʔ nu nari bariŋ. person old one 1SG.OBJ pull ‘An old man is pulling me.’ (lr-adn-20110407-02-SurreyVideos1st&2ndPers)

(163)

Na ʔari putaŋ. 1SG.SBJ 3.OBJ bump ‘I bump into him.’ (lr-adn-20110405-01-SurreyVideos&1st2ndPers)

Although the third person object pronouns are not inherently singular or plural, the non-reflexive pronoun ʔari can be preceded by supi to emphasize or clarify plurality.

Adang

(164)

Sa [supi ʔari]NP 3.OBJ 3SG.SBJ 3PL ‘S/he hit them.’

261

bɛh. hit

The third person non-reflexive pronoun can also be adnominal in usage, co-occurring with a proper name, as in (165), where ʔari is co-referential with the proper name Bain, and (166), where the ʔari is co-referential with the proper names Bain ɛ Nani ‘Bain and Nani’. (165)

Sa [Bain ʔari]NP 3SG.SBJ Bain 3.OBJ ‘S/he hit Bain.’

(166)

Ni [Bain ɛ 1PL.EXCL.SBJ Bain and ‘We hit Bain and Nani.’

bɛh. hit

Nani Nani

ʔari]NP 3.OBJ

bɛh. hit

The third person non-reflexive pronoun cannot co-occur adnominally with a common noun. Contrast the grammatical (167) with the ungrammatical (168). (167)

Sa sɔ ʔɔb bɛh. 3SG.SBJ 3.REFL.POSS woman hit ‘He hit his wife.’ (lr-adn-20120630-01-sketch)

(168) *sa 3SG.SBJ

sɔ 3.REFL.POSS

ʔɔb woman

ʔari 3.OBJ

bɛh hit (lr-adn-20120630-01-sketch)

The object pronominal paradigm includes a first person plural inclusive distributive pronoun which indicates that each person was acted upon individually. Contrast the distributive in (169) with the non-distributive first person inclusive plural in (170). (169)

Sa tari 3SG.SBJ 1PL.INCL.DISTR.OBJ ‘S/he hit each one of us.’

(170)

Sa piri 3SG.SBJ 1PL.INCL.OBJ ‘S/he hit (all of ) us.’

bɛh. hit

bɛh. hit

262

Laura C. Robinson & John W. Haan

An object pronoun is used in left dislocation, where it occurs in initial position and is followed by the definite determiner ho. In (171), the left-dislocated object pronoun refers to the object of the transitive verb bɛh ‘hit’, while in (172), the left-dislocated object pronoun refers to a subject. (171)

Nari ho, Bika so (sa) 1SG.OBJ DEF Bika FOC.SBJ 3SG.SBJ ‘As for me, it is Bika who hit me.’

(172)

Iri 2PL.OBJ

ho, DEF

(i) 2PL.SBJ

dɔi hill

(nari) 1SG.OBJ

lɔl climb/move.along

bɛh. hit

ɛ and

niri ho, (ni) mol lɔl. 1PL.EXCL.OBJ DEF 1PL.EXCL.SBJ river climb/move.along ‘As for you, you climb the hills, and as for us, we’ll walk along the rivers.’ Left-dislocated object pronouns can also occur with the subject focus marker so (173) or the object focus marker fe (174), as appropriate. (173)

Iri so, (i) sam don. 2PL.OBJ FOC.SBJ 2PL.SBJ go shop ‘It is you who are going shopping.’

(174)

niri fe, sa 1PL.EXCL.OBJ FOC.OBJ 3SG.SBJ ‘It was us whom s/he hit.’

(niri) 1PL.EXCL.OBJ

bɛh. hit

The third person pronouns we have labeled REFL can have either a reflexive or reciprocal interpretation. In (175), the reflexive usage is illustrated. Contrast (175) with (176), where a non-reflexive pronoun is used and the object pronoun cannot co-index the subject. (175)

John sari bɛh. John 3.REFL.OBJ hit ‘John hit himself.’

(176)

John ʔari bɛh. John 3.OBJ hit ‘John hit him.’

Adang

263

Examples (177) and (178) illustrate reciprocal meanings. (177)

Namɛi sarii putaŋ. person 3.REFL.OBJ bump ‘People bump into each other.’ (lr-adn-20110407-02-SurreyVideos1st&2ndPers)

(178)

[ʔai person

lɔtɛ male

pir CLF

alɔ]i two

sɔi-padomaŋ 3.REFL.POSS-eyeglasses

hiŋ take.off

ɛ med sarii tiliŋ. and take 3.REFL.OBJ trade ‘Two men take off their glasses and trade each other.’ (lr-adn-20110407-01-FuncVs) Surprisingly, the 3.REFL pronoun is used in left-dislocation where there is no obvious subject antecedent (179). Using a non-reflexive third person pronoun in left-dislocation is ungrammatical, as in (180). Haan (2001: 162) suggests that this is because “the focus or topic part of a sentence is itself a clause with an understood subject” so the 3.REFL pronoun is co-referential with that understood subject. (179)

Sari ho, na (ʔari) 3.REFL.OBJ DEF 1SG.SBJ 3.OBJ ‘As for him, I will hit (him).’

(180)

*ʔari 3.OBJ

ho, DEF

na 1SG.SBJ

(ʔari) 3.OBJ

bɛh. hit

bɛh hit

5.3 Numbered pronouns When a pronoun is modified by den ‘how many’ or a number higher than one, the formative naŋ is attached to the pronoun. This could be considered a classifier for pronouns. The numbered pronouns are used for subjects in (181) and (182). (181)

Sanaŋ den John bɛh? 3PL.NMBR how.many John hit ‘How many of them hit John?’

264

Laura C. Robinson & John W. Haan

(182)

ʔai child

lɔtɛ male

kaʔai small

nu one

ɛ and

ʔɔ 3.POSS

bel dog

sanaŋ 3PL.NMBR

alɔ two

upapaʔ nu ba botol om mi ho ʔɛ-dun eh. frog one REL bottle inside in DEF 3.ALL-look PROG ‘A small boy and his dog; the two of them are looking at a frog that is inside in a bottle.’ (frog story, Baird n.d.) Numbered pronouns are also used for objects (183 and 184). (183)

John inaŋ den bɛh? John 2PL.NMBR how.many hit ‘How many of you did John hit?’

(184)

John ninaŋ John 1PL.EXCL.NMBR ‘John hit four of us.’

ut four

bɛh. hit

Numbered pronouns can co-occur with subject or object pronouns. The numbered pronoun follows the subject or object pronoun, as in examples (185) and (186) respectively. (185)

[I inaŋ tou] sam 2PL.SBJ 2PL.NMBR three go ‘Three of you go shopping.’

(186)

John [niri ninaŋ John 1PL.EXCL.OBJ 1PL.EXCL.NMBR ‘John hit three of us.’

don. shop

tou] three

bɛh. hit

They can also be used in combination with a personal name to form an associative plural, meaning ‘x and others’. In such cases, the personal name precedes the numbered pronoun, as in example (187). (187)

Heri [John ninaŋ tou] bɛh. Heri John 1PL.EXCL.NMBR three hit ‘Heri hit three of us, including John.’

The numbered pronouns are ungrammatical without a following number. Contrast the grammatical construction in (188) with the ungrammatical construction in (189).

Adang

(188)

265

Inaŋ tou sam don. shop 2PL.NMBR three go ‘You three go shopping.’

(189) *inaŋ 2PL.NMBR

sam go

don shop

5.4 Alone pronouns The alone pronouns from Table 13 can be used to indicate that a person or persons do something alone. They may co-index another argument, as sɔlɔ co-indexes ʔɔb tumɔʔ nu ‘an old woman’ in (190), or ʔai tumɔʔ nu ‘an old man’ in (191). (190)

(191)

Ɂɔb tumɔʔ nu sɔlɔ mih asal laugh woman old one 3.ALONE sit ‘An old woman is sitting by herself laughing.’ (lr-adn-20100714-01-SurreyVideosCore) Ɂai tumɔʔ nu lapaŋ ta tar person old one field on lie.down ‘An old man is lying in a field alone talking.’ (lr-adn-20100714-01-SurreyVideosCore)

eh. PROG

sɔlɔ 3.ALONE

huʔtaŋ chat

eh. PROG

The alone pronouns may also occur without an antecedent when referring to a subject, as indicated by the parentheses around the subject pronoun in (192). Note that alone pronouns can be plural, as in (192), indicating that multiple people are by themselves without other implied individuals. (192)

(I) ilɔ sam don. shop 2PL.SBJ 2PL.ALONE go ‘You go shopping by yourselves.’

When used to indicate the object of a transitive verb, alone pronouns must co-occur with object pronouns. Contrast the grammatical construction in (193) containing an object pronoun with the ungrammatical (194) lacking an object pronoun.

266 (193)

Laura C. Robinson & John W. Haan

John nɔlɔ nari bɛh. John 1SG.ALONE 1SG.OBJ hit ‘John hit me alone.’ (i.e., ‘John hit just me.’)

(194) *John John

nɔlɔ 1SG.ALONE

bɛh hit

6 Pronominal prefixes and valency increasing prefixes Adang has four sets of pronominal prefixes, given in Table 14. These prefixes are all etymologically related to the independent pronouns discussed in section 5. Adang also has an applicative prefix (discussed in 6.4) and a causative prefix (discussed in 6.5) Table 14: Adang pronominal prefixes

1SG 1PL.INCL.DISTR 1PL.EXCL 1PL.INCL 2SG 2PL 3 3.REFL

Object

Ablative

Allative 1

Allative 2

n(a)t(a)nipi(a)iʔ(a)s(a)-

neltelnielpielelielʔelsel-

nɛtɛniɛpiɛɛiɛʔɛsɛ-

nɔtɔniɔpiɔɔiɔʔɔsɔ-

6.1 Object prefixes The prefixes in the first column of Table 14 are used to mark inalienable possession (discussed in section 4.1) and the objects of certain transitive verbs. As mentioned in section 5, objects of transitive verbs in Adang can be indicated with either an object pronoun or an object prefix. The difference is primarily lexical. Certain verbs select independent object pronouns while others select pronominal prefixes. Fedden et al (2013) report that Adang verbs that usually take animate objects are more likely to be prefixing, but this is just a tendency, and speakers must memorize which verbs take prefixes versus independent object pronouns. Examples (195) and (196) both have animate objects, but the verb -abiʔiŋ ‘pull’ takes object prefixes, while the verb putaŋ ‘bump’ takes object pronouns.

Adang

(195)

Namɛ lɔtɛ nu n-abiʔiŋ. person male one 1SG.OBJ-pull ‘A man pulls me.’ (lr-adn-20110409-02-SurreyVideos)

(196)

A ʔari putaŋ. 2SG.SBJ 3.OBJ bump ‘You bump into him.’ (lr-adn-20110407-02-SurreyVideos1st&2ndPers)

267

Ditransitive verbs are also split between those which take pronominal prefixes (197) and those which take object pronouns (198). (197)

John sɛŋ t-ɛn. John money 1PL.INCL.DISTR.OBJ-give ‘John gave us money.’

(198)

Nina sɛŋ nari ta. Nina money 1SG.OBJ add ‘Nina made me owe some money.’ (lit., ‘Nina put some money on me.’)

For prefixing verbs, the pronominal prefix is present when there is a lexical noun phrase. In (199) and (200), the pronominal prefixes co-index the noun phrases Bain and ti bɔʔ matɛ nu ‘a large log’, respectively. (199)

Bel Baini ʔi-eh. dog Bain 3.OBJ-bite ‘A dog bit Bain.’

(200)

Ɂai ʔɔb [ti bɔʔ matɛ nu]i ʔi-abiʔiŋ eh. person female tree CLF large one 3.OBJ-pull PROG ‘A girl is pulling a large log.’ (lr-adn-20100607-01-SurreyVideosCore)

In this way, this object prefixes are similar to agreement for this class of verbs. However, an object prefix cannot co-occur with an object pronoun to indicate the same argument. For this reason, we do not consider them to be agreement. Example (201) is ungrammatical because the verb -eh ‘bite’ requires an object prefix, and so an object pronoun cannot be added. (201) *bel dog

nari 1SG.OBJ

n-eh 1SG.OBJ-bite

268

Laura C. Robinson & John W. Haan

6.2 Ablative prefixes Ablative prefixes increase the valency of the verb and indicate movement away from the referent of the ablative prefix for a purpose. Contrast (202), where tafuniŋ ‘hide’ occurs without an ablative prefix and (203) where an ablative prefix is added. (202)

ʔai lɔtɛ kadere ʔa-mot mi tafuniŋ. child male chair 3.INAL-back in hide ‘A boy is hiding behind a chair.’ (lr-adn-20100714-TopoRels)

(203)

Supi John ʔel-tafuniŋ eh. John 3.ABL-hide PROG 3PL ‘They are hiding from John.’

The movement away may be more metaphorical than physical, as in (204) and (205), where no physical motion is implied by the verb. (204)

Pi tel-mala. 1PL.INCL.SBJ 1PL.INCL.DISTR.ABL-shy ‘We are shy around each other.’

(205)

Ɂai ʔɔb mon ʔel-baroc. child female snake 3.ABL-afraid ‘A girl is afraid of (lit., ‘away from’) a snake.’ (lr-adn-20100607-01-SurreyVideosCore)

Ablative prefixes only occur with a limited number of verbs. We have identified only five verbs which can take ablative prefixes: papaɲ ‘imitate’, mala ‘shy’, baroc ‘afraid’, tafuniŋ ‘hide’, and tɛʔɛŋ ‘run’.

6.3 Allative prefixes There are two sets of allative prefixes in Adang. The first set, based on the paradigmatic vowel ɛ is homophonous with one set of possessive prefixes (see Section 4.1). Allative prefixes indicate movement toward the referent of the allative prefix with a purpose. In example (206), the subject purposefully rolls the ball towards ‘me’. In example (207), the allative prefix refers to na ‘thing’ and when negated, indicates that the subject did not move towards the ‘things’ with purpose.

Adang

(206)

Valdi bal nɛ-a-golaŋ. Valdi ball 1SG.ALL-CAUS-roll ‘Valdi rolled a ball to me.’

(207)

Rudy ʔɛ nai ʔai-dɛ ʔɛi-hɔʔ Rudy NEG thing 3.OBJ-eat 3.ALL-come ‘Rudy did not come to eat (things=food).’

269

nɛnɛ. NEG

The second set of allative prefixes is based on the paradigmatic vowel ɔ, and is very limited in distribution. We have found only three verbs which can take this set of allative prefixes: lap ‘search’ (208), lɔfɛ ‘call’ (209), and tain ‘release’ (208)

Ɂai tɛʔɛŋ s-ɔmaŋ ʔɔ-lap. child run 3.REFL.INAL-father 3.ALL-search ‘A child runs to her father.’ (lr-adn-20110409-02-SurreyVideos)

(209)

Namɛ nu ʔai tumɔʔ ʔɔ-lɔfɛ. person one person old 3.ALL-call ‘Someone calls to the old man.’ (lr-adn-20100607-02-SurreyVideosPeri)

Allative prefixes increase the valency of the verb. Contrast the intransitive use of bad ‘be happy’ in (210) with the bivalent use of the same verb in (211). (210)

Supi bad. happy 3PL ‘They are happy.’

(211)

Supi nɛ-bad. 1SG.ALL-happy 3PL ‘They are happy because of me.’

Allative prefixes can be combined with object prefixes on a single verb. In example (212), the first prefix is the allative, which co-indexes the subject, while the second prefix co-indexes the object. (212)

sɛi-ʔaj-tɛl. Ayi afɛj Ay ladder 3.REFL.ALL-3.OBJ-lift ‘Ay lifted a ladder towards herself.’

270

Laura C. Robinson & John W. Haan

6.4 Applicative prefix In addition to the ablative and allative prefixes, there is also an applicative prefix u- that is used to increase the valency of a verb. In example (213), the verb -tan ‘ask’ occurs without the applicative prefix. In (214), the applicative prefix is added along with a theme noun phrase kɔd habar ho ‘the new shirt’. (213)

Sa na-tan. 3SG.SBJ 1SG.OBJ-ask ‘S/he asked me.’

(214)

Uli kɔd habar ho u-na-tan. DEF APPL-1SG .OBJ-ask Uli shirt new ‘Uli asked me about the new shirt.’

The applicative prefix introduces noun phrases of various thematic roles. The introduced noun phrases are typically themes, as in (215), but they can also be goals (216) or beneficiaries (217). (215)

ʔai tumɔʔ nu mih na nu u-muniŋ. person old one sit thing one APPL-smell ‘An old man sits and smells (at) something.’ (lr-adn-20110409-02-SurreyVideos)

(216)

Roni nari u-hɔʔ. Roni 1SG.OBJ APPL-come ‘Roni came toward me (for a purpose).’

(217)

Ella Ani u-aʔɔmɛ. Ella Ani APPL-pray ‘Ella prayed for Ani.’

The introduced noun phrase can be omitted if it is understood from the context. In example (218), the introduced theme noun phrase has been omitted, but the sentence is still grammatical. (218)

Uli u-na-tan. Uli APPL-1SG.OBJ-ask ‘Uli asked me about it.’

Adang

271

The applicative prefix occurs outside of any pronominal prefixes, as in (218) and (219), where the pronominal prefixes occur between the applicative prefix and the verb root. (219)

Pi Lahtal ʔɔ kufaʔ u-ta-her. 3.POSS power APPL-1PL.INCL.DISTR.OBJ-amaze 1PL.INCL.SBJ God ‘We are amazed by God’s power.’ (lit., ‘We amaze ourselves with God’s power.’)

The introduced noun phrase can also be an object pronoun, as in (216) and (220). (220)

Ni iri u-matɛŋ. 1PL.EXCL.SBJ 2PL.OBJ APPL-talk ‘We talked about you.’

Example (221) illustrates the verb papaɲ ‘imitate’ with an ablative prefix, while (222) illustrates the same verb with an applicative prefix. In this case, there is little obvious difference in meaning. It may be the case that for the verb papaɲ ‘imitate’, an ablative prefix is used when the introduced argument is animate, but the applicative prefixes introduce animate arguments with many other verbs. (221)

Laka nel-papaɲ karaŋ. Laka 1SG.ABL-imitate always ‘Laka always imitates me.’

(222)

Rony nɔ lamɛ u-papaɲ. Rony 1SG.POSS walk APPL-imitate ‘Rony imitated my walk.’

6.5 Causative prefix The causative prefix a- can be combined with a limited number of intransitive verbs to increase their valency. We have identified only ten verbs that take the causative prefix: muj ‘fall down’, ʔɔl ‘fall over’, poʔ ‘shatter’, hid ‘break’, butuŋ ‘ruin’, tiniŋ ‘steady’, min ‘die’, mih ‘sit’, tɔh ‘stand’, and tar ‘lie down’. The examples below illustrate the verb muj ‘fall’ with a causative prefix (223) and without (224). The causative prefix introduces a causer argument which is realized as the

272

Laura C. Robinson & John W. Haan

subject, in this case the proper noun John. The causee is realized as the object, in this case the first person singular prefix. (223)

John ʔɛdai n-a-muj. John almost 1SG.OBJ-CAUS-fall ‘John almost made me fall down.’

(224)

Fa pir muj. coconut fruit fall ‘A coconut falls.’ (lr-adn-20100607-01-SurreyVideosCore)

The causative prefix is almost always combined with pronominal prefixes. The only known exception is the verb ʔɔl ‘fall’ (225), which cannot take a pronominal prefix when combined with a causative prefix. Example (226) is ungrammatical because a pronominal prefix has been added to the causative verb. (225)

John Semy a-ʔɔl am. John Semy CAUS-fall PFV ‘John made Semy fall over.’

(226) *John John

Semy Semy

ʔ-a-ʔɔl 3.OBJ-CAUS-fall

am PFV

If a pronominal noun phrase is added, the verb ʔɔl ‘fall’ takes independent object pronouns instead of object prefixes (227). (227)

John nari a-ʔɔl am. John 1SG.OBJ CAUS-fall PFV ‘John made me fall over.’

Four verbs have lexicalized causative suffixes when combined with the causative prefix. These are aminɛ ‘kill’ from min ‘die’, amihiŋ ‘cause to sit’ from mih ‘sit’, atuhuɲ/atuhuin ‘cause to stand’ from tɔh ‘stand’, and ataraŋ ‘cause to lie down’ from tar ‘lie down’. The examples below illustrate the causative (228) and non-causative (229) uses of the verb tɔh ‘stand’. (228)

Butiliŋ ho, namɛ med fɔi ta ʔ-atuhuin. bottle DEF person take stone on 3.OBJ-CAUS.stand ‘As for the bottle, someone stood it on a stone.’ (lr-adn-20110408-03-PositionalVs-Part1)

Adang

(229)

273

Sa seʔar u-ʔab tɔh. 3SG.SBJ door APPL-lean stand ‘S/he stood leaning on the door.’ (lr-adn-20100714-05-SurreyVerbs1)

There is also a syntactic causative construction in Adang. It can be used with these verbs as well as with verbs that do not take the causative prefix, as in (230) and (231). (230)

John na-nɔʔ ʔɔl. John 1SG.OBJ-affect fall.over ‘John made me fall over.’

(231)

Umi na-nɔʔ mud ʔa-dɛ. Umi 1SG.OBJ-affect citrus 3.OBJ-eat ‘Umi made me eat citrus fruits.’

7 Serial verb constructions Adang makes extensive use of serial verb constructions (SVCs) where multiple verbs are used together in a single predicate. Verbs in Adang SVCs share at least one argument. They also share negation, aspect, and modal adverbs. Conjunctions do not occur between verbs in an SVC, and clauses with SVCs have intonation similar to simple clauses and quite unlike conjoined clauses. These features of Adang SVCs are illustrated in section 7.1. Adang has both symmetrical SVCs, where the verbs are from open classes (discussed in section 7.2), and asymmetrical SVCs, where one of the verbs comes from a closed class (discussed in sections 7.3 through 7.7). Unlike in symmetrical SVCs, verbs in asymmetrical SVCs typically have a fixed order, with the closed-class verb preceding the open-class verb.

7.1 Properties of SVCs There are no conjunctions between verbs in an SVC. Example (232) contains two verb phrases within the SVC. The first is dopaŋ puɲ ‘with sticks’ and the second is bel bɛh ‘hit dogs’. There is a single intonation contour over the entire SVC. This is comparable to intonation of the monoverbal clause in (233), but quite different from that of the conjoined clause in (234), where there are clearly two separate intonation contours for the two conjoined verb phrases.

274

Laura C. Robinson & John W. Haan

(232)

Heri dopaŋ puɲ bel bɛh. Heri stick hold dog hit ‘Heri hit dogs with sticks.’

(233)

Heri bel bɛh. Heri dog hit ‘Heri hit dogs.’

(234)

Heri maʔ dopaŋ puɲ ɛ bel Heri perhaps stick hold and dog ‘Perhaps Heri held sticks and hit dogs.’

bɛh. hit

The verbs in an SVC share negation, as in (235) where the entire predicate is negated. (235)

Heri ʔɛ dopaŋ puɲ bel bɛh hold dog hit Heri NEG stick ‘Heri did not hit dogs with sticks.’

nɛnɛ. NEG

It is ungrammatical to negate part of an SVC, as in (236), but conjoined verb phrases can be negated separately, as in (237). (236) *Heri Heri (237)

dopaŋ stick

puɲ hold

ʔɛ NEG

bel dog

bɛh hit

nɛnɛ NEG

Heri ʔɛ dopaŋ puɲ nɛnɛ ɛ (ʔɛ) bel bɛh hold NEG and NEG dog hit Heri NEG stick ‘Heri did not hold sticks and did not hit dogs either.’

nɛnɛ. NEG

The verbs in an SVC share aspect marking. In (238), the progressive particle eh has scope over the entire SVC, whereas when two verb phrases are conjoined as in (239), the aspectual particle does not have scope over the verb that precedes the conjunction. Likewise, in (238) the modal adverb dai has scope over the entire SVC, whereas in (239) it only has scope over the verb that precedes the conjunction.

Adang

(238)

(239)

Rin dai hur med mɛja ta Rin actually spoon take table add ‘Rin is actually putting spoons on tables.’

meŋ put

275

eh. PROG

Rin dai hur med ɛ mɛja ta meŋ Rin actually spoon take and table add put ‘Rin just took spoons and is putting (them) on tables.’

eh. PROG

SVCs in Adang share arguments. Most SVCs in Adang share a single subject. In (240), the first person pronoun is the subject of both the verb tɔh ‘stand’ and the verb dun ‘look’. (240)

Na tɔh dun eh. 1SG.SBJ stand look PROG ‘I am standing looking.’ (lr-adn-20110407-02-SurreyVideos1st&2ndPers)

It is possible, however, for a later verb to take the object of an earlier verb as its subject. In example (241), the first person object of the first verb -ɛn ‘give’ is the subject of the later verb -dɛ ‘eat’. Likewise, in (242) the object of the first verb -hou ‘command’ is Rin, which is also the subject of the second verb dou ‘cook’. (241)

Ella ab n-ɛn ʔa-dɛ. Ella fish 1SG.OBJ-give 3.OBJ-eat ‘Ella gave me fish to eat.’

(242)

Roni Rin ʔa-hou ala Roni Rin 3.OBJ-command rice ‘Roni asked Rin to cook rice.’

dou. cook

SVCs can share objects, as in example (243), where sɛŋ ‘money’ is the object of both med ‘take’ and taroʔ ‘save’. (243)

John sɛŋ med taroʔ. John money take save ‘John saved/stored money.’

The verbs in an SVC may not all have the same valency. In example (244), the first verb sam ‘go’ is intransitive while the second verb fel ‘buy’ is transitive, with the noun bɛ ‘mango’ as its object. (244)

Heri sam bɛ fel. Heri go mango buy ‘Heri went to buy mangos.’

276

Laura C. Robinson & John W. Haan

Each verb in an SVC can take its own object, as in (245), where duir ‘knife’ is the object of puin ‘hold’ and ab ‘fish’ is the object of taʔoin ‘cut’. (245)

Namɛ duir puin ab taʔoin. person knife hold fish cut ‘Someone cuts fish with a knife.’ (lr-adn-20110408-02-FuncVs)

Up to eight verbs have been observed in Adang SVCs. Example (246) has eight verbs: -ra ‘be with’, med ‘take’, fit ‘carry’, mid ‘ascend’ (twice), lɛ ‘to, toward’, mi ‘be in, be at’, and taroʔ ‘save’. (246)

Toni Toni

na-ra 1SG.OBJ-with

sɛŋ money

med take

fit carry

mid ascend

ɛsɛl lɛ mid ʔoj mi taroʔ. storehouse to ascend basket in save ‘Toni and I took money up to a storehouse to save/store in a basket.’

7.2 Symmetrical SVCs Symmetrical SVCs, which contain two or more open-class verbs, are very common in Adang. Symmetrical SVCs can be used to indicate a series of events that take place in sequence, and the order of verbs in the SVC is iconic of the order the events take place, as in example (247), where the person sits first, then rests. (247)

Sa mih aɛr. rest 3SG.SBJ sit ‘He sat down and rested.’ (lr-adn-20110405-01-SurreyVideos&1st2ndPers)

Symmetrical SVCs can be used to elaborate on the manner of an event, as in (248), where the second verb hanɔʔaŋ ‘hurry’ elaborates on the manner of the first verb lamɛ ‘walk’. (248)

Roni lamɛ hanɔʔaŋ. Roni walk hurry ‘Roni walked hurriedly.’

Adang

277

7.3 Comitative SVCs The verb -ra ‘be with’ is used in SVCs to form a comitative construction, as in example (249). The verb -ra ‘be with’ is transitive and must take a pronominal prefix. Example (249) illustrates use of a comitative with an intransitive verb, while (250) illustrates a comitative with a transitive verb. (249)

Adi ni-ra Adi 1PL.EXCL.OBJ-with ‘Adi stayed with us.’

(250)

Ella na-ra ala Ella 1SG.OBJ-with rice ‘Ella ate rice with me.’

mih. sit/stay

mɛj cooked

ʔa-dɛ. 3.OBJ-eat

The verb -ra ‘be with’ must occur before the main verb in the SVC. Switching the order of verbs, as in (251), is ungrammatical. (251) *Adi Adi

mih sit/stay

ni-ra 1PL.EXCL.OBJ-with

7.4 Causative SVCs Causative SVCs in Adang can be constructed using three different verbs: -nɔʔ ‘affect’, ‑hou ‘command’, and -ɛn ‘give’. In causative SVCs, the object of the first verb is the subject of the second verb, as in (252), where the third person object of the causative verb ‑hou ‘command’ is the subject of the main verb -dɛ ‘eat’. (252)

Ella ʔa-hou mud ʔa-dɛ. Ella 3.OBJ-command citrus 3.OBJ-eat ‘Ella told him/her to eat citrus fruit.’

The causative verb must precede the open-class verb. Changing the order, as in (253), is ungrammatical. (253) *Ella Ella

mud citrus

ʔa-dɛ 3.OBJ-eat

ʔa-hou 3.OBJ-command

Example (254) illustrates use of a causative verb with the intransitive por ‘be quiet’ in, while (252) illustrates use with the transitive -dɛ ‘eat’.

278 (254)

Laura C. Robinson & John W. Haan

Sa sɔ-bel ʔa-hou por am. 3SG.SBJ 3.REFL.POSS-dog 3.OBJ-command quiet PFV ‘He told his dog to be quiet.’ (frog story, Baird n.d.)

When the verb -ɛn ‘give’ is used in a causative SVC, it means ‘help’, as in (255). (255)

Ella kod n-ɛn maŋ. Ella shirt 1SG.OBJ-give put.on ‘Ella helped me put on my shirt.’

7.5 Directional SVCs Another common type of asymmetrical SVC in Adang employs an intransitive directional verb. The directional verb precedes the open-class verb and indicates the direction of the event. The main verbs in this type of asymmetrical SVC can be intransitive (256), transitive (257) or ditransitive (258). (256)

Rin sam don. Rin go(far) shop ‘Rin went shopping.’

(257)

Na ma gelas med. 1SG.SBJ come glass take ‘I come take the glass.’ (lr-adn-20110420-01-MotionVerbs)

(258)

ʔai person

ʔɔb woman

nu one

sɛi water

gelas glass

nu one

ʔa-ra 3.OBJ-with

fa go.near

ʔai tumɔʔ nu ʔ-ɛn. person old one 3.OBJ-give ‘A woman (went and) gave a water glass to an old man.’ (lr-adn-20110409-01-FuncVs)

7.6 Instrumental SVCs The verb puin/puɲ ‘hold’ is used in an asymmetrical SVC to introduce instrument arguments. While verbs meaning ‘take’ are more commonly used for this function in other Alor-Pantar languages, using a ‘hold’ verb to indicate an instrument argument is very common cross-linguistically (Aikhenvald 2006: 26). In

Adang

279

(259), the verb puin ‘hold’ introduces the instrument duir ‘knife’. In Adang, an SVC with puin/puɲ is the only way to introduce an instrument argument. (259)

Namɛ duir puin napah tatɔʔ. person knife hold cloth cut ‘Someone cuts a cloth with a knife.’ (lr-adn-20110408-02-FuncVs)

In instrument SVCs, the verb puin/puɲ no longer retains its original semantics, as indicated by the stars preceding the translations with ‘hold’ in (260) and (261). (260)

Namɛ s-ataŋ puin ti bɔʔ tatɔʔ. person 3.REFL.INAL-arm hold tree CLF cut ‘A person cuts a stick with his/her (own) hand.’ *‘A person cuts a stick holding his/her (own) hand.’ (lr-adn-20110408-02-FuncVs)

(261)

Roni sɔ kufaʔ puɲ luduŋ Roni 3.REFL.POSS strength hold post ‘Roni lifted up posts with his strength.’ *‘Roni lifted up posts holding his strength.’

ʔa-tɛl. 3.OBJ-lift

7.7 Theme SVCs The verb med ‘take’ is used in an asymmetrical SVC to introduce a theme argument. The open-class verb follows the verb med ‘take’. Both verbs in this kind of SVC share a subject argument, as in example (262) where ʔai nu ‘a person/ someone’ is the subject of both med ‘take’ and meŋ ‘set’ or example (263), where the first person singular pronoun is the subject of both med ‘take’ and -ɛn ‘give’. (262)

ʔai nu suraʔ med meŋ am. PFV person one book take set ‘Someone set down a book.’ (lr-adn-20110408-02-FuncVs)

(263)

Na salim med ʔ-ɛn. 1SG.SBJ cloth take 3.OBJ-give ‘I give him/her a cloth.’ (lr-adn-20100820-01-SurreyVerbs)

The verb med is optional in these constructions. Both (264) and (265) are grammatical. Haan (2001: 387) notes that he has “discovered no pragmatic or

280

Laura C. Robinson & John W. Haan

semantic difference” between the two versions, but that the version with med ‘take’ “is very common in everyday speech”. SVCs with med ‘take’, then, do not appear to be valency increasing devices, but occur with verbs which are already semantically transitive or ditransitive. (264)

Ince sɛŋ med dɛc mi. Ince money take wallet put ‘Ince put some money into wallets.’

(265)

Ince sɛŋ dɛc mi. Ince money wallet put ‘Ince put some money into wallets.’

The verbs med ‘take’ and puin/puɲ ‘hold’ seem to be grammaticalizing along the path from true verbs towards case markers introducing theme and instrumental arguments, respectively.

8 Aspect marking In contrast to languages of central and eastern Alor, Adang has little grammatical marking of tense, aspect, and mood. Only aspect is marked grammatically in Adang. There are three aspectual particles, eh ‘progressive’, am ‘perfective’, and eham ‘inceptive’, all of which occur predicate-finally. The marking of aspect is optional, and unmarked predicates must be interpreted contextually. Example (266) shows use of the progressive particle with a verb, while example (267) shows its use in an equational construction, where it can occur with a nonverbal predicate. (266)

Napah tɛn nu mɛja ta hil eh. cloth red one table on hang PROG ‘A red cloth is hanging on a table.’ (lr-adn-20110408-PositionalVs)

(267)

Roni dai guru eh. Roni still teacher PROG ‘Roni is still a teacher.’

Example (268) shows the perfective particle with a dynamic verb, while example (269) shows a stative verb, and example (270) illustrates use of the particle with a non-verbal predicate.

Adang

(268)

ʔai nu suraʔ med meŋ am. PFV person one book take set ‘Someone set down a book.’ (lr-adn-20110408-02-FuncVs)

(269)

Ni-maŋ tumɔʔ biʔ 1PL.EXCL.INAL-father old very ‘Our father has become very old.’

(270)

281

am. PFV

Roni guru am. Roni teacher PFV ‘Roni has become a teacher.’

The aspectual particles occur as the final element of the clause, regardless of the part of speech of the preceding word. This is one reason we suggest they are particles, rather than suffixes. Placing the particle after the subject (271) or after a non-final verb (272) is ungrammatical. (271) *Roni Roni (272) *Rin Rin

PFV

am

guru teacher

sɛi water

bunɛ hot

eh PROG

na drink

With a relative clause, the particles occur sentence-internally, but clausefinally. In (273), the perfective particle am occurs at the end of the relative clause ba guru am ‘who has become a teacher’. (273)

[Roni [ba guru am]REL ho]NP sam sakolah. DEF go school Roni REL teacher PFV ‘Roni, who has become a teacher, went to school.’

These particles express aspect, and not tense. In example (274), the perfective am refers to the event of tree climbing, which has not yet taken place. The perfective is used because the main action of the clause takes place at a time when the tree climbing will have completed. (274) Paneʔ Manu fa ho lɔl am fe supi ʔa-bɔʔɔi. in.a.moment Manu coconut DEF climb PFV FOC.OBJ 3PL 3.OBJ-cut ‘It is in a moment, after Manu has climbed the coconut trees, that they will cut them.’

282

Laura C. Robinson & John W. Haan

There is also a predicate-final inceptive particle eham. It may be composed of both aspectual particles together. (275)

Roni guru matɛ eham. Roni teacher large INCP ‘Roni is about to become a principal.’ (lit., ‘large teacher’)

(276)

Sɛi bunɛ biʔ eham. water hot very INCP ‘The water is about to become very hot.’

9 Discussion In some respects, Adang is a very typical Alor-Pantar language. It is verb-final (section 3.1) and makes extensive use of SVCs (section 7). There is a distinction between alienable and inalienable nouns, and inalienable nouns are marked by possessive prefixes (section 4.1) which are homophonous with prefixes indicating undergoer arguments (section 6.1). Like other Alor-Pantar languages, it has a complex system of deictics, which depend on elevation (section 4.6). Adang is like Blagar and Klon, its closest neighbors, in employing a causative prefix (section 6.5) and an applicative prefix u- (section 6.4). Adang is like Teiwa and Blagar in having an accusative alignment system (section 5) and like all Alor-Pantar languages in allowing some objects of transitive verbs to be encoded with free pronouns while others are encoded with pronominal prefixes. Adang is unique in its nearly obligatory use of classifiers (section 4.3), which are optional in other Alor-Pantar languages.

References Aikhenvald, Alexandra Y. 2000. Classifiers: A typology of noun categorization devices. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. 2006. Serial verb constructions in typological perspective. In Alexandra Aikhenvald & R.M.W. Dixon (eds.), Serial verb constructions: A cross-linguistic typology, 1–68. Oxford: Oxford University Press. Baird, Louise. n.d. Pitungbang frog story. Unpublished ms. Fedden, Sebastian, Dunstan Brown, Greville Corbett, Gary Holton, Marian Klamer, Laura C. Robinson & Antoinette Schapper. 2013. Conditions on pronominal marking in the AlorPantar languages. Linguistics 51(1). 33–74.

Adang

283

Grinevald, Colette. 2004. Classifiers. In Booij, Geert, Christian Lehmann, Joachim Mugdan, & Stavros Skopeteas (eds.), Morphology: An international handbook on inflection and word-formation, 1016–1031. Berlin: Walter de Gruyter. Haan, John W. 2001. The grammar of Adang, a Papuan language spoken on the island of Alor, East Nusa Tenggara, Indonesia. PhD thesis, University of Sydney. Hamilton, Amanda, Jawee Perla & Laura C. Robinson. 2013. A psycholinguistic assessment of language loss in Eastern Indonesia: Evidence from the HALA project. In Mari C. Jones & Sarah Ogilvie (eds.) Keeping languages alive: Documentation, revitalization, and pedagogy, 16–28. Cambridge: Cambridge University Press. Holton, Gary, Marian Klamer, František Kratochvíl, Laura C. Robinson & Antoinette Schapper. 2012. The historical relation of the Papuan languages of Alor and Pantar. Oceanic Linguistics 51(1). 86–122.

Antoinette Schapper

6 Kamang 1

The language scene

286

2 2.1 2.2 2.3 2.4 2.5 2.5.1 2.5.2 2.5.3

287 Phonology 287 Vowels 289 Consonants 291 Phonotactics 292 Stress 294 Morphophonemics 294 Reduplication 296 Prefixes Suffixes, enclitics and gemination

3 3.1 3.1.1 3.1.2 3.2 3.3

Basic clausal syntax 297 299 Predicate types 300 Verbal predicates 302 Non-verbal predicates 304 The post-predicative slot 305 Elevational marking

4 4.1 4.2 4.2.1 4.2.2 4.3 4.4 4.5

Noun phrases 306 307 Attributes and relative clauses 308 The numeral phrase Numeral quantifiers and the numeral phrase 309 Non-numeral quantifiers 310 Demonstratives and articles 311 Possessors 313 NP-appositional slot

5 5.1 5.2 5.3 5.4 5.5

Pronouns 314 315 Basic pronouns 317 Reflexive pronouns 318 Possessive pronouns 319 Quantifying pronouns 320 Focus pronouns

297

6 Agreement and locational prefixes 321 322 6.1 The agreement (AGR) slot 323 6.1.1 Obligatorily prefixed verbs 324 6.1.2 Unobligatorily prefixed verbs 324 6.1.2.1 Valency-preserving prefixation

308

286

Antoinette Schapper

P-adding prefixation 327 The incorporated (INCORP) slot 328 mi- ‘IN’ 330 wo- ‘AT’ The self-benefactive (SBEN) slot

6.1.2.2 6.2 6.2.1 6.2.2 6.3 7

327

331

7.1 7.2 7.3 7.3.1 7.3.2

Suffixal morphology: dynamicity, aspect and dependency 332 marking 332 Verbal auxiliary suffix 334 Aspectual suffixes 336 Dependency-marking suffixes 336 Dependent uses 338 Independent uses

8 8.1 8.2 8.3 8.4 8.5 8.6

Complex predications and serial verb constructions 340 The light verb me 342 Aspectual serialization 344 Causative serialization 345 Resultative serialization 346 Adverbial serialization 347 Motion serialization

9

Discussion

References

340

348

349

1 The language scene1 Kamang (ISO 639-3 code: woi) is spoken by around 6,000 people in north-central Alor. The figure of 16,500 speakers given in the Ethnologue (Lewis 2009) is inflated. It subsumes several closely related languages under the label ‘Kamang’. In actuality, the Kamang language, as designated by speakers and their neighbors and as clearly divisible on linguistic grounds, is limited to the northwestern area of the region marked as Kamang on the Ethnologue map.2 There are three main dialects of Kamang: Western Kamang, Lowland Kamang and 1 Field work on Kamang was supported by a small grant from the Foundation for Endangered Languages and the project grant 08-EuroBABEL-OP-O25 from De Nederlandse Organisatie voor Wetenschappelijk Onderzoek. I am very grateful to Wim Stokhof for his generous and enthusiastic support of my work on Kamang. Many thanks go to him, as well as to Ger Reesink, Hein Steinhauer and Marian Klamer who all commented on early drafts of this sketch. All errors are of course mine. 2 See the map at URL: http://www.ethnologue.com/show_map.asp?name=IDN&seq=90. The Kamang-marked area on the Ethnologue’s map covers five groups: Kamang, Suboo, Moo, Tiee and Manetaa (also known as Manet or Manetee).

Kamang

287

Upland Kamang. Each is characterized by certain sound changes and by a range of distinct lexemes. Speakers of western Kamang typically also speak Abui and their dialect shows substantive lexical influence from that language. Traditionally, the Kamang people lived in the central mountainous area of the island. However, during the 1970s many villages were compelled by Indonesian administrators to move down to the lower lying coastal areas. According to community elders, Kamang culture has been considerably eroded since resettlement. Today, the Kamang language is severely endangered. Kamang children and adolescents are rarely more than passive speakers of their language. Whilst parents and elders may speak amongst themselves in Kamang, child directed speech is typically in the local variety of Malay. In the several domestic settings I have observed, even where children are spoken to in Kamang, they invariably answer in Malay. Language materials and preliminary description of Kamang grammar and culture are found in Stokhof (1976, 1977, 1978, 1979, 1982, 1983) under the name Woisika.3 More recently, aspects of Kamang morphosyntax have been discussed in Schapper (2011), Schapper & Klamer (2011), Schapper & Manimau (2011), and Klamer & Schapper (2012). The data in this sketch come from linguistic fieldwork undertaken by the author in 2010 and 2011 on the Lowland Kamang dialect as spoken in Atoitaa and Sama villages and the Upland Kamang dialect as spoken in Takailubui. The corpus is composed of a lexicon of approximately 1850 words and nine hours recording encompassing ten natural speech texts, numerous responses to different picture and video stimuli, and several hundred elicited sentences. I also make use of some examples taken from Stokhof’s work, as marked throughout the text. The glossing and morphological segmentation of these is, however, from my own analysis.

2 Phonology 2.1 Vowels Kamang has ten vowel phonemes, five short and five long (Table 1). Short and long vowels contrast chiefly in duration; in my data quality differences are found to be minimal (contra Stokhof 1979). 3 This name is not used here as it is recognized neither by the Kamang speakers themselves nor by other Alorese ethnolinguistic groups as a name for a language or people. ‘Woisika’ means ‘Stone Stack’ (< woi ‘stone’, sika ‘stack’) in Kamang and refers to a village within the Kamang area. Today, Woisika is a desa officially known as ‘Waisika’ due to an early spelling mistake on the part of Indonesian administrators.

288

Antoinette Schapper

Table 1: Kamang vowel phonemes

HIGH MID LOW

FRONT

BACK

i iː e eː

u uː o oː a aː

Minimal pairs illustrating the phonemic contrastiveness of long and short vowels are presented in (1). Long vowels are represented in the practical orthography used later in this sketch by the doubling of the vowel graph (e.g., /a:/ is represented as {aa}). Minimal pairs for vowel phonemes (1)

/i/ ≠ /iː/

:

/asi/ ‘faeces’

≠

/asiː/ ‘salt’

/e/ ≠ /eː/

:

/te/ ‘go up’

≠

/teː/ ‘dig’

/a/ ≠ /aː/

:

/apa/ ‘this’

≠

/apaː/ ‘shadow’

/o/ ≠ /oː/

:

/ko/ ‘yam’

≠

/koː/ ‘stay’

/u/ ≠ /uː/

:

/su/ ‘three’

≠

/suː/ ‘heavy’

Word-final short vowels are typically followed by a partial glottal closure [ ʔ ] when they occur in phrase final position with a falling pitch contour. This glottal is purely phonetic and drops away when the lexeme is suffixed or where the lexeme occurs in non-phrase-final positions with a continuing pitch contour. In Lowland Kamang there are a few lexical items in which non-final /u/ can be realized as either [u] or [i], as in the example lexemes in (2). Examination of cognates in related languages points to [u] being etymologically prior. Lexemes with alternations in realization of /u/ (2)

/sube/

‘chicken’

[sube] ~ [sibe]

/munaːu/

‘which’

[munaːw] ~ [minaːw]

/tulen/

‘divide’

[tulen] ~ [tilen]

/tui/

‘mouse’

[tui] ~ [tiː]

Final /i/ in bisyllabic words of the shape (C)VCi where V is either /a/ or /i/ are frequently dropped in all varieties of Kamang. Examples are given in (3).

Kamang

(3)

289

Lexemes with dropping of final /i/ /ami/ ‘breast’ [ami] ~ [am] /api/ ‘fish’ [api] ~ [ap] /wati/ ‘sun’ [wati] ~ [wat] /ili/ ‘water’ [ili] ~ [il] /iti/ ‘call’ [iti] ~ [it] /bini/ ‘seed’ [bini] ~ [bin]

2.2 Consonants Kamang has 14 consonant phonemes: six oral stops, three nasal stops, two oral fricatives, two glides and one liquid (Table 2). In addition, there are two glottal consonants (bracketed) which have only a very marginal phonemic status (discussed further below). Most consonant phonemes are represented orthographically with the graph that is identical to the IPA representations. Differences are noted in angled brackets.4 Table 2: Kamang consonant phonemes ALVEOLAR

BILABIAL

PLOSIVE FRICATIVE NASAL GLIDE LATERAL

p ɸ

b m w

t s

PALATAL

d†

VELAR k

GLOTTAL g‡

(ʔ) (h)

ŋ

n j l

† is used for the [r] allophone of /d/ ‡ is used for the [w] allophone of /g/; is used for the [j] allophone of /g/ and for /j/

Minimal pairs illustrating the contrastiveness of Kamang consonant phonemes are presented in (4):

4 These orthographic choices have been made in consultation with speakers of Kamang and are heavily influenced by Indonesian spelling norms. Nevertheless, they are judged by them to be the best possible representation of the sounds of the language and are adhered to here. Underlying forms of phonemes are provided for the reader in an extra line of glossing in the examples.

290

(4)

Antoinette Schapper

Minimal pairs for consonant phonemes /p/ ≠ /b/ ≠ /ɸ/ ≠/m/: /pe/ ‘pig’, /be/ ‘cricket’, /ɸe/ ‘go down’, /me/ ‘come’ /t/ ≠ /d/ ≠ /s/ ≠ /n/: /taː/ ‘ON’, /daː/ ‘song’, /saː/ ‘ko. trap’, /naː/ ‘NEG’ /k/ ≠ /g/: /ket~ket/ ‘k.o. sound’, /get/ ‘3.OTHER’ /k/ ≠ /ŋ/ ≠ /n/: /=ak/ ‘DEF’, /aŋ/ ‘that.NKWN’, /an/ ‘THUS’ /m/ ≠ /n/ ≠ /l/: /maː/ ‘depart’, /naː/ ‘NEG’, /laː/ ‘burn’ /b/ ≠ /w/: /=beː/ ‘also’, /weː/ ‘blood’ /g/ ≠ /j/ ≠/w/: /gaː/ ‘rice bushel/, /jaː/ ‘reach’, /waː/ ‘mushroom’

Comparative evidence shows that the marginal Kamang glottal consonants /ʔ/ and /h/ originate in historical allophones of /k/ and /s/ respectively. Synchronically, there are no minimal pairs contrasting /ʔ/ with /k/, and /h/ with /s/, and the distribution of both pairs is largely complementary. That is: /ʔ/ is limited to non-final codas after short vowels in Lowland Kamang dialect,5 while /k/ occurs almost exclusively in onsets and final codas, and only a few medial codas; /s/ occurs almost exclusively in onsets and only a handful of word-final codas, while /h/ is limited to final and non-final codas and changes to [s] when suffixed with /-iŋ/ ‘SETTING’, but not other suffixes with the shape -VC. In short, the glottal consonants can only be considered very weakly phonemic in Kamang. Four Kamang consonant phonemes have conditioned allophony. This is set out along with the conditioning environments of the allophones in (5). Unless otherwise mentioned, consonant phonemes closely conform to their IPA values. Consonant allophony (5) a. /d/ > [d] ~ [r ] / V_V > [d] / elsewhere b. /g/ > [w] > [g] ~ [j]

e.g., /kadiː/ ‘house’ [kadiː] ~ [kariː] /dum/ ‘child’ [dum]

/ _ [+back vowel] e.g., /gobaleː/ ‘too’ [wobaleː] / elsewhere /ga/ ‘3.AGT’ [ga] ~ [ya]

c. /k/ > [k] ~ [x] / V_V [k] / elsewhere

e.g., /bakai/ ‘break’ [bakaj] ~ [baxaj] /kai/ ‘cheer’ [kaj]

d. /w/ > [w] ~ [β] / V_V > [w] / elsewhere

e.g., /tewe/ ‘go up’ [tewe] ~ [teβe] /weː/ ‘blood’ [weː]

5 In Upland Kamang, the glottal stop phoneme is lost, but the preceding vowel shows compensatory lengthening, e.g., Lowland Kamang /-paʔta/ ‘wing’, but Upland Kamang /-pa:ta/ ‘wing’ (cf. Abui /-baːkai/ ‘wing’ where k is preserved).

Kamang

291

2.3 Phonotactics The vast majority of Kamang roots are of two syllables in length, but there also small numbers of monosyllabic and trisyllabic roots. Kamang allows syllables of the shape (C)V(C). Codas are, however, highly restricted in terms of the consonants that they can contain. Table 3 presents an overview of the distribution of the individual consonants. We see that bilabial consonants, voiced stops and oral fricatives do not occur in medial or final codas. The velar nasal is the only consonant to be restricted from occurring in onsets. Bracketed segments refer to historically restricted allophones (discussed in section 2.2). Table 3: Consonant phoneme distribution in roots Initial

Medial onset

Medial coda

Final

p

+

+

–

–

b

+

+

–

–

t

+

+

+

+

d

+

+

–

–

k

+

+

(ʔ)

+

g

+

+

–

–

ɸ

+

+

–

–

s

+

+

(h)

(h)

m

+

+

–

–

n

+

+

+

+

ŋ

–

–

+

+

l

+

+

+

+

w

+

+

–

–

j

+

+

–

–

Sequences of vowels are permitted in Kamang. In sequences of two vowels where the second vowel is a high vowel, /i/ or /u/, these are realized as offglides, [j] and [w] respectively. The first vowel in the sequence may be long or short. Examples of the observed combinations are given in (6) for vowel sequences with /i/ surfacing [j] and in (7) for vowel sequences with /u/ surfacing as [w].

292

(6)

(7)

Antoinette Schapper

Off-glides formed out of /i/ /ai/: /aɸai/ ‘gully’ /aːi/: /bai/ ‘great’ /ei/: /atei/ ‘seed’ /eːi/: /aleːi/ ‘k.o. garden’ /oi/: /woi/ ‘stone’ /oːi/: /moːi/ ‘sun’ /ui/: /bui/ ‘chop’ /uːi/: /wuːi/ ‘slave’

[afaj] [baːj] [atej] [aleːj] [woj] [moːj] [buj] [wuːj]

Off-glides formed out of /u/ /au/: /auka/ ‘dry’ /aːu/: /kaːu/ ‘kapok cotton’ /ou/: /kou/ ‘raw’

[awka] [kaːw] [kow]

Note that the vowel /u/ does not become a glide following a front vowel; the sequences /iu/, /i:u/ and /e:u/ are not attested in Kamang at all. The vowel sequence /eu/ is attested only once in the lexeme /ɸeuŋ/ ‘pigeon’, but the vowels syllabify into different syllables /ɸe.uŋ/. The sequence /oːu/ is also not attested, but its short counterpart is found (e.g., /kou/ ‘raw’ given in 7); this lack may just represent a sampling deficiency, rather than a systematic gap in the language. When the second vowel in a sequence is long, it does not become a glide: for instance, /a.iː/ ‘rain’ is realized as [a.ˈiː] and /auː/ ‘pestle’ as [a.ˈuː].

2.4 Stress Stress in Kamang is largely phonemic, with weight sensitivity playing a role in the stress assignment of only a portion of disyllabic roots. Although unpredictable, stress carries a low functional load in the language with no stressdifferentiated minimal pairs being found between monomorphemic words. As such, stress is not marked outside this sub-section. On dysllabic roots, stress assignment depends on the presence of heavy syllables in the word root. A heavy syllable is a syllable with a consonant (including glides) in the coda and/or a long vowel in the nucleus. Light syllables lack both codas and long vowels. Where a disyllabic root has a heavy syllable, stress is non-phonemic and regularly falls on the heavy syllable (8a). Where a disyllabic root is composed of only light syllables, stress is phonemic and is lexically assigned to either syllable (8b). Disyllabic roots in which both syllables are heavy are rare, but stress appears to be phonemic (8c).

Kamang

(8)

Stress on disyllablic a. LH /kamaŋ/ /iten/ /piaː/ /kateː/ /tapui/ /matei/

b.

c.

roots ‘Kamang’ ‘ripe’ ‘different’ ‘eat’ ‘crab’ ‘bat’

[ka.ˈmaŋ] [i.ˈten ~ i.ˈton] [pi.ˈaː] [ka.ˈteː] [taˈpuj] [maˈtej]

‘firmly’6 ‘fragrant’ ‘wing’ ‘trap’ ‘firefly’ ‘chili’

[ˈlik.ka] [ˈmuːl.ma] [ˈ-paʔ.ta ~ ˈ-paː.ta] [ˈkajta] [ˈkaŋ.pi] [ˈluːka]

HL

/likka/ /muːlma/ /-paʔta/ /kaita/ /kaŋpi/ /luːka/

ˈLL

/ˈsibe/ /ˈkine/ /ˈtama/ /ˈuda/ /ˈsela/ /ˈdaɸe/

‘chicken’ ‘knife’ ‘sea’ ‘bird sp.’ ‘saddle’ ‘needle’

[ˈsi.be ~ ˈsu.be] [ˈki.ne] [ˈta.ma] [ˈu.ra] [ˈse.la] [ˈda.ɸe]

LˈL

/suˈe/ /piˈe/ /waˈte/ /kuˈme/ /puˈna/ /kuˈda/

‘come’ ‘crocodile ‘coconut’ ‘snake’ ‘smoke’ ‘horse’

[su.ˈe ~ ˈseː] [pi.ˈe] [wa.ˈte] [ku.ˈme] [pu.ˈna] [ku.ˈra]

ˈHH

/ˈtaːŋsok/ /ˈuːtiŋ/ /ˈkaupai/ /ˈlaŋsal/ /ˈlaːwaŋ/

‘formerly’ ‘pot’ ‘buffalo’ ‘bird sp.’ ‘bee’

[ˈtaːŋ.sok] [ˈuː.tiŋ] [ˈkaw.paj] [ˈlaŋ.sal] [ˈlaː.waŋ]

HˈH

/maːˈɸaŋ/ /teːˈkaŋ/

‘brave’ ‘pole’

[maː.ˈɸaŋ] [teː.ˈkaŋ]

293

On the small number of trisyllabic roots in the corpus, stress is phonemic. Heavy syllables do not attract stress. Examples are given in (9). 6 This is one of only two items in the corpus where /k/ appears in a non-final coda. The other is the onomatopoeic /takkeː/ ‘house gecko’. Lowland Kamang /likka/ ‘strong’ is /lilka/ in Upland Kamang, and so the geminate /k/ appears to have arisen through an original liquid consonant /l/ harmonizing with the following /k/.

294

(9)

Antoinette Schapper

Stress on trisyllablic roots a. Initial /ˈumakiː/ /ˈtakabau/ /ˈsopale/ /ˈmaukubai/ b.

Medial

c.

Final

/kaˈwasa/ /taˈɸoda/ /kaˈkisiŋ/ /aˈwiŋga/ /tapkiˈdi/ /taoˈmaŋ/

‘quail’ ‘frog’ ‘snail’ ‘cat’

[ˈu.ma.kiː] [ˈta.ka.baw] [ˈso.pa.le] [ˈmaw.ku.baj]

‘rich’ ‘iron’ ‘ant sp.’ ‘door’ ‘spider’ ‘crazy’

[ka.ˈwa.sa] [ta.ˈɸo.ra] [ka.ˈki.siŋ/ [a.ˈwiŋ.ga]

[tap.ki.ˈri] [ta.o.ˈmaŋ]

Suffixes never cause stress shift, but prefixes can. All prefixes except the patientive series appear to attract stress (see Stokhof 1979: 87–88). It is only in the presence of such stress shifting prefixes that I have observed any stressdifferentiated words (but not roots!) in Kamang, for instance, /teːˈkaŋ/ ‘pole’ versus /teː-kaŋ/ [ˈteːkaŋ] CMN.DAT-good ‘be good to one another’ or /si=jaː/ [siˈjaː] 1PL.INCL.AGT=reach ‘we go, reach’ versus /si-jaː/ [ˈsijaː] 1PL.INCL.GEN-stool ‘our stool’.

2.5 Morphophonemics Kamang displays a wide range of morphophonemic changes triggered by the attachment of affixes and clitics to roots. In this section, I outline only the most frequent processes.

2.5.1 Reduplication Kamang has three patterns of reduplication: (i) CV reduplication of verbs; (ii) full reduplication of nouns and adjectives, and; (iii) irregular partial reduplication of numerals. CV reduplication applied to verbs denotes durative aspect and/or intensity of action depending on the lexical semantics of the reduplicated verb. In this kind of reduplication, the first two segments on the left edge of the word are reduplicated, as in (10).

Kamang

(10)

a. b.

CV reduplication of verbs /tadak/ > /ta~tadak/ [taˈtarak] RDP~close.together close.together /-sol/ begin

>

/-so~sol/ [-soˈsol] RDP~begin

295

‘very close together’

‘very beginning’

Where the V of the root is a front vowel, then the copy vowel of the reduplicant is /u/, as in (11a–b). Where the first V of the root is a long vowel, the copy vowel of the reduplicant is a short vowel, as in (11c). Similarly, where there is a diphthong (i.e., VG), only the V is copied in the reduplicant, as in (11d). (11)

a.

/ɸewe/ go.down

>

b.

/bisik/ snap.off

>

c.

/poː/ break.off

>

/waisin/ hear

>

d.

RDP ~go.down

/ɸu~ɸewe/

(*ɸeɸewe) ‘keep on going down’

/bu~bisik/ RDP ~snap.off

(*bibisik) ‘snap right off’

/po~poː/

(*poːpoː) ‘snap right off’

RDP ~break.off

/wa~waisin/ RDP ~hear

(*waiwaisin) ‘keep on listening’

Full reduplication applies to nouns and adjectives. With nouns it denotes plurality of kinds, as in (12). With adjectives it denotes intensification, as in (13).

(12)

a. b.

(13)

a. b.

Full reduplication of nouns /taːu/ > /taːu~taːu/ where RDP~where ‘wherever, different kinds of places’ /atoi/ bird

>

/atoi~atoi/ RDP~bird

‘different kinds of birds’

Full reduplication of adjectives /sosaŋ/ > /sosaŋ~sosaŋ/ RDP~handsome handsome

‘very handsome’

/buʔsei/ strong

‘very strong’

>

/buʔsei~buʔsei/ RDP~strong

Finally, numerals have an irregular pattern of partial reduplication to create distributive numerals. These are set out in Table 4. With the numerals ‘one’ to ‘three’, it is the initial (C)V that is reduplicated. With ‘four’, the initial CV.V

296

Antoinette Schapper

redupulicates, but with /i/ becoming a glide in the reduplicant such that it is only monosyllabic [bje ~ ]. The numerals with initial wesing (five to nine) repeat the CVCV [wesi ~ ]. Where a numeral is a complex expression (being made up of multiple independent numeral words), then only the last element of the complex numeral is reduplicated, as illustrated for ‘100’ to ‘1000’. Table 4: Kamang distributive numerals 1 by 1 2 by 2 3 by 3 4 by 4 5 by 5 6 by 6 7 by 7 8 by 8 9 by 9 10 by 10 20 by 20 100 by 100 1000 by 1000

/no~nok/ /o~ok/ [oˈ ʔok] /su~su/ /bie~biat/ [bjebiˈ jat] wesi~wesiŋ wesi~wesiŋnok wesi~wesiŋok wesi~wesiŋsu wesi~wesiŋbiat ataːk no~nok ataːk o~ok ataːk nok waːl no~nok ribu no~nok

2.5.2 Prefixation Agreement prefixes of the PATIENTIVE inflection have two forms, full and reduced (see section 4.4 and section 6.1 on the functions of the inflection on nouns and verbs respectively). Full forms retain the /a/ vowel in the singular of the paradigm and are used productively on consonant- and vowel-initial roots. Reduced forms have lost the /a/ vowel in the singular of the paradigm and are only found on a small number of vowel initial roots. There is no rule which can predict which vowel initial roots will take the full or the reduced forms of the inflection. Examples of both paradigms on vowel initial roots are provided in Table 5. Table 5: Forms of the PATIENTIVE prefixal paradigm

1SG 2SG 3 CMN 1PL.EXCL 1PL.INCL 2PL

FULL

/-ai/ ‘vagina’

/-oː/ ‘birth’

REDUCED

/-ouko/ ‘mother’

/-eh/ ‘bite’

/na-/ /a-/ /ga-/ /ta-/ /ni-/ /si-/ /i-/

/na-ai/ /a-ai/ /ga-ai/ /ta-ai/ /ni-ai/ /si-ai/ /i-ai/

/na-oː/ /a-oː/ /ga-oː/ /ta-oː/ /ni-oː/ /si-oː/ /i-oː/

/n-/ /Ø-/ /g-/ /t-/ /ni-/ /si-/ /i-/

/n-ouko/ /Ø-ouko/ /g-ouko/ /t-ouko/ /ni-ouko/ /si-ouko/ /i-ouko/

/n-eh/ /Ø-eh/ /g-eh/ /t-eh/ /ni-eh/ /si-eh/ /i-eh/

Kamang

297

2.5.3 Suffixation, encliticization and gemination A suffix or enclitic of the shape VC is reduced to a simple C when it attaches to a vowel final root, as illustrated in (14).

(14)

Suffix-enclitic reduction a. /ko:/ + /-ih/ > ‘stay’ ‘CSEQ’ b.

/pe/ ‘pig’

+

/=al/ ‘CONTR’

>

[ko:h] ‘stay such that. . .’ [pel] ‘a pig (not another animal)’

The enclitic =ou ‘RESTR’ also often reduces to simply /u/ when it attaches to a vowel-final root. Thus:

(15)

Reduction of =ou /kine/ + /=ou/ ‘knife’ ‘RESTR’

>

[kineu] ‘a knife (and only a knife)’

Gemination of final consonants occurs when a suffix or enclitic of the shape V(C) attaches to a consonant final root, as in (16). All word-final consonants except /ŋ/ can geminate on suffixation. In these cases, /ŋ/ is realized as [n], as in (16c).

(16)

Final consonant gemination a. /bilen/ + /-ih/ > [bilennih] ‘write such that. . .’ ‘write’ ‘CSEQ’ b.

/dum/ ‘child’

+

/=a/ ‘SPEC’

>

[dumma] ‘a (specific) child’

c.

/boŋ/ ‘tree’

+

/=ak/ ‘DEF’

>

[bonnak] ‘the tree’

3 Basic clausal syntax The basic clause in Kamang is composed of a predicate (PRED) and its arguments (ARG). A predicate is the clausal element that may host suffixal morphology (-AUX and -ASP) when in the final predicate position. Kamang predicates may be avalent (that is, take no argument), but more frequently are monovalent or

298

Antoinette Schapper

bivalent. An argument always precedes its predicate. The predicate may be followed by an item in the post-predicative slot (END). Thus, a basic, monopredicative clause in Kamang has the structure given in (17). Optional elements are given in brackets.

(17)

Mono-predicate clause structure [(ARGNP) (ARGNP) Pred(-AUX)(-ASP) (END)]CLAUSE

The examples in (18) illustrate the difference between a predicate and a postpredicative element. In (18a) the clausal negator naa ‘NEG’ occurs in the END slot following the monovalent verbal predicate, -bo’ra ‘die’, which is marked with the imperfective aspect suffix -si ‘IPFV’. Moving the aspectual suffix onto the clause-final, but non-predicative negator as in (18b) is ungrammatical. (18)

a.

Dumma gabo’rasi naa. dum=a ga-bo’da-si child=SPEC 3.PAT-die-IPFV ‘The child is not dying.’

naa NEG

b. *Dumma gabo’ra naasi. dum=a ga-bo’da naa-si child=SPEC 3.PAT-die NEG-IPFV A Kamang clause may also have more than one predicate (see section 8). In this case, final and non-final predicates are distinguished, as represented in the template in (19). A final predicate can host an aspectual suffix (-ASP), either with or without the addition of the verbal auxiliary suffix (-da ‘AUX’; see section 7.1). Non-final predicates cannot host any suffixes.

(19)

Multi-predicate clause structure [. . . (PREDNON-FINAL) (PREDNON-FINAL) PREDFINAL(-AUX)(-ASP) (END)]CLAUSE

Examples (20) and (21) illustrate the distinction in aspectual marking between non-final and final predicates. In each case only the final (third) predicate hosts an aspectual suffix; the suffixation of a non-final (first or second) predicates would result in ungrammaticality. (20)

Tama wooseh we yaasi. tama [goo-seh]PRED1 [we]PRED2 [yaa-si]PRED3 sea 3.AST-lower go depart-IPFV ‘The sea lowers and goes way out.’

Kamang

(21)

Peu yeeng see akmi lokma. pe=ou geeng [sue]PRED1 [ak=mi]PRED2 pig=RESTR 3.RESTR come here=IN ‘Pigs have come here and dug up the earth.’

299

[lok-ma]PRED3 dig.up-PFV

3.1 Predicate types Kamang distinguishes morphologically between verbal and non-verbal predicates. Verbal predicates are those predicates that directly host aspectual suffixes when in the final predicate position. A non-verbal predicate is one that must first be suffixed with the verbal auxiliary before it can host an aspectual suffix in the final predicate position (see section 7.1). For instance, compare the suffixing behavior of the monovalent verbal predicate poo ‘be broken’ in (22) and the adjectival predicate piila ‘be sharp’ in (23). In the (a) examples we see that both types can occur as simple predicates without any aspectual marking. However, in the (b) examples we see a difference between the behavior of the predicates when marked with the perfective aspectual suffix -ma ‘PFV’. While the verbal predicate directly hosts the suffix (22b), the adjectival predicate must be first suffixed with -da ‘AUX’ before it can host the same suffix (23b). Verbal predicate (22) a. Bong pa poo. bong=apa poo wood=THIS broken ‘This wood is broken.’ Adjectival predicate (23) a. Bong pa piila. bong=apa piila wood=THIS sharp ‘This wood is sharp.’

b. Bong pa pooma. bong=apa poo-ma wood=THIS broken-PFV ‘This wood is already broken.’

b. Bong pa piilarama. bong=apa piila-da-ma (*piila-ma) wood=THIS sharp-AUX-PFV sharp-PFV ‘This wood is already sharp.’

The types of verbal predicates and non-verbal predicates will be discussed and illustrated in section 3.1.1 and section 3.1.2 respectively. Discussion will be limited to mono-predicative clauses for now. Explicit treatment of complex multi-predicative clauses will be reserved for section 7.

300

Antoinette Schapper

3.1.1 Verbal predicates We saw above that verbal predicates are those predicates that directly host aspectual suffixes when in the final predicate position. Verbal predicates can be preliminarily divided into classes on the basis of three morphosyntactic properties: (i) valency: the number of arguments for which each verb subcategorizes, zero to three arguments, and; (ii) flagging of arguments: the obligatory marking of an argument NP with the light verb me ‘TAKE’, either P or T. (iii) argument agreement: the presence of an agreement prefix indexing an argument, one of S, P or R. Table 6 summarizes the verb classes that arise out of these criteria. Each of these verbal predicate types is discussed and illustrated below. Table 6: Overview of Kamang verb types’ argument encoding VERB

ARGUMENT

TYPE

Avalent Monovalent Bivalent Defective bivalent ~ trivalent

ENCODING

V NPS V NPS ARGS-V NPA NPP V NPA NPP ARGP-V NPA NPP me V NPA NPT me NPR ARGR-V

Avalent verbs form a small class which semantically denotes times of the day. Examples of two such verbs are given in (24). Kamang does not use a “dummy subject”, such as is necessary in the English versions of such predications. As they are without arguments, these verbs obviously do not have any argument flagging or agreement.

(24)

a.

Avalent verbs Itunma. itun-ma late.afternoon-PFV ‘It’s late afternoon.’

b.

Mitafeensi. mi-tafeeng-si IN-day.break-IPFV ‘It’s becoming day’

Monovalent verbs subcategorize for a single argument, S. There are two primary lexical classes of monovalent verbs, those that do not obligatorily prefix

Kamang

301

S and those that do, such as the verbs in (25a) and (25b) respectively. S is never flagged.

(25)

a.

Monovalent verbs Unprefixed S Markus mu’tanma. Markus mu’tan-ma Markus fall-PFV ‘Markus fell already.’

b.

Prefixed S Markus gamanteima. Markus ga-mantei-ma Markus 3.PAT-thirst-PFV ‘Markus is thirsty already.’

Bivalent verbs subcategorize for two arguments, A and P. There are two primary lexical classes, those that do not obligatorily prefix P and those that do. The unprefixing type is illustrated in (26a), while the P prefixing type is illustrated in (26b).

(26)

a.

Bivalent verbs Unprefixed P Na Markus bo’nama. na Markus bo’na-ma 1SG.AGT Markus hit-PFV ‘I have hit Markus.’

b.

Prefixed P Na Markus gatakma. na Markus ga-tak-ma 1SG.AGT Markus 3.PAT-see-PFV ‘I have seen Markus.’

The verbal predicates we have looked at thus far have had no argument flagging. So-called “defective” verbs are a minor class of polyvalent verbs (around 10 members) that require the light verb me ‘TAKE’ flagging one of their arguments (for more details of the status of me with these verbs, see Klamer & Schapper 2012). Trivalent verbs in Kamang are always defective. They invariably flag T with me and index R with an agreement prefix on the final verb. For instance, in (27a) and (27b) the verbs -tulen ‘divide’ and -n ‘give’ both have me flagging T and at the same time index R with locative and genitive inflections respectively (see section 6.1 on the different prefixal inflections).

(27)

Defective trivalent verbs a. A T R AGRR-V Markus patei me dumma wotulen. Markus patei me dum=a go-tulen Markus corn TAKE child=SPEC 3.LOC-divide ‘Markus divides corn amongst the children.’

302

Antoinette Schapper

b.

A T AGRR-V Maria falak me nen. Maria falak me ne-n Maria cloth TAKE 1SG.GEN-give ‘Maria gives me a cloth.’

Whereas all trivalent verbs are defective, only a handful of all the bivalent verbs in Kamang are defective. Defective bivalent verbs flag P with me, but never index P with an agreement prefix. The pattern is illustrated by the two defective verbs pidak ‘bury’ and mota ‘discard’ given in (28a) and (28b).

(28)

Defective bivalent verbs a. A P V Na yeok me mota. na ge-ok me mota 1SG.AGT 3.GEN-two TAKE discard ‘I throw the second one away.’ b.

A P Marten bakaa itonak me pidak. Marten bakaa iton=ak me Marten betel.nut ripe=DEF TAKE ‘Marten buries the betel nut.’

V pidak bury

The appearance and alignment of agreement prefixes on verbs is treated in depth in section 6.

3.1.2 Non-verbal predicates The definition of a non-verbal predicate given above is any predicate that cannot directly host aspectual suffixes. There are three major types of non-verbal predicates in Kamang: predicates headed by adjectives, predicates headed by postpositions, and nominal predicates. These elements act as predicates in their own right; Kamang has no copula with which to “support” non-verbal predications. Adjectival predicates are of two types. First, there are simple adjectives which take one argument, such as suu ‘heavy’ in (29). Second, there are labile adjectives which can take either no argument or a single argument. Semantically, labile adjectives depict properties that can be attributed both to a specific entity and to an ambient weather condition. For instance, the adjective kamal

Kamang

303

‘cold’ can be used with a single argument in predications about the temperature of an entity, such as ili ‘water’ in (30a). Equally, kamal can occur without an argument to express that the ambient temperature is cold, as in (30b). Often ambient weather predications occur with a postpositional phrase (PP) denoting the location at which the temperature is found, but this is not obligatory (see Schapper forthcoming a).

(29)

(30)

Simple adjectival predicate Taas pa suu. taas=apa suu (*suu-si bag=THIS heavy heavy-IPFV ‘This bag is heavy.’

a.

b.

/

*suu-ma) heavy-PFV

Labile adjectival predicate Ili pa kamal. ili=apa kamal (*kamal-si cold-IPFV water=THIS cold ‘This water is cold.’

/

*kamal-ma) cold-PFV

Buk taa mi kamal. buk taa mi kamal mountain above IN cold ‘It’s cold on the mountain tops.’

Adjectives do not ever occur with an obligatory agreement prefix indexing their argument. They can, however, optionally occur with a locative agreement prefix indexing their argument in order to express greater affectedness. This pattern is also found with monovalent verbs and is discussed in section 6.1.2.1. Postpositional predicates are headed by one of the four Kamang postpositions (mi ‘in’, taa ‘on’, ii ‘under’, and wo ‘at’) and take an NP complement (i.e., PP = [NP POST]PRED), as in (31).

(31)

Postpositional predicate Nal Aapui mi. nal Aapui mi (*mi-si IN IN -IPFV 1SG Apui ‘I’m in Apui.’

/

*mi-ma) IN- PFV

Nominal predicates consist simply of two NPs juxtaposed to one another, as in (32). Neither predicate type can directly host aspectual suffixes.

304

(32)

Antoinette Schapper

Nominal predicate Epaa almakang. e-paa almakang 2SG.GEN-father person ‘Your father is a human.’

(*almakang-si person-IPFV

/

*almakang-ma) person-PFV

See section 7.1 on the use of the verbal auxiliary -da allowing non-verbal predicates to host aspectual suffixes.

3.2 The post-predicative slot: negation and intensification The three items that can occur in the post-predicative slot are given in Table 7. Typically, only one item can occupy the slot at any given time.7 Table 7: Items in the post-predicate slot

naa gee borang

GLOSS

FUNCTION

‘NEG’ ‘PROH’ ‘VERY’

general clausal negator negative imperative clausal intensification

Clausal negation is expressed in Kamang by means of the negator naa ‘NEG’ following the predicate, as in (33). The same negator is used for all types of predicates. Negative imperatives are marked with gee, as in (34). (33)

(34)

Nal isei maa katee naa. nal isei maa katee 1SG game edible eat ‘I don’t eat meat.’

naa NEG

Ala wobalee posansi yee. al=a gobalee posan-si arrogant-IPFV 2SG=SPEC too ‘Don’t you be so arrogant!’

gee PROH

7 Kamang also has adverbs, but the class is much larger than those found in post-verbal slot. The class includes temporal, manner and aspectual adverbs. In contrast to the items discussed here they all occur before the (final) clausal predicate, such as wobalee in (34).

305

Kamang

Intensification of a clause is expressed with borang ‘VERY’. This item occurs with monovalent verbal and adjectival predicates, as in (35a) and (35b). (35)

a.

Verbal Predicate Ili tutunma borang. ili tutun-ma borang water fire.hot-PFV VERY ‘The water’s already really hot.’

b.

Adjectival Predicate Ili bebarama borang. ili beba-da-ma borang water hot-AUX-PFV VERY ‘The water’s already really hot.’

3.3 Elevational marking An omnipresent feature of Kamang spatial discourse is elevation marking. Events and motions are persistently located in space by means of “elevationals” and “elevational motion verbs’. Elevationals express that a location is at a certain elevation relative to the deictic center, while elevational motion verbs denote that a motion occurs over or from/to a particular elevation. The Kamang elevational terms are given in Table 8. Table 8: Kamang elevation terms

ELEVATION

DIRECTION †

DISTANCE ‡

LEVEL HIGH

DIRECT INDIRECT

NEAR FAR

LOW

DIRECT INDIRECT

NEAR FAR

Elevationals

Elevational motion verbs From DC

To DC

mung

we

me

tung mutung tumung

te wete tewe

taang metaang taangme

fung muhung fumung

fe wehe fewe

yaang yaangme

† Direction has to do with the angle of the path taken or referenced location relative to the angle of the slope. Using a DIRECT elevational term means that the path taken follows the angle of the slope directly (i.e., at its steepest), whilst an INDIRECT elevational term means that the path traverses across the angle of the slope or that the referenced location is off to the side of angle of the slope. ‡ Distance is concerned with whether the path taken is short or long or the referenced location is near or far. Thus, using a NEAR elevational term means traversing across a slope for a short distance, while using a FAR one traversing across a slope for a long distance.

306

Antoinette Schapper

While elevational motion verbs behave simply as verbs, the Kamang elevationals are notable in that they constitute their own word class and accordingly have their own peculiar syntax. They occur in three syntactic positions in the clause. The first is immediately before a verbal predicate where it denotes the location at or from which an event takes place, as in (36) where mutung denotes the location from which the calling takes place. In the second position the elevational follows a motion verb specifying the resultant location of the motion, as in (37) where the elevational tung follows its corresponding elevational verb te. The final position is before an NP to denote the location of the NP referent, as in (38) where tung gives the location of wuleh ‘slope’. (36)

Nok sue koo mutung woitisi. nok sue koo mutung go-iti-si 3.LOC-call-IPFV one come stay LEVEL ‘Somebody was calling him from over there.’ (from Stokhof 1979)

(37)

Nal te tung. nal te 1SG go.HIGH.DIRECT ‘I go up top.’

(38)

tung high.direct

Markus tung wuleh sama kawailama. Markus tung guleh sama kawaila-ma Markus HIGH.DIRECT slope middle stumble-PFV ‘Markus stumbled slope up (which is) up there.’

See Stokhof (2012: 33–47) for more discussion of Kamang elevation-marked items.

4 Noun phrases The template of the Kamang noun phrase (NP) is presented in (39). The NP is maximally composed of a head noun (NHEAD) followed by its attribute (ATTR), a numeral phrase (NUMP), a relative clause (RC), a demonstrative (DEM) and an article (ART). Adnominal possessors are expressed by an agreement prefix on the head noun; a possessor NP (outside the NP) may optionally co-occur with the prefix to explicitly identify the possessor. Finally, the Kamang NP can occur with a range of items co-referential with it in a slot outside the NP, called here the NP-appositional (APPOS) slot.

Kamang

(39)

307

Template of the Kamang NP PSRNP [AGRPSR-NHEAD ATTR NUMP RC DEM ART]NP APPOS

The various elements of the NP and the NP-appositional slots will be exemplified briefly in the following sections.

4.1 Attributes and relative clauses Attributes and relative clauses both modify a NHEAD, but they differ in their precise marking of it, as will be outlined in this section. Attributes are simple adjectives or stative monovalent verbs that provide information about characteristics of the referent. There are no morpho-syntactic differences in the behavior of attributes that are adjectival (40a) or verbal (40b). (See section 3.1 on the differing suffixal behavior of adjectival predicates and monovalent verbal predicates). (40) a.

Adjectival attribute alma puk person big ‘big person’

b.

Verbal attribute arita aleen leaf wilt ‘wilted leaf’

Attributes occur directly following the head noun, preceding numerals, numeral classifiers and relative clauses in the NP. In (41) we see the numeral nok ‘one’ following the attribute saak ‘old’. In (42) we see the attribute kiding ‘old’ precedes a relative clause. (41)

Kata saak nok yedumma ganeng. ga-neng [kata saak nok]NP ge-dum=a hind old one 3.GEN-child=SPEC 3.PAT-with ‘There was only a grown-up hind with her stag.’ (from Stokhof 1978)

(42) Dum kidingbo faa parong mi, kui yera suboitamah . . . suboita-ma-ih [dum kiding=bo faa padong mi]NP kui geda IN dog 3.CONTR jump-PFV-CSEQ child small=REL earth mud ‘Then the child who was in the mud, his dog jumped up such that. . .’ Relative clauses in Kamang are restrictive; they identify a referent by restricting the head noun’s scope of reference to a single entity or set of entities. The head of a relative clause appears at the front of the RC marked with =bo ‘REL’

308

Antoinette Schapper

(also a general clause linker). The head is gapped (signaled with Ø) inside the RC, as seen in (43) and (44). (43)

(44)

Dumbo wosol gaoo yaangme pie. [dum=bo go-sol Ø ga-oo yaangme]NP child=REL 3.LOC-begin 3.PAT-born exit ‘The child that first came out was a crocodile.’ Mateus yepateibo si aia wobun. Mateus [ge-patei=bo si Ø ai=a]NP Mateus 3.GEN-corn=REL 1PL.INCL.AGT take=SPEC ‘The corn of Mateus which we took is hidden.’

pie crocodile

go-bun 3.LOC-hidden

4.2 Quantification Kamang lacks a coherent class of quantifiers. Quantification is done by a range of items from multiple word classes. In this section I discuss those quantificational items not discussed elsewhere in this sketch. The plural word nung occurs in the NP appositional slot and is accordingly discussed in section 4.5. Quantifying pronouns are discussed in section 5.4.

4.2.1 Numeral quantifiers and the numeral phrase A numeral phrase is composed of a classifier (CLF) and a numeral (NUM). The classifier precedes the numeral. A classifier may be specific to a class of entities. We see in (45) that the head noun pe ‘pig’ can take the large animal classifier tuk (45a) or the default classifier uh (45b). (45)

a.

pe [tuk su]NUMP pig CLF three ‘three pigs’

b.

pe [uh su]NUMP pig CLF three ‘three pigs’

The NUMP has two positions: (i) its unmarked position within the NP closed off by the article on the left periphery (46a), and; (ii) a marked position where the NUMP is post-posed into the NP-appositional slot outside the NP (46b). The latter position is less frequent and pragmatically marked, functioning to topicalize the enumerated referent. See section 4.5 for more on this NP-external position.

Kamang

(46)

309

Syntactic positions of the NUMP a. sibe uh sua [sibe [uh su]NUMP=a]NP chicken CLF three=SPEC ‘the three chickens’ b.

sibea uh su [sibe=a]NP [uh su]NUMP chicken=SPEC CLF three ‘the chickens, the three ones’

Kamang has an inflecting numeral classifier used exclusively in human reference. The inflectional pattern of the human classifier -ning ‘CLF.HUM’ follows that of the PATIENTIVE inflection (section 6.1). Examples (47) and (48) show the human classifier with a 3rd person and a 1st person exclusive inflection respectively. (47)

Male ganing su oi yefe. male ga-ning su oi ge-fe woman 3.PAT-CLF.HUM three thither 3.GEN-go.down ‘Those three women went down there.’ (from Stokhof 1978)

(48)

Nining su kon ako. ni-ning su kon ako 1PL.EXCL.PAT-CLF.HUM three only be.here ‘We are only three people here.’ (from Stokhof 1978)

4.2.2 Non-numeral quantifiers Kamang does not have a syntactic class of non-numeral quantifiers; items denoting many, few, a little etc. are adjectives and occur in the ATTR slot of the NP, i.e., to the left of the NP-defining article. Non-numeral quantifiers such as adu ‘many/much’ occupy the ATTR slot within the NP and cannot be postposed out of it, as seen in (49). (49)

a.

sibe adu=a chicken many=SPEC ‘the many chickens’

b.

*sibe=a chicken=SPEC

adu many

310

Antoinette Schapper

Kamang also has a suffix marking associative plurality, -lee ‘ASSOC’. This suffix can occurs on kin terms or proper names, as in (50) and (51) respectively. Nouns marked by –lee cannot be modified by any other NP elements. Kamang associative plurality marking . . . , gerumlee see silanta malii. ge-dum-lee sue silanta malii 3.GEN-child-ASSOC arrive mourn mourn ‘. . . , her children and their associates come to mourn.’

(50)

(51)

Martenlee nat tak. Marten-lee n-at tak Marten-ASSOC 1SG-from run ‘Marten and his associates run away from me.’

4.3 Demonstratives and articles Table 9 provides an overview of Kamang’s demonstratives and articles. Kamang has a three-way demonstrative system, with the “that” demonstratives being distinguished from one another in terms of (non-)knowledge (‘KNWN’/‘NKNWN’) of the referent.8 The articles mark specificity (‘SPEC’) and definiteness (‘DEF’). Table 9: Kamang demonstratives and articles

Demonstratives

Articles

Form

Gloss

apa† pang ang =a =ak

‘this’ ‘that.KNWN’ ‘that.NKNWN’ ‘SPEC’ ‘DEF’

† This demonstrative can surface as pa adnominally. The full form (apa) can appear both adnominally and pronominally.

8 It is beyond the scope of this sketch to discuss demonstrative semantics in any detail. The “knowledge” difference between pang and ang can be summarized as follows: pang can refer to items accessible/proximal to the addressee, or to items which are either visible or wellknown to both speaker and addressee; ang can only refer to items outside the immediate sphere of the speech participants, that is, to items which are either not visible or whose identity is not well-known to the speaker and addressee.

Kamang

311

A demonstrative may occur in the NP both modifying a head noun (52a) or independently with no other NP constituents present (52b). (52)

a.

b.

Gera koo sukuu ang miilai. geda koo [sukuu ang]NP THAT. NKNWN 3.CONTR stay hole ‘He stared into that hole.’ Gera koo ang miilai. geda koo [ang]NP 3.CONTR stay THAT.NKNWN ‘He stared into that.’

mi-ilai IN -look

mi-ilai IN -look

As per the NP template, demonstratives and articles have distinct NP slots and can co-occur, as for instance in (53) and (54). (53)

(54)

Male saak anga koo yeeisi. [male saak ang=a]NP THAT.NKNWN = SPEC female old ‘That’s a woman laughing.’

koo stay

yeei-si laugh-IPFV

Si kadii pak woehsi. [si kadii apa=ak]NP go-eh-si 3.LOC-measure-IPFV 1PL.INCL house THIS=DEF ‘We will measure this here house.’

4.4 Possessors Kamang has a three-way possessive split. The three classes of nouns are: alienably possessed nouns, obligatorily possessed nouns, and inalienably possessed nouns. The three possessive classes of nominals are distinguished by (i) the inflection used to mark the possessor on the possessed noun, and (ii) the obligatoriness of a possessor. Table 10 presents an overview of the Kamang possessive classes.

312

Antoinette Schapper

Table 10: Kamang possessive classes Alienable possession

Obligatory possession

Inalienable possession

Marking of possessor

GENITIVE

GENITIVE/ COMPOUND

PATIENTIVE

Occurrence of possessor

optional

obligatory

obligatory

Alienably possessed nouns constitute an open class of nouns. The majority of nominals in the language belong to this class and new nouns are readily integrated into it. In this class, nouns may occur without the expression of a possessor, as in kui ‘dog’ (55a). Where a possessor is expressed, it is encoded with the genitive prefix (55b). (55)

a.

b.

Kui pang tini tuurama. kui pang tini dog that.KNWN dirty ‘That dog is all dirty.’

tuu-da-ma whole-AUX-PFV

Nekui pang tini tuurama. ne-kui pang tini 1SG.GEN-dog that.KNWN dirty ‘My dog is all dirty.’

tuu-da-ma whole-AUX-PFV

Inalienably possessed nouns are nouns which must occur with a possessor marked with a patentive prefix, as seen with -pai ‘belly’ in (56a). Such noun roots cannot exist independent of an inflection (56b). Inalienably possessed nouns constitute a closed class of nouns. Semantically, it is overwhelmingly comprised of human body part nouns, but there are also a few animal body part nouns, bodily and emotional condition nouns, kinship nouns and locative nouns. (56)

a.

Napai aru baai. na-pai adu baai 1SG.PAT-belly big great ‘My stomach is very big.’

b. *Pai belly

aru big

baai. great

Obligatorily possessed nouns (such as waa ‘egg’ in 57) are a small class of around a dozen nouns whose morpho-syntactic properties are intermediate

Kamang

313

between those of inalienably and alienably possessed nouns. Like inalienably possessed nouns, an obligatorily possessed noun must occur with a possessor. Depending on the lexical identity of the obligatorily possessed noun, there may be two ways of expressing the possessor. Like alienably possessed nouns, the possessor can be expressed using a prefix of the genitive inflection (57a). Alternatively, the noun can occur in a possessive compound with another noun expressing the possessor (57b). Simply omitting a possessor is not lexically permissible (57c). (57)

a.

ne-waa 1SG.GEN-egg ‘my egg’ (said by an anthropo-morphosized chicken)

b.

sibe waa chicken egg ‘chicken egg’

c. *waa egg

4.5 NP-appositional slot The appositive slot is not in the NP, but is, as the name suggests, syntactically apposite to the NP. Three kinds of item can occur in the NP-appositional slot: (i) a pronoun co-referent with the NP head (58a), (ii) a plural word quantifying the referent of the NP (58b), or (iii) a numeral phrase, also quantifying the referent of the NP (58c). The NP-external position of the slot is seen by the fact that items in the slot occur to the right of an article marking the periphery of the NP.

(58)

Items in the NP-appositional slot a. Pronoun almakang=ak gera 3.CONTR people=DEF ‘the {specific group of} people’ b.

Plural word almakang=ak nung PL people=DEF ‘the {multiple} people’

314

Antoinette Schapper

c.

Numeral phrase almakang=ak uh biat CLF four people=DEF ‘the people, four of them’

The apposition between an NP and an item in the NP-appositional slot is tight: there is no intonational break or pause between NP and the appositional item; no free items may intervene between them. Despite this tightness, NP enclitics, such as the restrictive focus enclitic =ou ‘RESTR’ in (59), still attach to the NP and not the item in the appositional slot, thereby illustrating the NPexternal status of the appositional slot. (59)

Peu yeeng see akmi lokma. ak=mi lok-ma [pe]NP=ou geengAPPOS sue come here=IN dig.up-PFV pig=RESTR 3.FOC ‘It was pigs that came here and dug up the earth.’

5 Pronouns Kamang has multiple paradigms of independent pronouns (Table 11). Pronouns show the same feature distinctions as agreement prefixes. That is, they distinguish 1st, 2nd, 3rd and COMMON persons.9 There is an additional inclusive and exclusive distinction in the 1st person plural. All but the quantifying (dual, universal and group) pronouns distinguish singular and plural numbers in all persons except for 3rd person. The majority of pronominal paradigms are not marked for grammatical relation. That is, the primary function of the different paradigms is not to mark grammatical roles like subject or object. The details of the use of the different pronominal paradigms will be discussed in the following sections.

9 The “common” person in Kamang is used variously for generic person ‘one(’s)’ and reciprocal ‘each other(’s)’. The related inflections in other TAP languages have similar labels, e.g., ‘distributive’ in Adang, or ‘all inclusive’ in Blagar.

315

Kamang

Table 11: Kamang pronominal paradigms BASIC PRONOUNS

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL CMN

REFLEXIVE

POSSESSIVE

AGENT

UNRESTRICTED

PRONOUNS

PRONOUNS

na a ga ni si i ta

nal al gal nil sil il tal

needu eedu geedu niidu siidu iidu teedu

neen een geen niin siin iin teen

QUANTIFYING PRONOUNS

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL CMN

ALONE

DUAL

ALL

GROUP

nallou allou gallou nillou sillou illou tallou

– – gannok ninnok sinnok innok tannok

– – gaima niima siima iima –

– – geifu niifu siifu iifu –

RESTRICTIVE

CONTRASTIVE

OTHER

neeng eeng geeng niing siing iing teeng

nera era gera nira sira ira tira

net et get nit sit it tet

FOCUS PRONOUNS

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL CMN

5.1 Basic pronouns The “basic” Kamang pronouns are so called because they do not carry any especially marked semantic content or pragmatic meaning as the other pronouns do. They are the most frequently used pronouns in Kamang discourse and can thus also be viewed as statistically “basic”. I distinguish two basic pronoun sets: (i) the “agent” pronouns, and (ii) the “unrestricted” pronouns. An agent pronoun (glossed with ‘AGT’) encodes a semantic agent, that is, a participant which is volitional and controlling; they occur in the syntactic role of either S (60a) or A (61a). In accordance with this restriction, agent pronouns cannot be used to encode either a non-agentive S (60b) or a P (61b).

316 (60)

(61)

Antoinette Schapper

a.

Agentive S Na maa. na maa 1SG.AGT walk ‘I walk.’

b.

Non-Agentive S *Na kawaila. na kawaila 1SG.AGT stumble ‘I stumble.’

a.

Agentive A Na falak sine. na falak sine 1SG.AGT cloth weave ‘I weave cloth.’

b.

P *Markus na bo’na. Markus na bo’na Markus 1SG.AGT hit Intended: ‘Markus hits me.’

‘Unrestricted’ pronouns are so called because they are not limited to encoding participants of any particular syntactic or thematic role. Unrestricted pronouns can occur in all contexts where the agent pronouns cannot. Examples of an unrestricted pronoun encoding the non-agentive S and the P of a main clause verb are given in (62a) and (62b) respectively. Pronouns of this paradigm are glossed simply with their person-number-clusivity feature status. (62)

a.

Non-agentive S Nal kawaila. nal kawaila 1SG stumble ‘I stumble.’

b.

P Markus nal bo’na. Markus nal bo’na Markus 1SG hit ‘Markus hits me.’

Kamang

317

Unrestricted pronouns can also encode agentive S and A, as in (63) and (64) respectively. (63)

(64)

Gal wehe me sukuu wooi gesilang. gal wehe me sukuu wo-oi 3 go.down come hole 3.LOC-towards ‘She went down into a hole.’ Genoklee gal Aribei saak pa ge’baasi. . . ge-nok-lee gal Aribei saak=apa Aribei old=this 3.GEN-friend-ASSOC 3 ‘Her friends said to old Aribei. . .’

ge-silang 3.GEN-descend

ge’-baa-si 3.SBEN-say-IPFV

An unrestricted pronoun can also encode peripheral roles such as those introduced by medial serial verbs or in postpositional phrases. In (65) unrestricted pronouns are used as the complement of the postposition mi. In (66) an unrestricted pronoun is used to encode the P of the medial verb me ‘TAKE’. Also unlike agent pronouns, unrestricted pronouns can encode the subjects of non-verbal clauses, as with the postpositional predicate in (67). (65)

Nok al misilang, nok nal misilang. nok al mi-silang nok nal mi-silang one 2SG IN-descend one 1SG IN-descend ‘One will follow you, one will follow me.’

(66)

Dumma gal me mu’tansi. dum=a gal me mu’tan-si TAKE fall-IPFV child=SPEC 3 ‘A child pushes him over.’

(67)

Nil karii mi. nil kadii mi. 1PL.EXCL house IN ‘We are in the house.’

5.2 Reflexive pronouns Reflexive pronouns reduce the valency of the verb and denote that the action is reflexive, that is, that the A is acting upon itself. In (68a) geedu ‘3.REFL’ denotes

318

Antoinette Schapper

that the single participant Leon is acting upon himself. A reflexive pronoun can also be used as an emphatic pronoun reinforcing the agency of A. For instance, in (68b) there are two participants and geedu emphasizes the agency of Leon in carrying out the event. The reflexive pronoun cannot be used to emphasize a P (68c). (68)

a.

Leon geeru bo’na. Leon geedu bo’na Leon 3.REFL hit ‘Leon hit himself.’

b.

Leon geeru nal bo’na. Leon geedu nal bo’na Leon 3.REFL 1SG hit ‘Leon hit me himself.’

c.

Nal Leon geeru bo’na. nal Leon geedu bo’na 1SG Leon 3.REFL hit ‘Me, Leon hit.’, *‘I hit Leon.’

5.3 Possessive pronouns Kamang possessive pronouns are true pronouns in the manner of English mine, yours etc (cf. possessor prefixes in section 4.4). The identity of the referent of the possessive pronoun must be retrieved anaphorically or from the discourse context, as for instance in (69) and (70). (69)

Anna yeen kou ai yen, niinbo balenna katee. an=a geen kou ai ge-n THUS=SPEC 3.POSS raw take 3.GEN-give balen=a cook=SPEC

niin=bo 1PL.EXCL.POSS=REL

katee eat

‘So theirs (which was) raw was given to them, ours that was cooked was eaten.’ (from Stokhof 1978) (70)

Topi pang een, apa neen. topi pang een apa neen. hat that.KNWN 2SG.POSS this 1SG.POSS ‘That hat is yours, this one is mine.’

Kamang

319

5.4 Quantifying pronouns Quantifying pronouns are special pronominal paradigms which express that their referents have a certain quantity. Those paradigms which highlight their referents’ quantity as being more than one lack singular number forms (i.e., 1st and 2nd person singular, following a similar inflection pattern to that of the human numeral classifier, section 4.2.1). Kamang has A pronominal paradigm denoting that the referents number precisely two. These “dual” pronouns are co-indexed by plural agreement prefixes, as in (71). (71)

Maau innok imauusi? maau innok i-mauu-si 2PL.GEN-war-IPFV who 2DU ‘Who are you two warring with?’

Kamang has a set of pronouns denoting ‘all X’. An example of an ‘all’ pronoun is given in (72). (72)

Siima mooia kate! siima mooi=a katee 1PL.INCL.all banana=SPEC eat ‘Let us all eat the bananas!’

An ‘alone’ pronoun expresses that its referent is without companion in the situation. The pronouns follow the pattern of unrestricted pronouns supplemented with /lou/. An example of an ‘alone’ pronoun is given in (73). (73)

Kui nallou noka’dima. kui nallou no-ka’di-ma dog 1SG.ALONE 1SG.LOC-bark-PFV ‘The dog barked at me alone.’

A group pronoun expresses that the referents act together as a group in the situation, as in (74). (74)

Iifu taa? iifu taa 2PL.GRP sleep ‘Are you sleeping together?’

320

Antoinette Schapper

5.5 Focus pronouns Focus pronouns encode referents whose identity is evaluated against a set of possible alternative participants or referents. A restrictive focus pronoun identifies its referent as precisely the one for whom the proposition expressed by the predicate holds. A restrictive focus pronoun encodes an argument with either an S or an A role, as in (75) and (76) respectively. (75)

Apa yeeng wet masela. apa geeng wet masela this 3.RESTR place beautiful ‘This is a beautiful place.’ (from Stokhof 1978)

(76)

Neeng Markus paanma. neeng Markus paan-ma 1SG.RESTR Markus kill-PFV ‘It was me who killed Markus.’

A contrastive focus pronoun refers to an entity in the discourse and contrasts it with another entity. For instance, in (77) geda ‘3.CONTR’ introduces the two (groups of) participants in the sentence and directly contrasts the actions of the one group going home with that of the other group who continued on up the mountain. (77)

Nung gera yeiyaa, nung gera wete. nung geda ge-iyaa, nung geda wete PL 3.CONTR 3.GEN-return PL 3.CONTR go.up ‘Some of them went home, some of them climbed up.’ (from Stokhof 1978)

An additive focus pronoun denotes that the referent is in addition to another participant. In other words, these pronouns introduce a further or additional participant into a situation or event. For instance, in (78) get ‘3.ADD’ is used to introduce a second participant into the second of two juxtaposed clauses. This ‘other’ participant wakes up the first participant already established as sleeping in the first clause. Additive focus pronouns are also frequently marked with the enclitic =bee ‘also’, as in (79). (78) Alma nok yetaa dii, kul nok yet sue gatan. alma nok ge-taa dii, kul nok get sue ga-tan human one 3.GEN-sleep lie again one 3.ADD come 3.PAT-wake ‘One person is lying sleeping, then another comes along and wakes him up.’

Kamang

321

(79) Nitbee isei maa kateesi. nit=bee isei maa katee-si 1PL.EXCL.ADD=also game cooked eat-IPFV ‘We also are eating meat.’

6 Agreement and locational prefixes There are three prefixal slots on a Kamang verb. The template is given in (80). The first slot, closest to the verb root, is the AGREEMENT SLOT (represented as “AGR” in the template) which takes an agreement prefix typically indexing either S or P, but not A. The second slot is the INCORPORATED SLOT (INCORP) hosting an incorporated postposition, either mi- or wo-. The final slot is the SELF-BENEFACTIVE SLOT (SBEN) where a self-benefactive inflectional prefix occurs indexing an agentive S or A.

(80)

Maximum verb prefix template SBENS/A-INCORP-AGRS/P-V

Note that there is one co-occurrence restriction between the INCORP slot and the AGR slot of the prefixal template: The prefix wo- never cooccurs in the INCORP slot with an agreement prefix of the LOCATIVE inflection in the AGR slot. This may be thought to indicate that the LOCATIVE inflection in fact occurs in the INCORP slot of the template. However, this analysis is rejected here, since it is not possible for a single root to have both a prefix of the LOCATIVE inflection alongside that of another inflectional agreement, as in (81). More likely, is that there is a historical relationship between the INCORP prefix wo- and the LOCATIVE inflection in the AGR slot (which is also realized as wo- in the 3rd person). Examples of the template are given below. In (81) the verb root -bei occurs with the SELF-BENEFACTIVE prefix ni’-, the INCORPORATED prefix wo- and the AGREEMENT prefix ta- (marking reciprocity). In (82) the verb root -apoo hosts a SELFBENEFACTIVE prefix ne’- and a INCORPORATED prefix mi-, while in (83) the verb root -pang hosts the INCORPORATED prefix mi- and a AGREEMENT prefix noo-. (81)

Ninnok ni’wotabei. ninnok ni’-wo-ta-bei 1DU.EXCL 1PL.EXCL-AT-CMN.PAT-promise ‘The two of us promise something to each other.’

322

Antoinette Schapper

(82)

Nal isei maa nal isei 1SG game ‘I stole meat

(83)

Ganeiak minoopanma. ga-nei=ak mi-noo-pang-ma 3.PAT-name=DEF IN-1SG.AST-forget-PFV ‘I forgot his name.’

ne’miapoo. maa ne’-mi-apoo edible 1SG.SBEN-IN-break.off away for myself.’

Whilst these examples show that it is possible for the prefixal slots to be filled simultaneously, it is actually relatively rare in Kamang for more than one slot to be occupied at a time. As such, in the following sections, I will discuss the slots of the Kamang prefixal template in isolation from one another.

6.1 The agreement (AGR) slot Table 12 gives an overview of the inflections of agreement prefixes appearing in the AGREEMENT SLOT of the template in (80). The PATIENTIVE and the GENITIVE inflections have already been discussed in their role marking possessors on different classes of nouns (Section 4.4). Table 12: Agreement prefixes of the AGREEMENT SLOT

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL CMN

PATIENTIVE (PAT )

GENITIVE (GEN)

LOCATIVE (LOC)

DATIVE (DAT )

DIRECTIVE (DIR)

ASSISTIVE (AST )

naaganisiita-

neegenisiite-

noowoniosioioto-

neeeegeeniisiiiitee-

naoaogaonioosiooiootao-

noooowoonioosiooiootoo-

Each of these prefixal inflections can index either S or P. That is, Kamang has a so-called ‘split-S alignment’ where the single argument of a monovalent verb (S) is sometimes coded like the agentive argument of a bivalent verb (A) and other times like the patientive argument of a bivalent verb (P). However, Kamang doesn’t just have one ‘split’ in the prefixal coding of S and P but multiple splits, since there are multiple inflections which can index S and P depending on the verb root. Yet not all AGR slot inflections can index these arguments.

Kamang

323

In addition to S, A and P as defined above, it is necessary to distinguish between P and applied P (APPL.P). While a ‘true’ P is the second subcategorized argument of an underived transitive verb, an applied P is a derived second argument introduced by an inflection onto a basically intransitive verb. In Table 13, we see that the PATIENTIVE inflection co-indexes S and P but not APPL.P. LOCATIVE and GENITIVE inflections index all three. The LOCATIVE and GENITIVE inflections each have only one verb root that indexes P, while the ASSISTIVE inflection has two members that index S and one that indexes a P. As such, DATIVE, DIRECTIVE and ASSISTIVE inflections are effectively limited to marking APPL.P. Table 13: Range of participants encoded by inflections in the AGR slot

S P APPL.P

PAT

GEN

LOC

DAT

DIR

AST

✓ ✓

✓ ✓ ✓

✓ ✓ ✓

(✓) ✓

(✓) ✓

(✓) (✓) ✓

The primary morphological divide between Kamang verbs is on the basis of lexical class. Two main classes can be differentiated: (i) Verbs with obligatory prefixation: an AGR prefix is required on the verb root in order to be well-formed, and there is no apparent semantic motivation for the choice of which inflection goes with which verb root (section 6.1.1). Roughly 35% of verb roots in the Kamang corpus are of this type. (ii) Verbs without obligatory prefixation: verb roots that do not require an AGR prefix to co-index one of their arguments, and where prefixes are added to these verb roots they have semantic transparent functions (section 6.1.2). Roughly 65% of verb roots in the Kamang corpus are of this type. The appearances of the individual prefixes in the different classes are discussed in the following sections.

6.1.1 Obligatorily prefixed verbs In Kamang, there is no flexibility in the choice of prefix on verb roots requiring a prefix to index either S or P. That is, verb roots with obligatorily prefixation are distributed into lexicalized inflectional classes. For instance, consider the examples in (84) to (86). In the (a) examples we see a monovalent verb encoding S with a prefix, while in the (b) examples we see a bivalent verb in which P is encoded with a prefix. In each of the example

324

Antoinette Schapper

sets, the PAT, LOC or GEN inflection cannot be left off the respective verb or replaced by a prefix of another inflection. There is no semantically transparent reason why one prefixal inflection is used with one root and another inflection with another root. The relation between prefix and root is simply lexically fixed. (84)

a.

S-marking PAT inflection Namaitansi. na-maitan-si 1SG.PAT-hunger-IPFV ‘I’m hungry.’

b.

P-marking PAT inflection Leon nataksi. Leon na-tak-si Leon 1SG.PAT-see-IPFV ‘Leon sees me.’

(85)

a.

S-marking GEN inflection Nelaitasi. ne-laita-si 1SG.GEN-shy-IPFV ‘I’m shy.’

b.

P-marking GEN inflection Leon nefaneesi. Leon ne-fanee-si Leon 1SG.GEN-shoot-IPFV ‘Leon shoots at me.’

(86)

a.

S-marking LOC inflection Nobieesi. no-biee-si 1SG.LOC-angry-IPFV ‘I’m angry.’

b.

P-marking LOC inflection Leon nowaasi. Leon no-waa-si Leon 1SG.LOC-cover-IPFV ‘Leon conceals me.’

The vast majority of root lexicalized prefixes are from the PAT (~50%), LOC (~35%) and GEN (~10%) inflections, as in the above examples. As stated already, there are only a handful of verbs ( ayaa irea ‘rain water’

Sawila

365

2.6.3 Reduplication Sawila nominal and verbal roots and stems may undergo reduplication. Partial reduplication of a root is not attested, but it is common in verbal stems, where often only the root undergoes reduplication. Table 6 lists the attested patterns and their meaning. Reduplication of nominal roots derives collective nouns; reduplication of verbal roots indicates extended duration or iterativity. Table 6: Sawila reduplication patterns

a.

b.

TYPE

EXAMPLE

RED [ROOT ]

manaa~manaa ta-na-noosa~noosa wii-tapo~tapo ta-li-diimu~diimu ga-mayii-ga~mayii

RED [STEM]

GLOSS

MEANING

village~COL

‘countryside’ ‘each individually’ ‘more and more’ ‘push each other’ ‘keep sounding’

UND- INV- ALONE~ COL APPL-add~ITER UND- APPL-press~ITER

3.I-sound~ITER

2.6.4 Fusion of verbal prefixes The last morphophonemic process to be discussed here is verbal prefix fusion. In rapid speech certain types of verbal prefixes may be fused, as in (25). Vowel harmony is usually part of the fusion process: the underlying forms of the fused prefix lee- are li- + ga-. (25) Masing leekala yaana niko. masing li-ga-kala yaana ni-ko [hunger APPL-3.I-hold]manner FIN.go 1PL.EXCL-stay ‘Ignoring the hunger we kept going’ [NB4.073] All instances of verbal prefix fusion involve an applicative prefix (discussed in section 6.2) followed by a person-marking prefix (6.1). The trigger of the vowel harmony (in bold) may be the applicative prefix or the verbal root, as can be seen in (26). (26) /li- ga- miˈti/ /ˌwiː-ga- ˈpi/ /ˌwiː-garaˈni/ /li- ga- ˈtaːni/ /ˈma- ga- ˈdiːla/

→ → → → →

[ˌleːmɪˈtɪ] [ˈweːpɪ] [ˌweːrɑˈnɪ] [lɛˈtaːnɪ] [mɔˈdiːlɑ]

li-ga-miti wii-ga-pi wii-ga-rani li-ga-taani ma-ga-diila

> > > > >

lee-miti wee-pi wee-rani le-taani mo-diila

‘he sits on it’ ‘make it with st.’ ‘give it to him’ ‘shot it’ ‘chase after it’

366

František Kratochvíl

2.7 Orthography The practical orthography used for Sawila in this sketch was designed by the Summer Institute for Linguistics team members Anderias Malaikosa, Isak Bantara, and Apris Malaikosa. The main concern was to stay as close as possible to the Standard Indonesian orthography and to avoid diacritics. The orthography has been used in the Sawila dictionary, corpus and in the ongoing Bible translation work. The system draws from the orthography of Standard Malay/Indonesian in using the graphemes y for [j], ng for [ŋ], and uy/ij for [y]. The vowel length is captured by doubling the grapheme, such as aa for [aː], as in naanu ‘not’, which is pronounced as [ˈnaːnʊ]. The primary challenge is how to capture the difference between pairs such as /ˈna.nu/ (1SG) vs. /ˈnaː.nu/ ‘not’, where the stressed syllable is non-final. Since the doubling of the vowel is used to mark a long vowel, non-final stressed short syllables have to be marked differently. Currently we distinguish pairs such as /ˈna.nu/ (1SG) vs. /ˈnaː.nu/ ‘not’ as nannu (1SG) vs. naanu ‘not’: we double the onset consonant immediately following the stressed short non-final syllable. Vowel harmony and dialectal differences are always captured in the orthography. So, for example, the combination of the words mi ‘take’ and naanu ‘not’ is written as mi neanu to mean ‘did not take’. Clitics are written together with their host words.

3 Basic clausal syntax This section discusses the structure of verbal (3.1), locative, existential, possessive predicates (3.2), non-verbal predicates (3.3), and structure of questions (3.4). Basic Sawila clauses consist of a predicate (PRED) preceded by its arguments (ARG), and followed by aspectual (ASP) and mood (MOOD) markers, as shown in (27). Sawila clause structure (27) ARG1 ARG2 PRED ASPECT MOOD

3.1 Verbal predicates Sawila verbal predicates can be divided according to their valence properties into two types, shown in Table 7. Monovalent predicates combine with a single

Sawila

367

argument (S), sometimes co-indexed on the verb (AGR). Bivalent predicates combine with two arguments (A stands for the agent-like argument, P for the patient-like argument). Trivalent predicates admit A, P, and E arguments and index the P argument on the verb. Arguments whose referent is recoverable from the context may be omitted. Monovalent and bivalent predicates can be extended with applicative prefixes and admit an additional E argument (see sections 6.2 and 6.4). Table 7: Overview of Sawila verbal predicate types VALENCE

REALIZATION

MONOVALENT

S S A A A A

BIVALENT

TRIVALENT

EXAMPLE

P P P P

V AGRS-V V AGRA-V AGRP-V E AGRP-V

(28) (29) (30) (31a) (31b) (33)

3.1.1 Monovalent predicates The single argument of a monovalent verb (S) always precedes the verb and may be realized as a noun phrase, free pronoun, or can be dropped if it is recoverable from the context. Third person pronouns may occur apposed to NPs, as can be seen in (28b). (28) a. Ayaakani iduni, ka ayaa me girra. girra [ayaakani]S iduni ka [ayaa]S me come certainly sky dark IRR rain ‘The sky is clouded, it will certainly rain.’ [D.Iduni.1] b. Kaluno se saaku girra kiikara kaana. kaluno [se saaku girra]S kiikara kaana FIN.now chicken old 3.AGT FIN .crow FIN .COMPL ‘The rooster has just crowed.’ [NB4.167] Person prefixes cannot be dropped. Throughout this section the arguments are indexed and where an argument has been omitted, this is indicated with a Ø. Person prefixes carry a subscript label indicating which argument is indexed.

368

František Kratochvíl

(29) a. Nirra ninee nipeati. niS-nee niS-paati [(niira)]S 1PL.EXCL.AGT 1PL.EXCL-consume 1PL.EXCL-eat ‘We (exclusive) eat and drink.’ [NB5.231] b. Nannu namiirine. c. Nikatu garammu. [ni-katu]S ga S-rammu [(nannu)]S naS-miirine 1SG.I-FIN.fear 1SG.II-dog 3.I-die 1SG ‘I am scared.’ [N12.45] ‘My dog died.’ [D.Ni1] 3.1.2 Bivalent predicates Bivalent predicates combine with two arguments, labeled here as A and P. Most commonly, the A argument precedes the P, as in (30a). As can be seen in (30b), third person pronoun girra can be apposed to the NP expressing the A argument. (30) a. Niya di patara ma maana. di [patara ma]P maana [ni-ya]A edible FIN.cook 1SG.II-mother still corn ‘My mother is (still) cooking rice.’ [N12.64] b. Ekatu girra namaddu paati. [e-katu girra]A [na-maddu]P paati 1SG.I-child bite 2SG.II-dog 3.AGT ‘Your dog bit my child.’ [CC.102] Sawila verbs may combine with person prefixes indexing either the A or P argument. As shown in (31), the verbs nee ‘consume’ and punni ‘hit’ may combine with the person prefix na- (1SG) to index the A and P argument respectively. Whether a verb takes person prefixes is determined lexically and will be discussed in section 6.4, however, there seems to be a general preference for indexing human participants. (31) a. Nannu tukku nanee niwuli. naA-nee niA-wuli [nannu]A [tukku]P palm.wine 1SG.I-consume 1SG.II-desire 1SG ‘I want to drink palm wine.’ [NB5.245] b. Girra napunni. [girra]A naP-punni 1SG.I-hit 3.AGT ‘He hit me.’ [D.Punni.1]

Sawila

369

Besides the unmarked AOV word order, illustrated in the examples above, is available when the A argument is backgrounded. The backgrounded A argument immediately precedes the verb. The P argument can be marked with the particle to, as shown in (32b). Focussed A arguments are marked by pronominal doubling (see section 5.4). OAV

(32) a. Niselu aning tii lee. tii lee [ni-salu]P [aning]A 1SG.II-yam NFIN.person dig PFV ‘My yams are all harvested.’ [N12.66] b. Seto girra baali, pito girra baali, weto girra baali. [girra]A baali pi=to girra baali we=to girra baali [se=to]P feed pig=too 3.AGT feed goat=too 3.AGT feed chicken=too 3.AGT ‘He was breeding chickens, pigs, and even goats.’ [TF.10]

3.1.3 Trivalent and augmented predicates Sawila verbs yaayi ‘ask for, request’ and -suwaara ‘show’ may combine with three arguments, termed here as A, P, and E. Sawila shows differential treatment of recipients (see Noonan 1992:121; Kittilä 2008). The third person recipient is treated as a P, while first and second person are treated as E and require an applicative prefix (see section 6.2). The third person P may require the inverse prefix na- (for details see 6.3) with yaayi ‘ask, request’ but not with -suwaara ‘show’. (33) a. Namaddu patama yaayi. [na-maddu]A [patama]P yaayi food ask 1SG.I-child ‘My child is asking for food.’ [D.Yaayi] b. Namaddu girra giya seena ganeyaayi. [seena]E ga P-na-yaayi [na-maddu girra]A [gi-ya]P 3.II-mother FIN.money 3.I-INV-ask 1SG.I-child 3.AGT ‘My child is asking her mother for money.’ [D.Yaayi] c. Namaddu girra seena aliyaayi tirea! aE-li-yaayi tirea [na-maddu girra]A [seena]P FIN .money 2SG. I-APPL-ask PROH 1SG.I-child 3.AGT ‘My child may not ask money from you.’ [D.Yaayi]

370

František Kratochvíl

Applicative prefixes can be used to extend the valence of many Sawila verbs (see sections 6.2 and 6.4 for detailed discussion).

3.2 Locative, existential, and possessive predicates Locations are encoded with locative verbs, spatial deictics, and positional words. Locative verbs take arguments and function as predicates but spatial deictics and positional words have to first be combined with a locative verb to become the predicate. Locative verbs are ma ‘be.PROX (adjacent)’, li ‘be.DIST (in open space)’, and mara ‘be.in (inside, exist)’. As can be seen in (34), ma is used when the location is in the proximity of the speaker, either physically or metaphorically. Ma is used with delimited locations such as houses, holes, forests, windows, doors, etc. When the locations are not in the speaker’s proximity or lack clear spatial limits, li ‘be.DIST’ is used. The verb mara can be used to indicate goal locations (35a), but sometimes the motion verb is repeated (35b). Mara is also used as an existential verb, as in (36). (34) a. Se mu annama gamiti wuno. se mu anna=ma ga-miti wuno FIN .DUR chicken tree above=be.PROX 3.I-sit ‘The chicken is sitting here on the tree.’ [LOC.5] b. Se mu annali gamiti. se mu anna=li ga-miti chicken tree above=be.DIST 3.I-sit ‘The chicken is sitting on the tree (over there).’ [LOC.3] (35) a. Nasuwi nearaasine mara kaana. na-suwi ne-’araasine mara kaana 1SG.I-arrive 1SG.II-FIN.house be.in FIN.COMPL ‘I have come back home.’ [NB5.122] b. Nayaa Kalambaasa nayaa. na-yaa Kalambaasa na-yaa 1SG.I-go.down FIN.place.name 1SG.I-go.down ‘I go down to Kalahabi.’ [D.Yaa.1] Spatial deictics and positional words are listed in Table 8. With the exception of oli and ali, spatial deictics and positional words combine with li

Sawila

371

‘be.DIST’ to denote locations; li is not used when directions are denoted, instead a directional or motion verb is used. Table 8: Sawila spatial deictics and other positional words DEICTIC

CLOSE

REMOTE

HORIZONTAL

oli ‘here’ ali ‘there’

mana ‘over there [visible]’ tana ‘over there [invisible]’ anna ‘above there’ yana ‘below there’

tawaaka ‘outside’

anna ‘top’ apa ‘below’ iduko ‘belly downwards’

VERTICAL

POSITIONAL WORDS

mirea ‘inside’

Existential predicates, such as mara ‘exist’ (36), awatu ‘be absent’ (37), or naanu ‘not exist’ (37), do not combine with person prefixes. (36) Gannu giriang mara. [gannu gi-’iria=ang]S mara 3 3II-water=DIST exist ‘That water [of the banana sp. storing water in its trunk] will be there.’ [SC.055] (37) Manaa dana angsi anina awatu. awatu manaa dana ang=si [anina]S village FIN.one DIST=TOP FIN.person absent ‘As for that village, there are no people.’ [D.Si.133] (38) Di araasing dana waka naanu. di [[araasing dana]S waka]RC=S naanu still NFIN.house FIN.one put.in.fire FIN.not.exist ‘There is no house that burned down.’ or ‘A single house did not burn down yet.’ [QNP.2] There is a dedicated free possessive pronoun, functioning often as a nominal predicate to express possession, as in (39) and (40). The genitive paradigm is homophonous and likely historically related to the agentive paradigm discussed in section 5.1.

372

František Kratochvíl

(39) Araasine ang nirra. araasine ang nirra FIN.house DIST 1SG .GEN ‘That house belongs to me.’ or ‘I own that house.’ [D.Ni.2] (40) Namaddu yaalara ama angana nirra we’arani, erra. na-maddu yaalara ama a=ngana nirra wii-a-rani erra 1SG.I-child FIN.female DIST.VIS DIST.DEF 1SG.AGT APPL-2SG.I-FIN.give 2SG.GEN ‘That daugher of mine, I will give you to have [as your wife].’ [KD.28] Sawila genitive pronouns are not used frequently, and tend to occur in the clause final position, as shown above. However, a few instances where a genitive pronoun expresses one of the arguments are attested, as in (41). (41) Kaluno waari niira wiitiila. kaluno waari niira wii-tiila FIN .now again 1PL. EXCL.GEN APPL-speak ‘I will retell it in ours [i.e. our language, Sawila].’ [KD.4]

3.3 Non-verbal predicates Noun phrases can be used a predicates without a copula, as in (42), where the NP ga-takka kaasala is the predicate. In (43), the predicate is a numeral phrase (NumP). (42) Namaddu nanna gatakka kaasala. [na-maddu nanna]S [ga-takka kaasala]predicate 1SG.I-child all 3.I-head hard ‘All my children are very naughty.’ [QNP.13] (43) Gamaddu ngana yaku. [ga-maddu ngana]S [yaku]predicate 3.I-child DEF two ‘They have got two children.’ [KM.4] Nominal predicates are compatible with the completive marker kaana, which indicates that the argument has assumed the state described by the predicate, as in (44a). The negation with naanu follows the pattern of verbal predicates (44b–c). Nominal predicates are compatible with the durative wuno (44d).

Sawila

373

Some nominal predicates, such as yaala murra ‘little girl’, are not compatible with the imperfective muno (44e). (44) a. Nannu guuru kaana. [nannu]S [guuru]predicate kaana teacher FIN .COMPL 1SG ‘I became a teacher.’ [N12.76] b. Nannu guuru naanu. [nannu]S [guuru]predicate naanu teacher FIN .not 1SG ‘I am not a teacher.’ [N12.77] c. Nidana di guuru naanu kaana. di [guuru]predicate naanu kaana [ni-dana]S FIN .not FIN .COMPL 1SG.II-friend still teacher ‘My friend has not become a teacher yet.’ [N12.75] d. Gannu kaana murra wuno. [gannu]S kaana [murra]predicate wuno FIN .DUR 3 may kid ‘She may still be a kid.’ [AB.66] e. *Nannu yaala murra muno. [nannu]S [yaala murra]predicate *muno female kid FIN. IPFV 1SG Intended: ‘I am still a young girl.’ [N12.69] Weather and time predicates are either nominal (morning, evening, night – these take no S argument) or describe the passage of the heavenly bodies either with a motion verb (sunrise) or with a stative verb (sun is hot, etc. – the heavenly body is the S argument). These predicates are compatible with a greater range of aspectual markers, such as the perfective lee, completive kaana, gapi ‘almost’ (45) or the imperfective muno (46). (45) Iduni, iduni lee, iduni lee kaana, iduni gapi. iduni iduni lee iduni lee kaana iduni gapi night night PFV night PFV FIN.COMPL night almost ‘It is night, it is already night, it is night, it’s almost night.’ [N12.35–36]

374

František Kratochvíl

(46) Iduni muno, yaang kaang-kaana! iduni muno yaang kaang-kaana night FIN.IPFV NFIN.travel FIN.well ‘It is still night, so travel carefully.’ [N12.41]

3.4 Negation The negator naanu occurs in the final position of the clause. Naanu can be also used as a negative existential verb ‘not exist, not be’, as discussed in section 3.2. (47) Siribisiang di gapi lee naanu. lee naanu [siribisi=ang]S di gaS-pi still 3.I-make PFV FIN.not work=DIST ‘That work is not ready yet.’ [WVP.132] (48) Gannu api paati neanu, ang noonine, di ganamasine neanu. naanu [gannu]A [api]P paati naanu ang noonine di gaS-na-masine 3 fish eat not because still 3.I-INV-FIN.hungry FIN.not ‘He is not eating fish because he is not hungry yet.’ [QVP.72] The prohibitive particle tirea occurs in imperatives (49) and suggestions (50). (49) Yamiiring tirea! yaS-miiring tirea 2PL.I-NFIN.fear PROH ‘Don’t you be scared!’ [KM.87] (50) Kaluno waari ineku tanamatataaku tirea! kaluno waari iS-n-yaku taE-na-ma-tataaku tirea FIN.now again 1PL. INCL-together-two DISTR .I-INV-APPL-meet PROH ‘Now let the two of us not meet again!’ [MC.33]

3.5 Questions and directives Questions in Sawila are marked with a final rise in intonation and in some cases with the question marker te. Sawila question words share the initial nV- syllable, likely related to the indexical root nV ‘like’, and occur either in-situ or fronted. Naadu ~ noodu ‘what’ is used to ask about things (51), namala ‘who’ about

Sawila

375

persons (52). To inquire about locations, noo is combined with the location verb li (53). When noo occurs pre-nominally, it inquires about the kind of the referent, as in (54). (51) Niya garammu, nannu ka naadu gapi? gaS-rammu [nannu]A ka [naadu]P gaP-pi [ni-ya]S 1SG soon what 3.I-make 1SG.II-mother 3.I-die ‘My mother died, what shall I do?’ [Q.35] (52) Erra tatuku ngana, namala Moona gapunni? [erra]S tatuku ngana [namala]A [Moona]P gaP-punni DEF who Moona 3.I-hit 2SG.AGT tell ‘Who did you say beat up Moona?’ [QVP.83] (53) Nooli araasing ma-ta-miti? nooli [araasing]E ma-taA-miti where NFIN.house APPL-DISTR.I-sit ‘Where is a house for us to stay the night?’ [N12.3] (54) Annu noo dara du dara? [annu]A [noo dara du]P dara which song PL sing 2SG ‘Which songs will you sing?’ [Q.46] Sawila wh-words can be used as indefinite pronouns. In (55), the question word noonine ‘how’ is used to refer to a description provided earlier on. (55) Anta noonine erra tatuku, amanto mara. anta noonine [erra]S tatuku [aman=to]S mara FIN .how 2SG.AGT talk NFIN . DIST.like=as.well be or ‘Or like you say, that’s also a possibility.’ [BDPK.37-038] Directives (imperatives, hortatives, and prohibitives, cf. Mauri and Sansò 2012) are marked by high pitch contour falling on the last syllable and regularly also by the mood particles da, di, dapa and tirea (the first three possibly historically related; for more details see 7.2). (56) a. Yanee yapaati dea! b. Ka nayaadi! ya-nee ya-paati da ka naS-yaa=di soon 1SG.I-go.down=REAL 2PL.I-consume 2PL.I-eat REAL ‘Please eat and drink!’ [D.Nee.1] ‘I better get going!’ [C2.13]

376

František Kratochvíl

c. Esaapang mi me dapa! me dapa [e-saapang]P mi 2SG.II-dipper take come PRIOR ‘Bring your dipper (before anything else)!’ [D1] d. Amayii tirea! tirea aS-mayii 2SG.I-sound PROH ‘Don’t make any noise!’ [FSYB.38]

4 Noun phrases Sawila noun phrases consist maximally of the head noun (N), which may be indexed for the possessor (POSS), and followed by a modifier (MOD), quantifier or numeral (QUANT), and a demonstrative (DEM), as shown in (57). Roots displaying alternation (discussed in 2.6.1) will appear in their full form in the phrase-final position and in their non-final form elsewhere. The topic marker gana precedes the NP, the other topic markers ba and si, and the focus marker po follow the NP. (57) NP → [POSS-N MOD NUM QUANT DEM] In the remainder of this section, I will discuss modification with demonstratives (4.1), modifiers (4.2), relative and nominalized clauses (4.3), quantifiers (4.4), and possessive constructions (4.5).

4.1 Demonstratives Sawila has a complex demonstrative paradigm, formed around a two-way contrast between proximate o and distal a forms, shown in Table 9. Syncretism can be observed in some of the forms derived with roots containing nasals. In terms of reference, there are distinct deictic, definite, spatial, and manner/kind forms, sometimes with visibility contrasts. Various types can be stacked, as will be shown later in this section. Forms in bold face are also attested in clause final position as nominalizers or evidential markers. There is a transparent morphological relationship with corresponding wh-words derived from the root noo ‘what’, listed in the rightmost column.

Sawila

377

Table 9: Sawila demonstratives ROOT

PROXIMAL (o +

na ‘VIS’

ong~om ‘PROX’ ona ‘PROX.VIS’

ngana ‘DEF’ wiina ‘DEF.NVIS’

ROOT )

ongana ‘PROX.DEF’ onine ‘PROX.like’

li ‘be.DIST’

oli ‘here’

DISTAL (a +

ROOT )

ang~am ‘DIST’ ama ‘DIST.VIS’ awa ‘DIST.NVIS’ angana ‘DIST.DEF’ awiina ‘DIST.DEF.NVIS’ amane ‘DIST.like’ amanu ‘DIST.like’ ali ‘there’

WH-WORD (noo +

ROOT )

noong ‘which’ noona ‘which’

nooyiing ‘which.NVIS’ noonine ‘how’ nooli ‘where’

Demonstratives ong (PROX) and ang (DIST) are often used with a pointing gesture. When they are part of an NP, they are found in its final slot, typically cliticizing to the previous word, as in (58). Besides the spatial interpretation, both demonstratives can also be interpreted anaphorically. (58) a. Araasingong nirra. b. Araasingang nirra. nirra [araasing=ang]NP nirra [araasing=ong]NP NFIN .house=PROX 1SG.GEN NFIN .house=DIST 1SG.GEN ‘This house belongs to me.’ [P.212] ‘That house belongs to me.’ [P.213] Their allomorphs om and am occur when the demonstratives are followed by the topic marker ba. The conversation fragment discussing the RM cigarette brand in (59) shows that both ong and ang can be used as third person pronouns: the cigarette brand is introduced by the speaker A, who refers to it in the following turn with the distal form am, undergoing assimilation triggered by the following topic marker ba. Speaker B takes up the topic and refers to the brand only with the proximal ong. (59) A: Eh, annu RM aning yoo? RM amba apaali-tulori lu muno. eh annu RM aning yoo RM am=ba INTER 2SG RM NFIN.person PART RM DIST=TOP apaali-tulori lu muno yesterday-two.days.ago that.time FIN.IPFV ‘So you are an RM guy (someone smoking only the RM brand cigarettes)? That RM is an old brand, from a long time ago.’ B: Ong kaluno mara naanu. ong kaluno mara naanu PROX FIN .now be FIN .not ‘This [brand] does not exist nowadays anymore.’ [BPDK.266–268]

378

František Kratochvíl

The demonstratives ona (PROX.VIS) and ama (DIST.VIS) refer to visible locations. The form ona (3 occurrences in the corpus) is shown in (60). The form ama is quite frequent (136 occurrences). In (60b), the speaker uses ama to point at the recorder in front of him and explains to a friend the red light on the recorder signals recording. The referents marked with ama do not have to be immediately visible, but a previous sighting is required, as can be seen in (60c), where the speakers discuss buildings on the other side of the island. (60) a. Katu pike ona ame dapa! a-me dapa [katu]S pike ona dog just PROX.VIS 2SG.I-come PRIOR ‘Dog, just come over here!’ [AL.29] b. Laampuro ike gate tana ama tamaka ba mi keana. ba]P laampuro ike ga-te tana [ama]A [ta-maka lamp red 3.I-lie same.time DIST.VIS DISTR.I-voice TOP mi kaana take FIN.COMPL ‘While the red light is on, our voices are recorded by that [thing].’ [NB4.47] c. Eh, bupati garaasine ama ali, kaantora ali! ali kaantora ali e [bupati (gi-’araasine]NP) ama INTER regent 3.II-FIN.house DIST.VIS there office there ‘No, there in the regent’s house, that’s where the office was.’ [BDPK.218] The root form na is not attested in noun phrases, but is used as a nominalizer. To distinguish it from other demonstratives, which can also be used in clause final position, the gloss VIS is kept. Subordinate clauses marked with na often indicate reason, as shown in (61). Insubordinate uses of na and other demonstrative roots although attested and documented in related languages (Kratochvíl 2011; Schapper and San Roque 2011), are not discussed in this sketch. (61) a. Araasing mirea iduni nea, laampuro mataani. araasing mirea iduni na laampuro mataani NFIN .house inside dark VIS lamp light ‘It is dark inside the house, so switch on the lamp.’ [D.Iduni.01]

Sawila

379

b. Girra aning kiibi nea, gimaarang laawa. girra aning kiibi na gi-maarang laawa 3.AGT NFIN.person pay VIS 3.II-NFIN.together search ‘She hired people to search together with her.’ [D.Kiibi.526] The distal visible ama has a non-visible counterpart awa (DIST.NVIS), which is used to refer to objects invisible to the addressee. In discourse, awa may mark new topics, as in (62). . . . There were two boys, friends, who together grew up and lived in the village. (62) Gidena awa, ang gannu gigako-giamiti ngeana suraata daadu du magako-magamiti. gi-dana awa ang gannu gi-ga-ko-gi-ga-miti ngana 3.II-3.I-stay-3.II-3.I-sit DEF 3.II-friend DIST.NVIS DIST 3 suraata-daadu du ma-ga-ko-ma-ga-miti PL APPL-3.I -stay-APPL-3.I-sit cards ‘That [other] friend, that one all he did was just play cards.’ [TF8] The use of the definite demonstrative ngana (DEF) is illustrated in (63) with a fragment taken from the beginning of a story, where the main characters are introduced. . . . They had two children, one was a girl, and the other was a boy. (63) Yaalara pike ngana madumma, imyaalara pike ngana ikisiila. [yaalara pike ngana]S madumma [imyaalara pike ngana]S ikisiila FIN .female small DEF senior FIN.male small DEF junior ‘The girl was older, the boy was younger.’ [KM4–6] The definite forms ongana and angana are used with pointing, as in (64), where the speaker is choosing between two objects, one in his own hand and one on the table in front of him. Both forms are used pronominally here. (64) Q: Noong li-a-na-maro? which APPL-2SG.I-INV-want ‘Which one do you want?’ A1: Ongana. PROX.DEF

‘(I want) this one (holding it in the hand).’

380

František Kratochvíl

A2: Angana. DIST.DEF

‘(I want) that one (somewhat further away).’ [N12.138–140] That both demonstratives may also be part of an NP can be seen in (65). (65) a. Araasine ongana nirra gapi nanapaniki boorana. araasine ongana nirra ga-pi na-na-paniki boorana FIN .house PROX.DEF 1SG .AGT 3.I -make 1SG .I-INV-pay.effort FIN .very ‘I have built this house with a lot of effort.’ [D.Paniki.4.17] b. Ada angana taling dapa warana. ada angana taling dapa warana firewood DIST.DEF NFIN.tie.up PRIOR FIN.carry.on.shoulder ‘That firewood has to be tied together before you carry it.’ [D.Taling.502] The definite demonstratives have their non-visible counterparts wiina (DEF.NVIS) and awiina (DIST.DEF.NVIS). In the fragment below, both the root wiina and the distal awiina are used to describe the importance of rain, in a discussion of the weather. While the speaker points to the visible clouds with ama, the rainwater and its benefits are marked with the contrasting non-visible forms. (66) Sakapullo ama angana ayaa giyeari, awiing taang-mading waari, we-pito gikaang wiina amane. sakapullo ama angana ayaa gi-yari awiing DIST.VIS DIST.DEF rain 3.II-root NFIN. DIST. DEF.NVIS cloud [taang-mading waari we-pi=to gi-kaang wiina]NP amane NFIN.plants again animals=as.well 3.II-NFIN.good DEF.NVIS DIST.like ‘Those clouds [there] are the source of rain that is to the benefit of both plants and animals.’ [S6] Yet another set of demonstratives points to properties of referents; such as their size, color, number, etc, similar in meaning to English so or such. The proximal form onine ‘PROX.like’ in (67) is used to point to the size of the fish, not to the fish itself. The fish is either directly visible or has to be clearly introduced in discourse.

Sawila

381

(67) Erra nooning gapi ka api waansa onine ataaku onu? erra nooning ga-pi ka api waansa onine a-taaku onu 3SG.I-make soon fish big PROX.like 2SG .I-obtain PROX.VIS 2SG.AGT how ‘How would you get a fish this big?’ [QVP.77] The distal amane, as shown in (68), refers to the size of a giant taro described earlier in the story. Besides their reference to nominal modifiers, the forms onine, amane, and amanu may also refer to manner. (68) Angmasi, waari gana alaata amane ama, ang di matataaku naanu du. ang=masi waari gana [alaata amane ama]NP ang di DIST =until again 3.TOP taro DIST.like DIST.VIS DIST still ma-tataaku naanu du APPL-find FIN .not ITER ‘But then that taro was so [big] that [such] has not been found before.’ [AL8] Sawila demonstratives can be stacked, in which case the definite demonstratives occur last, preceded by deictic or visible demonstratives, as in (66), and (68) above and in (69). (69) Kuda atu likka duama angana, galoomang mi. [[kuda atu likka du ama]NP angana]DP ga-loomang mi 3.I-pick.up take horse excrement very PL FIN.DIST DIST.DEF ‘[He] picked up [all] that horse shit there.’ [KS.15] Table 9 also contains the dectic forms oli ‘here’ and ali ‘there’. Unlike the other forms, these never modify an NP, but occur as an independent constituent in the clause, referring to locations.

4.2 Modifiers Modifiers are intransitive verbs denoting a descriptive property of the head noun; there is no distinct morphosyntactic class of adjectives in Sawila. Modifiers always follow the head noun (N), and their category determines their relative ordering. As can be seen in (70), the COLOR modifier siila ‘white’ has to precede the SIZE modifier likka ‘large’. SHAPE modifiers have to precede SIZE modifiers, as shown in (71).

382

František Kratochvíl

(70) Se siila likka dana angana iring. . . siilaMOD likkaMOD danaQUANT anganaDEM]NP iring [seN FIN.one DIST.DEF fly chicken white large ‘That large white rooster flew. . .’ [KM.78] (71) Plaastik loora lamu likka dana angana wiya lee. [plaastikN looraMOD lamuMOD likkaMOD danaQUANT anganaDEM]NP plastic straight long large FIN .one DIST.DEF wiya lee put.in PFV ‘[They] fill that large long straight plastic bag [with the pounded peppers].’ [MS.46] In nominalized structures, modifier verbs may combine with aspectual markers, as in (72). (72) Luuli aruka kaana ngana, kiira madang taang. . . arukaMOD kaanaASP ngana]NMLZ kiira madang taang [luuliS FIN .COMPL DEF flat.basket NFIN.EACH pound peppers dry ‘As soon as the peppers are dry, she pounds each basket. . .’ [MS.46]

4.3 Relative and nominalized clauses Sawila has internally-headed relative clauses. The shared argument (in bold face) – the head noun of the relative clause (NREL) – is the S or A argument of the RC, as in (73)–(74). (73) Imealar dana apee eesa gamiti galila dooka yaana. ga-miti]RC ga-lila dooka yaana [[imyaalar dana]S apee eesa NFIN.man FIN.one wall lean.on 3.I-sit 3.I-lift stand FIN.go ‘The man, who was leaning on the wall, got up and walked away.’ [S.AM.96] (74) Aning sibe wise nannu matataaku neanu. sibeP wise]RC nannu ma-tataaku naanu [aningA NFIN .person back.basket carry.on.back 1SG APPL-meet FIN .not ‘I haven’t met the person, who was carrying the basket.’ [CC.97]

Sawila

383

The reference of the NREL can vary in terms of number between the RC from the main clause (MC). In (75), the NREL aning refers to ‘people’ inside the RC, but when further modified with the numeral sundana ‘one’ in the MC, its reference is to a single person. (75) Aning baala gapi dooka, sundana me. baalaP ga-pi dooka]RC sundana me [aningA NFIN .person ball 3.I-make stand FIN .one come ‘One of the people who were playing [soccer] came.’ [MPI.FbPF.3] Most Sawila RCs are unmarked, but in some cases, the focus particle po (FOC) may follow them, as shown in (76). (76) Aning yaku boorana kinkala gamidde po, ganaliya naanu. [aning yaku boorana kinkala ga-midde]RC po ga-na-liya naanu NFIN .person two FIN .only FIN .jump.over 3.I-go.up FOC 3.I-INV-hit FIN.not ‘Those two men who jumped up were not hit [by the ball].’ [MPI.FbHoTF] In a very few cases, the NREL is the P (the patient-like argument of a transitive verb) argument of the RC, as in (77). It is presently unclear whether the presence of the inverse marker on the verb paati changes the syntactic status of aning dana into the syntactic subject of the RC and whether consequently only syntactic subjects can be relativized in Sawila. (77) Aning dana katu ganapaati po, nite gadea naanu. ga-dea naanu [[aning dana]P katuA ga-na-paati]RC po nite NFIN .person FIN .one dog 3.I-INV-bite FOC 1SG .POT 3.I-FIN .see FIN .not ‘I haven’t seen the person who was bitten by the dog.’ [CC.36] Additional information about participants can also be provided in nominalized clauses, marked by demonstratives, as in (78)–(79). A variety of demonstratives are attested, but ama and ngana occur most frequently. (78) Araasing likka dana tingkaata gamiti ama, di gamiti muno? [araasing likka dana tingkaata ga-miti ama]NMLZ NFIN .house large FIN.one FIN.storey 3.I-sit DIST.VIS di ga-miti muno FIN .IPFV still 3.I-sit ‘Is that big house with two floors still there [or was it demolished]? [BDPK.234]

384

František Kratochvíl

(79) Katu daapuruma dooka ngana, gatiira gamuri! dooka ngana]NMLZ ga-tiira ga-muri [katuS daapuru=ma FIN .kitchen=be.PROX stand DEF 3.I-drive.out 3.I-run dog ‘Chase the dog standing in the kitchen!’ [QVP.143] The argument status of the nominalized clause inside the main clause may be unclear, and require resumptive pronouns. In (80), the resumptive ang, which follows the pause, serves to activate the reference to the ‘uninfected chickens’. In (81), the barking dog is referred to with the resumptive pronoun gannu. Nominalizations have uses beyond supplying additional information about participants; they can be used to encode reason, or stimulus. (80) Se di mate gamantaani neanu du ama, ang nirre suwe girre. [se di mate ga-mantaani naanu du ama]NMLZ ang FIN.not PL DIST.VIS DIST chicken still disease 3.I-infect nirra suwe girra 1SG.AGT inject must ‘the chickens not infected by the disease, those I will vaccinate’ [BF.52] (81) Katu gamayii gako ngana, nekatu gannu neanu. [katu ga-mayii ga-ko ngana]NMLZ ne-katu gannu naanu 1SG.II-dog 3 FIN.not dog 3.I-sound 3.I-stay DEF ‘The barking dog is not my dog.’ [QNP.5]

4.4 Quantifiers and numerals The following quantifiers are common: nanna ‘all’ (85), -ninale ‘all’ (91), du ‘PL’ (86), pike ‘a bit’, likka ‘very’, luri ‘many’, and maarang ‘together’ (87), and numerals (88)–(90). The distribution of several Sawila quantifiers is a diagnostic for noun classes. Dana, the short form of the numeral sundana ‘one’ combines with count nouns and often also indicates specific reference (82). The original numeral meaning is also available, as can be seen in (83), where the form dan in the quantifier slot is followed by the demonstrative ang. The form dana can also be used as a quantifier meaning ‘together’, and may combine with other numerals, which must precede it, as shown in (84).

Sawila

385

(82) Imyaala saaku dana suwi. [imyaala saaku dana]NP suwi FIN .one arrive man old ‘A [grown-up] man arrived.’ [MPICW.1] (83) Aning danang gani ngena Yohanis. ngana]NP Yohanis [aning danQUANT=angDEM ga-ni NFIN .person NFIN .one=DIST 3.I-name DEF Yohanis ‘The name of that one man was John.’ [Mk1:7] (84) Imyaalara yaku dana angana dooka biseara gako. [imyaalara yakuNUM danaQUANT anganaDEM]NP dooka biseara ga-ko FIN .man two NFIN .one DIST.DEF stand FIN.talk 3.I-stay ‘Two men together are talking.’ [MPI.TBTB] The total quantifier nanna ‘all’ combines exclusively with animates (85). (85) a. [ga-maddu nanna]NP b. [*yu nanna]NP 3.I-child all granary all ‘all his children’ [QNP.12] intended reading: ‘all granaries’ The plural marker du combines with any countable noun and may even be used with nouns denoting professions such as polisi du ‘the police, policemen’, as in (86). (86) Polisi duba made niliweaka lamana. ni-li-waaka lamana [polisi du]NP=ba made come.up 1PL.EXCL-APPL-watch FIN.immediately police PL=TOP ‘The policemen, they immediately came up to guard us.’ [SOS2.11] The quantifier pike refers to a small quantity, as in (87), and may also quantify events. Quantifiers likka and luri precede the plural marker du, as shown in (87b). The quantifier likka ‘very’ intensifies the meaning of the quantifier luri ‘many’, but regularly also occurs alone as an existential quantifier, as in (87c) where reference is made to a particular type of banana. The quantifier maarang ‘together’ follows the quantified constituent, as in (87d), where it follows the pronoun tannu ‘all of us’.

386

František Kratochvíl

(87) a. Eroti padiki dang pike. e-roti padiki dang pike NFIN.one just 2SG.II-bread a.bit ‘Just a bit of your bread.’ [WWH.37] b. Aning likka luri du ama. aning likka luri du ama NFIN .person very many PL DIST.VIS ‘Very many bodies.’ [SOS3.69] c. maaka likka dana d. tannu maarang [tannu maarang]PRO [maaka likka dana]NP banana very FIN .one DISTR together ‘this banana [bunch]’ [KSWS.32] ‘all of us together’ Sawila NPs may contain numerals to indicate the exact number of referents. The system is quinary (base 5 in 6 and above) and decimal (base-ten), as can be seen in (88). The numeral riibu ‘1000’ is a Malay loan.6 Note also that the numeral yaku ‘two’ has an allomorph raku in ma-raku ‘twenty’. (88) sundana araasiku yootingyaku adaaku asaka

‘1’ ‘4’ ‘7 (5+2)’ ‘10’ ‘100’

yaku yootine yootingtuo adaaku maraku riibu

‘2’ ‘5’ ‘8 (5+3) ‘20’ ‘1000’

tua yootingsuno yotaraasiku adaaku matua

‘3’ ‘6 (5+1)’ ‘9 (5+4)’ ‘30’

The connector garising is used to link basic cardinals with decimals. In higher numbers, sundana ‘1’ is shortened to dana, as can be seen in (89). Ordinal numerals are derived with the prefix wii-. (89) Riibu dana asaka yootaraasiku adaaku maraku garising yaku. riibu dana asaka yootaraasiku adaaku ma-yaku ga-rising yaku ten APPL-two 3.I-increase two thousand FIN.one hundred nine ‘1922’ In enumeration, the numeral phrase (NumP) may sometimes be separated from the NP and function as a non-verbal predicate, as in (90). The NumP may contain an optional classifier-like constituent, such as aning ‘person’, or kiki ‘seed’. 6 Sawila numerals are cognate with the other AP languages except for sundana ‘one’.

Sawila

387

(90) Juraagana aning sundana, murra ngana (aning) yootingsuno. sundana]NumP [murra ngana]NP [juraagana]NP [aning FIN .captain NFIN .person FIN .one servant DEF [(aning) yootingsuno]NumP NFIN.person FIN .six ‘[We were issued our travel permits:] one captain, the crew six [men].’ [HM.084] A number of quantifiers, such as -ninale ‘all’ combine with person prefixes co-indexing the quantified entity. (91) shows that ga-ninale is not sensitive to animacy. (91) a. yaninale b. araasing mana ganinale [araasing mana ga-ninale]NP [ya-ninale]PRO 2PL.I-all house DIST 3.I-all ‘all of you’ [BDPK.7] ‘all the houses over there’ [L15] Forms -ning/ni/na- ‘together’ and -noosa ‘alone’, shown in (92), indicate a precise number; -waa ‘jointly’ is underspecified. The numeral root yaku undergoes shortening to ku. (92) a. gimale ganaku b. nanoosa kide [na-noosa]QuantP kide gi-male [ga-na=yaku]QuantP 3.I-together=two 1SG.I-alone only 3.II-servant ‘[the king] with his servant’ [KAS.8] ‘just me alone’ [Q.45] c. Ganaku tawaa garammu. ga-na=yaku [ta-waa]QuantP ga-rammu 3.I-die 3.I-together=two DISTR.I-same ‘Both of them alike died.’ [NB4.176] To inquire about the number of participants, the question word diini is placed in the quantifier slot of the NP, as shown in (93). (93) Q: Erra pi diini wusi? A: Nirra adaaku wusi! erra pi diini wusi nirra adaaku wusi 1SG.AGT ten shoot 2SG.AGT pig how.many shoot ‘How many pigs have you shot?’ ‘I have shot ten!’ [QVP.92–93]

388

František Kratochvíl

4.5 Possessors The possessive relationship can be marked on the possessed noun by four sets of person prefixes, shown in Table 10. In addition to possessor marking prefixes, there is also a genitive paradigm of free pronouns, discussed in sections 3.2 and 5. Table 10: Sawila possessive prefixes PERSON

a-SET

ang-SET

i-SET

e-SET

1SG 2SG

nangangtanggangningingying-

nietigini-/niyiyi-

ne-

3 1PL.EXCL 1PL.INCL 2PL

naatagani-/niyiyi-

POSSESSOR

INALIENABLE

DISTR

tege-

ALIENABLE

The a- and ang- sets are obligatory and mark inalienable possession. The ang- forms occur with /k/-initial roots and can be considered allomorphs of the former. The class of inalienably possessed nouns contains most body parts, and several kinship terms, such as -maddu ‘child’, -taamu ‘grandchild’, -mooka ‘cousin’, -naamu ‘husband’, -yo ‘wife’, -daatala ‘in-law’. (94) a. na-mang 1SG.I-neck ‘my neck’

b. a-limpuru 2SG.I-tongue ‘your tongue’

c. ga-yiiri 3.I-mouth ‘his mouth’

d. nang-kari e. ang-kalla f. gang-kise 2SG.I-shoulder 3.I-back 1SG.I-forehead ‘my forehead’ ‘your shoulder’ ‘his back’ The e- and i- sets mark alienably possessed nouns, which can be divided into two subclasses, as shown in (95). A relatively small set of nouns combines with the e- set (95a–d). The set contains inherently possessed kinship terms such as ne-pa ‘my father’. The great majority of Sawila nouns combines with the i- set prefixes, which are always optional (95e–h). (95) a. ne-pa b. *pa c. e-ya d. *ya father 2SG.II-mother mother 1SG.II-father ‘my father’ ‘your mother’

389

Sawila

e. ni-katu f. e-kine g. gi-male h. i-manaa 2SG.II-knife 3.II-servant 1PL.INCL-village 1SG.II-dog ‘my dog’ ‘your knife’ ‘his servant’ ‘our villages’ Relational concepts are expressed by simple juxtaposition of two nouns, in which the first noun specifies the type of the head, as in (96a–b). Part-whole relationships, such as in (96c), regularly require the possessive construction. (96) a. baasara araasing b. se atu c. arakeesa ga-apaala FIN.market NFIN.house chicken shit FIN .rice 3.I-chaff ‘market houses’ ‘chicken shit’ ‘rice chaff’ The Sawila possessive construction is used to construct predicates of bodily condition, emotion, cognition, and perception. The literal usage describing physical characteristics or bodily condition can be seen in (97). It is ungrammatical to raise the possessor and drop the noun denoting the body part, as shown in (97b), or to add a free pronoun after the body part noun (97c). (97) a.

Gawaari muko. ga-waari muko b. deaf 3.I-ear ‘He is deaf.’ [P.150]

*na-muko 1SG.I-deaf

c.

*ga-waari 3.I-ear

gannu 3

muko deaf

Bodily processes and states are expressed in a similar way, as can be seen in (98). The human experiencer is expressed only indirectly as a possessor. (98) a. Naniiri keana. b. Naniiri keankaana. na-niiri kaana na-niiri kaan~kaana 1SG.I-eye RDP~FIN.good 1SG.I-eye FIN.good ‘I am alive.’ [P.77] ‘I have woken up.’ [P.78]

5 Pronouns Sawila has three paradigms of pronouns, listed in Table 11. The number distinction is neutralized in the third person.7 Distributive forms are used with collective and generic reference. First person plural forms distinguish inclusive and 7 The number opposition in the unrestricted paradigm is not used consistently; the form gannu may have a plural reference.

390

František Kratochvíl

exclusive reference. The agentive paradigm shows formal irregularity in first plural exclusive, where the first syllable is lengthened. The genitive forms are homophonous and likely historically related to agentive forms. More details on their use can be found in section 3.2. The potentive paradigm displays variation too, presently not well understood. The CVta forms may be outcomes of language contact with the neighboring Wersing (see Schapper and Hendery, this volume). Table 11: Sawila free pronouns PERSON

AGENTIVE /GENITIVE

POTENTIVE

UNRESTRICTED

1SG 2SG 3

nirra erra girra tirra niira irra yirra

nite/neta ete/eta gite/geta tite/tita nite/niita ite/ita yite/yita

nannu annu gannu tannu ninnu innu yannu gandu

DISTR

1PL.EXCL 1PL.INCL 2PL 3PL

The distribution of the free pronouns in Sawila is according to a semantic alignment (Donohue and Wichmann 2008). The distinctions between these paradigms become obvious when alternations, which are common, are contrasted, as in (99)–(100). The agentive forms highlight volitional and controlling involvement of the agent; potentive forms indicate agent’s ability or a decreased degree of volition or control; the unrestricted forms admit all argument types and are the only possibility to realize P and SP arguments. (99) a. Nirre namade. nirre na-made 1SG.AGT 1SG.I-come.up ‘I shall come up.’ c. Nannu namade. nannu na-made 1SG.I-come.up 1SG ‘I am coming up.’ [NB4.25]

b. Nite namade? nite na-made 1SG.POT 1SG.I-come.up ‘Should I come up?’ d. Namade. na-made 1SG.I-come.up ‘I am coming up.’

(100) a. Nirra nasoorana. b. Nite nasoorana. nirra na-soorana. nite na-soorana 1SG.POT 1SG.I-FIN.angry 1SG.AGT 1SG.I-FIN.angry ‘I am [deliberately] raging.’ ‘I feel angry [capable to show anger].’

Sawila

391

c. Nannu nasoorana. d. Nasoorana. nannu na-soorana na-soorana 1SG.I-FIN.angry 1SG.I-FIN.angry 1SG ‘I am angry.’ [N12.22] ‘I am angry.’ [NB4.65] While some verbs are compatible with all three free pronoun paradigms, other verbs are more restricted, due to semantic incompatibility of the verb and either the agentive or potentive paradigm. Such behavior is typical of intransitive verbs whose single argument is not a semantic agent, as in (101). (101) a. Nannu nayaati. b. Nayaati. nannu na-yaati na-yaati 1SG.I-FIN.sick 1SG.I-FIN.sick 1SG ‘I am sick.’ [N12.027] ‘I am ill.’ [BP.228] d. Nannu narammu. nannu na-rammu 1SG.I-die 1SG ‘I am dying.’

c. Nirre nayaati. nirra na-yaati 1SG.AGT 1SG.I-FIN.sick ‘I pretend to be sick.’

e. Narammu. f. *Nirra narammu. na-rammu *nirra na-rammu 1SG.I-die 1SG.AGT 1SG.I-die ‘I am dying/dead.’ [N12.025]

All of the examples above show that free pronouns may cooccur with coreferential person prefixes. However, the distribution of person prefixes is not determined by the same semantic alignment; it is lexicalized and discussed separately in section 6.

5.1 Agentive paradigm Sawila agentive pronouns express semantic agents (both transitive A or intransitive SA) as in (102). Third person forms may occur in apposition to coreferential NPs, which they must always follow (102f). (102) a. Nirra tiimu masaara. b. Erra loora. [erra]S loora [nirra]A [tiimu]P ma-saara yard APPL-sweep 2SG.AGT FIN.right 1SG.AGT ‘I am sweeping the yard.’ [N12.171] ‘You are right.’ [NB5.215]

392

František Kratochvíl

c. Erra asiina dana tatuku. dana]P tatuku [erra]A [asiina 2SG.AGT FIN.fable FIN.one tell ‘You are telling a fable.’ [NB4.044] d. Girra namasaraana naanu. naanu [girra]A na-ma-saraana 1SG.I-APPL-FIN.ask FIN.not 3.AGT ‘He didn’t ask me.’ [NB4.018] e. Girra buku mi wiiyreani. girra buku mi wii-i-rani 3.AGT book take APPL-1PL.INCL-FIN.give ‘He gave us a book.’ [N12.173] f. Aning du girra parra laata. [aning du girra]A [parra]P laata NFIN .person PL 3.AGT field burn ‘People are burning fields.’ [NB4.015] In addition to agents, agentive pronouns can express forces, such as wiiki ‘flood’ in (103). (103) Wiiki girra giyaraasing gabungala. wiiki girra gi-’araasing ga-bungala flood 3.AGT 3.II-NFIN.house 3.I-wash.down ‘The flood washed down their houses.’ [N12.176] Under negation, agentive pronouns stress the active refusal of the actor to participate in the event, as in (104). Otherwise, unrestricted forms are preferred, as in (105). (104) a. Lalama nirre siribisi neanu. lalama [nirra]S siribisi naanu FIN .not tomorrow 1SG.AGT work ‘Tomorrow I will not work (because I have decided so).’ [N12.020] b. Lalama nannu siribisi neanu. lalama [nannu]S siribisi naanu work FIN .not tomorrow 1SG ‘Tomorrow I will not work (I have been told not to).’ [Q.030]

Sawila

393

(105) a. Nirra apunni. b. Nannu apunni neanu. [nannu]A (?nirra) a-punni naanu [nirra]A apunni 1SG 1SG.AGT 2SG.I-hit FIN.not 1SG.AGT 2SG.I-hit ‘I hit you.’ [P.105] ‘I did not hit you.’ [P.105] c. Ang nannu di mayiin neanu. mayiin naanu [ang]P [nannu]A (?nirra) di DIST 1SG 1SG.AGT still NFIN.hear FIN.not ‘That I haven’t heard before.’ [BDPK.091]

5.2 Potentive paradigm Potentive pronouns differentiate semantic agents that have the ability or intention to perform an action. As such, their grammatical function extends to mood marking, encoding unactualized situations (irrealis). (106) a. Nite namidde. b. Eta mayiina? nite na-midde eta mayiina 2SG.POT FIN.hear 1SG.POT 1SG.I-go.up ‘I am just leaving.’ [KM.074] ‘Can you hear?’ [BDPK.314] c. Nite anatulong naanu, arammu! naanu, a-rammu [nite]A a-na-tulong 1SG.POT 2SG.I-INV-NFIN.help FIN.not 2SG.I-die ‘I don’t intend to help you, you drop dead!’ [MT.067] Potentive forms are also regularly used to express experiencers, whether they are involved in the event voluntarily or not, as in (107). (107) a. Giya gite iyera. [gi-ya gite]A i-yara 3.II-mother 3.POT 1PL.INCL-laugh ‘Her mother was laughing at us.’ [MS.039] b. Ekatu girra namaddu peati ete gadea? [e-katu girra na-maddu paati]P [ete]A ga-dea 2SG.POT 3.I-FIN.see 2SG.II-dog 3.AGT 1SG.I-child bite ‘Did you happen to see your dog biting my child?’ [CC.102]

394

František Kratochvíl

5.3 Unrestricted paradigm Unrestricted pronouns are admitted to all argument positions (S, A, P) and regularly also encode possessors. They are the only paradigm that can express the transitive P argument, as shown in (108). (108) Nirra gannu gapunni. [nirra]A [gannu (*girra)]P ga-punni 3.AGT 3.I-hit 1SG.AGT 3 ‘I hit him.’ [N12.018] The unrestricted paradigm is also the default choice to realize the S of unaccusative verbs, as shown in (100)–(101) above and in (109) below. (109) a. Nannu miilana. [nannu]S (*nirra) miilana 1SG.AGT FIN.threaten 1SG ‘I am scared.’ [P064] b. Bangko gannu gayaati wanti. wanti [Bangko gannu]S (*girra) ga-yaati 3.I-FIN.sick REPORT Bangko 3 3.AGT ‘Bangko is reportedly ill.’ [CC.062]

5.4 Pronominal doubling Free pronouns realizing semantic agents (A or S) may undergo doubling. The first pronoun must come from the unrestricted paradigm, followed by the agentive or potentive form; the reverse order is ungrammatical. Doubling never occurs with P and SP arguments. Pronominal doubling is used to put the agent in focus, as can be seen in (110). (110) a. Aning mu tukki dooka, nannu nite gadea naanu. [aning mu tukki dooka]P [nannu nite]A ga-dea naanu 1SG.POT 3.I-see FIN.not person wood cut stand 1SG ‘I haven’t been able [but someone could] to see the person who was cutting the wood. ’ [CC.098]

Sawila

395

b. Kaana, nannu nirra nooning liteani? kaana [nannu nirra]A nooning li-taani FIN .good 1SG 1SG.AGT NFIN.how APPL-answer ‘Well, how should I answer then?’ [SOS3.085] c. Ninnu nirre yamsalu mading siida. [yaamusalu]P mading siida [ninnu niira]A NFIN .plant finish 1PL.EXCL 1PL.EXCL.AGT cassava ‘We have planted all cassava [others may be still planting].’ [D.Madine.1] Pronominal doubling is also available to topical agents, but the unrestricted pronoun is left-dislocated and followed by a pause. The topical status of the argument can be overtly marked, as shown in (111). (111) Nannu, nirraba we lee awa. lee awa [nannu] [nirra]A=ba we 1SG.AGT=TOP leave PFV DIST.NVIS 1SG ‘Me, I am actually out of [stories].’ [KSWS.3]

6 Verbal morphology Sawila verbs combine with prefixes which mark the relationship between the verb and its arguments. Besides two paradigms of person marking prefixes (AGR), there is one inverse prefix (INV) and three distinct applicative prefixes (APPL). The longest attested form contains five prefixes, whose ordering is shown in (112). Maximum verb stem (112) APPL-AGR-INV-APPL-AGR-V Which of the prefixes appears with a particular verbal root is lexicalized, allowing us to identify several inflectional subclasses, further discussed in section 6.4.

396

František Kratochvíl

6.1 Person marking prefixes There are two paradigms of person marking prefixes, which can be seen in Table 12.8 The two sets show a clear morphological correspondence with each other, differing just in the vowel quality in singular and being homophonous in plural. As discussed in section 4.5, the person prefixes also mark possessors on nouns. Table 12: Sawila person prefixes PERSON

SET

1SG 2SG

naatagani-/niyiyi-

DISTR

3

1PL.EXCL 1PL.INCL 2PL

I

SET

II

nietigini-/niyiyi-

Set I person prefixes mark all core argument types (S, A, P). First and third person forms can be seen in (113). (113) a. Napaati. b. Nannu tukku nanee niwuli. [nannu tukku]A naA-nee ni-wuli naS-paati 1SG palm.wine 1SG.I-consume 1SG.II-desire 1SG.I-eat ‘I eat’ [P.203] ‘I want to drink arak.’ [NB5.245] c. Katu pi gakuna. katu [pi]P ga P-kuna dog pig 3.I-bite.to.death ‘The dog bit the pig to death.’ [P.204] Distributive prefixes are used with generic reference (114a) and with reciprocals (114b–c). (114) a. Angana gidawana ngana tate. angana gi-dawana ngana taS-te DIST.DEF 3.II-FIN.medicine DEF DISTR. I-lie ‘For that the only remedy is that one sleeps.’ [BDPK.033] 8 The Sawila person prefix paradigm resembles systems found IN other languages of the AlorPantar family. The system is organized in a similar way as the reconstructed Trans-New Guinea systems – consonant indicates person, vowel records number (see Ross 2005; Kratochvíl et al. 2011).

Sawila

397

b. Ganaku tataani neanu. naanu ga-n-yaku taS-taani 3.I-together-two DISTR.I-FIN.fall FIN.not ‘They always stick together.’, lit. ‘They do not drop each other.’ [NB5.011] c. Imyaalara yaalara ganaku tanami. imyaalara yaalara ga-n-yaku ta-na-mi man female 3.I-together-two DISTR.I-INV-take ‘The man and the woman married each other.’ [NB5.079] While set I prefixes combine with the majority of Sawila verbs, set II prefixes are attested only with the following ten verbs: -ku ‘fart’, -laadi ‘ashamed’, -wuli ‘desire’, -suko ‘like’, -litinga ‘understand’, -wiita ‘call’, -tatuku ‘tell’, -tekuli ‘do on purpose, -mu ‘help’, -suwaara ‘show’. (115) Ganintuba gilaadi wiigawaari. ga-nin-tua=ba gi-laadi wii-ga-waari 3.I-together-three=TOP 3.II-ashamed APPL-3.I-return ‘All three of them returned ashamed.’ [WWH.047] (116) Gannu ganamulaana gibuku mi linisuwaara. gannu ga-na-mulaana gi-buku mi li-ni-suwaara 3 3.I-INV-FIN.not.want 3.II-book take APPL-1SG.II-show ‘He doesn’t want to show me his book.’ [NB5.249]

6.2 Applicative prefixes The Sawila applicative prefixes ma-, li-, and wii- have two main functions.9 The first one is to mark an increased or decreased effect of the event on the P argument. The second one is to extend the valence of the verb by adding a new argument, which will be labeled as E (for extension), corresponding to Dixon’s ‘AP-O’ (2012: 294ff). With monovalent verbs, the prefix ma- regularly indicates a greater intensity of the event but usually does not extend the valency, unlike the prefix li-, which adds a beneficiary or location to the argument structure. With verbs that do not combine with person prefixes, applicative prefixes attach directly to the root.

9 The applicative prefixes grammaticalized from the verbs ma ‘be.PROX’, li ‘be.DIST’, and wii ‘use’.

398

František Kratochvíl

(117) a. Laampuru ilo. b. Laampuru mailo. [laampuru]S ma-ilo [laampuru]S ilo APPL-bright lamp bright lamp ‘The lamp is bright.’ [A.022] ‘The lamp is brighter/turned up.’ [A.024] c. Laampuru liilo. laampuru li-ilo APPL-bright lamp ‘The lamp is shining for someone/at something.’ [P.204] If person prefixes are present, applicative prefixes must precede them, as in (118). The meaning of applicative prefixes remains similar; ma- again indicates a greater intensity, while li- and wii- add E denoting the stimulus. (118) a. Nasoorana. b. Erra gapi manasoorana. erra ga-pi ma-na-soorana naS-soorana 2SG.AGT 3.I-make APPL-1SG.I-FIN.angry 1SG.I-FIN.angry I am angry. [A.117] ‘You are making me more angry.’ [A.119] c. Nirre linesoorana. nirra li-na-soorana 1SG.AGT APPL-1SG.I-FIN.angry ‘I am angered about it.’ [A.118] d. Erra tulori napunni ngana, kaluno wiinesoorana. [erra tulori na-punni ngana]E kaluna wii-na-soorana FIN.now APPL-1SG. I-FIN .angry 2S.AGT two.days.ago 1SG.I-hit DEF ‘You hitting me that time, is making me angry now.’ [A.120] With bivalent verbs, as in (119), the prefix ma- indicates that the event is carried out with a greater effect on the same P. The prefixes li- and wii- promote locations and instruments to E arguments. The P argument luuku ‘hole’ may be dropped. (119) a. Nirre luuku tii. b. Nirre luuku matii. nirra [luuku]P ma-tii nirra [luuku]P tii dig 1SG.AGT hole APPL-dig 1SG.AGT hole ‘I dig a hole.’ [A.9] ‘I dig a hole deeper.’ [A.012] c. Nirre siripine litii. nirra [siripine]E li-tii APPL-dig 1SG.AGT FIN.root ‘I dig [a hole] along a root.’ [A.011]

Sawila

399

d. Nirre mu madiirana wiitii. nirra [mu madiirana]E wii-tii APPL-dig 1SG.AGT wood FIN.sharp ‘I dig [a hole] with a sharp wood.’ [A.014] On person-marked bivalent verbs, applicative prefixes also must attach before person prefixes, as can be seen in (120). The person-marking prefix cannot be dropped. (120) a. Girre niwiite. b. Girre maniwiite. girra ma-niP-wiita girra niP-wiita 3.AGT APPL-1SG.II-call 3.AGT 1SG.II-call ‘He is calling me.’ [A.035] ‘He keeps calling me.’ [A.037] c. Girre liniwiite. girra li-niP-wiita 3.AGT APPL-1SG.II-call ‘He is calling me to it [to take some fish].’ [A.036] As mentioned above, the prefix ma- indicates an increased effect on the P. With some verbs, the increased effect simply means repetition, as with the pairs ga-pi ‘build’ ~ ma-ga-pi ‘rebuild, repair’ in (121). The prefix li- may indicate a lesser effect, as in ga-pi ‘build’ ~ li-ga-pi ‘extend, build at’, where no E argument is introduced. (121) a. Nirra araasing gapi. nirra araasing ga-pi 1SG.AGT NFIN.house 3.I-make ‘I build a house.’ [A.5] b. Nirra araasing magapi. nirra [araasing]P ma-ga-pi 1SG.AGT NFIN.house APPL-3.I-make ‘I rebuild [repair] a house.’ [A.6] c. Nirra araasing ligapi. nirra [araasing]E li-ga-pi 1SG.AGT NFIN.house APPL-3.I-make ‘I extend [build at] a house.’ [A.7]

400

František Kratochvíl

d. Niaraasine mu wiigapi. [mu]E wii-ga-pi [ni-araasine]P 1SG.II-FIN.house wood APPL-3.I-make ‘My house is built with wood.’ [A.8] The applicative morphology is a productive part of Sawila grammar. As illustrated in (122a), the applicative prefix li- can combine with Malay loans such as paarlu ‘need’ (from Mly. perlu ‘need’) and with complex verbs such as maali=leawa ‘indulge oneself’ (122b). (122) a. Nannu iria lipaarlu boorana. nannu [iria]E li-paarlu boorana water APPL-need FIN.very 1SG ‘I really need water.’ [QVP.43] b. Gannu limalileawa. gannu li-mali=laawa APPL-tasty=search 3 ‘He always indulges himself.’ [A.71]

6.2.1 Applicative prefix maAs illustrated above, the applicative prefix ma- usually indicates the increased effect of an event on P. In some cases, however, the prefix is used to mark a surprising value of P argument, such as gibalisi ‘rice packet [boiled or steamed rice wrapped in leaves for human consumption, not for chickens], as in (123). (123) a. Alawurukola araaka tuku gako mana. alawurukola [araaka]P tuku ga-ko mana zebra.finch rice peck 3.I-stay over.there ‘Zebra finches are pecking rice over there.’ [D.Tuku.1] b. Se siila dana angana gibalisi matuku. se siila dana angana [gi-balisi]E ma-tuku chicken white FIN.one DIST.DEF 3.II-packet APPL-peck ‘The white rooster was pecking in his (steamed rice) package.’ [NB4.163] When attached to monovalent verbs describing properties, the prefix maindicates that the described property exceeds the standard.

Sawila

401

(124) a. Nannu marung pike. b. Nannu mamarung. [nannu]S ma-marung [nannu]S marung pike small just 1SG APPL-small 1SG ‘I am very small.’ [N12.184] ‘I am the smallest (among all).’ [N12.183] The Sawila prefix ma- only rarely adds a location to the argument structure, unlike in Kula, where this use of ma- is common (Williams and Donohue n.d.). In Sawila, such uses usually involve a shift in meaning, as can be seen in (125), where the meaning of miti shifts from ‘sit’ to ‘stay’. (125) a. Nepa gamiti. b. Nepa magamiti. ga-miti [ne-pa]A Ø ma-ga-miti [ne-pa]S 1SG.II-father APPL-3.I -sit 1SG.II-father 3.I-sit ‘My father is sitting.’ [A.51] ‘My father is staying [here].’ [A.53] c. Noong araasing matamiti. [noong araasing]E ma-ta-miti which NFIN.house APPL-DISTR.I-sit ‘In which house shall we stay?’ [N12.3] d. Nepa ligamiti. Ø li-ga-miti [ne-pa]A APPL-3.I-sit 1SG.II-father ‘My father is sitting [on it].’ [A.52]

6.2.2 Applicative prefix liThe applicative prefix li- promotes a non-argument to a core argument position and often replaces the original P argument, which has to be expressed in a separate clause, as can be seen in (126). (126) a. Girra we paasa. girra [we]P paasa 3.AGT goat bind ‘He has bound up the goat.’ b. Girre we mi aka lipeasa. girra [we]P mi [aka]E li-paasa 3.AGT goat take fence APPL-bind ‘He has tied the goat to the fence.’ [A.16]

402

František Kratochvíl

Applicative prefixation with li- is often accompanied by a slight meaning shift, such as king ‘cut’ ~ li-king ‘shave (i.e. cut a bit)’. (127) a. Iloola mi king mi isuwi maaka ka maana inee-ipeati. i-loola mi king mi i-suwi maaka ka maana take 1PL.INCL-arrive banana IRR FIN.cook 1PL.INCL-walk take cut ‘Having gone, cut, and brought back, we will cook and eat that banana.’ [SC.61] b. Napiini madi liking. li-king [na-piini madi]E 1SG.I-lip body.hair APPL-NFIN.cut ‘I shave my beard.’ [NB5.174] With some verbs, the applicative li- encodes a meaning similar to a partitive – it marks a partially affected P, as in (128), where examples (b,d) imply that the banana is not going to be entirely eaten and the church not going to be entirely painted. (128) a. Maaka itiina nanee lee. lee [maaka itiina]P na-nee banana ripe 1SG.I-consume PFV ‘I have eaten up a banana.’ [P.125] b. Maaka itiina linanee. [maaka itiina]E li-na-nee APPL-1SG.I -consume banana ripe ‘I ate a bit of banana.’ [N12.102] c. Imyaalara mura du putaraasine ceeta. imyaala mura du [putaraasine]P ceeta PL FIN .church FIN.paint boy ‘The boys are painting the church.’ [A.3] d. Imyaalara mura du putaraasine maceeta. imyaala mura du [putaraasine]E li-ceeta PL FIN.church APPL-FIN .paint boy ‘The boys continue to paint the church.’ [A.2] 6.2.3 Applicative prefix wiiThe applicative prefix wii- usually promotes an instrument to the E argument of the verb (i.e. the argument licensed by the applicative prefix), while the original P may be dropped.

Sawila

403

(129) a. Girre natang paasa. girra [na-tang]P paasa 3.AGT 1SG.I-arm bind ‘He has bound up my hand.’ [NB4.147] b. Karupo angana kape wiipaasa. [karupo angana]P [kape]E wii-paasa APPL-bind wound DIST.DEF rope ‘The wound has to be bound with a rope.’ [A.18] c. Saaku tiila masi, angana pang-kadeli. saaku tiila masi angana pang-kadeli DIST. DEF season.name old speak if ‘When old people speak, [they call] that [season] pang-kadeli.’ [SC.69] d. ‘Anjing’ ngana titatuku wiitiila masi? wii-tiila masi [anjing ngana]P [ti-tatuku]E DISTR. II-languge APPL-speak if dog (Mly) DEF ‘How do you say the [word] ‘dog’ in our language?’ In addition to instruments, the applicative prefix wii- always combines with the transfer verb -rani ‘give’ to add the transferred theme to its argument structure, as in (130). More details can be found in section 8.2. (130) Liiri nesu ngana: “We adaaku mi wiinarani!” liiri na-su ngana we adaaku mi wii-na-rani goat ten take APPL-1SG.I-FIN.give ruler 1SG.I-order DEF ‘The chief ordered me saying: “Give me ten goats!”’ [QVP.9]

6.3 Inverse prefix Sawila prefix na- indicates the presence of an affected human. Following Donohue’s (1997b) description of a cognate prefix in closely related Kula (referred to as Tanglapui in Donohue’s paper), the prefix is labeled INV (for inverse), but its distribution is not the same as in Tanglapui. Donohue (1997b: 105–107) states that Tanglapui na- is restricted to transitive affective verbs, i.e. ‘verbs in which the affected argument undergoes a change of state, or is adversely affected by the completion of the action’. The Tanglapui na- is used when third person acts on first or second person (3 > 1/2), or second person acts on first person (2 > 1),

404

František Kratochvíl

but not in cases where first person acts on second or third person (*1 > 2/3), or when second person acts on third (*2 > 3). The main function of the Sawila na- is to indicate the presence of an affected human argument in a marked scenario, (3 > 1/2, 2 > 1, or even 3/2/1 > 3human), as in (131). The human argument is always co-indexed with a set I person prefix. (131) a. Erra noonine tulante? erra noonine tulong=te 2SG.AGT FIN.how NFIN.help=Q ‘How will you help?’ [MS.40] b. Asire nanatulong dapa! a-sire na-na-tulong dapa 2SG.I-come.down 1SG.I-INV-NFIN.help PRIOR ‘Come down and help me!’ [MT.66] Sawila prefix na- is not restricted in terms of referential hierarchy; it is irrelevant whether the acting argument is ranked higher or the same on the referential hierarchy: second person acts on third in (132b), third person acts on third in (132c), distributive refers to first person in a reciprocal construction in (132d). In all cases, the function of na- is to disambiguate the role of the human participants, as to whether or not they are affected by the event. This is particularly obvious in cases where the human participant is affected only indirectly, as in (132c). The sentence describes a man taking a cup from a woman, who has to be indexed with the inverse prefix, not the theme ‘cup’. (132) a. Yaala saaku dawang mi, gatangmaala gawaali garammu. yaala saaku dawang mi ga-tangmaala NFIN.medicine take 3.I-daughter-in-law female old ga-waali ga-rammu 3.I-treat 3.I-die ‘The old lady took the medicine to poison her daughter-in-law.’ [QVP.67] b. Gannu game munteana ganami dapa game. gannu ga-me mun=tana ga-na-mi dapa ga-me 3 3.I-come IPFV=same.time 3.I-INV-take PRIOR 3.I-come ‘If he wants to come, take him and come.’ [QVP.63]

Sawila

405

c. Yaala saaku iria wunni dooka, imeala saaku me girra ganemi. yaala saaku iria wunni dooka imyaala saaku me lady water hold stand guy come girra ga-na-mi 3.AGT 3.I-INV-take ‘A lady was holding a cup; a guy who came in took [it] from her.’ [SAL2.38] d. Imyaalara dana nineku tanami. imyaalara dana ni-n-yaku ta-na-mi FIN.one 1PL .EXCL-together-two DISTR. I-INV-take man ‘I got married to a man.’ lit. ‘We took each other with one man.’ [N12.95] In Tanglapui the prefix na- is restricted to ‘transitive affective verbs’, but Sawila allows certain verbs to occur with and without the inverse prefix. When the alternations are contrasted, as in (133), the inverse prefix is used when the P argument is affected in an irreversible or not easily repeated way (scold vs. ask to marry). (133) a. Nannu Sawila tiila limei. nannu Sawila tiila limei Sawila speak learn 1SG ‘I learn to speak Sawila.’ [CC.94] b. Girra natiila. c. Girra nanatiila. girra na-tiila girra na-na-tiila 3.AGT 1SG.I-INV-speak 3.AGT 1SG.I-speak ‘He is scolding me.’ [NB4.70] ‘He asked to marry me.’ [N12.97] The verb wunni displays a similar alternation, but the meaning shift is greater (hold vs. hide vs. catch). (134) a. Nanala wunni kaang-kaana! naanala wunni kaang-kaana FIN.well something hold ‘Hold it properly!’ [D.Wunni.1] b. Nirra gawunni. c. Nirra ganawunni. nirra ga-wunni nirra ga-na-wunni 1SG.AGT 3.I-INV-catch 1SG.AGT 3.I-hide ‘He hid it away.’ [BP.122] ‘I caught him.’ [BP.123]

406

František Kratochvíl

Some bivalent verbs index the A argument when the P is non-human. However, with human Ps, the inverse prefix na- is required, as in (135). (135) a. Nidana pu gapaati. ni-dana pu ga-paati 1SG.II-friend betel.nut 3.I-eat ‘My friend is chewing betel nut.’ [N12.123] b. Waawa anapaati girre nea. waawa a-na-paati girra na crocodile 2SG.I-INV-eat certainly VIS ‘The crocodile will certainly eat you.’ [A.10] The inverse prefix does not occur with verbs that index their P argument (class III), such as kinyaala ‘cheat’, shown in (136). More information about this verb class can be found in section 6.4.3. (136) a. Girra nakinyaala. girra na-kinyaala (*na-na-kinyaala) 1SG.I-INV-cheat 3.AGT 1SG.I-cheat ‘He cheated me.’ [NB4.102] b. Nirra akinyaala. nirra a-kinyaala (*a-na-kinyaala) 1SG.AGT 2SG.I-cheat 2SG.I-INV-cheat ‘I cheated you.’ [N12.99] As can be seen in (137), na- also attaches to monovalent predicates describing states that have some adverse effect on the participant (a–d) or permanent state (e–f). Similar groups of intransitive verbs are also reported to combine with na- in Kula (Donohue 1997b: 110–111; Nicholas Williams pers. comm.). (137) a. Nanasaaku. b. Nanamate. na-na-saaku (*na-saaku/ *nannu saaku) na-na-mate 1SG old 1SG.I-INV-hurt 1SG.I-INV-old 1SG.I-old ‘I am old.’ [N12.71] ‘I feel sick.’ [P.112] c. Nanamasing. d. Nana’eesala. na-na-masing na-na-’eesala 1SG.I-INV-gasp.for.breath 1SG.I-INV-NFIN.hungry ‘I am hungry.’ [P.122] ‘I am exhausted.’ [WWH.34]

Sawila

407

e. Nannu nana’akana. f. Nannu nanamareeka. nannu na-na-’akana nannu na-na-mareeka 1SG.I-INV-black 1SG 1SG.I-INV-yellow 1SG ‘I am black (Papuan).’ [P.18] ‘I am white (Eurasian).’ [P.22]

6.4 Verbal inflectional classes Sawila verbal inflectional classes are distinguished by the attested combinatory behavior of verbal roots with verbal prefixes (person prefixes, applicative prefixes and the inverse prefix na-) discussed above. There are four major classes: classes I–III are relatively large, but class IV contains just about two dozen verbs. Table 13: Distribution of Sawila verbal prefixes CLASS

APPL 5

I

A

I

B

II

APPL

APPL

AGR4

INV 3

APPL2

S/P human

INV

Ehuman

INV

APPL

Ehuman

INV

APPL

INV /APPL

AGR1

S/(A) S/ P

III

A

Ehuman

III

B

E1-2.I

APPL

P 3.ii

IV

A

Ehuman

APPL

S

IV

B

Ehuman

APPL

IV

C

IV

D

APPL

Shuman SII

INV SI

Class I verbs do not index their core arguments but have to combine either with the inverse na- to index highly ranked Ps (Ia), or add an E argument with an applicative prefix (Ib). Class II verbs index the S/A argument and may take on applicative prefixes. Class III verbs obligatorily index the S/P argument and do not require the inverse prefix na- when the P is highly ranked on the referential hierarchy. Class IV verbs are monovalent, but display considerable variation in indexing the S argument. Most class IV verbs are compatible with applicative prefixes, but do not require the inverse prefix na- to index a human E. Whether the applicative prefixes are present or not has no consequence for the person prefixes in the first agreement slot (AGR1). Some Sawila verbs display defective or fossilized agreement. The verb magadiila ‘pursue’ is derived with the applicative prefix ma- attached before the third person prefix ga- on the root -diila ‘hang, hang on’, which is a class II verb. However, when ma-ga-diila ‘pursue, chase’ is inflected, the person agreement is no longer active, and the verb patterns as a class Ia verb, as shown in (138).

408

František Kratochvíl

(138) a. Nirra mura magadiila. b. Nirra ganamagadiila. nirra mura magadiila nirra ga-na-magadiila FIN .pursue 1SG.AGT 3.I-INV-FIN.pursue 1SG.AGT kid ‘I am chasing children.’ [P.178] ‘I am chasing him/her/them.’ c. Nirra anamagadiila. nirra a-na-magadiila 1SG.AGT 2SG.I-INV-FIN.pursue ‘I am chasing you.’ [P.182]

d. Girra nanamagadiila. girra na-na-magadiila 3.AGT 1SG.I-INV-FIN.pursue ‘He/she/they is/are chasing me.’ [P.180]

6.4.1 Class I verbs Class I verbs, both monovalent and bivalent, either appear bare or admit person and applicative prefixes. Class Ia admits person prefixes, indexing the human S/P, only in combination with the inverse na-. Class Ib requires an applicative prefix to be attached before the inverse na-. Examples of verbs belonging to Class I can be seen in (139). Class Ia (139) tabiila akana labaani maduro taaru alaata baali baleeta laawa luusi mi

‘wrong’ ‘black’ ‘go along’ ‘check, spy’ ‘oppose’ ‘make fall’ ‘feed’ ‘carry out’ ‘search’ ‘measure’ ‘take’

Class Ib aruko tee taani tubutili yaana yoodang ayatara kakara kupi maki/maky sumo

‘dry’ ‘paddle’ ‘fall’ ‘go around, tighten, wrap’ ‘go’ ‘stumble’ ‘pay, compensate’ ‘rinse’ ‘fold’ ‘save, store’ ‘watch’

An example of the use of the monovalent Ia class verb akana ‘black’ can be seen in (140). The inverse prefix na- is required to index the human S, but not needed with the inanimate S. (140) a. Gatakka-lakka akana. b. Nana’akana. na-na-’akana [ga-takka-lakka]S akana black 1SG.I-INV-black 3.I-head-leaf ‘His hair is black.’ ‘I am black.’

Sawila

409

The bivalent root maduro ‘check, spy on’ is contrasted in (141). The inverse na- appears when the P is human. (141) a. Namidde parra maduro. b. Dokotoora nanamaduro. [dokotoora]A na-na-maduro na-midde [parra]P maduro check doctor 1SG.I-INV-check 1SG.I-go.up field ‘I went up to check the fields.’ ‘I am examined by the doctor.’ Monovalent class Ib verbs denoting states, such as aruko ‘dry’ or bireasa ‘collapsed’ remain monovalent when they take on applicative prefixes, as in (142). (142) a. Ilu aruko. b. Karupo li’aruko. [ilu]S aruko [karupo]S li-’aruko APPL-dry river dry wound ‘The river is dry.’ ‘The wound healed, lit. dried out.’ c. Araasing bireasa. [araasing]S bireasa NFIN .house collapsed ‘The house collapsed.’ d. Araasing amba mabireasa boorana. ma-bireasa boorana [araasing am=ba]S NFIN .house DIST=TOP APPL-collapsed FIN .very ‘That house is getting ever more dilapidated.’ Monovalent class Ib denoting activities such as tee ‘paddle’ regularly increase their valency when combined with the applicative prefix, as in (143). (143) a. Tee di sipera yaang woo! tee di sipera yaang woo paddle REAL FIN.boat NFIN.go INC ‘Paddle so that the boat moves on!’ b. Kaadang matee midde. [kaadang]E ma-tee midde APPL-paddle go.up canoe ‘They paddled a canoe to the shore.’ Bivalent verbs belonging to class Ib may add an E argument, usually denoting a location, when an applicative prefix is attached.

410

František Kratochvíl

(144) a. Girra kaada luuting dana maki. girra kaada luuting dana maki 3.AGT cloth couple FIN.one save ‘She keeps some traditional clothes.’ b. Gimirea mamaki kaang-kaana. [gi-mirea]E ma-maki kaang-kaana 3.II-inside APPL-save FIN.well ‘She kept it inside her heart.’ When the E of a class Ib verb is human, the inverse na- is required. Verbs such as tataaku ‘find, meet’ display this pattern, as shown in (145). (145) a. Ang nooli tataaku? [ang]P nooli tataaku DIST where find ‘Where did you find it?’ [BDPK.299] b. Kaluno waari ineku tanamatataaku tirea! kaluno waari i-n-yaku ta-na-ma-tataaku tirea FIN .now again 1PL .INCL-together-two DISTR. I-INV-APPL-find PROH ‘From now on, let the two of us not meet again!’ [MC.33]

6.4.2 Class II verbs Class II verbs index the S/A argument. A number of verbs belonging to this class are listed in (146). Many of these verbs refer to what is known as middle situations (Kemmer 1993, 1994), including (i) grooming, (ii) body care, (iii) change in bodily posture, (iv) translational motion, (v) emotions, and (vi) cognition. Note that some of the class II verbs index the S/A argument obligatorily (marked with a hyphen), while for other verbs the indexing is optional. (146) -wile te -rammu we welaaka -diila laadi

‘bathe’ ‘lie, sleep’ ‘die’ ‘leave’ ‘look back’ ‘hang’ ‘ashamed’

kiita -pama miti -ko wiiwi pulo litinga

‘defecate’ ‘eat’ ‘sit, sit down’ ‘stay, dwell’ ‘turn away’ ‘blow’ ‘understand’

-ku nee dooka luku baabala -soorana -te

‘fart’ ‘consume, inhale’ ‘stand, stand up’ ‘bend, bow’ ‘stumble’ ‘be angry’ ‘know’

Sawila

411

Monovalent verbs belonging to class II index their S argument and admit applicative prefixes to add the E argument, as in (147). (147) a. Nawile wuno. b. Saabang wiigawile lee. wuno [saabang]E wii-gaS-wile lee naS-wile FIN . DUR soap APPL-3.I-bathe PFV 1SG.I-bathe ‘I am bathing.’ [D.Wile] ‘They washed themselves with soap.’ [BF.22] c. Oli gako. d. Koolang manako. [Koolang]E ma-naS-ko oli gaS-ko Kolana APPL-1SG. I-stay here 3.I-stay ‘He lives here.’ [D.Ko.1] ‘I stay in Kolana.’ [D.Ko.2] Monovalent verbs such as miti ‘sit’ can admit up to five prefixes and two E arguments, as can be seen in (148c), the maximum attested among Sawila verbal classes. (148) a. Amiti kaang-kaana. aS-miti kaang-kaana 2SG.I-sit FIN.well ‘Seat yourself comfortably.’ b. Adala suwi mu gawa maruna leemiti. maruna]E li-gaS-miti [adala]S suwi [mu ga-wa bird arrive tree 3.I-branch FIN.small APPL-3.I-sit ‘The bird sat on the tree branch.’ c. Sakiila wiitanamagamiti-gamiti amane. amane [sakiila]E2 wii-taE1-na-ma-gaS-miti~ga-miti APPL-DISTR.I -INV-APPL-3I-sit~VERBAL.PL DIST.like ant ‘They crowded around us as ants.’ [NB5.125] Bivalent verbs, such as the verbs of consumption nee ‘consume’ or paati ‘chew’, admit person prefixes indexing the A argument, as can be seen in (149). The meaning of paati shifts to ‘bite’ with human Ps, which require the inverse na- and pattern as class Ia. The inverse prefix does not occur with inanimate Ps and the person indexing is dropped to indicate habitual events but has to appear to refer to events that are actualized.

412

František Kratochvíl

(149) a. Tapaati. b. Katu nanapaati. [katu]A naP-na-paati taS-paati DISTR .I-eat dog 1SG.I-INV-bite ‘We are eating.’ [Paati.1] ‘The dog bit me.’ [Paati.2] c. Nidena pu gapaati. [pu]P gaA-paati [ni-dana]A 1SG.II-friend betel.nut 3.I-chew ‘My friend is chewing betel nut.’ d. Nidena pu paati. [pu]P paati [ni-dana]A 1SG.II-friend betel.nut chew ‘My friend (habitually) chews betel nut.’ Emotion and cognition verbs belonging to class II pattern like -yoopana ‘forget’ in (150). The applicative prefix li- adds an E argument, while ma- indicates the non-standard effect of the event. (150) a. Nayoopana. b. Nirra manayoopang padiki dana. [nirra]S ma-naS-yoopang padiki dana naS-yoopana 1SG.AGT APPL-1SG.I-NFIN.forget a.bit FIN.one 1SG.I-FIN.forget ‘I forgot.’ [A.113] ‘I have forgotten a bit.’ [A.116] c. Buku linayoopana. [buku]E li-naS-yoopana APPL-1SG.I -FIN .forget book ‘I forgot about the book.’ [P.135]

6.4.3 Class III verbs Sawila class III verbs obligatorily index the S/P argument and may extend their valency with applicative prefixes. A number of verbs belonging to this class, all of them bound roots requiring person prefixes marking S/P (marked with a hyphen), can be seen in (151). Many class III verbs denote events affecting the P argument. (151) -taala -ri -kinyaala -leng -miikana

‘stack up, praise’ ‘scold, argue’ ‘cheat’ ‘fight, attack’ ‘pull’

-yooka -mantaani -kuna -lila -mu

‘shake’ ‘hit, infect’ ‘bite to death’ ‘lift, pick up’ ‘help’

-mari -dea -su -loomana -punni

‘quarrel’ ‘see’ ‘order’ ‘collect, pick’ ‘hit’

Sawila

413

Class III verbs do not admit the inverse prefix na- with human Ps. Various types of P arguments of -lila ‘lift’ are shown in (152). (152) a. Nannu nalila penetua iloola yootingtua gimirea meka-ka. [nannu]P na-lila penetua iloola yootingtua gi-mirea ma-ga-ko~ga-ko 1SG.I-lift elders year eight 3.II-inside APPL-3.I-stay~VERBAL.PL 1SG ‘I was elected in the elders council for eight years.’ [NB5.238] b. Suwi arakeesa galila suwi, pitamanka. suwi [arakeesa]P ga-lila suwi pi ta-mangka 3.I-lift arrive pig DISTR.I-replace arrive FIN.rice ‘They brought rice and bartered it for pigs.’ [HM.5] c. Mootorlaut magalila tiwenti. ma-ga-lila tiwenti [mootorlaut]E boat.with.engine APPL-3.I-lift PART ‘They will load it on a boat [reportedly].’ [NB4.81] In (153), the verb -punni ‘hit’ is exemplified with a single P argument and with each of the three applicative prefixes, adding an E argument. (153) a. Giira napunni. b. Gangkari magapunni tirea! girra na-punni [gang-kari]E ma-ga-punni tirea 3.I-face APPL-3.I-hit PROH 3.AGT 1SG.I-hit ‘He hit me.’ [Punni.1] ‘Don’t hit him on his face!’ [Punni.2] c. Katu api tako ngana ligepunni! [katu api tako ngana]E li-ga-punni APPL-3.I-hit dog fish steal DEF ‘Hit the dog for stealing fish!’ [A.26] d. Girra gisi wara wiigapunni. [wara]E wii-ga-punni [girra]A [gi-si]P 3.II-body stone APPL-3.I-hit 3.AGT ‘He was hitting himself with stones.’ [D.Punni] When combined with affected human arguments, interpersonal variation can be observed with some class III verbs, as shown in (154). While some speakers prefer to use the applicative li- to add a human participant as an E argument, others use the inverse na- instead. It is yet unclear whether the inverse alternative should be considered as an instance of ‘double-indexing’ of the human P,

414

František Kratochvíl

or as a type of defective agreement (class Ia), analogous to -magadiila ‘chase’ discussed in (138). (154) a. Annu namiikana. b. Kape gate litamiikang! annu na-miikana [kape ga-te]E li-ta-miikang 1SG.I-FIN.pull rope 3.I-lie APPL-DISTR.I-NFIN.pull 2SG ‘You are pulling me.’ [SOB.3] ‘Pull on that rope!’ [M.124] c. Aning yaku talitamiikana. aning yaku ta-li-ta-miikana NFIN .person two DISTR. I-APPL-DISTR. I-FIN .pull ‘Two people pull each other.’ [C1p25o] d. Aning yaku tanetamiikang. aning yaku ta-na-ta-miikang NFIN.person two DISTR. I-INV- DISTR .I-NFIN.pull ‘Two people pull each other.’ [C1p25a] Several class III verbs, such as tatuku ‘tell’, wiita ‘call’, or tuwang ‘shout, announce’, do not index the inanimate P argument or may be used as intransitives without any pronominal indexing, as shown in (155). (155) a. Girra niwiita. b. Eya wiita na awaari. [e-ya]S wiita na a-waari [girra]A niP-wiita 1SG.II-call 2SG.II-mother call VIS 2SG.I-FIN.return 3.AGT ‘He called me.’ [D.Wiita] ‘Your mother calls you home.’ [D.Wiita] c. Erra asiing dan tatuku dapo! tatuku dapo [erra]A [asiing dang]P NFIN .one tell PRIOR 2SG.AGT fable ‘Tell me a fable!’ [D.Tatuku] d. Nepa gigawurine po, nitatuku naanu. ne-pa gi-ga-wurine po niP-tatuku naanu 1SG.II-father 3.II-3.I-dream FOC 1SG.II-tell FIN.not ‘My father didn’t tell me about his dream.’ [D.Tatuku] The verb -suwaara ‘show’ (single member of class IIIb) alternates the way the human recipient is expressed depending on its position in the noun phrase hierarchy (Silverstein 1976, 1981). While third person recipients can be expressed as Ps with set II prefixes, first and second person recipients are expressed as E arguments and require the applicative prefix li-, as in (156).

Sawila

415

(156) a. Nannu suraata mi ligisuwaara. nannu suraata mi li-giP-suwaara letter take APPL-3.II-show 1SG ‘I show him the letter.’ [D.Suwaara] b. Gannu ganamulaana gibuku mi nalisuwaara. gannu ga-na-mulaana gi-buku mi na E-li-suwaara 3 3.I-INV-FIN.not.want 3.II-book take 1SG.I-APPL-show ‘He doesn’t want to show me his book.’ [D.Suwaara] c. Nibuku mi alisuwaara nanamulaana. na-na-mulaana ni-buku mi aE-li-suwaara 1SG.II-book take 2SG.I-APPL-show 1SG.I-INV-FIN.not.want ‘I don’t want to show you my book.’ [D.Suwaara] 6.4.4 Class IV verbs Class IV verbs are monovalent, displaying great variation in indexing the S argument, subdividing the class of two dozen verbs into four subclasses. Class IVa contains three verbs: -lu ‘bark’, -waawang ‘remember’ and -welaaka ‘look back, consider’; both index the S argument with a set I prefix and admit applicative prefixes to add the E argument. No inverse prefix is required when the E argument has a human reference, as can be seen in (157b–c). (157) a. Liawaawang kaang-kaang! b. Alinawaawang. kaang-kaang aE-li-naS-waawang li-aS-waawang APPL-3.I -NFIN .remember NFIN .well 2SG.I-APPL-1SG.I-NFIN.remember ‘Remember this well!’ [D.Waawang] ‘I remember you.’ [D.Waawang] c. Nooningda niliawelaaka naanu? naanu nooning da niE-li-aS-welaaka NFIN.how REAL 1PL. EXCL-APPL-2SG.I -look.back NFIN .not ‘Why don’t you have any concern for us?’ [Mk4.37] There are about twenty IVb verbs in the Sawila corpus, many of which are bound roots (marked with a hyphen), never attested without applicative prefixes, as shown in (158). (158) -biina -diimu parisine -waaka

‘forbid’ ‘press, squeeze’ ‘chat, protest’ ‘watch, guard’

-biki -ilela pane sako

‘trust, rely on’ ‘look, browse’ ‘touch’ ‘embrace’

biseara luusi -saraana yaayi

‘talk, gossip’ ‘measure, consider’ ‘ask’ ‘ask for, request’

416

František Kratochvíl

Unlike class IVa verbs, class IVb verbs do not index the S argument but do admit applicative prefixes and index Es, as shown in (159). (159) a. Gannu du libiina. [gannu du]E li-biina PL APPL-forbid 3 ‘Those are forbidden.’ [D.Biina] b. Dokoteera nalibiina tabaaku maduli naanu. tabaaku maduli naanu [dokoteera]S naE-li-biina FIN.doctor 1SG.I-APPL-forbid tobacco smoke FIN.not ‘The doctor forbids me to smoke.’ [D.Biina] The same pattern can be observed with the verb -biki ‘trust, rely on’ which takes the applicative prefix li- to add the E argument. Only human Es are indexed, with set I prefixes. (160) a. Nayaati masi, namaddu du libiki. na-yaati masi [na-maddu du]E li-biki 1SG.I-child PL APPL-trust 1SG.I-sick if ‘If I fall ill, I will rely on my children.’ [D.Biki] b. Erra nalibiki tirea! tirea erra naE-li-biki 2SG.AGT 1SG.I-APPL-trust PROH ‘Do not rely on me! [I will not help you.]’ [D.Biki] Class IVc contains ten bound verbs, listed in (161). The subclass lacks a common semantic core, but several verbs form antonymous pairs (-kaali ‘refuse’ vs. -maro ‘agree’). (161) -kaali -madukona -masine -noosa-noosa -sirei

‘refuse, not want’ ‘hiccup’ ‘suffer hunger’ ‘be/do alone, individually’ ‘give birth’

-maro -marungki -mulaana -punara -waana

‘agree, want’ ‘care no longer’ ‘avoid, not want’ ‘menstruate’ ‘gather, do/be together’

Class IVc verbs require the inverse prefix na- if the S argument is to be indexed. Applicative prefixes can be attached before the person prefix, to add an E argument, as shown in (162).

Sawila

417

(162) a. Singko game ganakaali. Singko ga-me gaS-na-kaali Singko 3.I-come 3.I-INV-refuse ‘Singko refuses to come’ [QVP.25] b. Tatuku ako tupara gapi mananakaali. tatuku a-ko tupara ga-pi ma-naS-na-kaali talk 2SG.I-stay random 3.I-make APPL-1SG.I-INV-refuse ‘Your nonsense talk makes me even more reluctant [to listen to you].’ [A.124] Class IVd allows double indexing of the S argument (set I and II prefixes). This subclass contains only two verbs: -waawang ‘think’, and -wurine ‘dream’, exemplified in (163). (163) a. Nawaawang. naS-waawang 1SG.I-NFIN.think ‘I am thinking.’ [D.Waawang] b. Ninawaawang naanusi, annu ayaati. ni-na-waawang naanu=si annu a-yaati 2SG.I-sick 1SG.II-1SG.I-NFIN.think FIN.not=TOP 2SG ‘I don’t think you are sick.’ [D.Waawang] Sawila verbal classes are not entirely discrete but allow some fluidity: a small number of verbs display patterns fitting into several classes described above. The verb mayiina ‘hear, listen’ can pattern as a monovalent or bivalent verb of class Ib, leaving its S, A, and P arguments unindexed. (164) a. Eta mayiina? b. Yite mayiina! mayiina [yite]S mayiina [eta]S FIN .hear 2PL.POT FIN.hear 2sG.POT ‘Can you hear [it]?’ [NB4.41] ‘You listen!’ [D.Mayiina] c. Yannu luwara mayiina anta naanu? [yannu]A [luwara]P mayiina anta naanu news FIN.hear or FIN .not 2PL ‘Did you hear the news or not?’ [D.Mayiina]

418

František Kratochvíl

d. Murra pike du, asiing dana tatuku girra na limayiina. murra pike du [asiing dana tatuku girra na]E li-mayiina certainly VIS APPL-FIN.hear kid small PL fable FIN.one tell ‘Children, I will tell you a story, so listen to it.’ [D.Na] As shown in (165), -mayiina ‘hear’ also patterns as a verb from class Ia, requiring the inverse prefix na- to index its S argument and admitting applicative prefixes. (165) a. Nanamayiina. b. Linanamayiina dapa! li-na-na-mayiina dapa naS-na-mayiina APPL-1SG.I -INV-FIN.hear PRIOR 1SG.I-INV-FIN.hear ‘I hear, let me hear.’ [N12.182] ‘Let me hear about it!’ [A.129] c. Mananamayiina asire wuno. ma-na-na-mayiina a-sire wuno APPL-1SG .I- INV-FIN .hear 2SG.I -NFIN .come.down FIN .DUR ‘I have been hearing repeatedly that you were coming.’ [A.130]

7 Predicate modifying particles Sawila predicates can be modified with particles encoding aspect (7.1) and mood (7.2). A generalized template is given in (166). Any of the particles can be followed by the intensifier boorang ‘very, indeed’, which also combines with nouns and nominal modifiers (7.3). Template of predicate modifiers (166) MOOD1 VERB VERB.PL ASPECT1 ASPECT2 NEGATION MOOD2 INTENSIFIER The distribution of the mood particles is sensitive to the syntactic position of the clause in the sentence, as either medial or final, as will be illustrated in section 7.2. The verbal number (pluractionality) is discussed in Kratochvíl (2014).

7.1 Postverbal aspect particles Aspectual modification is achieved by particles, discussed in this section, and auxiliaries discussed separately in section 8.6. In both cases, the aspectual word follows the modified verb. An overview of the aspectual particles is given in Table 14.

Sawila

419

Table 14: Sawila aspectual particles FORM

GLOSS

FUNCTION

HISTORICAL SOURCe

EXAMPLES

lee kaana wuno muno mawo woo

PFV

perfective completive durative imperfective perfect inchoative

lee ‘finish’ kaang ‘good’ wunni ‘hold’ unclear unclear woo ‘hit’

(167a, d) (167b–d) (168) (169) (170) (171)

COMPL DUR IPFV PRF INC

The perfective lee immediately follows the verb and may be followed by other markers (no instances attested for PFV+INC). The completive kaana can be followed by mawo (PRF) but never by lee (PFV). Durative wuno may be followed by muno (IPFV). The perfective lee in (167a) indicates that an event is viewed as a simple whole. The completive kaana indicates that the event has been carried out thoroughly and to its completion, as in (167b–c). The perfective and completive are not in complementary distribution, but if they co-occur, the perfective has to precede the completive, as in (169d). (167) a. Gawaawang duba yarammu lee. ga-waawang du=ba ya-rammu lee PL=TOP 2PL. I-die PFV 3.I-think ‘They thought that you had already died.’ [D.12] b. Nibaadu ikeba sideari kaana. ni-baadu ike=ba sideari kaana FIN .COMPL 1SG.II-shirt red=TOP torn ‘My red shirt is torn up.’ [D.6] c. Girra tatuku sinnang kaana. girra tatuku sinnang kaana stop FIN .COMPL 3.AGT talk ‘He has stopped talking.’ [NB4.135] d. Laampuru, paada-maadana madiika lee kaana. laampuru paada-maadana madiika lee kaana PFV FIN .COMPL lamp lamp glow ‘(All) lamps and lights were lit.’ [KM.103] The durative wuno marks ongoing events and is compatible with a wide range of predicates including nominal predicates (see section 3.3).

420

František Kratochvíl

(168) a. Nawile wuno. b. Ame dapo! Name wuno. na-wile wuno a-me dapo na-me wuno 2SG.I-come PRIOR 1SG.I-come FIN.DUR 1SG.I-bathe FIN.DUR ‘I am bathing.’ [D.Wuno.1] ‘Come here!’ ‘I am already coming!’ [P.99] c. Wadi-piisa galo lapoong gapulo wuna. wadi-piisa ga-lo lapoong ga-pulo wuno sun-ascend 3.I-follow NFIN.breeze 3.I-blow FIN.DUR ‘The wind is blowing from the east.’ [D.Galo.1] The imperfective muno marks events that have not reached their final boundary, as illustrated in (169), being either existential or locative predicates. The marker cannot occur with some nominal predicates, but can be used with weather and time predicates, as discussed in section 3.3. (169) a. Ekatu nooli muno? e-katu noo=li muno 2SG.II-dog where=be.DIST FIN.IPFV ‘Where is your dog?’ [QVP.98] b. Ang ganiiri-uupuru ampo Toombangma gate muno. ang ga-niiri-uupuru am=po Toombang=ma ga-te muno DIST 3.I -grave DIST=FOC Tombang=be.PROX 3.I-lie FIN. IPFV ‘Those [men], their graves are (still) in Tombang.’ [SOS3.32] The perfect marker mawo indicates that the event ended in the past. (170) a. Nannu maninge-pike mawo. b. Gannu maninge-pike. nannu maninge-pike mawo gannu maninge-pike 1SG small-very.little PRF 3 small-very.little ‘I was very young [then].’ [NB5.214] ‘She is very young.’ [NB5.214] The inchoative particle woo occurs in encouragements (weak imperatives), such as (171a). In self-directed encouragements woo is replaced by di combined with the adverbial ka, as in (171b). (171) a. Ayiina tatuku woo! b. Ka nayaadi! a-yiina tatuku woo ka na-yaa=di INC soon 1SG.I-go.down=REAL 2SG.I-FIN.self tell ‘Tell something yourself!’ [BDPK.81] ‘I better get going!’ [C2.13]

Sawila

421

7.2 Mood Sawila predicates may be marked for mood, i.e. the difference between actualized and non-actualized situations (Mauri and Sansò 2012: 147). The Sawila mood marking is idiosyncratic, i.e. the mood is not only encoded by mood particles but also by potentive pronouns. The mood marking is organized in a non-joint and non-binary system; the occurrence of mood particles is not bound to any other grammatical category and besides realis and irrealis, other moods are distinguished (cf. Palmer 2007: 145–161). Finally, Sawila mood marking is restricted to certain clause types and constructions. The mood particle inventory is given in Table 15; the potentive pronouns are discussed in 5.2. Table 15: Sawila mood particles FORM

GLOSS

FUNCTIOn

SYNTAX

EXAMPLE

da di dapo tirea ka ara

REAL

realis, directive realis, directive prioritive, condition prohibitive future irrealis

V MOOD

(172), (174), (175) (173)–(174) (176)–(177) (49)–(50) (178) (179)

REAL PRIOR PROH IRR IRR

V MOOD V MOOD V MOOD MOOD V V MOOD AUX.ASP

Some of the markers accumulate several grammatical functions in some of their uses (cf. Palmer 2007: 150). In addition to marking mood, particles da, di, and dapa also specify the chronological order and resemble therein conjunctions: di encodes sequence, da both sequence and overlap, dapa condition. Actualized situations (realis) are marked by the particles da and di in two clause types: declaratives and directives. In declaratives, both particles are restricted to linked clauses; they occur in medial (non-final) clauses and in addition to mood also distinguish chronological sequence and overlap (cf. Palmer 2007: 152–154; Mithun 2006: 176–177). Clauses marked with da are followed by a clause denoting actualized situations (present or past; overlap or sequence), as can be seen in (172). (172) a. Nayaada gari mi made girra. na-yaa=da ga-ri mi made girra 1SG.I-go.down=REAL 3I-tell take come.up must ‘I am going down and telling them (that) they should bring it up.’ [BDPK.358]

422

František Kratochvíl

b. Nepa gasuwi wooda tatong yootine. ne-pa ga-suwi woo=da tatong yootine 1SG.II-father 3.I-arrive INC=REAL NFIN.half FIN.five ‘My father came at half past four. (lit. My father arrived and it was half past four.)’ [D.Tatong.1] c. Arammu kaanda, kaluno analuuki! kaluno [Ø]A a-na-luuki aS-rammu kaan=da NFIN.compl=REAL FIN .now 2SG.IP-INV-bury 2SG.I-die ‘You died, (and so) now [I] bury you!’ [AB.37] Medial clauses marked with di are followed by clauses denoting non-actualized situations. The chronological order of the linked clauses is that of a sequence, where the second clause describes purpose or intended future situations, as in (173). (173) a. Nirra siiking mi api suwedi, lidiirana. nirra siiking mi api suwe=di li-diirana take fish skewer=REAL APPL-FIN.grill 1SG.AGT fork ‘I skewered a fish on a bamboo fork to grill it.’ [D.Suwe.1] b. Iria tawiledi, tamidde parang luuki. iria ta-wile=di ta-midde parang luuki water DISTR.I-bathe=REAL DISTR.I-go.up dead.body bury ‘We wash ourselves and then we will go up to the funeral.’ [D.Tawile.1] Besides declaratives, particles da and di occur in directives (imperatives, hortatives, or permissives). Sawila groups directives with other realis situations, displaying a cross-linguistically less common pattern (Mauri and Sansò 2012: 151). In directives, da is restricted to the sentence-final position, but di is attested in both medial and final clauses. Medial uses may have been extended through insubordination to sentence-final uses (Evans 2007, 2009). (174) a. Erra etaama nineku ninetulongdi, alaataong patingda. niA-n-yaku [erra]A [e-taama]P 2SG.AGT 2SG.II-grandparent 1PL.EXCL-together-two niP-na-tulong=di [Ø]A [alaata=ong]P pating=da taro=PROX pull.out=REAL 1PL.EXCL-INV-NFIN.help=REAL ‘Help your grandparents (so that [we] get) to harvest the taro.’ [AL.18]

Sawila

423

b. Sipeera libaraasadi, takiida madung tirea! sipera li-baraasa=di takiida madung tirea FIN .boat APPL-balance=REAL leaning sink PROH ‘Balance the boat, it must not turn over and sink!’ [D.Baraasa.1] c. Hoi! Yamedi nalomuridi! hoi ya-me=di na-lo-muri=di hey 2PL.I-come=REAL 1SG.I-follow-run=REAL ‘Hey, come, follow me!’ [Mk1:17] The particle da has developed a discourse function in foregrounding constituents within the clause, such as question words and manner phrases, as can be seen (175). Similar discourse function of realis is also documented in Teiwa (Klamer 2012:223–227). (175) a. Beekalelle! Nooli munda ete mite? [Ø]P mi=te beekalelle nooli mung=da [ete]A INTER where NFIN.IPFV=REAL 2SG.POT take=Q ‘You fool! Where is it that you can get [your food]?’ [MC.29] b. Nooningda annu apama anakaali? nooning=da annu a-pama a-na-kaali NFIN.how=REAL 2SG 2SG.I-eat 2SG.I-INV-refuse ‘Why (is it that) you refuse to eat?’ [D.Nooning.216] c. Annu mapadikidea saakara! annu ma-padiki=da saakara APPL-tweak=REAL FIN .offended 2SG ‘You are easily offended. (lit. You are teased a bit and get offended.)’ [BDPK.160] The prioritive particle dapo/dapa marks predicates prior in chronological order to predicates that follow them. Such predicates describe conditions that have to be met before the second event can occur. Besides predicates, dapo combines with adverbials and nominals to mark time (176c) or object (176d) required for the event to be actualized. The range of use of dapo/dapa corresponds to the Alor Malay markers dulu (baru).

424

František Kratochvíl

(176) a. Gannu gawile dapa anda, nanamaro ineku iwe sikola. [gannu]S gaS-wile dapa anda naS-na-maro 1SG.I-INV-agree 3 3.I-bathe PRIOR so iS-n-yaku iS-we sikola 1PL.INCL-together-two 1PL.INCL-leave school ‘After he had a shower, I would want us to go to school together.’ [CC.87] b. Erra tamisi dapa, mu dinni. erra tamisi dapa mu dinni 2SG.AGT strong PRIOR wood lift.up ‘You have to be strong to lift a log.’ [D.Tamisi.1] c. A: Lalama ayaa liiri araasinema siribisi anta naanu? lalama a-yaa liiri araasine=ma siribisi tomorrow 2SG.I-go.down ruler FIN.house=be.PRX work anta naanu FIN.not or A: ‘Will you go work for the king tomorrow or not?’ B: Yoo, lalama dapa nayaa, kalunosi nisi mateda nate. yoo lalama dapa na-yaa kaluno=si yeah tomorrow PRIOR 1SG.I-go.down FIN.now=TOP ni-si mate=da na-te 1SG.II-body hurt=REAL 1SG.I-lie B: ‘Yes, tomorrow I will go, (but) today I am tired and will rest.’ [Q42–43] d. Iyaala-takura kaang dapa galongmuri. iyaala-takura kaang dapa ga-longmuri way-crossroad NFIN.good PRIOR 3.I-follow ‘Follow the good path.’ [D.Iyala.1] In the final clause, dapa marks situations certain to be actualized in the future and bound to occur before anything else. This use naturally extends to directives, in which dapa can be followed by the inchoative particle woo, and is therefore not in the sentence-final slot. Similarly to the Indonesian dulu, Sawila dapa can be used to fend off directives by pointing out situations with even higher priority.

Sawila

425

(177) a. Roti ali dapa woo! b. Nawe na-wata dapa. roti ali dapa woo na-we na-wata dapa 1SG.I-leave 1SG.I-urinate PRIOR bread buy PRIOR INC ‘Buy some bread!’ [D.Ali.1] ‘I am going to pee first.’ [P.11] c. Ninanamarosi, gannu gawile dapa. ni-na-na-maro=si gannu ga-wile dapa 3.I-bathe PRIOR 1SG.II-1SG.I-INV-agree=TOP 3 ‘He will bathe when I want it.’ [CC.81] The irrealis particle ka precedes the predicate and indicates unactualized future situations where there is a strong expectation that they will occur. Ka occurs in clauses describing impending events (178a–b), resolutions for the future (178c), predictions based on provided and acceptable evidence (178d), and in warnings (178e). Ka also occurs in non-factive complements of speech verbs (178f), as well as in questions targeting the future (178g). (178) a. Ninnu ka ninee-nipeati gapi. ninnu ka ni-nee=ni-paati gapi 1PL.EXCL IRR 1PL.EXCL-drink=1PL.EXCL-eat almost ‘We are about to eat.’ [QVP.27] b. Nannu kinkala ka nana’eesala. nannu kinkala ka na-na-’eesala FIN .jump IRR 1SG .I- INV-tired 1SG ‘I am almost tired of jumping.’ [P.220] c. Nannu ka waari siribisi neanu. nannu ka waari siribisi naanu IRR again work FIN.not 1SG ‘I will not work again.’ [Q.32] d. Ayaakani iduni, ka ayaa me girra. ayaakani iduni ka ayaa me girra sky dark IRR rain come certainly ‘The sky is clouded, it will certainly rain.’ [D.Ayaakani.1] e. Erra wiitaani tirea angma ka bireasa! erra wii-taani tirea ang=ma ka bireasa 2SG.AGT APPL-FIN.release PROH DIST=when IRR split ‘Don’t throw it over or it would break.’ [D.Ka.1]

426

František Kratochvíl

f. Gannu nari ngana, gayo waari ka gamade. gannu na-ri ngana ga-yo waari ka ga-made 3.I-wife again IRR 3.I-come.up 3 1SG.I-tell DEF ‘He told me that his wife will/should come.’ [Q.56] g. Gannu ka lalama gamade? gannu ka lalama ga-made IRR tomorrow 3.I-come.up 3 ‘Will he come tomorrow?’ [Q.39] The irrealis particle ara has to precede aspectual particles and auxiliaries, as shown in (179).10 It marks unactualized situations (often expected to occur in the past or present). (179) a. Ineaku tatuku ngana, erra seeng tasu ara mawo. i-ning-yaku tatuku ngana erra seeng tasu DEF 2SG.AGT NFIN.money send 1PL.INCL-together-two talk ara mawo IRR

PRF

‘We made a deal that you would send money (but you didn’t).’ [QVP.16] b. Gannu nari ngana, gasuwi ara gako. gannu na-ri ngana ga-suwi ara ga-ko 1SG.IP-tell DEF 3.IS-arrive IRR 3.IS-keep [3]A ‘He told me that he would be coming.’ [QVP.2]

7.3 The intensifier particle boorana The intensifier particle boorana ‘only, very, really’, cognate with the Wersing borong (Schapper and Hendery this volume), occurs in the sentence final position (180a), following all other post-verbal particles, including the negator naanu (180b). (180) a. Kaluno arammu, kaluno edamara napaati boorana. kaluno aS-rammu kaluno [e-damara]P naA-paati boorana FIN .now 2SG .I-die FIN.now 2SG. II-liver 1SG.I-eat FIN.very ‘Now you will die, this is really my chance to eat your heart.’ [MC.15] 10 According to some speakers, who criticize its usage as code-mixing, the irrealis marker ara is a loan from Kula spoken in the neighboring Kiralela.

Sawila

427

b. Amba seena yaniina naanu boorana. yaA-niina naanu boorana am=ba [seena]P DIST=TOP FIN .money 2PL.I -FIN .have FIN .not FIN .very ‘You people surely had no money at all.’ [BDPK.112]

8 Serial verb constructions This section deals briefly with multi-verb constructions. In terms of headedness, two types can be distinguished, namely symmetrical and asymmetrical constructions. In symmetrical constructions (8.1), none of the participating verbs can be identified as the syntactic or semantic head of the complex. The ordering is best captured by a list of procedural templates associated with particular complex event types, as has been found for other Papuan languages (see Pawley and Bulmer 2011). Asymmetrical constructions come in a great variety. In augmented predicates (8.2) one of the verbs, belonging to a closed set, realizes an additional argument. Directional serial verbs contain a motion verb describing the direction or path of the event (8.3). Resultative and manner serialization closely resemble each other (8.4–8.5). Aspectual serialization is discussed in (8.6).

8.1 Narrative serialization Narrative serial verb constructions describe complex events – individual verbs carve out a part of the macro-event and their linear ordering is in iconic relation to the course of the event. For example in (181), where the verbs gayaru ‘push it’, gamidde ‘go up’, and gate ‘lie’ describe how a ladder is pushed back to the large granary house. Granaries are built on stilts with the floor of the house at about a man’s height, so one needs a ladder to get inside. The ladder is put up each time one goes in or out and then placed back to keep mice out. The event is a conceptual equivalent of closing the door. As shown in (182), a single verb equivalent also exists. (181) Nirra adiing gayaru gamidde gate. nirra adiing ga-yaru ga-midde ga-te 1SG.AGT NFIN.ladder 3.I-push.forward 3.I-go.up 3.I-lie ‘I closed the (floor) door.’ lit. ‘I pushed the ladder up into the house.’ [DC.1]

428

František Kratochvíl

(182) Nirra adiing kooya. nirra adiing kooya 1SG.AGT NFIN.ladder push.up ‘I closed the (floor) door.’, lit. ‘I pushed the ladder up in the house.’ [N12.191] In (183), the addressee is to accompany the speaker, which is captured by two verbs – the loan verb antara ‘accompany, come with’ and the motion verb yaa ‘go down’, co-indexed for the first person plural. (183) Erra nana’antara iyea. erra na-na-’antara i-yaa 2SG.AGT 1SG.I-INV-accompany 1PL.INCL-go.down ‘You have to accompany me going down together.’ [MC.10]

8.2 Serial verb constructions adding arguments In addition to applicatives, additional arguments can be expressed by using serial verb constructions to describe complex events, such as those involving external cause, target, or transferred theme. The verb -pi ‘make’ is used to introduce external cause (184–185), -dea ‘see’ introduces experiencers, stimuli, and targets (186–187), wii ‘use’ instruments (188), and mi ‘take’ themes of transfer or locomotion (189–190). Sawila verbs, such as -rammu ‘die’ or -yaati ‘sick’ are strictly monovalent. The only way to realize the causer of the death or disease is to create a serial verb with -pi ‘make’. In (185), the phrase tapi tarammu refers to ‘being killed’, while the verb gataani describes the declaration of that punishment and takes the serial verb tapi tarammu as its argument. (184) Girra gapi nayaati. girra ga-pi na-yaati AGT 3. I -make 1 SG .I-FIN .sick 3. ‘He made me sick.’, lit. ‘He made it [that] I am sick.’ [P.116] (185) Amang lee dakalla, tapi tarammu gataani. amang lee da=kalla [ta-pi ta-rammu]P ga-taani NFIN . DIST.like PFV REAL=subsequently DISTR. I-make DISTR. I-die 3.IP-release ‘After that happened so, they sentenced us to be killed.’ [NB5.206]

Sawila

429

The experiencer of anger is co-indexed on the class IV.a verb -soorang ‘angry’. Serialization with -dea ‘see’ is needed to realize the stimulus, as in (186). (186) a. Nirra adea nasoorana. b. Girra nadea gasoorana. nirra a-dea na-soorana girra na-dea ga-soorana 3.AGT 1SG.I-FIN.see 3.I-FIN.angry 1SG.AGT 2SG.I-FIN.see 1SG.I-FIN.angry ‘I am angry with you.’ [NB4.066] ‘He is angry with me.’ [NB4.066] The verb yara ‘laugh’ also combines with ‘see’ to express the stimulus or target (187). (187) a. Erra nadi yara. b. Nirra gadi yara. erra na-di yara nirra ga-di yara 1SG.AGT 3.I-NFIN.see NFIN.laugh 2SG.AGT 1SG.I-NFIN.see laugh ‘You are laughing at me.’ [NB4.071] ‘I am laughing at him.’ [NB4.071] Instruments are expressed as arguments of the verb wii, which is also the source of one of the applicative prefixes. An example of its use can be seen in (188), where the breaking is done with bare hands – ga-tang wii. (188) Yaalar dang gatang wii mu gawingkili gapunni poo. yaalar dang ga-tang wii mu ga-wingkili ga-punni poo NFIN.female NFIN .one 3.I-arm use wood 3.I-knot 3.I-hit break ‘A woman broke a stick with her bare hands.’ [AM002.0026] In Alor-Pantar languages, reflexes of the proto-Alor-Pantar *med ‘take’ are used to realize themes (T) of transfer and locomotion (Klamer and Schapper 2012:183). The Sawila reflex mi is serialized with the verb describing the transfer/ motion, as in (189). In several Alor-Pantar languages, reflexes of *ma ‘come’ are flagging the T (Klamer and Schapper (2012:199). Sawila uses the applicative prefix wii- (< wiya~wii ‘use’) in its transfer construction. The transfer verb -rani ‘give’ displays person marking of the recipient (R) fused with the applicative wii- as wee-. In combination with other person prefixes, the fused prefix takes another shape, as can be seen in (190). Finally, the verb -rani ‘give’ can be serialized with other verbs than mi ‘take’ to indicate the benefactor (BEN), as in (191).

430

František Kratochvíl

(189) Dawang makala mi eayaa yanang weerani. [yana=ang]R wii-ga R-rani [dawang makala]T mi a-yaa NFIN.medicine FIN .bitter take 2SG .I-go.down DIST. L=DIST APPL-3.I-FIN.give ‘Deliver the poison to him down there.’ [KS.046] (190) Girra buku mi wiireni. [girra]A [buku]T mi wii-iR-rani book take APPL-1PL.INCL-FIN.give 3.AGT ‘He gave us a book.’ [A.19] (191) Gatiimu naanu girra boorang magapi wiirani. ga-tiimu naanu girra boorang ma-ga-pi wii-a BEN-rani FIN .not 3.AGT NFIN .only APPL-3.I-make APPL-2SG .I-FIN .give 3.I-pay ‘You will not pay, he will just fix it for you [for free].’ [BDPK.346]

8.3 Directional serialization The direction of a motional event may be specified by another verb. Labaani ‘pass, traverse’ indicates the trajectory and takes a nominal complement, as in (192). (192) Bangko nearaasine labaani midde. Bangko ne-’araasine labaani midde go.up Bangko 1SG.II-FIN.house pass ‘Bangko is going up along the upper side my house.’ [NB4.062] There is a set of directional verbs describing the direction of an event. The basic vertical verbs midde ‘go up’, made ‘come up’, yaa ‘go down’, and sire ‘come down’ only refer to climbing or descending of a slope at any angle in the most direct line downwards or upwards. The paths with any oblique angle are referred to with the forms glossed with ‘at an oblique angle’, which are historically compounds. The morpheme ma- in mamade and masire preserves the vowel of the original proto-Alor-Pantar *mai ‘come’ (Holton et al. 2012: 115); the modern Sawila reflex is me. The morphemes made ‘come up’ and midde ‘go up’ likely originate in the proto-Alor-Pantar *mid ‘climb, go up’ (Holton et al. 2012: 115). The entire system is shown in Figure 1, including the spatial deictics (see also section 3.2). The verb midde ‘go up’ indicates upward direction (193), the verb yaa ‘go down’ is used to indicate downward direction (194). Other verbs such as me

Figure 1: Sawila directional verbs and spatial deictics

Sawila

431

432

František Kratochvíl

‘come’ and we ‘leave’ are also used. In (193), the verb mada ‘climb’ may refer to motion in any direction but is specified by midde ‘go up’ as upwards. (193) Amada midde nitiing lamiti. li-a-miti [a-mada1 midde2]motion ni-tiing 1SG.II-back APPL-2SG.I-sit 2SG.I-climb go.up ‘Climb up and sit on my back!’ [NB4.98] Example (194), which comes from a fable, describes a son putting his father in the grave. The verb taani ‘release’ is specified for direction by yaa ‘go down’ (as V2) and by dooka ‘stand’ (as V3) for resulting position. All three verbs together translate as ‘bury’. (194) Gabokela luuku mataani yaa dooka. yaa2 dooka3 ga-bokela luuku ma-taani1 3.I-embrace grave APPL-release go.down stand ‘He embraced him and released him (standing) down in the grave.’ [NB4.117] Some combinations are highly conventionalized, such as (195), where the directional sire ‘come down’ has lost some of its spatial meaning and is conventionalized for the modern concept of ‘sending’, usually of goods from East Alor ‘down’ to Kalabahi. (195) Aning tasusire kaana? aning ta-su=sire kaana NFIN .person DISTR. I-order=NFIN.come.down FIN.COMPL ‘Did they send it yet?’ [NB5.100]

8.4 Manner serialization Serialization expressing manner can be seen in (196). When the manner verb occurs before the lexical verb, it refers to participant-oriented manner, as in (196a– c). Manner verbs located after the lexical verb are event-oriented (196d). (196) a. Masing leekala yaana niko. [masing li-ga-kala]manner yaana ni-ko NFIN .food APPL-3.I-hold FIN .go 1PL. EXCL-stay ‘Resisting the hunger we kept going.’ [NB4.72]

Sawila

433

b. Maaka itiina anoosa anee! maaka itiina [a-noosa]manner a-nee! 2SG.I-consume banana ripe 2SG.I-alone ‘Eat the banana yourself!’ [MT.63] c. Annu mapadiki masaakara. annu [ma-padiki]manner ma-saakara APPL-tweak APPL-FIN.offended 2SG ‘You are easily offended.’ lit. ‘You are offended from being tweaked.’ [NB4.67] d. Girra pata ma maana padoota boorana. girra pata ma maana [padoota boorana]manner 3.AGT corn edible FIN.cook long FIN .very ‘She is cooking the food for very long time.’ [NB4.1] Manner, reason, and causes may be expressed by separate clauses requiring the lexical verb to take on an applicative prefix, as in (197). (197) a. Nimuku manimiti. [ni-muku]manner ma-ni-miti 1PL.EXCL-stupid APPL-1PL.EXCL-sit ‘We were sitting unconcerned.’, lit. ‘We were sitting stupidly.’ [N12.201] b. Tukku nanee linawaala. [tukku na-nee]manner li-na-waala palm.wine 1SG.I-consume APPL-1SG.I-drunk ‘I got drunk drinking palm wine.’ [N12.24]

8.5 Resultative serialization Serial verb constructions are frequently used to encode resultatives. It is always the V2 that describes the resulting state. For person-marking verbs, co-indexed participants are marked with subscript “i” in the gloss line. (198) Girra napunni mate boorana. girra na-punni [mate boorana]result 3.AGT 1SG.I-hit hurt FIN .very ‘He hit me [to the extent] that it hurts.’ [P.115]

434

František Kratochvíl

(199) Gasoorana gakiina gapi dooka. ga i-pi dooka]result ga-soorana [ga-kiinai 3.I-FIN.angry 3.I-FIN.breath 3.I-make stand ‘They gasped in rage.’, lit. ‘they were [so] angry [it] made their breath stand.’ [NB5.115] Resultatives may be also encoded with a set of posture verbs, which refer to the end result. The most commonly used verbs are dooka ‘stand’ (200), miti ‘sit’ (201), and -te ‘lie’ (202). Each of the verbs may be co-indexed for the participant assuming the resulting state. (200) Oningba Koolang mara mang idooka. oning=ba Koolang mara mang [iA-dooka]result NFIN. PROX.be.like=TOP Kolana be.in NFIN.arrive 1PL.INCL-stand ‘Around this time [of the day] we arrived in Kolana.’ [BDPK.161] (201) Nirra wara taaba gamiti. nirra warai taaba [ga i-miti]result 1SG.AGT stone pile.up 3.I-sit ‘I have piled up stones.’ [NB4.78] (202) Seto yaati, wiikipo sire gate. [ga-te]result se=to yaati wiikii=po sire swim=too FIN.bad flood=FOC NFIN.come.down 3.Ii-lie ‘He could not swim, a flood inundated [the river bed].’ [MT.12] The verb kasa ‘come off’ serializes with a small set of verbs to denote events that brought about some permanent change and where no further continuation is possible. Besides the verb tuku ‘peck’, kasa occurs in the corpus with the verbs pareesa ‘investigate < Mly. periksa’, li-kiipa ‘block out, lock up’, ga-pi garammu ‘slaughter off’, ma-ga-ko ‘settle in, occupy’, mading ‘plant out (the whole field)’, pake ‘use up < Mly. pakai’, and li-ga-diika ‘shave off’. (203) Balisi angba se girra tuku kasa. girra]A tuku [kasa]result [balisi ang=ba]P [se packet DIST=TOP chicken 3.AGT peck come.off ‘That (steamed rice) package, the chicken pecked (it) completely.’ [NB4.165]

Sawila

435

8.6 Aspectual serialization Two Sawila posture verbs have grammaticalized into aspectual auxiliaries placed after the lexical verb. Ongoing events are marked with -ko ‘stay’, which is co-indexed for the A/S argument of the main predicate, as in (204). The form ga-ko retains its literal meaning of ‘stay in a place, live, dwell’. (204) a. Nannu buku limei nako. b. Gannu buku limei gako. gannui buku limei ga i-ko nannui buku limei na i-ko book read 1SG.I-stay 3 book read 3.I-stay 1SG ‘I am reading a book.’ [N12.203] ‘He is reading a book.’ [Q.50] Momentary events, limited to a short time spam, can be marked with dooka ‘stand’. The verb dooka occurs with its original meaning ‘stand’ as well. The prospective auxiliary ga-pi marks the impending occurrence of an event. (205) a. Gannu daamata laka dooka. gannu daamata laka dooka FIN .door open stand 3 ‘He is opening the door.’ lit. ‘He stands opening the door.’ [Q.52] b. Gannu yaang gapi dooka. gannu yaang ga-pi dooka NFIN.go 3.I-make stand 3 ‘He is about to leave.’ lit. ‘He stands making-it going.’ [Q.48]

9 Discussion Sawila is a language with an interesting phonological system where the interaction between stress and length requires further study. The language has innovated a system of final and non-final roots, whose distribution depends on their syntactic position. The system is somewhat similar to that of Kaera (Klamer, this volume), Wersing (Schapper & Hendery, this volume) and Kula (Williams, p.c.). Their relationship with the echo vowel systems reported in various languages of Timor is not presently understood, but appears to be an areal feature of East Timor, Kisar, and East Alor.11 11 Echo vowels have been reported in several languages of the neighboring islands of Timor and Kisar: Fataluku (Engelenhoven 2009: 240–241), Fataluku and Waima’a (Himmelmann 2012), Makalero (Huber 2011: 80–86), and Meher (Samloy 1998: 10). The insertion of echo vowel

436

František Kratochvíl

In Sawila noun phrases, the head noun occurs initially and may be followed by its modifiers. Demonstratives occur in the final position. Sawila seems to have only internally headed relative clauses, allowing only S and A as shared arguments. Sawila has three pronominal paradigms, and two series of person prefixes indexing both possessors and arguments, related in a transparent manner. Two of the pronominal paradigms mark agents, while the third is not restricted to any argument type. The unrestricted paradigm can be used in pronominal doubling. The person prefixes distinguish roughly inalienably possessed nouns from alienably possessed, although many relational nouns are marked as inalienable. The paradigm shows some influence from the contact with Wersing (cf. Schapper & Hendery, this volume). The combinatorics of verbal prefixes (person, applicative, and inverse prefixes) is taken here to define four inflectional subclasses. Only two out of four classes share a semantic core. Class II contains verbs that refer to what is known as middle situations (Kemmer 1993, 1994), while class III contains verbs that combine with ‘affected’ arguments. Similar to Kula, Sawila has an inverse prefix used to mark the presence of an affected human undergoer (both S and P). Sawila predicates are marked for aspect and mood and may join into serial verb constructions to (i) describe complex events, (ii) add arguments, (iii) specify direction, (iv) manner, (v) result, or (vi) aspect.

References Anderson, Stephen R., Lea Brown, Alice Gaby & Jacqueline Lecarme. 2006. Life on the Edge: There’s Morphology There After All! Lingue e linguaggio 1: 33–48. Balle, Misriani, & Stuart Cameron. 2012. The interplay of quantifiers and number in Helong grammar. Presented at the 12th International Conference on Austronesian Linguistics (12 ICAL) in Denpasar. Brown, Lea. 1997. Nominal mutation in Nias. In Cecilia Odé & Wim Stokhof (eds.), Proceedings of the 7th International Conference on Austronesian Linguistics, 95–413. Amsterdam/ Atlanta: Rodopi. Dixon, Robert M. W. 2012. Basic linguistic theory: Further grammatical topics. Vol. 3. Oxford: Oxford University Press. is treated as a phonological process producing the preferred open syllables. In Makalero, echo vowels are added to closed syllables in the final position of a phonological phrase and the addition of the echo vowel is generally followed by a compensatory lengthening of the preceding vowel (Huber 2011:83–86). The echo-vowel insertion may be historically related to the emergence of metathesis found in West Timorese languages such as Helong (Balle and Cameron 2012), or in the Uab Meto dialect complex (Grimes et al. 2012).

Sawila

437

Donohue, Mark. 1997a. Developments of affectedness marking in the languages of Alor (split, splat). Talk presented at the University of California at Berkeley. Berkeley, California. 25 September 1997. Donohue, Mark. 1997b. Inverse in Tanglapui. Language and linguistics in Melanesia: journal of the Linguisitic Society of Papua New Guinea 27: 101–118. Donohue, Mark & Søren Wichmann. 2008. The typology of semantic alignment. Oxford: Oxford University Press. Engelenhoven, Aone van. 2009. On derivational processes in Fataluku, a non-Austronesian language in East-Timor. In Leo Wetzels (ed.), The linguistics of endangered languages: contributions to morphology and morphosyntax, 333–362. Utrecht: LOT. Evans, Nicholas. 2007. Insubordination and its uses. In Irina Nikolaeva (ed.), Finiteness: Theoretical and empirical foundations, 366–431. Oxford: Oxford University Press. Evans, Nicholas. 2009. Insubordination and the grammaticalisation of interactive presuppositions. Paper presented at the Methodologies in determining morphosyntactic change conference, Museum of Ethnography, Osaka, March 2009. Grimes, Charles, Heronimus Bani & Agustinho Caet. 2012. Quantifying things in the grammars of Amarasi (Eastern Indonesia) and Baikeno (Timor-Leste). Presented at the 12th International Conference on Austronesian Linguistics, Denpasar. Himmelmann, Nikolaus. 2012. Prosody in language description: Taking spoken language seriously. Paper presented at the International Workshop Current Trends of Linguistic Research of Indigenous Languages in Indonesia. Research Institute for Languages and Cultures of Asia and Africa. Tokyo University of Foreign Studies. Tokyo, Japan. 17–18 February 2012. Huber, Juliette. 2011. A grammar of Makalero: a Papuan language of East Timor. Utrecht: LOT. Iosad, Pavel. 2010. Right at the left edge: initial consonant mutations in the languages of the world. In Jan Wohlgemuth & Michael Cysouw (eds.), Rethinking universals: how rarities affect linguistic theory, 105–137. Berlin/New York: Walter de Gruyter. Kemmer, Suzanne. 1993. The middle voice (Typological Studies in Language 23). Amsterdam/ Philadelphia: John Benjamins. Kemmer, Suzanne. 1994. Middle voice, transitivity, and the elaboration of events. In Barbara A. Fox & Paul J. Hopper (eds.), Voice: Form and function (Typological Studies in Language 27), 179–230. Amsterdam/Philadelphia: John Benjamins. Kittilä, Seppo. 2008. Animacy effects on differential goal marking. Linguistic Typology 12(2). 245–268. Klamer, Marian. 2012. Reality status in Teiwa (Papuan). Language Sciences 34. 216–228. Klamer, Marian & Antoinette Schapper. 2012. ‘Give’ constructions in the Papuan languages of Timor-Alor-Pantar. Linguistic Discovery 10(3). 174–207. Kratochvíl, František. 2011. Discourse-structuring functions of Abui demonstratives. In Foong Ha Yap, Karen Grunow-Hårsta & Janick Wrona (eds.), Nominalization in Asian languages: Diachronic and typological perspectives, 757–788. Amsterdam/Philadelphia: John Benjamins. Kratochvíl, František. 2014. Number in Abui and Sawila. In Marian Klamer & František Kratochvíl (eds.), Number and Quantity in East Nusantara, 127–155. Canberra: Asia-Pacific Linguistics. Kratochvíl, František, Sebastian Fedden, Gary Holton, Marian Klamer, Laura C. Robinson & Antoinette Schapper. 2011. Pronominal systems in Alor-Pantar languages: Development and diachronic stability. Paper read at the 20th International Conference on Historical Linguistics, Osaka, Japan, 28 July, 2011. Mauri, Caterina & Andrea Sansò. 2012. The reality status of directives and its coding across languages. Language Sciences 34. 147–170.

438

František Kratochvíl

Mithun, Marianne. 2006. The languages of native North America. Cambridge: Cambridge University Press. Noonan, Michael. 1992. A grammar of Lango. Berlin: Mouton de Gruyter. Palmer, Frank R. 2007. Mood and modality. Cambridge: Cambridge University Press. Pawley, Andrew & Ralph Bulmer. 2011. A dictionary of Kalam with ethnographic notes. Canberra, A.C.T.: Pacific Linguistics, School of Culture, History and Language, College of Asia and the Pacific, The Australian National University. Ross, Malcolm. 2005. Pronouns as a preliminary diagnostics for grouping Papuan languages. In Andrew Pawley & R. Attenborough (eds.), Papuan pasts: Investigations into the cultural, linguistic and biological history of the Papuan speaking peoples, vol. 30, 15–66. Canberra: Pacific Linguistics. Samloy, Semuel. 1998. Struktur bahasa Kisar. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa, Departemen Pendidikan dan Kebudayaan. Silverstein, Michael. 1976. Hierarchy of features and ergativity. In Robert M.W. Dixon (ed.), Grammatical categories in Australian languages, 112–171. Canberra: Australian Institute of Aboriginal Studies. Silverstein, Michael. 1981. Case marking and the nature of language. Australian Journal of Linguistics 1. 227–244. Schapper, Antoinette & Lila San Roque. 2011. Demonstratives and non-embedded nominalisations in three Papuan languages of the Timor-Alor-Pantar family. Studies in Language 35(2). 380–408.

Antoinette Schapper & Rachel Hendery

8 Wersing 1

The language scene

440

2 2.1 2.2 2.2.1 2.2.2 2.3 2.4 2.4.1 2.4.2 2.5 2.5.1 2.5.2 2.6

Phonology 442 442 Vowels 443 Consonants Consonant phoneme minimal pairs 445 Consonant allophony 446 Stress 447 Phonotactics 449 Epenthetic vowels 450 Echo vowels 450 Morphophonemics 450 Prefixation 452 Metathesis 452 Orthography

444

3 Basic clausal syntax 453 453 3.1 Predicate types 453 3.1.1 Verbal predicates 455 3.1.2 Nominal predicates 457 3.1.3 Elevational predicates 3.1.4 Predications expressing location and existence 458 3.1.4.1 Locative verb ming 3.1.4.2 Posture verbs and the expression of existence 460 3.2 Negation 461 3.3 Questions 462 3.4 Imperatives and related clause types 4 4.1 4.2 4.2.1 4.2.2 4.3 4.4

Noun phrases 464 464 Attributes and relative clauses 466 Quantification 466 Numeral quantifiers 468 Non-numeral quantifiers 469 Determiners 470 Possession

5 5.1 5.2 5.3

Pronouns 471 473 Third person A pronoun Agentive and oblique pronouns 476 Pragmatic pronouns

474

458 459

440

Antoinette Schapper & Rachel Hendery

Quantificational pronouns 477 479 Possessive pronouns

5.4 5.5 6 6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.3 6.4

Agreement and applicative prefixes 479 481 Agreement slot 1 481 Agreement on intransitive verbs 483 Agreement on transitive verbs 485 Applicative prefixes 486 le487 mi488 Agreement slot 2 Agreement slot 3: Inceptive prefixation

7 7.1 7.1.1 7.1.2 7.2 7.2.1 7.2.2 7.2 7.3 7.4 7.5

Serial verb constructions 490 490 Instrumental serialization 491 ong ‘use’ 492 medi ‘take’ 492 Aspectual serialization 492 Continuative -ko 493 Completive & quantitative lai 494 Manner serialization 495 Motion serialization 495 Causative serialization 496 Resultative serialization

8 8.1 8.2 8.2.1 8.2.2 8.2.3

Realis and aspect marking 497 497 Realis suffix -a 499 Aspectual markers 499 de ‘IPFV’ 500 kana ‘PFV’ 501 ge ‘PROSP’

9

Discussion

References

488

501

502

1 The language scene1 Wersing (ISO 639-3 code: kvw) differs from most other languages discussed in this book in that it is not found in one single contiguous settlement area, but 1 We would like to thank Hans Retebana, Benny Delpada and Emilie T.B. Wellfelt for their invaluable assistance. In compiling this sketch, Hans proved to be the most excellent language teacher we could have hoped for. Benny was a brilliant deputy and people organizer.

Wersing

441

rather is scattered amongst other languages along the north, south and eastern coasts of central-eastern Alor. On the north coast, there are two distinct settlements: one centered on the village of Adagai in the west and the other around Taremana in the east. These villages are in close contact with Abui and Kamang speakers respectively and villagers typically have some competency in those languages. On the east coast, the Wersing settlement is centered on the village Kolana, containing a mix of Wersing and Sawila speakers. On the south coast, Wersing villages neighbor Sawila-, Kula- and Manet-speaking villages which are typically situated upland from the Wersing ones. According to local traditions, the settlements in Pureman on the south coast constitute the original Wersing settlement, with the settlement on the east coast having occurred later and from there spreading further onto the north coast. There are no reliable sources of how many Wersing speakers there are. The most recent edition of the Ethnologue (Lewis et al. 2013) estimates there to be around 4,000 speakers. In the villages, Wersing appears to be a vibrant language, with children acquiring it in places such as Kolana and Pureman (Emilie Wellfelt pers. comm.). Wersing is part of the low-level East Alor sub-group, within which it groups by itself, with Montane-East-Alor containing its neighbours, Kula and Sawila, as a sister node (cf. Schapper & Klamer forthcoming, Holton et al. 2012: 114). The Wersing lexicon also exhibits influence from Austronesian languages of north coast Timor, an area with which the Wersing people have customary relations (Wellfelt & Schapper 2013). Some examples of Wersing borrowings from Timor are sirwis ‘work’ < Tetun serewisu; aluli ‘sacred anchor’ < Tokodede luli ‘sacred’; baniak ‘guest’ < Tetun bainaka ‘guest’, etc. Previous work on Wersing has been limited. We have identified the following work: Mark Donohue conducted several days of fieldwork in Kolana in the 1990s; this resulted in a manuscript of a phonological sketch on the Wersing dialect of that place (Donohue 1997) as well as scattered examples in published works (e.g., Donohue 2004: 226 on metathesis, 2008: 66–67 on alignment). A Wersing reader was produced in the mid-2000s by Fredik Langko, a native speaker of Kolana Wersing and retired primary school teacher, in order to support local content curriculum (muatan lokal) teaching (Langko n.d.). In 2010 Gary Holton collected two 400-item wordlists in Kalabahi, the capital of Alor, with Wersing speakers of the Kolana dialect. Recently we also obtained a Wersing version of the gospel of Mark translated by Anderias Malaikosa (n.d.).2

2 Anderias passed away suddenly in 2011. He worked as a bible translator for the GMIT-Unit Bahasa & Budaya in Kupang, but had only completed the one gospel. He was not only a native speaker of Wersing and Sawila, but also a gifted linguist.

442

Antoinette Schapper & Rachel Hendery

The sketch presented here is based primarily on two weeks of fieldwork conducted by the authors in 2012 with a single speaker of the Pureman dialect, Hans Retebana, originally from the village of Peitoko on Alor’s south-eastern coast within the desa of Pureman. The collected corpus contains (i) several hundred elicitated utterances, (ii) a lexicon of ~1200 items, (iii) responses to five MPI stimuli tasks, and (iv) seven narrative texts. A further two days of elicitation were conducted by Schapper with Retebana in Kalabahi, the capital of Alor, in July 2013. Other existing materials from other dialects (as outlined above) have only been used in a very limited way in compiling this sketch. In the few instances, where we refer to the Mark gospel, these are marked with [MRK] followed by chapter and verse number.

2 Phonology 2.1 Vowels Wersing has a five vowel system, as set out in Table 1. Unlike most other languages of central-eastern Alor, Wersing has no phonemic vowel length contrast. Table 1: Wersing vowel phonemes

HIGH MID LOW

FRONT

BACK

i ε a

u ɔ

Minimal pairs and sets illustrating the contrastiveness of the five vowel phonemes are given in (1). (1)

/ba/ ‘DEF’ /-kui/ ‘fart’ /ta/ ‘pound’

≠ ≠ ≠

/bɔ/ ‘EMPH’ /kɔi/ ‘basket’ /tε/ ‘CONJ’ ≠ /ti/ ‘lie down’ ≠ /tu/ ‘three’

All vowels appear in all positions, i.e., initially, medially and finally. Initially /ɔ/ is often realized with a preceding glide onset, thus /ɔŋ/ ‘use’ and /ɔta/ ‘coconut’ can be realized [ɔŋ ~ wɔŋ] and [ɔta ~ wɔta] respectively.

Wersing

443

We have observed one instance of a sequence of like vowels in /laadu/ ‘lean’. In this, the vowels may be realized either distinctly with an intervening epenthetic glottal stop or as a single long vowel [laʔadu ~ la:du]. It remains to be determined how many other instances of this are present in the language, but they are likely to be few. Both raising and lowering processes optionally take place word-finally. The vowel /a/ can be realized as [e], for example /'nakar/ ‘previously’ can be pronounced ['nakara], i.e. with an echo vowel, or ['nakare]. The vowel /ɔ/ can be raised to [u], as in the emphatic marker /bɔ/ [bɔ ~ bu]. And the vowel /ε/ can be raised to [i], for example the prospective marker ge is sometimes realized as [gi]. Final high vowels following a consonant can also be dropped. For instance, in the corpus we have /j-'wεri/ ‘2PL-ear’ realized as [jǝ'wεri ~ jǝ'wεr] and /'tnanasu/ ‘1PL.INCL-each’ realized as [ta'nanasu ~ ta'nanas]. In sequences of two vowels where the second vowel is a high vowel, /i/ or /u/, these are realized as off-glides, [j] and [w] respectively. Examples are given in (1).

(1)

Phonetic diphthongs /ai/ ‘fish sp.’ [aj] /gεi/ ‘later’ [gεj] /kɔi/ ‘basket’ [kɔj] /uin/ ‘capture’ [ujŋ] [nawŋ] /naun/ ‘NEG’ /kou/ ‘pond’ [kow]

2.2 Consonants Wersing has 13 consonant phonemes, as listed in Table 2. Of these, /ɲ/, /r/ and /j/ are marginal phonemes, with many fewer appearances in the lexicon than other consonant phonemes. Orthographic representations of phonemes that differ from IPA symbols are given in brackets.

444

Antoinette Schapper & Rachel Hendery

Table 2: Wersing consonant phonemes ALVEOLAR

BILABIAL PLOSIVE FRICATIVE NASAL APPROXIMANT TRILL LATERAL

p

b

t s

PALATAL

d

m w

VELAR k

g

ɲ j

n r l

2.2.1 Consonant phoneme minimal pairs The plosive phonemes are /p b t d k g/. Minimal pairs and near-minimal pairs that illustrate the phonemic voicing contrasts are given in (2): (2)

/pa/ /dε/ /'gεta/

‘again’, ‘beneath’ ‘IPFV’ ‘3.AGT’

≠ ≠ ≠

/ba/ /=tε/ /'kεtε/

‘DEF’ ‘CONJ’ ‘thunder’

Some minimal pairs showing contrasts between the plosives and other modes of articulation are given in (3). (3)

/-bir/ /ti/ /-dai/

‘mouth’ ‘lie down’ ‘come up’

≠ ≠ ≠

/wir/ /=si/ /-nai/

‘candlenut’ ‘CONJ’ ‘eat’ ≠ /lai/ ‘finished’

The phonemes /d/ and /r/ are almost completely in complementary distribution. For the most part /d/ only occurs in onsets, and when it is in clusters it can only be the initial element, while /r/ for the most part only occurs in codas or as the second element of onset clusters. However, we analyze these as two separate phonemes, because there are exceptions that mean that there is overlap in the distribution of the phones. In particular, there are many instances of medial onsets with /r/ rather than the expected /d/. In this environment we find several near-minimal pairs for /d/ and /r/, as in (4). (4)

/a'rak/ /'kari/ /a'rudi lemi/ /a'mira/

‘rice’ ‘basket’ ‘woman’s brother’ ‘milk’

≠ ≠ ≠ ≠

/a'dai/ /'ladi/ /a'dur/ /a'kida/

‘tens’ ‘lose’ ‘fireplace ash’ ‘there’

Wersing

445

The lateral /l/ contrasts with both /d/ initially and /r/ medially, as can be seen in (5). (5)

/-dai/ /'sεri/

‘come up’ ‘bone’

≠ ≠

/lai/ /'sεli/

‘finished’ ‘float’

The nasals /m/ and /n/ contrast with each other both in onset and coda positions. These two also contrast only initially with the palatal nasal, as this phoneme is restricted to the beginning of 1st person plural person forms (agreement prefixes and pronouns).3 Examples of the nasal phoneme contrasts are given in (6). (6)

/-nai/ /mam/

‘eat’ ‘chew’

≠ ≠

/mai/ /-man/

‘come’ ≠ /ɲai/ ‘1PL.EXCL.TOP’ ‘cheek’

The phoneme /w/ occurs before all vowels and also (phonemically) before some consonants, for example in /wdi/ ‘sun’, realized as [widi] or [wədi] (cf. the discussion in section 2.4.1 below).

2.2.2 Consonant allophony Conditioned allophony occurs in the circumstances shown in (7).4 (7) /b/ > [b] ~ [ß] / V_V > [b] / elsewhere /g/ > [g] ~ [γ] / V_V > /d/ > > /r/ > >

[g] / elsewhere [d] ~ [ɾ] / V_V [d] / elsewhere [r] ~ [ɾ] / V_V [r] / elsewhere

e.g., /o'ba/ ‘PROX.DEM’ [ʔɔ'ßa ~ ʔɔ'ba] e.g., /bɔ/ ‘EMPH’ [bɔ ~ bu] e.g., /lε-g-mit/ ‘APPL-3-sit’ [lεgi'mit ~ lεγi'mit ~ lεgə'mit ~ lεγi'mit] e.g., /g-tεn/ ‘3-come.up’ [gə'tεŋ] e.g., /wɔd/ ‘pound’ ['wɔdɔ ~ 'wɔɾɔ] e.g., /gdir/ ‘3-see’ [g'dir] e.g., /g-'kurεn/ ‘3-bottom’ [gə'kurεŋ ~ gə'kuɾεŋ] e.g., /'wɔrkdεd/ ‘frog’ ['wɔrkεdεdε]

3 This phoneme seems likely to have developed out of the pTAP 1st person plural exclusive prefix *ni- (Holton et al. 2012). There is one exceptional appearance of /ɲ/ in one corpus: /mˈleɲa/ ‘yesterday’. 4 Voiceless plosives are occasionally aspirated word-initially, but this is neither phonemic nor phonetically conditioned.

446

Antoinette Schapper & Rachel Hendery

/n/ > > > /s/ >

[ŋ] / in codas [ŋ] ~ [n] / _[+alveolar] [n] / elsewhere [s] ~ [ʃ] / _i#

> [s] / elsewhere

e.g., e.g., e.g., e.g.,

/'asin/ ‘story’ ['asiŋ]5 /'dεintε/ ‘where’ ['dεiŋtε ~ 'dεintε] /nɔ/ ‘one’ [nɔ] /n-a'malɔsi/ ‘1SG-not.want’ [na'malɔsi ~ na'malɔʃ] e.g., /sob/ ‘house’ [sɔb]

Allophones between vowels include sequences of VCV across word boundaries, as shown in (8). (8)

[asina γai lara] /asin=a gai lar-a/ story=SPEC 3.A tell-REAL ‘They tell a story.’

We also observed nasal assimilation of n > m before labials across word boundaries. For instance, e.g., /tεn bɔ/ ‘first EMPH’ can be realized as [tεm bu] or [tεŋ bu]. Note that the nasal allophony rules are ordered. That is, nasal assimilation instead of the usual word-final velarization occurs across word boundaries when the following word begins with a labial.

2.3 Stress Stress is realized as a combination of volume, pitch and length, but does not affect vowel quality. At this stage, we are unable to describe the stress system in general, as there seem to be complex interactions between word-level, phraselevel and sentence-level stress patterns. Some non-exhaustive examples of Wersing stress patterns are given in Table 3, showing that stress can occur on a variety of syllables.

5 We have two exceptions to these nasal allophony rules in our database: (i) kengel /keŋal/ ‘jump’ violates the rule that sees velar nasals only in codas. This exception appears to be the result of a reduction of an historical cluster of /n/ and /k/; in Sawila we find kinkala ‘jump’, suggesting a proto-East Alor form *kenkal for ‘jump’; (ii) anta /'anta/ ‘or’ violates the rule that the velar nasal should be found in codas. The form ['anta] is itself not impossible given the rule that sees velar nasals assimilate to the place of a following alveolar consonant. However, we would expect here an underlying velar nasal (i.e., */'aŋta/) and thus the possibility of a velar nasal realisation in this environment (i.e., *['aŋta]) which we do not find.

Wersing

447

Table 3: Examples of Wersing stress

DISYLLABIC TRISYLLABIC

QUADRISYLLABIC

Syllable

Examples

1st 2nd 1st 2nd 3rd 3rd

/'naida/ ‘1SG.TOP’ ['najda], /'lailɔl/ ‘walk’ ['lajlɔl] /pe'kɔ/ ‘gecko’ [pɛ'kɔ], /a'nin/ ‘person’ [a'niŋ] /'malkori/ ‘monkey’ ['malkɔri] /a'kɔlin/ ‘dolphin’ [a'kɔliŋ] /ada'lun/ ‘bird sp.’ [ada'luŋ], /wɔrkɔ'mu/ ‘bird sp.’ [wɔrkɔ'mu] /sεma'rεki/ ‘seagull’ [sεma'rεki]

Stress only falls on roots. Prefixes are unstressed. In some cases, for example the agreement prefixes, the vowel is only an epenthetic schwa or echo vowel, and these are never stressed. So, for example, we find the forms in (9): (9)

/j-'wεri/ /t-'nanasu/ /g-mai/

‘2PL-ear’ ‘1PL.INCL-each’ ‘3-come’

[jǝ'wεri ~ jε'wεri] [tǝ'nanasu ~ ta'nanasu] [ga'mai ~ gǝ'mai]

The avoidance of prefixal stress is not simply because of the underlying lack of vowel in prefixes. For instance, the applicative prefix le-, which does have an underlying vowel, is also always unstressed, as in (10). (10)

/na-lε-tɔdɔra/ /lε-wakar/

‘1SG-APPL-aslant’ ‘APPL-climb’

[nalε'tɔdɔɾa] [lε'wakar]

Despite the unpredictable nature of stress in Wersing, the functional load it carries appears to be minimal; we have identified no stress-differentiated minimal pairs.

2.4 Phonotactics Monosyllabic words have maximally the structure CCVVC. Disyllabic words have maximally the structure CCVC.CVC. The observed syllable types are set out in Table 4.

448

Antoinette Schapper & Rachel Hendery

Table 4: Attested Wersing word shapes Syllable no. Monosyllabic

Disyllabic

Shape

Examples

V(V)(C)

/a/ [a] ‘oh’, /ai/ [aj] ‘fish sp.’, /uin/ [ujŋ] ‘capture’

CV(C)

/nɔ/ [nɔ] ‘one’, /tai/ [taj] ‘1PL.INCL’, /gεi/ [gεj] ‘later’, /kɔi/ [kɔj] ‘basket’, /naun/ [nawŋ] ‘NEG’

CCV(C)

/slain/ [slajŋ] ‘muslim’, /blɔin/ [blɔjŋ] ‘write’, /brain/ [brajŋ] ‘brave’

V.CV(C)

/o-'ba/ [ɔ.'ba] ‘PROX-DEM’, /a'sin/ [a.'siŋ] ‘story’, /i.'nεm/ [i.nεm] ‘shell’

CV.CV(C)

/'mεdi/ ['mε.di] ‘take’, /'lɔmu/ ['lɔ.mu] ‘talk’, /galun/ [ga.'luŋ] ‘wave’, /di.'nεl/ [di.nεl], ‘window’

CCV.CV(C)

/'brukin/ ['bru.kiŋ] ‘machine’ /'plita/ [pli.ta] ‘sharp-REAL’

(C)VC.CV(C)

/'lokman/ ['lok.maŋ] ‘head’, /'namsin/ ['nam.siŋ] ‘1SGhungry’, /'anta/ ['an.ta] ‘or’

There are several phonotactic restrictions. First, /r/ occurs word-initially in only three words: ribu ‘thousand’, roti ‘monitor lizard’ and raka ‘rice’. Phonetically, word-initial onset clusters are most commonly plosives followed by a liquid phoneme, /l/ or /r/. We do find clusters of two plosives occasionally, such as gdai or gdira, but more frequently speakers separate the two plosives with an epenthetic vowel (see section 2.4.2). Across syllable boundaries we seem to find the combinations that we would predict based on coda and onset distributions, as shown in some examples in (11). (11)

nasal + fricative nasal + plosive rhotic + plosive plosive + nasal

/n-'amsin/ /'deintɛ/ /wɔrkɔ'mu/ /g-'lɔkman/

‘1SG-hungry’ ‘where’ ‘bird sp.’ ‘3-head’

['nam.siŋ] ['dejŋ.tɛ ~ 'dejn.tɛ] [wɔr.kɔ.'mu] [gǝ.'lɔk.maŋ]

While we do not have any examples of plosive + plosive across syllable boundaries in our corpus, we cannot rule out that it is possible, as syllables beginning in plosives and syllables ending in plosives (realized without an echo vowel) do both exist, although the latter is infrequent. The absence of plosive + plosive across syllable boundaries may therefore be attributable to the rarity of plosive codas in general. The preferred surface syllable structure in Wersing is a simple CV syllable. However, as we have seen in this section, Wersing has many underlying syllable shapes which do not fit with this. Two phonotactic processes are used in Wersing to convert underlying non-CV syllables into surface CV syllables. These are vowel

Wersing

449

epenthesis and echo vowel insertion, discussed in section 2.4.1 and section 2.4.2 respectively.

2.4.1 Epenthetic vowels Word-initial consonant clusters #CCV(C) are brought into line with the preferred CV syllable shape through vowel epenthesis between C1 and C2. The epenthetic vowel may be [ə] or harmonized with the following phonemic vowel. Examples include: (12)

/tlam/ /j-'weri/ /plit/ /mlɔs/ /mlul/

‘six’ ‘2PL-ear’ ‘sharp’ ‘shark’ ‘banana’

[tə'lam ~ ta'lam] [jǝ'wεri ~ jε'wεri] [pǝ'lit ~ pi'lit] [mǝ'lɔs ~ mɔ'lɔs] [mǝ'lul ~ mu'lul]

Although the onsets are rarely if ever realized without at least a short schwa, this cannot merely be a case of vowel lenition in unstressed syllables, as this full vowel ~ schwa alternation only takes place when the vowel is the same as the vowel in the following syllable. We therefore analyze words such as those in (12) as having an underlying initial consonant cluster. Consonant clusters across syllable boundaries (i.e., root with the shape (C) VC.CV(C)) may also be broken up by vowel epenthesis, but not obligatorily. Where these clusters are broken up, the epenthetic vowel may be a schwa or it may be harmonised with the stressed vowel of the root. In the few examples we have in our corpus, the stressed vowel is that of the first underlying syllable, as in the examples in (13). (13)

/g-'lɔkman/ /'sirwis/

‘3-head’ ‘work’

[gǝ'lɔkmaŋ ~ gǝ'lɔkǝmaŋ ~ gǝ'lɔkɔmaŋ] ['sirwis ~ 'sirǝwis ~ 'siriwis]

Vowel epenthesis also occurs in combination with agreement prefixation. See section 2.5.1 in relation to this. There is, however, one instance of vowel epenthesis with a prefix that requires special mention. The Wersing root for ‘give’ is composed of two parts with an intervening agreement prefix: wa-AGR-nir ‘give to AGR’. With this root, we find that the epenthetic vowel that separates the consonantal AGR prefix from the following consonant can, instead of being schwa, harmonise with either the vowel of the preceding syllable (i.e., wa-) or the following syllable (i.e., -nir). This gives the following possible realizations of the

450

Antoinette Schapper & Rachel Hendery

3rd person inflection of /wa-g-nir/ ‘give-3-give’: [wagǝ'nir ~ waga'nir ~ wagi'nir]. It remains to be seen whether this phenomenon or both pro-active and retro-active harmonization is more widely present in the language.

2.4.2 Echo vowels Word-final codas can be brought into line with the preferred open syllable shape by means of an echo vowel. An echo vowel is harmonized with the preceding phonemic vowel, as in (14). (14)

/pis/ /bεb/ /war/ /pɔk/ /sɔb/ /tuk/

‘mango’ ‘palm top’ ‘and’ ‘little’ ‘house’ ‘short’

['pis ~ 'pisi] ['bεb ~ 'bεbε] ['war ~ 'wara] ['pɔk ~ 'pɔkɔ] ['sɔb ~ 'sɔbɔ ~ 'sɔbu] ['tuk ~ tuku]

Unlike an epenthetic vowel, an echo vowel is never schwa (e.g., pis ‘mango’ can be realized as [pisi] but not *[pisə]). The echo vowel is not always, however, an exact ‘echo’ of the preceding vowel. Notice that raising of the final vowel applies in sob ‘house’ in (14). Not all final Cs precipitate echo vowel insertion to the same extent. For example, /lar/ ‘tell’, /min/ ‘be at’, /pur/ ‘garden’, and /tam/ ‘grandchild’ are always found in our corpus without echo vowels. By contrast, /pok/ ‘little’ and /pis/ ‘mango’, for example, are found both with and without echo vowels, but more often with; the conditioning factors for the (non-) appearance of the echo vowel remains to be determined by means of a statistical analysis of a fuller corpus. In other words, liquids and nasals generally remain as word-final codas, but other underlying consonant codas tend to be converted to CV.

2.5 Morphophonemics 2.5.1 Prefixation The paradigms of Wersing agreement prefixes are summarised in Table 5. The prefixes in AGREEMENT SLOT 2 and AGREEMENT SLOT 3 show no allophonic variation and are not discussed further here (for details of their use, see section 6.3 and 6.4 respectively). Prefixes in AGREEMENT SLOT 1 are divided into two paradigms, the choice of which is dependent on the shape of the root.

Wersing

451

Table 5: Wersing agreement prefixes AGREEMENT SLOT 1

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

Usually

On a- roots only

nagɲ tj-

nεεgεnitεji-

AGREEMENT SLOT 2

AGREEMENT SLOT 3

naaganitaji-

nεεgεnitεji-

On most roots (verbs and nouns) including those beginning with /a/, the first paradigm is used in AGREEMENT SLOT 1. On vowel initial roots, the prefixes are realized as per their underlying forms. On consonant initial roots, the consonantal prefixes (i.e., all but 2SG) surface as an unstressed CV syllable, where the V is realized as either a schwa or a vowel harmonized with the nearest root vowel. Examples are provided in Table 6. Table 6: Prefixation of consonant initial roots /-tan/ ‘hand’ 1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

n-tan a-tan g-tan ɲ-tan t-tan j-tan

/-mit/ ‘sit’ [na'taŋ ~ nə'taŋ] [a'taŋ] [ga'taŋ ~ gə'taŋ] [ɲa'taŋ ~ ɲə'taŋ] [ta'taŋ ~ tə'taŋ] [ja'taŋ ~ jə'taŋ]

n-mit a-mit g-mit ɲ-mit t-mit j-mit

[ni'mit ~ nə'mit] [a'mit] [gi'mit ~ gə'mit] [ɲi'mit ~ ɲə'mit] [ti'mit ~ tə'mit] [ji'mit ~ jə'mit]

Roots consisting of a single vowel /a/ have their own agreement prefix forms in AGREEMENT SLOT 1, distinct from those found on other roots, including vowel-initial ones. Table 7 sets out the inflections of the two roots, one noun and one verb, which take these distinct inflections. Note also that this inflectional paradigm is identical to that occurring in AGREEMENT SLOT 3. Table 7: Prefixation of -a roots

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

-a ‘mother’

-a ‘come down’

nε-a ε-a gε-a ni-a tε-a ji-a

nε-a ε-a gε-a ni-a tε-a ji-a

[nε'a] [ε'a] [gε'a] [ni'a] [tε'a] [ji'a]

452

Antoinette Schapper & Rachel Hendery

2.5.2 Metathesis Wersing has productive metathesis associated with suffixation of /-a/ ‘REAL’ and encliticization of /=a/ ‘SPEC’ onto a root of the shape CV1CV2. In this environment, the final vowel metathesizes with the preceding consonant to CV1V2C-. Examples from our corpus include: (15)

[ganing weting getaita] *getatia /ganin wetin g-tati-a/ 3-stand-REAL 3CLF:HUM five ‘There are five people standing.’

(16)

[hans sauka] /hans saku=a/ Hans elder=SPEC ‘Mr Hans’

*sakua

High vowel metathesis is also observed in borrowings, for instance, /baniak/ ‘guest’ < Tetun bainaka ‘guest’ and /brain/ ‘brave’ < Malay berani ‘brave’.

2.6 Orthography There is as yet no working practical orthography for Wersing. In this sketch, we use an orthography close to the Indonesian orthography, as this is the preference of the speakers with whom we have worked. The five vowel phonemes are represented with the graphs {a, e, i, o, u}. The epenthetic schwa vowel is represented by {e}. Consonant phonemes are noted with the graph that is identical to the IPA representations of their underlying form, with two exceptions: {y} is used for the phoneme /j/, and {ng} is used for the velar allophone of the nasal /n/. Stress although phonemic bears a low functional load and as such will not normally be marked in the orthography used here. Throughout the sketch, we represent each utterance in its surface form as we heard it and provide the underlying forms of the words in a separate line of glossing. Thus, the underlying root /tlam/ ‘six’ may be represented as either telam for realization with an epenthetic schwa [təlam] or talam for realization with an epenthetic vowel harmonized with the root [talam].

Wersing

453

3 Basic clausal syntax Like all other TAP languages, Wersing has a basic SV/APV clause order, (17) and (18). Only a few items can follow a predicate, such as the clausal intensifier (17) and clausal negator (18). Free (pro)nominal elements encoding arguments are readily elided when the referent can be retrieved from the discourse context.

(17)

(18)

INTES S V Lemi poko gumur borong. lmi pok g-mur boron male small 3-run very ‘The boy really ran away.’ A P V Naida lemi poko gopoing nanu. naida lmi pok g-poin 1SG.TOP male small 3-hit ‘I didn’t hit the boy.’

NEG nanu not

In the following subsections, we describe basic clausal elements and clause types in Wersing: predicate types (section 3.1), clausal negation (section 3.2), and non-declarative clause types, namely questions and imperatives (section 3.3). Discussion is limited here to mono-predicative clauses. Multi-predicative serial verb constructions (SVCs) are discussed in section 7.

3.1 Predicate types Basic predicate types in Wersing can be differentiated on the basis of their ability to host the realis suffix -a ‘REAL’ when in final predicate position. This suffix is discussed in detail in section 8.1. While non-verbal predicates cannot host the realis suffix in any position, final verbal predicates in Wersing can host it. In our corpus we identify three kinds of predicate: (i) verbal predicates (section 3.1.1), (ii) nominal predicates (section 3.1.2), and (iii) elevational predicates (section 3.1.3).

3.1.1 Verbal predicates There are three syntactic types of verbal predicates in Wersing: one-place, twoplace and three-place verbs. Whilst one and two-participant verbs can be simply

454

Antoinette Schapper & Rachel Hendery

characterized as intransitive and transitive, three-participant verbs are more complex and cannot, as we will see below, easily be made to carry the label “ditransitive”. In this section, we will not deal with the agreement properties of verbs; this is dealt with in section 6.1. Wersing does not have a distinction between verbs and adjectives. Property words are simply one-place or intransitive verbs with adjective-like stative semantics rather than dynamic semantics. Both semantic types of one-place verbs can equally host -a ‘REAL’, as seen with tuk ‘short’ and lailol ‘walk’ in (19) and (20). See section 6.1.1 on agreement on intransitive verbs. (19)

Gaida tuka. gaida tuk-a 3.TOP short-REAL ‘He is short.’

(20)

Gaida lailola. gaida lailol-a 3.TOP walk-REAL ‘He is walking.’

Two-place or transitive verbs are many and also host -a ‘REAL’, as seen on taing ‘reach’ and -poing ‘hit’ in (21) and (22). Note the difference in agreement behavior of the two verbs. See section 6.1.2 for more details. (21)

Neta Kolana taina. neta Kolana tain-a 1SG.AGT KOLANA reach-REAL ‘I reached Kolana.’

(22)

Neta David gopoina. neta David g-poin-a 1SG.AGT DAVID 3-hit-REAL ‘I hit David.’

Finally, Wersing has a handful of three-place verbal predicates. Three-place predicates in Wersing always have a (semi-)fossilised applicative prefix. In the case of -nir ‘give’ (23), the applicative is wa-, a prefix which appears on no other root in the language. In the case of -soi ‘send’ (24), the applicative is le-, a prefix

Wersing

455

with productive functions elsewhere in the language (see section 6.2.1).6 With these verbs, the R typically stands unmarked directly before the predicate, while T is flagged by the serial verb medi ‘take’, (23a) and (24a). However, Wersing also allows medi to be omitted with these verbs and T to be unflagged, (23b) and (24b). Thus, Wersing three-place verbs can be perhaps considered “optionally ditransitive”, in that they typically use an additional verb ‘take’ to code T, but not obligatorily.

(23)

a.

b.

(24)

a.

b.

Flagging of T Neta buku medi Markus waganira. neta buku medi Markus wa-g-nir-a Markus APPL-3-give-REAL 1SG.AGT book take ‘I gave a book to Markus.’ No flagging of T Neta buku Markus waganira. neta buku Markus wa-g-nir-a 1SG.AGT book Markus APPL-3-give-REAL ‘I gave a book to Markus.’ Flagging of T Neta surat medi nepa lesoia. neta surat medi n-pa 1SG-father 1SG.AGT letter take ‘I sent a letter to my father.’ No flagging of T Neta surat nepa lesoia. neta surat n-pa 1SG.AGT letter 1SG-father ‘I sent my father a letter.’

le-soi-a APPL-send-REAL

le-soi-a

APPL-send-REAL

3.1.2 Nominal predicates Nominal predicates are found in equative clauses. These are clauses where one entity is being equated with another by means of juxtapositioning of two NPs

6 We see lesoi ‘send’ as consisting of an applicative le- on the root soi, a verb historically related to the unprefixed verb soi ‘order’.

456

Antoinette Schapper & Rachel Hendery

with one another. The second NP constitutes the predicate. Examples of this type of predicate are provided in (25) to (27). Note that Wersing, like other TAP languages, has no copula. (25)

Topi ba neweng. [topi ba]NP [newen]NP hat DEF 1SG.POSS ‘That hat’s mine.’

(26)

Gaida gedeina David. [gaida g-dein=a]NP 3.TOP 3-name=SPEC ‘His name is David.’

(27)

[David]NP David

Naida Peitoko wasi. [naida]NP [Peitok wasi]NP 1SG.TOP Peitoko local Literally, ‘I am a local of Peitoko’, i.e., ‘I am from Peitoko’.

At first glance, it may appear that nominal predicates do in fact allow the realis suffix. For example, the simple nominal predicate guru ‘teacher’ in (28a) can be marked with the enclitic article =a ‘SPEC’ (28b), a morpheme homophonous with -a ‘REAL’. We cannot, however, analyze the clause final morpheme in (28b) as the realis suffix. This is because we can, for instance, find the morpheme within the scope of a negator (28c), which is impermissible for the realis suffix (see section 8.1). (28)

a.

Naida guru. naida guru 1SG.TOP teacher ‘I am a/the teacher.’

b.

Naida gurua. naida guru=a 1SG.TOP teacher=SPEC ‘I am a teacher.’

c.

Naida gurua nanu. naida guru=a 1SG.TOP teacher=SPEC ‘I am not a teacher.’

nanu NEG

Wersing

457

3.1.3 Elevational predicates As in other TAP languages, Wersing has a set of elevational terms distinguishing three levels. The elevationals and the matching elevational motion verbs are set out in Table 8. Table 8: Wersing elevation terms Elevational

LEVEL HIGH LOW

mona tona yona

Elevational motion verb From DC

To DC

-wai -mid -a

-mai -dai -sir

Wersing elevationals can be used as one-place predicates encoding the location of a NP referent at an elevation (higher, lower or equal with) relative to the speaker. Example (29) illustrates this predicative use. It is not possible to mark such an elevational predicate with the realis suffix (i.e., *toan-a, see section 2.5.2 on metathesis). (29)

Sobo ba tona. sob ba [tona]PRED house DEF HIGH ‘The house is up there.’

The elevationals also have non-predicative uses where they locate an action or an entity by their elevation. In these contexts the elevational follows the clausal element(s) over which it has scope. In (30) the elevational mona follows the NP headed by pei ‘pig’ and denotes the elevation of the pig at the time of its still breathing. In (31) yona follows the verbal predicate aki ‘call’ and denotes the elevation at which the calling takes place.

(30)

NP scope Pei ba mona de geking sesai. pei ba mona de g-kin sesai pig DEF LEVEL IPFV 3-breathe breath ‘The pig (that is) over there is still breathing.’

458

(31)

Antoinette Schapper & Rachel Hendery

Predicate scope David aki yona. David aki yona David call LOW ‘David calls (from) down there.’

3.1.4 Predications expressing location and existence Aside from the elevational predicates, Wersing has a range of verbal predicates used to express location. See section 6.2 on applicatives for further, non-predicative ways of encoding locations in Wersing.

3.1.4.1. Locative verb ming The most common locational predicate is the verb ming /min/, as in (32) and (33). In verb serialization constructions, where ming follows a motion verb, it denotes the location which is the result of the motion.7 This is seen in the second occurence of ming in (33). (32) Gaida Lantok mina. gaida Lantok min-a 3.TOP Lantoka be.located-REAL ‘He is in Lantoka.’ (33) Naida sobo minte nemur neya lang mina. naida sob min=te n-mur n-ya lan min-a 1SG.TOP house be.located=CONJ 1SG-run 1SG-go seaside be.located-REAL ‘I was in the house and then ran down to the beach.’ As can be seen from the above examples, ming has general locative semantics. More specific locative relations are denoted by using ming in combination with a locative noun, such as mira ‘inside’ (34) and motumo ‘beneath’ (35). (34)

Pisi ba cangkir gemira mina. pis ba cankir g-mira 3-inside mango DEF cup ‘The mango is inside the cup.’

min-a be.located-REAL

7 This is a common construction in related languages, see Schapper (2011b).

Wersing

(35)

Pede pilit motumo mina ong. pde plit motumo min=a machete sharp beneath be.located=SPEC ‘Use the sharp knife which is underneath.’

459

on use

Finally, ming occurs prefixed with the deictic particles o- ‘PROX’ and a- ‘DIST’8, as in (36) and (37). (36)

(37)

Nyaida omin oba. nyaida o-min 1PL.EXCL.TOP PROX-be.located ‘We are here now.’

o-ba

PROX-DEM

Buku amina. buku a-min-a book DIST-be.located-REAL ‘The book’s there.’

3.1.4.2. Posture verbs and the expression of existence Wersing lacks a dedicated set of existential verbs. Instead Wersing uses posture verbs to express the existence of an entity. For non-humans, the choice of posture verb is determined by the conventional or customary posture of the referent, as in (38) and (39). (38)

Dring nor tu gemit. drin nor tu g-mit honey master three 3-sit ‘There are three bees.’ Literally, ‘Three bees sit’.

(39)

Neta nemur, aser ba getia gepai nemida. neta n-mur aser ba g-ti=a 1SG.AGT 1SG-run mountain REL 3-lie=SPEC

g-pai 3-make

n-mid-a 1SG-go.up-REAL

‘I run along and am forced to go uphill by the mountains that are there.’ Literally, ‘I run along and the mountains that lie (there) make me go up.’ For humans, the choice of posture verb is a matter of whatever posture (e.g., sitting or standing) the referent is in at the time of the utterance. It is notable, 8 These particles also appear in the forms of the demonstratives, o-ba ‘PROX-DEM’ and a-ba ‘DIST-DEM’. See section 4.3.

460

Antoinette Schapper & Rachel Hendery

however, that Wersing speakers seem to avoid the use of posture verbs to express the existence of humans.

3.2 Negation Unlike most other TAP languages, Wersing has two distinct post-predicative negators: nanu and naung /naun/. The distinction between them is one of expected polarity. Compare the two clauses in (40). Nanu is the unmarked negator (40b), while naung is used for negation in contexts where there is a background expectation of positive polarity (40b), for instance, in response to a question like You’re walking too, yeah? (40)

a.

Naida lailol nanu. naida lailol nanu 1SG.TOP walk NEG ‘I’m not walking.’

b.

Naida lailol naung. naida lailol naun 1SG.TOP walk NEG ‘I’m not walking.’

Neither of these negators is limited to a particular type of predicate. They can both be freely used not only with verbs as in (40), but also with nominal and elevational predicates, as in (41) and (42) respectively. (41)

(42)

a.

Naida guru nanu. naida guru nanu 1SG.TOP teacher NEG ‘I’m not a teacher.’

b.

Naida guru naung. naida guru naun 1SG.TOP teacher NEG ‘I’m not a teacher (contrary to expectation).’

a.

Sobo ba tona nanu. sob ba tona nanu house DEF HIGH NEG ‘The house is not up there.’

Wersing

b.

461

Sobo ba tona naung. sob ba tona naun house DEF HIGH NEG ‘The house is not up there (contrary to expectation).’

In addition, Wersing has a special negator, sining /snin/ that is used in negative imperatives. Like the other Wersing negators, sining occurs postpredicatively, as in (43). (43)

Asorong sining. a-soron snin 2SG-angry NEG ‘Don’t be angry.’

Wersing also has two negative interjections: nau ‘no’ and the more emphatic naungka ‘no way, never’.

3.3 Questions Wersing polar interrogatives are formed with the usual declarative word order plus a tag clitic =e on the final verb (44a). Such a question can be answered with yo ‘yes’ or nau ‘no’, followed optionally by the declarative positive or negative statement, as illustrated in (44b). The realis suffix -a is not possible in the declarative response to a polar interrogative (44c). (44)

a.

Aida lailole? aida lailol=e 2SG.TOP walk=TAG ‘Are you walking?’

b.

Yo, naida lailol. yo naida lailol yes 1SG.TOP walk ‘Yes, I am walking.’

c. *Yo, naida lailola. yo naida lailol-a yes 1SG.TOP walk-REAL ‘Yes, I am walking.’

462

Antoinette Schapper & Rachel Hendery

Questions with interrogative words use one of the interrogatives in Table 9. The interrogatives ‘when’ and ‘why’ seem to be, at least historically, multimorphemic. The word deing is otherwise a pluralizing word, and the clitic =te in other contexts is a clause conjunction. Table 9: Wersing interrogatives tro deingte nalte mel tron trongte ~ trong paite

where when what who how why

Interrogatives typically occur in the position of the questioned consitituent or in the topical clause-initial position. Some examples of interrogatives from this table are given in (45) to (48). (45)

David gedai tro migoko ge? David g-dai tro mi-g-ko ge David 3-come.up where APPL-3-stay PROSP ‘If David comes, where is he going to stay?’

(46)

Deingte David gedai ge? deinte David g-dai ge when David 3-come.up PROSP ‘When is David going to come?’

(47)

Mel genanal oba? mel g-nanal o-ba who 3-possessions PROX-DEM ‘Whose stuff is this?’

(48)

Trong nisir ge? tron n-sir ge how 1SG-descend PROSP ‘How should I come down?’

3.4 Imperatives and related clause types Imperatives take the form of a standard clause with second person A or S, that is, the verb is marked for second person if this would usually be the case for

Wersing

463

that verb with a second person A or S. Examples of imperatives with a verb that agrees with the second person and one that does not are shown in (49a & b). The presence of a free pronoun is optional, and is illustrated in (49c). (49)

a.

Waganir telam kida. wa-g-nir tlam kida APPL-3-give six just ‘Give him just six.’

b.

Yemai teya sirwis belet. y-mai t-ya sirwis blet 2PL-come 1PL.INCL-go work do.work ‘You come and we will go and work.’

c.

Eta awai pisa geleki. eta a-wai pis=a g-leki 2SG.AGT 2SG-go mango=SPEC 3-count ‘You go and count the mangos.’

Similarly, first person imperatives are constructed as for any clause with a first person plural S or A. Both second and first person imperatives can occur with or without an emphatic particle bo. Examples with bo are given in (50) to (52). (50)

Amidi bo! a-mid bo 2SG-go.up EMPH ‘Climb up!’

(51)

Tanaing dra bo! tnain dra 1PL.INCL.all sing ‘Let’s sing.’

(52)

bo EMPH

Tawaisi gemaksebi teng bo. t-wai=si g-maksebi 1PL.INCL-go=CONJ 3-ask ‘Let’s go and ask him first.’

ten first

bo EMPH

464

Antoinette Schapper & Rachel Hendery

4 Noun phrases The template for the Wersing noun phrase (NP) is given in (53). It is much like those in other TAP languages. Modifiers follow the head noun of the NP (NHEAD). They are an attribute (ATTR), a numeral (NUM) or the plural word (PL), and a relative clause (RC). Right-most in the NP is a determiner (DET). An adnominal possessor (PSR) is marked by a prefix on NHEAD and this may be co-indexed by a separate NP encoding the possessor.

(53)

Template of Wersing NP (PSRNP) [PSR-NHEAD ATTR NUM/PL RC DET]NP

Examples (54) to (56) present examples of the different NP slots in Wersing. The slots will be further discussed and illustrated in the following subsections.

(54)

(55)

(56)

PSR-NHEAD ATTR PL DET gis kebai deina g-is kebai dein=a 3-content young PL=SPEC ‘their young flesh’ NHEAD RC DET blek ba mede ayang legatati ba blek ba med ayan le-g-tati REL table top APPL-3-stand tin ‘that tin that is on top of the table’

ba DEF

NHEAD ATTR NUM mulul merek weting mlul mrek wetin banana ripe five ‘five ripe bananas’

4.1 Attributes and relative clauses Wersing has slots in the NP for both an attribute and a relative clause. An illustration of the slots is given in (57) in the NP headed by pede /pde/ ‘machete’. An attribute is a simple intransitive verb that directly follows the NP head, in this example pilit /plit/ ‘sharp’. A relative clause is a clause that modifies the referent

Wersing

465

of head noun and attribute, in this example the locative predication motumo ming ‘be underneath’. (57)

David, pede pilit motumo mina ong. David [pde plitATTR [motumo min]RC=a]NP David machete sharp beneath be.at=SPEC ‘David, use the sharp machete that’s down below.’

on use

Unlike an attribute, a relative clause may be more than a simple intransitive verb. However, where there is only a single intransitive verb modifying a head noun and no other diagnostic elements in the NP, it is not possible to differentiate an attribute and a RC from one another. For instance, the intransitive verb braing /brain/ ‘brave’ in (58a) has an ambiguous status and can be interpreted as either an attribute or an RC. An attribute can be differentiated from an RC through marking of the head; the RC head marker ba ‘REL’ can be used to restrict the identity of the referent, as in (58b).9 The explicit marking of a predicate as being in a relative clause with ba ‘REL’ means that a distinct attribute can be included in the phrase, as with akeng /aken/ ‘black’ in (58c). This must be considered an attribute since it cannot be preceded by ba ‘REL’ either together or distinct from the RC predicate braing, (58d) and (58e) respectively. (58)

a.

kutual braing ba kutual brain ba dog brave DEF ‘that brave dog’ or ‘that dog (which) is brave’

b.

kutual ba braing ba kutual ba brain ba REL brave DEF dog ‘that dog which is brave’

c.

kutual akeng ba braing ba kutual aken ba brain ba dog black REL brave DEF ‘that black dog which is brave’

9 The reader will observe that ba ‘REL’ is homophonous with the definite article ba ‘DEF’. See Schapper (2007) on the relationship between determiners and historical development of these particles in TAP languages and Hendery (2012:52–56, 127–128) for a discussion of the relationship between determiners and relative clause markers cross-linguistically.

466

Antoinette Schapper & Rachel Hendery

d. *kutual ba akeng braing ba kutual ba aken brain REL black brave dog

ba DEF

e. *kutual ba akeng ba braing ba kutual ba aken ba brain REL black REL brave dog

ba DEF

4.2 Quantification Number is not marked on nouns in Wersing and may be signalled optionally by a range of quantificational items. Whilst numeral quantifiers in Wersing form a well-defined class (section 4.2.1), non-numeral quantifiers have diverse morphosyntax (section 4.2.2).

4.2.1 Numeral quantifiers The basic numeral quantifiers in Wersing are given in Table 10. The system is quinary-decimal. Wersing dialects differ in where the quinary numerals begin: Wersing Pureman has a reflex of etymological PTAP *talam ‘six’, while Wersing Kolana has a complex quinary numeral ‘5+1’ for ‘6’ (probably due to contact with Sawila; Schapper & Klamer forthcoming). The higher numeral bases are aska ‘100’ and ribu ‘1000’. Wersing does not appear to have any numeral classifiers. Table 10: Wersing numerals 1

/no/

2 3 4 5

/yoku/ /tu/ /arasoku/ /wetin/

6 7 8 9 10

/tlam/ (Pureman form) /wetin no/ (Kolana form) /wetin yoku/ /wetin tu/ /wetin arasoku/ /adayoku/

By means of several additional morphemes, numerals are combined to form higher numerals. For instance, consider the numeral in (59). In this, yoku ‘two’ is marked with mi- to denote multiplication of the preceding numeral adayoku ‘ten’ (cf. section 6.2.2 for other uses of mi-). Tu ‘three’ is added to the total by means of the addition word geresing /gresin/ ‘plus’.

Wersing

(59)

sob adayoku miyoku geresing tu sob adayoku mi-yoku MULTIPLY-two house ten ‘twenty-three houses’

gresin plus

467

tu three

The numeral no ‘one’ has a special status in Wersing, being able to combine with other numerals without the numerical value ‘1’. In examples (60) and (61) we see that no ‘one’ follows the numerals yoku ‘two’ and telam /tlam/ ‘six’ respectively. Here no denotes that the enumerated referents form a single unit in their actions. (60)

Muduk yoku no ire. mduk yoku no ire bat two one fly ‘Two bats are flying.

(61)

Peter awi telam no geti. pter awi tlam no g-ti corn cob six one 3-lie ‘There are six cobs of corn.’

Similar unifying uses of the numeral ‘one’ are known from Bunaq (Schapper 2010: 104) and Sawila (Kratochvíl this volume). Wersing does not for the most part have numeral classifiers. However, as in other Alor languages, there is an obligatory inflecting classifier used when enumerating human referents above one. Table 11 sets out the inflections and examples of the inflections in use are given in (62) and (63). Table 11: Wersing inflecting human numeral classifier 1PL.EXCL 1PL.INCL 2PL 3

(62)

/nnin ~ nanin/ /tnin ~ tanin/ /ynin ~ yanin/ /gnin ~ ganin/

Saku mei ganing weting sibai takada gisir. takada [saku mei ganin wetin]NP sibai basket carry.head old female 3.HUM:CLF five ‘Five women carrying baskets down on their heads.’

g-sir 3-go.down

468 (63)

Antoinette Schapper & Rachel Hendery

Tening tu tawai. [tnin tu]NP 1PL.INCL.HUM:CLF three ‘We three go across.’

t-wai 1PL.INCL-go.across

The human classifiers are derived from the noun aning /anin/ ‘human’. The relationship between the noun and the classifier is still evident in the fact that when the head of the NP is aning, no classifier can be used alongside a numeral. Compare the grammaticality of (64a) and (64b). (64)

a.

aning telam anin tlam human six ‘six people’

b. *aning ganing telam anin ganin human 3.HUM:CLF

tlam six

4.2.2 Non-numeral quantifiers Non-numeral quantifiers are of several kinds in Wersing. First, some intransitive verbs are used to express the quantificational concepts of ‘much, many’ (65a) and ‘a little, some’ (65b).

(65)

a.

Verbal quantifiers aning kang bal anin kan bal human good many ‘many people’

b.

asir poko asir pok salt little ‘little bit of salt’

Second, Wersing has a plural word deing /dein/ glossed ‘PL’.10 The plurality marked by deing is vague, and means that the plural cannot be used in combination with other number and non-numeral quantifiers in Wersing. There is, however, no restriction on the type of referent; the plural word can be used to signal the plurality of a human (66), animal (67) or inanimate referent (68). 10 Wersing deing does not reflect the PAP plural word *noŋ (Schapper & Klamer 2011). Deing appears to be related to the noun -deing ‘name’ and to be a calque from Timorese languages, many of which use the word for ‘name’ for expressing plurality (see Schapper 2011a).

Wersing

469

(66) . . . , saku deing bias ol tamu poko deina gepai getai. . . saku dein bias ol tamu pok dein=a g-pai g-tai usually child grandchild small PL=SPEC 3-make 3-sleep adult PL ‘. . . , the adults would usually [do it] to make the children and grandchildren sleep. . .’ (67) Nekarbau wari newai deing na yeta legadar. n-karbau wari n-wai dein=na yeta le-gdar 2PL.AGT APPL-guard 1SG-buffalo and 1SG-goat PL=FOC ‘You watch out for my buffalo and my goats.’ (68)

Kiki deing aso gemira susa. kik dein aso g-mira susa also 3-inside suffer flower PL ‘The flowers were also suffering.’

Wersing has an additional plural word, naing /nain/, marking associative plurality. This form is known to us from the Mark gospel, marking personal names, as in (69) and (70). Note the similarity between naing and the root on which the ‘all’ pronouns are built (see Table 13). (69)

Petrus naing gaumeng gapang gepai. Petrus nain g-aumen g-pan g-pai ASSOC 3-fear 3-dead 3-make Peter ‘Peter and the others were afraid to die.’ [MRK009:002]

(70)

Yesus naing lailol gewai Kapernaum taing. Yesus nain lailol g-wai Kapernaum tain Kapernaum reach Jesus ASSOC walk 3-go ‘Jesus and the others walked onto Kapernaum.’ [MRK009:033]

Finally, Wersing has pronominal series which are used for quantification in certain contexts. They are discussed in section 5.4.

4.3 Determiners Wersing has four determiners: two articles =a ‘SPEC’ and ba ‘DEF’ and two demonstratives, a-ba ‘DIST-DEM’ and o-ba ‘PROX-DEM’.11 There is only one determiner 11 Wersing demonstratives appear to be formed using the definite article ba plus one of the two deictic morphemes o- ‘PROX’ and a- ‘DIST’ (cf. section 3.1.4.1).

470

Antoinette Schapper & Rachel Hendery

slot in Wersing and multiple determiners cannot mark a single NP (e.g.,*=a o-ba ‘SPEC PROX-DEM’). The articles are dependent elements; they only occur adnominally and cannot stand on their own in the NP. The semantic difference between articles appears to be one of specificity-definiteness. The article =a marks specificity, while ba marks definiteness. The difference between the two articles is chiefly obvious in that =a can occur within a NP marked with the numeral no ‘one’, as in (71), whereas ba cannot (*no ba ‘one DEF’). (71) Botil yoku no gebir ong medi, pangsa botil noa gekureng ong medi. botil yoku no g-bir on medi pan=sa [botil no=a]NP bottle two one 3-mouth use take CONC=CONJ bottle one=SPEC g-kuren on 3-bottom use

medi take

‘Two bottles are put upside down (lit. on their mouths), while a single [bottle] is put upright (lit. on its bottom).’ Wersing demonstratives occur both adnominally and pronominally, as with a-ba ‘DIST-DEM’ in (72) and (73) respectively. (72)

Kutual aba namalosi. kutual a-ba n-amalosi DIST-DEM 1SG-not.want dog ‘I don’t want that dog.’

(73)

Aba ong mada namalosi. a-ba on mad n-amalosi DIST-DEM use cook 1SG-not.want ‘I don’t want to cook with that.’

Semantically, the demonstratives contrast in distance from the speaker, o-ba ‘PROX-DIST’ referring to an entity in the immediate vicinity of the speaker and a-ba ‘DIST-DEM’ to an entity outside that zone. Discourse and non-spatial uses of the demonstratives remain to be studied.

4.4 Possession Possessors in Wersing are marked on the possessed noun by means of a prefix (74a). An optional NP denoting a possessor (and coindexing the possessive prefix) may precede the possessum in Wersing (74b).

Wersing

(74)

a.

gedurki g-durki 3-mouse ‘its / their mouse’

b.

471

akamau gedurki akamau g-durki cat 3-mouse ‘the cat’s mouse’

Wersing does not have a distinction between alienable and inalienable possessed nouns as most other TAP languages do. Whilst the vast majority of TAP languages use distinct possessive markers for alienable and inalienable possessed nouns, Wersing uses one and the same prefixal set for the expression of all possessors on all nouns. There is, however, a distinction between nouns that are obligatorily possessed and nouns that are only optionally possessed. That is, Wersing has a closed subset of nouns which never occur without a possessive prefix, and this contrasts with a large open class which can occur without. As in the “inalienable” possessive classes in other languages, obligatorily possessed nouns refer to body parts and a few kin relations. Table 12 presents two examples of each class. Table 12: Examples of possessive class inflections OBLIGATORILY

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

POSSESSED NOUNS

OPTIONALLY

POSSESSED NOUNS

/-tan/ ‘hand’

/-pa/ ‘father’

/sob/ ‘house’

/kutual/ ‘dog’

n-tan a-tan g-tan ni-tan t-tan yi-tan

n-pa a-pa g-pa ni-pa t-pa yi-pa

n-sob a-sob g-sob ni-sob t-sob yi-sob

n-kutual a-kutual g-kutual ni-kutual t-kutual yi-kutual

See section 5.5 for the discussion of possessive pronouns in Wersing.

5 Pronouns Like other central-eastern Alorese languages, Wersing has multiple sets of independent pronouns. Table 13 presents the ten pronominal paradigms found in our data. Wersing pronouns are used to mark a range of functions, including agentivity-patientivity, pragmatics, quantity and possession. Pronominal paradigms have the same feature distinctions as agreement prefixes as well as similar forms initially (see section 6; Table 14). Pronouns differentiate singular and plural in the 1st and 2nd person. The third person makes no number distinc-

472

Antoinette Schapper & Rachel Hendery

tion. In the 1st person plural there is an additional inclusive-exclusive distinction. 1st person plural inclusive pronouns (and prefixes) are also used in the encoding of reciprocals in Wersing. Syntactically, pronouns stand in for noun phrases, as in (75a). In the 3rd person, a pronoun can also stand apposite to an NP, as in (75b). (75)

a.

Genuku unang lewena. gnuk unan le-wena louse APPL-search 3.DU ‘The two of them searched for lice.’

b.

Aning genuku unang lewena. aning gnuk unan le-wena louse APPL-search person 3.DU ‘Those two people searched for lice.’

Pronouns in Wersing can also be stacked. In our corpus, an agentive or oblique pronoun is always the right-hand of two stacked pronouns, as in (76) and (77) respectively. The left-hand pronoun may be a quantificational pronoun, as with the dual pronoun in (76), or a pragmatic pronoun, as with the focus pronoun in (77). (76)

Yenuk yeta yeweri mana. ynuk yeta y-weri mana 2PL.AGT 2PL-ear turn.on 2DU ‘You two you turn on your ears.’

(77)

Newaia anta nanu, newe nadi kida. n-wai-a anta nanu newe nadi NEG 1SG.FOC 1SG.OBL 1SG-go-REAL or ‘Whether I go or not, it’s up to me alone.’

kida very

In section 5.1, we discuss gai, a special pronoun for encoding 3rd person that stands outside of a paradigm. Following this, we discuss the normal paradigmatic pronouns by function: agentive and oblique pronouns in section 5.2, pragmatic pronouns in section 5.3, quantificational pronouns in section 5.4, and possessive pronouns in section 5.5.

473

Wersing

Table 13: Wersing pronominal paradigms

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

AGENT

OBLIQUE

TOPIC

FOCUS

ALONE

neta eta geta nyeta teta yeta

nadi adi gadi nyadi tadi yadi

naida aida gaida nyaida taida yaida

newe ewe gewe nyewe tewe yewe

nenasu enasu genasu nyenasu tenasu yenasu

INDEPENDENT

DUAL

ALL †

GROUP ‡

POSSESSIVE

nenal enal genal nyenal tenal yenal

– – gnuk nynuk tnuk ynuk

– – gnaing nynaing tnaing ynaing

– – gawi nyawi tawi yawi

neweng eweng geweng nyeweng teweng yeweng

† The root of this pronominal paradigm, -naing, is also found in free form; naing is used to mark associative plurality (see section 4.2.2). ‡ The root of this pronominal paradigm, -awi, is also found in free form; awi is a verb meaning ‘whole’.

5.1 Third person A pronoun When a verbal clause has two 3rd person participants (A and P), the A participant is frequently encoded with gai ‘3.A’. We do not observe the use of gai where A or P is, for instance, a 1st person. Gai is included where the P is fronted. Indeed all examples in the corpus which have two 3rd person participants and a fronted P have their A marked by gai, indicating that it may be obligatory to mark the A in these contexts. Such obligatoriness is also suggested by the fact that marking of A with gai is not limited to possibily ambiguous contexts. Example (78) is potentially ambiguous in that, if David were not marked with gai, we would have no way to know that Ela was not A. Only the inclusion of gai with David precludes this interpretation. However, gai also marks A in (79) where the P is inanimate. In this, the possibility of interpreting tas ‘bag’ as an A is very slight, were gai not present, because of the semantic improbability of an agentive inanimate referent. (78)

Ela saku ba David gai kamar miguwuing. [Ela saku ba]P [David gai]A kamar DEF David 3.A room Ela old ‘Ela is hidden in the room by David.’ *‘David is hidden in the room by Ela.’

mi-g-wuni IN -3-hide

474 (79)

Antoinette Schapper & Rachel Hendery

Tas turu pona ol poko gai semel boli. [ol pok gai]A semel [tas tur pon=a]P lift bag heavy thus=SPEC child little 3.A ‘A heavy bag like this cannot be lifted by a child.’ */#‘A child cannot be lifted by a heavy bag like this.’

boli bad

It is also possible for gai to mark a 3rd person A where P is not fronted, as in (80) and (81). The pragmatics of this marking are not well understood and further research is needed using a larger text corpus. (80)

(81)

David gai Sem gopoing. [David gai]A [Sem]P Sem David 3.A ‘David hit Sem.’ *‘Sem hit David.’

g-poin 3-hit

Munu gai ira misai diakal borong. diakal [mun gai]A [ira]P mi-sai water APPL-swim quickly snake 3.A ‘The snake swims in the water very fast.’

boron very

Gai is treated as a pronoun because like other 3rd person pronouns, it can either stand in apposition to an NP (as seen in the examples thus far) or it can stand in for an NP, as in (82) and (83). (82) Asina gai lar. [asin=a]P [gai]A lar tell story=SPEC 3.A ‘Stories were told by them.’ (83) Nelokomang pilit oba, gai bebe ba tona gamangtaingsa, . . . ba]P tona g-mantain=sa n-lokman plit o-ba [gai]A [beb palm.top DEF HIGH 3-hook=CONJ 1SG-head sharp PROX-DEM 3.A ‘this sharp head of mine, it will get stuck in the top of the palm tree. . .’

5.2 Agentive and oblique pronouns The most frequently used pronouns in our corpus of Wersing are the AGENT and OBLIQUE paradigms.

Wersing

475

AGENT pronouns can be used to code participants that are considered to be agentive. This means that agentive Ss can be encoded with these pronouns, such as with the intransitive verbs sui ‘whistle’ in (84) and mid ‘go up’ in (85). (84)

. . . , neta suisa. . . neta sui=sa 1SG.AGT whistle=CONJ ‘. . . and I will whistle. . .’

(85) Nyeta yel iyut golol lailol nyemidi pur taing. nyeta yel iyut g-lol lailol ny-mid pur tain 1PL.EXCL.AGT path narrow 3-follow walk 1PL.EXCL-go.up garden until ‘We follow the narrow paths to the garden.’ AGENT pronouns also encode As of both two- and three-place verbs, as in (86) and (87) respectively. We see in comparing the examples in (86) that being an A is sufficient for “AGENT” coding: whilst the A of ‘hit’ is clearly agentive (86a), the A of ‘smell’ is an experiencer and lacks the prototypical agent properties (86b). (86)

(87)

a.

Neta Benny gopoing. neta Benny g-poin 1SG.AGT Benny 3-hit ‘I hit Benny.’

b.

Bongkou mada ba neta lemuing. bongkou mad ba neta cassava cooked DEF 1SG.AGT ‘I smell the cooked cassava.’

le-muin APPL-smell

Neta buku medi Markus wagenir. neta buku medi Markus wa-g-nir 1SG.AGT book take Markus APPL-3-give ‘I give Markus a book.’

OBLIQUE pronouns can be used for all participants that cannot be encoded with AGENT pronouns. That is, they encode participants with roles that are nonagentive, including Ps (88) and non-agentive Ss (89).

476 (88)

Antoinette Schapper & Rachel Hendery

a.

Kutual nadi pasi. kutual nadi pasi dog 1SG.OBL bite ‘A dog bites me’

b.

Kapo genor ganing gedai gemai topoing tadi belas ge. kapo g-nor g-anin g-dai g-mai t-poin lest 3-master 3-person 3-come.up 3-come 1PL.INCL-hit tadi belas ge 1PL.INCL.OBL smack PROSP ‘Lest some of the owners come to give us a hiding.’

(89)

Adi modo pante leyer noko. adi mod pan=te le-yer CONC = CONJ APPL-laugh 2SG.OBL fall ‘You fall and I laugh at (you).’

n-ko 1SG-PROG

5.3 Pragmatic pronouns The pragmatically marked pronouns are the TOPIC and FOCUS paradigms. A topic pronoun denotes what an utterance is about. In our corpus, topic pronouns are always subjects (i.e., S or A). Topic pronouns appear to always occur at the beginning of an intonation unit, as in the first position of a clause after a final falling intonation in (90), or at the beginning of a syntactic clause, as after a clausal conjunction in (91). (90)

a.

Aida Wersing. aida Wersin 2SG.TOP Wersing ‘You (will teach) Wersing.’

b.

Naida Abui gelomu ong getenara, . . . naida Abui g-lomu on g-tenar 1SG.TOP Abui 3-language use 3-teach ‘I will teach Abui.’

(91) . . . neta suisa, aida amasining wota ayang leamiti ako. neta sui=sa aida a-masnin ota ayan le-a-mit a-ko 1SG.AGT whistle=CONJ 2SG.TOP 2SG-quiet coconut top APPL-2SG-sit 2SG-stay ‘I will whistle such that you will stay quiet on top of the coconut.’

Wersing

477

Topic pronouns lose their final syllable when followed by the plural word deing /dein/ ‘PL’, as in (92). (92)

Tai deina omina. taida dein=a 1PL.INCL PL=SPEC ‘We all are here.’

o-min-a

PROX-be.at-REAL

Focus pronouns are used to restrict the reference to precisely the individual (s) at issue in the discourse. A focus pronoun encoding A is often preceded by a fronted P (93), but this is not obligatory for the pronoun’s use (94). (93)

Nau, sobo ba, newe gapai. nau sob ba newe g-pai no house DEF 1SG.FOC 3-make ‘No, as for that house, I made it.’

(94)

David gewe nalegedara. David gewe na-le-gdar David 3.FOC 1SG-APPL-guard ‘It is David who looks after me.’

An S argument can also be encoded with a focus pronoun, as in (95). (95)

Aning no gewe bisara. anin no gewe bisar person one 3.FOC speak ‘It is one person who speaks.’

Whilst there are no instances of a P encoded by a focus pronoun, it remains to be confirmed with negative evidence that focus pronouns are restricted to S and A.

5.4 Quantificational pronouns Quantificational pronouns express that their referent have a particular number of participants. Paradigms of quantificational pronouns are divided into two types: those that have singular and plural forms (ALONE and INDEPENDENT paradigms), and those that only have plural forms (DUAL, ALL and GROUP paradigms).

478

Antoinette Schapper & Rachel Hendery

Alone and independent pronouns express that their referents acted individually. The semantic difference between the two pronominal paradigms is one of exclusivity. An alone pronoun (glossed with ‘ALONE’) expresses that the referent(s) is/are entirely without companion, i.e. that no one acted or experienced as the referent did (96). By contrast, an independent pronoun (glossed with ‘INDEP’) expresses that referent(s) acted entirely on their own without aid from anyone else who may have been present or similarly engaged (97). (96)

Bong ba genal poda. bon ba genal pod-a tree DEF 3.ALONE broken-REAL ‘That tree alone was broken.’

(97)

Aya keng kum no genasu lakal. aya ken kum no genasu lakal rain cloth round one 3.INDEP hang ‘A fluffy cloud floats by itself (in the sky).’

Dual pronouns are used where the speaker wants to be clear that the referents are exactly two in number, as in (98). The use of dual pronouns for two participants is, however, not obligatory. (98)

Yinuk yer. ynuk yer laugh 2.DU ‘The two of you laughed.’

All pronouns (glossed with ‘all’) express that every participant is included in the reference, as in (99). A pronoun of this paradigm can also be marked with -le, as in (100). This suffix does not appear on nouns or any other pronominal series in Wersing; it is likely cognate with the Kamang associative plural marker -lee (see section 4.2 in the Kamang sketch). (99)

(100)

Tanaing dra bo! tnain dra 1PL.INCL.all sing ‘Let’s all sing.’

bo EMPH

Genaingle kamar mingte nanal temekeng. gnain-le kamar min=te nanal t-meken room be.in=CONJ thing RECP-exchange 3.all-PL ‘All of those who are in the room exchange things.’

Wersing

479

Group pronouns (glossed with ‘GRP’) express that the referents act together as a group (101). (101)

Nyawi nyimit o! nyawi ny-mit 1PL.EXCL.GRP 1PL.EXCL-sit ‘Let’s sit together!’

o EXCLAM

5.5 Possessive pronouns Like other pronouns, a possessor pronoun substitutes for a whole possessed NP, and does not merely encode the possessor. The identity of the possessed item itself must be retrieved from the discourse context. Examples of possessive pronouns are provided in (102) and (103). (102)

Keng ba neweng. ken ba newen cloth DEF 1SG.POSS ‘The cloth’s mine.’

(103)

Akamau durki wing gewena. akamau durki win gewen=a cat mouse grab 3.POSS=SPEC ‘The cat grabs the mouse (and makes it) his.’

6 Agreement and applicative prefixes The primary template for prefixation of the Wersing verb is given in (104). There are three prefixal slots on a Wersing verb. The first slot, closest to the verb root, is the AGREEMENT SLOT 1 (AGR1). The second or middle slot is the APPLICATIVE SLOT (APPL) where one of the two Wersing applicatives can appear. The third and outermost slot is the AGREEMENT SLOT 2 (AGR2). AGR1 is the primary agreement site on the verb in Wersing. AGR2 is never filled when there is no item in APPL. The two agreement slots are not mutually exclusive, but are rarely both filled at the same time.

(104)

Primary Wersing verb prefixal template AGR2-APPL-AGR1-V

480

Antoinette Schapper & Rachel Hendery

Wersing also has a secondary prefixation paradigm that is used for marking inceptive aspect. In this, a distinct series of agreement prefixes attaches to the verb in what is called here AGREEMENT SLOT 3 (AGR3). The AGR1 slot may be filled where required alongside AGR3, but not with the APPL and the AGR2 slots. This is set out in the template in (105).

(105)

Secondary Wersing verb prefixal template AGR3-AGR1-V

Table 14 presents the forms of the prefixes that can fill each slot. In the following sections, we will discuss and illustrate the slots of verbal prefixes in Wersing further. See section 2.5.1 for information on the morphophonemics of prefixation. The use of agreement prefixes with noun roots was discussed in section 4.4 in connection with nominal possession and will not be discussed further here.12 Table 14: Wersing verbal prefixes by slot AGREEMENT SLOT 3

AGREEMENT SLOT 2

APPLICATIVE SLOT †

AGREEMENT SLOT 1

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

APPL APPL

1SG 2SG 3 1PL.EXCL 1PL.INCL 2PL

neegenitayi-

naaganitayi-

lemi-

nagnyty-

† The prefix wa- on the verb -nir ‘give’ strictly could be said to occur in this slot also (see section 3.1.1). We do not, however, include it in this table, because it is a lexically specified prefix and does not occur on any verbs apart from -nir.

It should be noted that agreement prefixes in Wersing are not pronominal in the sense that they replace or exclude free nominal elements. A participant indexed by an agreement prefix can simultaneously be encoded by free nominal elements, either a pronoun (106) or, in the case of 3rd person referent, an NP (107). 12 The account given here of Wersing agreement differs from that in Donohue (2008: 66–67). In our data, we find no evidence for multiple contrastive prefixal series. The form nadi- glossed as ‘1SG.ABS.III’ by Donohue appears to be a misinterpretation of the oblique pronoun nadi (see section 5.2) as a bound form. The form nai- glossed by Donohue as ‘1SG.ABS.II’ does not occur in our data except as an apparent contraction of the prefix series na-le- ‘1SG-APPL-’ on the verb skau ‘hug’. (Note that similar contractions of agreement and applicative prefixes are observed in Sawila). Our account is, however, preliminary. On collection and consideration of a larger corpus, it may turn out that there is indeed a “AI” series of prefixes in Wersing.

Wersing

(106)

(107)

481

Agreement prefix coindexed by pronoun Eta awai. eta a-wai 2SG.AGT 2SG-go ‘Off you go.’ Agreement prefix coindexed by NP Naida pis ba geleki. naida [pis ba]NP g-leki 3-count 2SG.TOP mango DEF ‘I count the mangos.’

6.1 Agreement slot 1 6.1.1 Agreement on intransitive verbs Intransitive verbs divide into two primary classes: those that cannot host a prefix in AGR1 and those that can. The division between these two is lexical and not semantic, as a comparison between examples of members of the two classes in Table 15 shows. Table 15: Wersing intransitive verbs and prefixal possibilites No prefix possible in AGR1 /aki/ /awaki/ /burik/ /dra/ /er/ /kum/ /kur/ /lailol/ /lakal/ /len/ /mdain/ /mdit/ /mdok/ /mod/ /takawai/ /wakar/ /wekel/ /yer/

‘call’ ‘yawn’ ‘snap’ ‘sing’ ‘crawl’ ‘round’ ‘grow’ ‘walk’ ‘hang’ ‘fruit’ ‘smart’ ‘sick’ ‘burp’ ‘fall over’ ‘fight’ ‘climb’ ‘flow’ ‘laugh’

Prefix possible in AGR1 /aumen/ /ber/ /ket/ /kui/ /ladi/ /mai/ /masnin/ /mid/ /mit/ /mur/ /pan/ /posin/ /seser/ /soron/ /tai/ /tain/ /tati/ /wel/

‘scared’ ‘strong’ ‘defecate’ ‘fart’ ‘shy’ ‘come’ ‘hungry’ ‘go up’ ‘sit’ ‘run’ ‘dead’ ‘pretty’ ‘shake’ ‘angry’ ‘sleep’ ‘fall down’ ‘stand’ ‘bathe’

482

Antoinette Schapper & Rachel Hendery

Within the class of intransitive verbs which can take a prefix in AGR1, there is a further division between those verbs which always take an AGR1 prefix and those that have differential S-marking based on animacy. Most verbs taking an AGR1 prefix belong to the first subset; they invariably have a prefix co-indexing their S participant. For instance, the verb tati ‘stand’ in (108) has a prefix both when S is inanimate (108a) as well as animate (108b). (108)

a.

b.

Ira botil yoku no getati. ira botil yoku no water bottle two one ‘Two bottles stand.’

g-tati 3-stand

Mei poko getati gebir ak. mei pok g-tati g-bir ak female small 3-stand 3-mouth open.mouth ‘A girl stands with her mouth open.’

A small number of verbs in our corpus show differential S-marking. The semantics of differential S-marking verbs often mean that inanimate Ss are infrequent and difficult to obtain in elicitation. Their status as differential S-marking verbs only becomes apparent in certain constructions. For instance, a motion verb such as mai ‘come’ readily occurs with an animate S in a mono-predicative clause such as (109a). The same mono-predicative clause with an inanimate S is not accepted by speakers (109b). The rejection of such examples is presumably due to the unlikeliness of an inanimate putting itself in motion. However, it is not because of the impossibility of having an inanimate S, as we see in the SVC in (109c). Here we have a “switch-subject” SVC in which beting /betin/ ‘bamboo’ is the P of the first verb ong /on/ ‘use’ and the S of second verb mai and which has causative meaning ‘take the bamboo (and make it) come’ i.e. ‘bring’ (see section 7.1 for more on such SVCs). (109)

a.

Daud gamai. Daud g-mai David 3-come ‘David came.’

b.

#/*Beting mai. betin mai bamboo.sp come Intended: ‘The bamboo came.’

483

Wersing

c.

Beting ba ong mai bong ba ladu. betin ba on mai bon ba laadu bamboo.sp DEF use come tree DEF lean ‘The bamboo is brought and leant on the tree.’

In short, there are three types of intransitive verbs in Wersing: (i) those never indexing S with an AGR1 prefix; (ii) those always indexing S with an AGR1 prefix, and; (iii) those indexing animate Ss, but not inanimate Ss with a prefix in AGR1 slot. The first and second of these subclasses are well represented in our corpus. The third class has only a handful of members, but we suspect that a larger corpus of natural texts would bring many more differential S-marking verbs to light.

6.1.2 Agreement on transitive verbs We observe three distinct agreement patterns in the AGR1 slot on transitive verbs. These are: (i) no prefix in AGR1, (ii) prefix in AGR1 for P, and (iii) prefix in AGR1 for A. Table 16 presents some examples of each type from our corpus. Table 16: Examples of Wersing transitive verbs and prefixal possibilites No prefix in AGR1 ‘take’ ‘reach’ ‘do work’ ‘hunt’ ‘be at’ ‘lift’

/medi/ /tain/ /blet/ /lau/ /min/ /semel/

Prefix in AGR1 for P /-dir/ /-lol/ /-maksebi/ /-pai/ /-poin/ /-wuni/

‘see’ ‘follow’ ‘ask’ ‘make, do’ ‘hit’ ‘hide’

Prefix in AGR1 for A /-nai/ /-tme/ /-amalɔsi/ /-mra/ /-pasi/

‘drink/eat’ ‘hear’ ‘not want’ ‘want’ ‘chew’

Sub-classes (i) and (ii) have many members in our corpus. In (110), we illustrate how the class (i) transitive verb medi ‘take’ has no prefix regardless of the animacy of P. In (111), we show how the class (ii) transitive verb -dir ‘see’ has a prefix for both animate and inanimate Ps.

(110)

a.

Class (i) transitive verb Naida ol poko medi. naida ol pok medi 1SG.TOP child little take ‘I fetched the child.’

484

Antoinette Schapper & Rachel Hendery

b.

(111)

a.

b.

Naida ira medi. naida ira medi 1SG.TOP water take ‘I fetched water.’ Class (ii) transitive verb Naida ol poko gidir. naida ol pok g-dir 1SG.TOP child little 3-see ‘I saw the child.’ Naida ira gidir. naida ira 1SG.TOP water ‘I saw water.’

g-dir 3-see

Sub-class (iii) transitive verbs take an AGR1 prefix for A. The class is small; our corpus has around a dozen verbs in the class. Semantically, verbs of this class have an experiencer as A, such as -nai ‘drink/eat’ (112) and -tme ‘hear’ (113). However, by no means all verbs with experiential As belong to this class (e.g., the class (ii) transitive verb -dir ‘see’ above).

(112)

(113)

Class (iii) transitive verbs Newe ira nenai. newe ira n-nai 1SG.FOC water 1SG-eat ‘I drink water.’ Geme teteme. g-me t-tme 3-voice 1PL.INCL-hear ‘We hear his voice.’

Our understanding of the alignment of the Wersing verb thus far can be summarized as follows. All three arguments (A, P or S) can be indexed on the verb by a prefix in AGR1 slot. But each verb only allows one of these arguments to be marked on the verb. There is some differerential marking of S on the basis of animacy. It remains to be determined whether there is any differential marking of P or even A.

Wersing

485

6.2 Applicative prefixes Wersing has two (productive) applicative prefixes: le- and mi-. These are most commonly prefixed on intransitive verbs. When the verb has a prefix for S in AGR1, it is not affected by the prefixation of the applicative (compare 114a and 114b). Without the applicative, the 2nd participant cannot be included in the clause (114c). (Note that cankir in these examples is an Indonesian loan). (114)

a.

Cangkir ba gemita. cankir ba g-mit-a DEF 3-sit-REAL cup ‘There’s the cup’, literally, ‘The cup sits’.

b.

Cangkir ba mede legemita. cankir ba med le-g-mit-a DEF table APPL-3-sit-REAL cup ‘There’s the cup on the table.’, literally, ‘The cup sits on the table.’

c. *Cangkir ba mede gemita. cankir ba med g-mit-a DEF table 3-sit-REAL cup Occasionally, we do find an applicative prefixed onto a transitive verb. In this case, the applied location takes the place of the P and the original P is “bumped” to be the P of medi ‘take’, as in (115). (Compare with three-place verbs described in section 3.1.1). (115) Lemi murung poko no gai bong ba medi gatal legopoing lemang podo. lmi murun pok no gai bon ba medi g-tal le-g-poin male small little one 3.A tree DEF take 3-thigh APPL-3-hit leman pod long broken ‘A small boy takes a stick and hits it against his thigh until it is broken.’ Whilst applicativization is the main function of these prefixes, each has additional non-applicativizing functions that will also be discussed and illustrated briefly in the following sections.

486

Antoinette Schapper & Rachel Hendery

6.2.1 LeWe have idenitified three applicativizing uses for le-. The first is to add a participant, typically human, towards whom the action denoted by the verb is directed. For instance, yer ‘laugh’ and aloi ‘bark’ are basically intransitive verbs (116a) and (117a). When prefixed with le-, a participant who is being laughed or barked at is added to the clause (116b) and (117b). (116)

(117)

a.

Naida yer noko. naida yer n-ko 1SG.TOP laugh 1SG-stay ‘I was laughing.’

b.

Naida Benny leyer noko. naida Benny le-yer 1SG.TOP Benny APPL-laugh ‘I was laughing at Benny.’

n-ko 1SG-stay

a.

Kutual aloi. kutual aloi dog bark ‘The dog barks.’

b.

Kutual ol pok deing lealoi. kutual ol pok dein le-aloi APPL- bark dog child small PL ‘The dog barks at the children.’

The second applicativising use is to add a location on which the S participant is positioned. The locative relation denoted by le- is placement on top of (118), above (119) or attachment to a vertical surface (120). (118)

Buku ba mede legatai. buku ba med le-g-tai book DEF table APPL-3-lie.down ‘The book is on top of the table.’

(119)

Aya keng kum no atoku bobung legemit. aya ken kum no atok bobun le-g-mit APPL-3-sit rain cloth round one earth peak ‘There is a round cloud above a mountain peak.’

Wersing

(120)

487

Keng badu ba kepe lelakala. ken badu ba kpe le-lakal-a cloth shirt DEF rope APPL-hang-REAL ‘The clothes are hanging from a line.’

Finally, le- can also introduce a cause for the event denoted by the verb. This use is particularly common on spontaneous bodily action verbs, such as with dudung /dudun/ ‘cough’ in (121) and mokor /mokor/ ‘vomit’ in (122). (121)

Naida tebaku ba ledudung. naida tbaku ba le-dudun 1SG.TOP cigarette DEF APPL-cough ‘I cough on account of smoking.’

(122)

Durki gapangte naida gemelasur ba lemokor. durki g-pan=te naida g-mlasur ba le-mokor rat 3-dead=CONJ 1SG.TOP 3-rotten DEF APPL-vomit ‘A mouse has died and I vomit on account of its stench.’

6.2.2 MiThe prefix mi- in its most frequent use is a locative applicative denoting a containment relationship. That is, mi- denotes that the S is located within or around the location introduced by the applicative prefix, as in (123) and (124) respectively. (123)

Wai aka mira migetati. wai aka mira mi-g-tati goat fence inside APPL-3-stand ‘The goat is inside the fence.’

(124)

Pita ba lilin getiting-titing mipas geti. pita ba lilin g-titin~titin mi-pas g-ti ribbon DEF candle 3-middle~RDP APPL-tie 3-put ‘The ribbon is tied around the middle of the candle.’

Mi- also has a use which does not add a locative participant, but instead denotes further movement into a state, that is, an increase in the state experienced by S, as in (125) and (126).

488

Antoinette Schapper & Rachel Hendery

(125)

David gai gepai minosorong. David gai g-pai mi-n-soron David 3.A 3-make APPL-1SG-angry ‘David makes me angrier.’

(126)

Taida oba misus. taida o-ba mi-sus 1PL.INCL.TOP PROX-DEM APPL-suffer ‘Here we are in more difficulties.’

6.3 Agreement slot 2 The third and outermost slot is the AGREEMENT SLOT 2 (AGR2). AGR2 is only filled when the applicative prefix le- occurs in the APPL slot. Where AGR1 and AGR2 are both filled, they always have distinct referents. The referent of the AGR2 prefix is always the same as the P applied by le-. In (127a) we see that the prefix n- in AGR1 slot indexes the S of -mur ‘run’. In (127b) the applicative prefix le- allows a second participant, the cause of the motion, the speaker’s mother, to be added to the clause. We see that this second participant can be encoded either with or without a prefix in AGR2, (compare 127b and 127c). The difference between examples such as (127b) without an AGR2 prefix and (127c) with an AGR2 prefix remains for future investigation. (127)

a.

Naida nemura. naida n-mur-a 1SG.TOP 1SG-run-REAL ‘I run.’

b.

Naida nea lenemura. naida ne-a le-n-mur-a 1SG.TOP 1SG-mother APPL-1SG-run-REAL ‘I run from my mother.’

c.

Naida nea galenemura. naida ne-a ga-le-n-mur-a 1SG.TOP 1SG-mother 3-APPL-1SG-run-REAL ‘I run from my mother.’

6.4 Agreement slot 3: Inceptive prefixation Wersing also allows marking for inceptive aspect, which identifies the beginning stage of an action or situation. Marking for this aspect is by means of AGREEMENT SLOT 3 (AGR3). An inceptive prefix is never obligatory.

Wersing

489

Compare the marking of the verb -tati ‘stand’ in (128). In (128a) where the verb shows the normal agreement pattern, we have a straightforward sequence of events whereby the cat moves into a standing position and then jumps on its prey. In (128b) where the AGR3 prefix is present on -tati, the act of standing is only in its first phase: the cat holds itself in a half-standing position, poised to jump. (128) a. Akamau getati ba, kengel gewai durki wing gewena. akamau g-tati ba kenel g-wai durki win gewen=a cat 3-stand DEF jump 3-go mouse grab 3.POSS=SPEC ‘The cat stands and jumps to catch the mouse.’ b. Akamau gegetati ba, kengel gewai durki wing gewena. akamau ge-g-tati ba kenel g-wai durki win gewen=a cat 3INCP-3-stand DEF jump 3-go mouse grab 3.POSS=SPEC ‘The cat is half standing ready to jump to catch the mouse.’ Inceptive aspect in Wersing thus refers to an as yet unrealized situation. Accordingly, it is incompatible with realis aspect marking, as in (129a). The clause can only be grammatical either without the inceptive marking (129b) or without the realis marking (129c). (129)

a. *Nenemita. ne-n-mit-a 1SG.INCP-1SG-sit-REAL b.

Nemita. n-mit-a 1SG-sit-REAL ‘I’m sitting.’

c.

Nenemit. ne-n-mit 1SG.INCP-1SG-sit ‘I’m half sitting, getting into a sitting position.’

In the examples we have looked at thus far inceptive aspect marking occurs on posture verbs with an animate S and therefore there is always an agreement prefix present on the verb in our examples. It is, however, also possible to find inceptive prefixation on verbs that do not typically co-index their S with a prefix.

490

Antoinette Schapper & Rachel Hendery

For example, the verb lailol ‘walk’ normally occurs without a prefix as in (130a). It can occur with inceptive aspect marking from AGR3 (130b), but cannot occur with any other prefixes such as in AGR1 as in (130c). (130)

a.

Naida lailol. naida lailol 1SG.TOP walk ‘I’m walking.’

b.

Naida nelailol. naida ne-lailol 1SG.TOP 1SG.INCP-walk ‘I’m about to get walking.’

c. *Naida nenelailol. naida ne-n-lailol 1SG.TOP 1SG.INCP-1SG-walk Intended: ‘I’m about to get walking.’

7 Serial verb constructions Like other TAP languages, Wersing makes use of serial verb constructions for a range of purposes. A serial verb construction in Wersing is where more than one verbal predicate occurs within the scope of a single realis suffix under a single intonation contour. In this section, we give a preliminary account of the most frequently occurring in the language. Further investigation into the language will no doubt reveal other serializations that are not present in our small corpus. See section 8.1 on the realis suffix.

7.1 Instrumental serialization In Wersing an NP encoding an instrument can be introduced into a clause by means of one of two verbs: ong /on/ ‘use’ or medi /medi/ ‘take’. In the following subsections, we will look at the different serialization contexts in which these two verbs occur.

Wersing

491

7.1.1 ong ‘use’ As an independent clausal verb ong /on/ means ‘use’, illustrated in (131). There are similarly used cognates in other eastern Alor languages, but elsewhere in the family cognates have meanings such as ‘do’ and ‘make’ (e.g., Abui on ~ ong ‘make’ Kratochvíl 2007: 375, Bunaq hoqon ‘make’ and on ‘do’ Schapper 2010: 491–495). (131)

Nepa bong koli ba, pede pilit ong. n-pa bon koli ba pde plit on 1SG-father tree cut.down DEF machete sharp use ‘When my father cuts wood, he uses a sharp machete.’

Like the other languages of eastern Alor, Wersing also uses ong to encode instruments. In general terms, the element introduced by ong expresses the means by which an event is achieved, whether this is a tool (132), vehicle (133), or a part of the item used to achieve its final state (134). (132)

Lemi poko kinai ong keng ba gepesi burika. ba g-pesi burik-a lmi pok [kinai on]INSTR ken man little knife use cloth DEF 3-cut snap-REAL ‘A young man cuts the cloth with a knife until it breaks.’

(133) Mademang mi newaka ong lailol Peitoko gelaitaing newai Pureman. Mademang mi [n-wak=a on] lailol Peitoko g-laitain Mademang LOC 1SG-foot=SPEC use walk Peitoko 3-go.via n-wai Pureman 1SG-go Pureman ‘I go by foot from Mademang to Pureman via Peitoko.’ (134)

Botil yoku no gebir ong medi, pangsa botil noa gekureng ong medi. botil yoku no [g-bir on] medi pan=sa bottle two one 3-mouth use take CONC=CONJ botil bottle

no=a one=SPEC

[g-kuren 3-bottom

on] use

medi take

‘Two bottles are put on their ends, while one is put on its bottom.’ There are also idiomatic verb serializations such as on mai ‘bring’ (lit: ‘use come’), illustrated in (135).

492 (135)

Antoinette Schapper & Rachel Hendery

Beting ba ong mai bong ba ladu. betin ba on mai bon ba ladu bamboo.sp DEF use come tree DEF lean ‘The bamboo is brought and leant on the tree.’

7.1.2 medi ‘take’ As an independent verb medi means ‘take’, as in (136). (136)

David, pede pilit motumo mina medi. David pde plit motumo min=a medi David machete sharp beneath be.at=SPEC take ‘David, take the sharp machete that’s down below.’

In combination with other verbs, it can be used to encode the means by which an event is achieved, as in (137). This is a similar function to ong, and the exact semantic range of each remains a question for future research. (137)

a.

Sobo ba aka medi lebelebel. sob ba [aka medi] le-belebel APPL- encircle house DEF fence take ‘The house is surrounded with a fence.’

b.

Gaida gekureng pilit ba medi atoku wodo geya. gaida [g-kuren plit ba medi] atok earth 3.TOP 3-bottom sharp DEF take ‘He spiked the ground with his sharp bum.’

wod pound

g-ya 3-go

As was mentioned and illustrated already in 3.1.1 medi is frequently (although optionally) used to encode the third participant of a three-place predicate, and as discussed in section 6.2, this is also possible when the three-place predicate is formed by the affixation of le- to a transitive verb. We do not find ong used in these contexts.

7.2 Aspectual serialization 7.2.1 Continuative -ko The verb -ko is used as an independent verb to mean ‘stay, live’ and agrees with the S/A (138).

Wersing

(138)

a.

b.

493

Independent uses of -ko Getamu de goko. g-tamu de g-ko 3-grandparent IPFV 3-stay ‘His grandparent is still living.’ Naida noko yona. naida n-ko yona 1SG.TOP 1SG-stay LOW ‘I am living/staying down there.’

In serialization, -ko is used to mark continuative aspect, that is, denoting persistent or repetitive action (139). This serial verb maintains agreement with the subject and follows the semantically main verb of the clause.

(139)

a.

b.

Serial uses of -ko David wortel alai goko. David wortel alai g-ko David carrot slice 3-stay ‘David slices the carrot repeatedly.’ Naida aki noko. naida aki n-ko 1SG.TOP call 1SG-stay ‘I call out repeatedly.’

7.2.2 Completive & quantitative lai Lai is used as an independent clausal verb to mean ‘finished’, as in (140).

(140)

Independent use of lai Basar ba lai. basar ba lai market DEF finish ‘The market was finished.’

First, it is used in serialization to mark completive aspect. Lai follows the semantically main verb of the clause, as in (141) and (142).

494

Antoinette Schapper & Rachel Hendery

(141)

Andreas ba gewili lai. Andreas ba g-wil lai Andreas DEF 3-bathe finish ‘Andreas finished washing.’

(142)

Mei poko kander legemiti lai, . . . mei pok kander le-g-mit lai APPL-3-sit finish female little chair ‘The girl finished sitting, . . .’ (i.e., she stood up)

In serialization with verbs denoting a state, lai denotes not that the state is finished, but that it has been entered into completely, as in (143) and (144). (143)

Keng badu ba inali lai. ken badu ba inali lai cloth shirt DEF clean finish ‘The clothes are completely clean.’

(144)

Kutual ba ira ko mitaingte ires lai. kutual ba ira ko mi-tain=te ires lai DEF water pond APPL-fall=CONJ wet finish dog ‘The dog fell into the pond and then was totally wet.’

7.2 Manner serialization Manner serialization is used to express the manner in which an action is performed. In this, the verb expressing manner is always intransitive, while the action verb can be intransitive or transitive. Where the action verb is intransitive, the manner verb occurs directly before it, as with pok ‘small’ in (145). Where the action verb is transitive, the manner verb occurs directly before the P, as with tebel ‘wrong’ in (146). (145)

Gaida poko er. gaida pok er 3.TOP little crawl ‘He crawled a little way.’

(146)

Ol pok deing tebel peter mada. ol pok dein tebel pter mad wrong corn cook child small PL ‘The small children were cooking corn the wrong way.’

Wersing

495

7.3 Motion serialization Two types of motion serialization are observed in the corpus. Both are found throughout the TAP languages. First, there is motion-action serialization. As in other TAP languages, this involves a motion verb followed by a verb denoting an action which the motion led up to, as in (147) and (148). (147)

Neya pei lau ge. n-ya pei lau ge 1SG-go pig hunt PROSP ‘I’m going to go hunt pigs.’

(148)

Yemai teya sirwis belet. y-mai t-ya sirwis blet 2PL-come 1PL.INCL-go work do.work ‘You come and we will go do work.’

Second, directional serialization is where one verb denotes the direction in which the action denoted by another verb is done. The directional verb provides information about the orientation of an event as either towards or away from the deictic center and as either up, down or on the same elevation as the deictic center. The directional verb directly follows the action verb for which it provides the orientation. Directional verbs are always intransitive, while the preceding action verbs can be either intransitive (149) or transitive (150). (149) Tekiling ba ire gesir durki wing laia, war ire gemida. tkilin ba ire g-sir durki win lai-a war ire g-mid-a eagle DEF fly 3-come.down mouse grab finish-REAL and fly 3-go.up-REAL ‘The eagle flies down, grabs the mouse and then goes up again.’ (150)

Ota bong ba modo, gesiriping genaing beteng gedai. ota bon ba mod g-sripin gnain beten g-dai 3-root 3.ALL pull.out 3-go.up coconut tree DEF fall ‘When the coconut tree fell, all its roots were pulled up.’

7.4 Causative serialization Causative serialization in Wersing makes use of the 3rd person inflection of the verb -pai ‘make’ followed by an intransitive verb denoting the caused situation.

496

Antoinette Schapper & Rachel Hendery

The causer is encoded as the A of -pai. This is indicated by the marking of the causer with gai (see section 5.1) in (151). There is no referent for the P of -pai; the g- is a dummy prefix. The causee is encoded as the S of the following verb, -ladi ‘shy’ and -seser ‘shake’ in (151) and (152) respectively. (151)

Ol pok deing gai gepai neladi. ol pok dein gai g-pai n-ladi 3.A 3-make 1SG-shy child small PL ‘The children make me shy.’ Literally, ‘The children make it (that) I am shy.’

(152)

Oto lailol, gepai neseser. oto lailol g-pai n-seser car walk 3-make 1SG-shake ‘The car drives along and makes me shake.’ Literally, ‘The car travels (and) makes it (that) I am shaken.’

The Wersing causative SVC differs from that in other TAP languages where the causee is encoded as the P of the causative verb (e.g., Kamang -baa, Bunaq -ini).

7.5 Resultative serialization Resultative serialization minimally involves a verb denoting an event and a following verb denoting the situation resulting from that event. A result is typically encoded with an intransitive verb and the event leading up to it by a transitive. For instance, tmes and burik in (153) and (154) denote the results of the hitting (/-poin/) and pulling (/-meken/) events respectively. (153)

Saku poko gai wortel ba geteng ong gopoing temes. saku pok gai wortel ba g-ten on g-poin tmes squashed adult small 3.A carrot DEF 3-hand use 3-hit ‘A young adult hits the carrot with his hand until it is squashed.’

(154)

Lemi poko gai kepe ba gemekeng burika. lmi pok gai kpe ba g-meken male small 3.A rope DEF 3-pull ‘A boy pulls on the rope until it snaps.’

burik-a snapped-REAL

Wersing

497

We identify a second kind of resultative construction involving locations resulting from motion events. This was discussed briefly in section 3.1.4.1.

8 Realis and aspect marking Grammaticalized TAME marking in Wersing is limited to: (i) a suffix marking realis mood on final verbal predicates (section 8.1), and; (ii) aspectual particles marking imperfective, perfective and prospective aspect (section 8.2). There do not appear to be grammaticalized morphemes marking tense or evidentiality in Wersing. Time can be given in a clause by means of temporal adverbs (e.g., mlenya ‘yesterday’, mdali ‘tomorrow’ etc.).

8.1 Realis suffix -a Final verbal predicates can be marked with the realis suffix -a in Wersing. Nonfinal verbal predicates cannot be marked with the realis suffix. As such, in (148) the only verb in the three verb SVC that can host -a is lai (155a); suffixing -a on the non-final verbs kengel and wing produces ungrammatical utterances, as in (155b) and (155c).13 (155)

a.

Akamau kengel durki wing laia. akamau kengel durki win lai-a cat jump mouse grab finish-REAL ‘The cat jumped and grabbed the mouse.’

b. *Akamau kengel durki wina lai. akamau kengel durki win-a cat jump mouse grab-REAL c. *Akamau kengela durki wing lai. akamau kengel-a durki win cat jump-REAL mouse grab

lai finish lai finish

The status of -a marking realis is seen in its inability to occur on verbs in clauses with two characteristics. Firstly, -a ‘REAL’ cannot mark a verb in a clause 13 Note that (155c) could be made grammatical by placing an intonation break after kengel-a ‘jump-REAL’. However, such a break would mean that the sentence is bi-clausal, with kengel being the final verb of the first clause.

498

Antoinette Schapper & Rachel Hendery

marked with one of the negators discussed in section 3.2. Compare the grammaticality of the clauses in (156). In (156a) we see that -a can occur on the verb ming /min/ ‘be at’. This verb can also be negated with nanu (156b), but the realis suffix cannot then also occur on the verb at the same time (156c). Other negators are similarly not permissible (156d–156e). (156)

a.

Naida sobo mina. naida sob min-a 1SG.TOP house be.located-REAL ‘I am in the house.’

b.

Naida sobo ming nanu. naida sob min 1SG.TOP house be.located ‘I am not in the house.’

nanu NEG

c. *Naida sobo mina nanu. naida sob min-a 1SG.TOP house be.located-REAL

NEG

d. *Naida sobo mina naung. naida sob min-a 1SG.TOP house be.located-REAL

NEG

e. *Naida sobo mina sining. naida sob min-a 1SG.TOP house be.located-REAL

PROH

nanu

naun

snin

Secondly, -a cannot mark a verb in a clause marked with the prospective aspect marker (section 8.2.3). Again compare the grammaticality of the clauses in (157). Whilst either -a ‘REAL’ or ge ‘PROSP’ can each mark a clause (157a and 157b respectively), they cannot occur together in the same clause (157c). However, -a ‘REAL’ can be used with other kinds of aspect markers so long as they do not entail an irrealis meaning (e.g., perfective in 150d or imperfective in 157e). See section 8.3 on the non-cooccurrence of inceptive and realis marking. (157)

a.

David geya pei laua. David g-ya pei lau-a David 3-go pig hunt-REAL ‘David’s gone hunting pigs.’

Wersing

b.

499

David geya pei lau ge. David g-ya pei lau ge David 3-go pig hunt PROSP ‘David’s going to go and hunt pigs.’

c. *David geya pei laua ge. David g-ya pei lau-a ge David 3-go pig hunt-REAL PROSP ‘David’s going to go and hunt pigs.’ d.

David geya pei laua kana. David g-ya pei lau-a kana David 3-go pig hunt-REAL PFV ‘David has already gone to hunt pigs.’

e.

David de geya pei laua. David de g-ya pei lau-a David IPFV 3-go pig hunt-REAL ‘David is gone hunting pigs.’

In Wersing marking a verb with the realis suffix does not appear to be obligatory in any circumstances. At this stage, we do not understand what semantic-pragmatic factors underpin the use of the realis suffix.

8.2 Aspectual markers Wersing has a three-way aspectual distinction of imperfective, perfective and prospective aspect. The markers of these aspects do not co-occur, but do to some extent interact with the realis suffix. The individual properties of each aspect marker are discussed in the following sections.

8.2.1 de ‘IPFV’ The imperfective aspect marker in Wersing is de ‘IPFV’. It occurs in a pre-verbal position. Typically de occurs directly before the predicate, kang /kan/ ‘good’ in (158). Occasionally, however, de occurs before an NP, as with the subject king /kin/ ‘breath’ in (159).

500

Antoinette Schapper & Rachel Hendery

(158)

David gemira de kana. David g-mira de kan-a David 3-inside IPFV good-REAL ‘David still has a good heart.’

(159)

Pei ba mona de geking sesai. pei ba mona de g-kin sesai pig DEF LEVEL IPFV 3-breath breathe ‘The pig (that is) over there is still breathing.’

When de occurs in a negative marked clause, the overall meaning is ‘not yet’. This is illustrated in (160). (160)

Lampu ba de gepai medal naungte de iding. lampu ba de g-pai mdal naun=te DEF IPFV 3-make shine NEG = CONJ lamp ‘The lamp is not yet alight so it’s still dark.’

de IPFV

idin. dark

8.2.2 kana ‘PFV’ The perfective aspect marker in Wersing is kana ‘PFV’. It is used in reference to an event that has ‘already’ been achieved by a reference time. For instance, kana in (161b) denotes that, once the head has become stuck in the ground (161a), the situation of not being able to pull it out was then achieved. (161)

a.

b.

Gekureng pilit ba medi atoku wodo geya. g-kuren plit ba medi atok wod 3-bottom sharp DEF take earth spike ‘He spiked the ground with his sharp bum.’

g-ya 3-go

Podote beteng boli kana. podot=te beten boli kana long.time=CONJ pull.out bad PFV ‘In no time he couldn’t pull out [his head].’

The perfective marker readily combines with the serial verb lai ‘finish’ (section 7.2.2), as in (162). Here the perfective marker denotes that the finishing is complete and lai that the calling is completed.

Wersing

(162)

501

Naida aki lai kana. naida aki lai kana 1SG.TOP call finish PFV ‘I finished calling already.’

No cognate perfective or semantically similar marker is known in the family. It may be that kana historically comes from kang ‘good’ suffixed with -a ‘REAL’.

8.2.3 ge ‘PROSP’ In Wersing prospective aspect is marked with the postverbal particle ge ‘PROSP’. The prospective aspect marker signals that the event denoted by a clause takes place at an undetermined future time, (163) and (164). The irrealis status of clauses marked with ge and their incompatibility with -a ‘REAL’ has already been discussed in section 8.1. (163)

David gedai tro migoko ge? David g-dai tro mi-g-ko ge David 3-come.up where APPL-3-stay PROSP ‘When David comes, where is he going to stay?’

(164)

Naida ira ba nenai teng ge. naida ira ba n-nai 1SG.TOP water DEF 1SG-eat ‘I will drink that water later.’

ten

ge

FIRST

PROSP

9 Discussion This sketch by no means represents the last word on Wersing, but it does permit us to place the language within the family. Wersing is in many respects simpler than its relatives on Alor, but is average within the Timor-Alor-Pantar family in terms of its phonological and morpho-syntactic complexity. Many of its distinctive features it shares in common with its neighbors and nearest relatives, Kula and Sawila. Wersing has a straightforward segmental phonology (sections 2.1 and 2.2), but surprisingly elaborate phonotactics (section 2.4): processes of echo-vowel insertion (section 2.4.2) and vowel epenthesis (section 2.4.1) work to create surface syllables of CV from underlying consonant clusters and closed syllables.

502

Antoinette Schapper & Rachel Hendery

Echo vowel insertion such as Wersing displays likely represents the historical antecedent of the final versus non-final word alternations that are observed in neighboring Kula and Sawila. The simple five-vowel system of Wersing (section 2.1) is also notable in that it contrasts with the majority of other languages in Alor. Historically, there appears to have been a five vowel system with a length contrast for each vowel in the languages of central-eastern Alor (Abui, Kamang, Sawila, Wersing, Kula). While Abui, Kamang and Sawila maintain the length contrast, Wersing has lost it, with historical short unstressed vowels being deleted. The result of this is the many underlying consonant clusters that we observe in Wersing today. The high vowel metathesis found in Wersing (section 2.5.2) is, however, an areal feature shared with Kula and Sawila. Whilst its pronominal paradigms are typical in number and function (section 5), the Wersing prefixal agreement system lacks the semantic flexibility and multiple distinct agreement paradigms found in central and western Alor (section 6). However, like the other eastern Alor languages, it allows prefixing not only of S and P but also of A (section 6.1.2). At the same time, Wersing is distinct from these in having a separate agreement paradigm and associated slot on the verb for marking inceptive aspect (section 6.4) and in its lack of the (remnant) inverse marking that is found in Kula and Sawila. Differential marking of S on the basis of animacy in Wersing is unique in the family (section 6.1.1), with other languages such as Teiwa and Bunaq largely restricting differential marking to P. Wersing shares suffixal realis marking on final verbs with Sawila and Kula (section 8.1), but lacks an elevational system containing diagonal terms (section 3.1.3) and complex demonstrative paradigm (section 4.3) such as they have.

References Donohue, Mark. 1997. Kolana: syllables, (small) words and CV. Unpublished manuscript, The Australian National University. Donohue, Mark. 2004. Typology and Linguistic Areas. Oceanic Linguistics 43(1): 221–239. Donohue, Mark. 2008. Semantic alignment systems: what’s what, and what’s not. In Mark Donohue & Søren Wichmann, The Typology of Semantic Alignment, 24–76. Oxford: Oxford University Press. Hendery, Rachel. 2012. Relative clauses in time and space. Amsterdam: John Benjamins. Holton, Gary, Marian Klamer, František Kratochvíl, Laura Robinson & Antoinette Schapper. 2012. The historical relation of the Papuan languages of Alor and Pantar. Oceanic Linguistics 51(1): 86–122. Kratochvíl, František. 2007. A Grammar of Abui. PhD dissertation, Leiden University, Leiden.

Wersing

503

Langko, Fredik. n.d. Bahasa Daerah Kolana: Diri Sendiri. Unpublished manuscript. Kalabahi & Kolana, Alor Indonesia. Lewis, M. Paul, Gary F. Simons, and Charles D. Fennig (eds.). 2013. Ethnologue: Languages of the World, Seventeenth edition. Dallas, Texas: SIL International. Online version: http:// www.ethnologue.com. Malaikosa, Anderias. n.d. Yesus Sakku Geleworo Kana. Unpublished manuscript. UBB-GMIT Kupang, Indonesia. Schapper, Antoinette. 2007. Grammaticalization of demonstratives in the languages of Timor, Alor and Pantar. Paper presented at the Fifth International East Nusa Tenggara Conference, Kupang, 1–3 August 2007. Schapper, Antoinette. 2010. Bunaq, a Papuan language of central Timor. PhD thesis, The Australian National University, Canberra. Schapper, Antoinette. 2011a. Crossing the border: Historical and linguistic divides among the Bunaq in central Timor. Wacana, Journal of the Humanities of Indonesia 13(1): 29–49. Schapper, Antoinette. 2011b. Iconicity of sequence in the coding of source and goal in two Papuan languages. Linguistics in the Netherlands 2011: 101–113. Schapper, Antoinette & Marian Klamer. 2011. Plural words in the Alor-Pantar languages. Proceedings of Conference on Language Documentation and Linguistic Theory 3. Schapper, Antoinette & Marian Klamer. forthcoming. Numeral systems in Alor-Pantar languages. In Marian Klamer (ed.), The Alor-Pantar languages: History and Typology. Berlin: Language Science Press. http://langsci-press.org/. Wellfelt, Emilie & Antoinette Schapper. 2013. Enriching the archival picture: The use of local sources in assessing the nature and impact of historical contact. Paper presented as part of the Panel on East Timor, the Portuguese, and the Indonesian Archipelago – Historical Sources and Methodological Reconsiderations at the International Convention of Asia Scholars 8, Macau, 24–27 June.

Index Abui 6, 10, 11, 12, 13, 14, 15, 287, 290, 348, 441, 491, 502 Abui-Papuna 6 accusativity, see morphological alignment Adang 11, 12, 13, 100, 103, 145, 216, 221– 283, 314 adjectives 178–179, 188, 303 – and derivation 155, 161, 172, 217, 303, 325 – and reduplication 155, 156, 157, 156–159, 178, 294, 295, 333 – lack of adjective class 14, 120, 307, 325, 381, 454 – separate class of adjectives 55, 166, 178– 179, 302, 309 adverbs 49–51, – aspect marking 89, 141, 143, 304 – deictic 159 – manner 111, 217, 304 – modality 111, 273 – temporal 304, 497 agriculture 7, 9, 26, 353 alignment, see morphological alignment Alor-Pantar subgroup 3, 5, 6, 7, 8, 10, 12 – features typical of Alor-Pantar languages 27, 100, 145, 232, 282, 355, 429 – differences from Alor-Pantar languages 94, 95, 100, 224, 278 – see also proto Alor-Pantar apocope 34 applicatives 13 – auxiliary 87 – historical development of 429, 454, 455 – in verb classes 407, 408–409, 411–416, 418, 485–488 – prefixes 270–271, 327–328, 397–403, 407, 479 apposition – locationals and NPs 70 – pronouns and NPs 53, 54, 74–5, 313–314, 367, 368, 391, 472 – quantifiers and NPs 56, 308, 313 articles 59, 66, 306, 308, 309, 310–311, 456, 465, 469–470

aspect – atelic 203 – continuative 43, 89, 141, 143–144, 156, 342–343, 492–493 – imperfective 89–90, 143–144, 298, 334– 335, 342, 343, 373, 419–420, 498–499 – inceptive 91–94, 280, 282, 480, 488– 490, 498 – neutral 203 – perfective 84, 89–90, 143–144, 204, 280–281, 334–335, 342 419–420, 499–501 – progressive 91, 93–94, 164, 203, 213– 214, 274, 280, 342–343, 373 – stative 78, 193, 217, 280, 307, 325, 334– 335, 342, 373, 454 – telic 203 associative plural, see plural number marking attribution/nominal attributes 14, 55, 63, 119, 120, 163, 247–248, 306, 307 Austronesian 1, 3, 7, 9, 10, 152, 216, 361 benefactives/beneficiaries 169, 171–172, 270, 321, 331–332, 397, 429–430 bilingualism/multilingualism 3, 26, 149, 222, 353 Bird’s Head 216, 222 Blagar 6, 10, 12, 13, 15 – comparison with other languages 98, 100, 103, 132, 145, 147–219, 258, 282, 314 body part nouns see possession Bol 6 borrowing see loans Bukalabang 11, 152 Bunaq 1, 3, 10, 11 – comparison with other Alor-Pantar languages 12, 13, 14, 15, 467, 491, 496, 502 case 14, 119, 280 causatives 166, 168, 172, 198, 273, 341 – in serial verb constructions 139, 277–278, 341–342, 344–345, 495–496

506

Index

– prefixes 13, 192, 193–195, 197, 266, 271– 273 – suffixes 13, 132, 141, 160 – unmarked 145 change, see language change classifiers – noun 248–252 – numeral 54, 56–57, 119–124, 248–252, 263, 308–309, 386, 467–468 – possessive 14, 15 clitics 211, 297, 462 – allomorphs 29, 33, 34 – and postpositions 173, 199, 200, 349 – and stress 38–39, 155 – article 452, 456 – demonstratives 377 – emphatic 186 – focus 314, 320 – numeral 180 – possessive 41, 60–63 – questions 461, 462 color terms 120, 380, 381 complex clauses – dependent clauses 13, 336–338 – insubordination 338, 339, 378, 422 – relative clauses 180–181, 253–256, 281, 307–308, 382–383, 464–466 – subordinate clauses 378 compounds 34, 42–43, 67, 161–162, 163– 164, 174, 178, 312, 313 – historical development 29, 170, 236–237, 430 – in morpho-phonology 34, 105 – in serial verb constructions 80, 430 – numbers 180 – question words 176–177 – stress 232 consonants – geminates 12, 28, 30, 32–33, 34, 37, 41– 42, 293, 297 – unusual phonemes 12 – voicing distinction 12, 27, 28, 100, 152, 153, 227–228, 289, 354, 444 consonant clusters 12 – within syllables 37, 38, 356 – word-medially 154, 231, 356, 446

– word-initially 34, 103, 153, 356, 358, 444, 448, 449 constituent order 14, 44, 65, 109–110, 280, 232, 233, 369, 368–369, 453 – chronological order 422, 423 – in modifiers 250, 381–382 – in numbers 180 – in pronouns 394 – in questions 461 – in serial verb constructions 273, 276, 277, 381, 427 contact – with Malay, see Malay – with other languages 7, 9, 26, 149, 216, 390, 436, 441, 466 coordination in NPs 164, 247 copulas, lack of 46, 112, 120, 233, 302, 372, 456 definiteness 120, 242–243, 253–256, 262, 310, 376, 379–381, 469–470, 151 deictic systems – contrastiveness 15, 125, 136, 186, 245– 246, 315 – distance 15, 305, 470 – elevation 15, 50, 68, 158, 495 – elevational paradigms 57, 69, 83, 162, 181, 217, 236, 305–306, 457 – knowledge 15, 310 – slope 15, 83, 305, 306, 430 – specificity 57–59, 310, 384, 470 – visibility 15, 50, 57–58, 156, 159, 181, 214, 310, 371, 377–381 demonstratives, see deictic systems derivational morphology, verbal – causative 166, 172, 193–195 – intentional 155 – other 160, 161, 163, 176, 204, 332 determiners 14, 57, 181–182, 256–257, 465, 469–470 dialect chains 222, 223 Dolabang 11, 149, 150, 151 Dorit 98 dual number marking 88–89, 184, 190–192, 315, 319, 472, 473

Index

eastern Indonesia 1, 9 elevationals, see deictic systems endangerment 149, 222, 230, 287, 353 equational constructions 46, 67, 112, 233– 234, 280 ergativity, see morphological alignment evidentials 94, 203, 206, 207, 376, 497 existential constructions 47, 112–114, 125, 137–138, 234–235, 370–371, 374, 420, 459–460 Fataluku 1, 3, 7, 10, 11, 12, 13, 14, 15, 435 focus 254, 376, 383 – additive 320–321 – and movement 111, 116 – and questions 115, 116, 240–242 – contrastive 245, 320 – focus pronouns 245, 315, 320, 472, 473, 477 – pronominal doubling 369, 394 – restrictive focus 314, 320 – subject/object focus 240–242, 256, 262, 369 fourth person 62, 71–74, 87, 88 gender, lack of 14, 54, 119 grammaticalization – in serial verb constructions 139–140 – verb to applicative 397 – verb to aspectual auxiliary 435 – verb to case marker 280 – verb to postposition 15, 236 Hamap 222, 223 history of the language family 6–9 Holocene 9 incorporation – postpositions 198–202, 321, 327–331 – pronouns 170 Indonesian West Timor 1, 10 insubordination, see complex clauses interrogatives, see questions intonation 94, 112, 121, 173, 175, 242, 374, 476, 497 – constructions in single intonation contour 44, 45, 80, 137, 273, 314, 490

507

juxtaposition 180, 233, 303, 320, 389, 455 Kabola 10, 222, 223 Kaera 12, 13, 97–145, 216, 258, 360, 435 Kamang 10, 11, 12, 13, 14, 15, 285–349, 502 – and multilingualism/contact 3, 258, 353, 441, 478, 496 Kilaara 353 Kisar Island 1, 3, 10, 435 Klon 11, 12, 13, 14, 15, 216, 282 Kui 6, 11, 12 Kula 3, 11, 12, 13, 14, 15, 258, 348, 353, – mutilingualism with Sawila 353 – loans and (dis)similarities with Sawila 353, 355, 357, 401, 403, 406, 426, 435, 436 – loans and similarities with Wersing 441, 501, 502 – see also Tanglapui Lamma 3, 11, 25, 26, 30, 34 Lamotena 353 language change 149, 153, 154, 222, 225, 226, language contact, see contact left dislocation 44, 236, 262, 263, 395 light verbs 15, 213, 300–301, 340–341 loans 33, 152, 155, 175, 356, 361, 426, 428, 441, 485 – from Malay 180, 355, 359, 360–361, 386, 400, 423, 452 – in phonology 33, 154, 227, 355, 358 locationals – postpositions 173, 199 – nouns 62, 68–70, 333 – prefixes 13, 303, 321–323, 325–326, 344, 486, 487 – serial verb constructions 83 – verbs 112, 118, 119, 235–237, 370–372, 458–459, 465 Loona 353 Lowland Kamang 286, 287, 288, 290, 293 Makalero 1, 7, 11, 12, 13, 14, 15, 435, 436 Makasae 1, 3, 7, 10, 11, 12, 13, 15 Malay 216, 366 – loans from, see loans

508

Index

– shift to/preferred language status 6, 26, 222, 287, 353 Manet/Manetaa/Manetee 3, 286, 353, 441 Meher 3, 435 migration 3, 26, 149 Moo 3, 286 mood 366, 375, 418, 421–427 – irrealis 90, 393, 422, 425–426 – realis 422–424, 497–499 morphological alignment – accusative alignment 13, 128, 258 – alignment splits 14, 128, 322 – inverse alignment 14, 395, 403–408 – semantic alignment 13, 14, 74, 390 multilingualism, see bilingualism Nedebang 6, 27, 33, 98 negation 14, 47, 52, 113–114, 174–175, 238– 239, 298, 304, 374, 460–461 – in serial verb constructions 80, 274 – negative imperative 52–53, 115, 239, 304, 461 networks, see dialect chains New Guinea 1, 2, 7, 9 nominalization 13, 55, 127, 163, 217, 382– 384 number marking, see dual, plural numeral classifiers, see classifiers numerals 122–124, 156, 159, 180, 249, 295– 296, 308–309, 314, 384–387, 466–468 Oirata 3, 10, 11, 12, 13 Papuan 1, 2, 3, 9, 10, 15, 16, 18, 427 pitch-accent 13 plural number marking – associative plural 14, 55, 264, 310, 469, 473 – distributive plural 64 – in pronouns 48, 80, 108, 128, 191 – plural possessive pronouns 61, 62, 65, 125, 182 – prefixes 62, 124, 171, 191, 197, 319 – reduplication 295 – separate marker 54, 122, 260–261, 308, 313, 385, 462, 468–469

possession – alienability 15, 60–61, 68, 126, 182, 184, 243, 246, 312, 388 – body part nouns 14, 61–62, 64, 125–126, 183, 244–245, 312, 388–389, 471 – inalienability 15, 60–64, 126–127, 182– 184, 243–244, 312–313, 388 – kinship terms 14, 62, 244, 312, 388 – no alienable/inalienable distinction 471 postpositions 14, 47–49, 69–70, 71, 117– 119, 172–174, 199, 303, 317, 321 – grammaticalization from verbs 15, 85, 213 prohibition 164, 175, 239, 374, 375, 421 proto Alor-Pantar 100, 108, 145, 223, 227, 357, 360, 361, 429, 430, 468 proto East Alor 446 proto Timor-Alor-Pantar 7, 9, 182, 445, 466 Pura 11, 149, 150, 153, 160, 170 quantification 55–6, 122–124, 180, 252– 253, 308–309, 319, 384–387, 466– 469, 477–479 – see also numerals questions/question words 115–117, 175–177, 240–242, 374–375, 423, 425, 461–462 reciprocals 87, 182, 189–190, 192, 257, 262–263, 314, 396–397, 404, 472 reduplication 43, 155–159, 162, 167, 178, 217, 294–296, 365 reflexives 62, 71–3, 247, 257, 260–263, 315, 317–318 relative clauses 180–181, 253–256, 281, 307–308, 382–384, 464–466 resultatives 345–346, 433–434, 496–497 right dislocation 44 rock art 9 Saliira 353 Salimana 353 Sawila 12, 13, 14, 15, 351–438, 441 – comparison with other languages 258, 348, 446, 466, 467, 480, 501, 502 specificity, in deictics 50, 57, 58–59, 310, 384, 470 stress 12, 13

Index

– final 106 – non-phonemic 106, 107, 154–155, 358– 359, 446–447 – penultimate 38, 154–155 – phonemic 38, 231, 292–294 – weight-sensitive 292–294 Suboo 3, 286 subordination, see complex clauses syllable structure 12, 36–38, 105–106, 153– 154, 230–231, 291–292, 358–359, 447–449 Tanglapui 3, 403, 405 – see also Kula Teiwa 11, 12, 13, 14, 15, 98, 99 – comparison with other languages 27, 100, 103, 145, 282, 423, 502 – tense implied 91 – unmarked 89, 280, 497 Tiee 3, 286 Timor Leste (East Timor) 1, 14, 435 Timor subgroup 6, 7, 16 tone 13 topic 263, 376–377, 379 – and pronoun doubling 395 – fronting 110, 462 – postposing 308 – pronouns 473, 476–477 Trans New Guinea 7, 9, 396 transitivity, see valency tri-syllabic words 37, 105, 291, 293, 294, 359, 447 Upland Kamang 287, 290, 293 valency – avalent verbs 297, 300 – labile verbs 133–134 – bivalent verbs 297–298, 300–302, 368– 369 – monovalent verbs 165, 297–298, 299, 300–301, 367–368

509

– trivalent/ditransitive verbs 45–46, 139, 169, 233, 267, 300–301, 367–370, 454, 455 valency increasing 197, 266–273, 409 verb classes 13, 76–80, 129–133, 167–169, 300–302, 323–326, 407–418, 481–484 vowels – diphthongs 35–36, 151–152, 154–155, 225, 229–230, 295, 355, 357, 358, 443 – diphthongs at morpheme boundaries 14 – echo vowels 12, 435, 436, 443, 447, 448– 449, 450–451, 501–502 – epenthesis 12, 37–38, 40, 223, 358, 447, 448, 449, 450, 452, 501 – long vowels, phonetic 36, 151, 230, 231, 290 – long vowels, phonemic 12, 27, 103, 355, 442 – metathesis 12, 436, 441, 452, 502 – tense/lax distinction 224, 228, 230, 355– 356 vowel harmony 107–109, 126, 193, 196, 363–364, 365, 449–450, 451, 452 vowel sequences 12, 31, 36, 104–105, 152, 154, 230, 291–292, 357, 443 Wersing 3, 6, 11, 12, 13, 14, 439–503 – comparison with other languages 145, 426, 435 – neighboring languages/language contact 352, 353, 357, 390, 436 West Bomberai 9 Western Kamang 286, 287 Western Pantar 3, 6, 11, 12, 13, 15, 23–95 – comparison with other languages 100, 103, 145, 258 – neighboring languages/language contact 98 Woisika 3, 10, 287 word order, see constituent order